Acupuncture as an Adjunctive Therapy for Depression: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

Summary

This study presents a systematic review and meta-analysis of randomized controlled trials evaluating the efficacy and safety of acupuncture for major depressive disorder as a standalone and adjunctive therapy. It aims to inform evidence-based clinical decision-making and support the integration of acupuncture into comprehensive depression management strategies.

Abstract

Major depressive disorder is a leading cause of global disability. Limitations of pharmacological treatments, including delayed onset, incomplete response, and adverse effects, have increased interest in integrative approaches. Although acupuncture is widely used as a complementary therapy for depression, its efficacy and safety remain uncertain. This systematic review and meta-analysis evaluated the efficacy and safety of acupuncture, alone or as an adjunct to selective serotonin reuptake inhibitors (SSRIs), in reducing depressive symptoms. Eight English and Chinese databases were searched from inception to June 30, 2025, for randomized controlled trials (RCTs) evaluating manual or electroacupuncture for primary depression. Comparators included sham acupuncture, waitlist, psychotherapy, or SSRIs. Two reviewers independently conducted study selection, data extraction, and risk-of-bias assessment using Cochrane ROB 2.0. Pooled mean differences (MDs) with 95% confidence intervals (CIs) were calculated using random-effects models, and the certainty of evidence was assessed using GRADE. Seventy-six RCTs involving 5,864 participants were included. Acupuncture was associated with greater reductions in depressive symptom severity compared with non-active controls, although heterogeneity was substantial. As an adjunct to SSRIs, acupuncture demonstrated additional short-term improvement at four weeks (MD −3.02; 95% CI −3.49 to −2.56; moderate-certainty evidence). Electroacupuncture showed potential for earlier symptomatic improvement, although the certainty of evidence was limited. Adjunctive acupuncture was also associated with fewer treatment-related adverse events and modest improvements in quality of life. The overall certainty of evidence ranged from very low to moderate; therefore, these findings should be interpreted cautiously, and the true effects may differ substantially from the observed estimates. Acupuncture, particularly as an adjunct to SSRI therapy, may offer short-term benefits in depressive symptoms and tolerability but should not be considered a replacement for standard treatment. Further high-quality, rigorously designed trials with standardized protocols and prolonged follow-up are required to clarify its role in depression care.

Introduction

Depression is a complex and disabling mental health disorder characterized by persistent low mood, anhedonia, cognitive impairment, and substantial psychosocial dysfunction¹. Despite advances in pharmacological and psychological interventions, treatment outcomes remain suboptimal, with high rates of partial response, relapse, and chronicity^2,3. It represents a major global public health challenge, affecting approximately 264 million individuals worldwide, as reported by the Global Burden of Disease Study, and ranking among the leading contributors to years lived with disability and non-fatal disease burden⁴. Beyond its clinical manifestations, depression imposes a substantial socioeconomic burden through increased healthcare utilization, reduced workplace productivity, and elevated risk of comorbid medical conditions⁵.

Pharmacotherapy, particularly selective serotonin reuptake inhibitors (SSRIs), remains the most commonly prescribed first-line treatment for depression. However, SSRIs have several clinically important limitations. Therapeutic onset is often delayed, requiring several weeks before meaningful symptom improvement occurs, and approximately one-third of patients do not achieve full remission. In addition, treatment adherence may be compromised by adverse effects, including gastrointestinal disturbances, dizziness, weight gain, sleep disturbances, and sexual dysfunction. These limitations highlight the need for additional safe and effective treatment strategies that can complement conventional antidepressant therapy and improve clinical outcomes^6,7.

Acupuncture has increasingly attracted attention as a potential integrative intervention for depressive disorders. Rooted in Traditional Chinese Medicine, acupuncture is traditionally conceptualized as regulating the flow of qi through meridian pathways. From a contemporary biomedical perspective, accumulating evidence suggests that acupuncture may influence neurochemical, neuroendocrine, inflammatory, and autonomic pathways relevant to mood regulation. Experimental and clinical studies indicate that acupuncture may modulate key neurotransmitters, including serotonin, norepinephrine, dopamine, and gamma-aminobutyric acid. Additional proposed mechanisms include regulation of the hypothalamic–pituitary–adrenal axis, attenuation of neuroinflammatory processes, enhancement of neuroplasticity, and modulation of brain regions involved in emotional processing, such as the prefrontal cortex, amygdala, and hippocampus^8,9.

Over the past two decades, numerous randomized controlled trials (RCTs) have evaluated acupuncture for depression, either as a standalone intervention or as an adjunct to antidepressant therapy. Many of these studies—conducted predominantly in East Asia, with a growing number in Western settings—have reported improvements in depressive symptom severity, quality of life, and treatment tolerability¹⁰. However, findings across individual trials remain inconsistent, reflecting substantial heterogeneity in study design, sample size, acupuncture techniques, treatment duration, comparator interventions, and outcome measures¹¹. Methodological limitations, including inadequate reporting of randomization and blinding procedures, further complicate the interpretation of the available evidence¹².

Several systematic reviews and meta-analyses have attempted to synthesize evidence on acupuncture for depression; however, many have been limited by restricted database coverage, narrow inclusion criteria, or insufficient differentiation between acupuncture modalities and comparator treatments^{12,13,14,15,16}. In particular, earlier reviews often pooled heterogeneous pharmacological comparators, which may obscure the specific integrative role of acupuncture when used alongside SSRIs^13,16. Additionally, few prior analyses have systematically evaluated time-dependent treatment effects or assessed evidence certainty using contemporary frameworks such as Grading of Recommendations Assessment, Development and Evaluation (GRADE)^12,15.

To address these limitations, the present systematic review and meta-analysis aimed to comprehensively evaluate the efficacy and safety of manual and electro-acupuncture for primary depression based on RCTs. Specifically, this study sought to: (1) assess the effectiveness of acupuncture compared with sham acupuncture, waitlist controls, psychotherapy, and SSRIs; (2) determine whether acupuncture provides additional benefit when used as an adjunct to SSRI therapy; (3) evaluate the safety and tolerability of acupuncture relative to pharmacotherapy; and (4) assess the certainty of evidence using the standardized Cochrane Risk of Bias Tool version 2.0 (ROB 2.0) risk-of-bias assessment tool and the GRADE framework. By synthesizing data from 76 RCTs involving 5,864 participants, this review provides an updated and integrative assessment of the potential role of acupuncture in depression management and highlights key priorities for future research^{16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91}.

Protocol

Ethics approval was not required for this study because it is a systematic review and meta-analysis based exclusively on previously published studies and did not involve direct participation of human subjects, collection of new data, or access to identifiable patient information.

Study design
This systematic review and meta-analysis was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines (Supplementary File 1). The review protocol was prospectively registered in the PROSPERO database (registration number: CRD42024526811).

Literature search
A comprehensive literature search was conducted in PubMed, MEDLINE, Embase, the Cochrane Central Register of Controlled Trials (CENTRAL), China National Knowledge Infrastructure (CNKI), Wanfang Database, China Biomedical Literature Database (CBM), and VIP Chinese Journal Full-Text Database (VIP). All databases were searched from inception to June 30, 2025. The detailed search strategy is provided in Supplementary File 2.

Inclusion criteria
Studies were included if they met the following criteria:

1. Population: Participants diagnosed with primary depression according to ICD-10, CCMD-3, DSM-IV, or DSM-5 criteria. Studies including mixed psychiatric populations were eligible only when outcome data for participants with primary depression could be extracted separately. Studies in which depression-specific data could not be isolated were excluded.
2. Study design: RCTs. Only studies explicitly described as randomized or employing recognized random allocation methods (e.g., random number tables or computer-generated randomization sequences) were included. Quasi-randomized studies employing allocation methods such as alternation, date of birth, hospital record number, or other non-random procedures were excluded.
3. Interventions and comparators: The intervention group received acupuncture alone or acupuncture combined with SSRIs. Control groups included sham acupuncture, waitlist, psychotherapy, or SSRIs. Manual acupuncture was defined as needle insertion with manual stimulation only. Electro-acupuncture was defined as acupuncture delivered with electrical stimulation through inserted needles. Sham acupuncture included non-penetrating sham devices, superficial needling, or needling at non-acupuncture points. Psychotherapy included structured psychological interventions such as cognitive behavioral therapy or counseling. SSRI treatment included any selective serotonin reuptake inhibitor administered according to standard clinical practice.
4. Outcomes: Primary outcomes included Hamilton Depression Rating Scale (HAMD), Self-Rating Depression Scale (SDS), and WHOQOL-BREF scores. Adverse effects were assessed using the Side Effect Rating Scale for Antidepressants (SERS). All reported assessment time points were extracted. For meta-analysis, outcomes were grouped into predefined intervals (0–3 weeks, 4 weeks, 6 weeks, and 8–12 weeks) to facilitate comparison across studies.

Exclusion criteria
Studies were excluded if they met any of the following criteria:

1. Incomplete or unavailable data for quantitative analysis: Studies with incomplete or unavailable data required for quantitative synthesis were excluded. Studies with partially missing outcome data were included when sufficient quantitative information was available to calculate effect estimates. When the necessary data could not be obtained from the published report, the study was excluded from the meta-analysis.
2. Duplicate publications.
3. Clearly inappropriate study design or major methodological flaws: Studies with clearly inappropriate study designs or major methodological flaws were excluded. Major methodological flaws included non-random allocation, absence of a control group, serious baseline imbalance between study groups, substantial deficiencies in outcome reporting, inability to verify primary outcome data, or inadequate information required for quantitative analysis.
4. Missing or incomplete primary outcome data.

Study selection
All retrieved records were imported into EndNote X9 for deduplication and screening. Two reviewers (JG and NW) independently screened titles and abstracts, followed by full-text evaluation according to the predefined eligibility criteria. Discrepancies were resolved through discussion, with arbitration by additional reviewer (PL) when necessary. Trials with unclear or non-random allocation methods were excluded to ensure methodological rigor. The literature search and study selection process are shown in Figure 1.

Figure 1. Study selection process and identification of eligible studies. PRISMA 2020 flow diagram illustrating the study selection process for the systematic review and meta-analysis. Records were identified through searches of eight electronic databases, screened after duplicate removal, assessed for eligibility through full-text review, and included according to predefined inclusion and exclusion criteria. A total of 76 randomized controlled trials involving 5,864 participants were included in the qualitative synthesis and quantitative meta-analysis. Abbreviations: CBM, China Biomedical Literature Database; CNKI, China National Knowledge Infrastructure; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses. Please click here to view a larger version of this figure.

Data extraction
Two reviewers (JG and NW) independently extracted data using a standardized data extraction form (Supplementary File 3). Extracted information included study characteristics (first author, publication year, sample size, and randomization method), participant characteristics (age and diagnostic criteria), intervention details, outcome measures, and time points of outcome assessment. The characteristics of the included studies are summarized in Supplementary Table 1. Any discrepancies were resolved through discussion, with arbitration by a third reviewer (PL) when necessary.

Assessment of risk of bias in included studies
The risk of bias of included RCTs was assessed using the Cochrane Risk of Bias Tool version 2.0 (ROB 2.0). This tool evaluates bias across domains including the randomization process, deviations from intended interventions, missing outcome data, measurement of outcomes, and selection of reported results. Each study was rated as having low risk, some concerns, or high risk of bias. Two reviewers (JG and NW) independently performed the assessments, and disagreements were resolved through discussion or consultation with a third reviewer (PL).

Statistical analysis
Statistical analyses were performed using RevMan version 5.4. For continuous outcomes, mean differences (MDs) with corresponding 95% confidence intervals (CIs) were calculated.

Statistical heterogeneity was assessed using the Chi-square (χ²) test and the I² statistic. Heterogeneity was interpreted as low (I² < 50%), moderate (50%–75%), or substantial (>75%). Random-effects meta-analyses were performed using the DerSimonian–Laird random-effects model.

When substantial heterogeneity (I² > 50%) was identified, subgroup analyses were conducted to explore potential sources of heterogeneity, including type of depression assessment scale (HAMD-17, HAMD-24, HAMD-NS), type of acupuncture intervention (manual versus electro-acupuncture), and treatment duration and follow-up time points. Predefined treatment-duration categories were 0–3 weeks, 4 weeks, 6 weeks, and 8–12 weeks. Follow-up analyses were performed according to the nearest reported assessment interval. Summary results of the primary, subgroup, and sensitivity analyses are presented in Supplementary Table 2.

Sensitivity analyses were performed by applying alternative statistical models (fixed-effects versus random-effects) and effect measures (mean difference versus standardized mean difference) to evaluate the robustness of pooled estimates.

If heterogeneity could not be adequately explained, results were interpreted cautiously, and the certainty of evidence was downgraded accordingly. Publication bias was assessed using funnel plots when at least 10 studies were available.

Certainty of evidence
The certainty of evidence for each outcome was assessed using the GRADE framework. Evidence from randomized controlled trials was initially rated as high certainty and subsequently downgraded based on five domains: risk of bias, inconsistency, indirectness, imprecision, and publication bias. The overall certainty of evidence was classified as high, moderate, low, or very low. The certainty-of-evidence assessments for all major outcomes are summarized in Table 1.

Table 1: Summary of Findings and GRADE certainty assessment for acupuncture and electroacupuncture interventions in patients with primary depression. Pooled effect estimates from randomized controlled trials comparing acupuncture, electroacupuncture, and their combinations with selective serotonin reuptake inhibitors (SSRIs) against psychotherapy, waitlist control, sham acupuncture, or SSRI treatment alone. Outcomes include Hamilton Depression Rating Scale (HAMD) scores at different treatment durations, Self-Rating Depression Scale (SDS) scores, and adverse effects assessed using the Side Effect Rating Scale for Antidepressants (SERS). For each outcome, the table presents the assessment of risk of bias, inconsistency, indirectness, imprecision, and publication bias according to the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework, together with the number of participants and studies, pooled effect estimates expressed as mean differences (MDs) with 95% confidence intervals (CIs), and the overall certainty of evidence. Certainty ratings are categorized as high (⊕⊕⊕⊕), moderate (⊕⊕⊕◯), low (⊕⊕◯◯), or very low (⊕◯◯◯). Please click here to download this Table.

The interpretation of results was explicitly guided by the GRADE ratings, with findings supported by low or very low certainty evidence interpreted cautiously and described using non-causal, uncertainty-aware language.

Results

Results of search
A total of 20,229 records were identified through electronic database searches. After removal of 9,567 duplicate records using EndNote, 10,662 records remained for screening. No duplicate studies were identified after full-text screening. Following title and abstract screening, 10,572 records were excluded based on the predefined inclusion and exclusion criteria. The full texts of 90 potentially eligible articles were assessed for eligibility, of which 14 were excluded according to the predefined exclusion criteria. Ultimately, 76 RCTs were included in the meta-analysis. The detailed study selection process is illustrated in Figure 1.

Characteristics of the included studies
The 76 included RCTs involved 5,864 participants diagnosed with primary depression. Baseline characteristics indicated that most participants had mild to moderate depression, with baseline HAMD scores generally ranging from 20 to 35. Regarding interventions, 33 trials evaluated manual acupuncture alone, 10 trials evaluated electro-acupuncture alone, 23 trials assessed manual acupuncture combined with SSRIs, and 10 trials examined electro-acupuncture combined with SSRIs. The most frequently used acupuncture points were GV20 (Baihui) (54 trials), LR3 (Taichong) (39 trials), SP6 (Sanyinjiao) (38 trials), PC6 (Neiguan) (37 trials), and GV24 (Yintang) (35 trials). Among control interventions, sham acupuncture was used in 5 trials, while the majority of studies employed SSRIs as the comparator. In sham-controlled trials, four studies used non-penetrating sham acupuncture, whereas one trial used stimulation at non-acupuncture points. HAMD scores were reported in all included trials, using different versions of the scale (HAMD-24, HAMD-17, and HAMD-NS). Fourteen trials reported SDS outcomes, three trials assessed quality of life using the WHOQOL-BREF scale, and safety outcomes were reported in approximately 15 trials. Detailed characteristics of the included studies are provided in Supplementary Table 1.

Risk of bias in included studies
Risk-of-bias assessments for the included RCTs are summarized in Supplementary Figures S1 and S2. Overall, only a minority of trials were judged to be at low risk of bias, while the majority were rated as having some concerns, primarily related to inadequate reporting of allocation concealment, incomplete descriptions of randomization procedures, and potential selective reporting. Concerns regarding allocation concealment were common, with many trials failing to clearly describe the methods used to ensure adequate concealment. Among studies that reported allocation procedures, methods such as sealed opaque envelopes, central randomization, and computer-generated randomization were employed; however, the quality and transparency of reporting varied considerably across studies. Blinding was generally limited due to the nature of acupuncture interventions. While participant blinding was not feasible in many trials, lack of blinding of outcome assessors may have introduced detection bias, particularly for subjective outcome measures such as HAMD scores. In addition, selective reporting bias was difficult to assess in several trials due to the absence of published protocols or prospective trial registration. Overall, these methodological limitations contribute to uncertainty in the pooled effect estimates and should be carefully considered when interpreting the findings.

Effects of interventions
Acupuncture versus psychotherapy, waitlist, or sham acupuncture.
Outcome: HAMD scores after 0–12 weeks of treatment:
Three randomized controlled trials reported HAMD outcomes at the end of treatment. Acupuncture was associated with greater reductions in depressive symptoms, as reflected by reductions in both HAMD-24 scores (MD −7.62, 95% CI −14.00 to −1.24; I² = 86%) and HAMD-17 scores (MD −3.80, 95% CI −6.83 to −0.77), compared with psychotherapy, waitlist, or sham acupuncture. The pooled subgroup analysis indicated that acupuncture reduced overall HAMD scores relative to control interventions (MD −5.80, 95% CI −9.06 to −2.55; I² = 88%). However, substantial heterogeneity was observed. Forest plots are shown in Figure 2.

Figure 2. Effect of acupuncture versus control interventions on depressive symptom severity. Forest plot comparing acupuncture with control interventions (psychotherapy, waitlist, or sham acupuncture) for Hamilton Depression Rating Scale (HAMD) outcomes at the end of treatment. Subgroup analyses are presented according to HAMD scale version (HAMD-24 and HAMD-17), together with the pooled overall effect estimate. Squares represent individual study or subgroup effect estimates, with horizontal lines indicating 95% confidence intervals (CIs). Diamonds represent pooled effect estimates. Negative mean differences favor acupuncture. Statistical heterogeneity was assessed using the I² statistic. Abbreviations: CI, confidence interval; HAMD, Hamilton Depression Rating Scale; HAMD-17, 17-item Hamilton Depression Rating Scale; HAMD-24, 24-item Hamilton Depression Rating Scale; IV, inverse variance. Please click here to view a larger version of this figure.

Outcome: SDS scores after 0–3 months of treatment
Five trials involving 208 participants reported SDS outcomes. At the end of treatment, acupuncture was associated with greater reductions in SDS scores compared with control interventions (MD −9.46, 95% CI −10.49 to −8.43; I² = 7%). Forest plots are provided in Supplementary File 4 (Analysis 1.2).

Acupuncture versus SSRI.
Outcome: HAMD scores after 0–3 weeks of treatment
Sixteen randomized controlled trials reported HAMD outcomes after 0–3 weeks of treatment. Acupuncture was associated with lower HAMD-17 scores (MD −1.49, 95% CI −2.35 to −0.64; I² = 43%). No significant differences were observed for HAMD-24 scores (MD −1.73, 95% CI −3.54 to 0.09; I² = 69%) or HAMD-NS scores (MD −1.62, 95% CI −3.61 to 0.37; I² = 75%). The pooled subgroup analysis demonstrated a modest overall reduction in HAMD scores favoring acupuncture (MD −1.40, 95% CI −2.16 to −0.63; I² = 63%), although moderate heterogeneity was observed. Forest plots are shown in Supplementary File 4 (Analysis 2.1).

Outcome: HAMD scores after 4 weeks of treatment
Twenty randomized controlled trials reported outcomes at 4 weeks. Acupuncture was associated with lower HAMD-24 scores (MD −4.11, 95% CI −6.71 to −1.50; I² = 79%), HAMD-17 scores (MD −1.89, 95% CI −3.10 to −0.67; I² = 23%), and HAMD-NS scores (MD −1.77, 95% CI −3.12 to −0.42; I² = 74%). The pooled estimate across subgroups indicated a small but statistically significant benefit of acupuncture compared with SSRIs (MD −2.48, 95% CI −3.56 to −1.40; I² = 77%), with substantial heterogeneity. Forest plots are provided in Supplementary File 4 (Analysis 2.2).

Outcome: HAMD scores after 6 weeks of treatment
Eleven trials involving 900 participants reported outcomes at 6 weeks. No significant differences were observed in HAMD-24 scores (MD −0.32, 95% CI −1.70 to 1.05; I² = 40%) or HAMD-NS scores (MD −0.21, 95% CI −1.36 to 0.94; I² = 0%). However, acupuncture was associated with lower HAMD-17 scores (MD −1.71, 95% CI −2.58 to −0.84; I² = 14%). The pooled analysis demonstrated a small overall benefit favoring acupuncture (MD −0.88, 95% CI −1.71 to −0.55; I² = 54%), indicating moderate heterogeneity. Forest plots are shown in Supplementary File 4 (Analysis 2.3).

Outcome: HAMD scores after 8–12 weeks of treatment
Five randomized controlled trials reported outcomes at 8–12 weeks. No significant differences were observed in HAMD-24 or HAMD-17 scores. A significant reduction was observed in HAMD-NS scores (MD −3.41, 95% CI −6.70 to −0.12; I² = 80%). The pooled analysis indicated a modest benefit favoring acupuncture (MD −2.33, 95% CI −4.26 to −0.40; I² = 76%), with substantial heterogeneity. Forest plots are shown in Supplementary File 4 (Analysis 2.4).

Outcome: Adverse effects
Four trials involving 245 participants assessed adverse effects using SERS. Acupuncture was associated with fewer adverse effects compared with SSRIs (MD −3.35, 95% CI −5.26 to −1.45; I² = 92%). However, heterogeneity was high, and the findings should be interpreted cautiously. Forest plots are shown in Supplementary File 4 (Analysis 2.5).

Acupuncture plus SSRI versus SSRI.
Outcome: HAMD scores after 0–3 weeks of treatment
Twenty-one randomized controlled trials reported HAMD scores as the primary outcome. Subgroup analyses showed that acupuncture combined with SSRIs was associated with greater reductions in HAMD-24 scores (MD −2.12, 95% CI −2.61 to −1.63; I² = 0%) and HAMD-17 scores (MD −3.04, 95% CI −3.24 to −2.83; I² = 22%) compared with SSRI treatment alone. The pooled subgroup estimate indicated a modest overall benefit of adjunctive acupuncture in reducing HAMD scores (MD −2.90, 95% CI −3.09 to −2.71; I² = 38%). Forest plots are presented in Supplementary File 4 (Analysis 3.1).

Outcome: HAMD scores after 4 weeks of treatment
Seventeen randomized controlled trials reported HAMD outcomes after 4 weeks of treatment. Subgroup analyses demonstrated that acupuncture plus SSRI was associated with lower HAMD-24 scores (MD −2.37, 95% CI −3.60 to −1.41; I² = 72%), HAMD-17 scores (MD −3.30, 95% CI −4.06 to −2.55; I² = 50%), and HAMD-NS scores (MD −2.57, 95% CI −3.71 to −1.43; I² = 0%) compared with SSRI treatment alone. The pooled estimate across subgroups confirmed a statistically significant reduction in HAMD scores favoring adjunctive acupuncture (MD −2.92, 95% CI −3.42 to −2.41; I² = 52%), although moderate heterogeneity was observed. Forest plots are shown in Figure 3.

Figure 3. Effect of adjunctive acupuncture plus SSRI therapy versus SSRI therapy alone on depressive symptom severity at 4 weeks. Forest plot comparing acupuncture combined with selective serotonin reuptake inhibitor (SSRI) therapy versus SSRI therapy alone for Hamilton Depression Rating Scale (HAMD) outcomes after 4 weeks of treatment. Subgroup analyses are presented according to HAMD scale version (HAMD-24, HAMD-17, and HAMD-NS), together with the pooled overall effect estimate. Squares represent individual study effect estimates, with horizontal lines indicating 95% confidence intervals (CIs). Diamonds represent pooled subgroup and overall effect estimates. Negative mean differences favor acupuncture plus SSRI therapy. Statistical heterogeneity was assessed using the I² statistic. Abbreviations: CI, confidence interval; HAMD, Hamilton Depression Rating Scale; HAMD-17, 17-item Hamilton Depression Rating Scale; HAMD-24, 24-item Hamilton Depression Rating Scale; HAMD-NS, Hamilton Depression Rating Scale–Neurosis Scale; IV, inverse variance; SSRI, selective serotonin reuptake inhibitor. Please click here to view a larger version of this figure.

Outcome: HAMD scores after 6 weeks of treatment
Twenty-one trials involving 1,545 participants reported HAMD outcomes after 6 weeks of treatment. Subgroup analyses showed reductions in HAMD-24 scores (MD −2.36, 95% CI −3.05 to −1.67; I² = 54%), HAMD-17 scores (MD −2.72, 95% CI −3.37 to −2.06; I² = 78%), and HAMD-NS scores (MD −3.37, 95% CI −5.05 to −1.24; one trial, 60 participants) compared with SSRI treatment. The pooled analysis demonstrated a significant overall benefit of acupuncture plus SSRI (MD −2.58, 95% CI −3.04 to −2.13; I² = 70%), with substantial heterogeneity. Forest plots are provided in Supplementary File 4 (Analysis 3.3).

Outcome: HAMD scores after 8–12 weeks of treatment
Nine randomized controlled trials reported HAMD outcomes after 8–12 weeks of treatment. Subgroup analyses showed reductions in HAMD-24 scores (MD −2.98, 95% CI −5.23 to −0.74; I² = 91%), HAMD-17 scores (MD −1.95, 95% CI −2.79 to −1.12; I² = 0%), and HAMD-NS scores (MD −4.49, 95% CI −6.41 to −2.56; I² = 79%) compared with SSRI treatment alone. The pooled estimate indicated a modest overall benefit of adjunctive acupuncture (MD −2.97, 95% CI −4.04 to −1.90; I² = 84%), with substantial heterogeneity. Forest plots are shown in Supplementary File 4 (Analysis 3.4).

Outcome: SDS scores
Six studies reported SDS outcomes. The pooled analysis showed no statistically significant difference between acupuncture plus SSRI and SSRI alone (MD −4.89, 95% CI −10.13 to 0.34; I² = 95%), indicating considerable heterogeneity. Forest plots are presented in Supplementary File 4 (Analysis 3.5).

Outcome: Adverse effects
Eleven studies involving 1,074 participants reported SERS outcomes. Overall, acupuncture combined with SSRI was associated with greater reductions in adverse effect scores compared with SSRI treatment alone (MD −2.87, 95% CI −4.25 to −1.50; I² = 96%). However, heterogeneity was substantial, and the findings should be interpreted cautiously. Forest plots are shown in Supplementary File 4 (Analysis 3.6).

Outcome: WHOQOL-BREF
Three trials assessed quality of life using the WHOQOL-BREF scale. Acupuncture plus SSRI was associated with improvements in overall quality of life (MD 0.34, 95% CI 0.16 to 0.53; I² = 0%), general health (MD 0.45, 95% CI 0.18 to 0.71; I² = 26%), and psychological health (MD 1.39, 95% CI 0.46 to 2.32; I² = 67%), with no significant differences observed in physical health, social relationships, or environmental domains. Forest plots are provided in Supplementary File 4 (Analysis 3.7).

Electro-acupuncture versus SSRI.
Outcome: HAMD scores after 0–3 weeks of treatment
Five randomized controlled trials reported HAMD outcomes after 0–3 weeks of treatment. Electro-acupuncture was associated with lower HAMD-17 scores (MD −1.75, 95% CI −3.46 to −0.04; one trial, 129 participants). No significant differences were observed for HAMD-24 scores (MD −4.08, 95% CI −9.74 to 1.58; I² = 92%) or HAMD-NS scores (MD 1.00, 95% CI −0.07 to 2.07; one trial, 80 participants). The pooled analysis showed no significant overall difference in HAMD scores between electro-acupuncture and SSRIs (MD −2.53, 95% CI −6.07 to 1.01; I² = 94%), with substantial heterogeneity. Forest plots are presented in Supplementary File 4 (Analysis 4.1).

Outcome: HAMD scores after 4 weeks of treatment
Eight randomized controlled trials reported outcomes at 4 weeks. Subgroup analyses showed no significant differences in HAMD-24 scores (MD −1.49, 95% CI −3.70 to 0.71; I² = 58%), HAMD-17 scores (MD −1.21, 95% CI −3.67 to 1.25; I² = 73%), or HAMD-NS scores (MD −1.08, 95% CI −9.51 to 7.35; I² = 99%). The pooled estimate confirmed no significant overall effect (MD −1.15, 95% CI −3.93 to 1.62; I² = 95%), indicating substantial heterogeneity. Forest plots are shown in Supplementary File 4 (Analysis 4.2).

Outcome: HAMD scores after 6 weeks of treatment
Seven trials involving 570 participants reported HAMD outcomes at 6 weeks. No significant differences were observed for HAMD-24 scores (MD −1.00, 95% CI −2.21 to 0.20; I² = 44%) or HAMD-17 scores (MD 1.23, 95% CI −0.45 to 2.19; one trial, 130 participants). One trial reported a significant advantage for SSRIs in HAMD-NS scores (MD 4.10, 95% CI 3.33 to 4.87; one trial, 80 participants). The pooled analysis showed no significant difference overall (MD 0.31, 95% CI −2.00 to 2.62; I² = 93%), with substantial heterogeneity. Forest plots are presented in Supplementary File 4 (Analysis 4.3).

Outcome: HAMD scores after 8–12 weeks of treatment
Two randomized controlled trials reported outcomes after 8–12 weeks of treatment. One trial showed a significant reduction in HAMD-24 scores favoring electro-acupuncture (MD −2.24, 95% CI −4.02 to −0.46), while another trial reported a significant reduction in HAMD-NS scores (MD −5.12, 95% CI −6.24 to −4.00). Due to the limited number of studies, a pooled analysis was not performed.

Outcome: SDS scores
Three trials reported SDS outcomes. The pooled analysis showed no significant difference between electro-acupuncture and SSRI treatment (MD 1.41, 95% CI −2.63 to 5.45; I² = 57%), indicating moderate heterogeneity. Forest plots are shown in Supplementary File 4 (Analysis 4.5).

Electro-acupuncture plus SSRI versus SSRI.
Outcome: HAMD scores after 0–3 weeks of treatment
Three randomized controlled trials reported outcomes at 0–3 weeks. Electro-acupuncture plus SSRI was associated with lower HAMD-24 scores (MD −5.60, 95% CI −6.43 to −4.77) and HAMD-17 scores (MD −5.18, 95% CI −6.67 to −3.69; I² = 0%) compared with SSRI treatment alone. The pooled analysis confirmed an overall reduction in HAMD scores favoring electro-acupuncture plus SSRI (MD −5.44, 95% CI −6.27 to −4.61; I² = 0%). Forest plots are shown in Figure 4.

Figure 4. Effect of adjunctive electro-acupuncture plus SSRI therapy versus SSRI therapy alone on depressive symptom severity at 0–3 weeks. Forest plot comparing electro-acupuncture combined with selective serotonin reuptake inhibitor (SSRI) therapy versus SSRI therapy alone for Hamilton Depression Rating Scale (HAMD) outcomes during the early treatment period (0–3 weeks). Subgroup analyses are presented according to HAMD scale version (HAMD-24 and HAMD-17), together with the pooled overall effect estimate. Squares represent individual study effect estimates, with horizontal lines indicating 95% confidence intervals (CIs). Diamonds represent pooled subgroup and overall effect estimates. Negative mean differences favor electro-acupuncture plus SSRI therapy. Statistical heterogeneity was assessed using the I² statistic. Abbreviations: CI, confidence interval; HAMD, Hamilton Depression Rating Scale; HAMD-17, 17-item Hamilton Depression Rating Scale; HAMD-24, 24-item Hamilton Depression Rating Scale; IV, inverse variance; SSRI, selective serotonin reuptake inhibitor. Please click here to view a larger version of this figure.

Outcome: HAMD scores after 4 weeks of treatment
Three randomized controlled trials reported outcomes at 4 weeks. Subgroup analyses showed reductions in HAMD-24 scores (MD −3.60, 95% CI −4.35 to −2.85; one trial, 101 participants) and HAMD-17 scores (MD −5.39, 95% CI −6.86 to −3.92; I² = 0%). The pooled analysis confirmed a benefit of adjunctive electro-acupuncture (MD −4.19, 95% CI −5.44 to −2.95; I² = 0%), with low heterogeneity. Forest plots are shown in Figure 5.

Figure 5. Effect of adjunctive electro-acupuncture plus SSRI therapy versus SSRI therapy alone on depressive symptom severity at 4 weeks. Forest plot comparing electro-acupuncture combined with selective serotonin reuptake inhibitor (SSRI) therapy versus SSRI therapy alone for Hamilton Depression Rating Scale (HAMD) outcomes after 4 weeks of treatment. Subgroup analyses are presented according to HAMD scale version (HAMD-24 and HAMD-17), together with the pooled overall effect estimate. Squares represent individual study effect estimates, with horizontal lines indicating 95% confidence intervals (CIs). Diamonds represent pooled subgroup and overall effect estimates. Negative mean differences favor electro-acupuncture plus SSRI therapy. Statistical heterogeneity was assessed using the I² statistic. Abbreviations: CI, confidence interval; HAMD, Hamilton Depression Rating Scale; HAMD-17, 17-item Hamilton Depression Rating Scale; HAMD-24, 24-item Hamilton Depression Rating Scale; IV, inverse variance; SSRI, selective serotonin reuptake inhibitor. Please click here to view a larger version of this figure.

Outcome: HAMD scores after 6 weeks of treatment
Eight trials involving 530 participants reported outcomes at 6 weeks. Electro-acupuncture plus SSRI was associated with lower HAMD-17 scores (MD −2.69, 95% CI −4.04 to −1.35; I² = 50%), while no significant difference was observed for HAMD-24 scores (MD −0.15, 95% CI −0.36 to 0.07; I² = 0%). The pooled analysis demonstrated a modest overall benefit favoring electro-acupuncture plus SSRI (MD −1.86, 95% CI −3.19 to −0.53; I² = 83%), with substantial heterogeneity. Forest plots are provided in Supplementary File 4 (Analysis 5.3).

Outcome: HAMD scores after 8–12 weeks of treatment
Two randomized controlled trials reported outcomes after 8–12 weeks. One trial showed a reduction in HAMD-24 scores favoring electro-acupuncture plus SSRI (MD −2.24, 95% CI −4.00 to −0.48), while another trial reported improvement in HAMD-17 scores (MD −5.24, 95% CI −6.95 to −3.53). Due to the limited number of studies, a pooled analysis was not performed.

Outcome: Adverse effects
Three studies involving 163 participants assessed SERS outcomes. Electro-acupuncture plus SSRI was associated with greater reductions in adverse effect scores compared with SSRI alone (MD −2.06, 95% CI −2.93 to −1.20; I² = 6%), indicating low heterogeneity. Forest plots are shown in Supplementary File 4 (Analysis 5.5).

Outcome: WHOQOL-BREF
Three trials reported WHOQOL-BREF outcomes. Improvements were observed in overall quality of life (MD 0.46, 95% CI 0.23 to 0.65; I² = 0%), general health (MD 0.38, 95% CI 0.16 to 0.59; I² = 0%), and psychological health (MD 1.40, 95% CI 0.94 to 1.87; I² = 0%), with no significant differences observed in physical health, social relationships, or environmental domains. Forest plots are shown in Supplementary File 4 (Analysis 5.6).

Reporting bias.
When at least 10 trials were available, funnel plots were examined to assess potential reporting bias. The results are presented in Supplementary File 5.

Sensitivity analysis.
Sensitivity analyses indicated that results were not robust for SDS outcomes in the comparisons of acupuncture plus SSRI versus SSRI and electro-acupuncture versus SSRI, as well as for HAMD scores at 6 weeks in the electro-acupuncture versus SSRI comparison. For all other comparisons, sensitivity analyses produced consistent results, supporting the overall stability of the primary findings. Summary results of the primary, subgroup, and sensitivity analyses are presented in Supplementary Table 2.

GRADE assessment .
Using the GRADE framework, 7 outcomes were rated as moderate quality, 13 as low quality, and 1 as very low quality. Evidence was downgraded primarily due to risk of bias (12 outcomes), inconsistency (15 outcomes), and potential publication bias (9 outcomes). Overall, the certainty of evidence ranged from very low to moderate, indicating that the true effects may differ from the observed estimates. The detailed Summary of Findings is presented in Table 1.

Across the included randomized controlled trials, acupuncture alone and acupuncture combined with SSRIs were generally associated with reductions in depressive symptom severity compared with control interventions, although the magnitude and consistency of effects varied across comparisons and time points. Electro-acupuncture combined with SSRIs also demonstrated favorable effects in several analyses. However, substantial heterogeneity and limitations in study quality were observed in a number of outcomes, and the certainty of evidence ranged from very low to moderate. These findings should therefore be interpreted in conjunction with the risk-of-bias and GRADE assessments.

Data Availability:
All data generated or analyzed during this study are included in this published article and its supplementary information files. Supporting materials include the PRISMA 2020 checklist (Supplementary File 1), detailed database search strategies (Supplementary File 2), the standardized data extraction form used during study selection and data collection (Supplementary File 3), forest plots of all quantitative meta-analyses (Supplementary File 4), funnel plots for assessment of reporting and publication bias (Supplementary File 5), characteristics of included randomized controlled trials (Supplementary Table 1), pooled meta-analysis and sensitivity analysis results (Supplementary Table 2), characteristics of previously published systematic reviews (Supplementary Table 3), and risk-of-bias assessments (Supplementary Figures S1 and S2). No additional datasets were generated or analyzed beyond those reported in the article and supplementary materials.

Supplementary File 1. PRISMA 2020 checklist for reporting of the systematic review and meta-analysis. This supplementary file provides the completed Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 checklist for the present systematic review and meta-analysis. The checklist outlines the reporting standards followed across all sections of the manuscript, including the Title, Abstract, Introduction, Methods, Results, Discussion, and Other Information. Each checklist item is mapped to its corresponding location within the manuscript to demonstrate adherence to PRISMA 2020 guidelines. This file is provided to enhance transparency, completeness of reporting, and reproducibility of the review process.Please click here to download this file.

Supplementary File 2. Database search strategies used for the systematic review and meta-analysis. This supplementary file provides the complete electronic search strategies used to identify randomized controlled trials evaluating acupuncture for primary depression. The search was conducted across major English-language and Chinese-language databases, including PubMed, Embase, Web of Science, the Cochrane Library, China National Knowledge Infrastructure (CNKI), Wanfang Data, VIP Database, and SinoMed. Search terms combined controlled vocabulary and free-text keywords related to depression, acupuncture, and randomized or controlled clinical trial design, as appropriate for each database interface. Both manual acupuncture and electroacupuncture terms were included. Chinese-language searches used equivalent subject headings and keyword combinations for acupuncture and depressive disorders. The complete search syntax and the number of records retrieved from each database are provided. This file is included to ensure transparency, reproducibility, and completeness of the literature search process.Please click here to download this file.

Supplementary File 3. Standardized data extraction form used for study selection and data collection. This supplementary file presents the standardized data extraction form used during study selection and data collection for the systematic review and meta-analysis. Variables extracted included study characteristics, participant demographics, diagnostic criteria, intervention and comparator details, treatment duration, outcome measures, follow-up information, adverse events, and risk-of-bias assessments. Data extraction was performed independently by two reviewers, and discrepancies were resolved through discussion and consensus. The form is provided to enhance transparency, reproducibility, and consistency of the review process.Please click here to download this file.

Supplementary File 4. Forest plots of meta-analyses evaluating acupuncture interventions for primary depression. This supplementary file presents the forest plots generated for all quantitative meta-analyses included in the systematic review and meta-analysis. Comparisons include acupuncture versus control interventions, acupuncture plus selective serotonin reuptake inhibitors (SSRIs) versus SSRIs alone, electro-acupuncture versus SSRIs, and electro-acupuncture plus SSRIs versus SSRIs alone. Outcomes include Hamilton Depression Rating Scale (HAMD) scores at multiple assessment time points, Self-Rating Depression Scale (SDS) scores, Side Effect Rating Scale for Antidepressants (SERS) scores, and WHOQOL-BREF quality-of-life outcomes. Forest plots display pooled effect estimates with corresponding 95% confidence intervals and measures of statistical heterogeneity. This supplementary file is provided to support transparency and facilitate evaluation of the quantitative synthesis.Please click here to download this file.

Supplementary File 5. Funnel plots for assessment of potential reporting and publication bias. This supplementary file presents funnel plots used to assess potential reporting and publication bias for meta-analyses evaluating acupuncture interventions for primary depression. Funnel plots were generated for outcomes and comparisons in which a sufficient number of randomized controlled trials were available for assessment. Comparisons include acupuncture versus psychotherapy, waitlist, or sham acupuncture; acupuncture versus selective serotonin reuptake inhibitors (SSRIs); acupuncture plus SSRIs versus SSRIs alone; electro-acupuncture versus SSRIs; and electro-acupuncture plus SSRIs versus SSRIs alone. Outcomes include Hamilton Depression Rating Scale (HAMD) scores at multiple assessment time points, Self-Rating Depression Scale (SDS) scores, and adverse-effect outcomes. Funnel plots were visually inspected for asymmetry as a potential indicator of small-study effects, reporting bias, or publication bias. This supplementary file is provided to support evaluation of the robustness and reliability of the pooled meta-analysis findings.Please click here to download this file.

Supplementary Table 1. Characteristics of included randomized controlled trials. Characteristics of the 76 randomized controlled trials included in the systematic review and meta-analysis evaluating acupuncture for primary depression. Information includes diagnostic criteria, sample size, participant age, intervention and comparator groups, outcome measures, and assessment time points. Interventions comprised manual acupuncture (MA), electroacupuncture (EA), acupuncture combined with selective serotonin reuptake inhibitors (SSRIs), and electroacupuncture combined with SSRIs. Comparator groups included sham acupuncture (SA), placebo acupuncture (PA), cognitive behavioral therapy (CBT), waitlist controls, and antidepressant medications. Outcomes included Hamilton Depression Rating Scale (HAMD) scores, Self-Rating Depression Scale (SDS) scores, quality-of-life measures, and adverse-event assessments.Please click here to download this file.

Supplementary Table 2. Summary of primary and sensitivity analyses for meta-analysis outcomes. Summary of pooled effect estimates for primary, subgroup, and sensitivity analyses comparing acupuncture interventions with control interventions in patients with primary depression. Comparisons included acupuncture versus psychotherapy, waitlist, sham acupuncture, or SSRIs; acupuncture plus SSRIs versus SSRIs alone; electroacupuncture versus SSRIs; and electroacupuncture plus SSRIs versus SSRIs alone. Outcomes included HAMD scores at multiple treatment durations, SDS scores, and adverse-effect outcomes. Results are presented as mean differences (MDs) or standardized mean differences (SMDs) with corresponding 95% confidence intervals (CIs). Sensitivity analyses using alternative effect measures and statistical models were performed to assess the robustness of the findings.Please click here to download this file.

Supplementary Table 3. Characteristics and findings of previously published systematic reviews evaluating acupuncture for depression. Summary of previously published systematic reviews and meta-analyses evaluating acupuncture for depression. Information includes publication date, study design, number and location of included studies, quality assessment methods, intervention and comparator characteristics, reported outcomes, and principal conclusions. Acupuncture interventions included manual acupuncture (MA), electroacupuncture (EA), laser acupuncture (LA), and acupuncture combined with antidepressant therapy. Comparator interventions included sham acupuncture (SA), placebo acupuncture (PA), psychotherapy, waitlist controls, usual care, and antidepressant medications. This table provides context for the current systematic review by summarizing the existing evidence base and highlighting differences in methodology and reported conclusions across prior reviews.Please click here to download this file.

Supplementary Figure S1. Risk-of-bias assessment of individual randomized controlled trials. Risk-of-bias judgments for each included randomized controlled trial assessed using the Cochrane Risk of Bias Tool version 2.0 (ROB 2.0). Each study was evaluated across five domains: randomization process (D1), deviations from intended interventions (D2), missing outcome data (D3), measurement of the outcome (D4), and selection of the reported result (D5). Overall risk-of-bias judgments are presented for each study. Green symbols indicate low risk of bias, yellow symbols indicate some concerns, and red symbols indicate high risk of bias. Please click here to download this file.

Supplementary Figure S2. Summary of risk-of-bias assessments across included randomized controlled trials. Summary of risk-of-bias judgments across all included randomized controlled trials assessed using the Cochrane Risk of Bias Tool version 2.0 (ROB 2.0). Bars represent the percentage of studies classified as low risk of bias, some concerns, or high risk of bias for each risk-of-bias domain and for the overall risk-of-bias assessment. Domains include the randomization process, deviations from intended interventions, missing outcome data, measurement of outcomes, and selection of reported results. Please click here to download this file.

Discussion

From 10,572 screened records, this systematic review and meta-analysis identified 76 randomized controlled trials involving 5,864 participants with primary depression. Across included studies, depressive symptom severity was primarily assessed using validated scales, including HAMD, SDS, and WHOQOL-BREF. The findings suggest that acupuncture may be associated with reductions in depressive symptom severity, particularly when used as an adjunct to SSRI therapy, and may offer advantages in treatment tolerability. When compared with psychotherapy, waitlist controls, or sham acupuncture, acupuncture may provide greater short-term improvements in depressive symptoms; however, these findings were characterized by substantial heterogeneity and generally low certainty of evidence, warranting cautious interpretation. In contrast, acupuncture used alone showed no consistent superiority over SSRIs across time points, indicating that its role as a standalone treatment remains uncertain. Importantly, when acupuncture was combined with SSRIs, consistent reductions in HAMD scores were observed at multiple time points, with the most robust effects seen at four weeks (MD −3.02; 95% CI −3.49 to −2.56; moderate-certainty evidence). While these findings suggest a potential benefit of adjunctive acupuncture, the presence of heterogeneity and variability in study design should be considered when interpreting these results. Electro-acupuncture appeared to confer earlier symptomatic improvement during the initial treatment phase, particularly within the first two to three weeks. However, this observation is based on limited evidence and was affected by methodological heterogeneity and imprecision, indicating that the true effect remains uncertain^92,93. Across comparisons, acupuncture—especially when combined with SSRIs—may be associated with fewer antidepressant-related adverse effects and modest improvements in quality-of-life measures⁹⁴. These findings should be interpreted in the context of previous systematic reviews evaluating acupuncture for depression, which are summarized in Supplementary Table 3. However, given that the certainty of evidence ranged from very low to moderate across outcomes, these findings should be interpreted cautiously. Interpretations were guided by GRADE principles, and outcomes with low or very low certainty evidence were described using uncertainty-aware, non-causal language.

Despite the use of randomized controlled designs, the overall methodological quality of the included studies was variable. Common limitations included inadequate reporting of random sequence generation, allocation concealment, and prospective trial registration. While blinding of participants is inherently challenging in acupuncture research, insufficient blinding of outcome assessors may have introduced detection bias in some trials. These methodological limitations likely contributed to the substantial heterogeneity observed across several analyses, including those with high I² values, and may have affected the robustness of pooled estimates. Clinical heterogeneity was further driven by variability in acupuncture protocols, including differences in acupoint selection, stimulation techniques, session frequency, and treatment duration, as well as variation in SSRI type and dosage. Outcome heterogeneity was also evident due to the use of multiple depression rating scales and inconsistent assessment time points. Although subgroup and sensitivity analyses were conducted to explore potential sources of variability, residual heterogeneity persisted, limiting confidence in the consistency and generalizability of treatment effects. Only a small number of included trials employed sham acupuncture as a control, and these interventions varied in design, including non-penetrating techniques and stimulation at non-acupoint locations. Importantly, sham acupuncture is not a physiologically inert placebo, as even minimal needle contact or superficial stimulation may activate sensory and neurobiological pathways. In addition to these physiological responses, non-specific factors such as patient expectations, placebo effects, and the therapeutic interaction between patient and practitioner may contribute substantially to clinical outcomes. Consequently, the observed differences between acupuncture and sham interventions may underestimate or obscure the specific effects of acupuncture, complicating interpretation of efficacy^95,96. Given the limited number and methodological variability of sham-controlled trials included in this review, the distinction between specific and non-specific effects remains uncertain. Future studies employing rigorously standardized sham controls, improved blinding strategies, consistent outcome assessment methods, and adequately powered sample sizes are needed to better isolate the specific therapeutic effects of acupuncture and reduce uncertainty in effect estimates⁹⁷. Overall, acupuncture—particularly as an adjunct to antidepressant therapy—was associated with improvements in depressive symptoms and treatment tolerability across multiple outcomes. However, given that several outcomes were supported by low or very low certainty evidence, these findings should be interpreted with caution, and the true effect may differ substantially from the observed estimates.

The biological mechanisms underlying the antidepressant effects of acupuncture have not been fully elucidated. Available experimental and clinical evidence suggests that acupuncture may modulate central monoaminergic neurotransmission, including serotonin, norepinephrine, dopamine, and gamma-aminobutyric acid, all of which play key roles in mood regulation. Additional proposed mechanisms include regulation of the hypothalamic–pituitary–adrenal axis, attenuation of neuroinflammatory processes, enhancement of neuroplasticity, and modulation of autonomic nervous system activity⁹⁸. Neuroimaging and preclinical studies further suggest that acupuncture may influence brain regions implicated in emotional processing, such as the hippocampus, prefrontal cortex, and amygdala. Depression has been associated with reduced hippocampal volume and impaired stress regulation, and acupuncture has been shown in experimental models to affect hippocampal signaling pathways, including the JNK signaling cascade. While these findings provide biological plausibility, their direct translation to clinical outcomes remains uncertain and requires confirmation through well-designed translational and clinical studies^99,100,101.

From an integrative medicine perspective, the findings of this review suggest that acupuncture may serve as a useful adjunct to conventional antidepressant therapy, particularly for patients who experience suboptimal response or adverse effects with SSRIs alone. The observed improvements in tolerability and quality of life may be clinically meaningful, especially within patient-centered care frameworks where treatment adherence and overall well-being are key considerations. However, given the heterogeneity of included studies and the overall low to moderate certainty of evidence, acupuncture should be considered a complementary option within a broader, individualized treatment strategy rather than a universal intervention for depression. Shared decision-making, careful consideration of patient preferences, and integration with established pharmacological and psychological therapies remain essential to optimize clinical outcomes.

The findings highlight the need for high-quality, rigorously designed randomized controlled trials to clarify the role of acupuncture in the management of depression. Future studies should prioritize prospective trial registration, transparent reporting in accordance with CONSORT and STRICTA guidelines, and the use of standardized acupuncture protocols to enhance reproducibility and reduce clinical heterogeneity. Adequate sample size calculations, robust allocation concealment, and blinding of outcome assessors are essential to minimize bias. Given the substantial heterogeneity and variability in study design observed in the current evidence base, future trials should aim to standardize key intervention characteristics, including acupoint selection, stimulation techniques, treatment frequency, and duration. In addition, the incorporation of rigorously designed sham controls is critical to better distinguish specific therapeutic effects from non-specific and contextual influences. Long-term follow-up is needed to determine the durability and sustainability of treatment effects, particularly given the chronic and recurrent nature of depressive disorders. Furthermore, future research should extend beyond symptom-based outcomes to include functional recovery, quality of life, cost-effectiveness, and patient-reported outcomes, which are central to integrative and patient-centered care. Several limitations should be considered when interpreting these findings. First, the majority of included trials were conducted in China, which may limit the generalizability of the results to other populations and healthcare settings. Although both English and Chinese databases were systematically searched, language bias cannot be excluded, as studies with positive findings may be more likely to be published in local-language journals. Additionally, variability in reporting standards across Chinese-language trials may have affected methodological transparency and risk-of-bias assessment. Publication bias is also a concern, particularly given the predominance of small, single-center trials and the relative scarcity of negative studies. Second, culturally specific factors may have influenced treatment outcomes. In settings where acupuncture is widely accepted and integrated into routine clinical practice, patient expectations, practitioner expertise, and the therapeutic context may enhance observed effects. These contextual influences may not be directly transferable to regions where acupuncture is less commonly practiced. Third, only manual and electro-acupuncture were included, excluding other commonly used modalities such as auricular acupuncture or acupressure, which may limit the comprehensiveness of the findings. Fourth, long-term outcomes could not be adequately assessed due to limited follow-up data. Finally, variability in methodological quality, reporting practices, and substantial heterogeneity across studies limits the strength of causal inferences and underscores the need for cautious interpretation of the results.

This systematic review and meta-analysis suggests that acupuncture, particularly when used as an adjunct to selective serotonin reuptake inhibitors, may be associated with short-term reductions in depressive symptom severity and improved treatment tolerability compared with pharmacotherapy alone. The most consistent benefits were observed during the early phases of treatment, particularly at four weeks, while electroacupuncture may confer earlier symptomatic improvement in some comparisons. In contrast, evidence supporting acupuncture or electroacupuncture as a standalone treatment for depression remains inconsistent and is limited by methodological heterogeneity. Across outcomes, the certainty of evidence ranged from very low to moderate; therefore, these findings should be interpreted with caution, and the true effects may differ substantially from the observed estimates. Overall, these findings suggest that acupuncture may serve as a potential integrative adjunct within comprehensive depression care rather than a replacement for conventional therapy. However, given the variability in study design, substantial heterogeneity, and limitations in evidence certainty, firm conclusions regarding efficacy cannot be drawn. Future research should prioritize rigorously designed, adequately powered randomized controlled trials with standardized acupuncture protocols, improved control conditions, transparent reporting, and longer follow-up to clarify the durability and generalizability of treatment effects.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Cochrane Risk of Bias Tool Version 2.0 (ROB 2.0)	Cochrane Collaboration	RRID:SCR_018214	Used for risk-of-bias assessment of included randomized controlled trials.
EndNote X9 (reference management software)	Clarivate Analytics	RRID:SCR_014001	Used for reference management, deduplication, and study screening.
GRADEpro Guideline Development Tool (GRADEpro GDT)	McMaster University / Evidence Prime	RRID:SCR_018239	Used to generate GRADE certainty-of-evidence assessments and Summary of Findings tables.
PRISMA 2020 Checklist	N/A	N/A	Reporting guideline used for conducting and reporting the systematic review and meta-analysis.
PROSPERO database	Centre for Reviews and Dissemination, University of York	N/A	International prospective register of systematic reviews (Registration No. CRD42024526811).
Review Manager (RevMan) Version 5.4	Cochrane Collaboration	RRID:SCR_003581	Used for statistical analyses and meta-analysis.
Standardized data extraction form	Author-developed	N/A	Used for extraction of study characteristics, intervention details, comparator information, outcome measures, and adverse-event data. The form is provided in Supplementary File 3.