Aromatase inhibitors (letrozole) for subfertile women with polycystic ovary syndrome

Sebastian Franik; Stephanie M Eltrop; Jan AM Kremer; Ludwig Kiesel; Cindy Farquhar

doi:10.1002/14651858.CD010287.pub3

. 2018 May 24;2018(5):CD010287. doi: 10.1002/14651858.CD010287.pub3

Aromatase inhibitors (letrozole) for subfertile women with polycystic ovary syndrome

Sebastian Franik ^1,^✉, Stephanie M Eltrop ¹, Jan AM Kremer ², Ludwig Kiesel ¹, Cindy Farquhar ³

Editor: Cochrane Gynaecology and Fertility Group

PMCID: PMC6494577 PMID: 29797697

Abstract

Background

Polycystic ovary syndrome (PCOS) is the most common cause of infrequent periods (oligomenorrhoea) and absence of periods (amenorrhoea). It affects about 4% to 8% of women worldwide and often leads to anovulatory subfertility. Aromatase inhibitors (AIs) are a class of drugs that were introduced for ovulation induction in 2001. Since about 2001 clinical trials have reached differing conclusions as to whether the AI letrozole is at least as effective as the first‐line treatment clomiphene citrate (CC).

Objectives

To evaluate the effectiveness and safety of aromatase inhibitors for subfertile women with anovulatory PCOS for ovulation induction followed by timed intercourse or intrauterine insemination (IUI).

Search methods

We searched the following sources from inception to November 2017 to identify relevant randomised controlled trials (RCTs): the Cochrane Gynaecology and Fertility Group Specialised Register, the Cochrane Central Register of Controlled Trials, MEDLINE, Embase, PsycINFO, Pubmed, LILACS, Web of Knowledge, the World Health Organization (WHO) clinical trials register and Clinicaltrials.gov. We also searched the references of relevant articles. We did not restrict the searches by language or publication status.

Selection criteria

We included all RCTs of AIs used alone or with other medical therapies for ovulation induction in women of reproductive age with anovulatory PCOS.

Data collection and analysis

Two review authors independently selected trials, extracted the data and assessed risks of bias. We pooled studies where appropriate using a fixed‐effect model to calculate odds ratios (ORs) and 95% confidence intervals (CIs) for most outcomes, and risk differences (RDs) for ovarian hyperstimulation syndrome (OHSS). The primary outcomes were live birth and OHSS. Secondary outcomes were clinical pregnancy, miscarriage and multiple pregnancy. We assessed the quality of the evidence for each comparison using GRADE methods.

Main results

This is a substantive update of a previous review. We identified 16 additional studies for the 2018 update. We include 42 RCTs (7935 women). The aromatase inhibitor letrozole was used in all studies.

Letrozole compared to clomiphene citrate (CC) with or without adjuncts followed by timed intercourse

Live birth rates were higher with letrozole (with or without adjuncts) compared to clomiphene citrate (with our without adjuncts) followed by timed intercourse (OR 1.68, 95% CI 1.42 to 1.99; 2954 participants; 13 studies; I² = 0%; number needed to treat for an additional beneficial outcome (NNTB) = 10; moderate‐quality evidence). There is high‐quality evidence that OHSS rates are similar with letrozole or clomiphene citrate (0.5% in both arms: risk difference (RD) −0.00, 95% CI −0.01 to 0.00; 2536 participants; 12 studies; I² = 0%; high‐quality evidence). There is evidence for a higher pregnancy rate in favour of letrozole (OR 1.56, 95% CI 1.37 to 1.78; 4629 participants; 25 studies; I² = 1%; NNTB = 10; moderate‐quality evidence). There is little or no difference between treatment groups in the rate of miscarriage by pregnancy (20% with CC versus 19% with letrozole; OR 0.94, 95% CI 0.70 to 1.26; 1210 participants; 18 studies; I² = 0%; high‐quality evidence) and multiple pregnancy rate (1.7% with CC versus 1.3% with letrozole; OR 0.69, 95% CI 0.41 to 1.16; 3579 participants; 17 studies; I² = 0%; high‐quality evidence). However, a funnel plot showed mild asymmetry, indicating that some studies in favour of clomiphene might be missing.

Letrozole compared to laparoscopic ovarian drilling

There is low‐quality evidence that live birth rates are similar with letrozole or laparoscopic ovarian drilling (OR 1.38, 95% CI 0.95 to 2.02; 548 participants; 3 studies; I² = 23%; low‐quality evidence). There is insufficient evidence for a difference in OHSS rates (RD 0.00, 95% CI −0.01 to 0.01; 260 participants; 1 study; low‐quality evidence). There is low‐quality evidence that pregnancy rates are similar (OR 1.28, 95% CI 0.94 to 1.74; 774 participants; 5 studies; I² = 0%; moderate‐quality evidence). There is insufficient evidence for a difference in miscarriage rate by pregnancy (OR 0.66, 95% CI 0.30 to 1.43; 240 participants; 5 studies; I² = 0%; moderate‐quality evidence), or multiple pregnancies (OR 3.00, 95% CI 0.12 to 74.90; 548 participants; 3 studies; I² = 0%; low‐quality evidence).

Additional comparisons were made for Letrozole versus placebo, Selective oestrogen receptor modulators (SERMS) followed by intrauterine insemination (IUI), follicle stimulating hormone (FSH), Anastrozole, as well as dosage and administration protocols.

There is insufficient evidence for a difference in either group of treatment due to a limited number of studies. Hence more research is necessary.

Authors' conclusions

Letrozole appears to improve live birth and pregnancy rates in subfertile women with anovulatory polycystic ovary syndrome, compared to clomiphene citrate. There is high‐quality evidence that OHSS rates are similar with letrozole or clomiphene citrate. There is high‐quality evidence of no difference in miscarriage rates or multiple pregnancy rates. There is low‐quality evidence of no difference in live birth and pregnancy rates between letrozole and laparoscopic ovarian drilling, although there were few relevant studies. For the 2018 update, we added good‐quality trials, upgrading the quality of the evidence.

Plain language summary

Aromatase inhibitors for subfertility treatment in women with polycystic ovary syndrome

Review question: Cochrane authors examined the evidence about aromatase inhibitors (AIs) for subfertile women with polycystic ovary syndrome (PCOS).

Background: PCOS is the most common cause of infrequent or absent menstrual periods, and affects about 4% to 8% of women worldwide. It often causes anovulatory subfertility (subfertility related to failure to ovulate). AIs are used to make ovulation happen. Since about 2001 clinical trials have reached differing conclusions as to whether the AI letrozole is at least as effective for treating subfertility as the most commonly used treatment, clomiphene citrate (CC).

Study characteristics: The review includes clinical studies where participants were randomly assigned to the intervention or to the comparison group (randomised controlled trials, RCTs). Our review includes 42 RCTs with 7935 women. In all trials the aromatase inhibitor used was letrozole. Comparators included CC, which was used in 25 of the RCTs, and laparoscopic ovarian drilling (a surgical technique to puncture the membrane surrounding the ovary), which was used in five RCTs. Several studies included other treatments in one or both arms.

Key results: Letrozole appears to improve live birth and pregnancy rates compared to CC when used to cause ovulation and timed intercourse. The quality of this evidence was moderate and seems to be reliable. There appeared to be no difference for miscarriage rate or multiple pregnancy rate. There appeared to be no difference between letrozole and laparoscopic ovarian drilling for any observed outcomes, although there were few relevant studies. Ovarian hyperstimulation syndrome (OHSS), a serious adverse event of hormonal stimulation, was a very rare event and in most studies it did not occur. The evidence is current to January 2018.

Quality of the evidence: The overall quality of the evidence ranged from moderate to high. Some studies in favour of clomiphene citrate may never have been published. It appears that studies that reported live births report higher clinical pregnancy rates in the letrozole group than studies that failed to report live births. This suggests that results might be somewhat less favourable to letrozole if all studies reported live births.

Summary of findings

Background

Description of the condition

Polycystic ovary syndrome (PCOS) is the most common cause of infrequent periods (oligomenorrhoea) and absence of periods (amenorrhoea), affecting about 4% to 8% of women worldwide in their fertile years (Abu 2012). Many of these women are subfertile, but for most of them it just takes longer to become pregnant naturally and only a small percentage need fertility treatment.

The mechanisms causing PCOS are very complex and the exact pathogenesis remains unknown, but some of the symptoms are believed to be caused by abnormal levels of the pituitary hormones luteinizing hormone (LH) and of the male hormones (androgens) which interfere with the normal function of the ovaries (Azziz 2006).

The diagnosis can be made based on the 'Rotterdam criteria 2003', jointly proposed by the European Society for Human Reproduction and Embryology and the American Society for Reproductive Medicine (Rotterdam 2003). The woman must have two of the following three criteria to be diagnosed with PCOS:

Oligoovulation (infrequent ovulation) or anovulation (absence of ovulation), or both
High male hormone levels (hyperandrogenism) diagnosed either clinically (excessive hair growth, hirsutism) or biochemically (raised serum testosterone levels)
Ovaries which appear to be polycystic on vaginal sonogram, defined by the presence of 12 or more antral follicles in an ovary or an ovarian volume of more than 10 mL. Antral follicles are defined as measuring between 2 and 9 mm in diameter.

Other definitions of PCOS include the National Institutes of Health Criteria (NIH), defined in 1990. They include only the presence of clinical and/or biochemical hyperandrogenism and oligo/amenorrhoea anovulation (Zawadski 1992).

The Androgen Excess Society defines PCOS as hyperandrogenism with ovarian dysfunction or polycystic ovaries (Azziz 2006).

Description of the intervention

There are many possible options for treatment of subfertility in women with anovulatory PCOS.

Clomiphene citrate (CC) is a selective oestrogen receptor modulator (SERM), and is the most common medication used for treating the condition. It was first introduced in 1960 for treatment of World Health Organization (WHO) type II anovulation (a type of subfertility where hormone levels stay normal) in subfertile women, and has been the first‐line treatment ever since. CC is given orally and is relatively safe and inexpensive, but there are also adverse effects associated with it, such as negative changes in endometrium and cervical mucus due to the down‐regulation of oestrogen receptors that might impair implantation after successful induction of ovulation (Casper 2006).

Aromatase inhibitors (AIs) are a newer class of drugs that were introduced for ovulation induction in 2001 by Mitwally and Casper (Mitwally 2001). Since about 2001 data from many clinical trials have been collected and there is evidence that the AI letrozole might be as effective as CC, but the outcome data vary. AIs are administered orally, but due to their short half‐life elimination time of 48 hours there are fewer adverse effects on oestrogen target tissues such as the endometrium and cervix compared with CCs (Baruah 2009; Jirge 2010; Samani 2009). Despite evidence of effectiveness and safety in well‐designed large randomised controlled trials (RCTs), letrozole is still used off‐label for ovulation induction, since it has not been approved by the US Food and Drug Administration (FDA) for this indication (Amer 2017; Legro 2014). A 2005 study (Biljan 2005), including 150 babies, raised some concerns about the teratogenicity of letrozole, but there were major methodological flaws in this study as the intervention group was not well controlled. Two other large studies, including 911 and 470 infants respectively, compared the use of letrozole to CC and spontaneously‐conceiving women. Both reported no higher levels of minor or major congenital malformations or cardiac abnormalities in newborns after use of letrozole for ovulation induction (Forman 2007; Tulandi 2006).

Due to the short half‐life elimination time of letrozole it should be completely cleared out of the system before implantation takes place. Some clinicians recommend testing the blood levels of ß‐hCG prior to treatment with letrozole to exclude pregnancy (Casper 2011). CC and AIs are usually both given for five days, starting on day three of the cycle. The dose for CC ranges from 50 mg to 150 mg a day, and for letrozole from 2.5 mg to 7.5 mg a day (Lee 2011).

Since many women with PCOS experience insulin resistance or impaired glucose tolerance, metformin and other insulin‐sensitising agents were thought to be a superior drug for treatment of ovulation induction (Velázquez 1997). However, the latest version of the Cochrane Review on oral agents for ovulation induction concludes that the use of metformin and other insulin sensitising agents as an adjunct is limited and might be favourable only in women who are resistant to CC alone (Brown 2016).

Human menopausal gonadotropins (hMG) were introduced into clinical practice in 1961 for ovulation induction. They exert a central role in ovulation induction in CC‐resistant subfertile normogonadotropic anovulatory women (Lunenfeld 2004). However, women with PCOS are at particular risk for complications such as ovarian hyperstimulation syndrome (OHSS) and multiple pregnancies, and a low‐dose step‐up protocol was introduced to reach the follicle stimulating hormone (FSH) threshold gradually in order to minimise the risks of OHSS and multiple pregnancies (White 1996). Use of FSH for ovulation induction in women with PCOS appears to be safe and effective (Homburg 2011).

For all the above‐mentioned drugs for ovulation induction, follicular growth should be monitored during a stimulation cycle to reassure effectiveness and also to minimise the occurrence of adverse events, such as multiple pregnancy (Von Hofe 2015).

Finally, another possible option for ovulation induction in cases of CC resistance is laparoscopic ovarian diathermy (or drilling, LOD), during which the damaging of localised areas in the ovarian cortex and stroma seems to have similar success rates compared with gonadotropin therapy (Farquhar 2002). It is not fully understood how the partial destruction of the ovary results in follicle development and ovulation induction (Farquhar 2012). However, long‐term outcomes of a study with eight to 12 years of follow‐up indicate that LOD is safe and effective (Nahuis 2011).

How the intervention might work

AIs down‐regulate the production of oestrogen by inhibiting the cytochrome P450 isoenzymes 2A6 and 2C19 of the aromatase enzyme complex (Cole 1990). They inhibit the negative feedback loop of oestrogen in the hypothalamus, and result in stronger gonadotropin‐releasing hormone (GnRH) pulses. The elevated levels of GnRH stimulate the pituitary gland to produce more FSH, which induces development of follicles in the ovaries. Because AIs do not deplete oestrogen receptors, in contrast to CC, the central feedback mechanism remains intact and as the dominant follicle grows and oestrogen levels rise, normal negative feedback occurs centrally. This results in suppression of FSH and the smaller‐growing follicles will undergo atresia, leading to a single dominant follicle and mono‐ovulation (ovulation of a single egg) in most cases (Casper 2006; Lee 2011). Therefore, by leaving the central mechanism intact, the AIs might lower the risk of high multiple ovulation and OHSS compared to CC.

Why it is important to do this review

Because evidence for and against the effectiveness and safety of AIs has fluctuated over the last decade, and new data based on recent RCTs have become available, an update of the existing Cochrane Review was necessary to provide up‐to‐date information for daily practice.

This review evaluates the effectiveness and safety of AIs compared to other agents for ovulation induction or laparoscopic ovarian drilling, to provide evidence about whether or not AIs should be used in subfertile women with PCOS who are trying to conceive.

Objectives

To evaluate the effectiveness and safety of aromatase inhibitors for subfertile women with anovulatory PCOS for ovulation induction followed by timed intercourse or intrauterine insemination (IUI).

Methods

Criteria for considering studies for this review

Types of studies

We considered randomised controlled trials (RCTs) for inclusion in the review. We excluded cross‐over trials unless phase one data were available separately.

Types of participants

Women of reproductive age with anovulatory PCOS (WHO type II anovulation in women with normogonadotropic normoestrogenic anovulation), diagnosed according to the Rotterdam Criteria (Rotterdam 2003), the NIH consensus criteria (Zawadski 1992) or the AES criteria (Azziz 2009).

Exclusion criteria  We excluded RCTs of women with hyperprolactinaemia or Cushing’s syndrome, or both. We also excluded trials covering women with WHO type I anovulation (hypogonadotropic hypogonadal anovulation). Women in this group have amenorrhoea, low or low‐normal serum FSH concentrations and low serum estradiol concentrations due to decreased hypothalamic secretion of gonadotropin‐releasing hormone (GnRH) or pituitary unresponsiveness to GnRH. We excluded studies using methods other than ovulation induction followed by intercourse or IUI, for example 'in vitro fertilisation' (IVF).

Types of interventions

We considered for inclusion aromatase inhibitors for ovulation induction, alone or in conjunction with medical adjuncts, e.g. metformin or FSH, followed by sexual intercourse or IUI in women with anovulatory subfertility. AIs were compared to each other and to other choices of treatment, including CC, tamoxifen, recombinant and urinary gonadotropin (FSH), insulin‐sensitising agents such as metformin, and laparoscopic ovarian drilling.

Types of outcome measures

Primary outcomes

Effectiveness:

1. Live birth rate by woman randomised, defined as delivery of a live fetus after 20 completed weeks of gestational age.

Adverse effects:

2. OHSS rate by woman randomised, as defined by the study authors.

Secondary outcomes

3. Clinical pregnancy rate by woman randomised, defined as the presence of a gestational sac on ultrasound.

4. Miscarriage rate by woman randomised, defined as the involuntary loss of a clinical pregnancy before 20 weeks of gestation, including partial loss of a multiple pregnancy.

5. Miscarriage rate by pregnancies, defined as the involuntary loss of a clinical pregnancy before 20 weeks of gestation, including partial loss of a multiple pregnancy.

6. Multiple pregnancy rate by woman randomised, defined as more than one intrauterine pregnancy, confirmed by ultrasound or delivery.

Search methods for identification of studies

We searched for all published and unpublished RCTs of the use of AIs for ovulation induction in anovulatory women with PCOS. We consulted the Cochrane Gynaecology and Fertility Group Information Specialist. We used the following search strategy, without language restriction:

Electronic searches

1. Cochrane Gynaecology and Fertility Group Specialised Register (CGFG) (inception to 6 November 2017; Appendix 1)

2. The Cochrane Central Register of Controlled Trials (CENTRAL) (inception to 6 November 2017; Appendix 2)

3. MEDLINE(R) In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE(R) Daily and Ovid MEDLINE(R) (inception to 6 November 2017; Appendix 3)

4. Embase (inception to 6 November 2017; Appendix 4)

5. PsycINFO (inception to 6 November 2017; Appendix 5)

We combined the MEDLINE search with the Cochrane highly sensitive search strategy for identifying randomised trials which appears in theCochrane Handbook of Systematic Reviews of Interventions (Cochrane Handbook) (Version 5.1.0, Section 6.4.11). We combined the Embase search with trial filters developed by the Scottish Intercollegiate Guidelines Network (SIGN) (www.sign.ac.uk/mehodology/filters.html#random). There was no language restriction in these searches.

Searching other resources

We checked the references of relevant systematic reviews and RCTs obtained by the search and contacted experts in the field and manufacturers of aromatase inhibitors, to pick up any additional, relevant data. We also searched the databases of the WHO, clinicaltrials.gov, Web of Knowledge, PubMed and LILACS, Google and Google Scholar, up to January 2018.

Data collection and analysis

We conducted data collection and analysis in accordance with theCochrane Handbook (Higgins 2011).

Selection of studies

For this update of the review, two review authors (SF and SE ) independently selected the trials to be included, in accordance with the aforementioned criteria. We excluded trials from the review if they made comparisons other than those specified above. Studies from non‐English language journals were translated if necessary. If a trial was published more than once, we only included the most complete and up‐to‐date data. We contacted authors of primary studies if papers did not contain enough information to enable an accurate assessment of eligibility for inclusion. We provide a list of excluded studies, showing the reasons for exclusion in the Characteristics of excluded studies table.

Data extraction and management

For this update of the review, two review authors (SF and SE) independently extracted the data, resolving any disagreements by recourse to a third party. We used a data extraction form designed and piloted by the review authors. All data collected for our analyses were dichotomous. If studies had multiple publications, we included only the main trial report. The review authors contacted study investigators to resolve any data queries, as required.

Assessment of risk of bias in included studies

We assessed the included studies for risks of bias, using the Cochrane 'Risk of bias' tool. We evaluated seven domains of possible biases:

Random sequence generation
Allocation concealment
Blinding of participants and personnel
Blinding of outcome assessment
Incomplete outcome data
Selective reporting
Other potential bias

We judged the different types of bias using the criteria from the Cochrane Handbook Table 8.5.d: Criteria for judging risk of bias in the ‘Risk of bias’ assessment tool (Higgins 2011). Two review authors (SF and SE) checked these domains of bias independently and rated them as being at high, low or unclear risk of bias. The assessments were compared and any disagreements resolved by consensus or by discussion with a third review author (CF). The conclusions are presented in the ’Risk of bias’ table and were incorporated into the interpretation of the review findings by means of sensitivity analyses.

Measures of treatment effect

Where dichotomous data measures were used, we have expressed the results in the control and intervention groups of each study as odds ratios (ORs) with 95% confidence intervals (CIs). For the very rare outcome OHSS we have used a risk difference (RD) analysis to allow CIs for the difference in percentage points. Based on the specified outcomes, there were no continuous data measures.

Unit of analysis issues

The primary analysis was by woman randomised. The secondary outcome of miscarriage rate was also analysed by pregnancies. We contacted authors of studies that used cycles as the denominator rather than women, for additional information; if we could not obtain it we did not include the trial in the analysis. If there were multiple cycles, the unit of analysis remained as the woman randomised. We used only the first phase of cross‐over‐trials in the analysis, as successful treatment prevents a cross‐over. We treated multiple live births as one event.

Dealing with missing data

If data were missing from included studies, we contacted the investigators to request the relevant missing data. If this was not possible, we imputed individual values for the primary and secondary outcomes. In participants without a reported outcome we assumed that live births had not occurred. For other outcomes, we analysed only the available data. We subjected any imputation to sensitivity analysis. We analysed the data on an intention‐to‐treat (ITT) basis, as far as possible.

Assessment of heterogeneity

We tested the results of the included studies for heterogeneity by measuring the scatter in the data points on the graph and the overlap in their CIs. We used the I² statistic, which describes the percentage of total variation across the trials that is due to heterogeneity rather than to chance (Higgins 2011). The values of the I² statistic lie between 0% (no heterogeneity) and 100% (extreme heterogeneity). We take values above 50% to indicate moderate heterogeneity and we explored them within sensitivity and subgroup analyses.

Assessment of reporting biases

In view of the difficulty of detecting and correcting for publication bias and other reporting biases, we aimed to minimise their potential impact by ensuring a comprehensive search for eligible studies and by being alert to duplication of data. We compared all outcome measures stated in the Methods section to the outcomes reported in the Results section, to ensure comparability. If there were more than 10 trials included in a comparison, we produced a funnel plot to test for reporting bias.

Data synthesis

We used a fixed‐effect model to combine the data from the primary studies if they were sufficiently similar. We conducted statistical analysis with Review Manager 5, in accordance with the guidelines for statistical analysis developed by Cochrane (Higgins 2011).

Our comparisons were:

Letrozole compared to placebo;
Letrozole compared to other ovulation induction agents followed by intercourse;
Letrozole compared to other ovulation induction agents followed by IUI;
Letrozole compared to laparoscopic ovarian drilling;
Letrozole compared to FSH;
Letrozole compared to anastrozole;
Comparison of different administration protocols of letrozole;
Dosage studies of letrozole.

Increases in the odds of an outcome, either beneficial (e.g. live birth) or detrimental (e.g. OHSS) are shown in the forest plots of the meta‐analysis to the right of the centre‐line.

Subgroup analysis and investigation of heterogeneity

We performed subgroup analysis for primary outcomes only, to evaluate the evidence for a study population with an average body mass index (BMI) above 25 compared to women with an average BMI below 25 within their study group. We conducted a further subgroup analysis comparing women with no previous treatment for ovulation induction to women that were CC‐resistant. We intended to perform subgroup analyses on other parameters, such as the age of the woman, the duration of subfertility and the duration and drug dosages, but this was not possible due to the lack of data.

Sensitivity analysis

We conducted a sensitivity analysis for the primary outcomes to evaluate whether the conclusions are robust to arbitrary decisions made about the eligibility and analysis of studies. This analysis includes consideration of whether the review conclusions would have differed if:

Eligibility was restricted to studies without high or unclear risk of bias;
We had used a random‐effects model;
We had implemented alternative imputation strategies;
The summary effect measure had been the risk ratio instead of the odds ratio.

Overall quality of the body of evidence: 'Summary of findings' table

We generated 'Summary of findings' tables using GRADEpro software. The first table evaluates the overall quality of the body of evidence for the main review comparison (aromatase inhibitors compared to selective oestrogen receptor modulators, with or without adjuncts), using GRADE criteria, i.e. study limitations (risk of bias), consistency of effect, imprecision, indirectness and publication bias. Judgements about evidence quality (high, moderate, low or very low) were justified, documented and incorporated into the reporting of results for each outcome.Two review authors independently evaluated the overall quality of the evidence for the main outcomes of the review (live birth rate, miscarriage rate, clinical pregnancy rate, multiple pregnancy rate, and OHSS by woman randomised). We produced a second ’Summary of findings’ table for the comparison 'Letrozole compared to laparoscopic ovarian drilling'. There are no 'Summary of findings' tables for the remaining comparisons of the review, because we considered them less clinically important. The results of these comparisons are discussed within the text of the review.

Results

Description of studies

Results of the search

The previous version of this review included 26 trials. The searches for the 2018 review update resulted in the retrieval of 38 full‐text papers (Figure 1). We included 16 new studies (Characteristics of included studies). We excluded 10 new studies (Characteristics of excluded studies). Two new studies are awaiting classification (Characteristics of studies awaiting classification); we have contacted their authors and still await a response. We have moved five studies from ongoing to excluded (NCT01679574; NCT01793038; NCT01431352;Palihawadana 2015; Sarvi 2010), and three studies from ongoing to included (Amer 2017; Liu 2017; Wu 2016). Four studies have been moved from awaiting classification to excluded (Al‐Hussaini 2014; NCT01577017; NCT00610077; Sharma 2010). We classify one new study as ongoing (NCT03009838; Characteristics of ongoing studies).

Included studies

Study design and setting

We include 42 parallel‐designed randomised controlled trials (RCTs) in this 2018 updated review.

The studies were done in different parts of the world:

China (Chen 2016; Liu 2015; Liu 2017; Wu 2016)
Egypt (Abdellah 2011; Abu Hashim 2010a; Abu Hashim 2010b; Badawy 2008; Badawy 2009a; Badawy 2009b; El‐Gharib 2015; El‐Khayat 2016; Elgafor 2013; Hassan 2017; Hendawy 2011; Ibrahim 2017; Moussa 2016; Selim 2012)
India (Begum 2009; Ganesh 2009; Kamath 2010; Kar 2012; Ray 2012; Roy 2012; Sh‐El‐Arab Elsedeek 2011)
Iraq (Al‐Omari 2004; Sharief 2015)
Iran (Davar 2011; Dehbashi 2009; Foroozanfard 2011; Ghahiri 2016; Ghomian 2015; Ramezanzadeh 2011; Seyedoshohadaei 2016; Sohrabvand 2006; Zarei 2015; Zeinalzadeh 2010)
Turkey (Atay 2006; Bayar 2006; Nazik 2012)
United Kingdom (Amer 2017)
USA (Legro 2014)

The following different settings recruited women into the trials:

Not stated (Atay 2006; Ray 2012; Roy 2012; Selim 2012).
Subfertility clinic (Amer 2017; Begum 2009; Davar 2011; Dehbashi 2009; Foroozanfard 2011; Ghahiri 2016; Ganesh 2009; Kar 2012; Nazik 2012; Ramezanzadeh 2011; Sh‐El‐Arab Elsedeek 2011; Seyedoshohadaei 2016; Sohrabvand 2006; Zeinalzadeh 2010).
Outpatient clinic (Abu Hashim 2010a; Abu Hashim 2010b; Badawy 2008; Badawy 2009a; Badawy 2009b; Bayar 2006; El‐Gharib 2015; Ibrahim 2017).
Department of obstetrics and gynaecology (Al‐Omari 2004; Chen 2016; Elgafor 2013; Legro 2014).
Division of reproductive endocrinology (Kamath 2010).
Maternity and child hospital (Seyedoshohadaei 2016; Zarei 2015).
University hospital (El‐Khayat 2016; Ghomian 2015; Hassan 2017; Hendawy 2011; Liu 2015; Liu 2017; Moussa 2016; Wu 2016).
Women's health institute (Abdellah 2011).

Drs Abu Hasim and Badawy confirmed to us that their five studies, conducted from 2008 until 2010, were independent and did not include the same women.

Participants

The studies included 7935 women who were subfertile due to anovulatory PCOS. The ages of the women ranged from 18 to 40 years.

Interventions

2/42 studies compared letrozole versus placebo (Kamath 2010; Zarei 2015).

26/42 studies compared letrozole to other ovulation induction agents followed by intercourse (Abu Hashim 2010b; Amer 2017; Atay 2006; Badawy 2009b; Bayar 2006; Begum 2009; Chen 2016; Davar 2011; Dehbashi 2009; El‐Gharib 2015; El‐Khayat 2016; Foroozanfard 2011; Ghahiri 2016; Hendawy 2011; Legro 2014; Liu 2017; Moussa 2016; Nazik 2012; Ray 2012; Roy 2012; Selim 2012; Seyedoshohadaei 2016; Sharief 2015; Sh‐El‐Arab Elsedeek 2011; Sohrabvand 2006; Wu 2016).

3/42 studies compared letrozole to other ovulation induction agents followed by IUI (Ganesh 2009; Kar 2012; Zeinalzadeh 2010).

5/42 studies compared letrozole versus laparoscopic ovarian drilling (Abdellah 2011; Abu Hashim 2010a; Elgafor 2013; Ibrahim 2017; Liu 2015).

1/42 studies compared letrozole versus FSH (Hassan 2017).

2/42 studies compared letrozole versus anastrozole (Al‐Omari 2004; Badawy 2008).

2/42 studies compared different administration protocols of letrozole (Badawy 2009a; Ghomian 2015).

1/26 studies compared different doses of letrozole (Ramezanzadeh 2011).

(See Characteristics of included studies)

Outcomes

16/42 studies reported live birth rate by woman randomised (Abdellah 2011; Abu Hashim 2010a; Abu Hashim 2010b; Amer 2017; Bayar 2006; Begum 2009; Dehbashi 2009; Foroozanfard 2011; Legro 2014; Liu 2015; Liu 2017; Kamath 2010; Ray 2012; Roy 2012; Seyedoshohadaei 2016; Wu 2016).

21/42 studies reported OHSS rate by woman randomised (Abu Hashim 2010a; Abu Hashim 2010b; Badawy 2008; Badawy 2009a; Badawy 2009b; Bayar 2006; Begum 2009; Chen 2016; El‐Khayat 2016; Foroozanfard 2011; Ganesh 2009; Ghahiri 2016; Hassan 2017; Kamath 2010; Legro 2014; Nazik 2012; Ramezanzadeh 2011; Roy 2012; Selim 2012; Zarei 2015; Zeinalzadeh 2010).

41/42 studies reported clinical pregnancy rate by woman randomised (Abdellah 2011; Abu Hashim 2010a; Abu Hashim 2010b; Al‐Omari 2004; Amer 2017; Atay 2006; Badawy 2008; Badawy 2009a; Badawy 2009b; Bayar 2006; Begum 2009; Chen 2016; Davar 2011; Dehbashi 2009; El‐Gharib 2015; El‐Khayat 2016; Elgafor 2013; Foroozanfard 2011; Ganesh 2009; Ghahiri 2016; Ghomian 2015; Hassan 2017; Ibrahim 2017; Kamath 2010; Kar 2012; Legro 2014; Liu 2015; Liu 2017; Moussa 2016; Nazik 2012; Ramezanzadeh 2011; Ray 2012; Roy 2012; Selim 2012; Sh‐El‐Arab Elsedeek 2011; Seyedoshohadaei 2016; Sharief 2015; Sohrabvand 2006; Wu 2016; Zarei 2015; Zeinalzadeh 2010).

30/42 studies reported miscarriage rate by woman randomised and by pregnancies (Abdellah 2011; Abu Hashim 2010a; Abu Hashim 2010b; Badawy 2008; Badawy 2009a; Badawy 2009b; Bayar 2006; Begum 2009; Chen 2016; Davar 2011; Dehbashi 2009; El‐Khayat 2016; Elgafor 2013; Foroozanfard 2011; Ganesh 2009; Ghahiri 2016; Hassan 2017; Ibrahim 2017; Kamath 2010; Kar 2012; Liu 2015; Liu 2017; Nazik 2012; Ramezanzadeh 2011; Ray 2012; Roy 2012; Seyedoshohadaei 2016; Sohrabvand 2006; Wu 2016; Zarei 2015).

30/42 studies reported multiple pregnancy rate by woman randomised (Abdellah 2011; Abu Hashim 2010a; Abu Hashim 2010b; Al‐Omari 2004; Amer 2017; Atay 2006; Badawy 2008; Badawy 2009a; Badawy 2009b; Bayar 2006; Begum 2009; Chen 2016; Dehbashi 2009; El‐Khayat 2016; Foroozanfard 2011; Ganesh 2009; Hassan 2017; Hendawy 2011 Kamath 2010; Kar 2012; Legro 2014; Liu 2015; Nazik 2012; NCT00610077; Ramezanzadeh 2011; Roy 2012; Selim 2012; Sharief 2015; Wu 2016; Zeinalzadeh 2010).

(See Characteristics of included studies)

Excluded studies

We exclude 30 trials from the review. Sixteen of these were identified for the 2018 update (Al‐Hussaini 2014; Al‐Shaikh 2017; Angel 2014; Azmoodeh 2015; Khanna 2013; Li 2016; NCT01431352; NCT01577017; NCT01679574; NCT01793038; Pakrashi 2014; Palihawadana 2015; Pourali 2017; Sharma 2010; Xi 2015; Yun 2015). The primary reasons for exclusion of the studies were inclusion criteria, interventions, inability to obtain study data and no randomisation. (See Characteristics of excluded studies)

Risk of bias in included studies

See Characteristics of included studies; Figure 2; Figure 3.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Allocation

Thirty studies were at low risk of selection bias related to sequence generation. They used computer randomisation, a random‐number table or lottery. The remaining 12 studies did not fully describe their method of randomisation and the contacted authors did not respond, so we rated them at unclear risk of this bias (Figure 2).

Fourteen studies were at low risk of selection bias related to allocation concealment. They used sequentially‐numbered, sealed (opaque) envelopes and the list was kept by a third party during the procedure. The other 28 studies did not describe allocation concealment sufficiently and the authors did not respond to our emails, so we rated them at unclear risk of bias (Figure 2).

Blinding

Seven out of 42 studies described the blinding of participants and personnel, and were thus rated tat low risk of performance bias. Twenty‐five studies did not mention blinding of participants of personnel and the authors did not respond to our emails, so we rated them at unclear risk of bias. Ten studies stated that there was no blinding of participants or personnel or both, and were at high risk of bias (Figure 2).

Eleven of 42 studies reported that the outcome assessors were blinded and were therefore at low risk of bias. Twenty‐three studies did not mention blinding of outcome assessors and the authors did not respond to our email contact, so were rated at unclear risk of bias. Four studies were at high risk of detection bias, because they reported that the outcome assessors were not blinded. Another four studies were also at high risk of bias because the participants were not blinded and it is therefore not plausible that the outcome assessors were blinded. (Figure 2).

Incomplete outcome data

Thirty‐eight of 42 studies included all or nearly all women they had randomised (more than 90%) and were therefore at low risk of attrition bias. Two studies were at unclear risk of attrition bias. One study had peculiar group numbers and none of the other biases were addressed, so we tried without success to contact the authors (Ray 2012). Another study did not report how many women were originally randomised (Sharief 2015). Two studies were at high risk of attrition bias: one study because 13 of 80 women were not analysed. Four women were excluded after randomisation due to an ovarian cyst on sonography on day three. Nine more women were lost to follow‐up without any reasons given (Ramezanzadeh 2011). The second study excluded 28 of 268 women from analysis; 13 were lost to follow‐up, three were excluded, and 12 had complications (Liu 2017; Figure 2).

Selective reporting

Thirty‐nine of the 42 included studies reported the outcomes they had stated in the Methods section, and we therefore judged them to be at low risk of bias. In three of the 42 studies only a few outcomes were presented and the contacted authors did not respond, so we rated them at unclear risk of reporting bias (Figure 2).

Other potential sources of bias

In one study there were substantial baseline differences in age and duration of infertility between the two groups and we deemed the risk of bias to be high (Nazik 2012). In another study, the methods were not well described and the clinical trial registration number led to a different trial (Zarei 2015). We found no potential sources of within‐study bias in the other 40 studies (Figure 2).

Effects of interventions

See: Table 1; Table 2

Summary of findings for the main comparison. Letrozole with or without adjuncts compared to clomiphene citrate (CC) with or without adjuncts for subfertile women with polycystic ovary syndrome.

Letrozole with or without adjuncts compared to clomiphene citrate (CC) with or without adjuncts for subfertile women with polycystic ovary syndrome
Patient or population: subfertile women with polycystic ovary syndrome  Setting: fertility clinics  Intervention: letrozole with or without adjuncts followed by timed intercourse  Comparison: CC with or without adjuncts followed by timed intercourse
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	No of participants  (studies)	Certainty of the evidence  (GRADE)	Comments
Outcomes	Risk with CC with or without adjuncts	Risk with letrozole with or without adjuncts	Relative effect  (95% CI)	No of participants  (studies)	Certainty of the evidence  (GRADE)	Comments
Live birth rate	214 per 1000	314 per 1000  (279 to 352)	OR 1.68  (1.42 to 1.99)	2954  (13 RCTs)	⊕⊕⊕⊝  Moderate ^a
Ovarian hyperstimulation syndrome rate	5 per 1000	5 per 1000  (5 to 5)	RD 0.00  (−0.01 to 0.00)	2536  (12 RCTs)	⊕⊕⊕⊕  High
Clinical pregnancy rate	264 per 1000	359 per 1000  (330 to 390)	OR 1.56  (1.37 to 1.78)	4629  (25 RCTs)	⊕⊕⊕⊝  Moderate ^b
Miscarriage rate by pregnancies	201 per 1000	191 per 1000  (150 to 240)	OR 0.94  (0.70 to 1.26)	1210  (18 RCTs)	⊕⊕⊕⊕  High
Multiple pregnancy rate	17 per 1000	13 per 1000  (7 to 21)	OR 0.69  (0.41 to 1.16)	3579  (17 RCTs)	⊕⊕⊕⊕  High
*The risk in the intervention group (and its 95% confidence interval) is based on the mean risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RD: Risk difference: OR: Odds ratio;
GRADE Working Group grades of evidence  High certainty: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Letrozole compared to laparoscopic ovarian drilling compared to placebo for subfertile women with polycystic ovary syndrome
Patient or population: Subfertile women with polycystic ovary syndrome  Setting: Fertility clinics  Intervention: Letrozole with or without adjuncts followed by timed intercourse  Comparison: Laparoscopic ovarian drilling (LOD) followed by timed intercourse
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	No of participants  (studies)	Certainty of the evidence  (GRADE)	Comments
Outcomes	Risk with LOD	Risk with letrozole	Relative effect  (95% CI)	No of participants  (studies)	Certainty of the evidence  (GRADE)	Comments
Live birth rate	236 per 1000	299 per 1000  (227 to 385)	OR 1.38  (0.95 to 2.02)	548  (3 RCTs)	⊕⊕⊕⊝  Low ^a,b
Ovarian hyperstimulation syndrome rate	0 per 1000	0 per 1000  (0 to 0)	RD 0.00  (−0.01 to 0.01)	260  (1 RCT)	⊕⊕⊝⊝  Low ^c
Clinical pregnancy rate	284 per 1000	336 per 1000  (271 to 408)	OR 1.28  (0.94 to 1.74)	774  (5 RCTs)	⊕⊕⊕⊝  Low ^a,b
Miscarriage rate by pregnancies	145 per 1000	101 per 1000  (49 to 196)	OR 0.66  (0.30 to 1.43)	240  (5 RCTs)	⊕⊕⊕⊝  Moderate ^a
Multiple pregnancy rate	0 per 1000	0 per 1000  (0 to 0)	OR 3.00  (0.12 to 74.90)	548  (3 RCTs)	⊕⊕⊝⊝  Low ^a,b
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RD: Risk difference; OR: Odds ratio;
GRADE Working Group grades of evidence  High certainty: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Date	Event	Description
6 November 2017	New search has been performed	The review has been updated. We conducted a new search on 6th November 2017. We include 16 new studies in the 2018 update ( Amer 2017; Chen 2016; El‐Gharib 2015; El‐Khayat 2016; Ghahiri 2016; Ghomian 2015; Hassan 2017; Hendawy 2011; Ibrahim 2017; Liu 2015; Liu 2017; Moussa 2016; Seyedoshohadaei 2016; Sharief 2015; Wu 2016; Zarei 2015) We added one study as ongoing (NCT03009838).
6 November 2017	New citation required but conclusions have not changed	There is no change in the conclusions of this review.

Date	Event	Description
24 September 2014	Amended	This review has been amended. A new search was conducted on 18.9.14 and new ongoing studies added.
17 July 2014	Amended	Addition of new data made available for Legro 2014. New secondary outcome has been added: Miscarriage rate per pregnancy
26 February 2014	Amended	Correction of search date in Abstract and Methods sections

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Live birth rate	1	36	Odds Ratio (M‐H, Fixed, 95% CI)	3.17 [0.12, 83.17]
2 Ovarian hyperstimulation syndrome rate	2	167	Risk Difference (M‐H, Fixed, 95% CI)	0.00 [‐0.05, 0.05]
3 Clinical pregnancy rate	2	167	Odds Ratio (M‐H, Fixed, 95% CI)	2.88 [1.08, 7.66]
4 Miscarriage rate by woman randomised	2	167	Odds Ratio (M‐H, Fixed, 95% CI)	1.60 [0.26, 9.89]
5 Miscarriage rate by pregnancies	1	20	Odds Ratio (M‐H, Fixed, 95% CI)	0.55 [0.07, 4.56]
6 Multiple pregnancy rate	1		Odds Ratio (M‐H, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Live birth rate	13	2954	Odds Ratio (M‐H, Fixed, 95% CI)	1.68 [1.42, 1.99]
1.1 AIs versus clomiphene citrate	8	1646	Odds Ratio (M‐H, Fixed, 95% CI)	1.79 [1.42, 2.25]
1.2 AI versus clomiphene + metformin	1	250	Odds Ratio (M‐H, Fixed, 95% CI)	1.05 [0.60, 1.81]
1.3 Aromatase inhibitor + metformin compared to clomiphene + metformin	2	194	Odds Ratio (M‐H, Fixed, 95% CI)	1.70 [0.89, 3.23]
1.4 Aromatase inhibitor + FSH compared to clomiphene + FSH	1	120	Odds Ratio (M‐H, Fixed, 95% CI)	1.18 [0.53, 2.61]
1.5 AIs versus clomiphene + estradiol valerate	1	100	Odds Ratio (M‐H, Fixed, 95% CI)	1.48 [0.54, 4.06]
1.6 AIs +/‐ berberine versus berberine	1	644	Odds Ratio (M‐H, Fixed, 95% CI)	1.94 [1.33, 2.84]
2 Live birth rate by BMI	11	2774	Odds Ratio (M‐H, Fixed, 95% CI)	1.69 [1.42, 2.02]
2.1 BMI > 25	7	1678	Odds Ratio (M‐H, Fixed, 95% CI)	1.67 [1.34, 2.09]
2.2 BMI < 25	4	1096	Odds Ratio (M‐H, Fixed, 95% CI)	1.73 [1.31, 2.28]
3 Live birth rate by first‐ or second‐line treatment	13		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only
3.1 No previous ovulation induction	4	1089	Odds Ratio (M‐H, Fixed, 95% CI)	1.61 [1.22, 2.14]
3.2 CC‐resistant women	4	344	Odds Ratio (M‐H, Fixed, 95% CI)	1.78 [1.08, 2.93]
3.3 Unclear or mixed study cohort	5	1521	Odds Ratio (M‐H, Fixed, 95% CI)	1.71 [1.35, 2.16]
4 Impact of allocation bias for live birth rate	13		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only
4.1 Unclear risk of allocation	8	1031	Odds Ratio (M‐H, Fixed, 95% CI)	1.90 [1.42, 2.54]
4.2 Low risk of allocation	5	1923	Odds Ratio (M‐H, Fixed, 95% CI)	1.58 [1.29, 1.95]
5 Impact of detection bias for live birth rate	13	2954	Odds Ratio (M‐H, Fixed, 95% CI)	1.68 [1.42, 1.99]
5.1 High risk of detection	1	64	Odds Ratio (M‐H, Fixed, 95% CI)	2.6 [0.83, 8.13]
5.2 Low risk of detection	7	2083	Odds Ratio (M‐H, Fixed, 95% CI)	1.63 [1.33, 1.99]
5.3 Unclear risk of detection	5	807	Odds Ratio (M‐H, Fixed, 95% CI)	1.76 [1.27, 2.44]
6 Impact of attrition bias for live birth rate	13		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only
6.1 Unclear risk of attrition	1	147	Odds Ratio (M‐H, Fixed, 95% CI)	2.04 [0.93, 4.50]
6.2 Low risk of attrition	11	2539	Odds Ratio (M‐H, Fixed, 95% CI)	1.69 [1.41, 2.03]
6.3 High risk of attrition	1	268	Odds Ratio (M‐H, Fixed, 95% CI)	1.46 [0.85, 2.50]
7 Ovarian hyperstimulation syndrome rate	12	2536	Risk Difference (M‐H, Fixed, 95% CI)	‐0.00 [‐0.01, 0.00]
7.1 AIs versus clomiphene citrate	9	2010	Risk Difference (M‐H, Fixed, 95% CI)	‐0.00 [‐0.01, 0.00]
7.2 AI versus clomiphene + metformin	1	250	Risk Difference (M‐H, Fixed, 95% CI)	0.0 [‐0.02, 0.02]
7.3 Aromatase inhibitor + hMG versus clomiphene + hMG	2	276	Risk Difference (M‐H, Fixed, 95% CI)	0.0 [‐0.04, 0.04]
8 Ovarian hyperstimulation syndrome rate per BMI	11		Risk Difference (M‐H, Fixed, 95% CI)	Subtotals only
8.1 BMI > 25	6	1851	Risk Difference (M‐H, Fixed, 95% CI)	‐0.00 [‐0.01, 0.00]
8.2 BMI < 25	5	605	Risk Difference (M‐H, Fixed, 95% CI)	0.0 [‐0.02, 0.02]
9 Clinical pregnancy rate	25	4629	Odds Ratio (M‐H, Fixed, 95% CI)	1.56 [1.37, 1.78]
9.1 AIs versus clomiphene citrate	17	2930	Odds Ratio (M‐H, Fixed, 95% CI)	1.50 [1.28, 1.76]
9.2 AI versus clomiphene + metformin	1	250	Odds Ratio (M‐H, Fixed, 95% CI)	1.01 [0.60, 1.71]
9.3 Aromatase inhibitor + metformin versus clomiphene + metformin	3	294	Odds Ratio (M‐H, Fixed, 95% CI)	1.86 [1.05, 3.29]
9.4 Aromatase inhibitor + hMG versus clomiphene + hMG	2	276	Odds Ratio (M‐H, Fixed, 95% CI)	1.37 [0.82, 2.27]
9.5 AIs versus tamoxifen	2	135	Odds Ratio (M‐H, Fixed, 95% CI)	1.58 [0.64, 3.90]
9.6 AIs versus clomiphene + estradiol valerate	1	100	Odds Ratio (M‐H, Fixed, 95% CI)	2.47 [0.94, 6.46]
9.7 AIs ± berberine versus berberine	1	644	Odds Ratio (M‐H, Fixed, 95% CI)	2.15 [1.48, 3.13]
10 Impact of allocation bias for clinical pregnancy rate	23		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only
10.1 Unclear risk of allocation	16	1907	Odds Ratio (M‐H, Fixed, 95% CI)	1.77 [1.43, 2.18]
10.2 Low risk of allocation	7	1912	Odds Ratio (M‐H, Fixed, 95% CI)	1.36 [1.12, 1.65]
11 Miscarriage rate by woman randomised	18	3754	Odds Ratio (M‐H, Fixed, 95% CI)	1.39 [1.07, 1.81]
11.1 AIs versus clomiphene citrate	11	2190	Odds Ratio (M‐H, Fixed, 95% CI)	1.37 [0.97, 1.93]
11.2 AI versus clomiphene + metformin	1	250	Odds Ratio (M‐H, Fixed, 95% CI)	1.03 [0.25, 4.23]
11.3 Aromatase inhibitor + metformin versus clomiphene + metformin	3	294	Odds Ratio (M‐H, Fixed, 95% CI)	1.21 [0.52, 2.82]
11.4 Aromatase inhibitor + hMG versus clomiphene + hMG	2	276	Odds Ratio (M‐H, Fixed, 95% CI)	0.84 [0.31, 2.27]
11.5 AIs versus clomiphene + estradiol valerate	1	100	Odds Ratio (M‐H, Fixed, 95% CI)	12.21 [0.66, 226.97]
11.6 AIs +/‐ berberine versus berberine	1	644	Odds Ratio (M‐H, Fixed, 95% CI)	1.63 [0.87, 3.04]
12 Miscarriage rate by pregnancies	18	1210	Odds Ratio (M‐H, Fixed, 95% CI)	0.94 [0.70, 1.26]
12.1 AIs versus clomiphene citrate	11	705	Odds Ratio (M‐H, Fixed, 95% CI)	0.96 [0.65, 1.42]
12.2 AI versus clomiphene + metformin	1	85	Odds Ratio (M‐H, Fixed, 95% CI)	1.03 [0.24, 4.40]
12.3 Aromatase inhibitor + metformin versus clomiphene + metformin	3	79	Odds Ratio (M‐H, Fixed, 95% CI)	0.80 [0.32, 2.02]
12.4 Aromatase inhibitor + hMG versus clomiphene + hMG	2	104	Odds Ratio (M‐H, Fixed, 95% CI)	0.67 [0.23, 1.96]
12.5 AIs versus clomiphene + estradiol valerate	1	24	Odds Ratio (M‐H, Fixed, 95% CI)	8.13 [0.39, 167.90]
12.6 AIs +/‐ berberine versus berberine	1	213	Odds Ratio (M‐H, Fixed, 95% CI)	0.88 [0.43, 1.80]
13 Multiple pregnancy rate	17	3579	Odds Ratio (M‐H, Fixed, 95% CI)	0.69 [0.41, 1.16]
13.1 AIs versus clomiphene citrate	13	2409	Odds Ratio (M‐H, Fixed, 95% CI)	0.61 [0.32, 1.16]
13.2 AI versus clomiphene + metformin	1	250	Odds Ratio (M‐H, Fixed, 95% CI)	0.14 [0.01, 2.82]
13.3 Aromatase inhibitor + hMG versus clomiphene + hMG	2	276	Odds Ratio (M‐H, Fixed, 95% CI)	0.94 [0.29, 3.05]
13.4 AIs +/‐ berberine versus berberine	1	644	Odds Ratio (M‐H, Fixed, 95% CI)	4.53 [0.24, 84.46]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Ovarian hyperstimulation syndrome rate	2	1494	Risk Difference (M‐H, Fixed, 95% CI)	‐0.00 [‐0.01, 0.00]
1.1 AI versus Clomiphene	1	107	Risk Difference (M‐H, Fixed, 95% CI)	‐0.02 [‐0.07, 0.03]
1.2 AI versus Clomiphene +rFSH and rFSH only	1	1387	Risk Difference (M‐H, Fixed, 95% CI)	‐0.00 [‐0.01, 0.00]
2 Clinical pregnancy rate	3	1597	Odds Ratio (M‐H, Fixed, 95% CI)	1.71 [1.30, 2.25]
2.1 AI versus Clomiphene	2	210	Odds Ratio (M‐H, Fixed, 95% CI)	2.09 [0.97, 4.53]
2.2 AI versus Clomiphene +rFSH and rFSH only	1	1387	Odds Ratio (M‐H, Fixed, 95% CI)	1.66 [1.23, 2.22]
3 Miscarriage rate by woman randomised	2	1490	Odds Ratio (M‐H, Fixed, 95% CI)	1.22 [0.62, 2.40]
3.1 AI versus Clomiphene	1	103	Odds Ratio (M‐H, Fixed, 95% CI)	0.32 [0.01, 8.06]
3.2 AI versus Clomiphene +rFSH and rFSH only	1	1387	Odds Ratio (M‐H, Fixed, 95% CI)	1.32 [0.66, 2.65]
4 Miscarriage rate by pregnancies	2	260	Odds Ratio (M‐H, Fixed, 95% CI)	0.76 [0.37, 1.57]
4.1 AI versus Clomiphene	1	15	Odds Ratio (M‐H, Fixed, 95% CI)	0.10 [0.00, 3.09]
4.2 AI versus Clomiphene +rFSH and rFSH only	1	245	Odds Ratio (M‐H, Fixed, 95% CI)	0.85 [0.40, 1.79]
5 Multiple pregnancy rate	3	1597	Odds Ratio (M‐H, Fixed, 95% CI)	1.03 [0.49, 2.13]
5.1 AI versus Clomiphene	2	210	Odds Ratio (M‐H, Fixed, 95% CI)	3.48 [0.14, 87.49]
5.2 AI versus Clomiphene +rFSH and rFSH only	1	1387	Odds Ratio (M‐H, Fixed, 95% CI)	0.94 [0.44, 2.03]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Live birth rate	3	548	Odds Ratio (M‐H, Fixed, 95% CI)	1.38 [0.95, 2.02]
2 Ovarian hyperstimulation syndrome rate	1		Risk Difference (M‐H, Fixed, 95% CI)	Totals not selected
3 Clinical pregnancy rate	5	774	Odds Ratio (M‐H, Fixed, 95% CI)	1.28 [0.94, 1.74]
3.1 AI versus LOD	4	628	Odds Ratio (M‐H, Fixed, 95% CI)	1.30 [0.93, 1.83]
3.2 AI + metformin versus LOD	1	146	Odds Ratio (M‐H, Fixed, 95% CI)	1.20 [0.60, 2.39]
4 Miscarriage rate by woman randomised	5	774	Odds Ratio (M‐H, Fixed, 95% CI)	0.81 [0.38, 1.70]
4.1 AI versus LOD	4	628	Odds Ratio (M‐H, Fixed, 95% CI)	0.69 [0.29, 1.63]
4.2 AI + metformin versus LOD	1	146	Odds Ratio (M‐H, Fixed, 95% CI)	1.35 [0.29, 6.27]
5 Miscarriage rate by pregnancies	5	240	Odds Ratio (M‐H, Fixed, 95% CI)	0.66 [0.30, 1.43]
5.1 AI versus LOD	4	191	Odds Ratio (M‐H, Fixed, 95% CI)	0.54 [0.22, 1.33]
5.2 AI + metformin versus LOD	1	49	Odds Ratio (M‐H, Fixed, 95% CI)	1.21 [0.24, 6.09]
6 Multiple pregnancy rate	3	548	Odds Ratio (M‐H, Fixed, 95% CI)	3.00 [0.12, 74.90]

Methods	Randomised controlled trial Duration and location of the trial: quote: "The present study was conducted from July 15, 2007, to February 28, 2010, at the Women's Health Center, Assiut University, Assiut, Egypt, after approval was received from the Ethics Committee of Assiut University.“
Participants	Inclusion criteria: all participants met the Rotterdam consensus criteria for the diagnosis of PCOS. Other inclusion criteria included primary or secondary infertility because of anovulation for at least 1 year and clomiphene resistance. Clomiphene resistance was defined as lack of ovulation after 6 consecutive induction cycles with 50 mg of CC, then with 150 mg of CC each day for 5 days in each cycle. The male partner of each participant was required to have a normal result on semen analysis and each woman was required to have patent tubes on hysterosalpingography or on a diagnostic laparoscopy. Exclusion criteria: exclusion criteria included age below 20 years or above 35 years; hormonal treatment within 3 months prior to the study; hyperprolactinaemia (morning plasma prolactin concentration 30 ng/ml or more); any other endocrine, hepatic, or renal disorder; presence of an organic pelvic mass; and a history of abdominal surgery that might have caused pelvic factor infertility. Number of women randomised: 147, 74 in the letrozole group and 73 in the LOD group Number of women analysed: 70 in the letrozole group and 70 in the LOD group Number of withdrawals/exclusions/loss to follow‐up and reasons: 7 women were lost to follow‐up. Number of centres: 1, Women’s Health Center, Assiut University, Assiut Age (y): group A letrozole: 23.9 ± 3.2, group B LOD: 23.6 ± 3.2 BMI (kg/m²): group A letrozole: 27.3 ± 2.6, group B LOD: 27.1 ± 2.6 Duration of infertility (y): group A letrozole: 4.2 ± 1.7, group B LOD: 4.2 ± 1.7 Country: Egypt
Interventions	Group A: letrozole, 5 mg/day given orally for 5 days during cycle days 3 ‐ 7 for up to 6 cycles Group B: LOD, triple‐puncture laparoscopy, monopolar diathermy, needle electrode set at 40 W pressed against border of ovary for 4 sec to achieve penetration depths of 7 ‐ 8 mm, punctured at 4 ‐ 6 points
Outcomes	Primary outcomes: ovulation rate Secondary outcomes: endometrial thickness on the day of hCG injection, rates of clinical pregnancy, spontaneous abortion, live birth and multiple pregnancies
Notes	Ethical approval: yes, the study was approved by Mansoura University Hospital Research Ethics Committee. Informed consent: yes, all participants gave informed consent before inclusion in the trial. Source of funding: not stated Conflicts of interest: quote:“Conflict of interest statement: We declare that we have no conflict of interest” Authors contacted about information on OHSS Power calculation: quote: "the sample size required to detect a 25% difference between the 2 groups with a power of 80% was estimated to be 68 patients per group."
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was performed using a computer‐generated random numbers table.
Allocation concealment (selection bias)	Low risk	Allocation concealment was achieved using serially‐numbered opaque envelopes that were only opened once the interventions were assigned.
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	Not mentioned
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Not mentioned
Incomplete outcome data (attrition bias)   All outcomes	Low risk	7 women lost to follow‐up, but similar (3 vs 4) in both groups; losses due to noncompliance
Selective reporting (reporting bias)	Low risk	All expected outcomes reported
Other bias	Low risk	None

Methods	Randomised controlled clinical trial Duration and location of the trial: quote: "The study comprised of 260 women with CC‐resistant PCOS among those attending the Outpatient Clinic in Mansoura University Hospitals, and a private practice setting in the period from August 2006 to March 2009.“
Participants	Inclusion criteria: infertile women with CC resistance and PCOS based on the Rotterdam criteria 2003. Patent fallopian tubes proved by hysterosalpingography and normal semen analysis for their partners according to the modified criteria of WHO. Exclusion criteria: other causes of infertility, age over 40 years, BMI > 35, contraindication to general anaesthetic, previous history of LOD and women who had received metformin, gonadotropin, oral contraceptives or other hormonal drugs during the preceding 6 months. Women who intended to start a diet or a specific programme of physical activity were also excluded. Number of centres: 2, outpatient clinic in Mansoura University hospitals and a private practice setting Number of women randomised: 260, 128 in the letrozole group and 132 in the LOD group Number of women analysed: 128 in the letrozole group and 132 in the LOD group Number of withdrawals/exclusions/loss to follow‐up and reasons: none Age (y): group A letrozole: 27.3 ± 2.6, group B LOD: 26.4 ± 2.4 BMI (kg/m²): group A letrozole: 26.4 ± 3.3, group B LOD: 26.6 ± 3.6 Duration of infertility (y): group A letrozole: 4.3 ± 1.11, group B LOD: 4.5 ± 1.24 Country: Egypt
Interventions	Group A: letrozole, 2.5 mg/day orally given for 5 days starting from day 3 of the cycle Group B: LOD, laparoscopy was performed using 3‐puncture technique. Each ovary was cauterised at 4 points, each for 4 s at 40 W for a depth of 4 mm with a mixed current, using a monopolar electrosurgical needle.
Outcomes	Primary outcome: ovulation rate Secondary outcomes: midcycle endometrial thickness (mm), biochemical pregnancy/cycle, clinical pregnancy/participant, biochemical miscarriage/cycle, clinical miscarriage/participant and live birth rates
Notes	Ethical approval: yes, the study was approved by Mansoura University Hospital Research Ethics Committee. Informed consent: yes, all participants gave informed consent before inclusion in the trial. Source of funding: not stated Conflicts of interest: quote:“Conflict of interest statement: We declare that we have no conflict of interest” Power calculation: quote: "Sample size was calculated based on the fact that with an expected rate of ovulation of 70% in the LOD group we needed 244 women to show an absolute increase of 15% in ovulation rate in the letrozole group, with a power of 80% at confidence interval of 95% using a two tailed chi‐square test with a 5% significance level (type alfa error)." We had contact with Prof. Abu Hashim, and all questions were answered in detail.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Women were randomised according to a computer‐generated random‐numeric table prepared by an independent statistician.
Allocation concealment (selection bias)	Low risk	Concealment of treatment allocation was done by using sealed opaque envelopes that were given to a third party (nurse) who assigned participants to study arms.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Quote: "Once allocated, the treatment was revealed to both the investigator and the patient."
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Quote: "Once allocated, the treatment was revealed to both the investigator and the patient."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No dropouts reported
Selective reporting (reporting bias)	Low risk	All expected outcomes reported
Other bias	Low risk	None

Methods	Randomised double blind clinical trial Duration and location of the trial: not stated
Participants	Inclusion criteria: non‐fertile clomiphene‐resistant women with PCOS Exclusion criteria: tubal, peritoneal and uterine causes of infertility were excluded by laparoscopic hysterosalpingogram or by ultrasonography. Specific endocrine abnormalities and male‐factor causes for infertility were also excluded. Participants had to end clomiphene treatment at least 2 months before enrolment. Number of centres: quote: "The study was done in the Baghdad teaching hospital/ Medical city which is a tertiary ref. hospital affiliated with Baghdad Med college/ University of Baghdad." (Email) Number of women randomised: 22 in the letrozole group and 18 in the anastrozole group Number of women analysed: 22 in the letrozole group and 18 in the anastrozole group Number of withdrawals/exclusions/loss to follow‐up and reasons: 0 Age (y): group A letrozole: 28.4 ± 5.18, group B anastrozole: 25.56 ± 6.26 BMI (kg/m²): group A letrozole: 29.95 ± 3.73, group B anastrozole: 27.90 ± 5.29 Duration of infertility (y): group A letrozole: 3.95 ± 2.70, group B anastrozole: 4.50 ± 3.61 Country: Iraq
Interventions	Group A: letrozole 2.5 mg/day orally given for 5 days during cycle days 3 ‐ 7 Group B: anastrozole 1 mg/day orally given for 5 days during cycle days 3 ‐ 7 Treatment was continued for 3 months. When ovulation or pregnancy did not occur, the same treatment protocol was used with the doubling of the first dose for a maximum of 2 treatment cycles.
Outcomes	Primary outcomes: ovulation rate/cycle, endometrial thickness (mm) measured on day of hCG administration Secondary outcomes: multiple pregnancy rate, pregnancy rate/cycle, E2 (pmol/l), progesterone (nmol/l), LH (U/l), number and size of follicles, pulsatility index, day of hCG administration
Notes	Ethical approval: quote: "Ethical approval was obtained from the Iraqi Board for medical specialization/ Scientific committee" (email contact) Informed consent: quote: "written consent was obtained from all patients" (email contact) Source of funding: quote: "The study was partially funded by the Iraqi Board for medical specialization as well as the Drug Scientific Office of the Iraqi Ministry of Health." Power calculation: not reported We had email contact with Dr. Al‐Omari, but there was no further information available about the outcomes.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Actually, we just put all envelopes in a box, mixing them then the patient herself selected one." (email with Dr. Al‐Omari)
Allocation concealment (selection bias)	Unclear risk	Quote: "My associate informed me that for randomisation we distributed blank envelops containing the medications at our Gyn.clinic on twice weekly basis." (email with Dr. Al‐Omari)
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Participants were not blinded
Blinding of outcome assessment (detection bias)   All outcomes	High risk	It is not plausible that outcome assessors were blinded if participants were not
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No dropouts reported
Selective reporting (reporting bias)	Low risk	All expected outcomes reported
Other bias	Low risk	None

Methods	Randomised double‐blind controlled clinical trial Duration and location of the trial: quote: "This study was conducted at the Fertility Unit, Derby Teaching Hospitals NHS Foundation Trust between April 2007 and June 2014.“
Participants	Inclusion criteria: quote: "eligible participants were women aged 18 – 39 years with BMI ≤ 35 kg/m2, anovulatory infertility, and a diagnosis of PCOS based on Rotterdam consensus (two of three criteria: oligo‐/anovulation, hyperandrogaenemia and sonographic appearance of polycystic ovaries) (Rotterdam ESHRE/ASRM‐sponsored PCOS consensus workshop group, 2004). Diagnosis of oligo‐/anovulation was based on a menstrual pattern of oligo‐/amenorrhoea (cycle > 35 days) and/or a low mid‐luteal serum progesterone concentration. Hyperandrogenaemia was diagnosed either clinically (acne/hirsutism) or biochemically (testosterone ≥ 2.5 nmol/l or free androgen index [FAI] ≥ 5). Ultrasound criteria included ≥ 12 follicles (2 – 9 mm) and/or an ovarian volume of > 10 ml (Jonard et al., 2003). All participants had proven patency of at least one fallopian tube and normal semen analysis of their male partners (WHO, 1999)." Exclusion criteria: quote: "We excluded patients who have received OI within 6 months and those with uncontrolled thyroid disease or hyperprolactinaemia. Patients with marked hyperandrogaenemia were screened for adult onset congenital adrenal hyperplasia (by measuring serum 17‐ α ‐hydroxyl‐progesterone concentration) and Cushing syndrome (by measuring urinary free cortisol)." Number of women randomised: 159 women were randomised in total; 79 to CC, 80 to letrozole Number of women analysed: all women randomised were also analysed in the ITT analysis. Number of withdrawals/exclusions/loss to follow‐up and reasons: 3 women in the CC group discontinued treatment due to failing to attend; also 3 women discontinued treatment in the letrozole arm (1 due to social reasons, 1 failed to attend, 1 withdrew consent). Number of centres: this was a single centre, 2‐arm double‐blind RCT. Age (y): letrozole: 28.3 (4.4) vs CC: 28.1 (4.2) BMI (kg/m²): letrozole: 27.5 (23.4 ‐ 32.2) vs CC: 27.7 (23.0 ‐ 31.0) Duration of infertility (y): 1.5 (1.0 ‐ 2.0) for both groups Country: United Kingdom
Interventions	Group A: letrozole was prescribed (by the senior investigator, SA) orally daily for 5 days starting on Days 2 – 4 of a menstrual period or a progestogen‐induced bleed (medroxy‐progesterone acetate 10 mg twice daily for 5 days). The starting dose was 1 tablet/day (letrozole 2.5 mg) and if ovulation was not achieved, the dose would be doubled in the second cycle. Group B: CC was prescribed daily for 5 days starting on days 2 – 4 of a menstrual period or a progestogen‐induced bleed (medroxy‐progesterone acetate 10mg twice daily for 5 days). The starting dose was 1 tablet/day (50 mg) and if ovulation was not achieved, the dose would be doubled in the second cycle. Participants who failed to ovulate on the maximum dose (2 tablets) or to conceive after 6 ovulatory cycles were crossed over to the other drug (after a 6‐week wash‐out period) following the same procedures as with the first drug.
Outcomes	Primary outcomes: clinical pregnancy (diagnosed by ultrasonographic visualisation of a gestational sac) rate per participant on primary treatment (before the cross‐over). Secondary outcomes: secondary outcomes included ovulation, live birth, pregnancy by ovulating participant, pregnancy by strata, mono‐ovulation, endometrial development (thickness and grades), pregnancy outcome and pregnancy complications. Other outcomes included pregnancy and live birth rates on secondary and overall (primary and secondary) treatments.
Notes	Ethical approval: the trial was approved by West Midlands Research Ethics Committee (Reference: 07/MRE07/5) and by the Medicines and Healthcare Products Regulatory Agency (MHRA). Informed consent: all participants gave written informed consent and the trial was monitored by the sponsor. Source of funding: it was sponsored by the University of Nottingham. Power calculation: quote: "to detect a clinically significant difference of 20% between the previously reported pregnancy rate of CC (˜35%) and letrozole with a two‐sided 5% significance level and power of 80%, a sample size of 212 participants (106 per arm) was required (Dickey and Holtkamp, 1996;Kousta et al., 1997; Imani et al., 2002)."
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "An independent pharmacist randomly allocated participants to letrozole or CC, in 1:1 ratio according to a randomisation list created by the trial statistician using NQuery Advisor v6.0 software. Randomization was stratified by patients’ BMI (non‐obese < 30 kg/m² and obese 30 – 35 kg/m²) using mixed block sizes."
Allocation concealment (selection bias)	Low risk	Quote: "An independent pharmacist randomly allocated participants to letrozole or CC, in 1:1 ratio."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "Investigators, patients, outcome assessors and the statistician were blinded to the allocation of participants."
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "Investigators, patients, outcome assessors and the statistician were blinded to the allocation of participants."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Quote: "Intention‐to‐treat (ITT) analysis included all randomised subjects, regardless of whether or not they received the study drug. Per protocol (PP) analysis included all randomised subjects who received the study drug and were not lost to follow‐up. Participants who were lost to follow‐up were assumed neither to be pregnant nor to have given LB in the ITT analysis."
Selective reporting (reporting bias)	Low risk	All expected outcomes reported
Other bias	Low risk	None

Methods	Randomised controlled study Duration and location of the trial: quote: "The study comprised 438 women (1063 cycles) with PCOS among those attending the Fertility Outpatient Clinic in Mansoura University Hospitals, Mansoura University, Egypt, and private practices in the period from January 2004 and September 2006.“
Participants	Inclusion criteria: infertile women with clomiphene‐resistant PCOS, diagnosis of PCOS based on the 2003 Rotterdam criteria. Normal serum PRl, TSH and 17OH‐P. Exclusion criteria: infertility caused by fallopian tube problems, infertility problems caused by male partner Number of centres: 2, outpatient clinic in Mansoura University Hospitals and a private practice setting Number of women randomised: 110 in the short letrozole group and 108 in the long letrozole group Number of women analysed: 110 in the short letrozole group and 108 in the long letrozole group Number of withdrawals/exclusions/loss to follow‐up and reasons: 0 Age (y): group A short letrozole: 25.1 ± 3.2, group B long letrozole: 25.3 ± 3.0 BMI (kg/m²): group A short letrozole: 33.9 ± 3.1, group B long letrozole: 34.2 ± 2.6 Duration of infertility (y): not stated Country: Egypt
Interventions	Group A: letrozole orally given, 5 mg/day for 5 days, from cycle days 3 ‐ 7 Group B: letrozole orally given, 2.5 mg/day for 10 days, from cycle days 3 ‐ 12
Outcomes	Primary Outcomes: number of growing and mature follicles, serum E2 (pg/mL), serum progesterone (ng/mL), and endometrial thickness (mm) Secondary Outcomes: pregnancy rate, miscarriage rate, multiple pregnancies, OHSS
Notes	Ethical approval: yes, the study was approved by the hospital ethics research committee. Informed consent: yes, all participants gave informed consent before inclusion in the trial Source of funding: not stated Conflicts of interest: quote:"All authors have nothing to disclose" Power calculation: quote: "Sample size calculation, using StatCalc 3.02 computer package 8Acastat software, Leesburg, VA), showed that each arm should contain at least 103 patients to have 80% power of the study at 95% confidence interval (CI)." We contacted the authors, but information on live birth was not collected.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were randomly allocated into two treatment groups using a computer‐generated random table: short letrozole group and long letrozole group."
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	Not stated
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Not stated
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No dropouts reported
Selective reporting (reporting bias)	Low risk	All stated outcomes were reported
Other bias	Low risk	None

Methods	Prospective randomised trial Duration and location of the trial: quote: "The study comprised 218 women (444 cycles) with clomiphene‐resistant PCOS among those attending the gynecology outpatient clinic in Mansoura University Hospitals, Egypt, and a private practice setting in the period from December 2005 to December 2007.“
Participants	Inclusion criteria: diagnosis of PCOS based on the 2003 Rotterdam criteria. All women had patent fallopian tubes proved by hysterosalpingography and their partners had normal semen analysis parameters according to the modified criteria of the WHO. All participants had normal serum prolactin, TSH and 17‐OH progesterone. Exclusion criteria: not stated Number of centres: multiple, University teaching hospital and private practices Number of women randomised: 218 in the letrozole group and 220 in the CC group Number of women analysed: 218 in the letrozole group and 220 in the CC group Number of withdrawals/exclusions/loss to follow‐up and reasons: 0 Age (y): group A letrozole: 27.1 ± 3.2, Group B clomiphene citrate: 29.3 ± 2.9 BMI (kg/m²): group A letrozole: 28.1 ± 3.2, group B CC: 27.1 ± 3.1 Duration of infertility (y): not reported Country: Egypt
Interventions	Group A: letrozole orally given 5 mg/day for 5 days from cycle days 3 ‐ 7 Group B: clomiphene citrate orally given 100 mg/day for 5 days from cycle days 3 ‐ 7
Outcomes	Primary Outcomes: number of growing and mature follicles, the concentrations of serum E2 (pg/mL) and progesterone (ng/mL), and the endometrial thickness (mm) Secondary Outcomes: ovulation rate, ovarian hyperstimulation rate, pregnancy rate, miscarriage rate, multiple pregnancy rate
Notes	Ethical approval: yes, the study was approved by the hospital research ethics committee. Informed consent: yes, all participants gave informed consent before inclusion in the trial Source of funding: the study was self‐funded Power calculation: not stated Authors were contacted via email and gave all information, but they did not measure the live birth.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomly allocated using a computer‐generated random table into 2 groups
Allocation concealment (selection bias)	Low risk	Quote: "Allocation concealment was done by sequentially numbered opaque sealed envelopes opened by the chief nurse" (by email contact with authors)
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Not blinded, email with Prof. Badawy
Blinding of outcome assessment (detection bias)   All outcomes	High risk	It is not plausible that outcome assessors were blinded if participants were not
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No dropouts reported
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported, Live birth was not measured
Other bias	Low risk	None

Methods	Randomised non‐blinded controlled trial Duration and location of the trial: quote: "The study was conducted in a private infertility care setting as a randomized controlled trial between August 2004 and December 2005.“
Participants	Inclusion criteria: infertile women with PCOS diagnosed by the Rotterdam criteria 2003 who failed to ovulate by taking 100 mg of CC/day for 5 days in 2 consecutive cycles Exclusion criteria: women with hyperprolactinaemia, thyroid disorder, male‐factor infertility, known or suspicious tubal factor infertility (endometriosis and pelvic inflammatory disease), and unexplained infertility were excluded from the study. Number of centres: 1, private infertility care setting Number of women randomised: 32 in each group Number of women analysed: 32 in each group Number of withdrawals/exclusions/loss to follow‐up and reasons: 0 Age (y): group A letrozole: 25.5 ± 4.0, group B clomiphene citrate: 26.1 ± 3.6 BMI (kg/m²): group A letrozole: 22.7 ± 2.8, group B clomiphene citrate: 23.6 ± 3.2 Duration of infertility (y): group A letrozole: 2.7 ± 1.1, group B clomiphene citrate: 2.6 ± 1.1 Country: India
Interventions	Group A: letrozole, 7.5 mg/day orally given for 5 days from cycle days 3 ‐ 7 Group B: clomiphene citrate, 150 mg/day orally given for 5 days from cycle days 3 ‐ 7
Outcomes	Primary Outcomes: ovulation and pregnancy rate Secondary Outcomes: follicular development by day 16 (mm), serum E2 on day of hCG (pg/mL), endometrial development by day 16 (mm), serum progesterone on day 21 (ng/mL), multiple pregnancies, OHSS cases. Live birth rate was provided by email contact.
Notes	Ethical approval: yes, the study protocol was approved by the institutional review board (IRB) of Dhaka medical college. Informed consent: yes, participants were counselled and informed consent was obtained before recruitment. Source of funding: the study was self‐funded. Power calculation: a study population of 57 women was calculated, considering an average of 60% of PCOS women are associated with insulin resistance, allowing an alfa value of 0.05. Authors were contacted by email, and additional information was provided.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was done by lottery method. They put the name of letrozole and CC in a sealed opaque envelope. By calculating sample size they made 64 pieces of paper, 32 for letrozole and 32 for CC.
Allocation concealment (selection bias)	Unclear risk	Quote: "All unleveled envelop were put together and the patients drew one piece of envelop from them. Then we opened the envelop to see the name of the drug." (by email contact with Prof. Rashida)
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Quote: "There was no blinding" (by email contact with Prof. Rashida)
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Quote: "There was no blinding" (email with Prof. Rashida)
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No dropouts
Selective reporting (reporting bias)	Low risk	All outcomes reported
Other bias	Low risk	None

Methods	Single‐blind randomised clinical trial Duration and location of the trial: quote: "In this single blind randomized trial, 148 ovarian cycles were studied in 100 clomiphene‐ resistance patients with PCOS who were chosen among 250 PCOS patients attending the Research and Clinical Center for Infertility, Shahid Sadoughi University of Medical Sciences, Yazd, Iran during the years 2007‐2008.“
Participants	Inclusion criteria: women who received 150 mg CC daily for 3 cycles and failed to become pregnant, and were diagnosed with anovulatory PCOS based on Rotterdam 2003. Exclusion criteria: quote: “We excluded patients with liver and kidney dysfunction, cardiovascular disease, diabetics, and those who use metformin or drugs affecting insulin secretion and clomiphene citrate in recent 2 cycles.” Number of centres: 1, research and clinical centre for infertility, Shahid Sadoughi University of Medical Sciences, Yazd Number of women randomised: 100 women, 50 in group A met‐letrozole, 50 in group B met‐CC Number of women analysed: 48 in group A met‐letrozole, 50 in group B met‐CC Number of withdrawals/exclusions/loss to follow‐up and reasons: 2, experienced side effects with metformin before letrozole was started Age (y): group A metformin‐letrozole: 28.5 ± 3.1, group B metformin‐clomiphene citrate: 29.6 ± 3.5 BMI (kg/m²): group A met‐letrozole: 29.0 ± 3.8, group B met‐CC: 29.2 ± 2.9 Duration of infertility (y): group A met‐letrozole: 3.8, group B met‐CC: 3.8 Country: Iran
Interventions	Group A: metformin 1500 mg daily for 6 ‐ 8 weeks, followed by 5 mg letrozole daily orally given for 5 days during cycle days 3 ‐ 7 if pregnancy did not occur Group B: metformin 1500 mg daily for 6 ‐ 8 weeks, followed by 100 mg CC daily orally given for 5 days during cycle days 3 ‐ 7 if pregnancy did not occur
Outcomes	E2 (pg/mL) on day of hCG administration, number of follicles > 18 mm in diameter, endometrial thickness on day of hCG administration (mm), clinical pregnancy rate, miscarriage rate
Notes	SEthical approval: yes, the study was approved by ethical board of Shahid Sagoughi University of Medical Sciences, Yazd. Informed consent: no, at least nothing written about it – authors contacted Source of funding: quote: “the study was fully supported and funded by Shahid Sadoughi University of Medical Sciences, Yazd, Iran” Power calculation: quote: "In this study, 50 cases were needed in each group so as to gain a significant difference of 22% in pregnancy rate at a significant level of 5% and a power of 80%" We contacted Dr Davar by email to get additional information, but we did not get a response.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was done using a random‐numbers table
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	Not stated who was blinded in this single‐blinded trial
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Not stated who was blinded in this single‐blinded trial
Incomplete outcome data (attrition bias)   All outcomes	Low risk	2 participants lost to follow‐up due to side effects experienced with metformin before letrozole was started.
Selective reporting (reporting bias)	Low risk	All outcomes reported stated in the protocol
Other bias	Low risk	None

Methods	Double‐blind randomised study Duration and location of the trial: quote: "During the period of February 2004 through November 2006, 100 patients with PCOS who attended the outpatient infertility clinics at Shiraz University of Medical Sciences participated in the present study.“
Participants	Inclusion criteria: infertility for at least 1 year, diagnosis of PCOS by the Rotterdam criteria 2003, having patent tubes on hysterosalpingogram, and normal semen analysis of the husband Exclusion criteria: participants must not have received any other medication for ovulation induction before enrolment into the study. Number of centres: 1, outpatient infertility clinics at Shiraz University of Medical Sciences Number of women randomised: 100 women, 50 in each group Number of women analysed: 100 women, 50 in each group Number of withdrawals/exclusions/loss to follow‐up and reasons: 0 Age (y): group A letrozole: 23.6 ± 2.9, group B CC: 24.3 ± 3.4 BMI (kg/m²): group A letrozole: 27.5 ± 4.6, group B CC: 27.1 ± 3.6 Duration of infertility (y): group A letrozole: 2.0 ± 1.3, group B CC: 2.3 ± 1.9 Country: Iran
Interventions	Group A: letrozole, 5 mg/day orally given for 5 days during cycle days 3 ‐ 7 Group B: clomiphene citrate, 100 mg/day orally given for 5 days during cycle days 3 ‐ 7
Outcomes	Total number of follicles with diameter ≥ 14 mm, endometrial thickness on the day of hCG injection, pregnancy rate, miscarriage rate, multiple pregnancy rate, live birth rate
Notes	Ethical approval: yes, quote: “The study was approved by the Institutional Ethics Committee of the University.” Informed consent: yes, quote: “An informed written consent was obtained from each patient” Source of funding: not stated Conflicts of interest: quote: "Conflicts of interest: None declared" Power calculation: not stated Authors contacted about randomisation, allocation, and information about OHSS
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not stated how it was done
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Only the pharmacist knew the name of the medication that had been taken by the participants.
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Only the pharmacist knew the name of the medication that had been taken by the participants.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No participants excluded or lost to follow‐up
Selective reporting (reporting bias)	Low risk	All expected outcomes reported
Other bias	Low risk	None

Methods	Randomised double‐blind controlled trial Duration and location of the trial: quote: "A prospective double‐blind randomized controlled trial was conducted at the Teaching University Hospital of Cairo University, Cairo, Egypt, between August 1, 2013, and December 31, 2014.“
Participants	Inclusion criteria: quote: "eligible women were younger than 40 years, had primary or secondary infertility associated with PCOS, and had not ovulated in response to three cycles of 150 mg clomiphene citrate every day for 5 days from day 3 of the menstrual cycle. PCOS diagnoses were based on the revised 2003 Rotterdam consensus criteria. The presence of at least two of the following characteristics was considered diagnostic of PCOS: oligo ovulation or anovulation; hyperandrogenism; and polycystic ovaries detected using vaginal ultrasonography." Exclusion criteria: quote: "exclusion criteria were other factors of infertility, diabetes mellitus, hypertension, liver or kidney malfunction, heart disease, urinary symptoms, persistent hyperprolactinaemia, thyroid dysfunction, gonadotropin induction, and previous ovarian drilling. Male factor infertility was defined as a sperm count of less than 15 × 10^6/mL, a total motility of less than 40%, or normal morphology of less than 4%. Tubal factor infertility was confirmed by hysterosalpingography." Number of women randomised: 100 women were randomised. Number of women analysed: 100 women were analysed, 50 in each group Number of withdrawals/exclusions/loss to follow‐up and reasons: none Number of centres: single centre Age (y): CC 26.6 ± 2.9; letrozole 25.8 ± 3.6 BMI (kg/m²): CC 26.6 ± 2.7; letrozole 26.5 ± 2.8 Duration of infertility (y): CC 3.1 ± 1.4; letrozole 2.7 ± 1.6 Country: Egypt
Interventions	Group A: (control group) received 100 mg clomiphene citrate, given as 2 x 50 mg tablets daily for 5 days from the third day of the menstrual cycle. Group B: 5 mg letrozole, given as 2 x 2.5 mg tablets daily for 5 days from the third day of the menstrual cycle. Participants in both groups also received metformin and pioglitazone, which was taken daily as 1 tablet containing 850 mg metformin and 15 mg pioglitazone, for 10 days starting from the first day of the menstrual cycle.
Outcomes	Primary outcome measure: cumulative ovulation rate (proportion of cycles in which ovulation occurred in the whole follow‐up period).  Other outcome measures: number of follicles ≥ 18 mm in size, endometrial thickness and serum estradiol levels on the day of hCG administration, serum progesterone level on day 21, and rate of clinical pregnancy (at least 1 intrauterine gestational sac detected)
Notes	Ethical approval: approved by the research ethics committee of the teaching University Hospital of Cairo University Informed consent: all participants signed a written informed consent form. Source of funding: Cairo University Power calculation: Previous data indicated that the ovulation rate in group A would be 62%. If the ovulation rate for the letrozole, metformin, and pioglitazone (experimental) group was 87% (previous unpublished data from the study unit), a total of 47 women would have to be recruited to each group to ensure a sufficiently powered study. Assuming an attrition of 10%, the total number of patients to be recruited was 50 per group. Intention‐to‐treat analyses were planned.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Enrolled women were randomly allocated using computer‐generated random number tables (block size four)"
Allocation concealment (selection bias)	Low risk	Quote: "Opaque sealed envelopes containing group allocations were prepared at a separate location every 24 hours. These envelopes were sent to an assigned nurse, who opened them before commencing ovulation induction"
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "Participants, the staff who conducted follow‐up, and data analysts were masked to the allocation to avoid bias"
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "Participants, the staff who conducted follow‐up, and data analysts were masked to the allocation to avoid bias"
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No participants were lost to follow‐up
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported
Other bias	Low risk	None

Methods	Randomised controlled trial Duration of the trial: quote: "The study was conducted during the period from 1st August 2015 to 30th March 2016.“
Participants	Inclusion criteria: quote: "age >20 and <35 years old, patients were diagnosed to have PCOS criteria of diagnosis of PCO. Normal HSG and their partners had normal semen analysis according to WHO criteria (WHO, 2010) and CC‐resistant. If patients fail to respond to 150 mg/day for 5 days for 3 consecutive cycles, they are considered as CC‐resistant." Exclusion criteria: quote: "Age less than 20 yr or more than 35 yr, non‐PCOS, and those Patients with poor ovarian reserve i.e. hyperprolactinaemia, hypo and hyperthyroidism, diabetic patients and Cushing’s syndrome were excluded, non‐classical congenital adrenal hyperplasia, current or previous (within the last 6 months) use of oral contraceptives, glucocorticoids, antiandrogens, antidiabetic or antiobesity drugs, or other hormonal drugs, any subject was affected by either neoplastic, metabolic, hepatic, or cardiovascular disorder or other concurrent medical illness (i.e. diabetes, renal disease, or malabsorptive disorders) were excluded, pelvic diseases, previous pelvic surgery, suspected peritoneal factor infertility, tubal infertility and male factor infertility were excluded with a hysterosalpingogram and with semen analysis, respectively." Number of women randomised: 80 women, 40 within each group Number of women analysed: 80 women, 40 within each group Number of withdrawals/exclusions/loss to follow‐up and reasons: none Number of centres: single‐centre trial Age (y): LOD group 28.8 ± 3.1 vs let group 29.7 ± 3.7 BMI (kg/m²): LOD group 29.1 ± 1.6, letrozole group 29.2 ± 1.7 Duration of infertility (y): no mean ± SD reported Country: Egypt
Interventions	Group A: quote: "In group A, laparoscopy was performed under intravenous general anaesthesia with the patient in a supine position. A 5 mm incision was made in the navel, through which a long sheath punctured into the abdominal cavity, and the inflatable pneumoperitoneum was placed. Another two 5‐mm incisions were made on the right and left lower abdomen and the surgical instruments were inserted into the abdominal cavity. The patient was adjusted into a position with the head high up, the pelvic organs were exposed and a comprehensive exploration of the pelvic organs was made, focusing on the structure and position of the adjacent organs of the bilateral ovaries. Once immobilized, each ovary was cauterized at 4–6 points, using a monopolar electrosurgical needle, according to the size of each ovary. Following cauterization, a bilateral tubal hydrotubation with methylene blue was performed. During the procedure. The pelvis was irrigated using physiological saline. Ringer’s solution plus dexamethasone was added into the abdominal cavity to avoid adhesion. The total duration of the procedure, as well as any intra‐operative or post‐operative complications, was noted." Group B: quote: "In group B, 2.5 mg twice daily LE oral tablets were administered on the 3rd day of menses and then every day for 5 days. Treatment was repeated for up to six cycles if the patient failed to ovulate, the patients were followed‐up for 6 months after the treatment in both groups."
Outcomes	Pregnancy rate, abortion rate, ovulation, regular cycles, ovarian volume, antral follicle count (AFC)
Notes	Ethical approval: This study was approved by Minia University Ethical Committee Informed consent: quote: “Informed consent was obtained from all participating women after the nature and purpose of the study had been explained to them and were fully understood” Source of funding: quote: “We have not received any funding from any corporate body or pharmaceutical company. “ Power calculation: none reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomization was achieved via the use of a randomisation number allocated prior to dosing, once eligibility had been determined, and a randomisation schedule was produced by an interactive voice response system vendor."
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Participants were not blinded
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "Once the patients had been allocated to one of the two groups, the treatment was revealed to the investigator; however, the doctor responsible for performing the transvaginal ultrasound follow up assessment was blinded to the treatment groups."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	All women randomised were also analysed.
Selective reporting (reporting bias)	Unclear risk	We found no reporting of outcomes in a study protocol or Methods section
Other bias	Low risk	None

Methods	Randomised double‐blind placebo‐controlled trial Duration and location of the trial: quote: "This trial was conducted in a university teaching hospital between 2007 and 2009.“
Participants	Inclusion criteria: women with PCOS and clomiphene resistance who were being treated with ovulation induction. Additionally, women had to have a normal hormone profile and a male partner with normal semen parameters by WHO criteria. Normal hormone profile was defined as a FSH level of < 12 IU/L, serum prolactin level of < 25 ng/mL, and a thyroid‐stimulating hormone (TSH) value between 0.3 and 4.5 µIU/mL. Exclusion criteria: women with other endocrine disorders such as Cushing syndrome, and congenital adrenal hyperplasia Number of centres: 1, Reproductive Medicine Unit, Christian Medical College, Vellore, Tamil Nadu, India Number of women randomised: 18 in each group Number of women analysed: 17 in each group Number of withdrawals/exclusions/loss to follow‐up and reasons: 2 lost to follow‐up before treatment started Age (y): group A letrozole: 25.6 ± 3.6, group B placebo: 25.7 ± 3.7 BMI (kg/m²): group A letrozole: 26.1 ± 3.7, group B placebo: 24.7 ± 4.2 Duration of infertility (y): group A letrozole: 5.2 ± 3.2, group B placebo: 3.6 ± 2.2 Country: India
Interventions	Group A: letrozole, orally given 2.5 mg/day for 5 days from cycle days 2 ‐ 6 Group B: placebo, also given for 5 days from cycle days 2 ‐ 6
Outcomes	Primary Outcome: ovulation rate Secondary Outcomes: live birth rate, OHSS rate, pregnancy rate, miscarriage rate, multiple pregnancy rate, endometrial thickness (mm), day 21 serum progesterone (nmol/L), number of participants with mature follicle (%)
Notes	Ethical approval: yes, the protocol of the study was approved by the institutional review board Informed consent: yes, written informed consent was obtained from each participant Source of funding: not stated Conflicts of interest: quote: "The Authors have nothing to disclose" Power calculation: quote: "Our literature pointed to a 75% ovulation rate when 2.5 mg of letrozole was used in women with PCOS who had clomiphene resistance. We hypothesized an ovulation rate of 60& with letrozole and 10% with placebo. On this basis, a sample size of 17 women in each arm (80% and alpha .05 for a two‐sided test) was calculated." Contacted authors about OHSS rate and how randomisation and allocation concealment were done in detail. All information provided
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomly distributed using a computer‐generated randomisation sequence in blocks of 6, into 2 groups.
Allocation concealment (selection bias)	Low risk	Allocation concealment was done by using consecutively‐numbered sealed opaque envelopes containing the treatment packets.
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	The randomisation code was maintained by the pharmacy department, which revealed the group assignments at the end of the trial.
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	The code was revealed after the statistical analysis had been performed.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	1 women in each group was lost to follow‐up, after randomisation and before treatment started.
Selective reporting (reporting bias)	Low risk	All outcomes reported
Other bias	Low risk	None

Methods	Randomised double‐blind multicentre trial Duration of the trial: quote: "Enrollment began in February 2009 and was completed in January 2012.“
Participants	Inclusion criteria: women with PCOS defined by the Rotterdam criteria and at least 1 patent fallopian tube and normal uterine cavity, and a male partner with sperm concentration of > 14 million/mL Exclusion criteria: quote: "We will exclude subjects with medical conditions that represent contraindications to CC,letrozole and/or pregnancy or who are unable to comply with the study procedures." Number of centres: multicentre trial Number of women randomised: 750, 374 in the letrozole group and 376 in the CC group Number of women analysed: 750, 374 in the letrozole group and 376 in the CC group Number of withdrawals/exclusions/loss to follow‐up and reasons: 0 Age (y): group A letrozole: 29 ± 5, group B CC: 28 ± 4 BMI (kg/m²): group A letrozole: 35 ± 10, group B CC: 35 ± 9 Duration of infertility (y): not reported Country: USA
Interventions	Group A: letrozole, orally given 2.5 mg/day for 5 days during cycle days 3 ‐ 7 Group B: clomiphene citrate, orally given 100 mg/day for 5 days during cycle days 3 ‐ 7
Outcomes	Live birth, ovulation rate, clinical pregnancy rate, miscarriage rate, multiple pregnancy rate
Notes	Ethical approval: quote:"The institutional review board at each centre approved the protocol, and all participants (women and their male partners) gave written informed consent." Informed consent: quote: "The institutional review board at each centre approved the protocol, and all participants (women and their male partners) gave written informed consent." Source of funding: quote: "The study is funded through a cooperative agreement by the Eunice Kennedy ShriverNational Institutes of Child Health and Human Development (NICHD)" Power calculation: a sample size of 300 subjects in each arm of the randomisation yields 81% statistical power to prospectively demonstrate a 0.10 absolute difference in live birth proportions between treatment arms (0.20 for CC and 0.30 for letrozole) using the Pearson’s Chi² test with a 2‐sided significance level of 0.05
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Almac statisticians will generate the randomisation scheme for the study."
Allocation concealment (selection bias)	Low risk	Third‐party allocation
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "In order to maintain the double‐blind, CC and letrozole will be over encapsulated and packaged in identically appearing numbered study kits (using AlmacClinical Services, Durham NC) which will then be directly shipped to each clinical site."
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "In order to maintain the double‐blind, CC and letrozole will be over encapsulated and packaged in identically appearing numbered study kits (using AlmacClinical Services, Durham NC) which will then be directly shipped to each clinical site. The randomisation scheme (including block size) will be disclosed to the DCC data manager, but not to any RMN investigators or staff, including the Protocol Lead Investigator."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No dropouts were reported
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported
Other bias	Low risk	None

PERMALINK

Aromatase inhibitors (letrozole) for subfertile women with polycystic ovary syndrome

Sebastian Franik

Stephanie M Eltrop

Jan AM Kremer

Ludwig Kiesel

Cindy Farquhar

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Effectiveness:

Adverse effects:

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Overall quality of the body of evidence: 'Summary of findings' table

Results

Description of studies

Results of the search

1.

Included studies

Study design and setting

Participants

Interventions

Outcomes

Excluded studies

Risk of bias in included studies

2.

3.

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Summary of findings for the main comparison. Letrozole with or without adjuncts compared to clomiphene citrate (CC) with or without adjuncts for subfertile women with polycystic ovary syndrome.

Summary of findings 2. Letrozole compared to laparoscopic ovarian drilling for subfertile women with polycystic ovary syndrome.

1. Letrozole compared to placebo

1.1. Analysis.

1.2. Analysis.

1.3. Analysis.

1.4. Analysis.

2. Letrozole compared to clomiphene citrate (CC) with or without adjuncts, followed by intercourse

Primary outcomes

2.1 Live birth

4.

2.1. Analysis.

Methods	A partly‐randomised controlled clinical trial. Duration and location of the trial: not stated
Participants	Inclusion criteria: infertile women with PCOS, diagnosis based on the 2003 Rotterdam criteria Exclusion criteria: women with ovarian or adnexal surgery, hypothyroidism, hyperprolactinaemia, bilateral tubal occlusion diagnosed with hysterosalpingography and unexplained infertility, and those with follicles greater than 10 mm Number of centres: 1, infertility polyclinic of Atatürk University Medical Faculty Erzurum Number of women randomised: 31 in group A, 33 in group B Number of women analysed: 31 in group A, 33 in group B Number of withdrawals/exclusions/loss to follow‐up and reasons: 0 Age (y): group A letrozole: 25.6 ± 4.5, group B CC: 27.8 ± 6.2 BMI (kg/m²): group A letrozole: 24.7 ± 3.6, group B CC: 24.9 ± 4.8 Duration of infertility (y): group A letrozole: 3.4 ± 3.0, group B CC: 4.4 ± 3.6 Country: Turkey
Interventions	Group A: letrozole, orally given 2.5 mg/day for 5 days during cycle days 3 ‐ 7 Group B: clomiphene citrate, orally given 100 mg/day for 5 days during cycle days 3 ‐ 7
Outcomes	Primary Outcomes: ovulation rate and pregnancy rate Secondary Outcomes: ovarian hyperstimulation syndrome rate, miscarriage rate, multiple pregnancy rate, number of follicles on day of hCG (≥ 17 mm), E2 (pg/mL) on hCG day, endometrial thickness (mm), other side effects
Notes	Ethical approval: yes, quote: “Ethical approval was obtained from the institutional review board of Atatürk University Medical Faculty in order to conduct this study.” Informed consent: quote: "Instead of written consent verbal approval was obtained from the patients prior to study begin and treatment" ‐ correspondence with Dr. Hakan Nazik Source of funding: quote: "This study was done by researchers without any funding" Power calculation: not stated All questions were answered by Dr. Hakan Nazik
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: “The patients were randomly allocated using a computer random list into first and second groups”
Allocation concealment (selection bias)	Unclear risk	Quote: “The patients were randomly allocated using a computer random list into first and second groups”
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Quote: "There was no blinding in our study" (email with Dr. Hakan Nazik)
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Quote: "There was no blinding in our study" (email with Dr. Hakan Nazik)
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No dropouts reported
Selective reporting (reporting bias)	Low risk	All expected outcomes reported
Other bias	High risk	Participants in Group 2 letrozole were significantly younger and had a significantly shorter duration of infertility.

Methods	Comparative randomised phase III open‐labelled trial Duration and location of the trial: quote: "A comparative, prospective, phase III, open labelled trial study was conducted in the Eden Hospital, Medical College Kolkata between January 2008 and December 2009.“
Participants	Inclusion criteria: infertile women aged 20 ‐ 35 with PCOS diagnosis based on the Rotterdam criteria 2003 Exclusion criteria: women with hyperprolactinaemia, thyroid disorder, male‐factor infertility, known or suspicious tubal‐factor infertility (endometriosis and pelvic inflammatory disease). Also women with a history of liver and kidney failure, cardiovascular diseases, diabetes, or women who consumed metformin or drugs affecting insulin secretion or CC in the previous 2 months Number of centres: 1, Eden Hopsital, Mecial College Kolkata Number of women randomised: 147; group A letrozole: 69, group B CC: 78 Number of women analysed: group A letrozole: 69, group B CC: 78 Number of withdrawals/exclusions/loss to follow‐up and reasons: 0 Age (y): group A letrozole: 28 (19 ‐ 35), group B CC: 29 (20 ‐ 35) BMI (kg/m²): group A letrozole: 28.8 (23.2 ‐ 34.6), group B CC: 28.5 (24.2 ‐ 33.6) Duration of infertility (y): group A letrozole: 2.2, group B CC: 2.4 (SD or range not given) Country: India
Interventions	Group A: letrozole, 2.5 mg/day given orally for 5 days from cycle day 3 ‐ 7 Group B: clomiphene citrate, 100 mg/day given orally for 5 days from cycle day 3 ‐ 7
Outcomes	Primary Outcomes: ovulation rate, average follicular diameter on day 16, number of mature follicles produced by cycle, mean estradiol level on the day of hCG administration, mean endometrial thickness, pregnancy rate Secondary Outcomes: miscarriage rate, live birth rate
Notes	Ethical approval: yes, the study protocol was approved by the ethical committee of Medical College Kolkata Informed consent: yes, participants were counselled and informed consent was obtained before recruitment Source of funding: quote: “Conflict of interest: the authors hereby declare that they have not received any financial support for this study and there is no conflict of interest.” Power calculation: not stated We contacted Dr. Ray by email about randomisation, allocation, blinding, MPR and OHSS, but he did not respond.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Unclear how randomisation was done
Allocation concealment (selection bias)	Unclear risk	Unclear how allocation was done
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	Not reported if anyone was blinded
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Not reported if anyone was blinded
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	No participants stated as lost, but 147 participants is an odd number to start with, and so are the groups of 69 and 78 respectively; authors contacted for protocol
Selective reporting (reporting bias)	Unclear risk	All expected outcomes were reported, but contacted authors for protocol
Other bias	Low risk	None

Methods	Randomised clinical trial Duration and location of the trial: quote: "This prospective randomized controlled trial was performed at a tertiary care hospital from January 2005 to January 2010.“
Participants	Inclusion criteria: women aged 20 ‐ 35 years having infertility for > 1 year, BMI < 28, and with anovulatory PCOS based on the Rotterdam 2003 criteria Exclusion criteria: quote: "in all patients, a comprehensive infertility work‐up was done. This included a tubal patency test, pelvic ultrasonography, husband semen analysis, and serum hormone measurements (FSH, LH, prolactin, estradiol, progesterone, and testosterone) on the 2nd to 5th day of the cycle. Patients having abnormality in any of these tests, which may be responsible for reproductive failure, were excluded from the study." Number of centres: 1, a tertiary care hospital in India Number of women randomised: 212 women; group A let: 104, group B CC: 108 Number of women analysed: letrozole group: 98, clomiphene group: 106 Number of withdrawals/exclusions/loss to follow‐up and reasons: 8 lost to follow‐up Age (y): group A letrozole: 26.1 ± 1.8, group B CC: 26.5 ± 1.3 BMI (kg/m²): group A letrozole: 25.8 ± 2.1, group B CC: 25.4 ± 1.6 Duration of infertility (y): group A letrozole: 6.4 ± 3.8, group B CC: 5.8 ± 3.1 Country: India
Interventions	Group A: letrozole, orally given in doses of 2.5 mg/day and 5 mg/day for 5 days during cycle days 3 ‐ 7 Group B: clomiphene citrate, orally given in doses of 50 mg/day and 100 mg/day for 5 days during cycle days 3 ‐ 7 Treatment was continued for 3 months.
Outcomes	The mean number of follicles, endometrial thickness, ovulatory cycle rate, conception rate, pregnancy outcome, miscarriage rate, multiple pregnancies and OHSS rate
Notes	Ethical approval: yes, the necessary ethical approval was taken from Institutional Review Board to conduct this study. Informed consent: yes, the participants were counselled, and informed consent was taken before randomisation. Source of funding: quote: “Source of support: Nil, Conflict of interest: None declared.” Power calculation: quote: "On basis of previous studies, to achieve a statistically valid comparison of pregnancy rates in the two groups, with a type I error of 0.05 and a power of 80%, a sample size of at least 40 women in each arm was required." We contacted Dr. Roy by email to get additional information, but he did not respond.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Online software was used to generate a random number table (www.randomization.com).
Allocation concealment (selection bias)	Unclear risk	Quote: "Randomisation codes (A, B) were packed into sealed opaque envelopes by an individual not involved in enrolment, treatment and follow‐up of subjects to ensure concealment of allocation. One resident had the responsibility for dispensing the trial drugs to the patient based on the unique randomisation code. At the end of allocation, the resident provided us with a randomisation list."
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	Not stated
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Not stated
Incomplete outcome data (attrition bias)   All outcomes	Low risk	8 losses to follow‐up of 112 participants
Selective reporting (reporting bias)	Low risk	All expected outcomes reported
Other bias	Low risk	None

Methods	Randomised controlled double‐blind trial Duration and location of the trial: not stated
Participants	Inclusion criteria: diagnosis of PCOS based on Rotterdam criteria provided that anovulation is 1 of the 2 required criteria Exclusion criteria: exclusion criteria were BMI > 35, presence of other causes of infertility, > 5 years infertility duration and known poor response to either drugs in previous cycles. Cases found to have baseline ovarian cysts or endometrial pathology were also excluded. Number of centres: 1, an infertility unit of a university hospital Number of women randomised: 124; group A letrozole: 62, group B CC: 62 Number of women analysed: group A letrozole: 59, group B CC: 57 Number of withdrawals/exclusions/loss to follow‐up and reasons: 3 in the letrozole and 5 in the CC group were reported as lost to follow‐up, but no further explanation given Age (y): group A letrozole: 25.0 ± 3.1, group B CC: 25.0 ± 3.6 BMI (kg/m²): group A letrozole: 27.7 ± 3.5, group B CC: 29.2 ± 3.5 Duration of infertility (y): Not reported Country: India
Interventions	Group A: letrozole, 5 mg/day orally given for 5 days, cycle days not given Group B: clomiphene citrate, 100 mg/day orally given for 5 days, cycle days not given
Outcomes	Pregnancy rate, ovulation rate, endometrial thickness (mm), mid‐luteal progesterone level (ng/mL), number of follicles ≥ 12 mm
Notes	Ethical approval: yes, institutional review board (IRB) approval was obtained for the study. Informed consent: yes, informed consent was taken from all included cases Source of funding: not reported, also no conflicts of interest given. Power calculation: not stated We contacted Dr. Sheik‐el‐Arab Elsedeek by email about allocation concealment, blinding of outcome assessors, information on live birth, miscarriage rate, OHSS, multiple pregnancies and funding/COI, but he did not respond.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were randomised using computer generated tables to undergo one cycle of CC or let induction."
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	Not stated if personnel were blinded and how the participants were blinded
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Quote: "Both patients and sonographers were blinded to this allocation." ‐ Unclear if the other outcome assessors were blinded
Incomplete outcome data (attrition bias)   All outcomes	Low risk	8 participants lost to follow‐up, reasons unknown
Selective reporting (reporting bias)	Unclear risk	Only pregnancy was reported; we contacted the authors to get study protocol.
Other bias	Low risk	None

Methods	Single‐blinded randomised clinical trial Duration and location of the trial: quote: "In this single‐blind randomized clinical trial, 120 ovarian cycles were studied in 60 clomiphene‐resistant patients with PCOS, who were chosen among 115 PCOS patients attending the infertility clinic of Vali‐e‐Asr Hospital (Tehran, Iran) during the years 2003–2004.“
Participants	Inclusion criteria: women with PCOS who had failed to become pregnant after 3 courses of 150 mg clomiphene citrate (considered as clomiphene‐resistant), whereas the values of hormonal tests were normal. Tests: thyroid function, prolactin level, hysterosalpingography and husband’s sperm analysis Exclusion criteria: women with a history of liver and kidney failure, cardiovascular disease, diabetes (based on criteria set by the American Diabetic Association) or women who consumed metformin or drugs affecting insulin secretion or clomiphene citrate in the previous 2 months Number of centres: 1, infertility clinic of Vali‐e‐Asr Hospital, Tehran Number of women randomised: 60; group A met‐letrozole: 30, group B met‐CC: 30 Number of women analysed: group A met‐letrozole: 29, group B met‐CC: 30 Number of withdrawals/exclusions/loss to follow‐up and reasons: 1 because she got pregnant after met treatment before letrozole was started Age (y): group A met‐letrozole: 28.2 ± 3.1 , group B met‐CC: 29.6 ± 3.5 BMI (kg/m²): group A met‐letrozole: 30.0 ± 4.8, group B met‐CC: 30.2 ± 3.9 Duration of infertility (y): group A met‐letrozole: 3.8, group B met‐CC: 3.8 Country: Iran
Interventions	Group A: metformin 500 mg x 3/d for 6 ‐ 8 weeks. If pregnancy did not occur, 2.5 mg letrozole from cycle days 3 ‐ 7 was given orally. Group B: metformin 500 mg x 3/d for 6 ‐ 8 weeks. If pregnancy did not occur, 100 mg CC from cycle days 3 ‐ 7 was given orally. Treatment was continued for 2 cycles.
Outcomes	Endometrial thickness on day of hCG administration (cm), N of follicles > 18 mm in diameter, Mean total estradiol level on day of hCG administration (pM/L), mean estradiol level by mature follicle (pM/l), regular menses after metformin, adverse effects of metformin, live birth rate, pregnancy rate, miscarriage rate
Notes	Ethical approval: yes, consent from the deputy of research and the medical ethics committee of Tehran University of Medical Sciences Informed consent: not obtained because quote: “it was the routine treatment protocol and it was just put in the frame of a structured study” (email with Dr Farnaz Sohrabvand) Source of funding: quote: "No funding was necessary" (email with Dr Farnaz Sohrabvand) Power calculation: not stated Authors were contacted about live birth, multiple pregnancies, OHSS per woman randomised, informed consent, funding. No data available on live birth, multiple pregnancies and OHSS. Information retrieved about informed consent and funding (email with Dr Farnaz Sohrabvand)
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "A series of blind envelopes numbered from 1 to 60 had been prepared. Each patient was invited to pull out an envelope and was placed by the clinic secretary in either the metformin‐letrozole group (number 1‐30) or in the metformin‐CC group (31‐60)."
Allocation concealment (selection bias)	Unclear risk	Quote: "A series of blind envelopes numbered from 1 to 60 had been prepared. Each patient was invited to pull out an envelope and was placed by the clinic secretary in either the metformin‐letrozole group (number 1‐30) or in the metformin‐CC group (31‐60)."
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Participants were not blinded.
Blinding of outcome assessment (detection bias)   All outcomes	High risk	It is not plausible that outcome assessors were blinded if participants were not
Incomplete outcome data (attrition bias)   All outcomes	Low risk	One participant was excluded due to pregnancy after start of metformin treatment
Selective reporting (reporting bias)	Low risk	All outcomes were reported
Other bias	Low risk	None

Methods	Randomised clinical trial Duration of the trial: quote: "The trial was started on October 2009. Owing to the expiration of the study drug (berberine and matching placebo), the data safety and monitoring board decided to stop enrollment in November 2013“
Participants	Inclusion criteria: quote: "Chinese women with PCOS attempting to get pregnant were eligible if they fulfilled the following criteria: 1) age 20–40 years; 2) diagnosis of PCOS according to two of the three Rotterdam 2003 criteria, including oligo‐ovulation or anovulation, clinical and/or biochemical signs of hyperandrogenism, and/or polycystic ovaries; 3) at least one open fallopian tube and normal uterine cavity documented by hysterosalpingography, sonohysterography, or diagnostic laparoscopy within the past 3 years; 4) a male partner with sperm concentration of 15 million/mL and motility of 40% in at least one ejaculate; and 5) at least 1 year of infertility." Exclusion criteria: quote: "Subjects were excluded if they used hormonal drugs or other medications, including Chinese herbal prescriptions, in the past 3 months; had known severe organ dysfunction or mental illness; were pregnant, post‐miscarriage, postpartum, or breastfeeding within the past 6 weeks; or had congenital adrenal hyperplasia, clinically suspected Cushing syndrome, or an androgen‐secreting neoplasm." Number of women randomised: 644 women Number of women analysed: 644 women were analysed, 215 in group A, 214 in group B, 215 in group C. Number of withdrawals/exclusions/loss to follow‐up and reasons: 16/215 (7.4%) in the letrozole group, 25/214 (11.7%) in the berberine group, and 15/215 (7.0%) in the combination group (P = 0.16). Reasons for withdrawal were similar among the 3 groups (P = 0.16 for the 3 groups; P = 0.19 for lost to follow‐up; P = 0.88 for dropout; P = 1.0 for protocol violations; and P = 0.33 for adverse events). Number of centres: multicentre trial, 19 hospitals Age (y): group A 27.8 ± 3.6; group B 27.8 ± 3.7; group C 27.8 ± 3.6 BMI (kg/m²): group A 24.8 ± 4.5; group B 24.5 ± 4.1; group C 25.1 ± 5.0 Duration of infertility (months): group A 32.7 ± 24.0; group B 28.5 ± 21.6; group C 29.8 ± 21.3 Country: China
Interventions	Group A: 2.5 mg (1 tablet) of letrozole on days 3 – 7 of the first 3 treatment cycles. This dose was increased to 5 mg letrozole (2 tablets) or 2 tablets of letrozole placebo on days 3 – 7 of the last 3 treatment cycles if not pregnant Group B: berberine was administered orally at a daily dose of 1.5 g for 6 months. Group C: letrozole and berberine were administered in the same doses as reported above.
Outcomes	Cumulative live births, ovulation rate, conception rate, clinical pregnancy rate, multiple pregnancy rate, abortion rate, pregnancy complications, adverse events from study medications
Notes	Ethical approval: the Institutional Review Boards at participating hospitals approved the protocol. Informed consent: every participant gave written informed consent. Source of funding: supported by National Public Welfare Projects for Chinese Medicine (200807021) of China, National Key Discipline of Chinese Medicine in Gynecolog, 2009–14, Heilongjiang Province Foundation for Outstanding Youths (JC200804), Intervention for Polycystic Ovary Syndrome Based on Traditional Chinese Medicine Theory–‘‘TianGui Shi Xu’’ (2011TD006), and National Clinical Research Base in Chinese Medicine, 2009–14, at First Affiliated Hospital, Heilongjiang University of Chinese Medicine. The funding sources had no involvement in the study design, the collection, analysis, and interpretation of data, the writing of the report, or in the decision to submit the article for publication. Power calculation: the sample size calculation was based on anticipated live birth rate.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The randomisation was performed through a web‐based computer program (http://210.76.97.192:8080/cjbyj) operated by an independent data coordinating centre, the Institute of Basic Clinical Medicine of the China Academy of Chinese Medical Sciences. The randomisation was stratified by the participating sites."
Allocation concealment (selection bias)	Low risk	Randomisation was operated by an independent data co‐ordinating centre
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Participants, investigators, physicians taking care of the participants, laboratory technicians, and data analysers were blinded to the assignments.
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Participants, investigators, physicians taking care of the participants, laboratory technicians, and data analysers were blinded to the assignments.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	7% – 12% were lost to follow‐up, had protocol violations or adverse events. ITT analysis was performed.
Selective reporting (reporting bias)	Low risk	None, study protocol was published prior to participant enrolment in the study
Other bias	Low risk	None

Study	Reason for exclusion
Akbari 2012	No PCOS
Al‐Hussaini 2014	Only short conference abstract; we did not find a published study article. We were unable to retrieve any information from study authors.
Al‐Shaikh 2017	Not a RCT
Angel 2014	RCT, but not about women with PCOS
Anwary 2012	Not a RCT
Azargoon 2012	Not a RCT
Azmoodeh 2015	No PCOS
Badawy 2009c	Not a RCT
Baruah 2009	Quasi‐randomised trial (quote: "Based on attendance order, patients with odd numbers were given letrozole and those with even numbers were given CC")
El Bigawy 2008	Quasi‐randomised trial
Foroozanfard 2013	Not randomised for clomiphene or letrozole
Khanna 2013	Not a RCT, no PCOS
Li 2016	Not randomised for letrozole
Mittal 2004	Not a RCT
Nahid 2012	Suspected quasi‐randomisation based on attendance order
NCT00610077	No published data found, last updated on clinicaltrials.gov in 2008. No response from study authors.
NCT01315912	Not a RCT
NCT01431352	For the 2018 update, there was no update on clinicaltrials.gov. No study data available, and no response from authors.
NCT01577017	No study results found, unable to obtain more information from study authors
NCT01679574	Study should be finished. We contacted authors by email found through Google because no contact information was written in the study protocol: abd_ellah98@yahoo.com. For the 2018 update, there was no update on clinicaltrials.gov and still no response from the author.
NCT01793038	Unable to obtain information from study authors. No study data found, no update on clinicaltrials.gov.
Ozdemir 2013	RCT, but not about women with PCOS
Pakrashi 2014	Not randomised for letrozole
Palihawadana 2015	Women with PCOS were excluded from the study
Pourali 2017	Women with PCOS were excluded from the study
Sarvi 2010	Not randomised for letrozole
Sharma 2010	Only conference abstract available, full article could not be found. Contact address of authors unknown
Xi 2015	Not a RCT
Yang 2008	Not a RCT (Quote: "the allocation depended on the patients' choice" ‐ translated by Prof Taixiang Wu)
Yun 2015	Not a RCT

Methods	Randomised clinical study
Participants	15 infertile women with polycystic ovarian syndrome
Interventions	Women were randomised into 3 treatment groups: In group 1, continuous metformin was used at the dose of 850 mg/tid/day for 6 months; afterwards, daily 2.5 mg letrozole between 3 and 7 days of the menstrual cycle was added to the metformin therapy. Group 2 participants received only daily 2.5 mg letrozole between days 3 and 7 of the menstrual cycle. Group 3 participants received daily 100 mg clomiphene citrate only between days 3 and 7 of the menstrual cycle.
Outcomes	Unclear
Notes	The article was written in Turkish, and we were not able to have it translated properly.

Methods	Randomised clinical trial
Participants	100 infertile women with PCOS referred to Asali Hospital and private clinic in 2008
Interventions	The women were randomised into 2 groups of 50 that were treated with 5 mg letrozole or 100 mg clomiphene citrate from day 3 to 7 of the menstrual cycle.
Outcomes	Outcomes: Pregnancy rate
Notes	The article was written in Persian, but we were not able to have it translated properly.

Methods	Randomised controlled trial
Participants	110 infertile women diagnosed as polycystic ovary syndrome (PCOS) aged 20 ‐ 35 distributed randomly
Interventions	55 women will receive letrozole 2.5 mg twice daily orally from the 2nd to the 6th day of the cycle for 3 successive cycles. 55 women will receive clomiphene citrate 50 mg twice daily orally from the 2nd to the 6th day of the cycle for 3 successive cycles.
Outcomes	Rate of ovulation assessed by number of mature follicles produced per cycle. Serum progesterone level on day 21 ( assessed up to 24 weeks). Mean endometrial thickness ( assessed up to 24 weeks). Chemical pregnancy ( assessed up to 24 weeks). Ongoing pregnancy ( assessed up to 24 weeks).
Notes	We found no study data, although trial completed since 2016 on clinicaltrials.gov