Brazilian Journal of Anesthesiology 2022;72(3): 322−330

ORIGINAL INVESTIGATION

Perioperative goal-directed fluid management using

noninvasive hemodynamic monitoring in gynecologic
oncology
a, a a a
Gunes O. Yildiz *, Gulsum O. Hergunsel , Gokhan Sertcakacilar , Duygu Akyol ,
b
Sema Karakaş , Zafer Cukurova a

a
University of Health Sciences Istanbul, Bakirkoy Dr. Sadi Konuk Training and Research Hospital, Department of Anesthesiology and
Intensive Care, _Istanbul, Turkey
b
University of Health Sciences Istanbul, Bakirkoy Dr. Sadi Konuk Training and Research Hospital, Department of Gynecological
Oncology, _Istanbul, Turkey

Received 9 February 2021; accepted 26 December 2021

Available online 1 February 2022

KEYWORDS Abstract
Fluid therapy; Background: Intraoperative fluid management is important for the prevention of perioperative
Hemodynamic morbidity and mortality. Our study aimed to investigate the perioperative feasibility and benefits
monitoring; of Goal-Directed Fluid Management (GDFM) using noninvasive hemodynamic monitoring in gyne-
Goal-directed fluid cologic oncology patients with acute blood loss and severe fluid loss. We assessed the effects of
therapy; GDFM on hemodynamics, organ perfusion, complications, and mortality outcomes.
Gynecologic debulk- Methods: This randomized prospective study included 104 patients over the age of 18 years,
ing surgery including 56 patients with endometrial cancer and 48 patients with ovarian cancer who had open
surgery. The anesthetic approach was standardized for all patients. We compared the periopera-
tive results of the subjects who were randomized into GDFM (n = 51) and Liberal Fluid Manage-
ment (LFM) (n = 53) groups using a computer program.
Results: The median perioperative crystalloid replacement (2000 vs. 2700; p < 0.001) and total
volume of fluid (2260 vs. 3200; p < 0.001) were lower in the GDFM group compared to the LFM
group. The hemodynamic findings and the HCO3 and lactate levels of the GDFM group did not sig-
nificantly change perioperatively. The heart rate, mean arterial pressure, and HCO3 levels of the
LFM group decreased and serum lactate levels increased perioperatively. The hospitalization
rate in ICU (7.8% vs. 28.3%; p = 0.010), rate of patients with comorbidity conditions indicated in
ICU (2% vs. 17%; p = 0.024), and rate of complications (17.6% vs. 35.8%; p = 0.047) were lower in
the GDFM group compared to the LFM group.
Conclusion: The amount of intraoperatively administered crystalloid solution and complication
rates were significantly lower in gynecologic oncologic surgery patients who received GDFM.

* Corresponding author.
E-mail: gunesozlemyildiz@gmail.com (G.O. Yildiz).

https://doi.org/10.1016/j.bjane.2021.12.012
0104-0014/© 2022 Sociedade Brasileira de Anestesiologia. Published by Elsevier Editora Ltda. This is an open access article under the CC BY-
NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 Brazilian Journal of Anesthesiology 2022;72(3): 322−330

Besides, hemodynamic findings, and lactate levels of the GDFM group did not change significantly
during the perioperative period.
© 2022 Sociedade Brasileira de Anestesiologia. Published by Elsevier Editora Ltda. This is an open
access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/
4.0/).

Introduction for major abdominal oncologic surgery (expected duration

≥ 2 h) were evaluated. The exclusion criterias were as fol-
Intraoperative fluid management is important for the preven- lows: patients younger than 18 years, Body Mass Index (BMI)
tion of peri- and postoperative complications, morbidity, and ≥ 35 kg.m 2, laparoscopic surgery, peripheral artery dis-
mortality, particularly in gynecologic oncological operations ease, chronic obstructive pulmonary disease, obstructive
involving multi-organ resection for tumors. Surgery can lead sleep apnea syndrome, arrhythmia, advanced kidney or liver
to significant fluid loss due to prolonged peritoneal air expo- failure, decompensated heart failure (ejection fraction of
sure, significant blood loss, and acute drainage of tumor- < 30%), and not being able to tolerate a tidal volume
associated ascites. Intraoperative drainage of ascites is fol- of 8 mL.kg 1 in mechanical ventilation (for accuracy of the
lowed by reduced fluids in vessels, which may require the Pleth Variability Index). Twenty three patients who met the
administration of large volumes of fluid for replacement.1 exclusion criteria were excluded from the study. A total
While planning fluid management, the anesthesiologist needs of 104 patients, including 56 patients with endometrial can-
to ensure adequate blood flow for adequate and safe tissue cer and 48 patients with ovarian cancer, who met the inclu-
oxygenation and the capacity of the cardiac output to meet sion criteria were included in the study (Figure 1).
metabolic requirements. The study was planned as per the 2013 Brazil revision of
Goal-Directed Fluid Management (GDFM) includes the the Helsinki Declaration. All patients signed informed con-
evaluation of fluid sensitivity and cardiac functions using sent forms. The study was granted ethical approval by the
dynamic and static parameters to measure cardiac output in local ethics committee (2019\152) and was registered as a
order to optimize oxygen supply to tissues during surgery. clinical trial (NCT03956901).
Randomized controlled trials and meta-analyses have Assuming an alpha of 0.05, a power of 0.80, and consis-
reported superior outcomes for GDFM in terms of organ dam- tent with previous reports7 with 10% lower complication
age, mortality, wound healing, length of hospital stay, and rate in the GDFM group, the estimated sample size was at
length of ICU stay compared to Liberal Fluid Management least 50 patients in both protocol groups. Age and body mass
(LFM).2,3 The established advantages of GDFM compared to index were evaluated as potential confounders. According
LFM and the subsequent financial benefits have led to GDFM to this, patients were single-blindly assigned to the targeted
being recommended as the standard of care by a large num- fluid therapy or liberal fluid therapy branches by computer
ber of scientific communities and being included in treat- stratified randomization analysis with a ratio of 1:1 in the
ment guidelines.4,5 STATA program (StataCorp LLC, Texas, USA). Anesthesiolo-
Patients undergoing gynecologic oncologic surgery are gists were not blinded to group assignments. However,
prone to significant fluid shift and blood loss and therefore patients, surgeons, PACU and surgical nurses, and those who
are at risk for hypovolemia, end-organ hypoperfusion, and provided the data analysis were blinded. The data of the
adverse postoperative outcomes.6 Although GDFM has been patients were obtained by saving them on the memory card
frequently investigated in patients undergoing major of the monitor at the end of the case.
abdominal surgery, it has been rarely investigated in
patients undergoing major gynecologic surgery. In addition,
intravascular volume depletion and acute hemorrhage are Anesthesia initiation
important problems in gynecological oncology cases. In
these cases, management of perioperative GDFM with mini- The anesthetic approach was standardized for all patients.
mal therapeutic procedures is essential. Therefore, in this Patients were allowed clear fluid intake up to 2 hours and
study, we investigated the effects of GDFM on hemodynamic solid food intake up to 6 hours before the operation. In the
morbidity using PVI and investigated the effects of GDFM on operating room, heart rate, oxygen saturation, Central
mortality. Venous Pressure (CVP), and invasive and noninvasive arterial
pressure measurements were made with a Datex S/5 moni-
tor (Datex OhmedaÒ , GE Healthcare, Chicago, IL, USA). In
Methods the GDFM group, hemodynamic fluid responsiveness was
achieved by Pleth Variability Index (PVI) monitoring with a
Study design and population finger sensor (in the arm without invasive arterial cannula-
tion, using the finger probe from the 4th finger and covered
This prospective study was conducted between May 2019 in an opaque manner) (Masimo Corporation, Radikal 7, USA).
and December 2019 at the Gynecological Oncology Clinic of Vascular access was opened with a 16−18G catheter for all
the Health Sciences University Istanbul Bakirkoy Dr. Sadi patients. Before anesthesia induction, 0.03 mg.kg 1 midazo-
Konuk Training and Research Hospital. One hundred and lam was administered, and epidural catheter was inserted
twenty seven adult patients with an American Society of for postoperative analgesia. No medication was adminis-
Anesthesiologists (ASA) physical status of II−III presenting tered to the patients through the intraoperative epidural

323
 G.O. Yildiz, G.O. Hergunsel, G. Sertcakacilar et al.

Figure 1 Trial profile. (GDFM, Goal-Directed Fluid Management; LFM, Liberal Fluid Management; MAP, Mean Arterial Pressure; NE,
Norepinephrine, PVI, Pleth Variability Index).

catheter. All patients had an arterial line for arterial blood Maintenance fluid therapy was continued in patients with
gas monitoring. MAP > 65 mmHg. If Mean Arterial Pressure (MAP) was <
After anesthesia induction with 1 mg.kg 1 fentanyl and 65 mmHg, the patient was administered a 250 mL bolus of
2−3 mg.kg 1 propofol, the patient was curarized crystalloid solution and 1−2 mg bolus Norepinephrine (NE).
with 0.6 mg.kg 1 rocuronium and connected to a ventilator. Afterwards, MAP was followed at 5-minute intervals. If MAP
Mechanical ventilation was set to the volume-controlled was still < 65 mmHg after subsequent measurements, the
mode with I:E 1:2, tidal volume 8 mL.kg 1 of ideal body patient was given 250 mL of crystalloid solution. In continu-
weight, and respiratory frequency adjusted according to ing hemodynamic follow-ups, if MAP < 65 mmHg despite
end-tidal CO2 (ECO2 35−45 mmHg). Mechanical ventilation mini fluid challenge, NE 0.1−1 mg.kg 1.min 1 (from a solu-
was maintained with a fresh gas flow rate of 3 L.min 1 tion at a concentration of 0.08 mg.mL 1) infusion was
with 40% oxygen. Anesthesia was maintained with 1 MAC sev- started and MAP adjusted to > 65 mmHg.
oflurane, 0.01−0.2 mg.kg 1.min 1 remifentanil infusion,
and rocuronium as needed. GDFM application
A crystalloid fluid bolus (Ringer’s lactate solution) of 500 mL
Perioperative fluid management was infused together with the induction of anesthesia,
followed by 2 mL.kg 1 crystalloid fluid infusion. If the PVI
LFM application was < 13% and MAP > 65 mmHg, the fluid infusion of the
A crystalloid fluid bolus (Ringer’s lactate solution) of 500 mL patients was continued at the adjusted dose. If PVI was
was infused together with the induction of anesthesia, fol- < 13% and MAP was < 65 mmHg, fluid infusion was continued,
lowed by a maintenance infusion of 4−8 mL.kg 1.h 1. and a 1−2 mg bolus NE was administered. Similar to the LFM

324
 Brazilian Journal of Anesthesiology 2022;72(3): 322−330

Table 1 Preoperative demographic, clinical, and laboratory findings.

Variables All population GDFM LFM p

n = 104 n = 51 n = 53
Age, years 57.1§12.3 55.5§12.5 58.6§12.1 0.191
BMI, kg.m 2 31.4§9.0 30.3§6.0 32.4§11.1 0.245
ASA, n (%)
II 69 (66.3) 35 (68.6) 34 (64.2) 0.681
III 35 (33.7) 16 (31.4) 19 (35.8)
Surgery time, minutes 307.5 (120‒550) 300 (120‒550) 350 (145‒510) 0.098
Diagnosis, n (%)
Endometrial 57 (54.8) 27 (52.9) 30 (56.6) 0.844
Ovarian 47 (45.2) 24 (47.1) 23 (43.4)
Types of procedures, n (%)
Endometrial n = 57 n = 27 n = 30
Hysterectomy 57 (100) 27 (100) 30 (100) -
Bilateral BSO 57 (100) 27 (100) 30 (100) -
PPALND 57 (100) 27 (100) 30 (100) -
Ovarian n = 47 n = 24 n = 23
PPALND 44 (93.6) 22 (91.7) 22 (95.7)
DPS/R 13 (27.7) 8 (33.3) 5 (21.7) 0.574
Bowel resection 8 (17.0) 5 (20.8) 3 (13.0) 0.747
LWR 4 (8.5) 2 (8.3) 2 (8.7) 0.999
Splenectomy 4 (8.5) 2 (8.3) 2 (8.7) 0.999
VATS 2 (4.3) ‒ 2 (8.7) 0.451
Ascites fluid, n (%)
No 83 (79.8) 41 (80.4) 42 (79.2) 0.999
Yes 21 (20.2) 10 (19.6) 11 (20.8)
Hemoglobin, g.dL 1 11.4§1.6 11.4§1.5 11.5§1.7 0.724
Albumine, mg.dL 1 37.6§4.6 37.1§4.9 38.0§4.2 0.349
Creatinine, mg.dL 1 0.6 (0.3−9.4) 0.6 (0.3−9.4) 0.6 (0.3−1.7) 0.554
Numerical variables with normal distribution were shown as mean § standard deviation. Numerical variables that do not show normal dis-
tribution are shown as median (min−max). Categorical variables were shown as numbers (%).
ASA, American Society of Anesthesiologists physical status; BMI, Body Mass Index; Ca, Cancer; GDFM, Goal-Directed Fluid Management;
LFM, Liberal Fluid Management; ICU, Intensive Care Unit; PPALND, Pelvic Paraaortic Lymph Node Dissection; DPS/R, Diaphragmatic Perito-
neal Stripping/Resection; LWR, Liver Wedge Resection; VATS, Video-Assisted Thoracoscopic Surgery.

group, MAP was followed at 5-minute intervals. If PVI was or 20% human albumin at 3 mL.kg 1 were used targeting
< 13% and MAP was < 65 mmHg after subsequent measure- PVI < 13%.9 In the LFM group, the amount of FFP and albumin
ments, a repeat dose of NE bolus was administered. In were determined according to MAP, presence of tachycardia,
repeated measurements, NE infusion of 0.1−1 mg.kg 1. and urine output findings. In these groups, the choice
min 1 (from a solution with a concentration of 0.08 mg. between using balanced crystalloids or colloids (albumin
mL 1) was started in patients with MAP < 65 mmHg. If PVI 100 mL) was left to the attending anesthesiologist.
was > 13% and MAP was < 65 mmHg, the patient was admin- Following perioperative hemodynamic stabilization, we
istered a 250 mL crystalloid fluid bolus. If MAP < 65 mmHg recorded PVI, blood gas values, perioperative blood loss,
persisted, 1−2 mg bolus NE was administered. If the PVI and the amount of administered fluids at T0 (15 minutes
was > 13% and MAP was < 65 mmHg in subsequent measure- after anesthesia induction), T1 (hour 1), T2 (hour 2), and
ments, a 250 mL crystalloid bolus was repeated for the last after extubation. At the end of the operation, the patients
time. NE infusion of 0.1−1 mg.kg 1.min 1 (0.08 mg.mL 1 were assessed to decide between extubation or admission to
solution) was started in patients with MAP < 65 mmHg on the ICU according to hemodynamic findings. NE infusion was
repeated measurements, and fluid infusion was continued evaluated and decided considering ongoing hemodynamic
until PVI was < 13%. The total fluid administered was instability, respiratory distress, massive blood transfusion,
recorded. and additional comorbidities (uncontrolled diabetes, uncon-
In patients with bleeding, if arterial blood gas, a target trolled hypertension) for ICU admission.
hemoglobin concentration within the target range (7−9 g. On the postoperative 24th hour, we recorded fever, creat-
dL 1), and PVI > 13%, they received erythrocyte suspen- inine, length of ICU stay, and length of hospital stay (the
sion.8 If more than 1,000 mL of ascitic fluid was aspirated in length of hospital stay was defined as the postoperative fol-
patients with ascites in the GDFM group, to maintain oncotic low-up in bed until discharge). The patients were reached
pressure, 1−2 units of Fresh Frozen Plasma (FFP) by phone on postoperative day 30 to confirm survival.

325
 G.O. Yildiz, G.O. Hergunsel, G. Sertcakacilar et al.

Statistical analysis the LFM group. Other perioperative clinical findings were
not significantly different for the GDFM and LFM groups
Data were analyzed using SPSS 20 for Windows (IBM Corp., (Table 2).
Armonk, NY, USA). The normal distribution of the data was The mean Heart Rate (HR), mean MAP, and median serum
evaluated by the Kolmogorov-Smirnov test. The normally lactate levels were similar for the two groups at T0 and T1,
distributed variables were presented as mean § standard whereas they were higher in the LFM group at T2. Other
deviation while the non-normally distributed variables were perioperative laboratory findings were not significantly dif-
presented as median (min−max). Categorical variables were ferent for the two groups (Table 3).
presented as numbers and percentages. We used Student’s The decrease in mean HR and mean MAP and the increase
t-test for the intergroup comparison of normally distributed in median lactate levels and median base excess were signifi-
variables and the Mann-Whitney U test for the intergroup cantly higher in the LFM group. The changes in other perioper-
comparison of non-normally distributed variables. Chi- ative laboratory parameters were not significantly different
Square and Fisher’s exact tests were used for the intergroup (Table 3).
comparison of categorical variables. The changes in labora- When preoperative and postoperative mean hemoglobin,
tory findings before and after the operation were evaluated mean albumin, and median creatinine levels were com-
using the Generalized Linear Mixed Model (GLMM) for the pared, and the changes observed were similar in the GDFM
analysis of longitudinal data with repeated measures for all and LFM groups (Table 4).
patients and the two groups (GDFM and LFM). Values of p < The postoperative findings are shown in detail in Table 5.
0.05 were considered statistically significant. The hospitalization rate in ICU (7.8% vs. 28.3%; p = 0.010),
rate of patients with comorbidities in ICU (2% vs. 17%;
p = 0.024), and rate of complications (17.6% vs. 35.8%;
Results p = 0.047) were lower in the GDFM group compared to the
LFM group. Other postoperative events and mortality out-
The mean age of the subjects was 57.1 § 12.3 years, and comes were not statistically different (Table 5).
66.3% (n = 69) were ASA II. There was no significant differ-
ence between the groups in terms of demographic charac-
teristics and preoperative clinical and laboratory findings. Discussion
Table 1 presents the relevant data in detail.
The median amount of crystalloid solution (2000 vs. 2700; In this cohort of gynecologic oncologic surgery patients, we
p < 0.001) and median total volume of fluid (2260 vs. 3200; applied a pulse contour analysis-guided hemodynamic and
p < 0.001) were lower in the GDFM group compared to static management protocol. We found that GDFM provides

Table 2 Perioperative clinical findings.

Variables All population GDFM LFM p

n = 104 n = 51 n = 53
Blood loss, mL 400 (50−3000) 350 (50−3000) 500 (50−2000) 0.484
Urine output, mL 150 (25−1500) 150 (25−1500) 175 (50−1100) 0.169
Administered NE infusion, n (%) 9 (8.7) 2 (3.9) 7 (13.2) 0.182
Intravenous fluid replacement
Crystalloid solution, n (%) 104 (100.0) 51 (100.0) 53 (100.0) ‒
Volume, mL 2500 (812−6000) 2000 (812−5000) 2700 (1000−6000) < 0.001a
Colloid solution, n (%) 49(47.1) 20(39.2) 29(54.7) 0.122
Volume, mL 500 (500−1000) 500 (500−1000) 500 (500−1000) 0.204
Total fluid volume, mL 2750 (812−8400) 2260 (812−7400) 3200 (1500−8400) < 0.001a
Erythrocyte replacement, n (%) 0.074
None 74 (71.2) 39 (76.5) 35 (66.0)
1 unit 8 (7.7) 6 (11.8) 2 (3.8)
2 units 13 (12.5) 3 (5.9) 10 (18.9)
3 units 9 (8.7) 3 (5.9) 6 (11.3)
Fresh Frozen Plasma, n (%) 0.335
None 67 (64.4) 35 (68.6) 32 (60.4)
1 unit 5 (4.8) 4 (7.8) 1 (1.9)
2 units 19 (18.3) 8 (15.7) 11 (20.8)
3 units 6 (5.8) 3 (5.9) 3 (5.7)
4 units 5 (4.8) 1 (2.0) 4 (7.5)
5 units 2 (1.9) ‒ 2 (3.8)
Numerical variables with normal distribution were shown as mean § standard deviation. Numerical variables that do not show normal dis-
tribution are shown as median (min−max). Categorical variables were shown as numbers (%).
GDFM, Goal-Directed Fluid Management; LFM, Liberal Fluid Management.
a
p < 0.05 shows statistical significance.

326
 Brazilian Journal of Anesthesiology 2022;72(3): 322−330

Table 3 Perioperative laboratory findings.

Variables All population GDFM LFM p

n = 104 n = 51 n = 53
1
Hemoglobin, g.dL
T0 10.8§1.5 11.0§1.5 10.7§1.6 0.356
T1 10.8§1.6 10.9§1.5 10.6§1.7 0.333
T2 10.7§1.7 10.9§2.0 10.6§1.4 0.322
Postextubation 11.0§1.3 11.1§1.4 11.0§1.2 0.808
pr 0.140 0.845 0.102
pd 0.805
Hearth rate, beats.min 1
T0 78.4§15.4 79.6§17.6 77.3§13.1 0.444
T1 68.9§12.2 71.9§14.3 66.0§9.1 0.014a
T2 69.1§12.5 72.3§13.3 66.2§10.9 0.012a
Postextubation 84.1§14.3 89.3§15.7 79.2§10.8 <0.001a
pr < 0.001a < 0.001a < 0.001a
pd 0.016a
Mean arterial pressure, mmHg
T0 79.2§14.2 82.7§18.0 75.8§8.0 <0.001a
T1 77.8§14.4 83.5§15.1 69.3§6.6 <0.001a
T2 75.9§13.9 82.6§15.5 69.4§8.1 <0.001a
Postextubation 85.1§14.9 93.2§14.3 77.3§10.9 <0.001a
pr < 0.001a 0.003a < 0.001a
pd 0.003a
pH
T0 7.5§0.1 7.5§0.1 7.5§0.1 0.777
T1 7.4§0.1 7.4§0.1 7.4§0.1 0.279
T2 7.4§0.1 7.4§0.1 7.4§0.1 0.107
Postextubation 7.4§0.1 7.4§0.1 7.4§0.1 0.198
pr < 0.001a < 0.001a < 0.001a
pd 0.138
HCO3, mmoL.L 1
T0 24.5§2.3 24.7§2.5 24.3§2.1 0.335
T1 24.0§3.6 24.9§4.2 23.2§2.6 0.013a
T2 23.9§3.3 24.8§3.9 23.2§2.4 0.013a
Postextubation 23.4§3.0 24.0§3.2 23.0§2.7 0.070
pr 0.004a 0.361 < 0.001a
pd 0.048a
Lactate, mEq.L 1
T0 1.1 (0.1‒3.2) 1.1 (0.5‒2.8) 1.1 (0.1‒3.2) 0.698
T1 1.1 (0.5‒3.5) 1.1 (0.5‒3.3) 1.1 (0.5‒3.5) 0.573
T2 1.2 (0.4‒6.3) 1.2 (0.4‒3.8) 1.4 (0.6‒6.3) 0.042a
Postextubation 1.7 (0.7‒6.5) 1.8 (0.7‒6.4) 1.8 (0.7‒6.5) 0.665
pr < 0.001a < 0.001a < 0.001a
pd 0.047a
Base excess, mEq.L 1
T0 -0.1((-8.9)‒(6.9)) -0.1((-4.1)‒(5.3)) 0.1((-8.9)‒(6.9)) 0.610
T1 -0.9((-14)‒(6.9)) -0.3((-4.9)‒(5.4)) -1.3((-14)‒(6.9)) 0.107
T2 -0.5((-11)‒(4.5)) -0.3((-6.7)‒(4.5)) -0.6((-11)‒(3.6)) 0.105
Postextubation -1.3((-13)‒(7.5)) -0.4((-5.1)‒(5.9)) -1.7((-13.0)‒(7.5)) 0.192
pr < 0.001a 0.121 < 0.001a
pd 0.032a
Numerical variables with normal distribution were shown as mean § standard deviation. Numerical variables that do not show normal dis-
tribution are shown as median (min−max). Categorical variables were shown as numbers (%).
HCO3, Bicarbonate; GDFM, Goal-Directed Fluid Management; LFM, Liberal Fluid Management; pH, power of Hydrogen, T0, perioperative
baseline, T1, 1-hour later; T2, 2-hour later.
a
p < 0.05 shows statistical significance. pr, Statistical difference of changes in laboratory findings in the group. pd, Statistical difference
of changes in laboratory findings between groups.

327
 G.O. Yildiz, G.O. Hergunsel, G. Sertcakacilar et al.

Table 4 Preoperative vs. postoperative laboratory findings.

Variables GDFM (n = 51) pr LFM (n = 53) pr Dp

1
Hemoglobin, g.dL
Preoperative 11.4§1.5 < 0.001a 11.5§1.7 < 0.001a 0.818
Postoperative 10.5§1.3 10.6§1.3
Albumin, mg.dL 1
Preoperative 37.1§4.9 < 0.001a 38.0§4.2 < 0.001a 0.230
Postoperative 28.8§3.5 28.6§4.5
1
Creatinine, mg.dL
Preoperative 0.6 (0.3−9.4) 0.989 0.6 (0.3−1.7) 0.101 0.605
Postoperative 0.7 (0.3−8.6) 0.7 (0.4−2.3)
Numerical variables with normal distribution were shown as mean § standard deviation. Numerical variables that do not show normal dis-
tribution are shown as median (min−max). Categorical variables were shown as numbers (%).
GDFM, Goal-Directed Fluid Management; LFM, Liberal Fluid Management; pH, power of Hydrogen; WBC, White Blood Cell; CRP, C-Reactive
Protein.
a
p < 0.05 shows statistical significance. pr, Statistical difference of changes in laboratory findings in the group. Dp, Statistical differ-
ence of changes in laboratory findings between groups.

favorable tissue oxygenation (such as stable serum lactate Goal-directed fluid therapy suggests that reducing fluid
values) and reduces the amount of perioperative fluid administration when hemodynamic targets are met can help
administration due to its hemodynamic effects (such as sta- avoid inadequate fluid replacement.11 Our results are con-
ble MAP) compared to LFM, and GDFM was associated with sistent with the literature. In our study, we administered a
reduced ICU admissions and postoperative complications. crystalloid solution to all patients, but the amount of admin-
Excess preoperative fluid administration in major abdom- istered crystalloid solution was lower in the GDFM group
inal and thoracic surgeries is associated with postoperative compared to the LFM group. The use of crystalloid solutions
complications. Liberal and restrictive fluid management alone may result in inadequate intravenous fluid replace-
have been shown to be hazardous for high-risk surgical ment and iatrogenic fluid loading,12 which may place
patients due to the risk of hypovolemia.10 Therefore, for the patients at risk. Moreover, the GDFM method can reduce the
possibility of safely limiting and personalizing fluid adminis- risk of tissue edema that may result from excessive crystal-
tration in high-risk surgical patients, GDFM is being recom- loid fluid administration.13 At the same time, the proportion
mended as the standard of care in relevant guidelines.4,5 of patients given both colloid solution and FFP in the LFM

Table 5 Event and mortality findings after the operation.

Variables All population (n = 104) GDFM (n = 51) LFM (n = 53) p

Hospitalization in ICU, n (%) 19 (18.3) 4 (7.8) 15 (28.3) 0.010a
Duration of stay in ICU, days 1 (1−10) 2 (1−10) 1 (1−9) 0.242
Indications of ICU, n (%)
Comorbidity 10 (9.6) 1 (2.0) 9 (17.0) 0.024a
Closer hemodynamic monitoring 5 (4.8) 1 (2.0) 4 (7.5) 0.383
Inotropic therapy 11 (10.6) 2 (3.9) 9 (17.0) 0.065
Hemodynamic instability 3 (2.9) 2 (3.9) 1 (1.9) 0.973
Duration of stay in hospital, days 8 (0−30) 7 (2−30) 9 (2−28) 0.078
Fever, n (%) 8 (7.7) 4 (7.8) 4 (7.5) 0.999
Complication, n (%) 28 (26.9) 9 (17.6) 19 (35.8) 0.047a
Wound infection 15 (14.4) 6 (11.8) 9 (17.0) 0.579
Pulmonary edema 4 (3.8) ‒ 4 (7.5) 0.136
Evisceration 3 (2.9) 2 (3.9) 1 (1.9) 0.973
Acute renal failure 2 (1.9) 1 (2.0) 1 (1.9) 0.999
Retroperitoneal hematoma 2 (1.9) 1 (2.0) 1 (1.9) 0.999
Postoperative ileus 2 (1.9) 1 (2.0) 1 (1.9) 0.999
Urethral stricture 1 (1.0) ‒ 1 (1.9) 0.999
Anastomotic leak 1 (1.0) ‒ 1 (1.9) 0.999
Mortality, n (%) 1 (1.0) ‒ 1(1.9) 0.999
Numerical variables that do not show normal distribution are shown as median (min−max). Categorical variables were shown as numbers
(%).
GDFM, Goal-Directed Fluid Management; LFM, Liberal Fluid Management; ICU, Intensive Care Unit.
a
p < 0.05 shows statistical significance.

328
 Brazilian Journal of Anesthesiology 2022;72(3): 322−330

group was partially higher than the GDFM group. This situa- conclusion in a meta-analysis that complementary hydration
tion supports the need for more fluid replacement to provide alone does not affect wound infection rate, regardless of any
hemodynamic stability in the LFM group, in line with the lit- specific hemodynamic goal.22 However, when the amount of
erature.7,9 In addition, although the proportion of patients perioperatively applied fluids is aimed at increasing subcuta-
with ascitic fluid was similar in both groups, these patients neous oxygen tension, a greater amount of collagen in wound
corresponded to half of the patients given colloid solution healing has been observed.23 In particular, anastomotic leak,
and FFP. It has been suggested that FFP transfusion in pulmonary edema, and urethral stricture were observed only
patients with malignant ascites may help restore hemody- in the LFM group. A recent study has shown that patients
namic stability.14 treated with a non-guided restrictive fluid administration pro-
Dynamic variables based on ventilation-induced changes tocol had a higher risk of postoperatively developing acute
in cardiac output help predict fluid response. Timely inter- kidney injury and that this was associated with lower intrao-
ventions based on the individual needs of the patient can perative urine output.24 Large amounts of IV fluid replace-
prevent hypotension and hemodynamic imbalance.15,16 ment can exacerbate capillary leakage associated with the
Arterial pulse pressure variation is cited as one of the systemic inflammatory response to surgery and may contrib-
most useful dynamic variables in guiding volume replace- ute to the development of intestinal anastomotic edema.25
ment.15,17 In our study, perioperative decreases in MAP and Therefore, the complications that were observed only in the
HR were significantly higher in the LFM group compared to LFM group may have been due to the unsteady mucosal perfu-
the GDFM group. The benefits of GDFM may be associated sion that resulted from unstable fluid therapy.26,27 However,
with the administration of fluids at the right time, helping in our study, the distribution of complication types deter-
avoid hypovolemia and hypoperfusion. Another advantage of mined in both groups did not differ significantly. In our study,
the GDFM protocol is using NE as the first-line intervention the complication rates in the GDFM group were consistent
for hypotension, independently of stroke volume variation.18 with GDFM applications presented in similar gynecologic sur-
In our study, the comparative norepinephrine usage rate was gery cohorts.12,28 One patient died during the 30-day follow-
approximately 1:3 for the GDFM and LFM groups, even though up. This patient was in the LFM group and died due to a post-
this finding was not statistically significant. These findings sup- operative ileus complication. Postoperative ileus is a common
port the idea that the risk of hypotension is lower for GDFM, and complication of intra-abdominal surgeries. However, it is sug-
the patient is less likely to require NE intervention. gested that targeted fluid therapy may not be an effective
The perioperative tissue perfusion outcomes of GDFM are strategy in reducing the risk of postoperative ileus.29 Also, a
superior to those of LFM approaches. When oxygen supply is longer-term follow-up would possibly change the postopera-
insufficient, pyruvate cannot enter the Krebs cycle and is tive mortality rate.
shunted to lactate. Therefore, serum lactate levels reflect The major limitations of the present study are the rela-
anaerobic cellular metabolism that results from tissue hyp- tively small number of patients and the results are based on
oxia.19 Serum lactate levels increase when oxygen supply is a single center experience. In addition, we evaluated
insufficient to meet tissue metabolism requirements. Latent only 30-day short term survival outcomes. Another impor-
hypoperfusion without clinical symptoms of shock elevates tant limitation is the intensive care indications due to the
serum lactate levels.20 These are solid parameters for the preoperative comorbidities of the patients. The higher ICU
assessment of inadequate oxygen perfusion. In our study, hospitalization rate observed especially in the LFM group
serum lactate levels did not change in the GDFM group but may be due to comorbidity including ICU indication. On the
increased in the LFM group. Forget et al.16 demonstrated other hand, the BMI level was slightly higher in the LFM
that GDFM with PVI applied during major abdominal surgery group. In our study, calculating the tidal volume based on
was associated with lower lactate levels and that these ideal body weight, is appropriate to exclude patients with
patients required significantly less crystalloid fluid adminis- very high BMI. Various metabolic and neurohormonal
tration. The results of our cohort of patients undergoing changes commonly associated with overweight or obesity
gynecologic oncological surgery suggest that GDFM is supe- may have contributed to abnormalities in cardiac morphol-
rior to LFM as demonstrated by more stable serum lactate ogy and function.30 One of the important limitations of our
levels and reduced fluid replacement. study is that muscle relaxation was not monitored. We used
Morbidity and mortality are significantly higher in patients rocuronium as a muscle relaxant in our patients. Monitoring
who undergo high-risk surgeries. Most of these patients have and maintenance of a constant level of muscle paralysis
clinically significant dehydration preoperatively, and patients could have minimized the possibility that muscular contrac-
lose varying volumes of fluid intraoperatively. Insufficient tion influenced on lung mechanics.
intravascular volume reduces pulse volume and cardiac out- We conclude that the clinical use of GDFM in gynecologic
put, resulting in inadequate tissue and organ perfusion. This oncological surgery is beneficial for limiting the amount of
can result in the development of serious complications, lon- intraoperatively administered crystalloid solution without
ger-term hospitalization, and, without the appropriate inter- causing hypovolemia, hypotension, or serum lactate eleva-
ventions, even death. For this reason, intraoperative fluid tion. This may be associated with rapid recovery of bowel
therapy is important both for surveillance and for preventing function, wound healing, and reduced ICU admission, length
postoperative complications. In our study, ICU admission and of hospital stay, and complication rates.
complication rates were lower in the GDFM group, consistent
with the literature.2,10,12,21 Also, the complication rate of the
LFM group was almost double that of the GDFM group. In our Conflicts of interest
study, the distribution of patients with wound infection in the
LFM and GDFM groups was similar. This is consistent with the The authors declare no conflicts of interest.

329
 G.O. Yildiz, G.O. Hergunsel, G. Sertcakacilar et al.

References reduces lactate levels and improves fluid management. Anesth

Analg. 2010;111:910−4.
1. Edwards MR, Grocott MP. Perioperative fluid and electrolyte 17. Lopes MR, Oliveira MA, Pereira VOS, Lemos IPB, Auler Jr JOC,
therapy. Young. 2015;60:7. Michard F. Goal-directed fluid management based on pulse pres-
2. Giglio MT, Marucci M, Testini M, Brienza N. Goal-directed hae- sure variation monitoring during high-risk surgery: a pilot ran-
modynamic therapy and gastrointestinal complications in major domized controlled trial. Crit Care. 2007;11:R100.
surgery: a meta-analysis of randomized controlled trials. Br J 18. Monnet X, Jabot J, Maizel J, et al. Norepinephrine increases
Anaesth. 2009;103:637−46. cardiac preload and reduces preload dependency assessed by
3. Gomez-Izquierdo JC, Feldman LS, Carli F, Baldini G. Meta-analy- passive leg raising in septic shock patients. Crit Care Med.
sis of the effect of goal-directed therapy on bowel function 2011;39:689−94.
after abdominal surgery. Br J Surg. 2015;102:577−89. 19. Jansen TC, van Bommel J, Bakker J. Blood lactate monitoring in
4. Powell-Tuck J, Gosling P, Lobo DN, et al. British Consensus critically ill patients: a systematic health technology assess-
Guidelines on Intravenous Fluid Therapy for Adult Surgical ment. Crit Care Med. 2009;37:2827−39.
Patients ‒ GIFTASUP. March 2011. 20. Friedman G, Berlot G, Kahn RJ, Vincent JL. Combined meas-
5. European Society of Anaesthesiology. Perioperative goal- urements of blood lactate concentrations and gastric intra-
directed therapy protocol summary. mucosal pH in patients with severe sepsis. Crit Care Med.
6. Gan TJ, Soppitt A, Maroof M, et al. Goal-directed intraoperative 1995;23. 1184-93.
fluid administration reduces length of hospital stay after major 21. Calvo-Vecino JM, Ripolle s-Melchor J, Mythen M, et al. Effect of
surgery. Anesthesiology. 2002;97:820−6. goal-directed haemodynamic therapy on postoperative compli-
7. Russo A, Aceto P, Grieco DL, et al. Goal-directed hemodynamic cations in low-moderate risk surgical patients: a multicentre
management in patients undergoing primary debulking gynae- randomised controlled trial (FEDORA trial). British J Anaesth.
cological surgery: A matched-controlled precision medicine 2018;120:734−44.
study. Gynecol Oncol. 2018;151:299−305. 22. Kabon B, Akça O, Taguchi A, et al. Supplemental intravenous
8. Kozek-Langenecker SA, et al. Management of severe periopera- crystalloid administration does not reduce the risk of surgical
tive bleeding: guidelines from the European Society of Anaes- wound infection. Anesth Analg. 2005;101:1546−53.
thesiology: First update 2016. Eur J Anaesthesiol. 2017;34 23. Hartmann M, Jonsson K, Zederfeldt B. Effect of tissue perfusion
(6):332−95. and oxygenation on accumulation of collagen in healing
9. Feldheiser A, Pavlova V, Bonomo T, et al. Balanced crystalloid wounds. Randomized study in patients after major abdominal
compared with balanced colloid solution using a goal-directed operations. Eur J Surg. 1992;158:521−6.
haemodynamic algorithm. Br J Anaesth. 2013;110:231−40. 24. Gomez-Izquierdo JC, Trainito A, Mirzakandov D, et al. Goal-
10. Giglio MT, Marucci M, Testini M, Brienza N. Goal-directed hae- directed fluid therapy does not reduce primary postoperative
modynamic therapy and gastrointestinal complications in major ileus after elective laparoscopic colorectal surgery. Anesthesiol-
surgery: a meta-analysis of randomized controlled trials. Br J ogy. 2017;127:36−49.
Anaesth. 2009;103. 637-46. 25. Walker KG, Bell SW, Rickard MJFX, et al. Anastomotic leakage is
11. Prowle JR, Chua H-R, Bagshaw SM, Bellomo R. Clinical review: predictive of diminished suvival after potentially curative
volume of fluid resuscitation and the incidence of acute kidney resection for colorectal cancer. Ann Surg. 2004;240:255.
injury-a systematic review. Critical Care. 2012;16:230. 26. Buchs NC, Gervaz P, Secic M, Bucher P, Mugnier-Konrad B, Morel
12. McKenny M, Conroy P, Wong A, et al. A randomised prospective P. Incidence, consequences, and risk factors for anastomotic
trial of intra-operative oesophageal Doppler-guided fluid dehiscence after colorectal surgery: a prospective monocentric
administration in major gynaecological surgery. Anaesthesia. study. Int J Colorectal Dis. 2008;23:265−70.
2013;68:1224−31. 27. Kimberger O, Arnberger M, Brandt S, et al. Goal-directed col-
13. Holte K, Sharrock N, Kehlet H. Pathophysiology and clinical loid administration improves the microcirculation of healthy
implications of perioperative fluid excess. Br J Anaesth. and perianastomotic colon. Anesthesiology. 2009;110:496−504.
2002;89:622−32. 28. Russo A, Aceto P, Grieco DL, et al. Goal-directed hemodynamic
14. Hunsicker O, Fotopoulou C, Pietzner K, et al. Hemodynamic management in patients undergoing primary debulking gynae-
consequences of malignant ascites in epithelial ovarian cancer cological surgery: A matched-controlled precision medicine
surgery*: a prospective substudy of a randomized controlled study. Gynecologic Oncology. 2018;151:299−305.
trial. Medicine (Baltimore). 2015;94:e2108. 29. Arslan-Carlon V, Tan KS, Dalbagni G, et al. Goal-directed versus
15. Zimmermann M, Feibicke T, Keyl C, et al. Accuracy of stroke standard fluid therapy to decrease {leus after open radical cys-
volume variation compared with pleth variability index to pre- tectomy: A Prospective Randomized Controlled Trial. Anesthesi-
dict fluid responsiveness in mechanically ventilated patients ology. 2020;133:293−303.
undergoing major surgery. Eur J Anaesthesiol. 2010;27:555−61. 30. Alpert MA, Omran J, Bostick BP. Effects of Obesity on cardiovas-
16. Forget P, Lois F, De Kock M. Goal-directed fluid management cular hemodynamics, cardiac morphology, and ventricular func-
based on the pulse oximeter−derived pleth variability index tion. Curr Obes Rep. 2016;5:424−34.

330