$Background:$ Short-term survival benefits of endovascular repair (EVAR) versus open repair of intact abdominal aortic aneurysms have been shown in randomised trials, but this early survival benefit is lost within a few years. The long-term survival benefit of EVAR is unclear.

$Methods:$ Between 1999 and 2004, 1252 patients from 37 centres with aneurysms ≥5.5cm diameter, suitable for either EVAR or open repair, were randomly assigned 1:1 to either EVAR (n=626) or open repair (n=626) using computer-generated sequences of randomly permuted blocks stratified by centre. The primary analysis compared rates of total and aneurysm-related death until mid-2015; re-interventions were a secondary outcome.

$Findings:$ Over a mean of 12.7 years (maximum 15.8-years) of follow-up there was no evidence of a difference in total or aneurysm-related mortality between the randomised groups. An early survival benefit in the EVAR group was lost and beyond 8-years of follow-up there was increased aneurysm-related and total mortality, adjusted hazard ratios 5.82 (95%CI 1.64,20.65; p=0.006) and 1.25 (95%CI 1.00,1.56; p=0.05), respectively: the hazard ratios increased further for per-protocol analyses. The increased aneurysm-related mortality beyond 8 years was mainly attributable to secondary aneurysm sac rupture, with increased cancer mortality also observed in the EVAR group. Overall, aneurysm re-intervention rates were higher in the EVAR than in the open repair group, 4.1 and 1.7 per 100 person-years respectively, p<0.001 with re-interventions occurring throughout follow-up.

$Interpretation:$ Early EVAR survival benefit was followed by inferior late survival and graft durability compared to open repair. Inadequate late surveillance may have contributed to these findings.