Meta-analysis of phase-specific survival after elective endovascular versus surgical repair of abdominal aortic aneurysm from randomized controlled trials and propensity score-matched studies

Objective: Although endovascular aneurysm repair (EVAR) for abdominal aortic aneurysm (AAA) significantly decreases perioperative mortality compared with open surgical repair (OSR), we have not concluded superiority between EVAR and OSR beyond the perioperative period. The aim of this study was to compare phase-specific survival after EVAR vs OSR. Methods: The review was conducted according to the Preferred Reporting Items for Systematic Reviews and MetaAnalyses guideline. Embase and MEDLINE were searched up to November 2019 to identify randomized controlled trials and propensity score-matched studies that investigated >= 2-year all-cause mortality (primary outcome) after EVAR vs OSR for intact infrarenal AAA. For each study, the hazard ratio (HR) with 95% confidence interval (CI) of mortality for EVAR vs OSR was calculated using survival curves for the following specific phases: early term (0-2 years after repair), midterm (2-6 years after repair), long term (6-10 years after repair), and very long term (>= 10 years after repair). The risk ratio (RR) in the perioperative (in-hospital or 30-day) period was also extracted. Phase-specific HRs or RRs were separately pooled using the random effects model. Sensitivity analyses were performed by removing one study at a time to confirm that our findings were not derived from any single study. Funnel plot asymmetry was also examined using the linear regression test. Results: Our search identified four randomized controlled trials and seven propensity score-matched studies enrolling a total of 106,243 AAA patients assigned to EVAR (n = 53,123) or OSR (n = 53,120). The mortality after EVAR compared with OSR was significantly lower in the perioperative period (RR, 0.39; 95% CI, 0.29-0.51; P <.00001) and similar in the earlyterm period (HR, 0.93; 95% CI, 0.84-1.03; P =.16). Notably, significantly higher mortality was observed in the EVAR group compared with the OSR group in the midterm period (HR, 1.15; 95% CI, 1.03-1.29; P =.01). However, similar mortality was observed between the EVAR group and the OSR group in the long-term (HR, 1.06; 95% CI, 0.96-1.17; P =.27) and very-longterm (HR, 1.17; 95% CI, 0.93-1.47; P =.19) periods. In sensitivity analyses, the significant benefit of EVAR in the perioperative period and that of OSR in the midterm period were not changed. No funnel plot asymmetry was identified in all analyses. Conclusions: Compared with OSR, EVAR was associated with lower perioperative mortality and higher mortality in the midterm period for intact infrarenal AAA. The superiority of EVAR was absent in the early-term period, and the inferiority of EVAR in the midterm period disappeared in the long-term and very-long-term periods.