CHAPTER 8

TEMPERATURE MANAGEMENT IN CARDIAC SURGERY

Laurie K. Davies and Heather Reed

TEMPERATURE MANAGEMENT IN CARDIAC SURGERY

The use of hypothermia as an adjunct to the treatment of a wide variety of disorders has been
advocated for centuries. Lowered body temperature has been employed to combat cancer,
infection, trauma, and central nervous system diseases, and as a regional method to induce
anesthesia for amputation (1,2). However, it was not until 1950 that Bigelow et al. (3)
demonstrated longer tolerance to inflow occlusion in hypothermic animals than in their
normothermic counterparts. This work led to the first clinical application of hypothermia in
cardiac surgery. Lewis and Taufic (4) used surface cooling to 28°C with 5.5 minutes of inflow
occlusion to facilitate successful closure of an atrial septal defect in a 5-year-old child. In 1952,
Gibbon (5) introduced the pump oxygenator to clinical practice, and in 1958, Sealy et al. (6)
used hypothermia in conjunction with the cardiopulmonary bypass (CPB) circuit for
intracardiac repairs. The use of the pump oxygenator and hypothermia has allowed cardiac
surgery to flourish. Complex lesions are repaired routinely with remarkably low mortality.
However, it is now recognized that inadvertent hyperthermia occurs relatively frequently in
the perioperative setting. Evidence is accumulating that elevated temperatures may be
deleterious in both animals and humans (7,8). A better understanding of the principles of
temperature management may maximize the benefits to our patients while minimizing
potential complications.

PHYSIOLOGY OF HYPOTHERMIA

One of the main difficulties in devising a reasonable strategy for the application of
hypothermia in humans is the fact that they are naturally homeothermic beings. Humans and
other homeothermic species have very effective homeostatic systems, which ensure that the
body temperature remains consistently near 37°C regardless of changes in environmental
temperatures. This tight regulation of temperature is accomplished by multiple mechanisms.
Cold is sensed by the thermoreceptors in the skin, which then causes the hypothalamus to
trigger a strong sympathetic nervous system response. Vasoconstriction of skin vessels, which
decreases convective heat loss, occurs simultaneously with vasodilation of the skeletal muscle
vascular beds, which augments muscular activity to produce heat by tensing and shivering. The
endocrine system is activated, oxygen consumption is increased, and heart rate, cardiac
output, and blood pressure are elevated. Because of the complexity of these interactions, one
can appreciate the difficulty in understanding the physiologically appropriate response to the
unnatural state of induced hypothermia in humans. One must extrapolate from animal studies,
biochemical equations, accidental hypothermia survivors, and normal organ temperature
gradients to try and develop the most effective management strategy when using deliberate
hypothermia.

Rationale for the Use of Hypothermia

Why is hypothermia often employed during CPB? The obvious purpose of using hypothermia is
to provide a degree of organ (and organism) protection and safety margin during CPB.
Hypothermia exerts its protective effect by multiple mechanisms. The most obvious
mechanism is a reduction in metabolic rate and oxygen consumption (9) (Fig. 8.1). This
metabolic suppression may not explain all the protective effects observed. Hypothermia also
helps preserve high-energy phosphate stores and reduces excitatory neurotransmitter release,
which is especially important to central nervous system protection (10,11). Normally, ischemic
neuronal cells rapidly release neuroexcitatory amines, especially glutamate (12). The
accumulation of these excitotoxic amines causes the opening of calcium channels and
activation of multiple destructive enzymatic systems. Hypothermia attenuates this excitotoxic
cascade, helping to prevent calcium entry into the cell and restricting membrane permeability
(13) (Fig. 8.2).

Figure 8.1 Whole-body oxygen consumption as a function of body temperature in dogs made
hypothermic by surface cooling.(From Kirklin JW. Cardiac surgery. New York: Churchill
Livingstone, 1993:61–127, with permission.)
Figure 8.2 Line plot of time-course changes in the perfusate levels of dopamine (pmol/ml) in
animals whose intraischemic brain temperature was maintained at 36°C (n = 10) (●—●), 33°C
(n = 4) (°—°), and 30°C (n = 8) (*—*). The data presented are mean ± SEM. Statistical
significance was assessed by two-way ANOVA. In animals whose intraischemic brain
temperature was maintained at 36°C, a massive increase in dopamine levels was observed
(*significantly higher than control values, p < 0.01). In animals whose intraischemic brain
temperature was maintained at 33°C or 30°C, a significant reduction in the ischemia-induced
increased dopamine levels was demonstrated (a, significantly lower than the corresponding
level in the 36°C group, p < 0.01).

Alteration of temperature causes a change in the reaction rate of all biochemical processes,
especially enzymatic reactions. This temperature dependence of reaction rates has been
described by the concept of Q10, which is defined as the increase or decrease in reaction rates
or metabolic processes in relation to a temperature change of 10°C. For instance, the reaction
rate of a process with a Q10 of 2 will double with a 10°C increase in temperature or be halved
with a drop of 10°C. Most reactions, including total body oxygen consumption, have a Q10 of 2
to 3 (9).
Some biochemical processes, especially those localized to cell membranes, show an abrupt
change in reaction rates at certain critical temperatures. This has been termed a phase
transition and is thought to be the result of a change in the cell membrane from a fluid to a gel
state (14). In mammalian tissues, phase transitions often occur at approximately 25°C to 28°C
and may disturb cell homeostasis. Biophysical processes such as osmosis and water diffusion
are also affected by temperature. Typically, a linear change of approximately 3%/10°C is seen.
Therefore, this effect is minimal at clinical levels of hypothermia. However, if the freezing point
of water is approached, ice is formed in the tissue, a condition that is not tolerated. The
solutes concentrate in a hyperosmolar manner in the residual nonfrozen water, causing
marked fluid shifts and membrane disruption. Mammalian tissue does not regain function on
thawing from a frozen state. For this reason, there is a limit to the beneficial effects of
hypothermia.

Figure 8.3 Nomogram of an equation expressing the relation of oxygen consumption to

perfusion flow rate at different temperatures in animals. The x represents the perfusion flow
rates used clinically at these temperatures.(From Kirklin JW. Hypothermia, circulatory arrest,
and cardiopulmonary bypass. In: Kirklin JW, Barratt-Boyes BG, eds. Cardiac surgery. New York:
Churchill Livingstone, 1993:61–127, with permission.)

In cardiac surgery, CPB in conjunction with systemic hypothermia allows lower pump flows,
better myocardial protection, less blood trauma, and better organ protection than does
normothermic perfusion (15). Oxygen needs predictably fall with lowered temperature. It was
recognized early that lowered bypass flows could be employed in this setting and still provide
adequate perfusion, as assessed by mixed venous oxygen tension and return of organ function
following bypass. Relating oxygen consumption (VO2) to perfusion flow rate at various
temperatures can also be valuable in assessing adequacy of tissue perfusion (Fig. 8.3). At a
given temperature, a fall in VO2 with a decrease in flow rate implies a flow-limited VO2,
indicating that oxygen delivery is not adequate. Hickey and Hoar (16) have shown in humans
that a reduction in flow rate from 2.1 to 1.2 L/min/m2 of body surface area at 25°C does not
alter VO2 or tissue perfusion. Slogoff et al. (17) were unable to correlate low flows (<40
mL/kg/min) or pressures (<50 mm Hg) during bypass in which moderate hypothermia and
hemodilution were used with postoperative renal or central nervous system dysfunction.
Lower perfusion flow rates allow better visualization by the surgeon. Venous return from the
bronchial, pulmonary, and noncoronary collateral vessels is also decreased. Because this
returning blood is at systemic temperature, it can inappropriately warm the heart when
cardioplegia-induced myocardial hypothermia is at a less-than-systemic temperature and can
jeopardize myocardial protection. Blood trauma is minimized because of both the lower pump
flows and the hemodilution employed during bypass. Because the etiology of most central
nervous system damage on bypass may be embolic in origin (18), lower bypass flows can
minimize these focal insults. Systemic hypothermia also provides some margin of safety for
organ protection if equipment failure occurs or circulatory arrest must be employed.

Acid–Base Alteration with Temperature Change

One of the more often discussed aspects of clinical hypothermia is the appropriate acid–base
management strategy during hypothermia. To understand acid–base regulation during
hypothermia, it is helpful to consider several example conditions. “Normal” values for pH and
PCO2 are usually thought of as 7.40 pH units and 40 mm Hg, respectively. However, it must be
kept in mind that these values are appropriate only at 37°C in blood. There is a temperature-
dependent spectrum of “normal” values, depending on the temperature of the specific site in
the body. For example, arterial blood leaves the heart at 37°C with a pH of 7.40 and a PCO2 of
40 mm Hg. When that same blood perfuses working skeletal muscle, where the ambient
temperature may be 40°C, it will have a pH of approximately 7.35 and a somewhat higher
PCO2, despite a constant CO2 content before any respiratory exchange with the muscle. In
contrast, the same arterial blood perfusing exposed skin, where temperature may be 20°C in
cool weather, will have a pH of 7.65 and a proportionately lower PCO2, again with no change
in CO2 content. Stated another way, a sample of arterial blood held in a gas-tight syringe will
have a PCO2 that varies directly with temperature, and a pH that varies inversely with
temperature, despite constant CO2 content.

This change in PCO2 with temperature is a consequence of the change in solubility of gases in
liquids with change in temperature. As a general rule, decreasing the temperature of a liquid
increases the solubility of a given gas in that liquid and therefore decreases the partial
pressure of that gas while the overall content of the gas in the liquid remains constant.
Increasing the temperature of a liquid increases the kinetic energy of the molecules in the
liquid, which both increases the tendency of dissolved gas molecules to leave the liquid
(decreased solubility) and increases the partial pressure of those gas molecules remaining in
the liquid. This concept is intuitive to anyone who has opened a container of warm carbonated
beverage and observed that its tendency to “fizz” (CO2 bubbling out of solution) is much
greater than that of a cold beverage.
The relation of pH to temperature is somewhat more complex than the change in PCO2 with
temperature. There is clearly a direct relation between the concentration of H+ [H+] in a water
solution and the CO2 content of that solution. The higher the CO2, the more H+ in solution,
and hence lower the pH. This is largely caused by the tendency of CO2 to combine chemically
with water to produce carbonic acid, which then dissociates in solution to yield H+.
Accordingly, one might expect that the change in pH of blood with temperature is caused by
the changes in PCO2 with temperature, as discussed in the preceding text. It is true that
cooling an anaerobic blood sample decreases both the PCO2 and the [H+]. However, and very
importantly, the change in [H+] and pH that occurs with change in temperature is independent
of a change in CO2 content, and therefore does not depend on the change in CO2 solubility
with the temperature change.

All acids and bases, including water, exist in solution in equilibrium between the undissociated
form and the ionized components of the parent molecule. The dissociation constant (K) is the
equilibrium ratio of the product of the concentrations of the ionized components to the
unionized component. For water at 25°C, the equilibrium dissociation equation is as follows:

K(H2O, 25°C) = [H+] × [OH−]/[H2O] = 1.08 × 10−14

Because, at equilibrium in water, [H+] equals [OH−], and because the concentration of H2O is
essentially 1, [H+] equals [OH−] equals √10−14, or 10−7. Because the definition of pH is the
negative log10 of [H+], the pH of pure water at equilibrium at 25°C is 7. This pH value has come
to be called the neutral pH, or pN, of water. Temperature change has a significant effect on the
tendency of molecules in solution to dissociate. In thermodynamic terms, the increased kinetic
energy associated with increased temperature promotes dissociation, whereas decreased
temperature has the opposite effect. For example, within the temperature range seen in
clinical CPB (approximately 15°C–40°C), the dissociation constant of water increases from
0.451 × 10−14 to 2.919 × 10−14. These values of KH2O relate to a change in [H+] from
approximately 67 nmol/L at 15°C to 170 nmol/L at 40°C. This nearly 3-fold change in [H+] is
solely a consequence of the effects of temperature change on dissociation. Pure water may be
considered the simplest weak acid–base solution. The major importance of these concepts is
that water is the fundamental solvent of all biologic systems, and the dissociation of virtually
all weak acids and bases in biologic solutions follows the same pattern as that described for
water.

The behavior of body fluids (intracellular and extracellular, intravascular and extravascular) is
far more complex than the simple scenario described earlier for water, but biologic fluids
behave much like water in terms of the intrinsic temperature-related changes in the
dissociation constants of the many weak acids and weak bases, of which they are composed.
The amino acids in proteins, the simple sugars in polysaccharides, the fatty acids in lipids, and
the major buffer systems, all follow this same basic pattern. As the temperature decreases, the
tendency to dissociate decreases, and the concentrations of the ionized components (H+ and
R−) also decrease.
Figure 8.4 Blood pH of various ectothermic species and the pH of neutral water as a function of
body temperature.(From Rahn H. Body temperature and acid–base regulation [Review Article].
Pneumonologie 1974;151:87–94, with permission.)

At normal body temperature (37°C), blood and tissue fluids are alkaline (lower [H+] and
correspondingly higher pH) relative to water at the same temperature. A number of buffer
systems create and maintain this relative alkalinity so that the ratio of [OH−] to [H+] remains
constant at approximately 16:1 despite temperature variation. As temperature changes, the
intrinsic dissociation of these buffer systems also changes to maintain the ratio of [OH−] to
[H+] constant. Therefore, the intrinsic pH shift of blood and tissue fluid parallels the pNH2O as
temperature changes, and this relative alkalinity remains constant in comparison with water
(19) (Fig. 8.4).
Figure 8.5 Changes in arterial pH and PCO2 as blood at 37°C arrives at skin and exercising
muscles at temperatures of 25°C and 41°C, respectively. Neutrality of water (pN) changes in
parallel with changes in blood pH. Therefore, the relative alkalinity of the blood or the ratio
between [OH−] and [H+] ions remains constant.(From Rahn H. Body temperature and acid–
base regulation [Review Article]. Pneumonologie 1974;151:87–94, with permission.)
Figure 8.6 Arterial and intracellular pH as a function of body temperature in ectothermic
animals. Intracellular pH closely follows the neutral pH of water; relative arterial alkalinity is
maintained at all temperatures.(From White FN, Weinstein Y. Carbon dioxide transport and
acid–base regulation during hypothermia. In: Utley JR, ed. Pathophysiology and techniques of
cardiopulmonary bypass, Vol. II. Baltimore: Williams & Wilkins, 1983:40–48, with permission.)

A major buffering system responsible for this constant relation of blood and tissue fluid pH to
pN, with temperature change, is the imidazole moiety of the amino acid histidine, which is
commonly found in body proteins. The pKa of this component of histidine is close to 7.0 at
body temperature, a property that confers potent buffering capacity for maintaining a
constant ratio of [H+] to [OH−] despite significant changes in the absolute concentration of
each as temperature varies. These considerations are applicable to the previous example of
arterial blood with a constant CO2 content perfusing tissues with different temperatures. The
observed shift in pH with cooling follows the pNH2O, and the buffering capacity of the
imidazole moiety of histidine preserves the constant relative alkalinity and ratio of [OH−] to
[H+] in the blood (Fig. 8.5). In cold-blooded vertebrates, the blood pH–temperature curve also
runs parallel to the pH of neutral water. Intracellular pH has also been measured in various
animals and shows changes with temperature identical to those that have been described for
pNH2O (20) (Fig. 8.6). The intracellular pH parallels the pN and blood pH slopes with
temperature changes and differs from the extracellular pH by a constant but species-specific
factor of approximately −0.6 to −0.8 pH units. Therefore, at 37°C, intracellular pH is
approximately 6.8 to 6.9 and so the [H+] is somewhat higher. In a manner similar to the
observed change in blood pH with temperature, the reaction kinetics of numerous respiratory
enzyme systems (lactate dehydrogenase, Na+-K+-ATPase [sodium–potassium adenosine
triphosphatase], acetyl CoA carboxylase, fatty acid synthetase, NADH [reduced nicotinamide
adenine dinucleotide] cytochrome c reductase, and succinate cytochrome c reductase) all
show optimal catalytic function with temperature change when the pH of the reaction medium
parallels the temperature-mediated pNH2O change (21).

This constant internal milieu is accomplished, as has been mentioned, predominantly by the
buffering capacity of the imidazole group of the amino acid histidine. As temperature changes,
the imidazole groups in protein change pKa in parallel with the pN of water. The ratio of the
unprotonated histidine imidazole groups to H+, a value known in the world of chemistry as
alpha, remains constant, total CO2 also remains constant, and the pH changes as per the
changes in temperature. The term α-stat has come to indicate an acid–base management
strategy in which the net charge (dissociation) of proteins remains constant as temperature
changes. Typically, this is managed during CPB by keeping total CO2 stores constant and
allowing pH and PaCO2 to follow their thermodynamically mediated dissociation changes with
changes in temperature. In other words, during cooling, exogenous CO2 is not added to the
system when following the α-stat strategy.

The alternative method of acid–base strategy is termed pH-stat. With this method, pH is the
value that is maintained constant at varying temperatures. Obviously, if the pH-stat strategy is
used when blood is cooled, CO2 must be added to maintain a PaCO2 of 40 and a pH of 7.40.
Extracellular and intracellular ratios of [OH−] to [H+] are altered and the total CO2 stores are
elevated.

Why might one strategy be chosen over another? During the first two decades of hypothermic
CPB, pH-stat management with the addition of 5% CO2 to the oxygenator gas flow was used
almost exclusively. An understanding of the expected changes in pH with temperature seemed
to be lacking, and CO2 was thought to be beneficial for cerebral vasodilation and maintenance
of cerebral blood flow (CBF). In the last 30 years, this practice has been questioned, and many
institutions have shifted toward an α-stat management protocol. It is only recently that
sufficient data have accumulated to enable a rational decision to be made for employing the
more suitable of the two strategies in a given situation.
On a theoretical basis, α-stat management may be preferable in certain situations.
Maintenance of constant intracellular electrochemical neutrality appears to be essential for
normal cellular function (22). Intracellular metabolic intermediates of high-energy phosphates
can be depleted if there are changes in the intracellular pH and these metabolites lose their
charged state. These substrates are then free to diffuse across lipid membranes. Most
enzymes depend on optimal pH for their function. Electrochemical neutrality is also important
in maintaining the Donnan equilibrium across cellular membranes to allow normal intracellular
anion concentrations and water content (23).

Poikilothermic animals, whose tissues must function optimally despite wide variations in
temperature, follow an α-stat acid–base strategy. On the other hand, hibernating mammals
appear to maintain a pH-stat strategy, with constant (temperature-corrected) blood pHa and
PCO2 (22). These animals hypoventilate as they hibernate, the tissue CO2 stores increase, and
intracellular pH becomes acidotic in most tissues. This acidotic state causes a further
depression of metabolism that teleologically may be useful by further decreasing the energy
consumption of nonfunctioning tissues, such as skeletal muscle, gastrointestinal tract, and
higher brain centers. In contrast, active tissues, such as heart and liver, adopt a different
strategy by actively extruding H+ across their cell membranes to maintain intracellular pH at or
near the values predicted by the α-stat methodology. Therefore, hibernating mammals are
able to vary their intracellular-to-extracellular pH gradient differently in different tissues,
depending on the state of metabolic activity of the tissue. Functionally, this provides different
types of acid–base regulation in different tissues, depending on the metabolic activity of the
tissue. The first noticeable change associated with arousal from hibernation is
hyperventilation. This depletes the CO2 stores, raises intracellular pH, and increases the
metabolic rate. The animal reverts to an overall α-stat pH control pattern during awakening,
which allows tissues to regain optimal function. Therefore, in hibernating mammals, the issue
of acid–base maintenance is not clearly defined because intracellular acid–base regulation can
be independent of blood regulation both within and among different tissues in the same
animal.

Despite the preceding discussion, the practical question of how acid–base status should be
regulated during hypothermic bypass in humans remains open. Some animal studies suggest
that α-stat acid–base management is beneficial in terms of myocardial protection. McConnell
et al. (24) evaluated α-stat regulation during hypothermia in dogs and demonstrated that
significant elevations in coronary blood flow, left ventricular oxygen consumption, and lactate
utilization occurred with maintenance of a pH of 7.7 at 28°C (α-stat) in comparison with a pH
of 7.4 (pH-stat). There was also a significant increase in peak ventricular pressure when a
standard preload was applied. Poole-Wilson and Langer (25) demonstrated a greater
contractility in hypothermic perfused papillary muscle when the pH of the perfusate was more
alkaline than 7.4. They also demonstrated a rapid fall in myocardial tension in addition to
changes in Ca2+ flux when the perfusate PaCO2 was increased (26). On the other hand, Sinet
et al. (27) found no effect of pH on the performance of isolated rat heart. The myocardium is
often not perfused but is purposely made ischemic to facilitate cardiac surgery. In this setting,
alkalinization of the blood before ischemia has been shown to decrease the development of
acidosis in coronary sinus blood and improve contractility on reperfusion (28). It also appears
that the pH of the blood reperfusing the heart may be critical to the recovery of ventricular
performance. Becker et al. (29) studied the myocardial effects of an acid–base strategy in
which alkalinization greater than that of α-stat was used. They found improvements in
myocardial performance after 1 hour of circulatory arrest and cardioplegia, with moderate
alkalinization in comparison with α-stat. Acid–base management also appears to be important
in cardiac electrophysiology. Swain et al. (30) showed the electrical stability of the heart to be
increased, with less spontaneous ventricular fibrillation, when α-stat blood regulation was
compared with pH-stat. Kroncke et al. (31) found a 40% incidence of ventricular fibrillation in
patients cooled to 24°C during pH-stat management and a 20% incidence in those managed
with α-stat.

The appropriate acid–base management for optimal cerebral perfusion has also been
questioned. Clearly, CBF decreases significantly with hypothermia. Cerebral metabolic rate also
decreases during hypothermic bypass. The response of the cerebral circulation to changes in
PaCO2 is preserved, at least during moderate hypothermia (32); therefore, α-stat management
will result in lower cerebral flows than those seen with pH-stat management. However,
because of the lowered metabolic demands, a lower CBF may be appropriate and indicative of
a maintained coupling of blood flow and metabolic demand. Govier et al. (33) demonstrated
intact autoregulation in humans using an α-stat strategy at temperatures ranging from 21°C to
29°C. Murkin et al. (34) showed coupling of CBF and metabolism that was independent of
cerebral perfusion pressure (CPP) within the range of 20 to 100 mm Hg when α-stat
management was employed. In contrast, cerebral autoregulation was abolished and CBF
varied with perfusion pressure when pH-stat strategy was used (Fig. 8.7). It has been argued
that the CBF during pH-stat hypothermia actually represents excessive blood flow and may be
detrimental. Unnecessarily high blood flows may put the brain at risk of damage by
microemboli or high intracranial pressure. With deep hypothermia (i.e., temperatures <20°C),
the normal vascular responses are lost and CBF becomes pressure dependent (35,36). At deep
hypothermic temperatures, coupling of cerebral flow and metabolism is also lost. It is
important to note, however, that the responses of CBF and CMRO2 (cerebral metabolic rate of
oxygen) are quantitatively different at deep hypothermic conditions. CBF decreases linearly
with the decrease in temperature, whereas CMRO2 drops exponentially. The net result is that
CBF becomes more luxuriant at deep hypothermic temperatures. At normothermia, the mean
ratio of CBF to CMRO2 is 20:1, and at deep hypothermia, the ratio increases to 75:1 (37). This
situation is important in the context of low-flow CPB. At very low temperatures, data indicate
that pump flow rates may be reduced to as little as 10 mL/kg/min before flow becomes
inadequate for cerebral metabolic requirements (38).
Figure 8.7 Simple linear regression of cerebral blood flow (CBF) versus cerebral perfusion
pressure (CPP) or cerebral oxygen consumption (CMRO2) for temperature-corrected and
temperature-uncorrected groups. Upper panel: There is no significant correlation between CBF
and CMRO2 in the temperature-corrected group (A1), whereas CBF significantly correlates
with CMRO2 in the temperature-uncorrected group (B1). Lower panel: CBF is significantly
correlated with CPP in the temperature-corrected group (A2), whereas CBF is independent of
CPP in the temperature-uncorrected group (B2).(From Murkin JM, Farrar JK, Tweed, WA, et al.
Cerebral autoregulation and flow/metabolism coupling during cardiopulmonary bypass: the
influence of PaCO2. Anesth Analg 1987;66:825–832, with permission.)

On a microcirculatory level, some evidence suggests that α-stat management may be

beneficial to the brain. Norwood et al. (39) studied the brains of hypothermic dogs perfused
with anoxic blood and found a decrease in extent and magnitude of lesions when the
perfusate had a higher pH. Acidic perfusate enhanced the extent of the lesions. On the other
hand, Priestley et al. (40) compared neurologic and histologic outcome in a survival piglet
model of deep hypothermic circulatory arrest. They showed that the pH-stat group
demonstrated better neurologic performance and less severe functional disability scores than
those piglets in the α-stat group. Histologic injury was also more severe in the α-stat group.
Theoretically, hypocarbia (and increased pH) result in a leftward shift of the oxyhemoglobin
dissociation curve, which causes oxygen to be less readily available to the tissues. However,
more oxygen is dissolved in the plasma during hypothermia, so that these two effects tend to
cancel out each other. The relatively low CBF during α-stat management has still been shown
to be in excess of cerebral metabolic needs (34).

In adults, the preponderance of evidence suggests either that CO2 management on CPB does
not matter or that an α-stat strategy is advantageous. Bashein et al. (41) examined the
influence of pH management in 86 adults in whom mild hypothermia was utilized
(approximately 30°C). They found no difference in cardiac or neuropsychologic outcome
regardless of acid–base management. It is important to realize, however, that it would have
been unlikely for them to be able to demonstrate a difference in this study under the
conditions of the study. The differences in PCO2 between the two groups amounted only to
approximately 6 to 7 mm Hg, and the degree of hypothermia was not very profound. Contrast
the scenario in their study to that of a patient in deep hypothermia, for whom the difference in
PCO2 between the two strategies approaches 80 mm Hg! In addition, their analysis looked for
differences in mean group performances rather than changes in individual patient
performance, a methodology that may have decreased the sensitivity of the study to detect
any existing difference. Three randomized prospective studies of moderate hypothermia in
adults have demonstrated that postoperative neurologic or neuropsychological outcome is
slightly, but consistently, better with α-stat management (42,43,44). The notion that α-stat
management may be beneficial in this setting in adults makes sense if one considers that the
most likely mechanism for neurologic injury in these patients is probably emboli. Therefore, α-
stat management would be expected to provide lower CBFs that are more aligned with the
cerebral metabolic rate and a lesser embolic load.

Although acid–base management is probably not as important when moderate hypothermic

temperatures are used, it may be critical in the setting of deep hypothermia. Proponents of the
α-stat method suggest that unnecessarily high blood flows (with pH-stat management) may
put the brain at risk for damage from microemboli, cerebral edema, or high intracranial
pressure, or may actually predispose to an adverse redistribution of blood flow (“steal”) away
from marginally perfused areas in patients with cerebrovascular disease. On the other hand,
proponents of the pH-stat strategy suggest that enhanced CBF may be helpful in improving
cerebral cooling before the initiation of circulatory arrest. In fact, total CBF is increased, global
cerebral cooling is enhanced, and brain blood flow is redistributed during pH-stat
management. An increased proportion of CBF is distributed to deep brain structures
(thalamus, brainstem, and cerebellum) when pH-stat management is used (45) (Figs. 8.8 and
8.9). However, other data suggest that cerebral metabolic recovery after circulatory arrest may
be better with the α-stat method than with the pH-stat mode. This variation in results has led
some authors to advocate a crossover strategy in which a pH-stat approach is used during the
first 10 minutes of cooling to provide maximal cerebral metabolic suppression, followed by an
α-stat strategy to remove the severe acidosis that accumulates during profound hypothermia
during pH-stat. This approach appears to offer maximal metabolic recovery in animals (46) (Fig.
8.10).
Figure 8.8 Regional blood flow in the brain during α-stat or pH-stat blood gas management. BS,
baseline on normothermic cerebral blood pressure (CBP); HT, at the end of 30 minutes of
hypothermic cardiopulmonary bypass (CPB); R5, 5 minutes after initiation of reperfusion and
rewarming after 1 hour of circulatory arrest; N0, after 45 minutes of rewarming, when
normothermia was achieved; N3, after 3 hours of reperfusion at normothermia. *p < 0.01, #p <
0.05 for group difference.(From Aoki M, Nomura F, Stromski ME, et al. Effects of pH on brain
energetics after hypothermic circulatory arrest. Ann Thorac Surg 1993;55:1093–1103, with
permission.)

Figure 8.9 Intracerebral distribution of blood flow: α-stat versus pH-stat. Rep(5), 5 minutes
after initiation of reperfusion and rewarming after 1 hour of circulatory arrest; NT(0), after 45
minutes of rewarming, when normothermia was achieved; NT(180), after 180 minutes of
reperfusion at normothermia. “+” indicates p < 0.05 for the within-group increase by paired t
test; “–” indicates p < 0.05 for the within-group decrease by paired t test.(From Aoki M,
Nomura F, Stromski ME, et al. Effects of pH on brain energetics after hypothermic circulatory
arrest. Ann Thorac Surg 1993;55:1093–1103, with permission.)

Figure 8.10 This figure demonstrates the effects of three different cooling strategies (α-stat,
pH-stat, and a crossover of pH-stat followed by α-stat) on cerebral metabolic suppression
before deep hypothermic circulatory arrest (DHCA) and the recovery of cerebral metabolism
after DHCA. The addition of CO2 (i.e., the pH-stat strategy) provides better cerebral metabolic
suppression before DHCA, but cerebral metabolic recovery after DHCA is poor. Initial cooling
with a pH-stat strategy followed by conversion to α-stat before DHCA results in the greatest
cerebral metabolic recovery. CMRO2, cerebral metabolic rate of oxygen.(From Kern FH,
Greeley WJ. pH-stat management of blood gases is not preferable to α-stat in patients
undergoing brain cooling for cardiac surgery. J Cardiothorac Vasc Anesth 1995;9:215–218, with
permission.)
Figure 8.11 Cortical oxygen saturation (ScO2) during DHCA in the pH-stat and α-stat groups.
Mean ± SD, eight animals per group. *p < 0.05 between groups. The ScO2 half-life during arrest
was significantly greater in the pH-stat than in the α-stat group.(From Kurth CD, O’Rourke MM,
O’ Hara IB. Comparison of pH-stat and α-stat cardiopulmonary bypass on cerebral oxygenation
and blood flow in relation to hypothermic circulatory arrest in piglets. Anesthesiology
1998;89:110–118, with permission.)

The choice of acid–base management may be particularly important in the subgroup of

pediatric patients with aortopulmonary collaterals, for whom cerebral cooling is problematic. It
appears that the addition of CO2 during cooling enhances cerebral perfusion and improves
cerebral metabolic recovery (47). Kurth et al. (48) demonstrated in a piglet model two
mechanisms by which pH-stat management may be beneficial to the brain. They showed that
pH-stat increases the rate of brain cooling and that the rate of depletion of brain oxygen
during deep hypothermic circulatory arrest (DHCA) is considerably slower with pH-stat
management than with α-stat management (Fig. 8.11). In addition to the increase in CBF seen
with elevated CO2, one would also expect to see an increase in pulmonary vascular resistance
and a decrease in pulmonary blood flow with the pH-stat strategy. In fact, in a randomized
clinical trial of 40 cyanotic children, Sakamoto et al. (49) demonstrated a reduction in the
systemic–pulmonary collateral circulation in the pH-stat group. They also showed improved
cerebral oximetry values and lower lactate levels in the patients in the pH-stat group. No
neurologic deficits were noted in the patients from either group.

A recent randomized single-center trial in human infants younger than 9 months found that
the infants who were managed with a pH-stat strategy had a trend toward better short-term
outcomes than those managed with the α-stat strategy (50). The pH-stat group had a shorter
recovery time to first electroencephalographic activity and a tendency to fewer
electroencephalographically manifested seizures. In this study, within the subset of infants
with transposition of the great vessels, those assigned to pH-stat tended to have a higher
cardiac index despite a lower requirement for inotropic agents, less frequent acidosis (p =
0.02) and hypotension (p = 0.05), and a shorter duration of mechanical ventilation and
intensive care unit stay (p = 0.01). Although the number of infants studied in this investigation
was small, their findings challenge the concept that α-stat management is more physiologic
and protective during deep hypothermia in infants than pH-stat management.

A follow-up paper examined the developmental and neurologic outcome in the same children
at 2 to 4 years of age (51). They found no difference in the neurodevelopmental outcomes
with α-stat versus pH-stat. Interestingly enough, there were some differences noted,
depending on which type of lesion was being repaired. In the cyanotic subgroup (transposition
and Tetralogy of Fallot), a slightly higher (but not statistically significant) Mental Development
Index score was found in those patients managed using pH-stat. However, in the ventricular
septal defect subgroup, the patients on pH-stat scored significantly worse. These data must be
interpreted with caution because the number of patients in each subgroup was small and
there were differences in age at the time of surgery, depending on the diagnosis.

Why is there an apparent difference in outcome between adults and children relative to pH
management? It may relate to differences in the mechanism of brain injury on CPB. In adults,
emboli appear to play a prominent role in adverse neurologic outcome (52). It is therefore
postulated that the reduced CBF associated with α-stat management may be protective by
limiting the dispersion of cerebral microemboli. On the other hand, the mechanism of injury in
children may relate more to hypoperfusion or activation of excitotoxic pathways (53). If a pH-
stat strategy is employed, the increase in CBF may be beneficial in ensuring complete brain
cooling and slowing oxygen consumption, thereby increasing the tolerance of the brain for
DHCA.

Alterations in Organ Function

Hypothermia causes a decrease in blood flow to all organs of the body. However, some areas
experience a greater decline than others. Skeletal muscle and the extremities have the
greatest reduction in flow, followed by the kidneys, splanchnic bed, heart, and brain. Despite
this decrease in flow, differences in the arteriovenous oxygen content are seen to either
decrease or remain unchanged, which implies that the oxygen supply is adequate to meet the
metabolic requirements.

With cooling, heart rate decreases but contractility remains stable or may actually increase.
Dysrhythmias become more frequent as temperature decreases and may include nodal,
premature ventricular beats, atrioventricular block, atrial and ventricular fibrillation, and
asystole. The mechanism of this dysrhythmogenic effect is unknown but may involve
electrolyte disturbances, uneven cooling, and autonomic nervous system imbalance. Because
coronary blood flow is well preserved during hypothermia, it is unlikely that myocardial
hypoxia plays a role in the genesis of these dysrhythmias.
The pulmonary system is characterized by a progressive decrease in ventilation as the
temperature is lowered. Physiologic and anatomic dead space increases during dilation of the
bronchi by cold. Gas exchange is largely unaffected.

The kidneys show the largest proportional decrease in blood flow of all the organs.
Hypothermia increases renal vascular resistance, with diminished outer and inner cortex blood
flow and oxygen delivery. Tubular transport of sodium, water, and chloride are decreased, and
the ability to concentrate becomes impaired. Tubular reabsorption is decreased. Urine flow
may be increased with hypothermia, but this effect can be masked by the stress-induced
release of antidiuretic hormone (ADH). The ability of the hypothermic kidney to handle glucose
is impaired, and glucose often appears in the urine. Hemodilution in combination with
hypothermic CPB improves renal blood flow and protects the integrity of the renal tubules
postoperatively.

In general, significant hepatic injury with hypothermic CPB is rare. Hepatic arterial blood flow is
reduced in proportion to the fall in cardiac output. The most significant effect of hypothermia
is the decrease in metabolic and excretory function of the liver. Obviously, drug actions and
requirements will be modified by this change in liver function. With rewarming, hepatic
efficiency reverts to normal.

Marked hyperglycemia is often a feature of hypothermic CPB. Endogenous insulin production

is decreased, and glycogenolysis and gluconeogenesis may be increased because of increases
in catecholamines. Even if exogenous insulin is administered, its efficacy is reduced during
hypothermia, and hyperglycemia may develop.

It is difficult to separate the effects of hypothermia from those of hemodilution and CPB.
Tissue water content is increased during hypothermic bypass, primarily as a consequence of
hemodilution (54). Cell swelling and edema occur, which may be related to an accumulation of
sodium and chloride within cells secondary to a decrease in reaction rates of membrane Na+ -
K+ -ATPase (15). Hypothermia decreases free water clearance and causes a decrease in plasma
potassium and an increase in osmolarity.

Hypothermia causes marked changes in the peripheral circulation. Systemic and pulmonary
vascular resistance typically rise with cooling below 26°C (55). This increase in vascular
resistance relates to increases in blood viscosity and catecholamines, hemoconcentration, cell
swelling, and perhaps active vasoconstrictor substances in the lung. In addition, arteriovenous
shunts appear at low temperatures (56) and may cause a further diminution in tissue oxygen
delivery. The increase in blood viscosity occurs because of fluid shifts, with loss of plasma
volume from capillary leak and cell swelling. The red blood cell volume remains unchanged
although the hematocrit rises. Red blood cell aggregation and rouleaux formation can occur,
further impeding blood flow. These changes can be somewhat attenuated by adequate
anesthesia, hemodilution, heparinization, and the use of vasodilators. Hypothermia also causes
thrombocytopenia by a reversible sequestration of platelets in the portal circulation.
The hormonal response to hypothermia depends on the level of anesthesia. Nonanesthetized
subjects demonstrate a marked sympathetic response to cold. This response can be almost
ablated if deep anesthesia is used. After deep hypothermia and total circulatory arrest, a
massive release of catecholamines occurs (57), which may contribute to the impaired cerebral
perfusion found by Greeley et al. (58). Corticosteroid release is suppressed with long-term
hypothermia below 28°C, but appears to be normal during short periods of hypothermia (59).
Complement activation occurs during CPB and is associated with neutrophil activation.
Respiratory complications correlate with the degree of complement activation (60).
Hypothermia, hemodilution, and heparin reduce complement activation and subsequent
neutrophil response and may protect patients from harmful sequelae. Circulating bradykinin
increases during hypothermia and CPB and may contribute to altered vascular permeability
and circulatory instability (61).

CLINICAL USE OF HYPOTHERMIA

Currently, hypothermia is used most commonly in cardiac surgery, although its use has also
been described for major vascular procedures, intracranial surgery, and removal of hepatic and
renal tumors. In most cardiac procedures, mild to moderate systemic hypothermia (>25°C) is
used for its protective effects, as previously described. More profound selective myocardial
hypothermia is also often used during aortic cross-clamping to aid in the preservation of
ischemic myocardium. Myocardial hypothermia is typically obtained in two ways: by coronary
perfusion with cold cardioplegic solution, and by topical application of an ice slush or cold
pericardial lavage. The optimal temperature for myocardial protection is controversial;
however, most studies have demonstrated superior protection at temperatures as low as 2°C
to 4°C, as long as freezing temperatures are avoided, alkalosis is present, and the heart is
promptly arrested during cooling (62).

Deep Hypothermic Circulatory Arrest

The most dramatic application demonstrating the protective effects of hypothermia is in

DHCA. Systemic temperatures of 20°C to 22°C or less are used to allow cessation of the
circulation for periods up to 40 to 60 minutes, often without detectable organ injury (9). DHCA
can be used in a variety of situations. In pediatric cardiac surgical patients (particularly those
weighing <8–10 kg), the repair of complex congenital cardiac lesions is often facilitated by the
asanguineous surgical field provided with circulatory arrest. It is often used in procedures
requiring occlusion of multiple cerebral vessels, particularly repair of aortic arch aneurysms. It
may be used to enhance surgical exposure and speed in procedures that could lead to
uncontrollable hemorrhage.

Central Nervous System Effects of DHCA

The brain is the organ at greatest risk for injury; this fact limits the duration of “safe” arrest
time. Cerebral metabolic activity is decreased with temperature, but it never ceases
altogether, even at temperatures approaching 0°C. As mentioned earlier, the protective effect
of hypothermia may involve more than just a reduction in cerebral metabolic rate. A Q10 of
2.7 would predict only a “safe” arrest time of approximately 15 minutes at 20°C. Clinical and
experimental evidence indicate, however, that 30 to 45 minutes is typically tolerated, so there
appears to be a disproportionate cerebral protective effect to profound levels of hypothermia.
Other factors, such as extracellular pH, may play a role. Swain et al. (63) showed that
hypothermia significantly increases the tissue energy state and intracellular pH in both heart
and brain. This increase in high-energy phosphate levels may partially explain the beneficial
effects of hypothermia on organ tolerance to ischemia. On the other hand, it may be that
cerebral oxygen consumption decreases in a nonlinear manner and more precipitously with
profound hypothermia than was previously thought. Michenfelder and Milde (64) showed a
change in Q10 from 2.23 between 37°C and 27°C to 4.53 between 27°C and 14°C. They
postulated that this marked drop in oxygen consumption at lower temperatures could be
explained by a primary effect of hypothermia on integrated neuronal function (as shown by
suppression of the electroencephalogram).

The rate of cooling also appears to be important in the occurrence of brain injury. Wide
gradients between body and perfusate temperature in dogs correlated with brain cell necrosis
and death (65). The optimal site for temperature monitoring is controversial, but it must be
remembered that gradients exist among the different regions (66) (Fig. 8.12). Monitoring
multiple sites to ensure uniform cooling before circulatory arrest is advisable. Coselli et al. (67)
suggested using electroencephalographic monitoring to determine the ideal depth of cooling
for safe circulatory arrest. They advocated using electroencephalographic silence as the
appropriate end point in cooling, but found that no peripheral body temperature consistently
predicts this level of hypothermia. There was a wide variation in temperature among body
sites when electroencephalographic silence occurred.
Figure 8.12 Average temperature [±SEM (standard error of mean)] of arterial cannula,
myocardium, cerebral cortex, nasopharynx, and rectum during 40 minutes of cooling and 90
minutes of rewarming under cardiopulmonary bypass in six pigs.(From Stefaniszyn HJ, Novick
RJ, Keith FM, et al. Is the brain adequately cooled during deep hypothermic cardiopulmonary
bypass? Curr Surg 1983;40:294–297, with permission.)
A consequence of ischemia and anoxia is the “no-reflow” phenomenon. The cerebral
microcirculation can shut down multifocally, causing incomplete reperfusion when flow is
resumed. The etiology of this problem is not completely understood, but may involve
increased blood viscosity, vascular smooth-muscle contraction resulting from increased
extracellular potassium, and precapillary shunting (68). It can occur with or without total
circulatory arrest and can be prevented by hypothermia (39). Microscopic cellular damage in
the brain occurs to some degree following hypothermia to 18°C, regardless of whether
pulsatile or nonpulsatile perfusion or total circulatory arrest is employed (69).

The main concern with the use of DHCA is the potential harmful effects on the organ most at
risk, the brain. As mortality decreases with improvements in technique, the question of an
effect on later intellectual development after hypothermic bypass, with or without circulatory
arrest, becomes critical. Several studies have shown evidence of a decreased intelligence
quotient and developmental capacity related to the duration of circulatory arrest (70,71,72).
However, other studies have not been able to demonstrate an adverse effect on intellectual
capacity and development when circulatory arrest times were less than 60 minutes at
nasopharyngeal temperatures of approximately 20°C (73,74,75). It is difficult to interpret many
of these studies because of the difficulty in defining an appropriate control group. Blackwood
et al. (76) used each child as his or her own control and found no difference between
preoperative and postoperative scores with arrest intervals as long as 74 minutes. In 1993,
Newburger et al. (77) reported a greater neurologic risk to neonates and infants with DHCA
than with continuous low-flow bypass. They reported modest but statistically significant
reductions in psychomotor developmental performance in those children in the circulatory
arrest group (78). Many investigators have tried to determine a “safe” duration for DHCA, but
the answer is still not known. Extensive clinical experience suggests that periods of DHCA as
long as 60 minutes are often well tolerated. Although patients vary widely, the data of
Newburger et al. suggest minimal adverse effects on psychomotor test results with circulatory
arrest times of approximately 35 minutes at 18°C.

Choreoathetosis

Choreoathetosis has been reported in 1% to 20% of children undergoing DHCA (74,79).

Choreoathetosis usually appears 2 to 6 days postoperatively and generally lessens in severity
with time. However, in severe cases, choreoathetoid movements or generalized hypotonia
may persist indefinitely. Choreoathetosis has also been occasionally observed following
continuous CPB, especially with profound hypothermia (10°C–12°C) (80). Numerous etiologies
have been proposed, including hyperglycemia (81), uneven cooling (77), the no- reflow
phenomenon (39,82), dopaminergic neurotransmitter alterations (83), and cerebral excitatory
amino acid neurotoxicity (84). It is generally thought that this hyperkinetic movement disorder
is a result of injury to basal ganglia, although often the pathology is not detectable by
conventional cranial computed tomography or magnetic resonance imaging (MRI). It appears
that there may also be an age-related phenomenon with regard to choreoathetosis. Data from
Boston Children’s Hospital suggest that the most vulnerable period starts at 6 to 9 months and
ends after 5 to 6 years (85). A mild, transient form of the condition was observed to develop in
younger children, whereas older children manifested a severe, persistent disorder with a high
mortality. The Boston group also noted that choreoathetosis is most likely to occur in children
with significant systemic–pulmonary collateral vessels. Perhaps the phenomenon of “steal”
from the cerebral circulation to the pulmonary arteries, particularly in the setting of
inadequate brain cooling, contributes to this problem.

Seizures

Seizures following CPB in neonates are much more common than in adults. Seizures have been
reported to occur clinically in as many as 20% of neonates following CPB (86,87). Seizures
detected by electroencephalogram occur even more commonly than clinically observed
seizures, a fact to be kept in mind if it becomes necessary to induce pharmacologic paralysis in
the postoperative period. The seizures are generally self-limited and can occur irrespective of
whether circulatory arrest has been used. Some series have reported no long-term adverse
sequelae, whereas others have suggested a decrement in psychomotor developmental
performance in addition to neurologic and MRI abnormalities (88,89,90). The long-term
prognosis for these children is still unclear. The question of whether seizures themselves
actually add to the damage, or are just a reflection of severe underlying brain pathology, also
arises. At present the answer is unknown, and further research is necessary in this important
area. Also, the relative involvement of hypothermia or other elements of CPB has not been
determined.

Therefore, the question of a “safe” circulatory arrest time is complex and cannot be answered
with certainty. Hypothermia can delay, but not prevent, the appearance of metabolic and
structural changes during ischemia that lead to functional neurologic impairment. A
nomogram has been devised that, although not rigorously defined, provides a best estimate of
the safe circulatory arrest times at three temperatures (9) (Fig. 8.13). Although patients vary
widely, the data of Newburger et al. (77) in children suggest minimal adverse effects on
psychomotor testing with circulatory arrest times of approximately 35 minutes at 18°C.
Figure 8.13 Nomogram of an estimate (not rigorously derived) of the probability of “safe” total
circulatory arrest (absence of structural or functional damage) according to the arrest time at
nasopharyngeal temperatures of 37°C, 28°C, and 18°C.(From Kirklin JW. Cardiac surgery. New
York: Churchill Livingstone, 1993:61–127, with permission.)

Central Nervous System Cooling and Rewarming

Figure 8.14 Scatter plot of duration of core cooling (minutes of cardiopulmonary bypass before
DHCA) and developmental index for infants with “short” periods of core cooling (<20 minutes).
The best-fit regression line and its 95% confidence interval (CI) are shown. DHCA, deep
hypothermic circulatory arrest.(From Bellinger DC, Wernovsky G, Rappaport LA, et al. Cognitive
development of children following repair of transposition of the great arteries using deep
hypothermic circulatory arrest. Pediatrics 1991;87:701–707, with permission.)

Because deep hypothermia and very low-flow rates or circulatory arrest will most likely remain
important tools in cardiac surgery, especially in congenital heart surgery and aortic arch repair,
much effort has been expended in trying to understand better the physiology involved. Newer
monitoring and protective strategies are being evolved in an attempt to improve cerebral
outcome in these patients. One of the areas of focus has been cooling methods before
circulatory arrest. The rate of cooling and the efficiency of brain cooling are important factors
in neurologic protection. There is marked variability in cerebral cooling, and sufficient time
must be allotted before circulatory arrest to ensure that there is uniform brain cooling. One
group of investigators reported slower brain cooling in one-third of the neonates studied by
jugular venous saturation monitoring (91). It appears that cooling periods of less than 20 to 25
minutes before the initiation of circulatory arrest with α-stat regulation are associated with a
lower developmental quotient in neonates undergoing DHCA (92) (Fig. 8.14). Uniform cerebral
cooling may also be problematic in certain groups of patients. Cyanotic patients with
aortopulmonary collaterals appear to be at increased risk for neurologic injury during CPB (93).
A reduction in total and regional measures of CBF occurs in these patients, resulting in less
efficient cerebral cooling. Because the brain is acidotic and a tremendous metabolic debt
develops during DHCA, the method of reestablishing CPB and rewarming may also be critically
important. Some intriguing data suggest that a period of cold reperfusion with delayed
rewarming after reestablishment of CPB may be beneficial (94). These authors demonstrated
an improvement in CBF velocity in the group of infants in whom rewarming was delayed by 10
minutes. It is also becoming clear that too aggressive rewarming, with hyperthermic
overshoot, is deleterious to neurologic outcome and should be avoided. Martin et al. (95) and
Mora et al. (96) reported an increased risk for neurologic injury in adult patients actively
warmed to 37°C, with a CPB circuit water bath maintained at 39°C to 40°C, in comparison with
patients allowed to “drift” to systemic temperatures of 33°C to 36°C in a more tepid approach
to warm heart bypass. They suggested that many, if not all, of their “normothermic” patients
experienced cerebral hyperthermia (>37°C), which may potentiate ischemic pathologic
processes in the brain during cardiac operations. Animal data have demonstrated that even
small (2°C) increases in brain temperature can exacerbate neuronal injury, with changes in
blood–brain permeability, increases in postischemic release of glutamate, and increased
mortality (97,98). Elevated temperatures must be strenuously avoided, not just in the
operating room but also in the postoperative period. Several authors have demonstrated that
hyperthermia occurs commonly after cardiac surgery in both adults and children (99,100,101).
The etiology is not entirely related to CPB per se, because patients having off-pump coronary
artery bypass surgery also commonly have fever postoperatively. Aggressive cooling measures
should be taken because this hyperthermia has been shown to be deleterious to outcome.
Grocott et al. (7) carefully studied 300 patients with hourly postoperative temperatures and
cognitive testing before and 6 weeks after surgery. They demonstrated that the maximum
postoperative temperature was associated with a greater degree of postoperative cognitive
dysfunction. Of course, it is not entirely clear whether the fever observed postoperatively is in
itself deleterious or whether it is a manifestation of an exaggerated inflammatory response
with subsequent brain injury. Further research is necessary in this area.

Tepid Temperature Management

Over the last few decades, an interest in tepid CPB has arisen. There are many theoretical and
clinically evident advantages and disadvantages detailed in the literature. Several variables
come into play when comparing multiple studies utilizing tepid CPB to normothermia, mild
hypothermia, and moderate hypothermia. This section will attempt to summarize the current
data available. Based on a recent consensus paper on hypothermic circulatory arrest (102), we
will use the following definitions for the degree of hypothermia throughout this section:
profound hypothermia as ≤14°C, deep hypothermia as 14.1°C to 20°C, and moderate
hypothermia as 20.1°C to 28°C. The article defines mild hypothermia as 28.1°C to 34°C. For this
section, we will consider mild hypothermia as 28.1°C to 32°C and consider tepid temperatures,
often termed “drift”, to range from 32.1°C to 36°C and normothermia as ≥36°C.

In the early 1990s, Tonz et al. (103,104) described theoretical objections to warm CPB
including extended cellular and humoral activation, decreased tolerance to ischemia resulting
in more evident effects of nonhomogeneous perfusion, and additional blood trauma due to
higher flow rates. Although logical, the evidence available does not support these objections.
One retrospective review comparing normothermia to moderate hypothermia found a
significant decrease in intubation times with the use of normothermic CPB. In addition, there
was no difference in mortality, myocardial infarction rates, or use of an intra-arterial balloon
pump for support. This study also found significantly shorter CPB times with the use of warm
cardioplegia during normothermic CPB compared to the use of cold cardioplegia, although
significantly higher doses of warm cardioplegia were necessary (104). Two small prospective
studies compared coronary artery bypass graft (CABG) and valve procedures utilizing moderate
hypothermia to normothermia (103,105). Lehot et al. focused on hemodynamic and hormonal
responses and found significantly longer cross-clamp and CPB times in the moderate
hypothermia group compared to the normothermic group. Hemodynamically, tachycardia was
significantly greater in the moderate hypothermia group at the termination of CPB. Although
systemic vascular resistance (SVR) was significantly lower in the normothermic group,
concentrations of norepinephrine and epinephrine were significantly higher in the
normothermic group. There were no other significant humoral responses found between the
two groups (105). A decrease in SVR was also described by others (103,104,105). Tonz et al.
(103) reported significant hemodynamic changes including a lower SVR and higher cardiac
index immediately postoperatively in the normothermic group, but this difference disappeared
by postoperative day 1 (POD 1). In this study, the total volume of blood shed through chest
and pleural tubes was significantly higher in the moderate hypothermia group. This group also
had an increased use of blood products including packed red blood cells and fresh frozen
plasma. There was no significant difference in inflammatory mediators or blood chemistries in
this study (103).

Engelman et al. (106) randomized 291 CABG patients to moderate hypothermia, tepid
temperatures, or normothermia. They found no significant difference among groups in
mortality, length of stay, or atrial arrhythmia. A subset of 53 patients was further analyzed to
evaluate for fibrinolysis potential. Of these, 42 patients were tested for preoperative and
postoperative prekallikrein activity. There was a significantly increased degradation of
prekallikrein activity with increasing temperatures. The authors state that this indicates
increasing fibrinolysis at normothermia, although from their larger patient population of 291
patients, they were unable to show any significant difference in blood loss or product use at
the three different temperatures. In 2002, Guadino et al. (107) evaluated hemostatic and
inflammatory activation in 113 CABG patients at moderate hypothermia or normothermia.
There was no significant difference in mortality, myocardial infarction, blood loss, or product
use at the different temperatures. There was no difference in hemostatic markers including
fibrinogen concentration, plasminogen activator inhibitor, and coagulation labs. They were
also unable to detect a significant difference in inflammatory markers including C-reactive
protein, interleukin-6, white blood cell, neutrophil, and monocyte counts.

In the last two decades, several prospective studies (some small and some well-powered) have
compared neurologic outcomes after CPB at various temperatures. One prospective study
from Duke University randomized 300 CABG patients to receive either tepid CPB (35.5°C–
36.5°C) or mild hypothermia (28°C–30°C) (108). Both groups received intermittent cold
cardioplegia, and the maximal perfusate temperature for rewarming was 38°C. All patients
underwent preoperative cognitive testing and the patients in the mild hypothermia group
were found to have slightly decreased performance preoperatively. There was no significant
difference in neurocognitive decline at 6 weeks. In another well-powered study, Heyer et al.
(109) found no difference in postoperative neurologic or neuropsychometric performance in
99 CABG patients randomized to moderate hypothermia or tepid CPB. Guadino et al. reports
that although the prevalence of neurologic events was similar between groups in his study,
there was an association with superior extension of brain damage and worse neurologic
outcomes, specifically in those patients who suffered intraoperative strokes with
normothermic CPB. Additionally, there was 6.6% mortality among the hypothermic group,
whereas the normothermic group had a 50% mortality rate in patients with intraoperative
stroke (110). Mora et al. (111) compared 138 patients undergoing CABG procedures with
moderate hypothermia or tepid temperatures. They report a significantly higher rate of new
perioperative central neurologic deficits in the tepid group. At baseline, the tepid group had a
significantly higher mean age which may have contributed. In addition, cardioplegia
administration varied between the two groups. The tepid group received continuous
retrograde cardioplegia after initial antegrade administration while the moderate hypothermia
group received only intermittent antegrade and retrograde cardioplegia.

In another article, using the same study population of 300 CABG patients described above, the
Duke group compared the change in creatinine clearance at different temperatures (112).
There were no significant differences in the change in creatinine clearance or the lowest
creatinine clearance values between the mild hypothermia and tepid CPB groups. Based on
these results, the authors state that there is no renal protective effect from hypothermia
although only creatinine clearance was evaluated in this study. Although maximal perfusate
temperature was well controlled in this study, there was a higher mean duration of
temperatures >37°C and an increased mean temperature >37°C in the mild hypothermia group
(108,112). There is strong evidence to support the detrimental effects of hyperthermia on
neurocognitive function and acute kidney injury (AKI) (113,114,115,116) and the need to
rewarm patients increases their risk of hyperthermia. A 2002 article also suggests improved
outcomes with slower rewarming (114). This study compared 165 patients rewarmed with
perfusate temperatures either 2°C or 6°C apart and found lower peak temperatures and a
significantly greater improvement in cognitive performance with slower rewarming. A recent
observational study of 1,393 patients describes a 34% increase in likelihood of AKI for every 10
minutes of arterial outlet temperatures >37°C (116).

Analysis of the varying studies leaves one somewhat confused as to the appropriate
temperature management for CPB. It appears that the majority of studies do not show a
marked difference in outcomes by using hypothermia for routine cases. There may be some
limited benefit to lower temperatures in minimizing neurologic injury in the setting of a stroke.
So there may be some benefit to letting the patient’s temperature drift, using a tepid
approach. However, strict attention to a careful rewarming strategy is imperative to avoid
inadvertent hyperthermia, which is likely to exacerbate any injury.

Cardiac surgery has advanced remarkably during the last 50 years. Hypothermia has
contributed substantially toward improving patient outcome in selected cases. Current efforts
must be directed toward defining methods of maximizing cerebral protection and refining
critical techniques.

KEY Points

● Acid–base balance is significantly affected by hypothermia.

● Biologic fluids are water solutions of weak acids and weak bases.
● Hypothermia decreases the tendency for weak acids and bases to dissociate in
solution.
● Blood pH is maintained at moderately alkaline values (0.4 pH units) relative to water.
● Maintenance of a constant ratio of [OH−] to [H+] (16:1) as temperature decreases
allows optimum function of many respiratory enzymes.
● α-stat regulation preserves the ratio of [OH−] to [H+] with change in temperature and
produces an alkaline shift with cooling.
● pH-stat regulation maintains an absolute constant [H +] regardless of temperature,
and requires added H+, usually as CO2, with cooling.
● The solubility of gases in biologic fluids increases with hypothermia.
● At a constant CO2 content, Paco2 decreases as temperature falls.
● Appropriate acid–base strategy during hypothermic CPB may be different between
children and adults based on different mechanisms of cerebral injury.
● pH-stat may be beneficial in infants to increase CBF and allow more efficient cooling.
● α-stat appears advantageous in adults by limiting the microembolic load to the brain.
● Protection of the brain during deep hypothermia (temperature <20°C) may be best
accomplished with a mixed acid–base strategy:
● pH-stat during the initial cooling phase.
● α-stat during reperfusion, rewarming, and termination of CPB.
● Hypothermia causes a decrease in blood flow to all vascular beds in proportion to the
reduced metabolic demands.
● The most profound effects occur in skeletal muscles and the extremities, followed by
the kidneys, splanchnic bed, heart, and brain.
● Hyperthermia occurs commonly after cardiac surgery and should be aggressively
treated because it may worsen ischemic damage.
TRADUÇÃO

MANEJO DA TEMPERATURA NA CIRURGIA CARDÍACA

O uso da hipotermia como auxiliar no tratamento de uma grande variedade de doenças

tem sido defendido há séculos. A temperatura corporal reduzida tem sido empregada
para combater o câncer, infecções, traumas e doenças do sistema nervoso central, e
como um método regional para induzir anestesia para amputação (1,2). No entanto, foi
somente em 1950 que Bigelow et al. (3) demonstraram maior tolerância à oclusão de
afluência em animais hipotérmicos do que em seus homólogos normotérmicos. Este
trabalho levou à primeira aplicação clínica da hipotermia na cirurgia cardíaca. Lewis e
Taufic (4) utilizaram resfriamento de superfície a 28°C com 5,5 minutos de oclusão de
afluência para facilitar o fechamento bem-sucedido de um defeito do septo atrial em
uma criança de 5 anos de idade. Em 1952, Gibbon (5) introduziu o oxigenador de
bomba na prática clínica, e em 1958, Sealy et al. (6) utilizaram a hipotermia em
conjunto com o circuito de bypass cardiopulmonar (CPB) para reparos intracardíacos.

O uso do oxigenador de bomba e da hipotermia permitiu que a cirurgia cardíaca

florescesse. Lesões complexas são reparadas rotineiramente com mortalidade
notavelmente baixa. No entanto, agora é reconhecido que a hipertermia inadvertida
ocorre com relativa frequência no ambiente perioperatório. A evidência está se
acumulando de que temperaturas elevadas podem ser prejudiciais tanto em animais
quanto em humanos (7,8). Uma melhor compreensão dos princípios do manejo da
temperatura pode maximizar os benefícios para nossos pacientes, minimizando
complicações potenciais.

FISIOLOGIA DA HIPOTERMIA

Uma das principais dificuldades em elaborar uma estratégia razoável para a aplicação da
hipotermia em humanos é o fato de que eles são seres naturalmente homeotérmicos.
Humanos e outras espécies homeotérmicas têm sistemas homeostáticos muito eficazes,
que garantem que a temperatura corporal permaneça consistentemente próxima a 37°C,
independentemente das mudanças nas temperaturas ambientais. Esta regulação rigorosa
da temperatura é realizada por múltiplos mecanismos. O frio é percebido pelos
termorreceptores na pele, o que então faz com que o hipotálamo desencadeie uma forte
resposta do sistema nervoso simpático. A vasoconstrição dos vasos da pele, que diminui
a perda de calor convectivo, ocorre simultaneamente com a vasodilatação dos leitos
vasculares dos músculos esqueléticos, o que aumenta a atividade muscular para produzir
calor por tensão e tremores. O sistema endócrino é ativado, o consumo de oxigênio é
aumentado, e a frequência cardíaca, o débito cardíaco e a pressão arterial são elevados.

Devido à complexidade dessas interações, é possível apreciar a dificuldade em entender

a resposta fisiologicamente apropriada ao estado não natural da hipotermia induzida em
humanos. É preciso extrapolar a partir de estudos em animais, equações bioquímicas,
sobreviventes de hipotermia acidental e gradientes de temperatura de órgãos normais
para tentar desenvolver a estratégia de manejo mais eficaz ao usar a hipotermia
deliberada.
Justificativa para o Uso da Hipotermia

Por que a hipotermia é frequentemente empregada durante o CPB? O propósito óbvio

de usar a hipotermia é fornecer um grau de proteção e margem de segurança para órgãos
(e para o organismo) durante o CPB. A hipotermia exerce seu efeito protetor por
múltiplos mecanismos. O mecanismo mais óbvio é a redução na taxa metabólica e no
consumo de oxigênio (9) (Fig. 8.1). Esta supressão metabólica pode não explicar todos
os efeitos protetores observados.

A hipotermia também ajuda a preservar os estoques de fosfato de alta energia e reduz a

liberação de neurotransmissores excitatórios, o que é especialmente importante para a
proteção do sistema nervoso central (10,11). Normalmente, as células neuronais
isquêmicas liberam rapidamente aminas neuroexcitatórias, especialmente o glutamato
(12). O acúmulo dessas aminas excitotóxicas causa a abertura dos canais de cálcio e a
ativação de múltiplos sistemas enzimáticos destrutivos. A hipotermia atenua essa
cascata excitotóxica, ajudando a prevenir a entrada de cálcio na célula e restringindo a
permeabilidade da membrana (13) (Fig. 8.2).

A alteração da temperatura causa uma mudança na taxa de reação de todos os processos

bioquímicos, especialmente as reações enzimáticas. Essa dependência da taxa de reação
em relação à temperatura tem sido descrita pelo conceito de Q10, que é definido como o
aumento ou diminuição nas taxas de reação ou nos processos metabólicos em relação a
uma mudança de temperatura de 10°C. Por exemplo, a taxa de reação de um processo
com um Q10 de 2 dobrará com um aumento de 10°C na temperatura ou será reduzida
pela metade com uma queda de 10°C. A maioria das reações, incluindo o consumo total
de oxigênio do corpo, tem um Q10 de 2 a 3 (9).

Alguns processos bioquímicos, especialmente aqueles localizados nas membranas

celulares, mostram uma mudança abrupta nas taxas de reação em certas temperaturas
críticas. Isso tem sido chamado de transição de fase e é considerado o resultado de uma
mudança na membrana celular de um estado fluido para um estado de gel (14). Em
tecidos de mamíferos, as transições de fase ocorrem frequentemente a aproximadamente
25°C a 28°C e podem perturbar a homeostase celular. Processos biofísicos como a
osmose e a difusão da água também são afetados pela temperatura.

Normalmente, uma mudança linear de aproximadamente 3%/10°C é vista. Portanto, este

efeito é mínimo em níveis clínicos de hipotermia. No entanto, se o ponto de
congelamento da água for atingido, o gelo é formado no tecido, uma condição que não é
tolerada. Os solutos se concentram de maneira hiperosmolar na água residual não
congelada, causando marcadas mudanças de fluidos e ruptura da membrana. O tecido de
mamíferos não recupera a função ao descongelar de um estado congelado. Por esta
razão, há um limite para os efeitos benéficos da hipotermia.

Na cirurgia cardíaca, o CPB em conjunto com a hipotermia sistêmica permite fluxos de

bomba mais baixos, melhor proteção miocárdica, menos trauma sanguíneo e melhor
proteção de órgãos do que a perfusão normotérmica (15). As necessidades de oxigênio
caem previsivelmente com a temperatura reduzida. Foi reconhecido precocemente que
fluxos de bypass reduzidos poderiam ser empregados neste cenário e ainda fornecer
perfusão adequada, conforme avaliado pela tensão de oxigênio venoso misto e pelo
retorno da função do órgão após o bypass. Relacionar o consumo de oxigênio (VO2)
com a taxa de fluxo de perfusão em várias temperaturas também pode ser valioso na
avaliação da adequação da perfusão tecidual (Fig. 8.3). A uma dada temperatura, uma
queda no VO2 com uma diminuição na taxa de fluxo implica um VO2 limitado pelo
fluxo, indicando que a entrega de oxigênio não é adequada. Hickey e Hoar (16)
mostraram em humanos que uma redução na taxa de fluxo de 2,1 para 1,2 L/min/m2 de
área de superfície corporal a 25°C não altera o VO2 ou a perfusão tecidual. Slogoff et
al. (17) não conseguiram correlacionar fluxos baixos (<40 mL/kg/min) ou pressões (<50
mm Hg) durante o bypass, no qual hipotermia moderada e hemodiluição foram usadas,
com disfunção renal ou do sistema nervoso central pós-operatória.

Taxas de fluxo de perfusão mais baixas permitem uma melhor visualização pelo
cirurgião. O retorno venoso dos vasos brônquicos, pulmonares e colaterais não
coronarianos também é diminuído. Como este sangue que retorna está em temperatura
sistêmica, ele pode aquecer inadequadamente o coração quando a hipotermia
miocárdica induzida por cardioplegia está em uma temperatura menor do que a
sistêmica e pode comprometer a proteção miocárdica. O trauma sanguíneo é
minimizado tanto pelos fluxos de bomba mais baixos quanto pela hemodiluição
empregada durante o bypass. Como a etiologia da maioria dos danos ao sistema nervoso
central no bypass pode ser de origem embólica (18), fluxos de bypass mais baixos
podem minimizar esses insultos focais. A hipotermia sistêmica também fornece alguma
margem de segurança para a proteção de órgãos se ocorrer falha do equipamento ou se a
parada circulatória precisar ser empregada.

Alteração Ácido-Base com Mudança de Temperatura

Um dos aspectos mais discutidos da hipotermia clínica é a estratégia de manejo ácido-

base apropriada durante a hipotermia. Para entender a regulação ácido-base durante a
hipotermia, é útil considerar várias condições de exemplo. Os valores "normais" para
pH e PCO2 são geralmente considerados 7,40 unidades de pH e 40 mm Hg,
respectivamente. No entanto, deve-se ter em mente que esses valores são apropriados
apenas a 37°C no sangue. Existe um espectro de valores "normais" dependente da
temperatura, dependendo da temperatura do local específico no corpo.

Por exemplo, o sangue arterial sai do coração a 37°C com um pH de 7,40 e um PCO2
de 40 mm Hg. Quando esse mesmo sangue perfunde o músculo esquelético em
atividade, onde a temperatura ambiente pode ser de 40°C, ele terá um pH de
aproximadamente 7,35 e um PCO2 um pouco mais alto, apesar de um conteúdo
constante de CO2 antes de qualquer troca respiratória com o músculo. Em contraste, o
mesmo sangue arterial que perfunde a pele exposta, onde a temperatura pode ser de
20°C em clima frio, terá um pH de 7,65 e um PCO2 proporcionalmente mais baixo,
novamente sem mudança no conteúdo de CO2. Dito de outra forma, uma amostra de
sangue arterial mantida em uma seringa estanque ao gás terá um PCO2 que varia
diretamente com a temperatura, e um pH que varia inversamente com a temperatura,
apesar do conteúdo constante de CO2.

Essa mudança no PCO2 com a temperatura é uma consequência da mudança na

solubilidade de gases em líquidos com a mudança na temperatura. Como regra geral,
diminuir a temperatura de um líquido aumenta a solubilidade de um determinado gás
nesse líquido e, portanto, diminui a pressão parcial desse gás, enquanto o conteúdo geral
do gás no líquido permanece constante. Aumentar a temperatura de um líquido aumenta
a energia cinética das moléculas no líquido, o que aumenta a tendência das moléculas de
gás dissolvidas de deixar o líquido (solubilidade diminuída) e aumenta a pressão parcial
dessas moléculas de gás que permanecem no líquido. Esse conceito é intuitivo para
qualquer pessoa que tenha aberto uma garrafa de bebida carbonatada quente e
observado que sua tendência a "efervescer" (o CO2 borbulhando para fora da solução) é
muito maior do que a de uma bebida fria.

A relação do pH com a temperatura é um pouco mais complexa do que a mudança no

PCO2 com a temperatura. Há claramente uma relação direta entre a concentração de H+
[H+] em uma solução de água e o conteúdo de CO2 dessa solução. Quanto maior o
CO2, mais H+ em solução, e, portanto, menor o pH. Isso é em grande parte causado
pela tendência do CO2 de se combinar quimicamente com a água para produzir ácido
carbônico, que então se dissocia em solução para produzir H+. De acordo com isso,
pode-se esperar que a mudança no pH do sangue com a temperatura seja causada pelas
mudanças no PCO2 com a temperatura, conforme discutido no texto anterior. É verdade
que o resfriamento de uma amostra de sangue anaeróbia diminui tanto o PCO2 quanto o
[H+]. No entanto, e muito importante, a mudança em [H+] e pH que ocorre com a
mudança de temperatura é independente de uma mudança no conteúdo de CO2 e,
portanto, não depende da mudança na solubilidade do CO2 com a mudança de
temperatura.

Todos os ácidos e bases, incluindo a água, existem em solução em equilíbrio entre a

forma não dissociada e os componentes ionizados da molécula-mãe. A constante de
dissociação (K) é a razão de equilíbrio do produto das concentrações dos componentes
ionizados para o componente não ionizado. Para a água a 25°C, a equação de
dissociação de equilíbrio é a seguinte: K(H2O, 25°C) = [H+] × [OH−]/[H2O] = 1,08 ×
10−14. Como, em equilíbrio na água, [H+] é igual a [OH−], e como a concentração de
H2O é essencialmente 1, [H+] é igual a [OH−] é igual a √10−14, ou 10−7. Como a
definição de pH é o log10 negativo de [H+], o pH da água pura em equilíbrio a 25°C é
7. Este valor de pH passou a ser chamado de pH neutro, ou pN, da água.

A mudança de temperatura tem um efeito significativo na tendência das moléculas em

solução de se dissociarem. Em termos termodinâmicos, a energia cinética aumentada
associada ao aumento da temperatura promove a dissociação, enquanto a temperatura
diminuída tem o efeito oposto. Por exemplo, dentro da faixa de temperatura vista no
CPB clínico (aproximadamente 15°C–40°C), a constante de dissociação da água
aumenta de 0,451 × 10−14 para 2,919 × 10−14. Esses valores de KH2O se relacionam a
uma mudança em [H+] de aproximadamente 67 nmol/L a 15°C para 170 nmol/L a
40°C. Esta mudança de quase 3 vezes em [H+] é unicamente uma consequência dos
efeitos da mudança de temperatura na dissociação.

A água pura pode ser considerada a solução mais simples de ácido-base fraca. A
principal importância desses conceitos é que a água é o solvente fundamental de todos
os sistemas biológicos, e a dissociação de virtualmente todos os ácidos e bases fracos
em soluções biológicas segue o mesmo padrão que o descrito para a água. O
comportamento dos fluidos corporais (intracelulares e extracelulares, intravasculares e
extravasculares) é muito mais complexo do que o cenário simples descrito
anteriormente para a água, mas os fluidos biológicos se comportam muito como a água
em termos das mudanças intrínsecas relacionadas à temperatura nas constantes de
dissociação dos muitos ácidos e bases fracas, dos quais são compostos. Os aminoácidos
em proteínas, os açúcares simples em polissacarídeos, os ácidos graxos em lipídios e os
principais sistemas de tamponamento, todos seguem este mesmo padrão básico. À
medida que a temperatura diminui, a tendência à dissociação diminui, e as
concentrações dos componentes ionizados (H+ e R−) também diminuem.

Na temperatura corporal normal (37°C), o sangue e os fluidos teciduais são alcalinos

(menor [H+] e pH correspondentemente mais alto) em relação à água na mesma
temperatura. Vários sistemas de tamponamento criam e mantêm essa alcalinidade
relativa para que a razão de [OH−] para [H+] permaneça constante em
aproximadamente 16:1, apesar da variação de temperatura. À medida que a temperatura
muda, a dissociação intrínseca desses sistemas de tamponamento também muda para
manter a razão de [OH−] para [H+] constante. Portanto, a mudança de pH intrínseca do
sangue e do fluido tecidual é paralela ao pNH2O à medida que a temperatura muda, e
essa alcalinidade relativa permanece constante em comparação com a água (19) (Fig.
8.4).

Um importante sistema de tamponamento responsável por essa relação constante do pH

do sangue e do fluido tecidual com o pN, com a mudança de temperatura, é a fração
imidazólica do aminoácido histidina, que é comumente encontrada em proteínas do
corpo. O pKa desse componente da histidina é próximo a 7,0 na temperatura corporal,
uma propriedade que confere potente capacidade de tamponamento para manter uma
razão constante de [H+] para [OH−] apesar de mudanças significativas na concentração
absoluta de cada um à medida que a temperatura varia.

Essas considerações são aplicáveis ao exemplo anterior de sangue arterial com um

conteúdo constante de CO2 perfundindo tecidos com diferentes temperaturas. A
mudança observada no pH com o resfriamento segue o pNH2O, e a capacidade de
tamponamento da fração imidazólica da histidina preserva a alcalinidade relativa
constante e a razão de [OH−] para [H+] no sangue (Fig. 8.5). Em vertebrados de sangue
frio, a curva de pH-temperatura do sangue também corre paralela ao pH da água neutra.
O pH intracelular também foi medido em vários animais e mostra mudanças com a
temperatura idênticas às que foram descritas para o pNH2O (20) (Fig. 8.6). O pH
intracelular é paralelo às inclinações do pN e do pH do sangue com as mudanças de
temperatura e difere do pH extracelular por um fator constante, mas espécie-específico,
de aproximadamente −0,6 a −0,8 unidades de pH. Portanto, a 37°C, o pH intracelular é
de aproximadamente 6,8 a 6,9 e, portanto, o [H+] é um pouco mais alto.

De maneira semelhante à mudança observada no pH do sangue com a temperatura, a

cinética de reação de numerosos sistemas enzimáticos respiratórios (lactato
desidrogenase, Na+-K+-ATPase [adenosina trifosfatase sódio-potássio], acetil CoA
carboxilase, sintase de ácidos graxos, NADH [dinucleotídeo de adenina e nicotinamida
reduzido] citocromo c redutase e succinato citocromo c redutase) mostram uma função
catalítica ótima com a mudança de temperatura quando o pH do meio de reação é
paralelo à mudança de pNH2O mediada pela temperatura (21). Este meio interno
constante é realizado, como já foi mencionado, predominantemente pela capacidade de
tamponamento do grupo imidazol do aminoácido histidina. À medida que a temperatura
muda, os grupos imidazólicos na proteína mudam de pKa em paralelo com o pN da
água. A razão dos grupos imidazólicos de histidina não protonados para H+, um valor
conhecido no mundo da química como alfa, permanece constante, o CO2 total também
permanece constante e o pH muda de acordo com as mudanças na temperatura.
O termo

α-stat passou a indicar uma estratégia de manejo ácido-base na qual a carga líquida
(dissociação) das proteínas permanece constante à medida que a temperatura muda.
Tipicamente, isso é gerenciado durante o CPB mantendo os estoques totais de CO2
constantes e permitindo que o pH e o PaCO2 sigam suas mudanças de dissociação
mediadas termodinamicamente com as mudanças na temperatura. Em outras palavras,
durante o resfriamento, o CO2 exógeno não é adicionado ao sistema ao seguir a
estratégia α-stat.

O método alternativo de estratégia ácido-base é denominado

pH-stat. Com este método, o pH é o valor que é mantido constante em temperaturas

variáveis. Obviamente, se a estratégia pH-stat for usada quando o sangue for resfriado, o
CO2 deve ser adicionado para manter um PaCO2 de 40 e um pH de 7,40. As razões
extracelulares e intracelulares de [OH−] para [H+] são alteradas e os estoques totais de
CO2 são elevados.

Por que uma estratégia seria escolhida em detrimento da outra? Durante as duas
primeiras décadas de CPB hipotérmico, o manejo pH-stat com a adição de 5% de CO2
ao fluxo de gás do oxigenador foi usado quase que exclusivamente. A compreensão das
mudanças esperadas no pH com a temperatura parecia estar faltando, e o CO2 era
considerado benéfico para a vasodilatação cerebral e manutenção do fluxo sanguíneo
cerebral (CBF). Nos últimos 30 anos, essa prática foi questionada, e muitas instituições
mudaram para um protocolo de manejo α-stat. Foi apenas recentemente que dados
suficientes se acumularam para permitir que uma decisão racional fosse tomada para
empregar a mais adequada das duas estratégias em uma determinada situação.

Em uma base teórica, o manejo α-stat pode ser preferível em certas situações. A
manutenção da neutralidade eletroquímica intracelular constante parece ser essencial
para a função celular normal (22). Os intermediários metabólicos intracelulares de
fosfatos de alta energia podem ser esgotados se houver mudanças no pH intracelular e
esses metabólitos perderem seu estado carregado. Esses substratos então podem se
difundir livremente através das membranas lipídicas. A maioria das enzimas depende de
um pH ótimo para sua função. A neutralidade eletroquímica também é importante para
manter o equilíbrio de Donnan através das membranas celulares para permitir
concentrações normais de ânions intracelulares e conteúdo de água (23).

Animais poiquilotérmicos, cujos tecidos devem funcionar de forma ideal apesar de

amplas variações de temperatura, seguem uma estratégia ácido-base α-stat. Por outro
lado, mamíferos hibernando parecem manter uma estratégia pH-stat, com pHa e PCO2
de sangue constantes (corrigidos pela temperatura) (22). Esses animais hipoventilam
enquanto hibernam, os estoques de CO2 do tecido aumentam e o pH intracelular se
torna acidótico na maioria dos tecidos. Este estado acidótico causa uma depressão
adicional do metabolismo que teleologicamente pode ser útil ao diminuir ainda mais o
consumo de energia de tecidos não funcionantes, como músculo esquelético, trato
gastrointestinal e centros cerebrais superiores. Em contraste, tecidos ativos, como
coração e fígado, adotam uma estratégia diferente, extrudindo ativamente H+ através de
suas membranas celulares para manter o pH intracelular nos valores ou perto dos
valores previstos pela metodologia α-stat.
Portanto, mamíferos hibernando são capazes de variar seu gradiente de pH intracelular
para extracelular de forma diferente em diferentes tecidos, dependendo do estado de
atividade metabólica do tecido. Funcionalmente, isso fornece diferentes tipos de
regulação ácido-base em diferentes tecidos, dependendo da atividade metabólica do
tecido. A primeira mudança perceptível associada ao despertar da hibernação é a
hiperventilação. Isso esgota os estoques de CO2, eleva o pH intracelular e aumenta a
taxa metabólica. O animal reverte para um padrão geral de controle de pH α-stat durante
o despertar, o que permite que os tecidos recuperem a função ótima. Portanto, em
mamíferos hibernando, a questão da manutenção ácido-base não é claramente definida
porque a regulação ácido-base intracelular pode ser independente da regulação do
sangue tanto dentro quanto entre diferentes tecidos no mesmo animal.

Apesar da discussão anterior, a questão prática de como o estado ácido-base deve ser
regulado durante o bypass hipotérmico em humanos permanece aberta. Alguns estudos
em animais sugerem que o manejo ácido-base α-stat é benéfico em termos de proteção
miocárdica. McConnell et al. (24) avaliaram a regulação α-stat durante a hipotermia em
cães e demonstraram que elevações significativas no fluxo sanguíneo coronário,
consumo de oxigênio do ventrículo esquerdo e utilização de lactato ocorreram com a
manutenção de um pH de 7,7 a 28°C (α-stat) em comparação com um pH de 7,4 (pH-
stat). Houve também um aumento significativo na pressão ventricular de pico quando
um pré-carga padrão foi aplicada. Poole-Wilson e Langer (25) demonstraram uma maior
contratilidade em músculo papilar perfundido hipotérmico quando o pH do perfusato
era mais alcalino do que 7,4. Eles também demonstraram uma queda rápida na tensão
miocárdica, além de mudanças no fluxo de Ca2+ quando o PaCO2 do perfusato foi
aumentado (26). Por outro lado, Sinet et al. (27) não encontraram nenhum efeito do pH
no desempenho do coração isolado de rato.

O miocárdio muitas vezes não é perfundido, mas é propositadamente tornado isquêmico

para facilitar a cirurgia cardíaca. Neste cenário, a alcalinização do sangue antes da
isquemia foi mostrada para diminuir o desenvolvimento de acidose no sangue do seio
coronário e melhorar a contratilidade na reperfusão (28). Também parece que o pH do
sangue que reperfunde o coração pode ser crítico para a recuperação do desempenho
ventricular. Becker et al. (29) estudaram os efeitos miocárdicos de uma estratégia ácido-
base na qual a alcalinização maior do que a de α-stat foi usada. Eles encontraram
melhorias no desempenho miocárdico após 1 hora de parada circulatória e cardioplegia,
com alcalinização moderada em comparação com α-stat.

O manejo ácido-base também parece ser importante na eletrofisiologia cardíaca. Swain

et al. (30) mostraram que a estabilidade elétrica do coração aumentava, com menos
fibrilação ventricular espontânea, quando a regulação sanguínea α-stat era comparada
com a pH-stat. Kroncke et al. (31) encontraram uma incidência de 40% de fibrilação
ventricular em pacientes resfriados a 24°C durante o manejo pH-stat e uma incidência
de 20% naqueles gerenciados com α-stat.

O manejo ácido-base apropriado para a perfusão cerebral ideal também foi questionado.
Claramente, o CBF diminui significativamente com a hipotermia. A taxa metabólica
cerebral também diminui durante o bypass hipotérmico. A resposta da circulação
cerebral às mudanças no PaCO2 é preservada, pelo menos durante a hipotermia
moderada (32) ; portanto, o manejo α-stat resultará em fluxos cerebrais mais baixos do
que os vistos com o manejo pH-stat. No entanto, devido às demandas metabólicas
reduzidas, um CBF mais baixo pode ser apropriado e indicativo de um acoplamento
mantido do fluxo sanguíneo e da demanda metabólica. Govier et al. (33) demonstraram
autorregulação intacta em humanos usando uma estratégia α-stat em temperaturas que
variam de 21°C a 29°C. Murkin et al. (34) mostraram acoplamento de CBF e
metabolismo que era independente da pressão de perfusão cerebral (CPP) dentro da
faixa de 20 a 100 mm Hg quando o manejo α-stat foi empregado. Em contraste, a
autorregulação cerebral foi abolida e o CBF variou com a pressão de perfusão quando a
estratégia pH-stat foi usada (Fig. 8.7).

Argumentou-se que o CBF durante a hipotermia pH-stat na verdade representa um fluxo

sanguíneo excessivo e pode ser prejudicial. Fluxos sanguíneos desnecessariamente altos
podem colocar o cérebro em risco de danos por microêmbolos ou alta pressão
intracraniana. Com hipotermia profunda (ou seja, temperaturas <20°C), as respostas
vasculares normais são perdidas e o CBF torna-se dependente da pressão (35,36). Em
temperaturas hipotérmicas profundas, o acoplamento do fluxo cerebral e do
metabolismo também é perdido. É importante notar, no entanto, que as respostas do
CBF e CMRO2 (taxa metabólica cerebral de oxigênio) são quantitativamente diferentes
em condições de hipotermia profunda. O CBF diminui linearmente com a diminuição da
temperatura, enquanto o CMRO2 cai exponencialmente. O resultado líquido é que o
CBF se torna mais luxuoso em temperaturas hipotérmicas profundas. Na normotermia, a
razão média de CBF para CMRO2 é de 20:1, e na hipotermia profunda, a razão aumenta
para 75:1 (37). Esta situação é importante no contexto de CPB de baixo fluxo. Em
temperaturas muito baixas, os dados indicam que as taxas de fluxo da bomba podem ser
reduzidas para apenas 10 mL/kg/min antes que o fluxo se torne inadequado para os
requisitos metabólicos cerebrais (38).

Em um nível microcirculatório, algumas evidências sugerem que o manejo α-stat pode

ser benéfico para o cérebro. Norwood et al. (39) estudaram os cérebros de cães
hipotérmicos perfundidos com sangue anóxico e encontraram uma diminuição na
extensão e magnitude das lesões quando o perfusato tinha um pH mais alto. O perfusato
ácido aumentou a extensão das lesões. Por outro lado, Priestley et al. (40) compararam o
resultado neurológico e histológico em um modelo de sobrevivência de porquinho de
parada circulatória hipotérmica profunda. Eles mostraram que o grupo pH-stat
demonstrou melhor desempenho neurológico e pontuações de deficiência funcional
menos graves do que os porquinhos no grupo α-stat. A lesão histológica também foi
mais grave no grupo α-stat.

Teoricamente, a hipocarbia (e o pH aumentado) resultam em um desvio para a esquerda

da curva de dissociação da oxi-hemoglobina, o que faz com que o oxigênio seja menos
prontamente disponível para os tecidos. No entanto, mais oxigênio é dissolvido no
plasma durante a hipotermia, de modo que esses dois efeitos tendem a se anular. O CBF
relativamente baixo durante o manejo α-stat ainda foi mostrado como sendo em excesso
das necessidades metabólicas cerebrais (34).

Em adultos, a preponderância das evidências sugere que o manejo de CO2 no CPB não
importa ou que uma estratégia α-stat é vantajosa. Bashein et al. (41) examinaram a
influência do manejo do pH em 86 adultos nos quais a hipotermia leve foi utilizada
(aproximadamente 30°C). Eles não encontraram nenhuma diferença no resultado
cardíaco ou neuropsicológico, independentemente do manejo ácido-base. No entanto, é
importante perceber que teria sido improvável que eles pudessem demonstrar uma
diferença neste estudo sob as condições do estudo. As diferenças no PCO2 entre os dois
grupos totalizaram apenas aproximadamente 6 a 7 mm Hg, e o grau de hipotermia não
foi muito profundo. Contrasta-se o cenário em seu estudo com o de um paciente em
hipotermia profunda, para quem a diferença no PCO2 entre as duas estratégias se
aproxima de 80 mm Hg! Além disso, sua análise procurou por diferenças no
desempenho médio do grupo, em vez de mudanças no desempenho do paciente
individual, uma metodologia que pode ter diminuído a sensibilidade do estudo para
detectar qualquer diferença existente.

Três estudos prospectivos randomizados de hipotermia moderada em adultos

demonstraram que o resultado neurológico ou neuropsicológico pós-operatório é
ligeiramente, mas consistentemente, melhor com o manejo α-stat (42,43,44). A noção
de que o manejo α-stat pode ser benéfico neste cenário em adultos faz sentido se
considerarmos que o mecanismo mais provável para a lesão neurológica nesses
pacientes é provavelmente êmbolos. Portanto, o manejo α-stat seria esperado para
fornecer CBFs mais baixos que estão mais alinhados com a taxa metabólica cerebral e
uma menor carga embólica.

Embora o manejo ácido-base provavelmente não seja tão importante quando

temperaturas hipotérmicas moderadas são usadas, ele pode ser crítico no cenário de
hipotermia profunda. Proponentes do método α-stat sugerem que fluxos sanguíneos
desnecessariamente altos (com manejo pH-stat) podem colocar o cérebro em risco de
danos por microêmbolos, edema cerebral ou alta pressão intracraniana, ou podem, na
verdade, predispor a uma redistribuição adversa do fluxo sanguíneo ("roubo") para
longe de áreas marginalmente perfundidas em pacientes com doença cerebrovascular.
Por outro lado, os proponentes da estratégia pH-stat sugerem que o CBF aprimorado
pode ser útil para melhorar o resfriamento cerebral antes do início da parada
circulatória. De fato, o CBF total é aumentado, o resfriamento cerebral global é
aprimorado e o fluxo sanguíneo cerebral é redistribuído durante o manejo pH-stat. Uma
proporção aumentada de CBF é distribuída para estruturas cerebrais profundas (tálamo,
tronco cerebral e cerebelo) quando o manejo pH-stat é usado (45) (Figs. 8.8 e 8.9).

No entanto, outros dados sugerem que a recuperação metabólica cerebral após a parada
circulatória pode ser melhor com o método α-stat do que com o modo pH-stat. Esta
variação nos resultados levou alguns autores a defender uma estratégia de cruzamento
na qual uma abordagem pH-stat é usada durante os primeiros 10 minutos de
resfriamento para fornecer supressão metabólica cerebral máxima, seguida por uma
estratégia α-stat para remover a acidose grave que se acumula durante a hipotermia
profunda durante o pH-stat. Esta abordagem parece oferecer a máxima recuperação
metabólica em animais (46) (Fig. 8.10).

Figura 8.8 Fluxo sanguíneo regional no cérebro durante o manejo α-estatístico ou pH-
estatístico dos gases sanguíneos.

BS, linha de base na pressão sanguínea cerebral (PSBC) normotérmica; HT, no final de
30 minutos de bypass cardiopulmonar (BCP) hipotérmico; R5, 5 minutos após o início
da reperfusão e reaquecimento após 1 hora de parada circulatória; N0, após 45 minutos
de reaquecimento, quando a normotermia foi alcançada; N3, após 3 horas de reperfusão
na normotermia.
*p < 0,01, #p < 0,05 para a diferença de grupo. (De Aoki M, Nomura F, Stromski ME,
et al. Effects of pH on brain energetics after hypothermic circulatory arrest. Ann Thorac
Surg 1993;55:10931103, com permissão.)

Figura 8.9 Distribuição intracerebral do fluxo sanguíneo: α-estatístico versus pH-

estatístico.

Rep(5), 5 minutos após o início da reperfusão e reaquecimento após 1 hora de parada

circulatória; NT(0), após 45 minutos de reaquecimento, quando a normotermia foi
alcançada; NT(180), após 180 minutos de reperfusão na normotermia.

● indica p < 0,05 para o aumento dentro do grupo por teste t pareado;

● indica p < 0,05 para a diminuição dentro do grupo por teste t pareado. (De Aoki
M, Nomura F, Stromski ME, et al. Effects of pH on brain energetics after
hypothermic circulatory arrest. Ann Thorac Surg 1993;55:10931103, com
permissão.)

Figura 8.10 Esta figura demonstra os efeitos de três estratégias de resfriamento

diferentes (α-estatística, pH-estatística e uma transição de pH-estatística seguida por α-
estatística) na supressão metabólica cerebral antes da parada circulatória hipotérmica
profunda (PCHP) e a recuperação do metabolismo cerebral após a PCHP.

A adição de CO2 (ou seja, a estratégia pH-estatística) fornece uma melhor supressão
metabólica cerebral antes da PCHP, mas a recuperação metabólica cerebral após a
PCHP é fraca. O resfriamento inicial com uma estratégia pH-estatística, seguido pela
conversão para α-estatística antes da PCHP, resulta na maior recuperação metabólica
cerebral. CMRO2, taxa metabólica cerebral de oxigênio. (De Kern FH, Greeley WJ.
pH-stat management of blood gases is not preferable to α-stat in patients undergoing
brain cooling for cardiac surgery. J Cardiothorac Vasc Anesth 1995;9:215–218, com
permissão.)

Figura 8.11 Saturação cortical de oxigênio (ScO2) durante a PCHP nos grupos pH-
estatístico e α-estatístico.

Média ± DP, oito animais por grupo. *p < 0,05 entre os grupos. A meia-vida da ScO2
durante a parada foi significativamente maior no grupo pH-estatístico do que no grupo
α-estatístico. (De Kurth CD, O’Rourke MM, O’ Hara IB. Comparison of pH-stat and α-
stat cardiopulmonary bypass on cerebral oxygenation and blood flow in relation to
hypothermic circulatory arrest in piglets. Anesthesiology 1998;89:110–118, com
permissão.)

A escolha do manejo ácido-base pode ser particularmente importante no subgrupo de

pacientes pediátricos com colaterais aortopulmonares, para quem o resfriamento
cerebral é problemático. Aparentemente, a adição de CO2 durante o resfriamento
melhora a perfusão cerebral e a recuperação metabólica cerebral. Kurth et al.
demonstraram em um modelo com leitões dois mecanismos pelos quais o manejo pH-
estatístico pode ser benéfico para o cérebro. Eles mostraram que o pH-estatístico
aumenta a taxa de resfriamento cerebral e que a taxa de depleção de oxigênio cerebral
durante a parada circulatória hipotérmica profunda (PCHP) é consideravelmente mais
lenta com o manejo pH-estatístico do que com o manejo α-estatístico (Fig. 8.11). Além
do aumento do FSC visto com CO2 elevado, também se esperaria ver um aumento na
resistência vascular pulmonar e uma diminuição no fluxo sanguíneo pulmonar com a
estratégia pH-estatística. De fato, em um ensaio clínico randomizado com 40 crianças
cianóticas, Sakamoto et al. demonstraram uma redução na circulação colateral
sistêmico-pulmonar no grupo pH-estatístico. Eles também mostraram valores de
oximetria cerebral melhorados e níveis de lactato mais baixos nos pacientes do grupo
pH-estatístico. Não foram notados déficits neurológicos nos pacientes de nenhum dos
grupos. Um estudo recente, randomizado e de centro único, em bebês com menos de 9
meses, descobriu que os bebês que foram manejados com uma estratégia pH-estatística
tiveram uma tendência a melhores resultados de curto prazo do que aqueles manejados
com a estratégia α-estatística. O grupo pH-estatístico teve um tempo de recuperação
mais curto para a primeira atividade eletroencefalográfica e uma tendência a menos
convulsões manifestadas eletroencefalograficamente. Neste estudo, dentro do
subconjunto de bebês com transposição das grandes artérias, aqueles designados para o
pH-estatístico tenderam a ter um índice cardíaco mais alto apesar de uma menor
necessidade de agentes inotrópicos, menos acidose frequente (p = 0,02) e hipotensão (p
= 0,05), e uma duração mais curta da ventilação mecânica e da permanência na unidade
de terapia intensiva (p = 0,01). Embora o número de bebês estudados nesta investigação
tenha sido pequeno, suas descobertas desafiam o conceito de que o manejo α-estatístico
é mais fisiológico e protetor durante a hipotermia profunda em bebês do que o manejo
pH-estatístico. Um artigo de acompanhamento examinou o resultado do
desenvolvimento e neurológico nas mesmas crianças com 2 a 4 anos de idade. Eles não
encontraram nenhuma diferença nos resultados do neurodesenvolvimento com α-
estatístico versus pH-estatístico. Curiosamente, algumas diferenças foram observadas,
dependendo do tipo de lesão que estava sendo reparada. No subgrupo cianótico
(transposição e Tetralogia de Fallot), foi encontrado um escore ligeiramente mais alto
(mas não estatisticamente significativo) no Índice de Desenvolvimento Mental naqueles
pacientes manejados usando pH-estatístico. No entanto, no subgrupo de defeito do septo
ventricular, os pacientes em pH-estatístico obtiveram pontuações significativamente
piores. Esses dados devem ser interpretados com cautela porque o número de pacientes
em cada subgrupo foi pequeno e houve diferenças na idade no momento da cirurgia,
dependendo do diagnóstico.

Por que existe uma aparente diferença no resultado entre adultos e crianças em relação
ao manejo do pH? Pode estar relacionado a diferenças no mecanismo de lesão cerebral
no BCP. Em adultos, os êmbolos parecem desempenhar um papel proeminente no
resultado neurológico adverso. É, portanto, postulado que o FSC reduzido associado ao
manejo α-estatístico pode ser protetor ao limitar a dispersão de microêmbolos cerebrais.
Por outro lado, o mecanismo de lesão em crianças pode estar mais relacionado à
hipoperfusão ou à ativação de vias excitotóxicas. Se uma estratégia pH-estatística é
empregada, o aumento no FSC pode ser benéfico para garantir o resfriamento cerebral
completo e retardar o consumo de oxigênio, aumentando assim a tolerância do cérebro
para a PCHP.

Alterações na Função dos Órgãos

A hipotermia causa uma diminuição do fluxo sanguíneo para todos os órgãos do corpo.
No entanto, algumas áreas experimentam um declínio maior do que outras. O músculo
esquelético e as extremidades têm a maior redução no fluxo, seguidos pelos rins, leito
esplâncnico, coração e cérebro. Apesar dessa diminuição no fluxo, as diferenças no
conteúdo de oxigênio arteriovenoso são vistas para diminuir ou permanecer inalteradas,
o que implica que o suprimento de oxigênio é adequado para atender às necessidades
metabólicas.

Com o resfriamento, a frequência cardíaca diminui, mas a contratilidade permanece

estável ou pode realmente aumentar. As disritmias tornam-se mais frequentes à medida
que a temperatura diminui e podem incluir batimentos nodais, batimentos ventriculares
prematuros, bloqueio atrioventricular, fibrilação atrial e ventricular, e assistolia. O
mecanismo desse efeito disritmogênico é desconhecido, mas pode envolver distúrbios
eletrolíticos, resfriamento irregular e desequilíbrio do sistema nervoso autônomo. Como
o fluxo sanguíneo coronário é bem preservado durante a hipotermia, é improvável que a
hipóxia miocárdica desempenhe um papel na gênese dessas disritmias.

O sistema pulmonar é caracterizado por uma diminuição progressiva na ventilação à

medida que a temperatura é reduzida. O espaço morto fisiológico e anatômico aumenta
durante a dilatação dos brônquios pelo frio. A troca gasosa é amplamente inalterada.

Os rins mostram a maior diminuição proporcional no fluxo sanguíneo de todos os

órgãos. A hipotermia aumenta a resistência vascular renal, com diminuição do fluxo
sanguíneo e entrega de oxigênio no córtex externo e interno. O transporte tubular de
sódio, água e cloreto é diminuído, e a capacidade de concentração fica prejudicada. A
reabsorção tubular é diminuída. O fluxo urinário pode ser aumentado com a hipotermia,
mas esse efeito pode ser mascarado pela liberação de hormônio antidiurético (ADH)
induzida pelo estresse. A capacidade do rim hipotérmico de lidar com a glicose é
prejudicada, e a glicose frequentemente aparece na urina. A hemodiluição em
combinação com o BCP hipotérmico melhora o fluxo sanguíneo renal e protege a
integridade dos túbulos renais no pós-operatório.

Em geral, a lesão hepática significativa com BCP hipotérmico é rara. O fluxo sanguíneo
arterial hepático é reduzido em proporção à queda no débito cardíaco. O efeito mais
significativo da hipotermia é a diminuição na função metabólica e excretória do fígado.
Obviamente, as ações e requisitos de medicamentos serão modificados por essa
mudança na função hepática. Com o reaquecimento, a eficiência hepática volta ao
normal.

A hiperglicemia acentuada é frequentemente uma característica do BCP hipotérmico. A

produção endógena de insulina é diminuída, e a glicogenólise e a gliconeogênese podem
ser aumentadas devido a aumentos nas catecolaminas. Mesmo se a insulina exógena for
administrada, sua eficácia é reduzida durante a hipotermia, e a hiperglicemia pode se
desenvolver. É difícil separar os efeitos da hipotermia dos da hemodiluição e do BCP.

O conteúdo de água do tecido é aumentado durante o bypass hipotérmico,

principalmente como consequência da hemodiluição. O inchaço celular e o edema
ocorrem, o que pode estar relacionado a um acúmulo de sódio e cloreto dentro das
células secundário a uma diminuição nas taxas de reação da Na+-K+-ATPase de
membrana (adenosina trifosfatase sódio-potássio). A hipotermia diminui a depuração de
água livre e causa uma diminuição do potássio plasmático e um aumento da
osmolaridade.
A hipotermia causa mudanças acentuadas na circulação periférica. A resistência
vascular sistêmica e pulmonar tipicamente aumenta com o resfriamento abaixo de 26°C.
Esse aumento na resistência vascular está relacionado a aumentos na viscosidade do
sangue e catecolaminas, hemoconcentração, inchaço celular e talvez substâncias
vasoconstritoras ativas no pulmão. Além disso, shunts arteriovenosos aparecem em
baixas temperaturas e podem causar uma diminuição adicional na entrega de oxigênio
aos tecidos. O aumento na viscosidade do sangue ocorre devido a mudanças de fluidos,
com perda de volume plasmático por vazamento capilar e inchaço celular. O volume de
glóbulos vermelhos permanece inalterado, embora o hematócrito aumente. A agregação
de glóbulos vermelhos e a formação de rouleaux podem ocorrer, impedindo ainda mais
o fluxo sanguíneo. Essas mudanças podem ser um tanto atenuadas por anestesia
adequada, hemodiluição, heparinização e o uso de vasodilatadores. A hipotermia
também causa trombocitopenia por uma sequestração reversível de plaquetas na
circulação portal.

A resposta hormonal à hipotermia depende do nível de anestesia. Indivíduos não

anestesiados demonstram uma resposta simpática acentuada ao frio. Essa resposta pode
ser quase aniquilada se a anestesia profunda for usada. Após hipotermia profunda e
parada circulatória total, ocorre uma liberação massiva de catecolaminas , o que pode
contribuir para a perfusão cerebral prejudicada encontrada por Greeley et al.. A
liberação de corticosteroides é suprimida com hipotermia de longo prazo abaixo de
28°C, mas parece ser normal durante curtos períodos de hipotermia. A ativação do
complemento ocorre durante o BCP e está associada à ativação de neutrófilos.
Complicações respiratórias se correlacionam com o grau de ativação do complemento.
Hipotermia, hemodiluição e heparina reduzem a ativação do complemento e a
subsequente resposta de neutrófilos e podem proteger os pacientes de sequelas
prejudiciais. A bradicinina circulante aumenta durante a hipotermia e o BCP e pode
contribuir para a permeabilidade vascular alterada e a instabilidade circulatória.

USO CLÍNICO DA HIPOTERMIA

Atualmente, a hipotermia é usada mais comumente na cirurgia cardíaca, embora seu uso
também tenha sido descrito para grandes procedimentos vasculares, cirurgia
intracraniana e remoção de tumores hepáticos e renais. Na maioria dos procedimentos
cardíacos, a hipotermia sistêmica leve a moderada (>25°C) é usada por seus efeitos
protetores, conforme descrito anteriormente. Uma hipotermia miocárdica seletiva mais
profunda também é frequentemente usada durante o clampeamento aórtico para auxiliar
na preservação do miocárdio isquêmico. A hipotermia miocárdica é tipicamente obtida
de duas maneiras: por perfusão coronária com solução cardioplégica fria e por aplicação
tópica de uma pasta de gelo ou lavagem pericárdica fria. A temperatura ideal para a
proteção miocárdica é controversa; no entanto, a maioria dos estudos demonstrou
proteção superior em temperaturas tão baixas quanto 2°C a 4°C, desde que as
temperaturas de congelamento sejam evitadas, a alcalose esteja presente e o coração
seja prontamente parado durante o resfriamento.

Parada Circulatória Hipotérmica Profunda

A aplicação mais dramática que demonstra os efeitos protetores da hipotermia é na

PCHP. Temperaturas sistêmicas de 20°C a 22°C ou menos são usadas para permitir a
cessação da circulação por períodos de até 40 a 60 minutos, muitas vezes sem lesão
orgânica detectável. A PCHP pode ser usada em uma variedade de situações. Em
pacientes pediátricos de cirurgia cardíaca (particularmente aqueles com peso <8-10 kg),
o reparo de lesões cardíacas congênitas complexas é frequentemente facilitado pelo
campo cirúrgico sem sangue proporcionado pela parada circulatória. É frequentemente
usada em procedimentos que requerem oclusão de múltiplos vasos cerebrais,
particularmente o reparo de aneurismas do arco aórtico. Pode ser usada para melhorar a
exposição cirúrgica e a velocidade em procedimentos que podem levar a uma
hemorragia incontrolável.

Efeitos da PCHP no Sistema Nervoso Central

O cérebro é o órgão com maior risco de lesão; esse fato limita a duração do tempo de
parada segura. A atividade metabólica cerebral é diminuída com a temperatura, mas
nunca cessa completamente, mesmo em temperaturas próximas de 0°C. Como
mencionado anteriormente, o efeito protetor da hipotermia pode envolver mais do que
apenas uma redução na taxa metabólica cerebral. Um Q10 de 2,7 preveria apenas um
tempo de parada seguro de aproximadamente 15 minutos a 20°C. No entanto,
evidências clínicas e experimentais indicam que 30 a 45 minutos são tipicamente
tolerados, então parece haver um efeito protetor cerebral desproporcional a níveis
profundos de hipotermia. Outros fatores, como o pH extracelular, podem desempenhar
um papel. Swain et al. mostraram que a hipotermia aumenta significativamente o estado
de energia do tecido e o pH intracelular tanto no coração quanto no cérebro. Esse
aumento nos níveis de fosfato de alta energia pode explicar parcialmente os efeitos
benéficos da hipotermia na tolerância do órgão à isquemia. Por outro lado, pode ser que
o consumo cerebral de oxigênio diminua de maneira não linear e mais abruptamente
com a hipotermia profunda do que se pensava anteriormente. Michenfelder e Milde
mostraram uma mudança no Q10 de 2,23 entre 37°C e 27°C para 4,53 entre 27°C e
14°C. Eles postularam que essa queda acentuada no consumo de oxigênio em
temperaturas mais baixas poderia ser explicada por um efeito primário da hipotermia na
função neuronal integrada (como mostrado pela supressão do eletroencefalograma). A
taxa de resfriamento também parece ser importante na ocorrência de lesão cerebral.
Gradientes amplos entre a temperatura corporal e a do perfusato em cães se
correlacionaram com necrose de células cerebrais e morte.

O local ideal para o monitoramento da temperatura é controverso, mas deve-se lembrar

que os gradientes existem entre as diferentes regiões. O monitoramento de múltiplos
locais para garantir um resfriamento uniforme antes da parada circulatória é
aconselhável. Coselli et al. sugeriram o uso de monitoramento eletroencefalográfico
para determinar a profundidade ideal de resfriamento para uma parada circulatória
segura. Eles defenderam o uso do silêncio eletroencefalográfico como o ponto final
apropriado no resfriamento, mas descobriram que nenhuma temperatura corporal
periférica prevê consistentemente esse nível de hipotermia. Houve uma ampla variação
de temperatura entre os locais do corpo quando o silêncio eletroencefalográfico ocorreu.

Figura 8.12 Temperatura média [SEM (erro padrão da média)] da cânula arterial,
miocárdio, córtex cerebral, nasofaringe e reto durante 40 minutos de resfriamento e 90
minutos de reaquecimento sob bypass cardiopulmonar em seis porcos.
(De Stefaniszyn HJ, Novick RJ, Keith FM, et al. Is the brain adequately cooled during
deep hypothermic cardiopulmonary bypass? Curr Surg 1983;40:294-297, com
permissão.)

Uma consequência da isquemia e anóxia é o fenômeno de não-refluxo. A

microcirculação cerebral pode desligar multifocalmente, causando reperfusão
incompleta quando o fluxo é retomado. A etiologia desse problema não é
completamente compreendida, mas pode envolver aumento da viscosidade sanguínea,
contração do músculo liso vascular resultante do aumento do potássio extracelular e
shunting pré-capilar. Pode ocorrer com ou sem parada circulatória total e pode ser
prevenida pela hipotermia. O dano celular microscópico no cérebro ocorre em algum
grau após a hipotermia para 18°C, independentemente de ser empregada perfusão
pulsátil ou não pulsátil ou parada circulatória total.

A principal preocupação com o uso da PCHP são os potenciais efeitos prejudiciais no

órgão de maior risco, o cérebro. À medida que a mortalidade diminui com as melhorias
na técnica, a questão de um efeito no desenvolvimento intelectual posterior após o
bypass hipotérmico, com ou sem parada circulatória, torna-se crítica. Vários estudos
mostraram evidências de uma diminuição do quociente de inteligência e da capacidade
de desenvolvimento relacionados à duração da parada circulatória. No entanto, outros
estudos não foram capazes de demonstrar um efeito adverso na capacidade intelectual e
no desenvolvimento quando os tempos de parada circulatória foram inferiores a 60
minutos em temperaturas nasofaríngeas de aproximadamente 20°C. É difícil interpretar
muitos desses estudos por causa da dificuldade em definir um grupo de controle
apropriado. Blackwood et al. usaram cada criança como seu próprio controle e não
encontraram nenhuma diferença entre as pontuações pré-operatórias e pós-operatórias
com intervalos de parada tão longos quanto 74 minutos. Em 1993, Newburger et al.
relataram um risco neurológico maior para neonatos e bebês com PCHP do que com
bypass contínuo de baixo fluxo. Eles relataram reduções modestas, mas estatisticamente
significativas, no desempenho do desenvolvimento psicomotor naquelas crianças do
grupo de parada circulatória.

Muitos investigadores tentaram determinar uma duração segura para a PCHP, mas a
resposta ainda não é conhecida. A extensa experiência clínica sugere que períodos de
PCHP tão longos quanto 60 minutos são frequentemente bem tolerados. Embora os
pacientes variem amplamente, os dados de Newburger et al. sugerem efeitos adversos
mínimos nos resultados dos testes psicomotores com tempos de parada circulatória de
aproximadamente 35 minutos a 18°C.

Coreoatetose

A coreoatetose foi relatada em 1% a 20% das crianças submetidas a PCHP. A

coreoatetose geralmente aparece 2 a 6 dias após a cirurgia e geralmente diminui em
gravidade com o tempo. No entanto, em casos graves, os movimentos coreoatetoideos
ou a hipotonia generalizada podem persistir indefinidamente. A coreoatetose também
foi ocasionalmente observada após BCP contínuo, especialmente com hipotermia
profunda (10°C-12°C). Numerosas etiologias foram propostas, incluindo hiperglicemia ,
resfriamento irregular , o fenômeno de não-refluxo , alterações de neurotransmissores
dopaminérgicos e neurotoxicidade de aminoácidos excitatórios cerebrais.
Geralmente, pensa-se que esse distúrbio de movimento hipercinético é resultado de
lesão nos gânglios da base, embora a patologia muitas vezes não seja detectável por
tomografia computadorizada craniana convencional ou ressonância magnética (RM).
Parece que também pode haver um fenômeno relacionado à idade em relação à
coreoatetose. Dados do Boston Children's Hospital sugerem que o período mais
vulnerável começa aos 6 a 9 meses e termina após 5 a 6 anos. Uma forma leve e
transitória da condição foi observada em crianças mais jovens, enquanto crianças mais
velhas manifestaram um distúrbio grave e persistente com alta mortalidade. O grupo de
Boston também observou que a coreoatetose é mais provável de ocorrer em crianças
com vasos colaterais sistêmico-pulmonares significativos. Talvez o fenômeno de
"roubo" da circulação cerebral para as artérias pulmonares, particularmente no cenário
de resfriamento cerebral inadequado, contribua para esse problema.

Convulsões

As convulsões após o BCP em neonatos são muito mais comuns do que em adultos. As
convulsões foram relatadas clinicamente em até 20% dos neonatos após o BCP. As
convulsões detectadas por eletroencefalograma ocorrem ainda mais comumente do que
as convulsões observadas clinicamente, um fato a ser lembrado se for necessário induzir
paralisia farmacológica no período pós-operatório. As convulsões são geralmente
autolimitadas e podem ocorrer independentemente de a parada circulatória ter sido
usada. Algumas séries relataram ausência de sequelas adversas a longo prazo, enquanto
outras sugeriram um decréscimo no desempenho do desenvolvimento psicomotor, além
de anormalidades neurológicas e de RM. O prognóstico a longo prazo para essas
crianças ainda não está claro. A questão de saber se as convulsões por si só realmente
aumentam o dano, ou se são apenas um reflexo de uma patologia cerebral subjacente
grave, também surge. No momento, a resposta é desconhecida, e mais pesquisas são
necessárias nesta área importante. Além disso, o envolvimento relativo da hipotermia ou
de outros elementos do BCP não foi determinado.

Portanto, a questão de um tempo de parada circulatória seguro é complexa e não pode

ser respondida com certeza. A hipotermia pode atrasar, mas não prevenir, o
aparecimento de mudanças metabólicas e estruturais durante a isquemia que levam a um
prejuízo neurológico funcional. Um nomograma foi elaborado que, embora não seja
rigorosamente definido, fornece uma estimativa de melhores tempos de parada
circulatória segura em três temperaturas. Embora os pacientes variem amplamente, os
dados de Newburger et al. em crianças sugerem efeitos adversos mínimos nos testes
psicomotores com tempos de parada circulatória de aproximadamente 35 minutos a
18°C.

Figura 8.13 Nomograma de uma estimativa (não rigorosamente derivada) da

probabilidade de parada circulatória total segura (ausência de dano estrutural ou
funcional) de acordo com o tempo de parada em temperaturas nasofaríngeas de 37°C,
28°C e 18°C.

(De Kirklin JW. Cardiac surgery. New York: Churchill Livingstone, 1993:61-127, com
permissão.)

Resfriamento e Reaquecimento do Sistema Nervoso Central

Figura 8.14 Gráfico de dispersão da duração do resfriamento central (minutos de bypass
cardiopulmonar antes da PCHP) e índice de desenvolvimento para bebês com períodos
curtos de resfriamento central (<20 minutos).

A linha de regressão de melhor ajuste e seu intervalo de confiança de 95% (IC) são
mostrados. PCHP, parada circulatória hipotérmica profunda. (De Bellinger DC,
Wernovsky G, Rappaport LA, et al. Cognitive development of children following repair
of transposition of the great arteries using deep hypothermic circulatory arrest.
Pediatrics 1991;87:701-707, com permissão.)

Como a hipotermia profunda e as taxas de fluxo muito baixas ou a parada circulatória

provavelmente permanecerão ferramentas importantes na cirurgia cardíaca,
especialmente na cirurgia cardíaca congênita e no reparo do arco aórtico, muito esforço
foi gasto tentando entender melhor a fisiologia envolvida. Novas estratégias de
monitoramento e proteção estão sendo desenvolvidas em uma tentativa de melhorar o
resultado cerebral nesses pacientes. Uma das áreas de foco tem sido os métodos de
resfriamento antes da parada circulatória. A taxa de resfriamento e a eficiência do
resfriamento cerebral são fatores importantes na proteção neurológica. Há uma
variabilidade acentuada no resfriamento cerebral, e tempo suficiente deve ser alocado
antes da parada circulatória para garantir que haja um resfriamento cerebral uniforme.
Um grupo de investigadores relatou um resfriamento cerebral mais lento em um terço
dos neonatos estudados por monitoramento de saturação da veia jugular. Parece que
períodos de resfriamento de menos de 20 a 25 minutos antes do início da parada
circulatória com regulação α-estatística estão associados a um quociente de
desenvolvimento mais baixo em neonatos submetidos a PCHP.

O resfriamento cerebral uniforme também pode ser problemático em certos grupos de

pacientes. Pacientes cianóticos com colaterais aortopulmonares parecem estar em maior
risco de lesão neurológica durante o BCP. Ocorre uma redução nas medidas totais e
regionais do FSC nesses pacientes, resultando em um resfriamento cerebral menos
eficiente.

Como o cérebro está acidótico e uma tremenda dívida metabólica se desenvolve durante
a PCHP, o método de restabelecimento do BCP e o reaquecimento também podem ser
criticamente importantes. Alguns dados intrigantes sugerem que um período de
reperfusão fria com reaquecimento retardado após o restabelecimento do BCP pode ser
benéfico. Esses autores demonstraram uma melhoria na velocidade do FSC no grupo de
bebês em que o reaquecimento foi atrasado em 10 minutos. Também está se tornando
claro que o reaquecimento muito agressivo, com sobreaquecimento hipertérmico, é
deletério para o resultado neurológico e deve ser evitado.

Martin et al. e Mora et al. relataram um risco aumentado de lesão neurológica em

pacientes adultos ativamente aquecidos a 37°C, com um banho-maria de circuito de
BCP mantido a 39°C-40°C, em comparação com pacientes que foram autorizados a
atingir temperaturas sistêmicas de 33°C-36°C em uma abordagem mais tépida para o
bypass do coração quente. Eles sugeriram que muitos, se não todos, de seus pacientes
normotérmicos experimentaram hipertermia cerebral (>37°C), que pode potenciar
processos patológicos isquêmicos no cérebro durante operações cardíacas. Dados de
animais demonstraram que mesmo pequenos aumentos (2°C) na temperatura cerebral
podem exacerbar a lesão neuronal, com mudanças na permeabilidade da barreira
hematoencefálica, aumentos na liberação pós-isquêmica de glutamato e aumento da
mortalidade. Temperaturas elevadas devem ser vigorosamente evitadas, não apenas na
sala de cirurgia, mas também no período pós-operatório.

Vários autores demonstraram que a hipertermia ocorre comumente após cirurgia

cardíaca tanto em adultos quanto em crianças. A etiologia não está inteiramente
relacionada ao BCP por si só, porque pacientes que fazem cirurgia de revascularização
do miocárdio sem bomba também costumam ter febre no pós-operatório. Medidas
agressivas de resfriamento devem ser tomadas porque essa hipertermia demonstrou ser
deletéria para o resultado. Grocott et al. estudaram cuidadosamente 300 pacientes com
temperaturas pós-operatórias horárias e testes cognitivos antes e 6 semanas após a
cirurgia. Eles demonstraram que a temperatura máxima pós-operatória foi associada a
um maior grau de disfunção cognitiva pós-operatória. É claro que não está totalmente
claro se a febre observada no pós-operatório é por si só deletéria ou se é uma
manifestação de uma resposta inflamatória exagerada com subsequente lesão cerebral.
Mais pesquisas são necessárias nesta área.

Manejo da Temperatura Tépida

Nas últimas décadas, surgiu um interesse no BCP tépido. Existem muitas vantagens e
desvantagens teóricas e clinicamente evidentes detalhadas na literatura. Várias variáveis
entram em jogo ao comparar múltiplos estudos que utilizam BCP tépido com
normotermia, hipotermia leve e hipotermia moderada. Esta seção tentará resumir os
dados atuais disponíveis. Com base em um artigo de consenso recente sobre parada
circulatória hipotérmica , usaremos as seguintes definições para o grau de hipotermia ao
longo desta seção: hipotermia profunda como ≤14°C, hipotermia profunda como 14,1°C
a 20°C e hipotermia moderada como 20,1°C a 28°C. O artigo define hipotermia leve
como 28,1°C a 34°C. Para esta seção, consideraremos a hipotermia leve como 28,1°C a
32°C e consideraremos as temperaturas tépidas, frequentemente chamadas de
"derivação", para variar de 32,1°C a 36°C e a normotermia como ≥36°C.

No início dos anos 1990, Tonz et al. descreveram objeções teóricas ao BCP quente,
incluindo ativação celular e humoral estendida, diminuição da tolerância à isquemia
resultando em efeitos mais evidentes de perfusão não homogênea e trauma sanguíneo
adicional devido a taxas de fluxo mais altas. Embora lógico, as evidências disponíveis
não apoiam essas objeções. Uma revisão retrospectiva comparando normotermia com
hipotermia moderada encontrou uma diminuição significativa nos tempos de intubação
com o uso de BCP normotérmico. Além disso, não houve diferença na mortalidade, nas
taxas de infarto do miocárdio ou no uso de um balão intra-arterial para suporte. Este
estudo também encontrou tempos de BCP significativamente mais curtos com o uso de
cardioplegia quente durante o BCP normotérmico em comparação com o uso de
cardioplegia fria, embora doses significativamente mais altas de cardioplegia quente
fossem necessárias.

Dois pequenos estudos prospectivos compararam enxerto de bypass da artéria coronária

(CABG) e procedimentos de válvula utilizando hipotermia moderada à normotermia.
Lehot et al. se concentraram nas respostas hemodinâmicas e hormonais e encontraram
tempos de clampeamento e BCP significativamente mais longos no grupo de hipotermia
moderada em comparação com o grupo normotérmico. Hemodinamicamente, a
taquicardia foi significativamente maior no grupo de hipotermia moderada no término
do BCP. Embora a resistência vascular sistêmica (RVS) fosse significativamente menor
no grupo normotérmico, as concentrações de norepinefrina e epinefrina foram
significativamente maiores no grupo normotérmico. Nenhuma outra resposta humoral
significativa foi encontrada entre os dois grupos. Uma diminuição na RVS também foi
descrita por outros. Tonz et al. relataram mudanças hemodinâmicas significativas,
incluindo uma RVS mais baixa e um índice cardíaco mais alto imediatamente no pós-
operatório no grupo normotérmico, mas essa diferença desapareceu no primeiro dia pós-
operatório (DPO 1). Neste estudo, o volume total de sangue derramado através dos
tubos de tórax e pleurais foi significativamente maior no grupo de hipotermia moderada.
Este grupo também teve um aumento no uso de produtos sanguíneos, incluindo glóbulos
vermelhos embalados e plasma fresco congelado. Não houve diferença significativa em
mediadores inflamatórios ou exames de sangue neste estudo.

Engelman et al. randomizaram 291 pacientes com CABG para hipotermia moderada,
temperaturas tépidas ou normotermia. Eles não encontraram diferença significativa
entre os grupos em mortalidade, tempo de permanência ou arritmia atrial. Um
subconjunto de 53 pacientes foi analisado para avaliar o potencial de fibrinólise. Desses,
42 pacientes foram testados para atividade de pré-calicreína pré-operatória e pós-
operatória. Houve uma degradação significativamente aumentada da atividade da pré-
calicreína com o aumento das temperaturas. Os autores afirmam que isso indica um
aumento da fibrinólise na normotermia, embora em sua população de pacientes maior
de 291 pacientes, eles não tenham conseguido mostrar nenhuma diferença significativa
na perda de sangue ou no uso de produtos nas três temperaturas diferentes.

Em 2002, Guadino et al. avaliaram a ativação hemostática e inflamatória em 113

pacientes com CABG em hipotermia moderada ou normotermia. Não houve diferença
significativa em mortalidade, infarto do miocárdio, perda de sangue ou uso de produtos
nas diferentes temperaturas. Não houve diferença em marcadores hemostáticos,
incluindo concentração de fibrinogênio, inibidor do ativador de plasminogênio e
laboratórios de coagulação. Eles também não conseguiram detectar uma diferença
significativa em marcadores inflamatórios, incluindo proteína C-reativa, interleucina-6,
glóbulos brancos, neutrófilos e contagens de monócitos.

Nas últimas duas décadas, vários estudos prospectivos (alguns pequenos e outros bem-
sucedidos) compararam os resultados neurológicos após o BCP em várias temperaturas.
Um estudo prospectivo da Duke University randomizou 300 pacientes com CABG para
receber BCP tépido (35,5°C-36,5°C) ou hipotermia leve (28°C-30°C). Ambos os grupos
receberam cardioplegia fria intermitente, e a temperatura máxima do perfusato para
reaquecimento foi de 38°C. Todos os pacientes foram submetidos a testes cognitivos
pré-operatórios e os pacientes no grupo de hipotermia leve foram encontrados com um
desempenho ligeiramente diminuído no pré-operatório. Não houve diferença
significativa no declínio neurocognitivo em 6 semanas. Em outro estudo bem-sucedido,
Heyer et al. não encontraram diferença no desempenho neurológico ou
neuropsicométrico pós-operatório em 99 pacientes com CABG randomizados para
hipotermia moderada ou BCP tépido.

Guadino et al. relatam que, embora a prevalência de eventos neurológicos fosse

semelhante entre os grupos em seu estudo, houve uma associação com extensão
superior de dano cerebral e piores resultados neurológicos, especificamente naqueles
pacientes que sofreram acidentes vasculares cerebrais intraoperatórios com BCP
normotérmico. Além disso, houve 6,6% de mortalidade entre o grupo hipotérmico,
enquanto o grupo normotérmico teve uma taxa de mortalidade de 50% em pacientes
com acidente vascular cerebral intraoperatório. Mora et al. compararam 138 pacientes
submetidos a procedimentos de CABG com hipotermia moderada ou temperaturas
tépidas. Eles relatam uma taxa significativamente maior de novos déficits neurológicos
centrais perioperatórios no grupo tépido. Na linha de base, o grupo tépido tinha uma
idade média significativamente mais alta, o que pode ter contribuído. Além disso, a
administração de cardioplegia variou entre os dois grupos. O grupo tépido recebeu
cardioplegia retrógrada contínua após a administração anterógrada inicial, enquanto o
grupo de hipotermia moderada recebeu apenas cardioplegia anterógrada e retrógrada
intermitente.

Em outro artigo, usando a mesma população de estudo de 300 pacientes com CABG
descrita acima, o grupo Duke comparou a mudança na depuração de creatinina em
diferentes temperaturas. Não houve diferenças significativas na mudança na depuração
de creatinina ou nos valores mais baixos de depuração de creatinina entre os grupos de
hipotermia leve e BCP tépido. Com base nesses resultados, os autores afirmam que não
há efeito protetor renal da hipotermia, embora apenas a depuração de creatinina tenha
sido avaliada neste estudo. Embora a temperatura máxima do perfusato fosse bem
controlada neste estudo, houve uma duração média mais alta de temperaturas >37°C e
uma temperatura média aumentada >37°C no grupo de hipotermia leve. Há fortes
evidências para apoiar os efeitos prejudiciais da hipertermia na função neurocognitiva e
lesão renal aguda (LRA) , e a necessidade de reaquecer os pacientes aumenta o risco de
hipertermia. Um artigo de 2002 também sugere melhores resultados com reaquecimento
mais lento. Este estudo comparou 165 pacientes reaquecidos com temperaturas de
perfusato separadas por 2°C ou 6°C e encontrou temperaturas de pico mais baixas e
uma melhoria significativamente maior no desempenho cognitivo com reaquecimento
mais lento. Um estudo observacional recente de 1.393 pacientes descreve um aumento
de 34% na probabilidade de LRA para cada 10 minutos de temperaturas de saída arterial
>37°C.

A análise dos vários estudos deixa-nos um pouco confusos quanto ao manejo de

temperatura apropriado para o BCP. Parece que a maioria dos estudos não mostra uma
diferença acentuada nos resultados ao usar hipotermia para casos de rotina. Pode haver
algum benefício limitado em temperaturas mais baixas para minimizar a lesão
neurológica no cenário de um acidente vascular cerebral. Portanto, pode haver algum
benefício em deixar a temperatura do paciente "derivar", usando uma abordagem tépida.
No entanto, a atenção rigorosa a uma estratégia de reaquecimento cuidadosa é
imperativa para evitar hipertermia inadvertida, que provavelmente exacerba qualquer
lesão.

A cirurgia cardíaca avançou notavelmente nos últimos 50 anos. A hipotermia contribuiu

substancialmente para melhorar o resultado do paciente em casos selecionados. Os
esforços atuais devem ser direcionados para a definição de métodos para maximizar a
proteção cerebral e refinar técnicas críticas.

Pontos Chave

● O equilíbrio ácido-base é significativamente afetado pela hipotermia.

● Fluidos biológicos são soluções aquosas de ácidos e bases fracos.
● A hipotermia diminui a tendência de ácidos e bases fracos se dissociarem em
solução.
● O pH do sangue é mantido em valores moderadamente alcalinos (0,4 unidades
de pH) em relação à água.
● A manutenção de uma razão constante de [OH-] para [H+] (16:1) à medida que
a temperatura diminui permite a função ideal de muitas enzimas respiratórias.
● A regulação α-estatística preserva a razão de [OH-] para [H+] com a mudança de
temperatura e produz uma mudança alcalina com o resfriamento.
● A regulação pH-estatística mantém um [H+] constante e absoluto,
independentemente da temperatura, e requer H+ adicionado, geralmente como
CO2, com o resfriamento.
● A solubilidade dos gases em fluidos biológicos aumenta com a hipotermia.
● Em um conteúdo constante de CO2, o PaCO2 diminui à medida que a
temperatura cai.
● A estratégia ácido-base apropriada durante o BCP hipotérmico pode ser
diferente entre crianças e adultos com base em diferentes mecanismos de lesão
cerebral.
● O pH-estatístico pode ser benéfico em bebês para aumentar o FSC e permitir um
resfriamento mais eficiente.
● O α-estatístico parece vantajoso em adultos ao limitar a carga microembólica
para o cérebro.
● A proteção do cérebro durante a hipotermia profunda (temperatura <20°C) pode
ser melhor realizada com uma estratégia ácido-base mista:
● pH-estatístico durante a fase inicial de resfriamento.
● α-estatístico durante a reperfusão, reaquecimento e término do BCP.
● A hipotermia causa uma diminuição no fluxo sanguíneo para todos os leitos
vasculares em proporção às demandas metabólicas reduzidas.
● Os efeitos mais profundos ocorrem nos músculos esqueléticos e nas
extremidades, seguidos pelos rins, leito esplâncnico, coração e cérebro.
● A hipertermia ocorre comumente após cirurgia cardíaca e deve ser tratada
agressivamente porque pode piorar o dano isquêmico
CHAPTER 9

SURGICAL MYOCARDIAL PROTECTION

Jakob Vinten-Johansen and Neil J. Thomas

INTRODUCTION

The science of myocardial protection had its beginnings in the 1950s, and has enjoyed a robust
and often controversial development since. Great strides were made in the 1980s and 1990s
which shaped the cornerstones of the cardioplegia strategy that we largely use today
(crystalloid or blood, warm, tepid or cold, continuous or intermittent, antegrade or
retrograde). Mortality has substantially decreased over the past decades as a result of applying
cornerstone as well as novel cardioprotective strategies and new technologies.

However, the patient of the 1980s and 1990s during which cardioplegia was developed has
morphed considerably in the 2000s. Active lifestyles have changed to more sedentary lifestyles
beginning in the early school years with less emphasis placed on physical activity, and more
distractions from active lifestyles imposed by the digital-age gadgets. This, coupled with the
availability of fast food with high caloric, high fat, and high sodium content, has conspired to
radically change our body habitus and biology; indeed, diabetes, obesity,
hypercholesterolemia, and the metabolic syndrome are more prevalent today. These more
prevalent comorbidities have impacted the efficacy of modern surgical therapies and the
outcomes (1) of myocardial protection strategies. Experimental animal models in which
effective cardioplegia strategies were developed decades ago, and observed to be effective in
the healthier patients with normal myocardial tissue substrate preoperatively in which they
were initially tested clinically seems to apply in fewer and fewer patients today. The “one
formula-fits-all” strategies of the past must be rethought in the current era. With patients
living longer as well, advancing age and the more senescent myocardium can display altered
responsiveness to some therapeutics. So we may ask, is myocardial protection developed in
the 1980s and 1990s being used as effectively on the patient of the 2010s? And, furthermore,
the yardstick by which our results are measured is changing in the world of health care today,
where a 30-day surgical mortality rate may no longer be the gold standard. Our results must
now be viewed in terms of longer-term, event-free survival and with measure of long-term
myocardial performance after an index event or surgery.

The concepts of cardioprotective strategies have changed significantly as our understanding of

the pathophysiologic determinants of ischemia, and particularly the complex mechanisms
involved in reperfusion, have expanded. Early empiric, largely physiologic observations have
given way to complex biochemical and molecular interactions, which in some cases have
reinforced physiologic intuition and in other cases have redirected our thinking, but has
nonetheless provided a scientific underpinning to a more molecular approach in developing
target-specific strategies of myocardial protection. This chapter will discuss the pathobiology
of ischemia-reperfusion injury, provide the basis by which the severity and duration of evolving
ischemic conditions set the stage for reperfusion and oxygen-related injury and its impact on
affected tissues as a basis for understanding cardioprotective strategies to attenuate injury. It
will also discuss new, emerging concepts of myocardial protection that have developed out of
a better understanding of the pathobiology of ischemia-reperfusion injury, and a growing
appreciation that the preoperative condition of the tissue may be equally, if not more
important, than any intraoperative strategy employed in previously normal tissue. It will also
discuss the all-elusive “golden ring” of arresting the heart in a nondepolarized state.

The goal of this chapter is to provide the reader with scientific and practical knowledge of
myocardial protection that can be applied to design appropriate strategies to meet the need of
specific patients, their unique demographics and clinical scenarios, and to provide flexibility in
the approach based on both preoperative condition and intraoperative events. It will also
provide some guided reading to enhance the background of relevant topics

HISTORICAL PERSPECTIVES ON MYOCARDIAL PROTECTION AND REPERFUSION INJURY

The need to protect the myocardium from intraoperative and postoperative damage was
realized before the development of cardiopulmonary bypass (CPB). Temporary inflow
occlusion techniques to limit blood flow to the heart permitted the repair of simple
intracardiac defects, and in many cases the occlusion time could be increased by adding
moderate systemic hypothermia (30°C) (2). After the development of CPB by Gibbon and
colleagues in 1954 (3,4), profound systemic hypothermia could be used safely to provide a
hemodynamic safety net, and to provide protection of the heart and other body organs by
permitting longer periods of ischemic cardiac arrest for the repair of more complicated lesions.

The development of extracorporeal circulation techniques allowed cardiac procedures to

proceed without threat of hemodynamic collapse. However, performing complicated
procedures on the heart was made difficult by its continuous motion and blood in the field.
Ischemia was known to achieve cardiac arrest (ischemic or anoxic arrest) through the
depletion of high-energy phosphate stores on which cardiac contraction depends. However,
myocardial ischemia led to necrosis and contractile dysfunction upon reperfusion, the worst of
which was an irreversible “stone” heart (5,6). The conundrum of surgeon convenience at the
expense of acceptable outcomes provided a driving force for innovation.

Once the severe complications of ischemic arrest were recognized, chemical methods were
used to arrest the heart. Based on reports from Ringer in 1883 that high concentrations of
potassium would arrest the heart (7), Melrose and associates in 1955 (8) and 1957 (9) used
hyperkalemia to induce asystole during CPB. The combination of immediate arrest and
hypothermia would ostensibly reduce energy (i.e., oxygen) demands of the heart, and thereby
avoid injury imposed by hypothermia alone. Melrose made a hyperkalemic solution by mixing
potassium citrate solution (77 mM) with 18 mL of blood, which was then infused by hand-held
syringe into the aortic root proximal to the aortic cross-clamp. This hyperkalemic arresting
solution was later called “cardioplegia” solution by Lam et al. (10) and Sergeant et al. (11).
Other contemporaries of Melrose used alternative arresting agents, such as acetylcholine
alone (10,11) or in combination with other drugs (12).

Although asystole was indeed achieved with the Melrose potassium citrate solution, the
concept of elective cardiac arrest was abandoned in the United States for some 20 years
following reports of high morbidity and pathologic complications in chemically arrested hearts
(13). Chemical cardioplegia was supplanted by alternative techniques of arresting the heart or
limiting blood flow in the surgical field, including ventricular fibrillation, intermittent aortic
cross-clamping, and profound systemic (14) or topical hypothermia (15). Laboratory studies
demonstrated that topical hypothermia protected inadequately against intraoperative injury,
which resulted in postoperative myocardial subendocardial necrosis, and postischemic
metabolic and functional depression (16,17,18). In the late 1960s, reports described scattered
myocardial or subendocardial necrosis in patients who had died after otherwise technically
successful cardiac surgical operations, suggesting that the current techniques of myocardial
protection were inadequate (19,20). This set the stage for a later re-emergence of chemical
cardioplegia in the United States.

During the surgeon-imposed moratorium on the use of the Melrose solution, cardioplegia
solutions continued development and clinical use in Europe. Bretschneider et al. (16) and
Kirsch et al. (17) used crystalloid formulations which mimicked intracellular ionic
concentrations (intracellular crystalloid solutions). The Bretschneider HTK (histidine,
tryptophan, α-ketoglutarate)-ketoglutarate) solution used low calcium, low sodium, and
procaine with histidine buffers to achieve a nondepolarized arrest, rather than hyperkalemia-
induced depolarized arrest. In the 1970s, Hearse and associates (18,21) at St. Thomas’ Hospital
in London developed an extracellular hyperkalemic solution based on a rigorous series of
experiments that established a scientific understanding of the determinants of injury incurred
during global ischemia. Braimbridge introduced St. Thomas’ cardioplegia solution into clinical
use in 1975 (22). St. Thomas’ solution was later modified based on further experimental
studies to the St. Thomas’ solution No. 2 (23), an intracellular ionic concentration,
normocalcemic, and hyperkalemic (16 mmol/L) solution; it is currently marketed as Plegisol in
the United States. The Bretschneider HTK solution is currently marketed as Custodiol HTK
solution.

Chemical cardioplegia was revived in the United States by Gay and Ebert (24) and Tyers et al.
(25,26), who found that the constituents in the Melrose formulation were inappropriate rather
than the concept of hyperkalemic chemical cardioplegia itself being erroneous. These reports
popularized the use of potassium-based cardioplegia to achieve electromechanical arrest. In
the late 1970s and early 1980s, Follette and colleagues from Dr. Buckberg’s laboratory (see
reference 224 below) introduced the concept of cold hyperkalemic blood cardioplegia. The
physiologic attributes of blood, including its superior buffering capacity, endogenous oxygen
radical scavengers, detoxifying substances, and superior oxygen-transport capacity, has
established this concept as a popular strategy that has facilitated surgery and vastly improved
patient outcomes. Today, a variety of both crystalloid and blood cardioplegia solutions are
used clinically worldwide to achieve elective cardiac arrest and a bloodless field.

The cornerstones of myocardial protection that had been laid by the end of the early 1980s
are: (1) rapid chemical arrest to conserve high-energy phosphate stores and reduce ischemia,
(2) hypothermia to reduce the metabolic rate, (3) adjunctive agents to reduce the effects of
ongoing ischemia (anti-ischemic agents), and (4) continuous or intermittent delivery to restore
or maintain tissue oxygenation. Further developments in cardioplegia-related myocardial
protection strategies focused on (1) optimizing delivery (retrograde, gentle cardioplegia
delivery pressures), (2) overcoming the adverse effects of hypothermia, (3) identifying
metabolic enhancements (ATP generation and pH management) and buffers, (4) managing
calcium (Ca2+) accumulation, and (5) understanding the pathophysiology of ischemia and
reperfusion, which would drive further developments of strategies to avoid these mechanisms
of both ischemic and postischemic patterns of injury. There has also been a persistent search
for agents that induce polarized arrest which overcome the disadvantages of hyperkalemia
(reviewed in Dobson et al. (27) and Maruyama et al. (28)). With the recent rapid developments
in molecular and cellular biology, experimental focus has shifted away from the organ and
physiologic aspects toward cellular and molecular areas to improve the understanding of the
effects of ischemia-reperfusion injury, inflammation, and cell death processes. Understanding
the mechanisms involved in the transition to cell death would direct developments in Ca2+
handling and preventing Ca2+ dyshomeostasis, the role of oxidant injury, and a newly
recognized role of mitochondria in determining the transition from reversible to irreversible
injury, and whether the cell pursues a necrotic pathway or apoptotic pathway to cell death.
Finally, the concept of conditioning, which recruits complex endogenous biologic mechanisms
of self-protection against stressors such as ischemia-reperfusion injury, oxidants and
apoptosis, has been incorporated into pretreatment (preconditioning), intraoperative
(preconditioning), and reperfusion (postconditioning) phases of cardiac surgery. Although
human beings have been evolving for hundreds of thousands of years, it has only been in the
past 60 years or less that the human heart has been subjected to the kind of induced ischemia
and surgical reperfusion discussed in these pages.

PATHOPHYSIOLOGY OF SURGICAL ISCHEMIA-REPERFUSION INJURY

Because many factors affect postoperative outcomes of hearts with various and complex
preoperative pathophysiologic features, developing strategies that adequately protect all
hearts during cardiac surgery is often problematic. The heart under consideration may present
a complex picture of preoperative disease, variable age, gender, metabolic and cellular
dysfunction, global or regional ischemia (both acute and chronic), each of which contribute to
differing vulnerabilities to ischemia and reperfusion injuries. However, adequate myocardial
protection must be based on a sound scientific basis. Obviously, a single cardioprotective
strategy may be inadequate to suit all surgeon preferences and target all patient pathologies.
Hence, the surgical team must often mold the composition of the cardioplegic solution or its
modality of delivery to meet the requirements of the patient’s pathologic profile as presented.
The measures taken to protect the heart during elective cardiac arrest should represent a
balance struck between the requirements of the heart during aortic cross-clamp or
reperfusion, with appropriate adjustments made for special considerations (hypertrophy,
diffuse coronary disease), and avoid distraction from the surgical procedure. Ischemia and
reperfusion injury (29) and the choice of cardioplegia technique (30) represent primary
contributors to patient outcomes.

Ischemic Injury

When Does Ischemia and Reperfusion Injury Occur during Cardiac Surgery?
Figure 9.1. Events occurring before the institution of cardiopulmonary bypass, during delivery
of cardioplegia and at reperfusion that predispose the myocardium to ischemia and
reperfusion injuries. Arrhythmias, ventricular fibrillation (VF) or severe hypotension can cause
antecedent global ischemia. During the delivery of cardioplegia, the composition of the
solution can cause edema, further ischemia if oxygen and nutrients are inadequate to meet
ongoing albeit lower demands, and Ca2+ concentration is either too high or too low. Coronary
obstructions, air or low cardioplegia infusion pressure can lead to maldistribution of solution.
Too high infusion pressure can lead to microvascular injury and edema. Potassium itself has
negative effects secondary to placing the cell in a depolarized state. Ischemia may be
encountered if air emboli, graft kinks, tight anastomoses, or systemic hypotension cause
inadequate distribution of blood flow. Finally, reperfusion injury can cause reversible (stunning,
arrhythmias, edema) or irreversible (necrosis, apoptosis) injury.

It is critical for the surgical team to understand that the presence of acute, ongoing and
evolving or progressing ischemia must impact the protection strategy. Ischemia-reperfusion
injury can occur in the surgical setting at three major time points during the surgical procedure
(summarized in Fig. 9.1).

(a) Before CPB has been instituted or cardioplegia solution has been delivered:
“Unprotected” clinical ischemia before CPB usually results from the presenting acute coronary
syndrome and can be exacerbated by hypotension, arrhythmias (e.g., atrial or ventricular
fibrillation), and other hemodynamic changes (e.g., cardiogenic shock, coronary artery spasm).
In most cases, acute ischemia is active in both the distribution of the principally involved artery
(cardiologists call this the “infarct artery”) but also distributions in which preexistent collateral
flow may be compromised. Patients referred for bypass surgery in these circumstances almost
always have multi-vessel disease patterns making the myocardial substrate dangerous from
the standpoint of the severity and duration of the syndrome, which can be extremely difficult if
not impossible to determine. Reperfusion and reoxygenation injury occur when these
conditions change abruptly before effective measures are implemented. Recent attention has
been paid to avoiding over-oxygenating such patients systemically, and this will be discussed in
more detail in later sections.

(b) At the initiation of cardiopulmonary bypass: At the moment bypass is initiated, blood is
immediately diverted from the right atrial drainage system into a venous reservoir where it is
then pumped through the oxygenator. This configuration does more to the heart than merely
oxygenate venous blood and send it back to the ascending aorta. As the heart is emptied into
the venous side of the pump, the oxygen demand of the metabolizing myocardium decreases
by 50%, but the high oxygen content of blood extracorporeally oxygenated makes the tissue
vulnerable to reoxygenation damage if no steps are taken to mitigate this injury. This might be
viewed as an abrupt interruption of antecedent ischemic conditions. While this is one of the
principal goals of bypass (to give the heart a “rest”), it must be recognized that the heart is also
vulnerable to a pattern of injury that has been, at best, underappreciated in the past. Ihnken et
al. (31) suggested as early as 1998 that initiation of bypass is also, perhaps paradoxically, a
dangerous moment for the heart. In this work, he observed clear, reproducible biochemical
evidence of injury in samples of blood drawn from the coronary sinus. Since Buckberg (32)
published his work on the reoxygenation injury in 1995, it has been observed that abrupt over-
exposure of the ischemic heart to oxygen increases injury in a pO2-dependent pattern of injury
that can be mitigated by altering the oxygen tension of the perfusate.

(c) During the “protected” period of arrest and delivery of cardioplegia: Ischemia or
reperfusion injury can occur during the cardioplegia phase. Ischemia can develop or be
worsened under several circumstances (1) the unintentional maldistribution of cardioplegia
solution distal to stenotic or totally occluded coronary arteries, (2) between intermittent
infusions of cardioplegia solution, (3) during interruption of continuous cardioplegia strategies,
(4) inadequate delivery of retrograde cardioplegia to areas of the right heart whose venous
blood drains directly into the Thebesian system, or areas that are simply under-filled due to
malposition of a retrograde delivery device. Though the cardioplegia phase is considered
“protected time,” the potential for worsening of, or the development of, ischemic conditions
followed by reperfusion injury can occur with the initial administration of cardioplegia and
with each subsequent delivery of cardioplegia solution [(re)perfusion] for a variety of reasons:
(1) excessive pressure (microvascular injury), (2) low onconicity favoring extravasation of fluids
into the extracellular space, (3) formation of oxygen radicals from over-oxygenation and
oxidative stress, and (4) other mechanisms of reperfusion injury as discussed below. While
delivery of cardioplegia should provide protection, this phase also offers opportunity for
additional injury to occur unintentionally. Any surgeon who has been faced with postoperative
low-cardiac output and who suspects inadequate protection knows this to be true.

d) After release of the cross-clamp and attempts to reanimate the heart: Additional ischemia
may be encountered at the release of the cross-clamp when (1) coronary blood flow is
impaired through kinked grafts or (2) tight anastomoses (3) air emboli are present in the
coronary arteries, (4) ventricular fibrillation occurs at low perfusion pressures, or (5) poor
ventricular performance or dysrhythmias cause hypotension once the heart is converted and
off bypass. Clearly, when the clamp is removed reperfusion injury can occur as blood flow is
restored and oxygen exposure is reestablished after any of the above events. In terms of
surgically induced ischemia, this is intuitively the major time when one thinks of reperfusion
injury occurring.

Understanding the pathophysiologic mechanisms of myocardial injury occurring at these

intervals and applying the principles of myocardial protection will help avoid injury induced by
the very techniques intended to preserve the myocardium. Unlike the cardiologist, the surgical
team has the opportunity to intervene directly to control many facets of ischemia and
reperfusion injury, and to limit at some point the mechanisms causing injury, thereby more
favorably directing the outcome of the postischemic heart. The pathophysiology of ischemia
and reperfusion is described in the following sections. Interventions targeting these various
pathologic processes are discussed under Myocardial Protection Therapy.

Determinants of Ischemic Injury

Ischemia encountered either before arrest or during cardioplegic arrest, is a major cause of
postcardioplegia injury. Ischemia sets the stage for reperfusion injury; without ischemia you
cannot have reperfusion injury. Ischemia is defined as inadequate energy supply to meet the
current demands, that is, an energy supply/demand ratio of less than 1; normally, the O2
supply/demand ratio is >1. Since the adenosine triphosphate (ATP) necessary to support the
energetic needs of the heart is provided predominantly by oxidative phosphorylation, with
scant generation of high-energy phosphates through anaerobic metabolism, the heart is nearly
an obligate aerobic tissue. Therefore, the energy consumption of the heart can be
approximated by the oxygen consumption. The limited availability of anaerobic energy-
generating pathways mandates that O2 be in constant supply. A very brief discussion of the
determinants of O2 supply and demand follows.

Oxygen supply: Ischemia may be caused by either a decrease in O2 supply relative to demands
(supply ischemia) or an increase in O2 demands relative to supply (demand ischemia). O2
supply is determined by the O2 extraction (arterial-coronary venous O2 difference) and
coronary blood flow. Since O2 extraction by the heart is normally ~75%, and its physiologic
limit is approximately 95%, only limited amounts of oxygen may be obtained through
additional extraction. The majority of O2 is therefore supplied by adjustments in coronary
blood flow, which under normal conditions may increase 4- to 5-fold to meet demands. This
O2 is largely present as hemoglobin-bound, with only a minor fraction (~1.5%) dissolved in
plasma. This limited dissolved O2 will become important when we discuss crystalloid
cardioplegia solutions in which oxygen availability is limited to that dissolved in solution.

Oxygen demand: Overall myocardial oxygen demand of the normal working heart is
determined by the work of the entire heart. The majority of this overall O2 consumption is
attributed to utilization by the left ventricle. In a nutshell, the determinants of left ventricular
O2 demand include pressure–volume work or wall stress, heart rate, temperature, inotropic
state, basal metabolism, and ionic homeostatic mechanisms (ATP-dependent pumps) required
to re-equilibrate ionic balance after electromechanical activity. In the immediate postoperative
heart, energy demands may be related to oxidative energy diverted to myocyte repair.

The energy related to maintain ionic equilibrium (~5% of total demand) is an ongoing demand
even when the heart is not contracting since ATP-dependent pumps are constantly adjusting
calcium and sodium influx, as well as potassium efflux. This energy demand of ionic
equilibration is important in determining ongoing oxygen demands in the arrested heart, as
discussed below. How oxygen demands are changed by hypothermia is discussed later.

Severity and Duration of Ischemia in Cardiac Surgery

The potential for myocardial injury is related, in part, to the duration of ischemia. The time to
onset of irreversible ischemic injury will depend on many factors, including the severity of
ischemia, myocardial temperature, ambient energy demands, and collateral blood flow.
However, irreversible injury can become apparent after as little as 30 to 45 minutes of
coronary occlusion in the working myocardium. However, shorter durations of global ischemia
can result in mild to severe systolic and diastolic dysfunction without irreversible tissue
necrosis (left ventricular “stunning” (33)). Therefore, the duration of ischemia, both preceding
CPB and intraoperatively after application of the cross-clamp, has often been used as a
predictor of postoperative myocardial injury in experimental models, and is one of the
underlying considerations in the correlation between total ischemic time (CPB, cardioplegia)
and clinical outcomes. Because the surgical team often does not have control over the
duration of warm ischemia (particularly antecedent ischemia caused by coronary occlusive
disease), other aspects of elective ischemic time are more often targeted for modification to
reduce the severity of ischemia and hence postischemic injury. Therefore, strategies such as
(1) initiating rapid asystole, (2) venting the left ventricle to reduce consequences of chamber
distension resulting in increased wall stress, and (3) imposing cardiac hypothermia all
specifically target the reduction of myocardial energy/oxygen demands, thereby increasing the
limits of “safe” ischemic time during aortic cross-clamping. With these cardioprotective
strategies in place, the duration of aortic clamping that can be safely imposed can be increased
from as little as 15 to 45 minutes to several hours using hypothermic cardioplegia (Fig. 9.2).
Reducing to almost zero the determinants of myocardial work and oxygen demand is the sine
qua non of myocardial protection. In the section on future directions, eliminating ischemic
conditions without additional oxygen-related injury and changing the milieu more gradually
will be discussed.

Figure 9.2. Left ventricular performance measured by in situ Starling curves (A) or end-systolic
pressure–volume relations (conductance catheter) (B) after 45 minutes of normothermic
unprotected global ischemia and reperfusion. C: Left ventricular stroke work index in patients
before (Baseline) or after (pCPB) 15 minutes, or 4 and 24 hours postcardiopulmonary bypass
(CPB). The dashed line between 4 and 24 hours after discontinuation of CPB represents further
decreases in cardiac performance secondary to onset of irreversible injury.(A, adapted from
Rosenkranz ER, Buckberg GD. Myocardial protection during surgical coronary reperfusion. J Am
Coll Cardiol 1983;1:1235–1246, with permission.)

The determinants of surgical ischemic injury can be summarized as:

● The duration and severity of antecedent” and “unprotected” ischemia: Although it has
been recognized for decades that ischemia time (duration) impacts tissue injury and
postischemic salvage and function, an increasing appreciation for the caveat that the
 manner in which the tissue is reperfused significantly alters the outcome. It is now
increasingly understood that reperfusion injury can occur by merely eliminating
ischemic conditions abruptly or in the presence of an inappropriate oxygen gradient. In
the case of the imposition of aortic cross-clamping, there is total elimination of native
flow with the exception of that supplied by noncoronary collateral flow. Global
contractile function may be impaired before necrosis or apoptosis are evident
(“stunning”). It is the reduction of myocardial oxygen demand that protects the heart
in this circumstance.
● The elimination of electromechanical activity during cross-clamping is paramount
because its mechanical activity increases O2 demands even in the presence of
hypothermia (discussed further in the section on Hypothermia).
● Myocardial temperature: Temperature influences metabolism by the “Q10 effect,” in
which the tissue metabolic rate decreases by half for each 10°C decrease in
temperature; conversely, myocardial metabolism doubles when the heart rewarms by
10°C. Hypothermia can buy biologic time during ischemia.
● The nutritional and stress status of the heart: For example, catecholamines increase
the O2 demands by increasing contractility and exerting an O2 wasting effect.
● Presentation of comorbidities such as hyperlipidemia which increase the vulnerability
to ischemia-reperfusion injury for any given duration of ischemia. This effect is
independent of energy demands, but is influenced more by mechanisms exerted at
reperfusion, that is, the capacity to generate O2 radicals, a pro-inflammatory state of
the endothelium and inflammatory cells, and calcium handling by the tissue.

Determinants of Reperfusion Injury

Although reperfusion of the myocardium is the ultimate goal of both surgical and nonsurgical
revascularization, abrupt reperfusion carries with it the potential for extending postischemic
injury to cardiomyocytes, coronary vascular endothelium, and the microvasculature. Indeed,
reperfusion injury associated with rapid changes in circulatory conditions may be responsible
for ~50% (34) or more of the ultimate infarct size, and is an important contributor to
postsurgical mortality and morbidity as well (29). We can define surgical reperfusion injury as
an incremental increase in the pathology that is observed after ischemic injury alone.
Reperfusion injury extends or accelerates damage from that observed during ischemia alone,
and occurs after the onset of reperfusion (i.e., cardioplegia) or reperfusion (i.e., cross-clamp
removal). Viewed another way, reperfusion injury can occur at the instant of the elimination of
ischemic conditions with reestablishment of flow. Hence, reperfusion injury can be viewed as
representing an increment of injury that had begun during ischemia but progresses or
manifests during the reperfusion phase. Other investigators have described de novo injury
starting at reflow or, in a newer construct, at the time of the elimination of ischemia (even if
that is accomplished by a reduction in oxygen demand). Reperfusion injury can be categorized
as either reversible or irreversible. (1) Reversible injury includes temporary contractile
dysfunction (“stunning”) in the absence of morphologic injury or necrosis, or (2) irreversible,
which includes necrosis and apoptosis. Biomarker enzymes such as creatine kinase (CK) or
cardiac troponin T (cTnT) or I (cTnI) released by disrupted cells have been observed to be
elevated during reperfusion rather than during ischemia. In the past, debate has often been
spirited regarding whether reperfusion injury even exists (35,36,37,38) and whether
reperfusion injury kills cells that are salvageable at the end of ischemia. However, a consensus
has developed in both the surgical, cardiology, and cardiovascular research communities that
reperfusion is indeed an important component of postischemic cardiac injury in animal models
(39) and in humans (34,40) and that alterations in reperfusion strategies can mitigate injury.

Among the more important mechanisms in the pathophysiology of myocardial ischemia-

reperfusion injury are (1) the duration and severity of antecedent ischemia, (2) oxygen radicals
generated not only by neutrophils but also by myocytes and vascular endothelium, (3) sodium
(Na+) and Ca2+ influx and loss of normal intracellular Ca2+ homeostasis, (4) neutrophils and
neutrophil-derived products that constitute a generalized inflammatory response to
reperfusion, and (5) mitochondrial dysfunction and opening of the mitochondrial permeability
transition pore (mPTP). Interestingly, the presence of hyperoxic or even relatively hyperoxic
conditions can ramp up or trigger all of these mechanisms. As such, these mechanisms do not
operate independently, nor are they mutually exclusive. Importantly, reperfusion injury is a
malleable process that can be attenuated by utilizing a number of strategies targeting these
various perpetrators of injury. In the operating room during cardiac surgery, and more and
more frequently in the modern era in the interventional lab, various forms of circulatory
support techniques and cardioprotective drugs are employed. The realization that circulatory
support can manipulate reperfusion therapy by altering oxygen supply/demand relationships is
critical for any operator that employs it. Such pumps provide an avenue for both the reduction
of oxygen demand and for the delivery of therapeutic agents via the pump, which can be
administered intravenously or by using cardioplegia as a vehicle. Using cardioplegia offers a
unique opportunity to modify the conditions of reperfusion (hydrodynamics, temperature,
route of delivery) and the composition of the solution (blood, crystalloid, pH, metabolic
substrates, hypocalcemia, oxygen, pharmaceuticals) that target the pathophysiologic
mechanisms of injury. It is now also recognized that the initiation of circulatory support in it
itself offers an opportunity to mitigate injury by paying attention to conditions at the moment
support is begun and ischemic conditions are eliminated.

Reactive Oxygen Species (ROS) and Reactive Nitrogen Species (RNS)

Molecular oxygen is relatively inert by virtue of the shared electron pair in the outer molecular
shell. However, oxygen must be viewed as a pharmaceutical and potentially toxic agent in
cardiovascular medicine and surgery. ROS are radicals or unstable oxygen molecules derived
from oxygen. Important oxygen radicals are superoxide anion (−O2) and hydroxyl anion
(•OH).OH). However, other oxygen-related molecules that are not technically radicals are
considered biologically reactive nonetheless. An example of this would be hydrogen peroxide
(H2O2). In addition, hypochlorous acid (HOCl) and chloramines derived from neutrophils exert
potent oxidative injury to biologic tissues.

Free-radical molecules that are derived from nitrogen are termed RNS and include nitric oxide
(NO•), peroxynitrite (ONOO–), nitrogen dioxide (NO2•) and nitrosyl hydride (HNO). In vascular
endothelium, NO• is constitutively generated by endothelial nitric oxide synthase (eNOS or
NOS-3). eNOS activity is dependent on Ca2+-calmodulin, flavin, NADPH, and
tetrahydrobiopterin (BH4). NO• is generated by an electron from the guanadino nitrogen of L-
arginine to incorporate oxygen to generate NO • and the byproduct L-citrulline. NO• has an
extremely short half-life (seconds) in vivo, and binds tightly to hemoglobin. OONO− is formed
by a nonenzymatic biradical reaction between NO• and −O2. NO• has been linked to biologic
signaling in cardioprotection, particularly anti-inflammatory effects, endothelial function,
vasorelaxation (hence its early name as endothelial-derived relaxing factor, EDRF),
neurotransmission, immune regulation and defense mechanisms.

NO• homeostasis is key in normal regulation of blood flow and responses to stress, and in
preventing endothelial dysfunction. NO• homeostasis is a balance between NO • generation,
bioavailability, and degradation or quenching by oxidants such as −O2. NO • has pleiotropic
effects in the cardiovascular system by paracrine effects: (1) NO • is a physiologic regulator of
blood pressure and flow, and counteracts local and systemic vasoconstrictors such as
endothelin (ET-1); (2) it inhibits neutrophil adhesion to vascular endothelium; (3) it inhibits
platelet aggregation; (4) it reduces myocardial oxygen consumption (41); (5) it has positive
lusitropic and inotropic effects; (6) it scavenges −O2 and therefore is directly antioxidant; and
(7) it is anti-proliferative. The effects on oxygen consumption and inotropy are somewhat
controversial (42,43). The role of NO• in ischemia-reperfusion injury and cardioprotection is
discussed in more detail under the section on endothelial dysfunction and NO • donors.

Reactive Oxygen Species

The oxygen radical is highly reactive with a broad range of biologic substances including sugars,
amino acids, phospholipids, and DNA. With such a large catalog of biologic targets, ROS and
their metabolites are potentially toxic to cells, and at the same time they are critical in the host
defense mechanism and bactericidal activities of neutrophils and phagocytes (44) and
participate in cardioprotective signaling. In this dualistic role of ROS as good guy–bad guy, it is
proposed that low-level production of ROS, particularly from mitochondria, acts as a signal for
cardioprotective conditioning responses (preconditioning, postconditioning) and stimulation of
cardioprotective kinase pathways, while large-scale generation of ROS from inflammatory cells
or dysfunctional mitochondria is deleterious. In the ischemic-reperfused myocardium, ROS-
induced injury includes peroxidation of lipid components of cellular membranes leading to
damage to sarcolemmal cell membrane and other membranous organelles such as
mitochondria and sarcoplasmic reticulum (45). ROS also cause the mPTP to open when
generated in an environment of Ca2+ accumulation and normalized pH (conditions achieved
during reperfusion), which leads to a collapse of the membrane potential, loss of metabolic
capacity, and release of pro-apoptotic factors. This is discussed in more detail below under the
role of mitochondria in reperfusion injury. The vascular endothelium is particularly vulnerable
to ROS-mediated injury, which manifests as impairment of vascular endothelial function
through attenuated production of vasoactive and anti-inflammatory autacoids such as
adenosine (46,47,48) and NO• (49), and increased production of pro-inflammatory and
vasoconstrictive molecules. These injury mechanisms contribute significantly to postischemic
dysfunction, dysrhythmias, and morphologic injury such as necrosis and apoptosis.

Sources of Oxygen Radicals

Both enzymatic and nonenzymatic reactions produce oxygen radicals. The major reactions and
their sources are shown in Figure 9.3. ROS are generated by a number of sources, including
activated neutrophils, cardiac myocytes, mitochondria, and the vascular endothelium.
Figure 9.3. The species and sources of reactive oxygen species (ROS) generated by neutrophils,
cardiomyocytes and their mitochondria, and vascular endothelium. SOD, superoxide
dismutase; CAT, catalase; GP, glutathione peroxidase; MPO, myeloperoxidase; HOCl,
hypochlorous acid; e−1, electron; NOX, NAD(P)H oxidase.

Neutrophils: Activated neutrophils represent the major source of ROS in the heart under
stress, including −O2, •OH, and HOCl species. Neutrophils are activated by various chemotactic
factors stimulated by either CPB or ischemia-reperfusion, including the complement
anaphylatoxins C3a and C5a, f-Met-Leu-Phe, tumor necrosis factor alpha (TNFα), interleukin-8
(IL-8), and platelet-activating factor. These factors trigger a “respiratory burst” of ROS. CPB
(independent of ischemia) activates the complement cascade (50,51), with the subsequent
activation and adherence of neutrophils (52) to the vascular endothelium primarily, which
subsequently triggers the release of cytotoxic products, all of which contribute to tissue injury
in the ischemic-reperfused myocardium (52,53). During the respiratory burst that occurs
seconds after stimulation by cytokines or ischemia-reperfusion, the −O2 in the neutrophil is
produced by the membrane-associated NADP oxidase, which transfers an electron to
molecular oxygen (Fig. 9.3, reviewed in Sheppard et al. (54)). Production of H2O2 by
neutrophils or the reduction of −O2 to H2O2 by superoxide dismutase (SOD) provides
substrate for generation of •OH, or the oxidant HOCl by myeloperoxidase. HOCl reacts with low
molecular weight amines to give rise to lipophilic chloramines, which can promote membrane
lipid peroxidation.

Cardiomyocytes: Xanthine oxidase (XO) is a major source of ROS from cardiomyocytes.

Normally, XO exists mostly in the dehydrogenase form and uses nicotinamide adenine
dinucleotide (NAD) as an electron acceptor. Xanthine dehydrogenase is converted to the
oxidant-generating enzyme XO during ischemia. Both hypoxanthine and xanthine accumulate
during ischemia. Hypoxanthine is converted to xanthine by XO, which is in turn converted to
uric acid by the same enzyme, which forms −O2. During ischemia, however, two events favor
the xanthine oxidase reaction: (1) hypoxanthine accumulates as a result of the sequential
catabolism of purine high-energy phosphates (ATP, ADP, AMP) and deamination of adenosine,
and (2) the dehydrogenase form of the enzyme is converted to the superoxide-producing
oxidase form by a protease activated by increased levels of intracellular calcium (55). XO is
reported to be localized to the vascular endothelium in substantial quantities (56). However,
the importance of XO to the pathologic generation of ROS in humans remains controversial
(57).

Mitochondria: Mitochondria are a major noninflammatory site of ROS generation (58). ROS are
generated at complex I (NADH/ubiquinone oxidoreductase) and complex III
(ubiquinol/cytochrome c oxidoreductase). The latter site catalyzes the conversion of O2 to
−O2• by a single-electron transfer. −O2 can be converted in the mitochondrial matrix to H2O2
by matrix superoxide dismutase. Since mitochondria do not have catalase, the H2O2 can be
degraded by nonenzymatic reaction with glutathione (GSH) by glutathione peroxidase or by
enzymatic reaction with thioredoxin reductase. The ROS species target cysteine residues on
the protein moieties of the mitochondrial membrane and polyunsaturated fatty acids, and
cause intramolecular cross-linkages and protein aggregates. The •OH causes peroxidation of
membrane lipids. ROS induce Ca2+ release from the mitochondria, which may cause an
imbalance in Ca2+ handling and stimulation of Ca2+-dependent proteases, nucleases and
phosphatases leading to functional and morphologic damage to the cell.

Vascular endothelium: Superoxide anion is produced by vascular endothelium by XO as

described above for cardiomyocytes (Fig. 9.3). A more important source of −O2 from the
vascular endothelium may be the NAD(P)H (reduced NAD phosphate) oxidase system. This
enzyme is a membrane-bound, flavin-containing NADH/NADPH-dependent oxidase present in
endothelial and vascular smooth-muscle cells. The molecular structure, function, and
regulation are somewhat similar to those of the neutrophil enzyme system, requiring assembly
of cytosolic components onto the membrane components before −O2 generation can occur.
However, −O2 production in the presence of NO • produced by endothelium favors the
biradical neutralization and elimination of NO • and the formation of peroxynitrite (ONOO–).
The elimination of NO• impairs not only vascular relaxation and autoregulatory control of
postischemic myocardial perfusion, but also fails to inhibit neutrophil-mediated postischemic
inflammatory responses.

Other sources of ROS include the oxidation of catecholamines, metabolism of arachidonic acid
to peroxy compounds, and generation of •OH by the cyclooxygenase and lipoxygenase
pathways, and by the metal-catalyzed Haber–Weiss reactions. The latter series of reactions
may be important in the setting of CPB because they can be recruited by neutrophil activation
during CPB (59). In addition, iron-containing products released by hemolysis may facilitate this
reaction.
When are Oxygen Radicals Generated?

Myocardial ischemia favors the generation of ROS. However, the myocardium has a cadre of
the endogenous antioxidants superoxide dismutase, catalase, reduced glutathione (GSH),
glutathione peroxidase, glutathione reductase. Normally, upregulation of these antioxidants
and phase 2 enzymes reduces oxidant-mediated injury (60), but they are depleted or can be
overwhelmed during ischemia, thereby attenuating the natural defense mechanisms of the
heart (61). Therefore, ischemia creates the biochemical setting for oxyradical production and
tissue vulnerability to ROS-mediated damage. However, the primary substrate, oxygen, is in
limited supply during ischemia, and is not readily available until reperfusion. Oxygen radicals
are not seen in great abundance until the onset of reperfusion. A clinical study by Prasad et al.
(62) show that oxygen radicals are released during and 24 hours after CPB in CABG patients.
Indirect evidence to support this reperfusion oxidative burst comes from a wealth of data in
which inhibitors or scavengers of oxygen radicals, or anti-neutrophil agents, exerted beneficial
effects when administered during reperfusion.

In the surgical setting, the myocardium may be vulnerable to ROS-induced injury at several
points when oxygen is introduced into the myocardium: (1) the initial delivery of cardioplegia
into the heart with antecedent unprotected ischemia, or initial delivery of oxygenated
cardioplegia to a newly revascularized myocardial segment; (2) intermittent reinfusions of
cardioplegia if a multidose strategy is being used; and (3) release of the aortic cross-clamp to
initiate reperfusion after the cardioplegia phase. The quantity of ROS generated during cardiac
surgery depends on the severity of the preceding ischemic injury, activation and recruitment of
neutrophils, the level of oxygen in cardioplegic solutions or blood oxygenated by the
extracorporeal circuit, and the status of endogenous scavengers and inhibitors potentially
depleted during ischemia. Blood cardioplegia theoretically provides a greater oxygen substrate
for oxygen radical production than do crystalloid solutions. However, plasma-dissolved oxygen
may be the pool from which oxygen is drawn for the generation of oxygen radicals, and the
oxygen content in this compartment may not differ dramatically between blood and crystalloid
solutions. The oxygen in blood cardioplegia may be reduced with “normoxic” or even
“deoxygenated” blood cardioplegia (63) or during reperfusion (64,65). In addition, blood
cardioplegia may offer the advantage of endogenous antioxidants (superoxide dismutase,
catalase, glutathione, and albumin) that are not present in crystalloid cardioplegia solutions.

Intracellular Na+ and Ca2+ Dyshomeostasis

Under normal physiologic conditions, intracellular Ca2+ concentrations are maintained at low
concentrations (<200 nmol) against a 5,000-fold trans-sarcolemmal gradient by voltage-gated
Ca2+ channels that remain closed until activation during phase 2 of the action potential. The
regulation of intracellular and intramitochondrial Ca2+ is tightly linked to regulation of
intracellular Na+. Intracellular Na+ concentration is normally controlled by the ATP-dependent
Na+/K+ pump which moves Na+ outward and K+ inward against their respective concentration
gradients, and the energy-independent Na+/H+ exchanger which normally removes one H+
from the cytosol in exchange for one Na+ moving into the cytosol. Further Na+ accumulation
mainly occurs through the Na+/Ca2+ exchanger which normally operates in the forward
direction during diastole to extrude intracellular Ca2+ in exchange for influx of Na+ into the
cytosol.

In ischemic-reperfusion myocardium, Ca2+ influx occurs through multiple pathways: (1)

diffusion through overt membrane disruptions or opened Ca2+ channels, fueled by the
differential concentration gradient; (2) reversal or inhibition of the Na+/Ca2+ exchanger driven
by the intracellular accumulation of Na+, (3) attenuated cyclical Ca2+ uptake and release from
the sarcoplasmic reticulum and mitochondria, and (4) facilitation of entry via α-adrenoreceptor
activation and cyclic adenosine monophosphate (AMP)-dependent processes, which increase
during postischemic catecholamine release.

The accumulation of intracellular Na+ and Ca2+ during ischemia-reperfusion are linked events
leading to Ca2+ overload which stimulates Ca2+-dependent proteases that damage
membranes and ion transport mechanisms at the moment of re-exposure to molecular
oxygen, opening of the mPTP, depletion of high-energy phosphate stores, the development of
local or global contracture (“stone heart”), and triggers the transition to cell death through
both apoptosis and necrosis. The balanced Ca2+ transport systems are potentially injured by
exposure to oxygen radicals, cytokines, and activated complement during CPB and reperfusion.
Na+ and Ca2+ dyshomeostasis occurs to differing degrees during ischemia and during
reperfusion, as described below.

Ischemia

During ischemia, glycolysis generates an abundance of H+ and only a limited amount of ATP (2
moles/mole glucose) compared to oxidative phosphorylation (36 moles ATP/mole glucose).
The accumulation of intracellular H+ stimulates activity of the Na+/H+ exchanger (NHE1) in
favor of H+ efflux and Na+ influx; at this time, the reduced availability of ATP decreases Na+/K+
ATPase activity. Therefore, both the attenuated Na+/K+ pump activity and increased activity of
the Na+/H+ exchanger result in intracellular Na+ accumulation. Additional Na+ may enter the
cell through sarcolemmal nonactivating Na+ channels, for example, the “window currents”
generated during ischemia-induced cellular depolarization. However, some evidence suggests
that the accumulation of H+ in both the extracellular and intracellular spaces attenuates the
H+ gradient, thereby reducing but not eliminating the participation of the Na+/H+ exchange
mechanism in intracellular Na+ accumulation during ischemia (66). However, the accumulation
of Na+ in the cytosol during ischemia would favor reversal of the energy-independent
Na+/Ca2+ exchanger in favor of extruding Na+ in exchange for Ca2+ influx, thereby favoring its
accumulation in the cytosol. Entry of Ca2+ into the mitochondria through the Ca2+ uniporter
depends on the Ca2+ electrochemical gradient and the presence of a mitochondrial
transmembrane potential. Ca2+ enters the mitochondria during early ischemia, but further
increases and overload are prevented by dissipation of the mitochondrial transmembrane
potential (67). In addition, the unregulated activation of Ca2+-dependent protease enzymes
such as calpains is inhibited by intracellular acidosis, and is delayed until intracellular pH is
renormalized during reperfusion (68). Therefore, Ca2+ accumulation and Ca2+-mediated
damage occurs to a greater extent during reperfusion (69,70,71) as discussed below.
Reperfusion

Na+ and Ca2+ dyshomeostasis during reperfusion is a significant factor contributing to

postischemic myocardial injury and functional impairment (72,73). The ion dyshomeostasis
described above sets the stage for explosive reperfusion/reoxygenation damage. It is
increasingly recognized that abrupt re-exposure to oxygen itself is the trigger for damage.
Excessive gradients of perfusate-to-tissue oxygen tension can be mitigated. In addition, the
events leading to abnormal Ca2+ handling which determine cell survival versus cell death are
as follows:

● Rapid renormalization of intracellular pH: Reperfusion washes out lactate and re-
establishes the intracellular–extracellular H+ gradient (extracellular < intracellular)
which allows the forward motion of the Na+/H+ exchanger favoring removal of H+
(renormalization of intracellular pH) and subsequent intracellular accumulation of Na+;
other factors contributing to normalization of intracellular pH are activation of
lactate/H+ co-transport mechanism and activation of the NaHCO3 co-transporter.
● Renormalization of intracellular Na+: During the initial phase of reperfusion, ATP is
regenerated, and if the Na+/K+ ATPase is not damaged by proteolytic actions of
calpain enzymes, intracellular Na+ levels are renormalized and the linked Na+/H+
exchange—Na+/Ca2+ exchange systems will not be activated. Otherwise, if the Na+/K+
ATPase pump is damaged, intracellular Na+ accumulation will engage the Na+/Ca2+
exchange mechanism with net intracellular accumulation of Ca2+.
● Ca2+ uptake and release by the sarcoplasmic reticulum (SR): The SR is the main Ca2+
storage site of the cardiomyocytes. Ca2+ in the micro-domain proximate to the closely
co-localized SR and interfibrillar mitochondria is cyclically taken up by the SR (lowering
of cytosolic concentration) during diastole, and Ca2+ is released into that micro-
domain by the ryanodine receptors (RyR) during systole. During both unprotected and
protected ischemia, sequestration of Ca2+ into the sarcoplasmic reticulum competes
for a limited amount of ATP. This paradox creates a cycle whereby a high intracellular
Ca2+ level binds high-energy phosphates to further limit ATP availability, which then
attenuates the ability of the SR to remove the Ca2+ actively from the cytoplasm,
resulting in a further increase in cytosolic Ca2+. If other Ca2+ handling mechanisms are
not engaged, the release of Ca2+ from the SR RyR may lead to overload. This activity of
the SR mechanism may be related to the severity of ischemia, being less so during
short periods of ischemia.

So, what events in the balance of Ca2+ regulation determines if a cell lives or dies during
reperfusion? Some evidence suggests that the timing of normalization of pH and Ca2+ handling
mechanisms are important. If pH normalization occurs before Ca2+ handling mechanisms are
restored, then Ca2+ overload, in conjunction with ROS generation and a normalized
intracellular pH environment, will trigger opening of the mPTP causing cell death. However, if
Ca2+ handling mechanisms are restored before pH is normalized, for example, by maintaining
tissue acidosis during early reperfusion, then Ca2+ overload can be avoided. Interestingly,
venous blood-based perfusate has been preliminarily observed to mitigate both over-
oxygenation during early reperfusion and delayed normalization of pH, thereby creating a
“permissive acidemia.”
The Inflammatory and Innate Immunity Responses to Ischemia-Reperfusion and
Cardiopulmonary Bypass

Injury mechanisms of ischemia-reperfusion and CPB involve inflammatory and innate immune
responses. Cardiomyocytes generate pro-inflammatory cytokines, chemokines and adhesion
molecules that direct the emigration of inflammatory cells into ischemic-reperfused tissue
regions and amplify the inflammatory response. The following sections will describe the
inflammatory process involving neutrophils, pro-inflammatory cytokines, and complement,
and the innate immune response involving toll-like receptors (TLRs).

Neutrophils and Neutrophil-Mediated Injury

Neutrophils play a critical role in the pathogenesis of myocardial reperfusion injury (74,75,76).
The inflammatory component of reperfusion injury requires an interaction between activated
neutrophils and vascular endothelium during the early moments of reperfusion, which is
prerequisite to activation of the full local inflammatory cascade. A well-orchestrated sequence
of events mediates the interaction between adhesion molecules on neutrophils and
endothelium. These adhesion molecules are categorized into three families:

● The selectins (P-selectin, L-selectin, E-selectin) are glycoproteins involved in the

interactions between neutrophils and the vascular endothelium in the early moments
of reperfusion (77,78,79). P-selectin is stored in Weibel–Palade bodies in endothelial
cells, and its surface expression is triggered by pro-inflammatory mediators such as
oxygen radicals, thrombin, complement components, cytokines, histamine, and H2O2
released during ischemia-reperfusion and CPB. IL-8 released from hypoxic endothelial
cells also increases the adhesiveness of endothelial cells for neutrophils (80). P-selectin
surface expression peaks after 10 to 20 minutes of reperfusion, and P-selectin is
subsequently shed to soluble fragments in blood. In contrast to P-selectin, L-selectin is
constitutively expressed on the surface of neutrophils and may be the counterligand
for P-selectin during early reperfusion (81). Loose adherence and “rolling” of
neutrophils on endothelium involves interaction with a glycoprotein sialyl Lewisx or a
high-affinity glycoprotein ligand termed P-selectin glycoprotein ligand-1 (PSGL-1) (82).
The third member of the selectin family, E-selectin, is expressed on the surface of
endothelial cells. It is expressed later in reperfusion (4–6 hours) and may therefore be
involved in later events, such as infarct extension.
● The β2-integrins (CD11/CD18 complex) are a family of glycoproteins located on
external membranes of neutrophils. There are three distinct α-chains (CD11a, CD11b,
CD11c) and a common β-subunit. Surface expression of perhaps the major complex
CD11b/CD18 is triggered by a number of pro-inflammatory mediators, including
circulating and endothelium-derived (acylated) platelet-activating factor. After initial
tethering of neutrophils by endothelial P-selectin, firm adherence is mediated by
interaction with CD11b/CD18 and its counterligand ICAM-1 on the endothelium.
● The third family of adhesion molecules is the immunoglobulin superfamily, including
ICAM-1 (intercellular adhesion molecule-1), VCAM-1 (vascular cell adhesion molecule-
1), and PECAM-1 (platelet–endothelial cell adhesion molecule-1). ICAM-1, the
counterligand for CD11/CD18 on neutrophils, is constitutively expressed on
 endothelial cells and is upregulated by cytokines 2 to 4 hours after myocardial
ischemia-reperfusion (83), coincident with the upregulation of CD11/CD18.
Subsequently, a high-affinity bond between the neutrophil and endothelial cell occurs
via integrins (CD11/CD18) on neutrophils and ICAM-1 and ICAM-2 on endothelial cells.
Once the neutrophil has firmly attached to the endothelial surface, it migrates through
the endothelial cell junctions to the myocardium, partly mediated by IL-8. The
shedding of L-selectin from the neutrophil and the presence of PECAM on the
endothelium may be required for neutrophil transendothelial migration. The
prerequisite nature of this interaction between neutrophils and adhesion molecules
makes anti-neutrophil and anti-adhesion therapy a potentially effective strategy, as
these agents intervene at proximal points in the cascade (discussed below).

Evidence supporting the rapid accumulation of neutrophils within ischemic-reperfused

myocardium following the neutrophil–endothelial cell interactions was appreciated not only by
our laboratory, but also by Sommers and Jennings (84). We (85,86,87,88,89) and others (90)
have demonstrated an association between the extent of postischemic injury and neutrophil
accumulation within ischemic-reperfused myocardium as well as a reduction of postischemic
injury associated with neutrophil suppression. As shown in Figure 9.4, Lefer et al. (91)
demonstrated that neutrophil adherence to coronary artery endothelium occurs in advance of
accumulation in extravascular sites, representing the lag time involved in transmigration of
intravascular neutrophils. In addition, neutrophil accumulation occurs in advance of the
development of necrosis, consistent with active neutrophil participation in cell demise. Brix-
Christensen et al. (92) reported that neutrophils accumulated in the hearts and other organs of
neonatal pigs placed on CPB. Therefore, neutrophil accumulation occurs in both postischemic
and post-CPB hearts.

Figure 9.4. The postischemic time course of endothelial injury, contractile dysfunction,
neutrophil (PMN) adherence and accumulation in tissue, necrosis, and apoptosis. These events
are indicated on the y-axis as a percent completion of that event (0% not started, 100% is
complete). Note that PMN adherence parallels that of endothelial dysfunction representing
PMN-mediated damage. Note also that accumulation in tissue lags behind adherence
representing emigration time. (Data obtained, in part, from Zhao ZQ, Nakamura M, Wang NP,
et al. Dynamic progression of contractile and endothelial dysfunction and infarct extension in
the late phase of reperfusion. J Surg Res 2000;94:133–144.) Apoptosis is a slower process than
necrosis as observed by Zhao et al. (565).(Adapted in part from Lefer AM, Tsao PS, Lefer DJ, et
al. Role of endothelial dysfunction in the pathogenesis of reperfusion injury after myocardial
ischemia. FASEB J 1991;5:2029–2034.)

The importance of neutrophils in postbypass reperfusion injury has been well-established in

both animal and human models. The myocardium is exposed to neutrophils during the delivery
of intermittent or continuous blood cardioplegia, or during reperfusion (after cross-clamp
removal) following either crystalloid or blood cardioplegia. Schwartz et al. (93) showed that
CPB circuitry directly “primes” neutrophils by depositing C3b on extracorporeal tubing, leading
to neutrophil activation and sequestration. Further evidence for the activation of neutrophils
following CPB was suggested by the finding of increased neutrophil surface expression of
CD11b/CD18 and decreased surface expression of L-selectin on cells circulated in simulated
CPB circuits (94,95). Finn et al. (96) noted in patients undergoing CPB that a rise in circulating
IL-8 begins at reperfusion with rewarming of the patient after systemic hypothermia, and
closely correlates with the rise in neutrophil count and circulating elastase. Similarly, Schwartz
et al. (93) and Gessler et al. (97) found higher levels of the pro-inflammatory mediators soluble
IL-6 and IL-8 6 hours after bypass than before bypass.

The time course for the upregulation of neutrophils and various adhesion molecules is equally
important in designing treatment strategies for patients undergoing extracorporeal bypass. In
a clinical study, Galiñanes et al. (98) investigated the perioperative temporal changes in the
expression of various neutrophil surface adhesion molecules, concluding that downregulation
of PECAM-1 is an early indicator of neutrophil activation and may therefore represent a target
for therapy aimed at reducing the inflammatory response associated with CPB.

The role of temperature in the activation of neutrophils or other inflammatory mediators

following CPB in humans has been investigated. Hypothermia has been reported to attenuate
expression of adhesion molecules in some models of tissue injury (99,100). Some investigators
have reported minimal or no difference in neutrophil activation and expression of neutrophil
adhesion molecules between normothermic (34°C–37°C) and hypothermic (28°C–30°C) CPB
(101,102), whereas others note an attenuation in the expression of IL-8 and neutrophil
elastase activity during normothermic CPBsurgery (103). Hypothermia has been shown to
delay, but not prevent, the increased expression of adhesion molecules such as CD11b and
CD11c (101). When measured 4 hours after bypass, the levels of IL-1α, soluble ICAM-1 (sICAM-
1), and elastase are higher after normothermic CPB than after hypothermic CPB (104,105). In
conclusion, further research is required to determine the most appropriate temperature of
cardioplegic administration to minimize the deleterious consequences of neutrophil-mediated
damage. Interventions designed to attenuate neutrophil-related myocardial injury following
CPB will be discussed later.
Complement in CPB and Myocardial Reperfusion Injury

The complement cascade, particularly the alternative pathway stimulated by contact

activation, is activated independently by myocardial ischemia-reperfusion and CPB, and
contributes importantly to the pathologic sequelae of CPB (106). Complement fragments such
as the anaphylatoxins C3a and C5a are generated and released both systemically and locally by
the ischemic-reperfused myocardium (107). The activation and migration of neutrophils to
ischemic-reperfused myocardium requires stimulation by chemotactic factors, most notably
complement fragments (C3a and C5a), eicosanoid products such as leukotriene B4, and
platelet-activating factor (108,109). Complement fragments (C5a) can be found in lymph from
ischemic myocardium (110), are temporally associated with reperfusion (111), and correlate
with increased neutrophil accumulation in reperfused myocardium. Furthermore, C5a directly
stimulates neutrophil −O2 production and enhances adherence to coronary artery
endothelium (112), whereas C3 activates monocyte adherence (113). The chemoattractant and
chemotactic properties of C5a cause progressive generalized vascular leukosequestration (114)
during CPB. The interaction between blood and extracorporeal surfaces also stimulates
deposition of C3b (115), which in turn amplifies the complement alternative pathway to
release C3a and C5a, and subsequently form the membrane attack complex (C5b 9) (115).
Accordingly, plasma levels of the anaphylatoxin C5a increase in patients undergoing CPB
(115,116), which activates neutrophils to generate −O2 and promotes adhesion to endothelial
cells, stimulates degranulation of mast cells, stimulates vascular smooth-muscle contraction,
and increases vascular permeability. Complement activation constitutes an early-phase
response to CPB, which then progresses to a second, slower response phase in which de novo
synthesis of proteins (E-selectin, ICAM, IL-8) and expression of adhesion molecules on cell
surfaces occurs. The magnitude of the complement response has been linked to the duration
of CPB and the type of circuitry used (116). The combined stimuli of myocardial ischemia-
reperfusion and CPB may synergize to exacerbate postischemic neutrophil-mediated
inflammatory responses in patients undergoing CPB for myocardial revascularization.

Actions of Toll-like Receptors

Recent evidence has shown that activation of TLRs is involved in the pathogenesis of
reperfusion injury (117,118). TLRs are transmembrane receptors expressed on endothelial
cells, cardiomyocytes, platelets, monocytes, and neutrophils. Cardiomyocytes participate in
the inflammatory and innate immune response to ischemia-reperfusion by releasing cytokines
(TNFα, IL-1β, IL-6, IL-8, IFN-γ), chemokines, and by expressing cell surface adhesion molecules.
TLRs are pattern recognition receptors (PRRs) that recognize pathogen-associated molecular
patterns (PAMPs) such as LPS or viral RNA on foreign pathogens, and sense endogenous
danger molecules and signals from injured cells. TLRs also recognize and are activated by
endogenous ligands such as danger-associated molecular patterns (DAMPS) (e.g., heat shock
proteins, HSP70) induced in cardiac tissue during stress (119) and cardiac surgery (120,121).
Inducible HSP70 has been found in plasma from CABG patients immediately after surgery
(122). ROS may also stimulate TLRs. Ten to eleven human TLRs have been identified; (123,124)
TLR receptor-1, -2, -4, -5, and -6 are located on the cell surface, with TLR-2 and TLR-4 being the
most involved in ischemia-reperfusion of the group of TLRs; TLR-2 works with other TLRs in co-
receptor cooperativity to recognize diacyl- and triacyl-lipoproteins, while TLR4 recognizes LPS
from gram-negative bacteria and DAMPs such as HSP70. Neutrophils constitutively express all
TRLs except TLR3.

The stimulation of TLRs during ischemia-reperfusion and cardiac surgery can lead to cell injury.
Dybdahl et al. (122) reported the release of HSP70 into the circulation of CABG patients and a
concomitant increase in monocyte TLR-2 and TLR-4 one day after surgery. Human monocytes
exposed to HSP70 released IL-6 and TNFα. Stimulation of TLRs by interaction with DAMPs
during ischemia-reperfusion is transduced by numerous molecular signaling pathways, which
ultimately results in the activation and translocation of NF-κB to the nucleus, where it
stimulates gene expression and release of cytokines and innate inflammatory factors. ROS
generated during reperfusion also stimulate TLR-dependent (most likely TLR-4) generation of
NF-κB through the PI3-K/Akt pathway, which suggests cross talk between the TLR and
reperfusion injury survival kinase (RISK) pathways. Inhibitors of TLRs, intermediary pathways,
or NF-κB will attenuate ischemia-reperfusion injury.

The Mitochondrial Permeability Transition Pore as a Determinant of Cell Death

Mitochondria are best known for being the power house of the cell, where ATP is generated
through oxidative phosphorylation. However, the mitochondria have been shown to function
as a “molecular switch” that governs cardiomyocyte viability; indeed it is a critical fulcrum
point in the transition from reversible to irreversible injury. Mitochondria can direct the cell to
pursue either an apoptotic or a necrotic pathway to cell death. The key event of this molecular
life/death switch is opening of the mPTP. This pore is a transmitochondrial membrane voltage-
dependent and Ca2+-dependent high-conductance channel located in the inner mitochondrial
membrane that permits water and solutes with a mass up to 1.5 kDa to enter the
mitochondrion. Under normal physiologic conditions, the mPTP is in a closed state. It opens in
response to high levels of intracellular ROS, Ca2+ and inorganic phosphate when mitochondria
matrix and intracellular pH are in the normal range. The mPTP remains closed during ischemia
because intracellular acidosis and the intracellular levels of Mg2+ and ADP counteract the
opening triggered by ROS and Ca2+. However, the mPTP opens during the early moments of
reperfusion (125) when there is an accumulation of ROS, Ca2+ and inorganic phosphate, and
when there is a rapid renormalization of intracellular and mitochondrial matrix pH (126). The
opening of the mPTP increases the permeability of mitochondria to water and small solutes,
causing swelling of the matrix and collapse of the mitochondrial transmembrane potential that
drives oxidative phosphorylation and ATP generation. mPTP opening also stimulates the
release of the pro-apoptotic factors cytochrome c and apoptosis-inducing factor (AIF). The
switch function comes into play based on the availability of intracellular ATP; if ATP is present
then the cell pursues an apoptotic pathway because apoptosis requres ATP. If ATP levels are
depleted then the cell pursues a necrotic pathway. With prolonged antecedent ischemia mPTP
is irreversible, but milder or short periods of ischemia trigger transient reversible opening
(127).
CHAPTER 9

SURGICAL MYOCARDIAL PROTECTION

Jakob Vinten-Johansen and Neil J. Thomas

INTRODUCTION

The science of myocardial protection had its beginnings in the 1950s, and has enjoyed a robust
and often controversial development since. Great strides were made in the 1980s and 1990s
which shaped the cornerstones of the cardioplegia strategy that we largely use today
(crystalloid or blood, warm, tepid or cold, continuous or intermittent, antegrade or
retrograde). Mortality has substantially decreased over the past decades as a result of applying
cornerstone as well as novel cardioprotective strategies and new technologies.

However, the patient of the 1980s and 1990s during which cardioplegia was developed has
morphed considerably in the 2000s. Active lifestyles have changed to more sedentary lifestyles
beginning in the early school years with less emphasis placed on physical activity, and more
distractions from active lifestyles imposed by the digital-age gadgets. This, coupled with the
availability of fast food with high caloric, high fat, and high sodium content, has conspired to
radically change our body habitus and biology; indeed, diabetes, obesity,
hypercholesterolemia, and the metabolic syndrome are more prevalent today. These more
prevalent comorbidities have impacted the efficacy of modern surgical therapies and the
outcomes (1) of myocardial protection strategies. Experimental animal models in which
effective cardioplegia strategies were developed decades ago, and observed to be effective in
the healthier patients with normal myocardial tissue substrate preoperatively in which they
were initially tested clinically seems to apply in fewer and fewer patients today. The “one
formula-fits-all” strategies of the past must be rethought in the current era. With patients
living longer as well, advancing age and the more senescent myocardium can display altered
responsiveness to some therapeutics. So we may ask, is myocardial protection developed in
the 1980s and 1990s being used as effectively on the patient of the 2010s? And, furthermore,
the yardstick by which our results are measured is changing in the world of health care today,
where a 30-day surgical mortality rate may no longer be the gold standard. Our results must
now be viewed in terms of longer-term, event-free survival and with measure of long-term
myocardial performance after an index event or surgery.

The concepts of cardioprotective strategies have changed significantly as our understanding of

the pathophysiologic determinants of ischemia, and particularly the complex mechanisms
involved in reperfusion, have expanded. Early empiric, largely physiologic observations have
given way to complex biochemical and molecular interactions, which in some cases have
reinforced physiologic intuition and in other cases have redirected our thinking, but has
nonetheless provided a scientific underpinning to a more molecular approach in developing
target-specific strategies of myocardial protection. This chapter will discuss the pathobiology
of ischemia-reperfusion injury, provide the basis by which the severity and duration of evolving
ischemic conditions set the stage for reperfusion and oxygen-related injury and its impact on
affected tissues as a basis for understanding cardioprotective strategies to attenuate injury. It
will also discuss new, emerging concepts of myocardial protection that have developed out of
a better understanding of the pathobiology of ischemia-reperfusion injury, and a growing
appreciation that the preoperative condition of the tissue may be equally, if not more
important, than any intraoperative strategy employed in previously normal tissue. It will also
discuss the all-elusive “golden ring” of arresting the heart in a nondepolarized state.

The goal of this chapter is to provide the reader with scientific and practical knowledge of
myocardial protection that can be applied to design appropriate strategies to meet the need of
specific patients, their unique demographics and clinical scenarios, and to provide flexibility in
the approach based on both preoperative condition and intraoperative events. It will also
provide some guided reading to enhance the background of relevant topics

HISTORICAL PERSPECTIVES ON MYOCARDIAL PROTECTION AND REPERFUSION INJURY

The need to protect the myocardium from intraoperative and postoperative damage was
realized before the development of cardiopulmonary bypass (CPB). Temporary inflow
occlusion techniques to limit blood flow to the heart permitted the repair of simple
intracardiac defects, and in many cases the occlusion time could be increased by adding
moderate systemic hypothermia (30°C) (2). After the development of CPB by Gibbon and
colleagues in 1954 (3,4), profound systemic hypothermia could be used safely to provide a
hemodynamic safety net, and to provide protection of the heart and other body organs by
permitting longer periods of ischemic cardiac arrest for the repair of more complicated lesions.

The development of extracorporeal circulation techniques allowed cardiac procedures to

proceed without threat of hemodynamic collapse. However, performing complicated
procedures on the heart was made difficult by its continuous motion and blood in the field.
Ischemia was known to achieve cardiac arrest (ischemic or anoxic arrest) through the
depletion of high-energy phosphate stores on which cardiac contraction depends. However,
myocardial ischemia led to necrosis and contractile dysfunction upon reperfusion, the worst of
which was an irreversible “stone” heart (5,6). The conundrum of surgeon convenience at the
expense of acceptable outcomes provided a driving force for innovation.

Once the severe complications of ischemic arrest were recognized, chemical methods were
used to arrest the heart. Based on reports from Ringer in 1883 that high concentrations of
potassium would arrest the heart (7), Melrose and associates in 1955 (8) and 1957 (9) used
hyperkalemia to induce asystole during CPB. The combination of immediate arrest and
hypothermia would ostensibly reduce energy (i.e., oxygen) demands of the heart, and thereby
avoid injury imposed by hypothermia alone. Melrose made a hyperkalemic solution by mixing
potassium citrate solution (77 mM) with 18 mL of blood, which was then infused by hand-held
syringe into the aortic root proximal to the aortic cross-clamp. This hyperkalemic arresting
solution was later called “cardioplegia” solution by Lam et al. (10) and Sergeant et al. (11).
Other contemporaries of Melrose used alternative arresting agents, such as acetylcholine
alone (10,11) or in combination with other drugs (12).

Although asystole was indeed achieved with the Melrose potassium citrate solution, the
concept of elective cardiac arrest was abandoned in the United States for some 20 years
following reports of high morbidity and pathologic complications in chemically arrested hearts
(13). Chemical cardioplegia was supplanted by alternative techniques of arresting the heart or
limiting blood flow in the surgical field, including ventricular fibrillation, intermittent aortic
cross-clamping, and profound systemic (14) or topical hypothermia (15). Laboratory studies
demonstrated that topical hypothermia protected inadequately against intraoperative injury,
which resulted in postoperative myocardial subendocardial necrosis, and postischemic
metabolic and functional depression (16,17,18). In the late 1960s, reports described scattered
myocardial or subendocardial necrosis in patients who had died after otherwise technically
successful cardiac surgical operations, suggesting that the current techniques of myocardial
protection were inadequate (19,20). This set the stage for a later re-emergence of chemical
cardioplegia in the United States.

During the surgeon-imposed moratorium on the use of the Melrose solution, cardioplegia
solutions continued development and clinical use in Europe. Bretschneider et al. (16) and
Kirsch et al. (17) used crystalloid formulations which mimicked intracellular ionic
concentrations (intracellular crystalloid solutions). The Bretschneider HTK (histidine,
tryptophan, α-ketoglutarate)-ketoglutarate) solution used low calcium, low sodium, and
procaine with histidine buffers to achieve a nondepolarized arrest, rather than hyperkalemia-
induced depolarized arrest. In the 1970s, Hearse and associates (18,21) at St. Thomas’ Hospital
in London developed an extracellular hyperkalemic solution based on a rigorous series of
experiments that established a scientific understanding of the determinants of injury incurred
during global ischemia. Braimbridge introduced St. Thomas’ cardioplegia solution into clinical
use in 1975 (22). St. Thomas’ solution was later modified based on further experimental
studies to the St. Thomas’ solution No. 2 (23), an intracellular ionic concentration,
normocalcemic, and hyperkalemic (16 mmol/L) solution; it is currently marketed as Plegisol in
the United States. The Bretschneider HTK solution is currently marketed as Custodiol HTK
solution.

Chemical cardioplegia was revived in the United States by Gay and Ebert (24) and Tyers et al.
(25,26), who found that the constituents in the Melrose formulation were inappropriate rather
than the concept of hyperkalemic chemical cardioplegia itself being erroneous. These reports
popularized the use of potassium-based cardioplegia to achieve electromechanical arrest. In
the late 1970s and early 1980s, Follette and colleagues from Dr. Buckberg’s laboratory (see
reference 224 below) introduced the concept of cold hyperkalemic blood cardioplegia. The
physiologic attributes of blood, including its superior buffering capacity, endogenous oxygen
radical scavengers, detoxifying substances, and superior oxygen-transport capacity, has
established this concept as a popular strategy that has facilitated surgery and vastly improved
patient outcomes. Today, a variety of both crystalloid and blood cardioplegia solutions are
used clinically worldwide to achieve elective cardiac arrest and a bloodless field.

The cornerstones of myocardial protection that had been laid by the end of the early 1980s
are: (1) rapid chemical arrest to conserve high-energy phosphate stores and reduce ischemia,
(2) hypothermia to reduce the metabolic rate, (3) adjunctive agents to reduce the effects of
ongoing ischemia (anti-ischemic agents), and (4) continuous or intermittent delivery to restore
or maintain tissue oxygenation. Further developments in cardioplegia-related myocardial
protection strategies focused on (1) optimizing delivery (retrograde, gentle cardioplegia
delivery pressures), (2) overcoming the adverse effects of hypothermia, (3) identifying
metabolic enhancements (ATP generation and pH management) and buffers, (4) managing
calcium (Ca2+) accumulation, and (5) understanding the pathophysiology of ischemia and
reperfusion, which would drive further developments of strategies to avoid these mechanisms
of both ischemic and postischemic patterns of injury. There has also been a persistent search
for agents that induce polarized arrest which overcome the disadvantages of hyperkalemia
(reviewed in Dobson et al. (27) and Maruyama et al. (28)). With the recent rapid developments
in molecular and cellular biology, experimental focus has shifted away from the organ and
physiologic aspects toward cellular and molecular areas to improve the understanding of the
effects of ischemia-reperfusion injury, inflammation, and cell death processes. Understanding
the mechanisms involved in the transition to cell death would direct developments in Ca2+
handling and preventing Ca2+ dyshomeostasis, the role of oxidant injury, and a newly
recognized role of mitochondria in determining the transition from reversible to irreversible
injury, and whether the cell pursues a necrotic pathway or apoptotic pathway to cell death.
Finally, the concept of conditioning, which recruits complex endogenous biologic mechanisms
of self-protection against stressors such as ischemia-reperfusion injury, oxidants and
apoptosis, has been incorporated into pretreatment (preconditioning), intraoperative
(preconditioning), and reperfusion (postconditioning) phases of cardiac surgery. Although
human beings have been evolving for hundreds of thousands of years, it has only been in the
past 60 years or less that the human heart has been subjected to the kind of induced ischemia
and surgical reperfusion discussed in these pages.

PATHOPHYSIOLOGY OF SURGICAL ISCHEMIA-REPERFUSION INJURY

Because many factors affect postoperative outcomes of hearts with various and complex
preoperative pathophysiologic features, developing strategies that adequately protect all
hearts during cardiac surgery is often problematic. The heart under consideration may present
a complex picture of preoperative disease, variable age, gender, metabolic and cellular
dysfunction, global or regional ischemia (both acute and chronic), each of which contribute to
differing vulnerabilities to ischemia and reperfusion injuries. However, adequate myocardial
protection must be based on a sound scientific basis. Obviously, a single cardioprotective
strategy may be inadequate to suit all surgeon preferences and target all patient pathologies.
Hence, the surgical team must often mold the composition of the cardioplegic solution or its
modality of delivery to meet the requirements of the patient’s pathologic profile as presented.
The measures taken to protect the heart during elective cardiac arrest should represent a
balance struck between the requirements of the heart during aortic cross-clamp or
reperfusion, with appropriate adjustments made for special considerations (hypertrophy,
diffuse coronary disease), and avoid distraction from the surgical procedure. Ischemia and
reperfusion injury (29) and the choice of cardioplegia technique (30) represent primary
contributors to patient outcomes.

Ischemic Injury

When Does Ischemia and Reperfusion Injury Occur during Cardiac Surgery?
Figure 9.1. Events occurring before the institution of cardiopulmonary bypass, during delivery
of cardioplegia and at reperfusion that predispose the myocardium to ischemia and
reperfusion injuries. Arrhythmias, ventricular fibrillation (VF) or severe hypotension can cause
antecedent global ischemia. During the delivery of cardioplegia, the composition of the
solution can cause edema, further ischemia if oxygen and nutrients are inadequate to meet
ongoing albeit lower demands, and Ca2+ concentration is either too high or too low. Coronary
obstructions, air or low cardioplegia infusion pressure can lead to maldistribution of solution.
Too high infusion pressure can lead to microvascular injury and edema. Potassium itself has
negative effects secondary to placing the cell in a depolarized state. Ischemia may be
encountered if air emboli, graft kinks, tight anastomoses, or systemic hypotension cause
inadequate distribution of blood flow. Finally, reperfusion injury can cause reversible (stunning,
arrhythmias, edema) or irreversible (necrosis, apoptosis) injury.

It is critical for the surgical team to understand that the presence of acute, ongoing and
evolving or progressing ischemia must impact the protection strategy. Ischemia-reperfusion
injury can occur in the surgical setting at three major time points during the surgical procedure
(summarized in Fig. 9.1).

(a) Before CPB has been instituted or cardioplegia solution has been delivered:
“Unprotected” clinical ischemia before CPB usually results from the presenting acute coronary
syndrome and can be exacerbated by hypotension, arrhythmias (e.g., atrial or ventricular
fibrillation), and other hemodynamic changes (e.g., cardiogenic shock, coronary artery spasm).
In most cases, acute ischemia is active in both the distribution of the principally involved artery
(cardiologists call this the “infarct artery”) but also distributions in which preexistent collateral
flow may be compromised. Patients referred for bypass surgery in these circumstances almost
always have multi-vessel disease patterns making the myocardial substrate dangerous from
the standpoint of the severity and duration of the syndrome, which can be extremely difficult if
not impossible to determine. Reperfusion and reoxygenation injury occur when these
conditions change abruptly before effective measures are implemented. Recent attention has
been paid to avoiding over-oxygenating such patients systemically, and this will be discussed in
more detail in later sections.

(b) At the initiation of cardiopulmonary bypass: At the moment bypass is initiated, blood is
immediately diverted from the right atrial drainage system into a venous reservoir where it is
then pumped through the oxygenator. This configuration does more to the heart than merely
oxygenate venous blood and send it back to the ascending aorta. As the heart is emptied into
the venous side of the pump, the oxygen demand of the metabolizing myocardium decreases
by 50%, but the high oxygen content of blood extracorporeally oxygenated makes the tissue
vulnerable to reoxygenation damage if no steps are taken to mitigate this injury. This might be
viewed as an abrupt interruption of antecedent ischemic conditions. While this is one of the
principal goals of bypass (to give the heart a “rest”), it must be recognized that the heart is also
vulnerable to a pattern of injury that has been, at best, underappreciated in the past. Ihnken et
al. (31) suggested as early as 1998 that initiation of bypass is also, perhaps paradoxically, a
dangerous moment for the heart. In this work, he observed clear, reproducible biochemical
evidence of injury in samples of blood drawn from the coronary sinus. Since Buckberg (32)
published his work on the reoxygenation injury in 1995, it has been observed that abrupt over-
exposure of the ischemic heart to oxygen increases injury in a pO2-dependent pattern of injury
that can be mitigated by altering the oxygen tension of the perfusate.

(c) During the “protected” period of arrest and delivery of cardioplegia: Ischemia or
reperfusion injury can occur during the cardioplegia phase. Ischemia can develop or be
worsened under several circumstances (1) the unintentional maldistribution of cardioplegia
solution distal to stenotic or totally occluded coronary arteries, (2) between intermittent
infusions of cardioplegia solution, (3) during interruption of continuous cardioplegia strategies,
(4) inadequate delivery of retrograde cardioplegia to areas of the right heart whose venous
blood drains directly into the Thebesian system, or areas that are simply under-filled due to
malposition of a retrograde delivery device. Though the cardioplegia phase is considered
“protected time,” the potential for worsening of, or the development of, ischemic conditions
followed by reperfusion injury can occur with the initial administration of cardioplegia and
with each subsequent delivery of cardioplegia solution [(re)perfusion] for a variety of reasons:
(1) excessive pressure (microvascular injury), (2) low onconicity favoring extravasation of fluids
into the extracellular space, (3) formation of oxygen radicals from over-oxygenation and
oxidative stress, and (4) other mechanisms of reperfusion injury as discussed below. While
delivery of cardioplegia should provide protection, this phase also offers opportunity for
additional injury to occur unintentionally. Any surgeon who has been faced with postoperative
low-cardiac output and who suspects inadequate protection knows this to be true.

d) After release of the cross-clamp and attempts to reanimate the heart: Additional ischemia
may be encountered at the release of the cross-clamp when (1) coronary blood flow is
impaired through kinked grafts or (2) tight anastomoses (3) air emboli are present in the
coronary arteries, (4) ventricular fibrillation occurs at low perfusion pressures, or (5) poor
ventricular performance or dysrhythmias cause hypotension once the heart is converted and
off bypass. Clearly, when the clamp is removed reperfusion injury can occur as blood flow is
restored and oxygen exposure is reestablished after any of the above events. In terms of
surgically induced ischemia, this is intuitively the major time when one thinks of reperfusion
injury occurring.

Understanding the pathophysiologic mechanisms of myocardial injury occurring at these

intervals and applying the principles of myocardial protection will help avoid injury induced by
the very techniques intended to preserve the myocardium. Unlike the cardiologist, the surgical
team has the opportunity to intervene directly to control many facets of ischemia and
reperfusion injury, and to limit at some point the mechanisms causing injury, thereby more
favorably directing the outcome of the postischemic heart. The pathophysiology of ischemia
and reperfusion is described in the following sections. Interventions targeting these various
pathologic processes are discussed under Myocardial Protection Therapy.

Determinants of Ischemic Injury

Ischemia encountered either before arrest or during cardioplegic arrest, is a major cause of
postcardioplegia injury. Ischemia sets the stage for reperfusion injury; without ischemia you
cannot have reperfusion injury. Ischemia is defined as inadequate energy supply to meet the
current demands, that is, an energy supply/demand ratio of less than 1; normally, the O2
supply/demand ratio is >1. Since the adenosine triphosphate (ATP) necessary to support the
energetic needs of the heart is provided predominantly by oxidative phosphorylation, with
scant generation of high-energy phosphates through anaerobic metabolism, the heart is nearly
an obligate aerobic tissue. Therefore, the energy consumption of the heart can be
approximated by the oxygen consumption. The limited availability of anaerobic energy-
generating pathways mandates that O2 be in constant supply. A very brief discussion of the
determinants of O2 supply and demand follows.

Oxygen supply: Ischemia may be caused by either a decrease in O2 supply relative to demands
(supply ischemia) or an increase in O2 demands relative to supply (demand ischemia). O2
supply is determined by the O2 extraction (arterial-coronary venous O2 difference) and
coronary blood flow. Since O2 extraction by the heart is normally ~75%, and its physiologic
limit is approximately 95%, only limited amounts of oxygen may be obtained through
additional extraction. The majority of O2 is therefore supplied by adjustments in coronary
blood flow, which under normal conditions may increase 4- to 5-fold to meet demands. This
O2 is largely present as hemoglobin-bound, with only a minor fraction (~1.5%) dissolved in
plasma. This limited dissolved O2 will become important when we discuss crystalloid
cardioplegia solutions in which oxygen availability is limited to that dissolved in solution.

Oxygen demand: Overall myocardial oxygen demand of the normal working heart is
determined by the work of the entire heart. The majority of this overall O2 consumption is
attributed to utilization by the left ventricle. In a nutshell, the determinants of left ventricular
O2 demand include pressure–volume work or wall stress, heart rate, temperature, inotropic
state, basal metabolism, and ionic homeostatic mechanisms (ATP-dependent pumps) required
to re-equilibrate ionic balance after electromechanical activity. In the immediate postoperative
heart, energy demands may be related to oxidative energy diverted to myocyte repair.

The energy related to maintain ionic equilibrium (~5% of total demand) is an ongoing demand
even when the heart is not contracting since ATP-dependent pumps are constantly adjusting
calcium and sodium influx, as well as potassium efflux. This energy demand of ionic
equilibration is important in determining ongoing oxygen demands in the arrested heart, as
discussed below. How oxygen demands are changed by hypothermia is discussed later.

Severity and Duration of Ischemia in Cardiac Surgery

The potential for myocardial injury is related, in part, to the duration of ischemia. The time to
onset of irreversible ischemic injury will depend on many factors, including the severity of
ischemia, myocardial temperature, ambient energy demands, and collateral blood flow.
However, irreversible injury can become apparent after as little as 30 to 45 minutes of
coronary occlusion in the working myocardium. However, shorter durations of global ischemia
can result in mild to severe systolic and diastolic dysfunction without irreversible tissue
necrosis (left ventricular “stunning” (33)). Therefore, the duration of ischemia, both preceding
CPB and intraoperatively after application of the cross-clamp, has often been used as a
predictor of postoperative myocardial injury in experimental models, and is one of the
underlying considerations in the correlation between total ischemic time (CPB, cardioplegia)
and clinical outcomes. Because the surgical team often does not have control over the
duration of warm ischemia (particularly antecedent ischemia caused by coronary occlusive
disease), other aspects of elective ischemic time are more often targeted for modification to
reduce the severity of ischemia and hence postischemic injury. Therefore, strategies such as
(1) initiating rapid asystole, (2) venting the left ventricle to reduce consequences of chamber
distension resulting in increased wall stress, and (3) imposing cardiac hypothermia all
specifically target the reduction of myocardial energy/oxygen demands, thereby increasing the
limits of “safe” ischemic time during aortic cross-clamping. With these cardioprotective
strategies in place, the duration of aortic clamping that can be safely imposed can be increased
from as little as 15 to 45 minutes to several hours using hypothermic cardioplegia (Fig. 9.2).
Reducing to almost zero the determinants of myocardial work and oxygen demand is the sine
qua non of myocardial protection. In the section on future directions, eliminating ischemic
conditions without additional oxygen-related injury and changing the milieu more gradually
will be discussed.

Figure 9.2. Left ventricular performance measured by in situ Starling curves (A) or end-systolic
pressure–volume relations (conductance catheter) (B) after 45 minutes of normothermic
unprotected global ischemia and reperfusion. C: Left ventricular stroke work index in patients
before (Baseline) or after (pCPB) 15 minutes, or 4 and 24 hours postcardiopulmonary bypass
(CPB). The dashed line between 4 and 24 hours after discontinuation of CPB represents further
decreases in cardiac performance secondary to onset of irreversible injury.(A, adapted from
Rosenkranz ER, Buckberg GD. Myocardial protection during surgical coronary reperfusion. J Am
Coll Cardiol 1983;1:1235–1246, with permission.)

The determinants of surgical ischemic injury can be summarized as:

● The duration and severity of antecedent” and “unprotected” ischemia: Although it has
been recognized for decades that ischemia time (duration) impacts tissue injury and
postischemic salvage and function, an increasing appreciation for the caveat that the
 manner in which the tissue is reperfused significantly alters the outcome. It is now
increasingly understood that reperfusion injury can occur by merely eliminating
ischemic conditions abruptly or in the presence of an inappropriate oxygen gradient. In
the case of the imposition of aortic cross-clamping, there is total elimination of native
flow with the exception of that supplied by noncoronary collateral flow. Global
contractile function may be impaired before necrosis or apoptosis are evident
(“stunning”). It is the reduction of myocardial oxygen demand that protects the heart
in this circumstance.
● The elimination of electromechanical activity during cross-clamping is paramount
because its mechanical activity increases O2 demands even in the presence of
hypothermia (discussed further in the section on Hypothermia).
● Myocardial temperature: Temperature influences metabolism by the “Q10 effect,” in
which the tissue metabolic rate decreases by half for each 10°C decrease in
temperature; conversely, myocardial metabolism doubles when the heart rewarms by
10°C. Hypothermia can buy biologic time during ischemia.
● The nutritional and stress status of the heart: For example, catecholamines increase
the O2 demands by increasing contractility and exerting an O2 wasting effect.
● Presentation of comorbidities such as hyperlipidemia which increase the vulnerability
to ischemia-reperfusion injury for any given duration of ischemia. This effect is
independent of energy demands, but is influenced more by mechanisms exerted at
reperfusion, that is, the capacity to generate O2 radicals, a pro-inflammatory state of
the endothelium and inflammatory cells, and calcium handling by the tissue.

Determinants of Reperfusion Injury

Although reperfusion of the myocardium is the ultimate goal of both surgical and nonsurgical
revascularization, abrupt reperfusion carries with it the potential for extending postischemic
injury to cardiomyocytes, coronary vascular endothelium, and the microvasculature. Indeed,
reperfusion injury associated with rapid changes in circulatory conditions may be responsible
for ~50% (34) or more of the ultimate infarct size, and is an important contributor to
postsurgical mortality and morbidity as well (29). We can define surgical reperfusion injury as
an incremental increase in the pathology that is observed after ischemic injury alone.
Reperfusion injury extends or accelerates damage from that observed during ischemia alone,
and occurs after the onset of reperfusion (i.e., cardioplegia) or reperfusion (i.e., cross-clamp
removal). Viewed another way, reperfusion injury can occur at the instant of the elimination of
ischemic conditions with reestablishment of flow. Hence, reperfusion injury can be viewed as
representing an increment of injury that had begun during ischemia but progresses or
manifests during the reperfusion phase. Other investigators have described de novo injury
starting at reflow or, in a newer construct, at the time of the elimination of ischemia (even if
that is accomplished by a reduction in oxygen demand). Reperfusion injury can be categorized
as either reversible or irreversible. (1) Reversible injury includes temporary contractile
dysfunction (“stunning”) in the absence of morphologic injury or necrosis, or (2) irreversible,
which includes necrosis and apoptosis. Biomarker enzymes such as creatine kinase (CK) or
cardiac troponin T (cTnT) or I (cTnI) released by disrupted cells have been observed to be
elevated during reperfusion rather than during ischemia. In the past, debate has often been
spirited regarding whether reperfusion injury even exists (35,36,37,38) and whether
reperfusion injury kills cells that are salvageable at the end of ischemia. However, a consensus
has developed in both the surgical, cardiology, and cardiovascular research communities that
reperfusion is indeed an important component of postischemic cardiac injury in animal models
(39) and in humans (34,40) and that alterations in reperfusion strategies can mitigate injury.

Among the more important mechanisms in the pathophysiology of myocardial ischemia-

reperfusion injury are (1) the duration and severity of antecedent ischemia, (2) oxygen radicals
generated not only by neutrophils but also by myocytes and vascular endothelium, (3) sodium
(Na+) and Ca2+ influx and loss of normal intracellular Ca2+ homeostasis, (4) neutrophils and
neutrophil-derived products that constitute a generalized inflammatory response to
reperfusion, and (5) mitochondrial dysfunction and opening of the mitochondrial permeability
transition pore (mPTP). Interestingly, the presence of hyperoxic or even relatively hyperoxic
conditions can ramp up or trigger all of these mechanisms. As such, these mechanisms do not
operate independently, nor are they mutually exclusive. Importantly, reperfusion injury is a
malleable process that can be attenuated by utilizing a number of strategies targeting these
various perpetrators of injury. In the operating room during cardiac surgery, and more and
more frequently in the modern era in the interventional lab, various forms of circulatory
support techniques and cardioprotective drugs are employed. The realization that circulatory
support can manipulate reperfusion therapy by altering oxygen supply/demand relationships is
critical for any operator that employs it. Such pumps provide an avenue for both the reduction
of oxygen demand and for the delivery of therapeutic agents via the pump, which can be
administered intravenously or by using cardioplegia as a vehicle. Using cardioplegia offers a
unique opportunity to modify the conditions of reperfusion (hydrodynamics, temperature,
route of delivery) and the composition of the solution (blood, crystalloid, pH, metabolic
substrates, hypocalcemia, oxygen, pharmaceuticals) that target the pathophysiologic
mechanisms of injury. It is now also recognized that the initiation of circulatory support in it
itself offers an opportunity to mitigate injury by paying attention to conditions at the moment
support is begun and ischemic conditions are eliminated.

Reactive Oxygen Species (ROS) and Reactive Nitrogen Species (RNS)

Molecular oxygen is relatively inert by virtue of the shared electron pair in the outer molecular
shell. However, oxygen must be viewed as a pharmaceutical and potentially toxic agent in
cardiovascular medicine and surgery. ROS are radicals or unstable oxygen molecules derived
from oxygen. Important oxygen radicals are superoxide anion (−O2) and hydroxyl anion
(•OH).OH). However, other oxygen-related molecules that are not technically radicals are
considered biologically reactive nonetheless. An example of this would be hydrogen peroxide
(H2O2). In addition, hypochlorous acid (HOCl) and chloramines derived from neutrophils exert
potent oxidative injury to biologic tissues.

Free-radical molecules that are derived from nitrogen are termed RNS and include nitric oxide
(NO•), peroxynitrite (ONOO–), nitrogen dioxide (NO2•) and nitrosyl hydride (HNO). In vascular
endothelium, NO• is constitutively generated by endothelial nitric oxide synthase (eNOS or
NOS-3). eNOS activity is dependent on Ca2+-calmodulin, flavin, NADPH, and
tetrahydrobiopterin (BH4). NO• is generated by an electron from the guanadino nitrogen of L-
arginine to incorporate oxygen to generate NO • and the byproduct L-citrulline. NO• has an
extremely short half-life (seconds) in vivo, and binds tightly to hemoglobin. OONO− is formed
by a nonenzymatic biradical reaction between NO• and −O2. NO• has been linked to biologic
signaling in cardioprotection, particularly anti-inflammatory effects, endothelial function,
vasorelaxation (hence its early name as endothelial-derived relaxing factor, EDRF),
neurotransmission, immune regulation and defense mechanisms.

NO• homeostasis is key in normal regulation of blood flow and responses to stress, and in
preventing endothelial dysfunction. NO• homeostasis is a balance between NO • generation,
bioavailability, and degradation or quenching by oxidants such as −O2. NO • has pleiotropic
effects in the cardiovascular system by paracrine effects: (1) NO • is a physiologic regulator of
blood pressure and flow, and counteracts local and systemic vasoconstrictors such as
endothelin (ET-1); (2) it inhibits neutrophil adhesion to vascular endothelium; (3) it inhibits
platelet aggregation; (4) it reduces myocardial oxygen consumption (41); (5) it has positive
lusitropic and inotropic effects; (6) it scavenges −O2 and therefore is directly antioxidant; and
(7) it is anti-proliferative. The effects on oxygen consumption and inotropy are somewhat
controversial (42,43). The role of NO• in ischemia-reperfusion injury and cardioprotection is
discussed in more detail under the section on endothelial dysfunction and NO • donors.

Reactive Oxygen Species

The oxygen radical is highly reactive with a broad range of biologic substances including sugars,
amino acids, phospholipids, and DNA. With such a large catalog of biologic targets, ROS and
their metabolites are potentially toxic to cells, and at the same time they are critical in the host
defense mechanism and bactericidal activities of neutrophils and phagocytes (44) and
participate in cardioprotective signaling. In this dualistic role of ROS as good guy–bad guy, it is
proposed that low-level production of ROS, particularly from mitochondria, acts as a signal for
cardioprotective conditioning responses (preconditioning, postconditioning) and stimulation of
cardioprotective kinase pathways, while large-scale generation of ROS from inflammatory cells
or dysfunctional mitochondria is deleterious. In the ischemic-reperfused myocardium, ROS-
induced injury includes peroxidation of lipid components of cellular membranes leading to
damage to sarcolemmal cell membrane and other membranous organelles such as
mitochondria and sarcoplasmic reticulum (45). ROS also cause the mPTP to open when
generated in an environment of Ca2+ accumulation and normalized pH (conditions achieved
during reperfusion), which leads to a collapse of the membrane potential, loss of metabolic
capacity, and release of pro-apoptotic factors. This is discussed in more detail below under the
role of mitochondria in reperfusion injury. The vascular endothelium is particularly vulnerable
to ROS-mediated injury, which manifests as impairment of vascular endothelial function
through attenuated production of vasoactive and anti-inflammatory autacoids such as
adenosine (46,47,48) and NO• (49), and increased production of pro-inflammatory and
vasoconstrictive molecules. These injury mechanisms contribute significantly to postischemic
dysfunction, dysrhythmias, and morphologic injury such as necrosis and apoptosis.

Sources of Oxygen Radicals

Both enzymatic and nonenzymatic reactions produce oxygen radicals. The major reactions and
their sources are shown in Figure 9.3. ROS are generated by a number of sources, including
activated neutrophils, cardiac myocytes, mitochondria, and the vascular endothelium.
Figure 9.3. The species and sources of reactive oxygen species (ROS) generated by neutrophils,
cardiomyocytes and their mitochondria, and vascular endothelium. SOD, superoxide
dismutase; CAT, catalase; GP, glutathione peroxidase; MPO, myeloperoxidase; HOCl,
hypochlorous acid; e−1, electron; NOX, NAD(P)H oxidase.

Neutrophils: Activated neutrophils represent the major source of ROS in the heart under
stress, including −O2, •OH, and HOCl species. Neutrophils are activated by various chemotactic
factors stimulated by either CPB or ischemia-reperfusion, including the complement
anaphylatoxins C3a and C5a, f-Met-Leu-Phe, tumor necrosis factor alpha (TNFα), interleukin-8
(IL-8), and platelet-activating factor. These factors trigger a “respiratory burst” of ROS. CPB
(independent of ischemia) activates the complement cascade (50,51), with the subsequent
activation and adherence of neutrophils (52) to the vascular endothelium primarily, which
subsequently triggers the release of cytotoxic products, all of which contribute to tissue injury
in the ischemic-reperfused myocardium (52,53). During the respiratory burst that occurs
seconds after stimulation by cytokines or ischemia-reperfusion, the −O2 in the neutrophil is
produced by the membrane-associated NADP oxidase, which transfers an electron to
molecular oxygen (Fig. 9.3, reviewed in Sheppard et al. (54)). Production of H2O2 by
neutrophils or the reduction of −O2 to H2O2 by superoxide dismutase (SOD) provides
substrate for generation of •OH, or the oxidant HOCl by myeloperoxidase. HOCl reacts with low
molecular weight amines to give rise to lipophilic chloramines, which can promote membrane
lipid peroxidation.

Cardiomyocytes: Xanthine oxidase (XO) is a major source of ROS from cardiomyocytes.

Normally, XO exists mostly in the dehydrogenase form and uses nicotinamide adenine
dinucleotide (NAD) as an electron acceptor. Xanthine dehydrogenase is converted to the
oxidant-generating enzyme XO during ischemia. Both hypoxanthine and xanthine accumulate
during ischemia. Hypoxanthine is converted to xanthine by XO, which is in turn converted to
uric acid by the same enzyme, which forms −O2. During ischemia, however, two events favor
the xanthine oxidase reaction: (1) hypoxanthine accumulates as a result of the sequential
catabolism of purine high-energy phosphates (ATP, ADP, AMP) and deamination of adenosine,
and (2) the dehydrogenase form of the enzyme is converted to the superoxide-producing
oxidase form by a protease activated by increased levels of intracellular calcium (55). XO is
reported to be localized to the vascular endothelium in substantial quantities (56). However,
the importance of XO to the pathologic generation of ROS in humans remains controversial
(57).

Mitochondria: Mitochondria are a major noninflammatory site of ROS generation (58). ROS are
generated at complex I (NADH/ubiquinone oxidoreductase) and complex III
(ubiquinol/cytochrome c oxidoreductase). The latter site catalyzes the conversion of O2 to
−O2• by a single-electron transfer. −O2 can be converted in the mitochondrial matrix to H2O2
by matrix superoxide dismutase. Since mitochondria do not have catalase, the H2O2 can be
degraded by nonenzymatic reaction with glutathione (GSH) by glutathione peroxidase or by
enzymatic reaction with thioredoxin reductase. The ROS species target cysteine residues on
the protein moieties of the mitochondrial membrane and polyunsaturated fatty acids, and
cause intramolecular cross-linkages and protein aggregates. The •OH causes peroxidation of
membrane lipids. ROS induce Ca2+ release from the mitochondria, which may cause an
imbalance in Ca2+ handling and stimulation of Ca2+-dependent proteases, nucleases and
phosphatases leading to functional and morphologic damage to the cell.

Vascular endothelium: Superoxide anion is produced by vascular endothelium by XO as

described above for cardiomyocytes (Fig. 9.3). A more important source of −O2 from the
vascular endothelium may be the NAD(P)H (reduced NAD phosphate) oxidase system. This
enzyme is a membrane-bound, flavin-containing NADH/NADPH-dependent oxidase present in
endothelial and vascular smooth-muscle cells. The molecular structure, function, and
regulation are somewhat similar to those of the neutrophil enzyme system, requiring assembly
of cytosolic components onto the membrane components before −O2 generation can occur.
However, −O2 production in the presence of NO • produced by endothelium favors the
biradical neutralization and elimination of NO • and the formation of peroxynitrite (ONOO–).
The elimination of NO• impairs not only vascular relaxation and autoregulatory control of
postischemic myocardial perfusion, but also fails to inhibit neutrophil-mediated postischemic
inflammatory responses.

Other sources of ROS include the oxidation of catecholamines, metabolism of arachidonic acid
to peroxy compounds, and generation of •OH by the cyclooxygenase and lipoxygenase
pathways, and by the metal-catalyzed Haber–Weiss reactions. The latter series of reactions
may be important in the setting of CPB because they can be recruited by neutrophil activation
during CPB (59). In addition, iron-containing products released by hemolysis may facilitate this
reaction.
When are Oxygen Radicals Generated?

Myocardial ischemia favors the generation of ROS. However, the myocardium has a cadre of
the endogenous antioxidants superoxide dismutase, catalase, reduced glutathione (GSH),
glutathione peroxidase, glutathione reductase. Normally, upregulation of these antioxidants
and phase 2 enzymes reduces oxidant-mediated injury (60), but they are depleted or can be
overwhelmed during ischemia, thereby attenuating the natural defense mechanisms of the
heart (61). Therefore, ischemia creates the biochemical setting for oxyradical production and
tissue vulnerability to ROS-mediated damage. However, the primary substrate, oxygen, is in
limited supply during ischemia, and is not readily available until reperfusion. Oxygen radicals
are not seen in great abundance until the onset of reperfusion. A clinical study by Prasad et al.
(62) show that oxygen radicals are released during and 24 hours after CPB in CABG patients.
Indirect evidence to support this reperfusion oxidative burst comes from a wealth of data in
which inhibitors or scavengers of oxygen radicals, or anti-neutrophil agents, exerted beneficial
effects when administered during reperfusion.

In the surgical setting, the myocardium may be vulnerable to ROS-induced injury at several
points when oxygen is introduced into the myocardium: (1) the initial delivery of cardioplegia
into the heart with antecedent unprotected ischemia, or initial delivery of oxygenated
cardioplegia to a newly revascularized myocardial segment; (2) intermittent reinfusions of
cardioplegia if a multidose strategy is being used; and (3) release of the aortic cross-clamp to
initiate reperfusion after the cardioplegia phase. The quantity of ROS generated during cardiac
surgery depends on the severity of the preceding ischemic injury, activation and recruitment of
neutrophils, the level of oxygen in cardioplegic solutions or blood oxygenated by the
extracorporeal circuit, and the status of endogenous scavengers and inhibitors potentially
depleted during ischemia. Blood cardioplegia theoretically provides a greater oxygen substrate
for oxygen radical production than do crystalloid solutions. However, plasma-dissolved oxygen
may be the pool from which oxygen is drawn for the generation of oxygen radicals, and the
oxygen content in this compartment may not differ dramatically between blood and crystalloid
solutions. The oxygen in blood cardioplegia may be reduced with “normoxic” or even
“deoxygenated” blood cardioplegia (63) or during reperfusion (64,65). In addition, blood
cardioplegia may offer the advantage of endogenous antioxidants (superoxide dismutase,
catalase, glutathione, and albumin) that are not present in crystalloid cardioplegia solutions.

Intracellular Na+ and Ca2+ Dyshomeostasis

Under normal physiologic conditions, intracellular Ca2+ concentrations are maintained at low
concentrations (<200 nmol) against a 5,000-fold trans-sarcolemmal gradient by voltage-gated
Ca2+ channels that remain closed until activation during phase 2 of the action potential. The
regulation of intracellular and intramitochondrial Ca2+ is tightly linked to regulation of
intracellular Na+. Intracellular Na+ concentration is normally controlled by the ATP-dependent
Na+/K+ pump which moves Na+ outward and K+ inward against their respective concentration
gradients, and the energy-independent Na+/H+ exchanger which normally removes one H+
from the cytosol in exchange for one Na+ moving into the cytosol. Further Na+ accumulation
mainly occurs through the Na+/Ca2+ exchanger which normally operates in the forward
direction during diastole to extrude intracellular Ca2+ in exchange for influx of Na+ into the
cytosol.

In ischemic-reperfusion myocardium, Ca2+ influx occurs through multiple pathways: (1)

diffusion through overt membrane disruptions or opened Ca2+ channels, fueled by the
differential concentration gradient; (2) reversal or inhibition of the Na+/Ca2+ exchanger driven
by the intracellular accumulation of Na+, (3) attenuated cyclical Ca2+ uptake and release from
the sarcoplasmic reticulum and mitochondria, and (4) facilitation of entry via α-adrenoreceptor
activation and cyclic adenosine monophosphate (AMP)-dependent processes, which increase
during postischemic catecholamine release.

The accumulation of intracellular Na+ and Ca2+ during ischemia-reperfusion are linked events
leading to Ca2+ overload which stimulates Ca2+-dependent proteases that damage
membranes and ion transport mechanisms at the moment of re-exposure to molecular
oxygen, opening of the mPTP, depletion of high-energy phosphate stores, the development of
local or global contracture (“stone heart”), and triggers the transition to cell death through
both apoptosis and necrosis. The balanced Ca2+ transport systems are potentially injured by
exposure to oxygen radicals, cytokines, and activated complement during CPB and reperfusion.
Na+ and Ca2+ dyshomeostasis occurs to differing degrees during ischemia and during
reperfusion, as described below.

Ischemia

During ischemia, glycolysis generates an abundance of H+ and only a limited amount of ATP (2
moles/mole glucose) compared to oxidative phosphorylation (36 moles ATP/mole glucose).
The accumulation of intracellular H+ stimulates activity of the Na+/H+ exchanger (NHE1) in
favor of H+ efflux and Na+ influx; at this time, the reduced availability of ATP decreases Na+/K+
ATPase activity. Therefore, both the attenuated Na+/K+ pump activity and increased activity of
the Na+/H+ exchanger result in intracellular Na+ accumulation. Additional Na+ may enter the
cell through sarcolemmal nonactivating Na+ channels, for example, the “window currents”
generated during ischemia-induced cellular depolarization. However, some evidence suggests
that the accumulation of H+ in both the extracellular and intracellular spaces attenuates the
H+ gradient, thereby reducing but not eliminating the participation of the Na+/H+ exchange
mechanism in intracellular Na+ accumulation during ischemia (66). However, the accumulation
of Na+ in the cytosol during ischemia would favor reversal of the energy-independent
Na+/Ca2+ exchanger in favor of extruding Na+ in exchange for Ca2+ influx, thereby favoring its
accumulation in the cytosol. Entry of Ca2+ into the mitochondria through the Ca2+ uniporter
depends on the Ca2+ electrochemical gradient and the presence of a mitochondrial
transmembrane potential. Ca2+ enters the mitochondria during early ischemia, but further
increases and overload are prevented by dissipation of the mitochondrial transmembrane
potential (67). In addition, the unregulated activation of Ca2+-dependent protease enzymes
such as calpains is inhibited by intracellular acidosis, and is delayed until intracellular pH is
renormalized during reperfusion (68). Therefore, Ca2+ accumulation and Ca2+-mediated
damage occurs to a greater extent during reperfusion (69,70,71) as discussed below.
Reperfusion

Na+ and Ca2+ dyshomeostasis during reperfusion is a significant factor contributing to

postischemic myocardial injury and functional impairment (72,73). The ion dyshomeostasis
described above sets the stage for explosive reperfusion/reoxygenation damage. It is
increasingly recognized that abrupt re-exposure to oxygen itself is the trigger for damage.
Excessive gradients of perfusate-to-tissue oxygen tension can be mitigated. In addition, the
events leading to abnormal Ca2+ handling which determine cell survival versus cell death are
as follows:

● Rapid renormalization of intracellular pH: Reperfusion washes out lactate and re-
establishes the intracellular–extracellular H+ gradient (extracellular < intracellular)
which allows the forward motion of the Na+/H+ exchanger favoring removal of H+
(renormalization of intracellular pH) and subsequent intracellular accumulation of Na+;
other factors contributing to normalization of intracellular pH are activation of
lactate/H+ co-transport mechanism and activation of the NaHCO3 co-transporter.
● Renormalization of intracellular Na+: During the initial phase of reperfusion, ATP is
regenerated, and if the Na+/K+ ATPase is not damaged by proteolytic actions of
calpain enzymes, intracellular Na+ levels are renormalized and the linked Na+/H+
exchange—Na+/Ca2+ exchange systems will not be activated. Otherwise, if the Na+/K+
ATPase pump is damaged, intracellular Na+ accumulation will engage the Na+/Ca2+
exchange mechanism with net intracellular accumulation of Ca2+.
● Ca2+ uptake and release by the sarcoplasmic reticulum (SR): The SR is the main Ca2+
storage site of the cardiomyocytes. Ca2+ in the micro-domain proximate to the closely
co-localized SR and interfibrillar mitochondria is cyclically taken up by the SR (lowering
of cytosolic concentration) during diastole, and Ca2+ is released into that micro-
domain by the ryanodine receptors (RyR) during systole. During both unprotected and
protected ischemia, sequestration of Ca2+ into the sarcoplasmic reticulum competes
for a limited amount of ATP. This paradox creates a cycle whereby a high intracellular
Ca2+ level binds high-energy phosphates to further limit ATP availability, which then
attenuates the ability of the SR to remove the Ca2+ actively from the cytoplasm,
resulting in a further increase in cytosolic Ca2+. If other Ca2+ handling mechanisms are
not engaged, the release of Ca2+ from the SR RyR may lead to overload. This activity of
the SR mechanism may be related to the severity of ischemia, being less so during
short periods of ischemia.

So, what events in the balance of Ca2+ regulation determines if a cell lives or dies during
reperfusion? Some evidence suggests that the timing of normalization of pH and Ca2+ handling
mechanisms are important. If pH normalization occurs before Ca2+ handling mechanisms are
restored, then Ca2+ overload, in conjunction with ROS generation and a normalized
intracellular pH environment, will trigger opening of the mPTP causing cell death. However, if
Ca2+ handling mechanisms are restored before pH is normalized, for example, by maintaining
tissue acidosis during early reperfusion, then Ca2+ overload can be avoided. Interestingly,
venous blood-based perfusate has been preliminarily observed to mitigate both over-
oxygenation during early reperfusion and delayed normalization of pH, thereby creating a
“permissive acidemia.”
The Inflammatory and Innate Immunity Responses to Ischemia-Reperfusion and
Cardiopulmonary Bypass

Injury mechanisms of ischemia-reperfusion and CPB involve inflammatory and innate immune
responses. Cardiomyocytes generate pro-inflammatory cytokines, chemokines and adhesion
molecules that direct the emigration of inflammatory cells into ischemic-reperfused tissue
regions and amplify the inflammatory response. The following sections will describe the
inflammatory process involving neutrophils, pro-inflammatory cytokines, and complement,
and the innate immune response involving toll-like receptors (TLRs).

Neutrophils and Neutrophil-Mediated Injury

Neutrophils play a critical role in the pathogenesis of myocardial reperfusion injury (74,75,76).
The inflammatory component of reperfusion injury requires an interaction between activated
neutrophils and vascular endothelium during the early moments of reperfusion, which is
prerequisite to activation of the full local inflammatory cascade. A well-orchestrated sequence
of events mediates the interaction between adhesion molecules on neutrophils and
endothelium. These adhesion molecules are categorized into three families:

● The selectins (P-selectin, L-selectin, E-selectin) are glycoproteins involved in the

interactions between neutrophils and the vascular endothelium in the early moments
of reperfusion (77,78,79). P-selectin is stored in Weibel–Palade bodies in endothelial
cells, and its surface expression is triggered by pro-inflammatory mediators such as
oxygen radicals, thrombin, complement components, cytokines, histamine, and H2O2
released during ischemia-reperfusion and CPB. IL-8 released from hypoxic endothelial
cells also increases the adhesiveness of endothelial cells for neutrophils (80). P-selectin
surface expression peaks after 10 to 20 minutes of reperfusion, and P-selectin is
subsequently shed to soluble fragments in blood. In contrast to P-selectin, L-selectin is
constitutively expressed on the surface of neutrophils and may be the counterligand
for P-selectin during early reperfusion (81). Loose adherence and “rolling” of
neutrophils on endothelium involves interaction with a glycoprotein sialyl Lewisx or a
high-affinity glycoprotein ligand termed P-selectin glycoprotein ligand-1 (PSGL-1) (82).
The third member of the selectin family, E-selectin, is expressed on the surface of
endothelial cells. It is expressed later in reperfusion (4–6 hours) and may therefore be
involved in later events, such as infarct extension.
● The β2-integrins (CD11/CD18 complex) are a family of glycoproteins located on
external membranes of neutrophils. There are three distinct α-chains (CD11a, CD11b,
CD11c) and a common β-subunit. Surface expression of perhaps the major complex
CD11b/CD18 is triggered by a number of pro-inflammatory mediators, including
circulating and endothelium-derived (acylated) platelet-activating factor. After initial
tethering of neutrophils by endothelial P-selectin, firm adherence is mediated by
interaction with CD11b/CD18 and its counterligand ICAM-1 on the endothelium.
● The third family of adhesion molecules is the immunoglobulin superfamily, including
ICAM-1 (intercellular adhesion molecule-1), VCAM-1 (vascular cell adhesion molecule-
1), and PECAM-1 (platelet–endothelial cell adhesion molecule-1). ICAM-1, the
counterligand for CD11/CD18 on neutrophils, is constitutively expressed on
 endothelial cells and is upregulated by cytokines 2 to 4 hours after myocardial
ischemia-reperfusion (83), coincident with the upregulation of CD11/CD18.
Subsequently, a high-affinity bond between the neutrophil and endothelial cell occurs
via integrins (CD11/CD18) on neutrophils and ICAM-1 and ICAM-2 on endothelial cells.
Once the neutrophil has firmly attached to the endothelial surface, it migrates through
the endothelial cell junctions to the myocardium, partly mediated by IL-8. The
shedding of L-selectin from the neutrophil and the presence of PECAM on the
endothelium may be required for neutrophil transendothelial migration. The
prerequisite nature of this interaction between neutrophils and adhesion molecules
makes anti-neutrophil and anti-adhesion therapy a potentially effective strategy, as
these agents intervene at proximal points in the cascade (discussed below).

Evidence supporting the rapid accumulation of neutrophils within ischemic-reperfused

myocardium following the neutrophil–endothelial cell interactions was appreciated not only by
our laboratory, but also by Sommers and Jennings (84). We (85,86,87,88,89) and others (90)
have demonstrated an association between the extent of postischemic injury and neutrophil
accumulation within ischemic-reperfused myocardium as well as a reduction of postischemic
injury associated with neutrophil suppression. As shown in Figure 9.4, Lefer et al. (91)
demonstrated that neutrophil adherence to coronary artery endothelium occurs in advance of
accumulation in extravascular sites, representing the lag time involved in transmigration of
intravascular neutrophils. In addition, neutrophil accumulation occurs in advance of the
development of necrosis, consistent with active neutrophil participation in cell demise. Brix-
Christensen et al. (92) reported that neutrophils accumulated in the hearts and other organs of
neonatal pigs placed on CPB. Therefore, neutrophil accumulation occurs in both postischemic
and post-CPB hearts.

Figure 9.4. The postischemic time course of endothelial injury, contractile dysfunction,
neutrophil (PMN) adherence and accumulation in tissue, necrosis, and apoptosis. These events
are indicated on the y-axis as a percent completion of that event (0% not started, 100% is
complete). Note that PMN adherence parallels that of endothelial dysfunction representing
PMN-mediated damage. Note also that accumulation in tissue lags behind adherence
representing emigration time. (Data obtained, in part, from Zhao ZQ, Nakamura M, Wang NP,
et al. Dynamic progression of contractile and endothelial dysfunction and infarct extension in
the late phase of reperfusion. J Surg Res 2000;94:133–144.) Apoptosis is a slower process than
necrosis as observed by Zhao et al. (565).(Adapted in part from Lefer AM, Tsao PS, Lefer DJ, et
al. Role of endothelial dysfunction in the pathogenesis of reperfusion injury after myocardial
ischemia. FASEB J 1991;5:2029–2034.)

The importance of neutrophils in postbypass reperfusion injury has been well-established in

both animal and human models. The myocardium is exposed to neutrophils during the delivery
of intermittent or continuous blood cardioplegia, or during reperfusion (after cross-clamp
removal) following either crystalloid or blood cardioplegia. Schwartz et al. (93) showed that
CPB circuitry directly “primes” neutrophils by depositing C3b on extracorporeal tubing, leading
to neutrophil activation and sequestration. Further evidence for the activation of neutrophils
following CPB was suggested by the finding of increased neutrophil surface expression of
CD11b/CD18 and decreased surface expression of L-selectin on cells circulated in simulated
CPB circuits (94,95). Finn et al. (96) noted in patients undergoing CPB that a rise in circulating
IL-8 begins at reperfusion with rewarming of the patient after systemic hypothermia, and
closely correlates with the rise in neutrophil count and circulating elastase. Similarly, Schwartz
et al. (93) and Gessler et al. (97) found higher levels of the pro-inflammatory mediators soluble
IL-6 and IL-8 6 hours after bypass than before bypass.

The time course for the upregulation of neutrophils and various adhesion molecules is equally
important in designing treatment strategies for patients undergoing extracorporeal bypass. In
a clinical study, Galiñanes et al. (98) investigated the perioperative temporal changes in the
expression of various neutrophil surface adhesion molecules, concluding that downregulation
of PECAM-1 is an early indicator of neutrophil activation and may therefore represent a target
for therapy aimed at reducing the inflammatory response associated with CPB.

The role of temperature in the activation of neutrophils or other inflammatory mediators

following CPB in humans has been investigated. Hypothermia has been reported to attenuate
expression of adhesion molecules in some models of tissue injury (99,100). Some investigators
have reported minimal or no difference in neutrophil activation and expression of neutrophil
adhesion molecules between normothermic (34°C–37°C) and hypothermic (28°C–30°C) CPB
(101,102), whereas others note an attenuation in the expression of IL-8 and neutrophil
elastase activity during normothermic CPBsurgery (103). Hypothermia has been shown to
delay, but not prevent, the increased expression of adhesion molecules such as CD11b and
CD11c (101). When measured 4 hours after bypass, the levels of IL-1α, soluble ICAM-1 (sICAM-
1), and elastase are higher after normothermic CPB than after hypothermic CPB (104,105). In
conclusion, further research is required to determine the most appropriate temperature of
cardioplegic administration to minimize the deleterious consequences of neutrophil-mediated
damage. Interventions designed to attenuate neutrophil-related myocardial injury following
CPB will be discussed later.
Complement in CPB and Myocardial Reperfusion Injury

The complement cascade, particularly the alternative pathway stimulated by contact

activation, is activated independently by myocardial ischemia-reperfusion and CPB, and
contributes importantly to the pathologic sequelae of CPB (106). Complement fragments such
as the anaphylatoxins C3a and C5a are generated and released both systemically and locally by
the ischemic-reperfused myocardium (107). The activation and migration of neutrophils to
ischemic-reperfused myocardium requires stimulation by chemotactic factors, most notably
complement fragments (C3a and C5a), eicosanoid products such as leukotriene B4, and
platelet-activating factor (108,109). Complement fragments (C5a) can be found in lymph from
ischemic myocardium (110), are temporally associated with reperfusion (111), and correlate
with increased neutrophil accumulation in reperfused myocardium. Furthermore, C5a directly
stimulates neutrophil −O2 production and enhances adherence to coronary artery
endothelium (112), whereas C3 activates monocyte adherence (113). The chemoattractant and
chemotactic properties of C5a cause progressive generalized vascular leukosequestration (114)
during CPB. The interaction between blood and extracorporeal surfaces also stimulates
deposition of C3b (115), which in turn amplifies the complement alternative pathway to
release C3a and C5a, and subsequently form the membrane attack complex (C5b 9) (115).
Accordingly, plasma levels of the anaphylatoxin C5a increase in patients undergoing CPB
(115,116), which activates neutrophils to generate −O2 and promotes adhesion to endothelial
cells, stimulates degranulation of mast cells, stimulates vascular smooth-muscle contraction,
and increases vascular permeability. Complement activation constitutes an early-phase
response to CPB, which then progresses to a second, slower response phase in which de novo
synthesis of proteins (E-selectin, ICAM, IL-8) and expression of adhesion molecules on cell
surfaces occurs. The magnitude of the complement response has been linked to the duration
of CPB and the type of circuitry used (116). The combined stimuli of myocardial ischemia-
reperfusion and CPB may synergize to exacerbate postischemic neutrophil-mediated
inflammatory responses in patients undergoing CPB for myocardial revascularization.

Actions of Toll-like Receptors

Recent evidence has shown that activation of TLRs is involved in the pathogenesis of
reperfusion injury (117,118). TLRs are transmembrane receptors expressed on endothelial
cells, cardiomyocytes, platelets, monocytes, and neutrophils. Cardiomyocytes participate in
the inflammatory and innate immune response to ischemia-reperfusion by releasing cytokines
(TNFα, IL-1β, IL-6, IL-8, IFN-γ), chemokines, and by expressing cell surface adhesion molecules.
TLRs are pattern recognition receptors (PRRs) that recognize pathogen-associated molecular
patterns (PAMPs) such as LPS or viral RNA on foreign pathogens, and sense endogenous
danger molecules and signals from injured cells. TLRs also recognize and are activated by
endogenous ligands such as danger-associated molecular patterns (DAMPS) (e.g., heat shock
proteins, HSP70) induced in cardiac tissue during stress (119) and cardiac surgery (120,121).
Inducible HSP70 has been found in plasma from CABG patients immediately after surgery
(122). ROS may also stimulate TLRs. Ten to eleven human TLRs have been identified; (123,124)
TLR receptor-1, -2, -4, -5, and -6 are located on the cell surface, with TLR-2 and TLR-4 being the
most involved in ischemia-reperfusion of the group of TLRs; TLR-2 works with other TLRs in co-
receptor cooperativity to recognize diacyl- and triacyl-lipoproteins, while TLR4 recognizes LPS
from gram-negative bacteria and DAMPs such as HSP70. Neutrophils constitutively express all
TRLs except TLR3.

The stimulation of TLRs during ischemia-reperfusion and cardiac surgery can lead to cell injury.
Dybdahl et al. (122) reported the release of HSP70 into the circulation of CABG patients and a
concomitant increase in monocyte TLR-2 and TLR-4 one day after surgery. Human monocytes
exposed to HSP70 released IL-6 and TNFα. Stimulation of TLRs by interaction with DAMPs
during ischemia-reperfusion is transduced by numerous molecular signaling pathways, which
ultimately results in the activation and translocation of NF-κB to the nucleus, where it
stimulates gene expression and release of cytokines and innate inflammatory factors. ROS
generated during reperfusion also stimulate TLR-dependent (most likely TLR-4) generation of
NF-κB through the PI3-K/Akt pathway, which suggests cross talk between the TLR and
reperfusion injury survival kinase (RISK) pathways. Inhibitors of TLRs, intermediary pathways,
or NF-κB will attenuate ischemia-reperfusion injury.

The Mitochondrial Permeability Transition Pore as a Determinant of Cell Death

Mitochondria are best known for being the power house of the cell, where ATP is generated
through oxidative phosphorylation. However, the mitochondria have been shown to function
as a “molecular switch” that governs cardiomyocyte viability; indeed it is a critical fulcrum
point in the transition from reversible to irreversible injury. Mitochondria can direct the cell to
pursue either an apoptotic or a necrotic pathway to cell death. The key event of this molecular
life/death switch is opening of the mPTP. This pore is a transmitochondrial membrane voltage-
dependent and Ca2+-dependent high-conductance channel located in the inner mitochondrial
membrane that permits water and solutes with a mass up to 1.5 kDa to enter the
mitochondrion. Under normal physiologic conditions, the mPTP is in a closed state. It opens in
response to high levels of intracellular ROS, Ca2+ and inorganic phosphate when mitochondria
matrix and intracellular pH are in the normal range. The mPTP remains closed during ischemia
because intracellular acidosis and the intracellular levels of Mg2+ and ADP counteract the
opening triggered by ROS and Ca2+. However, the mPTP opens during the early moments of
reperfusion (125) when there is an accumulation of ROS, Ca2+ and inorganic phosphate, and
when there is a rapid renormalization of intracellular and mitochondrial matrix pH (126). The
opening of the mPTP increases the permeability of mitochondria to water and small solutes,
causing swelling of the matrix and collapse of the mitochondrial transmembrane potential that
drives oxidative phosphorylation and ATP generation. mPTP opening also stimulates the
release of the pro-apoptotic factors cytochrome c and apoptosis-inducing factor (AIF). The
switch function comes into play based on the availability of intracellular ATP; if ATP is present
then the cell pursues an apoptotic pathway because apoptosis requres ATP. If ATP levels are
depleted then the cell pursues a necrotic pathway. With prolonged antecedent ischemia mPTP
is irreversible, but milder or short periods of ischemia trigger transient reversible opening
(127).
CHAPTER 9

SURGICAL MYOCARDIAL PROTECTION

Jakob Vinten-Johansen and Neil J. Thomas

INTRODUCTION

The science of myocardial protection had its beginnings in the 1950s, and has enjoyed a robust
and often controversial development since. Great strides were made in the 1980s and 1990s
which shaped the cornerstones of the cardioplegia strategy that we largely use today
(crystalloid or blood, warm, tepid or cold, continuous or intermittent, antegrade or
retrograde). Mortality has substantially decreased over the past decades as a result of applying
cornerstone as well as novel cardioprotective strategies and new technologies.

However, the patient of the 1980s and 1990s during which cardioplegia was developed has
morphed considerably in the 2000s. Active lifestyles have changed to more sedentary lifestyles
beginning in the early school years with less emphasis placed on physical activity, and more
distractions from active lifestyles imposed by the digital-age gadgets. This, coupled with the
availability of fast food with high caloric, high fat, and high sodium content, has conspired to
radically change our body habitus and biology; indeed, diabetes, obesity,
hypercholesterolemia, and the metabolic syndrome are more prevalent today. These more
prevalent comorbidities have impacted the efficacy of modern surgical therapies and the
outcomes (1) of myocardial protection strategies. Experimental animal models in which
effective cardioplegia strategies were developed decades ago, and observed to be effective in
the healthier patients with normal myocardial tissue substrate preoperatively in which they
were initially tested clinically seems to apply in fewer and fewer patients today. The “one
formula-fits-all” strategies of the past must be rethought in the current era. With patients
living longer as well, advancing age and the more senescent myocardium can display altered
responsiveness to some therapeutics. So we may ask, is myocardial protection developed in
the 1980s and 1990s being used as effectively on the patient of the 2010s? And, furthermore,
the yardstick by which our results are measured is changing in the world of health care today,
where a 30-day surgical mortality rate may no longer be the gold standard. Our results must
now be viewed in terms of longer-term, event-free survival and with measure of long-term
myocardial performance after an index event or surgery.

The concepts of cardioprotective strategies have changed significantly as our understanding of

the pathophysiologic determinants of ischemia, and particularly the complex mechanisms
involved in reperfusion, have expanded. Early empiric, largely physiologic observations have
given way to complex biochemical and molecular interactions, which in some cases have
reinforced physiologic intuition and in other cases have redirected our thinking, but has
nonetheless provided a scientific underpinning to a more molecular approach in developing
target-specific strategies of myocardial protection. This chapter will discuss the pathobiology
of ischemia-reperfusion injury, provide the basis by which the severity and duration of evolving
ischemic conditions set the stage for reperfusion and oxygen-related injury and its impact on
affected tissues as a basis for understanding cardioprotective strategies to attenuate injury. It
will also discuss new, emerging concepts of myocardial protection that have developed out of
a better understanding of the pathobiology of ischemia-reperfusion injury, and a growing
appreciation that the preoperative condition of the tissue may be equally, if not more
important, than any intraoperative strategy employed in previously normal tissue. It will also
discuss the all-elusive “golden ring” of arresting the heart in a nondepolarized state.

The goal of this chapter is to provide the reader with scientific and practical knowledge of
myocardial protection that can be applied to design appropriate strategies to meet the need of
specific patients, their unique demographics and clinical scenarios, and to provide flexibility in
the approach based on both preoperative condition and intraoperative events. It will also
provide some guided reading to enhance the background of relevant topics

HISTORICAL PERSPECTIVES ON MYOCARDIAL PROTECTION AND REPERFUSION INJURY

The need to protect the myocardium from intraoperative and postoperative damage was
realized before the development of cardiopulmonary bypass (CPB). Temporary inflow
occlusion techniques to limit blood flow to the heart permitted the repair of simple
intracardiac defects, and in many cases the occlusion time could be increased by adding
moderate systemic hypothermia (30°C) (2). After the development of CPB by Gibbon and
colleagues in 1954 (3,4), profound systemic hypothermia could be used safely to provide a
hemodynamic safety net, and to provide protection of the heart and other body organs by
permitting longer periods of ischemic cardiac arrest for the repair of more complicated lesions.

The development of extracorporeal circulation techniques allowed cardiac procedures to

proceed without threat of hemodynamic collapse. However, performing complicated
procedures on the heart was made difficult by its continuous motion and blood in the field.
Ischemia was known to achieve cardiac arrest (ischemic or anoxic arrest) through the
depletion of high-energy phosphate stores on which cardiac contraction depends. However,
myocardial ischemia led to necrosis and contractile dysfunction upon reperfusion, the worst of
which was an irreversible “stone” heart (5,6). The conundrum of surgeon convenience at the
expense of acceptable outcomes provided a driving force for innovation.

Once the severe complications of ischemic arrest were recognized, chemical methods were
used to arrest the heart. Based on reports from Ringer in 1883 that high concentrations of
potassium would arrest the heart (7), Melrose and associates in 1955 (8) and 1957 (9) used
hyperkalemia to induce asystole during CPB. The combination of immediate arrest and
hypothermia would ostensibly reduce energy (i.e., oxygen) demands of the heart, and thereby
avoid injury imposed by hypothermia alone. Melrose made a hyperkalemic solution by mixing
potassium citrate solution (77 mM) with 18 mL of blood, which was then infused by hand-held
syringe into the aortic root proximal to the aortic cross-clamp. This hyperkalemic arresting
solution was later called “cardioplegia” solution by Lam et al. (10) and Sergeant et al. (11).
Other contemporaries of Melrose used alternative arresting agents, such as acetylcholine
alone (10,11) or in combination with other drugs (12).

Although asystole was indeed achieved with the Melrose potassium citrate solution, the
concept of elective cardiac arrest was abandoned in the United States for some 20 years
following reports of high morbidity and pathologic complications in chemically arrested hearts
(13). Chemical cardioplegia was supplanted by alternative techniques of arresting the heart or
limiting blood flow in the surgical field, including ventricular fibrillation, intermittent aortic
cross-clamping, and profound systemic (14) or topical hypothermia (15). Laboratory studies
demonstrated that topical hypothermia protected inadequately against intraoperative injury,
which resulted in postoperative myocardial subendocardial necrosis, and postischemic
metabolic and functional depression (16,17,18). In the late 1960s, reports described scattered
myocardial or subendocardial necrosis in patients who had died after otherwise technically
successful cardiac surgical operations, suggesting that the current techniques of myocardial
protection were inadequate (19,20). This set the stage for a later re-emergence of chemical
cardioplegia in the United States.

During the surgeon-imposed moratorium on the use of the Melrose solution, cardioplegia
solutions continued development and clinical use in Europe. Bretschneider et al. (16) and
Kirsch et al. (17) used crystalloid formulations which mimicked intracellular ionic
concentrations (intracellular crystalloid solutions). The Bretschneider HTK (histidine,
tryptophan, α-ketoglutarate)-ketoglutarate) solution used low calcium, low sodium, and
procaine with histidine buffers to achieve a nondepolarized arrest, rather than hyperkalemia-
induced depolarized arrest. In the 1970s, Hearse and associates (18,21) at St. Thomas’ Hospital
in London developed an extracellular hyperkalemic solution based on a rigorous series of
experiments that established a scientific understanding of the determinants of injury incurred
during global ischemia. Braimbridge introduced St. Thomas’ cardioplegia solution into clinical
use in 1975 (22). St. Thomas’ solution was later modified based on further experimental
studies to the St. Thomas’ solution No. 2 (23), an intracellular ionic concentration,
normocalcemic, and hyperkalemic (16 mmol/L) solution; it is currently marketed as Plegisol in
the United States. The Bretschneider HTK solution is currently marketed as Custodiol HTK
solution.

Chemical cardioplegia was revived in the United States by Gay and Ebert (24) and Tyers et al.
(25,26), who found that the constituents in the Melrose formulation were inappropriate rather
than the concept of hyperkalemic chemical cardioplegia itself being erroneous. These reports
popularized the use of potassium-based cardioplegia to achieve electromechanical arrest. In
the late 1970s and early 1980s, Follette and colleagues from Dr. Buckberg’s laboratory (see
reference 224 below) introduced the concept of cold hyperkalemic blood cardioplegia. The
physiologic attributes of blood, including its superior buffering capacity, endogenous oxygen
radical scavengers, detoxifying substances, and superior oxygen-transport capacity, has
established this concept as a popular strategy that has facilitated surgery and vastly improved
patient outcomes. Today, a variety of both crystalloid and blood cardioplegia solutions are
used clinically worldwide to achieve elective cardiac arrest and a bloodless field.

The cornerstones of myocardial protection that had been laid by the end of the early 1980s
are: (1) rapid chemical arrest to conserve high-energy phosphate stores and reduce ischemia,
(2) hypothermia to reduce the metabolic rate, (3) adjunctive agents to reduce the effects of
ongoing ischemia (anti-ischemic agents), and (4) continuous or intermittent delivery to restore
or maintain tissue oxygenation. Further developments in cardioplegia-related myocardial
protection strategies focused on (1) optimizing delivery (retrograde, gentle cardioplegia
delivery pressures), (2) overcoming the adverse effects of hypothermia, (3) identifying
metabolic enhancements (ATP generation and pH management) and buffers, (4) managing
calcium (Ca2+) accumulation, and (5) understanding the pathophysiology of ischemia and
reperfusion, which would drive further developments of strategies to avoid these mechanisms
of both ischemic and postischemic patterns of injury. There has also been a persistent search
for agents that induce polarized arrest which overcome the disadvantages of hyperkalemia
(reviewed in Dobson et al. (27) and Maruyama et al. (28)). With the recent rapid developments
in molecular and cellular biology, experimental focus has shifted away from the organ and
physiologic aspects toward cellular and molecular areas to improve the understanding of the
effects of ischemia-reperfusion injury, inflammation, and cell death processes. Understanding
the mechanisms involved in the transition to cell death would direct developments in Ca2+
handling and preventing Ca2+ dyshomeostasis, the role of oxidant injury, and a newly
recognized role of mitochondria in determining the transition from reversible to irreversible
injury, and whether the cell pursues a necrotic pathway or apoptotic pathway to cell death.
Finally, the concept of conditioning, which recruits complex endogenous biologic mechanisms
of self-protection against stressors such as ischemia-reperfusion injury, oxidants and
apoptosis, has been incorporated into pretreatment (preconditioning), intraoperative
(preconditioning), and reperfusion (postconditioning) phases of cardiac surgery. Although
human beings have been evolving for hundreds of thousands of years, it has only been in the
past 60 years or less that the human heart has been subjected to the kind of induced ischemia
and surgical reperfusion discussed in these pages.

PATHOPHYSIOLOGY OF SURGICAL ISCHEMIA-REPERFUSION INJURY

Because many factors affect postoperative outcomes of hearts with various and complex
preoperative pathophysiologic features, developing strategies that adequately protect all
hearts during cardiac surgery is often problematic. The heart under consideration may present
a complex picture of preoperative disease, variable age, gender, metabolic and cellular
dysfunction, global or regional ischemia (both acute and chronic), each of which contribute to
differing vulnerabilities to ischemia and reperfusion injuries. However, adequate myocardial
protection must be based on a sound scientific basis. Obviously, a single cardioprotective
strategy may be inadequate to suit all surgeon preferences and target all patient pathologies.
Hence, the surgical team must often mold the composition of the cardioplegic solution or its
modality of delivery to meet the requirements of the patient’s pathologic profile as presented.
The measures taken to protect the heart during elective cardiac arrest should represent a
balance struck between the requirements of the heart during aortic cross-clamp or
reperfusion, with appropriate adjustments made for special considerations (hypertrophy,
diffuse coronary disease), and avoid distraction from the surgical procedure. Ischemia and
reperfusion injury (29) and the choice of cardioplegia technique (30) represent primary
contributors to patient outcomes.

Ischemic Injury

When Does Ischemia and Reperfusion Injury Occur during Cardiac Surgery?
Figure 9.1. Events occurring before the institution of cardiopulmonary bypass, during delivery
of cardioplegia and at reperfusion that predispose the myocardium to ischemia and
reperfusion injuries. Arrhythmias, ventricular fibrillation (VF) or severe hypotension can cause
antecedent global ischemia. During the delivery of cardioplegia, the composition of the
solution can cause edema, further ischemia if oxygen and nutrients are inadequate to meet
ongoing albeit lower demands, and Ca2+ concentration is either too high or too low. Coronary
obstructions, air or low cardioplegia infusion pressure can lead to maldistribution of solution.
Too high infusion pressure can lead to microvascular injury and edema. Potassium itself has
negative effects secondary to placing the cell in a depolarized state. Ischemia may be
encountered if air emboli, graft kinks, tight anastomoses, or systemic hypotension cause
inadequate distribution of blood flow. Finally, reperfusion injury can cause reversible (stunning,
arrhythmias, edema) or irreversible (necrosis, apoptosis) injury.

It is critical for the surgical team to understand that the presence of acute, ongoing and
evolving or progressing ischemia must impact the protection strategy. Ischemia-reperfusion
injury can occur in the surgical setting at three major time points during the surgical procedure
(summarized in Fig. 9.1).

(a) Before CPB has been instituted or cardioplegia solution has been delivered:
“Unprotected” clinical ischemia before CPB usually results from the presenting acute coronary
syndrome and can be exacerbated by hypotension, arrhythmias (e.g., atrial or ventricular
fibrillation), and other hemodynamic changes (e.g., cardiogenic shock, coronary artery spasm).
In most cases, acute ischemia is active in both the distribution of the principally involved artery
(cardiologists call this the “infarct artery”) but also distributions in which preexistent collateral
flow may be compromised. Patients referred for bypass surgery in these circumstances almost
always have multi-vessel disease patterns making the myocardial substrate dangerous from
the standpoint of the severity and duration of the syndrome, which can be extremely difficult if
not impossible to determine. Reperfusion and reoxygenation injury occur when these
conditions change abruptly before effective measures are implemented. Recent attention has
been paid to avoiding over-oxygenating such patients systemically, and this will be discussed in
more detail in later sections.

(b) At the initiation of cardiopulmonary bypass: At the moment bypass is initiated, blood is
immediately diverted from the right atrial drainage system into a venous reservoir where it is
then pumped through the oxygenator. This configuration does more to the heart than merely
oxygenate venous blood and send it back to the ascending aorta. As the heart is emptied into
the venous side of the pump, the oxygen demand of the metabolizing myocardium decreases
by 50%, but the high oxygen content of blood extracorporeally oxygenated makes the tissue
vulnerable to reoxygenation damage if no steps are taken to mitigate this injury. This might be
viewed as an abrupt interruption of antecedent ischemic conditions. While this is one of the
principal goals of bypass (to give the heart a “rest”), it must be recognized that the heart is also
vulnerable to a pattern of injury that has been, at best, underappreciated in the past. Ihnken et
al. (31) suggested as early as 1998 that initiation of bypass is also, perhaps paradoxically, a
dangerous moment for the heart. In this work, he observed clear, reproducible biochemical
evidence of injury in samples of blood drawn from the coronary sinus. Since Buckberg (32)
published his work on the reoxygenation injury in 1995, it has been observed that abrupt over-
exposure of the ischemic heart to oxygen increases injury in a pO2-dependent pattern of injury
that can be mitigated by altering the oxygen tension of the perfusate.

(c) During the “protected” period of arrest and delivery of cardioplegia: Ischemia or
reperfusion injury can occur during the cardioplegia phase. Ischemia can develop or be
worsened under several circumstances (1) the unintentional maldistribution of cardioplegia
solution distal to stenotic or totally occluded coronary arteries, (2) between intermittent
infusions of cardioplegia solution, (3) during interruption of continuous cardioplegia strategies,
(4) inadequate delivery of retrograde cardioplegia to areas of the right heart whose venous
blood drains directly into the Thebesian system, or areas that are simply under-filled due to
malposition of a retrograde delivery device. Though the cardioplegia phase is considered
“protected time,” the potential for worsening of, or the development of, ischemic conditions
followed by reperfusion injury can occur with the initial administration of cardioplegia and
with each subsequent delivery of cardioplegia solution [(re)perfusion] for a variety of reasons:
(1) excessive pressure (microvascular injury), (2) low onconicity favoring extravasation of fluids
into the extracellular space, (3) formation of oxygen radicals from over-oxygenation and
oxidative stress, and (4) other mechanisms of reperfusion injury as discussed below. While
delivery of cardioplegia should provide protection, this phase also offers opportunity for
additional injury to occur unintentionally. Any surgeon who has been faced with postoperative
low-cardiac output and who suspects inadequate protection knows this to be true.

d) After release of the cross-clamp and attempts to reanimate the heart: Additional ischemia
may be encountered at the release of the cross-clamp when (1) coronary blood flow is
impaired through kinked grafts or (2) tight anastomoses (3) air emboli are present in the
coronary arteries, (4) ventricular fibrillation occurs at low perfusion pressures, or (5) poor
ventricular performance or dysrhythmias cause hypotension once the heart is converted and
off bypass. Clearly, when the clamp is removed reperfusion injury can occur as blood flow is
restored and oxygen exposure is reestablished after any of the above events. In terms of
surgically induced ischemia, this is intuitively the major time when one thinks of reperfusion
injury occurring.

Understanding the pathophysiologic mechanisms of myocardial injury occurring at these

intervals and applying the principles of myocardial protection will help avoid injury induced by
the very techniques intended to preserve the myocardium. Unlike the cardiologist, the surgical
team has the opportunity to intervene directly to control many facets of ischemia and
reperfusion injury, and to limit at some point the mechanisms causing injury, thereby more
favorably directing the outcome of the postischemic heart. The pathophysiology of ischemia
and reperfusion is described in the following sections. Interventions targeting these various
pathologic processes are discussed under Myocardial Protection Therapy.

Determinants of Ischemic Injury

Ischemia encountered either before arrest or during cardioplegic arrest, is a major cause of
postcardioplegia injury. Ischemia sets the stage for reperfusion injury; without ischemia you
cannot have reperfusion injury. Ischemia is defined as inadequate energy supply to meet the
current demands, that is, an energy supply/demand ratio of less than 1; normally, the O2
supply/demand ratio is >1. Since the adenosine triphosphate (ATP) necessary to support the
energetic needs of the heart is provided predominantly by oxidative phosphorylation, with
scant generation of high-energy phosphates through anaerobic metabolism, the heart is nearly
an obligate aerobic tissue. Therefore, the energy consumption of the heart can be
approximated by the oxygen consumption. The limited availability of anaerobic energy-
generating pathways mandates that O2 be in constant supply. A very brief discussion of the
determinants of O2 supply and demand follows.

Oxygen supply: Ischemia may be caused by either a decrease in O2 supply relative to demands
(supply ischemia) or an increase in O2 demands relative to supply (demand ischemia). O2
supply is determined by the O2 extraction (arterial-coronary venous O2 difference) and
coronary blood flow. Since O2 extraction by the heart is normally ~75%, and its physiologic
limit is approximately 95%, only limited amounts of oxygen may be obtained through
additional extraction. The majority of O2 is therefore supplied by adjustments in coronary
blood flow, which under normal conditions may increase 4- to 5-fold to meet demands. This
O2 is largely present as hemoglobin-bound, with only a minor fraction (~1.5%) dissolved in
plasma. This limited dissolved O2 will become important when we discuss crystalloid
cardioplegia solutions in which oxygen availability is limited to that dissolved in solution.

Oxygen demand: Overall myocardial oxygen demand of the normal working heart is
determined by the work of the entire heart. The majority of this overall O2 consumption is
attributed to utilization by the left ventricle. In a nutshell, the determinants of left ventricular
O2 demand include pressure–volume work or wall stress, heart rate, temperature, inotropic
state, basal metabolism, and ionic homeostatic mechanisms (ATP-dependent pumps) required
to re-equilibrate ionic balance after electromechanical activity. In the immediate postoperative
heart, energy demands may be related to oxidative energy diverted to myocyte repair.

The energy related to maintain ionic equilibrium (~5% of total demand) is an ongoing demand
even when the heart is not contracting since ATP-dependent pumps are constantly adjusting
calcium and sodium influx, as well as potassium efflux. This energy demand of ionic
equilibration is important in determining ongoing oxygen demands in the arrested heart, as
discussed below. How oxygen demands are changed by hypothermia is discussed later.

Severity and Duration of Ischemia in Cardiac Surgery

The potential for myocardial injury is related, in part, to the duration of ischemia. The time to
onset of irreversible ischemic injury will depend on many factors, including the severity of
ischemia, myocardial temperature, ambient energy demands, and collateral blood flow.
However, irreversible injury can become apparent after as little as 30 to 45 minutes of
coronary occlusion in the working myocardium. However, shorter durations of global ischemia
can result in mild to severe systolic and diastolic dysfunction without irreversible tissue
necrosis (left ventricular “stunning” (33)). Therefore, the duration of ischemia, both preceding
CPB and intraoperatively after application of the cross-clamp, has often been used as a
predictor of postoperative myocardial injury in experimental models, and is one of the
underlying considerations in the correlation between total ischemic time (CPB, cardioplegia)
and clinical outcomes. Because the surgical team often does not have control over the
duration of warm ischemia (particularly antecedent ischemia caused by coronary occlusive
disease), other aspects of elective ischemic time are more often targeted for modification to
reduce the severity of ischemia and hence postischemic injury. Therefore, strategies such as
(1) initiating rapid asystole, (2) venting the left ventricle to reduce consequences of chamber
distension resulting in increased wall stress, and (3) imposing cardiac hypothermia all
specifically target the reduction of myocardial energy/oxygen demands, thereby increasing the
limits of “safe” ischemic time during aortic cross-clamping. With these cardioprotective
strategies in place, the duration of aortic clamping that can be safely imposed can be increased
from as little as 15 to 45 minutes to several hours using hypothermic cardioplegia (Fig. 9.2).
Reducing to almost zero the determinants of myocardial work and oxygen demand is the sine
qua non of myocardial protection. In the section on future directions, eliminating ischemic
conditions without additional oxygen-related injury and changing the milieu more gradually
will be discussed.

Figure 9.2. Left ventricular performance measured by in situ Starling curves (A) or end-systolic
pressure–volume relations (conductance catheter) (B) after 45 minutes of normothermic
unprotected global ischemia and reperfusion. C: Left ventricular stroke work index in patients
before (Baseline) or after (pCPB) 15 minutes, or 4 and 24 hours postcardiopulmonary bypass
(CPB). The dashed line between 4 and 24 hours after discontinuation of CPB represents further
decreases in cardiac performance secondary to onset of irreversible injury.(A, adapted from
Rosenkranz ER, Buckberg GD. Myocardial protection during surgical coronary reperfusion. J Am
Coll Cardiol 1983;1:1235–1246, with permission.)

The determinants of surgical ischemic injury can be summarized as:

● The duration and severity of antecedent” and “unprotected” ischemia: Although it has
been recognized for decades that ischemia time (duration) impacts tissue injury and
postischemic salvage and function, an increasing appreciation for the caveat that the
 manner in which the tissue is reperfused significantly alters the outcome. It is now
increasingly understood that reperfusion injury can occur by merely eliminating
ischemic conditions abruptly or in the presence of an inappropriate oxygen gradient. In
the case of the imposition of aortic cross-clamping, there is total elimination of native
flow with the exception of that supplied by noncoronary collateral flow. Global
contractile function may be impaired before necrosis or apoptosis are evident
(“stunning”). It is the reduction of myocardial oxygen demand that protects the heart
in this circumstance.
● The elimination of electromechanical activity during cross-clamping is paramount
because its mechanical activity increases O2 demands even in the presence of
hypothermia (discussed further in the section on Hypothermia).
● Myocardial temperature: Temperature influences metabolism by the “Q10 effect,” in
which the tissue metabolic rate decreases by half for each 10°C decrease in
temperature; conversely, myocardial metabolism doubles when the heart rewarms by
10°C. Hypothermia can buy biologic time during ischemia.
● The nutritional and stress status of the heart: For example, catecholamines increase
the O2 demands by increasing contractility and exerting an O2 wasting effect.
● Presentation of comorbidities such as hyperlipidemia which increase the vulnerability
to ischemia-reperfusion injury for any given duration of ischemia. This effect is
independent of energy demands, but is influenced more by mechanisms exerted at
reperfusion, that is, the capacity to generate O2 radicals, a pro-inflammatory state of
the endothelium and inflammatory cells, and calcium handling by the tissue.

Determinants of Reperfusion Injury

Although reperfusion of the myocardium is the ultimate goal of both surgical and nonsurgical
revascularization, abrupt reperfusion carries with it the potential for extending postischemic
injury to cardiomyocytes, coronary vascular endothelium, and the microvasculature. Indeed,
reperfusion injury associated with rapid changes in circulatory conditions may be responsible
for ~50% (34) or more of the ultimate infarct size, and is an important contributor to
postsurgical mortality and morbidity as well (29). We can define surgical reperfusion injury as
an incremental increase in the pathology that is observed after ischemic injury alone.
Reperfusion injury extends or accelerates damage from that observed during ischemia alone,
and occurs after the onset of reperfusion (i.e., cardioplegia) or reperfusion (i.e., cross-clamp
removal). Viewed another way, reperfusion injury can occur at the instant of the elimination of
ischemic conditions with reestablishment of flow. Hence, reperfusion injury can be viewed as
representing an increment of injury that had begun during ischemia but progresses or
manifests during the reperfusion phase. Other investigators have described de novo injury
starting at reflow or, in a newer construct, at the time of the elimination of ischemia (even if
that is accomplished by a reduction in oxygen demand). Reperfusion injury can be categorized
as either reversible or irreversible. (1) Reversible injury includes temporary contractile
dysfunction (“stunning”) in the absence of morphologic injury or necrosis, or (2) irreversible,
which includes necrosis and apoptosis. Biomarker enzymes such as creatine kinase (CK) or
cardiac troponin T (cTnT) or I (cTnI) released by disrupted cells have been observed to be
elevated during reperfusion rather than during ischemia. In the past, debate has often been
spirited regarding whether reperfusion injury even exists (35,36,37,38) and whether
reperfusion injury kills cells that are salvageable at the end of ischemia. However, a consensus
has developed in both the surgical, cardiology, and cardiovascular research communities that
reperfusion is indeed an important component of postischemic cardiac injury in animal models
(39) and in humans (34,40) and that alterations in reperfusion strategies can mitigate injury.

Among the more important mechanisms in the pathophysiology of myocardial ischemia-

reperfusion injury are (1) the duration and severity of antecedent ischemia, (2) oxygen radicals
generated not only by neutrophils but also by myocytes and vascular endothelium, (3) sodium
(Na+) and Ca2+ influx and loss of normal intracellular Ca2+ homeostasis, (4) neutrophils and
neutrophil-derived products that constitute a generalized inflammatory response to
reperfusion, and (5) mitochondrial dysfunction and opening of the mitochondrial permeability
transition pore (mPTP). Interestingly, the presence of hyperoxic or even relatively hyperoxic
conditions can ramp up or trigger all of these mechanisms. As such, these mechanisms do not
operate independently, nor are they mutually exclusive. Importantly, reperfusion injury is a
malleable process that can be attenuated by utilizing a number of strategies targeting these
various perpetrators of injury. In the operating room during cardiac surgery, and more and
more frequently in the modern era in the interventional lab, various forms of circulatory
support techniques and cardioprotective drugs are employed. The realization that circulatory
support can manipulate reperfusion therapy by altering oxygen supply/demand relationships is
critical for any operator that employs it. Such pumps provide an avenue for both the reduction
of oxygen demand and for the delivery of therapeutic agents via the pump, which can be
administered intravenously or by using cardioplegia as a vehicle. Using cardioplegia offers a
unique opportunity to modify the conditions of reperfusion (hydrodynamics, temperature,
route of delivery) and the composition of the solution (blood, crystalloid, pH, metabolic
substrates, hypocalcemia, oxygen, pharmaceuticals) that target the pathophysiologic
mechanisms of injury. It is now also recognized that the initiation of circulatory support in it
itself offers an opportunity to mitigate injury by paying attention to conditions at the moment
support is begun and ischemic conditions are eliminated.

Reactive Oxygen Species (ROS) and Reactive Nitrogen Species (RNS)

Molecular oxygen is relatively inert by virtue of the shared electron pair in the outer molecular
shell. However, oxygen must be viewed as a pharmaceutical and potentially toxic agent in
cardiovascular medicine and surgery. ROS are radicals or unstable oxygen molecules derived
from oxygen. Important oxygen radicals are superoxide anion (−O2) and hydroxyl anion
(•OH).OH). However, other oxygen-related molecules that are not technically radicals are
considered biologically reactive nonetheless. An example of this would be hydrogen peroxide
(H2O2). In addition, hypochlorous acid (HOCl) and chloramines derived from neutrophils exert
potent oxidative injury to biologic tissues.

Free-radical molecules that are derived from nitrogen are termed RNS and include nitric oxide
(NO•), peroxynitrite (ONOO–), nitrogen dioxide (NO2•) and nitrosyl hydride (HNO). In vascular
endothelium, NO• is constitutively generated by endothelial nitric oxide synthase (eNOS or
NOS-3). eNOS activity is dependent on Ca2+-calmodulin, flavin, NADPH, and
tetrahydrobiopterin (BH4). NO• is generated by an electron from the guanadino nitrogen of L-
arginine to incorporate oxygen to generate NO • and the byproduct L-citrulline. NO• has an
extremely short half-life (seconds) in vivo, and binds tightly to hemoglobin. OONO− is formed
by a nonenzymatic biradical reaction between NO• and −O2. NO• has been linked to biologic
signaling in cardioprotection, particularly anti-inflammatory effects, endothelial function,
vasorelaxation (hence its early name as endothelial-derived relaxing factor, EDRF),
neurotransmission, immune regulation and defense mechanisms.

NO• homeostasis is key in normal regulation of blood flow and responses to stress, and in
preventing endothelial dysfunction. NO• homeostasis is a balance between NO • generation,
bioavailability, and degradation or quenching by oxidants such as −O2. NO • has pleiotropic
effects in the cardiovascular system by paracrine effects: (1) NO • is a physiologic regulator of
blood pressure and flow, and counteracts local and systemic vasoconstrictors such as
endothelin (ET-1); (2) it inhibits neutrophil adhesion to vascular endothelium; (3) it inhibits
platelet aggregation; (4) it reduces myocardial oxygen consumption (41); (5) it has positive
lusitropic and inotropic effects; (6) it scavenges −O2 and therefore is directly antioxidant; and
(7) it is anti-proliferative. The effects on oxygen consumption and inotropy are somewhat
controversial (42,43). The role of NO• in ischemia-reperfusion injury and cardioprotection is
discussed in more detail under the section on endothelial dysfunction and NO • donors.

Reactive Oxygen Species

The oxygen radical is highly reactive with a broad range of biologic substances including sugars,
amino acids, phospholipids, and DNA. With such a large catalog of biologic targets, ROS and
their metabolites are potentially toxic to cells, and at the same time they are critical in the host
defense mechanism and bactericidal activities of neutrophils and phagocytes (44) and
participate in cardioprotective signaling. In this dualistic role of ROS as good guy–bad guy, it is
proposed that low-level production of ROS, particularly from mitochondria, acts as a signal for
cardioprotective conditioning responses (preconditioning, postconditioning) and stimulation of
cardioprotective kinase pathways, while large-scale generation of ROS from inflammatory cells
or dysfunctional mitochondria is deleterious. In the ischemic-reperfused myocardium, ROS-
induced injury includes peroxidation of lipid components of cellular membranes leading to
damage to sarcolemmal cell membrane and other membranous organelles such as
mitochondria and sarcoplasmic reticulum (45). ROS also cause the mPTP to open when
generated in an environment of Ca2+ accumulation and normalized pH (conditions achieved
during reperfusion), which leads to a collapse of the membrane potential, loss of metabolic
capacity, and release of pro-apoptotic factors. This is discussed in more detail below under the
role of mitochondria in reperfusion injury. The vascular endothelium is particularly vulnerable
to ROS-mediated injury, which manifests as impairment of vascular endothelial function
through attenuated production of vasoactive and anti-inflammatory autacoids such as
adenosine (46,47,48) and NO• (49), and increased production of pro-inflammatory and
vasoconstrictive molecules. These injury mechanisms contribute significantly to postischemic
dysfunction, dysrhythmias, and morphologic injury such as necrosis and apoptosis.

Sources of Oxygen Radicals

Both enzymatic and nonenzymatic reactions produce oxygen radicals. The major reactions and
their sources are shown in Figure 9.3. ROS are generated by a number of sources, including
activated neutrophils, cardiac myocytes, mitochondria, and the vascular endothelium.
Figure 9.3. The species and sources of reactive oxygen species (ROS) generated by neutrophils,
cardiomyocytes and their mitochondria, and vascular endothelium. SOD, superoxide
dismutase; CAT, catalase; GP, glutathione peroxidase; MPO, myeloperoxidase; HOCl,
hypochlorous acid; e−1, electron; NOX, NAD(P)H oxidase.

Neutrophils: Activated neutrophils represent the major source of ROS in the heart under
stress, including −O2, •OH, and HOCl species. Neutrophils are activated by various chemotactic
factors stimulated by either CPB or ischemia-reperfusion, including the complement
anaphylatoxins C3a and C5a, f-Met-Leu-Phe, tumor necrosis factor alpha (TNFα), interleukin-8
(IL-8), and platelet-activating factor. These factors trigger a “respiratory burst” of ROS. CPB
(independent of ischemia) activates the complement cascade (50,51), with the subsequent
activation and adherence of neutrophils (52) to the vascular endothelium primarily, which
subsequently triggers the release of cytotoxic products, all of which contribute to tissue injury
in the ischemic-reperfused myocardium (52,53). During the respiratory burst that occurs
seconds after stimulation by cytokines or ischemia-reperfusion, the −O2 in the neutrophil is
produced by the membrane-associated NADP oxidase, which transfers an electron to
molecular oxygen (Fig. 9.3, reviewed in Sheppard et al. (54)). Production of H2O2 by
neutrophils or the reduction of −O2 to H2O2 by superoxide dismutase (SOD) provides
substrate for generation of •OH, or the oxidant HOCl by myeloperoxidase. HOCl reacts with low
molecular weight amines to give rise to lipophilic chloramines, which can promote membrane
lipid peroxidation.

Cardiomyocytes: Xanthine oxidase (XO) is a major source of ROS from cardiomyocytes.

Normally, XO exists mostly in the dehydrogenase form and uses nicotinamide adenine
dinucleotide (NAD) as an electron acceptor. Xanthine dehydrogenase is converted to the
oxidant-generating enzyme XO during ischemia. Both hypoxanthine and xanthine accumulate
during ischemia. Hypoxanthine is converted to xanthine by XO, which is in turn converted to
uric acid by the same enzyme, which forms −O2. During ischemia, however, two events favor
the xanthine oxidase reaction: (1) hypoxanthine accumulates as a result of the sequential
catabolism of purine high-energy phosphates (ATP, ADP, AMP) and deamination of adenosine,
and (2) the dehydrogenase form of the enzyme is converted to the superoxide-producing
oxidase form by a protease activated by increased levels of intracellular calcium (55). XO is
reported to be localized to the vascular endothelium in substantial quantities (56). However,
the importance of XO to the pathologic generation of ROS in humans remains controversial
(57).

Mitochondria: Mitochondria are a major noninflammatory site of ROS generation (58). ROS are
generated at complex I (NADH/ubiquinone oxidoreductase) and complex III
(ubiquinol/cytochrome c oxidoreductase). The latter site catalyzes the conversion of O2 to
−O2• by a single-electron transfer. −O2 can be converted in the mitochondrial matrix to H2O2
by matrix superoxide dismutase. Since mitochondria do not have catalase, the H2O2 can be
degraded by nonenzymatic reaction with glutathione (GSH) by glutathione peroxidase or by
enzymatic reaction with thioredoxin reductase. The ROS species target cysteine residues on
the protein moieties of the mitochondrial membrane and polyunsaturated fatty acids, and
cause intramolecular cross-linkages and protein aggregates. The •OH causes peroxidation of
membrane lipids. ROS induce Ca2+ release from the mitochondria, which may cause an
imbalance in Ca2+ handling and stimulation of Ca2+-dependent proteases, nucleases and
phosphatases leading to functional and morphologic damage to the cell.

Vascular endothelium: Superoxide anion is produced by vascular endothelium by XO as

described above for cardiomyocytes (Fig. 9.3). A more important source of −O2 from the
vascular endothelium may be the NAD(P)H (reduced NAD phosphate) oxidase system. This
enzyme is a membrane-bound, flavin-containing NADH/NADPH-dependent oxidase present in
endothelial and vascular smooth-muscle cells. The molecular structure, function, and
regulation are somewhat similar to those of the neutrophil enzyme system, requiring assembly
of cytosolic components onto the membrane components before −O2 generation can occur.
However, −O2 production in the presence of NO • produced by endothelium favors the
biradical neutralization and elimination of NO • and the formation of peroxynitrite (ONOO–).
The elimination of NO• impairs not only vascular relaxation and autoregulatory control of
postischemic myocardial perfusion, but also fails to inhibit neutrophil-mediated postischemic
inflammatory responses.

Other sources of ROS include the oxidation of catecholamines, metabolism of arachidonic acid
to peroxy compounds, and generation of •OH by the cyclooxygenase and lipoxygenase
pathways, and by the metal-catalyzed Haber–Weiss reactions. The latter series of reactions
may be important in the setting of CPB because they can be recruited by neutrophil activation
during CPB (59). In addition, iron-containing products released by hemolysis may facilitate this
reaction.
When are Oxygen Radicals Generated?

Myocardial ischemia favors the generation of ROS. However, the myocardium has a cadre of
the endogenous antioxidants superoxide dismutase, catalase, reduced glutathione (GSH),
glutathione peroxidase, glutathione reductase. Normally, upregulation of these antioxidants
and phase 2 enzymes reduces oxidant-mediated injury (60), but they are depleted or can be
overwhelmed during ischemia, thereby attenuating the natural defense mechanisms of the
heart (61). Therefore, ischemia creates the biochemical setting for oxyradical production and
tissue vulnerability to ROS-mediated damage. However, the primary substrate, oxygen, is in
limited supply during ischemia, and is not readily available until reperfusion. Oxygen radicals
are not seen in great abundance until the onset of reperfusion. A clinical study by Prasad et al.
(62) show that oxygen radicals are released during and 24 hours after CPB in CABG patients.
Indirect evidence to support this reperfusion oxidative burst comes from a wealth of data in
which inhibitors or scavengers of oxygen radicals, or anti-neutrophil agents, exerted beneficial
effects when administered during reperfusion.

In the surgical setting, the myocardium may be vulnerable to ROS-induced injury at several
points when oxygen is introduced into the myocardium: (1) the initial delivery of cardioplegia
into the heart with antecedent unprotected ischemia, or initial delivery of oxygenated
cardioplegia to a newly revascularized myocardial segment; (2) intermittent reinfusions of
cardioplegia if a multidose strategy is being used; and (3) release of the aortic cross-clamp to
initiate reperfusion after the cardioplegia phase. The quantity of ROS generated during cardiac
surgery depends on the severity of the preceding ischemic injury, activation and recruitment of
neutrophils, the level of oxygen in cardioplegic solutions or blood oxygenated by the
extracorporeal circuit, and the status of endogenous scavengers and inhibitors potentially
depleted during ischemia. Blood cardioplegia theoretically provides a greater oxygen substrate
for oxygen radical production than do crystalloid solutions. However, plasma-dissolved oxygen
may be the pool from which oxygen is drawn for the generation of oxygen radicals, and the
oxygen content in this compartment may not differ dramatically between blood and crystalloid
solutions. The oxygen in blood cardioplegia may be reduced with “normoxic” or even
“deoxygenated” blood cardioplegia (63) or during reperfusion (64,65). In addition, blood
cardioplegia may offer the advantage of endogenous antioxidants (superoxide dismutase,
catalase, glutathione, and albumin) that are not present in crystalloid cardioplegia solutions.

Intracellular Na+ and Ca2+ Dyshomeostasis

Under normal physiologic conditions, intracellular Ca2+ concentrations are maintained at low
concentrations (<200 nmol) against a 5,000-fold trans-sarcolemmal gradient by voltage-gated
Ca2+ channels that remain closed until activation during phase 2 of the action potential. The
regulation of intracellular and intramitochondrial Ca2+ is tightly linked to regulation of
intracellular Na+. Intracellular Na+ concentration is normally controlled by the ATP-dependent
Na+/K+ pump which moves Na+ outward and K+ inward against their respective concentration
gradients, and the energy-independent Na+/H+ exchanger which normally removes one H+
from the cytosol in exchange for one Na+ moving into the cytosol. Further Na+ accumulation
mainly occurs through the Na+/Ca2+ exchanger which normally operates in the forward
direction during diastole to extrude intracellular Ca2+ in exchange for influx of Na+ into the
cytosol.

In ischemic-reperfusion myocardium, Ca2+ influx occurs through multiple pathways: (1)

diffusion through overt membrane disruptions or opened Ca2+ channels, fueled by the
differential concentration gradient; (2) reversal or inhibition of the Na+/Ca2+ exchanger driven
by the intracellular accumulation of Na+, (3) attenuated cyclical Ca2+ uptake and release from
the sarcoplasmic reticulum and mitochondria, and (4) facilitation of entry via α-adrenoreceptor
activation and cyclic adenosine monophosphate (AMP)-dependent processes, which increase
during postischemic catecholamine release.

The accumulation of intracellular Na+ and Ca2+ during ischemia-reperfusion are linked events
leading to Ca2+ overload which stimulates Ca2+-dependent proteases that damage
membranes and ion transport mechanisms at the moment of re-exposure to molecular
oxygen, opening of the mPTP, depletion of high-energy phosphate stores, the development of
local or global contracture (“stone heart”), and triggers the transition to cell death through
both apoptosis and necrosis. The balanced Ca2+ transport systems are potentially injured by
exposure to oxygen radicals, cytokines, and activated complement during CPB and reperfusion.
Na+ and Ca2+ dyshomeostasis occurs to differing degrees during ischemia and during
reperfusion, as described below.

Ischemia

During ischemia, glycolysis generates an abundance of H+ and only a limited amount of ATP (2
moles/mole glucose) compared to oxidative phosphorylation (36 moles ATP/mole glucose).
The accumulation of intracellular H+ stimulates activity of the Na+/H+ exchanger (NHE1) in
favor of H+ efflux and Na+ influx; at this time, the reduced availability of ATP decreases Na+/K+
ATPase activity. Therefore, both the attenuated Na+/K+ pump activity and increased activity of
the Na+/H+ exchanger result in intracellular Na+ accumulation. Additional Na+ may enter the
cell through sarcolemmal nonactivating Na+ channels, for example, the “window currents”
generated during ischemia-induced cellular depolarization. However, some evidence suggests
that the accumulation of H+ in both the extracellular and intracellular spaces attenuates the
H+ gradient, thereby reducing but not eliminating the participation of the Na+/H+ exchange
mechanism in intracellular Na+ accumulation during ischemia (66). However, the accumulation
of Na+ in the cytosol during ischemia would favor reversal of the energy-independent
Na+/Ca2+ exchanger in favor of extruding Na+ in exchange for Ca2+ influx, thereby favoring its
accumulation in the cytosol. Entry of Ca2+ into the mitochondria through the Ca2+ uniporter
depends on the Ca2+ electrochemical gradient and the presence of a mitochondrial
transmembrane potential. Ca2+ enters the mitochondria during early ischemia, but further
increases and overload are prevented by dissipation of the mitochondrial transmembrane
potential (67). In addition, the unregulated activation of Ca2+-dependent protease enzymes
such as calpains is inhibited by intracellular acidosis, and is delayed until intracellular pH is
renormalized during reperfusion (68). Therefore, Ca2+ accumulation and Ca2+-mediated
damage occurs to a greater extent during reperfusion (69,70,71) as discussed below.
Reperfusion

Na+ and Ca2+ dyshomeostasis during reperfusion is a significant factor contributing to

postischemic myocardial injury and functional impairment (72,73). The ion dyshomeostasis
described above sets the stage for explosive reperfusion/reoxygenation damage. It is
increasingly recognized that abrupt re-exposure to oxygen itself is the trigger for damage.
Excessive gradients of perfusate-to-tissue oxygen tension can be mitigated. In addition, the
events leading to abnormal Ca2+ handling which determine cell survival versus cell death are
as follows:

● Rapid renormalization of intracellular pH: Reperfusion washes out lactate and re-
establishes the intracellular–extracellular H+ gradient (extracellular < intracellular)
which allows the forward motion of the Na+/H+ exchanger favoring removal of H+
(renormalization of intracellular pH) and subsequent intracellular accumulation of Na+;
other factors contributing to normalization of intracellular pH are activation of
lactate/H+ co-transport mechanism and activation of the NaHCO3 co-transporter.
● Renormalization of intracellular Na+: During the initial phase of reperfusion, ATP is
regenerated, and if the Na+/K+ ATPase is not damaged by proteolytic actions of
calpain enzymes, intracellular Na+ levels are renormalized and the linked Na+/H+
exchange—Na+/Ca2+ exchange systems will not be activated. Otherwise, if the Na+/K+
ATPase pump is damaged, intracellular Na+ accumulation will engage the Na+/Ca2+
exchange mechanism with net intracellular accumulation of Ca2+.
● Ca2+ uptake and release by the sarcoplasmic reticulum (SR): The SR is the main Ca2+
storage site of the cardiomyocytes. Ca2+ in the micro-domain proximate to the closely
co-localized SR and interfibrillar mitochondria is cyclically taken up by the SR (lowering
of cytosolic concentration) during diastole, and Ca2+ is released into that micro-
domain by the ryanodine receptors (RyR) during systole. During both unprotected and
protected ischemia, sequestration of Ca2+ into the sarcoplasmic reticulum competes
for a limited amount of ATP. This paradox creates a cycle whereby a high intracellular
Ca2+ level binds high-energy phosphates to further limit ATP availability, which then
attenuates the ability of the SR to remove the Ca2+ actively from the cytoplasm,
resulting in a further increase in cytosolic Ca2+. If other Ca2+ handling mechanisms are
not engaged, the release of Ca2+ from the SR RyR may lead to overload. This activity of
the SR mechanism may be related to the severity of ischemia, being less so during
short periods of ischemia.

So, what events in the balance of Ca2+ regulation determines if a cell lives or dies during
reperfusion? Some evidence suggests that the timing of normalization of pH and Ca2+ handling
mechanisms are important. If pH normalization occurs before Ca2+ handling mechanisms are
restored, then Ca2+ overload, in conjunction with ROS generation and a normalized
intracellular pH environment, will trigger opening of the mPTP causing cell death. However, if
Ca2+ handling mechanisms are restored before pH is normalized, for example, by maintaining
tissue acidosis during early reperfusion, then Ca2+ overload can be avoided. Interestingly,
venous blood-based perfusate has been preliminarily observed to mitigate both over-
oxygenation during early reperfusion and delayed normalization of pH, thereby creating a
“permissive acidemia.”
The Inflammatory and Innate Immunity Responses to Ischemia-Reperfusion and
Cardiopulmonary Bypass

Injury mechanisms of ischemia-reperfusion and CPB involve inflammatory and innate immune
responses. Cardiomyocytes generate pro-inflammatory cytokines, chemokines and adhesion
molecules that direct the emigration of inflammatory cells into ischemic-reperfused tissue
regions and amplify the inflammatory response. The following sections will describe the
inflammatory process involving neutrophils, pro-inflammatory cytokines, and complement,
and the innate immune response involving toll-like receptors (TLRs).

Neutrophils and Neutrophil-Mediated Injury

Neutrophils play a critical role in the pathogenesis of myocardial reperfusion injury (74,75,76).
The inflammatory component of reperfusion injury requires an interaction between activated
neutrophils and vascular endothelium during the early moments of reperfusion, which is
prerequisite to activation of the full local inflammatory cascade. A well-orchestrated sequence
of events mediates the interaction between adhesion molecules on neutrophils and
endothelium. These adhesion molecules are categorized into three families:

● The selectins (P-selectin, L-selectin, E-selectin) are glycoproteins involved in the

interactions between neutrophils and the vascular endothelium in the early moments
of reperfusion (77,78,79). P-selectin is stored in Weibel–Palade bodies in endothelial
cells, and its surface expression is triggered by pro-inflammatory mediators such as
oxygen radicals, thrombin, complement components, cytokines, histamine, and H2O2
released during ischemia-reperfusion and CPB. IL-8 released from hypoxic endothelial
cells also increases the adhesiveness of endothelial cells for neutrophils (80). P-selectin
surface expression peaks after 10 to 20 minutes of reperfusion, and P-selectin is
subsequently shed to soluble fragments in blood. In contrast to P-selectin, L-selectin is
constitutively expressed on the surface of neutrophils and may be the counterligand
for P-selectin during early reperfusion (81). Loose adherence and “rolling” of
neutrophils on endothelium involves interaction with a glycoprotein sialyl Lewisx or a
high-affinity glycoprotein ligand termed P-selectin glycoprotein ligand-1 (PSGL-1) (82).
The third member of the selectin family, E-selectin, is expressed on the surface of
endothelial cells. It is expressed later in reperfusion (4–6 hours) and may therefore be
involved in later events, such as infarct extension.
● The β2-integrins (CD11/CD18 complex) are a family of glycoproteins located on
external membranes of neutrophils. There are three distinct α-chains (CD11a, CD11b,
CD11c) and a common β-subunit. Surface expression of perhaps the major complex
CD11b/CD18 is triggered by a number of pro-inflammatory mediators, including
circulating and endothelium-derived (acylated) platelet-activating factor. After initial
tethering of neutrophils by endothelial P-selectin, firm adherence is mediated by
interaction with CD11b/CD18 and its counterligand ICAM-1 on the endothelium.
● The third family of adhesion molecules is the immunoglobulin superfamily, including
ICAM-1 (intercellular adhesion molecule-1), VCAM-1 (vascular cell adhesion molecule-
1), and PECAM-1 (platelet–endothelial cell adhesion molecule-1). ICAM-1, the
counterligand for CD11/CD18 on neutrophils, is constitutively expressed on
 endothelial cells and is upregulated by cytokines 2 to 4 hours after myocardial
ischemia-reperfusion (83), coincident with the upregulation of CD11/CD18.
Subsequently, a high-affinity bond between the neutrophil and endothelial cell occurs
via integrins (CD11/CD18) on neutrophils and ICAM-1 and ICAM-2 on endothelial cells.
Once the neutrophil has firmly attached to the endothelial surface, it migrates through
the endothelial cell junctions to the myocardium, partly mediated by IL-8. The
shedding of L-selectin from the neutrophil and the presence of PECAM on the
endothelium may be required for neutrophil transendothelial migration. The
prerequisite nature of this interaction between neutrophils and adhesion molecules
makes anti-neutrophil and anti-adhesion therapy a potentially effective strategy, as
these agents intervene at proximal points in the cascade (discussed below).

Evidence supporting the rapid accumulation of neutrophils within ischemic-reperfused

myocardium following the neutrophil–endothelial cell interactions was appreciated not only by
our laboratory, but also by Sommers and Jennings (84). We (85,86,87,88,89) and others (90)
have demonstrated an association between the extent of postischemic injury and neutrophil
accumulation within ischemic-reperfused myocardium as well as a reduction of postischemic
injury associated with neutrophil suppression. As shown in Figure 9.4, Lefer et al. (91)
demonstrated that neutrophil adherence to coronary artery endothelium occurs in advance of
accumulation in extravascular sites, representing the lag time involved in transmigration of
intravascular neutrophils. In addition, neutrophil accumulation occurs in advance of the
development of necrosis, consistent with active neutrophil participation in cell demise. Brix-
Christensen et al. (92) reported that neutrophils accumulated in the hearts and other organs of
neonatal pigs placed on CPB. Therefore, neutrophil accumulation occurs in both postischemic
and post-CPB hearts.

Figure 9.4. The postischemic time course of endothelial injury, contractile dysfunction,
neutrophil (PMN) adherence and accumulation in tissue, necrosis, and apoptosis. These events
are indicated on the y-axis as a percent completion of that event (0% not started, 100% is
complete). Note that PMN adherence parallels that of endothelial dysfunction representing
PMN-mediated damage. Note also that accumulation in tissue lags behind adherence
representing emigration time. (Data obtained, in part, from Zhao ZQ, Nakamura M, Wang NP,
et al. Dynamic progression of contractile and endothelial dysfunction and infarct extension in
the late phase of reperfusion. J Surg Res 2000;94:133–144.) Apoptosis is a slower process than
necrosis as observed by Zhao et al. (565).(Adapted in part from Lefer AM, Tsao PS, Lefer DJ, et
al. Role of endothelial dysfunction in the pathogenesis of reperfusion injury after myocardial
ischemia. FASEB J 1991;5:2029–2034.)

The importance of neutrophils in postbypass reperfusion injury has been well-established in

both animal and human models. The myocardium is exposed to neutrophils during the delivery
of intermittent or continuous blood cardioplegia, or during reperfusion (after cross-clamp
removal) following either crystalloid or blood cardioplegia. Schwartz et al. (93) showed that
CPB circuitry directly “primes” neutrophils by depositing C3b on extracorporeal tubing, leading
to neutrophil activation and sequestration. Further evidence for the activation of neutrophils
following CPB was suggested by the finding of increased neutrophil surface expression of
CD11b/CD18 and decreased surface expression of L-selectin on cells circulated in simulated
CPB circuits (94,95). Finn et al. (96) noted in patients undergoing CPB that a rise in circulating
IL-8 begins at reperfusion with rewarming of the patient after systemic hypothermia, and
closely correlates with the rise in neutrophil count and circulating elastase. Similarly, Schwartz
et al. (93) and Gessler et al. (97) found higher levels of the pro-inflammatory mediators soluble
IL-6 and IL-8 6 hours after bypass than before bypass.

The time course for the upregulation of neutrophils and various adhesion molecules is equally
important in designing treatment strategies for patients undergoing extracorporeal bypass. In
a clinical study, Galiñanes et al. (98) investigated the perioperative temporal changes in the
expression of various neutrophil surface adhesion molecules, concluding that downregulation
of PECAM-1 is an early indicator of neutrophil activation and may therefore represent a target
for therapy aimed at reducing the inflammatory response associated with CPB.

The role of temperature in the activation of neutrophils or other inflammatory mediators

following CPB in humans has been investigated. Hypothermia has been reported to attenuate
expression of adhesion molecules in some models of tissue injury (99,100). Some investigators
have reported minimal or no difference in neutrophil activation and expression of neutrophil
adhesion molecules between normothermic (34°C–37°C) and hypothermic (28°C–30°C) CPB
(101,102), whereas others note an attenuation in the expression of IL-8 and neutrophil
elastase activity during normothermic CPBsurgery (103). Hypothermia has been shown to
delay, but not prevent, the increased expression of adhesion molecules such as CD11b and
CD11c (101). When measured 4 hours after bypass, the levels of IL-1α, soluble ICAM-1 (sICAM-
1), and elastase are higher after normothermic CPB than after hypothermic CPB (104,105). In
conclusion, further research is required to determine the most appropriate temperature of
cardioplegic administration to minimize the deleterious consequences of neutrophil-mediated
damage. Interventions designed to attenuate neutrophil-related myocardial injury following
CPB will be discussed later.
Complement in CPB and Myocardial Reperfusion Injury

The complement cascade, particularly the alternative pathway stimulated by contact

activation, is activated independently by myocardial ischemia-reperfusion and CPB, and
contributes importantly to the pathologic sequelae of CPB (106). Complement fragments such
as the anaphylatoxins C3a and C5a are generated and released both systemically and locally by
the ischemic-reperfused myocardium (107). The activation and migration of neutrophils to
ischemic-reperfused myocardium requires stimulation by chemotactic factors, most notably
complement fragments (C3a and C5a), eicosanoid products such as leukotriene B4, and
platelet-activating factor (108,109). Complement fragments (C5a) can be found in lymph from
ischemic myocardium (110), are temporally associated with reperfusion (111), and correlate
with increased neutrophil accumulation in reperfused myocardium. Furthermore, C5a directly
stimulates neutrophil −O2 production and enhances adherence to coronary artery
endothelium (112), whereas C3 activates monocyte adherence (113). The chemoattractant and
chemotactic properties of C5a cause progressive generalized vascular leukosequestration (114)
during CPB. The interaction between blood and extracorporeal surfaces also stimulates
deposition of C3b (115), which in turn amplifies the complement alternative pathway to
release C3a and C5a, and subsequently form the membrane attack complex (C5b 9) (115).
Accordingly, plasma levels of the anaphylatoxin C5a increase in patients undergoing CPB
(115,116), which activates neutrophils to generate −O2 and promotes adhesion to endothelial
cells, stimulates degranulation of mast cells, stimulates vascular smooth-muscle contraction,
and increases vascular permeability. Complement activation constitutes an early-phase
response to CPB, which then progresses to a second, slower response phase in which de novo
synthesis of proteins (E-selectin, ICAM, IL-8) and expression of adhesion molecules on cell
surfaces occurs. The magnitude of the complement response has been linked to the duration
of CPB and the type of circuitry used (116). The combined stimuli of myocardial ischemia-
reperfusion and CPB may synergize to exacerbate postischemic neutrophil-mediated
inflammatory responses in patients undergoing CPB for myocardial revascularization.

Actions of Toll-like Receptors

Recent evidence has shown that activation of TLRs is involved in the pathogenesis of
reperfusion injury (117,118). TLRs are transmembrane receptors expressed on endothelial
cells, cardiomyocytes, platelets, monocytes, and neutrophils. Cardiomyocytes participate in
the inflammatory and innate immune response to ischemia-reperfusion by releasing cytokines
(TNFα, IL-1β, IL-6, IL-8, IFN-γ), chemokines, and by expressing cell surface adhesion molecules.
TLRs are pattern recognition receptors (PRRs) that recognize pathogen-associated molecular
patterns (PAMPs) such as LPS or viral RNA on foreign pathogens, and sense endogenous
danger molecules and signals from injured cells. TLRs also recognize and are activated by
endogenous ligands such as danger-associated molecular patterns (DAMPS) (e.g., heat shock
proteins, HSP70) induced in cardiac tissue during stress (119) and cardiac surgery (120,121).
Inducible HSP70 has been found in plasma from CABG patients immediately after surgery
(122). ROS may also stimulate TLRs. Ten to eleven human TLRs have been identified; (123,124)
TLR receptor-1, -2, -4, -5, and -6 are located on the cell surface, with TLR-2 and TLR-4 being the
most involved in ischemia-reperfusion of the group of TLRs; TLR-2 works with other TLRs in co-
receptor cooperativity to recognize diacyl- and triacyl-lipoproteins, while TLR4 recognizes LPS
from gram-negative bacteria and DAMPs such as HSP70. Neutrophils constitutively express all
TRLs except TLR3.

The stimulation of TLRs during ischemia-reperfusion and cardiac surgery can lead to cell injury.
Dybdahl et al. (122) reported the release of HSP70 into the circulation of CABG patients and a
concomitant increase in monocyte TLR-2 and TLR-4 one day after surgery. Human monocytes
exposed to HSP70 released IL-6 and TNFα. Stimulation of TLRs by interaction with DAMPs
during ischemia-reperfusion is transduced by numerous molecular signaling pathways, which
ultimately results in the activation and translocation of NF-κB to the nucleus, where it
stimulates gene expression and release of cytokines and innate inflammatory factors. ROS
generated during reperfusion also stimulate TLR-dependent (most likely TLR-4) generation of
NF-κB through the PI3-K/Akt pathway, which suggests cross talk between the TLR and
reperfusion injury survival kinase (RISK) pathways. Inhibitors of TLRs, intermediary pathways,
or NF-κB will attenuate ischemia-reperfusion injury.

The Mitochondrial Permeability Transition Pore as a Determinant of Cell Death

Mitochondria are best known for being the power house of the cell, where ATP is generated
through oxidative phosphorylation. However, the mitochondria have been shown to function
as a “molecular switch” that governs cardiomyocyte viability; indeed it is a critical fulcrum
point in the transition from reversible to irreversible injury. Mitochondria can direct the cell to
pursue either an apoptotic or a necrotic pathway to cell death. The key event of this molecular
life/death switch is opening of the mPTP. This pore is a transmitochondrial membrane voltage-
dependent and Ca2+-dependent high-conductance channel located in the inner mitochondrial
membrane that permits water and solutes with a mass up to 1.5 kDa to enter the
mitochondrion. Under normal physiologic conditions, the mPTP is in a closed state. It opens in
response to high levels of intracellular ROS, Ca2+ and inorganic phosphate when mitochondria
matrix and intracellular pH are in the normal range. The mPTP remains closed during ischemia
because intracellular acidosis and the intracellular levels of Mg2+ and ADP counteract the
opening triggered by ROS and Ca2+. However, the mPTP opens during the early moments of
reperfusion (125) when there is an accumulation of ROS, Ca2+ and inorganic phosphate, and
when there is a rapid renormalization of intracellular and mitochondrial matrix pH (126). The
opening of the mPTP increases the permeability of mitochondria to water and small solutes,
causing swelling of the matrix and collapse of the mitochondrial transmembrane potential that
drives oxidative phosphorylation and ATP generation. mPTP opening also stimulates the
release of the pro-apoptotic factors cytochrome c and apoptosis-inducing factor (AIF). The
switch function comes into play based on the availability of intracellular ATP; if ATP is present
then the cell pursues an apoptotic pathway because apoptosis requres ATP. If ATP levels are
depleted then the cell pursues a necrotic pathway. With prolonged antecedent ischemia mPTP
is irreversible, but milder or short periods of ischemia trigger transient reversible opening
(127).
CHAPTER 9

SURGICAL MYOCARDIAL PROTECTION

Jakob Vinten-Johansen and Neil J. Thomas

INTRODUCTION

The science of myocardial protection had its beginnings in the 1950s, and has enjoyed a robust
and often controversial development since. Great strides were made in the 1980s and 1990s
which shaped the cornerstones of the cardioplegia strategy that we largely use today
(crystalloid or blood, warm, tepid or cold, continuous or intermittent, antegrade or
retrograde). Mortality has substantially decreased over the past decades as a result of applying
cornerstone as well as novel cardioprotective strategies and new technologies.

However, the patient of the 1980s and 1990s during which cardioplegia was developed has
morphed considerably in the 2000s. Active lifestyles have changed to more sedentary lifestyles
beginning in the early school years with less emphasis placed on physical activity, and more
distractions from active lifestyles imposed by the digital-age gadgets. This, coupled with the
availability of fast food with high caloric, high fat, and high sodium content, has conspired to
radically change our body habitus and biology; indeed, diabetes, obesity,
hypercholesterolemia, and the metabolic syndrome are more prevalent today. These more
prevalent comorbidities have impacted the efficacy of modern surgical therapies and the
outcomes (1) of myocardial protection strategies. Experimental animal models in which
effective cardioplegia strategies were developed decades ago, and observed to be effective in
the healthier patients with normal myocardial tissue substrate preoperatively in which they
were initially tested clinically seems to apply in fewer and fewer patients today. The “one
formula-fits-all” strategies of the past must be rethought in the current era. With patients
living longer as well, advancing age and the more senescent myocardium can display altered
responsiveness to some therapeutics. So we may ask, is myocardial protection developed in
the 1980s and 1990s being used as effectively on the patient of the 2010s? And, furthermore,
the yardstick by which our results are measured is changing in the world of health care today,
where a 30-day surgical mortality rate may no longer be the gold standard. Our results must
now be viewed in terms of longer-term, event-free survival and with measure of long-term
myocardial performance after an index event or surgery.

The concepts of cardioprotective strategies have changed significantly as our understanding of

the pathophysiologic determinants of ischemia, and particularly the complex mechanisms
involved in reperfusion, have expanded. Early empiric, largely physiologic observations have
given way to complex biochemical and molecular interactions, which in some cases have
reinforced physiologic intuition and in other cases have redirected our thinking, but has
nonetheless provided a scientific underpinning to a more molecular approach in developing
target-specific strategies of myocardial protection. This chapter will discuss the pathobiology
of ischemia-reperfusion injury, provide the basis by which the severity and duration of evolving
ischemic conditions set the stage for reperfusion and oxygen-related injury and its impact on
affected tissues as a basis for understanding cardioprotective strategies to attenuate injury. It
will also discuss new, emerging concepts of myocardial protection that have developed out of
a better understanding of the pathobiology of ischemia-reperfusion injury, and a growing
appreciation that the preoperative condition of the tissue may be equally, if not more
important, than any intraoperative strategy employed in previously normal tissue. It will also
discuss the all-elusive “golden ring” of arresting the heart in a nondepolarized state.

The goal of this chapter is to provide the reader with scientific and practical knowledge of
myocardial protection that can be applied to design appropriate strategies to meet the need of
specific patients, their unique demographics and clinical scenarios, and to provide flexibility in
the approach based on both preoperative condition and intraoperative events. It will also
provide some guided reading to enhance the background of relevant topics

HISTORICAL PERSPECTIVES ON MYOCARDIAL PROTECTION AND REPERFUSION INJURY

The need to protect the myocardium from intraoperative and postoperative damage was
realized before the development of cardiopulmonary bypass (CPB). Temporary inflow
occlusion techniques to limit blood flow to the heart permitted the repair of simple
intracardiac defects, and in many cases the occlusion time could be increased by adding
moderate systemic hypothermia (30°C) (2). After the development of CPB by Gibbon and
colleagues in 1954 (3,4), profound systemic hypothermia could be used safely to provide a
hemodynamic safety net, and to provide protection of the heart and other body organs by
permitting longer periods of ischemic cardiac arrest for the repair of more complicated lesions.

The development of extracorporeal circulation techniques allowed cardiac procedures to

proceed without threat of hemodynamic collapse. However, performing complicated
procedures on the heart was made difficult by its continuous motion and blood in the field.
Ischemia was known to achieve cardiac arrest (ischemic or anoxic arrest) through the
depletion of high-energy phosphate stores on which cardiac contraction depends. However,
myocardial ischemia led to necrosis and contractile dysfunction upon reperfusion, the worst of
which was an irreversible “stone” heart (5,6). The conundrum of surgeon convenience at the
expense of acceptable outcomes provided a driving force for innovation.

Once the severe complications of ischemic arrest were recognized, chemical methods were
used to arrest the heart. Based on reports from Ringer in 1883 that high concentrations of
potassium would arrest the heart (7), Melrose and associates in 1955 (8) and 1957 (9) used
hyperkalemia to induce asystole during CPB. The combination of immediate arrest and
hypothermia would ostensibly reduce energy (i.e., oxygen) demands of the heart, and thereby
avoid injury imposed by hypothermia alone. Melrose made a hyperkalemic solution by mixing
potassium citrate solution (77 mM) with 18 mL of blood, which was then infused by hand-held
syringe into the aortic root proximal to the aortic cross-clamp. This hyperkalemic arresting
solution was later called “cardioplegia” solution by Lam et al. (10) and Sergeant et al. (11).
Other contemporaries of Melrose used alternative arresting agents, such as acetylcholine
alone (10,11) or in combination with other drugs (12).

Although asystole was indeed achieved with the Melrose potassium citrate solution, the
concept of elective cardiac arrest was abandoned in the United States for some 20 years
following reports of high morbidity and pathologic complications in chemically arrested hearts
(13). Chemical cardioplegia was supplanted by alternative techniques of arresting the heart or
limiting blood flow in the surgical field, including ventricular fibrillation, intermittent aortic
cross-clamping, and profound systemic (14) or topical hypothermia (15). Laboratory studies
demonstrated that topical hypothermia protected inadequately against intraoperative injury,
which resulted in postoperative myocardial subendocardial necrosis, and postischemic
metabolic and functional depression (16,17,18). In the late 1960s, reports described scattered
myocardial or subendocardial necrosis in patients who had died after otherwise technically
successful cardiac surgical operations, suggesting that the current techniques of myocardial
protection were inadequate (19,20). This set the stage for a later re-emergence of chemical
cardioplegia in the United States.

During the surgeon-imposed moratorium on the use of the Melrose solution, cardioplegia
solutions continued development and clinical use in Europe. Bretschneider et al. (16) and
Kirsch et al. (17) used crystalloid formulations which mimicked intracellular ionic
concentrations (intracellular crystalloid solutions). The Bretschneider HTK (histidine,
tryptophan, α-ketoglutarate)-ketoglutarate) solution used low calcium, low sodium, and
procaine with histidine buffers to achieve a nondepolarized arrest, rather than hyperkalemia-
induced depolarized arrest. In the 1970s, Hearse and associates (18,21) at St. Thomas’ Hospital
in London developed an extracellular hyperkalemic solution based on a rigorous series of
experiments that established a scientific understanding of the determinants of injury incurred
during global ischemia. Braimbridge introduced St. Thomas’ cardioplegia solution into clinical
use in 1975 (22). St. Thomas’ solution was later modified based on further experimental
studies to the St. Thomas’ solution No. 2 (23), an intracellular ionic concentration,
normocalcemic, and hyperkalemic (16 mmol/L) solution; it is currently marketed as Plegisol in
the United States. The Bretschneider HTK solution is currently marketed as Custodiol HTK
solution.

Chemical cardioplegia was revived in the United States by Gay and Ebert (24) and Tyers et al.
(25,26), who found that the constituents in the Melrose formulation were inappropriate rather
than the concept of hyperkalemic chemical cardioplegia itself being erroneous. These reports
popularized the use of potassium-based cardioplegia to achieve electromechanical arrest. In
the late 1970s and early 1980s, Follette and colleagues from Dr. Buckberg’s laboratory (see
reference 224 below) introduced the concept of cold hyperkalemic blood cardioplegia. The
physiologic attributes of blood, including its superior buffering capacity, endogenous oxygen
radical scavengers, detoxifying substances, and superior oxygen-transport capacity, has
established this concept as a popular strategy that has facilitated surgery and vastly improved
patient outcomes. Today, a variety of both crystalloid and blood cardioplegia solutions are
used clinically worldwide to achieve elective cardiac arrest and a bloodless field.

The cornerstones of myocardial protection that had been laid by the end of the early 1980s
are: (1) rapid chemical arrest to conserve high-energy phosphate stores and reduce ischemia,
(2) hypothermia to reduce the metabolic rate, (3) adjunctive agents to reduce the effects of
ongoing ischemia (anti-ischemic agents), and (4) continuous or intermittent delivery to restore
or maintain tissue oxygenation. Further developments in cardioplegia-related myocardial
protection strategies focused on (1) optimizing delivery (retrograde, gentle cardioplegia
delivery pressures), (2) overcoming the adverse effects of hypothermia, (3) identifying
metabolic enhancements (ATP generation and pH management) and buffers, (4) managing
calcium (Ca2+) accumulation, and (5) understanding the pathophysiology of ischemia and
reperfusion, which would drive further developments of strategies to avoid these mechanisms
of both ischemic and postischemic patterns of injury. There has also been a persistent search
for agents that induce polarized arrest which overcome the disadvantages of hyperkalemia
(reviewed in Dobson et al. (27) and Maruyama et al. (28)). With the recent rapid developments
in molecular and cellular biology, experimental focus has shifted away from the organ and
physiologic aspects toward cellular and molecular areas to improve the understanding of the
effects of ischemia-reperfusion injury, inflammation, and cell death processes. Understanding
the mechanisms involved in the transition to cell death would direct developments in Ca2+
handling and preventing Ca2+ dyshomeostasis, the role of oxidant injury, and a newly
recognized role of mitochondria in determining the transition from reversible to irreversible
injury, and whether the cell pursues a necrotic pathway or apoptotic pathway to cell death.
Finally, the concept of conditioning, which recruits complex endogenous biologic mechanisms
of self-protection against stressors such as ischemia-reperfusion injury, oxidants and
apoptosis, has been incorporated into pretreatment (preconditioning), intraoperative
(preconditioning), and reperfusion (postconditioning) phases of cardiac surgery. Although
human beings have been evolving for hundreds of thousands of years, it has only been in the
past 60 years or less that the human heart has been subjected to the kind of induced ischemia
and surgical reperfusion discussed in these pages.

PATHOPHYSIOLOGY OF SURGICAL ISCHEMIA-REPERFUSION INJURY

Because many factors affect postoperative outcomes of hearts with various and complex
preoperative pathophysiologic features, developing strategies that adequately protect all
hearts during cardiac surgery is often problematic. The heart under consideration may present
a complex picture of preoperative disease, variable age, gender, metabolic and cellular
dysfunction, global or regional ischemia (both acute and chronic), each of which contribute to
differing vulnerabilities to ischemia and reperfusion injuries. However, adequate myocardial
protection must be based on a sound scientific basis. Obviously, a single cardioprotective
strategy may be inadequate to suit all surgeon preferences and target all patient pathologies.
Hence, the surgical team must often mold the composition of the cardioplegic solution or its
modality of delivery to meet the requirements of the patient’s pathologic profile as presented.
The measures taken to protect the heart during elective cardiac arrest should represent a
balance struck between the requirements of the heart during aortic cross-clamp or
reperfusion, with appropriate adjustments made for special considerations (hypertrophy,
diffuse coronary disease), and avoid distraction from the surgical procedure. Ischemia and
reperfusion injury (29) and the choice of cardioplegia technique (30) represent primary
contributors to patient outcomes.

Ischemic Injury

When Does Ischemia and Reperfusion Injury Occur during Cardiac Surgery?
Figure 9.1. Events occurring before the institution of cardiopulmonary bypass, during delivery
of cardioplegia and at reperfusion that predispose the myocardium to ischemia and
reperfusion injuries. Arrhythmias, ventricular fibrillation (VF) or severe hypotension can cause
antecedent global ischemia. During the delivery of cardioplegia, the composition of the
solution can cause edema, further ischemia if oxygen and nutrients are inadequate to meet
ongoing albeit lower demands, and Ca2+ concentration is either too high or too low. Coronary
obstructions, air or low cardioplegia infusion pressure can lead to maldistribution of solution.
Too high infusion pressure can lead to microvascular injury and edema. Potassium itself has
negative effects secondary to placing the cell in a depolarized state. Ischemia may be
encountered if air emboli, graft kinks, tight anastomoses, or systemic hypotension cause
inadequate distribution of blood flow. Finally, reperfusion injury can cause reversible (stunning,
arrhythmias, edema) or irreversible (necrosis, apoptosis) injury.

It is critical for the surgical team to understand that the presence of acute, ongoing and
evolving or progressing ischemia must impact the protection strategy. Ischemia-reperfusion
injury can occur in the surgical setting at three major time points during the surgical procedure
(summarized in Fig. 9.1).

(a) Before CPB has been instituted or cardioplegia solution has been delivered:
“Unprotected” clinical ischemia before CPB usually results from the presenting acute coronary
syndrome and can be exacerbated by hypotension, arrhythmias (e.g., atrial or ventricular
fibrillation), and other hemodynamic changes (e.g., cardiogenic shock, coronary artery spasm).
In most cases, acute ischemia is active in both the distribution of the principally involved artery
(cardiologists call this the “infarct artery”) but also distributions in which preexistent collateral
flow may be compromised. Patients referred for bypass surgery in these circumstances almost
always have multi-vessel disease patterns making the myocardial substrate dangerous from
the standpoint of the severity and duration of the syndrome, which can be extremely difficult if
not impossible to determine. Reperfusion and reoxygenation injury occur when these
conditions change abruptly before effective measures are implemented. Recent attention has
been paid to avoiding over-oxygenating such patients systemically, and this will be discussed in
more detail in later sections.

(b) At the initiation of cardiopulmonary bypass: At the moment bypass is initiated, blood is
immediately diverted from the right atrial drainage system into a venous reservoir where it is
then pumped through the oxygenator. This configuration does more to the heart than merely
oxygenate venous blood and send it back to the ascending aorta. As the heart is emptied into
the venous side of the pump, the oxygen demand of the metabolizing myocardium decreases
by 50%, but the high oxygen content of blood extracorporeally oxygenated makes the tissue
vulnerable to reoxygenation damage if no steps are taken to mitigate this injury. This might be
viewed as an abrupt interruption of antecedent ischemic conditions. While this is one of the
principal goals of bypass (to give the heart a “rest”), it must be recognized that the heart is also
vulnerable to a pattern of injury that has been, at best, underappreciated in the past. Ihnken et
al. (31) suggested as early as 1998 that initiation of bypass is also, perhaps paradoxically, a
dangerous moment for the heart. In this work, he observed clear, reproducible biochemical
evidence of injury in samples of blood drawn from the coronary sinus. Since Buckberg (32)
published his work on the reoxygenation injury in 1995, it has been observed that abrupt over-
exposure of the ischemic heart to oxygen increases injury in a pO2-dependent pattern of injury
that can be mitigated by altering the oxygen tension of the perfusate.

(c) During the “protected” period of arrest and delivery of cardioplegia: Ischemia or
reperfusion injury can occur during the cardioplegia phase. Ischemia can develop or be
worsened under several circumstances (1) the unintentional maldistribution of cardioplegia
solution distal to stenotic or totally occluded coronary arteries, (2) between intermittent
infusions of cardioplegia solution, (3) during interruption of continuous cardioplegia strategies,
(4) inadequate delivery of retrograde cardioplegia to areas of the right heart whose venous
blood drains directly into the Thebesian system, or areas that are simply under-filled due to
malposition of a retrograde delivery device. Though the cardioplegia phase is considered
“protected time,” the potential for worsening of, or the development of, ischemic conditions
followed by reperfusion injury can occur with the initial administration of cardioplegia and
with each subsequent delivery of cardioplegia solution [(re)perfusion] for a variety of reasons:
(1) excessive pressure (microvascular injury), (2) low onconicity favoring extravasation of fluids
into the extracellular space, (3) formation of oxygen radicals from over-oxygenation and
oxidative stress, and (4) other mechanisms of reperfusion injury as discussed below. While
delivery of cardioplegia should provide protection, this phase also offers opportunity for
additional injury to occur unintentionally. Any surgeon who has been faced with postoperative
low-cardiac output and who suspects inadequate protection knows this to be true.

d) After release of the cross-clamp and attempts to reanimate the heart: Additional ischemia
may be encountered at the release of the cross-clamp when (1) coronary blood flow is
impaired through kinked grafts or (2) tight anastomoses (3) air emboli are present in the
coronary arteries, (4) ventricular fibrillation occurs at low perfusion pressures, or (5) poor
ventricular performance or dysrhythmias cause hypotension once the heart is converted and
off bypass. Clearly, when the clamp is removed reperfusion injury can occur as blood flow is
restored and oxygen exposure is reestablished after any of the above events. In terms of
surgically induced ischemia, this is intuitively the major time when one thinks of reperfusion
injury occurring.

Understanding the pathophysiologic mechanisms of myocardial injury occurring at these

intervals and applying the principles of myocardial protection will help avoid injury induced by
the very techniques intended to preserve the myocardium. Unlike the cardiologist, the surgical
team has the opportunity to intervene directly to control many facets of ischemia and
reperfusion injury, and to limit at some point the mechanisms causing injury, thereby more
favorably directing the outcome of the postischemic heart. The pathophysiology of ischemia
and reperfusion is described in the following sections. Interventions targeting these various
pathologic processes are discussed under Myocardial Protection Therapy.

Determinants of Ischemic Injury

Ischemia encountered either before arrest or during cardioplegic arrest, is a major cause of
postcardioplegia injury. Ischemia sets the stage for reperfusion injury; without ischemia you
cannot have reperfusion injury. Ischemia is defined as inadequate energy supply to meet the
current demands, that is, an energy supply/demand ratio of less than 1; normally, the O2
supply/demand ratio is >1. Since the adenosine triphosphate (ATP) necessary to support the
energetic needs of the heart is provided predominantly by oxidative phosphorylation, with
scant generation of high-energy phosphates through anaerobic metabolism, the heart is nearly
an obligate aerobic tissue. Therefore, the energy consumption of the heart can be
approximated by the oxygen consumption. The limited availability of anaerobic energy-
generating pathways mandates that O2 be in constant supply. A very brief discussion of the
determinants of O2 supply and demand follows.

Oxygen supply: Ischemia may be caused by either a decrease in O2 supply relative to demands
(supply ischemia) or an increase in O2 demands relative to supply (demand ischemia). O2
supply is determined by the O2 extraction (arterial-coronary venous O2 difference) and
coronary blood flow. Since O2 extraction by the heart is normally ~75%, and its physiologic
limit is approximately 95%, only limited amounts of oxygen may be obtained through
additional extraction. The majority of O2 is therefore supplied by adjustments in coronary
blood flow, which under normal conditions may increase 4- to 5-fold to meet demands. This
O2 is largely present as hemoglobin-bound, with only a minor fraction (~1.5%) dissolved in
plasma. This limited dissolved O2 will become important when we discuss crystalloid
cardioplegia solutions in which oxygen availability is limited to that dissolved in solution.

Oxygen demand: Overall myocardial oxygen demand of the normal working heart is
determined by the work of the entire heart. The majority of this overall O2 consumption is
attributed to utilization by the left ventricle. In a nutshell, the determinants of left ventricular
O2 demand include pressure–volume work or wall stress, heart rate, temperature, inotropic
state, basal metabolism, and ionic homeostatic mechanisms (ATP-dependent pumps) required
to re-equilibrate ionic balance after electromechanical activity. In the immediate postoperative
heart, energy demands may be related to oxidative energy diverted to myocyte repair.

The energy related to maintain ionic equilibrium (~5% of total demand) is an ongoing demand
even when the heart is not contracting since ATP-dependent pumps are constantly adjusting
calcium and sodium influx, as well as potassium efflux. This energy demand of ionic
equilibration is important in determining ongoing oxygen demands in the arrested heart, as
discussed below. How oxygen demands are changed by hypothermia is discussed later.

Severity and Duration of Ischemia in Cardiac Surgery

The potential for myocardial injury is related, in part, to the duration of ischemia. The time to
onset of irreversible ischemic injury will depend on many factors, including the severity of
ischemia, myocardial temperature, ambient energy demands, and collateral blood flow.
However, irreversible injury can become apparent after as little as 30 to 45 minutes of
coronary occlusion in the working myocardium. However, shorter durations of global ischemia
can result in mild to severe systolic and diastolic dysfunction without irreversible tissue
necrosis (left ventricular “stunning” (33)). Therefore, the duration of ischemia, both preceding
CPB and intraoperatively after application of the cross-clamp, has often been used as a
predictor of postoperative myocardial injury in experimental models, and is one of the
underlying considerations in the correlation between total ischemic time (CPB, cardioplegia)
and clinical outcomes. Because the surgical team often does not have control over the
duration of warm ischemia (particularly antecedent ischemia caused by coronary occlusive
disease), other aspects of elective ischemic time are more often targeted for modification to
reduce the severity of ischemia and hence postischemic injury. Therefore, strategies such as
(1) initiating rapid asystole, (2) venting the left ventricle to reduce consequences of chamber
distension resulting in increased wall stress, and (3) imposing cardiac hypothermia all
specifically target the reduction of myocardial energy/oxygen demands, thereby increasing the
limits of “safe” ischemic time during aortic cross-clamping. With these cardioprotective
strategies in place, the duration of aortic clamping that can be safely imposed can be increased
from as little as 15 to 45 minutes to several hours using hypothermic cardioplegia (Fig. 9.2).
Reducing to almost zero the determinants of myocardial work and oxygen demand is the sine
qua non of myocardial protection. In the section on future directions, eliminating ischemic
conditions without additional oxygen-related injury and changing the milieu more gradually
will be discussed.

Figure 9.2. Left ventricular performance measured by in situ Starling curves (A) or end-systolic
pressure–volume relations (conductance catheter) (B) after 45 minutes of normothermic
unprotected global ischemia and reperfusion. C: Left ventricular stroke work index in patients
before (Baseline) or after (pCPB) 15 minutes, or 4 and 24 hours postcardiopulmonary bypass
(CPB). The dashed line between 4 and 24 hours after discontinuation of CPB represents further
decreases in cardiac performance secondary to onset of irreversible injury.(A, adapted from
Rosenkranz ER, Buckberg GD. Myocardial protection during surgical coronary reperfusion. J Am
Coll Cardiol 1983;1:1235–1246, with permission.)

The determinants of surgical ischemic injury can be summarized as:

● The duration and severity of antecedent” and “unprotected” ischemia: Although it has
been recognized for decades that ischemia time (duration) impacts tissue injury and
postischemic salvage and function, an increasing appreciation for the caveat that the
 manner in which the tissue is reperfused significantly alters the outcome. It is now
increasingly understood that reperfusion injury can occur by merely eliminating
ischemic conditions abruptly or in the presence of an inappropriate oxygen gradient. In
the case of the imposition of aortic cross-clamping, there is total elimination of native
flow with the exception of that supplied by noncoronary collateral flow. Global
contractile function may be impaired before necrosis or apoptosis are evident
(“stunning”). It is the reduction of myocardial oxygen demand that protects the heart
in this circumstance.
● The elimination of electromechanical activity during cross-clamping is paramount
because its mechanical activity increases O2 demands even in the presence of
hypothermia (discussed further in the section on Hypothermia).
● Myocardial temperature: Temperature influences metabolism by the “Q10 effect,” in
which the tissue metabolic rate decreases by half for each 10°C decrease in
temperature; conversely, myocardial metabolism doubles when the heart rewarms by
10°C. Hypothermia can buy biologic time during ischemia.
● The nutritional and stress status of the heart: For example, catecholamines increase
the O2 demands by increasing contractility and exerting an O2 wasting effect.
● Presentation of comorbidities such as hyperlipidemia which increase the vulnerability
to ischemia-reperfusion injury for any given duration of ischemia. This effect is
independent of energy demands, but is influenced more by mechanisms exerted at
reperfusion, that is, the capacity to generate O2 radicals, a pro-inflammatory state of
the endothelium and inflammatory cells, and calcium handling by the tissue.

Determinants of Reperfusion Injury

Although reperfusion of the myocardium is the ultimate goal of both surgical and nonsurgical
revascularization, abrupt reperfusion carries with it the potential for extending postischemic
injury to cardiomyocytes, coronary vascular endothelium, and the microvasculature. Indeed,
reperfusion injury associated with rapid changes in circulatory conditions may be responsible
for ~50% (34) or more of the ultimate infarct size, and is an important contributor to
postsurgical mortality and morbidity as well (29). We can define surgical reperfusion injury as
an incremental increase in the pathology that is observed after ischemic injury alone.
Reperfusion injury extends or accelerates damage from that observed during ischemia alone,
and occurs after the onset of reperfusion (i.e., cardioplegia) or reperfusion (i.e., cross-clamp
removal). Viewed another way, reperfusion injury can occur at the instant of the elimination of
ischemic conditions with reestablishment of flow. Hence, reperfusion injury can be viewed as
representing an increment of injury that had begun during ischemia but progresses or
manifests during the reperfusion phase. Other investigators have described de novo injury
starting at reflow or, in a newer construct, at the time of the elimination of ischemia (even if
that is accomplished by a reduction in oxygen demand). Reperfusion injury can be categorized
as either reversible or irreversible. (1) Reversible injury includes temporary contractile
dysfunction (“stunning”) in the absence of morphologic injury or necrosis, or (2) irreversible,
which includes necrosis and apoptosis. Biomarker enzymes such as creatine kinase (CK) or
cardiac troponin T (cTnT) or I (cTnI) released by disrupted cells have been observed to be
elevated during reperfusion rather than during ischemia. In the past, debate has often been
spirited regarding whether reperfusion injury even exists (35,36,37,38) and whether
reperfusion injury kills cells that are salvageable at the end of ischemia. However, a consensus
has developed in both the surgical, cardiology, and cardiovascular research communities that
reperfusion is indeed an important component of postischemic cardiac injury in animal models
(39) and in humans (34,40) and that alterations in reperfusion strategies can mitigate injury.

Among the more important mechanisms in the pathophysiology of myocardial ischemia-

reperfusion injury are (1) the duration and severity of antecedent ischemia, (2) oxygen radicals
generated not only by neutrophils but also by myocytes and vascular endothelium, (3) sodium
(Na+) and Ca2+ influx and loss of normal intracellular Ca2+ homeostasis, (4) neutrophils and
neutrophil-derived products that constitute a generalized inflammatory response to
reperfusion, and (5) mitochondrial dysfunction and opening of the mitochondrial permeability
transition pore (mPTP). Interestingly, the presence of hyperoxic or even relatively hyperoxic
conditions can ramp up or trigger all of these mechanisms. As such, these mechanisms do not
operate independently, nor are they mutually exclusive. Importantly, reperfusion injury is a
malleable process that can be attenuated by utilizing a number of strategies targeting these
various perpetrators of injury. In the operating room during cardiac surgery, and more and
more frequently in the modern era in the interventional lab, various forms of circulatory
support techniques and cardioprotective drugs are employed. The realization that circulatory
support can manipulate reperfusion therapy by altering oxygen supply/demand relationships is
critical for any operator that employs it. Such pumps provide an avenue for both the reduction
of oxygen demand and for the delivery of therapeutic agents via the pump, which can be
administered intravenously or by using cardioplegia as a vehicle. Using cardioplegia offers a
unique opportunity to modify the conditions of reperfusion (hydrodynamics, temperature,
route of delivery) and the composition of the solution (blood, crystalloid, pH, metabolic
substrates, hypocalcemia, oxygen, pharmaceuticals) that target the pathophysiologic
mechanisms of injury. It is now also recognized that the initiation of circulatory support in it
itself offers an opportunity to mitigate injury by paying attention to conditions at the moment
support is begun and ischemic conditions are eliminated.

Reactive Oxygen Species (ROS) and Reactive Nitrogen Species (RNS)

Molecular oxygen is relatively inert by virtue of the shared electron pair in the outer molecular
shell. However, oxygen must be viewed as a pharmaceutical and potentially toxic agent in
cardiovascular medicine and surgery. ROS are radicals or unstable oxygen molecules derived
from oxygen. Important oxygen radicals are superoxide anion (−O2) and hydroxyl anion
(•OH).OH). However, other oxygen-related molecules that are not technically radicals are
considered biologically reactive nonetheless. An example of this would be hydrogen peroxide
(H2O2). In addition, hypochlorous acid (HOCl) and chloramines derived from neutrophils exert
potent oxidative injury to biologic tissues.

Free-radical molecules that are derived from nitrogen are termed RNS and include nitric oxide
(NO•), peroxynitrite (ONOO–), nitrogen dioxide (NO2•) and nitrosyl hydride (HNO). In vascular
endothelium, NO• is constitutively generated by endothelial nitric oxide synthase (eNOS or
NOS-3). eNOS activity is dependent on Ca2+-calmodulin, flavin, NADPH, and
tetrahydrobiopterin (BH4). NO• is generated by an electron from the guanadino nitrogen of L-
arginine to incorporate oxygen to generate NO • and the byproduct L-citrulline. NO• has an
extremely short half-life (seconds) in vivo, and binds tightly to hemoglobin. OONO− is formed
by a nonenzymatic biradical reaction between NO• and −O2. NO• has been linked to biologic
signaling in cardioprotection, particularly anti-inflammatory effects, endothelial function,
vasorelaxation (hence its early name as endothelial-derived relaxing factor, EDRF),
neurotransmission, immune regulation and defense mechanisms.

NO• homeostasis is key in normal regulation of blood flow and responses to stress, and in
preventing endothelial dysfunction. NO• homeostasis is a balance between NO • generation,
bioavailability, and degradation or quenching by oxidants such as −O2. NO • has pleiotropic
effects in the cardiovascular system by paracrine effects: (1) NO • is a physiologic regulator of
blood pressure and flow, and counteracts local and systemic vasoconstrictors such as
endothelin (ET-1); (2) it inhibits neutrophil adhesion to vascular endothelium; (3) it inhibits
platelet aggregation; (4) it reduces myocardial oxygen consumption (41); (5) it has positive
lusitropic and inotropic effects; (6) it scavenges −O2 and therefore is directly antioxidant; and
(7) it is anti-proliferative. The effects on oxygen consumption and inotropy are somewhat
controversial (42,43). The role of NO• in ischemia-reperfusion injury and cardioprotection is
discussed in more detail under the section on endothelial dysfunction and NO • donors.

Reactive Oxygen Species

The oxygen radical is highly reactive with a broad range of biologic substances including sugars,
amino acids, phospholipids, and DNA. With such a large catalog of biologic targets, ROS and
their metabolites are potentially toxic to cells, and at the same time they are critical in the host
defense mechanism and bactericidal activities of neutrophils and phagocytes (44) and
participate in cardioprotective signaling. In this dualistic role of ROS as good guy–bad guy, it is
proposed that low-level production of ROS, particularly from mitochondria, acts as a signal for
cardioprotective conditioning responses (preconditioning, postconditioning) and stimulation of
cardioprotective kinase pathways, while large-scale generation of ROS from inflammatory cells
or dysfunctional mitochondria is deleterious. In the ischemic-reperfused myocardium, ROS-
induced injury includes peroxidation of lipid components of cellular membranes leading to
damage to sarcolemmal cell membrane and other membranous organelles such as
mitochondria and sarcoplasmic reticulum (45). ROS also cause the mPTP to open when
generated in an environment of Ca2+ accumulation and normalized pH (conditions achieved
during reperfusion), which leads to a collapse of the membrane potential, loss of metabolic
capacity, and release of pro-apoptotic factors. This is discussed in more detail below under the
role of mitochondria in reperfusion injury. The vascular endothelium is particularly vulnerable
to ROS-mediated injury, which manifests as impairment of vascular endothelial function
through attenuated production of vasoactive and anti-inflammatory autacoids such as
adenosine (46,47,48) and NO• (49), and increased production of pro-inflammatory and
vasoconstrictive molecules. These injury mechanisms contribute significantly to postischemic
dysfunction, dysrhythmias, and morphologic injury such as necrosis and apoptosis.

Sources of Oxygen Radicals

Both enzymatic and nonenzymatic reactions produce oxygen radicals. The major reactions and
their sources are shown in Figure 9.3. ROS are generated by a number of sources, including
activated neutrophils, cardiac myocytes, mitochondria, and the vascular endothelium.
Figure 9.3. The species and sources of reactive oxygen species (ROS) generated by neutrophils,
cardiomyocytes and their mitochondria, and vascular endothelium. SOD, superoxide
dismutase; CAT, catalase; GP, glutathione peroxidase; MPO, myeloperoxidase; HOCl,
hypochlorous acid; e−1, electron; NOX, NAD(P)H oxidase.

Neutrophils: Activated neutrophils represent the major source of ROS in the heart under
stress, including −O2, •OH, and HOCl species. Neutrophils are activated by various chemotactic
factors stimulated by either CPB or ischemia-reperfusion, including the complement
anaphylatoxins C3a and C5a, f-Met-Leu-Phe, tumor necrosis factor alpha (TNFα), interleukin-8
(IL-8), and platelet-activating factor. These factors trigger a “respiratory burst” of ROS. CPB
(independent of ischemia) activates the complement cascade (50,51), with the subsequent
activation and adherence of neutrophils (52) to the vascular endothelium primarily, which
subsequently triggers the release of cytotoxic products, all of which contribute to tissue injury
in the ischemic-reperfused myocardium (52,53). During the respiratory burst that occurs
seconds after stimulation by cytokines or ischemia-reperfusion, the −O2 in the neutrophil is
produced by the membrane-associated NADP oxidase, which transfers an electron to
molecular oxygen (Fig. 9.3, reviewed in Sheppard et al. (54)). Production of H2O2 by
neutrophils or the reduction of −O2 to H2O2 by superoxide dismutase (SOD) provides
substrate for generation of •OH, or the oxidant HOCl by myeloperoxidase. HOCl reacts with low
molecular weight amines to give rise to lipophilic chloramines, which can promote membrane
lipid peroxidation.

Cardiomyocytes: Xanthine oxidase (XO) is a major source of ROS from cardiomyocytes.

Normally, XO exists mostly in the dehydrogenase form and uses nicotinamide adenine
dinucleotide (NAD) as an electron acceptor. Xanthine dehydrogenase is converted to the
oxidant-generating enzyme XO during ischemia. Both hypoxanthine and xanthine accumulate
during ischemia. Hypoxanthine is converted to xanthine by XO, which is in turn converted to
uric acid by the same enzyme, which forms −O2. During ischemia, however, two events favor
the xanthine oxidase reaction: (1) hypoxanthine accumulates as a result of the sequential
catabolism of purine high-energy phosphates (ATP, ADP, AMP) and deamination of adenosine,
and (2) the dehydrogenase form of the enzyme is converted to the superoxide-producing
oxidase form by a protease activated by increased levels of intracellular calcium (55). XO is
reported to be localized to the vascular endothelium in substantial quantities (56). However,
the importance of XO to the pathologic generation of ROS in humans remains controversial
(57).

Mitochondria: Mitochondria are a major noninflammatory site of ROS generation (58). ROS are
generated at complex I (NADH/ubiquinone oxidoreductase) and complex III
(ubiquinol/cytochrome c oxidoreductase). The latter site catalyzes the conversion of O2 to
−O2• by a single-electron transfer. −O2 can be converted in the mitochondrial matrix to H2O2
by matrix superoxide dismutase. Since mitochondria do not have catalase, the H2O2 can be
degraded by nonenzymatic reaction with glutathione (GSH) by glutathione peroxidase or by
enzymatic reaction with thioredoxin reductase. The ROS species target cysteine residues on
the protein moieties of the mitochondrial membrane and polyunsaturated fatty acids, and
cause intramolecular cross-linkages and protein aggregates. The •OH causes peroxidation of
membrane lipids. ROS induce Ca2+ release from the mitochondria, which may cause an
imbalance in Ca2+ handling and stimulation of Ca2+-dependent proteases, nucleases and
phosphatases leading to functional and morphologic damage to the cell.

Vascular endothelium: Superoxide anion is produced by vascular endothelium by XO as

described above for cardiomyocytes (Fig. 9.3). A more important source of −O2 from the
vascular endothelium may be the NAD(P)H (reduced NAD phosphate) oxidase system. This
enzyme is a membrane-bound, flavin-containing NADH/NADPH-dependent oxidase present in
endothelial and vascular smooth-muscle cells. The molecular structure, function, and
regulation are somewhat similar to those of the neutrophil enzyme system, requiring assembly
of cytosolic components onto the membrane components before −O2 generation can occur.
However, −O2 production in the presence of NO • produced by endothelium favors the
biradical neutralization and elimination of NO • and the formation of peroxynitrite (ONOO–).
The elimination of NO• impairs not only vascular relaxation and autoregulatory control of
postischemic myocardial perfusion, but also fails to inhibit neutrophil-mediated postischemic
inflammatory responses.

Other sources of ROS include the oxidation of catecholamines, metabolism of arachidonic acid
to peroxy compounds, and generation of •OH by the cyclooxygenase and lipoxygenase
pathways, and by the metal-catalyzed Haber–Weiss reactions. The latter series of reactions
may be important in the setting of CPB because they can be recruited by neutrophil activation
during CPB (59). In addition, iron-containing products released by hemolysis may facilitate this
reaction.
When are Oxygen Radicals Generated?

Myocardial ischemia favors the generation of ROS. However, the myocardium has a cadre of
the endogenous antioxidants superoxide dismutase, catalase, reduced glutathione (GSH),
glutathione peroxidase, glutathione reductase. Normally, upregulation of these antioxidants
and phase 2 enzymes reduces oxidant-mediated injury (60), but they are depleted or can be
overwhelmed during ischemia, thereby attenuating the natural defense mechanisms of the
heart (61). Therefore, ischemia creates the biochemical setting for oxyradical production and
tissue vulnerability to ROS-mediated damage. However, the primary substrate, oxygen, is in
limited supply during ischemia, and is not readily available until reperfusion. Oxygen radicals
are not seen in great abundance until the onset of reperfusion. A clinical study by Prasad et al.
(62) show that oxygen radicals are released during and 24 hours after CPB in CABG patients.
Indirect evidence to support this reperfusion oxidative burst comes from a wealth of data in
which inhibitors or scavengers of oxygen radicals, or anti-neutrophil agents, exerted beneficial
effects when administered during reperfusion.

In the surgical setting, the myocardium may be vulnerable to ROS-induced injury at several
points when oxygen is introduced into the myocardium: (1) the initial delivery of cardioplegia
into the heart with antecedent unprotected ischemia, or initial delivery of oxygenated
cardioplegia to a newly revascularized myocardial segment; (2) intermittent reinfusions of
cardioplegia if a multidose strategy is being used; and (3) release of the aortic cross-clamp to
initiate reperfusion after the cardioplegia phase. The quantity of ROS generated during cardiac
surgery depends on the severity of the preceding ischemic injury, activation and recruitment of
neutrophils, the level of oxygen in cardioplegic solutions or blood oxygenated by the
extracorporeal circuit, and the status of endogenous scavengers and inhibitors potentially
depleted during ischemia. Blood cardioplegia theoretically provides a greater oxygen substrate
for oxygen radical production than do crystalloid solutions. However, plasma-dissolved oxygen
may be the pool from which oxygen is drawn for the generation of oxygen radicals, and the
oxygen content in this compartment may not differ dramatically between blood and crystalloid
solutions. The oxygen in blood cardioplegia may be reduced with “normoxic” or even
“deoxygenated” blood cardioplegia (63) or during reperfusion (64,65). In addition, blood
cardioplegia may offer the advantage of endogenous antioxidants (superoxide dismutase,
catalase, glutathione, and albumin) that are not present in crystalloid cardioplegia solutions.

Intracellular Na+ and Ca2+ Dyshomeostasis

Under normal physiologic conditions, intracellular Ca2+ concentrations are maintained at low
concentrations (<200 nmol) against a 5,000-fold trans-sarcolemmal gradient by voltage-gated
Ca2+ channels that remain closed until activation during phase 2 of the action potential. The
regulation of intracellular and intramitochondrial Ca2+ is tightly linked to regulation of
intracellular Na+. Intracellular Na+ concentration is normally controlled by the ATP-dependent
Na+/K+ pump which moves Na+ outward and K+ inward against their respective concentration
gradients, and the energy-independent Na+/H+ exchanger which normally removes one H+
from the cytosol in exchange for one Na+ moving into the cytosol. Further Na+ accumulation
mainly occurs through the Na+/Ca2+ exchanger which normally operates in the forward
direction during diastole to extrude intracellular Ca2+ in exchange for influx of Na+ into the
cytosol.

In ischemic-reperfusion myocardium, Ca2+ influx occurs through multiple pathways: (1)

diffusion through overt membrane disruptions or opened Ca2+ channels, fueled by the
differential concentration gradient; (2) reversal or inhibition of the Na+/Ca2+ exchanger driven
by the intracellular accumulation of Na+, (3) attenuated cyclical Ca2+ uptake and release from
the sarcoplasmic reticulum and mitochondria, and (4) facilitation of entry via α-adrenoreceptor
activation and cyclic adenosine monophosphate (AMP)-dependent processes, which increase
during postischemic catecholamine release.

The accumulation of intracellular Na+ and Ca2+ during ischemia-reperfusion are linked events
leading to Ca2+ overload which stimulates Ca2+-dependent proteases that damage
membranes and ion transport mechanisms at the moment of re-exposure to molecular
oxygen, opening of the mPTP, depletion of high-energy phosphate stores, the development of
local or global contracture (“stone heart”), and triggers the transition to cell death through
both apoptosis and necrosis. The balanced Ca2+ transport systems are potentially injured by
exposure to oxygen radicals, cytokines, and activated complement during CPB and reperfusion.
Na+ and Ca2+ dyshomeostasis occurs to differing degrees during ischemia and during
reperfusion, as described below.

Ischemia

During ischemia, glycolysis generates an abundance of H+ and only a limited amount of ATP (2
moles/mole glucose) compared to oxidative phosphorylation (36 moles ATP/mole glucose).
The accumulation of intracellular H+ stimulates activity of the Na+/H+ exchanger (NHE1) in
favor of H+ efflux and Na+ influx; at this time, the reduced availability of ATP decreases Na+/K+
ATPase activity. Therefore, both the attenuated Na+/K+ pump activity and increased activity of
the Na+/H+ exchanger result in intracellular Na+ accumulation. Additional Na+ may enter the
cell through sarcolemmal nonactivating Na+ channels, for example, the “window currents”
generated during ischemia-induced cellular depolarization. However, some evidence suggests
that the accumulation of H+ in both the extracellular and intracellular spaces attenuates the
H+ gradient, thereby reducing but not eliminating the participation of the Na+/H+ exchange
mechanism in intracellular Na+ accumulation during ischemia (66). However, the accumulation
of Na+ in the cytosol during ischemia would favor reversal of the energy-independent
Na+/Ca2+ exchanger in favor of extruding Na+ in exchange for Ca2+ influx, thereby favoring its
accumulation in the cytosol. Entry of Ca2+ into the mitochondria through the Ca2+ uniporter
depends on the Ca2+ electrochemical gradient and the presence of a mitochondrial
transmembrane potential. Ca2+ enters the mitochondria during early ischemia, but further
increases and overload are prevented by dissipation of the mitochondrial transmembrane
potential (67). In addition, the unregulated activation of Ca2+-dependent protease enzymes
such as calpains is inhibited by intracellular acidosis, and is delayed until intracellular pH is
renormalized during reperfusion (68). Therefore, Ca2+ accumulation and Ca2+-mediated
damage occurs to a greater extent during reperfusion (69,70,71) as discussed below.
Reperfusion

Na+ and Ca2+ dyshomeostasis during reperfusion is a significant factor contributing to

postischemic myocardial injury and functional impairment (72,73). The ion dyshomeostasis
described above sets the stage for explosive reperfusion/reoxygenation damage. It is
increasingly recognized that abrupt re-exposure to oxygen itself is the trigger for damage.
Excessive gradients of perfusate-to-tissue oxygen tension can be mitigated. In addition, the
events leading to abnormal Ca2+ handling which determine cell survival versus cell death are
as follows:

● Rapid renormalization of intracellular pH: Reperfusion washes out lactate and re-
establishes the intracellular–extracellular H+ gradient (extracellular < intracellular)
which allows the forward motion of the Na+/H+ exchanger favoring removal of H+
(renormalization of intracellular pH) and subsequent intracellular accumulation of Na+;
other factors contributing to normalization of intracellular pH are activation of
lactate/H+ co-transport mechanism and activation of the NaHCO3 co-transporter.
● Renormalization of intracellular Na+: During the initial phase of reperfusion, ATP is
regenerated, and if the Na+/K+ ATPase is not damaged by proteolytic actions of
calpain enzymes, intracellular Na+ levels are renormalized and the linked Na+/H+
exchange—Na+/Ca2+ exchange systems will not be activated. Otherwise, if the Na+/K+
ATPase pump is damaged, intracellular Na+ accumulation will engage the Na+/Ca2+
exchange mechanism with net intracellular accumulation of Ca2+.
● Ca2+ uptake and release by the sarcoplasmic reticulum (SR): The SR is the main Ca2+
storage site of the cardiomyocytes. Ca2+ in the micro-domain proximate to the closely
co-localized SR and interfibrillar mitochondria is cyclically taken up by the SR (lowering
of cytosolic concentration) during diastole, and Ca2+ is released into that micro-
domain by the ryanodine receptors (RyR) during systole. During both unprotected and
protected ischemia, sequestration of Ca2+ into the sarcoplasmic reticulum competes
for a limited amount of ATP. This paradox creates a cycle whereby a high intracellular
Ca2+ level binds high-energy phosphates to further limit ATP availability, which then
attenuates the ability of the SR to remove the Ca2+ actively from the cytoplasm,
resulting in a further increase in cytosolic Ca2+. If other Ca2+ handling mechanisms are
not engaged, the release of Ca2+ from the SR RyR may lead to overload. This activity of
the SR mechanism may be related to the severity of ischemia, being less so during
short periods of ischemia.

So, what events in the balance of Ca2+ regulation determines if a cell lives or dies during
reperfusion? Some evidence suggests that the timing of normalization of pH and Ca2+ handling
mechanisms are important. If pH normalization occurs before Ca2+ handling mechanisms are
restored, then Ca2+ overload, in conjunction with ROS generation and a normalized
intracellular pH environment, will trigger opening of the mPTP causing cell death. However, if
Ca2+ handling mechanisms are restored before pH is normalized, for example, by maintaining
tissue acidosis during early reperfusion, then Ca2+ overload can be avoided. Interestingly,
venous blood-based perfusate has been preliminarily observed to mitigate both over-
oxygenation during early reperfusion and delayed normalization of pH, thereby creating a
“permissive acidemia.”
The Inflammatory and Innate Immunity Responses to Ischemia-Reperfusion and
Cardiopulmonary Bypass

Injury mechanisms of ischemia-reperfusion and CPB involve inflammatory and innate immune
responses. Cardiomyocytes generate pro-inflammatory cytokines, chemokines and adhesion
molecules that direct the emigration of inflammatory cells into ischemic-reperfused tissue
regions and amplify the inflammatory response. The following sections will describe the
inflammatory process involving neutrophils, pro-inflammatory cytokines, and complement,
and the innate immune response involving toll-like receptors (TLRs).

Neutrophils and Neutrophil-Mediated Injury

Neutrophils play a critical role in the pathogenesis of myocardial reperfusion injury (74,75,76).
The inflammatory component of reperfusion injury requires an interaction between activated
neutrophils and vascular endothelium during the early moments of reperfusion, which is
prerequisite to activation of the full local inflammatory cascade. A well-orchestrated sequence
of events mediates the interaction between adhesion molecules on neutrophils and
endothelium. These adhesion molecules are categorized into three families:

● The selectins (P-selectin, L-selectin, E-selectin) are glycoproteins involved in the

interactions between neutrophils and the vascular endothelium in the early moments
of reperfusion (77,78,79). P-selectin is stored in Weibel–Palade bodies in endothelial
cells, and its surface expression is triggered by pro-inflammatory mediators such as
oxygen radicals, thrombin, complement components, cytokines, histamine, and H2O2
released during ischemia-reperfusion and CPB. IL-8 released from hypoxic endothelial
cells also increases the adhesiveness of endothelial cells for neutrophils (80). P-selectin
surface expression peaks after 10 to 20 minutes of reperfusion, and P-selectin is
subsequently shed to soluble fragments in blood. In contrast to P-selectin, L-selectin is
constitutively expressed on the surface of neutrophils and may be the counterligand
for P-selectin during early reperfusion (81). Loose adherence and “rolling” of
neutrophils on endothelium involves interaction with a glycoprotein sialyl Lewisx or a
high-affinity glycoprotein ligand termed P-selectin glycoprotein ligand-1 (PSGL-1) (82).
The third member of the selectin family, E-selectin, is expressed on the surface of
endothelial cells. It is expressed later in reperfusion (4–6 hours) and may therefore be
involved in later events, such as infarct extension.
● The β2-integrins (CD11/CD18 complex) are a family of glycoproteins located on
external membranes of neutrophils. There are three distinct α-chains (CD11a, CD11b,
CD11c) and a common β-subunit. Surface expression of perhaps the major complex
CD11b/CD18 is triggered by a number of pro-inflammatory mediators, including
circulating and endothelium-derived (acylated) platelet-activating factor. After initial
tethering of neutrophils by endothelial P-selectin, firm adherence is mediated by
interaction with CD11b/CD18 and its counterligand ICAM-1 on the endothelium.
● The third family of adhesion molecules is the immunoglobulin superfamily, including
ICAM-1 (intercellular adhesion molecule-1), VCAM-1 (vascular cell adhesion molecule-
1), and PECAM-1 (platelet–endothelial cell adhesion molecule-1). ICAM-1, the
counterligand for CD11/CD18 on neutrophils, is constitutively expressed on
 endothelial cells and is upregulated by cytokines 2 to 4 hours after myocardial
ischemia-reperfusion (83), coincident with the upregulation of CD11/CD18.
Subsequently, a high-affinity bond between the neutrophil and endothelial cell occurs
via integrins (CD11/CD18) on neutrophils and ICAM-1 and ICAM-2 on endothelial cells.
Once the neutrophil has firmly attached to the endothelial surface, it migrates through
the endothelial cell junctions to the myocardium, partly mediated by IL-8. The
shedding of L-selectin from the neutrophil and the presence of PECAM on the
endothelium may be required for neutrophil transendothelial migration. The
prerequisite nature of this interaction between neutrophils and adhesion molecules
makes anti-neutrophil and anti-adhesion therapy a potentially effective strategy, as
these agents intervene at proximal points in the cascade (discussed below).

Evidence supporting the rapid accumulation of neutrophils within ischemic-reperfused

myocardium following the neutrophil–endothelial cell interactions was appreciated not only by
our laboratory, but also by Sommers and Jennings (84). We (85,86,87,88,89) and others (90)
have demonstrated an association between the extent of postischemic injury and neutrophil
accumulation within ischemic-reperfused myocardium as well as a reduction of postischemic
injury associated with neutrophil suppression. As shown in Figure 9.4, Lefer et al. (91)
demonstrated that neutrophil adherence to coronary artery endothelium occurs in advance of
accumulation in extravascular sites, representing the lag time involved in transmigration of
intravascular neutrophils. In addition, neutrophil accumulation occurs in advance of the
development of necrosis, consistent with active neutrophil participation in cell demise. Brix-
Christensen et al. (92) reported that neutrophils accumulated in the hearts and other organs of
neonatal pigs placed on CPB. Therefore, neutrophil accumulation occurs in both postischemic
and post-CPB hearts.

Figure 9.4. The postischemic time course of endothelial injury, contractile dysfunction,
neutrophil (PMN) adherence and accumulation in tissue, necrosis, and apoptosis. These events
are indicated on the y-axis as a percent completion of that event (0% not started, 100% is
complete). Note that PMN adherence parallels that of endothelial dysfunction representing
PMN-mediated damage. Note also that accumulation in tissue lags behind adherence
representing emigration time. (Data obtained, in part, from Zhao ZQ, Nakamura M, Wang NP,
et al. Dynamic progression of contractile and endothelial dysfunction and infarct extension in
the late phase of reperfusion. J Surg Res 2000;94:133–144.) Apoptosis is a slower process than
necrosis as observed by Zhao et al. (565).(Adapted in part from Lefer AM, Tsao PS, Lefer DJ, et
al. Role of endothelial dysfunction in the pathogenesis of reperfusion injury after myocardial
ischemia. FASEB J 1991;5:2029–2034.)

The importance of neutrophils in postbypass reperfusion injury has been well-established in

both animal and human models. The myocardium is exposed to neutrophils during the delivery
of intermittent or continuous blood cardioplegia, or during reperfusion (after cross-clamp
removal) following either crystalloid or blood cardioplegia. Schwartz et al. (93) showed that
CPB circuitry directly “primes” neutrophils by depositing C3b on extracorporeal tubing, leading
to neutrophil activation and sequestration. Further evidence for the activation of neutrophils
following CPB was suggested by the finding of increased neutrophil surface expression of
CD11b/CD18 and decreased surface expression of L-selectin on cells circulated in simulated
CPB circuits (94,95). Finn et al. (96) noted in patients undergoing CPB that a rise in circulating
IL-8 begins at reperfusion with rewarming of the patient after systemic hypothermia, and
closely correlates with the rise in neutrophil count and circulating elastase. Similarly, Schwartz
et al. (93) and Gessler et al. (97) found higher levels of the pro-inflammatory mediators soluble
IL-6 and IL-8 6 hours after bypass than before bypass.

The time course for the upregulation of neutrophils and various adhesion molecules is equally
important in designing treatment strategies for patients undergoing extracorporeal bypass. In
a clinical study, Galiñanes et al. (98) investigated the perioperative temporal changes in the
expression of various neutrophil surface adhesion molecules, concluding that downregulation
of PECAM-1 is an early indicator of neutrophil activation and may therefore represent a target
for therapy aimed at reducing the inflammatory response associated with CPB.

The role of temperature in the activation of neutrophils or other inflammatory mediators

following CPB in humans has been investigated. Hypothermia has been reported to attenuate
expression of adhesion molecules in some models of tissue injury (99,100). Some investigators
have reported minimal or no difference in neutrophil activation and expression of neutrophil
adhesion molecules between normothermic (34°C–37°C) and hypothermic (28°C–30°C) CPB
(101,102), whereas others note an attenuation in the expression of IL-8 and neutrophil
elastase activity during normothermic CPBsurgery (103). Hypothermia has been shown to
delay, but not prevent, the increased expression of adhesion molecules such as CD11b and
CD11c (101). When measured 4 hours after bypass, the levels of IL-1α, soluble ICAM-1 (sICAM-
1), and elastase are higher after normothermic CPB than after hypothermic CPB (104,105). In
conclusion, further research is required to determine the most appropriate temperature of
cardioplegic administration to minimize the deleterious consequences of neutrophil-mediated
damage. Interventions designed to attenuate neutrophil-related myocardial injury following
CPB will be discussed later.
Complement in CPB and Myocardial Reperfusion Injury

The complement cascade, particularly the alternative pathway stimulated by contact

activation, is activated independently by myocardial ischemia-reperfusion and CPB, and
contributes importantly to the pathologic sequelae of CPB (106). Complement fragments such
as the anaphylatoxins C3a and C5a are generated and released both systemically and locally by
the ischemic-reperfused myocardium (107). The activation and migration of neutrophils to
ischemic-reperfused myocardium requires stimulation by chemotactic factors, most notably
complement fragments (C3a and C5a), eicosanoid products such as leukotriene B4, and
platelet-activating factor (108,109). Complement fragments (C5a) can be found in lymph from
ischemic myocardium (110), are temporally associated with reperfusion (111), and correlate
with increased neutrophil accumulation in reperfused myocardium. Furthermore, C5a directly
stimulates neutrophil −O2 production and enhances adherence to coronary artery
endothelium (112), whereas C3 activates monocyte adherence (113). The chemoattractant and
chemotactic properties of C5a cause progressive generalized vascular leukosequestration (114)
during CPB. The interaction between blood and extracorporeal surfaces also stimulates
deposition of C3b (115), which in turn amplifies the complement alternative pathway to
release C3a and C5a, and subsequently form the membrane attack complex (C5b 9) (115).
Accordingly, plasma levels of the anaphylatoxin C5a increase in patients undergoing CPB
(115,116), which activates neutrophils to generate −O2 and promotes adhesion to endothelial
cells, stimulates degranulation of mast cells, stimulates vascular smooth-muscle contraction,
and increases vascular permeability. Complement activation constitutes an early-phase
response to CPB, which then progresses to a second, slower response phase in which de novo
synthesis of proteins (E-selectin, ICAM, IL-8) and expression of adhesion molecules on cell
surfaces occurs. The magnitude of the complement response has been linked to the duration
of CPB and the type of circuitry used (116). The combined stimuli of myocardial ischemia-
reperfusion and CPB may synergize to exacerbate postischemic neutrophil-mediated
inflammatory responses in patients undergoing CPB for myocardial revascularization.

Actions of Toll-like Receptors

Recent evidence has shown that activation of TLRs is involved in the pathogenesis of
reperfusion injury (117,118). TLRs are transmembrane receptors expressed on endothelial
cells, cardiomyocytes, platelets, monocytes, and neutrophils. Cardiomyocytes participate in
the inflammatory and innate immune response to ischemia-reperfusion by releasing cytokines
(TNFα, IL-1β, IL-6, IL-8, IFN-γ), chemokines, and by expressing cell surface adhesion molecules.
TLRs are pattern recognition receptors (PRRs) that recognize pathogen-associated molecular
patterns (PAMPs) such as LPS or viral RNA on foreign pathogens, and sense endogenous
danger molecules and signals from injured cells. TLRs also recognize and are activated by
endogenous ligands such as danger-associated molecular patterns (DAMPS) (e.g., heat shock
proteins, HSP70) induced in cardiac tissue during stress (119) and cardiac surgery (120,121).
Inducible HSP70 has been found in plasma from CABG patients immediately after surgery
(122). ROS may also stimulate TLRs. Ten to eleven human TLRs have been identified; (123,124)
TLR receptor-1, -2, -4, -5, and -6 are located on the cell surface, with TLR-2 and TLR-4 being the
most involved in ischemia-reperfusion of the group of TLRs; TLR-2 works with other TLRs in co-
receptor cooperativity to recognize diacyl- and triacyl-lipoproteins, while TLR4 recognizes LPS
from gram-negative bacteria and DAMPs such as HSP70. Neutrophils constitutively express all
TRLs except TLR3.

The stimulation of TLRs during ischemia-reperfusion and cardiac surgery can lead to cell injury.
Dybdahl et al. (122) reported the release of HSP70 into the circulation of CABG patients and a
concomitant increase in monocyte TLR-2 and TLR-4 one day after surgery. Human monocytes
exposed to HSP70 released IL-6 and TNFα. Stimulation of TLRs by interaction with DAMPs
during ischemia-reperfusion is transduced by numerous molecular signaling pathways, which
ultimately results in the activation and translocation of NF-κB to the nucleus, where it
stimulates gene expression and release of cytokines and innate inflammatory factors. ROS
generated during reperfusion also stimulate TLR-dependent (most likely TLR-4) generation of
NF-κB through the PI3-K/Akt pathway, which suggests cross talk between the TLR and
reperfusion injury survival kinase (RISK) pathways. Inhibitors of TLRs, intermediary pathways,
or NF-κB will attenuate ischemia-reperfusion injury.

The Mitochondrial Permeability Transition Pore as a Determinant of Cell Death

Mitochondria are best known for being the power house of the cell, where ATP is generated
through oxidative phosphorylation. However, the mitochondria have been shown to function
as a “molecular switch” that governs cardiomyocyte viability; indeed it is a critical fulcrum
point in the transition from reversible to irreversible injury. Mitochondria can direct the cell to
pursue either an apoptotic or a necrotic pathway to cell death. The key event of this molecular
life/death switch is opening of the mPTP. This pore is a transmitochondrial membrane voltage-
dependent and Ca2+-dependent high-conductance channel located in the inner mitochondrial
membrane that permits water and solutes with a mass up to 1.5 kDa to enter the
mitochondrion. Under normal physiologic conditions, the mPTP is in a closed state. It opens in
response to high levels of intracellular ROS, Ca2+ and inorganic phosphate when mitochondria
matrix and intracellular pH are in the normal range. The mPTP remains closed during ischemia
because intracellular acidosis and the intracellular levels of Mg2+ and ADP counteract the
opening triggered by ROS and Ca2+. However, the mPTP opens during the early moments of
reperfusion (125) when there is an accumulation of ROS, Ca2+ and inorganic phosphate, and
when there is a rapid renormalization of intracellular and mitochondrial matrix pH (126). The
opening of the mPTP increases the permeability of mitochondria to water and small solutes,
causing swelling of the matrix and collapse of the mitochondrial transmembrane potential that
drives oxidative phosphorylation and ATP generation. mPTP opening also stimulates the
release of the pro-apoptotic factors cytochrome c and apoptosis-inducing factor (AIF). The
switch function comes into play based on the availability of intracellular ATP; if ATP is present
then the cell pursues an apoptotic pathway because apoptosis requres ATP. If ATP levels are
depleted then the cell pursues a necrotic pathway. With prolonged antecedent ischemia mPTP
is irreversible, but milder or short periods of ischemia trigger transient reversible opening
(127).
CHAPTER 13

INFLAMMATORY RESPONSES TO CARDIOPULMONARY BYPASS

Karsten Bartels and Mustafa Zakkar

The development of the heart–lung machine—cardiopulmonary bypass (CPB) in the 1950s—

together with the development of cardiac catheterization and the use of heparin and
protamine, made open-heart surgery possible and led to enormous improvement in patient
longevity and quality of life (1). CPB imposes a complex set of nonphysiologic circumstances
during which patients are subjected to severe physiologic alterations. With current anesthetic
and surgical care paradigms, few serious adverse sequelae occur in most patients (2,3,4).
Moreover, multiple recent large-scale clinical trials comparing off-pump coronary artery
bypass grafting (CABG) surgery to conventional CPB (5,6,7,8) have demonstrated few
meaningful differences, calling into question the clinical relevance of inflammation attributable
to CPB alone (Table 13.1).

Table13.1 Summary of primary outcomes in recent clinical trials comparing off-pump CABG
surgery to conventional CABG using cardiopulmonary bypass

Inflammation is one of the oldest pathophysiologic concepts known to mankind—Galen and

Celsus are credited with describing the pentad of calor (heat), rubor (redness), tumor
(swelling), dolor (pain), and function laesa (altered function) over 2,000 years ago (9).
Immunologic responses provide protection from a wide array of external pathogens, irritants,
and endogenous substances. Historically, the immune system has been categorized into the
innate and the adaptive (or acquired) immune system (10,11). The core feature of the adaptive
immune system is its ability to establish memory from joint processing by lymphocytes and
antigen-presenting cells (11). In case of reexposure to a previously encountered antigen,
antibodies can then be rapidly generated. The innate immune system consists of
hematopoietic cells such as neutrophils, eosinophils, mast cells, macrophages, and natural
killer cells, as well as nonhematopoietic cells such as endo- and epithelial cells (10). In sterile
inflammation, these can be activated by damage-associated molecular patterns (DAMPs),
sterile particulates, and intracellular cytokines that are released from damaged cells (12). The
innate immune response is then further mediated by soluble factors such as complement,
cytokines/chemokines, and acute phase proteins. It should be noted that the innate and
adaptive pathways of immunity are highly interrelated and that crosstalk is common.

Activation of the immune system occurring in the context of CPB is multifactorial (Fig. 13.1)
(13)—a feature that makes design of experimental models to mimic cardiac surgery especially
challenging (14).
Figure13.1 Schematic diagram of the sequence of events by which cardiopulmonary bypass
(CPB) may lead to the development of systemic inflammatory response syndrome (SIRS).(From
Warltier DC, Laffey JG, Boylan JF, et al. The systemic inflammatory response to cardiac surgery:
implications for the anesthesiologist. Anesthesiology 2002;97(1):215–252. Copyright © 2002,
American Society of Anesthesiology. Reproduced with permission from Wolters Kluwer Health.)

THE SYSTEMIC INFLAMMATORY RESPONSE TO CARDIOPULMONARY BYPASS

Definition

At the outset, it is important to define what the systemic inflammatory response means both
in general and in the specific context of CPB. Both clinical and experimental studies indicate
that various infectious and noninfectious conditions induce a similar host response known as
“systemic inflammatory response syndrome” (SIRS). This concept was advanced by the
American College of Chest Physicians/Society of Critical Care Medicine (ACCP/SCCM)
consensus conference in 1992 (5,6,15,16). Two or more of the following clinical manifestations
must be fulfilled for the diagnosis:

● Body temperature higher than 38°C or lower than 36°C.

● Heart rate more than 90 beats/min.

● Respiratory rate more than 20/min or PaCO2 less than 32 mm Hg.

● Leukocyte count more than 12,000 cells/mm3 or less than 4,000 cells/mm3, or the
presence of more than 10% immature neutrophils.

The 2001 International Sepsis Conference pointed out the nonspecific nature of the SIRS
criteria and further specified the variables that are likely to be affected by an infectious
inflammatory stimulus (17). CPB can produce a “whole-body” noninfectious or “sterile”
inflammatory response (18). Inflammation that occurs after CPB could well be described as an
SIRS. However, discriminating the postoperative inflammatory component that derives from
the surgical procedure versus the exposure to CPB proves challenging. Systemic inflammatory
response after cardiopulmonary bypass (SIRAB) therefore includes inflammatory response
from other components of cardiac surgery such as sternotomy or thoracotomy. Organ
dysfunction associated with cardiac surgery can provoke the activation of a plethora of
proinflammatory cascades adding to SIRAB (19). Hypotensive episodes during CPB can lead to
hypoxia and ischemia, which are strong proinflammatory activators and can initiate more
inflammatory responses that can last long after the termination of bypass and manifest as a
vasoplegic state (20). While most patients experience few clinically adverse sequelae from
CPB, and only a minority develop severe hemodynamic changes or organ failure (21,22), length
of CPB has been independently associated with morbidity and mortality after cardiac surgery
(23).

Initiation of Systemic Inflammatory Response after Bypass

A number of injurious processes that impinge on both cellular and noncellular (humoral)
elements of blood initiate SIRAB. These processes generate microemboli, disrupt hemostasis,
and lead to a generalized whole-body inflammatory response. Most importantly, they set in
motion a sequence of cytokine-mediated events that activate vascular endothelium, allowing
further neutrophil-mediated inflammatory injury. The damaging effects of CPB are most
commonly attributed to contact of blood with the foreign surface of the extracorporeal circuit.
However, other factors such as altered arterial blood flow patterns, shear stress generated by
blood pumps, cardiotomy suction devices, tissue ischemia and reperfusion, hypothermia,
anemia, and heparin, may be equally important (24). This inflammatory response may
continue long after CPB discontinuation, depending partly on its magnitude (25).

The repeated passage of blood through the nonphysiologic extracorporeal circuit initiates a
“contact activation” response with deposition of fibrin and activation of contact protein
cascades upon exposure of the blood to the nonendothelialized surface of the CPB machine.
Four proteins are involved in the contact activation cascades: coagulation factors XII and XI,
prekallikrein, and high-molecular-weight kininogen (HMWK). This activation forms bradykinin
and converts plasminogen into plasmin. This in turn initiates fibrinolysis and can trigger the
classical complement cascade. Therefore, multiple inflammatory mediators are released upon
exposure to the extracorporeal circuit (26). These responses initiate a sequence of chemokine-
mediated events that activate vascular endothelium and potentiate neutrophil-mediated
injury, which can disrupt hemostasis and create a generalized inflammatory response.

Cellular Components of Blood

Red Blood Cells

Tissue oxygenation depends upon continuous oxygen delivery, most of which is carried by the
hemoglobin molecules in the red blood cells (RBCs). The extracorporeal circuit exposes RBCs to
a challenging environment, thereby altering their integrity and function. Rheologic shear stress
forces mechanically damage RBCs to reduce their deformability, which is important to
maintaining normal RBC microcirculatory flow. As a further consequence of membrane
distortion, RBCs are susceptible to the membrane attack complex (MAC) generated by the
activation of complement (18), leading to hemoglobin leakage into the plasma. Free plasma
hemoglobin may impair tissue function by increasing plasma oncotic pressure and viscosity.
The ability of RBCs to aggregate also diminishes as a consequence of the dilution of blood by
the CPB circuit priming solution. In addition to the obvious potential reduction in tissue
oxygenation, this can trigger endothelial activation pathways (27).

Vascular Endothelium

Figure13.2 Adherence of circulating neutrophils (A) is triggered by activation of the endothelial

cells, which causes the neutrophils to roll along the endothelium (B), a process mediated by
selectin interaction with carbohydrate ligands. Activation of the neutrophils (e.g., by
interleukin [IL]-8, C5a, or platelet activating factor [PAF]) results in activation of integrin
molecules (e.g., by leukocyte function antigen [LFA]-1) on the neutrophil surface. This results in
firm adhesion of the neutrophils to activated endothelium (C) through receptors belonging to
the Ig family (e.g., intercellular adhesion molecules [ICAMs]). Activated adherent neutrophils
extravasate through the endothelium into the tissue interstitial space (D).
Under resting conditions, vascular endothelium offers a relatively inert surface that regulates
the passage of intravascular substrates to the extravascular space and ensures unhindered
flow of cellular and serum components through the capillary network. Endothelial cells are
extremely sensitive to the physiologic trespass imposed by CPB and surgical manipulation (and
hypoxia, should it occur). Release of endogenous substances such as cytokines (IL-1β and
tumor necrosis factor [TNF]-α), thrombin, C′5a, and lipopolysaccharide (endotoxin) activates
endothelial cells to increase surface expression of E-selectin, endothelial leukocytes adhesion
molecule (ELAM), intercellular adhesion molecule-1 (ICAM-1), and vascular adhesion molecule
(VCAM), all of which mediate leukocyte recruitment (28,29,30) (Fig. 13.2).

Leukocytes

The recruitment of leukocytes from the circulation to an inflammatory site is a multistep

process (Fig. 13.2) involving members of several adhesion receptor families (31,32). CPB
activates vascular endothelium, leading to rapid expression of members of the selectin family
(33). These are transmembrane molecules, expressed on the surface of leukocytes and
activated endothelial cells. Selectins contain an N-terminal extracellular domain with structural
homology to calcium-dependent lectins, followed by a domain homologous to epidermal
growth factor, and two to nine consensus repeats (CR) similar to sequences found in
complement regulatory proteins. Each of these adhesion receptors is inserted through a
hydrophobic transmembrane domain and possesses a short cytoplasmic tail. During
inflammation, bloodstream leukocytes attach to endothelium via selectins, thereby impairing
microcirculatory flow, potentially culminating in transient standstill from leukocyte adhesive
“plugs” along the endothelium. The selectin family includes P-, E-, and L-selectin. P-selectin is
largely responsible for the rolling phase of the leukocyte adhesion cascade. Transmigration,
the next step, is mediated by adhesion molecule upregulation, including platelet endothelial
cell adhesion molecule (PECAM)-1, CD99, ICAM-1, CD11a/CD18, very late antigen (VLA)-4, and
endothelium-secreting chemoattractants (34).

Neutrophils

Neutrophils are key mediators of the systemic inflammatory response; their recruitment,
activation, and cytotoxic capability are essential to warding off infection. Neutrophil activation
through interaction with activated vascular endothelium may be responsible for much of the
clinical sequelae of SIRAB. Activation of neutrophils during CPB is shown by the loss of L-
selectin and the upregulation of CD11b/CD18 (Mac-1). Increased production of reactive oxygen
intermediate (35) and neutrophil-derived elastase and similar molecules have also been
reported (35,36). Studies in complement-deficient dogs suggest that at least some of the
neutrophil activation seen in CPB is due to complement activation (37,38). In addition, the
release of cytotoxic products by activated cells can cause direct cellular damage. These
cytotoxins include both preformed agents present in neutrophil granules and newly
synthesized molecules. These novel substances include leukotrienes (36) and reactive oxygen
intermediates (35). This synthesis can be readily detected by a dramatic increase in cellular
oxygen consumption. These species are largely responsible for the so-called ischemia-
reperfusion injury. Ischemic injury occurs when the blood supply to a tissue is impaired or
suboptimal, which may occur with CPB. The paradox, however, is that a more severe tissue
injury occurs when blood flow is restored upon reperfusion.

Monocytes

Peripheral monocytes are also involved in the systemic inflammatory process. They possess
migratory, chemotactic, pinocytic, and phagocytic activities, as well as receptors for IgG and C
′3b. Upon migration to tissue, they undergo further differentiation to become tissue
macrophages, which participate in both specific and nonspecific immune pathways.
Recruitment of these cells occurs as early as 1 hour following insult (39) and is thought to be
mediated mainly by monocyte chemoattractant protein-1 (MCP-1), although other factors such
as C′5a play an important role. Upon activation, they play a pivotal role in inflammation,
serving as effector cells secreting cytokines (40). Monocytes can exhibit both pro- and anti-
inflammatory properties, depending on the signals they receive, and it has been suggested
that these cells have the capacity to switch from one state to the other to participate in both
the induction and the resolution of inflammation (39).

Platelets

CPB is associated with a transient deficit in platelet function and number, which can impair
postoperative hemostasis (41). The normal platelet adheres to a damaged endothelial cell or
the subendothelial layer. The multimeric form of von Willebrand factor forms a bridge
between endothelium and platelet at the platelet glycoprotein (GpIb) receptor site. The
platelet then undergoes a conformational change to expose different glycoproteins, including
the GpIIb/IIIa complex, which can in turn bind to fibrinogen. Fibrinogen is an essential cofactor
in platelet adhesion and in platelet-to-platelet binding during irreversible aggregation. The
protein complex thrombospondin stabilizes this platelet aggregate. Concurrent thromboxane
A2 release produces local vasoconstriction and further platelet aggregation.

Numerous CPB-related factors contribute to platelet changes. These include physical factors
(e.g., hypothermia and shear forces), exposure to artificial surfaces (42), drugs, and
endogenous chemicals (43,44). Hemodilution contributes to initial thrombocytopenia (45), but
mechanical disruption, adhesion to the extracorporeal circuit, and sequestration in organs
explains a disproportionate drop (beyond hemodilution alone) in platelet count, which can be
30% to 50% below baseline (46). The response of platelets to CPB is complex and
multifactorial, including rapid consumption of platelets during bypass (45), decreased
reactivity to known agonists (44), increase in the concentration of the α-granule compounds in
plasma (47), and an increase in the stable metabolite of thromboxane A2 (thromboxane B2)
released from aggregating platelets (48). Morphologic changes, including spherical appearance
and the development of pseudopods, also occur during CPB (49). Prolonged bleeding time
after CPB correlates with duration of bypass; however, its time course and precise mechanism
remain unclear (50).

Platelets activated during CPB form conjugates among themselves and between platelets and
leukocytes. Activated platelets express P-selectin, which contributes to leukocyte conjugate
formation by binding PSGL-1 (51). Activated platelets use this P-selectin/PSGL-1 adhesion
pathway to stimulate conjoined monocytes, thereby leading to secretion of the
proinflammatory cytokines IL-1β, IL-8, and MCP-1 (52,53). P-selectin also induces tissue factor
expression and fibrin deposition by monocytes, thereby contributing to thrombus evolution
(54,55).

Coagulation System

Figure13.3 Simplified summary of the coagulation cascade.

The coagulation cascade can be activated by either intrinsic or extrinsic pathways, each of
which consists of a series of enzymes. Figure 13.3 shows a simplified version of these
pathways.

The intrinsic pathway begins with Factor XII activation upon contact of blood with collagen in
the damaged vascular wall or an artificial surface, and ends with the formation of fibrin
through a cascade including activated factors XI, IX, X, and thrombin. Exposure of blood to
nonvascular tissue cells expresses a protein called tissue factor, which binds to factor VII and
activates the extrinsic pathway. This in turn activates factor X. Once factor Xa is generated, the
remainder of the cascade is the same as the intrinsic pathway.

Coagulation cascade activation induces the fibrinolytic system to coactivate, which limits the
amount of fibrin clot formed and localizes clot formation to the injury site. CPB activates both
the clotting and fibrinolysis pathways. Not only does this cause fibrin formation and platelet
activation, but also it activates vascular endothelium. Thrombin-mediated endothelial cell
activation elicits leukocyte adhesion as a result of selectin expression. In addition, the factor
VIIa–tissue factor complex and factor Xa activate cells through protease-activated receptors,
and this, in turn, generates cellular responses similar to those mediated by thrombin activation
of protease-activated receptor-1 (56). CPB-induced increases in factor XIIa and kallikrein
activate fibrinolysis, which has been linked to increased postoperative bleeding time, blood
loss, and reduction in platelet adhesion and aggregation capabilities from redistribution of
glycoprotein Ib, and IIb/IIIa receptors (57,58).

Humoral Components of Blood

Kirklin and McGriffin (3) initially hypothesized that many of the deleterious effects of CPB
derived from exposure of blood to nonendothelial surfaces, which initiates a “whole-body
inflammatory response.” They characterized this response as activation of coagulation, the
kallikrein system, fibrinolysis, and complement (3,59). We now also recognize the importance
of cytokines (60) and the combined effects of these humoral cascades in the activation of
endothelial cells and neutrophil adhesion.

Inflammatory Cascades

The principal event is the activation of factor XII (Hageman factor), which stimulates a number
of inflammatory systems (Fig. 13.3). After surface contact, factor XII undergoes a
conformational change and attaches to the protein HMWK, which is involved in the early
stages of intrinsic pathway activation and is a precursor of bradykinin.

This complex attaches to the foreign surface and, after limited proteolysis, releases kallikrein,
bradykinin, and more factor XIIa. Factor XIIa can initiate the intrinsic coagulation cascade by
activating factor XI, which binds to the foreign surface and activates factor VII, thereby further
augmenting the intrinsic coagulation cascade. Factor XIIa also provides a positive feedback
loop by inducing prekallikrein to form kallikrein, which in turn activates factor XII to produce
more factor XIIa (61). Kallikrein activates neutrophils, which further activate inflammatory
cascades to produce oxygen free radicals and proteolytic enzymes. Furthermore, both
kallikrein and bradykinin stimulate the fibrinolytic system. Kallikrein stimulates plasmin
production by its action on prourokinase and bradykinin releases tissue-type plasminogen
activator from the endothelium.

Complement System

The complement system comprises one of the preeminent immunologic mechanisms involved
in the inflammatory process and it is activated by CPB (62). More than 30 complement system
proteins serve both as an immune response effector arm and as a primitive self- versus non-
self-recognition system. Complement system functions include mediating inflammation,
opsonization of antigenic particles, and causing membrane damage to pathogens.
Complement components interact such that the products of one reaction form the enzyme for
the next. Therefore, a small initial stimulus can trigger a cascade of biologic activation and
amplification. The fact that 5% to 10% of serum proteins are complement components attests
to its importance. Blood which collects in the pericardium has very high levels of Tissue Factor
which, if aspirated into the CPB equipment using the cardiotomy suction, can accelerate the
inflammatory response if reinfused back into the patient.
Figure13.4 The complement systems. C1q, C3b, and C5a represent key functional elements that
crosstalk in multiple directions. Formation of the MAC leads to lysis, but also leads to cytokine
secretion. MBL, mannan-binding lectin; MAC, membrane attack complex.

The complement system (Fig. 13.4) was previously interpreted as a cascading system with a
common final pathway forming the MAC. Current understanding views complement as a
dynamic network that contains several hubs such as C1q, C3b, and C5a, each of which is highly
integrated within other pathways (63). Complement activation not only forms the MAC but
also interacts complexly with other systems such as proinflammatory signaling pathways, lipid
metabolism, and adaptive immunity (Fig. 13.4). Complement inhibition has been suggested as
a promising target for immunomodulative therapy in the context of cardiac surgery (64).

Further Contributions to Systemic Inflammatory Response after Bypass

In addition to the inflammatory interactions that result from the bypass circuit, cardiogenic
shock and endotoxemia may compound the response (65). Cardiogenic shock can contribute to
SIRAB, and prolonged nonpulsatile perfusion or periods of circulatory arrest can also lead to
diffuse end-organ ischemia (66). The end-organ hypoxic insult likely causes endothelial cells,
circulating monocytes, and tissue-fixed macrophages to release cytokines and oxygen-derived
free radicals that drive this response. After resuscitation from shock and hypoxic end-organ
reperfusion, a systemic ischemic-reperfusion injury ensues (65).

Endotoxemia represents another form of inflammatory activation that results from

extracorporeal circulation and ischemia-reperfusion. Frequently detected in high
concentrations in the systemic circulation after CPB (67), endotoxin potently induces
macrophage TNF release, which likely explains its increased concentrations after CPB in some
patients. In a rodent CPB model, brain TNF-α levels were elevated in the CPB group compared
to nonoperative control animals. However, a similar effect was seen in animals that underwent
cannula placement without CPB (68). These data suggest that TNF-α release and associated
neurologic injury also occur from surgical manipulation without CPB (69).
IMMUNE RESPONSE AFTER CARDIOPULMONARY BYPASS

The cellular and humoral constituents of the adaptive immune system undergo changes in
both function and number after CPB. The concern is that postoperative infection partially
derives from CPB-induced loss of T- and B cells and associated immunoglobulin consumption.
This section examines the effect of CPB on adaptive immunity in general and T-cell function in
particular.

Immunodeficiency and Cardiopulmonary Bypass

Duration of CPB correlates with postoperative infections (70). CPB depresses immunologic
reactivity, thereby increasing susceptibility to perioperative infections and the potential for
septic shock (71).

Normally, relatively low levels of immunoglobulins and complement provide adequate

bacterial opsonization (72). However, if the load of infectious microorganisms increases, as
may occur in patients undergoing open-heart surgical procedures, then an imbalance may
develop in host defense homeostasis. Quantitative and qualitative exhaustion of humoral and
cell-mediated immune mechanisms may adversely affect clinical outcome when additional
injury occurs postoperatively in a patient with a downregulated immune system. It is therefore
unsurprising that opportunistic pathogens are frequently found in post-CPB patients who
experience complications such as renal failure or low cardiac output. Post-CPB patients exhibit
an increased risk for sepsis-related multisystem organ failure, which accounts for a high
proportion of perioperative fatalities (73).

Humoral Immunity

Serum levels of immunoglobulins and complement change markedly during CPB (21,74,75),
with a resultant reduction in host defenses, for example, reduced opsonization of bacteria in
vitro (76). Leukocyte counts fall beyond the affect of hemodilution at the onset of CPB as a
result of tissue sequestration after activation by anaphylatoxins C′3a and C′5a. After CPB,
granulocyte chemotaxis is impaired (77), which likely increases susceptibility to bacterial
infections. CPB also impairs leukocyte phagocytosis and metabolic function during and after
CPB (106).

Some investigations report unchanged B-cell numbers after CPB (18), whereas others
demonstrate a drop (107,108). Roth et al. reported that relative B-cell levels (percentage)
increase after CPB, with no change in their absolute number, which may result from a
combination of hemodilution and immunosuppression. The secretion of IgG, IgM, and IgA by B
cells in response to pokeweed mitogen diminishes after CPB. Recent work by Lante et al.
showed that IgE levels did not change after CPB until day 3; in contrast, both IgG and IgM
levels decreased significantly on day 1. IgM returned to baseline on day 5, at which time IgG
levels remained low. This was associated with a reduction in the B-cell count on day 1 and a
return to baseline by day 3. The bactericidal activity of serum is depressed after CPB. Whereas
complement is consumed during CPB, other components of humoral immunity decrease in
proportion to CPB hemodilution.
When immune proteins contact gaseous and foreign surfaces, denaturation ensues. Surface
depolarizing forces disrupt sulfhydryl and hydrogen bonds that stabilize the secondary and
tertiary structure of the protein molecules. This unfolds globular protein molecules, which may
create a randomly coiled molecule and expose previously masked chemical groups. The
macromolecules so formed then tend to flocculate. Plasma protein denaturation may also
coalesce serum lipids and increase plasma viscosity (75).

Mast Cells and Paneth Cells

Mast cells are sentinels that herald gut inflammation. In a rodent model of CPB and deep
hypothermic circulatory arrest, intestinal mast cells in conjunction with Paneth cells mediate
ischemia/reperfusion injury (Fig. 13.5) (78). The impact of unappreciated acute gut injury on
adverse outcomes after cardiac surgery is only recently being appreciated {Karhausen, 2014
#1308}. Since mast cells may be activated by complement, suppressing this activation may
prove valuable in attenuating CPB-induced ischemia/reperfusion injury.

Figure13.5 Paneth cells and intestinal mast cells release potent effectors to regulate local injury
and systemic inflammation after intestinal ischemia/reperfusion. Most prominently, the Paneth
cell-dependent pathway (blue) depends on release of interleukin (IL)-17 from Paneth cells
localized at the base of small-intestinal crypts. Mast cell responses (purple) use a number of
preformed and de novo synthesized products such as proteases (tryptase, mast cell protease
[MCP]-4), lipid mediators (leukotrienes, prostaglandins), and cytokines (tumor necrosis factor
[TNF]-α, IL-6).(Figure prepared by Annemarie B. Johnson, C.M.I., Medical Illustrator, Vivo
Visuals, Winston-Salem, NC, for Bartels K, Karhausen J, Clambey ET, et al. Perioperative organ
injury. Anesthesiology 2013;119:1474–1489. Copyright © 2013, American Society of
Anesthesiology. Reproduced with permission from Wolters Kluwer Health.)

Natural Killer Cells

Natural killer cells, a heterogeneous lymphocyte subpopulation that is neither T nor B, produce
cytotoxic responses in virus-infected cells and transformed target cells (e.g., tumor cells).
Decreases in both number and function of natural killer cells occur after CPB (79,80).

Reticuloendothelial System

The reticuloendothelial system comprises tissue macrophages in the spleen, lymph nodes,
lung, and liver. Its cells derive from blood-borne monocytes. Normal reticuloendothelial
function includes clearing the blood of bacteria, endotoxins, platelets, denatured proteins,
chylomicrons, plasma hemoglobin, thrombin, fibrin, fibrin degradation products,
thromboplastin, and plasminogen activator. CPB introduces microparticles into the
bloodstream, which in turn depresses reticuloendothelial system function (81).

Cells

T cells are lymphocytes that develop in the thymus, which is seeded by lymphocytic stem cells
from the bone marrow during embryonic development. Following T-cell development, mature
naive T cells leave the thymus and begin to spread throughout the body. These cells then
develop their T-cell antigen receptors and differentiate into the two major peripheral T-cell
subsets, one of which expresses the CD4+ marker (helper cells) and the other CD8+ (cytotoxic
cells). Proliferating T helper cells can differentiate into two major subtypes on the basis of the
specific cytokines they produce, known as TH1 and TH2. T helper cells play a central role in the
initiation and regulation of the acquired immune response.

T helper cells recognize antigen presented on the surface of antigen-presenting cells in

association with class II molecules encoded by the major histocompatibility complex (MHC). T-
cell activation requires other specific costimulatory signals generated by the antigen-
presenting cell. Cytotoxic T cells recognize antigen presented on the surface of antigen-
presenting cells in association with class I molecules encoded by the MHC.

T helper cells provide “help” in the form of lymphokine secretion. Such lymphokines help B
cells to divide, differentiate, and produce antibodies. Lymphokines are also required for the
development of leukocyte lines from hematopoietic stem cells and development of cytotoxic T
cells. They also cause activation of macrophages, allowing them to destroy the pathogens they
have taken up. Cytotoxic T cells are capable of destroying virus-infected target cells or
allogeneic (transplanted) cells.

Morphologic Correlation

Scanning electron microscopy shows profound alterations to the plasma membrane of the T
cells after CPB. The number of microvilli and lymphocyte surface folding both decrease.
Furthermore, after CPB, the membrane does not accommodate monoclonal microbeads.
Quality and Quantity

Phenotypic changes induced by CPB have been investigated using standard commercially
available leukocyte-labeling monoclonal antibodies together with flow cytometry. CD3+ and
CD4+ cells decrease and CD8+ cells increase, hence the CD4+/CD8+ ratio increases. Such
changes peak on postoperative day 1 but remain abnormal for approximately 1 week (72). T-
lymphocyte counts decrease after CPB (82). The capacity of the immune system to counteract
microbial infection decreases after cardiac surgery because of consumed (and unreplaced)
complement factors and decreased cellular immune elements such as neutrophils and natural
killer cells (79,80,81,82).

Mechanisms

The mechanism responsible for the decrease in circulating T lymphocytes after CPB remains
undefined. Elevated cortisol levels probably play an important postoperative
immunosuppressant role, yet this is unlikely to be the only factor in this complex
phenomenon. Elevation of serum corticosteroids may induce redistribution of circulating T
cells to lymphoid tissues, although this is usually mild and transient. Although serum cortisol
levels are elevated after surgery, the lack of correlation between changes in serum cortisol and
changes in lymphocyte number and function suggests that serum cortisol is not an etiologic
factor in postoperative T-cell dysfunction. Post-CPB reductions in T-cell activity are both
quantitative and qualitative, likely causes for which include hemodilution, intravascular-to-
extravascular fluid shifts, mechanical destruction, consumption, and redistribution among
bone marrow, lymphoid tissue, and peripheral blood (80,83).

THERAPEUTIC APPROACHES

Effective amelioration of SIRAB has yet to become a reality. Difficulty in maintaining adequate
perfusion pressure toward the end of bypass suggests a profoundly low systemic vascular
resistance, which is consistent with SIRAB. Vasoconstrictor and inotropic drugs constitute the
primary current treatment of this situation, although this does not address the
pathophysiology. To prevent or ameliorate SIRAB, the optimal strategy would be to develop a
CPB circuit that does not induce contact activation of blood components.

Pharmacologic Manipulation

Pharmacologic attempts to suppress the inflammatory response to cardiac surgery have failed
thus far to improve clinical outcomes (84). This is likely due to the fact that the immunologic
response may not be pathologic per se (as it is in autoimmune diseases), but rather represents
a secondary response to the altered physiology of CPB.

Corticosteroids

The physiologic effects of corticosteroids are numerous and widespread. They possess anti-
inflammatory properties and influence the water/electrolyte balance and the metabolism of
carbohydrate, protein, and lipid. Single-dose dexamethasone for cardiac surgery has failed to
affect 30-day mortality in a randomized, controlled clinical trial of 4,494 patients (85). A
subgroup of this trial underwent cognitive testing at 1 and 12 months following surgery and
failed to show any improvement in cognitive function (86). Similarly, preliminary results of the
Canadian Steroids In Cardiac Surgery Trial (SIRS Trial) (NCT00427388) have not been promising.
Further studies might investigate a role for steroid treatment using other doses, administration
protocols, or types of steroids.

Alternative Strategies

The C5 complement inhibitor pexelizumab has been studied in several clinical trials.
Pexelizumab was associated with only a 6.7% nonsignificant reduction in the primary
composite end point of death or myocardial infarction in the PRIMO-CABG II trial (87). These
results were surprising as the previous PRIMO-CABG I trial had suggested more favorable
outcomes (88,89). Furthermore, pexelizumab administration had no effect on global cognitive
outcomes after cardiac surgery, with the exception of a possible effect to reduce dysfunction in
the visuo-spatial domain (90). An exploratory analysis of the combined PRIMO I and II data
(7,353 patients) suggested a mortality benefit for certain high-risk cardiac surgery patients
(87). Furthermore, a meta-analysis of pexelizumab in ischemic heart disease suggested
benefits in patients undergoing CABG (91).

An interesting novel approach is represented by the administration of cyclosporine.

Cyclosporine inhibits the opening of the mitochondrial permeability transition pore and its
ability to prevent ischemia/reperfusion injury in patients undergoing aortic valve replacement
has been studied in a small population (61 patients) (92). Intravenous cyclosporine bolus
dosing reduced the area under the curve for postoperative cardiac troponin I postoperatively
by 35%. Whether these preliminary results will translate into improvement of more
meaningful clinical outcomes needs further study.

Mechanical Manipulation

Depletion of Leukocytes and Inflammatory Mediators

Conducting ultrafiltration at the termination of CPB, such as modified ultrafiltration (MUF),

aims to reverse hemodilution and decrease tissue edema and circulating inflammatory
mediators. Reducing the number of circulating leukocytes by the use of leukocyte-depleting
filters is another way to modify the inflammatory process. Some studies report reduced lung
injury and improved oxygenation (93), as well as reduced hospital length of stay (94).

Mechanical Considerations

Pulsatile perfusion may have beneficial effects with respect to hemodynamics,

microcirculation, and organ dysfunction, compared with nonpulsatile flow (95), as
demonstrated by many experimental and clinical studies (96,97,98,99). A recent meta-analysis
of eight studies including 970 patients suggested improved postoperative pulmonary function
in addition to shorter intensive care unit (ICU) and hospital stay (100). However, as Murphy et
al. (101) point out, there have been over 150 basic and clinical science investigations on the
use of pulsatile versus nonpulsatile technique for CPB and the controversy continues because
the results are inconclusive.
Membrane oxygenators currently predominate over bubble oxygenators for cardiopulmonary
bypass in developed nations, as they provide more titratable arterial PO2 values and less blood
trauma, especially with prolonged use. Modern membrane oxygenators provide blood flow to
surrounding fibers that are filled with “inspired” gases that equilibrate across the membrane
(102).

Coating of oxygenator membranes and CPB tubing and filters with heparin inhibits the release
of proinflammatory cytokines and reduces complement and platelet activation (103,104). A
meta-analysis including 3,434 patients found that heparin-bonded circuits decreased the
incidence of blood transfusion, sternal reexploration for bleeding, duration of mechanical
ventilation, and hospital length of stay. Effects on other outcomes were minimal (105).

The adverse impact of blood contact with extracorporeal surfaces, air–fluid interface, and cell
damage by cardiotomy suction associated with conventional bypass has led to the
development of minimized cardiopulmonary bypass circuits (mini-CPB). Such circuits use a
closed CPB system characterized by reduced surface area and priming volume, elimination of
cardiotomy suction, and prevention of air–blood contact. Studies comparing mini-CPB to
conventional CPB show delayed or reduced secretion of proinflammatory cytokines,
attenuated complement activation, and blunted leukocyte activation (106,107,108). The use of
mini-CPB was associated with less hemodilution and thus less blood transfusion (106,107,108).
It remains unclear if this approach improves major clinical outcomes, but interpretation is
complicated by the use of different mini-CPB systems and the heterogeneity of CPB protocols.

SUMMARY

It appears likely that the inflammatory and immunologic sequelae of CPB are not primarily
responsible per se for a large part of the morbidity and mortality seen postoperatively,
although they may contribute. This is highlighted by the lack of obvious deleterious effects of
CPB in the more recent off-pump/on-pump CABG prospective comparison trials. However,
perioperative inflammation assumes greater importance in longer, more complex surgery
performed on patients who are at extremes of age and who have significant comorbid
conditions. Furthermore, it appears as if prevention of inflammation through minimizing and
preempting iatrogenic injury is more likely to succeed than pharmacologic intervention after
the injury has occurred. While some substances have shown promise (e.g., the C5 complement
inhibitor pexelizumab), effects on meaningful clinical outcomes have been inconsistent.

KEY Points

● CPB is associated with a systemic inflammatory response that includes cellular and
noncellular elements.

● Results from recent large-scale clinical trials comparing off-pump CABG surgery to
conventional CABG question the clinical relevance of inflammation attributable to CPB
alone.
● The complement system is a dynamic network that is highly integrated within other
pathways, and is activated during CPB.

● Use of complement inhibitors (e.g., pexelizumab) remains a target for

immunomodulative therapy in the context of cardiac surgery.

● Mini-CPB show promise to attenuate the inflammatory response to CPB through

surface area and priming volume reductions, cardiotomy suction elimination, and air–
blood contact prevention.

● Minimizing occurrence of inflammation by preempting iatrogenic injury is more likely

to be successful than pharmacologic intervention after the injury has occurred.
CHAPTER 13

INFLAMMATORY RESPONSES TO CARDIOPULMONARY BYPASS

Karsten Bartels and Mustafa Zakkar

The development of the heart–lung machine—cardiopulmonary bypass (CPB) in the 1950s—

together with the development of cardiac catheterization and the use of heparin and
protamine, made open-heart surgery possible and led to enormous improvement in patient
longevity and quality of life (1). CPB imposes a complex set of nonphysiologic circumstances
during which patients are subjected to severe physiologic alterations. With current anesthetic
and surgical care paradigms, few serious adverse sequelae occur in most patients (2,3,4).
Moreover, multiple recent large-scale clinical trials comparing off-pump coronary artery
bypass grafting (CABG) surgery to conventional CPB (5,6,7,8) have demonstrated few
meaningful differences, calling into question the clinical relevance of inflammation attributable
to CPB alone (Table 13.1).

Table13.1 Summary of primary outcomes in recent clinical trials comparing off-pump CABG
surgery to conventional CABG using cardiopulmonary bypass

Inflammation is one of the oldest pathophysiologic concepts known to mankind—Galen and

Celsus are credited with describing the pentad of calor (heat), rubor (redness), tumor
(swelling), dolor (pain), and function laesa (altered function) over 2,000 years ago (9).
Immunologic responses provide protection from a wide array of external pathogens, irritants,
and endogenous substances. Historically, the immune system has been categorized into the
innate and the adaptive (or acquired) immune system (10,11). The core feature of the adaptive
immune system is its ability to establish memory from joint processing by lymphocytes and
antigen-presenting cells (11). In case of reexposure to a previously encountered antigen,
antibodies can then be rapidly generated. The innate immune system consists of
hematopoietic cells such as neutrophils, eosinophils, mast cells, macrophages, and natural
killer cells, as well as nonhematopoietic cells such as endo- and epithelial cells (10). In sterile
inflammation, these can be activated by damage-associated molecular patterns (DAMPs),
sterile particulates, and intracellular cytokines that are released from damaged cells (12). The
innate immune response is then further mediated by soluble factors such as complement,
cytokines/chemokines, and acute phase proteins. It should be noted that the innate and
adaptive pathways of immunity are highly interrelated and that crosstalk is common.

Activation of the immune system occurring in the context of CPB is multifactorial (Fig. 13.1)
(13)—a feature that makes design of experimental models to mimic cardiac surgery especially
challenging (14).
Figure13.1 Schematic diagram of the sequence of events by which cardiopulmonary bypass
(CPB) may lead to the development of systemic inflammatory response syndrome (SIRS).(From
Warltier DC, Laffey JG, Boylan JF, et al. The systemic inflammatory response to cardiac surgery:
implications for the anesthesiologist. Anesthesiology 2002;97(1):215–252. Copyright © 2002,
American Society of Anesthesiology. Reproduced with permission from Wolters Kluwer Health.)

THE SYSTEMIC INFLAMMATORY RESPONSE TO CARDIOPULMONARY BYPASS

Definition

At the outset, it is important to define what the systemic inflammatory response means both
in general and in the specific context of CPB. Both clinical and experimental studies indicate
that various infectious and noninfectious conditions induce a similar host response known as
“systemic inflammatory response syndrome” (SIRS). This concept was advanced by the
American College of Chest Physicians/Society of Critical Care Medicine (ACCP/SCCM)
consensus conference in 1992 (5,6,15,16). Two or more of the following clinical manifestations
must be fulfilled for the diagnosis:

● Body temperature higher than 38°C or lower than 36°C.

● Heart rate more than 90 beats/min.

● Respiratory rate more than 20/min or PaCO2 less than 32 mm Hg.

● Leukocyte count more than 12,000 cells/mm3 or less than 4,000 cells/mm3, or the
presence of more than 10% immature neutrophils.

The 2001 International Sepsis Conference pointed out the nonspecific nature of the SIRS
criteria and further specified the variables that are likely to be affected by an infectious
inflammatory stimulus (17). CPB can produce a “whole-body” noninfectious or “sterile”
inflammatory response (18). Inflammation that occurs after CPB could well be described as an
SIRS. However, discriminating the postoperative inflammatory component that derives from
the surgical procedure versus the exposure to CPB proves challenging. Systemic inflammatory
response after cardiopulmonary bypass (SIRAB) therefore includes inflammatory response
from other components of cardiac surgery such as sternotomy or thoracotomy. Organ
dysfunction associated with cardiac surgery can provoke the activation of a plethora of
proinflammatory cascades adding to SIRAB (19). Hypotensive episodes during CPB can lead to
hypoxia and ischemia, which are strong proinflammatory activators and can initiate more
inflammatory responses that can last long after the termination of bypass and manifest as a
vasoplegic state (20). While most patients experience few clinically adverse sequelae from
CPB, and only a minority develop severe hemodynamic changes or organ failure (21,22), length
of CPB has been independently associated with morbidity and mortality after cardiac surgery
(23).

Initiation of Systemic Inflammatory Response after Bypass

A number of injurious processes that impinge on both cellular and noncellular (humoral)
elements of blood initiate SIRAB. These processes generate microemboli, disrupt hemostasis,
and lead to a generalized whole-body inflammatory response. Most importantly, they set in
motion a sequence of cytokine-mediated events that activate vascular endothelium, allowing
further neutrophil-mediated inflammatory injury. The damaging effects of CPB are most
commonly attributed to contact of blood with the foreign surface of the extracorporeal circuit.
However, other factors such as altered arterial blood flow patterns, shear stress generated by
blood pumps, cardiotomy suction devices, tissue ischemia and reperfusion, hypothermia,
anemia, and heparin, may be equally important (24). This inflammatory response may
continue long after CPB discontinuation, depending partly on its magnitude (25).

The repeated passage of blood through the nonphysiologic extracorporeal circuit initiates a
“contact activation” response with deposition of fibrin and activation of contact protein
cascades upon exposure of the blood to the nonendothelialized surface of the CPB machine.
Four proteins are involved in the contact activation cascades: coagulation factors XII and XI,
prekallikrein, and high-molecular-weight kininogen (HMWK). This activation forms bradykinin
and converts plasminogen into plasmin. This in turn initiates fibrinolysis and can trigger the
classical complement cascade. Therefore, multiple inflammatory mediators are released upon
exposure to the extracorporeal circuit (26). These responses initiate a sequence of chemokine-
mediated events that activate vascular endothelium and potentiate neutrophil-mediated
injury, which can disrupt hemostasis and create a generalized inflammatory response.

Cellular Components of Blood

Red Blood Cells

Tissue oxygenation depends upon continuous oxygen delivery, most of which is carried by the
hemoglobin molecules in the red blood cells (RBCs). The extracorporeal circuit exposes RBCs to
a challenging environment, thereby altering their integrity and function. Rheologic shear stress
forces mechanically damage RBCs to reduce their deformability, which is important to
maintaining normal RBC microcirculatory flow. As a further consequence of membrane
distortion, RBCs are susceptible to the membrane attack complex (MAC) generated by the
activation of complement (18), leading to hemoglobin leakage into the plasma. Free plasma
hemoglobin may impair tissue function by increasing plasma oncotic pressure and viscosity.
The ability of RBCs to aggregate also diminishes as a consequence of the dilution of blood by
the CPB circuit priming solution. In addition to the obvious potential reduction in tissue
oxygenation, this can trigger endothelial activation pathways (27).

Vascular Endothelium

Figure13.2 Adherence of circulating neutrophils (A) is triggered by activation of the endothelial

cells, which causes the neutrophils to roll along the endothelium (B), a process mediated by
selectin interaction with carbohydrate ligands. Activation of the neutrophils (e.g., by
interleukin [IL]-8, C5a, or platelet activating factor [PAF]) results in activation of integrin
molecules (e.g., by leukocyte function antigen [LFA]-1) on the neutrophil surface. This results in
firm adhesion of the neutrophils to activated endothelium (C) through receptors belonging to
the Ig family (e.g., intercellular adhesion molecules [ICAMs]). Activated adherent neutrophils
extravasate through the endothelium into the tissue interstitial space (D).
Under resting conditions, vascular endothelium offers a relatively inert surface that regulates
the passage of intravascular substrates to the extravascular space and ensures unhindered
flow of cellular and serum components through the capillary network. Endothelial cells are
extremely sensitive to the physiologic trespass imposed by CPB and surgical manipulation (and
hypoxia, should it occur). Release of endogenous substances such as cytokines (IL-1β and
tumor necrosis factor [TNF]-α), thrombin, C′5a, and lipopolysaccharide (endotoxin) activates
endothelial cells to increase surface expression of E-selectin, endothelial leukocytes adhesion
molecule (ELAM), intercellular adhesion molecule-1 (ICAM-1), and vascular adhesion molecule
(VCAM), all of which mediate leukocyte recruitment (28,29,30) (Fig. 13.2).

Leukocytes

The recruitment of leukocytes from the circulation to an inflammatory site is a multistep

process (Fig. 13.2) involving members of several adhesion receptor families (31,32). CPB
activates vascular endothelium, leading to rapid expression of members of the selectin family
(33). These are transmembrane molecules, expressed on the surface of leukocytes and
activated endothelial cells. Selectins contain an N-terminal extracellular domain with structural
homology to calcium-dependent lectins, followed by a domain homologous to epidermal
growth factor, and two to nine consensus repeats (CR) similar to sequences found in
complement regulatory proteins. Each of these adhesion receptors is inserted through a
hydrophobic transmembrane domain and possesses a short cytoplasmic tail. During
inflammation, bloodstream leukocytes attach to endothelium via selectins, thereby impairing
microcirculatory flow, potentially culminating in transient standstill from leukocyte adhesive
“plugs” along the endothelium. The selectin family includes P-, E-, and L-selectin. P-selectin is
largely responsible for the rolling phase of the leukocyte adhesion cascade. Transmigration,
the next step, is mediated by adhesion molecule upregulation, including platelet endothelial
cell adhesion molecule (PECAM)-1, CD99, ICAM-1, CD11a/CD18, very late antigen (VLA)-4, and
endothelium-secreting chemoattractants (34).

Neutrophils

Neutrophils are key mediators of the systemic inflammatory response; their recruitment,
activation, and cytotoxic capability are essential to warding off infection. Neutrophil activation
through interaction with activated vascular endothelium may be responsible for much of the
clinical sequelae of SIRAB. Activation of neutrophils during CPB is shown by the loss of L-
selectin and the upregulation of CD11b/CD18 (Mac-1). Increased production of reactive oxygen
intermediate (35) and neutrophil-derived elastase and similar molecules have also been
reported (35,36). Studies in complement-deficient dogs suggest that at least some of the
neutrophil activation seen in CPB is due to complement activation (37,38). In addition, the
release of cytotoxic products by activated cells can cause direct cellular damage. These
cytotoxins include both preformed agents present in neutrophil granules and newly
synthesized molecules. These novel substances include leukotrienes (36) and reactive oxygen
intermediates (35). This synthesis can be readily detected by a dramatic increase in cellular
oxygen consumption. These species are largely responsible for the so-called ischemia-
reperfusion injury. Ischemic injury occurs when the blood supply to a tissue is impaired or
suboptimal, which may occur with CPB. The paradox, however, is that a more severe tissue
injury occurs when blood flow is restored upon reperfusion.

Monocytes

Peripheral monocytes are also involved in the systemic inflammatory process. They possess
migratory, chemotactic, pinocytic, and phagocytic activities, as well as receptors for IgG and C
′3b. Upon migration to tissue, they undergo further differentiation to become tissue
macrophages, which participate in both specific and nonspecific immune pathways.
Recruitment of these cells occurs as early as 1 hour following insult (39) and is thought to be
mediated mainly by monocyte chemoattractant protein-1 (MCP-1), although other factors such
as C′5a play an important role. Upon activation, they play a pivotal role in inflammation,
serving as effector cells secreting cytokines (40). Monocytes can exhibit both pro- and anti-
inflammatory properties, depending on the signals they receive, and it has been suggested
that these cells have the capacity to switch from one state to the other to participate in both
the induction and the resolution of inflammation (39).

Platelets

CPB is associated with a transient deficit in platelet function and number, which can impair
postoperative hemostasis (41). The normal platelet adheres to a damaged endothelial cell or
the subendothelial layer. The multimeric form of von Willebrand factor forms a bridge
between endothelium and platelet at the platelet glycoprotein (GpIb) receptor site. The
platelet then undergoes a conformational change to expose different glycoproteins, including
the GpIIb/IIIa complex, which can in turn bind to fibrinogen. Fibrinogen is an essential cofactor
in platelet adhesion and in platelet-to-platelet binding during irreversible aggregation. The
protein complex thrombospondin stabilizes this platelet aggregate. Concurrent thromboxane
A2 release produces local vasoconstriction and further platelet aggregation.

Numerous CPB-related factors contribute to platelet changes. These include physical factors
(e.g., hypothermia and shear forces), exposure to artificial surfaces (42), drugs, and
endogenous chemicals (43,44). Hemodilution contributes to initial thrombocytopenia (45), but
mechanical disruption, adhesion to the extracorporeal circuit, and sequestration in organs
explains a disproportionate drop (beyond hemodilution alone) in platelet count, which can be
30% to 50% below baseline (46). The response of platelets to CPB is complex and
multifactorial, including rapid consumption of platelets during bypass (45), decreased
reactivity to known agonists (44), increase in the concentration of the α-granule compounds in
plasma (47), and an increase in the stable metabolite of thromboxane A2 (thromboxane B2)
released from aggregating platelets (48). Morphologic changes, including spherical appearance
and the development of pseudopods, also occur during CPB (49). Prolonged bleeding time
after CPB correlates with duration of bypass; however, its time course and precise mechanism
remain unclear (50).

Platelets activated during CPB form conjugates among themselves and between platelets and
leukocytes. Activated platelets express P-selectin, which contributes to leukocyte conjugate
formation by binding PSGL-1 (51). Activated platelets use this P-selectin/PSGL-1 adhesion
pathway to stimulate conjoined monocytes, thereby leading to secretion of the
proinflammatory cytokines IL-1β, IL-8, and MCP-1 (52,53). P-selectin also induces tissue factor
expression and fibrin deposition by monocytes, thereby contributing to thrombus evolution
(54,55).

Coagulation System

Figure13.3 Simplified summary of the coagulation cascade.

The coagulation cascade can be activated by either intrinsic or extrinsic pathways, each of
which consists of a series of enzymes. Figure 13.3 shows a simplified version of these
pathways.

The intrinsic pathway begins with Factor XII activation upon contact of blood with collagen in
the damaged vascular wall or an artificial surface, and ends with the formation of fibrin
through a cascade including activated factors XI, IX, X, and thrombin. Exposure of blood to
nonvascular tissue cells expresses a protein called tissue factor, which binds to factor VII and
activates the extrinsic pathway. This in turn activates factor X. Once factor Xa is generated, the
remainder of the cascade is the same as the intrinsic pathway.

Coagulation cascade activation induces the fibrinolytic system to coactivate, which limits the
amount of fibrin clot formed and localizes clot formation to the injury site. CPB activates both
the clotting and fibrinolysis pathways. Not only does this cause fibrin formation and platelet
activation, but also it activates vascular endothelium. Thrombin-mediated endothelial cell
activation elicits leukocyte adhesion as a result of selectin expression. In addition, the factor
VIIa–tissue factor complex and factor Xa activate cells through protease-activated receptors,
and this, in turn, generates cellular responses similar to those mediated by thrombin activation
of protease-activated receptor-1 (56). CPB-induced increases in factor XIIa and kallikrein
activate fibrinolysis, which has been linked to increased postoperative bleeding time, blood
loss, and reduction in platelet adhesion and aggregation capabilities from redistribution of
glycoprotein Ib, and IIb/IIIa receptors (57,58).

Humoral Components of Blood

Kirklin and McGriffin (3) initially hypothesized that many of the deleterious effects of CPB
derived from exposure of blood to nonendothelial surfaces, which initiates a “whole-body
inflammatory response.” They characterized this response as activation of coagulation, the
kallikrein system, fibrinolysis, and complement (3,59). We now also recognize the importance
of cytokines (60) and the combined effects of these humoral cascades in the activation of
endothelial cells and neutrophil adhesion.

Inflammatory Cascades

The principal event is the activation of factor XII (Hageman factor), which stimulates a number
of inflammatory systems (Fig. 13.3). After surface contact, factor XII undergoes a
conformational change and attaches to the protein HMWK, which is involved in the early
stages of intrinsic pathway activation and is a precursor of bradykinin.

This complex attaches to the foreign surface and, after limited proteolysis, releases kallikrein,
bradykinin, and more factor XIIa. Factor XIIa can initiate the intrinsic coagulation cascade by
activating factor XI, which binds to the foreign surface and activates factor VII, thereby further
augmenting the intrinsic coagulation cascade. Factor XIIa also provides a positive feedback
loop by inducing prekallikrein to form kallikrein, which in turn activates factor XII to produce
more factor XIIa (61). Kallikrein activates neutrophils, which further activate inflammatory
cascades to produce oxygen free radicals and proteolytic enzymes. Furthermore, both
kallikrein and bradykinin stimulate the fibrinolytic system. Kallikrein stimulates plasmin
production by its action on prourokinase and bradykinin releases tissue-type plasminogen
activator from the endothelium.

Complement System

The complement system comprises one of the preeminent immunologic mechanisms involved
in the inflammatory process and it is activated by CPB (62). More than 30 complement system
proteins serve both as an immune response effector arm and as a primitive self- versus non-
self-recognition system. Complement system functions include mediating inflammation,
opsonization of antigenic particles, and causing membrane damage to pathogens.
Complement components interact such that the products of one reaction form the enzyme for
the next. Therefore, a small initial stimulus can trigger a cascade of biologic activation and
amplification. The fact that 5% to 10% of serum proteins are complement components attests
to its importance. Blood which collects in the pericardium has very high levels of Tissue Factor
which, if aspirated into the CPB equipment using the cardiotomy suction, can accelerate the
inflammatory response if reinfused back into the patient.
Figure13.4 The complement systems. C1q, C3b, and C5a represent key functional elements that
crosstalk in multiple directions. Formation of the MAC leads to lysis, but also leads to cytokine
secretion. MBL, mannan-binding lectin; MAC, membrane attack complex.

The complement system (Fig. 13.4) was previously interpreted as a cascading system with a
common final pathway forming the MAC. Current understanding views complement as a
dynamic network that contains several hubs such as C1q, C3b, and C5a, each of which is highly
integrated within other pathways (63). Complement activation not only forms the MAC but
also interacts complexly with other systems such as proinflammatory signaling pathways, lipid
metabolism, and adaptive immunity (Fig. 13.4). Complement inhibition has been suggested as
a promising target for immunomodulative therapy in the context of cardiac surgery (64).

Further Contributions to Systemic Inflammatory Response after Bypass

In addition to the inflammatory interactions that result from the bypass circuit, cardiogenic
shock and endotoxemia may compound the response (65). Cardiogenic shock can contribute to
SIRAB, and prolonged nonpulsatile perfusion or periods of circulatory arrest can also lead to
diffuse end-organ ischemia (66). The end-organ hypoxic insult likely causes endothelial cells,
circulating monocytes, and tissue-fixed macrophages to release cytokines and oxygen-derived
free radicals that drive this response. After resuscitation from shock and hypoxic end-organ
reperfusion, a systemic ischemic-reperfusion injury ensues (65).

Endotoxemia represents another form of inflammatory activation that results from

extracorporeal circulation and ischemia-reperfusion. Frequently detected in high
concentrations in the systemic circulation after CPB (67), endotoxin potently induces
macrophage TNF release, which likely explains its increased concentrations after CPB in some
patients. In a rodent CPB model, brain TNF-α levels were elevated in the CPB group compared
to nonoperative control animals. However, a similar effect was seen in animals that underwent
cannula placement without CPB (68). These data suggest that TNF-α release and associated
neurologic injury also occur from surgical manipulation without CPB (69).
IMMUNE RESPONSE AFTER CARDIOPULMONARY BYPASS

The cellular and humoral constituents of the adaptive immune system undergo changes in
both function and number after CPB. The concern is that postoperative infection partially
derives from CPB-induced loss of T- and B cells and associated immunoglobulin consumption.
This section examines the effect of CPB on adaptive immunity in general and T-cell function in
particular.

Immunodeficiency and Cardiopulmonary Bypass

Duration of CPB correlates with postoperative infections (70). CPB depresses immunologic
reactivity, thereby increasing susceptibility to perioperative infections and the potential for
septic shock (71).

Normally, relatively low levels of immunoglobulins and complement provide adequate

bacterial opsonization (72). However, if the load of infectious microorganisms increases, as
may occur in patients undergoing open-heart surgical procedures, then an imbalance may
develop in host defense homeostasis. Quantitative and qualitative exhaustion of humoral and
cell-mediated immune mechanisms may adversely affect clinical outcome when additional
injury occurs postoperatively in a patient with a downregulated immune system. It is therefore
unsurprising that opportunistic pathogens are frequently found in post-CPB patients who
experience complications such as renal failure or low cardiac output. Post-CPB patients exhibit
an increased risk for sepsis-related multisystem organ failure, which accounts for a high
proportion of perioperative fatalities (73).

Humoral Immunity

Serum levels of immunoglobulins and complement change markedly during CPB (21,74,75),
with a resultant reduction in host defenses, for example, reduced opsonization of bacteria in
vitro (76). Leukocyte counts fall beyond the affect of hemodilution at the onset of CPB as a
result of tissue sequestration after activation by anaphylatoxins C′3a and C′5a. After CPB,
granulocyte chemotaxis is impaired (77), which likely increases susceptibility to bacterial
infections. CPB also impairs leukocyte phagocytosis and metabolic function during and after
CPB (106).

Some investigations report unchanged B-cell numbers after CPB (18), whereas others
demonstrate a drop (107,108). Roth et al. reported that relative B-cell levels (percentage)
increase after CPB, with no change in their absolute number, which may result from a
combination of hemodilution and immunosuppression. The secretion of IgG, IgM, and IgA by B
cells in response to pokeweed mitogen diminishes after CPB. Recent work by Lante et al.
showed that IgE levels did not change after CPB until day 3; in contrast, both IgG and IgM
levels decreased significantly on day 1. IgM returned to baseline on day 5, at which time IgG
levels remained low. This was associated with a reduction in the B-cell count on day 1 and a
return to baseline by day 3. The bactericidal activity of serum is depressed after CPB. Whereas
complement is consumed during CPB, other components of humoral immunity decrease in
proportion to CPB hemodilution.
When immune proteins contact gaseous and foreign surfaces, denaturation ensues. Surface
depolarizing forces disrupt sulfhydryl and hydrogen bonds that stabilize the secondary and
tertiary structure of the protein molecules. This unfolds globular protein molecules, which may
create a randomly coiled molecule and expose previously masked chemical groups. The
macromolecules so formed then tend to flocculate. Plasma protein denaturation may also
coalesce serum lipids and increase plasma viscosity (75).

Mast Cells and Paneth Cells

Mast cells are sentinels that herald gut inflammation. In a rodent model of CPB and deep
hypothermic circulatory arrest, intestinal mast cells in conjunction with Paneth cells mediate
ischemia/reperfusion injury (Fig. 13.5) (78). The impact of unappreciated acute gut injury on
adverse outcomes after cardiac surgery is only recently being appreciated {Karhausen, 2014
#1308}. Since mast cells may be activated by complement, suppressing this activation may
prove valuable in attenuating CPB-induced ischemia/reperfusion injury.

Figure13.5 Paneth cells and intestinal mast cells release potent effectors to regulate local injury
and systemic inflammation after intestinal ischemia/reperfusion. Most prominently, the Paneth
cell-dependent pathway (blue) depends on release of interleukin (IL)-17 from Paneth cells
localized at the base of small-intestinal crypts. Mast cell responses (purple) use a number of
preformed and de novo synthesized products such as proteases (tryptase, mast cell protease
[MCP]-4), lipid mediators (leukotrienes, prostaglandins), and cytokines (tumor necrosis factor
[TNF]-α, IL-6).(Figure prepared by Annemarie B. Johnson, C.M.I., Medical Illustrator, Vivo
Visuals, Winston-Salem, NC, for Bartels K, Karhausen J, Clambey ET, et al. Perioperative organ
injury. Anesthesiology 2013;119:1474–1489. Copyright © 2013, American Society of
Anesthesiology. Reproduced with permission from Wolters Kluwer Health.)

Natural Killer Cells

Natural killer cells, a heterogeneous lymphocyte subpopulation that is neither T nor B, produce
cytotoxic responses in virus-infected cells and transformed target cells (e.g., tumor cells).
Decreases in both number and function of natural killer cells occur after CPB (79,80).

Reticuloendothelial System

The reticuloendothelial system comprises tissue macrophages in the spleen, lymph nodes,
lung, and liver. Its cells derive from blood-borne monocytes. Normal reticuloendothelial
function includes clearing the blood of bacteria, endotoxins, platelets, denatured proteins,
chylomicrons, plasma hemoglobin, thrombin, fibrin, fibrin degradation products,
thromboplastin, and plasminogen activator. CPB introduces microparticles into the
bloodstream, which in turn depresses reticuloendothelial system function (81).

Cells

T cells are lymphocytes that develop in the thymus, which is seeded by lymphocytic stem cells
from the bone marrow during embryonic development. Following T-cell development, mature
naive T cells leave the thymus and begin to spread throughout the body. These cells then
develop their T-cell antigen receptors and differentiate into the two major peripheral T-cell
subsets, one of which expresses the CD4+ marker (helper cells) and the other CD8+ (cytotoxic
cells). Proliferating T helper cells can differentiate into two major subtypes on the basis of the
specific cytokines they produce, known as TH1 and TH2. T helper cells play a central role in the
initiation and regulation of the acquired immune response.

T helper cells recognize antigen presented on the surface of antigen-presenting cells in

association with class II molecules encoded by the major histocompatibility complex (MHC). T-
cell activation requires other specific costimulatory signals generated by the antigen-
presenting cell. Cytotoxic T cells recognize antigen presented on the surface of antigen-
presenting cells in association with class I molecules encoded by the MHC.

T helper cells provide “help” in the form of lymphokine secretion. Such lymphokines help B
cells to divide, differentiate, and produce antibodies. Lymphokines are also required for the
development of leukocyte lines from hematopoietic stem cells and development of cytotoxic T
cells. They also cause activation of macrophages, allowing them to destroy the pathogens they
have taken up. Cytotoxic T cells are capable of destroying virus-infected target cells or
allogeneic (transplanted) cells.

Morphologic Correlation

Scanning electron microscopy shows profound alterations to the plasma membrane of the T
cells after CPB. The number of microvilli and lymphocyte surface folding both decrease.
Furthermore, after CPB, the membrane does not accommodate monoclonal microbeads.
Quality and Quantity

Phenotypic changes induced by CPB have been investigated using standard commercially
available leukocyte-labeling monoclonal antibodies together with flow cytometry. CD3+ and
CD4+ cells decrease and CD8+ cells increase, hence the CD4+/CD8+ ratio increases. Such
changes peak on postoperative day 1 but remain abnormal for approximately 1 week (72). T-
lymphocyte counts decrease after CPB (82). The capacity of the immune system to counteract
microbial infection decreases after cardiac surgery because of consumed (and unreplaced)
complement factors and decreased cellular immune elements such as neutrophils and natural
killer cells (79,80,81,82).

Mechanisms

The mechanism responsible for the decrease in circulating T lymphocytes after CPB remains
undefined. Elevated cortisol levels probably play an important postoperative
immunosuppressant role, yet this is unlikely to be the only factor in this complex
phenomenon. Elevation of serum corticosteroids may induce redistribution of circulating T
cells to lymphoid tissues, although this is usually mild and transient. Although serum cortisol
levels are elevated after surgery, the lack of correlation between changes in serum cortisol and
changes in lymphocyte number and function suggests that serum cortisol is not an etiologic
factor in postoperative T-cell dysfunction. Post-CPB reductions in T-cell activity are both
quantitative and qualitative, likely causes for which include hemodilution, intravascular-to-
extravascular fluid shifts, mechanical destruction, consumption, and redistribution among
bone marrow, lymphoid tissue, and peripheral blood (80,83).

THERAPEUTIC APPROACHES

Effective amelioration of SIRAB has yet to become a reality. Difficulty in maintaining adequate
perfusion pressure toward the end of bypass suggests a profoundly low systemic vascular
resistance, which is consistent with SIRAB. Vasoconstrictor and inotropic drugs constitute the
primary current treatment of this situation, although this does not address the
pathophysiology. To prevent or ameliorate SIRAB, the optimal strategy would be to develop a
CPB circuit that does not induce contact activation of blood components.

Pharmacologic Manipulation

Pharmacologic attempts to suppress the inflammatory response to cardiac surgery have failed
thus far to improve clinical outcomes (84). This is likely due to the fact that the immunologic
response may not be pathologic per se (as it is in autoimmune diseases), but rather represents
a secondary response to the altered physiology of CPB.

Corticosteroids

The physiologic effects of corticosteroids are numerous and widespread. They possess anti-
inflammatory properties and influence the water/electrolyte balance and the metabolism of
carbohydrate, protein, and lipid. Single-dose dexamethasone for cardiac surgery has failed to
affect 30-day mortality in a randomized, controlled clinical trial of 4,494 patients (85). A
subgroup of this trial underwent cognitive testing at 1 and 12 months following surgery and
failed to show any improvement in cognitive function (86). Similarly, preliminary results of the
Canadian Steroids In Cardiac Surgery Trial (SIRS Trial) (NCT00427388) have not been promising.
Further studies might investigate a role for steroid treatment using other doses, administration
protocols, or types of steroids.

Alternative Strategies

The C5 complement inhibitor pexelizumab has been studied in several clinical trials.
Pexelizumab was associated with only a 6.7% nonsignificant reduction in the primary
composite end point of death or myocardial infarction in the PRIMO-CABG II trial (87). These
results were surprising as the previous PRIMO-CABG I trial had suggested more favorable
outcomes (88,89). Furthermore, pexelizumab administration had no effect on global cognitive
outcomes after cardiac surgery, with the exception of a possible effect to reduce dysfunction in
the visuo-spatial domain (90). An exploratory analysis of the combined PRIMO I and II data
(7,353 patients) suggested a mortality benefit for certain high-risk cardiac surgery patients
(87). Furthermore, a meta-analysis of pexelizumab in ischemic heart disease suggested
benefits in patients undergoing CABG (91).

An interesting novel approach is represented by the administration of cyclosporine.

Cyclosporine inhibits the opening of the mitochondrial permeability transition pore and its
ability to prevent ischemia/reperfusion injury in patients undergoing aortic valve replacement
has been studied in a small population (61 patients) (92). Intravenous cyclosporine bolus
dosing reduced the area under the curve for postoperative cardiac troponin I postoperatively
by 35%. Whether these preliminary results will translate into improvement of more
meaningful clinical outcomes needs further study.

Mechanical Manipulation

Depletion of Leukocytes and Inflammatory Mediators

Conducting ultrafiltration at the termination of CPB, such as modified ultrafiltration (MUF),

aims to reverse hemodilution and decrease tissue edema and circulating inflammatory
mediators. Reducing the number of circulating leukocytes by the use of leukocyte-depleting
filters is another way to modify the inflammatory process. Some studies report reduced lung
injury and improved oxygenation (93), as well as reduced hospital length of stay (94).

Mechanical Considerations

Pulsatile perfusion may have beneficial effects with respect to hemodynamics,

microcirculation, and organ dysfunction, compared with nonpulsatile flow (95), as
demonstrated by many experimental and clinical studies (96,97,98,99). A recent meta-analysis
of eight studies including 970 patients suggested improved postoperative pulmonary function
in addition to shorter intensive care unit (ICU) and hospital stay (100). However, as Murphy et
al. (101) point out, there have been over 150 basic and clinical science investigations on the
use of pulsatile versus nonpulsatile technique for CPB and the controversy continues because
the results are inconclusive.
Membrane oxygenators currently predominate over bubble oxygenators for cardiopulmonary
bypass in developed nations, as they provide more titratable arterial PO2 values and less blood
trauma, especially with prolonged use. Modern membrane oxygenators provide blood flow to
surrounding fibers that are filled with “inspired” gases that equilibrate across the membrane
(102).

Coating of oxygenator membranes and CPB tubing and filters with heparin inhibits the release
of proinflammatory cytokines and reduces complement and platelet activation (103,104). A
meta-analysis including 3,434 patients found that heparin-bonded circuits decreased the
incidence of blood transfusion, sternal reexploration for bleeding, duration of mechanical
ventilation, and hospital length of stay. Effects on other outcomes were minimal (105).

The adverse impact of blood contact with extracorporeal surfaces, air–fluid interface, and cell
damage by cardiotomy suction associated with conventional bypass has led to the
development of minimized cardiopulmonary bypass circuits (mini-CPB). Such circuits use a
closed CPB system characterized by reduced surface area and priming volume, elimination of
cardiotomy suction, and prevention of air–blood contact. Studies comparing mini-CPB to
conventional CPB show delayed or reduced secretion of proinflammatory cytokines,
attenuated complement activation, and blunted leukocyte activation (106,107,108). The use of
mini-CPB was associated with less hemodilution and thus less blood transfusion (106,107,108).
It remains unclear if this approach improves major clinical outcomes, but interpretation is
complicated by the use of different mini-CPB systems and the heterogeneity of CPB protocols.

SUMMARY

It appears likely that the inflammatory and immunologic sequelae of CPB are not primarily
responsible per se for a large part of the morbidity and mortality seen postoperatively,
although they may contribute. This is highlighted by the lack of obvious deleterious effects of
CPB in the more recent off-pump/on-pump CABG prospective comparison trials. However,
perioperative inflammation assumes greater importance in longer, more complex surgery
performed on patients who are at extremes of age and who have significant comorbid
conditions. Furthermore, it appears as if prevention of inflammation through minimizing and
preempting iatrogenic injury is more likely to succeed than pharmacologic intervention after
the injury has occurred. While some substances have shown promise (e.g., the C5 complement
inhibitor pexelizumab), effects on meaningful clinical outcomes have been inconsistent.

KEY Points

● CPB is associated with a systemic inflammatory response that includes cellular and
noncellular elements.

● Results from recent large-scale clinical trials comparing off-pump CABG surgery to
conventional CABG question the clinical relevance of inflammation attributable to CPB
alone.
● The complement system is a dynamic network that is highly integrated within other
pathways, and is activated during CPB.

● Use of complement inhibitors (e.g., pexelizumab) remains a target for

immunomodulative therapy in the context of cardiac surgery.

● Mini-CPB show promise to attenuate the inflammatory response to CPB through

surface area and priming volume reductions, cardiotomy suction elimination, and air–
blood contact prevention.

● Minimizing occurrence of inflammation by preempting iatrogenic injury is more likely

to be successful than pharmacologic intervention after the injury has occurred.
CHAPTER 15

NEUROLOGIC EFFECTS OF CARDIOPULMONARY BYPASS

Eduardo S. Rodrigues and David J. Cook

This chapter will provide an overview of the neurologic effects of cardiac surgery and
cardiopulmonary bypass. First, the extent and nature of bypass-related neurologic
injury will be presented, emphasizing the importance of patient demographics and
comorbidities. In this context etiologies will be considered. Second, basic cerebral
physiology during cardiopulmonary bypass will be reviewed. Third, interventions
having the potential to reduce neurologic morbidity will be considered and fourth, the
applicability of neurologic monitoring techniques for cardiopulmonary bypass will be
briefly presented.

THE POPULATION AT RISK, DEMOGRAPHICS, AND COMORBIDITIES

Cardiac operations requiring cardiopulmonary bypass (CPB) are among the most
commonly performed surgical procedures in Europe and North America. Studies
describing more than 20,000 cardiac surgical patients indicate that significant
morbidity may be experienced by 20% to 25% of patients (1,2,3). These morbidities
include cardiac and pulmonary failure, neurologic injury, renal insufficiency, bleeding,
and infection. Complications have predictable adverse effects on duration and cost of
hospitalization (4,5,6) as well as on patient quality of life and functional capacity on
discharge (6,7,8).

When considering cardiac surgical neurologic morbidity, it is important to understand

that many adult cardiac surgical patients have preexisting risk factors for stroke and
cognitive impairment even without the added risk of cardiac surgery and CPB. Cardiac
or noncardiac surgery superimposes incremental risk. The recent European study of
cognitive outcomes after noncardiac surgery superbly illustrated this risk (9). In that
study, 1,218 elderly patients having major noncardiac surgery underwent
neurocognitive assessment preoperatively, before discharge, and at 3 months
postoperatively. A large (n = 321) nonsurgical control group was also enrolled. That
study reported a 26% incidence of cognitive dysfunction 1 week after surgery and a
10% incidence at 3 months. Cardiac surgery poses greater risks because cardiac
surgical patients experience more serious neurologic morbidity than age- and health-
matched controls undergoing noncardiac surgery (Table 15.1) (8). Consequently, it is
useful to regard the perioperative period as one of acute neurologic stress in a
subpopulation chronically at risk.
Table15.1 Severity of postoperative neuropsychological deterioration in CABG and
surgical control patients
Figure15.1 Incidence of type I and type II cerebral outcomes according to age.(Roach
GW, Kanchuger M, Mangano CM, et al. Adverse cerebral outcomes after coronary
bypass surgery. Multicenter Study of Perioperative Ischemia Research Group and the
Ischemia Research and Education Foundation Investigators. N Engl J Med
1996;335:1857–1863. Copyright © 1996 Massachusetts Medical Society. All rights
reserved.)

There is general agreement about preoperative patient-related factors that increase

the likelihood of perioperative stroke, but the role of many intraoperative factors is
much less clear. Preoperative risk factors include advanced age, history of prior
neurologic events, aortic atherosclerosis, low cardiac output states, atrial arrhythmias,
hypertension, and diabetes. Advanced age, the most dramatic risk factor, appears to
magnify the other risk factors. The importance of this risk factor cannot be
overemphasized, because over the last two decades the number of cardiac surgery
patients over age 60 doubled and the percentage of patients over age 70 increased 7-
fold (10). The proportion of cardiac surgical patients over age 80 will grow faster than
any other group. Age alone is an independent risk factor for stroke, with patients
under age 60 having ≤1% incidence of stroke across studies (1,10,11), while those over
age 70 have a 4% to 9% incidence of stroke or coma following operation (Fig. 15.1)
(1,10,11,12). This is important because the increasing age of the surgical population
and the successes provided by the initial heart surgery are increasing the incidence of
redo heart surgery (13) which is known to increase neurologic risk. In a recent
multicenter, retrospective study, Biancari et al. (14) evaluated the risk factors for
stroke in a population undergoing redo heart surgery. This study included seven
centers in three different European countries with a total of 741 patients. The overall
incidence of stroke was 6.5%, but extremely variable depending on patient gender and
perioperative events.

As the surgical population has aged, the proportion of patients with multiple risk
factors for neurologic injury has increased. Hypertension and diabetes occur in
approximately 55% and 25% of cardiac surgical patients (10,15,16,17,18,19). Fifteen
percent demonstrate ≥50% carotid stenosis and up to 13% may have had a TIA or prior
stroke (15,19). When compared to previous cardiac surgical patients, these
intercurrent diseases may double the risk of perioperative stroke for the present and
future patients (10,20).

The Cardiovascular Health Study (CHS) enhances appreciation for preexisting cerebral
morbidity and cerebral risk in the general population (21). In that investigation, a
community population of 3,360 individuals over 65 years of age underwent MRI. Of
these, 31% had “silent” cerebral infarcts (21). These infarcts were primarily subcortical,
lacunar, and strategically placed or small enough so that clinical signs of stroke were
not evident. However, detailed neurologic testing suggested the possibility that these
infarcts may have accounted for cognitive and gait abnormalities (21). These lesions
likely are related to chronic hypertension, which narrows the penetrating vessels
supplying deep white matter, thereby rendering these regions vulnerable to focal
ischemia (22,23). This type of cerebral vascular disease is clearly linked to late-onset
dementia (24,25).

These findings have implications for the cardiac surgical population. Of the 5,888 men
and women enrolled in the CHS, those who underwent MRI were “younger… and were
more likely never to have smoked and less likely to have prior cardiovascular disease,
hypertension …and diabetes than those who did not undergo scanning” (21,26). As
such, the cardiac surgical population probably has an incidence of preexisting “silent”
cerebral infarction in excess of 31%. One study of 31 neurologically asymptomatic
CABG patients reported a 16% incidence of thromboembolic infarcts and a 58%
incidence of lacunar infarcts on preoperative MRI (27).

In addition to the clear predisposition of chronic cerebrovascular disease to adverse

neurologic outcome in cardiac surgery, investigation into other patient factors that
may determine or modulate brain outcomes continues. Females tend to have worse
neurologic outcomes (14,28,29), perhaps in part because of referral bias and small
vessel size, but genetic sex-linked factors may also play a role (16,30). Another active
area of investigation is genomics. To date, only apolipoprotein E4 (ApoE4) has been the
subject of significant investigation in cardiac surgical outcomes (31,32,33,34,35).
ApoE4 is a protein that plays an important role in cholesterol transport and
metabolism, with three common alleles resulting in E2, E3, and E4 isoforms (36). The
apolipoprotein alleles define a spectrum of LDL (low-density lipoprotein) cholesterol
levels (37). Of the isoforms, the ApoE4 allele has a clear link to the development of
atherosclerosis, cerebrovascular disease, and dementia (37,38,39,40).

In cardiac surgical patients, the relationship of patient factor ApoE4 to cerebral

autoregulation (33), inflammatory cytokines (35), as well as markers of brain injury
S100 and neuron-specific enolase (NSE) has been investigated (31). Because the
absence of ApoE4 has been linked to better recovery from neurologic injury, the most
interesting investigation to date was by investigators from Duke University who
identified an association of the ApoE4 allele with worse cognitive outcomes in 65
patients who underwent cardiac surgery (32). However, the frequency of the E4 allele
was only 13%; so the study was underpowered. Additionally, the authors (32) point out
that an association of E4 and adverse neurologic outcomes, or markers of neurologic
outcomes (31,36), might simply be related to a greater burden of atherosclerotic
disease in the E4 patients rather than a more specific interaction of the gene and
neurologic outcome.

Neurologic Complications in Cardiac Surgery

Neurologic injury following cardiac surgery has been a source of concern since its
inception and this injury can be generally divided into frank stroke, encephalopathy,
and neurocognitive disorders. The incidence of major neurologic morbidity related to
cardiac surgery is 1% to 6% (5,11,14,16,41,42), although select populations may have a
clinically evident stroke risk as high as 8% to 9% (1,6). It is important to note that a
recent prospective study demonstrated that the intensity of evaluation affects
reported stroke rate; with greater scrutiny, up to 17% of the patients may be
diagnosed with postoperative stroke, and if cerebral ischemia is assessed by MRI, up to
32% to 54% of the patients will show evidence of cerebral insults (43,44).

Two prominent multicenter studies from the Multicenter Study of Perioperative

Ischemia (McSPI) group examined neurologic outcomes and their predictors in patients
undergoing CABG (5), or combined CABG and open-ventricle procedures (6). Morbid
neurologic outcomes were classified as type 1 (cerebral deaths, nonfatal strokes, and
new transient ischemic attacks (TIAs) or type 2 (new intellectual deterioration at
discharge or new seizures). Of 2,108 patients in the CABG study, 3.1% had type 1 and
3% had type 2 outcomes (5). The predictors of type 1 outcomes were aortic
atherosclerosis, a history of neurologic disease, and older age; while the predictors of
type 2 outcomes were age, systolic hypertension, pulmonary disease, and excessive
alcohol consumption (Table 15.2). In 273 patients undergoing combined CABG and
open-ventricle procedures, adverse neurologic outcomes occurred in 16% (8.4% type 1
and 7.3% type 2) (6). In that study, predictors of type 1 outcome included aortic
atherosclerosis or thrombus, while type 2 outcomes were predicted by aortic
atherosclerosis, a history of endocarditis, postoperative low cardiac output state, a
history of alcohol abuse, perioperative arrhythmias, and elevated systolic blood
pressure on admission (6). The latter study also demonstrates the adverse effect of
neurologic morbidity on ICU length of stay, total duration of hospitalization, and
discharge disposition.
Table15.2 Adjusted odds ratios for type I and type II cerebral outcomes associated with
selected risk factorsaaOdds ratios are for the risk of a type I or II outcome in patients
with the risk factor in question as compared with those without the risk factor. Odds
ratios have been adjusted for all the factors listed for each model. Excessive alcohol
consumption indicates previous hospitalization because of alcohol consumption or
alcohol withdrawal. Perioperative hypotension indicates a systolic blood pressure <80
mm Hg (during surgery but before cardiopulmonary bypass or after bypass) or <40 mm
Hg (during bypass) for more than 10 minutes. CABG denotes coronary artery bypass
graft.Roach GW, Kanchuger M, Mangano CM, et al. Adverse cerebral outcomes after
coronary bypass surgery. Multicenter Study of Perioperative Ischemia Research Group
and the Ischemia Research and Education Foundation Investigators. N Engl J Med
1996;335:1857–1863. Copyright © 1996 Massachusetts Medical Society. All rights
reserved.

In the cardiac surgical population, neurocognitive morbidity (i.e., intellectual function)

is much more frequent than frank neurologic morbidity (i.e., defects present on
neurologic physical examination). The incidence of neurocognitive deficits at discharge
ranges from approximately 10% to 80%, with 5% to 20% of deficits still being present 3
to 6 months after surgery (42,45,46,47,48,49,50). Because the results of cognitive
testing prior to discharge may be confounded by pain, drugs, and sleep deprivation
and by long-term CNS recovery capabilities, longer-term assessment is much more
important than predischarge assessment.

Large variations in the reported incidence of neuropsychological impairment result

from several factors. First, the incidence may vary 2-fold depending on whether the
investigation is retrospective or prospective (51). Results also differ depending on the
batteries of tests performed and the definition of an abnormal test result. Differences
may also arise secondary to the timing of test administration and to preoperative
variables such as preexisting deficit or impaired language skills (52,53). A 1995
consensus conference addressed these issues and better agreement now exists over
appropriate test batteries, test timing, and the definition of an abnormal result
(54,55,56).

In spite of methodologic differences, specific patterns of deficits are commonly

described. These include deficits in psychomotor speed, attention and concentration,
new learning ability, and short-term memory (Table 15.3) (8,49,50,56,57,58).
Ophthalmologic abnormalities have also been described in up to 25% of patients and
primitive reflexes may be seen in 39% (42). While a significant proportion of deficits
are clinically silent and become evident only on testing, Shaw et al (8) reported a 38%
incidence of significant intellectual impairment and a 10% incidence of overt disability
in 235 patients showing postoperative neuropsychological test score deterioration.
Table15.3 Outcome changes within each cognitive domainaaEach cognitive domain is
considered as a separate entity, and thus, an individual patient will have different
outcomes in different domains.McKhann GM, Goldsborough MA, Borowicz LM Jr, et al.
Cognitive outcome after coronary artery bypass: a one-year prospective study. Ann
Thorac Surg 1997;63:510–515.

The pattern of neurologic injury in children undergoing cardiac surgery differs from
that of adults, suggesting a different etiology. Nonstroke neurologic impairment in
children more often includes seizures, movement disorders, and developmental delays
(59). The incidence of these events in children is influenced by the use of deep
hypothermic circulatory arrest and by its duration when it occurs (60,61).

Epidemiologic, clinical, and neuroimaging investigations continue to refine the

understanding of postcardiac surgical cognitive outcomes. Historically, the use of CPB
has made cardiac surgery unique, and the working assumption has been that adverse
cognitive outcomes were directly related to that physiologic experience. However,
more recent data do not bear that out. If CPB is responsible for cognitive injury, one
must predict that patients undergoing off-pump coronary artery bypass grafting
(CABG) would have significantly different cognitive outcomes than those having CABG
surgery with bypass. While most reports have generally been small or nonrandomized,
one prominent, randomized comparison and meta-analysis and a sufficient number of
patients (62,63) demonstrated virtually no effect of CPB on cognitive decline. In that
report, the incidence of impairment 3 months postoperatively was 21% in the off-
pump group versus 29% in the on-pump group (p = 0.15). At 1 year in the off- and on-
pump groups, the incidence was 31% and 34%, respectively (p = 0.69) (62).
Additional data outside cardiac surgery have shed light on “postcardiac surgical”
cognitive dysfunction. A 25% to 30% incidence of cognitive impairment has also been
clearly demonstrated in older patients undergoing major vascular, orthopedic, and
thoracic surgeries (9). Finally, when Selnes and colleagues (64) compared cognitive
outcomes at 3 and 12 months in age- and health-matched cardiology and cardiac
surgical patients, they were unable to identify meaningful cognitive outcome
differences. Wahrborg and colleagues (65) made similar observations in a smaller
study. While the incidence of neurocognitive disorders in noncardiac surgical patients
is usually lower, this may be a function of the severity and duration of the stresses
associated with cardiac surgery rather than the uniqueness of the experience itself.

If cognitive outcomes are not determined by the presence or absence of CPB and if
similar cognitive deficits are seen in elderly noncardiac surgical patients and in age-
and health-matched cardiology patients, one might logically conclude that cognitive
decline is a function of the elderly brain and periprocedural factors rather than specific
intraoperative events in cardiac surgery. MRI results of the Cardiovascular Health
Survey suggest the possibility that perioperative stressors mask underlying disease, as
supported by cerebral imaging studies (21,26) and in the significant incidence of
preoperative cognitive dysfunction in cardiac surgical patients (66) (Table 15.4). This
conclusion has tremendous implications for redirecting the thrust of clinical
investigation in this area. If postcardiac surgical cognitive dysfunction does not
primarily result from intraoperative events, this would explain why attempts to reduce
adverse cognitive outcomes with perioperative interventions (including mean arterial
pressure [MAP] (67), bypass temperature (47,48,68,69), CO2 and glucose management
(70,71,72), pulsatility (72), pharmacologic neuroprotectants (73,74,75), and surgical
devices (76)) have not shown a consistent effect on cognitive outcomes. On the other
hand, improvements in surgical techniques have reduced the number of
neuropsychological deficits by one-half since 2005 (77,78).
Table15.4 Subclinical preoperative cognitive impairment

There are likely to be parallels with postoperative delirium in cardiac surgery, but
isolating the specific effect of cardiac surgery has not been evaluated by comparing
this complication in cardiac surgery to noncardiac surgery patients. The very high 26%
to 52% incidence of postoperative delirium after cardiac surgery is frequently
underappreciated. Not surprisingly, delirium is associated with poor outcomes
including long-term cognitive dysfunction, increased risk of stroke risk, and mortality. A
large fraction of these patients present with hypoactive delirium and if not actively
looked for this complication is vastly underrecognized. Unfortunately, strategies to
prevent this complication have been largely unsuccessful in the cardiac surgery
population (79) and with population aging, better understanding the etiology of
postoperative delirium is a compelling task ahead.

SURGICAL STRESS AND NEUROLOGIC RISK

Although the perioperative period contains atheroembolic risk, hypo- and

hypertension, anemia, arrhythmias, dehydration, and disorders of coagulation, the
contribution of these factors to neurologic risk and outcome remains difficult to
define. Since the incidence of frank neurologic injury in the surgical population is
relatively low, the contribution of individual physiologic factors to neurologic and
neuropsychological outcomes is difficult to determine. Because of the relatively high
incidence of postoperative neurocognitive disorders, these outcomes offer the
potential to be more revealing than neurologic morbidity. Nevertheless, with few
exceptions (58,77,78,80,81,82,83,84), surgical or medical interventions to date have
failed to demonstrate a clear effect on neuropsychological outcomes. It is also difficult
to isolate perioperative risk factors, because several risk factors are both common and
closely associated with each other. One should therefore regard neurologic morbidity
as resulting from the combination of a patient chronically at risk with a variety of
surgically related insults or stresses for which the vascular system is unable to
compensate.

In the absence of hypoxia, neurologic morbidity results from inadequate tissue

perfusion. In its simplest form, ischemia may result from embolic obstruction of a large
or small vessel, from inadequate flow through a fixed or variable stenosis, or from a
failure of collateral circulation. At the level of the microvasculature, inflammation and
endothelial dysfunction could also play a role in compromising oxygen delivery and
neuronal functional integrity (85,86,87,88).

Embolization

Most patients experience cerebral embolization during CPB (58,89,90,91). Transcranial

Doppler (TCD) and echocardiographic (58,89,92), retinal angiographic (93,94),
pathologic, and radiographic information (95,96,97,98,99) indicate that cerebral
embolization during CPB may be the primary cause of serious brain injury during
cardiac surgery. These embolic events are related to aortic atheromata (100,101),
platelet–fibrin and leukocyte aggregates (93,102,103), and by bubbles generated in the
CPB circuit or in the surgically exposed left-sided circulation (58,104). The CNS may be
exposed to hundreds or thousands of particulate and nonparticulate emboli and the
periods of embolic risk are usually associated with specific surgical events (Fig. 15.2)
(58,83,89,92,105). While air embolization is a primary source of TCD signals, some
authors have observed better cognitive outcomes by reducing embolization (58,77). In
the study by Hammon and colleagues (77), the combination of epiaortic scanning,
single-clamp technique, and increased left ventricular venting decreased cognitive
decline modestly. At 1-month assessment, the incidence was 18% in the treatment
group versus 29% in a historical control. However, neuroimaging studies question the
importance of air embolization as a significant determinant of neurologic outcome
(106).
Figure15.2 Number of embolic events per minute in a representative patient from
before insertion of the aortic cannula to after end of cardiopulmonary bypass (CPB). X-
clamp, cross-clamp.(van der Linden J, Casimir-Ahn H. When do cerebral emboli appear
during open heart operations? A transcranial Doppler study. Ann Thorac Surg
1991;51:237–241.)

In contrast to air, the effect of particulate embolization can be well documented

clinically by focal neurologic signs or by neuroimaging studies. Of the different
neuroimaging techniques, diffusion-weighted magnetic resonance imaging (DW-MRI) is
probably the best tool for detecting cerebral ischemia in the early postoperative
period. This highly sensitive modality can detect ischemic areas as small as 0.4 mm in
size (107). Furthermore, in contrast to conventional neuroimaging modalities, for
example, computed tomography (CT) and standard T1- and T2-weighted MRI,
impairment of water diffusion occurs minutes after onset of acute ischemia, resulting
in high signal intensity with DW-MRI (107,108). This allows DW-MRI to detect early
ischemia as well as distinguish acute and chronic lesions. Although most studies have
enrolled small numbers of patients, postoperative DW-MRI has typically demonstrated
new cerebral ischemic events in about 30% to 50% of patients (44,109,110,111,112).
The largest investigation to use DW-MRI was at the authors’ institution where scans
were done in 50 cardiac surgical patients; in that series, the incidence of cerebral
ischemic events was 32% and all events were embolic in nature (43).

Hypoperfusion

Although there are no compelling data that MAP management during CPB is a primary
determinant of neurologic outcome, regional cerebral hypoperfusion probably occurs
during CPB secondary to hypertensive, diabetic, or senile atherosclerotic disease. Both
hypertension and diabetes mellitus are independent risk factors for neurologic
complications following cardiac surgery (5,19,113,114) and these forms of vascular
disease are present in at least 50% of adult cardiac surgical patients. In addition to a
rightward shift of the autoregulatory curve (115), chronic hypertension may narrow
penetrating arteries, decrease collateral blood flow, and reduce ischemic tolerance
(23). Similarly, diabetes results in macroangiopathies and degenerative changes in
cerebral penetrating arteries, increases the likelihood of embolization (116), and alters
cerebral autoregulatory capacity (117,118). Therefore, the increased neurologic risk
associated with hypertension and diabetes may result from either increased
embolization or regional hypoperfusion. However, it is reasonable to assume that the
greater morbidity in these patients may be a function of the interaction of these two
processes.

Therefore, even without compelling evidence that perioperative MAP is a primary

determinant of neurologic outcome, regional hypoperfusion (either as a direct result of
vascular disease, or as the inability to compensate for the regional ischemia associated
with microembolization), should remain an essential component of our model for
understanding perioperative cerebral ischemia. Depending upon the patient’s
preexisting pathology and intraoperative events such as embolization, MAP may play a
role in this regional hypoperfusion and its outcome.

Inflammation

Our thinking about brain injury has moved beyond discussions of vascular obstruction
and hypoperfusion. During CPB, a variety of pathophysiologic mechanisms may impact
the vascular lining and these events may contribute to post-CPB encephalopathy.
Because the endothelium regulates vasomotor tone, thrombosis, fluid transport, and
the inflammatory response, alterations in endothelial function may be integral to
postbypass CNS integrity. Although interactions among inflammation, ischemia, and
the endothelium remain ill- defined, this is a fertile area of ongoing research.

CPB is associated with ischemia and reperfusion injury to the heart and lung as well as
a generalized inflammatory response. Ischemia and reperfusion are potent triggers for
activation of leukocytes and for leukocyte–endothelial or leukocyte–platelet–
endothelial binding (Fig. 15.3) (88,119, 120). Vascular integrity may then be impaired
either through capillary plugging (121,122) or by the liberation of free radicals,
hydrogen peroxide, and proteolytic enzymes upon leukocyte degranulation
(88,119,123). The aspiration of wound blood into the oxygenator via cardiotomy
suction amplifies the inflammatory response as this blood contains tissue factor which
is a potent activator of the complement cascade (124). Endothelial dysfunction has
been demonstrated following CPB reperfusion in pulmonary and coronary vessels
(125,126,127). CPB may also alter endothelial function in nonischemic tissues by a
variety of other mechanisms (128,129).
In addition to localized inflammation which may occur during CPB, exposure of the
blood to the bypass circuit results in activation of platelets, monocytes, and
neutrophils (103,130,131,132,133,134). The endothelium normally inhibits platelet
aggregation and leukocyte activation and binding (88,119,135,136). However, contact
activation, complement production, and the release of a variety of cytokines (137)
stimulate the endothelium to more actively bind circulating blood elements (138).
Therefore, activation of blood elements and the endothelium during CPB may
overwhelm the usual inflammatory homeostatic mechanisms. The extent of the
endothelial dysfunction as a result of CPB has been recently described in two small
studies. The first was a prospective nonrandomized trial of ten patients undergoing
aortic valve replacement in Sweden. Spinal fluid was sampled pre-, intra-, and
postoperatively and correlated with TCD and serum measurements of inflammatory
cytokines and proteins. They found marked elevations of inflammatory markers S110B
and GFAP without evidence of neuronal injury or microemboli. In addition, there was
evidence of blood–brain barrier breakdown (BBB), suggesting endothelial damage as a
result of inflammation (139). A second study enrolled 19 patients undergoing CABG or
valve replacement. Postoperative brain MRI demonstrated BBB breakdown in 50%
with a similar incidence of embolism (140).

The prospective observational cohort study performed by Mu et al. (141) identified an

association of greater inflammatory process (documented by cortisol levels) during the
perioperative period and worse cognitive function measured by a battery of seven
cognitive tests (55), 7 days after surgery. Not surprisingly, the inflammatory process
was also found to be associated with an increased incidence of postoperative delirium
and poorer outcomes by the same group in a previous study (142). Outside of surgery,
Yaffe et al. (143) studied 3,031 individuals over a 2-year period and documented the
association of inflammation, mainly interleukin-6 (IL-6) and C-reactive protein (CRP),
and worse long-term neurologic outcomes in the general population of well-
functioning elderly individuals. Again, this observation points to population-based
issues with the physiology of aging as the potential sources of poorer neurologic
outcomes in surgery.

It remains to be seen if CPB-related inflammation is sufficient to alter CNS endothelial

function in the absence of an ischemic substrate. The relative role of embolic
phenomena, localized hypoperfusion, and the influence of ischemia in other organs
also remain to be defined. The inflammatory response triggered by CPB may be
sufficient to magnify minor insults into pathologic ones (144).
Figure15.3 White blood cells (WBCs), especially neutrophils, may cause and/or
aggravate ischemic lesions by several mechanisms that may interact and amplify one
another.(Ernst E, Hammerschmidt DE, Bagge U, et al. Leukocytes and the risk of
ischemic diseases. JAMA 1987;257:2318–2324. Copyright 1987, American Medical
Association.)

CEREBRAL PHYSIOLOGY DURING CPB

For those who regularly work in cardiac surgery operating suites, it becomes easy to
take CPB for granted. However, the physiology of CPB is unique. It is only during bypass
that a mammalian species would acutely undergo up to a 20°C change in body
temperature, a rapid reduction in hematocrit (Hct), and a loss of pulsatile flow. It can
be difficult to define what is “normal” under these conditions.

Assuming a steady state of anesthesia, cerebral oxygen demand (CMRO2) during CPB
varies primarily in inverse proportion to brain temperature. Normal values at different
temperatures have been established. However, the same cannot be said of cerebral
blood flow (CBF). CBF is strongly influenced by CMRO2, PaCO2, Hct, and MAP. Each of
these variables may act independently or together to increase or decrease CBF. As
such, attempting to define “normal” CBF during bypass may be misleading or
counterproductive. Fundamentally, cerebral oxygen delivery (CDO2) must be sufficient
to support metabolic demand. Therefore, in order to determine the minimum MAP,
Hct, or pump flow required at a given temperature, it is best to examine the effect of
these variables on CDO2 at a given predicted CMRO2.
Figure15.4 Schematic representing the relationship of CMRO2 to cerebral O2 delivery
(or the determinants of CDO2).

Under nonbypass conditions, cerebral O2 delivery far exceeds cerebral O2 demand

such that CMRO2 remains independent of CDO2 over a broad range. When CDO2
delivery is progressively reduced, CMRO2 can be maintained by an increase in O2
extraction but once the capacity for oxygen extraction is exhausted, further decreases
in delivery will result in cerebral ischemia and CMRO2 will decrease. These
relationships are represented schematically in Figure 15.4.

Because CDO2 equals the product of CBF and arterial O2 content (CaO2), individual
determinants of CBF or CaO2 can serve as the abscissa in Figure 15.4. Temperature
change may shift the curve describing the CDO2–CMRO2 relationship upward or
downward, but at a stable temperature, the physiologic relationship is unaltered. The
minimum O2 delivery supporting oxygen demand is the “critical O2 delivery”
(145,146,147). Similarly, during CPB, the critical Hct (Hctcrit) or critical perfusion
pressure for brain can be defined as the minimum Hct or perfusion pressure
maintaining CMRO2 at a given temperature (148,149).

Determinants of CPB

CPB is typically associated with changes in body temperature and Hct. Significant
alterations in PaCO2 and MAP may also be experienced. While it is common for these
variables to change simultaneously, each has a predictable effect on cerebral
perfusion.

Temperature
In the context of CPB practice, temperature is the primary determinant of CBF, and it
has direct and indirect effects. Generally the brain regulates its flow in response to its
oxygen demand such that increases or decreases in CMRO2 are associated with
proportional changes in CBF. This relationship has been termed “flow-metabolism
coupling.”

The Q10, or CNS respiratory quotient, describes the increase in CMRO2 per 10°C
increase in temperature. Between 27°C and 37°C, this value is approximately 2.4 to
3.0. Therefore, a 10°C temperature reduction decreases metabolic rate more than
50%, and all other things being equal, CBF is reduced proportionately. The Q10 is
nonlinear, such that at more profound levels of hypothermia (15°C–27°C) the Q10 rises
significantly (150).

Figure15.5 Cerebral blood flow (CBF) (■) and metabolic rate for oxygen (CMRO2) (□)
(both 100 mL/g/min) versus temperature during CPB (Group I, 38°C; Group II, 28°C; and
Group III, 18°C) (n = 8 per group). Mean values for 10 additional animals at 33°C (n = 5)
and 22°C (n = 5) are also shown (CBF, %; CMRO2, +). (Cook DJ, Orszulak TA, Daly RC.
Minimum hematocrit at differing cardiopulmonary bypass temperatures in dogs.
Circulation 1998;98:II-170–II-175.)

As temperature decreases, hypothermia’s effects on CBF become more complex. In

addition to the change in metabolic rate, there is a change in blood rheologic
character, and probably in cerebral vascular responsiveness. Below approximately 22°C
to 23°C, CBF and metabolism appear to become uncoupled such that changes in CBF
do not track changes in CMRO2 (Fig. 15.5). This is of greatest relevance during
profound hypothermia in children in whom a cerebral “vasoparesis” has been
described (151,152). The physiologic or biophysical reasons for this vasoparesis remain
unknown.

Mean Arterial Pressure

Under non-CPB conditions, the healthy brain maintains CBF to an MAP of

approximately 50 to 55 mm Hg. Significant autoregulatory capacity may also be
preserved during CPB, but this depends in large part on how other CPB variables are
managed.

Since the determinants of cerebral perfusion are the same under bypass and
nonbypass conditions, one would anticipate that MAP during CPB would determine
cerebral perfusion. However, clinical studies in the 1980s deemphasized the role of
MAP, reporting that CBF is maintained with α-stat acid–base management to MAPs as
low as 20 to 35 mm Hg under hypothermic conditions (153,154,155). These results
were interpreted to mean that the cerebral autoregulatory curve was shifted leftward
during hypothermia (153,156).

4However, these clinical results indicating a leftward shift of the autoregulatory curve
must be viewed with caution. Those studies pooled measurements of MAP and CBF
from multiple patients under differing temperature, CO2, and Hct conditions. Because
of large between-patient variability in the physiologic determinants of CBF, the
resulting high variability in CBF virtually ensures that no correlation will be found
between MAP and CBF when single data points are pooled in that manner.

These limitations can be overcome in animal models, where multiple cerebral

physiologic measurements can be made at differing MAPs with extremely tight
physiologic control. Unlike in patients, the lower limits of autoregulation can also be
evaluated in animal models. In contrast to the clinical studies of the 1980s, laboratory
reports have not indicated a leftward shift of the autoregulatory curve during
hypothermic CPB (149,157,158,159).
Figure15.6 Cerebral oxygen delivery (CDO2) and CMRO2 versus mean arterial pressure
(MAP) during CPB at 33°C. Values (100 mL/g/min) are mean ± SD (*p < 0.05 vs. MAP of
60 mm Hg by repeated measure analysis of variance followed by Student–Neuman–
Keuls test).(Plöchl W, Cook DJ, Orszulak TA, et al. Critical cerebral perfusion pressure
during tepid heart surgery in dogs. Ann Thorac Surg 1998;66:118–124.)

Because a large part of CPB is now conducted with mild hypothermia (oft termed
“tepid”), Plochl et al. (149) examined critical cerebral perfusion pressure at 33°C in
dogs. The relationship between MAP, CBF, CDO2, and CMRO2 was described as
consisting of two parts, a pressure-independent portion and a pressure-dependent
portion. CBF and CDO2 were preserved at MAPs of 60 mm Hg and higher, while at 50
mm Hg or less, CBF, and more importantly CDO2, became pressure- dependent (Fig.
15.6). However, cerebral ischemia was not observed at an MAP of 50 mm Hg because
the reduction in CDO2 was compensated for by an increased cerebral O2 extraction.
While cerebral ischemia was first statistically significant at an MAP of 40 mm Hg, the
authors suggested that an MAP of 45 mm Hg is probably inadequate at 33°C, because
at 45 mm Hg some animals showed evidence of cerebral ischemia.

Laboratory studies in healthy animals indicate that relatively normal MAPs should be
maintained during CPB between 27°C and 37°C (149,157,159). Physiologically, an MAP
of at least 50 to 55 mm Hg appears desirable because at this pressure some safety
margin for the brain (increased O2 extraction) will exist because most patients must be
presumed to have cerebral vascular disease or hypertension.

Carbon Dioxide

While hypothermia increases ischemic tolerance, it also introduces unique physiologic

questions. One of the most practical has been the appropriate management of CO2.
While CO2 can be managed with pH-stat or α-stat techniques, discussions of which
approach is more “physiologic” are probably less productive than determining which
CO2 management strategy is most consistent with certain physiologic goals.

Carbon dioxide is one of the most potent determinants of CBF and most
(160,161,162,163), but not all (164), studies indicate that CO2 reactivity is preserved
during bypass. Depending on CO2 strategy, CBF may vary by more than 50%. Changes
in PaCO2 alter CBF largely independent of CMRO2, so like hemodilution, changes in
PaCO2 may alter the ratio of CBF to CMRO2 without indicating pathology.

Figure15.7 Response of cerebral blood flow (CBF) to changes in mean arterial pressure
(MAP) at three different levels of carbon dioxide tension (mean values of 33, 45, and 57
mm Hg).(Henriksen L. Brain luxury perfusion during cardiopulmonary bypass in
humans. A study of the cerebral blood flow response to changes in CO2, O2, and blood
pressure. J Cereb Blood Flow Metab 1986;6:366–378.)

Henriksen clearly demonstrated the effect of CO2 on CBF as well as the effect of PaCO2
on cerebral autoregulation during clinical CPB (160). Figure 15.7 shows that during
bypass an elevated PaCO2 (upper curve) is associated with a higher CBF for any given
MAP. Additionally, Henriksen’s data show that autoregulation is preserved with an α-
stat strategy between MAPs of 55 and 95 mm Hg (Fig. 15.7, lower curve), while CBF
becomes largely pressure passive when PaCO2 is elevated (Fig. 15.7, upper curve).
Subsequent clinical and animal studies have confirmed these findings
(162,164,165,166,167). This effect of CO2 is seen across different values of MAP,
CMRO2, and Hct.

Hematocrit

In the adult, CPB hemodilution typically reduces hemoglobin (Hgb) concentration (and
hence Hct) by a third. This reduces blood viscosity and vascular resistance and
increases CBF (148,160,168,169). This rheologically-mediated increase in CBF supports
cerebral oxygen delivery as Hct is reduced, so CMRO2 remains independent of Hct over
a broad range. However, with progressive hemodilution, CDO2 and then O2
consumption are compromised when the CBF response can no longer compensate for
the reduction in arterial oxygen content and when oxygen extraction capacity has been
exhausted. At this point, the Hctcrit is reached (Fig. 15.4) (148).

During hypothermic CPB, the increase in CBF occurring with hemodilution may be
offset by the decrease in CBF associated with the reduction in CMRO2 (168). As such,
during hypothermia, CBF may increase, decrease, or remain largely unchanged. The
net change in CBF is a function of both the magnitude of temperature change and Hct
reduction (168).

Hemodilution practice during CPB is relatively unique and this can lead to a
misunderstanding of CBF responses. Under non-CPB conditions, there is a relatively
fixed ratio (13,14,15,16,17) between CBF and CMRO2. Authors have described
increases in this ratio (uncoupling) during normothermic or moderately hypothermic
CPB (153,170,171) with a suggestion that this is pathophysiologic. While CBF–CMRO2
uncoupling may occur with profound hypothermia, a change in this ratio should be
expected with hemodilution (as with variations in PaCO2).

During normothermic CPB, the CBF-to-CMRO2 ratio is elevated; however, this elevated
ratio is not pathologic because the increase in CBF constitutes appropriate physiologic
compensation for the reduction in Hct. Accordingly, CDO2 does not change from the
prebypass state (168). For this reason, discussions of the CBF-to-CMRO2 ratio during
CPB can be misleading, whereas focusing instead on cerebral O2 delivery may provide
clarity by accounting for the effects of changing Hct.

While experience dictates that very low Hcts can be tolerated during CPB (172,173),
until recently these limits have not been defined. In a recent publication, Loor et al.
(174) prospectively evaluated 7,957 patients who were not transfused looking for
association of nadir Hgb and outcomes. There was linear, organ-dependent association
(myocardial injury, glomerular filtration ratio, ventilator support, hospital stay, and
mortality) between the degree of anemia and organ injury (Fig. 19) but no specific
cutoff at which the risk would sharply increase. Additionally, there was no evidence to
recommend a specific transfusion trigger for this patient population. Indeed, studies by
Swaminathan et al. in 2003 and a retrospective study in 2011 suggested that the
worsened renal outcomes associated with anemia were not improved by the use of
blood transfusions. This pathophysiology is complex because the severity of kidney
injury associated with blood transfusions might be even more severe in the anemic
patients being transfused when compared to the nontransfused patients (175,176).
Figure15.8 Hct–CMRO2 relationship at 38°C, 28°C, and 18°C. Forty value pairs per
group (n = 8 per group). Linear regression curves were drawn for data points above and
below a range of Hcts below which the transition from Hct independence to Hct
dependence occurs.(Cook DJ, Orszulak TA, Daly RC. Minimum hematocrit at differing
cardiopulmonary bypass temperatures in dogs. Circulation 1998;98:II-170–II-175.)

To determine temperature-dependent limits on hemodilution, our group placed dogs

on bypass at 38°C, 28°C, or 18°C and the cerebral effects of progressive hemodilution
were determined (148). We predicted that the critical Hct would be reduced in
proportion to the reduction in CMRO2; however, this was not found. Between 38°C
and 18°C, appropriate reductions in metabolic rate were demonstrated; however, the
critical Hct at 28°C was only about 3 units lower than that at 38°C. Similarly, at 18°C,
the CMRO2 was less than half that at 28°C, but the critical Hct at 18°C was only 3 to 4
units lower than that at 28°C (Fig. 15.8). Therefore, the leftward shift in the Hctcrit is
quite small and less than proportionate to the decrease in CMRO2. This occurs because
the CBF response to hemodilution attenuates with progressive hypothermia (148).

Finally, there is an impression that hemodilution increases tolerance for hypotension

because it increases organ blood flow. While hemodilution increases CBF, this is not
equivalent to saying that tolerance to hypotension is increased. While CBF may be
“normal” at a lower MAP, cerebral O2 delivery will be reduced. Following
hemodilution, autoregulatory curves can be constructed for brain, and although the
absolute levels of CBF shift upward with each level of hemodilution, the minimum
autoregulatory threshold for MAP remains about the same as in the nonhemodiluted
state (160).

CPB Flow and Brain Perfusion

To a large extent, asking whether pump flow or MAP is of greater importance during
bypass constitutes an artificial distinction. As in the intact circulation, changes in pump
flow during CPB are usually associated with changes in MAP. However, this
relationship may not be linear. At higher flows, reductions in vascular resistance may
dampen the effect of pump flow changes on MAP, while at low flows, MAP may be
reduced in direct proportion to a change in pump flow (159). A focus on the low-flow
range, where cerebral physiologic variables show pump flow dependence
(152,177,178), may lead to the conclusion that pump flow is a primary determinant of
cerebral perfusion. However, the effect of pump flow on cerebral perfusion is indirect.

These relationships were well demonstrated by Sadahiro et al. (159), who described
the effect of stepwise CPB flow reductions on MAP and CBF in dogs. In that report,
pump flow did not alter CBF until the MAP decreased below 50 mm Hg. Below that
point, pump flow, MAP, and cerebral perfusion diminished in nearly linear fashion.
Scrutiny of other reports examining the effects of pump flow on cerebral physiologic
variables show similar CNS dependency on pump flow when those reduced flows
induce MAPs below the autoregulatory threshold (152,179,180,181,182,183).

The primacy of perfusion pressure has been demonstrated under conditions where
MAP was normal (61 mm Hg) and pump flow was low (0.75 L/min/m2) as well as when
pump flow was normal (2.2 L/min/m2) and MAP was low (24 mm Hg) (180,184,185). In
that investigation, a low MAP with a normal pump flow was achieved by a large
femoral arterial to venous shunt. These studies showed that if MAP was maintained,
pump flow rate had no effect on CBF; however, reduced perfusion pressure reduced
CBF even if pump flow rate was normal. Data from both animal models
(158,182,183,186) and during clinical CPB (152,187,188,189) support those
conclusions. These results indicate that CPB flow is important to cerebral perfusion in
so much as it generates an MAP and that maintenance of CPB flow is not sufficient to
guarantee cerebral perfusion if MAP is reduced.

Pulsatility

Like the acute changes in temperature and Hct, which may occur during CPB, the loss
of pulsatile flow is a physiologic condition unique to CPB. As such, the effect of pulse
pressure on organ perfusion during CPB has been of interest. Nonpulsatile perfusion
has been reported to result in arteriolar closure (190) and a disturbance in the
coupling of blood flow and metabolism (171). Conversely, pulsatile flow has been
reported to improve CBF, microcirculatory perfusion, and tissue oxygen consumption,
and to facilitate the recovery of CBF following ischemia, low-flow CPB, and circulatory
arrest (191,192).

At least an equal number of reports have found no significant physiologic effect of

pulsatility during CPB. Hindman and colleagues found no difference in CBF or CMRO2
between pulsatile and nonpulsatile CPB in rabbits under normothermic (193) or
hypothermic (194) conditions. Sadahiro et al. (159) reported similar results in dogs as
did Cook and colleagues (186) under three temperature and CPB-flow conditions.
Conflicting reports about the effects of pulsatile CPB exist for essentially every variable
examined. Because convincing evidence of benefit is lacking and because of the
technical complexity and difficulties in generating a meaningful pulse pressure in the
systemic circulation, pulsatile CPB systems have not become a routine of practice.

CPB Duration

Two clinical reports have indicated that CBF during CPB decreases with bypass
duration (195,196). In one study, two CBF measurements were obtained 20 to 30
minutes apart during hypothermia and while nasopharyngeal temperature, PaCO2,
and MAP were stable, a 0.7% to 1%/min decrease in CBF was observed (196). A second
study from the same group concluded that CBF declines over time based on alterations
in CO2 responsiveness (195). Although decreases in CBF with time were shown in these
studies, there is evidence to suggest that these decreases in CBF may have been
related to other factors.

Hindman et al. (197) were unable to document an effect of CPB duration on CBF in a
rabbit model where brain temperature was directly measured. They speculated that
the previous clinical findings were a function of brain cooling not reflected in
nasopharyngeal temperature. A study in baboons also failed to show an effect of CPB
duration on CBF (185). The canine study by Johnston et al. (164) is particularly notable.
During CPB, blood flow to cerebral cortex and cerebellum were measured and
although some decrease was seen following rewarming, three measurements over 120
minutes of stable hypothermia demonstrated no decline in CBF under either α-stat or
pH-stat conditions. Subsequent clinical studies have also failed to show a decrease in
CBF with bypass time (168,198).

INTRAOPERATIVE MANAGEMENT AND NEUROLOGIC OUTCOME

Mean Arterial Pressure

A variety of early reports as well as a few from the 1980s (199,200) identified a
relationship between intraoperative MAP and neurologic outcome. This relationship
has not been identified in other clinical studies (5,7,51,57). However, it should be
emphasized that no investigation before 1995 was specifically designed to examine the
role of MAP in neurologic outcome. The studies lacked a control group, were often
retrospective, or were designed to examine the neurologic effects of other
interventions (51,71,201), and the effect of MAP on outcome was examined
secondarily. Additionally, neurologic follow-up was typically only short-term and the
patients were younger and had fewer comorbidities than our current surgical
population. These studies also used both α-stat and pH-stat management (71,201),
bubble and membrane oxygenators (7,201), and may not have used arterial line
filtration (7,51). Therefore, the applicability of all of these studies to our current
surgical population must be viewed with some caution.

The later multicenter study on cerebral outcomes following CABG described a 6%

incidence of neurologic deficit at discharge and reflects our current surgical population
(5). While perioperative hypotension was not identified as a risk factor for adverse
neurologic outcome, like earlier studies, patients were not randomized to differing
MAPs. To date, only two studies have done this in a prospective randomized manner
(67,202). Gold et al. (67) randomized 248 patients to either lower (50–60 mm Hg) or
higher (80–100 mm Hg) MAP during hypothermic (28°C–30°C) CPB and examined the
outcome at 6 months. Patient characteristics and CPB management were otherwise
equivalent and reflect current perfusion practice and patient risk profile. While
statistical differences were not shown for individual outcomes, the overall incidence of
cardiac and neurologic morbidity at 6 months was significantly reduced in the high
MAP group and there was a trend for every variable, except neurocognitive outcome,
to show better results in the higher MAP group (Table 15.5). The authors speculate
that their study size may have been underpowered for relatively low-frequency events.
A prospective randomized controlled trial by the same institution, performed 12 years
later, divided patients into two groups: a “high” MAP 80 mm Hg group and the
“custom” MAP group that used the patient preoperative MAP as the goal MAP during
CPB. Using this approach, there was no statistically significant difference in outcomes
between the groups, demonstrating that either approach is adequate. In pragmatic
analysis, it was noticed that in the MAP “custom” group, for patients in which the MAP
was not maintained in the target range, but 20 mm Hg below, the incidence of cardiac
and neurologic complications increased from 7.3% to 15.9%. This finding is then in
agreement with the earlier report (202).

Table15.5 Effect of MAP on neurologic cardiac outcomes in patients randomized to

lower or higher MAP during bypass The authors do not provide p values for each
variable.
Over the range of temperatures at which most adult CPB is practiced, physiologic data
indicate that cerebral oxygen delivery may become compromised at an MAP of
approximately 50 mm Hg. While physiologic measurements do not translate directly
into clinical outcomes, both the clinical report by Gold et al. (67) and the consecutive
study by Charlson (202) are highly suggestive and these studies are arguably the best
to have been conducted on the relationship between MAP and outcome. Finally, our
surgical population is older and has greater stroke risk factors than when many prior
“negative” outcome studies were done. Thus, while the data are not in agreement, it is
prudent to be conservative and assume that our patients will benefit from
maintenance of MAP in the autoregulatory range during CPB.

Temperature and Outcomes

Table15.6 Neurologic outcome and cardiopulmonary bypass temperature

The role of CPB temperature in neurologic outcome came to the forefront a decade
ago when two large, randomized outcome studies were published. The Warm Heart
Investigators (203) reported a study of 1,732 patients who underwent CPB at either
33°C to 37°C or 25°C to 30°C. In that study, the incidence of stroke at discharge was
equivalent between temperature groups (Table 15.6). However, in the same year,
Martin et al. (204) reported a trial of 1,001 patients undergoing CPB at either ≥35°C (n
= 493) or ≤28°C (n = 508). The Martin study reported a 3-fold greater incidence of
neurologic morbidity in the warm group.

Reports over the next year (205,206) identified a variety of demographic and
intraoperative management issues that may have accounted for the outcome
differences in the Toronto and Emory studies. They pointed out that the Toronto
population had fewer females, a smaller proportion of patients older than 70 years
(16% vs. 38%), and fewer reoperations. Differences in cardioplegia and aortic cross-
clamp management may also have placed the Emory patients at greater risk for
cerebral embolism (101,207,208,209). Finally, the warm group at Toronto was 2°C to
4°C cooler than the warm group at Emory (206).

Neurocognitive outcomes for part of these study populations have also been provided.
Complete testing was reported on 89 of the 1,001 patients enrolled at Emory (47).
Although CPB resulted in deficits in both temperature groups, there was no difference
in neurocognitive outcome between the warm and cold populations (47).
Neurocognitive outcomes were also reported on 153 patients from the Toronto study.
Like the Emory report, temperature management did not affect neurocognitive injury 3
months postoperatively (48,210).

A variety of other investigations have compared neurologic outcomes using different

CPB temperatures. Plourde et al. (211) randomized 62 patients to CPB at either 34°C to
35°C or 28°C CPB and did not find any neuropsychologic differences between groups at
the seventh postoperative day. Hvass et al. reported 100 patients operated on with a
warm-body (37°C)–cold-heart technique and documented a 1% incidence of stroke,
and Birdi et al. randomized 300 CABG patients to CPB at either 37°C, 32°C, or 28°C
(212,213), and found a 0% to 1% incidence of stroke with no difference among the
three groups (although the neurologic assessment was not described). Singh et al.
(214) compared 2,585 consecutive patients having CPB at 37°C to a historical cohort of
1,605 patients operated on with systemic hypothermia (25°C–30°C). In that study, the
stroke incidence was 1% and 1.3% in the warm and cold groups, respectively. Regragui
and colleagues (215) conducted neurocognitive testing on 96 patients randomized to
CPB at either 37°C, 32°C, or 28°C and found that the neurocognitive score was worst in
the 37°C group (n = 31), better in the 32°C group (n = 36), and equivalent between the
28°C (n = 29) and 32°C groups. From these results, the authors made an argument for
mild hypothermia. However, their paper was followed by a commentary that faulted
the study’s statistical power, neurocognitive testing methods, and data analysis.
Given the profound effects of temperature on cerebral oxygen demand and the
neuroprotective effect of hypothermia, it is reasonable to expect that lower CPB
temperatures would improve cognitive outcomes. However, such improvement has
not been demonstrated in randomized or nonrandomized trials (47,48,68,69). As such,
investigators have turned to examining rewarming rate and postoperative
temperature in determining cognitive outcomes. When the randomized trial of CPB
temperature did not show a difference in cognitive outcomes (68), Grocott and
colleagues (216) looked at postoperative temperature in the same patients and
published the data separately. They hypothesized that postoperative cerebral
hyperthermia was a determinant of cognitive dysfunction. In that study, cerebral
hyperthermia was defined in either of two ways: (1) the maximum temperature in the
first 24 hours; and/or (2) the area under the curve for temperature greater than 37°C.
Peak temperature was not a determinant of cognitive outcome while area under the
curve for temperature >37°C was associated “…albeit weakly, with a greater amount of
cognitive dysfunction” (216).

Data from the original randomized 2001 Grigore et al. (68) study were reported a third
time when a subset of that negative outcome trial was used to examine the effect of
rewarming speed on cognitive outcomes (82). The cognitive data were analyzed as a
continuous variable (better or worse) and as a dichotomous variable (defect present or
not). Neither univariate nor multivariate dichotomous analysis showed an effect of
rewarming speed on cognitive outcomes. However, the authors found that slow
rewarming analyzed as a continuous variable was associated with greater
improvement in cognitive performance at 6 weeks than conventional rewarming.

Although there are multiple publications interrogating the effect of perioperative

temperature management in cognitive outcomes, the weight of the evidence is far
weaker than the titles would suggest. Of two of those reporting a positive effect
(82,216), the strength of the association is tenuous, and the authors could have equally
chosen the opposite conclusion by another analysis included within those reports.
Furthermore, those positive trials consisted primarily of reporting additional study
periods (rewarming phase or postoperatively) from a larger negative randomized trial
(68). This is not to say that perioperative temperature management is unimportant,
only that the evidence for an effect of a perioperative management on cognitive
outcome is quite weak.

Glucose Management

In addition to the absolute temperature difference in the Toronto (203) and Emory
(217) neurologic outcome trials, the Emory group had significantly higher blood
glucose levels in their warm CPB group (205). This is of interest because elevations in
blood glucose aggravate neurologic ischemic injury in experimental models
(218,219,220). However, to date, no study has documented an independent effect of
glucose on neurologic outcome in clinical CPB (47,70,205,221,222,223). Although the
mean blood glucose in the “warm” Emory group may have exceeded 275 mg/dL (47), a
multivariate analysis did not identify blood glucose as a predictor of neurologic or
neurocognitive outcomes in that study (47,205). This is an area where more
investigation is needed. To date, only one prospective nonrandomized study that
included 200 consecutive patients has suggested an association of significant
hyperglycemia and increased incidence of postoperative neurologic complication
(224). One of the important recent investigations in regard to perioperative glucose
management in cardiac surgery is the randomized controlled study from the Mayo
Clinic group led by Gandhi (225). This investigation compared the tight glucose
management (90–110 mg/dL) versus conventional management (<200 mg/dL). The
intervention group had a higher incidence of deaths (4 vs. 0, p = 0.061) and strokes (8
vs. 1, p = 0.02) when compared with the group under traditional management. Despite
the tight glycemic control, this study did not identify significant hypoglycemia in the
treatment group; so this presumed etiology could not explain the worse outcomes in
the intervention group (226). Regardless of establishing an impact on neurologic
outcome, the Society of Thoracic Surgeons (STS) does provide guidance regarding
intraoperative glucose management in diabetic and nondiabetic patients. In summary,
the glucose levels should be kept below 180 mg/dL and insulin should be provided with
the use of continuous infusions since these provide better control than subcutaneous
dosing (226).

As previously noted, it appears that neurologic injury arises from a combination of

preexisting patient conditions and intraoperative factors. As with MAP, it is therefore
difficult to demonstrate that management of an isolated physiologic variable, like
blood glucose, determines patient outcome.

CO2 Management and Outcome

In the 1990s, studies from four countries examined the effects of PaCO2 management
on neurologic outcome in adults (71,72,162,227). In 1990, Bashein et al. (71)
randomized 86 patients to undergo hypothermic (30°C) CPB with either α-stat or pH-
stat management, and those authors also examined neurologic and neurocognitive
outcome at discharge and 7 months postoperatively. Neurologic impairment was
unrelated to PaCO2 at either short-term or long-term follow-up. Stephan et al. (162)
randomized 65 patients undergoing CPB at 26°C and conducted a neurologic exam on
the seventh postoperative day. The pH-stat and α-stat groups had a 29% and 7%
incidence of neurologic sequelae, respectively, and multivariate analysis identified only
PaCO2 as an etiologic factor in adverse neurologic outcome. In 1996, Patel and
colleagues (227) randomized 70 patients undergoing CPB at 28°C to pH-stat or α-stat
management and performed neuropsychological testing preoperatively and at 6 weeks
postoperatively. Using deterioration in three or more tests as the criterion for adverse
outcome, a greater proportion of patients in the pH-stat group had an adverse
outcome. In 1996, Murkin et al. (72) also reported on the neurologic effects of CO2
management in 316 patients undergoing CABG randomized according to CO2 strategy
and to pulsatile or nonpulsatile flow. Overall, the stroke rate in the two CO2 groups did
not differ (2.5%), nor did the incidence of neurologic or neurocognitive dysfunction at
7 days. Approximately half of the patients had CPB times >90 minutes, and in that
group only, cognitive dysfunction was less prevalent in α-stat (27%) than in pH-stat
patients (44%) at 2-month follow-up.

While all of these studies have limitations, there is a suggestion that pH-stat
management may be associated with a worsened neurologic outcome and that α-stat
CO2 management may be more important in populations at higher risk. The underlying
hypothesis is that reductions in CO2 should reduce cerebral embolization but this
effect may only be clinically evident when embolic risk is greater. The hypothesis
relating higher CO2 to embolization has been expressed since the early 1980s (228) but
was only recently confirmed in an animal model (229).

The pediatric literature has identified a different effect of CO2 on neurologic outcome.
The Boston Children’s group retrospectively reviewed their outcome data in children
undergoing cardiac surgery with profound hypothermia and found that the
introduction of α-stat management worsened neurologic outcome (59). They
speculated that the increased CBF associated with pH-stat management improved
cerebral cooling and neuroprotection in the pre–circulatory arrest period. Subsequent
clinical studies using jugular bulb saturation (230,231) and animal studies measuring
brain temperature intracellular pH and high-energy phosphates in animals
(232,233,234) have confirmed their hypothesis. These studies indicate that in the
context of profound hypothermia and circulatory arrest, pH-stat management may
have important metabolic advantages in addition to its facilitation of brain cooling.

From all of these results, it appears that the effect of PaCO2 on outcome is likely to be
a function of the type of cerebral stress introduced by the operation. In adults, α-stat
management is probably primarily important in reducing atheroembolic risk. In
children, who lack atherosclerosis, the primary neurologic stress is hypoperfusion; so
pH-stat management should be beneficial, at least in the setting of profound
hypothermia (systemic temperatures below 20°C). While these are reasonable working
hypotheses, the literature on the effect of CO2 management on outcomes in adults is
very scanty.

Other CPB Variables

The effects of other CPB variables, pulsatility, Hct, or flow rate on neurologic outcome
have received less attention. Two studies have failed to show an effect of pulsatility on
neurocognitive outcome (72,235). Aside from a single study in infants, the effect of
CPB Hct on neurologic outcome has only been addressed secondarily. Shaw’s study
related a drop in Hgb during the operation to worsened outcome, but this was in
relation to a vascular surgery control group which did not undergo hemodilution (8).
Short-term neurologic outcome has also been related to perioperative blood loss, but
not to CPB Hct. Finally, well-designed trials relating CPB flow rate to outcomes have
only been conducted in the pediatric CPB population. With profound hypothermia, low
flow is associated with a better outcome than circulatory arrest (60,181), but for adults
there is no evidence that CPB flow rate per se affects neurologic outcome.

INTERVENTIONS

Efforts to reduce neurologic morbidity in cardiac surgery currently proceed in multiple

directions. Broadly, these can be classified as surgical and technical changes in
practice, pharmacologic interventions, and physiologic management strategies.

Surgical and Technical

Figure15.9 Frequency of moderate or severe atherosclerosis of the ascending aorta

necessitating modifications in operative technique according to age.(Wareing TH,
Davila-Roman VG, Barzilai B, et al. Management of the severely atherosclerotic
ascending aorta during cardiac operations. A strategy for detection and treatment. J
Thorac Cardiovasc Surg 1992;103:453–462.)

A greater understanding of surgical causes of neurologic morbidity has resulted in

practice change. Recently, surgical clinical investigators show appreciation for the role
of atheroembolism in brain injury and are emphasizing techniques to reduce this
problem. Greater attention to the atherosclerotic aorta has led to the application of
epiaortic ultrasound for assessment and management of aortic cannulation (Fig. 15.9)
(17,101,236) as well as the single cross-clamp and “no touch” techniques
(237,238,239). In a recent retrospective study by the Emory group evaluating 565 total
patients undergoing coronary bypass, outcomes in patients undergoing the
revascularization on or off CPB were evaluated. It was observed that from 2002 to
2009, the use of epiaortic ultrasound increased from 45% in 2002 to 89% in 2009.
Furthermore, the use of any cross-clamp reduced from approximately 97.7% to 72.7%.
Among the on-pump cohort, the single-cross-clamp technique was associated with a
decreased risk of stroke compared with the double-clamp (cross-clamp plus partial
clamp) technique (odds ratio, 0.385; p = 0.044). The authors concluded that there has
been a change in their practice with more frequent use of epiaortic ultrasound and
more cases being performed without cross-clamp. These changes were due to
surgeons’ increased awareness of complications associated with manipulation of the
aorta (240). To date, multiple publications associate the use of epiaortic ultrasound
and lower postoperative stroke rates (241,242). An extension of the “no touch
technique,” together with the greater long-term durability of arterial grafting, has also
contributed to the use of arterial conduits for all CABG grafting (243).

Technology solutions are also being directed to reduce aortic instrumentation in

cardiac surgery. The European group of Emmert et al. (244) compared the use of
HEARTSTRING (Maquet Cardiovascular LLC, San Jose, CA) for proximal anastomosis,
with total arterial revascularization, off-pump CABG with partial cross-clamp, and on-
pump CABG patients in a total study size of 4,314 patients. In their analysis, patients
undergoing off-pump CABG using partial cross-clamp and patients undergoing the HS
approach had significantly lower incidence of stroke (0.7% vs. 2.3%; CI 95%, 0.16–0.90;
p = 0.04), and these results were similar to those of the control group that underwent
no-touch total arterial revascularization (stroke rate, 0.8%). In a separate investigation,
the use of the HS device in 1,380 CABG patients who were separated in four groups
according to the severity of atherosclerotic disease (grades I, II, III, and IV) in the
ascending aorta (diagnosed by the use of epiaortic ultrasound) was studied. The use of
the HS device reduced the predicted risk of stroke by 44%, with the theoretical benefit
less apparent in patients with grade I atherosclerotic disease compared with patients
with grade II or higher (245).

The logical extension of reducing aortic manipulation and instrumentation provided

impetus for off-bypass CABG (246,247). Smaller studies associating off-bypass CABG
with poorer overall outcomes have reduced the initial enthusiasm (248,249,250,251).
However, a recent publication from Emory University studying 12,079 CABG patients
for postoperative stroke, credits off-bypass technique with a marked decrease in
stroke rates (1.5% in patients having on-bypass surgery versus 0.6% in off-bypass
patients with a no-touch technique (p < 0.01) (252).

Aortic embolization has also been addressed by devices designed to deflect (252,253)
or trap (254,255) emboli liberated from the root. While these devices appear to serve
their intended purpose, an impact on the neurologic outcome has not been
demonstrated.
Echocardiographic assessment of ventricular “de-airing” has become a more prevalent
operating-room practice (256). Insufflation of CO2 into the chest wound has also been
advocated (257), because CO2 emboli should be rapidly absorbed without detrimental
effects. While air emboli constitute a primary source of TCD signals, it remains unclear
whether micro-air emboli influence neurologic outcome (83,106,258). CPB circuit
management such as hollow fiber oxygenators, arterial inflow “line” filters, and
minimizing transfusion of mediastinal shed blood may also impact neurologic
morbidity (58,104,259,260). The elimination of small debris and/or fat emboli is quite
challenging since the use these can pass through 20 µm in line filters and arterial filters
with smaller pores have excessive resistance (261). Some groups have developed
venous and arterial filtering systems to significantly reduce the emboli load generated
by the CPB (262). Biocompatibility of the CPB circuit constitutes another technical
factor that has the potential to impact outcomes. If the CPB surface can mimic the
endothelial glycocalyx, the generalized inflammatory response to CPB may be reduced.
Surface heparinization has been the first step; given the regulatory role of the
glycoproteins in neutrophil, platelet, and endothelial interaction (88,119,135,263),
biocompatibility may evolve into more specific glycoprotein surfaces.

Pharmacologic

In addition to surgical and technical interventions to reduce embolic risk,

pharmacologic neuroprotection represents a second line of investigation. At least
three forms of pharmacologic neuroprotection can be considered: metabolic
depressants, agents that inhibit different steps in the cellular ischemic pathway, and
antiadhesive agents.

Metabolic Suppression

The fundamental understanding of ischemia as a state of cerebral O2 delivery

insufficient for demand immediately leads to considerations of metabolic suppression
as a means of neuroprotection. Although this inference is obvious, suppressors of CNS
metabolism such as barbiturates have not become a routine part of our practice.
Animal models indicate that barbiturate therapy improves outcome in models of
incomplete ischemia, but the usefulness of barbiturate therapy for brain protection
during CPB has been difficult to demonstrate clinically.

The 1982 study of Slogoff et al. (51) randomized 204 cardiac surgical patients to a
thiopental (15 mg/kg) or a control group and evaluated neurologic outcome on the
first and fourth postoperative days. While the data suggested neuroprotection in the
thiopental group, the result was not statistically significant. A subsequent study from
the same institution randomized 182 patients undergoing open-ventricle procedures
to a control group or to receive burst suppression doses of thiopental throughout CPB
(264). As in the earlier study, patients approximated normothermia (>34°C); and a
bubble oxygenator was used with no arterial line filter. Neurologic evaluation on the
tenth postoperative day demonstrated a 7.5% incidence of neurologic defects in the
control group and 0% in the treatment group. This was the first demonstration of
barbiturate cerebral protection in humans.

To address the side effects associated with the high-dose thiopental infusion (the
requirement for greater inotropic and ventilatory support), Metz and Slogoff (265)
conducted a subsequent study where neurologic outcomes were compared in groups
receiving bolus thiopental before aortic declamping or thiopental throughout CPB.
Neurologic outcome was equivalent between groups and the authors concluded that
bolus thiopental administration without EEG monitoring offered the same
neuroprotection as the infusion technique. However, an accompanying editorial
pointed out that the lack of a control group required comparison of the bolus
thiopental group to a historical control and that several practice changes occurred
between the Nussmeier (264) and Metz studies which may have made the use of
Nussmeier’s (266) control group invalid for the later study.

Zaidan et al. (267) published a major study on barbiturate neuroprotection in 1991. In

that investigation, 300 patients undergoing CABG at 28°C with a membrane
oxygenator and an arterial line filter were randomly assigned to placebo or to
thiopental infusion sufficient to achieve an isoelectric EEG throughout CPB. A
neurologic examination was performed on the second and fifth postoperative days. As
in the previous studies, patients receiving thiopental required more inotropic support
(264), but significant neuroprotection was not provided. An accompanying editorial
speculated on reasons why the Nussmeier and Zaidan results differed (268).

The primary differences between the two studies were CPB temperature (28°C vs.
34°C), closed versus open ventricles, and the presence or absence of a membrane
oxygenator and arterial line filter. It is difficult to determine the relative importance of
these factors. Open-ventricle procedures are usually associated with poorer outcomes
and more air embolism, but the relationship between air embolism and
neurophysiologic or neurologic outcome has not been clearly defined (269,270,271).
Additionally, if atheroembolism is the primary cause of neurologic injury in cardiac
surgery, then the use of arterial filters and membrane oxygenators would not be
expected to explain the study differences. Whether CPB temperature differences are
sufficient to explain the outcome differences is also not discernible. Regardless, the
results of this combination of studies led the editors to conclude that “routine
thiopental therapy has no place in the management of patients undergoing CABG.”
Furthermore, they note that “demonstration of barbiturate protection in
normothermic, unfiltered, bubble-oxygenator bypass is not a reasonable argument for
barbiturate use during alternative bypass circumstances unless the appropriate trials
are performed” (268).
The introduction of propofol led to some interest in its applicability for
neuroprotection in cardiac surgery, but studies similar to Nussmeier’s or Zaidan’s have
not been performed. The effect of burst suppression doses of propofol on CBF and
CMRo2 was evaluated and the authors speculated that propofol may have a role in
reducing cerebral embolism during CPB via its reduction in CBF (272). Other
investigators have evaluated propofol’s effect on CBF velocity (273) or whether burst
suppression doses can ameliorate cerebral venous O2 desaturation during rewarming
(274). Given the body of work on thiopental in cardiac surgery, it is unlikely that large
outcome studies testing the effects of other cerebral metabolic depressants will be
conducted in the near future.

Agents That Inhibit Ischemic Pathways

Calcium Antagonists

Of a variety of neuroprotectant agents showing efficacy in animal models, only a few

have found their way into clinical trials in cardiac surgery. Intracellular accumulation of
calcium is one of the key factors leading to cell death in cerebral ischemia. In
nonbypass models, calcium channel blockers limit ischemic injury (275,276). These
agents have also been shown to improve neurologic outcome in human stroke trials
(277). In cardiac surgery, the effect of nimodipine on neurologic outcome was assessed
in a small trial by Forsman et al. (278). Thirty-nine patients undergoing cardiac surgery
were randomized to receive nimodipine or placebo and neurocognitive outcome was
determined at 6 months. Six of 28 patients (21%) showed deficits at 6 months.
Nimodipine-treated patients were described as having a slightly better outcome in
verbal fluency and visual retention, but the authors emphasize that study size limited
statistical power and prevented any definitive conclusions from being drawn (278).

A critical trial of calcium channel blockers in cardiac surgery was reported in 1996 (73).
Nimodipine neuroprotection was to be tested in a double-blind, randomized trial of
400 patients undergoing valve replacement, but the trial was interrupted at 150
patients because of greater morbidity and mortality in the nimodipine treatment
group. At termination, there was a 10.7% incidence of death in the nimodipine group
and a 1.3% incidence in the control group. Major bleeding occurred in 13.3% of
nimodipine patients versus 4.1% of control patients. Additionally, there were no
differences in neurocognitive outcomes between the two groups at 1 week, 1 month,
or 6 months.

The authors attributed the morbidity and mortality with nimodipine treatment to
bleeding complications and speculate that this might have resulted from a
combination of vasodilatation and the antiplatelet effects of the drug (73). The authors
do not extend their observations to other calcium antagonists, but based on these
results it will be difficult to justify another large trial of calcium channel blockers as
neuroprotectants in cardiac surgery.

Pegorgotein and Other Experimental Ischemic Pathway Inhibitors

More recently, the antioxidant pegorgotein was investigated in a randomized blinded

trial of the effect of the drug on neuropsychological and other outcomes (74). A
placebo group was compared to two pegorgotein doses in a total population of 67
patients. While the study size was small, there were no strokes and no differences in
cognitive outcomes between groups. The same research group also tested the
potential neuroprotectant effects of clomethiazole on CABG patients in another
randomized trial (75). Patients were randomized to either clomethiazole or placebo
and cognitive outcome was assessed at 4 to 7 weeks postoperatively. The study size
was larger than the pegorgotein trial, with testing completed in 219 patients (110 drug
and 109 controls); however, differences in cognitive outcomes could not be
demonstrated. Finally, a study designed to assess the potential neuroprotective effect
of S(+)-ketamine in 106 patients also failed to show any effect of a perioperative drug
treatment on cognitive outcomes (279).

A variety of other pharmacologic agents having the potential to inhibit ischemic

pathways have not reached clinical trials in cardiac surgery. These agents include
aminosteroid compounds (280), excitatory amino acid receptor antagonists (281), and
agents directed at the nitric oxide pathway (282,283,284).

Leukocyte Inhibition and Endothelial Protection: Antiadhesion Therapies

Investigations from the last decade show that multiple inflammatory mediators
contribute to the progression of ischemic injury (87,120,285). In particular, neutrophils
aggregate in the capillary beds of ischemic tissues and may contribute to plugging,
endothelial dysfunction, and tissue damage. Characterization of the receptors
responsible for neutrophil–endothelium binding is leading to the development of
specific monoclonal antibodies against these receptors as well as nonspecific inhibitors
of neutrophil–endothelium binding (286,287,288).

Monoclonal antibodies to the adhesion molecules mediating the polymorphonuclear–

endothelial binding have shown promise in laboratory models of transient focal
ischemia (86,289,290). Additionally, oligosaccharides and oligopeptides structurally
related to endothelial binding sites competitively inhibit neutrophil binding to the
endothelium (263,291,292,293,294). Application of these oligosaccharides can reduce
lung injury and myocardial reperfusion injury (292,294), and block lymphocyte
adhesion to cardiac endothelium in animal models (293). More specific glycoproteins
representing the selectin binding sites for leukocytes have also been shown to
decrease ischemic injury and increase CBF after transient focal ischemia (291). These
observations are interesting because the inhibitory oligosaccharide fragments are
structurally related to, and can be generated from, multiple fractionations of heparin
(295,296,297,298) and there is some evidence that heparin itself may reduce cerebral
ischemic injury by acting as an antiadhesion molecule (299). So far no randomized
controlled trials have shown relevant clinical benefit from pharmacologic interventions
in regard to the neurologic outcome. The perioperative neurologic insult is generally
multifactorial and interventions touching only one mechanism not surprisingly have so
far been proven to be ineffective (300).

Physiologic Interventions

Temperature

Hypothermia provides flexibility in surgical and perfusion practice. Cerebral

hypothermia attenuates the physiologic impact of reductions in perfusion pressure and
Hct, and extends the “safe” period of low-flow CPB and circulatory arrest. Hypothermia
reduces cerebral injury in both regional and global models of cerebral ischemia;
however, the magnitude of neuroprotection is not directly related to the reduction in
cerebral O2 demand (301,302,303).

There is a convincing body of evidence in the experimental stroke literature that even
small temperature differences have important effects on neurochemical,
neuropathologic, and neurophysiologic outcomes after ischemia (301,303,304,305). As
little as 2°C of hypothermia significantly attenuates brain injury (301,302). These
effects are probably related to reduction in the cascade of injury precipitated by an
ischemic insult and so are independent of the effect of temperature on metabolism
(301,302). Mild hypothermia attenuates the depletion of cerebral ATP following
ischemia and decreases the production of the excitatory neurotransmitter glutamate
(303,306); infarct volume is reduced, particularly in the neocortex, and neurologic
outcome may be improved (301,302,303).

The effect of small temperature differences on the EEG, cerebral metabolism,

excitatory amino acids (EEA), and cellular high-energy phosphates was recently
evaluated in a swine CPB model (306). When pigs were subjected to 20 minutes of
global ischemia at 37°C, 34°C, 31°C, or 28°C, Conroy and colleagues (306) found that
hypothermia to 28°C and 31°C facilitated recovery from ischemia but that cooling to
28°C did not provide greater protection than cooling to 31°C. At 34°C metabolic and
EEG recovery, EEA release and S100 (a sensitive marker of cerebral injury) levels were
intermediate between that observed at the colder temperatures and at 37°C.

In this context, the avoidance of cerebral hyperthermia deserves comment (307). Just
as 2°C to 3°C of hypothermia may offer significant brain protection in ischemia models,
2°C of hyperthermia significantly worsens outcome. Minamisawa et al. (301)
demonstrated the effect of mild hyperthermia on neuronal necrosis following 10
minutes of ischemia in a rat model. As ischemic temperature was increased from 35°C
to 39°C, the percentage of neurons damaged increased from approximately 15% to
80% (Fig. 15.10). This is clinically relevant, because cerebral temperatures >39°C have
been documented in patients during rewarming (308) and these high temperatures
may occur when cerebral embolic risk is greatest. From these reports, rewarming
practice is being changed to avoid high brain temperatures. Additionally, when CPB is
to be conducted “warm,” systemic temperatures have shifted toward mild
hypothermia (33°C–34°C) from strict normothermia (309).

Figure15.10 Influence of temperature on ischemia-induced neuronal necrosis in the

hippocampus (CAI sector and subiculum). Quantitative calculation of percentage of
damaged neurons was performed bilaterally in each animal. There was no statistically
significant difference between right and left hemispheres. Values are ±SE.(Modified
from Minamiasawa H, Smith ML, Siesjo BK. The effect of mild hyperthermia and
hypothermia on brain damage following 5, 10, and 15 minutes of forebrain ischemia.
Ann Neurol 1990;28(1):26–33.)

It has been suggested that the risk of cerebral hypothermia might be avoided if
patients are weaned from bypass with mild hypothermia in the core compartment.
Additionally, one could speculate that this might offer some cerebral protection in the
early postoperative period. This approach to rewarming has been evaluated in two
trials (81,310), and feasibility has been demonstrated. In an investigation of 13
patients, six served as controls and seven were rewarmed to a nasopharyngeal (NP)
temperature of 35°C during CPB, and then warmed further by a surface device for the
next 4 hours. Peak jugular bulb temperatures during CPB rewarming were lower in the
surface warming group although these values were equivalent in the two groups 4
hours after surgery (310). The second investigation was also randomized and larger
(81). In the trial by Nathan and colleagues (81), 223 patients were randomized to
undergo CPB with rewarming either to 37°C or 34°C NP temperature. The primary
endpoint of the trial was cognitive outcomes at 1 week, with ICU variables such as
chest tube output, intubation time, myocardial infarction, and infection as secondary
outcomes. At 1 week, 62% of the control patients had cognitive decline versus 48% in
the hypothermic group. ICU outcomes such as bleeding, myocardial infarction, and
intubation time did not differ between groups.

Although incomplete rewarming in the operating room is feasible, it may be

undesirable for many adult cardiac surgical patients and it is still not clear if the
theoretical benefit is offset by potential disadvantages. The problem with terminating
bypass with mild core hypothermia is that body temperature is likely to continue to
decrease postoperatively. The magnitude of this afterdrop can be related to the
temperature at the termination of bypass (311). If bypass is terminated with an NP or
venous return temperature of 37°C, core temperature is often near 35°C on leaving the
operating room. This occurs, in part, because bypass rewarming is not uniform. Blood
flow to muscle and viscera may be reduced by 50% to 80% during bypass (312);
therefore, temperature in many skeletal muscles may still be 32°C when NP or cardiac
temperature is 37°C at the end of CPB. As such, heat is shifted from the core to the
periphery after weaning from CPB and core temperature then drops. Additionally, after
bypass, the patient continues to lose heat to the environment. One would predict that
most patients weaning from CPB with a core temperature of 35°C will have an
unacceptably large whole-body thermal debt in the ICU. In noncardiac surgery,
postoperative hypothermia may contribute to reduced cardiac output and increased
SVR, a greater incidence of ischemia, increased respiratory demands, and more
bleeding (313,314). While the study by Nathan et al. (81) begins to address concerns
about postoperative hypothermia in cardiac surgical patients, the study was
underpowered to assess outcomes such as myocardial infarction; additionally, the
wide range in postoperative bleeding makes comparison difficult. Finally, hypothermia
is not compatible with early extubation. In the Nathan report, the mean intubation
time was about 17 hours; so this report does not allow us to assess the effect of
postoperative hypothermia on current practice.

Overall, the evidence indicates that tepid CPB (nasopharyngeal 33°C–35°C) (as
opposed to strict normothermia) in a relatively healthy patient population does not
appear to significantly increase neurologic risk. Awareness of the importance of
temperature management has moved broadly through the anesthesia and cardiac
surgical community such that the practice has matured and hyperthermia is probably
much less common now (315) than when cerebral hyperthermia was first reported
(308).

Physiologic Interventions to Reduce Embolization: CO2, Temperature, and CPB Flow

As detailed above, surgical attention to embolic risk and technical changes in the CPB
circuit decrease cerebral embolization. Physiologic interventions may also be relevant.
It has been suggested that increases in CBF as occur with pH-stat management or
normothermic CPB may result in a higher incidence of cerebral embolization. Similarly,
acute reductions in CBF during periods of embolic risk might reduce cerebral
embolization.

Figure15.11 The regional distribution of cerebral emboli in pigs with a mean PaCO2 of
52 mm Hg (■) or 27 mm Hg (□). Values are the mean ± SD emboli per gram (n = 10 in
each group). *p < 0.05.(Modified from Plöchl W, Cook, DJ. Quantification and
distribution of cerebral emboli during cardiopulmonary bypass in the swine: the impact
of PaCO2. Anesthesiology 1999;90(1):183–190.)

This was tested in a swine CPB model. The effect of PaCO2 on total and regional
cerebral embolization of microspheres was determined during normothermic CPB
(229). The study demonstrated that a 25-mm Hg reduction in PaCO2 during a 5-minute
period of embolic risk could reduce total cerebral embolization by more than 50% (Fig.
15.11). These results indicate that during periods of embolic risk, such as clamping and
unclamping of the aorta and the initial phases of ventricular ejection, a moderate
reduction in PaCO2 may significantly decrease cerebral embolization.

The same model has been used to examine the effect of temperature and bypass flow
on cerebral embolization. The investigators found that 10°C of hypothermia is effective
in reducing embolization but that the combination of hypothermia and hypocarbia did
not reduce cerebral embolization more than hypocarbia alone. The relationship
between CPB flow rate and cerebral embolization has also been examined (316). At a
stable MAP, an inverse relationship between cerebral embolization and pump flow was
described in dogs. Essentially, at lower pump flows, a greater proportion of that flow is
delivered to the brain. Therefore, if a fixed number of emboli enter the aortic root,
more emboli will be delivered to the brain if CPB flow is lower. The authors suggest
that elevating pump flow during periods of embolic risk may also reduce cerebral
embolization during clinical CPB.

Hypocarbia, hypothermia, and a high flow rate are all effective in reducing cerebral
embolization in an animal model. A combination of these physiologic interventions
during periods of embolic risk may reduce neurologic morbidity. A clinical outcome
study applying these physiologic interventions has yet to be conducted.

PaCO2 and Facilitated Cooling

While there is limited clinical evidence that pH-stat management may be associated
with worsened neurologic outcomes in adults, the opposite may be true in certain
pediatric populations. A series of reports from the Boston Children’s Hospital suggest
that elevation of PaCO2 during cooling may improve neurologic outcome in children
undergoing circulatory arrest (59,61). Elevation of CBF during the cooling phase may
result in improved cerebral protection (59,61). Second, the rightward shift of the
oxyhemoglobin dissociation curve with the more acidotic strategy might facilitate O2
transfer from Hgb. These hypotheses have largely been confirmed in a series of
laboratory and clinical studies (61,232,233,234). Given these results, elevating PaCO2
to achieve pH-stat physiology may serve as a useful physiologic intervention to
improve neuroprotection in select patient populations.

NEUROLOGIC MONITORING

Given the unusual physiologic conditions of CPB and incidence of neurologic injury,
there have been persistent efforts to assess the adequacy of cerebral O2 delivery
during CPB. Some current techniques include (1) measurement of venous
oxyhemoglobin saturation at the jugular bulb (SjvO2), (2) near-infrared optical
spectroscopy (NIRS), (3) TCD, and (4) electrophysiologic monitors, for example, EEG
and evoked potentials.

Sjvo2

Measurement of cerebral venous oxygen saturation offers clinical appeal. Under

nonbypass conditions, the SjvO2 provides an index of the adequacy of global cerebral
oxygenation and normal values are established (317,318). Additionally, fiberoptic
oximetry allows for continuous measurement, placement of a jugular bulb catheter is
relatively simple, and the measurement is immediately familiar. Nevertheless, SjvO2
monitoring in cardiac surgery has remained primarily a research tool. While SjvO2 can
provide useful trends in cerebral oxygenation, the technique is limited by the fact that
it is a measure of global oxygenation, so focal events may be undetected. SjvO2 is also
potentially difficult to interpret with progressive hypothermia where changes in the
P50 become important.

One of the first reports of SjvO2 during clinical CPB was by Croughwell et al. (319), who
documented SjvO2 < 50% or a PvO2 < 25 mm Hg in 23% of 133 patients. In the same
year, Nakajima and colleagues (320) reported continuous recording of SjvO2 in 12
patients and described an inverse relationship between CPB temperature and SjvO2 as
well as a relationship between desaturation and rewarming speed. The effect of CPB
temperature on SjvO2 was further examined in a randomized comparison of
normothermic and hypothermic CPB (321).

Perhaps the most prominent work on SjvO2 was that identifying a relationship
between SjvO2 and post-CPB cognitive dysfunction (322). In that report, intraoperative
SjvO2 and neurocognitive outcome at discharge were evaluated in 255 patients.
Seventeen percent of patients showed desaturation during CPB while 38%
demonstrated neurocognitive impairment at discharge. This neurocognitive
impairment was related to the patients’ baseline scores, their educational level, and
the SjvO2 during CPB.

Studies have continued to examine the determinants of cerebral venous desaturation

or have used SjvO2 as a monitor of cerebral O2 balance. Enomoto and colleagues (323)
reported that cerebral venous desaturation during rapid rewarming was a function of
CMRO2 increasing more quickly than CBF. Sapire and colleagues (324) examined SjvO2
during rewarming and found it to be a function of Hct, rewarming speed, and MAP.
Grubhofer et al. (325) also identified a dependency of SjvO2 on MAP during CPB. von
Knobelsdorff et al. (326) could not identify a relationship between rewarming speed
and SjvO2, but in that study, patients were rewarmed from 27°C to 36°C in either 7 or
15 minutes.

Dexter and Hindman (327) published a provocative theoretical analysis of SjvO2 during
hypothermia. They pointed out that the increase in Hgb O2 affinity with progressive
hypothermia necessitates that what constitutes a minimally acceptable SjvO2 must
increase as temperature decreases. The most interesting physiologic question posed
by this discussion is whether the high O2 affinity at low temperatures is sufficient to
compromise O2 delivery to tissues. While this has not been answered, another
technology may offer insight into this question.

Since the earliest reports of SjvO2 in cardiac surgery more than a decade ago, SjvO2
has not found its way significantly into clinical practice and has primarily been an
investigational tool. In the last few years, SjvO2 has been used to evaluate changes in
rewarming practice (310), anesthetic choice (328), off-pump CABG (329), and in PaCO2
management (330,331,332). The CO2 studies are of some note because of the
consistency of findings from multiple institutions. Ali and colleagues (330) found that
the effect of CO2 management had a greater effect on SjvO2 than patient
temperature. Kiziltan and colleagues (332) reported more favorable cerebral O2
balance with a higher PaCO2 during bypass and (331) suggested that low SjvO2 values
in the early postoperative period may be related to hypocarbia-related
vasoconstriction.

Near-Infrared Spectroscopy

Hgb undergoes a characteristic near-infrared absorption shift with O2 binding, so NIRS

has the potential to provide a continuous, noninvasive, transcutaneous assessment of
regional brain oxygenation (333,334). The Hgb emission spectra is a function of the
aggregate of arterial, venous, and capillary blood, so the absence of a clear correlation
between NIRS output and either arterial or venous oxyhemoglobin saturation during
CPB is not surprising (333,335,336). However, trends provided by NIRS are of interest.
Various reports show that NIRS is sensitive to changes in temperature, PaCO2, and Hct
as well as the cessation and reestablishment of CPB flow (Fig. 15.12)
(335,337,338,339). Additionally, the rate of NIRS desaturation during circulatory arrest
has been reported to be a function of temperature at arrest (slower at colder
temperatures) (339) as well as patient age, with the youngest patients being most
tolerant of global ischemia (338).

Figure15.12 Near-infrared spectroscopy (NIRS) record of the changes in

cerebrovascular hemoglobin oxygen saturation (+ScO2) illustrating the points at which
+ScO2 was captured and the deoxygenation curve during circulatory arrest. The patient
was 9 months old and was undergoing a hemi-Fontan operation. B, baseline at
normothermia; 1, on deep hypothermic cardiopulmonary bypass (cCPB); 2, at the end
of circulatory arrest (arrest); 3, on recirculation 3 min after resuming CPB (rCPB); 4,
normothermic CPB (wCPB); 5, after CPB (end). By definition, +ScO2 = 0 at baseline.
(Kurth CD, Steven JM, Nicolson SC. Cerebral oxygenation during pediatric cardiac
surgery using deep hypothermic circulatory arrest. Anesthesiology 1995;82:74–82.)

Other studies have reported that during hypothermia, NIRS tissue saturation and SjvO2
may move in opposite directions (335,336,340). This might be interpreted as impaired
O2 offloading by Hgb, but at present the technology is not sufficiently advanced to
determine if a low P50 could result in cerebral hypoxia.

While NIRS technology has great potential, its development is ongoing. Currently, (1)
quantitative measurements of HbO2 are not possible (334); (2) the region of
interrogation is not clearly defined; (3) there is no external standard against which
NIRS can be calibrated; and (4) scalp blood and skull may contaminate the output of
the devises, particularly in adults (341); and perhaps most importantly, (5) Hbg
saturation in the brain—whether venous, arterial, or capillary—may not reflect tissue
O2 utilization because of the high Hgb O2 affinity. Measurement of the redox state of
intracellular cytochrome oxidase (aa3) may solve the latter problem because the aa3
signal may be intimately related to intracellular high-energy phosphate concentration
(342,343,344). However, at present, the analysis of the aa3 signal remains highly
complex.

In the last several years, the most valuable report on the use of NIRS in adult cardiac
surgery was the meta-analysis reported by Taillefer and Denault (345). A review of
NIRS in adult cardiac surgery yielded 48 papers describing almost 6,000 patients. Of
the 48 reports, only 1 was a randomized controlled trial, no study reached level I
evidence, and only 1 study was level II. Seventy-five percent of studies were
determined to offer level V evidence. Studies included patients primarily undergoing
CABG surgery, but valve surgery, congenital, off-pump, and arch procedures were also
captured. The authors concluded that “…NIRS validity has not been clearly established
clinically along with its predictive value” (345). In addition to multiple problems in
study design in NIRS clinical reports, the authors outline 10 clinically relevant technical
limitations in the technology. Despite all the mentioned limitations, a randomized
controlled trial conducted by the Canadian group of researchers led by Dr. Murkin et
al. (346) demonstrated that the use of NIRS was associated with less cerebral
desaturation during CPB and the group of patients being monitored had lower
incidence of postoperative organ dysfunction, shorter ICU stay, and better survival
when compared to their control group.

No similar assessment of NIRS use has been performed in pediatric cardiac surgery,
although in small children and infants some of the technical limitations seen with NIRS
in adults are probably lessened. In particular, the presence of open fontanelles, thinner
skulls, and lesser extracerebral mass may improve its reliability. Although studies using
NIRS in pediatric heart surgery have been published (347,348), an impact of this
technology on clinical outcomes has not been demonstrated.

Transcranial Doppler

Under non-CPB conditions, TCD measurements of blood flow velocity in the middle
cerebral artery may show good correlation with measured CBF (349). The technique is
noninvasive and provides continuous measurements; so TCD has been evaluated as a
monitor of cerebral perfusion during CPB. Blood flow velocity may be sensitive to
temperature change, MAP, and pump flow as well as to PaCO2 and Hct, but for flow
velocity to be a reliable measure of CBF, the diameter of the insolated vessel must not
change (350). This condition may not be met during CPB, so the correlation of TCD flow
velocity and measured CBF is relatively poor (351,352,353). Nevertheless, TCD can
provide trending information, and the relative changes in flow velocity show a better
correlation with CBF than flow velocity measurements (352,354,355).

Although strict quantitative CBF measurements are not possible, TCD may have greater
application in pediatric CPB. Obtaining the temporal window is easier in infants and
children and this population may be subjected to profound reductions in CPB flows at
stable temperatures and Hcts. Under these conditions, TCD monitoring may be useful
to determine whether these reduced CPB flows and MAPs are sufficient to maintain
cerebral perfusion (356). The response of the cerebral circulation when flow is
reestablished following circulatory arrest can also be evaluated (357).

In adult cardiac surgery, TCD has found much greater use in emboli detection than
assessment of cerebral perfusion. Although not a cerebral monitor per se, TCD emboli
detection can be used to indirectly evaluate surgical technique (89,92) and CPB circuit
modifications (94,104,260). TCD could also serve as an independent predictor of
neurologic outcome (58,90,209), although TCD emboli counts have not translated into
consistent differences in embolization detected by neuroimaging (82,358,359) or with
cognitive outcomes (258,360).

While becoming more popular, the limits of emboli detection must also be considered
(361,362). The resolution for embolus size has a limit which depends on the physical
character of the embolic material. Microthrombi are difficult to detect against the
blood background signal, while air or plaque may provide an adequate signal even if
significantly smaller. In this regard, the inability to determine embolus composition
constitutes another limit of the current technology. Finally, the “gating” used to
separate signals from background will, in large part, determine the results. However,
automated signal detection and a technical consensus on microembolus detection
(363) will help limit this source of variability.

EEG/Evoked Potentials
In theory, electrophysiologic monitors should provide one of the best means for
determining adequacy of cerebral oxygenation. Although familiarity with
intraoperative EEG deriving from carotid endarterectomy experience dates back to the
1970s and its use has been periodically advocated for use during CPB, EEG monitoring
has never established a strong presence in that setting. This remains true even though
the technology has become simpler to use and advancements have been made in data
acquisition, processing, and display.

Over the last 40 years, numerous studies have used EEG during cardiac surgery and
certain EEG changes, such as slowing at the onset of CPB, have been demonstrated
with consistency (364,365,366). Nevertheless, a clear relationship between the EEG
output and intraoperative physiology or clinical outcome has been elusive. Most
studies advocating the use of EEG have been descriptive and have typically lacked a
control group. The available evidence indicates that intraoperative EEG lacks both
sensitivity and specificity. Patients may show an abnormal EEG during CPB and recover
without neurologic deficit while others may have a normal intraoperative EEG
examination and experience stroke. This is not surprising, as advances in data
processing cannot address the primary limitation of EEG. EEG only records superficial
cortical electrical activity. If most neurologic injuries are embolic in origin, the EEG
cannot be expected to detect focal ischemia occurring in small areas deep in the brain
(367). Secondarily, the background anesthetic and hypothermia may alter both the
power and frequency spectra of the EEG, as can CPB-related artifact (366).

In the study by Bashein and colleagues (368), 78 patients underwent CPB at 28°C to
32°C with EEG recording; neurocognitive status was determined preoperatively and at
8 days and 7 months postoperatively. Electrical noise contaminated the EEG in 40% of
patients in spite of extensive computational modeling, and there was no clear
relationship between EEG power or frequency and outcome (368).

Evoked potentials provide another means of assessing functional neurologic integrity

during CPB (369,370). Somatosensory evoked potentials (SSEP), in which a stimulus is
delivered peripherally and the integrity of its transmission from the peripheral nerve
through the spinal cord and to the sensory cortex is recorded, have been most
commonly used (371). A decrease in signal amplitude may represent ischemia, as can
an increase in signal latency. In the experimental setting, evoked potentials have been
successfully used as monitors for ischemia (178), although their use in the operating
room has been limited (372,373). Like EEG, evoked-potential amplitude and latency
are sensitive to hypothermia (374,375). Proper signals may also be difficult to obtain,
and hundreds of potentials must be averaged to separate signal from noise (369,376).
Finally, the definition of what constitutes an abnormality may not be agreed upon. In
clear cases of brachial plexus injury, a 3-standard-deviation change from the pre-CPB
measurement was required to identify neurologic injury (372). As such, the sensitivity
and specificity of evoked-potential monitoring may be at least as limited as EEG
monitoring.

Bispectral EEG analysis (BIS) has been the subject of some recent trials in cardiac
surgery, and generally the output of the BIS monitor behaves as one would predict for
a processed EEG (377,378). Patient temperature and anesthetic agents affect BIS (377).
However, it seems highly unlikely that ease of use will eliminate the inherent
limitations of EEG in preventing intraoperative neurologic injury.

S100

Although not a technique for cerebral monitoring, the astroglial and Schwann cell
protein S100 is of interest because serum and CSF levels of S100 increase following
brain damage from stroke, trauma, and subarachnoid hemorrhage (379). Therefore,
the relationship between S100 and CPB-related cerebral injury has been evaluated
(380). Observational studies describing the time course of S100 release have shown
that the protein is not detected prior to CPB and that peak levels occur
intraoperatively between the end of rewarming and the end of CPB (358,
381,382,383). These levels appear to be related to CPB time (359,382) and to patient
age (359,384). Intracardiac operations result in higher S100 levels than CABG
procedures (385). Arterial line filtration may reduce S100 levels (381), and there
appears to be some correlation between embolization and S100 levels (384). Patients
undergoing off-pump CABG have undetectable or fractionally raised levels (382).

The sensitivity or specificity of S100 as a marker for cerebral injury has not been
established, although it is clear that levels measured at the end of CPB are not specific
(Table 15.7). Unfortunately, some extracranial tissues can either release S100 or non-
S100 substances that are detected by current S100 assays, such that measured S100
values may not reflect cerebral events (386,387,388). A recent study using mass
spectroscopy, gel electrophoresis, and Western blot analysis demonstrated that a
variety of non-S100 proteins in the surgical field react with the S100 commercial
immunoassay (389). This leads to problems with test sensitivity and specificity. In a
study by Jönsson et al. (359), 93% of 515 consecutive patients had detectable S100 at
the end of surgery but less than 10% showed either stroke, encephalopathy, or
delayed awakening. In a study of 40 patients, 58% of patients showed elevated S100
levels 1 hour postoperatively but no patient demonstrated overt cerebral injury (381).
High levels of S100, 48 hours postoperatively, have more sensitivity and specificity for
cerebral injury than early postoperative measurements (359,
384,385,386,387,388,389,390), as S100 values that are very high 48 hours after surgery
reflect size and predict outcome of stroke (391,392). However, what therapeutic
intervention one might base on a 48-hour measurement remains unclear.
Table15.7 Impact of recorded variables on S100 protein release at the different
sampling intervalsaaMultiple linear regression analysis showing the variables
significantly influencing the S100 protein levels after cardiac operation with
extracorporeal circulation in the 515 patients studied. The presence of a diseased
ascending aorta was confirmed with palpation intraoperatively. Renal insufficiency was
defined as a preoperative serum creatinine level >2.2 mg/dL. Cerebrovascular
accident/transient ischemic attack (CVA/TIA) relates to a history of stroke or transient
ischemic attack. Age and perfusion time were continuous variables, whereas the other
variables were analyzed as dichotomous variables. All correlations found were
positive.Jönsson H, Johnsson P, Alling C, et al. Failure of intraoperative jugular bulb S-
100B and neuron-specific enolase sampling to predict cognitive injury after carotid
endarterectomy. Ann Thorac Surg 1998;65:1639–1644.

General Approach to the Patient That Developed Acute Stroke

Despite the best surgical technique, patient selection, preparation, and perioperative
management, unfortunately some patients will still develop acute stroke in the
postoperative period. For acute-stroke care, American Heart Association/American
Stroke Association society guidelines are available to provide organized and
coordinated care (393). For nonsurgical patients, the early administration of
intravenous thrombolytic r-tPA is recommended within a 4.5-hour window of initial
symptoms (393,394). Because intravenous thrombolysis is contraindicated in patients
with recent surgery (393), techniques that provide localized intra-arterial thrombolysis
or even mechanical clot removal can be options for postoperative patients (395,396).
These techniques also have a more generous treatment window (as far as 6–8 hours
after initial symptoms). Having a well-structured and coordinated stroke management
team is pivotal for the success (393) and this should be evaluated for every cardiac
surgery program.

CONCLUSIONS

In adults, postcardiac surgical brain injury can occur as a result of a variety of

neurologic risk factors. Surgery precipitates those risks and imposes additional
ischemic, physiologic, and inflammatory stresses for which the patient may be unable
to compensate.

Because neurologic morbidity derives from interaction of the patient and the surgery,
the greatest reductions in morbidity may come from the application of risk
stratification, technical surgical maneuvers, and the use of modern and appropriate
perfusion equipment. Other innovations such as off-bypass surgery, antiadhesion
therapies, and near-infrared monitoring are supportive and complementary to prevent
stroke or significant neurocognitive functional decline. For the patients who do go to
surgery, we must be comprehensive and practical in developing an intraoperative
management strategy. Because the risk for neurologic injury derives largely from
preexisting patient risk factors, multivariate analysis indicates that age and
atherosclerotic disease determine outcome.

While patient factors may determine outcome in multivariate analysis, this is not
equivalent to saying surgical, physiologic, and pharmacologic management are
unimportant. We have a great deal of understanding about hypertensive
cerebrovascular disease, regional blood flow in focal ischemia, and the potent effect of
temperature on the cascade of injury. This allows us to make well-informed choices
about physiologic management. Similarly, gas physiology can be applied to facilitate
absorption of bubbles or change the distribution of CBF during periods of embolic risk.
Echocardiography can tell us when it is appropriate to not clamp the aorta, remove an
aortic or left ventricular vent, and administer a barbiturate if circulatory arrest must
occur before adequate cooling.
Our understanding of “postcardiac surgical cognitive decline” is evolving. First, the best
available evidence suggests that chronic aortic and brachiocephalic atherosclerosis is a
primary determinant of the neurologic outcome of on- and off-pump cardiac surgery
and must guide the manipulation of these vessels. Second, if patients undergoing
interventional cardiology procedures have the same cognitive outcomes as cardiac
surgical patients, it appears that chronic cerebral vascular disease may constitute the
foundation of “postoperative” cognitive decline.

Finally, it appears that cognitive decline, and probably delirium, are nonspecific events
occurring in older patients following noncardiac as well as cardiac surgery. The
frequency and severity may be higher in cardiac surgery simply because it is more
physiologically stressful than noncardiac surgery.

Exquisite control of many physiologic variables is possible during CPB and in the
perioperative period. Until more effective interventions to improve neurologic
outcomes are more clearly identified, we will help our patients most by understanding
their risk factors and placing them in an anesthetic and surgical milieu that minimizes
the impact of surgical stressors while maximizing compensatory mechanisms.

KEY Points

● Age is an independent predictor of postoperative stroke. The proportion of

patients older than 80 years undergoing cardiac surgery will increase faster
than any other group.

● The conduct of CPB has a significant effect on the development of cognitive

decline following cardiac surgery. Patients with neurodegenerative changes are
more susceptible to this disabling condition.

● Up to 50% of cardiac surgery patients will develop postoperative delirium and

the incidence is often underappreciated. Strategies to prevent delirium have
been largely unsuccessful.

● In the absence of hypoxemia, neurologic morbidity results from inadequate

tissue perfusion; embolic phenomena, flow limitation by fixed obstruction, and
failure by collateral circulation. All these are aggravated by low MAP common
with CPB.

● Maintenance of normal pump flow is not enough for adequate CBF if MAP is
reduced.

● During CPB, the mean arterial pressure is a determinant of neurologic and

cardiac complications. It is probably prudent to maintain MAPs not less than 20
 mm Hg below preoperative MAP. It is also important to consider having the
MAP between 80 and 100 for patients with potential for critical arterial
stenosis.

● The STS recommends that glucose levels be maintained below 180 mg/dL, with
continuous insulin infusion for diabetic and nondiabetic patients.

● The effects of the Paco2 management (α-stat vs. pH-stat) can be beneficial or
detrimental depending on the cerebral circulatory circumstances of the
operation, but is primarily relevant below 27°C.

● Aortic cannulation guided by epiaortic ultrasound can improve neurologic

outcomes in high-risk patients by reducing embolic risk.

● Small changes in the patient’s temperature importantly effect the vulnerability

to ischemia. The cardiac surgery team should be vigilant to avoid hyperthermia
during rewarming.

● Despite the best care, patients will develop acute stroke; and having a
structured stroke team is pivotal for successful management.

● Fifty percent of cardiac surgery-related strokes seem to occur in the

postoperative period and further attention to reducing this risk is indicated.