Effectiveness and Safety of Acupuncture for Perimenopausal Depression: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

Objective To determine the effectiveness and safety of acupuncture for perimenopausal depression. Methods We searched the Cochrane Central Register of Controlled Trials, PubMed, EMBASE, CNKI, VIP Citation Databases, Wan Fang, and online trial registries such as ClinicalTrials.gov for randomized controlled trials (RCTs) assessing the efficacy and safety of acupuncture for perimenopausal depression. Literature screening, data extraction, and determination of the risk of bias were performed by two researchers independently. The extracted data were pooled and meta-analyzed using RevMan5.3 software. Results In total, 16 RCTs covering 1311 patients were enrolled. Overall, the results showed that acupuncture was more effective in the treatment of perimenopausal depression than antidepressants (OR = 2.68, 95% CI (1.84, 3.90), P < 0.00001). Furthermore, HAMD scores in the manual acupuncture group and electroacupuncture group were lower than those of antidepressants (manual acupuncture vs. antidepressants (MD = −2.35, 95% CI (−2.93, −1.77), P < 0.00001) and electroacupuncture vs. antidepressants (MD = −1.2, 95% CI (−1.92, −0.48), P=0.001)). Data analysis revealed that the treatment effect of acupuncture was more stable than that of antidepressants (MD = −2.4, 95% CI (−3.37, −1.43), P < 0.00001). Moreover, acupuncture was safer than antidepressants based on the incidence of adverse events (OR = 0.23, 95% CI (0.1, 0.52), P=0.0004). But acupuncture has no effect on estrogen levels (P ≥ 0.05). Conclusions Acupuncture for perimenopausal depression is safe and effective. Moreover, it has more stable long-term effects than antidepressants and hormone replacement therapy (HRT). We recommend acupuncture as a clinical treatment of perimenopausal depression.


Introduction
Perimenopausal women are at a higher risk of developing depression and other undesirable changes than at any other stage of women's lives [1,2]. e prevalence of perimenopausal depression is estimated to range between 4.7% and 41.8% [3][4][5]. Perimenopausal women are known to undergo numerous physiological changes that negatively impact their health physically and mentally. e onset of perimenopause is characterized by feelings of depression, despair, and sometimes suicidal thoughts. It is frequently accompanied by dysfunction of the endocrine and autonomic nervous systems, leading to hot flashes, night sweats, sleeping difficulty, reduced sex drive, and fluctuations in weight, energy, and cognitive ability [6].
However, the underlying causes of high-incidence perimenopausal depression are still unclear [7]. An increasing body of evidence suggests that the elevated risk of depression in perimenopausal women is at least in part, due to hormonal fluctuations, although this theory remains controversial [8][9][10][11]. Predisposing factors include sociodemographic factors, psychosocial factors, onset of menopausal symptoms, as well as marital and family status [6]. Conventionally, perimenopausal depression is managed pharmacologically using antidepressants or through psychotherapy [6]. Unfortunately, these therapeutic options are not always effective. Antidepressants are frequently associated with numerous negative side effects including mouth dryness, fatigue, drowsiness, weight gain, and sexual dysfunction, thereby contributing to low patient compliance rates [12,13]. Hormone replacement therapy (HRT) has also been used in the management of perimenopausal depression, but the effectiveness of this strategy remains uncertain [14,15]. erefore, safer and more effective therapeutic options are therefore needed for the management of this condition.
As a promising and effective therapy, acupuncture is gaining widespread global acceptance as a therapeutic option for the management of various health conditions [16]. It has been reported that depression ranks second in the top 10 indications for acupuncture [17]. Moreover, multiple studies and clinical trials have documented the effectiveness of acupuncture in the treatment of depression [18][19][20][21]. It should, however, be noted that the physiological-based perimenopausal depression is characterized by wide hormonal fluctuations, therefore differing from general depression [6]. Recent reports from randomized controlled trials (RCTs) on acupuncture indicate that the use of acupuncture in the management of for perimenopausal depression has also risen gradually risen [22].
In this study, we carried out a systematic review of published RCTs on the use of acupuncture in the management of perimenopausal depression, aiming to better understand its benefits in the management of this condition relative to conventional treatment options.

Protocol and Registration.
is systematic review was registered in PROSPERO as CRD42018114506 and was carried out following a previously described protocol [22]. In brief, the systematic review and meta-analysis was done in accordance with the guidelines outlined in the Cochrane Handbook for Systematic Reviews of Intervention [23] and reported in accordance with PRISMA guidelines [24]. Because this was a review of published studies and did not directly involve patients, ethical clearance and patient consents were not sought.

Study Selection (Inclusion and Exclusion Criteria)
2.2.1. Types of Studies. In this study, we included all RCTs on the use of acupuncture in the management of perimenopausal depression, whether the trials were blinded or not.

Types of Participants.
Studies on perimenopausal women diagnosed with depression were included. In this study, the perimenopausal phase was defined using the 2012 criteria described by the North American Menopause Society [25]. Diagnosis of depression in the studies analyzed met at least one of the following criteria: ICD-10 [26], DSM-5 ™ [27], and CCMD-3 [28]. Studies that included patients with other serious diseases, such as heart disease, stroke, and severe gynecological diseases, were excluded from our analyses.

Types of Interventions.
Only studies in which the only intervention within the experimental group was acupuncture (manual or electroacupuncture) were included in our analyses. Studies in which the experimental group was subjected to ear acupuncture, moxibustion, or pharmacologic interventions were not included in our analyses. Studies were included in which the control group had been subjected to various interventions including pharmacologic agents, placebo, and sham acupuncture and treatment courses of greater than 2 weeks. Other conditions, including nursing care, were consistent between the experiment and control groups.

Types of Outcome Measures.
Our analysis evaluated primary and secondary outcomes of acupuncture. e primary outcome computed two measures: the effective rate and the Hamilton depression scale (HAMD) score. As previously described, a HAMD score reduction rate ≥25% was regarded as an effective treatment [29]. e lower the HAMD score, the higher effect the interventions had. e secondary outcome was measured by evaluating the levels of estrogen (FSH, E2, and LH), the HAMD score during follow-up period, and the incidence of adverse events.

Study Selection and Data Extraction.
Results from the literature search were evaluated by two reviewers (XX and JY Z), who independently determined which articles met the inclusion as described in Section 2.2. Any disagreements and conflicting opinions about which studies to include were resolved by consensus with a third independent reviewer (YX J). Next, XX and JY Z extracted and entered the data from the selected studies into a predefined data template.
e extracted data included the study features, information about participants, intervention to which participants were subjected, outcome measures, adverse outcomes, and participants follow-up.

Assessment of Risk of Bias and Quality of Evidences.
To minimize reviewer bias, two researchers, XX and YX W, carried out independent evaluations of all the selected studies using a previously described collaboration tool recommended by the Cochrane Handbook [23]. ere were six points that had to be evaluated: random allocation, 2 Evidence-Based Complementary and Alternative Medicine allocation concealment, blinding, incomplete outcome data, selective outcome reporting, and other biases. e disagreement was also resolved by discussion. Finally, funnel plots were used to evaluate publication bias in the more than 10 selected studies as described before [30].

Data Analyses
2.6.1. Date Synthesis. Meta-analysis and data synthesis were carried out using RevMan5.3.5 software from Cochrane collaboration (http://www.cochrane.org). For analyses of dichotomous variables, we used risk ratios (OR) and 95% confidence intervals (95% CI). For analyses of continuous variables, we used mean differences (MD) and 95% confidence. For continuous variables with different units, standardized mean differences (SMD) and 95% confidence intervals (95% CI) were used. P values of P < 0.05 were considered to be statistically significant.

Assessment of Heterogeneity.
e Chi-squared test and the I 2 statistic were used for the determination of heterogeneity. In cases of low heterogeneity, within the acceptable range of P > 0.10, I 2 < 50%, the fixed effect model was used for data analysis. Otherwise, the random effect model was used.

Subgroup Analysis and Sensitivity
Analysis. In order to evaluate the possible sources of heterogeneity, subgroup analysis and sensitivity analyses were conducted. Subgroup analyses looked at the possible factors that might contribute to heterogeneity, such as intervention in the experimental group, control group, the duration of treatment or the standards of the reports under consideration.

Risk of Bias in Included Studies.
Of the 16 RCTs we analyzed, 8 did not report their methods of generating Evidence-Based Complementary and Alternative Medicine random sequences, and only 2 trials reported that allocation process was hidden. Because the analyzed RCTs focused on acupuncture, it was not possible to blind the acupuncturists. Only one of the RCTs studies was blinded to patients and to data evaluators. For the other studies, we did bias risk assessments based on implementation methods reported in the respective articles. Only 6 trials reported the reasons for participant withdrawal and dropout. Together, these factors made it hard to establish the integrity or validity of the reported outcomes. e risk of selective reporting and other bias was not found ( Figure 2). Full-text articles excluded (n = 16) Repeat publication (n = 9) (1) No relevant outcomes (n = 5) (2) Combining with other diseases (n = 2)   (Figure 4). Because the heterogeneity analysis revealed low heterogeneity (I 2 � 9%, P � 0.35), the fixed effect model was adopted.

HAMD Scores of the Follow-Up Period. Among the 16
RCTs we analyzed, 5 reported follow-up analysis [32,34,36,42,44], while 4 reported HAMD scores [32,34,42,44]. Because follow-up was carried out for different durations, we based our analysis of follow-up on the data for the fourth week of follow-up in the respective studies. is analysis revealed that the HAMD scores of the acupuncture (experimental) groups were significantly lower than in the control groups (P < 0.05) (MD � − 2.4, 95% CI (− 3.37, − 1.43), P < 0.00001). Because meta-analysis revealed mild heterogeneity between the 4 trials (I 2 � 35%, P � 0.02) (an acceptable level of heterogeneity), the fixed effect model was not adopted (Figure 6).

Safety of Acupuncture.
Of the 16 RCTs that we analyzed, 1 reported the treatment emergent symptom scale (TESS) score [38], while 9 others reported adverse events [32-34, 36, 39, 40, 44-46]. As shown in Table 2, none of the studies reported serious adverse outcomes in the acupuncture (experimental) group or the control. However, it was difficult to estimate adverse outcomes in 2 trials [33,34], and hence they excluded from further analysis. is analysis revealed that the incidence rate of adverse outcomes in patients treated with acupuncture was significantly lower than in the group treated with pharmacological agents (P < 0.05) (OR � 0.23, 95% CI (0.1, 0.52), P � 0.0004). Because meta-analysis revealed moderate heterogeneity (I 2 � 61%, P � 0.01) between the studies, the randomized control model was adopted (Figure 7).

Publication Bias.
Because our review and meta-analysis looked at more at more than 10 RCTs, we conducted a publication bias analysis by visual analysis of funnel plots. is analysis produced an almost symmetrical funnel plot distribution, indicating that there was no publication bias (Figures 8 and 9).

Summary of the Main Results.
Here, we report a metaanalysis of 16 RCTs that together, comprising 1311 Evidence-Based Complementary and Alternative Medicine participants. Our analysis of both the effective rate and HAMD score revealed that acupuncture achieved better outcomes in the management of perimenopausal depression-relative antidepressants and HRT. Additionally, followup analysis indicates that the benefits of acupuncture against perimenopausal depression last longer those of antidepressants. However, measurements of the levels of estrogenic hormones did not reveal statistically significant differences in those treated by acupuncture relative to pharmacologic antidepressants. Further research is therefore warranted to establish whether the observed acupunctural benefits are associated with changes in the levels of estrogen [47]. In traditional Chinese medicine (TCM), it is theorized that abnormal flow of Qi in a meridian is related to disease and that its natural flow is restorable through acupuncture at appropriate acupoints [48]. Multiple studies suggest that acupuncture has a modulatory, normalizing effect on the limbic-paralimbic-neocortical network (LPNN) of the brain     Evidence-Based Complementary and Alternative Medicine [49][50][51][52]. It is generally believed that the LPNN is related to sleep and emotions. Other studies also report that the antidepressant effect of acupuncture is significant and has a multitarget characteristic [53]. Acupuncture treats nervous system diseases by increasing the brain-derived neurotrophic factor level and involving multiple signal pathways, which may be related to amino acid metabolism and inflammatory pathways, especially the toll-like receptor signaling pathway [54].
Our meta-analysis suggests that, in the management of perimenopausal depression, acupuncture is relatively safer than pharmacological agents. e main adverse effects following acupuncture include subcutaneous bleeding and pain following the piercing. ese side effects can be reduced by choosing different acupoints. Additionally, patients with fear of needles experience needle sickness, which is characterized by anxiety, dizziness, palpitations, and sweating. ese side effects are temporary and can be reduced by training acupuncturists to psychologically prepare patients prior to starting the procedure. On the contrary, negative side effects of antidepressants are related to the respective drugs' mechanisms of action and most cannot be avoidable. It should, however, be noted that the risks associated with acupuncture could not be assessed with confidence as many of the RCTs in our study did not adequately report adverse effects of their interventions. Our analysis of heterogeneity between the RCT studies we reviewed, revealed mild heterogeneity (HAMD scores of therapeutic effect: I 2 � 36%, P � 0.08). Subsequent subgroup analysis, in which interventions were separated into manual acupuncture and electroacupuncture, exposed a markedly lower level of heterogeneity between the studies. Mild heterogeneity was also observed following assessment of the effectiveness of treatment during the follow-up period (I 2 � 35%, P � 0.02). e underlying causes of heterogeneity might be caused by differences in interventions. In addition,  Evidence-Based Complementary and Alternative Medicine the HAMD score is a subjective metric, and results are easily influenced by artificial errors. A safety evaluation revealed moderate heterogeneity (I 2 � 61%, P � 0.01), and the underlying causes might reflect the assessment method used and/or acupuncturist's professional skill.

Characteristics and Limitations.
We are aware of three previous systematic reviews (two in Chinese and one in English) addressing this topic [55][56][57]. However, these studies did not carry out meta-analysis, and the inclusion criteria were not sufficiently rigorous, causing high levels of heterogeneity between studies. is has made it difficult to accurately evaluate the safety and effectiveness of acupuncture. Unlike previous systematic reviews, we evaluated more recent clinical trials. In addition, we applied stricter inclusion and exclusion criteria, used widely accepted analyses tools, and performed meta-analysis. In order to limit our analysis to the use of body acupuncture in the management of perimenopausal depression, we limited the intervention in the experimental group to manual acupuncture and electroacupuncture only and excluded the studies on ear acupuncture, moxibustion, or pharmacologic interventions. We also evaluated the durability of acupunctural benefits over the follow-up period as well as the effects of the length treatment period. Moreover, we evaluated the safety of perimenopausal interventions by computing the incidence rate of adverse effects. e RCTs included in our analyses had several shortcomings including the generation of random numbers, distribution hiding, and blind methods. Of the studies we evaluated, only one was registered in ClinicalTrials.gov [32]. It was therefore difficult to estimate the risk of bias in this study. All the included studies were conducted in China. Moreover, we did not carry out manual retrieve, and the collection of gray literature might be insufficient.  Participants in the control groups of the RCTs we analyzed had been treated with antidepressants. None of the studies that we evaluated reported the use of placebo or sham acupuncture, limiting our ability to more confidently evaluate the effectiveness of acupuncture. However, this is a defect in the design of the RCTs we included, and it is also a place that needs to be improved in future clinical trials.

Conclusion.
Our analysis indicates that the use of acupuncture in the management of perimenopausal depression is safe, effective, and offers long-lasting benefits relative to antidepressants and HRT. Our data indicate that acupuncture is an effective option for the clinical management of perimenopausal depression. Because the existing studies are not of sufficiently high quality, higher quality RCTs are required to establish the clinical benefits and longterm effectiveness of acupuncture in the treatment of perimenopausal depression.

Abbreviations
HAMD: Hamilton depression scale OR: Risk ratio MD: Mean difference SMD: Standardized mean difference CI: Confidence interval. Evidence-Based Complementary and Alternative Medicine 11