Acupuncture and Moxibustion for Poststroke Depression: Systematic Review

Introduction

Poststroke depression (PSD) is a prevalent neuropsychiatric complication following stroke, characterized by depressed mood, loss of interest, and diminished quality of life during the recovery process [1]. PSD not only impairs patients’ quality of life but also exacerbates cognitive dysfunction, delays neurological recovery, and elevates the risk of recurrent stroke and mortality [2]. The pathogenesis of PSD remains incompletely understood but is thought to involve a combination of stroke-induced damage to emotional regulatory brain regions, resulting in neurotransmitter imbalances, reduced ability to perform daily functions, and insufficient social support [3].

Globally, the prevalence of PSD ranges from 17% to 34.3% [4]. Its incidence is influenced by factors such as age, gender, education level, marital and economic status, and stroke type [5]. The mainstays of Western medical management for PSD include pharmacotherapy, most notably selective serotonin reuptake inhibitors and norepinephrine-dopamine reuptake inhibitors, and psychotherapeutic approaches such as cognitive-behavioral and interpersonal therapies [6,7]. However, pharmacotherapy is frequently associated with adverse effects such as dry mouth, constipation, dizziness, somnolence, sexual dysfunction, and, in some patients, lack of efficacy or development of drug resistance [8-10].

Acupuncture, an integral part of traditional Chinese medicine, has demonstrated beneficial effects for PSD. Clinical evidence suggests that acupuncture can alleviate depressive symptoms, improve quality of life, and reduce the adverse effects induced by pharmacological agents [11,12]. The putative mechanisms include modulation of neurotransmitter systems, enhancement of cerebral blood flow, and regulation of immune function [13]. Numerous clinical trials and case studies point to acupuncture and moxibustion as effective interventions for PSD [14,15]. In recent years, multiple systematic reviews have synthesized the evidence on acupuncture and moxibustion for PSD, occasionally yielding inconsistent or conflicting results [16,17].

Therefore, this study presents an overview of systematic reviews that comprehensively summarizes the current state of research on acupuncture and moxibustion for PSD. By synthesizing existing evidence, it aims to provide a scientific foundation for clinical practice, clarify methodological limitations of previous studies, and propose recommendations for future research.

Methods

Protocol and Registration

The protocol for this overview of systematic reviews was registered with PROSPERO (International Prospective Register of Systematic Reviews; CRD42024576753). The reporting of this study adheres to PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 guidelines [18].

Data Sources and Search Strategy

A total of 8 electronic databases were searched: PubMed, Embase, Cochrane Library, Web of Science, China National Knowledge Infrastructure, Wanfang, VIP Database, and Chinese Biomedical Literature Database. The search included records published up to August 10, 2024. Keywords used were acupuncture, moxibustion, electroacupuncture, manual acupuncture, acupoint catgut embedding, systematic review, meta-analysis, and PSD. Detailed search strategies for each database are provided in Table S1 in Multimedia Appendix 1.

Eligibility Criteria

Studies were included if they met the following criteria: (1) systematic reviews or meta-analyses of randomized controlled trials (RCTs) or quasi-RCTs focused on acupuncture and moxibustion for PSD; (2) participants diagnosed with PSD, confirmed by computed tomography or magnetic resonance imaging, regardless of gender, age, ethnicity, education, or economic status; (3) interventions applied to the experimental group included acupuncture, moxibustion, electroacupuncture, manual acupuncture, acupoint catgut embedding, or transcutaneous electrical acupoint stimulation, alone or in combination with Western medicine or basic medications. Control groups received interventions other than acupuncture or moxibustion (eg, Western medicine, placebo, and sham acupuncture) to compare the efficacy of acupuncture with other treatment methods or placebo in the treatment of PSD; (4) main outcome indicators included overall effectiveness rate, Hamilton Depression Rating Scale (HAMD), Self-Rating Depression Scale (SDS), and the modified Edinburgh-Scandinavian stroke scale scores; and (5) no restrictions were imposed on the duration of the follow-up period (Table S2 in Multimedia Appendix 1 [16,19-41]).

Exclusion criteria were (1) systematic reviews or meta-analyses that did not include RCTs; (2) studies involving patients without a clear diagnosis of PSD; (3) control group interventions that included acupuncture; (4) duplicate publications; (5) protocols for systematic reviews, commentaries, or conference abstracts; and (6) studies lacking complete or essential information.

Literature Selection

All studies identified in the search were imported into EndNote (EndNote X20; Clarivate) for deduplication. Two independent reviewers (LM and CLX) screened titles and abstracts according to the inclusion and exclusion criteria and further independently assessed the full text of potentially eligible studies. Disagreements were resolved by discussion, with a third reviewer (XCT) involved when necessary.

Data Extraction

Data extraction was conducted independently by 2 reviewers (LM and CLX). Extracted data included (1) bibliographic details (title, authors, and year of publication); (2) number and sample size of included studies; (3) intervention methods, control interventions, outcome results, 95% CI, I² statistics, P values, and study conclusions; and (4) methodological quality assessments of the included studies. Any discrepancies were resolved by discussion with a third reviewer (XCT).

Quality Assessment

Methodological Quality

The included systematic reviews were assessed using the Assessment of Multiple Systematic Reviews-2 (AMSTAR 2) tool [42]. Items 2, 4, 7, 9, 11, 13, and 15 were considered critical domains. Reviews were categorized as “high” quality (no noncompliance on critical items), “moderate” quality (noncompliance on noncritical items only), “low” quality (1 critical item noncompliant), and “very low” quality (more than 1 critical item noncompliant).

In addition, we used the Cochrane Risk-of-Bias tool to reassess the risk of bias in the original RCTs included in the systematic reviews and performed statistical analyses on items with a high risk of bias, such as random sequence generation, allocation concealment, and blinding.

Reporting Quality

PRISMA 2020 [18] was used to evaluate reporting quality. Each of the 27 items was scored as 1 (fully reported), 0.5 (partially reported), or 0 (not reported), for a maximum score of 27. Scores <15 indicated “severely deficient” reporting, 15.5‐21 “somewhat deficient,” and 21.5‐27 “relatively complete” [43].

Evidence Quality

The Grading of Recommendations Assessment, Development and Evaluation (GRADE) system [44,45] was used to assess the quality of each outcome. Evidence was categorized as high, moderate, low, or very low, in accordance with Cochrane and contemporary guidelines [46]. Quality assessments were carried out independently by 2 reviewers (LM and XXH), with disagreements resolved through discussion or consultation with a third reviewer (XCT).

Heterogeneity Assessment

Statistical heterogeneity among studies was assessed using the I² statistic, quantifying the percentage of variability due to heterogeneity rather than chance [47,48]. I² values of 25%, 50%, and 75% were considered low, moderate, and high heterogeneity, respectively [49].

Data Synthesis and Analysis

A narrative synthesis was performed, focusing on study population, sample size, interventions, outcome indicators, relative effect sizes, I² statistics, P values, and main conclusions. Where possible, results were also summarized in a tabular form. All collected data were entered into Microsoft Excel 2019 for qualitative synthesis. Overlap of primary studies was assessed using the corrected covered area (CCA) method and calculated as follows: CCA (%)=(N-r)/(rc-r), where N is the total number of included observations from all reviews, r is the number of unique primary studies, and c is the number of systematic reviews. Interpretation was categorized as 0%‐5% (slight overlap), 6%‐10% (moderate overlap), 11%‐15% (high overlap), and >15% (very high overlap).

Results

Literature Search Results

A total of 251 records were screened for eligibility. After removing 107 duplicates, 144 articles remained. Upon review of titles and abstracts, 78 articles were excluded, with an additional 42 excluded after full-text assessment. Ultimately, 24 articles [16,19-41] met the inclusion criteria for this overview (Figure 1).

‎

Study Characteristics

The included studies comprised 24 systematic reviews and meta-analyses, with 11 papers published in English and the remaining in Chinese [20,21,24-26,28,29,31,33,37-39,41] (Table 1). Interventions in the experimental group included acupuncture, electroacupuncture, scalp acupuncture, alone or in combination with Western medicine. Comparator groups received Western medicine, sham acupuncture, placebo, or no treatment. All systematic reviews evaluated the methodological quality of their included studies, with 17 [16,20,23-26,28-30,32,34,35,38-41] using the Cochrane Collaboration Risk of Bias tool, 3 [19,21,33] using the Jadad scale, and 3 [27,31,37] combining both tools. A total of 21 systematic reviews [16,19,21,23-26,28-41] concluded that acupuncture was effective for the treatment of PSD, while 3 [20,22,27] reviews found insufficient evidence to support its benefits. Acupuncture was associated with fewer adverse events than antidepressants.

Methodological Appraisal

The methodological quality of the included reviews was assessed using the AMSTAR 2 tool. Every review had at least 1 critical domain rated as “very low” in methodological quality. In particular, major deficiencies were noted in protocol registration (Item 2), provision of excluded studies (Item 7), and comprehensive assessment of publication bias (Item 15; Figure 2). Out of the 24 included studies, 22 [16,19-26,28-39,41] were rated as “critically low” and 2 [27,40] as “low” in methodological quality. Notably, most studies (21/24) reporting positive findings for acupuncture had critically low methodological quality, while all 3 reviews [20,27,29] indicating insufficient evidence also had methodological shortcomings.

‎

Reporting Quality Appraisal

Reporting quality was evaluated using the PRISMA 2020 checklist. Scores for the 24 studies ranged from 14 to 23. One study [22] was classified as “severely deficient” (score <15), 21 studies [19-21,23-33,35-41] as “somewhat deficient” (score 15‐21), and 2 studies [34,40] as “relatively complete” (>21). Underreported items included data items (10), methods for evaluating quality of evidence for each outcome (15), presentation of evidence grading (22), registration and protocol (24), competing interests (26), and public information reporting (27) (Figure 3).

‎

Risk of Bias in Original Literature

After deduplication, 222 original studies remained. Of these, 46 articles (20.7%) were classified as high risk of bias, while the remainder were labeled as having “some concern.” Among the high-risk domains, random sequence generation constituted 8.6% (19/222), blinding of participants and personnel 7.7% (17/222), selective reporting 4.5% (10/222), and other biases 1.4% (3/222) (Figure 4). Further details can be found in Multimedia Appendix 2. in

‎

Classification of Evidence Quality

The quality of evidence for the outcome indicators, covering 6 different intervention types, ranged from very low to high. Common downgrading factors included high risk of bias, significant heterogeneity, small sample sizes, and inconsistency of results, as outlined in Table 2 [16,19-41].

Heterogeneity Assessment

Of the reported outcomes, 25 of 60 (42%) demonstrated no heterogeneity (I²<25%), 11 of 60 (18%) showed low heterogeneity (25% ≤I²<50%), 8 of 60 (13%) moderate heterogeneity (50% ≤I²<75%), and 16 of 60 (27%) high heterogeneity (I²≥75%). Meta-analyses with high heterogeneity primarily used clinical effective rate and HAMD score as outcomes (Table S2 in Multimedia Appendix 1) [16,19-41].

Overlap of Primary Studies

The primary research indicators we observed were mostly mild to moderate overlap (Table 3). A CCA <10% may suggest incomplete coverage of original studies in the field, with key clinical trials not being adequately included (eg, publication bias of negative results), and a lack of targeted evidence for subgroup populations (eg, different stroke types and severity of depression).

Discussion

Main Findings

Acupuncture therapy is increasingly used for PSD in clinical practice, primarily due to its capacity to mitigate depressive symptoms, enhance patients’ quality of life, and diminish the adverse side effects typically associated with pharmacological treatments. In recent years, the body of systematic reviews on acupuncture for PSD has grown substantially. This study undertook a comprehensive evaluation of the existing literature, focusing on methodological quality, reporting quality, and evidence quality, to elucidate current weaknesses and provide guidance for future research. Overall, our findings indicate that the included systematic reviews and meta-analyses generally experienced low methodological, reporting, and evidence quality, largely a result of the poor quality of primary studies, especially due to high risks of bias such as inadequate randomization and insufficient allocation concealment. Furthermore, the primary research indices demonstrated mainly mild to moderate overlap, suggesting that coverage of original studies in this field remains incomplete, leaving potentially pivotal clinical trials unincorporated.

Potential Mechanism

The pathogenesis of PSD is multifactorial. Two major hypotheses involve neurotransmitter imbalance and glutamate-mediated excitotoxicity [50]. The monoaminergic mechanism posits that lower levels of monoamines—such as serotonin, norepinephrine, and dopamine—are associated with depressive symptoms [51]. In addition, glutamate, the brain’s principal excitatory neurotransmitter, is often elevated in patients with PSD, reflecting a state of excitotoxicity and contributing to depressive pathology [52].

Acupuncture is believed to modulate these neurochemical disturbances. Animal and clinical studies suggest that acupuncture elevates levels of serotonin, norepinephrine, and DA in the brain, normalizes glutamate concentrations, and upregulates key neurotransmitter receptors with fewer desensitization issues compared with pharmacotherapy [53-55]. Beyond neurotransmitter effects, acupuncture exerts anti-inflammatory actions, reducing levels of proinflammatory cytokines and improving immunological balance—processes increasingly recognized as pivotal in PSD progression [56-61]. Furthermore, acupuncture appears to modulate the hypothalamic–pituitary–adrenal axis, helping to normalize neuroendocrine function, and may also reduce neuronal apoptosis, particularly within the hippocampus, which is closely linked to mood regulation [62-65].

Quality Summaries

The methodological quality of the included reviews was evaluated using the AMSTAR 2 system. The methodological strengths of the included reviews were the use of appropriate methods for outcome statistics (24/24, 100%), a satisfactory assessment of risk of bias (24/24, 100%), and study selection and data extraction in duplicate (15/24, 62.5%). Common methodological weaknesses included the failure to provide preregistered protocols (5/24, 20.8%), explain ROB in individual studies when interpreting results (7/24, 29.2%), provide a full list of excluded studies (0/24, 0%), and extract data on funding sources (0/24, 0%).

The methodological quality of reviews was systematically assessed using AMSTAR 2. All included reviews applied appropriate statistical methods and risk of bias assessments (24/24, 100%) and most performed duplicate selection and data extraction (15/24, 62.5%). However, major shortcomings were evident: only a minority preregistered protocols (5/24, 20.8%), few explained the risk of bias in interpreting results (7/24, 29.2%), none provided a list of excluded studies (0/24), and none extracted funding source data (0/24). The absence of protocol registration and transparency in study selection raises concerns about selective reporting.

According to the GRADE system, although acupuncture appeared more effective than controls for PSD, none of the 60 outcome variables achieved a high-quality evidence rating, and only 8 were rated as moderate quality. Most underlying RCTs were small, single-center studies, with risk of bias, mostly from inadequate blinding, being the most prominent factor for downgrading.

Regarding PRISMA 2020 criteria, the reporting quality was generally suboptimal: only 1 review [22] had major defects, 21 had certain deficiencies, and 2 [16,34] met standards for completeness. Notably, most reviews neglected protocol registration, subgroup or sensitivity analyses for heterogeneity, or evidence grading for outcomes. Several lacked key visualizations such as flow or bias charts, and some failed to report funding or conflicts of interest, compromising the transparency and objectivity of their results.

Outlook and Recommendations

Almost all of the included SRs exhibited low or very low methodological quality, indicating an urgent need to improve the conduct and reporting of these reviews. Acupuncture shows promise as a complementary approach for improving depressive symptoms after stroke. However, these findings highlight a pressing need for large, well-designed, randomized controlled trials and higher-quality systematic reviews. Assessments typically rely on scales such as HAMD, SDS, and the modified Edinburgh-Scandinavian stroke scale, but these are somewhat subjective and may introduce bias. Future research should emphasize blinding (of both patients and outcome assessors), clearer methodology, robust protocol registration, and transparency regarding funding and conflicts of interest, in order to improve the reliability and applicability of the evidence. In addition, the results showed that there was mainly mild to moderate overlap, and key clinical trials may not have been included. Follow-up studies need to increase the coverage of clinical trials and include multiple languages and countries.

Strengths and Limitations

A principal strength of this review lies in its focus on meta-analyses of RCTs, representing the highest tier of evidence, with rigorous eligibility criteria to minimize confounding. Nonetheless, several limitations should be acknowledged: (1) interventions were heterogeneous across studies, precluding quantitative synthesis; (2) the included systematic reviews generally exhibited low methodological, reporting, and evidence quality; (3) the lack of differentiation between different subtypes of PSD and TCM syndromes complicates the interpretation of the evidence for acupuncture in treating PSD; and (4) the subjective nature of AMSTAR 2 and PRISMA assessments could introduce rater bias.

Conclusions

Overall, the preponderance of current systematic reviews suggests a positive effect of acupuncture in the management of PSD. However, the evidence base is undermined by pervasive deficiencies in methodological rigor and quality of reporting. To advance the field and inform clinical decision-making, further high-quality multicenter RCTs and rigorously conducted systematic reviews are imperative.