Interventional outcomes for patients eligible for entry into the ARUBA clinical trial: a systematic review and meta-analysis

OBJECTIVE A Randomized Trial of Unruptured Brain Arteriovenous Malformations (ARUBA) suggested that medical management afforded outcomes superior to those following intervention for unruptured arteriovenous malformations (AVMs), but its findings have been controversial. Subsequent studies of AVMs that would have met the eligibility requirements of ARUBA have supported intervention for the management of some cases. The present meta-analysis was conducted with the object of summarizing interventional outcomes for ARUBA-eligible patients reported in the literature. METHODS A systematic literature search (PubMed, Web of Science, Google Scholar) for AVM intervention studies that used inclusion criteria identical to those of ARUBA (age >= 18 years, no history of AVM hemorrhage, no prior intervention) was performed. The primary outcome was death or symptomatic stroke. Secondary outcomes included AVM obliteration, hemorrhage, death, and poor outcome (modified Rankin Scale score >= 2 at final follow-up). Bias assessment was performed with the Newcastle-Ottawa Scale, and the results were synthesized as pooled proportions. RESULTS Of the 343 articles identified through database searches, 13 studies met the inclusion criteria, yielding an overall study cohort of 1909 patients. The primary outcome occurred in 11.2% of patients (pooled = 11%, 95% CI 8%-13%). The rates of AVM obliteration, hemorrhage, poor outcome, and death were 72.7% (pooled = 78%, 95% CI 70%-85%), 8.4% (pooled = 8%, 95% CI 6%-11%), 9.9% (pooled = 10%, 95% CI 7%-13%), and 3.5% (pooled = 2%, 95% CI 1%-4%), respectively. Annualized primary outcome and hemorrhage risks were 1.85 (pooled = 2.05, 95% CI 1.31-2.94) and 1.34 (pooled = 1.41, 95% CI 0.83-2.13) per 100 patient-years, respectively. CONCLUSIONS Intervention for unruptured AVMs affords acceptable outcomes for appropriately selected patients. The risk of hemorrhage following intervention compared favorably to the natural history of unruptured AVMs. The included studies were retrospective and varied in treatment and AVM characteristics, thereby limiting the generalizability of their data. Future studies from prospective registries may clarify patient, nidus, and intervention selection criteria that will refine the challenging management of patients with unruptured AVMs.