Complementary Therapies in Medicine (2014) 22, 954964

Available online at www.sciencedirect.com

ScienceDirect

j ourna l ho me pa g e: www.elsev ierhea l th .com/ journa ls /c t im

Safet al medicinein prexpeand t

Axel WJosef H

a Working GBerufsorganb UniversityAvailable on

KEYWOChinese medicinePregnancEmbryotFetotoxiTeratogeExperimHuman d

CorrespoE-mail ad

http://dx.d0965-2299/(http://creaegnancyRe-evaluation ofrimental data of two animal studieshe clinical experience

iebrechta,, Wilhelm Gausb, Simon Beckera,ummelsbergera, Kirsten Kuhlmanna

roup Safety of Chinese Herbal Medicine During Pregnancy of the Schweizerischeisation fr TCM and the Societas Medicinae Sinensis, Bundesallee 141, D-12161 Berlin, Germany

of Ulm, Institute of Epidemiology and Medical Biometry, D-89081 Ulm, Germanyline 19 August 2014

RDSherbal;y risks;oxicity;city;nicity;ental studies;ata

SummaryIntroduction: Chinese herbal medicine is an increasingly popular worldwide medical therapywhich also has an impact in pregnancy. However, the question of its drug safety during pregnancyremains unresolved. Potential problems include teratogenicity, abortion, perinatal toxicity, pre-and postnatal developmental abnormalities, and eventually an increased risk for carcinomas inthe offspring. Standard Materia Medica textbooks contain unreliable information when it comesto risks during pregnancy. Wang and co-workers conducted an experimental study (WS) on micein which they investigated the effects of 17 Chinese medicinals regarding embryotoxicity andfetotoxicity. All these drugs seemed to exhibit multiple signicant toxic effects. Another studyby Li and co-workers (LS) investigated the reproductive toxicity of Atractylodis macrocephalaeRhizoma in mice, rats and rabbits. They described an increased pre- and postnatal mortalityand, at high doses, congenital malformations. In an attempt to identify the risks of the testedmedicinals during pregnancy, we analysed these two experimental studies and compared theirresults with possible safety data for humans from two reviews of clinical studies on threatenedmiscarriage (AR and CR).Methods: We re-evaluated WS and LS in relation to accordance with internationally acceptedrules, equivalence to human dose, biometric accuracy, plausibility, and coherence. Eligiblestudies of the two reviews on threatened miscarriage were evaluated for specic pregnancyrisks concerning the 17 medicinals tested in WS and LS.Results: We found that WS does not conform to international ICH guidelines and includesmany inconsistencies, implausibilities and several severe biometrical aws. It reported a totalof 364 signicant events out of which 145 false signicant results are expected. The data-handling pointed to irregularities. Analysis of LS exhibited also many inconsistencies. The results

nding author. Tel.: +0049 30 8591067; fax: +0049 30 85999705.dress: axel.wiebrecht@tcm-embryotox.de (A. Wiebrecht).

oi.org/10.1016/j.ctim.2014.08.005 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND licensetivecommons.org/licenses/by-nc-nd/3.0/).y aspects of Chinese herb

Safety aspects of Chinese herbal medicine in pregnancy 955

regarding congenital malformations were statistically insignicant and are based on small casenumbers. Insofar as the safety data of the 17 medicinals were documented by eligible studiesof the two reviews, there was no indication of an increased abortion rate in humans. Fetalgrowth retardation was not observed in the human studies. For neonatal health and postnataldevelopment, there were sufcient safety data only for a few medicinals in the human studies.

Contents

IntroduMethodResults

ReReReRe

DiscussLimitatConcluConicAcknowAppendRefere

Introduc

Chinese hemedical thits ancientmodern pawhich Wesisfactory ctreatment form to mocrucial in therapy wifoetus, resment optithreatenedrent diseasmay also ior any inciment.

The quecult one important sible hazarpostnatal risk for car

StandarinformatioChinese mAs for teratogenicity, only small case numbers (0 to 109) were available from the human data.Conclusion: WS and LS are not reliable data sources for deriving pregnancy risks in humans for thetested Chinese medicinals. In addition, the results appear to contradict the outcomes observedin the treatment of humans. Regarding teratogenicity, for most Chinese medicinals, neither thesafety nor the risk during pregnancy can be denitively ascertained. Further studies on the risksof Chinese medicinals during pregnancy are urgently needed. 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CCBY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

ction.............................................................................................................. 955s.................................................................................................................. 956

................................................................................................................... 958-evaluation of WS................................................................................................. 958-evaluation of LS.................................................................................................. 958-evaluation of AR ................................................................................................. 958-evaluation of CR ................................................................................................. 959ion................................................................................................................ 959ions............................................................................................................... 963sion ............................................................................................................... 963t of interest statement............................................................................................ 963ledgements....................................................................................................... 963ix A. Supplementary data ...................................................................................... 963

nces ............................................................................................................... 963

tion

rbal medicine (CHM) is an increasingly popularerapy which is practiced world-wide.1,2 Despite

roots, Chinese medicine may be able to offertients treatment options especially in cases intern medicine has not been able to provide sat-linical results.3 In order to full its role as aoption for contemporary patients, CHM must con-dern safety requirements. These are particularlythe treatment of pregnant women because anyll affect the health of the developing embryo orpectively. Chinese medicine offers many treat-ons for pregnancy-related indications such as

abortion, hyperemesis gravidarum, or intercur-es. In treatment of patients for infertility, thismpact the course of an undiagnosed pregnancypient pregnancy occurring after successful treat-

inconsistent information when it comes to risks during preg-nancy. Examples of different safety classications of somewell-known standard textbooks46 are provided in Table 1.In Chen and Chen6 which focuses on Chinese pharmacology,terms such as embryotoxicity or fetotoxicity appear onlytwice (in relation to the plant substances Arecae Semen andArecae Pericarpium).

In order to improve the unsatisfactory situation concern-ing the available data on pregnancy risks of CHM, Wang andco-workers conducted an experimental study (the Wangstudy, WS) in mice in which they selected 17 Chinesemedicinals commonly used during pregnancy and adminis-tered them at different periods of pregnancy.7 Their effectsregarding embryotoxicity and fetotoxicity were then inves-tigated. The results caught the TCM community unaware.All the drugs investigated seemed to exhibit multiple signif-icant toxic effects for several periods of drug administration,especially regarding fetal resorptions, stillbirths, fetal andstion of safety during pregnancy is already a dif-for Western drugs due to insufcient data. Anconcern is potential teratogenic risks. Other pos-ds include abortion, perinatal toxicity, pre- anddevelopmental abnormalities, and an increasedcinomas for the child later in his or her lifetime.d Materia Medica textbooks providing excellentn about the properties, functions and actions ofedicinals contain unreliable and sometimes even

postnatal datogenicitywere foundRhizoma, ations were Herba, GlyRadix, Amargyi FoliuHowever, teffort to reeaths, postnatal growth retardation, and ter-. Signicant results for skeletal malformations

for Rehmanniae Radix praeparata, Chuanxiongnd Citri reticulati Pericarpium. Minor malforma-found for Cuscutae Semen, Dipsaci Radix, Taxillicyrrhizae Radix, Codonopsis Radix, Dioscoreaeomi Fructus, Chuanxiong Rhizoma, Artemisiaem and Citri reticulatae Pericarpium (Table 2).he results appear to show implausibilities and the-analyse the data was considered justied.

956 A. Wiebrecht et al.

Table 1 Statements regarding safety during pregnancy in various standard textbooks.

Medicinal pin yin Bensky4 Chen6 Hempen5

Arecae Semen bing lang ca ciCoicis SemCrataegi FLycopi HeMassa medPlantaginiSiegesbec

ci = contr gnanduring pre

Anotherof Atractylrabbits (theters and the adminito mice wgrowth, a san increaseat high dohydrops, andescribed.the malfor

To explhumans, arelevant auanalysis ofthreatenedAR).9 Accothe most cChina. AnomiscarriageThis reviewthose of AR

Althougals includeconclusionsnonethelesof the usedsafety datatal data ofthe discusspregnancy.

Methods

WS was anationally biometric LS was asscoherence.

AR was the 17 mewe includeof the reciently lon

ncy of W, fetal gr

ationd coudie-up tnimus, ntal w

exccaued aplicarmar of

eacted t of f ther ofertairisk 100 f advty. W

300s ra

newludet, 95isteden yi yi ren ructus shan zha ci rba ze lan ca icata ferm shen qu s Semen che qian zi ca kiae Herba xi xian cao -

aindicated during pregnancy, ca = use with caution during pregnancy.

study investigated the reproductive toxicityodis macrocephalae Rhizoma in mice, rats ande Li study, LS).8 It used the same param-time frames as in WS. This study found thatstration of Atractylodis macrocephalae Rhizomaas associated with signicantly reduced fetalignicantly prolonged duration of pregnancy andd prenatal and postnatal mortality. Furthermore,ses, congenital malformations (skeletal, fetald short ear anomaly) and fetal resorptions were

The study was vividly documented with photos ofmations in mice.ore the risks of using CHM during pregnancy in

group of authors with the participation ofthors from WS conducted a review and meta-

adverse outcomes in studies and case series on miscarriage (the adverse outcome review,rding to the authors, threatened miscarriage isommon indication for CHM during pregnancy inther evaluation of CHM in treating threatened

is the Cochrane review on this subject (CR).10

included for the most part different studies than.h the methodological quality of most of the tri-d in these two reviews does not permit denite

to be drawn, the data from eligible studies cans be used to evaluate potential pregnancy risks

medicinals. Through an analysis of the extracted on humans and comparison with the experimen-

the animal studies, this paper tries to add toion on the safety of Chinese medicines during

pregnaresultsin WS)to feta(in relWS) anonly stfollowthe mimationpostna

Weble bedescribnal apno infonumbe

Forextraccontexbers onumbeto a cnancy least rate oof safesize oftaneoufor allto excpercenhave lnalyzed in relation to accordance with inter-accepted rules, equivalence to human dose,accuracy, plausibility and coherence. Similarly,essed for equivalence of dosing, plausibility and

evaluated for specic pregnancy risks concerningdicinals tested in WS and LS. From this review,d those studies which implied a documentationspective pregnancy risk and ensured a suf-g observation time for this scope. The following

respectivelfurther dru

The samfrom CR. Fevaluated documenteanalyses ofthe combinthe numbeThe sum mwhich a laction is suppci ca al cica ci cica ci

cy, al = allowed, - = no indication of restrictions

risks were evaluated and compared with theS: abortion rate (in relation to fetal resorptionsl growth retardation and birth weight (in relationowth in WS), neonatal health and development

to postnatal deaths and postnatal weight gain inngenital malformations. To ensure a high validity,s were included which documented a sufcientime: for fetal growth retardation and birth weightm time was until delivery, for congenital malfor-eonatal health and development until the fourtheek.luded studies which are obviously not credi-se there were signicantly fewer side-effectss could be spontaneously expected, when exter-tions of CHM were investigated or if there wastion provided about the medicinals used or thepatients treated with the different medicinals.h study relevant to a particular question, wethe cases in which a certain medicinal in thea formula was used. Subsequently, the case num-

relevant studies were added. The sum was the documented cases for which data with respectn adverse effect were available. For each preg-and each particular medicinal, a number of atdocumented cases demonstrating no increasederse events was rated as a preliminary evidenceith regard to teratogenicity, however, a sample

was considered to be the minimum. The spon-te for major malformations lies at 3 to 4 percentborns.11 Three hundred documented cases allow

a 2.7-fold increase of this risk (at a power of 80 percent condence level).12 In our analysis, we

only drugs that have been tested in WS or LS,

y. In the clinical studies, most formulas includedgs that we did not analyze.e methodology was applied to eligible studiesor this analysis only the abortion rate could bebecause other adverse events are not adequatelyd. For those controlled trials that in the meta-

AR or CR had shown a signicant superiority ofed therapy compared with Western drugs alone,rs of cases for each medicinal were summed-up.ay indicate the number of documented cases ink of abortifacient effect of the medicinal in ques-orted.

Safety aspects

of Chinese

herbal m

edicine in

pregnancy

957

Table 2 Signicant results in WS7 regarding reproductive toxicology for different periods of pregnancy at which the medicinals were given.

Tested for period: FR CRL HL SO SkA MiA SB EPD LPD PWG PWG28

abc bc bc b cde c de de de de de

Medicinal English nameAtractylodis macrocephalae Largehead AtractylodisRhizoma Rhizome a,c b,c b,c b d d,e e d,eCuscutae Semen Chinese Dodder Seed b,c b,c b c e d,e dDipsaci Radix Himalayan Teasel Root b,c b,c b c e d,e d dTaxilli Herba Chinese Taxillus Twing a,b,c b,c c e d d,e dGlycyrrhizae Radix Liquorice Root a b,c b,c b c e d,eAstragali Radix Milkvetch Root a,c b b b e e d,ePaeoniae Radix alba White Paeony Root a b,c e e d d,e eAngelicae sinensis Radix Chinese Angelica a,b,c b,c c b d e d d,eEucommiae Cortex Eucommia Bark b b,c b,c b e d,eRehmanniae Radix praep. Steamed Rehmannia Root c b,c c b c e d,e dCodonopsis Radix Pilose Asiabell Root a,b c c c d dDioscoreae Radix Common Yam Rhizome b,c b,c c d d d,e d,eAmomi Fructus Villous Amomum Fruit a,b b,c c c e d d,eRehmanniae Radix Rehmannia Root c b e d,e d,e d,e d,eChuanxiong Rhizoma Szechuan Lovage Rhizome a,b,c b,c b,c d c e e d d,eArtemisiae argyi Folium Chinese Mugwort Leaf a,b b,c c c e dCitri reticulatae Pericarpium Tangerine Peel c b c c,d c d d,e

Period (a) gestational day (GD) 36, period (b) GD 68, period (c) GD 815, period (d) GD 15 until delivery, period (e) GD 0 until delivery.FR = fetal resorptions, CRL = crown-rump length, HL = head length, SO = somite, SkA = skeletal anomalies, MiA = other minor anomalies, SB = stillbirth, EPD = early postnatal death, LPD = latepostnatal death, PGW = postnatal weight gain, signicant results for at least one period (day 1, 7, 14, 28), PWG28 = postnatal weight gain, signicant results at postnatal day 28.

958 A. Wiebrecht et al.

Results

Re-evaluation of WS

In our analysis, we found that WS does not conform to theinternationthe preferrIn additiondevelopmenumber ofprobability

The autin a range a 2- or 3-fbasic studequivalentdrugs. For aEucommiaepared Rehmthe maxim

A closemedicinalsadministraticularly evthe entire for three mRhizoma, aperiods didistered thrfor two mand Angelitwo medicfor seven m(see Table appearing prolonged

WS apptakes. A tomedicinalsof merely group that group, morfor toxic egroups, dueabout 38 tigroup.

WS reposignicanceand 106 a signicant of signicasignicance

In Tableof all tabletests are poculated. Bucalculated,tests werelevel of 5%results of tUnfortunatmultiple st

For the statistical analysis, the authors have used thedata from every control group for 51 tests. The multiple useof the same control group requires an adjustment of thep-values. The authors did not consider this.

In addition to the biometric deciencies, the handling ofta apcephions

thewas

Theencrs inn-ditableserio

alua

casum dt the

rabbed inen tespoant ducisteradmheld

doseffecperi

cas caseis notal do sibirthl.se re, peis clbbits

or 6nly

resully insed al reumm

or fa in

alua

sultsan e eval ICH guideline13 which species the rat to beed animal species in testing reproductive toxicity., the usual number of cases in studies to test forntal toxicity is set at 20 dams,14 but in WS the

dams ranged from 4 to 10. In this respect, the of random errors is high.hors claimed that they maintained the dosageequivalent to the clinical dose for humans andold increase thereof, respectively. However, they dose slightly exceeded the maximum of the

dose listed in the Chinese Pharmacopeia for fournother ve drugs (Cuscutae Semen, Taxilli Herba,

Cortex, Rehmanniae Radix praep. and unpre-anniae Radix), the dose level clearly exceeded

um.15

r look at WS reveals many inconsistencies. For showing toxic effects at certain periods oftion during pregnancy, the toxicity should be par-ident if the medicinals were taken throughout

pregnancy. However, skeletal anomalies reportededicinals (Rehmanniae Radix praep., Chuanxiongnd Citri reticulatae Pericarpium) given in certain

not arise when these medicinals were admin-oughout the entire pregnancy. The same is trueedicinals (Atractylodis macrocephalae Rhizomacae sinensis Radix) associated with stillbirth, forinals associated with early postnatal death andedicinals associated with late postnatal death2). It seems implausible that the toxic effectsat one period of pregnancy can disappear afteradministration of the medicinal.ears to contain several serious biometrical mis-tal of 1751 animals were treated with Chinese

(study group) but the control group consisted46 animals. It is very likely that in a studyis approximately 38 times larger than the controle abnormalities should arise. Even if the ratesffects were the same in the study and control

to the very large difference in number of cases,mes more events can be expected for the study

rted 364 signicant events. Of these, 211 had a level of 5%, 47 had a signicance level of 1%signicance level of 0.1%. The large number ofevents at 0.1% is surprising because at this levelnce only 1/10 of the signicant events from a 1%

level is to be expected.s S3S7 of WS, 17 substances, 3 doses and (totals) 57 criteria are listed. Thus, 17 3 57 = 2907ssible. We presume that all these tests were cal-t even if only a portion of these tests had been

it may be assumed that the most promising included. Simply by applying the signicance, a total of 145 (2.907 5%) false signicanthe total of 364 signicant events are expected.ely, Wang et al. did not mention this problem ofatistical tests.

the damacrodeviatis thatdams tively.coincidnumbeand noa repuraises

Re-ev

In themaximand no(mice,was us

Whdose-rsignicwas readminwhen latter highertoxic nancy are 13but noresult postnawere nas stillcontro

Thechancewhich and rathe 3-case, o

Thetisticaare baa caus

In sotoxicRhizomstudy.

Re-ev

The remals cadverspears to be incorrect. All values for Atractylodisalae Rhizoma, including the means and standardare identical to those of LS. The only difference

number of cases was doubled; the number ofdoubled from 5 to 10 and from 4 to 8, respec-

same was done for the numbers of offspring. Ae can thus be ruled out. Nor is it likely that the

LS were halved. Doubling the number of casessclosure of using results from another study is not

research approach. The approach of this studyus doubts about the reliability of its ndings.

tion of LS

e of Atractylodis macrocephalae Rhizoma, theose of 12 g for humans15 was used as a baseline

median dose as claimed. Therefore, the 1- to 3-its) or 6-fold (rats) equivalent dose, respectively

the experimental study.he results with mice are analyzed, a generalnse relationship cannot be detected, even if onlyndings are considered. The crown-rump lengthed if Atractylodis macrocephalae Rhizoma wased during the period of gastrulation and increasedinistered in the organogenesis period. But the

true only for a 1-fold clinical dose, not ates. If Atractylodis macrocephalae Rhizoma hasts, the administration during the entire preg-od should permit a conclusion. However, therees of late postnatal death with the 2-fold dose,

with the 3-fold dose. Only one single signicantted for the highest dose: a reduced weight gain atay 7, but not at postnatal day 14, 21 or 28. Theregnicant differences for other parameters suchs, postnatal deaths and deformities compared to

sults are inconsistent and may be due merely torhaps caused by the small case number of 5 damsearly below the customary number of 20. In rats

also, there were no signicant differences when-fold clinical dose was administered. In this3 dams with their offspring were tested.lts regarding congenital malformations were sta-signicant. Results that were not signicant andon such small numbers of cases cannot establishlationship regarding teratogenicity.ary, no valid conclusions with respect to embry-etotoxic effects for Atractylodis macrocephalae

mice, rats and rabbits can be drawn from this

tion of AR

of the experimental studies WS and LS in ani-be contrasted with the human experiences ofents in the treatment of threatened abortion. If

Safety aspects of Chinese herbal medicine in pregnancy 959

the increased resorption rates of fetuses found in WS for allmedicinals except Dioscoreae Rhizoma were transferable tohumans, this would mean that the administration of thesemedicinals to pregnant women, even more a combination ofthose, would have inevitably led to serial abortions or still-births, restrials withstudies at and with Wthe resultsincreased r

For detmedicinalstrolled triacriteria. Tthis and tthe highesthat this sdocumentsspontaneounumber of umenting sshown in T

For 13 respectivewith case 100 to 150zoma), themedicinalscated for trPericarpiumRadix). Forexist, for sthe medicnumber ofabortion.

Growthhead lengtnals for onmedicinalsobserved inexceeded 1tae SemenAstragali RRadix, EucFructus). Fresults are

WS foudeath in mwere Taxillatae Perichumans, thment. Threfor this ev100 docummacroceph(Table 4 an

With reAR only oncase serieswomen tresponding t

observation period ends at delivery, it does not meet ourinclusion requirements. For individual medicinals, only smallnumbers of cases (0 to 109) are available (Table 4 and Sup-plementary Table S5). These are insufcient for a safetyassessment regarding congenital malformations. CHM was

isterh anted Thiss.

all men

datnce d

dehe tr mopmecumi Rad

alua

a meized

nd Wantlabseed notheumb* stus inc

resoole

for s of Aiod. dicine wn drs to d toie-o

numae R

Glyc, Rereaecise um v

ssio

thor of Cnterele thans.eve

s in pectively. In AR the abortion rate of controlled CHM alone was at 5 to 18.5% and in the case2 to 20%, with combined therapy at 0 to 22.3%,estern drugs alone at 15 to 33%. In this respect

are consistent and there is no evidence of anisk of abortion from Chinese medicinals.ermining the frequency at which the individual

are administrated to humans we considered con-ls and case control studies that met our inclusionhe study by Zhang (2000)* was not included inhe following analyses, even though it containst number of cases (n = 630). It was determinedtudy lacks sensitivity for adverse events as it

considerably fewer adverse events than can besly expected in uncomplicated pregnancy. The

cases treated for threatened miscarriage and doc-afety regarding an increased rate of abortion isable 3.of the total 17 medicinals tested in WS and LS,ly, in mice, documented experiences on humansnumbers of at least 100 are available (range of2 cases). For one medicinal (Chuanxiong Rhi-re are only 44 documented cases; for 3 more, no cases exist because these drugs are not indi-eatment of threatened abortion (Citri reticulatae, Artemisiae argyi Folium, and raw Rehmanniae

11 medicinals, case numbers of more than 500ix medicinals more than 1000. Thus, for most ofinals tested in WS and LS there is quite a high

cases documenting the safety with respect to

retardation (in terms of crown-rump length,h, somite) is present in WS for all tested medici-e or two of the testing periods. For documented

no evidence of fetal growth retardation was humans. For 11 medicinals, the number of cases00 (Atractylodis macrocephalae Rhizoma, Cuscu-, Dipsaci Radix, Taxilli Herba, Glycyrrhizae Radix,adix, Paeoniae Radix alba, Angelicae sinensisommiae Cortex, Dioscoreae Rhizoma and Amomior six medicinals, smaller case numbers or no

available (Table 4 and Supplementary Table S3).nd increased rates of early or late postnatalost of the tested medicinals. The only exceptionsli Herba, Eucommiae Cortex and Citri reticu-arpium. For the comparison with experiences one parameters were neonatal health and develop-e studies of AR fullled the requirements denedaluation. Relevant case numbers of more thanented cases were reached only for Atractylodisalae Rhizoma, Dipsaci Radix and Taxilli Herbad Supplementary Table S4).spect to teratogenicity, among the studies ofe unspecied malformation was reported in the

Zhou (2006)*. With a sample size of 40 pregnantated with Chinese medicine, this case (corre-o 2.5%) is not signicant. Because the studys

adminthe 6tconducweek. mation

Forimpairerencerefereweightever, tyear odeveloare doDipsac

Re-ev

In CR randomCHM asignicable providand ancase n(2002)apy wa

Thewere pas wellstudieup perfor mecarriagWesterappliebe usefetal din casecephalHerba,CortexDioscoof preminim

Discu

The aucationwere iexampto hum

Howsuch aed in the study conducted by He (1997)* betweend the 12th week of pregnancy, and in the studyby Wang and Li (2000)* between the 5th and 20th

partly exceeds the sensitive period for malfor-

tested medicinals, WS found a signicantt of postnatal weight gain on at least one ref-e. Taking into consideration only the last studiedate (postnatal day 28), there were signicantcits noted for seven medicinals. In humans, how-hree studies with an observation period of onere listed in Table 4 and S4 showed no postnatalntal disorders. Sufcient sample sizes of over 100ented for Atractylodis macrocephalae Rhizoma,ix and Taxilli Herba.

tion of CR

ta-analysis was undertaken which included ve controlled trials. The combination therapy ofestern drugs versus Western drugs alone was

y superior in terms of the main outcome vari-nce of abortion. Of these ve studies, one

o information about the administered medicinalsr used an external application. We calculated theers of the remaining three studies. As to Zhongsdy, only the case number for the combined ther-luded.ulting case numbers for individual medicinalsd with the case numbers of the AR meta-analysis,combined therapy of CHM and Western drugs. TwoR were not included due to an insufcient follow-Table 5 shows the number of documented casesals for which a signicantly reduced rate of mis-as shown as part of a combined therapy versusugs alone. This is a much stricter criterion thanthe case numbers in Table 3. These results can

argue against the transferability of a potentialff in mice to humans. Eleven drugs were presentbers of 100 or more, namely Atractylodis macro-hizoma, Cuscutae Semen, Dipsaci Radix, Taxilliyrrhizae Radix, Paeoniae Radix alba, Eucommiaehmanniae Radix praeparata, Codonopsis Radix,

Rhizoma, and Amomi Fructus. Due to the lackinformation for other medicinals, for these onlyalues that lie below 100 could be estimated.

n

s of WS attempted to identify risks for the appli-hinese medicinals during pregnancy. The authorssted in conducting a carefully planned study, forey tried to use doses akin to those administered

r, their work contains serious biometric awsthe scheduling and evaluation of the study and

960

A. W

iebrecht et

al.

Table 3 Number of cases on which single medicinals were used in relevant studies of AR9 showing no evidence of failure of intervention at threatened miscarriage.

Study 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Number of cases CM 54 45 100 44 40 131 56 58 118 58 68 47 86 60 30 305 67 34 40 61 41 1543Formula ZXBTD BSGTD JWATD TSPSP STP ZNBTF YSGCD TEAP STP STP BYD STP STP STP STP ATD ATD ATD STP STP WZDAtraclyl. macr. Rhz. 54 100 44 40 131 56 118 58 68 47 86 60 30 305 67 34 40 61 1399Cuscutae Sem. 54 45 100 44 40 131 56 58 118 58 68 47 86 60 30 305 67 34 40 61 1502Dipsaci Rd. 54 45 100 44 40 131 56 118 58 68 47 86 60 30 305 67 34 40 61 1444Taxilli Hb. 54 45 44 40 131 56 118 58 68 47 86 60 30 305 67 34 40 61 1344Glycyrrhizae Rd. 54 45 44 305 67 34 549Astragali Rd. 54 44 131 56 58 305 67 34 749Paeoniae Rd. alba 54 45 100 44 56 58 305 67 34 763Angelicae sinen. Rd. 54 44 58 305 67 34 562Eucommiae Cort. 45 100 44 40 56 118 58 47 86 60 30 305 67 34 40 61 41 1232Rehmanniae Rd. prp. 44 56 100Codonopsis Rd. 45 100 44 131 68 388Dioscoreae Rhz. 45 40 56 118 58 47 86 60 30 305 67 34 40 61 1047Amomi Fr. 54 100 40 118 58 47 86 60 30 40 61 694Chuanxiong Rhz. 44 44

Only studies were included that exhibited a sufcient long follow-up period. Only medicinals are presented that are tested in WS. Studies: (1) Song and Zhu (2007), (2) Li (2006), (3) Heand Che (2007), (4) Yue (2009), (5) Zhou (2006), (6) Xu (2008), (7) Ye and Qiu (2008), (8) Luo (2007), (9) Chou (2002), (10) Xu (2001), (11) Chen and Yun (1999), (12) Cui (1998), (13) Kang(1998), (14) Chen (1997), (15) He (1997), (16) Zhou (1997), (17) Zhu and Li (1992), (18) Tian and Li (1991), (19) Li (1989), (20) Wang and Wang (1987), (21) Wu (1987). Formulas: ATDAnTai Decoction, BSGTDBu Shen Gu Tai Dec., BYDBao Yun Dec., JWATDJiu Wei An Tai Dec., STPShou Tai Pill, TEAPTai Er An Pill, TSPSPTai Shan Pan Shi Pill, WZDWu Zi Dec.,YSGCDYi Shen Gu Chong Dec., ZNBTFZi Ni Bao Tai Formula, ZXBTDZhi Xue Bao Tai Dec. CM = Chinese herbal therapy (CHT) alone or combined medicine (CHT plus Western drugs).References of the studies are quoted in Supplementary Table S1.

Safety aspects of Chinese herbal medicine in pregnancy 961

Table 4 Sum of frequencies the various medicinals were used in relevant studies of AR9 apparently documenting safety withregard to the specic outcome.

Medicinal No increasedrate of abortion

Fetal growth orbirth weight

Neonat. healthand develop.

Congenital malform.and neonat. develop.

Atraclyl. macr. Rhz. 1399 562 176 109Cuscutae Sem. 1502 483 97 30Dipsaci Rd. 1444 562 176 109Taxilli Hb. 1344 613 176 109Glycyrrhizae Rd. 549 121 67 0Astragali Rd. 749 172 67 0Paeoniae Rd. alba 763 172 67 0Angelicae sinen. Rd. 562 121 67 0Eucommiae Cort. 1232 470 97 30Rehmanniae Rd. prp. 100 51 0 0Codonopsis Rd. 388 0 0 0Dioscoreae Rhz. 1047 429 97 30Amomi Fr. 694 416 30 30Rehmanniae Rd. 0 0 0 0Artemisiae argyi Fol. 0 0 0 0Chuanxiong Rhz. 44 0 0 0Citri reticulatae Peric. 0 0 0 0

Only medicinals are presented that are tested in WS. The number of cases is summed up from the relevant studies.Details are presented in Supplementary Tables S3S5.

Table 5 Number of cases of studies showing a signicant lower abortion rate for combined medicine (CHT plus Western drugs)versus Western drugs alone at threatened abortion.

Study Li (2006) He and Che(2007)

Yue (2009) Chen (2002) Lv (2007) Zhong (2002) Number ofcases Total

Number of cases CM 45 100 44 51 58 30 TotalAtraclyl. macr. Rhz. 100 44 n.d. 58 30 232Cuscutae Sem. 45 100 44 51 58 30 328Dipsaci Rd. 45 100 44 51 58 30 328Taxilli Hb. 45 44 51 58 30 228Glycyrrhizae Rd. 45 44 n.d. 58 147Astragali Rd. 44 n.d. 44Paeoniae Rd. alba 45 100 44 n.d. n.d. n.d. 189Angelicae sinen. Rd. 44 n.d. 44Eucommiae Cort. 45 100 44 n.d. 189Rehmanniae Rd. prp. 44 58 102Codonopsis Rd. 45 100 44 n.d. 58 30 277Dioscoreae Rhz. 45 58 30 133Amomi Fr. 100 n.d. n.d. n.d. 100Rehmanniae Rd. n.d. n.d. 0Artemisiae argyi Fol. n.d. n.d. 0Chuanxiong Rhz. 44 44Citri reticulatae Peric. n.d. n.d. 0Number of cases total is the number of cases from relevant AR9 and CR10 studies summed up.Studies exhibiting an insufcient follow-up were excluded. Only medicinals are presented that are tested in WS and LS. Some studiesuse more medicinals in addition to the basic formula, whose frequency is not dened (n.d.), case numbers are minimum numbershere. CM = CHT alone or combined medicine (CHT plus Western drugs). References of the studies are quoted in Supplementary Tables S1and S2. Studies from the AR meta-analysis.

Studies from the CR meta-analysis.

962 A. Wiebrecht et al.

irregularities in handling of data. The number of cases issmall. These factors prevent us from accepting the resultsas valid. The absence of a dose-response relationship makesthe results appear even more implausible. Thus, we con-clude that the study is not suitable for using the resultsto derive pmedicinalsinsignicandrawing vafetotoxicit

Several other reseause of Taximedicinal these disorof body wecompared The conclufor which examined lae Rhizomof Atractylorally in dgestation. found no sof Atractylrespect to obvious ma

In additdict the reas the admdocumentewere inclufetal growdevelopmeminimum tChinese medrugs are r1500. Thusunlikely tofurther supof threatenthe two mcant superWestern drber of 100 CHM for refrom thoseWestern drcases. Theby CHM.18

Fewer cquestions and postnaumented eoccurrenceof cases waOtherwise,clues of te

Generalimental anhumans is

metabolism, and sensitivity can differ greatly and giventhe differing embryogenesis can lead to different effects.Since there is a threshold dose for teratogenic effects,results from high doses often used in animal studies arenot generally predictive for therapeutic doses of humans.

are effecans.s in y ins, bu

preged agatibits

ers abryod coing is.21

natul anie anededncy atogerge curre

drues of anquenncieleastlforma (hions. Thf thhe c

drenegnase infolloally,cumese ptiosulther ess dre reenceof teienti

thisticet byile cl reheironalregnancy risks in humans for the tested Chinese. Similarly, because of several implausibilities andt results, LS cannot be considered qualied forlid conclusions with respect to embryotoxicity ory of Atractylodis macrocephalae.results of WS and LS differ from the outcomes ofrch. WS found developmental disorders with thelli Herba in mice embryos. Liu et al. tested thisin groups of 12 pregnant rats and did not detectders. There were neither divergences in termsight, body length, or tail length of the embryosto control nor were there skeletal anomalies.16

sions of LS clash with results from another studya detailed abstract was published.17 This studythe pregnancy risks of Atractylodis macrocepha-a in groups of 17 or 18 mice. An aqueous extractodis macrocephalae Rhizoma was administeredoses of 2, 8 and 32 g/kg from day 6 to 15 ofDistilled water served as the control. The studyignicant differences between the varying dosesodis macrocephalae Rhizoma and the control withthe incidence of fetal resorptions, dead fetuses,lformations or skeletal abnormalities.ion, the results of WS and LS appear to contra-sults obtained in treatment of humans, in so farinistration of the tested medicinals has beend in Chinese controlled trials or case series andded in AR.9 For the parameters abortion risk,th retardation, postnatal survival and postnatalnt, the number of at least 100 cases was rated aso make a preliminary statement on the safety ofdicinals. For the parameter abortion rate, most

eaching high numbers of cases of up to more than, administration of the medicinals in question is

increase the rate of abortion. This statement isported by studies from CR10 on the treatmented miscarriage. When combining the results ofeta-analyses with the strict criterion of signi-iority of CHM as adjunctive therapy compared tougs alone, eleven out of 17 drugs achieved a num-documented cases or more. A further review oncurrent miscarriage evaluating studies different

of CR showed that CHM or CHM combined withugs was superior to Western drugs alone in mostre was no evidence of an elevated abortion rate

ases are available for analysis to answer theof fetal growth retardation, postnatal deathstal developmental disorders. However, the doc-xperiences in humans appear to disprove the

of such effects. For teratogenicity, the numbers insufcient to ascertain the absence of this risk.

it is worth noting that in AR there are no positiveratogenicity for the medicinals tested.ly, the transferability of results from exper-imal studies to the therapeutic situation in

questionable.11,19 Species-related bioavailability,

There genic in humdefectgenicitrabbitduringregardally nefor rabhamstor emfor foooccurrhuman

Formentaprovidare nepregnaon terwith lathe occertainoutcomedge othe frepregna

At tal maRhizomformatsystemtions oeye. (TIn chiling princreamean

Usunot dothat thconsumAs a reand oteven lthat aexperiterms ern scduringrst noand no

Whnaturauntil ttraditiseveral substances which have proven terato-ts in animals which do not exhibit this property

For example, aspirin is known to cause cardiacrats and rhesus monkeys but shows no terato-

humans. Insulin is teratogenic in rats, mice andt is regarded the treatment of choice for diabetesnancy in humans.14 Of 165 compounds which ares non-teratogenic to humans, the tests were actu-ve only in 80 percent for monkeys, 70 percent, 50 percent for rats and in only 35 percent fornd mice.20 There is also evidence of teratogenictoxic and fetotoxic effects from animal studiesnstituents, e.g. for caffeine20 and for alkaloidsn potatoes which however, pose no known risk to23

ral medicines with a long history of use, experi-mal studies are likely to raise more questions thanswers. Instead, more data from human studies. Intervention trials on the safety of drugs duringare forbidden for ethical reasons. Human datanicity are derived primarily from cohort studies

numbers of cases in which a relationship betweennce of congenital malformations and the use of ag can be established. These studies document thef drugs that were taken without or despite knowl-

existing pregnancy and compare these data withcies of malformations with those of non-exposeds.

one cohort study has shown a risk of congeni-ations for the use of CHM in humans.24 Coptidisuang lian) was associated with increased mal-

of the nervous system and the external genitale formula an tai yin led to increased malforma-e musculoskeletal and connective tissue and theomposition of an tai yin varies between sources.)

whose mothers had taken Coptidis Rhizoma dur-ncy, the same authors were able to observe an

cancer cases, particularly of the CNS, after aw-up period of 14.9 years.25

the use of natural remedies during pregnancy isented. The reasons for this are, among others,remedies are considered less risky and that theirn frequently occurs without medical supervision., little objective data about their teratogenicitypregnancy risks are available.19,2628 And in fact,ata are available than for pharmaceutical drugslatively new on the market.11 Nevertheless, the

collected over time is of some value, even inratogenicity, since CHM has been used under mod-c supervision for about two generations. Even

time, most teratogens of Western drugs wered through case reports and clinical case studies

animal testing.19

onventional medicine often takes the view thatmedies are to be avoided during pregnancy

safety has been proven,29,30 others argue thatly-used remedies with few side-effects should not

Safety aspects of Chinese herbal medicine in pregnancy 963

be banned until there is clear evidence of risk.26 In the caseof dietary supplements, the European Food Safety Authoritiy(EFSA) has established a category in which a substance is pre-sumed to be safe and does not require further safety assaysif no adverse effects have been reported after long-term useon large podient and iadditional ation, i.e. would be tconclude tingredient years witho

On the detect selof Coptidisdemiologicdenitive closely obsable. Unfois a continof Chinesethere is corisks, evening pregnadrugs apprthe catego

Therefoshould onlcareful beapplied if during thetreatment,important gestationateratogensthis perioding the antstill be aff

Limitatio

One limitaeffects as won the typtered. Howitself wouldata is genthe standating medicherbal medused mediexperiencecombinatio

The dalow-qualityregarding bFurthermotimes durinphase for mtime of ad

some data are based on a high number of cases and are veryconsistent, which partly outweighs the limitations.

Conclusion

the ome r thunre

preling tof saon re

namncy ks oflogicly ney.

ict

thorthis w

owl

thorMedaritstruf Scy, atviewbergwas Deutny, theimnich

ndix

men in th.2014lude; inc.

enc

J. T2. tentu Codicintp://-Tradessedpulations31: Depending on the botanical ingre-ts uses, there are circumstances under which nodata are judged necessary for the safety evalu-a presumption of safety would be applied. Thishe case whenever available data would allow tohat exposure to known levels of the botanicalhas occurred in large population groups for manyut reported adverse effects.

other hand, experience alone is not sufcient todom-occurring risks. The obvious teratogenicity

Rhizoma was only recognized by a large epi-al study. Even if this cannot be regarded as astatement, this risk must be perceived and beerved as long as no mitigating data are avail-rtunately, this study is an exception and thereuing lack of knowledge about the teratogenicity

medicinals. For most of conventional drugs, too,nsiderable uncertainty regarding their pregnancy

for those that are sometimes administered dur-ncy. It is estimated that about 98 percent of theoved between 2000 and 2010 in the U.S. fall intory of undetermined teratogenic risk.32

re, as for chemical drugs, Chinese medicinalsy be administered to pregnant women after anet-risk assessment. Medicines should only bethere is an unequivocal indication, especially

rst trimester. The same applies for fertility which often extends into the pregnancy. Theperiod of organogenesis in humans is betweenl day 22 and 55. However, it is known that some

can cause malformations if administered prior to.14 For example, even one year after discontinu-i-psoriatic agent acitretin, every 20th child couldected by malformations.33,34

ns

tion of the risks assessment is the fact that theell as the side-effects of medicinals may depende of combination in which they are adminis-ever, to assess each individual combination byd require a great deal more data. This type oferally unavailable. Furthermore, in some studies,rd formula was modied by adding or subtrac-inals individually, a common practice in Chineseicine. At present, any evaluation of commonly-

cinals is only possible by relying on the sum ofs collected through the application in differentns.ta extracted from AR and CR are based on

studies. Therefore, only limited conclusionsoth efcacy and safety can be made from these.re, the Chinese medicinals were used at differentg pregnancy, in some cases outside the sensitivealformations. For a few studies no data for the

ministration are available. On the other hand,

Takingis of sthat foing an that aRegarddence situatidrugs,pregnathe risdemiourgentgenicit

Con

The auest in

Ackn

The auSocial the Chable inChristoatologfor reNuremstudy of the GermaDegerssis, Mu

Appe

Supplefound,j.ctim

*IncTable 1Table 2

Refer

1. Liu201con

2. Hirmehtandaccviewpoint that documented clinical experiencenotable value for safety evaluation, we believeose medicinals with higher case numbers show-markable course of pregnancy, the data suggestiminary positive safety statement is warranted.eratogenicity, there are insufcient data for evi-fety. Therefore, for most Chinese medicinals, themains similar to that of the majority of chemicalely, that neither the safety nor the risk duringcan be denitely ascertained. Further studies on

Chinese medicinals during pregnancy, chiey epi-al studies and documented administrations, areeded to investigate the issue of possible terato-

of interest statement

s indicate that they do not have conicts of inter-ork.

edgements

s thank Prof. Dr. Benno Brinkhaus, Institute foricine, Epidemiology and Health Economics at University Medical Center, Berlin, for valu-ctions for preparation of the manuscript, PD Dr.haefer, Institute for Pharmacovigilance and Ter-

the Charit University Medical Center, Berlining the manuscript, and Dr. Velia Wortman,, for help with the translation of the text. Thenancially supported by the Scientic Chaptersche rztegesellschaft fr Akupunktur, Munich,he Schweizerische Berufsorganisation fr TCM,, Switzerland, and the Societas Medicinae Sinen-, Germany.

A. Supplementary data

tary material related to this article can bee online version, at http://dx.doi.org/10.1016/.08.005.d studies from AR are cited in Supplementaryluded studies from CR are cited in Supplementary

es

CM exports set to rise at a healthy clip. China Dailyhttp://www.chinadaily.com.cn/bizchina/2012-02/10/14578000.htm (date accessed 2014/07/30).rporation. Hiru Corporation and traditional Chinesee market research report. Hiru Corporation; 2010.www.docstoc.com/docs/45250839/Hiru-Corporation-itional-Chinese-Medicine-Market-Research (date

2012/12/27).

964 A. Wiebrecht et al.

3. WHO Commission on Intellectual Property Innovation and PublicHealth. Traditional Chinese medicine could make health forone true. WHO Commission on Intellectual Property Innova-tion and Public Health; 2005. http://www.who.int/intellectualproperty/studies/Jia.pdf (date accessed 01/12/2013).

4. Bensky D, Clavey S, Stger E. Chinese herbal medicine. Materiamedica. 3rd ed. Seattle: Eastland Press; 2004.

5. Hempen CH, Fischer T. A materia medica for Chinese medicine:plants, minerals and animal products. 1st ed. Edinburgh:Churchill Livingstone; 2009.

6. Chen JK, Chen TT. Chinese medical herbology and pharmacol-ogy. 1st ed. Art of Medicine Press; 2004.

7. Wang CC, Li L, Tang LY, Leung PC. Safety evaluation of commonlyused Chinese herbal medicines during pregnancy in mice. HumReprod 2012;27:244856.

8. Li L, Tang LY, Man GC, et al. Potential reproductive toxic-ity of Largehead Atractylodes Rhizome, the most commonlyused Chinese medicine for threatened miscarriage. Hum Reprod2011;26:32808.

9. Li L, Dou LX, Neilson JP, Leung PC, Wang CC. Adverse out-comes oa system2012;18

10. Li L, Dofor thre2012;5:C

11. Mitchelleditor. P2005. p.

12. SchaefeC. ArznMnchen

13. ICH. ICHto reprfertilityWeb SiteGuideli

14. Bailey Jresearch

15. Chinese Peoplesence Pre

16. Liu MS, embryotguo Yous

17. Xu M, Tition of Rstudies (tor. TheChinesecology;

18. Yang GYthe trea

of randomized clinical trials. BMC Complement Altern Med2013;13:320.

19. Polifka JE, Friedman JM. Clinical teratology: identifying terato-genic risks in humans. Clin Genet 1999;56:40920.

20. Caffeine: FDA. deletion of GRAS status, proposed declarationthat no prior sanction exists, and use on an interim basis pendingadditional study. Fed Regist 1980;45:6981738.

21. Friedman M, Rayburn JR, Bantle JA. Developmental toxi-cology of potato alkaloids in the frog embryo teratogen-esis assayXenopus (FETAX). Food Chem Toxicol 1991;29:53747.

22. Kline BE, Von Elbe H, Dahle NA, Kupchan SM. Toxic effects ofpotato sprouts and of solanine fed to pregnant rats. Proc SocExp Biol Med 1961;107:8079.

23. Gafeld W, Keeler RF. Induction of terata in hamsters bysolanidane alkaloids derived from Solanum tuberosum. ChemRes Toxicol 1996;9:42633.

24. Chuang CH, Doyle P, Wang JD, et al. Herbal medicines usedduring the rst trimester and major congenital malformations:an analysis of data from a pregnancy cohort study. Drug Saf

6;29:53748.ang CH, Doyle P, Wang JD, et al. Herbal medicines dur-

pregm a 9;18ns Tbal 3;68gens elop3;68

Dogern TZ. Ugnan7;16rcus bal :111A Scanicants inm Matogeet Cmer

er te01.neite

Psomm 1f Chinese medicines used for threatened miscarriage:atic review and meta-analysis. Hum Reprod Update

:50424.u L, Leung PC, Wang CC. Chinese herbal medicinesatened miscarriage. Cochrane Database Syst RevD008510.

AA. Studies of drug-induced birth defects. In: Strom BL,harmacoepidemiology. Chichester: John Wiley & Sons;

p. 501514.r C, Spielmann H, Vetter K, Weber-Schndorfereimittel in Schwangerschaft und Stillzeit. 8th ed.: Urban&Fischer; 2012.

harmonised tripartite guideline: detection of toxicityoduction for medical products & toxicity to male

S5(R2); 2005. http://www.ich.org/leadmin/Public/ICH Products/Guidelines/Safety/S5 R2/Step4/S5 R2

ne.pdf (last update 2005)., Knight A, Balcombe J. The future of teratology

is in vitro. Biogenic Amines 2005;19:97145.Pharmacopoeia Commission. Pharmacopoeia of the

Republic of China, Vol. I. Beijing: China Medical Sci-ss; 2010.Li XH, Zhang H, et al. [Effects of herba taxilli on theic development in SD pregnant rats] (Chinese). Zhong-heng Yu Yichuan Zazhi 2013;21:1057, 58.an XY, Chong LL, et al. Developmental toxicity evalua-hizoma Atractylodis Macrocephalae: in-vivo and in-vitroAbstract S2-20). In: Chinese Society of Toxicology, edi-

abstracts book of the 5th National Congress of the Society of Toxicology. Beijing: Chinese Society of Toxi-2009. p. 334., Luo H, Liao X, Liu JP. Chinese herbal medicine fortment of recurrent miscarriage: a systematic review

20025. Chu

ingfro200

26. Johher200

27. Jurdev200

28. LowAlt

29. Li Dpre200

30. Maher105

31. EFSbotdie

32. AdaterGen

33. Lamaft108

34. Arzdesgranancy and childhood cancers: an analysis of datapregnancy cohort study. Pharmacoepidemiol Drug Saf:111920., Sibeko L. Pregnancy outcomes in women usingtherapies. Birth Defects Res B Dev Reprod Toxicol:5014.TM. Potential toxicities of herbal therapies in the

ing fetus. Birth Defects Res B Dev Reprod Toxicol:4968.

T. The use of botanicals during pregnancy and lactation.her Health Med 2009;15:548.se of traditional Chinese herbal medicines during earlycy in mainland China. Pharmacoepidemiol Drug Saf:9423.DM, Snodgrass WR. Do no harm: avoidance of

medicines during pregnancy. Obstet Gynecol 2005;922.ientic Committee. Guidance on safety assessment ofls and botanical preparations intended for use as ingre-

food supplements. EFSA J 2009;7:124968.P, Polifka JE, Friedman JM. Evolving knowledge of thenicity of medications in human pregnancy. Am J Med

Semin Med Genet 2011;157:17582. EJ. Embryopathy in infant conceived one yearrmination of maternal etretinate. Lancet 1988;2:

legramm. Schwangerschaft. Wartezeit nach Absetzenriasis-Retinoids Acitretin (NEOTIGASON). Arzneitele-997;28:58.

Safety aspects of Chinese herbal medicine in pregnancyRe-evaluation of experimental data of two animal studies and the cl...IntroductionMethodsResultsRe-evaluation of WSRe-evaluation of LSRe-evaluation of ARRe-evaluation of CR

DiscussionLimitationsConclusionConflict of interest statementAcknowledgementsAppendix A Supplementary dataReferences