Académique Documents
Professionnel Documents
Culture Documents
(Review)
This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library
2010, Issue 5
http://www.thecochranelibrary.com
Ulster Hospital, Belfast, UK. 3 Department of Obstetrics and Gynaecology, St Marys Hospital, London, UK
Contact address: Priscilla Devaseelan, Department of Obstetrics and Gynaecology, Craigavon Area Hospital, 68, Lurgan Road, Porta-
down, Northern Ireland, BT63, UK. pdevaseelan@yahoo.co.uk.
Citation: Devaseelan P, Fogarty PP, Regan L. Human chorionic gonadotrophin for threatened miscarriage. Cochrane Database of
Systematic Reviews 2010, Issue 5. Art. No.: CD007422. DOI: 10.1002/14651858.CD007422.pub2.
Copyright 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
ABSTRACT
Background
Miscarriage is a common occurrence in early pregnancy. Human chorionic gonadotrophin (hCG) is secreted by the syncytiotrophoblast.
It promotes the corpus luteum to secrete progesterone and helps in maintaining the pregnancy. Hence, there has been much interest
in the use of hCG to treat threatened miscarriage.
Objectives
To assess the efficacy and safety of human chorionic gonadotropins in the treatment of threatened miscarriage.
Search methods
We searched the Cochrane Pregnancy and Childbirth Groups Trials Register (February 2010), the Cochrane Central Register of
Controlled Trials (The Cochrane Library 2010, Issue 1), MEDLINE (1966 to 12 February 2010), EMBASE (1980 to 12 February
2010) and CINAHL (1989 to 12 February 2010). We also scanned the bibliographies of all located articles for any unidentified articles
and attempted to contact authors where necessary.
Selection criteria
All randomised controlled trials (RCTs) that assess the effectiveness of hCG in the treatment of threatened miscarriage compared
to placebo, no treatment of any other intervention, provided viability of the fetus has been confirmed by ultrasound before the
commencement of treatment.
Data collection and analysis
At least two authors assessed the trials for inclusion in the review and extracted the data.
Main results
Three studies (312 participants) were included in the review, one of which was of poor methodological quality. The meta-analysis
showed that there was no statistically significant difference in the incidence of miscarriage between hCG and no hCG (placebo or no
treatment) groups (Risk ratio (RR) 0.66; 95% confidence interval (CI) 0.42 to 1.05). When hCG and bed rest alone were compared,
there was a significant reduction in the risk of miscarriage (RR 0.47; 95% CI 0.27 to 0.82). This result should be interpreted with
caution, as one of the two trials from which this result is derived was of poor methodological quality. There was no report of adverse
effects of hCG on mother or baby. More good quality RCTs are urgently needed to assess the effects of hCG in threatened miscarriage.
Human chorionic gonadotrophin for threatened miscarriage (Review) 1
Copyright 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Authors conclusions
The current evidence does not support the routine use of hCG in the treatment of threatened miscarriage.
Threatened miscarriage is a term used when a woman who is less than 24 weeks pregnant with a live baby experiences bleeding. The
mother has vaginal bleeding, with or without abdominal pain and cramps, but the cervix is closed. Miscarriage is a source of great
physical and psychological distress and is very common. Causes include chromosomal abnormalities in the fetus, when the mother
is older, has uterine or endocrine abnormalities or has polycystic ovarian syndrome. Human chorionic gonadotrophin (also called
hCG) is a hormone produced by the placenta and is known to help maintain the pregnancy. Hence there has been much interest
in the use of hCG for treating threatened miscarriage with the aim of preserving the pregnancy.This review of three trials (including
312 participants), one of which was of poor quality, found no evidence that hCG can be used as effective treatment for threatened
miscarriage. There was no report on adverse effects of hCG on the mother or baby. More good-quality research is needed to study the
impact of hCG on miscarriage.
BACKGROUND
on the connection between polycystic ovarian syndrome, insulin
Miscarriage is pregnancy loss before 24 weeks of gestation (WHO
resistance and pregnancy loss ( Rai 2006).The presence of bacte-
1992). It is unfortunately common in early pregnancy and an es-
rial vaginosis in the first trimester has been associated with second
timated 15% to 20% of clinical pregnancies miscarry (Alberman
trimester miscarriage and preterm delivery (Hay 1994). Human
1992). A large majority of the pregnancies are lost before the
chorionic gonadotrophin (hCG) is a glycoprotein secreted by the
missed period and, therefore, referred to as biochemical preg-
syncytiotrophoblast of the placenta as early as eight days after the
nancies. Miscarriage can be associated with considerable physi-
luteinising hormone peak (Rosevear 1999). The concentration of
cal and psychological sequelae ranging from depression to intense
hCG gradually rises during early pregnancy until it reaches a peak
grief. Threatened miscarriage is defined as any vaginal bleeding
at about eight to 10 weeks of gestation, thereafter falling to a stable
before 24 weeks, with or without abdominal pain in the presence
new level for the rest of pregnancy. It is known that subnormal
of a viable fetus and closed cervix (Cunningham 2001). When
values of hCG are associated with a failing pregnancy or a miscar-
bleeding is slight or resolves, pregnancy may continue normally
riage. Ho et al in their study of hCG levels in normal and abnor-
but threatened miscarriage can be associated with a higher like-
mal pregnancies noted that abnormal pregnancies had a deficiency
lihood of adverse pregnancy outcome like prematurity (which is
of hCG production and the rate of rise and biological activity of
increased twofold), small-for-gestational-age babies (which is in-
hCG were lower (Ho 1997). hCG promotes the maintenance of
creased three-fold), and perinatal death (Sipila 1992; Williams
the corpus luteum during early pregnancy and helps it to secrete
1991).
progesterone to maintain the pregnancy. It has also been hypoth-
esised that hCG may be a placental link for the development of
local maternal immunotolerance (Kayisli 2003).
Of the many causes of miscarriage, the incidence of chromo-
somal abnormalities has been estimated to be 30% to 60%
(Simpson 1992), the most frequent cytogenetic abnormality being Owing to the above properties, hCG has been used in the treat-
trisomy, followed by polyploidy and monosomy X (Fritz 2001). ment of threatened miscarriage in an attempt to maintain the preg-
Other causes of miscarriage include advanced maternal age (Nybo nancy, but with conflicting results. Although the corpus luteum
Anderson 2000), congenital uterine abnormalities (Grimbizis serves its physiological purpose until 12 weeks, hCG has been used
2001), antiphospholipid antibodies (Rai 1997), thrombophilic beyond this period, with a view to preserve the pregnancy, with lit-
disorders (Rey 2003), maternal endocrine abnormalities like un- tle supportive evidence. This review aims to study all the available
controlled diabetes (Greene 1999), hyperprolactinemia (Garzia data on the effectiveness of hCG in the treatment of threatened
2004) and polycystic ovarian syndrome. Attention is now focused miscarriage.
Human chorionic gonadotrophin for threatened miscarriage (Review) 2
Copyright 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
OBJECTIVES Search methods for identification of studies
To assess the efficacy and safety of human chorionic gonadotropin
in the treatment of threatened miscarriage.
Electronic searches
We searched the Cochrane Pregnancy and Childbirth Groups Tri-
als Register by contacting the Trials Search Co-ordinator (Febru-
METHODS ary 2010).
The Cochrane Pregnancy and Childbirth Groups Trials Register
is maintained by the Trials Search Co-ordinator and contains trials
identified from:
Criteria for considering studies for this review 1. quarterly searches of the Cochrane Central Register of
Controlled Trials (CENTRAL);
2. weekly searches of MEDLINE;
3. handsearches of 30 journals and the proceedings of major
Types of studies conferences;
All randomised controlled trials (RCTs) that assessed the effective- 4. weekly current awareness alerts for a further 44 journals
ness of human chorionic gonadotropin (hCG) in the treatment plus monthly BioMed Central email alerts.
of threatened miscarriage compared to placebo, no treatment of Details of the search strategies for CENTRAL and MEDLINE,
any other intervention, provided viability of the fetus has been the list of handsearched journals and conference proceedings, and
confirmed by ultrasound before the commencement of treatment the list of journals reviewed via the current awareness service can
. be found in the Specialized Register section within the edito-
rial information about the Cochrane Pregnancy and Childbirth
Group.
Trials identified through the searching activities described above
Types of participants
are each assigned to a review topic (or topics). The Trials Search
Pregnant women less than 24 weeks gestation with threatened mis- Co-ordinator searches the register for each review using the topic
carriage, who had a fetal heartbeat demonstrated on ultrasound. list rather than keywords.
We included multiple pregnancies and women with a past history In addition, we searched:
of recurrent miscarriages. 1. the Central Register of Controlled Trials (The Cochrane
Library 2010, Issue 1): see Appendix 1 for details of search
strategy;
2. MEDLINE (1966 to 12 February 2010): see Appendix 2
Types of interventions
for details of search strategy;
hCG injections, used in any dose, single or multiple, regardless of 3. EMBASE (1980 to 12 February 2010): see Appendix 3 for
duration and frequency of administration compared with placebo, details of search strategy;
no treatment or any other intervention. 4. CINAHL (1989 to 12 February 2010): see Appendix 4 for
details of search strategy.
For dichotomous data, we presented results as risk ratio with 95% We carried out statistical analysis using the Review Manager soft-
confidence intervals. ware (RevMan 2008). We use fixed-effect inverse variance meta-
analysis for combining data where trials were examining the same
intervention, and the trials populations and methods were judged
Continuous data sufficiently similar.
We planned to express continuous differences between groups in
the meta-analysis as a mean difference and 95% confidence inter- Subgroup analysis and investigation of heterogeneity
val. However, we found no continuous data. There was insufficient information in the original trials to carry
out planned subgroup analyses by dose of hCG, and the effect of
hCG in early and late threatened miscarriage.
Unit of analysis issues
For future updates, should sufficient data become available, we
will conduct planned subgroup analyses classifying whole trials by
interaction tests as described by (Deeks 2001).
Cluster randomised trials
There were no cluster randomised trials. All trials included ran-
domised individuals. Sensitivity analysis
We carried out sensitivity analysis to explore the effect of trial
quality. This involved analysis of selection bias and attrition bias.
Crossover trials Of the three studies, only two were of good quality with adequate
There were no crossover trials. allocation concealment (Harrison 1993; Qureshi 2005), hence in
analysing our data, we categorised the studies into high quality and
not high quality (Suvonnakote 1986) and compared the outcome.
Quasi randomised trials
There were no quasi randomised trials.
RESULTS
Dealing with missing data
The three studies had a total of 312 participants. Nine subjects
were lost to follow up (Qureshi 2005) and we have carried out Description of studies
analysis for the remaining 303 participants. Only one study anal- See: Characteristics of included studies; Characteristics of excluded
ysed data on all participants with available data in the group to studies.
which they are allocated, regardless of whether or not they received See Characteristics of included studies; Characteristics of excluded
the allocated intervention (analysis on an intention-to-treat basis) studies.
(Qureshi 2005). We were unable to perform intention-to-treat
analysis for the other two studies included in this review because
of insufficient information from the original trials. Results of the search
For future updates, if in the original reports participants are not We included three studies in the review (Harrison 1993; Qureshi
analysed in the group to which they were randomised, and there is 2005; Suvonnakote 1986). For further details of included and
sufficient information in the trial report, we will attempt to restore excluded studies, see the tables of Characteristics of included
them to the correct group. studies and Characteristics of excluded studies.
Three trials met the inclusion criteria, involving 312 participants. No trials reported any adverse effect of hCG on the mother or
fetus.
Primary outcome
Miscarriage DISCUSSION
Meta-analysis of the effect of hCG compared to placebo or no There has been much interest in the use of hCG to prevent mis-
treatment showed no statistical significance on the effect of mis- carriage for many years. But the lack of good quality trials have
carriage (Risk ratio (RR) 0.66; 95% confidence interval (CI) 0.42 resulted in uncertainties in the use of hCG. Of the three trials
to 1.05); heterogeneity Chi2 = 2.16, see Analysis 1.1. When trials included in this review, only two were of good quality and one
of high quality were compared to those not of high quality, it did was methodologically poor in respect to method of randomisation
not make any impact on the effect of hCG on miscarriage. When and allocation concealment (Suvonnakote 1986), which made se-
hCG was compared to placebo, there was no significant differ- lection bias a possibility. Only one study performed an intention-
ence on the effect of miscarriage (RR 0.83; 95% CI 0.46 to 1.47); to-treat analysis, and there was insufficient information from the
heterogeneity Chi2 = 0.89, see Analysis 1.2. However, when hCG other two studies to perform a similar analysis.
was compared to bed rest, there was a significant reduction in the
risk of miscarriage (RR 0.47; 95% CI 0.27 to 0.82);heterogeneity This meta-analysis did not show any statistical significance in the
Chi2 = 0.49, see Analysis 1.3. This result should however be in- overall effect on miscarriage or birthweight in either the hCG or
terpreted with caution as it is influenced by two trials (Harrison no hCG (placebo or bed rest) groups. However, when hCG and
1993; Suvonnakote 1986), one of which was of poor methodolog- bed rest were compared, although there was a significant reduction
ical quality. in the miscarriage rate, this should be interpreted with caution
Due to a paucity of data in the included trials, subgroup analysis as the evidence was pooled from two trials, one of which was
for early and late miscarriage or effect of hCG by dose could not methodologically poor (Suvonnakote 1986). The apparent overall
be carried out. decrease in preterm birth in the hCG group was no longer apparent
when considering only the two trials of high quality. We there fore
recommend caution in interpreting these results.
Secondary outcomes
There was insufficient information in the trials included to per-
form planned subgroup analysis by dose of hCG or study the ef-
fect of hCG in early (less than 12 weeks) and late (more than 12
Preterm birth
weeks) threatened miscarriage.
When the effect of hCG was compared to placebo or no treatment
on preterm birth, there was a reduction in the risk of preterm No adverse effects of hCG were reported.
birth in the hCG arm (RR 0.47; 95% CI 0.17 to 1.33); hetero-
geneity Chi2 = 2.41. However this result should be interpreted
with caution as the analysis was influenced by a trial which was of AUTHORS CONCLUSIONS
methodologically poor quality (Suvonnakote 1986), see Analysis
2.1. However, when trials of high quality were grouped together
Implications for practice
(Harrison 1993; Qureshi 2005), there was no significant differ-
ence in the risk of preterm birth between hCG and placebo or no The current evidence does not support the routine use of hCG
treatment (RR 0.71; 95% CI 0.21 to 2.39); heterogeneity Chi2 = in the treatment of threatened miscarriage. There are no reported
0.72. adverse effects of hCG on the mother or baby.
REFERENCES
References to studies included in this review Harrison 1992 {published data only}
Harrison RF. Human chorionic gonadotrophin (hCG) in
Harrison 1993 {published data only} the management of recurrent abortion; results of a multi-
Harrison RF. A comparative study of human chorionic centre placebo-controlled study. European Journal of
gonadotropin, placebo, and best rest for women with early Obstetrics & Gynecology and Reproductive Biology 1992;47
threatened abortion. International Journal of Fertility & (3):1759.
Menopausal Studies 1993;38(3):1605.
Quenby 1994 {published data only}
Harrison RF. Miscarriage - hormonal treatment with
Quenby S, Farquharson RG. Human chorionic
hCG [Avortement: Traitement hormonal par les HCG].
gonadotropin supplementation in recurrent pregnancy loss:
Contraception, Fertilite, Sexualite 1991;19:3736.
a controlled trial. Fertility and Sterility 1994;62(4):70810.
Qureshi 2005 {published data only} Sandler 1979 {published data only}
Qureshi NS, Edi-Osagie EC, Ogbo V, Ray S, Hopkins RE. Sandler SW, Baillie P. The use of human chorionic
First trimester threatened miscarriage treatment with human gonadotropin in recurrent abortion. South African Medical
chorionic gonadotrophins: a randomised controlled trial. Journal 1979;55(21):8325.
BJOG: an international journal of obstetrics and gynaecology
Toth 2001 {published data only}
2005;112:153641.
Toth P. Clinical data supporting the importance of vascular
Suvonnakote 1986 {published data only} LH/hCG receptors of uterine blood vessels. Seminars in
Suvonnakote T. Treatment of threatened abortion with Reproductive Medicine 2001;19(1):5561.
human chorionic gonadotropin. Journal of the Medical
Toth 2003 {published data only}
Association of Thailand 1986;69(12):6548.
Toth P, Lukacs H, Gimes G, Sebestyen A, Valent S, Pasztor
References to studies excluded from this review N, et al.The role of LH/hCG receptors in premature
delivery. Magyar Noorvosok Lapja 2003;66(2):959.
Baillie 1973 {published data only} Ushiroyama 2006 {published data only}
Baillie P, Sandler SW, Kukard RF. Human chorionic Ushiroyama T. Efficacy of kampo medicine Xiong-Gui-Jiao-
gonadotropin in threatened abortion. South African Medical Ai-Tang, a traditional herbal medicine, in the treatment of
Journal 1973;47(29):12936. threatened abortion in early pregnancy. American Journal of
Czeizel 2008 {published data only} Chinese Medicine 2006;34(5):73140.
Czeizel AE, Dudas I, Gidai J, Horvath-Puho E. No effect of Utian 1974 {published data only}
human chorionic gonadotropin treatment due to threatened Utian WH. Human chorionic gonadotrophin in the
abortion in early pregnancy for birth outcomes. Central treatment of threatened and recurrent miscarriage. A
European Journal of Medicine 2008;3(1):716. preliminary reappraisal. South African Medical Journal
Eskes 1992 {published data only} 1974;48(18):76971.
Eskes TK. Human Chorionic Gonadotrophin (hCG) in the Vorster 1977 {published data only}
management of recurrent abortion. European Journal of Vorster CZ, Pannall PR, Slabber CF. Treatment of
Obstetrics & Gynecology and Reproductive Biology 1992;47 threatened and habitual abortion with human chorionic
(3):173. gonadotrophin. The role of serum human placental
Harrison 1985 {published data only} lactogen determination. South African Medical Journal
Harrison RF. Treatment of habitual abortion with human 1977;51(6):1656.
chorionic gonadotropin: results of open and placebo- Zhuchenko 1988 {published data only}
controlled studies. European Journal of Obstetrics & Zhuchenko PG, Lisovskaia TV, Gordiichuk VD. Effect
Gynecology and Reproductive Biology 1985;20(3):15968. of turinal and progesterone combined with chorionic
Human chorionic gonadotrophin for threatened miscarriage (Review) 8
Copyright 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
gonadotropin on cellular immunity during the treatment gonadotropin in normal and abnormal pregnancies. Early
of threatened abortion in the first and second trimesters. Pregnancy 1997;3(3):21324.
Akusherstvo i Ginekologiia 1988;12:247.
Kayisli 2003
Additional references Kayisli U, Selam B, Guzeloglu-Kayisli O, Demir R, Arici
A. Human chorionic gonadotropin contributes to maternal
Alberman 1992 immunotolerance and endometrial apoptosis by regulating
Alberman E. Spontaneous abortion: epidemiology. In: Fas-Fas ligand system. Journal of Immunology 2003;171(5):
Stabile I, Grudzinskas JG, Chard T editor(s). Spontaneous 230513.
abortion: diagnosis and treatment. London: Springer-Verlag, Nybo Anderson 2000
1992:920. Nybo Andersen AM, Wohlfahrt J, Christens P, Olsen J,
Cunningham 2001 Melbye M. Maternal age and fetal loss: population based
Cunningham FG. Reproductive success and failure. register linkage study. BMJ 2000;320(7251):170812.
Williams obstetrics. 21st Edition. London: McGraw-Hill,
Rai 1997
2001.
Rai R, Cohen H, Dave M, Regan L. Randomised controlled
Deeks 2001
trial of aspirin and aspirin plus heparin in pregnant women
Deeks JJ, Altman DG, Bradburn MJ. Statistical methods
with recurrent miscarriage associated with phospholipid
for examining heterogeneity and combining results from
antibodies (or antiphospholipid antibodies). BMJ 1997;
several studies in meta-analysis. In: Egger M, Davey Smith
314(7076):2537.
G, Altman DG editor(s). Systematic reviews in health care:
meta-analysis in context. London: BMJ Books, 2001. Rai 2006
Rai R, Regan L. Recurrent miscarriage. Lancet 2006;368
Fritz 2001
(9535):60111.
Fritz B, Hallermann C, Olert J, Fuchs B, Bruns M,
Aslan M, et al.Cytogenetic analyses of culture failures by RevMan 2008
comparative genomic hybridisation (CGH)-Re-evaluation The Nordic Cochrane Centre, The Cochrane Collaboration.
of chromosome aberration rates in early spontaneous Review Manager (RevMan). 5.0. Copenhagen: The Nordic
abortions. European Journal of Human Genetics 2001;9(7): Cochrane Centre, The Cochrane Collaboration, 2008.
53947. Rey 2003
Garzia 2004 Rey E, Kahn SR, David M, Shrier I. Thrombophilic
Garzia E, Borgato S, Cozzi V, Doi P, Bulfamante G, Persani disorders and fetal loss: a meta-analysis. Lancet 2003;361
L, et al.Lack of expression of endometrial prolactin in early (9361):9018.
implantation failure: a pilot study. Human Reproduction Rosevear 1999
2004;19(8):19116. Rosevear S. Bleeding in early pregnancy. In: James DK,
Greene 1999 Steer PJ, Weiner CP, Gonik B editor(s). High risk pregnancy:
Greene MF. Spontaneous abortion and major malformations management options. 2nd Edition. London: W.B. Saunders,
in women with diabetes mellitus. Seminars in Reproductive 1999:6189.
Endocrinology 1999;17:12736. Simpson 1992
Grimbizis 2001 Simpson JL. Aetiology of pregnancy failure. In: Stabile I,
Grimbizis GF, Camus M, Tarlatzis BC, Bontis JN, Devroey Grundzinskas JG, Chard T editor(s). Spontaneous abortion:
P. Clinical implications of uterine malformations and diagnosis and treatment. London: Springer-Verlag, 1992:
hysteroscopic treatment results. Human Reproduction 2148.
Update 2001;7(2):16174. Sipila 1992
Hay 1994 Sipila P, Hartikainen- Sorri Al, Oja H, Von Wendt L.
Hay PE, Lamont RF, Taylor-Robinson D, Morgan DJ, Ison Perinatal outcome of pregnancies complicated by vaginal
C, Pearson J. Abnormal bacterial colonisation of the genital bleeding. British Journal of Obstetrics and Gynaecology 1992;
tract and subsequent preterm delivery and late miscarriage. 99(12):95963.
BMJ 1994;308:2958. WHO 1992
Higgins 2008 World Health Organization. International statistical
Higgins JPT, Green S, editors. Cochrane Handbook for classification of diseases and related health problems. 10th
Systematic Reviews of Interventions Version 5.0.1 [updated Edition. Vol. 1, Geneva: WHO, 1992.
September 2008]. The Cochrane Collaboration, 2008. Williams 1991
Available from www.cochrane-handbook.org. Williams MA, Mittendorf R, Lieberman E, Monson RR.
Ho 1997 Adverse infant outcomes associated with first-trimester
Ho HH, OConnor JF, Nakajima ST, Tieu J, Overstreet vaginal bleeding. Obstetrics & Gynecology 1991;78(1):148.
JW, Lasley BL. Characterization of human chorionic
Indicates the major publication for the study
Harrison 1993
Interventions Treatment arm: 5000 IU of hCG twice weekly up to 12 weeks and thereafter once weekly
up to 16 weeks
Control arms:
1. placebo: identical injection regime containing placebo;
2. bedrest.
Outcomes Miscarriage.
Pregnancy complications like pre-ecclampsia, abruption.
Preterm labour.
Mode of delivery.
Birthweight.
Risk of bias
Incomplete outcome data addressed? Yes None were lost to follow up.
All outcomes
Qureshi 2005
Participants 183 women were randomised. 47 did not comply with the study protocol. 9 were lost
to follow up
Exclusions: all women with recurrent miscarriage (3 consecutive), genital tract anomalies,
underlying medical conditions like antiphospholipid syndrome, diabetes mellitus, etc)
Age: 27 years (18 to 41) in the placebo group and 28 years(18 to 40) in the hCG group
Location: Bolton, UK.
Timing and duration: 1997 to 2000.
Interventions HCG: 5000 IU intramuscularly every week from recruitment until 14 weeks.
Placebo: injection of normal saline of similar appearance, following the same regime
Outcomes Miscarriage.
Preterm birth.
Small for dates.
Mode of delivery.
Notes All patients entering the study had a viable fetus confirmed by ultrasound
Risk of bias
Incomplete outcome data addressed? Yes Analysis was by intention to treat. All pa-
All outcomes tients were accounted for
Free of selective reporting? Yes The study reported its pre-specified objec-
tives.
Free of other bias? No 47 patients did not comply with the study
protocol.
Suvonnakote 1986
Interventions Treatment arm: 2000 IU hCG injections thrice a week intramuscularly until bleeding
stopped.
Control arm: bed rest in hospital or home.
Outcomes Miscarriage.
Preterm birth.
Small for gestational age.
Notes All patients entering the study had gestational age and viability confirmed by ultrasound.
Author contacted regarding method of randomisation
Risk of bias
Adequate sequence generation? Unclear Quote: Patients were divided into two
groups. Comment: unclear how patients
were divided
Incomplete outcome data addressed? Yes none were lost to follow up. All patients
All outcomes were accounted for
Free of selective reporting? Yes All the pre-specified objectives of the study
were reported.
Czeizel 2008 Not a randomised controlled trial. This study is a retrospective population based study on hCG for threatened
miscarriage
Eskes 1992 This study deals with hCG for recurrent and not threatened miscarriage
Harrison 1985 This study deals with hCG for women with recurrent and not threatened miscarriage
Harrison 1992 This study deals with hCG for women with recurrent miscarriage and not those with threatened miscarriage
Quenby 1994 hCG was used for patients with recurrent miscarriage and not for threatened miscarriage
Sandler 1979 This study deals with hCG for recurrent and not threatened miscarriage
Toth 2001 Not a randomised controlled trial. hCG was used in combination with magnesium and progesterone
Toth 2003 Not a randomised controlled trial; hCG was used in conjunction with magnesium and progesterone
Ushiroyama 2006 hCG was compared with a traditional Chinese herbal medicine, but a hemostat was also given along with hCG
to arrest bleeding
Vorster 1977 Not a randomised controlled trial. Human placental lactogen was used to select patients for treatment
Zhuchenko 1988 Not a randomised controlled trial; hCG was used in combination with turinal and progesterone
No. of No. of
Outcome or subgroup title studies participants Statistical method Effect size
1 Miscarriage (all trials) 3 303 Risk Ratio (IV, Fixed, 95% CI) 0.66 [0.42, 1.05]
1.1 High quality trials 2 235 Risk Ratio (IV, Fixed, 95% CI) 0.69 [0.40, 1.19]
1.2 Not high quality trial 1 68 Risk Ratio (IV, Fixed, 95% CI) 0.60 [0.25, 1.40]
2 Miscarriage (hCG versus 2 215 Risk Ratio (IV, Fixed, 95% CI) 0.83 [0.46, 1.47]
placebo)
3 Miscarriage (hCG versus bed 2 108 Risk Ratio (IV, Fixed, 95% CI) 0.47 [0.27, 0.82]
rest)
3.1 High quality trial 1 40 Risk Ratio (IV, Fixed, 95% CI) 0.4 [0.20, 0.82]
3.2 Not high quality trial 1 68 Risk Ratio (IV, Fixed, 95% CI) 0.60 [0.25, 1.40]
No. of No. of
Outcome or subgroup title studies participants Statistical method Effect size
1 Preterm birth 3 303 Risk Ratio (IV, Fixed, 95% CI) 0.47 [0.17, 1.33]
1.1 High quality trials 2 235 Risk Ratio (IV, Fixed, 95% CI) 0.71 [0.21, 2.39]
1.2 Not high quality trial 1 68 Risk Ratio (IV, Fixed, 95% CI) 0.15 [0.02, 1.13]
No. of No. of
Outcome or subgroup title studies participants Statistical method Effect size
1 Small-for-gestational age 2 242 Risk Ratio (IV, Fixed, 95% CI) 0.89 [0.25, 3.25]
1.1 High quality trial 1 174 Risk Ratio (IV, Fixed, 95% CI) 0.82 [0.19, 3.57]
1.2 Not high quality trial 1 68 Risk Ratio (IV, Fixed, 95% CI) 1.19 [0.08, 18.31]
Study or subgroup hCG Bed rest Risk Ratio Weight Risk Ratio
n/N n/N IV,Fixed,95% CI IV,Fixed,95% CI
APPENDICES
HISTORY
Protocol first published: Issue 4, 2008
Review first published: Issue 5, 2010
DECLARATIONS OF INTEREST
None known.
INDEX TERMS