Journal of Cardiothoracic and Vascular Anesthesia 36 (2022) 10211028

Contents lists available at ScienceDirect

Journal of Cardiothoracic and Vascular Anesthesia

journal homepage: www.jcvaonline.com

Original Article

Paraplegia After Open Surgical Repair Versus

Thoracic Endovascular Aortic Repair for Thoracic
Aortic Disease: A Retrospective Analysis of Japanese
Administrative Data
Takeshi Umegaki, MD, PhD*, Susumu Kunisawa, MD, PhDy,
Kota Nishimoto, MD*, Yasufumi Nakajima, MD, PhD*, ,1
Takahiko Kamibayashi, MD, PhD*, Yuichi Imanaka, MD, PhDy
*
Department of Anesthesiology, Kansai Medical University Hospital, Osaka, Japan
y
Department of Healthcare Economics and Quality Management, Graduate School of Medicine, Kyoto Uni-
versity, Kyoto, Japan

Objectives: To comparatively examine the risk of postoperative paraplegia between open surgical descending aortic repair and thoracic endo-
vascular aortic repair (TEVAR) among patients with thoracic aortic disease.
Design: Retrospective cohort study.
Setting: Acute-care hospitals in Japan.
Participants: A total of 6,202 patients diagnosed with thoracic aortic disease.
Interventions: None.
Measurements and Main Results: The main outcome of this study was the incidence of postoperative paraplegia. Multiple logistic regression models,
using inverse probability of treatment weighting and an instrumental variable (ratio of TEVAR use to open surgical repair and TEVAR uses), showed
that the odds ratios of paraplegia for TEVAR (relative to open surgical descending aortic repair) were 0.81 (95% confidence interval: 0.42-1.59;
p = 0.55) in the inverse probability of treatment-weighted model and 0.88 (0.42-1.86; p = 0.75) in the instrumental-variable model.
Conclusions: There were no statistical differences in the risk of paraplegia between open surgical repair and TEVAR in patients with thoracic
aortic disease. Improved perioperative management for open surgical repair may have contributed to the similarly low incidence of paraplegia in
these two surgery types.
Ó 2021 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/)

Key Words: paraplegia; aortic diseases; open surgical aortic repair; thoracic endovascular aortic repair; Diagnosis Procedure Combination

THE CURRENT surgical options for thoracic aortic aneur- treatment strategies and improve patient prognoses.1 A national
ysms and dissections include open surgical aortic repair and tho- survey conducted by the Japanese Association for Thoracic Sur-
racic endovascular aortic repair (TEVAR). Despite advances in gery reported that 20,746 surgical procedures were performed for
surgical techniques, these procedures still are associated with thoracic and thoracoabdominal aortic diseases throughout Japan
high postoperative mortality rates, and there is a need to optimize in 2017, and that patient numbers gradually have increased over
time.1
1
Among the postoperative complications in patients with tho-
Correspondence to. Yuichi Imanaka, MD, PhD, Department of Healthcare
racic aortic disease, paraplegia is especially important, as it
Economics and Quality Management, Graduate School of Medicine, Kyoto
University, Yoshida Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan. entails severe physical impairment and reduced survivability.2
E-mail address: imanaka-y@umin.net (Y. Imanaka). Furthermore, paraplegia due to spinal cord dysfunction can

https://doi.org/10.1053/j.jvca.2021.07.043
1053-0770/Ó 2021 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/)
1022 T. Umegaki et al. / Journal of Cardiothoracic and Vascular Anesthesia 36 (2022) 10211028

prolong hospitalization,3 thereby increasing resource con- aortic aneurysm (I711), descending thoracic aortic aneurysm
sumption and the risk of iatrogenic conditions. (I712), abdominal aortic aneurysm (I713, I714), thoracoabdo-
Despite the devastating physiologic and psychologic impact minal aortic aneurysm (I715, I716), aortic rupture (I718), and
of postoperative paraplegia on patients, there is a lack of evi- aortic aneurysm (I719).
dence on the differential risk of this complication between the
two major surgical options for thoracic aortic disease in Japan. Patient Characteristics
To address this gap in the evidence, the authors comparatively
assessed the postoperative paraplegia risk between open surgi- The authors collected information on the following baseline
cal repair and TEVAR in patients with thoracic aortic disease, patient characteristics: age, sex, height, weight, primary diag-
using a large-scale multicenter administrative claims database. nosis, activities of daily living (ADL) score at admission and
discharge, admission type (scheduled v emergency), and
Methods Charlson Comorbidity Index (CCI) score. The authors also
acquired information on surgical procedures, use of cerebro-
Study Design and Data Source spinal fluid drainage (preoperative or postoperative), and use
of motor-evoked potential monitoring during hospitalization.
The authors conducted a retrospective cohort study of ADL scoring was based on the following ten items (maximum
patients with thoracic aortic aneurysm or aortic dissection who score: 20 points): feeding (0-2 points), transferring (0-3 points),
had undergone open surgical descending aortic repair or grooming (0-1 point), toileting (0-2 points), bathing (0-1 point),
TEVAR at 139 Japanese acute- care hospitals between April 1, walking on level ground (0-3 points), climbing stairs (0-2
2010, and March 31, 2019. Most acute-care hospitals in Japan points), dressing (0-2 points), bowel continence (0-2 points), and
generate administrative data for insurance claims using the urinary continence (0-2 points). Although the CCI originally
Diagnosis Procedure Combination (DPC) patient case-mix developed in 1987 has been used widely to predict one-year mor-
system. For this study, data were acquired from a DPC data- tality,4 the authors used the modified version developed by Quan
base containing clinical information and administrative data et al., as it was designed specifically for use in administrative
routinely collected from participating hospitals by the Quality data.5 Body mass index was calculated using the following equa-
Indicator/Improvement Project (QIP). Using these data, the tion: body weight (kg) £ height2 (m). Each patient’s height and
QIP analyzes a variety of indicators (eg, management effi- weight also were used to calculate body surface area with the
ciency, healthcare processes, and patient outcomes) and DuBois formula: (weight in kg .0425 £ height in cm
presents the findings as feedback to the participating hospitals. 0.725) £ 0.007184.
This study database includes information such as patient dem-
ographics (eg, age, sex, height, and weight), diagnoses, comor- Outcome Measures
bidities upon and after admission, surgical procedures, disease
severity, and treatments. This study was approved by the insti- The primary outcome measure was postoperative paraple-
tutional ethics committee of Kansai Medical University Hospi- gia. First, paraplegia was identified using the following Inter-
tal (approval number: 2017024). national Classification of Diseases, 10th Revision codes:
flaccid paraplegia (G820), paraplegia (G822), tetraplegia
Patient Selection (G825), paraplegia of lower limbs (G831), transverse spinal
cord injury (G838), spinal cord ischemia (G951), and disease
The authors first identified patients who were hospitalized of spinal cord (G959). Postoperative cases of paraplegia were
for a thoracic aortic aneurysm or aortic dissection of the identified as the occurrence of any of the above conditions
descending aorta and had undergone isolated open surgical after admission (ie, during hospital stay). Other outcome meas-
repair or isolated TEVAR during the study period. The DPC ures included length of intensive care unit stay, overall hospital
system uses Japan-specific procedural codes (K-codes) to iden- stay, discharge destination (home, other hospital, nursing
tify all procedures performed during hospitalization. For this home, others, or death), and in-hospital mortality. The authors
study, the two target procedures were identified using the K- also analyzed the incidences of the following postoperative
codes K5604 (open surgical repair of descending thoracic complications: cerebral hemorrhage, cerebral infarction, cere-
aorta) and K5611 (thoracic endovascular stent grafting). The bral hypoxia, and recurrent nerve paralysis. All outcomes were
study patients were divided into two groups according to these compared between the two surgery groups.
procedures. The authors excluded from analysis patients who
were younger than 20 on admission and patients who had Statistical Analysis
undergone open surgical thoracoabdominal aortic repair, a
combination of open surgery and stent grafting of the thora- Continuous variables were expressed as means and standard
coabdominal aorta, or a combination of ascending/arch repairs deviations (SDs), and categorical variables were expressed as
and stent grafting. In addition, the authors also excluded percentages. For univariate comparisons between the surgery
patients without records of any of the following descending groups (open surgical repair v TEVAR), continuous variables
aortic diseases, as indicated by International Classification of were assessed using Student t test, and categorical variables
Diseases, 10th Revision codes: dissecting aorta (I710), thoracic were assessed using the chi-squared test or Fisher exact test, as
 T. Umegaki et al. / Journal of Cardiothoracic and Vascular Anesthesia 36 (2022) 10211028 1023

appropriate. From these univariate analyses, the authors identi- height, weight, body mass index, body surface area, and
fied the characteristics significantly associated with postopera- admission type. In contrast, no differences were observed in
tive paraplegia for inclusion as covariates in the subsequent sex, ADL score at admission or discharge, or CCI score. The
logistic regression models. The propensity score for TEVAR open surgical repair group had significantly higher use of pre-
was calculated using a multiple logistic regression model con- operative cerebrospinal fluid drainage (14.9% v 5.1%, p <
ditioned on age, sex, body surface area, admission type, ADL 0.001) and motor-evoked potential monitoring (12.1% v 0.3%,
score at admission, and CCI score. Using inverse probability p < 0.001) than the TEVAR group.
of treatment weighting (IPTW) by the propensity score, the Table 2 summarizes the use of postoperative care and
authors calculated the odds ratio (OR) and 95% confidence patient outcomes in each group. The duration of mechanical
interval (CI) of TEVAR for postoperative paraplegia relative ventilation was longer in the open surgical repair group (mean
to open surgical repair in a binary logistic regression analysis § SD: 5.3 § 10.1 days) than in the TEVAR group (1.2 § 4.9
with a generalized estimating equation approach. Due to the days). Similarly, intensive care unit stay was longer in the
disparity in sample size between the two groups, the use of open surgical repair group (mean § SD: 6.3 § 5.2 days) than
propensity-score matching necessitated the exclusion of a in the TEVAR group (2.9 § 3.7 days), and the overall hospital
large proportion of the open surgical repair group. Accord- stay was longer in the open surgical repair group (38.8 § 38.1
ingly, IPTW was employed to allow the maximum use of all days) than in the TEVAR group (24.6 § 42.0 days). In-hospital
available data. mortality also was significantly higher in the open surgical
The authors also calculated the ORs and 95% CIs from an repair group (11.5%) than in the TEVAR group (5.0%).
instrumental variable analysis in which the instrumental vari- Table 2 also shows the incidence of postoperative complica-
able was the ratio of TEVAR use-to-open surgical repair use. tions in each group. The open surgical repair group had signifi-
Postoperative paraplegia again was used as the dependent vari- cantly higher incidences of cerebral hypoxia (p < 0.01) and
able, and TEVAR (reference: open surgical repair) was the recurrent nerve paralysis (p < 0.001) than the TEVAR group. In
independent variable of interest. Other patient characteristics contrast, the incidence of postoperative paraplegia was similar
were included as covariates in the instrumental variable analy- between the groups (open surgical repair group: 3.0%; TEVAR
sis if they had p values below 0.1 in the univariate analyses. P group: 2.3%; p = 0.36). The incidence of postoperative paraplegia
values below 0.05 (2-tailed) were regarded as statistically sig- also was similar between the groups among the survivors (open
nificant. All analyses were performed using SPSS version 26.0 surgical repair group: 3.0%; TEVAR group: 2.3%; p = 0.31) and
(IBM Japan, Ltd, Tokyo, Japan). nonsurvivors (open surgical repair group: 1.7%; TEVAR group:
2.3%; p = 0.62). The incidence of mortality or paraplegia was sig-
Results nificantly higher in the open surgical repair group than in the
TEVAR group (open surgical repair group: 14.2%; TEVAR
The patient selection process is shown in Figure 1. First, the group: 7.5%; p < 0.001). Among patients who developed postop-
authors identified 7,288 candidate patients with thoracic aortic erative paraplegia, the incidence of mortality was 6.7% in the
aneurysm or aortic dissection of the descending aorta. From open surgical repair group and 5.3% in the TEVAR group
these, the authors excluded patients who had undergone open (p = 0.58). Among patients with preoperative cerebrospinal drain-
surgical thoracoabdominal aortic repair (n = 673), a combina- age, the incidence of paraplegia was similar between the groups
tion of open surgery and stent grafting of the thoracoabdomi- (open surgical repair group: 6.5%; TEVAR group: 4.8%;
nal aorta (n = 129), and a combination of ascending/arch p = 0.36). The most common postoperative complication in both
repairs and stent grafting (n = 131). The remaining 6,355 groups was cerebral infarction (open surgical repair group: 7.4%;
patients were divided into an open surgical repair group TEVAR group: 6.4%; p = 0.39).
(n = 523) and a TEVAR group (n = 5,832). The following then The results of the logistic regression analyses using IPTW
were excluded from each group: patients without any records and the instrumental variable are presented in Table 3. TEVAR
of descending aortic diseases (18 patients in the open surgical did not show a significantly lower postoperative paraplegia
repair group and 130 patients in the TEVAR group) and risk than open surgical repair in the IPTW model (OR: 0.81;
patients who were younger than 20 years on admission (two 95% CI: 0.42-1.59; p = 0.55) or the instrumental variable
patients in the open surgical repair group and three patients in model (OR: 0.88; 95% CI: 0.42-1.86; p = 0.75).
the TEVAR group). The final analysis was performed using
503 patients in the open surgical repair group and 5,699 Discussion
patients in the TEVAR group.
The patient characteristics are presented in Table 1. The This large-scale multicenter cohort study compared the risk
study sample included 6,202 patients from 139 QIP participat- of postoperative paraplegia between two major surgical inter-
ing hospitals. The open surgical repair group and TEVAR ventions for thoracic aortic disease, using administrative data.
group comprised 503 patients from 94 hospitals and 5,699 The analysis did not detect any significant difference in the
patients from 129 hospitals, respectively. The mean patient risk of postoperative paraplegia between open surgical repair
ages were 65.0 years in the open surgical repair group and and TEVAR in Japanese hospitals.
73.0 years in the TEVAR group (p < 0.001). The univariate Paraplegia is a devastating complication following the sur-
analyses also found significant intergroup differences in gical repair of thoracic aortic disease. A previous multicenter
1024 T. Umegaki et al. / Journal of Cardiothoracic and Vascular Anesthesia 36 (2022) 10211028

Fig 1. Flowchart of patient selection. ICD-10, International Classification of Diseases, 10th Revision.

retrospective observational study in Japan found that paraple- occurred in 3.4% of open surgical repair patients and 2.3% of
gia occurred in 6.7% of 75 patients who underwent open surgi- TEVAR patients with type B aortic dissection.9 It is possible
cal repair of a descending thoracic aortic aneurysm between that the study patients included cases with permanent and tem-
2001 and 2002.6 In the United States, a single-center retrospec- porary paraplegia. However, recent studies have reported
tive observational study reported that paraplegia occurred in TEVAR to have lower paraplegia risks than open surgical
2.3% of 246 patients who underwent TEVAR of the descend- repair.10,11 The authors posit that the gradual reduction in post-
ing thoracic aorta between 2000 and 2008.7 Another single- operative neurologic complications may be due to improve-
center retrospective observational study in the United States ments in perioperative patient management and surgical
reported that paraplegia/paraparesis occurred in 8.3% of 2,012 procedures. Further large-scale prospective studies would
patients who underwent open descending and thoracoabdomi- broaden understanding of the occurrence of paraplegia follow-
nal aortic aneurysm repair between 1991 and 2012.8 In the ing open surgical repair.
study population, which did not include abdominal or thora- The pathophysiology of postoperative paraplegia due to spi-
coabdominal aortic aneurysms, paraplegia occurred in less nal cord injury is multifactorial and complex. Table 4 summa-
than 3% of both surgery groups. This was similar to the find- rizes the risk of spinal cord injury as described by three
ings of Lou et al., which reported that postoperative paraplegia previous studies.12-14 Spinal cord injury after TEVAR
 T. Umegaki et al. / Journal of Cardiothoracic and Vascular Anesthesia 36 (2022) 10211028 1025

Table 1 analyze age, emergency surgery, and CCI score (including

Patient Characteristics According to Surgery Group (n = 6,202) chronic pulmonary disease) as covariates, and could not exam-
Variables Open Surgical TEVAR p Value
ine these other variables. Additional studies, therefore, are
Repair (n = 503) (n = 5,699) needed to incorporate these other potential risk factors in their
analytical models to verify the results.
Number of 94 129 The authors observed several notable differences in postop-
hospitals
Patient
erative complications and outcomes between the surgery
characteristics groups. For example, the incidence of recurrent nerve paralysis
Age (y) 65.0 § 12.6 73.0 § 10.6 <0.001 in open surgical repair patients (2.9%) was six times that of
Male (%) 76.7 74.2 0.22 TEVAR patients (0.5%). Recurrent nerve paralysis can
Height (cm) 164.3 § 8.6 161.2 § 9.3 <0.001 weaken the cough reflex and prevent expectoration, thereby
Weight (kg) 64.6 § 13.9 60.5 § 12.6 <0.001
Body mass index 23.8 § 4.1 23.2 § 3.8 <0.001
increasing the risk of postoperative pulmonary complications
(kg/m2) such as pneumonia and atelectasis.15-17 As there is little infor-
Body surface area 1.7 § 0.2 1.6 § 0.2 <0.001 mation in the existing literature on recurrent nerve paralysis
(m2) after open surgical repair and TEVAR, there is a need for
ADL score at 15.8 § 7.7 15.8 § 7.4 0.98 research to explore this complication in greater detail. Next,
admission
ADL score at 17.4 § 5.8 17.6 § 5.2 0.49
in-hospital mortality in the open surgical repair patients
discharge (11.5%) was approximately twice that of TEVAR patients
Admission type (5.0%). A recent meta-analysis of type B aortic dissections
(%) found that TEVAR patients had better outcomes (in-hospital
Scheduled 67.4 73.5 <0.01 mortality, 30-day mortality, and long-term survival) than open
Emergency 32.6 26.5
Charlson
surgical repair patients.18 TEVAR also reportedly is associated
Comorbidity I with a lower mortality rate than open surgical repair for rup-
score (%) tured descending thoracic aneurysms.19 Similarly, a compari-
0 - - 0.07 son of TEVAR and open surgical repair for intact descending
1 38.0 38.5 thoracic aortic aneurysms found that the former had a signifi-
2 36.0 31.0
3 16.1 18.4
cantly lower 180-day mortality rate.20 The findings that in-hos-
4 10.0 12.2 pital mortality was lower in TEVAR patients were concordant
Aneurysm or with these previous studies. In contrast, Chiu et al. reported
dissection that there was a small but nonsignificant difference in long-
Aneurysm 55.3 71.2 <0.001 term survival between these surgical interventions, suggesting
Dissection 44.7 28.8
Cerebrospinal
that these survival rates gradually may converge over time.20
fluid drainage The results indicated that open surgical repair patients had
(%) spinal draining and motor-evoked potential monitoring more
Preoperative 14.9 5.1 <0.001 frequently than TEVAR patients. Few prospective studies
Postoperative 1.4 1.2 have examined the preventive effects of spinal drainage use on
Motor-evoked 12.1 0.3 <0.001
potential
spinal cord ischemia. In a study on TEVAR patients, Hnath
monitoring (%) et al. found that spinal cord ischemia occurred in 8% of
patients without cerebrospinal fluid drainage, whereas none of
Values are presented as mean § SD for continuous variables and percentages the patients with cerebrospinal fluid drainage experienced spi-
for categorical variables.
Abbreviations: ADL, activities of daily living; TEVAR, thoracic endovascular
nal cord ischemia.21 In a study of 155 TEVAR patients with
aortic repair. more than 12 cm of aortic coverage who underwent spinal fluid
drainage and had received a stent in the zone-3 region, only
one patient developed spinal cord injury.22 However, Hanna
reportedly is associated with prolonged intraoperative hypo- et al. reported that minor complications (such as spinal head-
tension and simultaneous closure of at least two independent ache, puncture-site bleeding, persistent cerebrospinal fluid
spinal cord vascular territories.12 The following risk factors leak, and subdural hematoma) occurred in 11.1% of 81
for spinal cord ischemia after TEVAR also have been identi- patients who received a preoperative lumbar drain.23 Spinal
fied: aneurysm rupture, shaggy aorta, chronic pulmonary dis- fluid drainage also has been linked with other complications
ease, use of a single-stage procedure, coverage of more than such as local infection, meningitis, and intracranial hemor-
nine segments, coverage of aortic segments that align with tho- rhage,24 demonstrating that this procedure is not without its
racic vertebral levels Th8 to Th12, minimum postoperative risks. The authors were able to identify the use and timing of
hemoglobin level, and number of postoperative patent segmen- cerebrospinal drainage in the patients, but the database did not
tal arteries.13 For open surgical repair, the extent of graft include records of blood pressure and cerebrospinal fluid pres-
replacement (including the distal part below Th9) and nondis- sure measurements. Permissive hypertension and collateral
secting aneurysms has been identified as determinants of post- revascularization may contribute to reducing the use of cere-
operative paraplegia.14 The present study only was able to brospinal drainage,25 and the high incidence of paraplegia in
1026 T. Umegaki et al. / Journal of Cardiothoracic and Vascular Anesthesia 36 (2022) 10211028

Table 2 Table 3
Use of Postoperative Care and Outcomes According to Surgery Group Results of Multiple Logistic Regression Analyses of Paraplegia Using Inverse
(n = 6,202) Probability of Treatment Weighting and an Instrumental Variable (n = 6,202)

Variables Open surgical TEVAR p Value Variables Odds Ratio 95% CI p Value
repair (n = 503) (n = 5,699)
IPTW
Postoperative care TEVAR (reference: open 0.81 0.42-1.59 0.55
Mechanical ventilation (d) 5.3 § 10.1 1.2 § 4.9 <0.001 surgical repair)
Renal replacement therapy (%) 8.2 3.2 <0.001 Dissection (reference: 0.59 0.31-1.13 0.11
Percutaneous cardiopulmonary 3.2 0.6 <0.001 aneurysm)
support (%) Motor-evoked potential 1.39 0.36-5.34 0.64
Outcomes monitoring
Paraplegia (%) 3.0 2.3 0.36 Preoperative cerebrospinal 2.41 0.99-5.80 0.05
Cerebral hemorrhage (%) 0.4 0.2 0.46 fluid drainage
Cerebral infarction (%) 7.4 6.4 0.39 Instrumental variable
Cerebral hypoxia (%) 1.0 0.2 <0.01 TEVAR (reference: open 0.88 0.42-1.86 0.75
Recurrent nerve paralysis (%) 3.0 0.5 <0.001 surgical repair)
ICU stay (days) 6.3 § 5.2 2.9 § 3.7 <0.001 Dissection (reference: 0.89 0.59-1.34 0.56
Overall hospital stay (days) 38.8 § 38.1 24.6 § 42.0 <0.001 aneurysm)
Discharge destination (%) Motor-evoked potential 1.16 0.33-4.09 0.82
Home 72.2 81.6 <0.001 monitoring
Other hospital 14.9 12.3 Preoperative cerebrospinal 2.51 1.49-4.25 <0.001
Nursing home 1.2 0.8 fluid drainage
Other destination 0.2 0.2 Emergency admission 1.58 1.09-2.30 0.02
Death 11.5 5.0 (reference: scheduled)
In-hospital mortality (%) 11.5 5.0 <0.001 Age 1.01 1.00-1.03 0.22
Body surface area 1.27 0.48-3.37 0.64
Values are presented as mean § SD for continuous variables and percentages Charlson Comorbidity Index 1.09 0.96-1.25 0.20
for categorical variables. score
Abbreviations: ICU, intensive care unit; TEVAR, thoracic endovascular aortic
repairs. Abbreviations: CI, confidence interval; IPTW, inverse probability of treatment
weighting; TEVAR, thoracic endovascular aortic repair.

patients after preoperative cerebrospinal drainage suggests that

patients with a high predicted risk of paraplegia tended to Although the TEVAR patients generally were older (mean age:
undergo cerebrospinal drainage. The sample included a com- 73.0 years) than the open surgical repair patients (65.0 years), the
paratively small number of patients with preoperative cerebro- groups showed similar ADL and CCI scores. The authors were
spinal drainage, which may indicate that physicians are also unable to ascertain the details of paraplegia (eg, specific
avoiding the use of cerebrospinal drainage in consideration of physical symptoms) or identify whether a specialist neurologist
severe neurologic and hemorrhagic complications. Next, a was involved in the clinical assessment of each patient with para-
report found that motor-evoked potential monitoring did not plegia. Accordingly, the authors did not include patients with par-
contribute to improving motor impairment in patients who had aparesis in the analysis, as these can have a more ambiguous
undergone descending and thoracoabdominal aortic repair.26 presentation. By excluding paraparesis patients with paralytic
However, the evidence for the usefulness of motor-evoked symptoms occurring from cerebral infarctions, this study focused
potential monitoring remains controversial, and some advocate on patients with spinal cord infarctions. In addition, the authors
this practice for spinal cord ischemia and injury.27,28 could not determine whether paraplegia was disease-related or
The Japanese Circulation Society’s Guidelines for Diagno- iatrogenic. Second, this study lacked surgical data such as blood
sis and Treatment of Aortic Aneurysm and Aortic Dissection loss, anesthesia time, cardiopulmonary bypass time, and operative
recommend TEVAR for patients with traumatic aortic injury, time. However, this study was not conducted to examine detailed
patients with complicated acute type B aortic dissection, and surgical factors, but was instead designed to examine the two sur-
situations in which there is available support from a surgical gical interventions as overall processes. Therefore, the use of sur-
team.29 Although the use of TEVAR has increased greatly, gery type as the variable of interest was suitable for the study
there are still hospitals that only can perform open surgical aim. Third, the administrative data lacked information on diag-
repairs due to inadequate manpower or expertise. However, nostic imaging, such as computed tomography, magnetic reso-
the data did not include the number of cardiac surgeons or their nance imaging, roentgenography, and echocardiography.
years of experience in each hospital, which may influence a Therefore, the authors did not have access to information on the
hospital’s capability to perform TEVAR. range of dissection and whether an aneurysm was infrarenal or
This study had the following limitations. First, patient-level suprarenal. In addition, although most paraplegia cases are likely
physiologic data (eg, leukocyte count, erythrocyte count, platelet to be due to spinal cord infarctions, a portion of these cases may
count, and hemoglobin levels) were not available in the DPC be due to other reasons such as spinal cord hemorrhage. Also, the
database, and the authors were unable to determine if there were authors could not exclude the possibility of other risk factors of
preoperative physiologic differences between the groups. paraplegia (eg, medical and surgical history) that were not
 T. Umegaki et al. / Journal of Cardiothoracic and Vascular Anesthesia 36 (2022) 10211028 1027

Table 4 3 New PW, Reeves RK, Smith E,  et al. International retrospective comparison
Risk of Spinal Cord Injury After Surgery of inpatient rehabilitation for patients with spinal cord dysfunction epidemiol-
ogy and clinical outcomes. Arch Phys Med Rehabil 2015;96:1080–7.
Open Surgical Repair TEVAR 4 Summary of Static/Dynamic Survey of Medical Institutions and Hospital
Report. Ministry of Health. Labour and Welfare; 2014.
Czerny et al.12 Prolonged intraoperative hypotension Simultaneous 5 Summary of Static/Dynamic Survey of Medical Institutions and Hospital
closure of at least two independent spinal cord Report. Ministry of Health. Labour and Welfare; 2008.
vascular territories 6 Yamauchi T, Takano H, Nishimura M, et al. Paraplegia and paraparesis
Hiraoka et al.13 Aneurysm rupture Shaggy aorta Chronic obstructive after descending thoracic aortic aneurysm repair: A risk factor analysis.
pulmonary disease Single-stage procedure Ann Thorac Cardiovasc Surg 2006;12:179–83.
Coverage of more than nine segments Coverage of 7 Preventza O, Wheatley GH 3rd, Williams J, et al. Identifying paraplegia
aortic segments that align with thoracic vertebral risk associated with thoracic endografting. Asian Cardiovasc Thorac Ann
levels Th8 to Th12 Minimum postoperative 2009;17:568–72.
hemoglobin level Number of postoperative patent 8 Tanaka A, Leonard SD, Sandhu HK, et al. Open descending and thora-
segmental arteries coabdominal aortic repairs in patients younger than 50 years old. Ann
Yamauchi et al.14 Extent of graft replacement including the distal part Thorac Surg 2019;108:693–9.
(below Th9) Nondissecting aortic aneurysm 9 Lou X, Chen EP, Duwayri YM, et al. The impact of thoracic endovascular
aortic repair on long-term survival in type B aortic dissection. Ann Thorac
Abbreviation: TEVAR, thoracic endovascular aortic repair.
Surg 2018;105:31–8.
10 Abraha I, Romagnoli C, Montedori A, et al. Thoracic stent graft versus sur-
gery for thoracic aneurysm. Cochrane Database Syst Rev 2016;6:
included in the covariates. Finally, open surgical repair often is
CD006796.
performed when a patient’s anatomy is not favorable for endovas- 11 Harky A, Kai Chan JS, Ming Wong CH, et al. Open versus endovascular
cular repair, but the authors could not account for this factor in repair of descending thoracic aortic aneurysm disease: A systematic review
the comparison of the two surgery groups. and meta-analysis. Ann Vasc Surg 2019;54:304–15;e5.
12 Czerny M, Eggebrecht H, Sodeck G, et al. Mechanisms of symptomatic
spinal cord ischemia after TEVAR: Insights from the European Registry of
Conclusions Endovascular Aortic Repair Complications (EuREC). J Endovasc Ther
2012;19:37–43.
This study was, to the authors’ knowledge, the first to 13 Hiraoka T, Komiya T, Tsuneyoshi H, et al. Risk factors for spinal cord
ischaemia after thoracic endovascular aortic repair. Interact Cardiovasc
observe that there are no significant differences in postopera- Thorac Surg 2018;27:54–9.
tive paraplegia risk between open surgical repair and TEVAR 14 Yamauchi T, Takano H, Nishimura M, et al. Paraplegia and paraparesis
in patients with thoracic aortic disease. Despite expectations after descending thoracic aortic aneurysm repair: A risk factor analysis.
that open surgical repair would be associated with a higher Ann Thorac Cardiovasc Surg 2006;12:179–83.
15 Hamdan AL, Moukarbel RV, Farhat F, et al. Vocal cord paralysis after
paraplegia risk, improvements in perioperative management
open-heart surgery. Eur J Cardiothorac Surg 2002;21:671–4.
may have contributed to the similarly low incidence of postop- 16 Hulscher JBF, van Sandick JW, Devriese PP, et al. Vocal cord paralysis
erative paraplegia in these two surgery types. after subtotal oesophagectomy. Br J Surg 1999;86:1583–7.
17 Filaire M, Mom T, Laurent S, et al. Vocal cord dysfunction after left lung
resection for cancer. Eur J Cardiothorac Surg 2001;20:705–11.
Acknowledgments 18 Li FR, Wu X, Yuan J, et al. Comparison of thoracic endovascular aortic
repair, open surgery and best medical treatment for type B aortic dissec-
This study was supported in part by a Grant-in-Aid for Sci- tion: A meta-analysis. Int J Cardiol 2018;250:240–6.
entific Research from the Japan Society for the Promotion of 19 Jonker FHW, Verhagen HJM, Lin PH, et al. Open surgery versus endovas-
cular repair of ruptured thoracic aortic aneurysms. J Vasc Surg
Science (Grant number: [A] 19H01075). The funders were not 2011;53:1210–6.
involved in study design, data collection, data analysis, manu- 20 Chiu P, Goldstone AB, Schaffer JM, et al. Endovascular versus open repair
script preparation, or the decision to publish. The authors had of intact descending thoracic aortic aneurysms. J Am Coll Cardiol
complete access to the study data. 2019;73:643–51.
21 Hnath JC, Mehta M, Taggert JB, et al. Strategies to improve spinal cord
ischemia in endovascular thoracic aortic repair: Outcomes of a prospective
Conflict of Interest cerebrospinal fluid drainage protocol. J Vasc Surg 2008;48:836–40.
22 Acher C, Acher CW, Marks E, et al. Intraoperative neuroprotective inter-
ventions prevent spinal cord ischemia and injury in thoracic endovascular
The authors report no financial relationships or conflicts of aortic repair. J Vasc Surg 2016;63:1458–65.
interest regarding the content herein. 23 Hanna JM, Andersen ND, Aziz H, et al. Results with selective preoperative
lumbar drain placement for thoracic endovascular aortic repair. Ann
Thorac Surg 2013;95:1968–74.
References 24 Fedorow CA, Moon MC, Mutch WAC, et al. Lumber cerebrospinal fluid
drainage for thoracoabdominal aortic surgery: Rationale and practical con-
1 Committee for Scientific Affairs, The Japanese Association for Thoracic siderations for management. Anesth Analg 2010;111:46–58.
Surgery, Shimizu H, Okada M, et al. Thoracic and cardiovascular surgeries 25 Weissler EH, Voigt SL, Raman V, et al. Permissive hypertension and collat-
in Japan during 2017: Annual report by the Japanese Association for Tho- eral revascularization may allow avoidance of cerebrospinal fluid drainage in
racic Surgery. Gen Thorac Cardiovasc Surg 2020;68:414–49. thoracic endovascular aortic repair. Ann Thorac Surg 2020;110:1469–75.
2 Bicknell CD, Riga CV, Wolfe JHN. Prevention of paraplegia during thora- 26 Yoshitani K, Masui K, Kawaguchi M, et al. Clinical utility of intraopera-
coabdominal aortic aneurysm repair. Eur J Vasc Endovasc Surg tive motor-evoked potential monitoring to prevent postoperative spinal
2009;37:654–60. cord injury in thoracic and thoracoabdominal aneurysm repair: An audit of
1028 T. Umegaki et al. / Journal of Cardiothoracic and Vascular Anesthesia 36 (2022) 10211028

the Japanese Association of Spinal Cord Protection in aortic surgery data- 28 Jellish WS. Motor-evoked potentials are an important determinant of spi-
base. Anesth Analg 2018;126:763–8. nal cord ischemic injury during aortic arch surgery. But can they be used
27 Hattori K, Yoshitani K, Kato S, et al. Association between motor-evoked exclusively? J Cardiothorac Vasc Anesth 2019;33:1843–4.
potentials and spinal cord damage diagnosed with magnetic resonance 29 Group JCS Joint Working. Guidelines for diagnosis and treatment of aortic
imaging after thoracoabdominal and descending aortic aneurysm repair. J aneurysm and aortic dissection (JCS 2011): Digest version. Circ J
Cardiothorac Vasc Anesth 2019;33:1835–42. 2013;77:789–828.