Endovascular repair of ruptured abdominal aortic aneurysm (rAAA) has rapidly gained popularity, but superior results may be biased by patient selection. The aim was to investigate whether suitability for endovascular repair predicted survival, irrespective of technique of repair. Two blinded investigators independently evaluated preoperative computed tomography angiograms of a consecutive cohort of patients with rAAA. Patients were categorized either 'suitable' or 'unsuitable' for endovascular repair, if assessments agreed. If assessments disagreed, they were classified 'borderline suitable'. Correlations between endovascular suitability and clinical outcome were adjusted for suspected confounding factors and tested for robustness using sensitivity analyses. A total of 248 patients with rAAA from January 2001 to December 2010 were included, of whom 237 (95·6 per cent) underwent open repair. Seventy patients (28·2 per cent) were classified as 'suitable' and 100 (40·3 per cent) as 'unsuitable' for endovascular repair; 63 (25·4 per cent) were considered 'borderline suitable'. Fifteen (6·0 per cent) could not be assessed and were included in the sensitivity analyses. The postoperative 30-day mortality rate was 15·3 per cent (38 deaths). Multiple logistic regression demonstrated that the odds of perioperative death increased 9·21 (95 per cent confidence interval 2·16 to 39·23) fold for 'unsuitable' rAAA (P = 0·003) and 6·80 (1·47 to 31·49) fold for 'borderline' rAAA (P = 0·014), compared with 'suitable' rAAA. This selection effect was robust across sensitivity analyses and sustained for at least 5 years of follow-up. Endovascular suitability was an independent and strongly positive predictor of survival after open repair of rAAA. Copyright © 2012 British Journal of Surgery Society Ltd. Published by John Wiley & Sons, Ltd.