Gonadotrophins for ovulation induction in women with polycystic ovary syndrome
Abstract
Background
Ovulation induction with follicle stimulating hormone (FSH) is a second‐line treatment in women with polycystic ovary syndrome (PCOS) who do not ovulate or conceive on clomiphene citrate.
Objectives
To compare the effectiveness and safety of gonadotrophins as a second‐line treatment for ovulation induction in women with clomiphene citrate‐resistant polycystic ovary syndrome (PCOS), and women who do not ovulate or conceive after clomiphene citrate.
Search methods
In January 2018, we searched the Cochrane Gynaecology and Fertility Group Specialised Register of Controlled Trials, CENTRAL, MEDLINE, Embase, PsycINFO, CINAHL, the World Health Organisation clinical trials register, Clinicaltrials.gov, LILACs, and PubMed databases, and Google Scholar. We checked references of in all obtained studies. We had no language restrictions.
Selection criteria
All randomised controlled trials reporting data on clinical outcomes in women with PCOS who did not ovulate or conceive on clomiphene citrate, and undergoing ovulation induction with urinary‐derived gonadotrophins, including urofollitropin (uFSH) in purified FSH (FSH‐P) or highly purified FSH (FSH‐HP) form, human menopausal gonadotropin (HMG) and highly purified human menopausal gonadotrophin (HP‐HMG), or recombinant FSH (rFSH), or continuing clomiphene citrate. We included trials reporting on ovulation induction followed by intercourse or intrauterine insemination. We excluded studies that described co‐treatment with clomiphene citrate, metformin, luteinizing hormone, or letrozole.
Data collection and analysis
Three review authors (NW, EK, and MvW) independently selected studies for inclusion, assessed risk of bias, and extracted study data. Primary outcomes were live birth rate per woman and multiple pregnancy per woman. Secondary outcomes were clinical pregnancy, miscarriage, incidence of ovarian hyperstimulation syndrome (OHSS) per woman, total gonadotrophin dose, and total duration of stimulation per woman. We combined data using a fixed‐effect model to calculate the risk ratio (RR). We summarised the overall quality of evidence for the main outcomes using GRADE criteria.
Main results
The review included 15 trials with 2387 women. Ten trials compared rFSH with urinary‐derived gonadotrophins (three compared rFSH with human menopausal gonadotrophin, and seven compared rFSH with FSH‐HP), four trials compared FSH‐P with HMG. We found no trials that compared FSH‐HP with FSH‐P. One trial compared FSH with continued clomiphene citrate.
Recombinant FSH (rFSH) versus urinary‐derived gonadotrophins
There may be little or no difference in the birth rate between rFSH and urinary‐derived gonadotrophins (RR 1.21, 95% confidence interval (CI) 0.83 to 1.78; five trials, N = 505; I² = 9%; low‐quality evidence). This suggests that for the observed average live birth per woman who used urinary‐derived FSH of 16%, the chance of live birth with rFSH is between 13% and 28%. There may also be little or no difference between groups in incidence of multiple pregnancy (RR 0.86, 95% CI 0.46 to 1.61; eight trials, N = 1368; I² = 0%; low‐quality evidence), clinical pregnancy rate (RR 1.05, 95% CI 0.88 to 1.27; eight trials, N = 1330; I² = 0; low‐quality evidence), or miscarriage rate (RR 1.20, 95% CI 0.71 to 2.04; seven trials, N = 970; I² = 0; low‐quality evidence). We are uncertain whether rFSH reduces the incidence of OHSS (RR 1.48, 95% CI 0.82 to 2.65, ten trials, n=1565, I² = 0%, very low‐quality evidence).
Human menopausal gonadotrophin (HMG) or HP‐HMG versus uFSH
When compared to uFSH, we are uncertain whether HMG or HP‐HMG improves live birth rate (RR 1.28, 95% CI 0.65 to 2.52; three trials, N = 138; I² = 0%; very low quality evidence), or reduces multiple pregnancy rate (RR 2.13, 95% CI 0.51 to 8.91; four trials, N = 161; I² = 0%; very low quality evidence). We are also uncertain whether HMG or HP‐HMG improves clinical pregnancy rate (RR 1.31, 95% CI 0.66 to 2.59; three trials, N = 102; I² = 0; very low quality evidence), reduces miscarriage rate (RR 0.33, 95% CI 0.06 to 1.97; two trials, N = 98; I² = 0%; very low quality evidence), or reduces the incidence of OHSS (RR 7.07, 95% CI 0.42 to 117.81; two trials, N = 53; very low quality evidence) when compared to uFSH.
Gonadotrophins versus continued clomiphene citrate
Gonadotrophins resulted in more live births than continued clomiphene citrate (RR 1.24, 95% CI 1.05 to 1.46; one trial, N = 661; I² = 0%; moderate‐quality evidence). This suggests that for a woman with a live birth rate of 41% with continued clomiphene citrate, the live birth rate with FSH was between 43% and 60%. There is probably little or no difference in the incidence of multiple pregnancy between treatments (RR 0.89, 95% CI 0.33 to 2.44; one trial, N = 661; I² = 0%; moderate‐quality evidence). Gonadotrophins resulted in more clinical pregnancies than continued clomiphene citrate (RR 1.31, 95% CI 1.13 to 1.52; one trial, N = 661; I² = 0%; moderate‐quality evidence), and more miscarriages (RR 2.23, 95% CI 1.11 to 4.47; one trial, N = 661; I² = 0%; moderate‐quality evidence). None of the women developed OHSS.
Authors' conclusions
There may be little or no difference in live birth, incidence of multiple pregnancy, clinical pregnancy rate, or miscarriage rate between urinary‐derived gonadotrophins and recombinant follicle stimulating hormone in women with polycystic ovary syndrome. For human menopausal gonadotropin or highly purified human menopausal gonadotrophin versus urinary follicle stimulating hormone we are uncertain whether one or the other improves or lowers live birth, incidence of multiple pregnancy, clinical pregnancy rate, or miscarriage rate. We are uncertain whether any of the interventions reduce the incidence of ovarian hyperstimulation syndrome. We suggest weighing costs and convenience in the decision to use one or the other gonadotrophin. In women with clomiphene citrate failure, gonadotrophins resulted in more live births than continued clomiphene citrate without increasing multiple pregnancies.
PICOs
Plain language summary
Gonadotrophins to induce ovulation in women with polycystic ovary syndrome (PCOS)
Review question
To compare the effectiveness and safety of gonadotrophins, hormones that regulate the reproductive system, as a second‐line treatment to stimulate ovulation in women with PCOS who do not ovulate or conceive on clomiphene citrate.
Background
Infertility due to ovulation disorders is the most common reason for women to seek counselling or treatment. These women are treated by stimulating ovulation with medication, so‐called 'ovulation induction'. This is usually done with tablets containing clomiphene citrate, as the first line of treatment. If women do not react to this medication, the most common second‐line treatment in these women is ovulation induction with gonadotrophins, which are injectable drugs. Various types of gonadotrophin have been developed: urinary‐derived products, available in purified (FSH‐P), and highly purified (FSH‐HP) form, and human menopausal gonadotrophin, also available in highly purified form (HP‐HMG). Finally, recombinant FSH (rFSH) was developed artificially to obtain even higher purity.
Women who do react, but do not conceive within six ovulatory clomiphene citrate cycles, may continue with clomiphene citrate or switch to gonadotrophins.
Study characteristics
The review includes 15 trials, covering 2387 women. Ten trials compared urinary‐derived gonadotrophins with rFSH. Of these, three trials compared rFSH with human menopausal gonadotrophin, and seven trials compared rFSH with FSH‐HP. Four trials compared FSH‐P with human menopausal gonadotrophin. One trial compared gonadotrophins with continued clomiphene citrate. We found no trials that compared rFSH with FSH‐P, or FSH‐HP with FSH‐P. The evidence is current to January 2018.
Key results
There may be little or no difference in live birth, multiple pregnancy, clinical pregnancy, or miscarriage rate between urinary‐derived gonadotrophins and recombinant FSH. We are uncertain whether human menopausal gonadotrophin or urinary follicle stimulating hormone improves pregnancy outcomes in women with PCOS. We are uncertain whether the interventions decrease the incidence of ovarian hyperstimulation syndrome.
When compared to continued treatment with clomiphene citrate, gonadotrophins resulted in more live births without increasing the rate of multiple pregnancies. Gonadotrophins resulted in more clinical pregnancies, but also in more miscarriages than clomiphene citrate, while there were no cases of ovarian hyperstimulation syndrome.
Quality of the evidence
The quality of the evidence was low to very low for outcomes from rFSH versus urinary gonadotrophins, and human menopausal gonadotrophin versus FSH‐P. The quality of the evidence was moderate for outcomes from gonadotrophins versus continued clomiphene citrate.
Ten of the fifteen studies included in this review reported a commercial sponsor.
Authors' conclusions
Summary of findings
| Recombinant follicle stimulating hormone versus urinary‐derived gonadotrophins for ovulation induction in women with polycystic ovarian syndrome | |||||
| Patient or population: women with polycystic ovarian syndrome (PCOS) undergoing ovulation induction Comparison: urinary‐derived gonadotrophins | |||||
| Outcomes | Anticipated absolute effects * (95% CI) | Relative effect | No of Participants | Quality of the evidence | |
| Risk with urinary‐derived gonadotrophins | Risk with rFSH | ||||
| Live birth rate per woman | 157 per 1000 | 190 per 1000 | RR 1.21 | 505 | ⊕⊕⊝⊝ |
| Incidence of multiple pregnancy (per woman) | 30 per 1000 | 25 per 1000 | RR 0.86 | 1368 | ⊕⊕⊝⊝ |
| Clinical pregnancy rate per woman | 239 per 1000 | 251 per 1000 | RR 1.05 | 1330 | ⊕⊕⊝⊝ |
| Miscarriage rate per woman | 47 per 1000 | 56 per 1000 | RR 1.20 | 970 | ⊕⊕⊝⊝ |
| Incidence of OHSS per woman | 22 per 1000 | 33 per 1000 | RR 1.48 | 1565 | ⊕⊝⊝⊝ |
| * 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). | |||||
| GRADE Working Group grades of evidence | |||||
| aDowngraded one level for imprecision around the absolute effect | |||||
| Human menopausal gonadotrophin or highly purified human menopausal gonadotrophin versus urinary follicle stimulating hormone for ovulation induction in women with polycystic ovarian syndrome | |||||
| Patient or population: women with polycystic ovarian syndrome (PCOS) undergoing ovulation induction Comparison: urinary follicle stimulating hormone (uFSH) | |||||
| Outcomes | Anticipated absolute effects * (95% CI) | Relative effect | No of Participants (studies) | Quality of the evidence | |
| Risk with uFSH | Risk with HMG or HP‐HMG | ||||
| Live birth rate per woman | 179 per 1000 | 230 per 1000 | RR 1.28 (0.65 to 2.52) | 138 | ⊕⊝⊝⊝ |
| Incidence of multiple pregnancy (per woman) | 23 per 1000 | 48 per 1000 | RR 2.13 | 161 | ⊕⊝⊝⊝ |
| Clinical pregnancy rate per woman | 203 per 1000 | 266 per 1000 | RR 1.31 | 102 | ⊕⊝⊝⊝ |
| Miscarriage rate per woman | 82 per 1000 | 27 per 1000 | RR 0.33 | 98 | ⊕⊝⊝⊝ |
| Incidence of OHSS per woman | No events c | 4/28 c | RR 7.07 | 53 | ⊕⊝⊝⊝ |
| * 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). | |||||
| GRADE Working Group grades of evidence | |||||
| aDowngarded two levels for serious imprecision around the absolute effect (wide CI and small sample size) d Downgraded one level for inconsistent definition or for the lack of definition of OHSS; two of four studies did not report this outcome | |||||
| Gonadotrophins compared to continued clomiphene citrate for ovulation induction | ||||||
| Patient or population: anovulatory women with clomiphene citrate‐failure | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | No of Participants | Quality of the evidence | Comments | |
| Risk with continued CC | Risk with gonadotrophins | |||||
| Live birth rate per woman | 413 per 1000 | 512 per 1000 | RR 1.24 | 661 | ⊕⊕⊕⊝ | |
| Incidence of multiple pregnancy per woman | 24 per 1000 | 21 per 1000 | RR 0.89 | 661 | ⊕⊕⊕⊝ | |
| Clinical pregnancy rate per woman | 446 per 1000 | 584 per 1000 | RR 1.31 | 661 | ⊕⊕⊕⊝ | |
| Miscarriages per woman | 33 per 1000 | 73 per 1000 | RR 2.23 | 661 | ⊕⊕⊝⊝ | There may be little or no difference when expressed per clinical pregnancy |
| Incidence of OHSS per woman | 0 per 1000 | 0 per 1000 | not estimable | 661 | ⊕⊕⊝⊝ | |
| *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). | ||||||
| GRADE Working Group grades of evidence | ||||||
| a Downgraded one level for risk of bias – no blinding performed b Downgraded one level for imprecision in result c Downgraded one level for inconsistency in outcome, i.e. there were more clinical pregnancies in the gonadotrophin group; there may be little or no difference when expressing miscarriage per clinical pregnancy | ||||||
Background
Description of the condition
Subfertility occurs in one in 10 couples world‐wide. In about one‐third of couples, this is based on polycystic ovary syndrome (PCOS). PCOS is characterised by oligo‐anovulation, clinical or biochemical hyperandrogenism, and polycystic ovaries (Rotterdam consensus group 2004a; Rotterdam consensus group 2004b). The syndrome affects approximately 6% to 10% of women of childbearing age.
Infertility due to chronic anovulation is the most common reason for women with PCOS to seek counselling or treatment. First line treatment for these women is ovulation induction with clomiphene citrate, with or without metformin. A recent review showed that letrozole is an effective alternative to clomiphene citrate (Franik 2018).
About 20% of women do not ovulate on clomiphene citrate, and require alternative or second‐line ovulation induction strategies. This failure to ovulate with clomiphene citrate is termed ‘clomiphene resistance’. The most common treatment in women with clomiphene citrate‐resistant PCOS is ovulation induction with gonadotrophins (Balen 2013), or laparoscopic electrocautery of the ovaries as an effective alternative treatment (Farquhar 2012).
Of the women ovulating on clomiphene citrate, only half of these women conceive within six months of treatment leads. If women fail to conceive with clomiphene citrate, despite regular ovulatory cycles, the term ‘clomiphene‐failure’ is used. Also in these women, clomiphene citrate or letrozole treatment is often changed to second‐line ovulation induction with gonadotrophins.
Description of the intervention
The strategy of stimulating follicle development and growth with exogenous gonadotrophins for ovulation induction in women with clomiphene citrate‐resistant PCOS or clomiphene citrate‐failure is well established.
Follicle‐stimulating hormone (FSH) is found in the pituitary gland, and circulates in the bloodstream in various molecular forms. This molecular heterogeneity is due to the variation in the structures of the carbohydrate moieties, in particular of sialic acid. It is the configuration of these carbohydrate moieties that determines the FSH isoform. The configuration depends on which glycosylation enzymes are available in the cell during synthesis (Wide 1997). Each molecular glycoform has a different molecular weight, net charge, circulating half‐life, and metabolic clearance (Baenziger 1988; Gray 1988; Stockell Hartree 1992; Wilson 1990). Gonadotrophins were originally extracted from pituitary glands (Gemzell 1958), and later from the urine of postmenopausal women (Lunenfeld 1960).
Over the last five decades, various urinary‐derived FSH products, or urofollitropins, have been developed. Menotropin (human menopausal gonadotrophin) has been available since the early 1960s and contains FSH, luteinising hormone (LH) and large quantities of potentially allergenic urinary proteins. Purified urofollitropin has been available since the mid‐1980s. Purified FSH is devoid of LH, but still contains urinary proteins. Highly purified urofollitropin has been available since the mid‐1990s and contains very small amounts of urinary proteins. The absence of urinary proteins reduces rare adverse reactions, such as local allergy or hypersensitivity (Albano 1996; Biffoni 1994). The most recent development in urinary gonadotrophins is highly purified menotropin (highly purified human menopausal gonadotrophin), containing equal amounts of FSH and LH activity.
To obtain even higher purity, gonadotrophins were developed with recombinant DNA technology (recombinant FSH) in 1988 (Howles 1996; Keene 1989). The production of recombinant FSH is independent of urine collection, thus guaranteeing a high availability of a biochemical pure FSH preparation that is free from LH and urinary protein contaminants. The production process also yields FSH with high specific bioactivity (roughly 100 times higher than for urine‐derived FSH products), minimal batch‐to‐batch discrepancies (Bergh 1999), and low immunogenicity. There is evidence that recombinant FSH has a higher bioactivity than urinary products (Andersen 2004).
At present, two preparations of recombinant FSH are available: follitropin alpha and follitropin beta. Both preparations are similar to pituitary and urinary FSH, although they show minor differences in the structure of the carbohydrate side chains, and contain more basic and fewer acidic isohormones than the urinary‐derived gonadotrophin preparations (De Leeuw 1996; Hard 1990; Lambert 1995).
Continued clomiphene citrate is taken for five days at the dose on which the woman ovulates. This is usually 50 mg, 100 mg, or 150 mg.
How the intervention might work
In the follicular phase of an ovulatory menstrual cycle, between 10 and 20 antral follicles develop. Of this cohort, one follicle will obtain dominance over the others, and will continue to grow until ovulation. In women with PCOS, this dominance does not occur. The aim of ovulation induction is to induce growth of preferably one follicle, and not more than three follicles. This can be accomplished by ovulation induction with gonadotrophins. Too forceful a regimen will result in overstimulation. and hence, in an increased risk of multiple pregnancy and ovarian hyperstimulation syndrome (OHSS); a stimulation regimen with too low a dosage of gonadotrophins will not result in a dominant follicle, and thereby, will not lead to ovulation.
Why it is important to do this review
Gonadotrophins are the standard drugs in medical ovulation induction for women with PCOS, who did not ovulate or conceive on clomiphene citrate. In women who do ovulate on clomiphene citrate, continued clomiphene citrate for another six cycles is an option. Knowlegde on effectiveness and safety of these treatment options will enable informed treatment decisions. The present review is an update and extension of two previous Cochrane reviews (Bayram 2001; Nugent 2000). Bayram 2001 compared rFSH with purified FSH and highly purified FSH; Nugent 2000 compared human menopausal gonadotrophin with purified FSH. No Cochrane review has yet compared human menopausal gonadotrophin with recombinant FSH in clomiphene citrate‐resistant women. Summarising the evidence on the effectiveness and safety of the various gonadotrophins will help gynaecologists and women to make informed decisions on the best regimen for ovulation induction.
Objectives
To compare the effectiveness and safety of gonadotrophins as a second‐line treatment for ovulation induction in women with clomiphene citrate‐resistant polycystic ovary syndrome (PCOS), and women who do not ovulate or conceive after clomiphene citrate.
Methods
Criteria for considering studies for this review
Types of studies
We included randomised controlled trials. We excluded quasi‐randomised controlled trials in which allocation was, for example, by alternation, reference to case record numbers, or to dates of birth. We also excluded cross‐over trials, which are not appropriate in this context (Vail 2003).
Types of participants
-
Subfertile women with clomiphene citrate‐resistant PCOS undergoing ovulation induction. We defined clomiphene citrate‐resistance as a failure to ovulate with clomiphene citrate doses of at least 100 mg/day for at least five days.
-
Subfertile women with PCOS and clomiphene citrate‐failure undergoing ovulation induction. We defined clomiphene citrate‐failure as a failure to conceive after three cycles of ovulation induction with clomiphene citrate.
-
Women with prior treatment with metformin with or without clomiphene citrate.
-
Women with prior treatment with electrocautery of the ovaries.
Types of interventions
-
Ovulation induction with recombinant follicle‐stimulating hormone (FSH) versus any other urinary gonadotrophin (human menopausal gonadotrophin, purified FSH, highly purified FSH)
-
Ovulation induction with highly purified FSH versus purified FSH
-
Ovulation induction with human menopausal gonadotrophin or highly purified human menopausal gonadotrophin versus purified FSH or highly purified FSH
-
Ovulation induction with gonadotrophins or continued clomiphene citrate
For all interventions, ovulation induction could be followed by intercourse or intrauterine insemination. We excluded trials involving co‐treatment with clomiphene citrate, metformin, luteinising hormone, letrozole or different gonadotrophins.
Types of outcome measures
Primary outcomes
1. Live birth rate per woman
2. Multiple pregnancy per woman
Secondary outcomes
3. Clinical pregnancy rate (per woman)
4. Miscarriage rate (per woman) or miscarriages per woman
5. Incidence of ovarian hyperstimulation syndrome (OHSS; (per woman))
6. Total gonadotrophin dose per woman (IU)
7. Total duration of stimulation per woman
Search methods for identification of studies
This review has drawn on the search strategy developed for the Cochrane Gynaecology and Fertility group (CGF) as a whole.
Electronic searches
Marian Showell (CGF Group Information Specialist) developed the search strategies. See Appendix 1, Appendix 2Appendix 3, Appendix 4, Appendix 5, Appendix 6.
1) We searched the following electronic sources:
-
Cochrane Gynaecology and Fertility Group specialised Register of Controlled Trials (searched 15 January 2018; Appendix 1)
-
Cochrane Central Register of Controlled Trials (CENTRAL Register of Studies Online (CRSO; searched 15 January 2018; Appendix 2))
-
MEDLINE (1946 to 15 January 2018; Appendix 3)
-
Embase (1980 to 15 January 2018; Appendix 4)
-
PsycINFO (1806 to 15 January 2018; Appendix 5)
-
CINAHL (1961 to 15 January 2018; Appendix 6)
2) Other electronic sources included:
-
US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (www.clinicaltrials.gov; searched 15 January 2018)
-
World Health Organization International Clinical Trials Registry Platform (apps.who.int/trialsearch/; searched 15 January 2018)
-
LILACS (Latin American and Caribbean Health Science Information database) and other Spanish and Portuguese language databases (pesquisa.bvsalud.org/portal/; 1982 to 15 January 2018)
-
OpenGrey for unpublished literature from Europe (www.opengrey.eu/; searched 15 January 2018)
Searching other resources
We searched the following conference abstracts:
-
American Society for Reproductive Medicine and Canadian Fertility and Andrology Society (ASRM/CFAS) Conjoint Annual Meeting (2001 to 2018), Abstracts of the Scientific Oral and Poster Sessions, Program Supplement;
-
European Society of Human Reproduction and Embryology (ESHRE) Annual Meeting (2001 to 2018), Abstracts of the Scientific Oral and Poster Sessions, Program Supplement.
We handsearched the references cited in all obtained studies. We searched PubMed and Google for any recent trials that had not yet been indexed in MEDLINE.
We asked Serono Benelux BV (Merck Group), Ferring, and IBSA, the manufacturers of gonadotrophins, for ongoing studies and unpublished data.
Data collection and analysis
Selection of studies
Three review authors (NW, EK, and MvW) independently examined the electronic search results for reports of possibly relevant trials, and retrieved these reports in full. All review authors independently applied the selection criteria to the trial reports, rechecking trial eligibility and resolving disagreements by discussion with the other authors.
Data extraction and management
Three review authors (NW, EK, and MvW) independently extracted the outcome data and information on funding, location, clinical and design details, and participants. We resolved any differences by discussion. We entered details of the studies into the 'Characteristics of included studies' table. We presented studies that appeared to meet the inclusion criteria but were excluded from the review in the 'Characteristics of excluded studies' table, briefly stating the reason for exclusion, but giving no further information.
Assessment of risk of bias in included studies
Three review authors (NW, EK, and MvW) extracted information regarding the risk of bias (threats to internal validity) under six domains (also see the Cochrane 'Risk of bias' assessment tool in Appendix 7; (Higgins 2011)). We resolved any differences by discussion.
1. Sequence generation. Evidence that an unpredictable random process was used.
2. Allocation concealment. Evidence that the allocation list was not available to anyone involved in the recruitment process.
3. Blinding of participants, clinicians, and outcome assessors. Evidence that knowledge of allocation was not available to those involved in subsequent treatment decisions or follow‐up efforts.
4. Completeness of outcome data. Evidence that any losses to follow‐up were low and comparable between groups.
5. Selective outcome reporting. Evidence that major outcomes had been reported in sufficient detail to allow analysis, independently of their apparent statistical significance.
6. Other potential sources. Evidence of miscellaneous errors or circumstances that might influence the internal validity of trial results.
We sought missing details from the authors of the original publications. We present all details in the 'Risk of bias' table following each included study.
Measures of treatment effect
We summarised all binary outcomes using relative risk ratio (RR) with a 95% confidence interval (CI). In cases of no events, we also calculated a risk difference (RD) with a 95% CI.
We treated ordinal scales, such as amount of gonadotrophin used and duration of ovarian stimulation, as continuous outcomes. We abstracted, calculated, or requested means and standard deviations and calculated the mean difference with 95% CI for these outcomes.
Unit of analysis issues
We expressed all outcomes per woman randomised, and multiple pregnancy per clinical pregnancy.
Dealing with missing data
Where there was insufficient information in the published report, we attempted to contact the authors for clarification. If missing data became available, we included them in the analysis. We anticipated that trials conducted over 10 years ago might not have data on live birth rates. We analysed data extracted from the trials on an intention‐to‐treat basis. Where randomised participants were missing from outcome assessment, we contacted the authors for additional data. If further data were not available, we assumed that missing participants had failed to achieve pregnancy and had not suffered any of the reported adverse events.
Assessment of heterogeneity
The presence of statistical heterogeneity of treatment effect among trials was determined using the I² statistic (Higgins 2003). We considered whether clinical and methodological characteristics of the included studies were sufficiently similar for meta‐analysis to provide a clinically meaningful summary. We assessed statistical heterogeneity by the measure of the I² statistic. We took an I² measurement greater than 50% to indicate substantial heterogeneity, in which case, we tested the effect of using a random‐effects model (Higgins 2011).
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 for duplication of data. If we had included 10 or more studies in an analysis, we had planned to use a funnel plot to explore the possibility of small‐study effects (a tendency for estimates of the intervention effect to be more beneficial in smaller studies).
Data synthesis
When multiple studies were available on a similar comparison, we used Review Manager 5.3 software to perform the meta‐analyses, using the Mantel‐Haenszel method with a fixed‐effect model (Review Manager 2014). For reporting purposes, we translated primary outcomes to absolute risks. We combined results for continuous outcomes using the mean difference.
Subgroup analysis and investigation of heterogeneity
If excessive heterogeneity existed within strata, we had planned to explore this informally using the clinical and design details recorded in the 'Characteristics of included studies' table.
-
Prospectively, we had planned to undertake three different stratifications of the primary outcomes: type of urinary gonadotrophin (human menopausal gonadotrophin, purified FSH and highly purified FSH); single or multiple cycles; sponsorship (commercial, non‐commercial (Lexchin 2003)).
Sensitivity analysis
We conducted sensitivity analyses for the primary outcomes to determine whether the conclusions were robust to arbitrary decisions made about study eligibility and analysis. These analyses included consideration of whether the review conclusions would have differed if:
-
we had used a random‐effects model
-
we had reported odds ratios rather than relative risk ratios
Overall quality of the body of evidence: 'Summary of findings' table
We generated 'Summary of findings' tables using GRADEpro software and Cochrane methods (GRADEpro GDT 2015; Higgins 2011). These tables present the overall quality of the body of evidence for main review outcomes (live birth, multiple pregnancy, clinical pregnancy, miscarriage, and OHSS) for the main review comparison (recombinant FSH versus urinary‐derived gonadotrophins) using GRADE criteria: study limitations (i.e. risk of bias), consistency of effect, imprecision, indirectness, and publication bias. We also presented tables for our other comparisons: human menopausal gonadotrophin or highly purified human menopausal gonadotrophin versus urinary FSH, and gonadotrophins versus continued clomiphene citrate. We justified judgements about evidence quality (high, moderate or low), documented them, and incorporated them into the reporting of results for each outcome.
Results
Description of studies
For details of the studies please see: Characteristics of included studies; Characteristics of excluded studies
Results of the search
For this update, we screened 588 titles and identified an additional five studies for eligibility assessment. From these five studies, we included one trial, we excluded two studies, and we listed two studies as studies awaiting classification.
See Figure 1.
Study flow diagram
Included studies
We included 15 trials in this update.
1. Ten studies compared the effects of recombinant follicle‐stimulating hormone (rFSH) versus urinary derived gonadotrophins (human menopausal gonadotrophin: Balen 2007; Platteau 2006; Revelli 2006; urinary follicle‐stimulating hormone (uFSH): Coelingh Bennink 1998; Feigenbaum 2001; Gerli 2004; Loumaye 1996; Szilágyi 2004; Taketani 2010; Yarali 1999). Loumaye 1996 was described in a review on human gonadotrophins produced by recombinant DNA technology. The authors of the 2001 Cochrane Review collected the data for this trial by personal communication, and we used them again in this update (Bayram 2001).
2. There were no studies that compared highly purified FSH with purified FSH.
3. Four studies compared purified FSH with human menopausal gonadotrophin (Gadir 1990: McFaul 1990; Sagle 1991; Seibel 1985). Gadir 1990 made an extra comparison with laparoscopic electrocautery of the ovaries.
4. One study compared gonadotrophins and continued clomiphene citrate during six cycles (Weiss 2018).
One trial also included normo‐ovulatory women with unexplained subfertility (Revelli 2006). For this review, we used only the data of women with polycystic ovary syndrome (PCOS). For Seibel 1985, we included pre‐cross‐over data.
Eight trials reported data on live birth, and thirteen trials reported on multiple pregnancy (Balen 2007; Coelingh Bennink 1998; Feigenbaum 2001; Gadir 1990; Gerli 2004; McFaul 1990; Platteau 2006; Revelli 2006; Sagle 1991; Seibel 1985; Taketani 2010; Weiss 2018; Yarali 1999).
All studies that compared types of gonadotrophins included women who were clomiphene citrate‐resistant; seven of them also included women with clomiphene citrate‐failure (Balen 2007; Coelingh Bennink 1998; Gerli 2004; Platteau 2006; Seibel 1985; Yarali 1999). The study that compared gonadotrophins with continuous included only women with clomiphene citrate‐failure (Weiss 2018). None of the women included in this review had been treated with electrocautery in the past. Ten trials analysed more than one cycle per woman, whereas five trials only analysed one cycle per woman (Balen 2007; Feigenbaum 2001; Platteau 2006; Revelli 2006; Taketani 2010). In four trials, intrauterine insemination was performed in some cases (Balen 2007; Gerli 2004; Platteau 2006; Weiss 2018). All trials used a low‐dose step‐up protocol, but the protocol used in Loumaye 1996 was unknown. Ten trials reported a commercial sponsor (Balen 2007; Coelingh Bennink 1998; Feigenbaum 2001; Loumaye 1996; Platteau 2006; Sagle 1991; Seibel 1985; Szilágyi 2004; Taketani 2010; Yarali 1999).
Six trials reported a power calculation (Balen 2007; Coelingh Bennink 1998; Loumaye 1996; Platteau 2006; Revelli 2006; Weiss 2018).
Excluded studies
We excluded six trials: one trial because the outcome measure was the effect of FSH on haemostasis (Ricci 2004); two studies because the outcome 'pregnancy' was not defined, and this outcome was only presented per cycle (Homburg 1990; Jacobs 1987); one study because it was a cross‐over design, and it was not possible to extract the pre‐cross‐over data per woman (Larsen 1990); one study had the wrong intervention (cotreatment with clomiphene citrate (Rashidi 2016)); and one study reported a wrong comparator (Zhou 2016).
Risk of bias in included studies
We summarised the risks of bias in the included studies in Figure 2 and Figure 3.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies
Risk of bias summary: review authors' judgements about each risk of bias item for each included study
Allocation
Allocation to the intervention or control group was adequately concealed in four trials (Balen 2007; Loumaye 1996; Platteau 2006; Weiss 2018). The allocation concealment was inadequate in two trials (Gadir 1990; Gerli 2004), and unclear in the remaining trials.
Blinding
Four trials were assessor‐blinded (Balen 2007; Coelingh Bennink 1998; Platteau 2006; Taketani 2010). Blinding was not performed in the remaining studies.
Incomplete outcome data
Two trials had a high risk of attrition bias (Seibel 1985; Szilágyi 2004). For another two trials, this was unclear (Loumaye 1996; Taketani 2010). All other trials had a low risk of bias for this domain.
Selective reporting
We rated six studies as having a low risk of selective reporting bias; eight as having an unclear risk of bias in this domain, and one study as having high risk (Szilágyi 2004).
Other potential sources of bias
We rated this as unclear for all studies. Some studies provided too few details to make a judgement. Within all the trials, the baseline characteristics appeared balanced over the two treatment groups. Only six of the 15 trials mentioned the duration of the trial (Balen 2007; Coelingh Bennink 1998; Loumaye 1996; Platteau 2006; Taketani 2010; Weiss 2018).
Effects of interventions
See: Summary of findings for the main comparison Recombinant follicle stimulating hormone versus urinary‐derived gonadotrophins for ovulation induction in women with polycystic ovarian syndrome; Summary of findings 2 Human menopausal gonadotrophin or highly purified human menopausal gonadotrophin versus urinary follicle stimulating hormone for ovulation induction in women with polycystic ovarian syndrome; Summary of findings 3 Gonadotrophins compared to continued clomiphene citrate for ovulation induction
1 Recombinant follicle‐stimulating hormone (rFSH) versus urinary‐derived gonadotrophins
1.1 Live birth rate per woman
Forest plot of comparison 1. Recombinant FSH (rFSH) versus urinary‐derived gonadotrophins (u‐gonadotrophins), outcome: 1.1 Live birth rate per woman by urinary gonadotrophins
Five trials, including 505 women, reported on live birth (Balen 2007; Feigenbaum 2001; Platteau 2006; Revelli 2006; Szilágyi 2004). After pooling the results, the overall risk ratio (RR) per woman was 1.21 (95% confidence interval (CI) 0.83 to 1.78; five RCTs, N = 505; I² = 9%, low‐quality evidence) indicating there may be little or no difference between the intervention and the comparison. Translated into absolute risks, this means that for a woman with a 16% chance of achieving a live birth with the use of urinary‐derived FSH, the chance of a live birth with the use of rFSH would be between 13% and 28%. Statistical heterogeneity for this outcome was low. The live birth rate varied from 16% to 40% in the rFSH group, and from 0% to 25% in the urinary gonadotrophin group.
When we divided the urinary‐derived gonadotrophins into subgroups (three trials compared rFSH versus highly purified human menopausal gonadotrophin, two trials compared rFSH versus highly purified FSH), we found that there may be little or no difference between the subgroups (P = 0.10). The RR for rFSH versus highly purified human menopausal gonadotrophin or human menopausal gonadotrophin was 1.04 (95% CI 0.68 to 1.57; three RCTs, N = 409; I² = 0%; low‐quality evidence), and for rFSH versus highly purified FSH was 2.66 (95% CI 0.95 to 7.43; two RCTs, N = 96; I² = 11%; low‐quality evidence).
1.2 Live birth rate per woman ‐ stratified by sponsor
All trials comparing rFSH and highly purified human menopausal gonadotrophin were sponsored by Ferring; the other two trials comparing rFSH and purified FSH did not report the sponsor. Therefore, subgrouped results per sponsor were similar to the gonadotrophin comparison, i.e. when we divided into subgroups, we found little or no difference between subgroups (P = 0.1; Analysis 1.2).
1.3 Incidence of multiple pregnancy per woman
Eight studies, including 1368 women, reported on multiple pregnancy (Balen 2007; Coelingh Bennink 1998; Feigenbaum 2001; Gerli 2004; Platteau 2006; Revelli 2006; Taketani 2010; Yarali 1999). There may be little or no difference in multiple pregnancy per woman between groups (RR 0.86, 95% CI 0.46 to 1.61; eight RCTs, N = 1368; I² = 0%; low‐quality evidence; Analysis 1.3).
When we divided the urinary‐derived gonadotrophins into subgroups (three trials compared rFSH versus highly purified human menopausal gonadotrophin, five trials compared rFSH versus highly purified FSH), there was no evidence of a difference between the subgroups (P = 0.34).
1.4 Incidence of multiple pregnancy per woman ‐ stratified per sponsor
When we subgrouped by sponsor, we found little or no difference between subgroups (P = 0.86; Analysis 1.4).
1.5 Clinical pregnancy rate per woman
Eight studies, including 1330 women, reported on clinical pregnancy (Balen 2007; Coelingh Bennink 1998; Feigenbaum 2001; Gerli 2004; Loumaye 1996; Platteau 2006; Taketani 2010; Yarali 1999). There may be little or no difference in clinical pregnancy (RR 1.05, 95% CI 0.88 to 1.27; eight RCTs, N = 1330; I² = 0%; low‐quality evidence; Analysis 1.5).
When we divided the urinary‐derived gonadotrophins into subgroups (three trials compared rFSH versus highly purified human menopausal gonadotrophin, five trials compared rFSH versus highly purified FSH), there was no evidence of a difference between the subgroups (P = 0.47). The RR for rFSH versus HP‐human menopausal gonadotrophin was 1.19 (95% CI 0.81 to 1.77, three RCTs, N = 409; I² = 0%; low‐quality evidence), and for rFSH versus highly purified FSH was 1.01 (95% CI 0.82 to 1.25; five RCTs, N = 921; I² = 0%; low‐quality evidence).
1.6 Incidence of multiple pregnancy per clinical pregnancy
We found that there may be little or no difference in multiple pregnancy per clinical pregnancy (RR 0.75, 95% CI 0.43 to 1.32; eight RCTs, 315 pregnancies; I² = 0%; Analysis 1.6).
1.7 Miscarriage rate per woman
Seven studies, including 970 women, reported on miscarriage (Balen 2007; Coelingh Bennink 1998; Gerli 2004; Loumaye 1996; Platteau 2006; Szilágyi 2004; Yarali 1999). There may be little or no difference in miscarriage rate (RR 1.20, 95% CI 0.71 to 2.04; seven RCTs, N = 970; I² = 0%; low‐quality evidence; Analysis 1.7).
When we divided the urinary‐derived gonadotrophins into subgroups (two trials compared rFSH versus highly purified human menopausal gonadotrophin, five trials compared rFSH versus highly purified FSH), we found no evidence of a difference between the subgroups (P = 0.71).
1.8 Incidence of ovarian hyperstimulation syndrome (OHSS) per woman
Forest plot of comparison 1. Recombinant FSH (rFSH) versus urinary‐derived gonadotrophins (u‐gonadotrophins), outcome: 1.8 Incidence of OHSS per woman by urinary gonadotrophins
Ten studies, including 1565 women, reported OHSS (Balen 2007; Coelingh Bennink 1998; Feigenbaum 2001; Gerli 2004; Loumaye 1996; Platteau 2006; Revelli 2006; Szilágyi 2004; Taketani 2010; Yarali 1999). After pooling the results, the overall RR for OHSS per woman was 1.48 (95% CI 0.82 to 2.65; 10 RCTs, N = 1565; I² = 0%; very low‐quality evidence), indicating we could not be certain whether rFSH reduced the incidence of OHSS (Analysis 1.8). This means that for a woman with a 2.2% chance of OHSS with the use of urinary‐derived gonadotrophins, the chance of OHSS with the use of rFSH would be between 1.2% and 9.6%. The OHSS rate varied from 0% to 20% in both groups.
When we divided the urinary‐derived gonadotrophins into subgroups (three trials compared rFSH versus highly purified human menopausal gonadotrophin, seven trials compared rFSH versus highly purified FSH), we found no evidence of a difference between the subgroups (P = 0.53). The RR for rFSH versus highly purified human menopausal gonadotrophin was 1.11 (95% CI 0.39 to 3.20; three RCTs, N = 409; I² = 0%; very low‐quality evidence), and for rFSH versus highly purified FSH was 1.67 (95% CI 0.82 to 3.39; seven RCTs, N = 1156; I² = 0%; very low‐quality evidence).
1.9 Mean total gonadotrophin dose per woman
We found that rFSH required a lower dose than urinary‐derived gonadotrophins to stimulate ovulation (MD ‐105.44 IU, 95% CI ‐154.21 to ‐56.68; six RCTs, N = 1046; I² = 81%). When we used a random‐effects model, in view of the high statistical heterogeneity, we found there may be little or no difference (MD ‐127.4 IU, 95% CI ‐258.06 to 3.26; Analysis 1.9).
1.10 Total duration of stimulation per woman (days)
We found that rFSH required a shorter time to stimulate ovulation than urinary‐derived gonadotrophins (MD ‐0.66 days, 95% CI ‐1.04 to ‐0.28; six RCTs, N = 1122; I² = 72%). When we used a random‐effects model, in view of the high statistical heterogeneity, we found there may be little or no difference (MD ‐0.80 days, 95% CI ‐1.66 to 0.05; Analysis 1.10).
2 Human menopausal gonadotrophin or highly purified human menopausal gonadotrophin versus urinary FSH (uFSH)
2.1 Live birth per woman
Three trials, including 138 women, reported on live birth (Gadir 1990; McFaul 1990; Sagle 1991). We are uncertain whether human menopausal gonadotrophin or highly purified human menopausal gonadotrophin improved live birth rate (RR 1.28, 95% CI 0.65 to 2.52; three RCTs, N = 138; I² = 0%; very low‐quality evidence; Analysis 2.1; Figure 6).
Forest plot of comparison 2. Human menopausal gonadotrophin (HMG) or highly purified HMG (HP‐HMG) versus urinary FSH (uFSH), outcome: 2.1 Live birth rate per woman
2.2. Incidence of multiple pregnancy per woman
We are uncertain whether human menopausal gonadotrophin or highly purified human menopausal gonadotrophin led to a higher multiple pregnancy rate per woman (RR 2.13, 95% CI 0.51 to 8.91; four RCTs, N = 161; I² = 0%; very low‐quality evidence; Analysis 2.2). As two of the four studies had no multiple pregnancies, we also calculated the risk difference (RD 0.03, 95% CI ‐0.05 to 0.11).
2.3 Incidence of multiple pregnancy per clinical pregnancy
We are uncertain whether human menopausal gonadotrophin or highly purified human menopausal gonadotrophin led to a higher multiple pregnancy rate per clinical pregnancy (RR 4.20, 95% CI 0.21 to 83.33; four RCTs, N = 161; I² = 0%; Analysis 2.3). As two of the four studies had no multiple pregnancies, we also calculated the risk difference (RD 0.11, 95% CI ‐0.22 to 0.45).
2.4 Clinical pregnancy rate per woman
One study reported clinical pregnancy rate per woman (Sagle 1991). McFaul 1990 presented pregnancy rates without defining this outcome. For this study, we calculated the clinical pregnancy rates by adding the number of live births to the number of miscarriages in each group. Seibel 1985 reported conception rates, which we used as clinical pregnancy rate. After pooling the data, we are uncertain whether human menopausal gonadotrophin or highly purified human menopausal gonadotrophin improved clinical pregnancy rate (RR 1.31, 95% CI 0.66 to 2.59; three RCTs, N = 102; I² = 0%; very low‐quality evidence; Analysis 2.4).
2.5 Miscarriage rate per woman
We are uncertain whether human menopausal gonadotrophin or highly purified human menopausal gonadotrophin reduced miscarriage rate (RR 0.33, 95% CI 0.06 to 1.97; two RCTs, N = 98; I² = 0%; very low‐quality evidence; Analysis 2.5).
2.6 Incidence of OHSS per woman
Two studies, including 53 women, reported OHSS (Sagle 1991; Seibel 1985). We are uncertain whether human menopausal gonadotrophin or highly purified human menopausal gonadotrophin reduced the incidence of OHSS (RR 7.07, 95% CI 0.42 to 117.81; two RCTs, N = 53; very low‐quality evidence; Analysis 2.6).
2.7 Mean total gonadotrophin dose per woman
Gadir 1990 and McFaul 1990 reported mean values for total doses, but they did not state standard deviations. Mean total does for human menopausal gonadotrophin or highly purified human menopausal gonadotrophin versus uFSH were 1568 IU versus 1478 IU in Gadir 1990, and 1770 IU versus 1995 IU in McFaul 1990. The authors reported that they found no significant difference between groups.
Sagle 1991 also reported no significant difference between groups. They reported values in mean total dose per cycle: human menopausal gonadotrophin or highly purified human menopausal gonadotrophin 1080 IU (range: 525 to 1950 IU) versus uFSH 1447.5 IU (range: 675 to 2887.5 IU).
2.8 Total duration of stimulation per woman (days)
McFaul 1990 reported no significant mean difference between human menopausal gonadotrophin (11.8 days) and uFSH (11.9 days). They did not provide standard deviations.
3 Gonadotrophins versus continued clomiphene citrate
One trial, including 661 women, measured all outcomes (Weiss 2018).
3.1 Live birth rate per woman
One trial, including 666 women reported on live birth (Weiss 2018). We found that gonadotrophins resulted in more live births than continued clomiphene citrate (RR 1.24, 95% CI 1.05 to 1.46; one trial, N = 661; moderate‐quality evidence; Analysis 3.1). This suggests that for a woman with a live birth rate of 41% with continued clomiphene citrate, the live birth rate with FSH was 43% to 60%.
3.2. Incidence of multiple pregnancy per woman
There is probably little or no difference in the multiple pregnancy rate per woman (RR 0.89, 95% CI 0.33 to 2.44; one trial, N = 661; moderate‐quality evidence; Analysis 3.2).
3.3 Incidence of multiple pregnancy per clinical pregnancy
There is probably little or no difference in the multiple pregnancy rate per clinical pregnancies (RR 0.68, 95% CI 0.25 to 1.84; one trial, N = 661, moderate‐quality evidence; Analysis 3.3).
3.4 Clinical pregnancy rate per woman
Gonadotrophins resulted in more clinical pregnancies than continued clomiphene citrate (RR 1.31, 95% CI 1.13 to 1.52; one trial, N = 661; moderate‐quality evidence; Analysis 3.4).
3.5 Miscarriage rate per woman
The number of miscarriages was higher in the group treated with gonadotrophins than in the clomiphene citrate group (RR 2.23, 95% CI 1.11 to 4.47; one trial, N = 661; low‐quality evidence). When expressed per clinical pregnancy, there was probably little or no difference in miscarriage rate (RR 1.70, 95% 0.86 to 3.36; Analysis 3.5)
3.6 Incidence of OHSS per woman
OHSS did not occur in any of the women, therefore, we could not calculate the RR. The estimate for the risk difference was (0.00, 95% CI ‐0.01 to 0.01; one trial, N = 661; low‐quality evidence; Analysis 3.6).
Discussion
Summary of main results
This review compared the effectiveness and safety of gonadotrophins as a second‐line treatment for ovulation induction in women with polycystic ovary syndrome (PCOS) who did not ovulate or conceive on clomiphene citrate. We found 10 studies that compared recombinant follicle‐stimulating hormone (rFSH) with urinary‐derived gonadotrophins, four trials that compared urinary FSH (uFSH) with human menopausal gonadotrophin, and one trial that compared gonadotrophins with continued clomiphene citrate. There may be little or no difference in pregnancy outcomes when rFSH was compared to urinary gonadotrophins as a whole. We are uncertain whether human menopausal gonadotrophin or highly purified human menopausal gonadotrophin improved pregnancy outcomes when compared with uFSH. We are uncertain whether there was any difference observed in ovarian hyperstimulation syndrome (OHSS) for any of the comparisons. We found no trials that compared rFSH and purified FSH, or highly purified FSH and purified FSH. The use of gonadotrophins resulted in higher live birth rates without increasing multiple pregnancy rates when compared to continued clomiphene citrate.
Overall completeness and applicability of evidence
For the trials that compared rFSH and urinary‐derived gonadotrophins, outcome data needed to make the planned comparisons were largely available; these trials were all published after 1996. The data from trials that compared rFSH and purified uFSH and highly purified uFSH were incomplete, probably because these trials had been published between 1985 and 1991, when there were no CONSORT or PRISMA guidelines, and clinical pregnancy or ovulation rates were still accepted endpoints. The outcome data for the gonadotrophin versus continued clomiphene citrate trial was complete.
Seven trials did not define the outcome OHSS. The remaining studies used very different definitions (see Characteristics of included studies). It is common to categorise cases of OHSS by three degrees; mild, moderate, or severe (Youssef 2014). Since this ranking was almost never used in the included studies of this review, it may be inappropriate to pool the data on OHSS. Also, different starting dosages were used, varying from 50 to 150 IU per day, with various criteria outlined to withhold an injection of human chorionic gonadotrophin. This may influence the incidence of OHSS, regardless of the type of gonadotrophin used. Nowadays, OHSS is not a common finding in ovulation induction. OHSS is mainly a complication that occurs after treatment with in vitro fertilisation (Youssef 2014).
The data on gonadotrophin dose used and duration of stimulation were never presented per woman randomised, and showed high statistical heterogeneity. Therefore, these outcomes are likely to be biased, and one should not draw conclusions on the basis of these data.
Four of the included studies comparing gonadotrophins used intrauterine insemination (IUI) in addition to ovulation induction with gonadotrophins. IUI may or may not have increased the pregnancy rate, but since these studies always provided IUI in both study arms, its effect on differential pregnancy rates was likely to be small. In the study comparing gonadotrophins with continued clomiphene citrate, women had also been randomised to IUI or intercourse. This study found little or no differences in the effect of IUI on any of the pregnancy outcomes (Weiss 2018).
For the studies comparing gonadotrophins, the included population represented women with PCOS who were either clomiphene citrate‐resistant or had failed to conceive with clomiphene citrate. The evidence is broadly applicable as a second‐line treatment for ovulation induction in these women. The study comparing gonadotrophins and continued clomiphene citrate included only women who had ovulated on previous clomiphene citrate cycles but failed to conceive.
Quality of the evidence
Using GRADE assessment, we found that evidence for most outcomes was of low to very low quality, due to the limited number of studies comparing gonadotrophins, small study size, statistical heterogeneity, and the risk of bias in the individual studies.
For the study comparing gonadotrophins with continuous clomiphene citrate, we assessed evidence for live birth and clinical pregnancy to be of moderate quality.
Potential biases in the review process
Strengths of this review include comprehensive systematic searching for eligible studies, rigid inclusion criteria for RCTs and data extraction, and independent analysis by three review authors. The possibility of publication bias was minimised by including both published and unpublished studies (such as abstracts from meetings). However, as with any review, we cannot guarantee that we found all eligible studies.
Agreements and disagreements with other studies or reviews
Our results are in line with the outcomes of the previous Cochrane review of Bayram 2001, in concluding that rFSH and urinary‐derived gonadotrophins are equally effective for ovulation induction in women with PCOS, in terms of ovulation rate, pregnancy rate, miscarriage rate, and multiple pregnancy rate. Our results are also in line with the outcomes of the previous Cochrane Review of Nugent 2000, who concluded that comparing FSH and human menopausal gonadotrophin showed little or no difference in pregnancy rates. Nugent 2000 did find a significant reduction in OHSS rate per cycle in women treated with purified FSH compared to human menopausal gonadotrophin. We focused on OHSS rate per woman, and found little or no difference, although only two trials were available for this analysis.
Bayram 2001 and Nugent 2000 did not evaluate the outcome of live birth. We found there may be little or no difference in live birth rate for the comparison of rFSH versus urinary gonadotrophins. We were uncertain whether human menopausal gonadotrophin or highly purified human menopausal gonadotrophin improved live birth rate when compared to uFSH.
Another review compared rFSH with urinary‐derived FSH products (Nahuis 2009). The authors found that follitropin alpha, beta, and urinary FSH products appeared to be similarly effective in live birth rates, and clinical, ongoing, and multiple pregnancy rates. Nahuis 2009 did not pool data on OHSS.
Weiss 2018 was the first to compare gonadotrophins and continuous clomiphene citrate in anovulatory women with clomiphene citrate‐failure.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies
Risk of bias summary: review authors' judgements about each risk of bias item for each included study
Forest plot of comparison 1. Recombinant FSH (rFSH) versus urinary‐derived gonadotrophins (u‐gonadotrophins), outcome: 1.1 Live birth rate per woman by urinary gonadotrophins
Forest plot of comparison 1. Recombinant FSH (rFSH) versus urinary‐derived gonadotrophins (u‐gonadotrophins), outcome: 1.8 Incidence of OHSS per woman by urinary gonadotrophins
Forest plot of comparison 2. Human menopausal gonadotrophin (HMG) or highly purified HMG (HP‐HMG) versus urinary FSH (uFSH), outcome: 2.1 Live birth rate per woman
Comparison 1 Recombinant follicle stimulating hormone (rFSH) versus urinary‐derived gonadotrophins (u‐gonadotrophins), Outcome 1 Live birth rate per woman by urinary gonadotrophins.
Comparison 1 Recombinant follicle stimulating hormone (rFSH) versus urinary‐derived gonadotrophins (u‐gonadotrophins), Outcome 2 Live birth rate per woman by sponsor.
Comparison 1 Recombinant follicle stimulating hormone (rFSH) versus urinary‐derived gonadotrophins (u‐gonadotrophins), Outcome 3 Multiple pregnancy per woman by urinary gonadotrophins.
Comparison 1 Recombinant follicle stimulating hormone (rFSH) versus urinary‐derived gonadotrophins (u‐gonadotrophins), Outcome 4 Multiple pregnancy per woman by sponsor.
Comparison 1 Recombinant follicle stimulating hormone (rFSH) versus urinary‐derived gonadotrophins (u‐gonadotrophins), Outcome 5 Clinical pregnancy rate per woman by urinary gonadotrophins.
Comparison 1 Recombinant follicle stimulating hormone (rFSH) versus urinary‐derived gonadotrophins (u‐gonadotrophins), Outcome 6 Incidence of multiple pregnancy per clinical pregnancy by urinary gonadotrophins.
Comparison 1 Recombinant follicle stimulating hormone (rFSH) versus urinary‐derived gonadotrophins (u‐gonadotrophins), Outcome 7 Miscarriage rate per woman by urinary gonadotrophins.
Comparison 1 Recombinant follicle stimulating hormone (rFSH) versus urinary‐derived gonadotrophins (u‐gonadotrophins), Outcome 8 Incidence of OHSS per woman by urinary gonadotrophins.
Comparison 1 Recombinant follicle stimulating hormone (rFSH) versus urinary‐derived gonadotrophins (u‐gonadotrophins), Outcome 9 Total gonadotrophin dose per woman (IU) by urinary gonadotrophins.
Comparison 1 Recombinant follicle stimulating hormone (rFSH) versus urinary‐derived gonadotrophins (u‐gonadotrophins), Outcome 10 Total duration of stimulation per woman (days) by urinary gonadotrophins.
Comparison 2 Human menopausal gonadotrophin (HMG) or highly purified HMG (HP‐HMG) versus urinary FSH (uFSH), Outcome 1 Live birth rate per woman.
Comparison 2 Human menopausal gonadotrophin (HMG) or highly purified HMG (HP‐HMG) versus urinary FSH (uFSH), Outcome 2 Multiple pregnancy per woman.
Comparison 2 Human menopausal gonadotrophin (HMG) or highly purified HMG (HP‐HMG) versus urinary FSH (uFSH), Outcome 3 Multiple pregnancy per clinical pregnancy.
Comparison 2 Human menopausal gonadotrophin (HMG) or highly purified HMG (HP‐HMG) versus urinary FSH (uFSH), Outcome 4 Clinical pregnancy rate per woman.
Comparison 2 Human menopausal gonadotrophin (HMG) or highly purified HMG (HP‐HMG) versus urinary FSH (uFSH), Outcome 5 Miscarriage rate per woman.
Comparison 2 Human menopausal gonadotrophin (HMG) or highly purified HMG (HP‐HMG) versus urinary FSH (uFSH), Outcome 6 Incidence of OHSS per woman.
Comparison 3 Gonadotrophins (FSH) versus continued clomiphene citrate (CC), Outcome 1 Live birth rate per woman.
Comparison 3 Gonadotrophins (FSH) versus continued clomiphene citrate (CC), Outcome 2 Multiple pregnancy per woman.
Comparison 3 Gonadotrophins (FSH) versus continued clomiphene citrate (CC), Outcome 3 Multiple pregnancy (per clinical pregnancy).
Comparison 3 Gonadotrophins (FSH) versus continued clomiphene citrate (CC), Outcome 4 Clinical pregnancy per woman.
Comparison 3 Gonadotrophins (FSH) versus continued clomiphene citrate (CC), Outcome 5 Miscarriages per woman.
Comparison 3 Gonadotrophins (FSH) versus continued clomiphene citrate (CC), Outcome 6 Incidence of OHSS per woman.
| Recombinant follicle stimulating hormone versus urinary‐derived gonadotrophins for ovulation induction in women with polycystic ovarian syndrome | |||||
| Patient or population: women with polycystic ovarian syndrome (PCOS) undergoing ovulation induction Comparison: urinary‐derived gonadotrophins | |||||
| Outcomes | Anticipated absolute effects * (95% CI) | Relative effect | No of Participants | Quality of the evidence | |
| Risk with urinary‐derived gonadotrophins | Risk with rFSH | ||||
| Live birth rate per woman | 157 per 1000 | 190 per 1000 | RR 1.21 | 505 | ⊕⊕⊝⊝ |
| Incidence of multiple pregnancy (per woman) | 30 per 1000 | 25 per 1000 | RR 0.86 | 1368 | ⊕⊕⊝⊝ |
| Clinical pregnancy rate per woman | 239 per 1000 | 251 per 1000 | RR 1.05 | 1330 | ⊕⊕⊝⊝ |
| Miscarriage rate per woman | 47 per 1000 | 56 per 1000 | RR 1.20 | 970 | ⊕⊕⊝⊝ |
| Incidence of OHSS per woman | 22 per 1000 | 33 per 1000 | RR 1.48 | 1565 | ⊕⊝⊝⊝ |
| * 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). | |||||
| GRADE Working Group grades of evidence | |||||
| aDowngraded one level for imprecision around the absolute effect | |||||
| Human menopausal gonadotrophin or highly purified human menopausal gonadotrophin versus urinary follicle stimulating hormone for ovulation induction in women with polycystic ovarian syndrome | |||||
| Patient or population: women with polycystic ovarian syndrome (PCOS) undergoing ovulation induction Comparison: urinary follicle stimulating hormone (uFSH) | |||||
| Outcomes | Anticipated absolute effects * (95% CI) | Relative effect | No of Participants (studies) | Quality of the evidence | |
| Risk with uFSH | Risk with HMG or HP‐HMG | ||||
| Live birth rate per woman | 179 per 1000 | 230 per 1000 | RR 1.28 (0.65 to 2.52) | 138 | ⊕⊝⊝⊝ |
| Incidence of multiple pregnancy (per woman) | 23 per 1000 | 48 per 1000 | RR 2.13 | 161 | ⊕⊝⊝⊝ |
| Clinical pregnancy rate per woman | 203 per 1000 | 266 per 1000 | RR 1.31 | 102 | ⊕⊝⊝⊝ |
| Miscarriage rate per woman | 82 per 1000 | 27 per 1000 | RR 0.33 | 98 | ⊕⊝⊝⊝ |
| Incidence of OHSS per woman | No events c | 4/28 c | RR 7.07 | 53 | ⊕⊝⊝⊝ |
| * 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). | |||||
| GRADE Working Group grades of evidence | |||||
| aDowngarded two levels for serious imprecision around the absolute effect (wide CI and small sample size) d Downgraded one level for inconsistent definition or for the lack of definition of OHSS; two of four studies did not report this outcome | |||||
| Gonadotrophins compared to continued clomiphene citrate for ovulation induction | ||||||
| Patient or population: anovulatory women with clomiphene citrate‐failure | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | No of Participants | Quality of the evidence | Comments | |
| Risk with continued CC | Risk with gonadotrophins | |||||
| Live birth rate per woman | 413 per 1000 | 512 per 1000 | RR 1.24 | 661 | ⊕⊕⊕⊝ | |
| Incidence of multiple pregnancy per woman | 24 per 1000 | 21 per 1000 | RR 0.89 | 661 | ⊕⊕⊕⊝ | |
| Clinical pregnancy rate per woman | 446 per 1000 | 584 per 1000 | RR 1.31 | 661 | ⊕⊕⊕⊝ | |
| Miscarriages per woman | 33 per 1000 | 73 per 1000 | RR 2.23 | 661 | ⊕⊕⊝⊝ | There may be little or no difference when expressed per clinical pregnancy |
| Incidence of OHSS per woman | 0 per 1000 | 0 per 1000 | not estimable | 661 | ⊕⊕⊝⊝ | |
| *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). | ||||||
| GRADE Working Group grades of evidence | ||||||
| a Downgraded one level for risk of bias – no blinding performed b Downgraded one level for imprecision in result c Downgraded one level for inconsistency in outcome, i.e. there were more clinical pregnancies in the gonadotrophin group; there may be little or no difference when expressing miscarriage per clinical pregnancy | ||||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Live birth rate per woman by urinary gonadotrophins Show forest plot | 5 | 505 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.21 [0.83, 1.78] |
| 1.1 rFSH versus HMG | 3 | 409 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.04 [0.68, 1.57] |
| 1.2 rFSH versus uFSH | 2 | 96 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.66 [0.95, 7.43] |
| 2 Live birth rate per woman by sponsor Show forest plot | 5 | 505 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.21 [0.83, 1.78] |
| 2.1 Ferring | 3 | 409 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.04 [0.68, 1.57] |
| 2.2 unknown | 2 | 96 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.66 [0.95, 7.43] |
| 3 Multiple pregnancy per woman by urinary gonadotrophins Show forest plot | 8 | 1368 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.86 [0.46, 1.61] |
| 3.1 rFSH versus HMG | 3 | 409 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.17 [0.49, 2.79] |
| 3.2 rFSH versus uFSH | 5 | 959 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.63 [0.25, 1.59] |
| 4 Multiple pregnancy per woman by sponsor Show forest plot | 8 | 1368 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.04 [0.67, 1.60] |
| 4.1 Ferring | 3 | 409 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.17 [0.49, 2.79] |
| 4.2 MSD ‐ Organon | 1 | 172 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.23 [0.68, 2.23] |
| 4.3 Merck ‐ Serono | 2 | 357 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.51 [0.14, 1.80] |
| 4.4 Unknown | 2 | 430 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.93 [0.19, 4.49] |
| 5 Clinical pregnancy rate per woman by urinary gonadotrophins Show forest plot | 8 | 1330 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.05 [0.88, 1.27] |
| 5.1 rFSH versus HMG | 3 | 409 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.19 [0.81, 1.77] |
| 5.2 rFSH versus uFSH | 5 | 921 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.01 [0.82, 1.25] |
| 6 Incidence of multiple pregnancy per clinical pregnancy by urinary gonadotrophins Show forest plot | 8 | 315 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.75 [0.43, 1.32] |
| 6.1 rFSH versus HMG | 3 | 81 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.99 [0.47, 2.09] |
| 6.2 rFSH versus uFSH | 5 | 234 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.57 [0.24, 1.35] |
| 7 Miscarriage rate per woman by urinary gonadotrophins Show forest plot | 7 | 970 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.20 [0.71, 2.04] |
| 7.1 rFSH versus HMG | 2 | 335 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.95 [0.24, 3.70] |
| 7.2 rFSH versus uFSH | 5 | 635 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.26 [0.71, 2.23] |
| 8 Incidence of OHSS per woman by urinary gonadotrophins Show forest plot | 10 | 1565 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.48 [0.82, 2.65] |
| 8.1 rFSH versus HMG | 3 | 409 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.11 [0.39, 3.20] |
| 8.2 rFSH versus uFSH | 7 | 1156 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.67 [0.82, 3.39] |
| 9 Total gonadotrophin dose per woman (IU) by urinary gonadotrophins Show forest plot | 6 | 1046 | Mean Difference (IV, Fixed, 95% CI) | ‐105.44 [‐154.21, ‐56.68] |
| 9.1 rFSH versus HMG | 2 | 335 | Mean Difference (IV, Fixed, 95% CI) | ‐283.94 [‐449.10, ‐118.78] |
| 9.2 rFSH versus uFSH | 4 | 711 | Mean Difference (IV, Fixed, 95% CI) | ‐88.40 [‐139.44, ‐37.36] |
| 10 Total duration of stimulation per woman (days) by urinary gonadotrophins Show forest plot | 6 | 1122 | Mean Difference (IV, Fixed, 95% CI) | ‐0.66 [‐1.04, ‐0.28] |
| 10.1 rFSH versus HMG | 2 | 335 | Mean Difference (IV, Fixed, 95% CI) | ‐2.28 [‐3.49, ‐1.07] |
| 10.2 rFSH versus uFSH | 4 | 787 | Mean Difference (IV, Fixed, 95% CI) | ‐0.49 [‐0.88, ‐0.09] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Live birth rate per woman Show forest plot | 3 | 138 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.28 [0.65, 2.52] |
| 2 Multiple pregnancy per woman Show forest plot | 4 | 161 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.13 [0.51, 8.91] |
| 3 Multiple pregnancy per clinical pregnancy Show forest plot | 3 | 22 | Risk Ratio (M‐H, Fixed, 95% CI) | 4.2 [0.21, 83.33] |
| 4 Clinical pregnancy rate per woman Show forest plot | 3 | 102 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.31 [0.66, 2.59] |
| 5 Miscarriage rate per woman Show forest plot | 2 | 98 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.33 [0.06, 1.97] |
| 6 Incidence of OHSS per woman Show forest plot | 2 | 53 | Risk Ratio (M‐H, Fixed, 95% CI) | 7.07 [0.42, 117.81] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Live birth rate per woman Show forest plot | 1 | 661 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.24 [1.05, 1.46] |
| 2 Multiple pregnancy per woman Show forest plot | 1 | 661 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.89 [0.33, 2.44] |
| 3 Multiple pregnancy (per clinical pregnancy) Show forest plot | 1 | 340 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.68 [0.25, 1.84] |
| 4 Clinical pregnancy per woman Show forest plot | 1 | 661 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.31 [1.13, 1.52] |
| 5 Miscarriages per woman Show forest plot | 1 | 661 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.23 [1.11, 4.47] |
| 6 Incidence of OHSS per woman Show forest plot | 1 | 661 | Risk Difference (M‐H, Fixed, 95% CI) | 0.0 [‐0.01, 0.01] |
