Repetitive Transcranial Magnetic Stimulation for Major Depressive Disorder: A Systematic Review and Network Meta‐Analysis
Yue Zhang, Jinyang Meng, Xiaobo Liu, Hong Tang, Dongling Zhong, Yuxi Li, Jialin Li, Juan Li, Yongxin Cui, Anren Zhang, Chen Xue, Rongjiang JinABSTRACT
Objective
This systematic review and network meta‐analysis (NMA) aimed to compare and rank the efficacy and acceptability of different repetitive transcranial magnetic stimulation (rTMS) modalities for major depressive disorder (MDD).
Methods
The following databases were searched from inception to October 8, 2025: PubMed, Embase, the Cochrane Library, Web of Science, Chinese Biomedical Database (CBM), Chinese National Knowledge Infrastructure (CNKI), Chinese Science and Technology Periodical Database (VIP), and Wanfang Database. In order to identify relevant unpublished studies and ongoing clinical trials, gray literature and trial registries were examined. The randomized controlled trials that investigated the effect of rTMS interventions for MDD were included. We included studies involving adults (aged ≥ 18 years) with a clinical diagnosis of unipolar MDD, which utilized at least two of the following rTMS modalities: low‐frequency left (LFL‐) rTMS, high‐frequency left (HFL‐) rTMS, low‐frequency right (LFR‐) rTMS, high‐frequency right (HFR‐) rTMS, bilateral (BL‐) rTMS, deep TMS (dTMS), theta burst stimulation (TBS), accelerated rTMS (aTMS), synchronized TMS (sTMS), or sham stimulation. The primary endpoints were efficacy (response rate) and acceptability (dropout rate). Additionally, the symptom remission rate was assessed as a secondary outcome. Two authors independently completed the article selection and data extraction. We also evaluated the risk of bias in the included studies with the Cochrane ROB 2.0 tool. The certainty of evidence for each outcome was assessed using the Confidence in Network Meta‐Analysis (CINeMA) framework. A conventional pairwise meta‐analysis was performed using RevMan software. A NMA was performed with Markov chain Monte Carlo simulations to facilitate indirect comparisons across different rTMS modalities. We applied funnel plots to examine publication bias and assessed the robustness of our findings through sensitivity analysis and subgroup analysis.
Results
A total of 141 trials comprising 10,587 participants with MDD were included. The aTMS (odds ratio [OR]: 4.14; 95% confidence interval [CI] 2.14, 8.01), TBS (OR: 4.47; 95% CI, 3.13, 6.37), BL (OR: 3.93; 95% CI, 2.56, 6.03), HFL (OR: 3.34; 95% CI, 2.80, 3.98), LFR (OR: 2.90; 95% CI, 2.13, 3.95) demonstrated significantly greater efficacy than sham control. Additionally, aTMS (OR: 4.44; 95% CI, 1.02, 19.29), TBS (OR: 4.78; 95% CI, 1.23, 18.62), BL (OR: 4.21; 95% CI, 1.06, 16.71) also showed superior efficacy compared to LFL. Furthermore, sTMS was less effective than TBS (OR: 0.27; 95% CI, 0.08, 0.94). In terms of achieving remission, TBS (OR: 4.81; 95% CI, 2.59, 8.95), BL (OR: 4.69; 95% CI, 2.27, 9.69), HFL (OR: 3.35; 95% CI, 2.34, 4.81), and LFR (OR: 3.27; 95% CI, 1.48, 7.24) were all significantly superior to sham control. The estimated relative ranking of treatments suggested that TBS (87.2%) might be the most efficacious among all rTMS modalities. The surface under the cumulative ranking curve (SUCRA) probabilities for acceptability were as follows: BL (80.1%), dTMS (78.1%), LFR (62.8%), HFL (56.5%), sham (50.0%), TBS (49.9%), sTMS (34.0%), aTMS (28.3%), LFL (10.4%). Our sensitivity analysis excluding high‐risk studies yielded largely consistent results with the primary analysis. In particular, the superiority of TBS over sham stimulation remained stable. Although the effect sizes were slightly reduced, the estimates remained statistically significant. In the treatment‐resistant depression (TRD) subgroup, TBS demonstrated the highest probability of response (SUCRA: 87.5%) and remission (SUCRA: 79.2%), with efficacy significantly superior to sham stimulation. In the non‐TRD subgroup, aTMS ranked best for response rate (SUCRA: 83.0%), while LFR achieved the highest remission rate (SUCRA: 81.5%; OR: 7.16). No significant differences in acceptability were observed between any active intervention and sham across all subgroups. The risk‐of‐bias assessment categorized 87 studies as high risk, 51 as having some concerns, and only three as low risk. CINeMA results showed that the comparison of BL versus sham had moderate confidence for remission rate, while other comparisons for response, remission, and acceptability were largely rated as low or very low confidence.
Conclusions
TBS had the highest probability of being the most effective rTMS modality for depression, but the confidence intervals overlapped across multiple comparisons, and the effect sizes were modest. High‐quality head‐to‐head trials are needed to confirm these probabilistic rankings.