2.1 Data sources and search strategy

The systematic reviews of repetitive transcranial magnetic stimulation in the treatment of Parkinson ’s disease were collected by searching PubMed, EMBASE, Cochrane Library, Web of Science, CNKI, VIP and Wanfang databases. The time range of all searches is from the creation of the database to March 1, 2024. The search terms included “Parkinson’ s disease”, “repetitive transcranial magnetic stimulation”, “transcranial magnetic stimulation”, “meta-analysis”, “systematic evaluation”, “systematic review”. In addition, we also manually searched reviews related to repetitive transcranial magnetic stimulation and Parkinson ’s disease. We use Boolean logic to formulate a retrieval formula, which is applicable to all databases. The search strategy is shown in S1 Table.

2.2 Eligibility criteria

The following inclusion criteria were used to identify eligible studies: (1) study types: published Meta-analysis/systematic review of the literature on rTMS for PD, limited to Chinese and English; (2) participants: patients with PD and were diagnosed according to any internationally recognized clinical guidelines [4]; (3) intervention: the experimental group was treated with rTMS, and the treatment parameters such as frequency and treatment site were not limited, or rTMS combined with other treatments, including PD conventional drug treatment, acupuncture and moxibustion, etc. The control group was treated with drugs, sham stimulation or blank control; (4) outcomes: reviews that assessed the motor (such as UPDRS III, walking performance, timed up-and-go test, etc.) and non-motor symptoms (such as depression, cognitive impairment, etc.) of PD as the main outcome measures were considered eligible. Measures of quality of life and activities of daily living were included if they were relevant to the assessment of PD symptoms.

Exclusion criteria: (1) studies without complete data, such as conference abstracts, letters or comments; (2) network meta analysis or indirect comparison; (3) not rTMS.

2.3 Data extraction and management

Two authors (LW. Z and YH.J) independently conducted literature search and data extraction, and a third author (H.X) made the decision in case of disputes. Import the retrieved systematic reviews into Endnote, delete duplicate content, and select literature that may meet the criteria by reading the title and abstract. Systematic review that met the criteria were ultimately included by reading the full text. Information was extracted from all included systematic reviews. The extracted information included: authors, number of studies, number of samples, treatment group, control group, rTMS basic information, risk assessment tools, adverse reactions, outcomes, and main conclusions.

2.4 Quality assessment

We used A Measurement Tool to Assess Systematic Reviews 2 (AMSTAR 2) to evaluate the methodological quality of SRs [24]. The tool includes a total of 16 items, of which items 2, 4, 7, 9, 11, 13, and 15 are considered critical items, and the rest are non-critical items. Each item is described with “Yes” and “No”. If there are no defects or only one non-critical item defect, the methodological quality is rated “High” and the conclusion of SRs is accurate and comprehensive. If there is more than one non-critical item defect but no critical item defect, the methodological quality is rated as “Moderate” and the conclusion of the SRs is accurate. If there is one critical item defect, the methodological quality is rated as “Low” and the conclusion of the systematic review is low. If more than one critical entry is rated as “No”, the methodological quality is “Very low” and the system evaluation decisions are inaccurate and incomplete. Two authors independently assessed the methodological quality of included SRs.

We used the Risk of Bias for Systematic Reviews (ROBIS) tool to assess the risk of bias (RoB) for SRs [25]. The evaluation process is divided into three phases: (1) assessing relevance; (b) phase 2 included four domain (study eligibility criteria, identification and selection of studies, collection and study appraisal, synthesis and findings); (3) risk of bias in the review. Two authors independently classified the risk levels into “low risk”, “high risk” and “unclear risk” and any disagreements were adjudicated by a third author.

We used GRADE to assess the quality of the primary results of SRs [26]. Five key factors that affect the quality of evidence include: study limitations, inconsistency of results, indirectness of evidence, imprecision and reporting bias. The quality of evidence for SRs was rated as “High” (not downgraded), “Moderate” (downgraded one level), “Low” (downgraded two levels), and “Very low” (downgraded three or more levels).

3.1 Literature retrieval results

Following the search strategy, a total of 81 systematic reviews (SRs) were identified. After excluding 40 duplicate papers through filtering and 8 papers based on title and abstract review, the remaining 33 SRs underwent full-text evaluation. Among these, 8 articles with interventions other than repetitive Transcranial Magnetic Stimulation (rTMS), 3 network meta-analyses, 4 articles comparing various TMS techniques, and 2 non-SRs were eliminated, resulting in a final inclusion of 16 articles. The literature screening process is shown in Fig 1.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Flow chart of the literature search and study selection process.

https://doi.org/10.1371/journal.pone.0313420.g001

3.2 Basic features of the included literature

The basic characteristics of the included 16 SRs are shown in Table 1 [27–42]. A total of 16 SRs met inclusion criteria, and the literature published period ranged from 2015 to 2022. The total number of included original studies was 260, with a maximum of 32 and a minimum of 5. Sample sizes ranged from 160 to 1084, with 3 SRs not reporting sample sizes [30, 31, 33]. The treatment in the control group was mainly PD routine treatment, routine treatment combined with sham stimulation, and only sham stimulation. The treatment of the experimental group was mainly rTMS combined with routine treatment, and a variety of rTMS. The basic characteristics of rTMS include a stimulation frequency ranging from 0.2 to 50 Hz, stimulation intensity varying from 70% to 120% of resting motor threshold (RMT), and a number of pulses ranging from 100 to 3000. Common stimulation sites are the left (or right) dorsolateral prefrontal cortex (lDLPFC or rDLPFC), left occipital cortex (LOC), inferior frontal gyrus (IFG), supplementary motor area (SMA), primary motor cortex (M1), premotor dorsal cortex (PMD), motor cortex (MC), and prefrontal cortex (PFC). For risk assessment, 12 SRs used the Cochrane risk of bias tool, 1 SR used the Jadad Quality Assessment Scale, 1SR used the Physiotherapy Evidence Database (PEDro) scale, 1 SR used the Oxford Center for Evidence Based Medicine levels, and one did not mention bias assessment. Overall, while many researchers suggest that rTMS shows promise in treating PD, further high-quality studies are needed to confirm these findings.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of the included systematic reviews.

https://doi.org/10.1371/journal.pone.0313420.t001

3.3 Evaluation results of methodological quality

Table 2 displays the results of evaluating methodological quality using the AMSTAR 2 tool. Of the 16 SRs, 2 (12.5%) were rated as high quality (29,40), 7 (43.75%) were rated as low quality [27, 28, 32, 34, 38, 39, 41], and 7 (43.75%) were rated as very low quality [30, 31, 33, 35–37, 42] (Fig 2A). Among the critical items, item 2 has a compliance rate of 18.75%, item 4 has a compliance rate of 68.75%, item 7 has a compliance rate of 75%, item 9 has a compliance rate of 75%, item 11 has a compliance rate of 68.75%, item 13 has a compliance rate of 81.25%, and item 15 has a compliance rate of 75%, thus most of the SRs are deficient in the critical items. Among the non-critical items, items 1, 3, 10, and 16 all met the standards, the compliance rate for items 5, 6, and 8 was 68.75%, and the compliance rate for items 12 and 14 was 62.5%.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Methodological quality, risk of bias, and evidence quality of included systematic reviews.

(A) AMSTAR 2; (B) Risk of bias; (C) evidence quality.

https://doi.org/10.1371/journal.pone.0313420.g002

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Methodological quality of included systematic reviews.

https://doi.org/10.1371/journal.pone.0313420.t002

3.4 Results of ROBIS evaluation

In the first stage of ROBIS, the relevance of the study topic was assessed, and all systematic reviews (SRs) were deemed to have a low risk of bias. Domain 1 evaluated study eligibility criteria, with all 11 SRs rated as low risk of bias [27–30, 32, 34, 38–42]. Domain 2 looked at the identification and selection of studies, revealing that 9 SRs had a low risk of bias [27–30, 32, 34, 38–42]. Domain 3 focused on the collection and evaluation of studies, finding that 11 SRs were at low risk of bias [27–30, 32, 34, 36–38, 40–42]. Domain 4 examined synthesis and findings, showing that 6 out of 16 SRs were rated as low risk of bias [27, 29, 34, 38–40]. In the final stage, the overall risk of bias of the reviews was considered, and 6 (37.5%) SRs were found to be at low risk of bias [27, 29, 34, 38–40]. Refer to Table 3 and Fig 2B for more detailed information.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Risk of bias of the included systematic reviews.

https://doi.org/10.1371/journal.pone.0313420.t003

3.5 Evidence quality level evaluation results

Using the GRADE system, 16 papers were analyzed for 62 outcome indicators. The findings revealed that there was no high quality evidence present in the SRs. Moderate quality evidence was found in 10 outcomes (16.13%), while low quality evidence was present in 21 outcomes (33.87%), and very low quality evidence in 31 outcomes (50%). Overall, the quality of evidence was deemed to be generally low. This was attributed to several factors: (1) Serious flaws in blinding, allocation concealment, and randomization in the included literature, impacting the validity of research results; (2) Heterogeneity in outcome indicators affecting result reliability; and (3) Asymmetric forest plot suggesting potential publication bias. Further details on the evaluation of evidence quality can be found in Table 4 and Fig 2C.

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Results of evidence quality with GRADE.

https://doi.org/10.1371/journal.pone.0313420.t004

3.6 Main outcome measures

Most SRs have focused on the efficacy and safety improvements of rTMS in patients with PD in the following areas: motor symptoms, depressive symptoms, cognitive function, ability to perform activities of daily living, and adverse effects.

4.1 Summary of the main results

We conducted a comprehensive descriptive analysis of 16 SRs of rTMS for PD, involving 260 original studies. Evidence from the majority of studies suggests that rTMS combined with conventional treatment for PD is more effective than conventional treatment or sham stimulation alone in improving motor symptoms in PD patients and does not cause serious adverse effects. Some of the SRs evidence also suggests that rTMS does not improve PD patients motor symptoms compared to conventional treatment. For non-motor symptoms, such as cognitive function and depressive symptoms, conclusions on rTMS for PD have not been consistent. The methodological quality of the SRs was assessed using the AMSTAR 2 tool, and the methodological quality of the SRs was variable, with 12.5% (2/16) of the SRs rated as high quality, 43.75% (7/16) rated as low quality, and 43.75% (7/16) rated as very low quality. The risk of bias was assessed using ROBIS and 37.5% (6/16) were rated as low risk. The GRADE results suggested that 16.13% (10/62) of the evidence was of moderate quality, 33.87% (21/62) of the evidence was of low quality and 50% (31/62) of the evidence was of very low quality.

According to AMSTAR 2, among the 16 SRs, 12.5% were rated as high quality, and the methodological quality of the remaining (87.5%) SRs was rated as low or very low quality. It is worth noting that in critical item 2, 81.25% (13/16) of SRs did not provide a previous protocol, which means that we cannot guarantee whether the protocol was strictly followed during the research report, thereby increasing the risk of bias. In critical item 4 and 7, 5 (31.25%) reports did not adequately provide search strategies, and 4 (25%) reports did not adequately provide exclusion lists. In addition, some of the SRs were unable to statistically combine results using appropriate methods, did not consider RoB when interpreting results, and did not adequately investigate publication bias. In non-critical items, 68.75% of reviewers performed duplicate literature screening, duplicate data extraction, and described the included studies in sufficient detail. 25% of the reviews did not provide satisfactory explanations and discussions when analyzing heterogeneity. In terms of risk of bias, all SRs rated low risk when assessing relevance. In the second stage, 19.23% of SRs rated high risk when assessing study eligibility criteria, 43.75% of SRs rated high risk when identifying and selecting studies, 31.25% of SRs rated high risk when data collection and study evaluation, and 62.5% of SRs rated high risk when synthesizing and concluding. In the third stage, after comprehensive review, only 37.5% of SRs were rated as low risk. Regarding the evaluation of the quality of the evidence, no high-quality evidence was found, 16.13% of the evidence was of medium quality, 33.87% of low-quality evidence, and 50% of very low-quality evidence. The risk of publication bias, inconsistency and imprecision constituted the most important determinants of evidence downgrading. Significant heterogeneity can lead to inconsistency, while limited sample sizes or too large confidence intervals can lead to imprecision. Notably, although most of the included SRs indicated that rTMS appears to be an effective treatment for PD, most authors did not draw clear conclusions due to the small sample size or low quality of the included trials. Therefore, to determine whether rTMS is beneficial for PD, more high-quality, large-sample RCT studies must be conducted.

rTMS was developed as an extension of TMS, utilizing a consistent stimulation intensity and frequency to target specific brain regions with a series of pulses [43, 44]. This technique is valuable for investigating the functional reorganization of neural networks. rTMS offers the benefits of bidirectional regulation of the cerebral cortex, ease of use, and non-invasiveness. A large number of studies have shown that rTMS can effectively alleviate motor symptoms, cognitive impairment and insomnia symptoms in patients with PD [36, 37, 44]. The therapeutic effects of rTMS on Parkinson’s disease could be linked to a variety of mechanisms. First, rTMS penetrates the skull to the cerebral cortex through magnetic pulses, stimulates the cerebral cortex and peripheral nerves to generate induction currents through the principle of electromagnetic induction, increases the excitability of neurons, prompts the synaptic activation of neurons in the inhibited state, and remodels the damaged neural pathways, which is conducive to the reconstruction of the function of cerebral cortical network, thus improving the symptoms of PD [44, 45]. Second, high-frequency rTMS improves cerebral blood flow, enhancing overall brain excitability and energy metabolism, which can benefit non-motor symptoms of PD, particularly in cognitive domains like language, memory, attention, and executive function [30, 32, 33]. Third, rTMS can induce synaptic structural and functional plasticity changes by influencing synaptic connectivity, neuronal morphology, synaptic gaps, and long-term potentiation (LTP) and long-term depression (LTD), leading to lasting neuroplasticity effects that contribute to long-term improvement in PD symptoms [46, 47]. Lastly, rTMS can activate the phosphorylation process, enhance biological signaling responses, promote neural cell proliferation and differentiation, and regulate the expression and function of various neurotrophic factors, highlighting its potential in neuroprotection and repair [48–50]. In summary, the multi-dimensional action mechanism of rTMS in treating PD is not only reflected in direct neurophysiological regulation, but also includes the enhancement of neuroplasticity and the optimization of the neurotrophic environment. Several studies have shown that rTMS applied to the cerebral cortex can increase oxygen levels in the striatum and thalamus, trigger dopamine release in the striatum, and reduce beta rhythmic neural oscillations in the thalamic nucleus. These findings indicate that the effects of rTMS stimulation may also impact other brain regions that are connected both structurally and functionally [51–53]. The primary motor cortex (M1), dorsolateral prefrontal cortex (DLPFC), and supplementary motor cortex (SMA) are currently the main stimulation target areas for improving motor symptoms of PD. Previous clinical studies have demonstrated that high-frequency rTMS stimulation of the M1 area yields significant improvements in bradykinesia and rigidity [19]. Conversely, low-frequency stimulation of the M1 area does not exhibit a notable impact on motor symptom enhancement. While an increasing number of studies support the efficacy of high-frequency M1 region stimulation in alleviating motor symptoms, its therapeutic benefits remain somewhat limited when compared to dopamine drug therapy [54]. Furthermore, high-frequency rTMS stimulation of the left DLPFC shows promise in ameliorating both motor and non-motor symptoms, particularly depression in PD [55]. Additionally, rTMS stimulation of the SMA has been found to improve freezing of gait [56].

4.2 Implications for future study

The included systematic evaluations were assessed using the AMSTAR-2, ROBIS, and GRADE tools, and the results showed limitations in currently published systematic evaluations of rTMS for the treatment of PD, identifying limitations that need to be improved in the future. Only three SRs submitted protocols prior to trial initiation. For the majority of SRs, it is essential to have their study plans pre-registered or published on reputable platforms such as PROSPERO or Cochrane. Additionally, it is crucial to take into account gray literature in the research process. Detailed search curation and exclusion lists should be provided for inclusion and selection of original studies. During search and data extraction, two reviewers should perform literature screening and data extraction, respectively. During the data analysis phase, it is imperative to employ suitable statistical methods to synthesize the findings. The risk of bias (RoB) must be meticulously assessed and factored into the interpretation of the results. Furthermore, a comprehensive examination of publication bias is necessary. In addressing heterogeneity, it is essential to offer a coherent explanation and engage in a thorough discussion to elucidate the observed variability. Many research designs lack proper reporting of blinding procedures, allocation concealment, and randomization methods, leading to increased risks of biases and heterogeneity. The limited number of high-quality randomized controlled trials on rTMS for PD can be attributed to the small number of included studies, low sample size, inconsistent treatment selection, data extraction, and result labeling. However, determining an ideal sample size requires consideration of a variety of statistical and clinical factors, including expected effect size, variability, significance level, and the statistical power of the study. Additionally, ethical and practical recruitment issues need to be considered. We recommend that future researchers consult statisticians when designing large-scale RCTs and use sample size calculation software or methods to determine appropriate sample sizes. Sample sizes used in similar studies, as well as information from clinical trial registries, can be consulted to guide sample size selection. Heterogeneity is further influenced by variations in rTMS parameters, combined therapies, individual differences, mental health, disease severity, and other factors.

4.3 Limitation

As an overview of rTMS in the treatment of PD, this study can provide a comprehensive evidence reference for clinical practice. In addition, through the evaluation process of AMSTAR-2, ROBIS and GRADE, it is found that SRs have obvious limitations, which may help guide future high-quality research. This overview has certain limitations. Assessment of quality is a subjective process, and researchers may have their own judgments about each factor. We can only provide a comprehensive and qualitative description of all the data and cannot make a quantitative evaluation. It is recommended that clinical personnel refer to real-life scenarios when using this data for clinical decision-making.