Interaction between frailty and sex on mortality after elective abdominal aortic aneurysm repair

Abstract

Background

Controversy exists surrounding gender outcome disparity and abdominal aortic aneurysm (AAA) repair. Previous reports have demonstrated worse outcomes for women undergoing open aneurysm repair (OAR); however, these differences are less evident with endovascular aneurysm repair (EVAR). Epidemiologic studies have documented that women score higher on most frailty assessment scales but paradoxically have longer life expectancy compared to men. The interaction of gender/frailty and the influence on outcomes and practice patterns surrounding EVAR and OAR is poorly described. This analysis characterizes the association of frailty/sex interactions on mortality as well as patient selection surrounding elective AAA repair in the Society for Vascular Surgery Vascular Quality Initiative.

Methods

All elective infrarenal AAA (EVAR + OAR; 2003-2017) cases were queried from the Vascular Quality Initiative database. Each patient was assigned a previously published modified frailty index (mFI) score derived from comorbidity and preoperative functional status data. Cox proportional hazard models, adjusted for statistically significant covariates, including procedural complexity, determined associations within full models and sex-stratified models.

Results

A total of 20,750 elective AAA cases were analyzed (EVAR 15,893 [77%]; OAR 4857 [23%]). Thirty-day mortality for EVAR and OAR was 0.7% (n = 115) and 3.5% (n = 169), respectively. Patients who died were significantly more likely to be older (EVAR, 78 vs 73 years; OAR, 74 vs 69 years; P < .0001), have larger AAA diameters (EVAR, 59 vs 56 mm; P = .005; OAR, 62 vs 59 mm; P = .001), higher mFI scores (EVAR, 3.2 vs 2.4; OAR, 3.1 vs 2.2; P < .0001), and be of female sex (EVAR hazard ratio = 1.66 [95% confidence interval, 1.10-2.52]; P = .007; OAR-1.43 [1.02-1.99]; P = .003). Significant differences in the gender distribution of frailty scores among EVAR patients were evident (mean mFI: male 2.42 vs female 2.34; P = .02), but no difference was detected for OAR (male 2.17 vs female 2.22; P = .38). The mFI was a strong independent predictor of mortality (30 days: EVAR hazard ratio = 1.36 [1.22-1.53] and OAR 1.46 [1.32-1.60]; 1 year: EVAR 1.32 [1.25-1.39] and OAR-1.38 [1.28-1.48]). There was no interaction between mFI and gender on the association with mortality. Across frailty strata, male patients were nearly twofold more likely to undergo either elective EVAR or OAR for an AAA below recommended minimum diameter thresholds (male, <5.5 cm; female, <5.0 cm). Greater mFI score did not alter OAR selection but was associated with less frequent EVAR of small AAA.

Conclusions

Given the strong association between frailty and postoperative mortality, mFI can be used as a predictive tool to aid in surgical planning of patients undergoing elective AAA repair. While mFI can predict postoperative mortality for both men and women, it does not account for the survival disparity between sexes, and further research is warranted to explain this difference. There appear to be significant gender differences in patient selection based on current Society for Vascular Surgery–endorsed treatment thresholds that may have important implications on the appropriateness of AAA care delivery nationally.