- Email: [email protected]
Use of extracorporeal bypass is associated with improved outcomes in open thoracic and thoracoabdominal aortic aneurysm repair
From the Society for Clinical Vascular Surgery
Use of extracorporeal bypass is associated with improved outcomes in open thoracic and thoracoabdominal aortic aneurysm repair Jahan Mohebali, MD, MPH,a Stephanie Carvalho, BA,a R. Todd Lancaster, MD, MPH,a Emel A. Ergul, MS,a Mark F. Conrad, MD, MSSc,a W. Darrin Clouse, MD,a Richard P. Cambria, MD,b and Virendra I. Patel, MD, MPH,c Boston and Brighton, Mass; and New York, NY
ABSTRACT Objective: There is no consensus on the use or benefit of extracorporeal circulation (EC) during aneurysm repair of the descending thoracic aorta (DTA) or thoracoabdominal aorta (TAA). We evaluated the role of EC during DTA or TAA aneurysm repair using U.S. Medicare data. Methods: Medicare (2004-2007) patients undergoing open repair of nonruptured DTA or TAA aneurysm were identified by International Classification of Diseases, Ninth Revision code. Specific exclusions included ascending aortic or arch repairs, concomitant cardiac procedures, and procedures employing deep hypothermic circulatory arrest. The impact of EC (code 3961) on early and late outcomes was analyzed using univariate analysis and multivariable regression. Survival was assessed using Kaplan-Meier analysis and Cox proportional hazards regression models. Results: There were 4230 patients who had repair of intact DTA or TAA aneurysms, 2433 (57%) of which employed EC. Differences in baseline clinical features of EC and non-EC patients showed that patients undergoing aortic reconstruction with EC were older (73 6 1 years vs 72 6 1 years; P ¼ .002), were more likely to be female (53% vs 47%; P < .001), and had more hypertension (56% vs 53%; P ¼ .02); they had less chronic obstructive pulmonary disease (28% vs 34%; P < .0001), peripheral vascular disease (5.7% vs 11.3%; P < .001), and chronic kidney disease (7.7% vs 5.5%; P ¼ .003). The 30-day mortality (9.7% for EC vs 12.2%; P ¼ .02) and any major complication (49% for EC vs 58%; P < .001) were significantly reduced with EC use. EC use was associated with a shorter length of stay (13.5 6 13 days vs 17.2 6 18 days; P < .01) and lower total hospital charges ($151,000 6 140,000 vs $180,000 6 190,000; P < .01) compared with non-EC patients. EC patients were more likely to be discharged home instead of to an extended care facility (67% vs 56%; P < .01). Multivariable regression modeling to adjust for baseline clinical differences showed EC to independently reduce the risk of operative mortality (odds ratio [OR], 0.80; 95% confidence interval [CI], 0.65-0.97; P ¼ .02), any complication (OR, 0.67; 95% CI, 0.59-0.76; P < .01), pulmonary complications (OR, 0.68; 95% CI, 0.59-0.79; P < .01), and acute renal failure (OR, 0.52; 95% CI, 0.44-0.61; P < .01). Long-term survival was higher (log-rank, P < .01) in EC patients at 1 year (81% 6 0.8% vs 73% 6 1%) and 5 years (67% 6 1% vs 52% 6 1%). Risk-adjusted Cox proportional hazards regression also showed that EC was independently associated with improved long-term survival (hazard ratio, 0.69; 95% CI, 0.63-0.74; P < .01). Conclusions: Although important clinical variables such as DTA or TAA aneurysm extent and spinal cord ischemic complications cannot be assessed with the Medicare database, EC use during open DTA and TAA aneurysm repair is associated with improved late survival and a significant reduction in operative mortality, morbidity, and procedural costs. These data indicate that EC should be a more widely applied adjunct in open DTA or TAA aneurysm repair. (J Vasc Surg 2018;-:1-7.) Keywords: Thoracoabdominal aortic aneurysm; Descending thoracic aortic aneurysm; Extracorporeal circulation; Atrial-femoral bypass
Aneurysmal dilation of the descending thoracic aorta (DTA) and thoracoabdominal aorta (TAA) most commonly results from primary cystic medial necrosis or secondary degeneration after type B dissection
and is classified on the basis of the extent of aortic involvement.1 Simple “clamp and sew” (CS) repair as championed by Crawford et al1 emphasizes expediency as the aorta is clamped and replaced with synthetic
From the Division of Vascular and Endovascular Surgery, Massachusetts
Correspondence: Virendra I. Patel, MD, MPH, Columbia University Medical
General Hospital, Harvard Medical School, Bostona; the Division of Vascular
Center, Herbert Irving Pavilion, 161 Fort Washington Ave, Fifth Fl, Ste 532,
Surgery, Steward Medical Group, St. Elizabeth’s Medical Center, Brightonb;
New York, NY 10032 (e-mail: [email protected]).
and the Division of Vascular Surgery and Endovascular Interventions,
The editors and reviewers of this article have no relevant financial relationships to
Columbia University Medical Center, Columbia University College of Physi-
disclose per the JVS policy that requires reviewers to decline review of any
cians and Surgeons, New York.c Author conflict of interest: none. Plenary presentation at the Forty-first Annual Symposium of the Society for Clinical Vascular Surgery, Miami, Fla, March 12-16, 2013.
manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2018 by the Society for Vascular Surgery. Published by Elsevier Inc. https://doi.org/10.1016/j.jvs.2017.12.072
1
2
Journal of Vascular Surgery
Mohebali et al
---
graft in a proximal to distal fashion.2 Thoracic aortic cross-clamping, however, results in profound physiologic derangement as distal tissues become ischemic. Furthermore, left ventricular preload and afterload rise acutely, resulting in increased left ventricular wall stress, stroke work, and, in turn, myocardial oxygen demand. Ischemia may result in postoperative liver dysfunction, gut necrosis, renal failure, and most devastatingly, paraplegia/ paralysis. Indeed, in his benchmark experience, Crawford found that cross-clamp time remained a significant predictor of early death even after multivariable adjustment.1 Consequently, a multitude of intraoperative adjuncts designed either to maintain end-organ perfusion or to minimize ischemic tissue metabolic demand during cross-clamping have been employed. Visceral perfusion can be achieved with a pump-powered cannula perfuser or by direct in-line (passive) mesenteric shunting.3-5 Complete or selective reimplantation of intercostal and lumbar vessels can be used to improve cord perfusion. As cross-clamping also increases cerebrospinal fluid (CSF) pressure and consequently decreases spinal perfusion pressure, CSF drainage can be used to improve cord perfusion and is the only paraplegia-mitigating adjunct supported by level I evidence.6 Finally, improved global perfusion distal to the cross-clamp can be achieved with the use of mechanical circulatory devices.6 These extracorporeal circulation (EC) techniques range from left atrial-femoral bypass to full cardiopulmonary bypass. Irrespective of technique, EC provides the added advantage of mitigating the increase in preload and afterload associated with proximal cross-clamping. Some studies, however, have demonstrated no significant advantage with EC compared with the simple CS technique,7 and many authors have noted that the increased intraoperative technical complexity associated with EC is not worth the perceived benefit. It becomes clear, then, that there is no standardized approach for the repair of DTA and TAA aneurysms, with the literature revealing conflicting data. Furthermore, most existing studies are limited to single institutions or single-surgeon series, which may fail to account for the various adjuncts and practice patterns, confounding the outcomes associated with use of CS vs EC. The purpose of this study, therefore, was to evaluate the effects of EC on open repair of DTA and TAA aneurysms across multiple institutions and surgeons using the U.S. Medicare database.
METHODS Study design and cohort. This retrospective cohort study of the Medicare population targeted Medicare patients undergoing intact open repair of DTA aneurysm or TAA aneurysm from 2004 to 2007. The Medicare Provider Analysis and Review specifically covered Medicare Part A claims and included a linkage to Vital
2018
ARTICLE HIGHLIGHTS d
d
d
Type of Research: Retrospective analysis of prospectively collected Medicare data Take Home Message: In 4230 patients receiving open descending thoracic or thoracoabdominal aortic aneurysm repair, use of extracorporeal circulation in 2433 patients resulted in lower major complication rate (49% vs 58%), shorter hospital stay, and lower total hospital charges. Recommendation: This study suggests that extracorporeal circulation should be used more frequently during open repair of descending thoracic and thoracoabdominal aortic aneurysms.
Statistics for mortality and long-term survival analysis. International Classification of Diseases, Ninth Revision (ICD-9) codes (Appendix) pertaining to nonruptured, open thoracic aortic repair were used to broadly identify the initial cohort. ICD-9 codes were then used to exclude patients lacking a diagnosis of DTA or TAA aneurysm and patients undergoing concomitant cardiac procedures including coronary revascularization, valve procedures, or procedures employing cardioplegia or deep hypothermic circulatory arrest. The remaining cohort was divided into two groups according to ICD-9 codes reflecting use of EC (ICD-9 3961) vs those that did not. Primary outcomes evaluated included 30-day mortality, long-term survival, and predictors of death or late mortality. Secondary outcomes included hospital and intensive care unit length of stay, systemic complications, and predictors of systemic complications. Acute renal failure was defined by ICD-9 code (Appendix). Institutional Review Board approval was obtained for the study. The need for informed consent of the patient was waived by the Institutional Review Board, given the use of deidentified administrative data for the study. Statistical analysis. All data were analyzed using the SAS statistical software package (version 9.3; SAS Institute, Cary, NC). A univariate cutoff of 0.1 was used for initial inclusion in multivariable models. Stepwise selection was used with an entrance and exit cutoff of .05 to prune the full models to arrive at a final, adjusted, multivariable model. Kaplan-Meier analysis was performed between the EC and no EC groups to determine longterm survival, with the log-rank test used to assess for significance. Cox proportional hazards regression was used to model survival after adjusting for significant covariates and confounders.
RESULTS Cohort characteristics. There were 18,282 patients identified as having nonruptured open thoracic aortic
Journal of Vascular Surgery Volume
-,
Number
Mohebali et al
3
-
aneurysm repairs in the U.S. Medicare population between 2004 and 2007. From these 18,282 patients, 10,658 were excluded for involving simultaneous cardiac revascularization, cardioplegia, or valve procedures; 116 were excluded for involving deep hypothermic circulatory arrest, and 3278 were excluded because they were lacking a diagnosis of DTA or TAA aneurysm. After these exclusions, the entire cohort of open DTA or TAA aneurysm repairs was composed of 4230 procedures. These procedures were then grouped into an EC group consisting of 2433 (57%) procedures and a no EC group consisting of 1797 (43%) procedures. Differences in baseline clinical features between patients who underwent procedures with or without EC are displayed in Table I. The EC group maintained a higher percentage of female patients and dissections (1.3%), whereas the no EC group included a higher percentage of TAA patients (64%) exhibiting chronic obstructive pulmonary disease, peripheral vascular disease, and chronic kidney disease (Table I). Perioperative outcomes. Use of EC resulted in a significantly reduced 30-day mortality rate (9.7% vs 12.2%; P ¼ .01) and a significantly reduced overall complication rate (49% vs 58%; P < .0001; Table II). The no EC group had a significantly higher increase in pulmonary complications (27% vs 21%; P < .0001), bleeding (19% vs 13%; P < .0001), and acute renal failure (24% vs 14%; P < .0001; Table II). EC resulted in a significantly shorter length of stay both in the intensive care unit (5 days vs 6 days; P < .0001) and in the hospital (9 days vs 11 days; P < .0001). In addition, significantly larger portions of patients with EC use during the procedure were discharged home vs discharged to an extended care facility (67% vs 56%; P < .0001). Multivariable regression modeling showed EC to independently reduce operative mortality (odds ratio [OR], 0.80; 95% confidence interval [CI], 0.65-0.97; P ¼ .02), any complication (OR, 0.67; 95% CI, 0.59-0.76; P < .0001), pulmonary complications (OR, 0.68; 95% CI, 0.59-0.79; P < .0001), and acute renal failure (OR, 0.52; 95% CI, 0.44-0.61; P < .0001) in adjusting for baseline clinical differences (Table III). Total hospital charges for patients with EC use were reduced compared with the total charge for patients who did not have EC ($110,000 vs $120,000; P < .0001; Table IV). Long-term survival. Cox proportional hazards regression showed that EC was independently associated with improved long-term survival (hazard ratio [HR], 0.69; 95% CI, 0.63-0.74; P < .0001), whereas postoperative complications were independently associated with worsened long-term survival (HR, 2.4; 95% CI, 2.1-2.7; P < .0001; Table V). Kaplan-Meier analysis of long-term survival showed a significantly (P < .01) higher survival in patients who had EC used during their procedure at both 1 year (81% vs 73%) and 5 years (67% vs 52%; Fig).
Table I. Clinical features of the study cohorts Variable
No EC (n ¼ 1797)
EC (n ¼ 2433)
72 6 8
73 6 9
Age, years, mean 6 SD Female Hypertension Coronary artery disease Cerebrovascular disease
.002
47
53
53
56
.02
24
23
.7
4.4
Chronic obstructive pulmonary disease
P value
34
5.4 28
.001
.17 <.0001
Peripheral vascular disease
11.3
5.7
<.0001
Chronic kidney disease
7.7
5.5
.003
EC, Extracorporeal circulation; SD, standard deviation. Values are reported as percentage unless otherwise indicated. Boldface P values represent statistical significance (P < .05).
Table II. Perioperative outcomes by cohort Variable
No EC (n ¼ 1797), %
30-Day mortality
12.2
Any complication
58
EC (n ¼ 2433), % 9.7 49
P value .01 <.0001
Cardiac
13
14
Pulmonary
27
21
<.0001
Bleeding
19
13
<.0001
Infectious Acute renal failure
7.7 24
5.9 14
.16
.02 <.0001
EC, Extracorporeal circulation. Boldface P values represent statistical significance (P < .05).
Table III. Independent effect of extracorporeal circulation (EC) on outcomes Variable Death
EC OR (95% CI)
P value
0.8 (0.65-0.97)
.02
0.67 (0.59-0.76)
<.0001
Pulmonary complications
0.68 (0.59-0.79)
<.0001
Infectious complications
0.72 (0.57-0.92)
.009
Bleeding complications
0.61 (0.52-0.73)
<.0001
Renal complications
0.52 (0.44-0.61)
<.0001
Any complication
CI, Confidence interval; OR, odds ratio. Multivariable models risk adjusted for age, sex, race, coronary artery disease, cerebrovascular disease, chronic kidney disease, and peripheral vascular disease. Boldface P values represent statistical significance (P < .05).
DISCUSSION The strategy employed for open surgical repair of DTA and TAA aneurysms remains variable. Whereas Crawford’s benchmark study using CS techniques1 as well as Svensson’s follow-up experience with 1509 patients8 demonstrated mortality rates <10%, the authors noted that “vexing problems” including paraplegia and renal failure remained unacceptably high. Among Crawford’s patients, 30-day postoperative mortality and “neuromuscular deficit” (paraplegia/paralysis) were significantly
4
Journal of Vascular Surgery
Mohebali et al
---
2018
Table IV. Length of stay and hospital charges by study cohort No EC (n ¼ 1797), mean 6 SD (median)
EC (n ¼ 2433), mean 6 SD (median)
180,000 6 190,000 (120,000)
151,000 6 140,000 (110,000)
Variable Total charges, $
14 6 13 (9)
P value <.0001
Hospital length of stay, days
17 6 18 (11)
<.0001
Intensive care unit length of stay, days
10 6 14 (6)
8 6 18 (5)
<.0001
Discharge to home
56%
67%
<.0001
EC, Extracorporeal circulation. Boldface P values represent statistical significance (P < .05).
Table V. Independent predictors of late mortality Variable EC
Death HR (95% CI)
P value
0.69 (0.63-0.74)
<.0001
Postoperative complications
2.4 (2.1-2.7)
<.0001
Cerebrovascular disease
2.4 (2.0-2.9)
<.0001
Chronic kidney disease
1.6 (1.4-2.0)
<.0001
Age (per year)
1.04 (1.03-1.04)
<.0001 <.0001
Chronic obstructive pulmonary disease
1.4 (1.2-1.5)
Peripheral vascular disease
1.3 (1.1-1.6)
.01
Nonwhite race
1.3 (1.1-1.5)
.002
Coronary artery disease
1.1 (1.01-1.3)
.04
EC, Extracorporeal circulation; HR, hazard ratio. Boldface P values represent statistical significance (P < .05).
associated with longer aortic cross-clamp times.1 Crawford additionally reported an overall paraplegia rate of 11%, noting an association with dissection and extent of disease. Indeed, the highest rates of paraplegia approaching 41% were seen among extent II patients with dissection.1 Crawford himself noted that “we have not been able to significantly prevent or reduce the incidence of these complications and have concentrated on rehabilitative therapy. There does not appear to be an immediate hope of solving this problem.”2 Furthermore, in consideration of distal perfusion adjuncts, the authors cited the work of Livesay et al,9 stating that “distal perfusion alone in a large number of cases subjected to extensive resection of the DTA has not altered the problem.” However, Livesay’s study used temporary bypasses and shunts to provide passive distal aortic perfusion rather than EC. Svensson’s later studies proved that left atrial-femoral bypass maintained blood supply to the spinal cord while the proximal anastomosis was being performed.8,10,11 Indeed, in their adoption of CSF drainage and distal aortic perfusion, Safi et al12 found that aortic cross-clamp time and even extent I disease were no longer significant predictors of neurologic deficit. In other words, EC essentially eliminated the cord ischemia associated with proximal descending thoracic aortic cross-clamping. These results laid the foundation for the theory that the spinal cord is
perfused by a collateral network of vessels supplied by multiple aortic branches. The collateral network concept was later confirmed by Backes et al13 and Jacobs et al14 using magnetic resonance angiography and by Griepp and Griepp,15 who formalized the concept by citing the increasing body of literature suggesting that routine sacrifice of segmental aortic branches perfusing the cord could be carried out as long as overall inflow to the network was maintained, as occurs with distal aortic perfusion. With regard to visceral perfusion, Safi’s group later demonstrated that distal aortic and consequently visceral perfusion also mitigates postoperative laboratory evidence of hepatic and visceral ischemia during open extent II aneurysm repair.5 Welborn et al16 similarly evaluated levels of plasma proinflammatory cytokines associated with multisystem organ injury, dysfunction, and critical illness among patients undergoing open TAA aneurysm repair. The authors clearly demonstrated that use of left atrial-femoral bypass significantly reduced levels of these factors to values seen in simple open infrarenal aortic aneurysm repair.16 Finally, Jacobs et al17 showed that renal and visceral ischemia could be significantly reduced by continuous perfusion during proximal aortic cross-clamping with consequent lower rates of postoperative renal failure. At our institution, the practice has similarly evolved from a CS technique18 using adjuncts such as epidural cooling, renal hypothermia, and CSF drainage with an aggressive intercostal reimplantation policy2 to one employing these adjuncts (without epidural cooling) in addition to distal perfusion through left atrial-femoral bypass and motor evoked potential monitoring (MEVP)19 for extent I to III thoracoabdominal aneurysms. Using a propensity matched analysis in a small cohort of these patients, we demonstrated that use of EC and MEVP resulted in lower early mortality and paraplegia rates compared with a CS technique.6 Our follow-up series with a larger cohort (n ¼ 485) demonstrated half the early postoperative mortality rate, a substantial decrease in acute renal failure requiring hemodialysis, and a profound decrease in permanent spinal cord injury for the EC/MEVP group compared with CS.19 In contrast, Coselli et al7 used a propensity matched
Journal of Vascular Surgery Volume
-,
Number
Mohebali et al
5
-
Fig. Kaplan-Meier curves for long-term survival after thoracic or thoracoabdominal aortic aneurysm repair by use of extracorporeal circulation (EC) and clamp and sew (CS) techniques.
multivariable logistic regression analysis to match patients on the basis of likelihood of receiving EC vs CS and found that after matching, there was no statistically significant difference in rates of 30-day mortality, paraplegia, or renal failure. We have found this to be true for our experience with type IV aneurysms only, as perioperative mortality and complication rates in these select patients are similar to open abdominal aortic aneurysm repair.20 Whereas our experience as well as the work of others clearly supports the use of EC for open repair of extent I to III thoracoabdominal aneurysm, controversy persists in the literature regarding the optimal surgical technique.7,21 Results are further confounded by the presence or absence of dissection, need for emergent repair,
extent of aneurysmal degeneration, use of non-EC adjuncts such as MEVP, spinal drainage, renal hypothermia, in-line mesenteric shunting, extent of intercostal reconstruction, and other surgeon-specific or institutionspecific practices that may not be captured in large databases. It becomes imperative, then, to consider the effects of EC across multiple institutions, populations of patients, practice patterns, and surgeons. Although a randomized controlled trial would certainly be ideal in helping to disentangle the benefit of EC from other adjuncts, to our knowledge, no such study exists as confirmed by the most recent Cochrane database review on the topic.22 The goal of this study, then, was to determine the effects of EC on surgical outcomes across multiple institutions and surgeons using the U.S. Medicare
6
Journal of Vascular Surgery
Mohebali et al
---
database. To our knowledge, this is the largest study examining the effects of EC during open repair of DTA aneurysm and TAA aneurysm to date. In this study, use of EC during open DTA aneurysm and TAA aneurysm repair was shown to have a favorable impact on perioperative outcomes, financial outcomes, and late outcomes. EC was beneficial in reducing perioperative mortality, pulmonary and renal morbidity, and length of stay and costs while allowing home discharge and improving late survival. Specifically, rates of renal failure for the EC group were reduced from 24% to 14%. Estrera et al23 routinely used EC and reported rates of renal dysfunction approaching 24%, whereas Coselli et al7 espoused no benefit with the use of EC as their rates of renal failure were not significantly different for the EC (7%) vs non-EC groups (7.6%). Our previous work demonstrated a trend toward lower rates of renal failure in the EC group (5.1% vs 11.4%); however, the results were not significant (P ¼ .063).19 As discussed previously, this variability could be the result of additional adjuncts, such as shunting and use of renal hypothermia or preservation fluid, that mitigate the effects of ischemia. Furthermore, variability in these results may be because of baseline levels of preoperative renal dysfunction, which multiple authors, including us, have demonstrated to be the greatest predictor of postoperative renal dysfunction. Nevertheless, this study shows that EC provides a clear benefit for postoperative renal function. With regard to 30-day mortality, this study shows that EC provided a clear benefit (9.7% vs 12.2%; P ¼ .01), which is similar to our previous work (4.0% vs 9.9%).19 Furthermore, whereas our study did not show significant differences in length of intensive care unit or hospital stay, the current study found a clear reduction in both of these variables for the EC group. Finally, the study results provide novelty in analyzing the effects of EC use on total hospital charges, which were significantly reduced for the EC group. Similarly, long-term mortality results of the study that demonstrate a protective effect of EC (OR, 0.69; 5-year survival of 67% 6 1% for EC vs 52% 6 1%) mirror our previous work demonstrating improved survival for left atrial-femoral bypass patients (4-year survival of 73% 6 6% for EC vs 60% 6 3%)19 after adjusting for other significant predictors of long-term mortality (HR, 0.590). These results are similar to those of other large series at centers of excellence.12,24 Therefore, unless it is contraindicated for technical or clinical reasons, these data support the growing body of literature indicating that EC should be more widely applied as an adjunct during open DTA aneurysm and TAA aneurysm repair.
2018
Medicare database did not allow us to account for the various additional adjuncts (renal hypothermia, mesenteric shunting, epidural cooling, MEVP) that may have been employed during these procedures. It is certainly possible that use of these adjuncts could confound results. Furthermore, the database lacks information on extent of aneurysm (I-IV), which has been predictive of spinal cord ischemia,2,3,8,25 postoperative renal failure,26 and death.3 In addition, the greatest benefit of EC has been the profound decrease in postoperative rates of spinal cord ischemia as demonstrated by multiple studies.4,6,19 Unfortunately, we were not able to evaluate for this devastating outcome as it is lacking in the database. Our previous work and the most recent work of Estrera et al,23 however, demonstrated that postoperative paraplegia/paralysis certainly increases longterm mortality, and it is therefore conceivable that the survival benefit noted in the EC group in this study could have been at least in part the result of lower rates of irreversible spinal cord ischemia. Finally, as with any retrospective study, there is undoubtedly unmeasured confounding in the form of selection bias pertaining to those patients for whom EC was chosen as an integral part of the repair. As noted before, the literature suggests that some surgeons or institutions preferentially employ EC in urgent cases, whereas some who routinely use EC may abandon it completely in the setting of emergent cases with rupture in favor of expedient salvage with CS (no EC) techniques. We did attempt to minimize the risk of this selection bias by excluding ruptured or emergent cases. Another source of clear selection bias is the practice of most large centers to selectively employ EC for more extensive aneurysm resections. We were unfortunately unable to adjust for the effects of this variable on outcomes as a result of database limitations as noted before. Finally, to the extent allowed by the Medicare database, we included known predictors of the postoperative outcomes analyzed in this study in a multivariable model to minimize the effects of selection bias.
CONCLUSIONS This study is the first to demonstrate a significant benefit in employing EC during DTA and TAA aneurysm repair across multiple institutions, surgeons, and populations of patients. Future databases and studies on the subject should seek to accurately capture and characterize the use of all intraoperative adjuncts, other patient and anatomic considerations that may affect the decision to employ EC, and all relevant postoperative outcomes and complications.
AUTHOR CONTRIBUTIONS Limitations and future studies. Whereas this study suggests that employing EC during DTA and TAA aneurysm repair has beneficial effects on mortality and postoperative renal failure, the inherent limitations of the U.S.
Conception and design: VP Analysis and interpretation: JM, RL, EE, MC, WC, RC, VP Data collection: SC Writing the article: JM, SC, VP
Journal of Vascular Surgery Volume
-,
Number
Mohebali et al
7
-
Critical revision of the article: JM, RL, EE, MC, WC, RC, VP Final approval of the article: JM, SC, RL, EE, MC, WC, RC, VP Statistical analysis: EE, VP Obtained funding: Not applicable Overall responsibility: VP
REFERENCES 1. Crawford ES, Crawford JL, Safi HJ, Coselli JS, Hess KR, Brooks B, et al. Thoracoabdominal aortic aneurysms: preoperative and intraoperative factors determining immediate and long-term results of operations in 605 patients. J Vasc Surg 1986;3:389-404. 2. Conrad MF, Crawford RS, Davison JK, Cambria RP. Thoracoabdominal aneurysm repair: a 20-year perspective. Ann Thorac Surg 2007;83:S856-61. 3. Coselli JS, Bozinovski J, LeMair SA. Open surgical repair of 2286 thoracoabdominal aortic aneurysms. Ann Thorac Surg 2007;83:S862-4. 4. Patel VI, Lancaster RT, Conrad MF, Cambria RP. Open surgical repair of thoracoabdominal aneurysmsdthe Massachusetts General Hospital experience. Ann Cardiothorac Surg 2012;1:320-4. 5. Safi HJ, Miller CC 3rd, Yawn DH, Iliopoulos DC, Subramaniam M, Harlin S, et al. Impact of distal aortic and visceral perfusion on liver function during thoracoabdominal and descending thoracic aortic repair. J Vasc Surg 1998;27:145-52. 6. Conrad MF, Ergul EA, Patel VI, Cambria MR, LaMuraglia GM, Simon M, et al. Evolution of operative strategies in open thoracoabdominal aneurysm repair. J Vasc Surg 2011;53:1195-201. 7. Coselli JS, LeMair SA, Conklin LD, Adams GJ. Left heart bypass during descending thoracic aortic aneurysm repair does not reduce the incidence of paraplegia. Ann Thorac Surg 2004;77:1298-303. 8. Svensson LG, Crawford ES, Hess KR, Coselli JS, Safi HJ. Experience with 1509 patients undergoing thoracoabdominal aortic operations. J Vasc Surg 1993;17:357-68. 9. Livesay JJ, Cooley DA, Ventemiglia RA, Montero CG, Warrian RK, Brown DM, et al. Surgical experience in descending thoracic aneurysmectomy with and without adjuncts to avoid ischemia. Ann Surg 1985;39:37-44. 10. Svensson LG, Patel V, Robinson MF, Ueda T, Roehm JO Jr, Crawford ES. Influence of preservation or perfusion of intraoperatively identified spinal cord blood supply on spinal motor evoked potentials and paraplegia after aortic surgery. J Vasc Surg 1991;13:355-65. 11. Svensson LG, Crawford ES, Patel V, McLean TR, Jones JW, DeBakey ME. Spinal cord oxygenation, intraoperative blood supply localization, cooling and function with aortic clamping. Ann Thorac Surg 1992;54:74-9. 12. Safi HJ, Miller CC 3rd, Huynh TT, Estrera AL, Porat EE, Winnerkvist AN, et al. Distal aortic perfusion and cerebrospinal fluid drainage for thoracoabdominal and descending thoracic aortic repair: ten years of organ protection. Ann Surg 2003;238:372-80. 13. Backes WH, Nijenhuis RJ, Mess WH, Wilmink FA, Schurink GW, Jacobs MJ. Magnetic resonance angiography of collateral blood supply to spinal cord in thoracic and thoracoabdominal aortic aneurysm patients. J Vasc Surg 2008;48:261-71. 14. Jacobs MJ, de Mol BA, Elenbaas T, Mess WH, Kalkman CJ, Schurink GW, et al. Spinal cord blood supply in patients with thoracoabdominal aortic aneurysm. J Vasc Surg 2002;35:30-7.
15. Griepp EB, Griepp RB. The collateral network concept: minimizing paraplegia secondary to thoracoabdominal aortic aneurysm resection. Tex Heart Inst J 2010;37:672-4. 16. Welborn MB, Oldenburg HS, Hess PJ, Huber TS, Martin TD, Rauwerda JA, et al. The relationship between visceral ischemia, proinflammatory cytokines, and organ injury in patients undergoing thoracoabdominal aortic aneurysm repair. Crit Care Med 2000;28:3191-7. 17. Jacobs MJ, Eijsman L, Meylaerts SA, Balm R, Legemate DA, de Haan P, et al. Reduced renal failure following thoracoabdominal aortic aneurysm repair by selective perfusion. Eur J Cardiothorac Surg 1998;14:201-5. 18. Cambria RP. Thoracoabdominal aortic aneurysm repair: how I do it. Cardiovasc Surg 1999;7:597-606. 19. Lancaster RT, Conrad MF, Patel VI, Cambria MR, Ergul EA, Cambria RP. Further experience with distal aortic perfusion and motor-evoked potential monitoring in the management of extent I-III thoracoabdominal aortic aneurysms. J Vasc Surg 2013;58:283-90. 20. Patel VI, Ergul E, Conrad MF, Cambria M, LaMuraglia GM, Kwolek CJ, et al. Continued favorable results with open surgical repair of type IV thoracoabdominal aortic aneurysms. J Vasc Surg 2011;53:1492-8. 21. Safi HJ, Miller CC 3rd, Subramaniam MH, Campbell MP, Iliopoulos DC, O’Donnell JJ, et al. Thoracic and thoracoabdominal aortic aneurysm repair using cardiopulmonary bypass, profound hypothermia, and circulatory arrest via left side of the chest incision. J Vasc Surg 1998;28:591-8. 22. Hsu CC, Kwan GN, van Driel ML, Rophael JA. Distal aortic perfusion during thoracoabdominal aneurysm repair for prevention of paraplegia. Cochrane Database Syst Rev 2012;3:CD008197. 23. Estrera AL, Sandhu HK, Charlton-Ouw KM, Afifi RO, Azizzadeh A, Miller CC 3rd, et al. A quarter century of organ protection in open thoracoabdominal repair. Ann Surg 2015;262:660-8. 24. Safi JH, Estrera AL, Miller CC, Huynh TT, Porat EE, Azizzadeh A, et al. Evolution of risk for neurologic deficit after descending and thoracoabdominal aortic repair. Ann Thorac Surg 2005;80:2173-9. 25. Wynn MW, Acher CW. A modern theory of spinal cord ischemia/injury in thoracoabdominal aortic surgery and its implications for prevention of paralysis. J Cardiothorac Vasc Anesth 2014;28:1088-99. 26. Attab M, Coselli JS. Renal and visceral protection in thoracoabdominal aortic surgery. J Thorac Cardiovasc Surg 2014;148:2963-6.
Submitted Sep 25, 2017; accepted Dec 21, 2017.
APPENDIX International Classification of Diseases, Ninth Revision (ICD-9) codes used in the study. Open thoracic aortic repair: 3845 Circulatory arrest: 39.63 Valve replacement: 3500-3504, 3510-3514, 3520-3528 Coronary artery bypass graft: 3610-3617 Extracorporeal circulation: 3961 Acute renal failure: 5845-5849, 586 Thoracoabdominal aneurysm: 4412 Dissection: 4410, 44100-44103
Use of extracorporeal bypass is associated with improved outcomes in open thoracic and thoracoabdominal aortic aneurysm repair Jahan Mohebali, MD, MPH,a Stephanie Carvalho, BA,a R. Todd Lancaster, MD, MPH,a Emel A. Ergul, MS,a Mark F. Conrad, MD, MSSc,a W. Darrin Clouse, MD,a Richard P. Cambria, MD,b and Virendra I. Patel, MD, MPH,c Boston and Brighton, Mass; and New York, NY
ABSTRACT Objective: There is no consensus on the use or benefit of extracorporeal circulation (EC) during aneurysm repair of the descending thoracic aorta (DTA) or thoracoabdominal aorta (TAA). We evaluated the role of EC during DTA or TAA aneurysm repair using U.S. Medicare data. Methods: Medicare (2004-2007) patients undergoing open repair of nonruptured DTA or TAA aneurysm were identified by International Classification of Diseases, Ninth Revision code. Specific exclusions included ascending aortic or arch repairs, concomitant cardiac procedures, and procedures employing deep hypothermic circulatory arrest. The impact of EC (code 3961) on early and late outcomes was analyzed using univariate analysis and multivariable regression. Survival was assessed using Kaplan-Meier analysis and Cox proportional hazards regression models. Results: There were 4230 patients who had repair of intact DTA or TAA aneurysms, 2433 (57%) of which employed EC. Differences in baseline clinical features of EC and non-EC patients showed that patients undergoing aortic reconstruction with EC were older (73 6 1 years vs 72 6 1 years; P ¼ .002), were more likely to be female (53% vs 47%; P < .001), and had more hypertension (56% vs 53%; P ¼ .02); they had less chronic obstructive pulmonary disease (28% vs 34%; P < .0001), peripheral vascular disease (5.7% vs 11.3%; P < .001), and chronic kidney disease (7.7% vs 5.5%; P ¼ .003). The 30-day mortality (9.7% for EC vs 12.2%; P ¼ .02) and any major complication (49% for EC vs 58%; P < .001) were significantly reduced with EC use. EC use was associated with a shorter length of stay (13.5 6 13 days vs 17.2 6 18 days; P < .01) and lower total hospital charges ($151,000 6 140,000 vs $180,000 6 190,000; P < .01) compared with non-EC patients. EC patients were more likely to be discharged home instead of to an extended care facility (67% vs 56%; P < .01). Multivariable regression modeling to adjust for baseline clinical differences showed EC to independently reduce the risk of operative mortality (odds ratio [OR], 0.80; 95% confidence interval [CI], 0.65-0.97; P ¼ .02), any complication (OR, 0.67; 95% CI, 0.59-0.76; P < .01), pulmonary complications (OR, 0.68; 95% CI, 0.59-0.79; P < .01), and acute renal failure (OR, 0.52; 95% CI, 0.44-0.61; P < .01). Long-term survival was higher (log-rank, P < .01) in EC patients at 1 year (81% 6 0.8% vs 73% 6 1%) and 5 years (67% 6 1% vs 52% 6 1%). Risk-adjusted Cox proportional hazards regression also showed that EC was independently associated with improved long-term survival (hazard ratio, 0.69; 95% CI, 0.63-0.74; P < .01). Conclusions: Although important clinical variables such as DTA or TAA aneurysm extent and spinal cord ischemic complications cannot be assessed with the Medicare database, EC use during open DTA and TAA aneurysm repair is associated with improved late survival and a significant reduction in operative mortality, morbidity, and procedural costs. These data indicate that EC should be a more widely applied adjunct in open DTA or TAA aneurysm repair. (J Vasc Surg 2018;-:1-7.) Keywords: Thoracoabdominal aortic aneurysm; Descending thoracic aortic aneurysm; Extracorporeal circulation; Atrial-femoral bypass
Aneurysmal dilation of the descending thoracic aorta (DTA) and thoracoabdominal aorta (TAA) most commonly results from primary cystic medial necrosis or secondary degeneration after type B dissection
and is classified on the basis of the extent of aortic involvement.1 Simple “clamp and sew” (CS) repair as championed by Crawford et al1 emphasizes expediency as the aorta is clamped and replaced with synthetic
From the Division of Vascular and Endovascular Surgery, Massachusetts
Correspondence: Virendra I. Patel, MD, MPH, Columbia University Medical
General Hospital, Harvard Medical School, Bostona; the Division of Vascular
Center, Herbert Irving Pavilion, 161 Fort Washington Ave, Fifth Fl, Ste 532,
Surgery, Steward Medical Group, St. Elizabeth’s Medical Center, Brightonb;
New York, NY 10032 (e-mail: [email protected]).
and the Division of Vascular Surgery and Endovascular Interventions,
The editors and reviewers of this article have no relevant financial relationships to
Columbia University Medical Center, Columbia University College of Physi-
disclose per the JVS policy that requires reviewers to decline review of any
cians and Surgeons, New York.c Author conflict of interest: none. Plenary presentation at the Forty-first Annual Symposium of the Society for Clinical Vascular Surgery, Miami, Fla, March 12-16, 2013.
manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2018 by the Society for Vascular Surgery. Published by Elsevier Inc. https://doi.org/10.1016/j.jvs.2017.12.072
1
2
Journal of Vascular Surgery
Mohebali et al
---
graft in a proximal to distal fashion.2 Thoracic aortic cross-clamping, however, results in profound physiologic derangement as distal tissues become ischemic. Furthermore, left ventricular preload and afterload rise acutely, resulting in increased left ventricular wall stress, stroke work, and, in turn, myocardial oxygen demand. Ischemia may result in postoperative liver dysfunction, gut necrosis, renal failure, and most devastatingly, paraplegia/ paralysis. Indeed, in his benchmark experience, Crawford found that cross-clamp time remained a significant predictor of early death even after multivariable adjustment.1 Consequently, a multitude of intraoperative adjuncts designed either to maintain end-organ perfusion or to minimize ischemic tissue metabolic demand during cross-clamping have been employed. Visceral perfusion can be achieved with a pump-powered cannula perfuser or by direct in-line (passive) mesenteric shunting.3-5 Complete or selective reimplantation of intercostal and lumbar vessels can be used to improve cord perfusion. As cross-clamping also increases cerebrospinal fluid (CSF) pressure and consequently decreases spinal perfusion pressure, CSF drainage can be used to improve cord perfusion and is the only paraplegia-mitigating adjunct supported by level I evidence.6 Finally, improved global perfusion distal to the cross-clamp can be achieved with the use of mechanical circulatory devices.6 These extracorporeal circulation (EC) techniques range from left atrial-femoral bypass to full cardiopulmonary bypass. Irrespective of technique, EC provides the added advantage of mitigating the increase in preload and afterload associated with proximal cross-clamping. Some studies, however, have demonstrated no significant advantage with EC compared with the simple CS technique,7 and many authors have noted that the increased intraoperative technical complexity associated with EC is not worth the perceived benefit. It becomes clear, then, that there is no standardized approach for the repair of DTA and TAA aneurysms, with the literature revealing conflicting data. Furthermore, most existing studies are limited to single institutions or single-surgeon series, which may fail to account for the various adjuncts and practice patterns, confounding the outcomes associated with use of CS vs EC. The purpose of this study, therefore, was to evaluate the effects of EC on open repair of DTA and TAA aneurysms across multiple institutions and surgeons using the U.S. Medicare database.
METHODS Study design and cohort. This retrospective cohort study of the Medicare population targeted Medicare patients undergoing intact open repair of DTA aneurysm or TAA aneurysm from 2004 to 2007. The Medicare Provider Analysis and Review specifically covered Medicare Part A claims and included a linkage to Vital
2018
ARTICLE HIGHLIGHTS d
d
d
Type of Research: Retrospective analysis of prospectively collected Medicare data Take Home Message: In 4230 patients receiving open descending thoracic or thoracoabdominal aortic aneurysm repair, use of extracorporeal circulation in 2433 patients resulted in lower major complication rate (49% vs 58%), shorter hospital stay, and lower total hospital charges. Recommendation: This study suggests that extracorporeal circulation should be used more frequently during open repair of descending thoracic and thoracoabdominal aortic aneurysms.
Statistics for mortality and long-term survival analysis. International Classification of Diseases, Ninth Revision (ICD-9) codes (Appendix) pertaining to nonruptured, open thoracic aortic repair were used to broadly identify the initial cohort. ICD-9 codes were then used to exclude patients lacking a diagnosis of DTA or TAA aneurysm and patients undergoing concomitant cardiac procedures including coronary revascularization, valve procedures, or procedures employing cardioplegia or deep hypothermic circulatory arrest. The remaining cohort was divided into two groups according to ICD-9 codes reflecting use of EC (ICD-9 3961) vs those that did not. Primary outcomes evaluated included 30-day mortality, long-term survival, and predictors of death or late mortality. Secondary outcomes included hospital and intensive care unit length of stay, systemic complications, and predictors of systemic complications. Acute renal failure was defined by ICD-9 code (Appendix). Institutional Review Board approval was obtained for the study. The need for informed consent of the patient was waived by the Institutional Review Board, given the use of deidentified administrative data for the study. Statistical analysis. All data were analyzed using the SAS statistical software package (version 9.3; SAS Institute, Cary, NC). A univariate cutoff of 0.1 was used for initial inclusion in multivariable models. Stepwise selection was used with an entrance and exit cutoff of .05 to prune the full models to arrive at a final, adjusted, multivariable model. Kaplan-Meier analysis was performed between the EC and no EC groups to determine longterm survival, with the log-rank test used to assess for significance. Cox proportional hazards regression was used to model survival after adjusting for significant covariates and confounders.
RESULTS Cohort characteristics. There were 18,282 patients identified as having nonruptured open thoracic aortic
Journal of Vascular Surgery Volume
-,
Number
Mohebali et al
3
-
aneurysm repairs in the U.S. Medicare population between 2004 and 2007. From these 18,282 patients, 10,658 were excluded for involving simultaneous cardiac revascularization, cardioplegia, or valve procedures; 116 were excluded for involving deep hypothermic circulatory arrest, and 3278 were excluded because they were lacking a diagnosis of DTA or TAA aneurysm. After these exclusions, the entire cohort of open DTA or TAA aneurysm repairs was composed of 4230 procedures. These procedures were then grouped into an EC group consisting of 2433 (57%) procedures and a no EC group consisting of 1797 (43%) procedures. Differences in baseline clinical features between patients who underwent procedures with or without EC are displayed in Table I. The EC group maintained a higher percentage of female patients and dissections (1.3%), whereas the no EC group included a higher percentage of TAA patients (64%) exhibiting chronic obstructive pulmonary disease, peripheral vascular disease, and chronic kidney disease (Table I). Perioperative outcomes. Use of EC resulted in a significantly reduced 30-day mortality rate (9.7% vs 12.2%; P ¼ .01) and a significantly reduced overall complication rate (49% vs 58%; P < .0001; Table II). The no EC group had a significantly higher increase in pulmonary complications (27% vs 21%; P < .0001), bleeding (19% vs 13%; P < .0001), and acute renal failure (24% vs 14%; P < .0001; Table II). EC resulted in a significantly shorter length of stay both in the intensive care unit (5 days vs 6 days; P < .0001) and in the hospital (9 days vs 11 days; P < .0001). In addition, significantly larger portions of patients with EC use during the procedure were discharged home vs discharged to an extended care facility (67% vs 56%; P < .0001). Multivariable regression modeling showed EC to independently reduce operative mortality (odds ratio [OR], 0.80; 95% confidence interval [CI], 0.65-0.97; P ¼ .02), any complication (OR, 0.67; 95% CI, 0.59-0.76; P < .0001), pulmonary complications (OR, 0.68; 95% CI, 0.59-0.79; P < .0001), and acute renal failure (OR, 0.52; 95% CI, 0.44-0.61; P < .0001) in adjusting for baseline clinical differences (Table III). Total hospital charges for patients with EC use were reduced compared with the total charge for patients who did not have EC ($110,000 vs $120,000; P < .0001; Table IV). Long-term survival. Cox proportional hazards regression showed that EC was independently associated with improved long-term survival (hazard ratio [HR], 0.69; 95% CI, 0.63-0.74; P < .0001), whereas postoperative complications were independently associated with worsened long-term survival (HR, 2.4; 95% CI, 2.1-2.7; P < .0001; Table V). Kaplan-Meier analysis of long-term survival showed a significantly (P < .01) higher survival in patients who had EC used during their procedure at both 1 year (81% vs 73%) and 5 years (67% vs 52%; Fig).
Table I. Clinical features of the study cohorts Variable
No EC (n ¼ 1797)
EC (n ¼ 2433)
72 6 8
73 6 9
Age, years, mean 6 SD Female Hypertension Coronary artery disease Cerebrovascular disease
.002
47
53
53
56
.02
24
23
.7
4.4
Chronic obstructive pulmonary disease
P value
34
5.4 28
.001
.17 <.0001
Peripheral vascular disease
11.3
5.7
<.0001
Chronic kidney disease
7.7
5.5
.003
EC, Extracorporeal circulation; SD, standard deviation. Values are reported as percentage unless otherwise indicated. Boldface P values represent statistical significance (P < .05).
Table II. Perioperative outcomes by cohort Variable
No EC (n ¼ 1797), %
30-Day mortality
12.2
Any complication
58
EC (n ¼ 2433), % 9.7 49
P value .01 <.0001
Cardiac
13
14
Pulmonary
27
21
<.0001
Bleeding
19
13
<.0001
Infectious Acute renal failure
7.7 24
5.9 14
.16
.02 <.0001
EC, Extracorporeal circulation. Boldface P values represent statistical significance (P < .05).
Table III. Independent effect of extracorporeal circulation (EC) on outcomes Variable Death
EC OR (95% CI)
P value
0.8 (0.65-0.97)
.02
0.67 (0.59-0.76)
<.0001
Pulmonary complications
0.68 (0.59-0.79)
<.0001
Infectious complications
0.72 (0.57-0.92)
.009
Bleeding complications
0.61 (0.52-0.73)
<.0001
Renal complications
0.52 (0.44-0.61)
<.0001
Any complication
CI, Confidence interval; OR, odds ratio. Multivariable models risk adjusted for age, sex, race, coronary artery disease, cerebrovascular disease, chronic kidney disease, and peripheral vascular disease. Boldface P values represent statistical significance (P < .05).
DISCUSSION The strategy employed for open surgical repair of DTA and TAA aneurysms remains variable. Whereas Crawford’s benchmark study using CS techniques1 as well as Svensson’s follow-up experience with 1509 patients8 demonstrated mortality rates <10%, the authors noted that “vexing problems” including paraplegia and renal failure remained unacceptably high. Among Crawford’s patients, 30-day postoperative mortality and “neuromuscular deficit” (paraplegia/paralysis) were significantly
4
Journal of Vascular Surgery
Mohebali et al
---
2018
Table IV. Length of stay and hospital charges by study cohort No EC (n ¼ 1797), mean 6 SD (median)
EC (n ¼ 2433), mean 6 SD (median)
180,000 6 190,000 (120,000)
151,000 6 140,000 (110,000)
Variable Total charges, $
14 6 13 (9)
P value <.0001
Hospital length of stay, days
17 6 18 (11)
<.0001
Intensive care unit length of stay, days
10 6 14 (6)
8 6 18 (5)
<.0001
Discharge to home
56%
67%
<.0001
EC, Extracorporeal circulation. Boldface P values represent statistical significance (P < .05).
Table V. Independent predictors of late mortality Variable EC
Death HR (95% CI)
P value
0.69 (0.63-0.74)
<.0001
Postoperative complications
2.4 (2.1-2.7)
<.0001
Cerebrovascular disease
2.4 (2.0-2.9)
<.0001
Chronic kidney disease
1.6 (1.4-2.0)
<.0001
Age (per year)
1.04 (1.03-1.04)
<.0001 <.0001
Chronic obstructive pulmonary disease
1.4 (1.2-1.5)
Peripheral vascular disease
1.3 (1.1-1.6)
.01
Nonwhite race
1.3 (1.1-1.5)
.002
Coronary artery disease
1.1 (1.01-1.3)
.04
EC, Extracorporeal circulation; HR, hazard ratio. Boldface P values represent statistical significance (P < .05).
associated with longer aortic cross-clamp times.1 Crawford additionally reported an overall paraplegia rate of 11%, noting an association with dissection and extent of disease. Indeed, the highest rates of paraplegia approaching 41% were seen among extent II patients with dissection.1 Crawford himself noted that “we have not been able to significantly prevent or reduce the incidence of these complications and have concentrated on rehabilitative therapy. There does not appear to be an immediate hope of solving this problem.”2 Furthermore, in consideration of distal perfusion adjuncts, the authors cited the work of Livesay et al,9 stating that “distal perfusion alone in a large number of cases subjected to extensive resection of the DTA has not altered the problem.” However, Livesay’s study used temporary bypasses and shunts to provide passive distal aortic perfusion rather than EC. Svensson’s later studies proved that left atrial-femoral bypass maintained blood supply to the spinal cord while the proximal anastomosis was being performed.8,10,11 Indeed, in their adoption of CSF drainage and distal aortic perfusion, Safi et al12 found that aortic cross-clamp time and even extent I disease were no longer significant predictors of neurologic deficit. In other words, EC essentially eliminated the cord ischemia associated with proximal descending thoracic aortic cross-clamping. These results laid the foundation for the theory that the spinal cord is
perfused by a collateral network of vessels supplied by multiple aortic branches. The collateral network concept was later confirmed by Backes et al13 and Jacobs et al14 using magnetic resonance angiography and by Griepp and Griepp,15 who formalized the concept by citing the increasing body of literature suggesting that routine sacrifice of segmental aortic branches perfusing the cord could be carried out as long as overall inflow to the network was maintained, as occurs with distal aortic perfusion. With regard to visceral perfusion, Safi’s group later demonstrated that distal aortic and consequently visceral perfusion also mitigates postoperative laboratory evidence of hepatic and visceral ischemia during open extent II aneurysm repair.5 Welborn et al16 similarly evaluated levels of plasma proinflammatory cytokines associated with multisystem organ injury, dysfunction, and critical illness among patients undergoing open TAA aneurysm repair. The authors clearly demonstrated that use of left atrial-femoral bypass significantly reduced levels of these factors to values seen in simple open infrarenal aortic aneurysm repair.16 Finally, Jacobs et al17 showed that renal and visceral ischemia could be significantly reduced by continuous perfusion during proximal aortic cross-clamping with consequent lower rates of postoperative renal failure. At our institution, the practice has similarly evolved from a CS technique18 using adjuncts such as epidural cooling, renal hypothermia, and CSF drainage with an aggressive intercostal reimplantation policy2 to one employing these adjuncts (without epidural cooling) in addition to distal perfusion through left atrial-femoral bypass and motor evoked potential monitoring (MEVP)19 for extent I to III thoracoabdominal aneurysms. Using a propensity matched analysis in a small cohort of these patients, we demonstrated that use of EC and MEVP resulted in lower early mortality and paraplegia rates compared with a CS technique.6 Our follow-up series with a larger cohort (n ¼ 485) demonstrated half the early postoperative mortality rate, a substantial decrease in acute renal failure requiring hemodialysis, and a profound decrease in permanent spinal cord injury for the EC/MEVP group compared with CS.19 In contrast, Coselli et al7 used a propensity matched
Journal of Vascular Surgery Volume
-,
Number
Mohebali et al
5
-
Fig. Kaplan-Meier curves for long-term survival after thoracic or thoracoabdominal aortic aneurysm repair by use of extracorporeal circulation (EC) and clamp and sew (CS) techniques.
multivariable logistic regression analysis to match patients on the basis of likelihood of receiving EC vs CS and found that after matching, there was no statistically significant difference in rates of 30-day mortality, paraplegia, or renal failure. We have found this to be true for our experience with type IV aneurysms only, as perioperative mortality and complication rates in these select patients are similar to open abdominal aortic aneurysm repair.20 Whereas our experience as well as the work of others clearly supports the use of EC for open repair of extent I to III thoracoabdominal aneurysm, controversy persists in the literature regarding the optimal surgical technique.7,21 Results are further confounded by the presence or absence of dissection, need for emergent repair,
extent of aneurysmal degeneration, use of non-EC adjuncts such as MEVP, spinal drainage, renal hypothermia, in-line mesenteric shunting, extent of intercostal reconstruction, and other surgeon-specific or institutionspecific practices that may not be captured in large databases. It becomes imperative, then, to consider the effects of EC across multiple institutions, populations of patients, practice patterns, and surgeons. Although a randomized controlled trial would certainly be ideal in helping to disentangle the benefit of EC from other adjuncts, to our knowledge, no such study exists as confirmed by the most recent Cochrane database review on the topic.22 The goal of this study, then, was to determine the effects of EC on surgical outcomes across multiple institutions and surgeons using the U.S. Medicare
6
Journal of Vascular Surgery
Mohebali et al
---
database. To our knowledge, this is the largest study examining the effects of EC during open repair of DTA aneurysm and TAA aneurysm to date. In this study, use of EC during open DTA aneurysm and TAA aneurysm repair was shown to have a favorable impact on perioperative outcomes, financial outcomes, and late outcomes. EC was beneficial in reducing perioperative mortality, pulmonary and renal morbidity, and length of stay and costs while allowing home discharge and improving late survival. Specifically, rates of renal failure for the EC group were reduced from 24% to 14%. Estrera et al23 routinely used EC and reported rates of renal dysfunction approaching 24%, whereas Coselli et al7 espoused no benefit with the use of EC as their rates of renal failure were not significantly different for the EC (7%) vs non-EC groups (7.6%). Our previous work demonstrated a trend toward lower rates of renal failure in the EC group (5.1% vs 11.4%); however, the results were not significant (P ¼ .063).19 As discussed previously, this variability could be the result of additional adjuncts, such as shunting and use of renal hypothermia or preservation fluid, that mitigate the effects of ischemia. Furthermore, variability in these results may be because of baseline levels of preoperative renal dysfunction, which multiple authors, including us, have demonstrated to be the greatest predictor of postoperative renal dysfunction. Nevertheless, this study shows that EC provides a clear benefit for postoperative renal function. With regard to 30-day mortality, this study shows that EC provided a clear benefit (9.7% vs 12.2%; P ¼ .01), which is similar to our previous work (4.0% vs 9.9%).19 Furthermore, whereas our study did not show significant differences in length of intensive care unit or hospital stay, the current study found a clear reduction in both of these variables for the EC group. Finally, the study results provide novelty in analyzing the effects of EC use on total hospital charges, which were significantly reduced for the EC group. Similarly, long-term mortality results of the study that demonstrate a protective effect of EC (OR, 0.69; 5-year survival of 67% 6 1% for EC vs 52% 6 1%) mirror our previous work demonstrating improved survival for left atrial-femoral bypass patients (4-year survival of 73% 6 6% for EC vs 60% 6 3%)19 after adjusting for other significant predictors of long-term mortality (HR, 0.590). These results are similar to those of other large series at centers of excellence.12,24 Therefore, unless it is contraindicated for technical or clinical reasons, these data support the growing body of literature indicating that EC should be more widely applied as an adjunct during open DTA aneurysm and TAA aneurysm repair.
2018
Medicare database did not allow us to account for the various additional adjuncts (renal hypothermia, mesenteric shunting, epidural cooling, MEVP) that may have been employed during these procedures. It is certainly possible that use of these adjuncts could confound results. Furthermore, the database lacks information on extent of aneurysm (I-IV), which has been predictive of spinal cord ischemia,2,3,8,25 postoperative renal failure,26 and death.3 In addition, the greatest benefit of EC has been the profound decrease in postoperative rates of spinal cord ischemia as demonstrated by multiple studies.4,6,19 Unfortunately, we were not able to evaluate for this devastating outcome as it is lacking in the database. Our previous work and the most recent work of Estrera et al,23 however, demonstrated that postoperative paraplegia/paralysis certainly increases longterm mortality, and it is therefore conceivable that the survival benefit noted in the EC group in this study could have been at least in part the result of lower rates of irreversible spinal cord ischemia. Finally, as with any retrospective study, there is undoubtedly unmeasured confounding in the form of selection bias pertaining to those patients for whom EC was chosen as an integral part of the repair. As noted before, the literature suggests that some surgeons or institutions preferentially employ EC in urgent cases, whereas some who routinely use EC may abandon it completely in the setting of emergent cases with rupture in favor of expedient salvage with CS (no EC) techniques. We did attempt to minimize the risk of this selection bias by excluding ruptured or emergent cases. Another source of clear selection bias is the practice of most large centers to selectively employ EC for more extensive aneurysm resections. We were unfortunately unable to adjust for the effects of this variable on outcomes as a result of database limitations as noted before. Finally, to the extent allowed by the Medicare database, we included known predictors of the postoperative outcomes analyzed in this study in a multivariable model to minimize the effects of selection bias.
CONCLUSIONS This study is the first to demonstrate a significant benefit in employing EC during DTA and TAA aneurysm repair across multiple institutions, surgeons, and populations of patients. Future databases and studies on the subject should seek to accurately capture and characterize the use of all intraoperative adjuncts, other patient and anatomic considerations that may affect the decision to employ EC, and all relevant postoperative outcomes and complications.
AUTHOR CONTRIBUTIONS Limitations and future studies. Whereas this study suggests that employing EC during DTA and TAA aneurysm repair has beneficial effects on mortality and postoperative renal failure, the inherent limitations of the U.S.
Conception and design: VP Analysis and interpretation: JM, RL, EE, MC, WC, RC, VP Data collection: SC Writing the article: JM, SC, VP
Journal of Vascular Surgery Volume
-,
Number
Mohebali et al
7
-
Critical revision of the article: JM, RL, EE, MC, WC, RC, VP Final approval of the article: JM, SC, RL, EE, MC, WC, RC, VP Statistical analysis: EE, VP Obtained funding: Not applicable Overall responsibility: VP
REFERENCES 1. Crawford ES, Crawford JL, Safi HJ, Coselli JS, Hess KR, Brooks B, et al. Thoracoabdominal aortic aneurysms: preoperative and intraoperative factors determining immediate and long-term results of operations in 605 patients. J Vasc Surg 1986;3:389-404. 2. Conrad MF, Crawford RS, Davison JK, Cambria RP. Thoracoabdominal aneurysm repair: a 20-year perspective. Ann Thorac Surg 2007;83:S856-61. 3. Coselli JS, Bozinovski J, LeMair SA. Open surgical repair of 2286 thoracoabdominal aortic aneurysms. Ann Thorac Surg 2007;83:S862-4. 4. Patel VI, Lancaster RT, Conrad MF, Cambria RP. Open surgical repair of thoracoabdominal aneurysmsdthe Massachusetts General Hospital experience. Ann Cardiothorac Surg 2012;1:320-4. 5. Safi HJ, Miller CC 3rd, Yawn DH, Iliopoulos DC, Subramaniam M, Harlin S, et al. Impact of distal aortic and visceral perfusion on liver function during thoracoabdominal and descending thoracic aortic repair. J Vasc Surg 1998;27:145-52. 6. Conrad MF, Ergul EA, Patel VI, Cambria MR, LaMuraglia GM, Simon M, et al. Evolution of operative strategies in open thoracoabdominal aneurysm repair. J Vasc Surg 2011;53:1195-201. 7. Coselli JS, LeMair SA, Conklin LD, Adams GJ. Left heart bypass during descending thoracic aortic aneurysm repair does not reduce the incidence of paraplegia. Ann Thorac Surg 2004;77:1298-303. 8. Svensson LG, Crawford ES, Hess KR, Coselli JS, Safi HJ. Experience with 1509 patients undergoing thoracoabdominal aortic operations. J Vasc Surg 1993;17:357-68. 9. Livesay JJ, Cooley DA, Ventemiglia RA, Montero CG, Warrian RK, Brown DM, et al. Surgical experience in descending thoracic aneurysmectomy with and without adjuncts to avoid ischemia. Ann Surg 1985;39:37-44. 10. Svensson LG, Patel V, Robinson MF, Ueda T, Roehm JO Jr, Crawford ES. Influence of preservation or perfusion of intraoperatively identified spinal cord blood supply on spinal motor evoked potentials and paraplegia after aortic surgery. J Vasc Surg 1991;13:355-65. 11. Svensson LG, Crawford ES, Patel V, McLean TR, Jones JW, DeBakey ME. Spinal cord oxygenation, intraoperative blood supply localization, cooling and function with aortic clamping. Ann Thorac Surg 1992;54:74-9. 12. Safi HJ, Miller CC 3rd, Huynh TT, Estrera AL, Porat EE, Winnerkvist AN, et al. Distal aortic perfusion and cerebrospinal fluid drainage for thoracoabdominal and descending thoracic aortic repair: ten years of organ protection. Ann Surg 2003;238:372-80. 13. Backes WH, Nijenhuis RJ, Mess WH, Wilmink FA, Schurink GW, Jacobs MJ. Magnetic resonance angiography of collateral blood supply to spinal cord in thoracic and thoracoabdominal aortic aneurysm patients. J Vasc Surg 2008;48:261-71. 14. Jacobs MJ, de Mol BA, Elenbaas T, Mess WH, Kalkman CJ, Schurink GW, et al. Spinal cord blood supply in patients with thoracoabdominal aortic aneurysm. J Vasc Surg 2002;35:30-7.
15. Griepp EB, Griepp RB. The collateral network concept: minimizing paraplegia secondary to thoracoabdominal aortic aneurysm resection. Tex Heart Inst J 2010;37:672-4. 16. Welborn MB, Oldenburg HS, Hess PJ, Huber TS, Martin TD, Rauwerda JA, et al. The relationship between visceral ischemia, proinflammatory cytokines, and organ injury in patients undergoing thoracoabdominal aortic aneurysm repair. Crit Care Med 2000;28:3191-7. 17. Jacobs MJ, Eijsman L, Meylaerts SA, Balm R, Legemate DA, de Haan P, et al. Reduced renal failure following thoracoabdominal aortic aneurysm repair by selective perfusion. Eur J Cardiothorac Surg 1998;14:201-5. 18. Cambria RP. Thoracoabdominal aortic aneurysm repair: how I do it. Cardiovasc Surg 1999;7:597-606. 19. Lancaster RT, Conrad MF, Patel VI, Cambria MR, Ergul EA, Cambria RP. Further experience with distal aortic perfusion and motor-evoked potential monitoring in the management of extent I-III thoracoabdominal aortic aneurysms. J Vasc Surg 2013;58:283-90. 20. Patel VI, Ergul E, Conrad MF, Cambria M, LaMuraglia GM, Kwolek CJ, et al. Continued favorable results with open surgical repair of type IV thoracoabdominal aortic aneurysms. J Vasc Surg 2011;53:1492-8. 21. Safi HJ, Miller CC 3rd, Subramaniam MH, Campbell MP, Iliopoulos DC, O’Donnell JJ, et al. Thoracic and thoracoabdominal aortic aneurysm repair using cardiopulmonary bypass, profound hypothermia, and circulatory arrest via left side of the chest incision. J Vasc Surg 1998;28:591-8. 22. Hsu CC, Kwan GN, van Driel ML, Rophael JA. Distal aortic perfusion during thoracoabdominal aneurysm repair for prevention of paraplegia. Cochrane Database Syst Rev 2012;3:CD008197. 23. Estrera AL, Sandhu HK, Charlton-Ouw KM, Afifi RO, Azizzadeh A, Miller CC 3rd, et al. A quarter century of organ protection in open thoracoabdominal repair. Ann Surg 2015;262:660-8. 24. Safi JH, Estrera AL, Miller CC, Huynh TT, Porat EE, Azizzadeh A, et al. Evolution of risk for neurologic deficit after descending and thoracoabdominal aortic repair. Ann Thorac Surg 2005;80:2173-9. 25. Wynn MW, Acher CW. A modern theory of spinal cord ischemia/injury in thoracoabdominal aortic surgery and its implications for prevention of paralysis. J Cardiothorac Vasc Anesth 2014;28:1088-99. 26. Attab M, Coselli JS. Renal and visceral protection in thoracoabdominal aortic surgery. J Thorac Cardiovasc Surg 2014;148:2963-6.
Submitted Sep 25, 2017; accepted Dec 21, 2017.
APPENDIX International Classification of Diseases, Ninth Revision (ICD-9) codes used in the study. Open thoracic aortic repair: 3845 Circulatory arrest: 39.63 Valve replacement: 3500-3504, 3510-3514, 3520-3528 Coronary artery bypass graft: 3610-3617 Extracorporeal circulation: 3961 Acute renal failure: 5845-5849, 586 Thoracoabdominal aneurysm: 4412 Dissection: 4410, 44100-44103