Impact of Treatment Delay on Outcome in the International Subarachnoid Aneurysm Trial

Background and Purpose— ISAT (International Subarachnoid Aneurysm Trial) demonstrated that 1 year after aneurysmal subarachnoid hemorrhage, coiling resulted in a significantly better clinical outcome than clipping. After 5 years, this difference did not reach statistical significance, but mortality was still higher in the clipping group. Here, we present additional analyses, reporting outcome after excluding pretreatment deaths. Methods— Outcome measures were death with or without dependency at 1 and 5 years after treatment, after exclusion of all pretreatment deaths. Treatment differences were assessed using relative risks (RRs). With sensitivity and exploratory analyses, the relation between treatment delay and outcome was analyzed. Results— After exclusion of pretreatment deaths, at 1-year follow-up coiling was favorable over clipping for death or dependency (RR, 0.77 [95% CI, 0.67–0.89]) but not for death alone (RR, 0.88 [95% CI, 0.66–1.19]). After 5 years, no significant differences were observed, neither for death or dependency (RR, 0.88 [95% CI, 0.77–1.02]) nor for death alone (RR, 0.82 [95% CI, 0.64–1.05]). Sensitivity analyses showed a similar picture. In good-grade patients, coiling remained favorable over clipping in the long-term. Time between randomization and treatment was significantly longer in the clipping arm (mean 1.7 versus 1.1 days;P <0.0001), during which 17 patients died because of rebleeding versus 6 pretreatment deaths in the endovascular arm (RR, 2.81 [95% CI, 1.11–7.11]).Conclusions— These additional analyses support the conclusion of ISAT that at 1-year follow-up after aneurysmal subarachnoid hemorrhage coiling has a better outcome than clipping. After 5 years, with pretreatment mortality excluded, the difference between coiling and clipping is not significant. The high number of pretreatment deaths in the clipping group highlights the importance of urgent aneurysm treatment to prevent early rebleeding.