Abstract
Objective
During open descending thoracic and thoracoabdominal aortic aneurysm (DTAA/TAAA) repair, we used a routine T8-T12 intercostal artery (ICA) reattachment strategy from July 2004 to June 2009 and after 2017, we used a selective ICA reattachment strategy (reattaching T8-T12 ICAs only when neuromonitor signals were lost) from July 2009 to 2016. This study reviewed our nearly 2-decade experience to assess the impact of 2 ICA reattachment strategies on spinal cord injury (SCI).
Methods
All open DTAA/TAAA repairs performed from July 2004 to June 2022 were included, except for cases without intraoperative cerebral spinal fluid drainage. Perioperative data were reviewed. Univariable and multivariable analyses and propensity matching for risk-adjusted effects of 2 strategies for ICA reattachment on SCI were used.
Results
In all, 375 patients were operated on with selective strategy and 584 with routine strategy. Age and prevalence of rupture and redo were similar in the 2 groups. The rate of operative mortality and immediate SCI was also similar (selective vs routine: mortality, 12.5% vs 12.3%; immediate SCI, 3.2% vs 2.2%). However, the incidence of delayed and permanent SCI was increased in the selective group (delayed, 10.4% vs 6.9%; permanent, 8.5% vs 5.3%). Multivariable analyses demonstrated selective strategy was a predictor of delayed and permanent SCI, along with TAAA extent II/III, and older age.
Conclusions
Two strategies of ICA reattachment did not impact the incidence of immediate SCI, which was infrequent, but the selective strategy was associated with greater rates of delayed permanent SCI. Reattachment of the ICAs within T8-T12 should be performed during open DTAA/TAAA.
Key Words: intercostal reattachment, cerebral spinal fluid drainage, descending thoracic aortic aneurysm, thoracoabdominal aortic aneurysm, spinal cord ischemia
Graphical Abstract
Summary of open descending and thoracoabdominal aortic repair techniques.
Central Message.
Aggressive intercostal artery reattachment reduced the incidence of delayed and permanent spinal cord injury after descending thoracic and thoracoabdominal aortic aneurysm repair.
Perspective.
Reattachment of the intercostal arteries within the T8-T12 level is recommended in patients undergoing descending thoracic and thoracoabdominal aortic aneurysm repair to spinal cord injury.
Spinal cord injury (SCI) is a life-disabling complication after the repair of descending thoracic and thoracoabdominal aortic aneurysms (DTAAs/TAAAs). SCI is seen in 5% to 10% of patients undergoing DTAA/TAAA repair1 and can occur immediately or in a delayed manner. Technical considerations and perioperative management protocols have been developed to prevent this devastating complication. Hypothermia, distal aortic perfusion, and cerebral spinal fluid drainage (CSFD) have become the mainstay of the intraoperative preventative measures for SCI during open distal aortic repair.2,3 The artery of Adamkiewicz, the dominant blood supply to the thoracic spinal cord, is often located in the T8-L1 level.4 Since the late 1990s, our group has been an advocate of routine reimplantation of patent T8-T12 intercostal arteries (ICAs) to prevent SCIs in distal aortic repair.5 We began using intraoperative somatosensory-evoked potentials (SSEPs) to monitor spinal cord function in mid-2001, and motor-evoked potentials (MEPs) were added to the neuromonitoring protocol in 2004. We found that intact SSEP and MEP signals had a strong negative predictive value of immediate SCI after DTAA/TAAA repairs.6 In 2009, we adopted selective reattachment of ICAs only when SSEP/MEP signals were lost during aortic repair, as we gained more confidence with the predictability of SCI with MEP/SSEP.2 However, we reviewed our data again in 2016, which was concerning for an increased incidence of SCIs with the selective ICA reattachment strategy.7 Thus, we switched back to our ICA management to routine reattachment in 2017. This study reviewed our nearly 2-decade experience to assess the impact of 2 ICA-reattachment strategies, routine versus selective, on SCI after open DTAA/TAAA repairs.
Methods
The Committee for Protection of Human Subjects, the local institutional review board, approved this study (approval number: HSC-MS-03-077, October 10, 2014). All open DTAA/TAAA repairs performed at our institution after routine MEP and SSEP use were included. Repairs performed without intraoperative CSFD were excluded, as it is the mainstay of our SCI prevention protocol and is only not used when contraindicated by infection, hostile lumbar spine or hemodynamic emergency. The study period was from July 2004 to June 2022. We used routine T8-T12 ICA reattachment strategy from July 2004 to June 2009, selective strategy (reattaching T8-T12 ICAs only when MEP/SSEP signals were lost) from July 2009 to December 2016, and routine strategy again from January 2017 to the present. The patients were divided for analysis into 2 groups, routine and selective, according to the used ICA strategies. Subanalysis was performed for era 1 to 3 (era 1, July 2004-June 2009; era 2, July 2009-December 2016, era 3 January 2017-June 2022) to address the heterogeneity across the interrupted time series. Indications for surgical interventions were aneurysm size exceeding 5 cm, aneurysm with pain, and rapid enlargement (>5 mm/year). Our surgical techniques for distal aortic repair have been described previously.2 Perioperative data were extracted from the prospective registry of our department. Paraplegia and paraparesis are combined and reported as SCI in this study.
Definitions
Paraplegia was defined as a neurologic deficit with the modified Tarlov score of 0-2, and paraparesis was described as a neurologic deficit with the modified Tarlov score of 3-4. Postoperative immediate SCI was defined as paraplegia or paraparesis observed upon awakening of the patient from anesthesia. Delayed SCI was defined as new-onset postoperative paraplegia or paraparesis that occurred after a period of an intact neurologic examination. Operative mortality included deaths within 30 days and any in-hospital death.
Operative Techniques
Our open surgical techniques of distal aortic repair are described elsewhere.2 To summarize, we expose the thoracic and thoracoabdominal aortic aneurysm through a modified thoracoabdominal incision. We routinely use the adjunct of distal aortic perfusion, moderate hypothermia, and CSFD for spinal cord protection. Intraoperative monitoring includes a Swan-Ganz catheter (Edwards Lifesciences), near-infrared spectroscopy, arterial line, and MEP/SSEP neuromonitoring. Left heart bypass is established using the left inferior pulmonary vein drainage and the left common femoral artery return. Distal aortic perfusion pressure is maintained above 60 mm Hg. CSFD is drained to keep intracranial pressure between 10 and 15 mm Hg. In the early 2000s, the anastomosis was performed cephalad to caudad: proximal aortic anastomosis; ICA reattachment; visceral/renal branch reattachment; and distal aortic anastomosis visceral/renal branch reattachment. The conduct of order shifted to proximal anastomosis; distal anastomosis; visceral/renal reattachment; and ICA reattachment for the following reasons: to reduce distal aortic perfusion time, which can be embolizing atheromatous plaque thrombus to the viscera/renal branches in retrograde fashion; establish pulsatile flow to pelvis sooner; and shorten the pump time. Patent ICAs within T8-T12 levels were temporarily occluded with 3-Fr Fogarty balloon catheters until reattachment. During the selective reattachment era, we ligated the T8 to T12 ICAs if MEP and SSEP signals remained intact. Regardless of the eras, once MEP and SSEP signals were lost, measures to increase spinal cord blood perfusion were initiated, including optimizing hematocrit, CSFD to maintain intracranial pressure less than 10 mm Hg, increasing distal aortic perfusion pressure greater than 80 mm Hg, re-establishing pulsatile flow to the pelvis, and expedited reattachment of ICAs within T8-T12. Reattachment techniques were mostly island reconstruction with end-to-end anastomosis fashion to the main body of the graft during the earlier study period (up to 2010), and loop reconstruction or separate bypass were used in the more recent era.
Statistical Methods
The routine to selective and back to routine ICA management policy periods produced a natural interrupted time series. Data across the 3 periods were analyzed and are presented in Table E1. Neurologic outcomes did not differ between the 2 routine time periods, and these were combined for analysis with propensity matching across the cohort to account for risk factor drift. MEP signal changes were included in the analysis but not SSEP to simplify the analysis because we have found that MEP is a more sensitive measure to detect early ischemic changes whereas SSEP is more valuable to detect advanced stages of spinal cord ischemia.8 Descriptive statistics were computed using standard univariate measures of frequency and central tendency. Student t test or Wilcoxon rank-sum test were used for hypothesis tests, depending on distributional assumptions. Normality was assessed by visual inspection of variable distributions and by very small P values for tests of heterogeneity of variances, and non-normal data were not transformed for propensity score modeling. Dichotomous bivariate measures of association were computed by contingency table methods by χ2 or Fisher exact test and with odds ratios (ORs) and test-based 95% confidence intervals. Propensity scores were computed using logistic regression with ICA management category as the dependent variable. Candidate-independent variables were screened by Spearman correlation and significant preoperative variables, as well as purposefully selected variables known to be associated with outcomes, were evaluated for model inclusion. Eligibility for match was constrained to the region of common support of the propensity score, such that patients matched into the analysis could have received either ICA strategy using greedy matching. Propensity-matched standardized mean difference <0.25 was considered negligible/good match. Standard logistic regression diagnostics and standardized difference plots were used to assess model specification. Hypothesis tests were considered statistically significant at 2-sided alpha <0.05. All data were analyzed using SAS software, version 9.4 (SAS Institute Inc).
Results
A total of 1101 operations were performed, with 142 excluded for CSF drain complications; 959 patients comprised the study population, including 375 operated on during the selective ICA reattachment strategy era and 584 during the routine ICA reattachment strategy era. Eleven patients were missing values needed to compute glomerular filtration rate (eg, weight, creatinine) and, hence, were not imputed for analysis. Preoperative patient demographics are summarized in Table 1. The median age was similar in the 2 groups (63 vs 63 years, P = .868). Aortic dissection, peripheral arterial disease, and end-stage renal disease on dialysis were more common in the selective era (selective vs routine: dissection, 56.2% vs 47.7%, P = .010; peripheral arterial disease, 25.6% vs 18.4%, P = .009; dialysis, 7.4% vs 3.7%, P = .017). The prevalence of rupture and redo cases were similar in the 2 eras (selective vs routine: rupture 9.1% vs 7.7%, P = .472; redo 26.1% vs 26.4%, P = .935).
Table 1.
Preoperative patient characteristics
| Patient variable | Overall N = 959 | Routine n = 584 | Selective n = 375 | P value |
|---|---|---|---|---|
| Age, y | 63 (43-59) | 63 (55-71) | 63 (52-73) | .868 |
| Female | 352 (36.7%) | 212 (36.3%) | 140 (37.3%) | .783 |
| Hypertension | 897 (93.5%) | 541 (92.5%) | 356 (94.9%) | .179 |
| Diabetes mellitus | 125 (13.0%) | 68 (11.6%) | 57 (15.2%) | .111 |
| COPD | 396 (41.2%) | 248 (42.5%) | 148 (39.5%) | .357 |
| Peripheral arterial disease | 204 (21.3%) | 108 (18.4%) | 96 (25.6%) | .009 |
| Coronary artery disease | 243 (25.3%) | 148 (25.3%) | 95 (25.3%) | .998 |
| History of stroke | 80 (8.3%) | 43 (7.3%) | 37 (9.8%) | .189 |
| eGFR <45 mL/min/1.83 m2 | 150 (15.8%)∗ | 63 (17.2%) | 87 (15.1%) | .386 |
| ESRD on dialysis | 50 (5.2%)∗ | 22 (3.7%) | 28 (7.4%) | .017 |
| HAD/CTD | 236 (24.6%) | 125 (21.4%) | 111 (29.6%) | .004 |
| Aortic dissection | 490 (51.0%) | 279 (47.7%) | 211 (56.2%) | .010 |
| Redo | 252 (26.2%) | 154 (26.4%) | 98 (26.1%) | .935 |
| Rupture | 79 (8.2%) | 45 (7.7%) | 34 (9.1%) | .472 |
| Emergent/urgent | 75 (7.8%) | 50 (8.6%) | 25 (6.6%) | .325 |
| Extent of aneurysm | ||||
| DTAA | 359 (37.4%) | 225 (38.5%) | 134 (35.7%) | .383 |
| TAAA extent I | 122 (12.7%) | 76 (13.0%) | 46 (12.3%) | .735 |
| TAAA extent II | 121 (12.6%) | 78 (13.4%) | 43 (11.5%) | .390 |
| TAAA extent III | 132 (13.8%) | 77 (13.2%) | 55 (14.7%) | .516 |
| TAAA extent IV | 162 (16.9%) | 92 (15.8%) | 70 (18.7%) | .240 |
| TAAA extent V | 63 (6.6%) | 36 (6.2%) | 27 (7.2%) | .594 |
Continuous variables are expressed as median (interquartile range), and categorical variables are expressed as number (%). COPD, Chronic obstructive pulmonary disease; eGFR, estimated glomerular filtration rate; ESRD, end-stage renal disease; HAD, hereditary aortic disease; CTD, connective tissue disorder; DTAA, descending thoracic aortic aneurysm; TAAA, thoracoabdominal aortic aneurysm.
Data missing in 8 patients in selective group and 6 patients in routine group.
Intraoperative data, including management of ICAs T8-T12, are shown in Table 2. In all, 20% to 23% of the ICAs were occluded in T8-T12 levels. There were no statistically significant differences between the 2 groups for prevalence of occluded ICAs, except for T8 (selective vs routine: 19.2% vs 25.5%, P = .023). MEP signal loss was more frequently seen in the selective era (selective vs routine: 58.4% vs 37.6%, P < .001), but the incidence of lost MEP signal at the conclusion of surgery were similar in 2 groups (selective vs routine: 4.0% vs 3.9%, P = 1.000). Ligation of T8-T12 ICAs occurred significantly less frequently in the routine ICA reattachment era compared with the selective ICA reattachment era.
Table 2.
Intraoperative data
| Patient variable | Overall N = 959 | Routine n = 584 | Selective n = 375 | P value |
|---|---|---|---|---|
| Originally occluded ICAs | ||||
| T8 | 221 (23.0%) | 149 (25.5%) | 72 (19.2%) | .023 |
| T9 | 210 (21.9%) | 129 (22.1%) | 81 (21.6%) | .858 |
| T10 | 195 (20.3%) | 113 (19.4%) | 82 (21.9%) | .345 |
| T11 | 196 (20.4%) | 110 (18.8%) | 86 (22.9%) | .125 |
| T12 | 203 (21.2%) | 114 (19.5%) | 89 (23.7%) | .119 |
| Ligation of ICAs | ||||
| T8 | 196 (20%) | 102 (17%) | 94 (25%) | .005 |
| T9 | 132 (13.8%) | 58 (9.9%) | 74 (19.7%) | <.001 |
| T10 | 106 (10.5%) | 40 (6.9%) | 66 (17.7%) | <.001 |
| T11 | 94 (9.8%) | 35 (5.9%) | 59 (15.7%) | <.001 |
| T12 | 116 (12.1%) | 58 (9.9%) | 58 (15.5%) | .010 |
| Any ligation in T8-T12 | 339 (35.3%) | 182 (31.1%) | 157 (41.8%) | .009 |
| Any MEP loss | 439 (45.7%) | 220 (37.6%) | 219 (58.4%) | <.001 |
| Lost MEP signals at the conclusion of surgery | 38 (0%) | 23 (3.9%) | 15 (4.0%) | 1.000 |
Categorical variables are expressed as number (%). ICA, Intercostal artery; MEP, motor-evoked potentials.
Postoperative outcomes are summarized in Table 3. Postoperative respiratory failure was more common in the selective era (selective vs routine, 42% vs 32%, P = .001). Operative mortality rate was similar in the 2 eras (selective vs aggressive: rupture 12.5% vs 12.3%, P = .960). The incidence of immediate SCI was similar in the 2 eras (selective vs routine, 3.2% vs 2.2%, P = .356). However, the incidence of delayed SCI showed an unadjusted trend toward increase in the selective era (selective vs routine, 10.4% vs 6.9%, P = .055). In addition, the incidence of permanent SCI was significantly greater in selective eras (selective vs routine, 8.5% vs 5.3%, P = .049). We also assessed the effect of the timing of ICA reconstructions (after proximal anastomosis vs at the end), which did not affect the incidence of any forms of SCI. Neurologic outcomes across the 2 routine implantation periods did not differ. Results by policy period are shown in Table E1. The incidence of any SCIs in DTAA, TAAA extent I, TAAA extent II, TAAA extent III, TAAA extent IV, and TAAA extent V were 13.8%, 20.0%, 30.0%, 26%, 7.5%, and 2.5%, respectively.
Table 3.
Postoperative outcomes
| Patient variable | Overall N = 959 | Routine n = 584 | Selective n = 375 | P value |
|---|---|---|---|---|
| Any SCIs | 103 (10.7%) | 52 (8.9%) | 51 (13.6%) | .025 |
| Immediate | 25 (2.6%) | 13 (2.2%) | 12 (3.2%) | .356 |
| Delayed | 79 (8.2%) | 40 (6.9%) | 39 (10.4%) | .055 |
| Permanent SCI | 63 (6.6%) | 31 (5.3%) | 32 (8.5%) | .049 |
| Respiratory failure | 347 (36.2%) | 188 (32.2%) | 159 (42.4%) | .001 |
| New permanent RRT | 87 (9.1%) | 57 (9.8%) | 30 (8.0%) | .420 |
| Stroke | 63 (6.6%) | 35 (6.0%) | 28 (7.4%) | .423 |
| Operative mortality | 119 (12.4%) | 72 (12.3%) | 47 (12.5%) | .920 |
Categorical variables are expressed as number (%). Respiratory failure = ventilation hours >48 hours, reintubation, or tracheostomy. SCI, Spinal cord injury; RRT, renal-replacement therapy.
Propensity matching produced an analytical set of 672 (336 per group) patients within the region of common support (Table E2 and Figure E1). Variables included in the model were hypertension, genetic or connective tissue disorder, end-stage renal disease on dialysis, aortic dissection, and peripheral arterial disease. Key outcomes are shown in Table 4. Delayed SCI was greater in selective than routine groups: 39 of 369 (10.6%) versus 23 of 369 (6.2%) respectively; OR 1.78, P < .034. Similarly, permanent SCI was 31 of 369 (8.4%) in selective versus 17 of 369 (4.6%) in routine; OR 1.90, P < .037. Immediate SCI and 30-day death were not different between the strategies.
Figure E1.
Plot of the variables for standardized mean differences included hypertension, genetic or connective tissue disorder, end-stage renal disease on dialysis, aortic dissection and peripheral arterial disease.
Table 4.
Multivariable analysis for spinal cord ischemia
| Patient variable | Odds ratio | 95% CI | P value |
|---|---|---|---|
| Any SCI | |||
| Selective era | 1.61 | 1.04-2.49 | .032 |
| TAAA extent II | 6.69 | 3.70-12.11 | <.001 |
| TAAA extent III | 3.48 | 2.06-5.88 | <.001 |
| Age | 1.03 | 1.01-1.05 | <.001 |
| Immediate SCI | |||
| Selective era | 1.51 | 0.67-3.37 | .320 |
| TAAA extent III | 2.97 | 1.25-7.09 | .014 |
| Emergent/urgent | 4.01 | 1.53-10.50 | .005 |
| Delayed SCI | |||
| Selective era | 1.71 | 1.06-2.77 | .028 |
| TAAA extent II | 7.70 | 4.08-14.5 | <.001 |
| TAAA extent III | 3.13 | 1.74-5.64 | <.001 |
| Age | 1.04 | 1.02-1.06 | <.001 |
| Permanent SCI | |||
| Selective era | 1.89 | 1.19-3.03 | .007 |
| TAAA extent II | 7.07 | 3.75-13.35 | <.001 |
| TAAA extent III | 4.24 | 2.45-7.34 | <.001 |
| Age | 1.04 | 1.02-1.06 | <.001 |
CI, Confidence interval; SCI, spinal cord injury; TAAA, thoracoabdominal aortic aneurysm.
Discussion
The study showed that the incidence of immediate SCI was acceptably low and similar in both routine and selective ICA reattachment strategies for open DTAA/TAAA repair, in both crude and propensity-matched analysis. Neuromonitoring enables us to detect the early signs of intraoperative SCI, allowing us to expedite the measures to improve spinal cord perfusion without a delay. On the contrary, delayed SCI and permanent SCI were more frequently observed in the selective ICA reattachment strategy compared with the routine ICA reattachment strategy, despite 96% of patients in both strategies having intact MEP signals at the conclusion of surgery. As previously reported, delayed SCI occurs as the result of the postoperative “second hit” to the spinal cord.9 All patients undergoing DTAA/TAAA repair experience the first hit to the spinal cord during aortic crossclamping. Even when there is no MEP/SSEP changes, the damage is not zero. Postoperatively, the second hit to the spinal cord occurs with drop in blood pressure.9 When T8-T12 ICAs are not reconstructed, the patient has a low reserve to the spinal cord perfusion. Thus, patients without ICA reattachment are more vulnerable to delayed SCI and less likely to recover, leading to the permanent SCI.
Etz and colleagues10 demonstrated that T8-T12 ICAs were not the only blood supply to the spinal cord and collateral network were also important supply source. They also reported ICA ligation was safe in patients with intact MEP and SSEP signals during open DTAA/TAAA repair.11 Other groups also did not find the benefit of ICA reattachment.12,13 Currently, most surgeons agree to the importance of reattachment of ICAs in extent TAAA II because they are at the greatest risk for SCI.14,15 We have shown previously that ICA implantation is beneficial in all aneurysm extents for which ICA orifices are exposed.5,7
We also assessed the impact of the timing of ICA reconstructions. It is a reasonable assumption that ICA reconstruction in the earlier timing would have less SCI but we were unable to find any significant findings. However, it may be the result of its biased nature because in the earlier time period of the study, ICAs were routinely reattached after the proximal anastomosis. In the more recent period, we usually reattached ICAs after all the anastomoses were completed to decrease the aortic clamp time and pump time. If there was a change in neuromonitoring, especially SSEP signal loss, the surrogate of advanced ischemia, the ICAs were reattached immediately. Thus, the reconstruction timing was heavily biased because patients with ICA reattachment after the proximal anastomosis in the recent era more likely had a severe degree of intraoperative spinal ischemia.
In the past 3 decades, our group has had many changes and evolutions in our techniques during open DTAA/TAAA repair (Figure 1). During the study period, side-arm femoral cannulation was initiated to minimize limb ischemia16 and protect renal injury, loop graft for intercostal artery instead of separate graft to expedite the surgery,2 and modification of “COPS” protocol, a protocol to treat SCI that occurred.17 To minimize the risk of SCI after open DTAA/TAAA repair, in the future our group will continue to use routine ICA reattachment strategy along with the 3 mainstays of SCI prevention adjunct: distal aortic perfusion; permissive hypothermia; and CSFD.
Figure 1.
Summary of open descending and thoracoabdominal aortic repair techniques. ICA, Intercostal artery; MEP, motor-evoked potentials; SSEP, somatosensory-evoked potentials; STAG, side-branched thoracoabdominal aortic graft; VP, visceral perfusion; CSFD, cerebral spinal fluid drain; DAP, distal aortic perfusion; EVAR, endovascular aortic repair.
Limitations
This study should be viewed with limitations, as it is a retrospective, nonrandomized, single-center study. The retrospective design of the current analysis restricts the generalizability of our findings, and center-specific biases were inherently present. ICA management varied across an interrupted time series in 3 policy periods, and details of individual artery management and how these relate to changes in evoked potentials are complex and difficult to assess in this project. Although the interrupted time series presents a sort of natural experiment, sample size constraints required risk factor balance by propensity methods rather than complete explication by policy period. Last, we have modified our intercostal artery reconstruction from side-to-side (main body of the graft to the island patch of intercostal arteries), end-to-end interposition graft (single pair of intercostal arteries), loop graft (side-to-side to the interposition graft and island patch), modified end-to-end interposition graft (multiple pairs of intercostal arteries), which may have different patency in the long-term. Because of the limited access to the follow-up imaging, analysis of intercostal artery patency was not available to the study, which may affect the outcomes.
Conclusions
Two strategies of ICA reattachment in open DTAA/TAAA repair did not impact the incidence of immediate SCI. However, selective ICA reattachment significantly increased delayed SCI, including permanent SCI (Figure 2). Routine reattachment of ICAs within T8-T12 is warranted to minimize the risk of postoperative SCI after DTAA/TAAA repair.
Figure 2.
Graphical abstract. Impact of ICA reattachment strategies on SCI after open descending and thoracoabdominal aortic aneurysm repair.
Conflict of Interest Statement
Drs Estrera and Sandhu are consultants for WL Gore. Dr Estrera is a speaker for Terumo Aortic. All other authors reported no conflicts of interest.
The Journal policy requires editors and reviewers to disclose conflicts of interest and to decline handling or reviewing manuscripts for which they may have a conflict of interest. The editors and reviewers of this article have no conflicts of interest.
Acknowledgments
We thank Troy Brown for editing and Chris Akers for illustrations.
Footnotes
Dr Miller received support from National Institutes of Health grant 1UL1-TR003167-01.
Drs Tanaka and Sandhu contributed equally to this article.
Appendix E1
Table E1.
Preoperative patient characteristics and outcomes by eras
| Patient variable | Era 1 N = 443 | Era 2 N = 375 | Era 3 N = 141 | P value |
|---|---|---|---|---|
| Age, y | 63 (43-59) | 63 (55-71) | 63 (52-73) | |
| Female | 162 (37%) | 140 (37%) | 50 (35%) | .923 |
| Hypertension | 410 (93%) | 356 (95%) | 131 (93%) | .365 |
| Diabetes mellitus | 53 (12%) | 57 (15%) | 15 (11%) | .258 |
| COPD | 197 (44%) | 148 (39%) | 51 (36%) | .143 |
| Peripheral arterial disease | 85 (19%) | 96 (26%) | 23 (16%) | .025 |
| Coronary artery disease | 126 (28%) | 95 (25%) | 22 (16%) | .010 |
| History of stroke | 38 (9%) | 37 (10%) | 5 (4%) | .067 |
| eGFR <45 mL/min/1.83 m2 | 74 (17%)∗ | 63 (17%) | 13 (9.3%) | .069 |
| ESRD on dialysis | 18 (4%)∗ | 28 (7%) | 4 (3%) | .036 |
| CTD/HAD | 81 (18%) | 111 (30%) | 44 (31%) | <.001 |
| Redo | 119 (27%) | 98 (26%) | 35 (25%) | .889 |
| Rupture | 32 (7%) | 34 (9%) | 13 (9%) | .570 |
| Emergent/urgent | 21 (5%) | 25 (7%) | 29 (21%) | <.001 |
| Extent of aneurysm | ||||
| DTAA | 179 (40%) | 134 (36%) | 46 (33%) | .172 |
| TAAA extent I | 51 (12%) | 46 (12%) | 25 (18%) | .147 |
| TAAA extent II | 50 (11%) | 43 (11%) | 28 (20%) | .020 |
| TAAA extent III | 53 (12%) | 55 (15%) | 24 (17%) | .256 |
| TAAA extent IV | 78 (18%) | 70 (19%) | 14 (10%) | .053 |
| TAAA extent V | 32 (7%) | 27 (7%) | 4 (3%) | .153 |
| Any MEP loss | 156 (35%) | 219 (58%) | 64 (45%) | <.001 |
| Any SSEP loss | 155 (35%) | 81 (22%) | 29 (21%) | <.001 |
| Originally occluded ICAs | ||||
| T8 | 115 (26%) | 72 (19%) | 734 (24%) | |
| T9 | 98 (22%) | 81 (22%) | 31 (22%) | |
| T10 | 87 (20%) | 82 (22%) | 26 (18%) | |
| T11 | 84 (19%) | 86 (23%) | 26 (18%) | |
| T12 | 83 (19%) | 89 (24%) | 31 (22%) | |
| Ligation of ICAs | ||||
| T8 | 67 (15%) | 94 (25%) | 35 (25%) | |
| T9 | 36 (8%) | 74 (20%) | 22 (16%) | |
| T10 | 23 (5%) | 66 (18%) | 17 (12%) | |
| T11 | 15 (4%) | 59 (16%) | 20 (14%) | |
| T12 | 29 (7%) | 58 (15%) | 29 (20%) | |
| Any ligation in T8-T12 | 116 (26%) | 157 (42%) | 68 (47%) | <.001 |
| Any SCI | 38 (9%) | 51 (14%) | 14 (10%) | .065 |
| Immediate SCI | 11 (2%) | 12 (3%) | 2 (1%) | .514 |
| Delayed SCI | 28 (6%) | 39 (10%) | 12 (9%) | .106 |
| Permanent SCI | 22 (5%) | 32 (9%) | 9 (6%) | .121 |
| Stroke | 30 (7%) | 28 (7%) | 5 (4%) | .270 |
| Operative mortality | 66 (15%) | 47 (13%) | 6 (4%) | .004 |
Continuous variables are expressed as median (interquartile range), categorical variables are expressed as number (%). P values are for general heterogeneity between the 3 eras. COPD, Chronic obstructive pulmonary disease; eGFR, estimated glomerular filtration rate; ESRD, end-stage renal disease; CTD, connective tissue disorder; HAD, hereditary aortic disease; DTAA, descending thoracic aortic aneurysm; TAAA, thoracoabdominal aortic aneurysm; MEP, motor-evoked potentials; SSEP, somatosensory-evoked potentials; ICA, intercostal artery; SCI, spinal cord injury.
Data missing in 5 patients in era 1, 8 patients in era 2, and 1 patient in era 3.
Table E2.
Perioperative outcomes after propensity match
| Patient variable | Overall N = 959 | Routine N = 584 | Selective N = 375 | P value | Routine PM N = 369 | Selective PM N = 369 | P value |
|---|---|---|---|---|---|---|---|
| Any MEP loss | 439 (45.7%) | 220 (37.6%) | 219 (58.4%) | <.001 | 141 (38.2%) | 214 (29%) | <.001 |
| Any SCIs | 103 (10.7%) | 52 (8.9%) | 51 (13.6%) | .025 | 50 (13.6%) | 30 (8.1%) | .018 |
| Immediate | 25 (2.6%) | 13 (2.2%) | 12 (3.2%) | .356 | 8 (2.2%) | 11 (3.0%) | .486 |
| Delayed | 79 (8.2%) | 40 (6.9%) | 39 (10.4%) | .055 | 23 (6.2%) | 39 (10.6%) | .034 |
| Permanent SCI | 63 (6.6%) | 31 (5.3%) | 32 (8.5%) | .049 | 17 (4.6%) | 31 (8.4%) | .037 |
| Respiratory failure | 347 (36.2%) | 188 (32.2%) | 159 (42.4%) | .001 | 116 (31.4%) | 154 (41.7%) | .004 |
| New permanent RRT | 87 (9.1%) | 57 (9.8%) | 30 (8.0%) | .420 | 40 (10.8%) | 30 (8.1%) | .209 |
| Stroke | 63 (6.6%) | 35 (6.0%) | 28 (7.4%) | .423 | 26 (7.1%) | 27 (7.3%) | .887 |
| Operative mortality | 119 (12.4%) | 72 (12.3%) | 47 (12.5%) | .920 | 43 (11.7%) | 47 (12.7%) | .653 |
Categorical variables are expressed as number (%). Respiratory failure = ventilation hours >48 hours, reintubation, or tracheostomy. PM, Propensity matched; MEP, motor-evoked potentials; SCI, spinal cord injury; RRT, renal-replacement therapy.
References
- 1.Gaudino M., Khan F.M., Rahouma M., et al. Spinal cord injury after open and endovascular repair of descending thoracic and thoracoabdominal aortic aneurysms: a meta-analysis. J Thorac Cardiovasc Surg. 2022;163(2):552–564. doi: 10.1016/j.jtcvs.2020.04.126. [DOI] [PubMed] [Google Scholar]
- 2.Estrera A.L., Sandhu H.K., Charlton-Ouw K.M., et al. A quarter century of organ protection in open thoracoabdominal repair. Ann Surg. 2015;262(4):660–668. doi: 10.1097/SLA.0000000000001432. [DOI] [PubMed] [Google Scholar]
- 3.Isselbacher E.M., Preventza O., Hamilton Black J., III, et al. 2022 ACC/AHA guideline for the diagnosis and management of aortic disease: a report of the American Heart Association/American College of Cardiology Joint Committee on Clinical Practice Guidelines. Circulation. 2022;146(24):e334–e482. doi: 10.1161/CIR.0000000000001106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Tanaka H., Ogino H., Minatoya K., et al. The impact of preoperative identification of the Adamkiewicz artery on descending and thoracoabdominal aortic repair. J Thorac Cardiovasc Surg. 2016;151(1):122–1228. doi: 10.1016/j.jtcvs.2015.07.079. [DOI] [PubMed] [Google Scholar]
- 5.Safi H.J., Miller C.C., III, Carr C., Iliopoulos D.C., Dorsay D.A., Baldwin J.C. Importance of intercostal artery reattachment during thoracoabdominal aortic aneurysm repair. J Vasc Surg. 1998;27(1):58–66. doi: 10.1016/s0741-5214(98)70292-7. discussion 66-68. [DOI] [PubMed] [Google Scholar]
- 6.Keyhani K., Miller C.C., III, Estrera A.L., Wegryn T., Sheinbaum R., Safi H.J. Analysis of motor and somatosensory evoked potentials during thoracic and thoracoabdominal aortic aneurysm repair. J Vasc Surg. 2009;49(1):36–41. doi: 10.1016/j.jvs.2008.08.005. [DOI] [PubMed] [Google Scholar]
- 7.Afifi R.O., Sandhu H.K., Zaidi S.T., et al. Intercostal artery management in thoracoabdominal aortic surgery: to reattach or not to reattach? J Thorac Cardiovasc Surg. 2018;155(4):1372–1378.e1. doi: 10.1016/j.jtcvs.2017.11.072. [DOI] [PubMed] [Google Scholar]
- 8.Tanaka A., Nguyen H., Dhillon J.S., et al. Reappraisal of the role of motor and somatosensory evoked potentials during open distal aortic repair. J Thorac Cardiovasc Surg. 2023;165(3):944–953. doi: 10.1016/j.jtcvs.2021.08.033. [DOI] [PubMed] [Google Scholar]
- 9.Azizzadeh A., Huynh T.T., Miller C.C., III, et al. Postoperative risk factors for delayed neurologic deficit after thoracic and thoracoabdominal aortic aneurysm repair: a case-control study. J Vasc Surg. 2003;37(4):750–754. doi: 10.1067/mva.2003.211. [DOI] [PubMed] [Google Scholar]
- 10.Etz C.D., Zoli S., Bischoff M.S., Bodian C., Di Luozzo G., Griepp R.B. Measuring the collateral network pressure to minimize paraplegia risk in thoracoabdominal aneurysm resection. J Thorac Cardiovasc Surg. 2010;140(6 suppl):S125–S130. doi: 10.1016/j.jtcvs.2010.07.040. discussion S142–S146. [DOI] [PubMed] [Google Scholar]
- 11.Etz C.D., Homann T.M., Luehr M., et al. Spinal cord blood flow and ischemic injury after experimental sacrifice of thoracic and abdominal segmental arteries. Eur J Cardiothorac Surg. 2008;33(6):1030–1038. doi: 10.1016/j.ejcts.2008.01.069. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Wynn M., Acher C., Marks E., Acher C.W. The effect of intercostal artery reimplantation on spinal cord injury in thoracoabdominal aortic aneurysm surgery. J Vasc Surg. 2016;64(2):289–296. doi: 10.1016/j.jvs.2015.12.060. [DOI] [PubMed] [Google Scholar]
- 13.Kobayashi K., Saito S. Impact of reconstructing intercostal artery on spinal cord circulation during open surgery for thoracoabdominal aortic aneurysm. Braz J Cardiovasc Surg. 2023;38(1):43–51. doi: 10.21470/1678-9741-2021-0219. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kouchoukos N.T., Kulik A., Haynes M., Castner C.F. Early outcomes after thoracoabdominal aortic aneurysm repair with hypothermic circulatory arrest. Ann Thorac Surg. 2019;108(5):1338–1343. doi: 10.1016/j.athoracsur.2019.04.014. [DOI] [PubMed] [Google Scholar]
- 15.Coselli J.S., Green S.Y., Price M.D., et al. Spinal cord deficit after 1114 extent II open thoracoabdominal aortic aneurysm repairs. J Thorac Cardiovasc Surg. 2020;159(1):1–13. doi: 10.1016/j.jtcvs.2019.01.120. [DOI] [PubMed] [Google Scholar]
- 16.Miller C.C., III, Grimm J.C., Estrera A.L., et al. Postoperative renal function preservation with nonischemic femoral arterial cannulation for thoracoabdominal aortic repair. J Vasc Surg. 2010;51(1):38–42. doi: 10.1016/j.jvs.2009.08.044. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Sandhu H.K., Evans J.D., Tanaka A., et al. Fluctuations in spinal cord perfusion pressure: a harbinger of delayed paraplegia after thoracoabdominal aortic repair. Semin Thorac Cardiovasc Surg. 2017;29(4):451–459. doi: 10.1053/j.semtcvs.2017.05.007. [DOI] [PubMed] [Google Scholar]