Gonadotrophins for idiopathic male factor subfertility

Types of studies

Randomised controlled trials of fertility treatments in which gonadotrophins were administered for the treatment of idiopathic male factor subfertility compared with placebo or no treatment. Trials in which couples received one or more cycles of fertility treatment after randomisation were included. We excluded quasi‐randomised trials (e.g. alternate randomisation, reference to hospital number or to date of birth), cross‐over trials if data before the cross‐over were not available and trials that did not report on outcomes of importance to the review.

Types of participants

Men with idiopathic male factor subfertility diagnosed by subnormal semen parameters as defined by the World Health Organisation (WHO) criteria, including oligospermia, teratospermia, asthenospermia and non‐obstructive azoospermia, or as defined by the study author.

Subnormal sperm parameters defined by WHO 1999 include (WHO 1999):

• sperm concentration ≤ 20 million sperm/mL;

• motility ≤ 50% motile sperm; and

• normal morphology ≤ 30%.

Types of interventions

Systemic administration of any type of FSH (urinary, purified or highly purified or recombinant) compared with placebo or no treatment.

Types of outcome measures

Electronic searches

We searched the following sources, without restriction by language or publication status and in consultation with the Cochrane Menstrual Disorders and Subfertility Group Trials Search Co‐ordinator:

• The Menstrual Disorders & Subfertility Group's Specialised Register of controlled trials (14 January 2013) for any trials with FSH or LH administration to subfertile men in the title, abstract or keywords section. See the Review Group for more details on the make‐up of the Specialised Register.

  • See Appendix 1.

• The Cochrane Central Register of Controlled Trials (CENTRAL; The Cochrane Library, Issue 12 of 12, 2012).

  • See Appendix 2.

• The following electronic databases, using Ovid software:

• • MEDLINE In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE(R) Daily and Ovid MEDLINE (1946 to 14 January 2013).

    • See Appendix 3.

  • EMBASE (1980 to week 2 2013).

    • See Appendix 4.

  • PsycINFO (1806 to January week 2 2013).

    • See Appendix 5.

The MEDLINE search was combined with the Cochrane highly sensitive search strategy for identifying randomised trials, which appears in the Cochrane Handbook for Systematic Reviews of Interventions (version 5.0.2; Chapter 6, 6.4.11).

The EMBASE search was combined with trial filters developed by the Scottish Intercollegiate Guidelines Network (SIGN) http://www.sign.ac.uk/mehodology/filters.html#random.

• ClinicalTrials.gov (19 January 2013) for federally and privately supported ongoing and registered trials and clinical trials including studies sponsored by the National Institutes of Health, other US federal agencies and private industry.

• The World Health Organisation International Trials Registry Platform (19 January 2013) at http://www.who.int/trialsearch/Default.aspx.

• The metaRegister of Controlled Trials, a major international searchable database of ongoing randomised controlled trials in all areas of healthcare, built by combining registers held by public, charitable and commercial sponsors of trials (http://controlled‐trials.com/mrct/). This database contains the National Research Register (NRR), entries from the Medical Research Council's Clinical Trials Register and details on reviews in progress collected by the NHS Centre for Reviews and Dissemination, as well as others.

• OpenSigle and OpenGrey (http://www.opengrey.eu/) for Grey literature from Europe (19 January 2013).

• The Database of Abstracts of Reviews of Effects (19 January 2013).

Searching other resources

The citation lists of relevant publications, review articles and abstracts of major scientific meetings and the bibliographies of included and excluded trials were also searched for additional trials.

Selection of studies

The search strategy described above was used to obtain titles and, where possible, abstracts of studies potentially relevant to the review. The titles and abstracts were screened by one review author (AMA), who discarded studies that were clearly ineligible but aimed to be overly inclusive rather than risk losing relevant studies. Two review authors (AMA, HGAl) independently assessed studies for inclusion in accordance with the mentioned criteria. Disagreements were resolved by consensus or through arbitration by the third review author (AMAS). Further information was sought from the authors when papers contained insufficient information to allow a decision to be made regarding eligibility.

Data extraction and management

All data extraction was performed independently by two of the three review authors (AMA, HGAI), using forms designed in accordance with Cochrane guidelines. The third review author (AMAS) resolved discrepancies. Extracted trial data included the following (Characteristics of included studies).

• Method of randomisation: randomly assigned by computer, random number tables or drawing of lots, or method not clear (e.g. stated but not further described). Quasi‐randomised trials were excluded from the review (e.g. allocation by hospital number or date of birth).

• Concealment of allocation: details recorded.

• Presence or absence of blinding to treatment allocation.

• Duration and type of follow‐up.

• Number of participants recruited, randomly assigned, excluded, analysed or lost to follow‐up.

• Location of trial: single‐centre or multi‐centre.

• Timing of trial.

• Whether an intention‐to‐treat analysis was done.

• Source of funding.

• Criteria for including participants and assessing outcomes.

Additional information on trial methodology and/or original trial data were sought from the authors of trials that appeared to meet the eligibility criteria but had aspects of methodology that were unclear, or for which the data were provided in a form unsuitable for meta‐analysis.

Assessment of risk of bias in included studies

All assessments of the risk of bias of trials were performed independently by two of the three review authors (AMA, AMAS), using the Cochrane risk of bias assessment tool (www.cochrane‐handbook.org). A third review author resolved discrepancies (HGAI).

Measures of treatment effect

Statistical analysis was performed in accordance with the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We used the number of events in the control and intervention groups of each study to calculate Peto odds ratios (ORs).

Live birth and pregnancy are considered positive consequences of treatment; therefore a higher proportion of women achieving these outcomes is considered a benefit. However, the outcome of adverse effects is a negative consequence; therefore higher numbers are considered detrimental. This needs to be taken into consideration when the summary graphs are viewed.

Studies varied in length of follow‐up after completion of gonadotrophin therapy. Because FSH therapy has a "short‐term effect" on fertility, we restricted our analysis to events that occurred spontaneously during FSH treatment and within 3 months after completion of FSH therapy, and in case of treatment of the female partner with IUI/IVF/ICSI, we restricted our analysis to the first cycle only.

Unit of analysis issues

To prevent unit of analysis errors when the outcome was presented for several time periods in the same trial (multiple treatment cycles), we included data only from the first time period.

Dealing with missing data

Data were extracted and analysed to allow an intention‐to‐treat analysis, defined as including in the denominator all originally randomly assigned participants. In addition, all participants were analysed according to their allocated groups, regardless of treatment eligibility, compliance or treatment given.

Assessment of heterogeneity

Homogeneity of the data from included trials was assessed by visual inspection of the outcomes tables and by use of the Chi² test (✗² test) for heterogeneity with a 10% level of statistical significance; a P value of 0.1 was selected as the cut‐off point for rejection of the null hypothesis of study homogeneity to limit type II errors. In addition, heterogeneity was quantified using the I² statistic, which describes the percentage of variability in effect estimates that is due to heterogeneity rather than sampling error (chance). An I² value > 50% may be considered to represent substantial heterogeneity.

Assessment of reporting biases

In the event that at least 10 included studies provided data for any given analysis, we planned to investigate publication bias by means of a funnel plot. If publication bias was suspected or detected by the aforementioned method, it would be confirmed or rejected using Egger's regression test.

Data synthesis

Although all included trials might be statistically homogeneous, differences in clinical parameters may be considerable (clinical heterogeneity). These differences were taken into account when the pooled results were analysed and interpreted. Clinical heterogeneity in infertility cannot be avoided because most centres use their own "materials and methods", which can vary along several parameters. If trials met the inclusion criteria and provided the same intervention for male factor subfertility, we considered it appropriate to pool their results.

We combined the data using a fixed effect model to calculate pooled Peto odds ratios (ORs) and 95% confidence intervals (CIs).

Subgroup analysis and investigation of heterogeneity

Subgroup analysis was performed according to treatment modality, whereby pregnancies were divided into spontaneous pregnancies resulting from natural intercourse and pregnancies that followed IUI or assisted reproduction (IVF/ ICSI) Figure 1. We also performed a subgroup analysis that was restricted to trials that enrolled 'normal' female partners (Figure 2). We planned to explore clinical and methodological differences between the studies if substantial statistical heterogeneity was noted.

Sensitivity analysis

We did not plan any sensitivity analyses.

Design and setting

All studies were parallel‐group randomised controlled trials.

Duration of follow‐up: 
 The duration of follow‐up varied between studies. Couples were followed up only during the treatment period (12 weeks) in one study (Baccetti 2004). In another study, the follow‐up period was extended by another 12 weeks after completion of therapy (Knuth 1987). In the Kamischke 1998 trial, participants were followed‐up during the treatment period, then for another 12 weeks, during which assessment examinations were done, and then for another 3 months, during which pregnancies were recorded. In the Foresta 2005 trial, couples were followed up during treatment and for 3 months after completion of treatment; then, depending on their sperm count, participants were either allocated directly or re‐randomly assigned into groups to receive either three cycles of IUI or one cycle of IVF/ICSI. One trial followed up with participants through a maximum of six IUI cycles (Matorras 1997). In the remaining trial, the duration of follow‐up was not reported (Paradisi 2006).

Sources of funding: 
 Two trials did not report sources of funding (Foresta 2005; Matorras 1997). The trial of Baccetti 2004 was funded by a grant from the Italian Ministry for Universities and Technological Research (2000); that of Kamischke 1998 was supported in part by the Federal Health Ministry (Bonn), the Deutsche Forschungsgemeinschaft (DFG, Bonn) and Ares‐Serono (Unterschleissheim, Germany); and in the study of Knuth 1987, human menopausal gonadotrophin (HMG)/human chorionic gonadotrophin (HCG) and placebo preparations were offered by Serono Co. The remaining trial was funded by Serono Co (supplied r‐hFSH and placebo injections) (Paradisi 2006).

Participants

Male partners

All trials included male partners with idiopathic subfertility based on subfertile semen parameters. Male subfertility was diagnosed in a different way in each trial: diagnostic criteria according to WHO 1987 criteria (Matorras 1997), WHO 1992 criteria (Kamischke 1998) or WHO 1999 criteria (Baccetti 2004); sperm count between 0.1 and 10 Mil/mL (Knuth 1987) and sperm count < 10 million/mL on at least three separate occasions (Foresta 2005); or poor semen quality in the form of moderate to severe oligoasthenozoospermia (range 1 to 15 10⁶/mL sperm concentration) (Paradisi 2006). The age of participants was comparable in the four trials (details in Characteristics of included studies).

Female partners

Female partners had normal fertility (as reported by authors) in five trials (Baccetti 2004; Foresta 2005; Kamischke 1998; Knuth 1987; Paradisi 2006). One trial reported female causes of subfertility in 50% of couples (14.9% tubal factor, 12.2% endometriosis, 14.9% ovulatory disorders, 2.0% hyperprolactinaemia and 6.0% mixed causes) (Matorras 1997).

Interventions

Male partners

HMG/HCG treatment was used in one study (Knuth 1987), purified or highly purified FSH was used in two studies (Baccetti 2004; Matorras 1997) and recombinant FSH was administered in three studies (Foresta 2005; Kamischke 1998; Paradisi 2006). HMG/HCG was given at a dose of 150 IU HMG three times a week and 2500 IU HCG twice a week for 13 weeks (Knuth 1987). Purified FSH was given at a dose of 150 IU/day for 12 weeks (Baccetti 2004; Matorras 1997). Recombinant FSH was given at a dose of 150 IU daily for 12 weeks (Kamischke 1998), 100 IU r‐hFSH IM on alternate days for 3 months (Foresta 2005) or 300 IU r‐hFSH SC daily for at least 4 months (Paradisi 2006). FSH therapy was compared with placebo in three trials (Kamischke 1998; Knuth 1987; Paradisi 2006) and with no treatment in the control group in the other three trials (Baccetti 2004; Foresta 2005; Matorras 1997).

Female partners

Female partners received no treatment in one trial (Knuth 1987), and ICSI was performed after ovarian stimulation, follicular monitoring and aspiration in another trial (Baccetti 2004). The trial of Kamischke 1998 did not state the exact numbers of participants who underwent treatment with assisted reproductive technologies (ART), although pregnancies were reported to occur, besides spontaneously, after IUI, IVF or ICSI. In the trial of Foresta 2005, no intervention was applied with female partners in the first 6 months (3 months of therapy and 12 weeks of follow‐up) of the trial; then participants were re‐randomly assigned to IVF, ICSI or IUI, depending on the sperm count. The re‐randomisation did not influence the data included in this review, but it invalidated the results of the last 3‐month period of the study because it interfered with the first randomisation and rendered the chances of pregnancy due to the treatment studied unequal between the two groups. Moreover, ART was used for longer than 3 months after completion of FSH, which we considered the limit of attributing pregnancies to treatment in our analysis. The trial by Matorras et al. (Matorras 1997) reported using ovarian stimulation, follicular monitoring and HCG administration before participants underwent IUI. The remaining trial did not report whether female partners received any fertility treatment or were monitored during the course of treatment of their male partners (Paradisi 2006).

Outcomes

• Live birth rate per couple randomly assigned.

• • Only one of the six included trials reported live birth rates (Paradisi 2006).

• Pregnancy rate per couple randomly assigned

• • The pregnancy rate was reported in all included trials. Pregnancy was diagnosed by ultrasound (US) 6 weeks after embryo transfer (Baccetti 2004; Matorras 1997), by US and beta‐HCG concentration increase (Kamischke 1998) or by beta‐HCG plasma level (Foresta 2005); one study did not mention the method of diagnosing pregnancy (Knuth 1987), and Paradisi 2006 reported only that all pregnancies went to term successfully.

• Miscarriage rate per pregnancy

• • None of the six trials reported on the miscarriage rate.

• Adverse events

• • Only one trial clearly addressed the side effects of the drugs used (Knuth 1987), another trial reported on adverse events in general (Matorras 1997) and a third trial reported that no adverse events occurred (Paradisi 2006).

Other outcomes

Other surrogate outcomes such as fertilisation rates and effects on sperm count, motility and morphology were not included in the analysis, as we excluded studies that reported only such surrogate outcomes without reporting pregnancy rates, and because such outcomes are intermediate ones that are of secondary importance compared with the participant‐oriented outcome (e.g. live birth or pregnancy rate).

Results of the search

A total of fifteen potentially eligible controlled trials were identified. Electronic searches identified 14 controlled trials that used gonadotrophin treatment for idiopathic male subfertility, and another trial was obtained via personal communication with the author (Foresta 2005) that was published later. All trials were thoroughly appraised for their eligibility to be included in the review, and their risk of bias was assessed.

Included studies

Six trials with 456 participants met the inclusion criteria (Baccetti 2004 with 44 participants, Foresta 2005 with 128 participants, Kamischke 1998 with 67 participants, Knuth 1987 with 39 participants, Matorras 1997 with 148 participants, Paradisi 2006 with 30 participants). Details of each study are provided in the Characteristics of included studies.

Excluded studies

Nine trials were excluded because they did not fulfil our inclusion criteria. One trial was excluded because of quasi‐randomisation (Ashkenazi 1999). One RCT was excluded because the authors compared two gonadotrophin regimens in hypogonadotrophic men (Bouloux 2003). Two RCTs were excluded because outcomes of the study did not include pregnancy rates (Foresta 1998; Foresta 2002). We contacted the authors of these trials to ask for data on pregnancy rates that were not published but received confirmation on the absence of such data. Another RCT was excluded because it was a partial cross‐over study, and because pregnancy rates were not reported (Ben‐Rafael 2000). The other four trials were excluded because all were non‐randomised (Caroppo 2003; Dirnfeld 2000; Iacono 1996; Thomalla‐Sauter 2001).

Allocation

Half the trials reported using a proper method of randomisation (Foresta 2005; Matorras 1997; Paradisi 2006). Two trials reported "true randomisation" by a third party without reporting any further details of the method (Baccetti 2004; Kamischke 1998). One trial was reported to be a randomised trial but did not report the method of randomisation used (Knuth 1987).

Allocation concealment was considered adequate in the three trials that used third party randomisation (Baccetti 2004; Foresta 2005; Kamischke 1998), but it was not reported in the remaining trials (Knuth 1987; Matorras 1997; Paradisi 2006).

Blinding

Half the trials were considered to be properly blinded (Kamischke 1998; Knuth 1987; Paradisi 2006), and one trial was open label (Matorras 1997). It was not clear from the description whether blinding was used in the remaining two trials (Baccetti 2004; Foresta 2005).

Incomplete outcome data

Three studies reported no withdrawals or dropouts in assessment of pregnancy rates (Baccetti 2004; Knuth 1987; Paradisi 2006). In the remaining studies, dropouts were variable (see Characteristics of included studies for further details).

We analysed the results on an intention‐to‐treat basis, adding the 16 excluded and dropout cases in the trial of Foresta 2005, the four dropout cases along with their corresponding groups in the Kamischke 1998 trial and the 12 excluded cases from the trial by Matorras 1997. The one excluded pregnant case in the Kamischke 1998 trial has been added to the treatment group (upon contacting the author, we were informed that the case belonged in the treatment group).

Selective reporting

None of the included trials had protocols available for review, but no differences were noted between the methods and results sections of the respective trial reports.

Other potential sources of bias

Only three of the included trials (Knuth 1987; Matorras 1997; Paradisi 2006) clearly reported homogeneity among baseline characteristics of included populations.

Gonadotrophins versus placebo/no treatment for participants with idiopathic male subfertility