Stratified log-rank P = .004. Bonferroni-corrected significance at P < 1.25 × 10−5. Shading indicates 95% confidence bands.
aNumber at risk value or difference between adjacent values is less than 6. Value is represented as a range to eliminate patient reidentification risk, according to mandatory Institute for Clinical Evaluative Sciences practice.
Stratified log-rank P = 1.06 × 10−7. Bonferroni-corrected significance set at P < 1.25 × 10−5. Shading indicates 95% confidence bands.
aNumber at risk value or difference between adjacent values is less than 6. Value is represented as a range to eliminate patient re-identification risk, according to mandatory Institute for Clinical Evaluative Sciences practice.
Stratified Gray test P = .68. Bonferroni-corrected significance set at P < 1.25 × 10−5. Shading indicates 95% confidence bands.
Stratified Gray test P = .69. Bonferroni-corrected significance set at P < 1.25 × 10−5. Shading indicates 95% confidence bands.
eTable 1. Summary of Institute for Clinical Evaluative Sciences Databases Used
eTable 2. Covariate and Outcome Definitions and Codes
eTable 3. Survival and Cumulative Incidence Rates in Unmatched and Matched Cohorts
eTable 4. Hazard Ratios From Cox Proportional Hazards Models for Matched and Inverse Probability of Treatment–Weighted Cohorts
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Salata K, Hussain MA, de Mestral C, et al. Comparison of Outcomes in Elective Endovascular Aortic Repair vs Open Surgical Repair of Abdominal Aortic Aneurysms. JAMA Netw Open. 2019;2(7):e196578. doi:10.1001/jamanetworkopen.2019.6578
¿Cuáles son los resultados a largo plazo de la reparación endovascular aórtica electiva de los aneurismas aórticos abdominales en comparación con los de la reparación quirúrgica abierta?
En este estudio de cohortes de base poblacional que incluyó a 17 683 pacientes que recibieron tratamiento electivo para aneurismas aórticos abdominales, no se hallaron diferencias estadísticamente significativas en la mortalidad por todas las causas a largo plazo entre la reparación aórtica endovascular y la reparación quirúrgica abierta durante un seguimiento máximo de 13,8 años.
La reparación aórtica endovascular no se asoció con un beneficio de supervivencia a largo plazo.
Knowledge regarding the long-term outcomes of elective treatment of abdominal aortic aneurysm (AAA) using endovascular aortic repair (EVAR) is increasing. However, data with greater than 10 years’ follow-up remain sparse and are lacking from population-based studies.
To determine the long-term outcomes of EVAR compared with open surgical repair (OSR) for elective treatment of AAA.
Design, Setting, and Participants
This retrospective, population-based cohort study used linked administrative health data from Ontario, Canada, to identify all patients 40 years and older who underwent elective EVAR or OSR for AAA repair from April 1, 2003, to March 31, 2016, with follow-up terminating on March 31, 2017. A total of 17 683 patients were identified using validated procedure and billing codes and were propensity score matched. Analysis was conducted from June 26, 2018, to January 16, 2019.
Elective EVAR or OSR for AAA.
Main Outcomes and Measures
The primary outcome was overall survival. Secondary outcomes were major adverse cardiovascular event–free survival, defined as being free of death, myocardial infarction, or stroke; reintervention; and secondary rupture.
Among 17 683 patients who received elective AAA repairs (mean [SD] age, 72.6 [7.8] years; 14 286 [80.8%] men), 6100 (34.5%) underwent EVAR and 11 583 (65.5%) underwent OSR. From these patients, 4010 well-balanced propensity score–matched pairs of patients were defined, with a mean (SD) age of 73.0 (7.6) years and 6583 (82.1%) men. In the matched cohort, the mean (SD) follow-up was 4.4 (2.7) years, and maximum follow-up was 13.8 years. The overall median survival was 8.9 years. Compared with OSR, EVAR was associated with a higher survival rate up to 1 year after repair (91.0% [95% CI, 90.1%-91.9%] vs 94.0% [95% CI, 93.3%-94.7%]) and a higher major adverse cardiovascular event–free survival rate up to 4 years after repair (69.9% [95% CI, 68.3%-71.3%] vs 72.9% [95% CI, 71.4%-74.4%]). Cumulative incidence of reintervention was higher among patients who underwent EVAR compared with those who underwent OSR at the 7-year follow-up (45.9% [95% CI, 44.1%-47.8%] vs 42.2% [95% CI, 40.4%-44.0%]). Survival analyses demonstrated no statistically significant differences in long-term survival, reintervention, and secondary rupture for patients who underwent EVAR compared with those who underwent OSR. Kaplan-Meier analysis suggested superior long-term major adverse cardiovascular event–free survival among patients who underwent EVAR compared with those who underwent OSR (32.6% [95% CI, 26.9%-38.4%] vs 14.1% [95% CI, 4.0%-30.4%]; stratified log-rank P < .001) during a maximum follow-up of 13.8 years.
Conclusions and Relevance
Endovascular aortic repair was not associated with a difference in long-term survival during more than 13 years’ maximum follow-up. The reasons for these findings will require studies to consider specific graft makes and models, adherence to instructions for use, and types and reasons for reintervention.
Endovascular aortic repair (EVAR) has changed the landscape of abdominal aortic aneurysm (AAA) treatment since its introduction in 1991.1 Randomized and population-based studies investigating EVAR vs open surgical repair (OSR) demonstrated superior perioperative survival as well as significant improvements in operative time, blood loss, transfusion requirements, cardiopulmonary complications, and reduced lengths of stay in intensive care units and hospitals in favor of EVAR.2-4 Consequently, EVAR has seen rapid uptake and has become the predominant approach to AAA management, as borne out by multiple Canadian,5 US,6,7 and European8 studies.
Despite its clear short-term superiority, early EVAR technology had unique complications, including endoleak, graft migration, graft thrombosis, and secondary rupture.9-12 Accordingly, patients who undergo EVAR require lifelong surveillance to identify and prevent these complications, and EVAR is associated with significantly higher reintervention rates.13 These are at least in part postulated to be the reasons for the eventual loss of the EVAR mortality benefit demonstrated in EVAR randomized clinical trials (RCTs).14-17 However, since the inception of these trials in the 1990s and early 2000s, EVAR endografts have undergone numerous manufacturing reiterations to address identified endograft-related complications. As a result, the longevity of the mortality benefit of contemporary EVAR may be underestimated by these trials. On the other hand, limited data are available to elucidate whether EVAR is safe in the long term. Presently, the Endovascular Aortic Repair 1 (EVAR-1) trial17 and Dutch Randomized Endovascular Aneurysm Management (DREAM)18 trial are the only trials to report longer than 10-year follow-up, and no population-based studies have been conducted with longer than 10 years of follow-up, to our knowledge.
This study aims to add to the limited literature regarding the long-term outcomes of EVAR. Specifically, the objective of this study is to assess the differences between EVAR and OSR for elective AAA repair in long-term survival, major adverse cardiovascular event (MACE)–free survival, reintervention, and secondary rupture. This developing body of knowledge will serve to better delineate the true mortality benefit of EVAR, considering accumulated experience, evolved endografts, and better knowledge of complications and their appropriate management.
The population of Ontario, Canada, is 14.2 million people. A retrospective, population-based cohort study of EVAR vs OSR for elective AAA management in Ontario, Canada, was performed using administrative health care data, in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies. The Ontario Ministry of Health and Long-Term Care records each publicly insured ambulatory, emergency, and inpatient health care system interaction requiring the use of an Ontario health card. The Institute for Clinical Evaluative Sciences, a prescribed entity governed under the Personal Health Information Protection Act, stores and manages these data. Data are linked together across primary sources using a unique encrypted identifier known as the Institute for Clinical Evaluative Sciences key number. Ontario administrative data are routinely used for population-level research and have previously been thoroughly described and validated.19-22 The specific databases used for this study are described in eTable 1 in the Supplement. This study was approved by the research ethics board at Sunnybrook Health Sciences Centre in Toronto, Ontario, Canada, and the requirement for informed consent was waived owing to the use of deidentified secondary data.
All elective EVARs and OSRs of AAA performed in Ontario, Canada, in patients 40 years and older from April 1, 2003, to March 31, 2016, with maximum follow-up terminating on March 31, 2017, were identified. This age cutoff was used to reduce the likelihood of contamination with connective tissue aneurysms. Combinations of the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Canadian Revision; Canadian Classification of Health Intervention; and Ontario Health Insurance Plan diagnostic, procedure, and billing codes were used according to a previously validated algorithm.22 This algorithm identified EVAR (infrarenal) and OSR (infrarenal, pararenal, and juxtarenal) with 96% and 95% positive predictive values, respectively. Patients with multiple AAA repair procedures listed on their index admission were excluded owing to an inability to establish appropriate event chronology and, thus, the primary procedure. Patients who relocated outside of Ontario or whose health cards expired prior to study conclusion were also excluded.
Baseline covariates included demographic characteristics, health care utilization, hospital and procedure characteristics, comorbidities, medications, and tracer variables (used to help demonstrate absence of residual confounding). A full list of covariates, definitions, corresponding codes, and references to respective validation studies is presented in eTable 2 in the Supplement. All baseline covariates were captured using a 5-year lookback window from the index procedure date, except those captured from the National Ambulatory Care Reporting System database, where a 3-year lookback window was used, as this database commenced in July 2000. All medication variables, except for fluoroquinolones, were measured using a 4-month lookback window to assess whether at least 1 prescription was recently filled, considering a 100-day maximum medication dispensing limit. In the province of Ontario, patients younger than 65 years are not eligible for publicly funded pharmaceutical care; as such, drug information is not captured for these individuals within administrative data. Patients younger than 65 years without drug information were coded as ineligible for the Ontario Drug Benefit program.
The primary outcome was overall survival. Secondary outcomes included MACE-free survival, defined as survival free of a composite of death, stroke, or myocardial infarction; any reintervention; and secondary rupture. These outcomes were defined according to previously validated codes where possible (eTable 2 in the Supplement). Patients were observed through administrative databases until the outcome of interest or end of the follow-up period on March 31, 2017, whichever happened first.
A propensity score–matched survival analysis was performed.23,24 The propensity score for repair approach was calculated using a logistic regression model incorporating all covariates as potential confounders. Patients who received EVAR or OSR were matched 1:1 using the greedy nearest neighbor method with a caliper width of 0.2 SD units. Balance of covariates was assessed using standardized differences, with differences less than 0.1 indicating good balance.25 Residual confounding was assessed using the distribution of tracer variables not used to specify the propensity score. Survival analyses were conducted using the Kaplan-Meier product limit method to assess differences in survival and MACE-free survival. Statistical differences were assessed using the stratified log-rank test with stratification on the matched pairs. To account for competing risks, cumulative incidence function analysis using a Fine and Gray model and stratified Gray test were conducted to assess differences in reintervention and secondary rupture. Statistical significance was set at a Bonferroni-corrected P value of 1.25 × 10−5 to adjust for the 4010 matched-pair comparisons.26,27 Survival and cumulative incidence rates and associated 95% CIs were obtained from corresponding survival and cumulative incidence functions at 30 days, annually for up to 10 years, as well as at maximum follow-up.
The calculated propensity score was used to conduct an inverse probability of treatment–weighted sensitivity analysis.28 The mean treatment effect in the treated weight was used, and balance of covariates was assessed using standardized mean differences as well as variance ratios, with ratio values between 0.5 and 2.0 indicating good balance.29,30 Cox regression models with a robust sandwich variance estimator were then used to calculate hazard ratios (HRs) for each of the outcomes.31 Cause-specific Cox models were fit for the reintervention and secondary rupture outcomes to account for competing risks. Proportionality of hazards was investigated using log-log survival plots and time-dependent covariates. Where hazards were not proportional, data were partitioned into multiple time intervals, and separate Cox models were built for each interval. Cutoff values were determined using plots of Schoenfeld residuals through time and tested as previously described. Statistical significance was set at a 2-tailed P value less than .05. All statistical analyses were conducted in SAS statistical software version 9.4 (SAS Institute). Analysis was conducted from June 26, 2018, to January 16, 2019.
The overall study population consisted of 17 683 patients who received elective AAA repairs (mean [SD] age, 72.6 [7.8] years; 14 286 [80.8%] men). Of these patients, 6100 (34.5%) underwent EVAR and 11 583 (65.5%) underwent OSR (Table). In the overall cohort, patients who underwent EVAR were more likely to be older, be men, reside in areas with populations more than 10 000 people, and use the health care system more frequently. Furthermore, patients who underwent EVAR were more likely to have comorbid congestive heart failure, diabetes, hypertension, and a Charlson Comorbidity Index score of 2 or more, while patients who underwent OSR were more likely to have underlying peripheral arterial disease. Patients older than 65 years who underwent EVAR were more likely to be prescribed statin, angiotensin-converting enzyme inhibitor or angiotensin receptor blocker, anticoagulant, or noninsulin antidiabetes medications or prednisone and were more likely to have a history of fluoroquinolone use. Additionally, EVAR procedures within the unmatched cohort were more commonly conducted in teaching hospitals or high-volume hospitals (defined as ≥10 AAA procedures per year) and were more common than OSR after 2009. The propensity matched–cohort included 4010 pairs of patients (mean [SD] age, 73.0 [7.6] years; 6583 [82.1%] men) who underwent EVAR or OSR and did not demonstrate any significant differences in baseline covariates or tracer variables.
Mean (SD) follow-up for the entire cohort was 5.5 (3.6) years with a maximum follow-up of 14.0 years. Among the matched cohort, mean (SD) follow-up was 4.4 (2.7) years with a maximum follow-up of 13.8 years. Median survival for the unmatched cohort was 9.4 years overall, 7.8 years among patients who underwent EVAR, and 9.8 years among patients who underwent OSR. In the matched cohort, median survival was 8.9 years overall and in each treatment group.
Within the matched cohort, EVAR was associated with a higher survival rate than OSR for up to 1 year after repair (94.0% [95% CI, 93.3%-94.7%] vs 91.0% [95% CI, 90.1%-91.9%]), and a higher MACE-free survival rate for up to 4 years after repair (72.9% [95% CI, 71.4%-74.4%] vs 69.9% [95% CI, 68.3%-71.3%]) (eTable 3 in the Supplement). The cumulative incidence of reintervention within 30 days was lower among patients who underwent EVAR compared with OSR (12.4% [95% CI, 11.4%-13.5%] vs 15.0% [95% CI, 13.9%-16.1%]) but higher after 7 years (45.9% [95% CI, 44.1%-47.8%] vs 42.2% [95% CI, 40.4%-44.0%]) (eTable 3 in the Supplement). No significant differences in the cumulative incidence of secondary rupture were demonstrated within the matched cohort at maximum follow-up (EVAR, 1.5% [95% CI, 0.9%-2.5%]; OSR, 0.8% [95% CI, 0.6%-1.2%]) (eTable 3 in the Supplement).
Kaplan-Meier survival analysis demonstrated no statistically significant differences between EVAR and OSR in long-term survival (at maximum follow-up: EVAR, 41.5% [95% CI, 37.7%-45.2%]; OSR, 26.9% [95% CI, 15.7%-39.5%]; stratified log-rank P = .004) (Figure 1). However, EVAR was associated with superior long-term MACE-free survival (at maximum follow-up: EVAR, 32.6% [95% CI, 26.9%-38.4%]; OSR, 14.1% [95% CI, 4.0%-30.4%]; stratified log-rank P = 1.06 × 10−7) (Figure 2). No statistically significant differences in long-term reintervention (at maximum follow-up: EVAR, 51.8% [95% CI, 48.8%-54.7%]; OSR, 49.6% [95% CI, 42.5%-56.3%]; stratified Gray test P = .68) (Figure 3) or secondary rupture (at maximum follow-up: EVAR, 1.5% [95% CI, 0.9%-2.5%]; OSR, 0.8% [95% CI, 0.6%-1.2%]; stratified Gray test, P = .69) were noted (Figure 4).
The inverse probability of treatment–weighted cohort was balanced on all covariates. Cox models demonstrated similar findings between the propensity score–matched cohort and inverse probability of treatment–weighted cohort (eTable 4 in the Supplement). Endovascular aortic repair was associated with lower risk of all-cause mortality within 45 days of the procedure (HR, 0.29; 95% CI, 0.22-0.40; P < .001). However, between 1 to 4 years’ follow-up, EVAR was associated with higher all-cause mortality (HR, 1.16; 95% CI, 1.02-1.32; P = .02). Similarly, EVAR was associated with a lower risk of MACE within 45 days of the procedure (HR, 0.32; 95% CI, 0.27-0.39; P < .001) but no subsequent differences outside of the perioperative period. The risk of reintervention within 30 days of repair was significantly lower among patients who underwent EVAR (HR, 0.74; 95% CI, 0.66-0.83; P < .001) but significantly higher from 30 days to 6 months (HR, 1.36; 95% CI, 1.12-1.65; P = .002) and after 2 years (HR, 1.11; 95% CI, 1.01-1.22; P = .03) (eTable 3 in the Supplement). There were no statistically significant differences in secondary rupture among the matched or sensitivity cohorts.
This population-based, retrospective cohort study involving 4010 pairs of propensity score–matched patients who underwent EVAR or OSR observed for longer than 13 years demonstrated no statistically significant differences in long-term survival, reintervention, or secondary rupture between patients who underwent EVAR vs those who underwent OSR. In contrast, EVAR was associated with higher MACE-free survival throughout follow-up.
The results of this study complement the limited existing long-term data comparing EVAR with OSR for elective AAA repair. To our knowledge, the longest reported follow-up of patients who underwent EVAR compared with those who underwent OSR to date comes from the EVAR-1 multicenter RCT17 of 1252 patients observed for a mean (SD) of 12.7 (1.5) years. That trial demonstrated superior all-cause mortality among patients who underwent EVAR for up to 6 months following repair, but this mortality benefit was lost after 6 months, and patients who underwent EVAR demonstrated a higher hazard for all-cause mortality after 8 years (adjusted HR [aHR], 1.25; 95% CI, 1.00-1.56; P = .048).17 Similarly, the 9-year results of the Standard Open Surgery Versus Endovascular Repair of AAA (OVER) RCT16 showed loss of the EVAR mortality benefit by 3 years (HR, 0.72; 95% CI, 0.51-1.00; P = .05). Additionally, in the longest and largest population-based follow-up of patients who underwent EVAR or OSR, to our knowledge, Schermerhorn et al32 demonstrated lower short-term mortality for patients who underwent EVAR within 30 days (aHR, 0.32; 95% CI, 0.29-0.35; P < .001) and from 30 to 90 days of repair (aHR, 0.64; 95% CI, 0.58-0.71; P < .001), with worse outcomes at up to 4 years (aHR, 1.17; 95% CI, 1.13-1.21; P < .001) and after 4 years (aHR, 1.05; 95% CI; 1.00-1.09; P = .03). These latter results were most congruent with the findings of our study.
With the accumulation of long-term data, it is becoming clearer that the superiority of EVAR may be limited to the short term. One prevalent explanation for this finding is the high reintervention rate associated with EVAR. The EVAR-1 trial17 reported a 15-year reintervention rate of 26% in the EVAR group and an aHR of 6.29 (95% CI, 3.09-12.78; P < .001) for any reintervention between 6 months and 4 years after EVAR, while the DREAM trial14 reported a reintervention rate of approximately 30% at 6 years follow-up, and the OVER trial16 reported a reintervention rate of approximately 20% at 9 years follow-up. However, the seminal EVAR vs OSR RCTs used considerable numbers of now obsolete grafts (approximately 50%). It follows, then, that contemporary, well-adjusted population-based studies of newer grafts with accumulated technical experience should demonstrate a prolongation of the EVAR mortality benefit and a reduction of reintervention rates. This was not the case in our study, as reintervention rates mirrored those demonstrated by the DREAM trial,14 with other population-based studies demonstrating similar results.32-35 Although many reinterventions associated with EVAR have been shown to be minor, a 2018 systematic review and meta-analysis36 demonstrated perioperative complication rates of 3.8% and AAA-associated mortality of 1.8% associated with the treatment of type II endoleak alone. Considering that approximately 50% of type II endoleaks have been reported to resolve spontaneously, overtreatment may be increasing EVAR mortality.37,38 Unfortunately, large, contemporary, population-based studies, including our study, do not offer direct insight into whether changes in graft design have reduced reintervention and mortality. In this study, the assumption that modern endografts were being used for EVAR is not unreasonable, given that the commencement of EVAR funding in Ontario coincides with Health Canada approval of the first endograft and that all approved endografts in Canada are considered modern. However, no information on specific endograft makes and models is available in Ontario administrative data to verify these assumptions.
Another related explanation for the persistent loss of EVAR mortality benefit may be the more liberal use of EVAR against manufacturer instructions for use (IFU). Studies have found that despite relatively stable AAA repair rates, the use of EVAR has increased.6,39-41 Furthermore, a 2018 study42 showed that these rates have increased in populations known to have more challenging anatomy, including women and elderly patients. A 2018 study by Herman et al43 of EVARs from 2005 to 2014 demonstrated IFU violations in 43.8% of patients undergoing elective EVAR and that non-IFU EVARs were associated with higher risk of graft-related adverse events (HR, 1.8; 95% CI, 1.05-3.10). Furthermore, a 2017 international study44 of elective AAA repair outcomes during 9 years confirmed worse EVAR perioperative mortality in octogenarians (1.8% vs 0.7%; P < .001) and women (1.9% vs 0.9%; P < .001) during the course of the study, suggesting that the long-term benefits of EVAR may be undermined by too-forceful application. Unfortunately, Ontario administrative data do not allow assessment of adherence to IFU or difficult anatomy, as anatomical characteristics are not available within these data sets.
Additionally, contemporary studies of EVAR compared with OSR may not be demonstrating improvement in EVAR outcomes owing to concomitant improvements in OSR outcomes through time. The initial differences in perioperative mortality may be decreasing with the development of multidisciplinary care settings, including experienced cardiovascular intensive care units, that optimize OSR outcomes. This study, including data from all centers performing EVAR and OSR in Ontario, Canada, whether tertiary, academic, or otherwise, demonstrated 30-day survival for OSR comparable with the experienced centers involved in the seminal RCTs, largely well-resourced tertiary care centers with well-developed AAA care programs. In 2015, Schermerhorn et al32 demonstrated earlier merging of OSR and EVAR survival curves when they compared mortality prior to 2005 with mortality after 2005. These findings suggest that OSR outcomes may also be improving, thus masking improvements in EVAR outcomes.
Endovascular repair has demonstrated a persistent loss of long-term mortality benefit and high reintervention rates in both RCTs and population-based studies. Consequently, economic analyses have demonstrated EVAR to be cost ineffective, which has led the National Institute for Health and Care Excellence45 in the United Kingdom to recommend against EVAR in its draft of AAA management guidelines. However, the best evidence used to define these guidelines is limited to the seminal RCTs, which are themselves limited by older endografts and early knowledge regarding complications and management. Furthermore, EVAR offers potential benefits other than reduced mortality, despite higher reintervention rates. Many studies have uniformly demonstrated lower perioperative cardiopulmonary complications, shorter lengths of stay in hospitals, and higher rates of discharge home, among others.2-4 Furthermore, mortality does not reflect the considerable impact of myocardial infarction and stroke on quality of life among survivors of myocardial infarctions and strokes.46-48 Consequently, a case can be made for extension of EVAR to younger, healthier patients with longer life expectancy as opposed to reserving it for the most comorbid patient populations or not using it at all. However, the safety of this approach will require even longer follow-up for patients who undergo EVAR than is currently available, to our knowledge. These data will ascertain whether any long-term sequelae, such as cancer, exist secondary to EVAR surveillance and permit greater understanding of the appropriate decision-making approach regarding endoleak and complication management in the context of evolved endograft architecture.
The findings of this study should be interpreted considering several limitations. First, the use of diagnostic, procedure, and billings codes to identify patients, covariates, and outcomes is necessary with the use of administrative data. These codes are subject to coding error and can lack granularity. To mitigate these concerns, validated codes were used for the identification of patients and variables where possible, and surrogate markers were used for important covariates that were not coded. For example, chronic obstructive pulmonary disease was used as a marker of significant smoking history. Still, information regarding anatomical and technical detail, graft types, adherence to IFU, and indications for procedures, imaging, and clinical follow-up does not exist within Ontario administrative data. Next, the use of Ontario administrative data may limit the generalizability of our findings to jurisdictions with similar single-payer health care systems. However, our findings were similar to those of the major RCTs, in which payment model should not influence outcomes. Additionally, although they are firmly established for the conduct of observational studies, propensity score methods only guarantee balance of treatment groups on the measured confounders and do not guarantee absence of residual confounding. However, the congruity of our findings with those of long-term RCTs suggests well-specified propensity score models and findings likely to be free of residual confounding.
This study found no statistically significant difference between outcomes after EVAR and OSR to repair AAA in long-term mortality during more than 13 years of follow-up. Similarly, there were no long-term differences in reintervention or secondary rupture between the AAA repair approaches. However, EVAR was associated with a higher long-term MACE-free survival rate. The long-term outcomes after EVAR do not appear to have improved more than OSR since the conduct of the seminal RCTs. The reasons for these findings will require further study, including consideration of specific graft makes and models, adherence to IFU, and types and reasons for reintervention.
Accepted for Publication: May 15, 2019.
Published: July 10, 2019. doi:10.1001/jamanetworkopen.2019.6578
Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Salata K et al. JAMA Network Open.
Corresponding Author: Mohammed Al-Omran, MD, MSc, Division of Vascular Surgery, Li Ka Shing Knowledge Institute, St Michael’s Hospital, 30 Bond St, Ste 7-074, Bond Wing, Toronto, ON M5B 1W8, Canada (firstname.lastname@example.org).
Author Contributions: Dr Salata had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Salata, Hussain, Greco, Aljabri, Mamdani, Verma, Al-Omran.
Acquisition, analysis, or interpretation of data: Salata, Hussain, de Mestral, Mamdani, Forbes, Bhatt, Al-Omran.
Drafting of the manuscript: Salata, Verma.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Salata, de Mestral, Mamdani.
Obtained funding: Salata.
Administrative, technical, or material support: Hussain, Aljabri, Mamdani, Al-Omran.
Supervision: de Mestral, Greco, Forbes, Verma, Al-Omran.
Conflict of Interest Disclosures: Dr Salata reported grants from the Canadian Institutes for Health Research Canada Graduate Scholarship (Master’s), Frank Goerc and Toronto Academic Vascular Specialists Scholarship, Ontario Graduate Scholarship, and James and Mari Rutka Surgeon Scientist Training Program Scholarship during the conduct of the study and grants from Physician Services Incorporated Foundation outside the submitted work. Dr Bhatt reported serving on the advisory boards of Cardax, Elsevier PracticeUpdate Cardiology, Medscape Cardiology, and Regado Biosciences; serving on the boards of directors of the Boston Department of Veterans Affairs (VA) Research Institute, Society of Cardiovascular Patient Care, and TobeSoft; serving as chair of the American Heart Association Quality Oversight Committee, National Cardiovascular Data Registry-MI Registry Steering Committee, VA Clinical Assessment, Reporting, and Tracking Research and Publications Committee; serving on data monitoring committees for the Cleveland Clinic, Duke Clinical Research Institute, Baim Institute for Clinical Research (formerly Harvard Clinical Research Institute, PORTICO trial, funded by St Jude Medical, now Abbott Laboratories), Mayo Clinic, Icahn School of Medicine at Mount Sinai (for the ENVISAGE trial, funded by Daiichi Sankyo), and Population Health Research Institute; receiving honoraria from the American College of Cardiology (senior associate editor, Clinical Trials and News, ACC.org; vice-chair, ACC Accreditation Committee), Baim Institute for Clinical Research (formerly Harvard Clinical Research Institute; RE-DUAL PCI clinical trial steering committee funded by Boehringer Ingelheim), Belvoir Publications (editor in chief, Harvard Heart Letter), Duke Clinical Research Institute (clinical trial steering committees), HMP Global (editor in chief, Journal of Invasive Cardiology), Journal of the American College of Cardiology (guest editor and associate editor), Population Health Research Institute (for the COMPASS operations committee, publications committee, steering committee, and US national coleader, funded by Bayer), Slack Publications (chief medical editor, Cardiology Today’s Intervention), Society of Cardiovascular Patient Care (secretary and treasurer), and WebMD (continuing medical education steering committees); serving as deputy editor for Clinical Cardiology; receiving research funding from Abbott Laboratories, Amarin Corporation, Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, Chiesi Farmaceutici, Eisai, Ethicon Inc, Forest Laboratories, Idorsia, Ironwood Pharmaceuticals, Ischemix, Eli Lilly and Company, Medtronic, Pfizer, PhaseBio Pharmaceuticals, Regeneron, Roche Holding, Sanofi, Synaptic Pharmaceuticals, and The Medicines Company; receiving royalties from Elsevier (editor, Cardiovascular Intervention: A Companion to Braunwald’s Heart Disease); serving as site coinvestigator at Biotronik, Boston Scientific, St Jude Medical (now Abbott Laboratories), and Svelte; serving as trustee for the American College of Cardiology; and conducting unfunded research for FlowCo Solutions, Merck and Co, Novo Nordisk, PLx Pharma, and Takeda Pharmaceutical. Dr Verma reported grants and personal fees from Amgen and Boehringher Ingelheim; personal fees from AstraZeneca, Bayer, Eli Lilly and Company, Janssen, Merck and Co, Novartis, Novo Nordisk, Sanofi, Servier, and Valeant; and grants from Bristol-Myers Squibb outside the submitted work. No other disclosures were reported.
Funding/Support: This study was supported by the Institute for Clinical Evaluative Sciences (ICES), which is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). This work was jointly funded by the Division of Vascular Surgery at St Michael’s Hospital, Toronto, Ontario, Canada, and funds from the Department of Surgery, King Saud University, Riyadh, Kingdom of Saudi Arabia. Dr Salata is supported in part by the Canadian Institutes of Health Research Canada Graduate Scholarship Master’s salary support award, the Goerc and Toronto Academic Vascular Specialists Surgeon Scientist Training Program Scholarship, and the Ontario Graduate Scholarship.
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: The opinions, results, and conclusions reported in this article are those of the authors and are independent from the funding sources. No endorsement by ICES or the Ontario MOHLTC was intended or should be inferred. Parts of this material are based on data or information compiled and provided by the Canadian Institute of Health Information. However, the analyses, conclusions, opinions, and statements expressed in the material are those of the authors and not necessarily those of the Canadian Institute of Health Information.
Additional Contributions: Atul Sivaswamy, MSc (ICES, Toronto, Ontario, Canada), helped with data set creation and analysis. No direct compensation was provided. However, administrative and analyst fees were paid to ICES for provision of the above services.