Microsurgery versus Microsurgery With Preoperative Embolization for Brain Arteriovenous Malformation Treatment: A Systematic Review and Meta-analysis

Neurosurgery 92:27–41, 2023

Preoperative embolization has traditionally been regarded as a safe and effective adjunct to microsurgical treatment of brain arteriovenous malformations (bAVM). However, there is currently no high-level evidence to ascertain this presumption.

OBJECTIVE: To compare the outcomes of microsurgery (MS) vs microsurgery with preoperative embolization (E + MS) in patients with bAVM through systematic review.

METHODS: We searched MEDLINE, PubMed, and Embase. The primary outcome was bAVM obliteration. Secondary outcomes were intraoperative bleeding (mL), complications, worsened modified Rankin Scale (mRS), and mortality. The pooled proportions of outcomes were calculated through the logit transformation method. The odds ratio (OR) of categorical data and mean difference of continuous data were estimated through the Mantel-Haenszel and the inverse variance methods, respectively.

RESULTS: Thirty-two studies met the eligibility criteria. One thousand eight hundred twenty-eight patients were treated by microsurgery alone, and 1088 were treated by microsurgery with preoperative embolization, respectively. The meta-analysis revealed no significant difference in AVM obliteration (94.1% vs 95.6%, OR = 1.15 [0.63-2.11], P = .65), mortality (1.7% vs 2%, OR = 0.88 [0.30-2.58], P = .82), procedural complications (18.2% vs 27.2%, OR = 0.47 [0.19-1.17], P = .10), worsened mRS (21.2% vs 18.5%, OR = 1.08 [0.33-3.54], P = .9), and intraoperative blood loss (mean difference = 182.89 [À87.76, 453.55], P = .19).

CONCLUSION: The meta-analysis showed no significant difference in AVM obliteration, mortality, complications, worse mRS, and intraoperative blood loss between MS and E + MS groups. For AVMs where MS alone has acceptable results, it is reasonable to bypass unnecessary preoperative embolization given higher postoperative complication risk.

%d bloggers like this: