Timing of microsurgical resection for ruptured brain arteriovenous malformations: a propensity score–matched analysis using prospective single-center registry data

Abstract

OBJECTIVE The optimal microsurgical timing in ruptured brain arteriovenous malformations (AVMs) is not well understood and is surrounded by controversy. This study aimed to elucidate the impacts of microsurgical resection timing on clinical outcomes. METHODS The authors retrieved and reviewed the records on all ruptured AVMs treated at their institution and registered in a nationwide multicenter prospective collaboration registry between August 2011 and August 2021. Patients were dichotomized into an early resection group (≤ 30 days from the last hemorrhagic stroke) and a delayed resection group (> 30 days after the last hemorrhagic stroke). Propensity score–matched analysis was used to compare long-term outcomes. The primary outcome was neurological status as assessed using the modified Rankin Scale (mRS). The secondary outcomes were complete obliteration rate, postoperative seizure, and postoperative hemorrhage. RESULTS Of the 3649 consecutive AVMs treated at the authors’ institution, a total of 558 ruptured AVMs were microsurgically resected and had long-term follow-up. After propensity score matching, 390 ruptured AVMs (195 pairs) were included in the comparison of outcomes. The mean (± standard deviation) clinical follow-up duration was 4.93 ± 2.94 years in the early resection group and 5.61 ± 2.56 years in the delayed resection group. Finally, as regards the distribution of mRS scores, short-term neurological outcomes were better in the delayed resection group (risk difference [RD] 0.3%, 95% CI −0.1% to 0.6%, p = 0.010), whereas long-term neurological outcomes were similar between the two groups (RD 0.0%, 95% CI −0.2% to 0.2%, p = 0.906). Long-term favorable neurological outcomes (early vs delayed: 90.8% vs 90.3%, p > 0.999; RD 0.5%, 95% CI −5.8% to 6.9%; RR 1.01, 95% CI 0.94–1.07) and long-term disability (9.2% vs 9.7%, p > 0.999; RD −0.5%, 95% CI −6.9% to 5.8%; RR 0.95, 95% CI 0.51–1.75) were also similar between these groups. In terms of secondary outcomes, postoperative seizure (early vs delayed: 8.7% vs 5.6%, p = 0.239; RD 3.1%, 95% CI −2.6% to 8.8%; RR 1.55, 95% CI 0.74–3.22), postoperative hemorrhage (1.0% vs 1.0%, p > 0.999; RD 0.0%, 95% CI −3.1% to 3.1%; RR 1.00, 95% CI 0.14–7.04), and hospitalization time (16.4 ± 8.5 vs 19.1 ± 7.9 days, p = 0.793) were similar between the two groups, whereas early resection had a lower complete obliteration rate (91.3% vs 99.0%, p = 0.001; RD −7.7%, 95% CI −12.9% to 3.1%; RR 0.92, 95% CI 0.88–0.97). CONCLUSIONS Early and delayed resection of ruptured AVMs had similar long-term neurological outcomes. Delayed resection can lead to a higher complete obliteration rate, although the risk of rerupture during the resection waiting period should be vigilantly monitored.