Effectiveness of low-frequency, non-navigated repetitive transcranial magnetic stimulation for obsessive-compulsive disorder: a systematic review and meta-analysis of randomized controlled trials

Background: Obsessive-compulsive disorder (OCD) imposes a substantial burden on patients and health systems alike: despite established pharmacological and psychological interventions, up to 40% of patients fail to achieve full remission with first-line treatment, driving persistent interest in alternative neuromodulation approaches. Among the neuromodulation strategies under investigation, low-frequency (predominantly 1 Hz), non-neuronavigated repetitive transcranial magnetic stimulation (rTMS) represents a surface-coil protocol class distinct from the FDA-cleared deep TMS approach for OCD, yet its effectiveness remains uncertain. Objective: To estimate the pooled effect of low-frequency (predominantly 1 Hz), non-neuronavigated rTMS versus sham stimulation on Yale-Brown Obsessive-Compulsive Scale (Y-BOCS) scores, and to determine whether stimulation target, treatment duration, or frequency systematically moderates that effect within this protocol class. Methods: This PRISMA 2020-compliant systematic review and meta-analysis (PROSPERO: CRD420261302647) searched PubMed, Embase, and Web of Science from inception through May 2025. Eligible studies were sham-controlled RCTs reporting Y-BOCS total scores; risk of bias was assessed using the Cochrane RoB 2.0 tool. The primary outcome was the between-group mean difference (MD) in Y-BOCS scores under a DerSimonian-Laird random-effects model, selected per pre-specified criteria (random effects if I² ≥50% or Q-test P ≤ 0.1). Pre-specified subgroup analyses examined stimulation target, treatment duration, and frequency. Results: Fourteen RCTs (N = 460; 252 active rTMS, 208 sham or control) were included. The primary pooled analysis yielded a mean difference (MD) in Y-BOCS scores of -0.69 (95% CI: -1.75 to 0.38; P = 0.21; I² = 58%), indicating no statistically significant benefit of rTMS over sham. Moderate heterogeneity was substantially reduced by exclusion of a single outlying trial, without altering the null direction of the estimate. All pre-specified subgroup analyses - by stimulation target (DLPFC vs. SMA), treatment duration (2-12 weeks), and stimulation frequency (1 Hz vs. 10 Hz) - were non-significant, and leave-one-out sensitivity analysis confirmed robustness across all 14 permutations. Conclusion: Pooled evidence from 14 RCTs does not support a statistically significant benefit of low-frequency, non-neuronavigated rTMS over sham for OCD symptom reduction (MD = -0.69, 95% CI: -1.75 to 0.38; P = 0.21). This finding is specific to low-frequency, non-neuronavigated surface-coil protocols and should not be interpreted as evidence against rTMS efficacy more broadly; approved deep TMS and neuronavigated protocols were outside the scope of this review. These results indicate that this low-frequency, non-navigated protocol class lacks demonstrated efficacy in controlled trials. Whether neuronavigation and cortical field verification can bridge this performance gap requires adequately powered trials designed for direct protocol comparison.