Forty-six studies (1,397 patients) were included in the review. There was one randomised controlled trial of endovascular repair versus open surgery (data was used only from 13 patients undergoing endovascular repair); the remainder were observational study designs.
The pooled result for mortality was 24.3% (95% CI 20.7 to 28.3%; 46 studies; Ι²= 49.4%, which indicated substantial heterogeneity). Meta-regression analysis showed the only statistically significant association was between endograft configuration (bifurcated approach) and a reduction in perioperative mortality (p=0.005; 36 studies). Further results of the meta-regression analysis were reported in the paper.
There was evidence to suggest the presence of publication bias. Re-analysis excluding the largest study did not substantially change the pooled result.