Association between imaging surveillance compliance and long-term outcomes after endovascular abdominal aortic aneurysm repair at Veterans Affairs hospitals.
Academic Article
Overview
abstract
OBJECTIVE: To assess the association between compliance with guideline-recommended annual imaging surveillance after endovascular aortic aneurysm repair (EVAR) and long-term outcomes. METHODS: Veterans who underwent EVAR between January 1, 2000, and December 31, 2023, in US Department of Veterans Affairs facilities were examined retrospectively. The exposure was imaging surveillance compliance, defined as at least one imaging study (computed tomography, ultrasound examination, or magnetic resonance imaging) per year after EVAR. Outcomes were all-cause mortality, reintervention, and rupture. Using a method called landmark analysis, surveillance compliance was assessed over a 2-year landmark period. Each patient was categorized as either noncompliant (no imaging obtained during the landmark period), partially compliant (imaging obtained in one year of the landmark period), or fully compliant (imaging obtained in both years of the landmark period). Kaplan-Meier survival curves evaluated each outcome in the 10 years after the landmark period. This analysis was repeated using different 2-year landmark periods spanning years 0 to 9 after EVAR, then again using a 3-year landmark period. Separately, we used Cox proportional hazard regression to evaluate the association between imaging compliance in a given year and outcomes during the following year. Models were adjusted for age, sex, race, US Department of Veterans Affairs priority group, and baseline Charlson Comorbidity Index score. RESULTS: We identified 27,792 veterans (mean age, 71.7 years; 82.8% White; 99.4% male) who underwent EVAR during the study period. Within the first decade of surveillance, 45.3% of veterans died, 21.1% had reinterventions, and 0.27% experienced late rupture. The number of patients by compliance category was 2430 noncompliant (13.2%), 4799 partially compliant (26.0%), and 11,228 fully compliant (60.8%). In the primary analysis where the landmark period was post-EVAR years 1 and 2, the median survival (95% confidence interval [CI]) for each group (noncompliant, partially, and fully) was 6.0 years (95% CI, 5.7-6.3 years), 6.3 (95% CI, 6.1-6.6 years), and 6.3 years (95% CI, 6.2-6.5 years), respectively. Freedom from reintervention [rupture] among surviving veterans was: noncompliant, 0.72 years (95% CI, 0.68-0.76 years) [1.00 years (95% CI, 1.00-1.00 years)]; partially compliant, 0.68 years (95% CI, 0.66-0.71 years) [0.99 years (95% CI, 0.98-0.99 years)]; and fully compliant, 0.69 years (95% CI, 0.67-0.70 years) [0.99 years (95% CI, 0.99-1.00 years)]. These findings were similar regardless of the landmark period's start or size. In the adjusted Cox proportional hazard models, compliance was associated with reintervention (adjusted hazard ratio, 1.33; 95% CI, 1.22-1.45), but not with mortality or rupture. CONCLUSIONS: This study found no association between less frequent imaging surveillance after EVAR and long-term outcomes of rupture or mortality. These results call into question the benefit of existing paradigms given the unclear clinical benefit and should guide policymakers in refining post-EVAR imaging surveillance recommendations.
