Background Guidelines regarding recommendations for surgical treatment of spontaneous cerebellar intracerebral haemorrhage (ICH) differ. We aimed to systematically review the literature to assess treatment strategies and outcomes.

Methods We searched PubMed and Embase between 1970 and 2019 for randomised or otherwise controlled studies and observational cohort studies. We included studies according to predefined selection criteria and assessed their quality according to the Newcastle-Ottawa Scale (NOS) and risk of bias according to a predefined scale. We assessed case fatality and functional outcome in patients treated conservatively or with haematoma evacuation. Favourable functional outcome was defined as a modified Rankin Scale score of 0–2 or a Glasgow Outcome Scale score of 4–5.

Results We included 41 observational cohort studies describing 2062 patients (40% female) with spontaneous cerebellar ICH. A total of 1171 patients (57%) underwent haematoma evacuation. Ten studies described a cohort of surgically treated patients (n=533) and 31 cohorts with both surgically and conservatively treated patients (n=638 and n=891, respectively). There were no randomised clinical trials nor studies comparing outcome between the groups after adjustment for differences in baseline characteristics. The median NOS score (IQR) was 5 (4–6) out of 8 points and the bias score was 2 (1–3) out of 8, indicative of high risk of bias. Case fatality at discharge was 21% (95% CI 17% to 25%) after conservative treatment and 24% (95% CI 19% to 29%) after haematoma evacuation. At ≥6 months after conservative treatment, case fatality was 30% (95% CI 25% to 30%) and favourable functional outcome was 45% (95% CI 40% to 50%) and after haematoma evacuation, case fatality was 34% (95% CI 30% to 38%) and 42% (95% CI 37% to 47%).

Conclusions Controlled studies on the effect of neurosurgical treatment in patients with spontaneous cerebellar ICH are lacking, and the risk of bias in published series is high. Due to substantial differences in patient characteristics between conservatively and surgically treated patients, and high variability in treatment indications, a meaningful comparison in outcomes could not be made. There is no good published evidence to support treatment recommendations and controlled, preferably randomised studies are warranted in order to formulate evidence-based treatment guidelines for patients with cerebellar ICH.