Natural history and management outcomes of patients with ruptured Spetzler-Martin grade IV and V brain arteriovenous malformations

OBJECTIVE The natural history of ruptured high Spetzler-Martin grade IV and V brain arteriovenous malformations (bAVMs) is underreported given the scarcity of this pathology, and decision-making for patients with bAVMs remains unclarified. In this study, the authors sought to shed light on this topic. METHODS Patients harboring ruptured high-grade bAVMs were identified from an institutional database spanning from 1990 to 2020. The authors examined outcomes of annual hemorrhagic risk in natural history and after treatment, follow-up hemorrhage rate, bAVM obliteration, follow-up modified Rankin Scale (mRS) score > 2, worsened mRS score, and mortality. RESULTS After reviewing the charts of 1066 patients without hereditary hemorrhagic telangiectasia, 84 patients with ruptured high-grade bAVMs were included in the study for analysis. For cortical bAVMs, the annual risk of hemorrhage during natural history was 2.68%. Surgery decreased the risk to 0.74%, while radiosurgery increased the risk to 5.35%, and embolization only increased the risk to 16.96%. For deep-seated high-grade bAVMs, the annual risk of hemorrhage during natural history was 8.37%. Radiosurgery decreased the risk to 3.11%, surgery decreased the risk to 5.25%, and embolization only increased the risk to 22.33%. Poisson regression analysis demonstrated that embolization only increased the risk of hemorrhage in cortical bAVMs (rate ratio 4.745, 95% CI 1.365–12.819; p = 0.005) and deep-seated bAVMs (rate ratio 6.290, 95% CI 0.997–21.932; p = 0.013). Logistic regression analysis showed that surgery (OR 52.000, 95% CI 8.083–1046.127; p = 0.004) and radiosurgery (OR 11.142, 95% CI 1.804–217.650; p = 0.029) were predictors of obliteration in cortical and deep-seated bAVMs, respectively. The proportions of patients experiencing a worsened mRS score, a follow-up mRS score > 2, and mortality were similar between conservative and treatment groups. CONCLUSIONS The natural history of cortical ruptured high-grade bAVMs bears a risk similar to that of incidental bAVMs, whereas deep-seated ruptured high-grade bAVMs have an increased risk of hemorrhage. With extremely prudent patient selection, surgery might be a viable option for cortical bAVMs to obliterate the bAVM and reduce hemorrhagic risk, while preserving functional status. Radiosurgery might be beneficial to lower hemorrhagic risk in deep-seated bAVMs. Embolization as a single modality should be avoided as it provides no benefit to reduce hemorrhagic risk.