The impact of therapeutic modality on outcomes following repair of ruptured intracranial aneurysms: an administrative data analysis

Publisher Website
Google Scholar
Abstract
Enrolling a selected sample of ruptured intracranial aneurysms, the International Subarachnoid Aneurysm Trial (ISAT) found endovascular coiling to be superior to microsurgical clipping. The performance of coiling in a more general population of ruptured aneurysms has not been adequately studied. Using provincial administrative data from Ontario, the authors conducted a retrospective cohort study of adult patients with subarachnoid hemorrhage (SAH) who underwent aneurysm repair. The exposure was defined as endovascular versus surgical aneurysm repair. The prespecified primary outcome was time to death or readmission for SAH. Data from the entire cohort were analyzed using a multivariable adjusted Cox proportional hazards model. Propensity scores were used to compare a matched subgroup of patients with aneurysms who had similar baseline characteristics. The potential impact of unmeasured confounding was assessed using sensitivity analysis. Between 1995 and 2004, 2342 aneurysms were clipped and 778 were coiled in Ontario. The proportion of aneurysms treated by coiling increased steadily over time. In the adjusted analysis of the entire cohort, endovascular coiling was associated with a significantly increased hazard of death or SAH readmission (hazard ratio 1.25 [95% CI 1.00–1.55], p = 0.04). Similar results were obtained from the propensity score matched analysis (hazard ratio 1.25 [95% CI 1.04–1.50], p = 0.02). Measures of procedural morbidity and mortality were not significantly different between groups. The results of the current analysis call into question the generalizability of the ISAT to all ruptured aneurysms. Given the limitations inherent in this form of analysis, further clinical studies—rigorously assessing the performance of endovascular therapy in patients with non-ISAT-like aneurysms—are indicated.

This publication has 20 references indexed in Scilit: