hum. reprod. update 2012 li 504 24

...........................................................................................................................

Adverse outcomes of Chinesemedicines used for threatenedmiscarriage: a systematic reviewand meta-analysisLu Li1, Li Xia Dou2,3, James P. Neilson3, Ping Chung Leung4

and Chi Chiu Wang1,*1Department of Obstetrics and Gynaecology, The Chinese University of Hong Kong, 1st Floor, Block E, Prince of Wales Hospital, Shatin,New Territories, Hong Kong 2School of Public Health, Peking University, Beijing, China 3Cochrane Pregnancy and Childbirth Group, Instituteof Translational Medicine, University of Liverpool, Liverpool, UK 4Institute of Chinese Medicine, The Chinese University of Hong Kong, Princeof Wales Hospital, Shatin, New Territories, Hong Kong

*Correspondence address. Tel: +852-2632-2810; Fax: +852-2636-0008; E-mail: [email protected] on March 3, 2012; resubmitted on April 20, 2012; accepted on May 2, 2012

table of contents

IntroductionThreatened miscarriageChinese medicines for threatened miscarriage

MethodsStudy characteristicsStudy designSearch strategyData extraction

ResultsIdentication of articlesStudy designs and methodsChinese medicines, type and complianceBaseline, exclusion and follow-upAdverse outcomes

Adverse effects and toxicity of Chinese medicinesFailure of the intervention and complicationsAdverse pregnancy outcomesAdverse perinatal outcomes

DiscussionSummary of evidenceLimitationsDifculties

Conclusions and recommendations

background: Threatened miscarriage is very common in early pregnancy. Chinese medicines have been widely used to prevent spon-taneous pregnancy loss. However, the safety of Chinese medicines is still unknown. A systematic review was performed to identify and de-scribe adverse events of Chinese medicines used for threatened miscarriage.

& The Author 2012. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved.For Permissions, please email: [email protected]

Human Reproduction Update, Vol.18, No.5 pp. 504524, 2012

Advanced Access publication on June 2, 2012 doi:10.1093/humupd/dms025

by guest on June 25, 2015http://hum

upd.oxfordjournals.org/D

ownloaded from

methods: Clinical studies of Chinese medicines for threatened miscarriage were selected. Primary outcomes were occurrence of adverseeffects or toxicity of Chinese medicines. Secondary outcomes were failure of treatment and adverse pregnancy and perinatal outcomes.

results: Thirty-two relevant articles included 9 randomized controlled trials, 1 quasi-randomized controlled trial and 2 controlled trialscomparing Chinese medicines alone or combined medicines with pharmaceuticals and 20 case series with no controls. Sample sizes of eachstudy were generally small. There was variation in Chinese medicine formulation, dosage and duration of treatment. In the pooled rando-mized controlled trials, dry mouth, constipation and insomnia (210%) and intervention failure (3.122.3%), diabetic complications (3%),preterm delivery (5%) and neurodevelopmental morbidity (1.8%) were recorded. Meta-analysis demonstrated that intervention failurewas signicantly lower in the combined Chinese medicines groups than in the Western medicines controls (relative risk 0.46; 95% con-dence interval: 0.300.70, I2 0%). No signicant differences were found between these groups for adverse effects and toxicity or foradverse pregnancy and perinatal outcomes.

conclusions: Studies varied considerably in design, interventions and outcome measures, therefore conclusive results remain elusive.In the absence of placebo-controlled trials, the safety of Chinese medicines for the treatment of threatened miscarriage is unknown. Rigorousscientic and clinical studies to assess the possible risks of Chinese medicines are needed.

Key words: Chinese medicines / miscarriage / adverse events / outcomes

Introduction

Threatened miscarriageThreatened miscarriage is a spontaneous miscarriage present withvaginal bleeding or any bloody vaginal discharge during the rst halfof pregnancy (Cunningham et al., 2005). It is the most common com-plication in early pregnancy. About a quarter of recognized pregnan-cies have this symptom and half of those will end in miscarriage(Everett, 1997). The causes of spontaneous miscarriage are diverseand comprise chromosomal, genetic, anatomical, immunological, hor-monal, infectious and psychological factors (Bulletti et al., 1996). Inmany cases, however, the specic underlying causes have not beenidentied (Orvieto et al., 1991). The risk of miscarriage increaseswith parity as well as maternal and paternal age. After the age of3035 years, fertility declines and the rate of spontaneous miscarriageincreases with a signicant social and economic impact (Brock andHolloway, 1990).

The vaginal bleeding in threatened miscarriage is frequently slightand will subside spontaneously. Clinical care typically includes pelvicexamination to exclude inevitable miscarriage, ultrasound to conrmviable intrauterine pregnancy and laboratory studies to monitorgrowth of the fetus. The vaginal bleeding may persist for days orweeks and progress no matter what is done. Expectant and medicalmanagement has not been shown to provide an additional benet;bedrest, progesterone or chorionic gonadotrophin supplements donot signicantly prevent pregnancy loss (Aleman et al., 2005; Hassand Ramsey, 2008; Devaseelan et al., 2010).

Chinese medicines for threatenedmiscarriageFor more than 5000 years, Traditional Chinese Medicine hasemployed a range of unique clinical practices to maintain health (Gio-vanni, 1997). The practices include acupuncture (inserting ne needlesinto specic points namely acupoints on the body), moxibustion(burning dried mugwort on the acupoints in conjunction with acupunc-ture), food therapy (individualizing dietary recommendations for ahealth condition), Chinese medicines (a collection of crude medicines,

prepared drugs in slices, patent medicines and simple preparations),cupping (creating vacuum in several glass spheres on the body),Qigong (breathing and meditation exercise) and Tuina (bodymassage akin to acupressure). Most of these approaches have beendisseminated to other countries since the sixth century BC (Unschuld,1985). Traditional Chinese Medicine is currently well accepted as amainstream of medical care throughout East Asia and is considereda complementary or alternative medicine in the Western world(Basics, 2010).

Chinese medicine is a common name for Chinese Materia Medica,which has therapeutic properties for medical treatment and healing(Read, 1976). It is considered as a primary modality of internal medi-cine in Traditional Chinese Medicine (Holland, 2000). NowadaysChinese medicines and other herbal medicines are commonly usedto promote maternal and fetal health and to relieve medical problemsduring pregnancy worldwide (Hemminki et al., 1991; Wang et al.,1995; Gharoro and Igbafe, 2000; Gibson et al., 2001; Hollyer et al.,2002; Mantani, 2003; Nordeng and Haven, 2004; Ong et al., 2005;Della et al., 2006; Chuang et al., 2009; Holst et al., 2009). Chinesemedicines are frequently used in threatened miscarriage (Li, 2011) inorder to prevent pregnancy loss (Liu, 2002).

Chinese medicines are regarded by the public and some health-careproviders as gentle and safe (Marcus and Snodgrass, 2005). Althoughthere is no scientic basis for the safety claim, 3050% of pregnantwomen use Chinese medicines to maintain good health and reducethe need for Western medication (Westfall, 2001; Ong et al.,2005). Despite the potential benecial effects of Chinese medicinesin threatened miscarriage (Li et al., 2012), no data are available toconrm the safety of Chinese medicines and no regulations havebeen established to monitor and control its clinical applications inpregnancy. In our animal studies of the commonly used Chinese med-icines for threatened miscarriage, we worried about the identied re-productive toxicity, including fetal resorption, growth restriction andcongenital malformations (Li et al., 2011; Wang et al., 2012). Wehypothesized that Chinese medicines may also be associated withadverse outcomes in human pregnancy. In this study, we havefocused on threatened miscarriage and aimed to identify and describethe adverse events of Chinese medicines by systematic review and

Chinese medicines for threatened miscarriage 505



ownloaded from

meta-analysis, which included adverse effects or toxicity of medicine,failure of intervention and adverse pregnancy and perinatal outcomes.This has not been done previously.

Materials and Methods

Study characteristicsThis systematic review focused on clinical studies involving women withthreatened miscarriage, the most common application of Chinese medicineduring pregnancy (Li, 2011). All published and unpublished studies that eval-uated the adverse effects of Chinese medicines for threatened miscarriagewere considered for review. Clinical studies included both uncontrolledstudies and controlled studies, which compared Chinese medicines or com-bined medicines to a comparison group receiving no therapy (placebo),minimal therapy (hospitalization or bedrest only) or another therapy/treat-ment, were considered. Outcome measures included adverse events inmothers and fetuses.

Study designStudies included both experimental and observational studies: (i) rando-mized control trials (RCTs), (ii) quasi-RCTs if the randomizationmethods were not adequate, (iii) controlled trials if no randomizationwas carried out; (iv) cohort studies if the study methods were not speci-ed or assessed by RCTs and (v) case series and case reports. Observa-tional studies were considered in this review because many adverse eventsare very uncommon and also take a long period to be observed. An inter-vention period of at least 1 week with either short-term or long-termfollow-up and Chinese medicines, or in conjunction with any other inter-vention, as an independent factor on outcomes of health, were includedfor this study. The intervention duration of at least 1 week was basedon capturing as many potentially effective interventional studies as pos-sible. The primary outcome was adverse effects or toxicity of theChinese medicines. Adverse effects and toxicity of Chinese medicineswere dened as harmful and undesired side effects and toxic effectsresulting from the treatment. Secondary outcomes were failure of theinterventions, and adverse pregnancy and perinatal outcomes. Failure ofintervention for threatened miscarriage was dened as inevitable miscar-riage presented with vaginal bleeding and pregnancy loss. Adverse preg-nancy outcomes were dened as pregnancy complications, and maternalmorbidity and mortality, such as preterm labor, post-date delivery, gesta-tional diabetics and hypertension and maternal death. Adverse perinataloutcomes were dened as prenatal and post-natal morbidity and mortal-ity, including intrauterine infection, asphyxia, neurodevelopmental conse-quence, congenital malformation and neonatal death. Studies wereexcluded if (i) no ultrasound examination was provided to conrm theviable pregnancy, (ii) complicated or recurrent miscarriage were includedfor intervention, (iii) various herbal agents and products were included,e.g. green tea and ginger, which are widely used as daily pharmanutrientsfor general health and (iv) the study was not published in either English orChinese.

Search strategyA systematic review of the literature was conducted in December 2011across the following databases: MEDLINE, PubMed, Cochrane Library,International Clinical Trials Registry Platform (ICTRP), EMB Reviews (in-cluding Cochrane Central Register of Controlled Trials (CENTRAL),Cochrane Database of Systematic Reviews, Cochrane Methodology Regis-ter, ACP journal club), EMBASE, CINAHL, NHS Evidence, Wiley Inter-Science, Chinese Clinical Trial Registry, China Journal Net and WanFangdatabase, with no time limit applied to any database. The reference lists

of review articles and eligible studies were hand searched to identifyother potentially eligible studies. Abstracts from conference proceedingswere also considered to identify further studies that were not yet pub-lished. The search terms, adapted when required for each database, toidentify Chinese medicines interventions in threatened miscarriage arepresented in Supplementary data, Table SI.

Data extractionTwo co-authors (L.L. and L.X.D.) independently extracted data from asubset of the articles into data summary tables. Data regarding thestudy design, population(s) studied, assessment of interventions,measure of adverse events (short-term and long-term adverse outcomes),timing of measurement(s) and prevalence of outcomes were extractedfrom each included study. Quality analysis was performed based on theCochrane Handbook of Systematic Reviews and Quality of Reporting ofMeta-Analysis checklist. A formal assessment of study quality was per-formed for each study using a standardized quality assessment formadopted from NewcastleOttawa Quality Assessment Scale (Stang,2010). The intent was to perform a quantitative synthesis of the datawhere possible, dependent on the quality and diversity of the data.Meta-analysis was performed for RCTs and quasi-RCTs, otherwise con-trolled trials were included when there were not enough RCTs orquasi-RCTs. Sensitivity analysis was carried out to explore the effect oftrial quality on overall result. Particularly, xed-effect meta-analyseswould be used for combining data in the absence of heterogeneity. Inthe presence of heterogeneity (I2 . 50%), where results were beingpooled from studies examining different interventions, or where it wasnot clear that the same outcome was being measured in all studies,random-effects meta-analyses would be used. Descriptive methods wereapplied to report and summarize the adverse outcomes of Chinese med-icines for case reports and case series.

Results

Identication of articlesFigure 1 shows the study selection according to PRISMA guidelines(Moher et al., 2009). In total, 591 publications on Chinese medicinesfor threatened miscarriage were identied from the database search,but none were obtained from other sources. After screening thetitles and abstracts, 65 commentary and 67 review articles wereexcluded. One hundred and eight-seven duplicated publicationswere also excluded. Full texts of remained 272 clinical studies wereassessed for study eligibility. Adverse effects of Chinese medicineswere not reported as an outcome measure in 232 studies. Adverseeffects were not due to the intervention (outcomes before interven-tion) in one study (An, 2000), pregnancies of gestational age over30 weeks were recruited in four studies (Wang et al., 2008; Jiang,2009; Meng, 2009; Bi, 2010), recurrent miscarriage was included inone study (Pan, 2008), different pharmaceuticals were compared inone study (Lu et al., 2007) and Chinese medicines were included inthe control group in one study (Zhang, 2009). These 240 studieswere also excluded, leaving 32 clinical studies with full records onthe adverse effects of Chinese medicines for threatened miscarriage.These were selected for quantitative or qualitative synthesis for thereview.

506 Li et al.



ownloaded from

Study designs and methodsTable I and Supplementary data, Table SII shows the characteristicsand main results of 32 selected clinical studies (total participants 3412). The studies, all from China, were published between 1987and 2010. There were 9 RCT studies (total participants 1542), 1quasi-RCT study (total participants 105), 2 non-randomized con-trolled trials (total participants 115) and 20 case series (totalparticipants 1650). No cohort studies, case control studies andcase reports were identied. Table II and Supplementary data,Table SV summarized the quality of the selected studies. Qualityscores were 46 out of 9 in all the RCTs, quasi-RCT and controlled

trials (Supplementary data, Table SIV). Two RCTs clearly indicated therandomization methods (Li et al., 2006; Zhang et al., 2000), while theother 7 RCTs did not report details of the randomization methods(Yang, 1992; Zhang et al., 2005; He and Che, 2007; Teng and Wu,2008; Zhao et al., 2008, 2010; Yue et al., 2009). The randomizationmethod of the quasi-RCTwas inadequate by the visiting date of the par-ticipants only (Song and Zhu, 2007). Two studies both did not mentionif any randomization method was used and were classied as non-randomized controlled studies (Gong and Chen, 1993; Zhou, 2006).The RCTs, quasi-RCT and non-randomized controlled studies con-sisted of two or three intervention groups, Chinese medicines alone,

Figure 1 Study exclusion and inclusion.




ownloaded from

.............................................................................................................................................................................................

Table I Summary of study characteristics and adverse events.

Study TCMintervention

WM comparison Adverseeffects andtoxicity

Interventionfailure andoutcomes

Adversepregnancyoutcomes

Adverse perinataloutcomes

Zhaoet al.(2010)

Formula: ZSRGF poDose: not reportedDosing: BIDDuration: until 10thgestational week

Progesterone imDose: 20 mgDosing: QDDuration: until 10thgestational week

Not reported Not reportedOutcome: inevitableabortion, missedabortion (overall14.8%)

Not reportedOutcome: pretermlabor

TCM 0%WM 0%

Yue et al.(2009)

Formula:TSPSP + STP poDose: 630 gDosing: BIDDuration: until 16thgestational week

Vitamin E poDose: 100 mgDosing: TIDDuration: until 16thgestational weekProgesterone imDose: not reportedDosing: PRNDuration: notreported

Not reported Combined TCM andWM 9%Outcome: notreportedWM 31%Outcome: therapeuticabortion (14.3%)

Combined TCM 0%WM not reported

Combined TCM 0%WM not reported

Teng andWu(2008)

Formula: BSYQD poDose: 520 gDosing: BIDDuration: 14 days

Progesterone imDose: 20 mgDosing: QDDuration: 14 days

Combined TCMand WM 0%WM notreported

Combined TCM andWM 7%Outcome: notreportedWM 20%Outcome: notreported

Not reported Not reported

Zhaoet al.(2008)

Formula: ZSRGF poDose: not reportedDosing: BIDDuration: until 14thgestational week

Progesterone imDose: 20 mgDosing: QDDuration: until 14thgestational week

Not reported Not reported TCM/WM 5%Outcome: stillbirth(groups not reported)

TCM/WM 5%Outcome: fetalventriculomegaly (groupsnot reported)

He andChe(2007)

Formula: JWATD poDose: 330 gDosing: BIDDuration: 10 days

Allylestrenol poDose: 5 mgDosing: BIDDuration: 10 days

Not reported Combined TCM andWM 8%Outcome: notreportedWM 20%Outcome: notreported

Combined TCM 3%Outcome: pretermdeliveryWM 2% Outcome:preterm delivery 1%Postdate pregnancy 1%

Combined TCM 0%WM 0%

Li et al.(2006)

Formula: BSGTD poDose: 530 gDosing: BIDDuration: 10 days

HCG imDose: 2000 UDosing: QDDuration: 10 daysProgesterone imDose: 20 mgDosing: QDDuration: 10 days

Not reported Combined TCM andWM 22.2%Outcome: notreportedWM 29.5%Outcome: notreported

Combined TCM andWM 3%Outcome: diabeticcomplicationsWM 3.7%Outcome: pretermdelivery

Combined TCM 0%WM 0%

Zhang(2005)

Formula: ZSYTP poDose: 5 gDosing: TIDDuration: 14 days

HCG imDose: 2000 UDosing: QDDuration: 14 days

TCM 10%Outcome:dry mouth 8%thick moss 4%constipation 4%insomnia 2%Combined TCMand WM: notreportedWM: notreported

TCM 13%Outcome: notreportedCombined TCM andWM 7.4%Outcome: notreportedWM 15%Outcome: notreported


Continued

508 Li et al.



ownloaded from

.............................................................................................................................................................................................

Table I Continued






Zhanget al.(2000)

TCM 1: YAD poDose: 615 gDosing: TIDDuration: 10 daysTCM 2: STP poDose: not reportedDosing: TIDDuration: 10 days

Progesterone imDose: 1020 mgDosing: QDDuration: 1 weekafter symptomsubsidedVitamin E poDose: 100 mgDosing: QDDuration: 1 weekafter symptomsubsided

Not reported TCM 1: 3.1%Outcome: notreportedTCM 2: 14%Outcome: notreportedWM 44%Outcome: notreported

Not reported TCM 1: 0%TCM 2: not reportedWM not reported

Yang(1992)

Formula: TSPSP poDose: not reportedDosing: not reportedDuration: notreported

Vitamin E,Progesterone,ChlordiazepoxideDose: not reportedDosing: not reportedDuration: notreported

Not reported Not reported Not reported TCM 1.8%Outcome:neurodevelopmentalmorbidityWM 7.4%Outcome:neurodevelopmentalmorbidity

Song andZhu(2007)

Formula: ZXBTD poDose: 210 gDosing: TIDDuration: 7 days

Progesterone imDose: 20 mgDosing: QDDuration: 7 daysVitamin E poDose: 100 mgDosing: TIDDuration: 7 days

TCM 0%WM 0%

TCM 18.5%Outcome: notreportedWM 56.9%Outcome: notreported

Not reported TCM 0%WM 0%

Zhou(2006)

Formula: STP+ SXPpoDose: 615 gDosing: QDDuration: 714 days

HCG imDose: 2000 UDosing: QDDuration: 14 daysProgesterone imDose: 20 mgDosing: QDDuration: 14 days

Not reported TCM 5%Outcome: notreportedCombined TCM andWM 0%Outcome: notreportedWM 30%Outcome: notreported

TCM 0%WM 0%Combined TCM andWM 5%Outcome: pretermdelivery

TCM 0%WM 0%Combined TCM and WM(5%)Outcome: neonatal deathdue to prematurity

GongandChen(1993)

Formula: GSATD poDose: not reportedDosing: QDDuration: notreported

Vitamin E poDose: 60 mgDosing: QDDuration: notreportedVitamin C poDose: 300 mgDosing: QDDuration: notreported

Not reported TCM 16.6%Outcome: notreportedCombined TCM andWM 10.4%Outcome: notreportedWM 33.3%Outcome: notreported


Gu et al.(2008)

Formula: LYND poDose: 10 mlDosing: TIDDuration: 1530days

No comparison TCM 0% TCM 11.4%Outcome: notreported


Xu(2008)

Formula: ZNBTF poDose: 520 gDosing: BIDDuration: 12months aftersymptom subsided

No comparison Not reported TCM 9.2%Outcome: notreported

Not reported TCM 0%

Continued




ownloaded from

.............................................................................................................................................................................................

Table I Continued






Ye andQiu(2008)

Formula: YSGCD poDose: 1030 gDosing: BIDDuration: 47 weeks


Not reported TCM 0%

Luo et al.(2007)

Formula: TEAP poDose: 10 gDosing: TIDDuration: 20 days


TCM 1.7%Outcome: pretermdelivery due tohypertension

TCM 0%

Chou(2002)

Formula:STP+WZD poDose: 1030 gDosing: TIDDuration: till fourthto fth month ofgestation/till delivery


TCM 0% TCM 0.9%Outcome: malformation(not specied)

Chenet al.(2001)

Formula: ZSYTP poDose: 5 gDosing: TIDDuration: .2 weeks

No comparison TCM 6.1%Outcome:nausea (3.9%)Dry mouth,anorexia,constipation(0.9%)

TCM 8.1%Outcome: notreported


Xu(2001)

Formula: STP poDose: 1030 gDosing: BIDDuration: 1 weekafter symptomsubsided


Not reported TCM 0%

Wangand Li(2000)

Formula: JPBSATDpoDose: 515 gDosing: BIDDuration: 310weeks


TCM 0% TCM 0%

Chenand Yun(1999)

Formula: BYD poDose: not reportedDosing: not reportedDuration: notreported


TCM 0% TCM 0%

Cui(1998)

Formula: STP poDose: 1015 gDosing: QDDuration: 12months


TCM 8.5%Outcome: pretermdelivery (6.4%)Premature rupture ofmembranes (2.1%)

TCM 8.5%Outcome: neonatal deathdue to prematurity (6.4%)Asphyxia (2.1%)

Kanget al.(1998)

Formula: STP poDose: not reportedDosing: BIDDuration: 12 weeksafter symptomsubsided


TCM 0% TCM 0%

Chen(1997)

Formula: STP poDose: 1030 gDosing: QDDuration: notreported

No comparison Not reported TCM 5%Outcome: notreported

Not reported TCM 0%

Continued

510 Li et al.



ownloaded from

pharmaceuticals alone and combined Chinese medicines and pharma-ceuticals. All pharmaceuticals were Western medicines, including pro-gesterone and oestrogen derivatives, HCG and vitamins C and E. Mostcase series only had a single Chinese medicines group, except one caseseries included combined Chinese medicines and pharmaceuticals astreatment group (Wang and Wang, 1987). All case series did nothave a control group for comparison.

Chinese medicines, type and complianceAmongst the Chinese medicines studied, there were 19 kinds ofChinese medicine formulae (Table I and Supplementary data, TableSV). Each formula contained wide varieties of 520 different individualChinese medicines. Some of the studies included more than one kindof formula in the intervention (Chou, 2002; Zhou, 2006; Yue et al.,

.............................................................................................................................................................................................

Table I Continued






He(1997)

Formula: STP poDose: 630 gDosing: QDDuration:1430 daysafter symptomsubsided


Not reported TCM 0%

Zhou(1997)

Formula: ATD poDose: 612 gDosing: BID/TIDDuration: notreported

No comparison Not reported Combined TCM andWM 4.6%Outcome: notreported

TCM 0.7%Outcome: pretermdelivery

TCM 1.3%Outcome: neonatal deathdue to prematurity (0.7%)Epilepsy (0.3%),neurodevelopmentalmorbidity (0.3%)

Wu andJi (1994)

Formula: STP poDose: 630 gDosing: BIDDuration: 12weeks/till delivery

No comparison TCM 0% TCM 9%Outcome: notreported


Zhu andLi (1992)

Formula: ATD poDose: 615 gDosing: BIDDuration: notreported


TCM 0% TCM 0%

Tian andLi (1991)

Formula: ATD poDose: 1030 gDosing: QDDuration: notreported


TCM 2.9%Outcome:oligohydramnios andstillbirth

TCM 0%

Li (1989) Formula: STP poDose: 330 gDosing: QDDuration: 15 days


TCM 0% Not reported

WangandWang(1987)

Formula: STP poDose: 1030 gDosing: BIDDuration: 1 monthon average


TCM 0% Not reported

Wu(1987)

Formula: WZD poDose: 1020 gDosing: QDDuration: notreported


Not reported TCM 2.4%Outcome: neonatal deathdue to aspirationpneumonia

Study location, designs, participants details and exclusion criteria are available in online Supplementary data, Table SII. TCM, Chinese medicines group; WM, pharmaceutical group(mainly Western medicines); combined TCM and WM, combined Chinese medicines and pharmaceutical group. ATD, An Tai Decoction; BSGTD, Bu Shen Gu Tai Decoction;BSYQD, Bu Shen Yi Qi Decoction; BYD, Bao Yun Decoction; GSATD, Gu Shen An Tai Decoction; JPBSATD, Jian Pi Bu Shen An Tai Decoction; JWATD, Jiu Wei An Tai Decoction;LYND, Le Yun Ning Decoction; STP, Shou Tai Pill; SXP, Shi Xiao Pill; TEAP, Tai Er An Pill; TSPSP, Tai Shan Pan Shi Pill; WZD, Wu Zi Decoction; YAD, Yun An Decoction; YSGCD, YiShen Gu Chong Decoction; ZNBTF, Zi Ni Bao Tai Formula; ZSRGF, Zi Shen Rou Gan Formula; ZSYTP, Zi Shen Yu Tai Pill; ZXBTD, Zhi Xue Bao Tai Decoction. Details of the Chinesemedicines are provided in Supplementary data, Table SIII. QD, quaque die; BID, bis in die; TID, ter in die; PRN, pro re nata; p.o., per os; i.m., intramuscular.




ownloaded from

.............................................................................................................................................................................................

Table II Summary of study qualities.

Study Design Qualityscorea

Summary Compliance Follow-upb Strength Weakness

Zhao et al.(2010)

RCT 4 Small studyRecruitment from July2007 to March 2009Randomized

100%completed

Short term:intervention failure andoutcomes onlyLong term: pregnancyand perinatal outcomes

RCTHighcomplianceBaselinecomparableLong-termfollow-up

Small sample sizePower calculation notreportedRandomization methodnot reportedAllocation method notreportedExclusion not reportedAdverse effects/toxicitynot studied

Yue et al.(2009)

RCT 4 Small studyRecruitment fromJanuary 2004 toDecember 2008Randomized

100%completed

Short term:intervention failure andoutcomes onlyLong term: pregnancyand perinatal outcomes


Small sample sizePower calculation notreportedRandomization methodnot reportedAllocation method notreportedCombined TCM and WMas subsetMore than one type ofTCM and WM includedAdverse effects/toxicitynot studied

Teng andWu(2008)

RCT 4 Small studyRecruitment fromApril 2007 toNovember 2007Randomized

95.6%completed

Short term: adverseeffects/toxicity andintervention failureLong term: no

RCTHighcompliance

Small sample sizePower calculation notreportedRandomization methodnot reportedAllocation method notreportedBaseline unclearCombined TCM and WMas subsetIntervention failureoutcomes not studiedNo long-term follow-up

Zhao et al.(2008)

RCT 4 Small studyRecruitment fromJanuary 2007 toJanuary 2008Randomized

100%completed

Short term: noLong term: pregnancyand perinatal outcomes


Small sample sizePower calculation notreportedRandomization methodnot reportedAllocation method notreportedNo short term follow-up

He andChe(2007)

RCT 4 Large studyRecruitment: notreportedRandomized

100%completed

Short term:intervention failure onlyLong term: pregnancyand perinatal outcomes

RCTLarge samplesizeHighcomplianceLong-termfollow-up(60%)

Power calculation notreportedRandomization methodnot reportedAllocation method notreportedBaseline unclearExclusion not reportedMixed rst and secondtrimestersCombined TCM and WMas subsetAdverse effects/toxicityand intervention failureoutcome not studied

Continued

512 Li et al.



ownloaded from

.............................................................................................................................................................................................

Table II Continued



Li (2006) RCT 5 Small studyRecruitment fromJanuary 2003 toSeptember 2005Randomized byrandom number table

100%completed



Small sample sizePower calculation notreportedAllocation method notreportedCombined TCM and WMas subsetMore than one type ofWM includedAdverse effects/toxicityand intervention failureoutcome not studied

Zhang(2005)

RCT 4 Small studyRecruitment fromJanuary 2002 toDecember 2004Randomized

100%completed

Short term: adverseeffects/toxicity andintervention failure onlyLong term: no

RCTHighcompliance

Small sample sizeBaseline unclearRandomization methodnot reportedAllocation method wasnot reportedCombined TCM and WMas subsetIntervention failureoutcome not studiedNo long-term follow ups

Zhanget al.(2000)

RCT 5 Large studyRecruitment: notreportedRandomized byrandom table

100%completed

Short term:intervention failure onlyLong term: perinataloutcomes only

RCTLarge samplesizeHighcomplianceBaselinecomparableLong-termfollow-up

Power calculation notreportedAllocation method notreportedExclusion not reportedMore than one type ofTCM and WM includedAdverse effects/toxicityand intervention failureoutcome not studiedNo pregnancy outcomes

Yang(1992)

RCT 6 Large studyRecruitment: notreportedRandomized

100%completed

Short term: noLong term: perinataloutcomes only

RCTLarge samplesizeHighcomplianceLong-termfollow-up

Power calculation notreportedRandomization methodnot reportedAllocation method notreportedBaseline unclearExclusion not reportedMore than one type ofWM includedNo short-term follow-upNo pregnancy outcomes

Song andZhu(2007)

Quasi-RCT 4 Small studyRecruitment fromOctober 2005 toAugust 2006Randomized byvisiting date

100%completed

Short term: adverseeffects/toxicity andintervention failure onlyLong term: perinataloutcomes only


Small sample sizePower calculation notreportedAllocation method notreportedMore than one type ofWM includedIntervention failureoutcomes not studiedNo pregnancy outcomes

Continued




ownloaded from

.............................................................................................................................................................................................

Table II Continued



Zhou(2006)

ControlledTrial

4 Small studyRecruitment fromJanuary 1997 to May2003

100%completed


HighcomplianceLong-termfollow-up

Small sample sizeBaseline unclearExclusion not reportedCombined TCM and WMas subsetPower calculation notreportedRandomization methodnot reportedAllocation method notreportedMore than one type ofTCM and WM studiedAdverse effects/toxicityand intervention failureoutcome not studied

Gong andChen(1993)

ControlledTrial

4 Small studyRecruitment: notreported

100%completed

Short term:intervention failure onlyLong term: no

Highcompliance

Small sample sizePower calculation notreportedBaseline unclearExclusion not reportedRandomization methodnot reportedAllocation method notreportedCombined TCM and WMas subsetMore than one type ofWM includedAdverse effects/toxicityand intervention failureoutcome not studiedNo long-term follow-up

Gu et al.(2008)

Case series Small studyRecruitment fromJune to December2006Prospective study

100%completed


Highcompliance

Low qualitySmall sample sizeAllocation method notreportedExclusion not reportedNo controlIntervention failureoutcome not studiedNo long-term follow ups

Xu (2008) Case series Small studyRecruitment from2000 to 2006Prospective study

100%completed



Low qualitySmall sample sizeAllocation method notreportedExclusion not reportedMixed rst and secondtrimestersNo controlAdverse effects/toxicityand intervention failureoutcome not studiedNo pregnancy outcomes

Continued

514 Li et al.



ownloaded from

.............................................................................................................................................................................................

Table II Continued



Ye andQiu(2008)

Case series Small studyRecruitment fromMarch 2000 to April2006Prospective study

100%completed



Low qualitySmall sample sizeAllocation method notreportedNo controlAdverse effects/toxicityand intervention failureoutcome not studiedNo pregnancy outcomes

Luo et al.(2007)

Case series Small studyRecruitment fromSeptember 2004 toDecember 2006Prospective study

100%completed



Low qualitySmall sample sizeAllocation method notreportedMixed rst and secondtrimestersExclusion not reportedNo controlAdverse effects/toxicityand intervention failureoutcome not studied

Chou(2002)

Case series Small studyRecruitment from1981 to 2001Prospective study

100%completed

Short term:intervention failure onlyLong term: pregnantand perinatal outcomes

No exclusionLong-termfollow-up

Low qualitySmall sample sizeAllocation method wasnot reportedBaseline unclearExclusion not reportedMore than one type ofTCM includedNo controlProlonged treatmentAdverse effects/toxicityand intervention failureoutcome not studied

Chenet al.(2001)

Case series Large studyRecruitment fromMarch to September2000Prospective study

100%completed


Large samplesizeHighcompliance

Low qualityAllocation method notreportedBaseline unclearNo controlIntervention failureoutcomes not studiedNo long-term follow ups

Xu (2001) Case series Small studyRecruitment: notreportedProspective study

100%completed



Low qualitySmall sample sizeAllocation method notreportedNo controlAdverse effects/toxicityand intervention failureoutcome not studiedNo pregnancy outcomes

Continued




ownloaded from

.............................................................................................................................................................................................

Table II Continued



Wang andLi (2000)

Case series Small studyRecruitment: notreportedProspective study

100%completed



Low qualitySmall sample sizeAllocation method notreportedExclusion unclearMixed rst and secondtrimestersNo controlAdverse effects/toxicityand intervention failureoutcome not studied

Chen andYun(1999)


100%completed



Low qualitySmall sample sizeAllocation method notreportedExclusion not reportedNo controlAdverse effects/toxicityand intervention failureoutcome not studied

Cui(1998)


100%completed



Low qualitySmall sample sizeBaseline unclearAllocation method notreportedNo controlAdverse effects/toxicityand intervention failureoutcome not studied

Kang et al.(1998)

Case series Small studyRecruitment fromJanuary 1993 toDecember 1997Prospective study

100%completed



Low qualitySmall sample sizeAllocation method notreportedMixed rst and secondtrimestersNo controlAdverse effects/toxicityand intervention failureoutcome not studied

Chen(1997)


100%completed



Low qualitySmall sample sizeBaseline unclearAllocation method notreportedNo controlAdverse effects/toxicityand intervention failureoutcome not studiedNo pregnancy outcomes

Continued

516 Li et al.



ownloaded from

.............................................................................................................................................................................................

Table II Continued



He (1997) Case series Small studyRecruitment from1989 to 1993Prospective study

100%completed

Shortterm: interventionfailure onlyLong term: perinataloutcomes only


Low qualitySmall sample sizeAllocation method notreportedExclusion not reportedNo controlAdverse effects/toxicityand intervention failureoutcome not studiedNo pregnancy outcomes

Zhou(1997)

Case series Large studyRecruitment fromJanuary 1988 toDecember 1994Prospective study

100%completed


Large samplesizeHighcomplianceLong-termfollow-up

Low qualityAllocation method notreportedExclusion not reportedNo controlCombined TCM and WMAdverse effects/toxicityand intervention failureoutcome not studied

Wu and Ji(1994)

Case series Small studyRecruitment fromDecember 1990 toMay 1994Prospective study

100%completed


Highcompliance

Low qualitySmall sample sizeAllocation method notreportedExclusion not reportedNo controlProlonged treatmentAdverse effects/toxicityand intervention failureoutcome not studiedNo long-term follow ups

Zhu and Li(1992)

Case series Small studyRecruitment fromApril 1987 to April1988Prospective study

100%completed



Low qualitySmall sample sizeAllocation method notreportedExclusion not reportedNo controlAdverse effects/toxicityand intervention failureoutcome not studied

Tian andLi (1991)

Case series Small studyRecruitment fromOctober 1988 to May1990Prospective study

100%completed



Low qualitySmall sample sizeMixed early and lategestationAllocation method notreportedNo controlAdverse effects/toxicityand intervention failureoutcome not studied

Li (1989) Case series Small studyRecruitment from1985 to 1987Prospective study

100%completed

Short term:intervention failure onlyLong term: pregnancyoutcomes only


Low qualitySmall sample sizeAllocation method notreportedExclusion not reportedNo controlAdverse effects/toxicityand intervention failureoutcome not studiedNo perinatal outcomes

Continued




ownloaded from

2009). In addition, the dosage of the Chinese medicines varied fromthe smallest 2 g in one pack per day (Song and Zhu, 2007) to thelargest 30 g in one pack per day (Li et al., 2006). The course of theregimes ranged from 7 days (Wu and Ji, 1994; Song and Zhu, 2007)up to term delivery (Wu and Ji, 1994; Chou, 2002), but sevenstudies did not report the duration of interventions (Wu, 1987; Tianand Li, 1991; Yang, 1992; Zhu and Li, 1992; Chen, 1997; Zhou,1997; Chen and Yun, 1999). For the pharmaceuticals used in RCTs,quasi-RCT and controlled trials, progesterone was commonly used to-gether with Vitamin E for a week (Zhang et al., 2000; Li et al., 2006;Song and Zhu, 2007; Teng and Wu, 2008; Zhao et al., 2008, 2010;Yue et al., 2009) or HCG for 12 weeks (Zhang et al., 2005; Zhou,2006). All studies reported 100% compliance to complete the inter-ventions, except Teng & Wus study (Teng and Wu, 2008) (Table II).

Baseline, exclusion and follow-upThe number of participants in most studies was ,100 (Table I andSupplementary data, Table SIII), only 5 studies included .200 subjects(Yang, 1992; Zhou, 1997; Chen et al., 2001; Zhang et al., 2000; Heand Che, 2007) and 2 studies included .300 subjects (Zhou, 1997;Zhang et al., 2000). The gestational age of participants in moststudies was ,12 weeks of pregnancy, but eight studies included preg-nancies from 12 to 20 weeks (Li, 1989; Tian and Li, 1991; Kang et al.,

1998; Wang and Li, 2000; Luo et al., 2007; Gu et al., 2008; Xu, 2008;Yue et al., 2009), two studies included up to gestational 30 weeks(Zhang et al., 2000; He and Che, 2007) and six studies did notreport the details (Wu, 1987; Yang, 1992; Chen, 1997; Cui, 1998;Chou, 2002; Zhou, 2006). Only 11 studies clearly excluded ectopicpregnancy, molar pregnancy, uterine abnormalities or other medicalcomplications in the subject recruitment (Tian and Li, 1991; Chen,1997; Cui, 1998; Kang et al., 1998; Chen et al., 2001; Zhang et al.,2005; Li et al., 2006; Song and Zhu, 2007; Teng and Wu, 2008;Zhao et al., 2008; Yue et al., 2009). In the other 21 studies, 11 didnot report the exclusion criteria (Zhu and Li, 1992; Wu and Ji,1994; Chen and Yun, 1999; He, 1997; Zhou, 1997, 2006; Wangand Li, 2000; Zhang et al., 2000; Xu, 2001; Chou, 2002; Luo et al.,2007), while the other 10 studies did not provide any information(Wang and Wang, 1987; Wu, 1987; Li, 1989; Yang, 1992; Gongand Chen, 1993; He and Che, 2007; Gu et al., 2008; Xu, 2008; Yeand Qiu, 2008; Zhao et al., 2008). Short-term and long-termoutcome measures were both included in most studies, except twostudies did not measure short-term outcomes (Yang, 1992; Zhaoet al., 2008), while six studies did not measure long-term outcomes(Gong and Chen, 1993; Wu and Ji, 1994; Chen et al., 2001; Zhanget al., 2005; Gu et al., 2008). For short-term outcomes, moststudies followed-up the subjects immediately or up to 2 weeks afterinterventions. For long-term outcomes, most studies followed-up

.............................................................................................................................................................................................

Table II Continued



Wang andWang(1987)


100%completed

Short term:intervention failure onlyLong term: pregnancyoutcomes only


Low qualitySmall sample sizeExclusion not reportedMixed rst and secondtrimestersCombined TCM and WMas subsetAllocation method notreportedNo controlAdverse effects/toxicityand intervention failureoutcome not studiedNo perinatal outcomes

Wu(1987)

Case series Small studyRecruitment date notreportedProspective study

100%completed



Low qualitySmall sample sizeAllocation method notreportedBaseline unclearExclusion not reportedNo controlAdverse effects/toxicityand intervention failureoutcome not studiedNo pregnancy outcomes

RCT, randomized controlled trial.aQuality score: total score obtained from the quality assessment according to NewcastleOttawa Quality Assessment Scale (see Supplementary data, Table SIV). Short-term follow-upinclude adverse effects and toxicity during intervention or intervention failure and outcomes; long-term follow-up include adverse pregnancy or perinatal outcomes.bParameters and durations of the long-term and short-term follow-up are available in Supplementary data, Table SV.

518 Li et al.



ownloaded from

the subjects around a month after intervention or until delivery,except for seven studies which prolonged the follow-up to yearsafter delivery (Yang, 1992; Zhu and Li, 1992; He, 1997; Wang andLi, 2000; Zhang et al., 2000; Xu, 2001, 2008).

Adverse outcomesIn these 32 selected trials, the reported adverse outcomes includedthe adverse effects and toxicity of Chinese medicines, failure and com-plications of the intervention, adverse pregnancy outcomes andadverse perinatal outcomes. Owing to the substantial diversity ofthe selected studies with respect to quality, methodology, interventionvariations, sample characteristics and outcome reporting, only a fewstudies meet the requirements for the planned quantitative data syn-thesis, comparing the adverse outcomes by meta-analysis. We alsoprovided a qualitative data synthesis, presenting the major results ofthe studies in the text and summary tables.

Adverse effects and toxicity of Chinese medicinesIn most studies, adverse effects and toxicity of the Chinese medicinesas one of the short-term outcome measures were not studied(Table I). Only 2 RCT studies with 130 participants (Zhang et al.,2005; Teng and Wu, 2008) reported adverse effects and toxicity ofcombined Chinese and Western medicines or Chinese medicinesalone, but not Western medicines alone. No adverse effects werereported under combined medicines, but minor adverse effectswere observed in 28% of women with Chinese medicine alonetreatment. The effects included dry mouth, constipation and insomnia.No adverse effects and toxicity were identied in the quasi-RCT ineither intervention (Song and Zhu, 2007). The controlled trials didnot report adverse effects and toxicity after interventions (Gong andChen, 1993; Zhou, 2006). No meta-analysis was performed.

Two case series reported no adverse effect and toxicity afterChinese medicines treatment (Wu and Ji, 1994; Gu et al., 2008).Only one case series identied 0.93.9% of some gastrointestinaleffects, including nausea, dry mouth, anorexia and constipation(Chen et al., 2001).

Failure of the intervention and complicationsTwo RCTs (Zhang et al., 2000; Zhang et al., 2005) and 1 quasi-RCT(Song and Zhu, 2007) with 778 participants compared Chinese med-icines alone with Western medicines alone. Failure of Chinese medi-cine treatment occurred in 3.118.5%, while failure of Westernmedicine treatment occurred in 1556.9%; however, none reportedthe details of the complications of threatened miscarriage after thefailure. In these three studies, there was signicant heterogeneity(x2 20.46, df 2, I2 90%, P, 0.0001), which indicated that themagnitude of the effects among the studies was not consistent, orthe data might be at risk of allocation bias. Meta-analysis showedthat Chinese medicine alone resulted in no signicant difference inintervention failures compared with Western medicine alone [relativerisk (RR) 0.24; 95% condence interval (CI): 0.061.00, I2 90%,random-effects model, evidence from three RCTs and quasi-RCTs,Fig. 2a). Sensitivity analysis was carried out to explore the impact ofthe quality of original trials, and the results showed no differencewhether the quasi-RCT study was included or not in themeta-analysis.

Five RCTs studies (Zhang et al., 2005; Li et al., 2006; He and Che,2007; Teng and Wu, 2008; Yue et al., 2009) with 482 participantscompared combined medicines with Western medicines. Failure ofcombined treatment occurred in 722.3%, while failure of Westernmedicine treatment occurred in 1531%; however, no details of thecomplications of threatened miscarriage were available. In these vestudies, there was no signicant heterogeneity (x2 2.86, df 4,I2 0%, P 0.58). Meta-analysis showed the incidence of interven-tion failure in combined medicines was signicantly lower than thatin Western medicines alone (RR 0.46; 95% CI: 0.300.70, I20%, xed-effect model, evidence from 5 RCTs and quasi-RCTs,Fig. 2a).

Two controlled trials (Gong and Chen, 1993; Zhou, 2006) with 115participants compared combined medicines, Chinese medicines aloneand Western medicines alone. The incidences of intervention failurewere recorded as 010.4, 516.6 and 3033%, respectively. Inthese two studies, there were no signicant heterogeneity in eithercomparison (x2 0.771.20, df 1, I2 017%, P 0.270.38).Meta-analysis showed that there was no signicant differencebetween Chinese medicines alone and Western medicines aloneinterventions (RR 0.30; 95% CI: 0.090.99, I2 0%, xed-effectmodel, evidence from two controlled trials, Fig. 2a) or between com-bined medicines and Western medicines alone interventions (RR 0.21; 95% CI: 0.060.67, I2 17%, xed-effect model, evidencefrom two controlled trials, Fig. 2a).

Twenty case series with 1532 participants studied Chinese medi-cines alone. Chinese medicines intervention failed to improve miscar-riage symptoms and prevent miscarriage in 220% cases; however,their clinical outcomes after the intervention failure had not been fol-lowed at all.

Adverse pregnancy outcomesAdverse pregnancy outcomes were only reported in ve RCT studies(Li et al., 2006; He and Che, 2007; Zhao et al., 2008; Yue et al., 2009;Zhao et al., 2010) and one controlled trial (Zhou, 2006). However,the incidents were not described separately in either interventiongroup in two RCT studies (Zhao et al., 2008, 2010) and in theWestern medicines group in another RCT study (Yue et al., 2009).One RCT study (He and Che, 2007) and the controlled trial (Zhou,2006) compared the adverse pregnancy outcomes in Chinese medi-cines alone and Western medicines alone; and the remaining oneRCT study compared the adverse pregnancy outcomes of combinedmedicines and Western medicines alone (Li et al., 2006). Theadverse pregnancy outcomes included preterm labour (1.75%),postdate delivery (1.7%) and diabetic pregnancy (3.7%). Resultsfrom one RCT comparing Chinese medicines alone versus Westernmedicines alone (He and Che, 2007) and one RCT comparing com-bined medicines versus Western medicines alone (Li et al., 2006) onthe adverse pregnancy outcome were not conclusive. Since insufcientRCT studies were available for meta-analysis, controlled trials werealso included for further analysis. Meta-analysis of a single RCTstudy (Li et al., 2006) and a single controlled trial (Zhou, 2006) onthe adverse pregnancy outcome of Chinese medicines alone andWestern medicines alone (RR 1.65, 95% CI: 0.2212.07, I2 0%,xed-effect model, evidence from 1 RCT and 1 controlled trial,Fig. 2b) or combined medicines and Western medicines alone(Fig. 2b) could not be conclusive.




ownloaded from

Figure 2 Adverse outcomes of Chinese medicines for threatened miscarriage. (A) Outcome: failure of the interventions. Study design: RCT andquasi-RCT . (B) Outcome: failure of the intervention. Study design: controlled trials. (C) Outcome: adverse pregnancy outcomes. Study design: RCTand controlled trials.

520 Li et al.



ownloaded from

Most case series did not report adverse pregnancy outcomes. Onlyseven cases series reported no adverse pregnancy outcome afterChinese medicine treatment (Wang and Wang, 1987; Li, 1989; Zhuand Li, 1992; Kang et al., 1998; Chen and Yun, 1999; Wang and Li,2000; Chou, 2002). Another four case series identied adverse preg-nancy outcomes after Chinese medicine treatment (Tian and Li, 1991;Zhou, 1997; Cui, 1998; Luo et al., 2007). Preterm deliveries were alsoidentied in three case series (Zhou, 1997; Cui, 1998; Luo et al., 2007)in 0.76.4%. Other adverse maternal outcomes were identied in twocase series, including premature rupture of membranes in 2.1% in onecase series (Cui, 1998) and stillbirths in 2.9% in other case series (Tianand Li, 1991).

Adverse perinatal outcomesAdverse perinatal outcomes were reported in seven RCT studies(Yang, 1992; Zhang et al., 2000; Li et al., 2006; He and Che, 2007;Zhao et al., 2008, 2010; Yue et al., 2009), one quasi-RCT study(Song and Zhu, 2007) and one controlled trial (Zhou, 2006).However, the incidents were not described in the control group inthree RCT studies (Zhang et al., 2000; Zhao et al., 2008; Yue et al.,2009). In four RCT and one quasi-RCT studies (Li et al., 2006; Heand Che, 2007; Song and Zhu, 2007; Zhao et al., 2008, 2010), noadverse perinatal outcomes were identied in either interventiongroup. No meta-analysis was performed. Neurodevelopmental mor-bidity was conrmed in one RCT study (Yang, 1992), the incidencewas 1.8 and 7.4% with Chinese medicines alone and with Westernmedicines alone, respectively. Neonatal death due to prematuritywas identied in the controlled study (Zhou, 2006), the incidentwas 5% in the combined medicines group, but not in the Chinesemedicines alone and Western medicines alone groups.

Most case series reported adverse perinatal outcomes afterChinese medicines treatment, except ve case series (Wang andWang, 1987; Li, 1989; Wu and Ji, 1994; Chen et al., 2001; Guet al., 2008). No adverse perinatal outcomes were reported in the11 case series, but neonatal death due to prematurity, asphyxia or in-fection were identied in three case series with an incidence rateranging from 0.7 to 6.4% (Wu, 1987; Zhou, 1997; Cui, 1998), whilean unspecied malformation was recorded in one case series withan incidence rate of 0.9% (Chou, 2002) and epilepsy and abnormal in-tellectual development in one case series with an incidence rate of0.3% (Zhou, 1997).

Discussion

Summary of evidenceIn this review, adverse events of Chinese medicines for threatenedmiscarriage, including adverse effects or toxicity, failure of interven-tion, adverse pregnancy and perinatal outcomes, were studied. Datawere pooled from studies that were clinically very heterogeneous;there were large differences across studies in study design, interven-tions and outcome measures. Furthermore, data were often statistic-ally herterogeneous as reected by an I2 value .50%. As aconsequence the results presented should be considered with care.Meta-analysis demonstrated that intervention failure was signicantlylower in the combined Chinese medicines groups than in theWestern medicines controls. No signicant differences were found

in adverse effects and toxicity, or in adverse pregnancy and perinataloutcomes.

This review identied the adverse maternal and perinatal outcomeof Chinese medicines for threatened miscarriage. Preterm prematurerupture of membranes (PPROM) was recorded in a study with inci-dence of 2.1%. Prematurity and associated neonatal mortality wererecorded in several studies with an incidence ranging from 0.7 to6.4%. It is well known that threatened miscarriage is associated withthe risk of suboptimal pregnancy outcome (van Oppenraaij et al.,2009; Saraswat et al., 2010). Women with threatened miscarriagehave a signicantly higher incidence of antepartum haemorrhage(overall 1.2%). They are also more likely to experience PPROM(overall 1.9%), preterm delivery (overall 11.0%) and to have babieswith an intrauterine growth restriction (overall 8.3%). First-trimesterbleeding is also associated with signicantly higher rates of perinatalmortality (overall 1.8%) and low and/or very low birthweight babies(overall 11.0%, odd ratios . 2.0). In comparison, after treatment ofthreatened miscarriage with Chinese medicines, no antepartum haem-orrhage, intrauterine growth restriction or low birthweights wererecorded and the prematurity rate was relatively lower, but thePPROM rate was slightly higher. The neonatal death rate due to pre-maturity with respect to rst trimester bleeding is 2.68.5% (Wil-liams et al., 1991), which is comparable to our current review.These results suggest that Chinese medicines treatment for threa-tened miscarriage may not be associated with an increased risk ofpreterm delivery and neonatal mortality.

On the other hand, malformations were reported in 0.9% of babieswhose mothers had taken Chinese medicines for threatened miscar-riage. Unfortunately, the details of the malformations were not speci-ed. Although at least half of threatened miscarriages are associatedwith chromosomal anomalies of the conceptus, the risk of birthdefects does not appear to be increased signicantly. The overall inci-dence rates of congenital anomalies in women with and withoutvaginal bleeding in rst trimester are 3.2 and 2.7%, respectively (Sar-aswat et al., 2010). The results suggest that Chinese medicines forthreatened miscarriage may not be associated with increased risk ofcongenital malformation.

LimitationsVery few clinical studies of Chinese medicines for threatened miscar-riage were eligible for this review. Over 90% of the identied studiesdid not include adverse effects as one of the study outcome measures.It may be because the incidence of adverse events was indeed too rareor because the awareness on the adverse effects was actually too low.In our qualitative analysis, a low rate of adverse events after maternalexposure to Chinese medicines for threatened miscarriage wasrecorded, but the incidences were not too low to be easily missed.Hence, lack of awareness on the safety issue of Chinese medicinesin general could be the main reason of limited studies being availablefor systematic review.

Apart from limited records in adverse outcomes, study designswere also restricted. No RCTs were placebo controlled and oneRCT had non-placebo controls (Yang, 1992). Of the RCTs, onlytwo were adequately randomized, while the other RCTs only men-tioned with participants were randomly received different treat-ments. The quasi-RCT used an inadequate randomization method




ownloaded from

by visiting date. All 20 case series studies had no controls for com-parison, allocation methods were not described and their qualityscores were not high. Furthermore, the study results were question-able. Though all the studies had high compliance and drop-out rateswere not high, sample sizes of the selected studies were still verysmall. Some important demographic data and study exclusion criteriawere not provided. Different studies used different Chinese medicineformulae to treat threatened miscarriage and also there were largevariations in the dose, dosing and duration of the interventionamongst the studies. Most studies followed-up the pregnancy until de-livery, but the outcome parameters in the pregnancy and perinatalcomplications were rather inconsistent. Few studies monitoredadverse effects and toxicity of Chinese medicines, and complicationof miscarriage was unknown. Owing to the limited number of RCTsand the clinical heterogeneity between studies, additional meta-analysis to evaluate the adverse effects of Chinese medicine for threa-tened miscarriage was not available.

DifcultiesUnlike Western herbalism, Chinese medicines include many animalmaterials and even mineraloid remedies as well as medicinal herbs(National Phamacopoeia Committee, 2005). In addition, Chinesemedicines are formulated and individualized. A typical formula maycontain 325 Chinese medicines. Most of the formula are processedby decoction in boiling water for hours and are orally administered as asoup (Scheid, 2002), though it can be supplied as powders, solublegranules and tablets nowadays. As each Chinese medicine has itsown properties and potential interactions, the application of formu-lized and individualized medication enhance the therapeutic actionsof some herbs and all the herbs collaborate in each other for the treat-ment. However, the adverse effects and toxicity of Chinese medicinesmay vary in different combinations, preparations and individuals. It isdifcult to identify which Chinese medicine contributes to a specicadverse effect.

Conclusions and recommendationsThis is the rst systematic review to evaluate the clinical studies ofChinese medicines for threatened miscarriages in identifying their po-tential adverse effects on mothers and fetuses. Unfortunately, the evi-dence regarding the use of Chinese medicines for threatenedmiscarriage on adverse effects/toxicity, and adverse pregnancy andperinatal outcomes, is limited. Studies vary considerably in design,interventions and outcome measures; therefore, conclusive resultsremain elusive. Rigorous scientic and clinical studies are necessaryto conrm the risk of Chinese medicines.

The active ingredients of the Chinese medicines are also chemicalsthat are similar to prescription pharmaceuticals. Chinese medicinesare not free of risk; similar to Western pharmaceutical medicines,they have the potential to cause adverse effects. Thus, Chinese med-icines in Traditional Chinese Medicine may not only result in maternalmanifestations that indirectly affect fetal health, but they may alsoharm the fetus directly. Despite variations in clinical practice andtherapeutic prescription, Chinese medication in Traditional ChineseMedicine should comply with the same modern pharmacological prin-ciples as Western Medicine. Chinese medicines may be benecial, butmay also adversely affect both mothers and fetuses in utero.

International regulations have not been designed or specied to cat-egorize these Chinese medicines for use in pregnancy. Detailed repro-ductive toxicity and pharmacotoxicity studies that assess the potentialrisk of Chinese medicines during pregnancy are still not available, justas conventional medications have not been thoroughly tested in preg-nancy. We hope that more comprehensive and systematic experi-ments will be conducted in the near future. Until more reliable andscientic research data become available, clinicians should appraiseboth the risks and benets before recommending these medicinesto pregnant women. Both Chinese and Western physicians should ex-plicitly document the history of the use of any Chinese medicationsduring pregnancy to prevent and recognize potential serious problemsassociated with their use. They should also encourage discontinuationof Chinese medications during pregnancy to avoid adverse eventsbefore sufcient safety evidence become available. We recommendmore systematic investigations of the safety implications of the useof Chinese medicines in animals. More studies and clinical trials inhumans with a larger sample size are also mandatory.

Supplementary dataSupplementary data are available at http://humupd.oxfordjournals.org/.

Authors rolesL.L. and L.X.D. are responsible as rst and second assessors for studyidentication, screening, eligibility and inclusion, data acquisition andstatistical analysis and drafting of the manuscript. C.C.W. is respon-sible as third assessor for study eligibility and inclusion, conceptionof the project, design of the study, research funding and drafting themanuscript. J.P.N. and P.C.L. were responsible for interpreting theresults and approving the nal manuscript.

FundingThe study is partially supported by the Health and Health Service Re-search Grant from Food and Health Bureau, Hong Kong Special Ad-ministration Region (HHSRF 06070511 and 10110901) to C.C.W.and P.C.L. L.L received the Hop Wai Scholarship and Zi Ying Scholar-ship from the Institute of Chinese Culture and Postgraduate StudentGrants for Overseas Academic Activities from the Graduate School,The Chinese University of Hong Kong to attend Cochrane trainingworkshops in Oxford and Freiburg and to study in the Cochrane Preg-nancy and Childbirth Group at University of Liverpool and the Yu ToSang Memorial Scholarship 2008/2009 and 2009/2010 from theChinese University of Hong Kong to pursue her PhD study.

Conict of interestNone.

ReferencesAleman A, Althabe F, Belizan J, Bergel E. Bed rest during pregnancy for preventing

miscarriage. Cochrane Database Syst Rev 2005;18:CD003576.

522 Li et al.



ownloaded from

An LY. Chinese medicines for threatened miscarriage. Henan J Chin Med 2000;15:5859 [Chinese].

Basics CAM. National Institutes of Health. U.S. Department of Health and HumanServices. NCCAM Publication No. D347, 2010.

Bi LM. Observations on the outcomes of treatment for threatened miscarriage. ChinCom Doc 2010;17:99 [Chinese].

Brock DJ, Holloway S. Fertility in older women. Lancet 1990;335:1470.Bulletti C, Flamigni C, Giacomucci E. Reproductive failure due to spontaneous

abortion and recurrent miscarriage. Hum Reprod Update 1996;2:118136.Chen CQ. Clinical observation study of shou tai pill for threatened miscarriage. Pract

Chin J Integr Trad West Med 1997;10:2186 [Chinese].Chen QF, Yun XY. Bao Yun decoction for threatened miscarriage. Maternal Child

Health Care Chin 1999;6:375 [Chinese].Chen Q, Zhao RL, Ma YY, Guo YY, Hu YF, Weng XY, Gai MY, Huang XH,

Zhou LH, Liu WJ et al. 231 clinical observations on Zishen Yutai pill forthreatened miscarriages. Chin J Perinat Med 2001;4:85 [Chinese].

Chou HG. Clinical observation on 112 cases of abortion treated by supplementedshoutai wuzi soup. J Youjiang Med College Nationalities 2002;2:301302 [Chinese].

Chuang CH, Chang PJ, Hsieh WS, Tsai YJ, Lin SJ, Chen PC. Chinese herbalmedicines use in Taiwan during pregnancy and the postpartum period: apopulation-based cohort study. Int J Nurs Stud 2009;46:787795.

Cui SH. Treatment on 47 cases of abortion. Shanxi J Trad Chin Med 1998;19:242[Chinese].

Cunningham G, Leveno KL, Bloom SL, Hauth JC, Gilstrap LC III, Wenstrom KD.Williams Obstetrics, 22nd edn. New York: Mcgraw-Hill Professional Publishing,2005.

Della AF, Angela D, Gemma W, Melissa B, Elizabeth M. Herbal medicine useduring pregnancy in a group of Australian women. BMC Pregnancy Childbirth2006;6:21.

Devaseelan P, Fogarty PP, Regan L. Human chorionic gonadotrophin for threatenedmiscarriage. Cochrane Database Syst Rev 2010;12:CD007422.

Everett C. Incidence and outcome of bleeding before the 20th week of pregnancy:prospective study from general practice. BMJ 1997;315:324.

Gharoro EP, Igbafe AA. Pattern of drug use amongst antenatal patients in Benin City,Nigeria. Med Sci Monit 2000;6:8487.

Gibson PS, Powrie R, Star J. Herbal and alternative medicine use during pregnancy: across-sectional survey. Obstet Gynecol 2001;97:S44S45.

GiovanniM. The Foundations of ChineseMedicine: AComprehensive Text for Acupuncturistsand Herbalists. Pennsylvania, USA: Churchill Livingstone Philadelphia, 1997.

Gong LL, Chen DY. Integrative treatment of threatened miscarriage and study ofbirth healthanalysis of 62 cases. Chin J Birth Health Heredity 1993;2:7576[Chinese].

Gu YT, Zhang GR, Gong G. 35 cases of Le Yun Ning decoction for threatenedmiscarriage. J Qiqihar Med College 2008;29:433 [Chinese].

Haas DM, Ramsey PS. Progestogen for preventing miscarriage. Cochrane DatabaseSyst Rev 2008;2:CD003511.

He YP. Shou Tai pill for threatened miscarriage. Guangzhou J Chin Med College 1997;14:8486 [Chinese].

He ZH, Che PF. Clinic curative effects on integrated Chinese and western medicinesfor threatened miscarriage & the childbearing results. Pract Chin J Integr Trad WestMed 2007;2:3233 [Chinese].

Hemminki E, Mantyranta T, Malin M, Koponen P. A survey on the use of alternativedrugs during pregnancy. Scand J Soc Med 1991;19:199204.

Holland A. Voices of Qi: An Introductory Guide to Traditional Chinese Medicine.California, USA: North Atlantic Books, 2000.

Hollyer T, Boon H, Georgousis A, Smith M, Einarson A. The use of CAM by womensuffering from nausea and vomiting during pregnancy. BMC Comp Altern Med 2002;2:16.

Holst L, Wright D, Haavik S, Nordeng H. The use and the user of herbal remediesduring pregnancy. J Altern Comp Med 2009;15:787792.

Jiang YH. Study of treatments for threatened miscarriage. Chin J Mod Drug Appl 2009;3:7273 [Chinese].

Kang ZY, Sun GL, Fu CH. Shou tai pill for threatened miscarriage86 cases. Henan JChin Med 1998;13:4344 [Chinese].

Li L, Dou L, Leung PC, Wang CC. Chinese herbal medicines for threatenedmiscarriage: review. Cochrane Database Syst Rev 2012;5:185.

Li L. Efcacy and safety of Chinese medicines for pregnancy. PhD Thesis. Hong Kong:The Chinese University of Hong Kong, 2011.

Li L,TangLY,ManGCW,YeungHY, LauBS, LeungPC,WangCC.Potential reproductivetoxicity of Largehead Atractylodes Rhizome, the most commonly used Chinesemedicine for threatened miscarriage. Hum Reprod 2011; 26:32803288.

Li L, Zhou ZX, Liu JY. Clinical studies of Bu Shen Gu Tai Decoction for threatenedmiscarriage. Guangxi Chin Med 2006;29:1718 [Chinese].

Li MY. Clinical study on shou tai pill for threatened miscarriage. Tianjin J Trad ChinMed 1989;3:6 [Chinese].

Liu RF. Clinical observation of integrated traditional and western medicine. Chin JIntegr Trad West Med 2002;22:6869 [Chinese].

Lu Z, Jin S, Chen H. Clinical observation on Chinese integrative medicine in thetreatment of threatened abortion. Chin Med Herald 2007;11:7374 [Chinese].

Luo HT, Zhao SP, Yuan HX. 58 cases of ultrasound observations on threatenedabortions with PPS. J Shanxi College Trad Chin Med 2007;30:26 [Chinese].

Mantani N. Effect of Seihai-to, a Kampo medicine, in relapsing aspiration pneumoniaan open-label pilot study. Phytomedicine 2003;9:195201.

Marcus DM, Snodgrass WR. Do no harm: avoidance of herbal medicines duringpregnancy. Obstet Gynecol 2005;105:11191122.

Meng XY. Outcomes of 156 patients with threatened miscarriage. Chin Com Doc2009;17:167 [Chinese].

Moher D, Liberati A, Tetzlaff J, Altman DG., The PRISMA Group. Preferredreporting items for systematic reviews and meta-analysis: the PRISMAstatement. PLoS Med 2009;6:e1000097.

National Phamacopoeia Committee. Chinese Pharmacopeia, 8th edn. Beijing:Chemical Industry Publishing House, 2005. Appendix 1111.

Nordeng H, Havnen G. Use of herbal drugs in pregnancy: a survey among 400Norwegian women. Pharmacoepidemiol Drug Safety 2004;13:371380.

Ong CO, Chan LY, Yung PB, Leung TN. Use of traditional Chinese herbal medicineduring pregnancy: a prospective survey. ActaObstet Gynecol Scand 2005;84:699700.

Orvieto R, Achiron A, Ben-Rafael Z, Achiron R. Intravenous immunoglobulintreatment for recurrent abortions caused by antiphospholipid antibodies. FertilSteril 1991;56:10131020.

Pan MQ. The effects of Chinese and Western medicines for threatened miscarriage.Chin West Med 2008;137:166168 [Chinese].

ReadBE (ed.).ChineseMaterialMedica. Taipei: SouthernMaterialsCenter, 19761977.Saraswat L, Bhattacharya S, Maheshwari A, Bhattacharya S. Maternal and perinatal

outcome in women with threatened miscarriage in the rst trimester: asystematic review. BJOG 2010;117:245257.

Scheid V. Chinese Medicine in Contemporary China: Plurality and Synthesis. Durham,North Carolina, USA: Duke University Press, 2002.

Song YL, Zhu LP. The fetus protection of Zhixue Baotai decoction on women ofearly threatened abortion with dark area surrounding pregnancy sac. Chin JIntegr Trad West Med 2007;27:10251028 [Chinese].

Stang A. Critical evaluation of the Newcastle-Ottawa scale for the assessment of thequality of nonrandomized studies inmeta-analyses.Eur J Epidemiol 2010;25:603605.

Teng J, Wu XQ. Bu Shen Yi Qi formula and Progesterone for threatenedmiscarriage. Master Thesis. Hubei Chin Med College, 2008, 110 [Chinese].

Tian ZL, Li YZ. Combined Chinese medicine and Western medicine forthreatened abortion on 90 patients. Chin J Integr Trad West Med 1991;1:2223[Chinese].

Unschuld PU. Medicine in China: A History of Ideas. CA, USA: University CaliforniaPress, 1985.

van Oppenraaij RH, Jauniaux E, Christiansen OB, Horcajadas JA, Farquharson RG,Exalto N; ESHRE Special Interest Group for Early Pregnancy (SIGEP). Predictingadverse obstetric outcome after early pregnancy events and complications: areview. Hum Reprod Update 2009;15:409421.

Wang CC, Li L, Tang LY, Leung PC. Safety evaluation of commonly used Chineseherbal medicines during pregnancy in mice. Hum Reprod 2012 doi:10.1093/humrep/des180.

Wang FS, Wang SQ. Clinical study on shou tai pill for threatened miscarriage61cases. Tianjin J Trad Chin Med 1987;1:10, 21, 22 [Chinese].

Wang XJ, Shi BZ, Lin GW, Wang Q. A survey on applications of drugs among 416pregnant women in ShangHai. Chin J Pharmacoepidemiol 1995;4:167169 [Chinese].

Wang YP, Lin AM, Dong X, Zhou YH. Observations of treatment for threatenedmiscarriage130 cases. Chin Pract Obstet Gynecol J 2008;12:936 [Chinese].

Wang ZM, Li ZF. Jian Pi Bu Shen An Tai decoction for threatened miscarriage. ShanxiJ Chin Med 2000;21:545 [Chinese].

Westfall RE. Herbal medicine in pregnancy and childbirth. Adv Ther 2001;18:4755.




ownloaded from

Willians MA, Mittendorf R, Lieberman E, Monson RR. Adverse infant outcomesassociated with rst-trimester vaginal bleeding. Obstet Gynecol 1991;78:1418.

Wu G, Ji X. Clinical observation on Shoutai pill for threatened miscarriages. HenanMed Info 1994;2:2728 [Chinese].

Wu ZG. Clinical observation on 41 cases of threatened abortion treated by shoutaipill the addition and subtraction. Central Plains Med J 1987;3:39 [Chinese].

Xu J. Chinese medicines for threatened miscarriage. ZheJiang J Integr Trad West Med2008;18:63 [Chinese].

Xu YH. Shou Tai pill for threatened miscarriage. Primary J Chin Materia Med2001:15:6061 [Chinese].

Yang CL. Study on intellectual development of children with maternal exposure totreatments of threatened miscarriage. Chin J Integr Trad West Med 1992;3:175[Chinese].

Ye XY, Qiu XL. Yi Shen Gu Chong decoction for early threatened miscarriage. J NewChin Med 2008;40:8586 [Chinese].

Yue P, Chen ZX, Jie FE. Combined Chinese medicines and western medicines forthreatened miscarriage. Gansu J Trad Chin Med 2009;22:4445 [Chinese].

Zhang LL. Combined Chinese medicines for threatened miscarriage. Chin Fore MedTrea 2009;34:89 [Chinese].

Zhang JF, Zhang YF, Liu GZ, Feng QJ. Clinical and experimental study on Yunangranule in treating threatened abortion. Chin J Integr Trad West Med 2000;4:251254 [Chinese].

Zhang R, Li GC, Chen ZL, Ma YL. Zi Shen Yu Tai Pill for threatened miscarriage.Chin Herb Med 2005;28:11441146 [Chinese].

Zhao W, Chen X, Zhang LM. Clinical study of serum CA125: Zi Shen Rou Ganformula for early threatened abortion. Sichuan J Trad Chin Med 2008;11:8889[Chinese].

Zhao W, Chen X, Shen WW, Zhang LM. Zi Shen Rou Gan formula for earlythreatened abortionpreliminary study of serum CA125. Liaoning J Trad ChinMed 2010;7:12821283 [Chinese].

Zhou Y. Clinical observation on 305 cases of threatened miscarriage treated byintegrated traditional Chinese and Western medicine. Pract J Integr Chin ModMed 1997;10:1614 [Chinese].

Zhou Y. Clinical observation on Shoutai Pill and Shixiao Powder for the treatment ofthreatened abortion with kidney deciency and blood stasis. J Guangzhou Uni TradChin Med 2006;23:2529 [Chinese].

Zhu YC, Li GW. An tai decoction for threatened miscarriage67 cases. Shanghai JChin Med 1992;1:1011 [Chinese].

524 Li et al.



ownloaded from

hum. reprod. update 2012 li 504 24

Documents