Background: Intracranial hemorrhage is a serious possible complication in patients with brain arteriovenous malformation (AVM). Several morphologic factors associated with hemorrhagic AVM presentation have been established, but their relevance for the risk of subsequent AVM hemorrhage remains unclear. Methods: The authors analyzed follow-up data on 622 consecutive patients from the prospective Columbia AVM database, limited to the period between initial AVM diagnosis and the start of treatment (i.e., any endovascular, surgical, or radiation therapy). Univariate and multivariate logistic regression and Cox proportional hazard models were applied to analyze the effect of patient age, gender, AVM size, anatomic location, venous drainage pattern, and associated arterial aneurysms on the risk of intracranial hemorrhage at initial presentation and during follow-up. Results: The mean pretreatment follow-up was 829 days (median: 102 days), during which 39 (6%) patients experienced AVM hemorrhage. Increasing age (hazard ratio [HR] 1.05, 95%CI 1.03 to 1.08), initial hemorrhagic AVM presentation (HR 5.38, 95%CI 2.64 to 10.96), deep brain location (HR 3.25, 95%CI 1.30 to 8.16), and exclusive deep venous drainage (HR 3.25, 95%CI 1.01 to 5.67) were independent predictors of subsequent hemorrhage. Annual hemorrhage rates on follow-up ranged from 0.9%for patients without hemorrhagic AVM presentation, deep AVM location, or deep venous drainage to as high as 34.4%for those harboring all three risk factors. Conclusions: Hemorrhagic arteriovenous malformation (AVM) presentation, increasing age, deep brain location, and exclusive deep venous drainage appear to be independent predictors for AVM hemorrhage during natural history follow-up. The risk of spontaneous hemorrhage may be low in AVMs without these risk factors. Background: Intracranial hemorrhage is a serious possible complication in patients with brain arteriovenous malformation (AVM). Several Morphologic Factors associated with hemorrhagic AVM presentation have been established, but their relevance for the risk of subsequent AVM hemorrhage remains unclear. Methods: The authors yet follow -up data on 622 consecutive patients from the prospective Columbia AVM database, limited to the period between initial AVM diagnosis and the start of treatment (ie, any endovascular, surgical, or radiation therapy). Univariate and multivariate logistic regression and Cox proportional hazard models were applied to analyze the effect of patient age, gender, AVM size, anatomic location, venous drainage pattern, and associated arterial aneurysms on the risk of intracranial hemorrhage at initial presentation and during follow-up. Results: The mean pretreatment follow-up was 829 days (median: 102 days), during which 39 (6%) patients experienced AVM hemorrha ge. Increasing age (hazard ratio [HR] 1.05, 95% CI 1.03 to 1.08), initial hemorrhagic AVM presentation (HR 5.38, 95% CI 2.64 to 10.96), deep brain location (HR 3.25, 95% CI 1.30 to 8.16) , and independent deep venous drainage (HR 3.25, 95% CI 1.01 to 5.67) were independent predictors of subsequent hemorrhage. Annual hemorrhage rates on follow-up ranged from 0.9% for patients without hemorrhagic AVM presentation, deep AVM location, or deep venous drainage to as high as 34.4% for those harboring all three risk factors. Conclusions: Hemorrhagic arteriovenous malformation (AVM) presentation, increasing age, deep brain location, and exclusive deep venous drainage appear to be independent predictors for AVM hemorrhage during natural history follow-up The risk of spontaneous hemorrhage may be low in AVMs without these risk factors.