Impact of repetitive transcranial magnetic stimulation combined with traditional Chinese Tuina for stroke: a systematic review and meta-analysis Authors: Haihua Xie, Bo Huang, Jia Jiang, Sihui Cao, Dejun Wang, Liang Zhang, Liang Peng, Mi Liu Published: 2025 ABSTRACT AbstractBackgroundStroke has emerged as one of the leading causes of long-term disability worldwide. Although the therapeutic effects of traditional Chinese Tuina and rTMS as monotherapies have been partially established, research on their combined application in stroke rehabilitation remains limited. This meta-analysis aims to investigate the effects of traditional Chinese Tuina combined with rTMS for stroke.MethodsNine databases, including EMBASE, PubMed, Cochrane Central Register of Controlled Trials, Web of Science, Alternative HealthWatch, VIP, Wanfang, CNKI and the Chinese Biomedical Literature Database, were searched from inception to August 5, 2025. Two authors independently performed literature screening, data extraction, and risk-of-bias assessment. Meta-analysis was conducted using STATA software.ResultsEleven studies involving 1,134 patients were included in this meta-analysis. The meta-analysis indicated that, compared to the control group, the intervention group demonstrated significantly higher motor function scores [p < 0.001] and life quality [p < 0.001]. The intervention group also exhibited lower depression scores [p < 0.001] and sleep scores [p < 0.001]. The meta-regression analysis results showed that coil type had a significant impact on quality of life [p < 0.001]; session count and stimulation frequency independently had a marginally significant impact on motor function [p = 0.09 < 0.1]; and quality of life [p = 0.06 < 0.1]. Subgroup analysis suggested that 24 rTMS sessions may yield the most favourable improvement in motor function for post-stroke patients [p < 0.001], while low-frequency and 8-coil demonstrate optimal efficacy in enhancing patients’ quality of life [p < 0.001; p < 0.001]. No publication bias was detected across the analyses.ConclusionsThe combination of rTMS and Tuina significantly improves motor function, enhances quality of life, alleviates depressive symptoms, and improves sleep quality for stroke. 24 rTMS sessions may yield the most favourable improvement in motor function. Moreover, low-frequency and 8-coil demonstrate optimal efficacy in enhancing patients’ quality of life. Introduction Stroke has emerged as one of the leading causes of long-term disability worldwide. According to the World Stroke Organization, disability-adjusted life years attributable to stroke are projected to rise from 144.8 million in 2020 to 189.3 million by 2050 [1], significantly impairing the ability of patients to perform daily activities and reducing their quality of life [2]. Concurrently, the global annual direct costs (treatment, rehabilitation expenses) and indirect costs (productivity losses) of stroke have exceeded $891 billion [1], imposing a substantial economic burden on families and societies. Among rehabilitation strategies for stroke, Traditional Chinese Tuina and repetitive transcranial magnetic stimulation (rTMS) are gaining increasing attention from researchers. As a cornerstone of traditional Chinese medicine, Tuina employs manual techniques to stimulate specific body regions or acupoints, aiming to unblock meridians, regulate Qi and blood circulation, resolve blood stasis, alleviate muscle tension, and modulate neurological function [3,4]. Previous studies have demonstrated that Traditional Chinese Tuina significantly improves motor function, reduces spasticity, and enhances quality of life in stroke patients [5,6]. In true, rTMS, a non-invasive, painless neurophysiological technique, improves neurological recovery of stroke patients by modulating cortical excitability and promoting neuroplasticity [7]. Recent studies have increasingly validated the efficacy of rTMS in addressing post-stroke motor dysfunction [8], cognitive impairment [9], depressive symptoms [10], sleep disorders [11] and activities of daily living (ADL) [12]. However, while the therapeutic effects of traditional Chinese Tuina and rTMS as monotherapies have been partially established [5,13–15], research on their combined application in stroke rehabilitation remains limited [16–18]. On one hand, existing studies are predominantly small-scale clinical trials lacking high-quality evidence to verify synergistic effects. On the other hand, significant heterogeneity exists across studies, including variations in patient inclusion criteria, intervention protocols, treatment parameters, and outcome measures, which compromises the reliability and generalizability of findings. Therefore, this study is designed to carry out a meta-analysis based on randomized controlled trials (RCTs), aiming to examine whether the combination of rTMS and Tuina therapy can yield significantly better improvements in the functional impairment and quality of life of stroke patients, as opposed to using rTMS or Tuina therapy alone. The study results are expected to provide evidence-based guidance for optimizing rehabilitation strategies for stroke patients and promote the integration of this combined therapy into clinical practice. Additionally, this study aims to provide direction for future research and contribute to ongoing progress in the field of stroke rehabilitation. Materials and methods This study strictly adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Supplementary file) [19], and has been registered on the PROSPERO platform with the registration number CRD42024618448 [20]. Inclusion criteria Study participants (P): Patients diagnosed with stroke according to established diagnostic criteria and confirmed via CT or MRI scans.Intervention (I): The intervention group received rTMS combined with traditional Chinese Tuina therapy. To enhance sample size and quality, studies where patients in intervention and control group additionally both received conventional rehabilitation therapy were also included.Control (C): The control group received rTMS alone or traditional Chinese Tuina therapy alone. Similarly, studies where patients in intervention and control group additionally both received conventional rehabilitation therapy were included to ensure methodological rigor and sample diversity.Outcome (O): Any validated assessment tools evaluating post-stroke functional impairments, including but not limited to: motor dysfunction, quality of life, sleep disturbances, dysphagia, and depression.Study type (S): This meta-analysis included registered randomized controlled trials published by August 5, 2025. No language restrictions were applied during literature screening. Study participants (P): Patients diagnosed with stroke according to established diagnostic criteria and confirmed via CT or MRI scans. Study participants (P): Patients diagnosed with stroke according to established diagnostic criteria and confirmed via CT or MRI scans. Intervention (I): The intervention group received rTMS combined with traditional Chinese Tuina therapy. To enhance sample size and quality, studies where patients in intervention and control group additionally both received conventional rehabilitation therapy were also included. Intervention (I): The intervention group received rTMS combined with traditional Chinese Tuina therapy. To enhance sample size and quality, studies where patients in intervention and control group additionally both received conventional rehabilitation therapy were also included. Control (C): The control group received rTMS alone or traditional Chinese Tuina therapy alone. Similarly, studies where patients in intervention and control group additionally both received conventional rehabilitation therapy were included to ensure methodological rigor and sample diversity. Control (C): The control group received rTMS alone or traditional Chinese Tuina therapy alone. Similarly, studies where patients in intervention and control group additionally both received conventional rehabilitation therapy were included to ensure methodological rigor and sample diversity. Outcome (O): Any validated assessment tools evaluating post-stroke functional impairments, including but not limited to: motor dysfunction, quality of life, sleep disturbances, dysphagia, and depression. Outcome (O): Any validated assessment tools evaluating post-stroke functional impairments, including but not limited to: motor dysfunction, quality of life, sleep disturbances, dysphagia, and depression. Study type (S): This meta-analysis included registered randomized controlled trials published by August 5, 2025. No language restrictions were applied during literature screening. Study type (S): This meta-analysis included registered randomized controlled trials published by August 5, 2025. No language restrictions were applied during literature screening. Exclusion criteria Studies that meet any of the following criteria will be excluded:Pre-clinical studies: Cannot be extrapolated to human clinical practice due to species differences.Conference papers: No peer review and unreliable data.Studies containing only experimental protocols: Lack of result data.Editorial material or review: No original data available.Retracted article: The data is unreliable.Case report: The sample size is exceedingly small.Studies for which data could not be obtained despite attempts to contact the authors through multiple channels. Pre-clinical studies: Cannot be extrapolated to human clinical practice due to species differences. Pre-clinical studies: Cannot be extrapolated to human clinical practice due to species differences. Conference papers: No peer review and unreliable data. Conference papers: No peer review and unreliable data. Studies containing only experimental protocols: Lack of result data. Studies containing only experimental protocols: Lack of result data. Editorial material or review: No original data available. Editorial material or review: No original data available. Retracted article: The data is unreliable. Retracted article: The data is unreliable. Case report: The sample size is exceedingly small. Case report: The sample size is exceedingly small. Studies for which data could not be obtained despite attempts to contact the authors through multiple channels. Studies for which data could not be obtained despite attempts to contact the authors through multiple channels. Retrieval strategy A search approach that combines subject headings and free terms was employed. The subject headings selected are ‘Stroke’, ‘Cerebral Infarction’, ‘Cerebral Haemorrhage’, ‘Transcranial Magnetic Stimulation’, ‘Tuina’, ‘Massage’, ‘Acupressure’, ‘Meridians’, and ‘Manipulation’. Meanwhile, free terms from each database will be integrated to conduct a full-text search, aiming to comprehensively retrieve eligible literature. The following nine databases will be searched: EMBSE, PubMed, Cochrane Central Register of Controlled Trials (CENTRAL), Web of Science, Alternative HealthWatch (Alt HealthWatch), China Science and Technology Journal Database (VIP), Wanfang, Chinese National Knowledge Infrastructure (CNKI), and China Biomedical Literature Database were searched. The Supplementary Table 1 provide the search terms and search strategies for all databases. Study selection and data extraction The two investigators (XHH and HB) first independently reviewed the title and abstract, screened studies based on inclusion and exclusion criteria, and cross-checked with each other to complete the preliminary screening. The other two researchers (JJ and CSH) then downloaded and independently read the full text of all potentially eligible literature for further evaluation, while recording specific reasons for exclusion. For all disagreements, the third author (LM) will make the final decision. The above screening was carried out in Note Express software (Beijing Aegean Music Technology Co., LTD., Beijing, China). Data extraction was carried out in a two-person, double-extraction mode, which was independently completed by two investigators (HB and JJ), and included sample size, mean age, sex, stroke type, course, intervention methods for each group, duration of treatment, and observation outcomes. The primary outcomes were motor function and quality of life, and the secondary outcomes were depression and sleep quality. Risk of bias assessment and evidence quality assessment Two investigators (HB and CSH) independently assessed the risk of bias in the included studies based on the Cochrane Bias risk assessment tool provided in the Cochrane Handbook of Systematic Reviews of Interventions [21]. The bias assessment tool covers seven areas: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other bias. The Grading of Recommendations Assessment, Development, and Evaluation (GRADE) was utilized to evaluate evidence quality [22]. In the event of a disagreement, the third researcher (JJ) makes the final decision. In addition, the two reviewers cross-check each other to eliminate any unnecessary errors. Data synthesis If fewer than three studies were included, descriptive analysis was performed. STATA 17 (Stata Corp LLC4905 Lakeway Drive, College Station, TX, USA) software was used for meta-analysis, and forest figures were used to present the data synthesis results. Post-intervention data from the intervention and control group were extracted for meta-analysis to assess the effect of rTMS combined with traditional Chinese Tuina therapy. Standardized mean difference and 95%CI were used to describe continuous data. I2 statistics and Galbraith plots were used for heterogeneity analysis. If I2=0, the fixed effect model was selected. Otherwise, the random effects model is adopted. Subgroup analysis of session counts (20 sessions, 24 sessions, and 30 sessions), stimulation frequencies (low-frequency rTMS and high-frequency rTMS), types of coils (8-coil and O-coil), and univariate meta-regression analysis were conducted. In addition, Egger’s test and Begg’s test were utilized to assess publication bias, while sensitivity analysis was performed to evaluate the robustness of the results. Results Search process and results A total of 810 literatures were searched, and after primary screening and re-screening, 11 studies were finally included [16–18,23–30], with 1134 participants. The literature screening process is shown in Figure 1. Flow chart. General information of included studies General information on the included studies is shown in Table 1. A total of 1134 stroke patients were included in this meta-analysis, of which 711 (62.70%) were male and 423 (37.30%) were female. The mean age range was 53.24–73.1 years in the intervention group and 54.24–74.9 years in the control group. A total of eight studies reported the duration of the patients’ disease, whereas the remaining three studies did not provide this information. In all the studies, the intervention groups received the combination of Tuina and rTMS. In the control groups, six studies utilized rTMS intervention, three studies had patients in the control groups receive Tuina treatment, and two studies set up separate control arms for rTMS and Tuina, respectively. As for outcomes, Bai YJ and Sun LM each employed two distinct outcomes to assess motor function, while all other studies utilized only a single outcome to evaluate the relevant functional aspects. Table 1.Characteristics of included studies.StudySample sizeParticipantsInterventionOutcomesMean age(T/C)Sex(M/F)course(month/day)TMS Stimulation parametersType of coilStimulation siteIntervention groupControl groupBai YJ (2021)112T: 57.84 ± 5.03C: 56.69 ± 4.72T: 31/25C: 34/22T: 40.14 ± 8.53dC: 39.08 ± 8.16d80% RMT, 10Hz, 20min, 1 time/d, 5 d/week, 6 weeks8-coilIpsilateral dorsolateral prefrontal cortexrTMS + TuinarTMS①⑤⑥Sun LM(2024)112T: 53.24 ± 5.92C: 55.27 ± 5.88T: 39/17C: 35/21T: 78.50 ± 12.33dC: 78.42 ± 11.85d90% RMT, 1Hz, 20min,1 time/d, 5 d/week, 4 weeks8-coilContralateral cerebral cortical arearTMS + TuinarTMS③④⑦Jia J(2021)42T: 58.35 ± 3.07C: 57.16 ± 2.61T: 16/5C: 13/8T: 3.25 ± 1.56mC: 2.46 ± 1.68m80% RMT, 1Hz, 23min58s, 1 time/d, 6 d/week, 4 weeks8-coilC3/C4 arearTMS + TuinarTMS③⑦⑧Liu W(2021)80T: 58.83 ± 3.37C: 56.26 ± 4.35T: 21/19C: 24/16T: 3. 65 ± 1.68mC: 2. 86 ± 1.25m80% RMT, 1Hz, 20min, 1 time/d, 6 d/week, 4 weeks8-coilM1 arearTMS + TuinarTMS⑤⑦⑧Dai X(2020)86T: 68.92 ± 2.14C: 68.81 ± 2.12T: 23/20C: 24/19Unreported50-100% RMT, 1Hz, 1 time/d, 3 d/week, 8 weeksO-coilIpsilateral cerebral cortical motor arearTMS + TuinaTuina②④⑦Zhou L (2023)90T: 56.23 ± 6.004C1: 56.23 ± 5.817C2: 57.67 ± 5.927T: 20/10C1: 19/11C2: 18/12T: 5.67 ± 2.309mC1: 5.40 ± 2.159mC2: 5.70 ± 2.395m80% RMT, 10Hz, 20min, 1 time/d, 5 d/week, 3 weeksO-coilLeft dorsolateral prefrontal cortexrTMS + TuinaC1: rTMSC2: Tuina⑦⑨Wang SY(2020)169T: 54.82 ± 5.27C: 55.62 ± 5.41T: 52/33C: 50/34T: 3.42 ± 0.83mC: 3.38 ± 0.91m80% RMT, 5Hz, 28min, 1 time/d, 5 d/week, 8 weeksO-coilLeft dorsolateral prefrontal loberTMS + TuinarTMS⑨⑫Zhang HL(2022)180T: 59.23 ± 8.74C: 58.12 ± 9.23T: 54/36C: 57/33T: 4.26 ± 1.34mC: 4.45 ± 1.41m5Hz, 30min, 1 time/d, 5 d/week, 8 weeksO-coilLeft prefrontal arearTMS + TuinarTMS⑨⑩Huang HQ(2022)80T: 73.1 ± 6.5C: 74.9 ± 6.9T: 31/9C: 33/7Unreported90% RMT, 1Hz, 10min, 1 time/d, 14d/course, 3 coursesUnreportedContralateral pharyngeal cortical representative arearTMS + TuinaTuina⑪Jiang MJ(2025)86T: 55.21 ± 5.37C1: 54.24 ± 6.87C2: 55.39 ± 7.47T: 20/9C1: 19/9C2: 19/10T: 32.23 ± 15.46dC1: 31.17 ± 13.69dC2: 31.52 ± 12.40d90% RMT, 1Hz, 20min, 1 time/d, 6 d/week, 4 weeksO-coilRight dorsolateral prefrontal cortexrTMS + TuinaC1: rTMSC2: Tuina⑨⑩Pei JJ(2022)97T: 67.31 ± 6.81C: 67.28 ± 6.77T: 30/19C: 29/19Unreported80% RMT, 10Hz, 20min, 1 time/d, 5 d/week, 4 weeks8-coilBilateral dorsolateral frontal loberTMS + TuinaTuina⑩Note: T/C: Test group/Control group; RMT:resting motor threshold; ①Berg scale; ②Fugl-Meyer balanced function score(FMA-B); ③Fugl-Meyer upper extremity assessment scale(FMA-U); ④Fugl-Meyer lower extremity assessment scale(FMA-L); ⑤Fugl-Meyer assessment scale(FMA); ⑥Wolf motor function test(WFTT); ⑦MBI score; ⑧Grip strength; ⑨Hamilton depression scale(HAMD); ⑩Pittsburgh sleep quality index(PSQI); ⑪ Water swallowing test; ⑫ the MOS item short from health survey (SF-36). Characteristics of included studies. Note: T/C: Test group/Control group; RMT:resting motor threshold; ①Berg scale; ②Fugl-Meyer balanced function score(FMA-B); ③Fugl-Meyer upper extremity assessment scale(FMA-U); ④Fugl-Meyer lower extremity assessment scale(FMA-L); ⑤Fugl-Meyer assessment scale(FMA); ⑥Wolf motor function test(WFTT); ⑦MBI score; ⑧Grip strength; ⑨Hamilton depression scale(HAMD); ⑩Pittsburgh sleep quality index(PSQI); ⑪ Water swallowing test; ⑫ the MOS item short from health survey (SF-36). Results of bias risk assessment for included study data 10 studies used completely randomized design, and one study did not. Regarding allocation concealment, a total of four studies reported the way of allocation concealment, and the remaining seven studies did not report. Two studies reported detection bias, and nine studies were poorly described. None of the 11 studies reported reporting bias, but their data were complete. Details can be found in Supplementary Figure 1. Meta-analysis results Four outcomes, namely motor function, quality of life, depression, and sleep function, have sufficient data for meta-analysis. Two outcomes, namely swallowing function and balance, have insufficient data for descriptive analysis. Motor function The 7 RCTs reported motor function scores after treatment in both groups. A higher score indicates superior motor function performance. Although the results of the Galbraith plot do not indicate heterogeneity among the studies (Figure 2(A)), the statistic I2 = 65.09% indicates a certain degree of heterogeneity. Therefore, a random effects model was used for analysis in this study. The results showed that the scores of the intervention group receiving rTMS combined with Tuina were higher than those of the control group [SMD= 5.91, 95%CI (4.50, 7.33), p < 0.001, Figure 3(A)]. Galbraith plots. (A) Motor function; (B) Quality of life; (C) Depression; (D) Sleep quality. Meta-analysis. (A) Motor function; (B) Quality of life; (C) Depression; (D) Sleep quality. Quality of life Quality of life scores were reported in seven RCTs. A higher score indicates superior life quality. Although all data points in the Galbraith plot fall within the regression line (Figure 2(B)), the statistic I2 = 80.13% indicates a certain degree of heterogeneity. Therefore, this study thus employed a random-effects model for the analysis. The results showed that the score of life quality in the intervention group was significantly higher than that in the control group after intervention [SMD= 9.31, 95%CI (6.31, 12.31), p < 0.001, Figure 3(B)]. Depression Depression scores were reported in the 6 RCTs. A higher score indicates greater severity of depressive symptoms. Although the results of the Galbraith plot do not indicate heterogeneity among the studies (Figure 2(C)), the statistic I2 = 82.75% indicates a certain degree of heterogeneity. Therefore, a random effects model was used for analysis in this study. The results showed that the depression score of the intervention group was significantly lower than that of the control group [SMD= −3.61, 95%CI (−4.46, −2.75), p < 0.001, Figure 3(C)]. Sleep quality The 4 RCTs reported sleep scores after treatment in both groups. A higher score indicates greater severity of sleep quality. Although all data points in the Galbraith plot fall within the regression line (Figure 2(D)), the statistic I2 = 63.59% indicates a certain degree of heterogeneity. Therefore, this study thus employed a random-effects model for the analysis. The results showed that the sleep function score of the intervention group was significantly lower than the control group [SMD= −2.82, 95%CI (−3.39, −2.24), p < 0.001, Figure 3(D)]. Other function Two studies evaluated the balance function of stroke patients using the Berg scale and the Fugl-Meyer Balance Function score scale, respectively [18,28]. The research results all show that rTMS combined with Tuina can significantly improve the balance function of patients. Only one study reported on swallowing function [23]. The results indicated that post-intervention, the experimental group demonstrated significantly higher scores on the water swallowing compared to the control group. Additionally, the overall response rate was notably higher in the experimental group. Subgroup analysis Session count Subgroup analysis based on motor function revealed significant heterogeneity in treatment efficacy across different session count subgroups [Q = 7.48, p = 0.02, Figure 4(A)]. Among all subgroups, the intervention group demonstrated significantly superior motor function recovery compared to the control group [20-sessions: SMD= 4.29, 95%CI (3.01, 5.57), p < 0.001; 24-sessions: SMD= 7.14, 95%CI (5.35, 8.93), p < 0.001; 30-sessions: SMD= 6.42, 95%CI (4.46, 8.38), p < 0.001]. The 24-session subgroup exhibited the largest SMD, indicating the most pronounced improvement in motor function. Subgroup analysis of session count. (A) Motor function; (B) Quality of life. Subgroup analysis based on quality of life revealed significant heterogeneity in treatment efficacy across different session count subgroups [Q = 25.97, p < 0.001, Figure 4(B)]. Among all subgroups, the intervention group demonstrated significantly superior quality of life compared to the control group [15-sessions: SMD= 6.32, 95%CI (4.26, 8.39), p < 0.001; 20-sessions: SMD= 16.11, 95%CI (12.77, 19.45), p < 0.001; 24-sessions: SMD= 10.29, 95%CI (5.61, 14.97), p < 0.001; 40-sessions: SMD= 5.76, 95%CI (1.18, 10.34), p = 0.01]. The 20-session subgroup exhibited the largest SMD, indicating the most pronounced improvement in quality of life. Stimulation frequency The results of subgroup analysis indicated that different stimulation frequencies exert comparable effects on motor function [Q = 0.26, p = 0.61, Figure 5(A)]. However, significant differences were observed among various stimulation frequencies regarding the improvement of patients’ quality of life [Q = 4.22, p = 0.041, Figure 5(B)]. In the comparison of high-frequency rTMS (HF-rTMS) combined with Tuina therapy versus monotherapy, the intervention group showed significantly better motor function recovery and quality of life in stroke patients compared with the control group [SMD = 6.42, 95% CI (4.46, 8.38), p < 0.001; SMD = 6.23, 95% CI (4.34, 8.11), p < 0.001]. Similarly, low-frequency rTMS (LF-rTMS) combined with Tuina therapy also yielded significantly greater improvements in motor function and quality of life compared to monotherapy [SMD = 5.73, 95% CI (3.96, 7.50), p < 0.001; SMD = 12.27, 95% CI (6.82, 17.73), p < 0.001]. The LF-rTMS subgroup exhibited the largest SMD, indicating the most pronounced improvement in quality of life. Subgroup analysis of stimulation frequency. (A) Motor function; (B) Quality of life. Types of coil Subgroup analysis showed that different coil type exerts comparable effects on motor function [Q = 0.94, p = 0.33, Figure 6(A)]. However, there are significant differences among various coil types in terms of improving patients’ quality of life [Q = 27.41, p < 0.001, Figure 6(B]]. In the 8-coil subgroup, the intervention group demonstrated significantly superior motor function recovery and life quality in stroke patients compared to the control group [SMD = 5.68, 95% CI (4.09, 7.27), p < 0.001; SMD = 15.32, 95% CI (12.50, 18.13), p < 0.001]. Similarly, the O-coil subgroup also showed significantly greater motor function and life quality improvement than the control group [SMD = 6.71, 95% CI (5.38, 8.04), p < 0.001; SMD = 7.07, 95% CI (5.79, 8.34), p < 0.001]. The 8-coil subgroup exhibited the largest SMD, indicating the most pronounced improvement in quality of life. Subgroup analysis of coli type. (A) Motor function; (B) Quality of life. Univariate meta-regression Univariate meta-regression analysis was employed to further explore the sources of heterogeneity. The analysis results showed that differences in coil type had a significant impact on quality of life [coefficient= −8.25, p < 0.001]; differences in session count had a marginally significant impact on motor function [coefficient = 0.25, p = 0.09 < 0.1]; and differences in stimulation frequency also had a marginally significant impact on quality of life [coefficient= −0.65, p = 0.06 < 0.1]. Details can be found in Supplementary Table 2. Publication bias Egger’s test and Begg’s test were utilized to evaluate publication bias. The results revealed that, in the current meta-analysis focusing on motor function [Egger’s test: t = 1.78, p = 0.135 and Begg’s tests: z = 1.20, p = 0.230, Figure 7(A)], quality of life [Egger’s test: t = 0.78, p = 0.472 and Begg’s tests: z = 0.90, p = 0.368, Figure 7(B)], depression [Egger’s test: t = 1.00, p = 0.372 and Begg’s tests: z = 0.38, p = 0.707, Figure 7(C)], and sleep quality [Egger’s test: t= −0.35, p = 0.760 and Begg’s tests: z = 0.34, p = 0.734, Figure 7(D)], neither Egger’s test nor Begg’s test revealed evidence of publication bias. Egger’s test plot and Begg’s test. (A) Motor function; (B) Quality of life; (C) Depression; (D) Sleep quality. Sensitivity analysis The leave-one-out sensitivity analysis demonstrated that the effect sizes for motor function, life quality, depression, and sleep quality remained within the original confidence intervals after sequentially excluding each study (Figure 8). This indicates that the meta-analysis results for these outcomes are robust and reliable. Sensitivity analysis. (A) Motor function; (B) Quality of life; (C) Depression; (D) Sleep quality. Evidence quality The GRADE was utilized to evaluate the quality of evidence. The results demonstrated that for the meta-analyses focusing on motor function, quality of life, and depression, the evidence quality was rated as moderate; whereas for the meta-analysis concerning sleep, the evidence quality was deemed low. Details are provided in Supplementary Table 3. Discussion Our comprehensive analysis reveals that the combined intervention of rTMS and Tuina exhibits significantly greater efficacy in improving motor function, enhancing quality of life, alleviating depressive symptoms, and enhancing sleep quality for stroke patients compared with monotherapy. Furthermore, the coil type, stimulation frequency, and session count of rTMS all influence the therapeutic outcomes. Specifically, 24 rTMS sessions may yield the most favorable improvement in motor function for post-stroke patients. Moreover, low-frequency stimulation and 8-coil demonstrates optimal efficacy in enhancing patients’ quality of life. Restoring motor function is a cornerstone of stroke rehabilitation. Our findings align with previous studies [31–34]. For instance, Barros Galvão et al. [35] investigated the effects of physical therapy combined with 1 Hz rTMS, which targeted the contralesional cortex, in chronic stroke patients with upper - limb spasticity. By employing the Fugl–Meyer Assessment scale, they found that the combined intervention group exhibited significantly more pronounced improvements in motor function. Tuina employs manual techniques to stimulate specific acupoints or body regions, aiming to unblock meridians, regulate Qi and blood circulation, and restore musculoskeletal balance, thereby improving muscle strength and joint mobility [36]. Meanwhile, rTMS modulates cortical excitability and promotes neuroplasticity, accelerating motor recovery [37]. The synergistic application of these modalities establishes a brain-limb synergistic regulatory mechanism, leading to more pronounced motor improvements. Additionally, our study reveals that 24 rTMS sessions may represent the optimal parameter for enhancing motor function, a conclusion that shows a certain degree of alignment with findings from previous studies [38,39]. In the study conducted by Yuan Peng et al. [40], 60 patients with ischemic stroke were enrolled, with 45 of them undergoing rTMS intervention. The results demonstrated that 20 rTMS sessions yielded the most pronounced therapeutic effects, presumably due to its enhancement of functional connectivity between the bilateral hemispheres and within the non-affected hemisphere’s cortico-cortical pathways. However, the conclusions drawn from that study do not entirely concur with our findings. Upon in-depth analysis of the discrepancies, it is plausible that the relatively small sample size, the inclusion of only a specific subtype of stroke patients, and variations in the intervention approach compared to our study collectively contributed to the observed heterogeneity between the studies. Furthermore, considering the marginally significant differences in our statistical results, future research should incorporate larger sample sizes and higher-quality studies for meta-analysis. Improving the quality of life and enhancing the happiness index of stroke patients represent the ultimate objectives pursued in clinical rehabilitation practice. Our study has found that, compared to monotherapy (treatment with rTMS alone or Tuina therapy alone), the combined therapy of rTMS and Tuina demonstrates significantly superior efficacy in enhancing the quality of life of stroke patients, a conclusion that aligns with findings from previous research [16,29]. Specifically, the study conducted by Sun LM et al. [29] included 112 stroke patients who were divided into an experimental group and a control group. Patients in the experimental group received combined rTMS and Tuina therapy, while those in the control group received rTMS alone. The results revealed that, after four weeks of intervention, the Modified Barthel Index scores of patients in the experimental group were significantly higher than those in the control group after intervention. Furthermore, subgroup analysis results further indicated that rTMS interventions employing low-frequency stimulation and an 8-coil configuration yielded the most optimal effects in improving patients’ quality of life. Xie G and colleagues [41] conducted a meta-analysis of 70 studies, which showed that early intervention with 1 Hz low-frequency rTMS in stroke patients could effectively improve their activities of daily living (ADL), and there was insufficient evidence to support the therapeutic effect of high-frequency rTMS on enhancing patients’ ADL. This research finding is relatively consistent with the conclusions of our study. Additionally, when using an 8-coil setup for rTMS treatment, the risk of triggering seizures is extremely low, resulting in a significantly enhanced safety profile [42]. It offers patients a more comfortable experience [43], thereby demonstrating higher applicability for clinical use. Depression is a prevalent psychological complication in stroke patients, severely hindering rehabilitation and quality of life. Hu et al. [44] assessed the effects of acupuncture combined with 5 Hz rTMS, which targeted the left dorsolateral prefrontal cortex, in patients with post-stroke depression. The patients were randomly divided into three groups: an acupuncture-only group, an rTMS-only group, and a combined therapy group. All groups received standard care (5 sessions per week for 4 weeks), and the HAMD showed that the combined group had significantly lower depression scores after the intervention compared with the monotherapy groups. Similar outcomes were reported by VanDerwerker et al. [45] Both Tuina and rTMS regulate neural networks [10,46–48], mitigating emotional tension and depressive symptoms, thereby improving mental health—a critical factor for holistic stroke recovery. In addition, our study also found that rTMS combined with Tuina therapy appears to have a positive effect on improving sleep quality in stroke patients. However, given the limited number of studies included in the analysis and the low quality of evidence assessed using the GRADE system, this conclusion requires further validation and exploration through high-quality studies. However, this study still has certain limitations. Firstly, the number of studies included is relatively small, and there are significant individual differences among patients, which may affect the accuracy and generalizability of the study results. Secondly, the long-term efficacy of the combined treatment remains unclear. Additionally, although the search strategy was formulated based on the team’s professional experience and underwent internal cross-validation, the lack of involvement of information experts (such as librarians) in the review process may pose a risk of undetected search biases. In view of this, future research should employ larger sample sizes, more rigorous study designs, and longer follow-up periods to validate the therapeutic effects. Meanwhile, information experts should be engaged to optimize the search strategy and expand the retrieval of grey literature, thereby further enhancing the quality of evidence. In conclusion, the integration of rTMS and Tuina demonstrates superior efficacy over monotherapy in addressing motor impairment, depression, and functional disability in stroke patients. We anticipate future investigations will further elucidate the mechanisms underlying this combined approach, thereby advancing stroke rehabilitation paradigms. Conclusion The combination of rTMS and Tuina significantly improves motor function, enhances quality of life, alleviates depressive symptoms, and improves sleep quality for stroke patients, with superior efficacy compared to monotherapy. 24 rTMS sessions may yield the most favorable improvement in motor function for post-stroke patients. Moreover, low-frequency and 8-coil demonstrates optimal efficacy in enhancing patients’ quality of life. Supplementary Material Supplementary Table 1.docx PRISMA_2020_checklist.docx Supplementary Table 2.docx Supplementary Table 3 .docx Supplementary Figure 1.jpg