Adjuvant Therapy of Oral Chinese Herbal Medicine for Menopausal Depression: A Systematic Review and Meta-Analysis

Objective. The aim of this meta-analysis was to evaluate the effectiveness of oral Chinese herbal medicine (OCHM) combined with pharmacotherapy for menopausal depression. Methods. The electronic databases were searched from their inception to December 25, 2016, comprising PubMed, Embase, Cochrane Central Register of Controlled Trials, Chinese National Knowledge Infrastructure (CNKI), Wanfang database, and Chinese Biomedical (CBM) database. Randomized controlled trials investigating the effectiveness of OCHM combined with pharmacotherapy for the people with menopausal depression were eligible. Risk of bias was evaluated according to the Cochrane handbook. Meta-analyses were performed to pool the effect size. Heterogeneity and publication bias were also examined. Results. Twenty-two RCTs with 1770 participants were included in the review. None of the studies used placebo as the control and the risk of bias was high in blinding the participants and personnel. Overall, the meta-analysis demonstrated that adjuvant therapy of OCHM was effective in reducing the Hamilton Rating Scale for Depression (HAMD) scores compared to pharmacotherapy (MD = −3.75; 95% CI = −5.22, −2.29; P < 0.00001). The meta-analysis also suggested that OCHM adjuvant therapy for menopausal depression was superior to pharmacotherapy in terms of response rate of reducing HAMD scores (RR = 1.17; 95% CI = 1.10, 1.25; I2 = 55%). Conclusions. OCHM may provide additional effectiveness to pharmacotherapy for the people with menopausal depression. RCTs including the placebo control were required to further determine the additional efficacy of OCHM for menopausal depression.


Introduction
Depression is the most common mental disorder in menopausal women, with a prevalence ranging from 26% to 41.8% [1][2][3][4][5]. Menopause represents a significant transition in the life of a woman, and it has been considered as a specific window of vulnerability to depression [6,7]. Women with menopausal depression are related to decreased quality of life, increased cardiovascular disease, and metabolic syndrome [8][9][10][11].
Antidepressants are considered the most common pharmacotherapy specifically for menopausal depression. However, the effect of antidepressants is unsatisfactory clinically, and the long-term use leads to considerable adverse effects. For example, selective serotonin reuptake inhibitors (SSRIs) are associated with sexual dysfunction, weight gain, and sleep disturbance [12][13][14]. Hormone therapy (HT) is widely used to alleviate menopause-related symptoms, and it has also been used in the treatment of menopausal depression [15,16]. However, its effectiveness was inclusive [17]. Therefore, alternative therapies with more benefits and fewer harms were in continuous demand.
Traditional Chinese medicine (TCM) is one of the oldest medicine systems in the world and has been widely used as a form of complementary and alternative medicine [18,19]. In China, oral Chinese herbal medicine (OCHM) is commonly used in combination with pharmacotherapy for menopausal depression and the associated clinical studies have been conducted [20,21]. However, most of the clinical 2 Evidence-Based Complementary and Alternative Medicine studies were of insufficient sample size and of varied quality of methodological design. Systematic reviews on OCHM for depression [22,23] and menopausal-related symptoms have already been published [24]. However, previous systematic reviews did not specifically evaluate the effectiveness of OCHM as an adjunctive therapy to pharmacotherapy for menopausal depression.
This systematic review was motivated by the large number of published clinical trials on OCHM combined with pharmacotherapy and the unresolved problems of pharmacotherapy. Our objective was to evaluate the effectiveness of OCHM combined with pharmacotherapy for menopausal depression.

Search Strategy.
We conducted and reported the systematic review according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement guidelines [25]. The electronic databases were searched from their inception to December 25, 2016, comprising PubMed, EMBASE, Cochrane Central Register of Controlled Trials, Chinese National Knowledge Infrastructure (CNKI), Wan Fang Database, Chinese Scientific Journal Database (VIP), and Chinese Biomedical (CBM) database. The combination of search terms was presence of menopause, depression, traditional Chinese medicine, and randomized controlled trial (RCT). There was no limit to the language or publication status. The search strategy is attached in the Appendix.

Eligibility Criteria
(1) Types of Studies. Only randomized controlled trials were eligible.
(2) Types of Participants. Menopausal women aged 40 to 60 years with depressive symptoms were eligible. Depression was diagnosed according to the Classification and Diagnosis of Mental Diseases (CCMD-3) (Psychiatry Branch of Chinese Medical Association) [26], the Diagnostic and Statistical Manual (DSM-IV) (American Psychiatric Association) [27,28], and the International Classification of Disease (ICD-10, WHO) (World Health Organization) [29].
(3) Types of Intervention and Control. Studies that compared OCHM plus pharmacotherapy to the same pharmacotherapy were included. All forms of OCHM (i.e., decoctions, tables, capsules, pills, and powders) were included. The pharmacotherapy included conventional antidepressants (paroxetine hydrochloride, fluoxetine hydrochloride, deanxit, and venlafaxine), HT (estradiol valerate tablets and tibolone), and a combination of antidepressants and HT. The treatment duration had to last for three weeks at least.
(4) Types of Outcome Measurements. The primary outcome was the total scores of Hamilton Rating Scale for Depression (HAMD). Clinically, HAMD scores have been commonly used to assess the severity of depression [30]. The secondary outcome was the response rate by reducing the scores of HAMD, Kupperman Menopausal Index (KMI), and adverse events (AEs).

Data Extraction and Quality Assessment.
Titles and abstracts of all retrieved studies were screened, and then full texts were reviewed for eligibility by two authors (Jiju Wang and Jian Liu) independently. The data was collected by using the predefined data extraction form, including the journal title, first author, year of publication, size of study, baseline characteristics of women (e.g., age, course of disease), methodological design, intervention strategy, treatment duration, and outcomes. The dataset was validated by the third reviewer (Xiaouyn Wang). Two reviewers independently evaluated the risk of bias by using the Cochrane Collaboration's tool [31]. Each domain of the bias was classified as a "low risk", "unclear", or "high risk".

Data Synthesis.
Meta-analysis was performed using Review Manager (the Cochrane Collaboration) software, Version 5.3.0. For dichotomous data, relative risk ratio (RR) with 95 % confidence intervals (CIs) was used to present the therapeutic effect; for continuous data, mean difference (MD) with 95 % CIs was used. The chi-square test and the Higgins I 2 test were used to assess heterogeneity. If heterogeneity was low (I2 < 50% or P > 0.1), the fixed effects model was used. If heterogeneity was high ( 2 > 50% or < 0.1), the random effect model was used and further analyses such as subgroup or sensitivity analysis were planned to clarify the source of heterogeneity.
Publication bias was visualized by the funnel chart. Begg's correlation test and Egger's intercept test were used to quantify the publication bias, with a significant level at 0.05.

Study Selection. Databases search found 1076 articles.
After duplicates were removed, 761 articles were screened by reading the title/abstract and 628 ineligible articles were removed. After reading the full texts of the remaining 133 articles, 22 trials  were finally included in the systematic review. The flowchart summarizes the screening process ( Figure 1).

Study Characteristics.
All included trials were conducted in China and they were published ranging from 2005 to 2016. The total sample size was 1777, including 901 women in the experimental group and 876 women in the control group. The participants ranged between 40 and 60 years. The most common diagnostic instrument was CCMD-3, and the combination of CCMD-3 and HAMD score was usually used as the inclusion criterion of RCTs. For outcome measurements, eighteen trials reported HAMD score [32-34, 36-40, 42-46, 49-53], eighteen calculated response rate by reducing HAMD scores [33, 35-39, 41-49, 51-53], and three trials reported KMI score [39,42,45]. AEs were reported in 10 trails [32, 37-39, 41, 43, 45, 46, 51]. The basic characteristics of the included studies are summarized in Table 1. Four forms of OCHM were investigated, consisting of decoctions, capsule, granule, and powder. Twenty-two formulas of OCHM were found. Bupleuri Radix (Chinese name: Chaihu) and Paeoniae Radix Alba (Chinese name: Baishao) were the herbs with top frequency. Details of OCHM in the included studies are summarized in Table 2. Control group included three categories of pharmacologic agents, antidepressants in 13 trials [32, 37-41, 44-49, 52], HT in 2 trials [42,53], and the combination of antidepressants and HT in 6 trials [33-36, 50, 51]. One trial did not specify the pharmacotherapy [43]. The treatment duration ranged from three to 12 weeks. Figures  2 and 3. For random sequence generation, six trials used random number tables [32,39,42,45,46,52], and 13 trials [33-35, 37, 38, 40, 41, 43, 47-50, 53] did not provide details of randomization. Three trials considered the order of visits as random, which were of high risks in selection bias [36,44,51]. None of the trials reported allocation concealment. Binding of participants, personnel, and outcome assessors was not applied in any studies. None of the studies had attribution bias. The risk of bias in selective reporting was unclear as none of the studies published their protocols.

Kupperman Menopausal Index (KMI) Scores. Three
RCTs [39,42,45] measured the effectiveness of OCHM for menopausal depression by using KMI scores. Meta-analysis showed OCHM combined with pharmacotherapy was more 4 Evidence-Based Complementary and Alternative Medicine

Publication
Bias. The funnel plots of HAMD scores and the response rate were asymmetrical (Figures 8 and 9). The regression analysis of Begg's rank correlation test and Egger's intercept test was statistically significant ( > 0.05), suggesting the presence of publication bias. The detection of publication bias was not available for other outcomes as the included studies were less than 10.

Statement of Principal Findings.
In the present study, we reviewed 22 RCTs involving a total of 1777 participants and assessed the add-on effects and safety of OCHM to pharmacotherapy in women with menopausal depression. Metaanalysis showed that the combination of OCHM and pharmacotherapy was more effective in improving menopausal depression. In addition, the incidence of AEs in the participants treated with OCHM adjunctive therapy was less than those without OCHM adjunctive therapy. However, the risk of bias of the included RCTs may affect the evidence certainty; particularly most studies did not blind the participants or personnel. hypothalamic pituitary ovarian (HPO) axis [54,55]. OCHM can also alleviate hippocampal neuron damage, inhibit early apoptosis of neurons, and increase the content of monoamine transmitters in brain tissue through the CREB-BDNF signaling pathway [56]. Furthermore, OCHM can regulate the hypothalamus pituitary adrenal (HPA) axis in menopausal depression rats and improve the behavior of model rats [57]. Bupleuri Radix (Chinese name: Chaihu) and Paeoniae Radix Alba (Chinese name: Baishao) were the herbs with top frequency. Bupleurum-saikoside, the main active ingredient of Bupleuri Radix, improved depression by regulating the monoamine neurotransmitters and BDNF in the brain [58].  Paeoniae Radix Alba improved depression by increasing the single amine neurotransmitter and adjusting the dysfunction of HPA axis [59].

Limitations and Implications of the Research.
There were several limitations in this study. Firstly, the quality of the included trials was generally poor. The risks of selection bias and reporting bias were unclear most of the time. And the risk of performance bias in almost all the RCTs was high. These biases may affect the results of meta-analysis. Secondly, the heterogeneity was observed across meta-analyses and it was not resolved by subgroup analysis. This can also reduce the evidence certainty. Thirdly, none of the RCTs included considered placebo as the control. Hence the current studies were unable to conclude the efficacy of OCHM. Although it is difficult to successfully produce a placebo to CHM because of its special characteristics such as the appearance, smell, and taste, recent progress of placebo making can be applied. For example, when researchers explored the preparation method of placebo to Moron Dan, they found that the flavor      characteristics and disintegration of the placebo made of soybean powder 100g, starch 100g, carbon black pigment 1g, and honey 70g were similar to the experimental drug [60]. Some researchers suggested that encrypted capsules could be also used as a placebo [61].

Conclusion
Adjuvant therapy of OCHM provided additional benefits to pharmacotherapy in the people with menopausal depression. More RCTs with a rigorous design, particularly applying placebo as the control as well as blinding the participants and personnel, are needed to confirm the efficacy of OCHM for menopausal depression.