Comparative effectiveness of unruptured cerebral aneurysm therapies: propensity score analysis of clipping versus coiling.

BACKGROUND AND PURPOSE Endovascular therapy has increasingly become the most common treatment for unruptured cerebral aneurysms in the United States. We evaluated a national, multi-hospital database to examine recent utilization trends and compare periprocedural outcomes between clipping and coiling treatments of unruptured aneurysms. METHODS The Premier Perspective database was used to identify patients hospitalized between 2006 to 2011 for unruptured cerebral aneurysm who underwent clipping or coiling therapy. A logistic propensity score was generated for each patient using relevant patient, procedure, and hospital variables, representing the probability of receiving clipping. Covariate balance was assessed using conditional logistic regression. Following propensity score adjustment using 1:1 matching methods, the risk of in-hospital mortality and morbidity was compared between clipping and coiling cohorts. RESULTS A total of 4899 unruptured aneurysm patients (1388 clipping, 3551 coiling) treated at 120 hospitals were identified. Following propensity score adjustment, clipping patients had a similar likelihood of in-hospital mortality (odds ratio [OR], 1.43; 95% confidence interval [CI], 0.49-4.44; P=0.47) but a significantly higher likelihood of unfavorable outcomes, including discharge to long-term care (OR, 4.78; 95% CI, 3.51-6.58; P<0.0001), ischemic complications (OR, 3.42; 95% CI, 2.39-4.99; P<0.0001), hemorrhagic complications (OR, 2.16; 95% CI, 1.33-3.57; P<0.0001), postoperative neurological complications (OR, 3.39; 95% CI, 2.25-5.22; P<0.0001), and ventriculostomy (OR, 2.10; 95% CI, 1.01-4.61; P=0.0320) compared with coiling patients. CONCLUSIONS Among patients treated for unruptured intracranial aneurysms in a large sample of hospitals in the United States, clipping was associated with similar mortality risk but significantly higher periprocedural morbidity risk compared with coiling.