Clin Endocrinol 2009.70(2).p.311-21

8/7/2019 Clin Endocrinol 2009.70(2).p.311-21

1/12

Clinical Endocrinology (2009) 70, 311321 doi: 10.1111/j.1365-2265.2008.03369.x

2009 The AuthorsJournal compilation 2009 Blackwell Publishing Ltd 311

O R I G I N A L A R T I C L E

BlackwellPublishingLtd

Clomiphene citrate, metformin or both as first-step approach

in treating anovulatory infertility in patients with polycystic

ovary syndrome (PCOS): a systematic review of head-to-head

randomized controlled studies and meta-analysis

Stefano Palomba*, Renato Pasquali, Francesco Orio Jr and John E. Nestler

*Chair of Gynecology & Obstetrics, University Magna Graecia of Catanzaro, Italy,

Chair of Endocrinology, University of

Bologna, Italy,

Endocrinology, Faculty of Exercise Sciences, University Parthenope of Naples, Italy,

Division of Endocrinology

and Metabolism, Virginia Commonwealth University, Richmond, VA, USA

Summary

Background

To date, no systematic review or meta-analysis has been

published of direct head-to-head studies comparing clomiphene

citrate (CC) vs.

metformin, or the combination of both drugs as

first-line therapy in anovulatory polycystic ovary syndrome (PCOS)

patients seeking pregnancy. The aim of the current paper was to

define, if possible, the best evidence-based recommendations regarding

the use of CC and/or metformin as the initial treatment of PCOS

women with anovulatory infertility.

Design

Systematic review and meta-analysis of the head-to-head

randomized controlled trials (RCTs) available in the literature.

Methods

A bibliographic search was performed using the following

bibliographic databases: Medline, EMBASE, Biological Abstracts,

Cochrane Controlled Trials Register and Cochrane Database of

Systematic Reviews. Reference lists of included studies, other

relevant review articles and textbooks were checked for additional

citations of interest.

Results

Four head-to-head RCTs were identified and qualified

for inclusion in the analysis. No difference in fertility improvement

was observed comparing CC with metformin (O

R

= 122, 95% CI

023655, P

= 0815), whereas a significant (

P


2/12

312

S. Palomba et al.

2009 The Authors

Journal compilation 2009 Blackwell Publishing Ltd, Clinical Endocrinology

, 70

, 311321

A previous meta-analysis

11

aimed to evaluate the effect on

live-birth rate of metformin, CC or both drugs in therapy nave

PCOS patients was published. Furthermore, some studies included

in that meta-analysis did not evaluate the live birth rate,

1316

whereas

in others

16,17

metformin pretreatment was scheduled. In addition,

the conclusions on metformin vs.

CC comparison were drawn using

data analysed with a fixed effects models,

17

even if a significant data

heterogeneity was detected in both pregnancy and live-birth rates.

Recently, a RCT

18

comparing CC, metformin or the combinationof both in Asian women with PCOS was published and that data has

not been included in any of the previous meta-analyses.

712

To date, no systematic review or meta-analysis was specifically

aimed to evaluate the reproductive efficacy of CC, metformin, or

the combination of both drugs as first-line therapy in anovulatory

PCOS patients seeking pregnancy using head-to-head comparisons.

Based on these considerations, the aim of the present report was to

search and review systematically all RCTs available in the literature

to define, if possible, the best evidence-based recommendations

regarding the use of these drugs administered alone or in com-

bination as initial treatment of PCOS women with anovulatory

infertility.

Materials and methods

Ethics

No institutional review board approval was required because only

published data were analysed.

Search strategy

A bibliographic search was performed using the following search

terms: anovulation, infertility, polycystic ovary syndrome, PCOS,

sterility; exposure: antiestrogens, clomiphene citrate, insulin-

sensitizing drug, metformin, ovulation induction; and outcome of

interest abortion, adverse events, compliance, live birth, menses,

menstruation, ovulation and pregnancy. To eliminate bias, no

study filter or language limit was applied.

The following bibliographic databases were searched: Medline

(from January 1966 to January 2008), EMBASE (from January

1980 to January 2008), Biological Abstracts, Cochrane Controlled

Trials Register and Cochrane Database of Systematic Reviews.

Reference lists of included studies, other relevant review articles

and textbooks were checked for additional citations of interest.

Titles and abstracts were screened and potential relevant articles

were identified. Bibliographies of review and retrieved studiesalso were searched for candidate articles. Successively, identified

articles were reviewed for inclusion and exclusion criteria.

The search was run every 3 weeks between January 2008 and

March 2008 to identify new articles.

Only RCTs were identified and head-to-head studies were

selected for the analysis. Specifically, retrospective, casecontrol,

nonrandomized and quasi-randomized trials and case reports/series

were excluded. Two independent reviewers, not blinded at any point

to the authors or sources of publication, identified and selected

the RCTs that met the inclusion criteria. Disagreements between the

two reviewers were resolved by discussion and final decision by the

first author (S.P).

Study populations

The population consisted of women with well defined diagnosis

of PCOS based on the presence of chronic oligo-anovulation

and/or clinical/biochemical hyperandrogenism and/or ovarian

morphology on ultrasonography [National Institute of Health(NIH)

19

and the European Society for Human Reproduction and

Embryology/American Society of Reproductive Medicine (ESHRE/

ASRM)

1

] or other not validated criteria and who were treated for

anovulatory infertility.

All selected studies included the following information:

demographic characteristics [specifically, age and body mass index

(BMI)], criteria used to diagnose PCOS (chronic oligo-anovulation,

oligo-amenorrhea, biochemical/clinical hyperandrogenism, polycystic

ovary at ultrasound), dosages and protocols used for administered

drugs, previous treatment(s) for inducing ovulation, presence of

glucose intolerance or diabetes mellitus, exclusion of male or tubal

factors of infertility and/or subfertility.

Interventions

Three different types of interventions were analysed: CC vs.

metformin, CC plus metformin vs.

CC, and CC plus metformin vs.

metformin. All dosages and schedules for CC and metformin admin-

istration are included in the present review.

Endpoints

Our primary end-point was the live birth rate.

20

Secondary end-points

were the rates of ovulation, pregnancy, abortion and discontinuation

for adverse events.

Studies quality

Quality of the included trials was assessed using the Cochrane

guidelines.

21

The following characteristics were assessed for each

study: allocation concealment, blinding, intention-to-treat (ITT)

analysis and follow-up. The allocation concealment was graded

as adequate (A), unclear (B), or inadequate (C) according to criteria

provided by the Cochrane Menstrual Disorders and Subfertility

Group.

22

Blinding was reported as yes/no/not reported for investi-

gators, patients, outcome assessors, or data analysts. The use of ITT

analysis was recorded and coded as yes/no. Finally, the follow-upperiod was also reported.

Statistical analysis

The outcome measures, defined as dichotomous data, were the pro-

portions of patients reporting ovulation, pregnancy, abortion, live

birth and discontinuation for adverse events, respectively. The analysis

of treatment effect was performed on an ITT basis considering

dropouts and missing data as treatment failures and per-protocol

basis considering the results from evaluated patients alone.


3/12

First-step approach for anovulatory PCOS

313

2009 The Authors


, 70

, 311321

Results for each study were expressed as OR with 95% CI

and combined for meta-analysis to calculate a pooled estimate of

treatment effect for each outcome across studies.

To measure the heterogeneity (the variation in study outcomes

between studies), we used Cochran Q

-test, which is calculated as the

weighted sum of squared differences among individual study effects

and the pooled effect across studies, with the weights being those

used in the pooling method. A Cochran Q

-test P = 005 represents

statistical homogeneity.

For data with statistical homogeneity, the MantelHaenszel

method was used for calculating the weighted summary OR under

the fixed effects model. On the other hand, the random-effects model

of meta-analysis was used in the presence of unexplained statisticalheterogeneity.

A P

-value < 005 of 95% CI not containing 10 OR was considered

statistically significant.

The statistical analysis was performed by use of StatsDirect

software (release 262, 19 February 2007).

Results

The flow-chart of the study selection according to Quality of Report-

ing of Meta-analyses (QUOROM) guidelines

23

is shown in Fig. 1.

Seven RCTs

1315,18,2426

met the initial eligibility criteria. Given

our primary end-point,

20

three studies

1315

were excluded, as the

treatment effectiveness on live-births was not analysed.

Thus, a total of four RCTs

18,2426

were included and analysed

in the present systematic review. Table 1 summarizes the quality of

the RCTs enclosed.

All four articles

18,2426

were published in the English language

between 2005 and 2008 in international journals of very good

quality, i.e. The Journal of Clinical Endocrinology and Metabolism,

The British Medical Journal, The New England Journal of Medicine

and Fertility and Sterility

, as defined by high impact factors and

their ranking in the Journal Citation Report (JCR). Two studies were

performed in Europe,

24,25

one in the United States (U.S),

26

and onein Asia.

18

The allocation concealment was adequately explained in

all studies. Treatment was blinded to both patients and investigators

in three studies,

2426

whereas in one study

18

neither investigators

nor patients were blinded to the treatments. In only one study,

26

data

analysis followed the ITT principle. Two studies

24,26

had a six-month

follow-up period followed by nine-month extension in pregnant

patients to evaluate live births, whereas only a six-month follow-up

was provided for the other two studies.

18,25

The characteristics of the included populations are summarized

in Table 2. Whereas, the treatments received were shown in Table 3.

Fig. 1 Trial flow chart.


4/12

314

S. Palomba et al.

2009 The Authors


, 70

, 311321

Outcomes

The analysis of treatment effect was performed both on an ITT basis

and per-protocol basis, and no difference was detected. Therefore,

only the meta-analyses on a per-protocol are shown.

CCvs.

metformin

Three RCTs evaluated the efficacy of CC vs.

metformin.

18,24,26

Descriptive data

In the study by Palomba et al

.,

24

the cumulative ovulation rate was

similar in women treated with CC or metformin (620%

vs.

840%,

respectively), whereas the pregnancy rate was higher (320% vs.

620%, respectively) and the abortion rate was lower (120% vs.

60%, respectively) in women treated with metformin compared

with those treated with CC. There was no statistical difference in live

births (180% vs.

52%, respectively) in women treated with CC vs.

metformin although a trend (

P = 007) favouring the metformin

group was present.

The subsequent study by Legro et al

.

26

reported markedly different

results. In this study, CC was markedly superior to metformin in

increasing cumulative ovulation rate (751% vs.

553, respectively),

pregnancy (297% vs.

120%, respectively) and live births (225% vs.

72%, respectively). There was no significant difference in abortions

(225% vs.

72%, respectively) between CC and metformin.

Lastly, in the study by Zain et al

.

18

the cumulative ovulation rate

was higher in CC than metformin (590% vs. 237%, respectively),

whereas no differences between groups was reported in pregnancy

(154% vs.

79%, respectively) and live births (154% vs.

79%,

respectively). No abortion was reported in any group.

The rate of discontinuation for adverse events was similar

between treatments in both studies by Palomba et al

.

24

(20% for

each treatment arm) and Legro et al

.

26

(19% vs.

29% for CC

and metformin group, respectively). No drop-out because of

drug-related side-effects was observed in the study by Zain et al

.

18

Meta-analysis

When the data of these studies

18,24,26

were combined, the effect

of metformin did not differ from CC with respect to cumulative

ovulation rate (OR = 155, 95% CI 040599, P

= 0527), pregnancy

rate (OR = 122, 95% CI 023655, P

= 0815), or live birth rate

(OR = 117, 95% CI 016861,P

= 0881) (Fig. 2). However, significant

(

P < 00001) heterogeneity was detected for all three end-points

(Fig. 2).

Similarly, the effect of metformin did not differ from CC with

respect to the rates of abortion (OR = 158, 95% CI 077325,

P

= 0219) or of discontinuation for adverse events (OR = 071,

95% CI 022225, P

= 0765) (Fig. 2). For these two parameters

no significant heterogeneity (

P

= 0219 and 0765, respectively) was

detected (Fig. 2).

Metformin plus CCvs.

CC

Three RCTs evaluated the efficacy of metformin plus CC vs.

CC.

18,25,26

Descriptive data

The combination of metformin plus CC was no more effective

than CC alone in inducing ovulation in the study by Moll et al

.

25

(640% vs. 719%, respectively) and by Zain et al.18

(684% vs. 590%,

respectively), whereas in the study by Legro et al.26

the combination

was significantly (P = 0041) more effective than CC alone (833%

vs. 751%, respectively).

Combination therapy was no better than CC alone with regard

to rates of pregnancy [(396% vs. 456%, respectively) and (383%

vs. 297%, respectively)], abortion [(117% vs. 105%, respectively)

Table 1. Quality assessment of the included RCTs

Study Year of publication Country Allocation concealment Blinding ITT Follow-up

A: Adequate A: Investigators

B: Unclear B: Patients

C: inadequate C: Outcome assessors

Palomba et al.24

2005 Italy A A: Yes No Six cycles plus 9-

B: Yes month extension for

C: Not reported pregnant patients

Moll et al.25

2006 The Netherlands A A: Yes No Six cycles

B: Yes

C: Not reported

Legro et al.26

2007 United States A A: Yes Yes Six cycles plus 9-

B: Yes month extension for

C: Yes pregnant patients

Zain et al.18

2008 Asia A A: No No Six cycles

B: No

C: No

ITT, intention-to-treat; RCTs, randomized controlled trials.


5/12

First-step approach for anovulatory PCOS 315

2009 The Authors

Journal compilation 2009 Blackwell Publishing Ltd, Clinical Endocrinology, 70, 311321

Fig. 2 Comparison CC vs. metformin.

Table2.

Characteristicsoftheincluded

populations

Study

Primaryend-point

Sample(n)

DiagnosisofPCOS

Age(years)

BMI(kg/m2)

Insulinresistance

Previousinfertility

treatments

Exclusionofothercausesforinfertility/sub-fertility

Palombaeta

l.2

4

Pregnancyrate

100

NIHcriteria(1992)

2034

Clin Endocrinol 2009.70(2).p.311-21

Documents

Transcript of Clin Endocrinol 2009.70(2).p.311-21