Systematic review and meta-analysis on mortality and functional outcome in patients with large cerebellar infarctions treated with neurosurgery

Background: Management of large cerebellar infarctions with potential malignant evolution is highly heterogeneous across physicians, and recommendations rely on low-evidence studies. Aim: We aimed to perform a systematic review and meta-analysis on patients with large cerebellar infarction undergoing neurosurgery, to study mortality and functional outcome, according to neurosurgical technique. Summary of review: We searched on PubMed and Embase according to predefined selection criteria and we assessed their quality according to a predefined risk of bias scale. Our primary outcomes were mortality and functional outcome rates. Favourable outcome was defined as a modified Rankin scale of 0-2, a Glasgow Outcome Scale of 4-5 or a Barthel Index >90%. Pooled rates were obtained using random effect model and heterogeneity was quantified using I2 statistics. Among 27 included studies (including 1173 patients), we studied the 662 patients undergoing neurosurgery. All studies were retrospective and observational; there was no randomized clinical trial (RCT). The median selection bias score was 5 (IQR 4-6). Mortality rate was estimated at 18% 95%CI 13%-24%, I² 58%. Among survivors, 64% achieved a favourable functional outcome 95%CI 51%-77%, I² 82%. Study design and heterogeneity in patients' characteristics limited a meaningful comparison of mortality and functional outcome according to neurosurgical techniques. Conclusion: High-quality evidence on neurosurgical treatment for large cerebellar infarctions remains limited. Our systematic review and meta-analysis, despite moderate risk of bias, suggest that neurosurgery may reduce mortality and improve functional outcomes. These findings support its potential benefit, but RCT are needed to confirm effectiveness and evaluate best surgical technique.