Acupuncture for mood disorders in perimenopausal women: A meta-analysis of randomized controlled trials

Abstract

Background:

Acupuncture reportedly improves mood disorders in perimenopausal women, but the effect is controversial. The aim of this study is to evaluate the effect of acupuncture in the treatment of mood disorders in perimenopausal women.

Methods:

Through the electronic retrieval from English and Chinese databases (PubMed, Embase, Cochrane Central Register of Controlled Trials, Web of Science, the China National Knowledge Infrastructure databases, Wangfang databases, VIP databases, and the Chinese Biomedical Literature databases), randomized controlled trial literature regarding acupuncture treatment of mood disorders in perimenopausal women till June 2024 was collected. All of the studies were available in English. Risk ratio, standardized mean differences (SMDs), 95% confidence intervals (CIs), and heterogeneity were calculated.

Results:

A total of 49 randomized controlled trials were included. Acupuncture was effective for mood disorders in perimenopausal women (risk ratio = 0.13; 95% CI = 0.09–0.16; P = .00; I² = 34.42%). In pooled analysis, the acupuncture group could significantly improve Hamilton Depression Scale score (SMD = −0.57; 95% CI = −0.70 to −0.44; P = .00; I² = 64.20%). In anxiety analysis with Hamilton Anxiety Scale and Self-Rating Anxiety Scale scores, there was significant improvement after treatment in Hamilton Anxiety Scale (SMD = −1.11; 95% CI = −0.79 to −0.44; P = .00; I² = 91.52%) and Self-Rating Anxiety Scale (SMD = −1.05; 95% CI = −1.87 to −0.23; P = .01; I² = 90.87%). Compared with drugs, acupuncture significantly improved the Kupperman Index score (SMD = −0.84; 95% CI = −1.23 to −0.46; P = .00; I² = 91.87%) and follicle-stimulating hormone level (SMD = −0.45; 95% CI = −0.84 to −0.05; P = .03; I² = 84.98%). There was no significant difference in luteinizing hormone level (SMD = −0.14; 95% CI = −0.32 to 0.04; P = .12; I² = 0.00%) and estradiol level (SMD = 0.14; 95% CI = −0.46 to 0.74; P = .64; I² = 92.69%).

Conclusions:

This meta-analysis demonstrates that acupuncture is effective in the treatment of mood disorders in perimenopausal women. However, due to some shortcomings in the methodology and quality of the included studies, more rigorously designed, large sample size and multicenter randomized controlled trials are needed to further support the above conclusions.

1. Introduction

Mood disorders in perimenopausal women are nonspecific psychological syndromes mainly characterized by depression and anxiety that occur in women before and after natural menopause.^[1] According to relevant reports, the incidence of mood disorders in perimenopausal women is more than 30%, and the incidence of anxiety and depression in perimenopausal women is 28.5% and 29.23%, respectively.^[2,3] This disease has a certain impact on women’s social function and quality of life.^[4] The pathogenesis of this disease is complex; some studies believe that when women are in the perimenopausal period, their estradiol (E₂) deficiency, follicle-stimulating hormone (FSH), and luteinizing hormone (LH) levels rise, making the hypothalamic-pituitary-gonad axis, hypothalamic-pituitary-adrenal (HPA) axis, and other endocrine axis imbalance. This leads to neuroendocrine dysfunction, enhance women’s social, environmental, physiological sensitivity, and vulnerability, thus causing mood disorders. There is no consensus on treatment like hormone replacement therapy, oral antidepressants, or hormone replacement combined with antidepressants. However, the long-term use of hormones and antidepressants has the side effects and potential risks of increasing the incidence of endometrial, breast, and ovarian cancer.^[5]

According to relevant studies, acupuncture treatment can effectively relieve perimenopausal mood disorders, and because of the cumulative effect of acupuncture, it can make the long-term prognosis effect relatively considerable. Compared with the adverse reactions brought by hormone therapy, acupuncture has the advantages of high safety and small adverse reactions. In addition, studies have confirmed that acupuncture therapy has a multi-target and multi-link benign regulatory effect, which can increase serum E₂, reduce the content of LH and FSH by regulating the function of the hypothalamic-pituitary-ovarian axis of the endocrine system of the body, so as to achieve the effect of treating the related symptoms of perimenopausal mood disorders.^[6,7]

In recent years, with the deepening of the research on acupuncture in the treatment of perimenopausal mood disorders,^[8,9] relevant literature has been published continuously. This study intends to update and expand the sample size with acupuncture as the intervention measure and drug therapy as the control. We attempted to conduct a comprehensive search and evidence evaluation of clinical research literature on acupuncture in the treatment of perimenopausal mood disorders, in order to clarify its safety and efficacy, and provide a certain reference for clinical treatment.

2. Methods

2.1. Protocol and registration

This meta-analysis was performed according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines and checklist.^[10] The protocol registration was performed with the PROSPERO international prospective register of systematic reviews and is available online (Registration number: CRD42024570396).^[11] The ethical approval is not necessary. This meta-analysis is a secondary study based on existing literature data and does not contain any data which can identify individual. And the data for this meta-analysis are sourced entirely from published research, and before publication, these original studies have undergone rigorous ethical reviews.

2.2. Selection criteria

In the meta-analysis, trials were included that met the following criteria: (1) Types of study. Randomized controlled trials (RCTs) are included, no matter if blinding method was used or not, and language of the essays was not restricted. (2) Patients’ ages and sources of cases were not restricted. But a definite diagnostic criterion was needed: the Diagnostic and Statistical Manual of Mental Disorders,^[12] the International Classification of Diseases and Related Health Problems (ICD-10) Chapter V: Mental Disorder Classification and diagnostic criteria,^[13] and Chinese Classification of Mental Disorders.^[14,15] Patients with other basic diseases, drug addiction, schizophrenia, dementia, or bipolar disorder were excluded. (3) The active group was given acupuncture therapy (acupuncture material, acupoint for treatment, operation procedure of acupuncture, needle retaining time, or course of treatment were not restricted). The control group was given medicine therapy (both Chinese and Western medicines are included), sham acupuncture, or placebo acupuncture.^[16] (4) One or more of the following outcomes were assessed: emotional assessment indicators, symptom scores, laboratory indicators, etc.

2.3. Search strategy

A comprehensive literature search was conducted in several databases (PubMed, Embase, Cochrane Central Register of Controlled Trials, Web of Science, the China National Knowledge Infrastructure databases, Wangfang databases, VIP databases, and the Chinese Biomedical Literature databases). The literature search will be conducted until the end of June 25, 2024 without language or date restrictions. Detailed search strategies are in Appendix S1, Supplemental Digital Content, https://links.lww.com/MD/P655.

2.4. Data extraction

Two investigators (YL and XM) independently reviewed the full manuscripts of eligible studies, removed duplicates, and screened the titles and abstracts of the selected studies for eligibility. Subsequently, the same 2 authors independently performed a full-text review of the selected studies, based on the inclusion and exclusion criteria. The primary outcomes are emotion-related scales (such as anxiety and depression), and data on the overall response rate, Kupperman Index (KMI), traditional Chinese medicine (TCM), LH, FSH, E₂, etc were also collected to evaluate the efficacy of acupuncture in hormone levels and related scales of women with perimenopausal mood disorders. They extracted information into an electronic database, including study period, study location, disease duration, treatment protocols that were compared, and number of patients assigned to each group. All references within the included studies were cross referenced for inclusion if they were missed by the initial search. Disagreements were resolved by discussion between the review authors or by consultation with another author (YX).

2.5. Assessment of risk of bias

The bias risk assessment of the included studies was evaluated using the bias risk assessment tool for RCTs recommended by Cochrane Manual 5.2.0. The following factors were assessed: random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and other bias. The risk of bias was assessed by 2 evaluators independently (YL, XM), and the results were cross-checked. In case of disagreement, discussion and resolution were conducted. And each included study was scored as being at a low, unclear, or high risk of bias.

2.6. Statistical analysis

Data analyses were performed with Stata 17.0 (StataCorp LLC, College Station) statistical software. The outcome index includes categorical variables and continuity variables. Risk ratio was used as the effect index for categorical variables. For the continuity variables, standardized mean differences (SMDs) with corresponding 95% confidence intervals (CIs) were utilized to standardize the results of studies to a uniform scale. Statistical heterogeneity was assessed with I² statistics: I² ≥ 50%, substantial heterogeneity; I² < 50%, low heterogeneity. If I² ≥ 50%, it indicates that statistical heterogeneity exists among study results, and further analysis of the source of heterogeneity is required. A random effects model was used because the studies included in this analysis were assumed to be random samples of all possible studies, thereby increasing the risk of heterogeneity. When I² ≥ 50%, factors that may affect heterogeneity, such as acupuncture frequency, acupuncture method, and so on, were taken as independent variables, and the effect size was taken as dependent variable. A regression model was established to obtain the regression coefficient; the regression coefficient and its statistical significance were analyzed to determine whether the factors were related to heterogeneity. Based on the results of meta-regression, subgroup analysis was performed on the corresponding independent variables to explore the sources of heterogeneity, and then to evaluate the treatment effects of patients in different subgroups.

The sensitivity analysis was to reanalyze the meta-analysis after removing individual studies, and evaluate the differences between the eliminated results and the original combined results. If the number of studies is ≥10, funnel plots are made to test publication bias, and Stata 17.0 software is used to perform the quantitative evaluation of Egger test by linear regression method, and the test level is α = 0.05.

3. Results

3.1. Study characteristics

The selection process for the studies is shown in a flow diagram in Figure 1. Forty-nine RCTs were included for meta-analysis.^[17–65] A total of forty-nine studies were included, among which 6 studies did not report on the the overall response rate.^{[23,25,43,50,63,65]} Seven of them did not report Hamilton Depression Scale (HAMD), a measure of depression outcomes.^{[17,20,40,47,51,56,64]} In terms of anxiety, 2 scales were used to assess it; 8 studies reported Hamilton Anxiety Scale (HAMA)^{[27,28,33,34,39,41,57,58]} and 5 studies reported Self-Rating Anxiety Scale (SAS).^{[21,27,34,39,57]} Twenty-two studies reported KMI^{[17–19,22,23,25–27,31,32,35,37–42,52,55,56,60,61]}; TCM syndrome score was reported in 8 studies.^{[19,32,37,38,42,54,55,60]} In terms of hormone levels, 8 articles reported hormone levels before and after FSH,^{[18,19,26,37,40,49,52,62]} 5 articles reported hormone levels before and after LH,^{[18,19,25,49,62]} and 7 articles reported hormone levels before and after E₂.^{[18,19,26,40,49,50,62]} Details of these studies are summarized in Table 1.

Figure 1.

Flow diagram of articles during selection process using PRISMA (Preferred Reporting Items for Systematic Meta-Analyses) guidelines. RCT = randomized controlled trial.

Table 1.

Summary of studies’ characteristics.

Study	Acupuncture group				Control group			Treatment frequency (wk)	Clinical evaluation	Adverse reaction	Study duration
	Sample size, n	Age (yr)	Intervention	Acupuncture method	Sample size, n	Age (yr)	Intervention
Huang[17]	30	Not mentioned	Acupuncture	Not mentioned	30	Not mentioned	Western medicine	6	①⑤⑥	Have mentioned	8 wk
Zheng et al[18]	60	52.27 ± 3.45	Acupuncture	Not mentioned	60	51.98 ± 3.14	Western medicine	3	①②⑤⑦⑧⑨	Not mentioned	3 mo
Wang[19]	31	50.16 ± 2.88	Acupuncture + ear pressure beans	Not mentioned	31	50.10 ± 2.74	Western medicine	5	①②⑤⑥⑦⑧⑨	Have mentioned	6 wk
Zhou et al[20]	60	48.4 ± 4.5	Acupuncture	Not mentioned	30	48.98 ± 4.66	Western medicine	3	①	Not mentioned	3 mo
Tian et al[21]	30	50.50 ± 4.88	Acupuncture	Have mentioned	30	50.03 ± 5.79	Ordinary acupuncture	7	①②④	Not mentioned	2 mo
Han[22]	30	50.03 ± 4.43	Acupuncture	Not mentioned	30	48.70 ± 4.93	Western medicine	3	①②⑤	Have mentioned	4 wk
Xiang et al[23]	23	49 ± 4	Acupuncture + thread burying	Not mentioned	23	48 ± 3	Ordinary acupuncture	1	②⑤	Not mentioned	8 wk
Liu[24]	29	50.36 ± 7.94	Acupuncture	Not mentioned	29	51.48 ± 7.63	Ordinary acupuncture	3	①②	Not mentioned	2 mo
Ding and Liu[25]	39	49.68 ± 3.90	Acupuncture	Have mentioned	39	49.50 ± 3.51	Western medicine	6	②⑤	Not mentioned	30 d
Zhang[26]	44	48.48 ± 5.39	Electro-acupuncture	Not mentioned	46	48.16 ± 4.15	Western medicine	7	①②⑤⑥⑦⑧⑨	Have mentioned	3 mo
Fan[27]	30	49.10 ± 2.60	Electro-acupuncture	Not mentioned	30	48.10 ± 2.47	Ordinary acupuncture	7	①③④⑤	Not mentioned	3 wk
Li and Dai[28]	30	Not mentioned	Electro-acupuncture	Not mentioned	39	Not mentioned	Western medicine	3	①②③	Not mentioned	6 wk
Sun et al[29]	21	50.29 ± 2.59	Electro-acupuncture	Not mentioned	21	49.86 ± 3.83	Western medicine	3	①②	Not mentioned	12 wk
Shi et al[30]	30	48.70 ± 1.99	Electro-acupuncture	Not mentioned	30	49.43 ± 1.87	Western medicine	3	①②	Not mentioned	12 wk
Wu et al[31]	30	49.60 ± 3.12	Electroacupuncture + Chinese medicine	Not mentioned	30	49.51 ± 3.23	Chinese medicine	6	①②⑤	Not mentioned	8 wk
Sun and Tao[32]	30	49 ± 2	Acupuncture	Have mentioned	30	49 ± 2	Ordinary acupuncture	6	①②⑤⑥	Have mentioned	6 wk
Liu[33]	32	Not mentioned	Acupuncture	Have mentioned	30	Not mentioned	Sham acupuncture	3	③	Have mentioned	4 wk
Xie[34]	32	50.22 ± 3.59	Acupuncture	Have mentioned	31	49.16 ± 3.25	Ordinary acupuncture	3	①③④	Have mentioned	4 wk
Wang and Ye[35]	30	49.20 ± 3.98	Acupuncture	Have mentioned	30	49.60 ± 3.90	Ordinary acupuncture	3	①②⑤	Have mentioned	8 wk
Wang et al[36]	30	49.6 ± 4.3	Acupuncture	Not mentioned	30	48.3 ± 4.7	Western medicine	3	①②	Have mentioned	4 wk
Zhao[37]	32	49.28 ± 1.98	Acupuncture	Have mentioned	32	49.64 ± 2.58	Ordinary acupuncture	5	①②⑤⑥	Have mentioned	4 wk
Lin[38]	31	Not mentioned	Acupuncture	Not mentioned	31	Not mentioned	Western medicine	3	①②⑤⑥	Have mentioned	4 wk
Mu[39]	32	47.89 ± 4.99	Acupuncture	Have mentioned	32	48.13 ± 4.01	Ordinary acupuncture	5	①③④⑤	Not mentioned	4 wk
Liang and Du[40]	30	47.9 ± 2.7	Acupuncture	Have mentioned	30	47.6 ± 2.8	Western medicine	2	①⑤⑦⑨	Not mentioned	4 wk
Bi[41]	33	50.91 ± 3.54	Acupuncture	Have mentioned	33	50.94 ± 3.94	Ordinary acupuncture	6	①②③⑤	Not mentioned	3 wk
Cai[42]	30	42.33 ± 6.75	Acupuncture	Have mentioned	30	41.05 ± 7.20	Ordinary acupuncture	6	①②⑤⑥	Not mentioned	2 wk
Qian et al[43]	33	Not mentioned	Acupuncture	Not mentioned	30	Not mentioned	Western medicine	5	②	Not mentioned	6 wk
Xing[44]	120	51.2 ± 5.4	Acupuncture	Have mentioned	120	49.5 ± 6.8	Western medicine	7	②	Have mentioned	6 wk
Chi and Zou[45]	30	51.63 ± 1.72	Acupuncture	Have mentioned	30	51.43 ± 1.62	Western medicine	7	①②	Have mentioned	4 wk
Lin[46]	42	50.3 ± 1.1	Electro-acupuncture	Not mentioned	45	51.1 ± 1.3	Western medicine	3	①②	Not mentioned	3 mo
Jiang and Qiu[47]	34	44.71 ± 3.50	Acupuncture + psychological counseling	Not mentioned	32	44.43 ± 2.88	Psychological counseling	3	①	Not mentioned	7 wk
Dong[48]	30	Not mentioned	Acupuncture	Have mentioned	30	Not mentioned	Western medicine	7	①②	Not mentioned	30 d
Zhang[49]	94	50.1 ± 2.7	Acupuncture	Have mentioned	94	49.8 ± 2.6	Western medicine	7	②⑦⑧⑨	Have mentioned	3 mo
Zhou and Hao[50]	30	51.8 ± 4.2	Acupuncture	Not mentioned	28	48.9 ± 3.8	Western medicine	6	①②	Not mentioned	6 wk
Zhou and Wu[51]	30	Not mentioned	Acupuncture	Not mentioned	28	Not mentioned	Western medicine	6	⑦⑨	Not mentioned	6 wk
Zhang et al[52]	30	50 ± 3		Not mentioned	30	49 ± 2	Western medicine	3	①②⑤	Not mentioned	8 wk
Bai and Yang[53]	20	47.22 ± 7.62	Acupuncture	Not mentioned	20	48.02 ± 7.85	Western medicine	3	①②	Not mentioned	8 wk
Niu and Wang[54]	41	54.1 ± 2.0	Acupuncture	Not mentioned	41	54.2 ± 2.1	Western medicine	5	①②⑤⑥	Have mentioned	6 wk
Ma[55]	30	50.73 ± 2.44	Acupuncture	Have mentioned	30	50.41 ± 2.70	Ordinary acupuncture	6	①②③⑤⑥	Not mentioned	4 wk
Kou[56]	31	48 ± 5	Acupuncture	Not mentioned	29	48 ± 5	Western medicine	7	①⑤	Not mentioned	23 d
Huang and Cui[57]	30	52.07 ± 2.82	Acupuncture + Western medicine	Not mentioned	30	53.20 ± 3.25	Western medicine	5	①③④	Have mentioned	4 wk
Chen and Zhou[58]	30	51.8 ± 4.2	Acupuncture	Not mentioned	30	48.9 ± 3.8	Western medicine	6	①②③	Not mentioned	6 wk
Qiang[59]	30	54.32 ± 3.29	Acupuncture	Not mentioned	30	54.0 ± 4.62	Western medicine	5	①②	Not mentioned	6 wk
Li[60]	32	50.59 ± 2.94	Acupuncture	Not mentioned	32	50.25 ± 2.71	Western medicine	6	①②⑤⑥	Not mentioned	12 wk
Wang[61]	35	47.76 ± 4.52	Acupuncture + thread burying	Not mentioned	33	47.32 ± 4.30	Western medicine	2	①②③⑤	Not mentioned	12 wk
Zheng et al[62]	40	50.78 ± 0.54	Acupuncture	Not mentioned	40	50.82 ± 0.58	Western medicine	5	①②⑦⑧⑨	Not mentioned	6 wk
Xu[63]	58	51.05 ± 2.04	Acupuncture + Western medicine	Not mentioned	54	51.27 ± 1.90	Western medicine	7	②	Have mentioned	4 wk
Lin and Sun[64]	40	Not mentioned	Acupuncture	Not mentioned	40	Not mentioned	Western medicine	5	①	Not mentioned	6 wk
Zhou et al[65]	30	50.03 ± 2.79	Acupuncture	Have mentioned	30	49.67 ± 2.98	Ordinary acupuncture	5	②	Not mentioned	4 wk

① total effective rate; ② HAMD, Hamilton Depression Scale; ③ HAMA, Hamilton Anxiety Scale; ④ SAS, Self-Rating Anxiety Scale; ⑤ KMI, Kupperman Index; ⑥ TCM syndrome score (traditional Chinese medicine syndrome score); ⑦ FSH, follicle-stimulating hormone; ⑧ LH, luteinizing hormone; ⑨ E₂, estradiol.

3.2. Risk of bias within studies

The RevMan 5.4 software recommended by the Cochrane Collaboration was used to evaluate the included RCTs according to the bias risk assessment tool. The results of the assessment for the risk-of bias in the included studies are summarized in Figure 2. A total of 49 literatures were included, and the overall quality of the literatures was low. Forty-eight literatures used computer-generated random sequences, which were considered appropriate. However, 2 studies did not clearly describe their randomization methods. Chi and Zou^[45] categorized the population by order of visit in selection bias, and reporting bias in this study was rated high. Lin^[46] did not explicitly mention the random method which lead to an unclear risk of selection bias. Allocation concealment methods were mentioned in 11 literatures.^{[18,19,27,28,32–34,38,52,53,61]} For instance, Li^[38] based on the chronological order of patient enrollment, corresponding envelopes were given to the patients, and they were treated according to the randomly assigned groups. This ensured the concealment of the allocation process. Six articles mentioned double blindness.^{[32–34,38,52,61]} The assessment of attrition bias and reporting bias for the included literature was rated as low risk. For instance, in the experimental design, Zhang et al^[52] clearly stated that the efficacy evaluation personnel were unaware of the study grouping scheme, and the data statisticians were not involved in the study design and implementation. Wang^[61] also clearly adopted the double-blind method in the experimental design. That is, the patients were unaware of the differences in efficacy between the 2 treatment options. The person executing the plan was different from the personnel conducting the efficacy assessment, and the evaluators were not aware of the patients’ groupings. The remaining literature did not incorporate the use of blinding in their experimental designs, so they are classified as “unclear.” Among the 49 included studies, the data integrity was good with no significant data loss. They were classified as “low.” In terms of selective reporting, for relevant indicators such as efficacy rate and emotion-related scales, all were fully reported, so they were classified as “low.” Regarding other biases, no specific mention was made, thus they were classified as “unclear.”

Figure 2.

Summary of the risk-of-bias assessment for included studies. (A) Risk of bias summary. (B) Risk of bias graph.

3.3. Results of individual studies

3.3.1. Analysis of mood disorders in perimenopausal women with acupuncture

3.3.1.1. Results for depression

At present, HAMD scale is widely used in the study of depression in perimenopausal women. Thirty-seven RCTs were included for this mood indicator. Compared with the control group, the ratio of HAMD score in the treatment group was significantly decreased (SMD = −0.57; 95% CI = −0.70 to −0.44; P = .00; I² = 64.20%), using a random effects model (Fig. 3). In order to further find the source of heterogeneity, meta-regression was performed on “acupuncture method,” “acupuncture frequency,” “number of acupoints,” “treatment duration,” and “needle insertion with retained needle,” and the results suggested that “acupuncture method” (P = .027), “acupuncture frequency” (P = .015), and “number of acupoints” (P = .014), were the source of heterogeneity (Table 2). Therefore, subgroup analysis was conducted with “acupuncture method,” “acupuncture frequency,” and “number of acupoints” as the entry point. The results were as follows: A subgroup analysis of “acupuncture method” is compared with the other 2 groups; there was high heterogeneity between studies in the general acupuncture group (using acupuncture only) (I² = 73.92%). For general acupuncture combined with other treatments (such as ear point pressure bean), heterogeneity was relatively low (I² = 45.33%). And there was heterogeneity among the general acupuncture combined with special needles (I² = 49.42%) (Fig. 4A). It was showed that the general acupuncture group, general acupuncture combined with other treatment groups, and acupuncture combined with special needles group are all beneficial to the control group in the treatment of perimenopausal mood disorders. But there was obvious heterogeneity in the general acupuncture group; the results require further exploration of more literature and mechanisms. A subgroup analysis of “acupuncture frequency”; the results showed that regardless of whether the acupuncture frequency was >5 times/week (I² = 69.16%) or ≤5 times/week (I² = 55.14%) (Fig. 4B), both subgroups had high heterogeneity. Further studies with higher quality literature are needed to investigate the impact of acupuncture frequency on the efficacy of acupuncture. A subgroup analysis was conducted on the acupuncture points, and the results showed that there was a high degree of heterogeneity in both subgroups (Fig. 4C).

Figure 3.

Forest plots of standardized mean difference with 95% CI in the HAMD for depression. Random effect models were used. CI = confidence interval, HAMD = Hamilton Depression Scale.

Table 2.

Meta-regression to find the source of heterogeneity in depression.

_ES	Coefficient	Std. err.	t	P >  t	95% confidence interval
Acupuncture method	−.1889796	.0816396	−2.31	.027	−.3554847	−.0224746
Acupuncture frequency	−.3639217	.1417965	−2.57	.015	−.6531175	−.0747258
Number of acupoints	.3708611	.1417763	2.62	.014	.0817063	.6600158
Treatment duration	−.0590426	.0829699	−0.71	.482	−.2282609	.1101756
Needle insertion with retained needl	.1363085	.2046355	0.67	.510	−.2810484	.5536654
_cons	−.2859228	.2569905	−1.11	.274	−.8100584	.2382128

Figure 4.

Forest plots of standardized mean difference with 95% CI in the HAMD for depression. Subgroup analysis of HAMD score was performed. Based on the different forms of acupuncture, they are divided into 3 subgroups, namely “general acupuncture,” “general acupuncture combined with other treatments,” and “acupuncture combined with special needles.” Random effect models were used. (A) Subgroup analysis of acupuncture methods. (B) Subgroup analysis of acupuncture frequency. (C) Subgroup analysis of number of acupoints. CI = confidence interval, HAMD = Hamilton Depression Scale.

3.3.1.2. Results for anxiety

HAMA and SAS scales are generally used to assess the change of anxiety. Regarding the HAMA scale, a total of 8 literatures were included.^{[27,28,33,34,39,41,57,58]} The meta-analysis showed significant heterogeneity between the studies (SMD = −1.11; 95% CI = −0.79 to −0.44; P = .00; I² = 91.52%) (Fig. 5), using a random effects model. According to the subgroup analysis of “weekly acupuncture frequency,” the outcomes are presented as follows: when the acupuncture frequency is ≤5 times/week, there was low heterogeneity among the studies (I² = 8.27%). In contrast, when the acupuncture frequency is >5 times per week, there was significant heterogeneity among studies (I² = 94.81%) (Fig. 6). Based on the results of the meta-analysis, the former group has a more stable curative effect. And the results suggested that weekly acupuncture frequency may be one of the factors leading to heterogeneity. Acupuncture can improve HAMA score in perimenopause. With regard to the group with acupuncture >5 times/week, more studies are needed to prove curative effect in the future. For the group of acupuncture frequency ≤5 times per week, the results are beneficial to intervention of perimenopausal anxiety symptoms.

Figure 5.

Forest plots of standardized mean difference with 95% CI in the HAMA for anxiety. Random effect models were used. CI = confidence interval, HAMA = Hamilton Anxiety Scale.

Figure 6.

Forest plots of standardized mean difference with 95% CI in the HAMA for anxiety. Subgroup analysis of HAMA score was performed. Random effect models were used. CI = confidence interval, HAMA = Hamilton Anxiety Scale.

Five studies were encompassed in the SAS score assessment, and meta-analysis revealed that acupuncture was beneficial to improve the SAS score in perimenopause; the heterogeneity was high (SMD = −1.05; 95% CI = −1.87 to −0.23; P = .01; I² = 90.87%) (Fig. 7). After deleting one article,^[54] when the meta-analysis was reanalyzed, it was found that the heterogeneity of the findings diminished considerably (SMD = −1.44; 95% CI = −1.87 to −1.09; P = .00; I² = 34.13%) (Fig. 8).

Figure 7.

Forest plots of standardized mean difference with 95% CI in the SAS for anxiety. Random effect models were used. CI = confidence interval, SAS = Self-Rating Anxiety Scale.

Figure 8.

The results of excluding a paper. Forest plots of standardized mean difference with 95% CI in the SAS for anxiety. Random effect models were used. CI = confidence interval, SAS = Self-Rating Anxiety Scale.

3.4. The overall response rate

According to the results of clinical studies, the marked effective (significant reduction in clinical symptoms and signs), effective (an effect of some improvement), and cured (complete disappearance of clinical symptoms and signs) were consolidated into the category of “effective,” while the rest were regarded as “ineffective.” The combined results were then analyzed using meta-analysis. Here are the aggregated results: the overall response rate scores in the acupuncture group were significantly decreased (risk ratio = 0.13; 95% CI = 0.09–0.16; P = .00; I² = 34.42%) as compared with the control group, and the corresponding indicated low heterogeneity (Fig. 9). By conducting meta-regression on the relevant factors (Table 3), the results indicate that the “acupuncture method” and “acupuncture frequency” are 2 factors that may contribute to the heterogeneity. Subgroup analysis was made on the related factors that may affect the heterogeneity, such as “acupuncture frequency” (P = .014) and “acupuncture method” (P = .006). The results revealed that the therapeutic effect of the general acupuncture group, general acupuncture combined with other treatment groups, and acupuncture combined with special needles group were all beneficial to the intervention of perimenopausal women with mood disorders. Meanwhile, the subgroup analysis with respect to the frequency of acupuncture revealed that all groups were effective in treating perimenopausal mood disorders (Figs. 10 and 11).

Figure 9.

Forest plots of risk ratio with 95% CI in the overall response rate. As one of the secondary outcome indicators, it assesses the overall therapeutic effect. Random effect models were used. CI = confidence interval.

Table 3.

Meta-regression to find the source of heterogeneity in the overall response rate.

_ES	Coefficient	Std. err.	t	P >  t	95% confidence interval
Acupuncture method	.0616795	.0212902	2.90	.006	.0185416	.1048175
Acupuncture frequency	−.0947691	.0365606	−2.59	.014	−.1688479	−.0206902
Number of acupoints	−.0117226	.0392645	−0.30	.767	−.09128	.0678348
Treatment duration	.0322226	.0227654	1.42	.165	−.0139045	.0783498
_cons	−.0116964	.0821299	−0.14	.888	−.1781074	.1547146

Figure 10.

Forest plots of standardized mean difference with 95% CI in the overall response rate. Subgroup analysis of the overall response rate was performed. Random effect models were used. CI = confidence interval.

Figure 11.

Forest plots of standardized mean difference with 95% CI in the overall response rate. Subgroup analysis of the overall response rate was performed. Random effect models were used. CI = confidence interval.

3.5. KMI score

In the context of the analysis regarding KMI score, 22 pieces of literature were incorporated. Meta-analysis showed that acupuncture could improve KMI score, and the heterogeneity of results was high (SMD = −0.84; 95% CI = −1.23 to −0.46; P = .00; I² = 91.87%) (Fig. 12). Meta-regression was performed on “weekly acupuncture frequency,” “acupuncture method,” “duration of illness,” and so on, and no heterogeneity was found. The results need to be confirmed by more studies.

Figure 12.

Forest plots of standardized mean difference with 95% CI in the KMI score. Random effect models were used. CI = confidence interval, KMI = Kupperman Index.

3.6. TCM syndrome score

According to the changes of TCM syndrome score before and after treatment, 8 literatures meeting the requirements were included. The results of meta-analysis (SMD = −0.78; 95% CI = −1.01 to −0.55; P = .00; I² = 36.93%) (Fig. 13) suggested that acupuncture could improve the scores of perimenopausal women in TCM mood disorders.

Figure 13.

Forest plots of standardized mean difference with 95% CI in the TCM syndrome score. Random effect models were used. CI = confidence interval, TCM = traditional Chinese medicine.

3.7. LH, FSH, and E₂

Analysis of LH content in the treatment of perimenopausal mood disorders by acupuncture. A total of 4 studies were included, and the meta-analysis results showed that there was no statistically significant difference in the effect of acupuncture on improving the LH levels in women with perimenopausal mood disorders (SMD = −0.14; 95% CI = −0.32 to 0.04; P = .12; I² = 0.00%) (Fig. 14).

Figure 14.

Forest plots of standardized mean difference with 95% CI in the LH. Random effect models were used. CI = confidence interval, LH = luteinizing hormone.

A total of 8 literatures were included to analyze FSH content; the meta-analysis showed significant heterogeneity among the studies (SMD = −0.45; 95% CI = −0.84 to −0.05; P = .03; I² = 84.98%) (Fig. 15), using a random effects model. According to the subgroup analysis of “weekly acupuncture frequency,” the results were as follows: when the frequency of acupuncture included in the study was ≤5 times per week, there was significant heterogeneity among the studies (I² = 85.47%), and more studies are needed to prove this in the future; when the frequency of acupuncture included in the study was >5 times per week, the study heterogeneity was low (I² = 49.29%) (Fig. 16). The results suggest that “acupuncture frequency” may be one of the factors leading to heterogeneity.

Figure 15.

Forest plots of standardized mean difference with 95% CI in the FSH. Random effect models were used. CI = confidence interval, FSH = follicle-stimulating hormone.

Figure 16.

Forest plots of standardized mean difference with 95% CI in the FSH. Subgroup analysis of FSH was performed. Random effect models were used. CI = confidence interval, FSH = follicle-stimulating hormone.

As for E₂, a total of 7 literatures were included into the meta-analysis. It showed that acupuncture had no statistical difference in E₂ content in the treatment of perimenopausal mood disorders (SMD = 0.14; 95% CI = −0.46 to 0.74; P = .64; I² = 92.69%) (Fig. 17).

Figure 17.

Forest plots of standardized mean difference with 95% CI in the E₂. Random effect models were used. CI = confidence interval, E₂ = estradiol.

3.8. Publication bias

The overall response rate was tested for publication bias, and the funnel plot suggested the existence of publication bias (Fig. 18A). Additionally, using the Egger test, we found significant publication bias (P = .000). The adjusted funnel plot after trim and fill is shown in Figure 18B; 14 articles were obtained, and 13 of which were not statistically significant.

Figure 18.

(A) Funnel plot to evaluate overall effectiveness (Egger test, P = .000). (B) Adjusted funnel plot after trim and fill also indicated the publication bias.

3.9. Sensitivity analysis

After removing a single study for sensitivity analysis, we performed a meta-analysis of the outcome index; the results were demonstrated (Fig. 19A–I). And the results from the sensitivity analyses were similar to those from the analyses with initial assessment.

Figure 19.

(A) Sensitivity analysis on the overall response rate. (B) Sensitivity analysis on HAMD. (C) Sensitivity analysis on HAMA. (D) Sensitivity analysis on SAS. (E) Sensitivity analysis on KMI. (F) Sensitivity analysis on TCM syndrome score. (G) Sensitivity analysis on FSH. (H) Sensitivity analysis on LH. (I) Sensitivity analysis on E₂. E2 = estradiol, FSH = follicle-stimulating hormone, HAMA = Hamilton Anxiety Scale, HAMD = Hamilton Depression Scale, KMI = Kupperman Index, LH = luteinizing hormone, SAS = Self-Rating Anxiety Scale, TCM = traditional Chinese medicine.

4. Discussion

This meta-analysis demonstrated that acupuncture can improve the symptoms of mood disorders in perimenopausal women. Nevertheless, due to some shortcomings in the methodology and quality of the included studies, more rigorous design and larger sample, multicenter RCTs were needed to further support these conclusions. The outcomes of the meta-analysis were as follows: (i) Acupuncture could improve mood disorders in perimenopausal women. In the context of perimenopausal women with depression, acupuncture was more effective than the control group in symptom improvement; at the same time, based on the results of the subgroup analysis, the effect of general acupuncture combined with other treatments (such as ear acupuncture, burying thread, etc) and general acupuncture combined with special needles group was greater than general acupuncture.^[66] Regarding perimenopausal women with anxiety, the meta-analysis findings indicated that acupuncture was conducive to the improvement of female perimenopausal anxiety symptoms. Moreover, as for HAMA index, the results showed that the frequency of weekly acupuncture might exert a certain influence on clinical efficacy, and the results reported that the group with a acupuncture frequency of ≤5 times/week might exhibit a relatively inferior outcome compared to the group with a acupuncture frequency of >5 times/week. (ii) Acupuncture was capable of enhancing the KMI score and the TCM syndrome score of perimenopausal women with emotional disorders; (iii) Acupuncture could improve the hormone levels of perimenopausal women’s emotional disorders to some extent, and it suggested that acupuncture could improve FSH levels in women. However, for LH and E₂, no statistically significant differences were observed in the treatment outcomes.

At present, in the treatment of mood disorders among perimenopausal women, Western medicine mainly adopts hormone therapy.^[72,73] When it comes to TCM treatment modalities, acupuncture emerges as a highly favored alternative. Based on previous published studies, and further conducting an in-depth analysis of acupuncture treatment for perimenopausal women’s emotional disorders^[67–69,74]: (i) In contrast to prior studies,^{[70,71,75,76]} this paper incorporated 49 RCTs, expanding and updating the sample size. Li et al^[75] and Guo ^[76] reported the difference in efficacy between acupuncture and Western medicine drugs in treating perimenopausal women’s emotional disorders, verifying the effectiveness of acupuncture treatment for the disease, but no in-depth exploration of the therapeutic approach was conducted. As far as this study was concerned, it delved into the heterogeneity of the results by examining potential factors that may contribute to it, such as acupuncture frequency, acupuncture techniques, disease duration, and so on. Furthermore, it conducted meta-regression and subgroup analysis to further explore the influence of different factors on the treatment effect of the disease. Sun et al,^[32] Wang and Ye,^[35] Zhao,^[37] and Ma^[55] employed specialized acupuncture therapies, such as “relocating and calming acupuncture method,” “Feng Teng Ba Fa,” “supplementing mother and purging son method,” etc. Compared with ordinary acupuncture, these acupuncture treatments demonstrated enhanced effectiveness; acupuncture combined with special acupuncture method can achieve better clinical efficacy, which was likely related to the special acupuncture method that could better stimulate acupoints and more accurately regulate the meridians, Qi and blood, etc. The specific mechanism needs to be verified by more studies in the future. Regarding the treatment frequency, 49 studies were included in this research, yet their acupuncture frequencies varied. For instance, Xiang et al^[23] controlled the acupuncture frequency at once per week, Zhang^[26] at twice per week, Zheng et al^[18] at 3 times per week, Wang^[19] at 5 times per week, Ding and Liu^[25]at 6 times per week, and Zhang^[26] at 7 times per week. According to the included study characteristics, the acupuncture frequency was subjected to subgroup analysis and classified into 2 categories: frequency ≤5 times/week and >5 times/week, aiming to approximately determine the appropriate acupuncture frequency. (ii) In previous studies, Zheng et al^[18,62]posited that acupuncture could enhance the E₂ levels of the research subjects by negatively regulating the hypothalamic-pituitary axis, influencing the release of gonadotropin-releasing hormone, and regulating monoamine neurotransmitters in the hypothalamus, thus alleviating clinical symptoms. Nevertheless, there was no statistical significance observed in FSH and LH. Wang^[19] and Zhang^[26] believed that the levels of FSH, LH, and E₂ of the research subjects could all be improved through acupuncture, and the therapeutic effect was remarkable. Liang and Du^[40] believed that acupuncture could increase the levels of FSH and E₂. Zhang^[49]believed that the therapeutic effect of acupuncture in improving female hormone levels was comparable to that of Western medicine, both of which could adjust the reproductive endocrine system of patients and thereby regulate their hormone levels. Zhou and Hao^[50] considered that acupuncture could improve the E₂ levels of patients, but there was no statistical significance in improving FSH. From the above-mentioned literature, it could be seen that acupuncture has different results in improving the hormone levels of women with perimenopausal mood disorders. Through meta-analysis, this study concluded that acupuncture can enhance the FSH levels of women with perimenopausal mood disorders, yet the improvement effect on LH and E₂ was not conspicuous. Regarding the underlying mechanism, the following considerations are made: (1) Discrepancies in neural regulatory pathways; the regulatory pathway of FSH secretion may have a more direct and intimate connection with the neural conduction pathway stimulated by acupuncture. In contrast, E₂ and LH, compared to FSH, are influenced by a greater number of factors and are relatively less variable. (2) Differences in hormone characteristics; the research subjects are women with perimenopausal mood disorders, and the FSH level usually increases significantly during this period, with obvious changes, and the results after acupuncture intervention are more pronounced. (3) Existing limitations in current studies: these include limitations in sample size and research methods (such as acupoint selection, acupuncture techniques, treatment frequency, etc), which contribute to the less obvious improvement in results.

Based on the outcomes of the subgroup analysis, the primary determinants influencing the heterogeneity of the results are identified as the frequency of acupuncture and the acupuncture method employed. It has been discerned that the integration of acupuncture with other traditional Chinese medical modalities (such as ear acupuncture and subcutaneous implantation), along with the utilization of specialized acupuncture techniques, yields more favorable clinical outcomes in the treatment of emotional disorders among perimenopausal women. According to current research, the onset of perimenopausal emotional disorders is related to low estrogen levels, perimenopausal symptoms, environmental stress, and genetic factors. The pathogenesis intricately involves disorders of the HPA axis, structural and functional changes within the limbic system, fluctuations in neurotransmitters such as 5-hydroxytryptamine, gamma-aminobutyric acid, and norepinephrine, and imbalanced release of inflammatory cytokines.^[69,70] Acupuncture exerts a positive effect on improving the HPA axis, serum melatonin, norepinephrine, and 5-hydroxytryptamine levels of patients. Additionally, other traditional Chinese medical treatments, such as ear acupuncture, subcutaneous implantation, and Chinese herbal medicine, also play a certain positive role in these aspects. By combining the approaches, the clinical symptoms of perimenopausal emotional disorders can be further alleviated. As for the effect of special acupuncture techniques, there is currently a lack of research on its clinical mechanism from a Western medicine perspective.^[78] From the vantage point of TCM, ordinary acupuncture therapy cannot fully consider individual differences in patients, and the therapeutic effect may be affected. Special acupuncture techniques utilize the unique TCM theory of differential diagnosis and treatment and are tailored to individual patients, which have the potential to further enhance the therapeutic outcomes. In terms of the frequency of acupuncture, in previous studies, Zhai^[77] posited that the long-term efficacy of electroacupuncture in improving symptoms related to perimenopause was related to the frequency of acupuncture. In some prior research, a high frequency of acupuncture was shown to be more efficacious, whereas in this particular study, a low frequency demonstrated better clinical efficacy. This disparity indicates that the efficacy resulting from the weekly acupuncture frequency cannot be simply determined based on the frequency alone. In the future, the publication of RCT studies with a view to higher treatment will have more scientific standards for the classification of acupuncture frequency, which can provide higher quality clinical data and better clinical guidance for acupuncture treatment of perimenopausal women with mood disorders.

In this study, a total of 49 pieces of literature were incorporated, the majority of which were of relatively low quality. The existing problems were as follows: (i) In terms of bias risk, most of the literature did not explicitly mention the allocation concealment method; the implementation of blinding was rather general, and to clarify this, there were relatively few high-quality double-blind studies. This might lead to a decrease in the validity of the results. (ii) The age of the literature is relatively old, which might result in a decline in data quality, and the obsolescence of the results, etc. It is recommended that the future-related clinical studies should be further strict in design: (i) explain the use of random assignment hiding and blind method as much as possible; (ii) strengthen the follow-up investigation of cases to better judge the long-term curative effect; (iii) develop a rigorous and complete high-quality randomized double-blind controlled study plan to ensure the accuracy and objectivity of data and reduce the risk of bias during the study implementation, so as to provide higher quality clinical trial data. At the same time, the specific mechanism underlying the treatment of this disease by acupuncture, as well as the mechanisms of differential therapeutic effects during the early menopausal transition, late menopausal transition, and early postmenopausal transition, have not been explored in depth.

Although this meta-analysis provides comprehensive evidence on the effectiveness of acupuncture in treating mood disorders in perimenopausal women, it must be acknowledged that there are some limitations that may affect the interpretation and application of the results. (i) The heterogeneity among studies was evaluated using I² inconsistency tests, in which values of >50% were considered to indicate high heterogeneity. Forest plots showed that there was significant heterogeneity among HAMD, HAMA, SAS, KMI, and FSH, suggesting significant differences among the research results. This heterogeneity might be due to differences in research design, intervention measures, or population characteristics. (ii) Most studies were conducted in urban settings with specific patient demographics, limiting the generalizability of the findings to other populations. The retrieval of gray literature (such as conference papers, unpublished research, etc) may not be comprehensive enough, which might also affect the completeness of the results. (iii) Although we conducted subgroup analyses for all the outcome indicators, due to the limited data and the incomplete nature of the information extracted, the results of some subgroup analyses may not be sufficiently robust.

The research results indicate that acupuncture can be regarded as the preferred or supplementary treatment method for mood disorders in perimenopausal women. It plays a significant role in alleviating symptoms such as depression and anxiety, and has produced remarkable effects in improving other specific indicators (such as KMI, FSH, etc). Clinically, based on the results of the meta-analysis, acupuncture’s potential benefits and limitations can be explained to women with perimenopausal mood disorders, ensuring that patients fully understand the relevant treatment methods and are able to make treatment choices that align with their personal health goals. Based on the clinical application of acupuncture, we further explore the groups that can benefit from acupuncture therapy. Based on the clinical application of acupuncture, we further explore the groups that can benefit from acupuncture therapy.

Based on the clinical situation, we believe that there may be several types of patients who could benefit from acupuncture therapy: (i) Women who wish to adopt non-pharmacological treatments, such as mood disorders in perimenopausal women, may be concerned about the side effects of drugs or their intolerance to drug treatment, and thus prefer nondrug treatments. Acupuncture, as a natural therapy with fewer side effects, can be an option; (ii) Women who have not responded well to current Western drug therapies can consider acupuncture as a physical treatment method, providing them with new treatment options; (iii) Women with multiple perimenopausal symptoms. Many clinical drugs have limited efficacy. Acupuncture, as a traditional Chinese treatment method, adopts the “holistic concept,” assessing the overall condition of the person and conducting comprehensive treatment. It not only alleviates patients’ mood disorders but also has advantages in improving hormone levels.

In conclusion, we believe that acupuncture can provide superior improvement for mood disorders in perimenopausal women. Acupuncture has the capacity to alleviate female anxiety and depression, while also enhancing the manifestations of perimenopausal syndrome. Compared with western medicine treatment, acupuncture treatment is safer. However, the determination of an optimal acupuncture treatment protocol remains an area that necessitates further in-depth clinical research in the future. Our current level of clinical research is limited; there is an urgent need for more rigorously designed, high-quality RCTs in the ensuing period to authenticate the efficacy of acupuncture in the treatment of perimenopausal mood disorders.

Author contributions

Writing – original draft: Yixuan Li.

Writing – review & editing: Yixuan Li, Yujie Xu, Lu Yang, Han Wu, Xian Ma.