and Purpose- Understanding the hemorrhage risks associated with brain arteriovenous malformations (AVMs) before and after stereotactic radiosurgery (SRS) is important. The aims of this multicenter, retrospective cohort study are to evaluate and compare the rates of pre- and post-SRS AVM hemorrhage and identify risk factors. Methods- We pooled AVM SRS data from 8 institutions participating in the International Radiosurgery Research Foundation. Predictors of post-SRS hemorrhage were determined

more »

... ng a multivariate logistic regression model. Pre- and post-SRS hemorrhage rates were compared using Fisher exact test. Ruptured and unruptured AVMs were matched in a 1:1 ratio using propensity scores, and their outcomes were compared. Results- The study cohort comprised 2320 AVM patients who underwent SRS. Deep AVM location (odds ratio, 1.86; 95% CI, 1.19-2.92; P=0.007), the presence of an AVM-associated arterial aneurysm (odds ratio, 2.44; 95% CI, 1.63-3.66; P<0.001), and lower SRS margin dose (odds ratio, 0.93; 95% CI, 0.88-0.98; P=0.005) were independent predictors of post-SRS hemorrhage. The post-SRS hemorrhage rate was lower for obliterated versus patent AVMs (6.0 versus 22.3 hemorrhages/1000 person-years; P<0.001). The AVM hemorrhage rate decreased from 15.4 hemorrhages/1000 person-years before SRS to 11.9 after SRS ( P=0.001). The outcomes of the matched ruptured versus unruptured AVM cohorts were similar. Conclusions- SRS appears to reduce the risk of AVM hemorrhage, although this effect is predominantly driven by obliteration. Deep-seated AVMs are more likely to rupture during the latency period after SRS. AVM-associated aneurysms should be considered for selective occlusion before SRS of the nidus to ameliorate the post-SRS hemorrhage rate of these lesions.