Chinese Herbal Medicine for Postpartum Depression: A Systematic Review of Randomized Controlled Trials

Background. Postpartum depression (PPD) does great harm to women following childbirth. The aim of this study was to conduct a systematic review of the literature to assess the efficacy and safety of CHM for the treatment of PPD. Methods. Published or ongoing registered trials were searched for from the inception of the various databases to December 31, 2015. Data extraction and methodology assessment were conducted independently by two researchers. RevMan 5.3 software was used to analyze the data. Results. Forty-seven registered clinical trials (RCTs) were identified and reviewed. The results showed CHM alone or in combination with routine treatments could reduce HAMD score, EPDS score, incidence of adverse events, TESS, and SERS. CHM combined with routine treatment was more effective in increasing serum estradiol levels and reducing progesterone levels than routine treatment alone. Meanwhile, pooled data revealed that MRLQS combined with routine treatments or MRLQS plus MSHS combined with routine treatments were more effective than other therapeutic methods in TCM. MRLQS plus MSHS alone was found to be an effective alternative when compared to routine treatments. Conclusions. This review suggested that CHM was safe and effective in the treatment of PPD. However, this could not be proven conclusively. To ensure evidence-based clinical practice, more rigorously designed trials are warranted.


Background
Childbirth is an important experience for women, during which great physical and psychological changes can occur.Postpartum depression (PPD) is a common complication of childbirth, mainly affecting a woman's mentality during the first postpartum year [1].It is accompanied by changes in thought and behavior and physical symptoms as well.It may negatively affect a woman's quality of life, limit her capacity to fulfill her maternal role, impair mother-infant bonding, and strain relationship with her partner.It even has negative influences on the child's emotional, cognitive, social, and behavioral developments [2][3][4][5].The prevalence of PPD was estimated to be much greater than 10%-15% [1].Some studies have reported that 60% or more of mothers of newborns were affected in certain areas [6,7].
The etiology of PPD is multifactorial.Psychological and sociological changes are important risk factors.A personal and family history of depression, poor marital relationship, poor social support, stressful life events, low socioeconomic status, unplanned/unwanted pregnancy, and regional culture may increase the risk for this disease [8][9][10].In addition, the reported trigger for the development of PPD refers to biological changes, such as estrogen and progesterone balance, hypothalamic-pituitary-adrenal axis, serotonergic neurotransmitter system, and certain genetic variations [11][12][13][14][15].
Due to differing causative factors, the treatments for PPD are diverse.Selective serotonin reuptake inhibitors (SSRIs) and serotonin-norepinephrine reuptake inhibitors (SNRIs) are antidepressants used as first-line agents.Hormone treatments are considered to be an alternative method 2 Evidence-Based Complementary and Alternative Medicine of pharmacotherapy.Psychotherapeutic and psychosocial interventions are the important nonpharmacotherapeutic methods used in clinics.Although various therapeutic approaches are available, sufficient evidence is not available to justify the use of any one of these approaches for PPD in clinical practice [1,[16][17][18][19].
Chinese herbal medicine (CHM), which is the most important component of traditional Chinese medicine (TCM), is widely used in China and is increasingly being used worldwide.According to TCM theory, the etiology of PPD mainly includes stagnation of the liver, stasis of the blood, deficiency of liver blood, phlegm-dampness, deficiency of kidney yin, deficiency of heart yin and blood, and deficiency in the heart and spleen [20][21][22][23][24][25][26].The clinical treatments of TCM should conduct syndrome differentiation (bian zheng) and project corresponding therapeutic methods to prescription Chinese medicine formula according to analysis of etiology and syndrome.Many clinical trials involving CHM for the treatment of PPD have been conducted and report promising results.Although a meta-analysis (involving 9 trials and 716 patients) based on integrated traditional Chinese and Western medicine for the treatment of PPD has been reported in a Chinese journal [27], its applicability is limited due to its focus on integrative medicine, and only the overall response rate was analyzed in the literature.A comprehensive summary of the safety and efficacy of CHM for the treatment of PPD does not currently exist.Therefore, we systematically reviewed the literature to assess this aspect and to justify the use of CHM for the treatment of PPD in clinical practice.

Methods
The protocol used for this systematic review was published in the PROSPERO database (http://www.crd.york.ac.uk/PROSPERO/display record.asp?ID=CRD42014013258).
2.1.Eligibility Criteria.Randomized controlled trials (RCTs) testing the effectiveness of CHM in the treatment of PPD in women were included.There was no limitation on language, blinding, or publication format.Quasi-randomized trials were excluded.A clear description of the diagnostic criteria used must have been included.The patients included in this review were at least 16 years of age.CHM included the use of a single herb, Chinese patent medicine, practitioner-prescribed herbal formula, or products extracted from natural herbs.There was no limitation on dosage, administration regimen, or duration of treatment.The control could be a placebo, no treatment, or routine treatments.Trials evaluating the use of CHM combined with routine treatments versus the routine treatments alone were also included.The primary outcome measured was improvement in depression at the end of treatment or at the time of follow-up.Valid assessment tools included the Hamilton Depression Rating Scale (HAMD), Edinburgh Postnatal Depression Score (EPDS), Self-Rating Depression Scale (SDS), Beck Depression Inventory (BDI), Postpartum Depression Screening Scale (PDSS), and adverse events associated with CHM.Additionally, some studies measured changes in the levels of estradiol and progesterone.

Search Strategy and Study Selection. Two authors (Zijie
Chen and Ning Yu) conducted a systematic search for published, unpublished, and ongoing trials.The Cochrane Central Register of Controlled Trials (CENTRAL), PubMed, Excerpta Medica database (EMBASE), Chinese Biomedical Literature Database (CBM), Chinese National Knowledge Infrastructure Database (CNKI), Chinese Science and Technology Periodical Database (VIP), and Wan Fang Database were searched from the date of their inception until December 31, 2015.The World Health Organization (WHO) International Clinical Trial Registry Platform (ICTRP) portal and the website of International Clinical Trial Registry by US National Institutes of Health (http://clinicaltrials.gov/) were searched for ongoing registered clinical trials.
The following search terms were used individually or combined: "maternal depression", "post partum depression", "postpartum depression", "puerperium depression", "postnatal depression", "Chinese traditional", "Chinese herb", "Oriental traditional", "herb", "herbal medicine", and "random".There were no restrictions on language, publication year, and publication status.The details were shown in Supplementary Table 1 in Supplementary Material available online at http://dx.doi.org/10.1155/2016/5284234.Titles and abstracts were screened to identify trials that potentially met the criteria of this study.The full texts of selected trials were reviewed by two authors (Zijie Chen and Ning Yu) to determine eligibility for inclusion.The reference lists for all eligible articles were perused to collect additional trials.Any discrepancies were resolved by consensus and discussion with a third party (Shuangqing Zhai) if necessary.

Data Extraction.
Data information was extracted independently by two authors (Keyu Yao and Yupeng Xi).A predesigned structured data extraction form was used for the eligible trials, which mainly included publication year, study type, random sequence generation, allocation concealment, blinding, diagnostic criteria, sample size, baseline characteristics of the participants, type of syndrome differentiation, therapeutic methods of TCM, intervention, control, outcomes, follow-up, and adverse events.The authors of the trials were contacted if any information was unclear or missing.Disagreements were resolved by discussion and consensus was reached with the help of a third party (Shuangqing Zhai).

Assessment of Risk of Bias.
The risk of bias in eligible trials was assessed independently by two authors (Keyu Yao and Yupeng Xi) according to the criteria described in the Cochrane Handbook version 5.1.0[73].The criteria mainly included random method, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other sources of bias.Each trial was classified as having a low, unclear, or high risk of bias.Any discrepancies were settled by discussion with a third party (Shuangqing Zhai).If necessary, the author of the trial was contacted to clarify any ambiguities.
Evidence-Based Complementary and Alternative Medicine 3 2.5.Data Analysis.RevMan 5.3 software was used to analyze the data.If the data were available, the following comparisons were performed: CHM versus no treatments, CHM versus placebo, and CHM with or without routine treatments versus routine treatments alone.The diversity of therapeutic methods used in TCM may be an important reason for the high heterogeneity in the interventions, and the subgroups were further analyzed in this review.A meta-analysis was performed if homogeneity in study design, participants, intervention, control, and outcome measures was acceptable.A random-effect model was used for the meta-analysis if the trials had significant heterogeneity ( 2 > 50%) [74].Otherwise, a fixed-effect model was used.The effect estimate was expressed as risk ratio (RR) with a 95% confidence interval (CI) for categorical data and as mean difference (MD) with a 95% CI for continuous data.Funnel plots were generated to detect publication bias when more than ten trials were identified [74].

Description of Studies.
The initial search yielded 17,640 records from seven electronic databases: CNKI ( = 2,237), CBM ( = 4,302), VIP ( = 5,245), Wan Fang ( = 5,853), PubMed ( = 3), CENTRAL ( = 0), and EMBASE ( = 0).After removing 9,132 records, which were duplicates from the different databases, 8,508 citations were screened.8,289 records were excluded for various reasons determined while reading the titles and abstracts.Full text of the remaining 219 articles was retrieved and assessed in detail for eligibility.An additional 172 trials were excluded due to inappropriate interventions, diagnostic uncertainty, basic/theory researches, review, duplicate publication, absence of randomized controlled trials, or nonavailability of outcomes.Ultimately, 47 RCT articles [21,26, met the inclusion criteria and were included in this systematic review.A flow chart detailing the search process and study selection is displayed in Figure 1.We also searched trial registries that identified only one ongoing clinical trial (http://clinicaltrials.gov/).However the trial (ClinicalTrials.govIdentifier: NCT01178008) focused on acupuncture for the treatment of PPD and did not meet our criteria for inclusion.

Methodological Quality.
All the trials provided limited information about design and methodology.The randomized allocation of participants among the groups was mentioned in all trials, but only 20 trials [21,26,30,33,35,38,41,45,49,50,52,54,57,60,63,65,67,68,70,71] described the specific methods used for sequence generation.These included random number table, drawing of lots, and randomized card law.Allocation concealment was only reported in one trial [57]; however it could not be judged whether this was properly conducted because of insufficient information.Only one trial [63] was of double-blind design, and the rest of trials did not provide any information regarding blinding.Information detailing withdrawal from the study was provided in 13 trials [29,37,39,43,45,47,52,54,57,63,67,68,72]; of these, 3 trials [45,47,67] reported that no participants dropped out, 3 trials [52,57,68] provided withdrawal reasons, and only one trial [57] conducted intention-to-treat analysis.To assess selective reporting, we searched the included registration information using trial registries.Registration information and the protocol used for the study were not accessible for all trials.Therefore, we assessed selective reporting by comparing the outcome measures mentioned in materials  and methods section with the outcomes reported in the results.20 trials [26,28,33,36,38,41,42,44,47,48,50,54,57,58,60,62,64,67,68,72] were evaluated as low risk because all of the reported outcomes were described in the methods, 4 trials [45,46,63,66] were evaluated as high risk because the outcomes reported in the results did not match prespecified outcomes, and 23 trials [29-32, 34, 35, 37, 39, 40, 43, 49, 51-56, 59, 61, 65, 67, 69-71] were evaluated as unclear because no description of outcomes was provided in the materials and methods section.Only one trial [68] conducted a pretrial estimation of required sample size.Consequently, all of the trials had a high risk of bias and were assessed to be of low quality (Figure 2 and Supplementary Figure 1).

EPDS Scores.
Six trials [28,41,42,48,57,68] provided data on EPDS score changes after treatment.Only one trial [57] compared CHM using MSHS with a placebo.The study results showed that MSHS was more effective than the placebo (MD −2.67, 95% CI −3.88 to −1.46) in reducing EPDS score.Two trials [42,68] which compared CHM with routine treatments conducted a random effects model for metaanalysis because of statistical heterogeneity of the results ( 2 = 90%).The pooled results showed significant beneficial effects of CHM as compared to routine treatments (MD −3.36, 95% CI −6.05 to −0.66).Subgroup analysis in these trials revealed that the use of MTQB plus MBASD (1 trial) [68] or MRLQS plus MSHS plus MBASD (1 trial) [42] resulted in a statistically significant reduction in EPDS score when compared with routine treatments alone.Three trials [28,41,48] compared CHM combined with routine treatments with routine treatments alone.These trials used MRLQS plus MSHS as the therapeutic approaches of CHM and used a random effects model for pooled analysis because of statistical heterogeneity ( 2 = 75%).The pooled result showed that CHM (MRLQS plus MSHS) combined with routine treatments had superior effectiveness when compared to routine treatments alone (MD −3.80, 95% CI −5.27 to −2.34) in reducing EPDS score (shown in Table 3).

Serum Estradiol.
Four trials [47,57,59,63] reported the effect on serum estradiol of CHM alone or in combination with routine treatments.Two trials [57,63] compared the effect of CHM with a placebo and used a random effects model for meta-analysis because of statistical heterogeneity ( 2 = 88%).The pooled result showed no statistically significant differences between the groups (MD 29.01, 95% CI −2.54 Evidence-Based Complementary and Alternative Medicine 13   to 60.56) in increasing serum estradiol.Subgroup analysis showed that MRLQS plus MTQB (1 trial) [63] was more effective than routine treatments alone.No significant difference was observed in another trial [57] that compared MSHS with a placebo.In comparing CHM combined with routine treatments to routine treatments alone, the data from 2 trials [47,59] was pooled and a fixed effects model was used because of no statistical heterogeneity ( 2 = 11%).The results revealed that CHM combined with routine treatments had superior effectiveness when compared with routine treatments alone (MD 38.36, 95% CI 35.38 to 41.33).The subgroup analysis revealed that MSHS combined with routine treatments (1 trial) [47] or MTQB combined with routine treatments (1 trial) [59] was more effective than routine treatments alone in increasing serum estradiol (shown in Table 3).

Progesterone.
Four trials [47,57,59,63] reported progesterone levels in patients receiving CHM alone or in combination with routine treatments.Two trials [57,63] compared effect of CHM to a placebo, and the data was pooled with a fixed effects model being used because of no statistical heterogeneity ( 2 = 0%).The results showed that CHM was more effective than a placebo in decreasing progesterone (MD −2.11, 95% CI −3.08 to −1.14).The subgroup analysis showed that MRLQS plus MTQB (1 trial) [63] was more effective than routine treatments alone, and no significant differences were observed in another trial [57] when MSHS was compared to a placebo.In comparing CHM combined with routine treatments to routine treatments alone, a random effects model was used for two pooled trials [47,59] because of statistical heterogeneity ( 2 = 88%).The results showed that CHM combined with routine treatments was more effective than routine treatments alone (MD −8.14, 95% CI −15.70 to −0.58).The subgroup analysis showed that MTQB combined with routine treatments (1 trial) [59] resulted in a statistically significant reduction in progesterone as compared to routine treatments alone.There were no significant differences observed in another trial [47] when MSHS was compared with a placebo (shown in Table 3).
Two trials [29,52] reported no adverse events occurring in CHM combined with routine treatments groups, and 2 trials [56,68] [26,33,37,39,61,64] utilized TESS, and 2 trials [38,44] utilized SERS to evaluate the adverse events.For better evaluation of the safety of TCM in the treatment of PPD, related data were analyzed.
Meanwhile, one trial [64] utilized TESS to compare MRLQS plus MSHS with routine treatments alone.The results showed that MRLQS plus MSHS was more effective than routine treatments alone.Five trials [33,37,39,61,69] compared CHM combined with routine treatments to routine treatments alone utilizing TESS.A random effects model was used to pool data because of statistical heterogeneity ( 2 = 71%).The results showed that CHM combined with routine treatments had superior effectiveness to routine treatments alone in reducing TESS (MD −1.80, 95% CI −2.46 to −1.14).Subgroup analysis of two trials [37,61] showed that MRLQS combined with routine treatments had superior effectiveness to routine treatments alone (MD −2.31, 95% CI −3.12 to −1.50).MTQB combined with routine treatments (1 trial) [39], MRLQS plus MTQB combined with routine treatments (1 trial) [69], and MRLQS plus MBASD combined with routine treatments (1 trial) [33] were all shown to be more effective than routine treatments alone.A pooled analysis of 2 trials [38,44] revealed that MRLQS combined with routine treatments had superior effectiveness to routine treatments alone in reducing SERS (MD −5.19, 95% CI −9.73 to −0.64) (shown in Table 3).

Publication Bias Assessment.
A funnel plot for HAMD score following treatment with MRLQS combined with routine treatments in the experimental group and routine treatments alone in the control group was conducted to investigate the publication bias.The result indicated that publication bias may have existed in this review (Figure 3).

Discussion
4.1.Principal Findings.PPD has become a public health problem concern, as it is harmful to postpartum women and their families.Women have an important responsibility in the parenting of newborns.Therefore, selecting a safe and effective method of treatment is of particular importance.This study conducted a meta-analysis to assess the safety and efficacy of treatments for PPD using the therapeutic methods of TCM.
This review evaluated 47 trials and the pooled results demonstrated that CHM treatments alone or in combination with routine treatments were effective in decreasing HAMD scores, EPDS scores, and incidence of adverse events.It indicated that CHM may be a safer and more effective treatment for PPD than routine treatments alone.A subgroup analysis was also conducted because the diversity of therapeutic methods of TCM used was an important factor influencing clinical effect.HAMD score was considered to be an important tool for evaluating the clinical efficacy of PPD.CHM plus routine treatments (76.60%, 36/47) was the most common treatment approach.According to pooled data from related trials, the results suggest that MRLQS combined with routine treatments is more effective than MRLQS plus MSHS combined with routine treatments or MTQB combined with routine treatments in reducing HAMD score when each is compared with routine treatments alone.In this review, we also observed that MRLQS plus MSHS alone had a better clinical efficacy than routine treatments alone for PPD.However, it was difficult to evaluate other therapeutic methods of TCM individually or in combination with routine treatments because the data could not be pooled for analysis.
In the TCM theory, depression belongs to the category of emotion and will diseases.It is mainly caused by emotional triggers and viscera weakness.Stagnation of liver-qi caused by multiple etiologies is the most important mechanism of depression because of the role of the liver in regulating emotion.Meanwhile, postpartum women had the physiological characteristics of deficiency syndrome and stasis syndrome according to the TCM theory [68].In the case of maternal frail state, excessive thinking exhausts the heart blood and spleen qi, which leads to deficiencies in both heart and spleen.Hypofunction of heart and spleen leads to mind (xin shen) dystrophy and causes or aggravates the symptoms of PPD.Stagnation of liver qi and deficiency of both heart and spleen are the principal pathogenic mechanisms of PPD.Therefore, stagnation of liver qi alone or in combination with deficiencies of both heart and spleen is the main type of syndrome differentiation (bian zheng fen xing) according to TCM theory.MRLQS plus MSHS is a potential treatment option besides MRLQS alone for the management of PPD.This matches the principal pathogenesis and syndrome differentiation according to Chinese medicine.This study found that MRLQS plus MSHS in combination with routine treatments was clinically effective when compared with routine treatments alone in reducing HAMD score.Meanwhile, mean difference (MD) value of MRLQS plus MSHS in combination with routine treatments was greater than MRLQS plus MSHS alone, which suggests that a combination of MRLQS plus MSHS with routine treatments may have synergistic effects.Meta-analysis of EPDS further confirmed the effectiveness of MRLQS plus MSHS in combination with routine treatments.In terms of safety, MRLQS in combination with routine treatments or MRLQS plus MSHS in combination with routine treatments was associated with fewer adverse effects than routine treatments alone according to the existing data.Therefore, based on our analysis of clinical research and TCM theory, the use of MRLQS in combination with routine treatments or MRLQS plus MSHS in combination with routine treatments appears to be a superior approach for the treatment of PPD.Additionally, because MRLQS plus MSHS alone is safer and more effective than routine treatments, MRLQS plus MSHS might be an alternative option for patients who cannot tolerate antidepressant drugs.
In the present review, 47.22% of the trials (17/36) utilized MRLQS in combination with routine treatments for the management of PPD.Xiao yao powder (11/17) was identified as the most frequently used Chinese herbal compound in related studies.A prior systematic review revealed the effectiveness of Xiao yao powder in the treatment of depression [75]; further, it indicated that Xiao yao powder may be effective in the treatment of PPD.In trials evaluating MRLQS plus MSHS, shugan jieyu capsule was the most frequently used Chinese herbal compound, and a prior systematic review also reported evidence supporting the effectiveness of shugan jieyu capsule for the treatment of depression [76].Although the present review confirmed the effectiveness of MRLQS or MRLQS plus MSHS and their corresponding Chinese herbal compounds, this information should be used cautiously.The therapeutic method of TCM and its corresponding Chinese herbal compound should be selected based on Chinese medicine syndrome differentiation (bian zheng).Due to the high risk of bias, the significant clinical heterogeneity observed between the trials, and the relatively small sample size, a definite conclusion still cannot be drawn.

Limitations of This
Review.There were some limitations in this review which contributed to the inconclusive results.First, all trials were of poor methodological quality, and there was a high risk of bias.Of the 47 trials, only 42.55% (20/47) provided information on the methods for sequence generation, and most of them provided insufficient information to ascertain whether the randomization was conducted properly.The remainder of the trials only mentioned that "the patients were randomized into two groups."Allocation concealment was mentioned in only one trial [57].Although the trial indicated that the random numbers had been placed in a sealed envelope, it was not mentioned whether the envelope was translucent.Only one trial [63] claimed doubleblind design and presample size estimation.The majority of the trials (72.34%) did not provide any information regarding withdrawal from the study, and only one trial [57] conducted intention-to-treat analysis.
Second, the therapeutic methods of TCM and Chinese herbal compounds investigated may be responsible for the considerable heterogeneity in this review.In the practice of TCM, the herbalist should evaluate the patients symptoms to conduct syndrome differentiation and then propose the corresponding therapeutic methods of TCM.The herbal formulation is prescribed based on the therapeutic methods of TCM.This is called syndrome differentiation and treatment (bian zheng lun zhi), which is the strength of TCM.According to TCM theory, one type of syndrome differentiation should match one type of fixed therapeutic method of TCM.Different types of Chinese herbal formulations are prescribed to create individualized treatment plans for the patients under the guidance of the relatively fixed therapeutic methods of TCM.Although the specifics of CHM may differ among the trials, the therapeutic methods are consistent.Therefore, the therapeutic methods of TCM are at the core of CHM treatment.However, only 19.15% of the trials (9/47) provided information on the patients' type of syndrome differentiation, only 38.30% of the trials (18/47) provided information on the therapeutic methods of TCM, and only 12.77% of the trials (6/47) contained both.Meanwhile, 39 different Chinese herbs were investigated in this review and could be divided into 8 categories of therapeutic methods of TCM.The formulations included herbal formulas (27 trials), powder (2 trials), granules (2 trials), pills (6 trials), capsules (5 trials), herbal extracts (3 trials), and liquids (2 trials).Therefore, it was difficult to draw a definite conclusion regarding which therapeutic method of TCM or Chinese herbal compound was the most effective.
Third, the majority of the trials did not provide followup information.Three trials provided follow-up information over a 3-6-month period; however, this is considered insufficient for assessment of the long-term effectiveness of CHM.
Finally, this review found that reporting on adverse events in the included trials was inadequate.No mention of adverse events was made by 21.78% of the trials (10/47).Although 78.72% (37/47) of the trials reported adverse events, the information was limited.No trials reported whether CHM could affect breast-feeding or have adverse effects on newborns.Therefore, the safety of CHM for the treatment of PPD could not be assessed.Deficiencies in this systematic review limit the reliability of the meta-analysis results.

Implications of Further Research.
To ensure evidencebased clinical practice, more rigorous multicenter, large, methodological design trials are warranted to provide higher quality evidence.The process of random sequence generation, adequate allocation concealment, double blinding, clear description of reasons for withdrawal, and intentionto-treat should be addressed.CHM should be conducted based on the syndrome differentiation pattern of PPD and the guidelines for TCM therapeutic methods, which have been previously reported.The adverse events associated with CHM should be described.Participants should be followed up for an extended period of time to assess the long-term effects of the treatment.Trials should adhere to CONSORT standards (http://www.consort-statement.org/) to promote internal and external validity.

Conclusion
In conclusion, the evidence indicates that CHM was effective in treating PPD; additionally, MRLQS combined with routine treatments or MRLQS plus MSHS combined with routine treatments was more effective than routine treatments alone.MRLQS plus MSHS alone might be an effective alternative to routine treatments.However, the evidence detailed in this review should be considered cautiously because of the high risk of bias and the poor methodological quality of the trials.Therefore, rigorously designed trials are needed for further evaluation of the safety and efficacy of CHM for the treatment of PPD.

Figure 1 :
Figure 1: Flow diagram of the search process and study selection for the systematic review.

Figure 2 :
Figure 2: Risk of bias graph: review authors' judgments about each risk of bias item presented as percentages across all included studies.

Table 1 :
Characteristics of included randomized controlled trials.
I: intervention group; C: control group; TCM: traditional Chinese medicine; DSM: diagnostic and statistical manual of mental disorders; CCMD: Chinese classification of mental disorders; ICD: international classification of diseases; NR: not reported; d: day; w: week; m: month.

Table 3 :
Effect estimates of CHM for PPD.