Hemorrhage risk after stereotactic radiosurgery of cerebral arteriovenous malformations

TO ANALYZE THE effect of stereotactic radiosurgery on the hemorrhage rate of arteriovenous malformations (AVMs), we reviewed the clinical and angiographic characteristics of 315 patients with AVMs before and after radiosurgery. One hundred ninety-six patients sustained 263 bleeds in 10,939 patient-years before radiosurgery, for an annual nonfatal hemorrhage rate of 2.4%. Clinical follow-up after radiosurgery was available in 312 patients (mean, 47 +/- 20 mo); follow-up greater than or equal to 24 months was obtained in 295 patients (94%). Twenty-one patients had AVM bleeds at a median of 8 months (range, 1-60 mo) after radiosurgery. Two additional patients had three aneurysmal bleeds (at 5, 27, and 32 mo, respectively) for a 7.4% total risk of hemorrhage per patient. The actuarial hemorrhage rate until AVM obliteration was 4.8% per year (95% confidence interval, 2.4-7.0%) during the first 2 years after radiosurgery and 5.0% per year (95% confidence interval, 2.3-7.3%) for the third to fifth years after radiosurgery. Multivariate analysis of clinical and angiographic factors demonstrated that the presence of an unsecured proximal aneurysm was associated with an increased risk of postradiosurgical hemorrhage (relative risk, 4.56; 95% confidence interval, 1.77-11.70%; P < 0.001). No AVM hemorrhages were observed after radiosurgery in seven patients with intranidal aneurysms. No protective effect against hemorrhage was observed in patients who received an ''optimal'' radiation dose (greater than or equal to 25 Cy to the AVM margin) compared with patients who received <25 Gy to the AVM margin (P = 0.36). No patient suffered a hemorrhage after angiography had confirmed complete obliteration (n = 140) or suffered from an early draining vein without residual nidus (n = 19). Stereotactic radiosurgery was not associated with a significant change in the hemorrhage rate of AVMs during the latency interval before obliteration. No protective benefit was conferred on patients who had incomplete nidus obliteration in early (<60 mo) follow-up after radiosurgery. AVM patients with unsecured proximal aneurysms should have aneurysms obliterated either before radiosurgery or at the time of surgical resection of their AVMs.