Repetitive transcranial magnetic stimulation for late-life depression: A systematic review and meta-analysis of randomized controlled trials

Abstract

Objective

Repetitive transcranial magnetic stimulation is considered a potential treatment of choice for late-life depression, but its efficacy remains unclear. Therefore, we conducted a systematic review and meta-analysis to investigate the efficacy of repetitive transcranial magnetic stimulation.

Methods

We systematically searched PubMed, Embase, Cochrane Library, and PsycINFO databases. Meta-analyses were performed to examine the improvement in depression measured by the Hamilton Depression Rating Scale score, response rate, and remission rate after repetitive transcranial magnetic stimulation treatment.

Results

A total of 10 trials were included. The pooled results revealed an improvement in the Hamilton Depression Rating Scale score (standardized mean difference:  0.51; 95% confidence interval: 0.33, 0.68), response rate (odds ratio: 2.62; 95% confidence interval: 1.37, 5.01), and remission rate (odds ratio: 6.88; 95% confidence interval: 2.93, 16.14). The pooled results of trials using repetitive transcranial magnetic stimulation with an intensity of ≥100% resting motor threshold revealed an improvement in the Hamilton Depression Rating Scale score (standardized mean difference:  3; 95% confidence interval: 1.87, 4.12) and response rate (odds ratio: 2.56; 95% confidence interval: 1.58, 4.13). No differences were observed in the pooled results of trials using repetitive transcranial magnetic stimulation with an intensity of <100% resting motor threshold.

Conclusion

Repetitive transcranial magnetic stimulation is effective in the treatment of late-life depression, especially with an intensity of ≥100% resting motor threshold. Future studies with larger sample sizes and long-term outcomes are still needed to provide more robust evidence.

INPLASY registration number: INPLASY202550075

Introduction

In older adults, late-life depression (LLD) is a common disorder with a primary diagnosis of major depression, dysthymia, or minor depression according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria. ¹ The prevalence of clinically significant depressive symptoms ranges from approximately 8% to 16% among community-dwelling older adults. ² Meanwhile, large-scale epidemiological investigations reported that the prevalence of major depressive disorder (MDD) in community samples of adults aged ≥65 years ranges from 1% to 5%. ³ Depression is associated with a higher risk of stroke and stroke-related mortality, with an increased all-cause mortality rate in the depressed population.^4,5 Older adults with LLD are more likely to have cognitive deficits, especially executive functional impairment, and are more likely to develop dementia subsequently. ⁶ Therefore, owing to its serious consequences on life quality, the management of LLD is an important public health issue.

Several treatment options are available for LLD, including antidepressants, psychotherapy, and physiotherapy.^7–9 Although antidepressants are widely used in older patients with depression, the rate of response to antidepressants is lower in patients with major LLD than in younger patients with depression. ¹⁰ The use of antidepressant drugs in people aged ≥65 years who have depression is associated with an increased risk of several adverse outcomes, such as blood pressure instability, all-cause mortality, stroke/transient ischemic attack, falls, fracture, epilepsy/seizures, and hyponatremia.^11,12 Psychosocial interventions, such as problem-solving therapy, cognitive behavioral therapy, and interpersonal therapy, are effective in the treatment of LLD. ⁸ However, evidence-based psychotherapies are rarely used correctly in the community, and most people cannot access them, especially in low-income countries. Although electroconvulsive therapy (ECT) demonstrates a rapid onset and high efficacy in treating geriatric depression, its utility in older patients is limited by cognitive side effects (e.g. memory impairment) and anesthesia-related risks.

In the past decades, repetitive transcranial magnetic stimulation (rTMS) has been a novel treatment option for depression, which is used to generate an electric current in the brain tissue immediately below the skull, altering the cortical excitation of this brain region and its interconnected brain network. ¹³ rTMS was approved by the Food and Drug Administration for the treatment of mild-to-moderate treatment-resistant depression (TRD) in 2008. Compared with other neuromodulation treatments for LLD, rTMS has unique advantages; it does not require anesthesia and is not associated with cognitive side effects.^14,15 Two meta-analyses showed that high-frequency rTMS favored sham-rTMS in patients with MDD.^16,17 However, a previous controlled study did not achieve a sufficient response to rTMS in the treatment of older patients with depression. ¹⁸ Recently, several studies have been conducted to examine the efficacy of rTMS in older patients with depression, which showed moderate effectiveness.^19–21 The inconsistency was limited probably due to the small sample size and the inconsistency in the treatment frequency, parameter settings, and site for rTMS. Therefore, there is an urgent need for a well-designed meta-analysis to pool the comprehensive clinical evidence to guide the application of rTMS in patients with LLD. Herein, we conducted a meta-analysis to review the literature systematically to evaluate the efficacy of rTMS in the treatment of LLD.

Methods

Search strategy

Four large electronic databases (PubMed, Embase, Cochrane Library, and PsycINFO) were searched to identify relevant clinical studies that investigated the efficacy of rTMS in treating LLD from inception to December 2024, without restrictions regarding language. The search was performed by using different combinations of the following keywords and medical subject headings (MeSH): “depression,” “late-life,” “geriatric,” and “transcranial magnetic stimulation.” In addition, we manually searched the reference lists of identified relevant reviews and included articles that might not have been captured in the database literature search.

Selection criteria

We included studies that met the following criteria: (a) the study design was a randomized controlled trial; (b) participants were older adult patients (male or female aged >55 years) who were diagnosed with depressive disorder according to the Structured Clinical Interview for DSM-IV (SCID) or International Classification of Diseases (10th edition) criteria as well as presented typical clinical symptoms of depression with scores above a validated cutoff on depressive rating scales, such as Hamilton Depression Rating Scale (HDRS); (c) patients in the intervention group received rTMS or combined therapy (rTMS plus antidepressant treatments) and were compared with those in the control group who received sham stimulation or usual treatment; and (4) the primary outcome was the effect of therapy measured by the mean change in the depression rating scale scores, remission rates, or response rates.

Data extraction

Two reviewers (P.-P.Y. and Y.-H.Z.) evaluated the eligibility of the relevant studies independently following the inclusion criteria described. Any disagreements on study selections were resolved through consensus or by discussion with the corresponding author (H.-L.L.). The following pieces of information were extracted from each study: (a) first author, year of publication, country, sample size, mean age of the participants, and diagnostic criteria for depression; (b) the depression scale used, other measure scales, and duration of intervention; (c) treatment strategy and rTMS parameter (stimulation site, intensity, and frequency) applied in the intervention and control groups. We extracted raw data of outcomes from each study, such as mean score, SD, number of participants, response rate, and remission rate. We attempted to contact the corresponding authors to request the missing data when the published article did not contain the complete data.

Risk-of-bias assessment

The same two reviewers used version 2 of the Cochrane risk-of-bias tool for randomized controlled trials (www.training.cochrane.org/handbook) to evaluate the quality of the included studies independently. The tool included seven domains: (a) random sequence generation; (b) allocation sequence concealment; (c) blinding of the participants and personnel; (d) blinding of outcome assessment; (e) incomplete outcome data; (f) selective outcome reporting; and (g) other sources of bias. The risk of bias was assessed and rated as low risk (when all domains were rated as low risk), high risk (when one or more domains were rated as high risk), or unclear risk (all other situations). Any disagreements between reviewers were resolved through discussion or consultation with a third reviewer (H.-L.L.).

Statistical analysis

We performed the meta-analysis using Reviewer Manager Software 5.4.1 (Cochrane Collaboration, Oxford, UK). We calculated the pooled estimates using the standardized mean difference (SMD) with 95% confidence intervals (CIs) for continuous outcomes and odds ratio (OR) with 95% CI for dichotomous outcomes. The heterogeneity across the included studies was assessed using the I² value, which measures the percentage of total variation across trials. An I² value >50.0% was recognized as significant heterogeneity. When significant heterogeneity existed, a random-effects model was used to analyze the pooled data; otherwise, a fixed-effects model was used.

Publication bias was evaluated via funnel plot asymmetry only when at least 10 studies were included in the meta-analysis. This study was registered post-study at INPLASY Registry (registration number INPLASY202550075).

Results

Study selection

We identified 271 records by searching 4 databases and reviewing relevant bibliographies. After excluding duplicates, 145 records remained. After screening the titles and abstracts, 122 records were excluded because they failed to meet the inclusion criteria. After reviewing the full text of the remaining 23 records, 3 were excluded because they lacked a suitable control group; 2 were excluded because the intervention mode was intermittent theta burst stimulation, which differs from rTMS; 1 was excluded as the participants were not exclusively patients with LLD; and 7 were excluded because they contained conference abstracts. Ultimately, 10 trials were included in this meta-analysis. The flowchart describing the selection process is shown in Figure 1.

Figure 1.

Flow diagram of the selection of the studies. LLD: late-life depression; rTMS: repetitive transcranial magnetic stimulation.

Study characteristics

The general characteristics of the included trials are described in Table 1. The included trials were published between 2001 and 2020. Of the ten included trials, three were conducted in the USA, two in Canada, one in Austria, and four in China. Two trials focusing on vascular depression, which has a similar pathogenesis to LLD, were also included for analysis.^22,23 Most trials employed a standard diagnostic criterion for depression; however, the trial by Leblhuber et al. ²⁰ did not mention the diagnostic criterion. The sample size ranged from 20 to 178 patients per trial, with a median of 56 patients. The mean age of the participants was >60 years. Overall, 358 patients were assigned to the rTMS group and 323 to the control group. Two trials established two rTMS intervention groups. A study by Jorge et al. ²² designed two experiments to compare the efficacy of two different total dose rTMS groups with two different sham stimulation groups, respectively, and a trial by Trevizol et al. ²⁷ compared the bilateral and unilateral rTMS groups with the same sham stimulation group. Consequently, the two studies contained two sets of experiments, indicated as (a) and (b), respectively, and are listed twice in Table 1. Most trials conducted rTMS treatment on the left dorsolateral prefrontal cortex (DLPC), except for the trial by Leblhuber et al., ²⁰ which chose the bilateral frontopolar cortex as the stimulation location, while the trial by Trevizol et al. ²⁷ conducted bilateral DLPC stimulation beside the left DLPC. The stimulation parameters among trials varied slightly. High-frequency (>10 Hz) stimulation was used to stimulate the left DLPC among the included trials, and low-frequency (1 Hz) stimulation was used to stimulate the right DLPC. The intensity of stimulation ranged from 30% to 120% and was set based on the resting motor threshold (rMT) in most trials. However, the trial by Leblhuber et al. ²⁰ set the rTMS parameter as a fixed intensity of 0.08 T with the 3-Hz stimulation frequency. The duration of rTMS treatment was 20 sessions for 4 weeks in 5 trials and was 10–18 sessions for 2 or 3 weeks in the other 5 trials. All trials used HDRS to evaluate the severity of depression before and after the treatment, and the rates of remission and response based on the change in HDRS scores were also assessed in some trials.^{19,22–24,26,27}

Table 1.

Characteristics of randomized controlled trials included in the systematic review of repetitive transcranial magnetic stimulation (rTMS) for late-life depression.

Trial	Country	Diagnostic criteria	Number of participants (N)		Age of participants (years), mean ± SD		rTMS location	rTMS parameter	Duration of treatment	Outcome measured
			rTMS group	Control group	rTMS group	Control group
Manes et al., 2001 18	USA	DSM-IV	10	10	60.5 ± 3.4	60.9 ± 2.0	Left DLPC	20 Hz with an intensity of 80% resting motor threshold, 2-s pulse train, 20 trains, 800 pulses per session	20 min/day, every weekday for 1 week	HDRS
Jorge et al., 2008 22 (a)Jorge et al., 2008 22 (b)	USA	DSM-IV	1533	1529	62.9 ± 7.264.3 ± 9.4	66.1 ± 1162.1 ± 8.5	Left DLPC	10 Hz with an intensity of 110% resting motor threshold, 6-s pulse train, 20 trains, 1200 pulses per session	1 session/day, 10-day period2 sessions/day, 10-day period	HDRS-17Response rateRemission rate
Narushima et al., 2010 23	USA	DSM-IV	32	11	63.4 ± 3.8	61.5 ± 2.5	Left DLPC	10 Hz with an intensity of 110% resting motor threshold, 6-s pulse train, 20 trains, 1200 pulses per session	Over 10 days for 2 weeks	HDRS-17Response rateRemission rate
Xie et al., 2015 24	China	ICD-10	35	26	65.3 ± 5.1	64.7 ± 4.2	Left DLPC	10 Hz with an intensity of 30% resting motor threshold	20 min a day, 5 times a week for 4 weeks	HDRS-17Response rate
Gu et al., 2017 25	China	CCMD-3	30	30	76.6 ± 8.3	76.1 ± 9.1	Left DLPC	10 Hz with an intensity of 70% resting motor threshold, 5-s pulse train, 1500 pulses per session	1 session/day, 5 days a week, for 4 weeks	HDRS-24
Qin et al., 2017 26	China	CCMD-3	80	98	70.0 ± 6.0	69.4 ± 6.0	Left DLPC	10 Hz with an intensity of 100% resting motor threshold, 4-s pulse train, 800 pulses per session	1 session/day, 5 days a week, for 4 weeks	HDRS-24Remission rateResponse rate
Kaster et al., 2018 19	Canada	DSM-IV	25	27	65.0 ± 5.5	65.4 ± 5.5	Left DLPC	18 Hz with an intensity of 120% resting motor threshold, 2-s pulse train, 6012 pulses per session	5 days per week for a total of 20 treatments over 4 weeks	HDRS -24Response rateRemission rate
Leblhuber et al., 2019 20	Austria	Not mentioned	19	10	71.9 ± 2.9	73.3 ± 2.7	Bilateral FPC	3 Hz, 0.08 T, 30 min per session	1 session/day, 10 sessions	HDRS-7
Trevizol et al., 2019 27 (a) Trevizol et al., 2019 27 (b)	Canada	DSM-IV	2011	1212	66.8 ± 5.866.1 ± 8.5	64.1 ± 3.764.1 ± 3.7	Bilateral DLPCLeft DLPC	1 Hz with an intensity of 120% resting motor threshold on the right DLPC (100 pulses per train, 465 pulses per session), followed by 10 Hz with the same intensity on the left DLPC10 Hz with an intensity of 120% resting motor threshold, 30 pulses per train, 750 pulses per session	5 sessions/week, over 3 weeks	HDRS-17 Response rateRemission rate
Dai et al., 2020 21	China	ICD-10	48	55	69.3 ± 8.7	67.2 ± 9.9	Left DLPC	10 Hz with an intensity of 100% resting motor threshold, 800 pulses per day	20 min/day, 5 times per week for 4 weeks	HDRS-17

CCMD: Chinese Classification and Diagnostic Criteria of Mental Disorders; DLPC: dorsolateral prefrontal cortex; FPC: frontopolar cortex; HDRS: Hamilton Depression Rating Scale; ICD-10: International Classification of Diseases, tenth edition.

Quality assessment

We evaluated the risk of bias for efficacy analysis in each included trial using the Cochrane risk-of-bias tool for randomized controlled trials. The risk of bias in the individual trials is shown in Figure 2, and the proportions of trials with low risk, unclear risk, and high risk of bias in each domain are shown in Figure 3. All included trials claimed randomization, but only two trials reported the process of randomization clearly. Allocation concealment was unclear in most trials, except for the trial by Kaster et al. ¹⁹ Four trials reported the blinding of the participants, which was unclear in the other six trials. On the contrary, six trials reported the blinding of the assessors, which was not mentioned in the other four trials. All trials reported completeness of follow-up for the outcome. Selective reporting was found only in two trials, which did not report the exact HDRS score post-rTMS intervention.^22,27 Other bias assessment showed that the trial by Trevizol et al. ²⁷ did not describe how the participants were blinded to different stimulation sites, and the characteristics of participants described in the trial by Xie et al. ²⁴ might not be sufficient.

Figure 2.

Risk-of-bias assessment for 10 trials.

Figure 3.

Risk-of-bias summaries of each risk-of-bias item presented as percentages across all included trials. rTMS: repetitive transcranial magnetic stimulation.

Efficacy of rTMS

Eight trials reported HDRS scores before and after the invention as the outcome measures. As shown in Figure 4(a), the overall pooled effect size of the reduction of HDRS scores showed a significant advantage of rTMS compared with the control group, with an SMD of 0.51 (95% CI: 0.33, 0.68; P < 0.00001) and no heterogeneity (I² = 0%; P = 0.64). In the meta-analysis of the six trials, which reported the response rates after the intervention, the results showed that patients receiving rTMS were more likely to achieve a significantly greater response rate than those receiving sham-rTMS (OR:  2.62; 95% CI: 1.37, 5.01; P = 0.004) (Figure 4(b)). The heterogeneity was small-to-moderate and not significant (I² = 34%; P = 0.16). For the remission rates after the intervention, the meta-analysis of five trials showed that patients receiving rTMS were more likely to achieve symptom remission than those receiving sham-rTMS (OR: 6.88; 95% CI: 2.93, 16.14; P < 0.00001) (Figure 4(c)), and no heterogeneity (I² = 0%; P = 0.80) was detected across these trials.

Figure 4.

Meta-analyses of the comparison of rTMS and control groups for efficacy outcome. (a) Forest plot of SMDs for the change in HDRS scores; (b) Forest plot of ORs for the response rate of depression based on the change in HDRS scores and (c) Forest plot of ORs for the remission rate of depression based on the change in HDRS scores. HDRS: Hamilton Depression Rating Scale; ORs: odds ratios; SMDs: standardized mean differences; rTMS: repetitive transcranial magnetic stimulation.

Subgroup and sensitivity analyses

We conducted a subgroup analysis to assess the efficacy outcome of rTMS according to different stimulation intensities. As all five trials reporting the remission rates after the intervention performed rTMS with an intensity of  ≥100% rMT, the subgroup analysis included the trials measuring efficacy outcome via HDRS scores and response rates. Among the trials measuring efficacy outcome via HDRS scores, four trials performed focal rTMS with an intensity of  ≥100% rMT and compared the results with those of sham-rTMS (SMD:  3.00; 95% CI: 1.87, 4.12; P < 0.0001; I² = 34%, with no significant heterogeneity) (Figure 5(a)), three trials performed focal rTMS with an intensity of <100% rMT and compared the results with those of sham-rTMS (SMD:  1.12; 95% CI: −0.16, 2.40; P = 0.09; I² = 46%, with moderate heterogeneity) (Figure 5(a)). A statistically significant difference was noted between the two subgroups (P = 0.03). In the trials measuring efficacy outcome by response rate, five trials performed rTMS with an intensity of  ≥100% rMT and compared the results with those of sham-rTMS (OR: 2.56; 95% CI: 1.58, 4.13; P = 0.0001, with no significant heterogeneity) (Figure 5(b)), and only one trial performed focal rTMS with an intensity of <100% rMT and compared the results with those of sham-rTMS. The robustness of the results was tested by sensitivity analysis. We removed each single trial in the abovementioned analyses sequentially, and the changes in the results and heterogeneity did not reach significance, suggesting that the results of our meta-analysis were stable.

Figure 5.

Subgroup analyses of rTMS and control groups for efficacy outcome: (a) Forest plot of SMDs for the change in HDRS scores for stimulation intensity and (b) Forest plot of ORs for the response rate of depression based on the change in HDRS scores for stimulation intensity. rTMS: repetitive transcranial magnetic stimulation; SMDs: standardized mean differences; HDRS: Hamilton Depression Rating Scale; ORs: odds ratios.

Discussion

In this systematic review, we conducted a meta-analysis to pool the currently available clinical evidence of applying rTMS for the treatment of LLD. Our meta-analysis revealed that rTMS is an effective treatment for improving depressive symptoms in patients with LLD. The improvement of depressive symptoms after the treatment assessed by the reduction of HDRS scores, response rate, and remission rate was significant. Although there were some differences in the intervention parameter among the included trials, such as stimulation sites, stimulation frequency, duration of treatment, and versions of HDRS, no significant heterogeneity was observed in the results of our meta-analysis. Most included trials only assessed the depressive symptoms at the end of the treatment, while the trial by Xie et al. ²⁴ reported that rTMS continued to relieve depressive symptoms until 2 weeks after the end of the treatment. Therefore, long-term follow-up trials need to be conducted to explore the potential of a long-lasting effect of rTMS.

In this meta-analysis, we mainly conducted subgroup analysis according to the different stimulation intensities of rTMS. The sensitivity analysis showed that after removing the trials according to potential moderators that may influence the results, such as stimulation location, number of treatment sessions, and study region, the changes in the results did not reach significance. Owing to the robustness of the results, no more subgroup analysis was conducted. As there were limited differences in rTMS parameters and duration among the included trials, subgroup analysis according to other stimulation parameters was not conducted. Most included trials used high-frequency rTMS to stimulate the left DLPC, representing an evidence-based treatment for this disorder. However, both high-frequency stimulation over the left DLPC and low-frequency stimulation over the right DLPC have shown antidepressant effects. ²⁸ For subgroup analysis of different stimulation intensities, there was inconsistency between the trials performing rTMS with an intensity of  ≥100% rMT and the trials performing rTMS with an intensity of <100% rMT. Compared with sham-rTMS, a beneficial effect measured by the change of HDRS scores or response rates was observed in the rTMS group with an intensity of  ≥100% rMT, which was not observed in the rTMS group with an intensity of <100% rMT. Therefore, rTMS with an intensity of  ≥100% rMT might be more effective in the treatment of LLD. As the number of trials performing rTMS with an intensity of <100% rMT is limited, a large number of follow-up studies are still needed to confirm the appropriate stimulation intensity.

Numerous studies have proved the efficacy of rTMS in the treatment of depression, which has been recommended for clinical application. ²⁹ rTMS was approved by the US Food and Drug Administration in 2008 as a treatment for TRD. ³⁰ Evidence from meta-analyses suggests that rTMS had a largely beneficial effect on patients with TRD, with significant improvement in depressive severity. ³¹ In another study, the prevalence of TRD was higher in older adults who often show lower efficacy of conventional antidepressants versus younger adults. ¹⁰ Therefore, rTMS could be an effective nonpharmacological intervention for LLD. The superiority of ECT, which is another effective treatment for major depression, seemed to be more apparent in patients with psychotic depression, while high-frequency rTMS was as effective as ECT in those with nonpsychotic depression. ³² In an uncontrolled study, high-frequency rTMS was found to be a safe, well-tolerated treatment and a useful adjunct to antidepressants in older patients with treatment-resistant depression. ³³ However, the evidence was still limited. A systematic review identified published literature on the antidepressant efficacy of rTMS for geriatric depression and found that rTMS treatment was safe and well-tolerated, showing encouraging efficacy results. ³⁴ Compared with this review, our research included more randomized controlled trials according to restricted included criteria and conducted a meta-analysis to assess the effect of different intensities of rTMS. Another systematic review performed a meta-analysis of the available evidence concerning rTMS efficacy in MDD treatment among older adults and demonstrated that rTMS is an effective and well-tolerated treatment. ³⁵ However, this review included studies that assessed patients who were only slightly older than 50 years, and some of these patients might not have been exclusively diagnosed with LLD. The results of response and remission rates of rTMS therapy from another review conducted by Zhang et al. ³⁶ were coincident with those of the current study. Furthermore, our research included a relatively larger number of trials, and our results considered the reduction in HDRS scores after rTMS treatment as the primary outcome, which was not reported in these previous reviews. Therefore, the results from our research provided more robust evidence for the application of rTMS in the clinical treatment of LLD.

Our research further demonstrated the efficacy of rTMS in the treatment of LLD based on previous studies; however, the detailed antidepressant mechanism of rTMS is still limited. Recent studies have indicated that rTMS may elicit antidepressant effects by modulating the level of inflammatory cytokines as well as a brain-derived neurotrophic factor (BDNF).^37–39 In addition, there are some pathophysiological mechanisms underlying rTMS in the treatment of depression, such as increased concentration of BDNF, increased glucose metabolism in the cortex, and altered brain connectivity in specific neural networks. ⁴⁰ In our research, the trial by Leblhuber et al. ²⁰ suggested an influence of rTMS on the enzyme phenylalanine hydroxylase, which plays a crucial role in the biosynthesis of neurotransmitter precursors related to geriatric depression. However, the specific mechanisms underlying the effect of rTMS on LLD remain unclear and require further attention. Understanding the mechanism of how rTMS affects LLD could be helpful in optimizing the specific stimulation target and parameters and providing a strong theoretical basis for clinical application.

There are several limitations of our research. First, the sample size of rTMS or control groups was small in some included trials; there were 10 groups from 5 trials that had fewer than 20 participants, which may affect the stability of the results. Second, most included trials did not clearly describe the methods of random sequence generation and allocation sequence concealment, which may increase the risk of bias in the results. Third, as the effectiveness of Egger’s publication bias test is commonly limited when the analysis involves fewer than 10 trials, we did not conduct the test in our research. Therefore, the potential publication bias might be neglected in this review, which may have affected the results. Fourth, the outcomes of each included trial, which were analyzed for differences in real vs. sham stimulation, were mostly evaluated immediately after the last rTMS treatment. Therefore, the long-term outcome after rTMS might not be evaluated adequately. Finally, owing to the limitations of the assessment of efficacy, the effects of rTMS on other symptoms, such as cognitive impairment, were not analyzed. Furthermore, post-study registration may have introduced bias to our results.

Conclusion

This systematic review and meta-analysis revealed that rTMS is effective in the treatment of LLD over a short term. In the subgroup analysis, rTMS with an intensity of  ≥100% rMT might be more effective than rTMS with an intensity of <100% rMT. The small number of studies and the lack of long-term follow-up are major limitations of this review. Further studies with larger sample sizes and long-term outcomes are needed in the future.

Data availability statement

The data supporting the findings of this study are available within the article.