OBJECTIVEThe optimal revascularization for pediatric moyamoya for reducing the incidence of future stroke events remains to be determined.METHODSThe authors conducted a search of PubMed, MEDLINE, Embase, and Web of Science electronic databases from inception until January 2019. The primary endpoint was the presence of a future ischemic stroke event. Secondary endpoints were angiographic outcomes as measured by postoperative Matsushima grade and clinical symptom persistence. Patients who underwent either direct or combined direct/indirect revascularization were classified into the direct cohort. Data from each study on presence of postoperative stroke events were used to generate standardized mean differences and 95% confidence intervals, which were combined using inverse variance-weighted averages of standardized mean differences in a random effects model.RESULTSTwenty-nine studies met the inclusion criteria for analysis, comprising 2258 patients (1011 males, mean age 8.3 ± 1.8 years) who were followed up clinically for an average of 71.4 ± 51.3 months. One hundred fifty-four patients underwent direct bypass alone, 680 patients underwent either direct or combined direct/indirect revascularization procedures, while 1424 patients underwent indirect bypass alone. The frequencies of future stroke events in patients undergoing direct bypass alone, combined bypass, or indirect bypass alone were 1 per 190.3 patient-years, 1 per 108.9 patient-years, and 1 per 61.1 patient-years, respectively, in each cohort. The estimated stroke rates were 9.0% with indirect revascularization, 4.5% with direct revascularization alone, and 6.0% with combined revascularization. A forest plot did not reveal any significant differences in the incidence of future stroke events or angiographic outcomes following direct-only, combined, or indirect-only revascularization.CONCLUSIONSDirect, indirect, and combined indirect/direct bypass techniques are all effective revascularization options for pediatric moyamoya disease. A relative paucity of granular studies—and inherent surgical selection bias—limits direct comparison between interventions. Suitably designed prospective cohort studies may be useful in identifying patients likely to receive benefit from specific procedures.