Chinese herbal medicine for premenstrual syndrome

Abstract

Background

Traditional Chinese herbal medicines are frequently used to treat premenstrual syndrome (PMS) in China. Until now, their efficacy has not been systematically reviewed.

Objectives

To evaluate the effectiveness and safety of traditional Chinese herbal medicines in the treatment of women with premenstrual syndrome.

Search methods

We searched MEDLINE (January 1950 to December 2007), EMBASE (January 1980 to December 2007), Chinese Biomedical Database (CBM) (January 1975 to December 2007), China National Knowledge Infrastructure (CNKI) (January 1994 to December 2007), and the VIP Database (January 1989 to December 2007).

Selection criteria

Randomised controlled trials (RCTs) studying the efficacy of traditional Chinese herbal medicine(s) for treatment of premenstrual syndrome were included.

Data collection and analysis

Two review authors telephoned the original authors of the RCTs to confirm the randomisation procedure and extracted and analysed data from the trials that met the inclusion criteria.

Main results

Two RCTs involving 549 women were included. One trial which was identified to be of higher methodological quality demonstrated the therapeutic effectiveness of Jingqianping granule. The other study was considered of lower quality due to the inherent risk of various biases in it. The two studies showed statistically significant differences in favour of taking Jingqianping granule compared with Xiaoyaowan in the elimination of symptoms during the proliferative and premenstrual phases (RR 3.50, 95% CI1.74 to 7.06). Women treated with Cipher decoction had a higher rate of recovery than those taking co‐vitamin B6 capsules (RR 48.99, 95% CI 3.06 to 783.99).

Authors' conclusions

It is rare in PMS management that efficacy claims are substantiated by clinical trials. One of the identified trials was well designed and reported on the effectiveness of Jingqianping in the treatment of premenstrual syndrome. However, currently there is insufficient evidence to support the use of Chinese herbal medicine for PMS and further, well‐controlled trials are needed before any final conclusions can be drawn.

Plain language summary

Herbal treatment for premenstrual syndrome

Herbal medicines are sometimes used for treating premenstrual syndrome (PMS). However, the effectiveness of this type of therapy has not be rigorously evaluated in randomised controlled trials.

The authors identified two trials that evaluated herbal medicines in PMS. One of these was a higher quality study that tested the traditional Chinese medicine decoction Jingqianping granule. This was shown to increase the rate of recovery from PMS symptoms. Because the formula for this herbal medicine was provided by the trialists themselves, the review authors recommend further trials to ensure that the results are reproducible with other formulations. Strong evidence in support of other herbal formulae for the treatment of PMS is currently lacking.

Summary of findings

Summary of findings for the main comparison. New Summary of findings table.

Study	Clinical recovery	No improvement
Qiao 2002	56/202 in intervention group; 8/101 in control group	8/202 in intervention group; 18/101 in control group
Zhu 2003	50/166 in intervention group; 0/80 in control group	2/166 in intervention group; 31/80 in control group

Background

Description of the condition

Premenstrual syndrome (PMS) is a common cyclic disorder for young and middle‐aged women that consistently occurs during the luteal phase of the menstrual cycle. It is characterised by emotional and physical symptoms and can manifest with a wide variety of symptoms, including depression, mood lability, abdominal pain, breast tenderness, headache, and fatigue (Dickerson 2003; Tempel 2001). The symptoms of PMS may be sufficiently severe to disrupt women's normal functioning, quality of life, and interpersonal relationships. More severe, mainly psychological symptoms such as depression, anxiety, and other affective symptoms which recur in the luteal phase of the menstrual cycle are classified as premenstrual dysphoric disorder (PMDD) (Dickerson 2003; Sundstrom 2003).

As many as 75% of women with regular menstrual cycles experience some symptoms of PMS, PMDD affects only 3% to 8% of women in this group (Steiner 2000). In addition, PMS is more prevalent among white women, smokers, obese, and younger women (Masho 2005).

The Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM‐IV) (APA 2005) requires at least five specified symptoms in order to diagnose premenstrual dysphoric disorder (PMDD); while the International Statistical Classification of Diseases, 10th Revision (ICD‐10) (WHO 2005) requires only one distressing symptom for a diagnosis of PMS. A diagnosis of PMS requires determining the timing of the symptoms in relation to menses, meaningful change between post‐ and premenstrual symptom severity, and clinically significant severity of the symptoms. The current diagnostic standard requires confirmation of subjective symptom reports by prospective daily diaries. Diagnostic criteria for PMS must recognise the broad range of symptoms, the temporal pattern of the symptoms, and the critical issue of symptom severity, which differentiates clinically significant PMS from normal menstrual cycle changes (Freeman 2003). The best tool to diagnose PMS is a daily symptoms rating calendar. To have the diagnosis of PMS, the symptoms must be severe enough to disrupt normal daily activities (Tempel 2001).

The pathogenesis of PMS remains uncertain; research suggests that altered regulation of neurohormones and neurotransmitters is involved (Dickerson 2003). Several biological mechanisms that underlie menstrually related symptoms have been proposed. They focus mostly on gonadal hormones, their metabolites, and interactions with neurotransmitters and neurohormonal systems such as serotonin, GABA, cholecystokinin, and the renin‐angiotensin‐aldosterone system. Altered responses of these systems to ovarian hormone fluctuations during the menstrual cycle, as well as increased sensitivity to changes in ovarian hormones, may contribute to menstrually related symptoms in vulnerable women. Core underlying mechanisms may be disrupted homeostasis and deficient adaptation (Halbreich 2003). Current research focusing on the biological background of premenstrual syndrome suggests that both psychological and physiological factors exert their effect through dysregulation of serotonergic function. The results of several studies point to a variation in the function of the serotonergic system throughout the menstrual cycle. Ovarian steroids have been found to profoundly influence the activity of the serotonergic system (Eriksson 1994; Gonda 2004; Joffe 1998; Rubinow 1998). Additionally, reduced serotonin uptake by platelets and reduced whole blood serotonin levels during the luteal phase have been demonstrated in women with PMS (Taylor 1984; Rapkin 1987; Steege 1992).

Hypomagnesaemia (low magnesium levels in the blood) has been implicated in the pathogenesis of PMS (Ventskivs'ka 2005). Moreover, high intake of fats and low intake of foods with high concentration of carbohydrate may be associated with premenstrual symptoms (Nagata 2004).

Description of the intervention

Many different treatments have been suggested as possible therapies for PMS. This is due to the uncertainty of its pathogenesis, the wide range of its manifestations, and a large placebo effect. As serotonin has been implicated in the pathogenesis of PMS, the therapeutic effectiveness of continuous and luteal phase dosing of selective serotonin reuptake inhibitors (SSRIs) has been evaluated (Dimmock 2000). SSRIs have emerged as first‐line therapy for PMS (Steiner 2006).

Ovarian function appears to play a fundamental role in PMS, accordingly treatment strategies designed to suppress ovulation have generally been found to be effective for treatment of menstrually related syndromes and symptoms. Gonadotrophin‐releasing hormone analogues (GnRHa) appear to provide an effective treatment of premenstrual syndrome (Backstrom 2003; Kouri 1998; Wyatt 2004).

In other studies, women with PMS who practised aerobic exercise reported fewer symptoms than participants in the control group (O'Brien 2000; Steege 1993). Dietary restrictions or supplements may also be useful in women with PMS (Kessel 2000; Moline 2000). Sodium restriction has been proposed to minimise bloating, fluid retention, breast swelling and tenderness. Caffeine restriction is recommended because caffeine intake is related to premenstrual irritability and insomnia. A systematic review of placebo‐controlled trials of evening primrose oil suggested lack of any benefit in PMS, although mild relief was demonstrated in women with breast tenderness (Budeiri 1996).

A randomised placebo‐controlled study reported there were significant improvements in the symptoms of negative feeling, pain, water retention, and total PMS symptoms in women receiving qigong therapy when compared to placebo controls (Jang 2004). Qigong consists primarily of meditation, relaxation, physical movement, mind‐body integration, and breathing exercises. Practitioners of qigong develop an awareness of qi sensations (energy) in their body and use their mind to guide the qi. When the practitioners achieve a sufficient skill level (master), they can direct or emit external qi for the purpose of healing others.

Some studies also indicate that acupuncture, homeopathy, aromatherapy, reflexology, Gingko biloba, kava kava, black cohosh, and agnus castus can relieve symptoms of PMS such as anxiety, depression, and irritability (Jones 2003; Tesch 2003; Yu 2005).

Other treatments for PMS, for which there is inconclusive evidence, include photic stimulation, cognitive behavioural therapy, relaxation therapy, vitamin B6, L‐tryptophan, stress reduction, spironolactone, and a complex carbohydrate drink. Although evidence for relief of PMS symptoms is inconclusive, it is reasonable to recommend these as healthy lifestyle changes which may have overall benefits (Douglas 2002; Girman 2003; Rapkin 2003).

How the intervention might work

Herbal medicines have long been used in traditional healing systems to treat conditions of particular interest to women, such as premenstrual syndrome (PMS) and menopausal symptoms. For a select number of phytomedicines, including evening primrose oil, black cohosh root extract, dong quai, and chaste tree berry, scientific investigation has identified the pharmacologically active constituents, mechanism of action, and clinical value. Based on the available evidence, chaste tree berry may be a reasonable treatment alternative for some women with PMS. Dong quai (in standard pinyin it is called dang gui) may be used to treat PMS when combined with traditional Chinese multiple‐herb formulations (Hardy 2000). The current literature supports the use of chaste berry for cyclical breast discomfort and premenstrual syndrome (Roemheld‐Hamm 2005).

Why it is important to do this review

Traditional Chinese medicine (TCM) has been used for thousands of years to treat PMS. The rule of TCM for PMS is 'Bian Zheng Lun Zhi': adding or reducing some herbs in a traditional preparation, so‐called 'demonstrated preparation', depending on the TCM 'Zheng' signs recognised by a TCM physician. At least eight types of 'Zheng' are recognised by TCM physicians, and various formulations can be used to treat the same type of 'Zheng'. Some new preparations for various 'Zheng xing' (TCM signs) have been developed, for example, 'gan yu qi ci' type PMS can be treated by using adjusted 'cai hu su gan san' (Lang 1996), 'pi shen yang xu' type PMS by adjusted 'fang ji huang qi tang' combined with 'you gui wan' (Zhu 1995). These formulations or preparations have been described as effective remedies and it has been suggested they could have clinical applications.

There is increasing public interest in, and use of, herbal medicine therapies which lie outside the 'mainstream' or traditional Western medical practice across the globe (HLSC 2000). There is evidence to indicate that not all herbs are risk‐free. There are concerns about adverse events, particularly allergic reactions, as well as Chinese herbal nephropathy (CHN) affecting the kidneys (Lampert 2002; Lord 2001; Nortier 2000). Hence, there is a need to systematically review herbal medicines for their effectiveness and safety.

Objectives

To evaluate the effectiveness and safety of traditional Chinese herbal medicines in the treatment of women with premenstrual syndrome.

Methods

Criteria for considering studies for this review

Types of studies

Only randomised controlled trials were included, without restriction on language or publication type. Pseudo‐RCTs were not be considered.

Types of participants

Women of reproductive age diagnosed with premenstrual syndrome who suffered with one or more symptoms occurring periodically during 0 to 14 days before menstruation, for three or more menstrual cycles. Women known to have medical problems such as hypothyroidism, hypoglycaemia, psychiatric disorders, or had bilateral oophorectomy or any type of cancer were excluded.

Types of interventions

Any form of herbal medicine was considered, including oral preparations, decoctions, injections, and tablets. Comparisons could consist of a placebo, no intervention, acupuncture, western medication, or any other intervention. Comparison of one kind of herbal medicine versus another herbal medicine was also included.

Types of outcome measures

Primary outcomes

Change of overall symptoms

The following tools were considered.

(1) PMS scoring systems 
 Any validated objective scoring system used for evaluating premenstrual symptoms, such as the Moos Menstrual Distress Questionnaire or the Daily Symptom Report were considered.

(2) Improvement of overall symptoms (dichotomous data)

If the above two outcome measures were not used in a study, we considered the outcome measures as follows:

‐ records made by participants or doctors on a chart or using a visual analogue scale;

‐ severity of symptoms reported, judged in accordance with the definitions of the numerical scales used in the charts.

Secondary outcomes

1. Quality of life 
 The Health Related Quality of Life (HRQOL) or other validated scales were considered in this review.

2. Adverse events 
 Events such as functional injury of the liver or kidney, nausea, vomiting, diarrhoea, and skin rash were recorded.

Search methods for identification of studies

We attempted to identify all relevant studies regardless of language or publication status (published, unpublished, in press, or in progress) (Appendix 1).

We searched: 
 (1) the Cochrane Menstrual Disorders and Subfertility Group Trials Register and the Cochrane Complementary Medicine Field Trials Register; 
 (2) the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library, Issue 4, 2007) (Appendix 2); 
 (3) MEDLINE (January 1966 to December 2007), EMBASE (January 1980 to December 2007) (Appendix 4), CISCOM (to December 2007), Chinese Biomedical Database (CBM) (January 1975 to December 2007), China National Knowledge Infrastructure (CNKI) (January 1979 to December 2007), VIP Database (January 1989 to December 2007), AMED (Appendix 4);

(4) ongoing trials in the Meta‐register of Controlled Trials, which includes the Medical Research Council Clinical Trials Directory, National Research Register, Chinese Clinical Trial Register, as well as other registers for ongoing trials;

(5) reference lists of the relevant trials and reviews identified;

(6) unpublished and on‐going trials, through correspondence with authors;

(7) major herbal treatment and obstetrics and gynaecology conference proceedings and poster abstracts (about this disease) over the last five years;

(8) for side‐effect studies and contacted various adverse reaction‐reporting bodies.

Data collection and analysis

The study selection and data analysis were undertaken by two authors (ZJ and WTX).

Selection of studies

From the search results, ZJ and YXZ scanned the identified results abstracts and relevant records. Full articles were retrieved for all potentially relevant trials. Since all the publications were in Chinese, WTX and ZJ were independently able to interview the original authors by telephone to identify what method was used to generate the allocation sequence. WTX selected the trials for inclusion. All identified reports were scrutinised to check for multiple publications of the same trials.

Data extraction and management

Using a piloted data extraction form, we extracted data on study characteristics including methods, participants, interventions, and outcomes. There were no disagreements between the authors.

For the included studies we extracted the formulation contents and herb names in three languages (Table 2, Table 3).

1. Herbs names in different language.

Pinying name	Latin name	English name
Baishao	Radix Paeoniae Alba	White Peony root
Baishu	Atractylodes macrocephala Koidz.	Largehead Atractylodes rhizome
Bohe	Herba Menthae	Field mint/peppermint
Chaihu	Radix Bupleuri	Chinese Thorowax root /Red Thorowax root
Chaozaoren	Stir‐baked Semen Ziziphi Spinosae	Stir‐baked Semen Ziziphi Spinosae
Chuanlianzi	Fructus Toosendan	Szechwan chinaberry fruit
Chuanxiong	Rhizoma Chuanxiong	Szechwan Lovage rhizome
Cuchaihu	Stir‐baked Radix Bupleuri with vinegar	Stir‐baked Radix Bupleuri with vinegar
Danggui	Radix Angelicae Sinensis	Angelica; angelica root
Fuling	Poria	Indian Bread/ Poria cocos/Poria cocos Wolf
Hehuanpi	Cortex Albizziae	Silktree albizzia bark
Juye	Citrus reticulata Blanco var. erythorosa H.H.Hu	Tangerine leaf
Shenglongchi	Dens Draconis	Dragon'sTeeth
Sigualuo	Retinervus Luffae Fructus	Retinerus Luffae Fructus
Xiangfu	Rhizoma Cyperi	Nutgrass Galingale rhizome
Yujin	Radix Curcumae	Aromatic Turmeric root‐tuber
Zhigancao	Radix Glycyrrhizae Preparata	Prepared Radix Glycyrrhizae
Zhike	Fructus Aurantii	Bitter Orange

2. Contents of the formulae in included studies.

Study ID	Contents
Qiao 2002	Did not provide any information about the contents of interventions but the drugs were the same as Wei 2006a, including the types and drug manufacture. Jingqianping granule: baishao, xiangfu, chuanxiong, zhike, chuanlianzi, etc; Xiaoyaowan: chaihu, danggui, baishao, baishu, fuling, zhigancao, bohe.
Wei 2006a	Jingqianping granule: baishao, xiangfu, chuanxiong, zhike, chuanlianzi, etc; it did not provide any information about the contents of xiaoyaowan (the control). We had indexed and found the contents of xiaoyaowan: chaihu, danggui, baishao, baishu, fuling, zhigancao, bohe.
Zhu 2003	Cipher Fang decoction: cuchaihu, baishao, yujin, juye, sigualuo, chaozaoren, hehuanpi, fuling, shenglongchi, etc. Vitamin B6 capsule: vitamin B6 20mg, vitamin B1 10mg, oryzanol 10mg.

Assessment of risk of bias in included studies

Risks of bias were assessed independently by at least two authors using the criteria that are described in the Cochrane Handbook for Systematic Reviews of Interventions 5.0.0 (Higgins 2008) and (Wu 2007).

The following characteristics were assessed.

Randomisation process: assessment for selection bias

A ‐ adequate sequence generation was reported using one of the following approaches: random number tables or computer‐generated random numbers, coin tossing or shuffling for generating the allocation sequence before the launch of the trial; trials considered eligible and with a low risk of selection bias.

B ‐ one of the adequate methods outlined in (A) was not specified and only mentioned 'random'; considered to have a moderate risk of selection bias.

C ‐ other methods of allocation, for example quasi‐randomisation, that appeared to have a high risk of bias; trials were excluded.

Allocation concealment process: assessment of selection bias

A ‐ adequate measures to conceal allocation: where the person who generated an allocation sequence did not attend to recruitment of the participants, and the allocation sequence was sealed in opaque envelopes or held in a locked computer, or another description that contained convincing elements of concealment; considered to have a low risk of selection bias.

B ‐ unclear: concealed trials in which the author did not report an approach to allocation concealment at all; considered as having a moderate risk of selection bias.

C ‐ inadequately‐concealed allocation that reported an approach that did not fall into one of the categories in (A).

D ‐ did not conceal allocation.

C and D were considered as having a high risk of selection bias. 

Level of blinding: assessment for performance bias and detection bias

A ‐ double blinding: participants and outcome assessor were masked and were considered as having a low risk of both performance and detection bias.

B ‐ single blinding of the outcome assessor was considered to have a moderate risk of both performance and detection bias. If participants were blind to group allocation but not the outcome assessor, the study was considered to have a high risk of detection bias.

C ‐ non‐blinding was considered to have a high risk of both performance and detection bias.

Measures of treatment effect

We have analysed the data using Review Manager (Version 5.0). We compared outcome measures for continuous data using the mean difference with 95% confidence interval. For binary data we used relative risks with 95% confidence interval.

Dealing with missing data

The risk of bias of missing data was estimated according to following method (Higgins 2008).

A ‐ low risk of bias: trials where few dropouts and losses to follow up were noted and an intention‐to‐treat analysis was possible.

B ‐ moderate risk of bias: trials which reported the rate of exclusions as about 10% whatever intention‐to‐treat analysis was used.

C ‐ high risk of bias: the rate of exclusions was at least 15%, or wide differences in exclusions between groups, whatever intention‐to‐treat analysis was used.

Assessment of heterogeneity

In the future, we will analyse the clinical and methodological heterogeneity if the included studies are similar enough. We will assess heterogeneity amongst trials to decide whether the data could be pooled by using the Chi² test and I² statistic. An I² < 25% means low heterogeneity, I² between 25% and 50% is considered moderate heterogeneity, and I² > 50% means notable heterogeneity. We will combine data using random‐effects model. If I² > 75%, which means substantive heterogeneity existed, the data will not be combined (Higgins 2008).

Assessment of reporting biases

The risk of reporting bias was assessed: refer to the description of Cochrane Handbook of Systematic Reviews of Interventions 5.0.0 (Higgins 2008).

No ‐ low risk of reporting bias: all outcomes were reported in detail.

Probably yes ‐ moderate risk of reporting bias: at least one of the outcomes was mentioned but not in detail.

Yes ‐ high risk of reporting bias: at least one outcome was not reported.

Data synthesis

We did not perform a combined analysis due to the use of different TCM therapy formulations and comparators in the included studies.

Potential publication bias was not tested because there were only two studies in this review.

Subgroup analysis and investigation of heterogeneity

We performed subgroup analysis based on the different traditional Chinese herbal medicine formulae used.

Sensitivity analysis

We did not conduct a sensitivity analysis as a combined analysis was not performed.

Results

Description of studies

A total of 17 trials that claimed to be randomised were identified. We successfully contacted seven of the trial authors by telephone. Of the seven trials, four were excluded because the trial authors misunderstood what was meant by random allocation and their trials were not real randomised controlled trials. Two remaining studies were identified as true RCTs (Qiao 2002; Zhu 2003).

Excluded studies

See 'Characteristics of excluded studies'.

Of the 17 published articles initially identified, nine articles were excluded as they were not randomised controlled comparisons (Li 2003; Pei 2005; Sun 2003; Wang 1997; Wang 2004; Yin 1999; Yue 2005; Zhang 2000; Zhang 2003). Outcome measures in Wei 2006 were not included in the criteria for this review. Excluded studies and the reasons for exclusion are listed in the 'Characteristics of excluded studies' table.

We were unable to contact the authors of five studies and these are listed in 'Studies awaiting classification' (Feng 1996; Guo 2004; Li 2002; Xie 1994; Yuan 2004).

Included studies

See 'Characteristics of included studies'.

Two RCTs fulfilled our inclusion criteria (Qiao 2002; Zhu 2003).

Design

Both trials used a parallel group design. One of the studies (Qiao 2002) had a multicentre, double‐dummy, double‐blind design.

Participants

The two studies recruited participants based on traditional Chinese medicine (TCM) signs. These studies included a total of 549 participants, age range 16 to 45 years. The trials authors of both studies reported that the diagnosis of PMS was made according to the "Regulation of clinical study of new TCM for PMS", however, there was no detailed description of this regulation.

Interventions

Qiao 2002 compared Jingqianping granule and Xiaoyaowan.

Zhu 2003 compared Cipher Fang decoction and vitamin B6 in a capsule.

Outcomes

The two studies (Qiao 2002; Zhu 2003) did not report outcomes using the tools defined in this review. Instead they reported on recovery, marked improvement, and no improvement based on TCM signs.

Risk of bias in included studies

See Figure 1; Figure 2.

1.

Methodological quality graph: review authors' judgements about each methodological quality item presented as percentages across all included studies.

2.

Methodological quality summary: review authors' judgements about each methodological quality item for each included study.

Randomisation

Both included studies were RCTs that used random number tables.This was clearly stated in the study by Zhu 2003 while it was only confirmed following contact with the authors for the Qiao 2002 study.

Allocation concealment

Neither of the of the included trials mentioned allocation concealment.

Following contact with the study authors, concealment was considered adequate in one of the included trials (Qiao 2002) as the random number allocation sequence was generated by a statistician not involved in participant recruitment.

Blinding

Single blinding was used in one study (Zhu 2003) while the Qiao 2002 study used a double‐dummy, double‐blind methodology hence the risk of both performance bias and detection bias was deemed very low.

Withdrawals, dropout, loss to follow‐up and intention to treat

Neither of the included trials reported on withdrawals, dropouts, or loss to follow up of participants.

Effects of interventions

See: Table 1

1. Change in premenstrual syndrome symptoms

(1) Rate of recovery (defined as disappearance of symptoms during the trial)

The two studies (Qiao 2002; Zhu 2003) reported recovery rates and showed statistically significant differences in elimination of symptoms between the intervention and control groups.

The rate of recovery in the group using Jingqianping granule was 27.7% (56/202) while it was 8% (8/101) in the Xiaoyaowan group (RR 3.50, 95% CI 1.74 to 7.06). This indicated that Jingqiaping granule was more effective than Xiaoyaowan in eliminating PMS symptoms (Qiao 2002).

Participants who took the Cipher decoction had a recovery rate of 30% (50/166) while there were no reported recoveries in the vitamin B6 capsule group (0/80) (RR 48.99, 95% CI 3.06 to 783.99). This suggested that Cipher decoction was more effective than vitamin B6 in eliminating PMS symptoms (Zhu 2003).

(2) Rate of no improvement (defined as symptom severity that remained unchanged during the trial)

Both studies also reported the rate of 'no improvement' and showed statistically significant differences between the intervention and control groups. Eight participants reported no improvement (4%, 8/202) in the Jingqianping granule group compared to 18 in the Xiaoyaowan group (18%, 18/101) (Qiao 2002).

Two participants reported no improvement in TCM signs (1.2%, 2/166) in the Cipher Fang decoction group compared to 31 in the vitamin B6 group (38.8%, 31/80) (Zhu 2003).

2. Time between commencing treatment and BBT or endocrine levels returning to normal

One study (Zhu 2003) reported on the changes in prolactin, estradiol, and progestin concentrations, before and after treatment. However, only a very small number of participants contributed to the data (14/166 in the experiment group and 10/80 in the control group). We therefore did not include these results in the analysis of this review.

3. Development of adverse events

Neither study reported any toxic reactions or adverse events.

Discussion

Only two trials (Qiao 2002; Zhu 2003) fulfilled the inclusion criteria. One study (Qiao 2002) was considered to be of higher methodological quality than the other with regard to study design and trial processing because a double‐dummy technique was used. Due to the potential selection, detection, and attrition biases in the Zhu 2003, the study was considered to be of a lower methodological quality. However, the formula of the experimental drug (Jingqianping granule) was prepared by the authors themselves and the drug was made by the authors' university. Similarly, the experimental drug used in the study by Qiao 2002 was also prepared by authors themselves. Therefore, independent validation of the findings of these trials is advisable.

None of the included trials used a placebo control and active controls were used instead. PMS is a condition with a substantial placebo response rate, hence it is scientifically important to include a placebo group in trials evaluating the therapeutic effectiveness of new medications. Moreover, it has been suggested that, by making comparisons with external placebo, the need for placebo control groups in new studies on patients could be minimised assuming that the novel medication will perform in the same way in a study with only active controls as it would have in a placebo‐controlled trial. However, evidence from 32 RCTs involving  7264 participants showed that the degree of improvement was nearly double in active‐controlled trials than with the same drugs and dosages in placebo‐controlled studies Woods 2005. Therefore, the high RRs identified in this review could be secondary to the lack of use of placebo controls.

None of the trials provided sufficient information about allocation concealment. After contacting the authors, it was judged that Qiao 2002 used adequate methodology to ensure concealment at the time of randomisation. In addition, none of the studies reported on their justification for the required sample size. Although the included trials seem to have recruited a large sample size (Qiao 2002 included 303 participants, 202:101; and Zhu 2003 included 166:80) it is important to have formal sample size calculations to ensure that the trial is of sufficient power to identify a difference, if one exists.

Surprisingly, several trials claimed to be RCTs but, after contacting the trial authors to enquire about the method of randomisation used, we found that most of the authors misunderstood the concept of randomisation or had forgotten the details of the methodology they employed; they had conducted the studies several years ago. Therefore, most of the studies of Chinese herbal medicine in the management of premenstrual syndrome (PMS) were inadequate to provide reliable therapeutic effect estimates because of poor study design and methodological quality.

Finally, although the diagnostic standard was mentioned in two studies a detailed description of the standard was not provided. In the absence of this key parameter we were unable to judge the validity of the studies and how the diagnosis of PMS was made for the study participants.

Authors' conclusions

Implications for practice.

Currently there is insufficient evidence on the efficacy of traditional Chinese medicine in the treatment of PMS.

Implications for research.

Further randomised clinical trials are required of traditional Chinese herbal medicine for the treatment of premenstrual syndrome. Independent validation of the therapeutic effectiveness of Jingqianping granule is advisable. It is crucial that investigators give considerable attention to the methods of randomisation, blinding, determining sample sizes, and the use of placebo‐controlled trials. The information on trial methodology should also be reported in detail, according to both CONSORT (Moher 2001) and CONSORT for TCM (Wu 2007 CONSORT).

What's new

Date	Event	Description
20 September 2010	Amended	Contact details updated.

History

Protocol first published: Issue 1, 2007
 Review first published: Issue 1, 2009

Date	Event	Description
30 April 2008	New search has been performed	Converted to new review format.
8 September 2006	New citation required and major changes	Substantive amendment

Notes

Approved by all of the authors.

Appendices

Appendix 1. Search strategy for electronic databases

Search strategy for the above electronic database.

1 Complementary Therapies 
 2 alternative medicine$.ti,ab. 
 3 Plant Extracts 
 4 plant extract$.ti,ab. 
 5 botanical extract$.ti,ab. 
 6 Plants, Medicinal 
 7 medicinal plant$.ti,ab. 
 8 Medicine, Kampo/ or Phytotherapy/ or Drugs, Chinese Herbal 
 9 Medicine, Chinese Traditional/ or Medicine, Oriental Traditional 
 10 (Chinese adj3 medicin$).ti,ab. 
 11 phytodrug$.mp. 
 12 phytomedicine$.ti,ab. 
 13 phytopharmaceutical$.ti,ab. 
 14 herbal medicine$.ti,ab. 
 15 (complementary adj3 medicine$).ti,ab. 
 16 non‐prescription drug$.mp. or Drugs, Non‐Prescription 
 17 (Chinese adj3 herb$).mp. 
 18 herbal remed$.mp. 
 19 herbal extract$.mp. 
 20 herbal preparation$.mp. 
 21 botanical preparation$.mp. 
 22 (herb$ adj3 mixture$).mp. 
 23 exp medicine, traditional/ or exp medicine, African traditional/ or exp medicine, arabic/ or exp medicine, unani/ or exp medicine, ayurvedic/ or exp medicine, kampo/ or exp medicine, oriental traditional/ or exp medicine, Chinese traditional/ or exp medicine, Tibetan traditional/ or exp shamanism 
 24 or/1‐23 
 25 premenstrual syndrome 
 26 ((premenstrual or pre‐menstrual or pre menstrual) adj syndrome$).tw. 
 27 ((premenstrual or pre‐menstrual or pre menstrual) adj tension).tw. 
 28 premenstrual dysphor$.tw. 
 29 late luteal.tw. 
 30 (luteal adj dysphor$).tw. 
 31 (pms or pmt or pmdd or llpd or llpdd).tw. 
 32 or/25‐31 
 33 24 and 32 
 34 randomised controlled trial.pt. 
 35 controlled clinical trial.pt. 
 36 Randomized controlled trials 
 37 random allocation 
 38 double‐blind method 
 39 single‐blind method 
 40 or/34‐39 
 41 clinical trial.pt. 
 42 exp clinical trials 
 43 (clin$ adj25 trial$).ti,ab,sh. 
 44 ((singl$ or doubl$ or tripl$ or trebl$) adj25 (blind$ or mask$)).ti,ab,sh. 
 45 placebos 
 46 placebo$.ti,ab,sh. 
 47 random$.ti,ab,sh. 
 48 Research design 
 49 or/41‐48 
 50 animal/ not (human/ and animal/) 
 51 40 or 49 
 52 51 not 50 
 53 33 and 52 
 54 from 53 keep 1‐39

Appendix 2. Key words

Keywords CONTAINS "premenstrual " or "premenstrual aggravation" or "premenstrual dysphoric disorder" or "premenstrual symptom scores" or "premenstrual symptoms" or "premenstrual syndrome" or  "premenstrual syndrome‐symptoms" or "PMDD" or "PMS" or Title CONTAINS "premenstrual " or "premenstrual aggravation" or "premenstrual dysphoric disorder" or "premenstrual symptom scores" or "premenstrual symptoms" or "premenstrual syndrome" or  "premenstrual syndrome‐symptoms" or "PMDD" or "PMS"

AND

Keywords CONTAINS "Chinese" or "Chinese herbal medicine" or "Chinese drugs" or "chinese herbal preparations" or "Chinese herbal remedy" or "Chinese traditional medicine" or "traditional Chinese medicine" or "traditional medicine" or "herbal preparations" or "herbal remedy" or "herbal supplement" or "herbal supplements" or Title CONTAINS  "Chinese" or "Chinese herbal medicine" or "Chinese drugs" or "chinese herbal preparations" or "Chinese herbal remedy" or "Chinese traditional medicine" or "traditional Chinese medicine" or "traditional medicine" or "herbal preparations" or "herbal remedy" or "herbal supplement" or "herbal supplements"

Appendix 3. AMED

1 exp chinese medicine/ or exp herbal medicine/ or exp oriental medicine/ or exp tibetan medicine/ (0) 
 2 exp Phytotherapy/ (1070) 
 3 exp alternative medicine/ (0) 
 4 herb$.tw. (9836) 
 5 (chinese adj5 medicine$).tw. (4694) 
 6 plant$.tw. (19442) 
 7 botanical.tw. (294) 
 8 (traditional adj5 medicine$).tw. (6403) 
 9 or/1‐8 (27813) 
 10 exp Premenstrual Dysphoric Disorder/ (0) 
 11 exp Premenstrual Syndrome/ (93) 
 12 Premenstrual Dysphoric Disorder$.tw. (0) 
 13 (PMDD or PMS).tw. (55) 
 14 late luteal.tw. (1) 
 15 (luteal adj5 dysphor$).tw. (0) 
 16 (pmt or llpd$).tw. (20) 
 17 premenstrual.tw. (145) 
 18 pre‐menstrual.tw. (9) 
 19 or/10‐18 (185) 
 20 9 and 19 (45) 
 21 from 20 keep 1‐45 (45)

Appendix 4. EMBASE

1 exp chinese medicine/ or exp herbal medicine/ or exp oriental medicine/ or exp tibetan medicine/ (14672) 
 2 exp Phytotherapy/ (3716) 
 3 exp alternative medicine/ (12096) 
 4 herb$.tw. (27873) 
 5 (chinese adj5 medicine$).tw. (6058) 
 6 plant$.tw. (127358) 
 7 botanical.tw. (1523) 
 8 (traditional adj5 medicine$).tw. (7549) 
 9 or/1‐8 (173861) 
 10 exp Premenstrual Dysphoric Disorder/ (411) 
 11 exp Premenstrual Syndrome/ (2702) 
 12 Premenstrual Dysphoric Disorder$.tw. (407) 
 13 (PMDD or PMS).tw. (2324) 
 14 late luteal.tw. (919) 
 15 (luteal adj5 dysphor$).tw. (90) 
 16 (pmt or llpd$).tw. (677) 
 17 premenstrual.tw. (2733) 
 18 pre‐menstrual.tw. (99) 
 19 or/10‐18 (6576) 
 20 9 and 19 (191) 
 21 Controlled study/ or randomised controlled trial/ (2692903) 
 22 double blind procedure/ (69337) 
 23 single blind procedure/ (7559) 
 24 crossover procedure/ (20294) 
 25 drug comparison/ (81252) 
 26 placebo/ (113542) 
 27 random$.ti,ab,hw,tn,mf. (408483) 
 28 latin square.ti,ab,hw,tn,mf. (1102) 
 29 crossover.ti,ab,hw,tn,mf. (35014) 
 30 cross‐over.ti,ab,hw,tn,mf. (11844) 
 31 placebo$.ti,ab,hw,tn,mf. (163643) 
 32 ((doubl$ or singl$ or tripl$ or trebl$) adj5 (blind$ or mask$)).ti,ab,hw,tn,mf. (114046) 
 33 (comparative adj5 trial$).ti,ab,hw,tn,mf. (15034) 
 34 (clinical adj5 trial$).ti,ab,hw,tn,mf. (566821) 
 35 or/21‐34 (3194701) 
 36 nonhuman/ (3067942) 
 37 animal/ not (human/ and animal/) (14468) 
 38 or/36‐37 (3071635) 
 39 35 not 38 (1885193) 
 40 20 and 39 (83) 
 41 from 40 keep 1‐83 (83)

Data and analyses

Comparison 1. Jinqianping granule versus Xiaoyaowan tablet.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Recovery	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
2 No improvement	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

1.1. Analysis.

Comparison 1 Jinqianping granule versus Xiaoyaowan tablet, Outcome 1 Recovery.

1.2. Analysis.

Comparison 1 Jinqianping granule versus Xiaoyaowan tablet, Outcome 2 No improvement.

Comparison 2. Cipher decoction versus vitamin B.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Recovery	1	246	Risk Ratio (M‐H, Fixed, 95% CI)	48.99 [3.06, 783.99]
2 No improvement	1	246	Risk Ratio (M‐H, Fixed, 95% CI)	0.03 [0.01, 0.13]

2.1. Analysis.

Comparison 2 Cipher decoction versus vitamin B, Outcome 1 Recovery.

2.2. Analysis.

Comparison 2 Cipher decoction versus vitamin B, Outcome 2 No improvement.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Qiao 2002.

Methods	Multicentre (6 hospitals) parallel design, double simulation. "Randomly allocated" was mentioned but no details about the method in original article. We telephoned the author: a random number table was used to generate the allocation sequence. Double blinding was used.
Participants	Women with premenstrual syndrome (PMS), with "Gan Qi" invasion. Diagnosed as having TCM signs of 'premenstrual tension syndrome', aged 18 to 40 years. 403 cases were divided into the intervention group (n=202), a control group (n=101), and an open treatment group (n=100). Status of disease for the treatment groups: light (22), moderate (133), severe (147).
Interventions	Jingqianping granule was used in the intervention group, oral 15 g tid, and xiaoyaowan placebo 9g bid; for two menstrual cycles. Xiaoyaowan for the control group, oral 9 g bid and Jingqianping granule placebo 15 g oral tid; for two menstrual cycles.
Outcomes	Observations for symptoms, menstrual cycle, BBT, body weight, endocrine. Recovery: TCM signs disappeared, BBT or endocrine normal; after treatment stopped, it did not relapse in three menstrual cycles, scores reduced more than 90%. Marked improvement: TCM signs markedly improved, BBT or endocrine almost normal; after treatment stopped, it did not worsen in three menstrual cycles, TCM sign scores reduced more than 60%. Improvement: main TCM signs improved, partial TCM signs relapsed after treatment stopped but better than before the treatment. TCM sign scores reduced more than 30% but less than 60%. No improvement: TCM signs had no significant change or got even worse, TCM sign scores reduced less than 30%. Safety. Adverse events.
Notes	This was a clinical study for new drug development in a phase III trial, approved by the State Food and Drug Administration (SFDA) of China. 100 persons were in the open treatment group but we do not know the results for this treatment. No statement on sample size calculation. The drug was made by the authors' university.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	No mention of the randomisation method.
Allocation concealment?	Low risk	The allocation sequence was generated by a statistician, and the sequences (four copies) were put in optic envelopes.
Blinding? All outcomes	Low risk	Double placebo technique was used.
Incomplete outcome data addressed? All outcomes	Unclear risk	Not mentioned.
Free of selective reporting?	Unclear risk	Not mentioned.
Free of other bias?	High risk	A potential conflict of interest existed because the Jingqianping granule was made by the authors' university.

Zhu 2003.

Methods	Parallel design. The published article described that a random numbers table was used to generate the allocation sequence. The participants were blinded to the drugs.
Participants	246 women with premenstrual tension syndrome, according to the diagnostic criteria: Le Jie, Gynecology and Obstetrics, 4th edition, Beijing: People's Medical Publishing House, 1997:341. 166 women in intervention group, aged 16 to 42 years; 80 women in control group, aged 16 to 40 years.
Interventions	Cipher Fang decoction was used in the intervention group, oral 20g tid, from the day BBT increased in the menstrual cycle for 12 days; repeated in the next menstrual cycle. Or, before menstruation oral 20g tid for 12 days; total of 3 menstrual cycles. Vitamin B6 capsule was given in the control group, oral 1 pill tid. The usage same as intervention.
Outcomes	Recovery: it did not relapse in three menstrual cycles after the treatment; TCM signs disappeared completely; BBT became normal; one or two of the hormones, xylose test, returned to normal. Marked improvement: main TCM signs disappeared or partial TCM signs improved more than 50%; sign scores reduced from serious to mild; BBT and PRL were markedly improved. Improvement: TCM signs were better; sign scores reduced from moderate to mild; BBT score increased 1 score, or temperature increased 0.1. No improvement. Change of hormones. BBT. Xylose excretory rate. Toxicity and adverse effects: liver and kidney function tests. No adverse effects were observed.
Notes	The formula was prepared by authors themselves. The Cipher Fang decoction was made by author's hospital. The Cipher Fang decoction contained cuchaihu, baishao, yujin, juye, sigualao, chaozaoren, hehuanpi, fuling, shenglongchi, etc. The control was a vitamin B6 capsule (vitamin B6 20mg, vitamin B1 10mg, oryzanol 10mg). There was no evidence on effect of vitamin B6 capsule for premenstrual syndrome (PMS). No statement on sample size calculation.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	No description about the method of randomisation.
Allocation concealment?	High risk	No description about allocation.
Blinding? All outcomes	High risk	No description.
Incomplete outcome data addressed? All outcomes	Unclear risk	Not mentioned.
Free of selective reporting?	Unclear risk	Not mentioned.
Free of other bias?	Unclear risk	Potential conflict of interest existed.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Li 2003	Claimed as RCT. We telephone interviewed the original author and learned that the method of allocation was according to the order of patients coming to hospital.
Pei 2005	Claimed as RCT. We telephone interviewed the original author and learned that the method of allocation was according to the order of patients coming to hospital.
Sun 2003	Claimed as RCT. We telephone interviewed the original author and learned that the method of allocation was according to the order of patients coming to hospital.
Wang 1997	Claimed as RCT. We telephone interviewed the original author and she said she had forgotten the method.
Wang 2004	Claimed as RCT. We telephone interviewed the original author and learned that the method of allocation was according to the order of patients coming to hospital.
Wei 2006	Hormone levels were used as outcomes in the study, which have very little relationship to PMS.
Yin 1999	Claimed as RCT. We telephone interviewed the original author and learned that the allocation method was performed optionally.
Yue 2005	Claimed as RCT. We telephoned the original author and learned that the patients were allocated according to the order of patients came to hospital.
Zhang 2000	We telephoned the original author and learned that the patients were allocated by authors optionally.
Zhang 2003	Claimed as RCT. We telephoned the original author and learned that this actually was a retrospective paper for a summary of the author's clinical experience.

Characteristics of studies awaiting assessment [ordered by study ID]

Feng 1996.

Methods	"Randomly allocated patients to two groups" mentioned.
Participants
Interventions
Outcomes
Notes	The original author cannot be contacted by telephone for confirmation of the randomisation method.

Guo 2004.

Methods	"Randomly allocated patients to two groups" mentioned.
Participants
Interventions
Outcomes
Notes	The original author cannot be contacted by telephone for confirmation of the randomisation method.

Li 2002.

Methods	"Randomly allocated patients to two groups" mentioned.
Participants
Interventions
Outcomes
Notes	The original author cannot be contacted by telephone for confirmation of the randomisation method.

Xie 1994.

Methods	"Randomly allocated patients to two groups" mentioned.
Participants
Interventions
Outcomes
Notes	The original author cannot be contacted by telephone for confirmation of the randomisation method.

Yuan 2004.

Methods	"Randomly allocated patients to two groups" mentioned.
Participants
Interventions
Outcomes
Notes	The original author cannot be contacted by telephone for confirmation of the randomisation method.

Differences between protocol and review

1. The outcomes measures were revised.

2. The standard of 'assessment of quality of included studies' had changed, given in the new version of the Cochrane Handbook of Systematic Reviews of Interventions 5.0.0.

Contributions of authors

Xiaoyan Chen, Taixiang Wu, and Khaled MK Ismail contributed to developing the protocol. Zheng Jing, Xunzhe Yang, Taixiang Wu, and Khaled MK Ismail contributed to developing the review.

Sources of support

Internal sources

• Chinese Cochrane Center, West China Hospital of Sichuan University, China.

External sources

• Chinese Medical Board of New York (CMB), USA.

Declarations of interest

None known

Edited (no change to conclusions)