Background: This work evaluates the consistency, effect size and incidence of persistent side effects of lesional neurosurgical interventions in the treatment of tremor due to Parkinson's disease (PD), essential tremor (ET), multiple sclerosis (MS) and midbrain lesions.
Methods: Systematic review and meta-analysis according to PRISMA-P guidelines. Random effects meta-analysis of standardised mean difference based on a peer-reviewed protocol (PROSPERO no. CRD42016048049).
Results: From 1249 abstracts screened, 86 peer-reviewed studies reporting 102 cohorts homogeneous for tremor aetiology, surgical target and technique were included.Effect on PD tremor was better when targeted at the ventral intermediate nucleus (V.im.) by radiofrequency ablation (RF) (Hedge's g: -4.15;) over V.im. by Gamma Knife (GK) (-2.2), subthalamic nucleus (STN) by RF (-1.12) and globus pallidus internus (GPi) by RF (-0.89). For ET MRI-guided focused ultrasound (MRIgFUS) ablation of the cerebellothalamic tract (CTT) (-2.35) and V.im. (-2.08) showed similar mean tremor reductions to V.im. ablation by RF (-2.42) or GK (-2.13). In MS V.im. ablation by GK (-1.96) and RF (-1.63) were similarly effective.Mean rates of persistent side effects after unilateral lesions in PD were 12.8% (RF V.im.), 13.6% (RF STN), 9.2% (RF GPi), 0.7% (GK V.im.) and 7.0% (MRIgFUS V.im.). For ET, rates were 9.3% (RF V.im.), 1.8% (GK V.im.), 18.7% (MRIgFUS V.im.) and 0.0% (MRIgFUS CTT), for MS 37.7% (RF V.im.) and for rubral tremor 30.3% (RF V.im.).
Conclusion: This meta-analysis quantifies safety, consistency and efficacy of lesional neurosurgical interventions for tremor by target, technique and aetiology.