Outcome reporting across randomized controlled trials evaluating potential treatments for male infertility: a systematic review.

Infertility affects 50 million couples globally ( Martinez et al. , 2012 ; Vander Borght and Wyns, 2018 ). Male factor infertility affects up to 18 million men worldwide ( Winters and Walsh, 2014 ; Agarwal et al. , 2015 ) and is recognized as a contributing factor in up to one-third of cases ( Thonneau et al. , 1991 ; Agarwal et al. , 2015 ; Tamrakar and Bastakoti, 2019 ). Treatments with the potential to reduce this health burden require robust evaluation. When assessing new treatments, randomized controlled trials (RCTs) are considered the gold-standard method to determine the efficacy and safety of potential treatments ( Liberati et al. , 2009 ). However, despite their potentially robust design, methodology and conduct, RCTs are only as meaningful as the outcomes they collect and report ( Ioannidis et al. , 2014 ; Duffy et al. , 2019 ). Complex issues, including a failure to consider the perspectives of people with fertility problems when selecting outcomes, variations in outcome definitions and measurement instruments as well as outcome reporting bias can make the selection, collection and reporting of outcomes challenging. The unique nature of male infertility research can add further complexity as outcomes will often need to consider three research participants, namely the male, his female partner or gestational carrier, and their subsequent offspring. Little is known about outcome reporting in male infertility clinical trials. To understand the heterogeneity in outcome reporting of RCTs in male infertility, and provide a basis for more consistent reporting to the highest possible standards, we undertook a systematic review of the outcomes and outcome measures reported by the 100 largest RCTs published over the last 10 years. Reporting on the outcomes, outcome measures and consistency of these outcomes across trials will enable us to identify how outcome reporting could be standardized in future trials. This will allow researchers to better understand the true efficacy of interventions assessed in RCTs to address male infertility.

This work received no funding.

We identified 1620 records. After excluding 12 duplicate records, 1608 titles and abstracts were screened to identify RCTs evaluating interventions for male infertility ( Fig. 1 ). We excluded 1411 records as they were either non-randomized studies, systematic reviews, did not report an intervention for male infertility or did not report on a male infertility cohort. Two independent reviewers evaluated the remaining 197 potentially relevant trials of which 175 were deemed to be relevant. From these, the largest 100 RCTs reporting data from 24 542 men (range: 80–2772 men) were used to identify and report outcomes ( Abdel-Maguid and Othman, 2010 ; Fang et al. , 2010 ; Fayez et al. , 2010 ; Kovacic et al. , 2010 ; Abdel-Meguid et al. , 2011 ; Azadi et al. , 2011 ; Balaban et al. , 2011 ; Figueira Rde et al. , 2011 ; Hafeez et al. , 2011 ; Safarinejad, 2011 ; Safarinejad et al. , 2011 ; Selice et al. , 2011 ; Turhan et al. , 2011 ; Wilding et al. , 2011 ; Amirzargar et al. , 2012 ; Colacurci et al. , 2012 ; El-Khayat et al. , 2012 ; Lee et al. , 2012 ; Mansour Ghanaie et al. , 2012 ; Parmegiani et al. , 2012 ; Rago et al. , 2012 ; Safarinejad et al. , 2012 ; Velaers et al. , 2012 ; Azizollahi et al. , 2013 ; De Vos et al. , 2013 ; Gopinath et al. , 2013 ; Kang et al. , 2013 ; Leandri et al. , 2013 ; Majumdar and Majumdar, 2013 ; Pan et al. , 2013 ; Worrilow et al. , 2013 ; Akin et al. , 2014 ; Karamahmutoglu et al. , 2014 ; Kolahdooz et al. , 2014 ; Moslemi Mehni et al. , 2014 ; Nematollahi-Mahani et al. , 2014 ; Pourmand et al. , 2014 ; Raigani et al. , 2014 ; Romany et al. , 2014 ; Wang et al. , 2014 ; Calogero et al. , 2015 ; Cyrus et al. , 2015 ; Ding et al. , 2015 ; ElSheikh et al. , 2015 ; Farrag et al. , 2015 ; Guo et al. , 2015 ; Haje and Naoom, 2015 ; Hou et al. , 2015 ; Peivandi et al. , 2015 ; Sikka et al. , 2015 ; Youssef and Abdalla, 2015 ; Hosseini et al. , 2016 ; Jin et al. , 2016 ; Nasr Esfahani et al. , 2016 ; Pan et al. , 2016 ; Park et al. , 2016 ; Bryniarski et al. , 2017 ; Guo et al. , 2017 ; Milardi et al. , 2017 ; Qu et al. , 2017 ; Rosety et al. , 2017 ; Taiyeb et al. , 2017 ; Hajizadeh Maleki and Tartibian, 2017a , b ; Babak et al. , 2018 ; Blomberg Jensen et al. , 2018 ; Bodin et al. , 2018 ; Busetto et al. , 2018 ; Habous et al. , 2018 ; Hajizadeh Maleki and Tartibian, 2018 ; Ketabchi and Salajegheh, 2018 ; Ketabchi et al. , 2018 ; Nasimi Doost Azgomi et al. , 2018 ; Sun et al. , 2018 ; Tsounapi et al. , 2018 ; Almekaty et al. , 2019 ; Chen et al. , 2019 ; De Geyter et al. , 2019 ; Hajizadeh Maleki et al. , 2019 ; Kızılay and Altay, 2019 ; Lin et al. , 2019 ; Mangoli et al. , 2019 ; Miller et al. , 2019 ; Tehrani et al. , 2019 ; Yetkinel et al. , 2019 ; Yu et al. , 2019 ; Zhao et al. , 2019 ; Chen et al. , 2020 ; Degirmenci et al. , 2020 ; Eslamian et al. , 2020 ; Hajizadeh Maleki and Tartibian, 2020 ; Hasanen et al. , 2020 ; Huang et al. , 2020 ; Joseph et al. , 2020 ; Karimi et al. , 2020 ; Kopets et al. , 2020 ; Liu et al. , 2020 ; Schisterman et al. , 2020 ; Bozhedomov et al. , 2021 ; Salas-Huetos et al. , 2021 ). PRISMA flow diagram outlining number of studies identified from our database search, number remaining after screening and the number of studies reporting interventions for male infertility included in our analysis. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta‐Analyses. Seventy-nine different treatments were reported across the 100 RCTs ( Table I ). These included trials reporting vitamin or dietary supplements or nutraceuticals (n = 18), surgical procedures (n = 18) and sperm selection or modification techniques (n = 22). Characteristics of the 100 largest trials included in this review evaluating interventions for male infertility. Multiarm trial. One hundred and four outcomes were reported across the included trials ( Tables II and III ). Primary outcomes reported in the 100 largest randomized trials evaluating interventions for male infertility. Secondary outcomes reported in the 100 largest randomized trials evaluating interventions for male infertility. Thirty-six different primary outcomes were reported by 47 trials with 13 of these 47 trials (28%) reporting a definition of these outcomes. Commonly reported primary outcomes included semen quality (16 trials; 34%), pregnancy rate (13 trials; 28%) and live birth (4 trials; 9%). Ninety-six trials reported 89 different secondary outcomes. Reported secondary outcomes were heterogeneous and included semen quality (52 trials; 54%), pregnancy rate (39 trials; 41%), pregnancy loss (9 trials, 10%) and live birth (9 trials; 10%) ( Table III ). Primary and secondary outcomes reported by the 25 largest RCTs are outlined in Table IV . Detailed primary and secondary outcomes reported in the largest 25 randomized trials revaluating interventions for male infertility. ● denotes if outcome included in trial. *Semen analysis includes: semen volume; sperm concentration; sperm motility; sperm morphology; sperm count; and total motile sperm count. Pregnancy outcomes includes: caesarean section; pre-eclampsia; gestational diabetes; gestational age at delivery; birth weight; small for gestational age; severe postpartum maternal morbidity (including postpartum haemorrhage, anaemia requiring transfusion, sepsis, seizure, HELLP [haemolysis, elevated level of liver enzymes, low platelet count] syndrome, and pre-eclampsia with pulmonary oedema), major neonatal complications (including structural malformations, chromosomal anomalies, bronchopulmonary dysplasia, necrotizing enterocolitis, severe intraventricular haemorrhage, periventricular leukomalacia, and retinopathy of prematurity), still-birth and neonatal death. USS, ultrasound scan. Nine trials defined live birth as a primary or secondary outcome in two different ways: birth >37 weeks’ gestation; and birth <37 weeks’ gestation. The remaining seven trials did not define this term. Pregnancy rate was reported by 51 trials as either a primary or secondary outcome, with 12 different definitions used by 21 trials. The remaining 26 trials did not define pregnancy rate, and 4 definitions were unclear. Definitions varied greatly, from a threshold of serum hCG >25 IU/l and the presence of a gestational sac on ultrasound scan (USS) to a viable foetus on transvaginal USS ( Table V ). Variation in outcome reporting definitions across the 100 largest trials evaluating interventions for male infertility. Three outcomes were selected to demonstrate variation across studies in how common outcomes were defined differently. The outcomes selected for this were semen analysis, pregnancy and live birth. PoC, product of conception; USS, ultrasound scan; WHO, World Health Organization. Semen quality was reported by 75 trials as either a primary or secondary outcome and there was a comparatively higher level of consensus between trials. A total of 57/75 trials defined these by the World Health Organization (WHO) criteria, having used either WHO 2010 (n = 32), WHO 1999 (n = 18), WHO 1992 (n = 3), WHO 1999 and 1992 (n = 1), in three the WHO semen analysis edition was not specified. Of the remaining studies, one used the Kruger strict morphology ( Ketabchi et al. , 2018 ) and the remaining 17 trials did not define this outcome ( Table V ). Not all of the studies included in our review used the most up to date edition of the WHO criteria available when conducting their trial. Studies defining semen analysis parameters using WHO 1992 criteria were commenced in 2012 and 2006 and could have utilized the WHO 1999 criteria when conducting the trial ( Selice et al. , 2011 ; Sikka et al. , 2015 ). A similar issue was identified with some studies defining semen analysis criteria using WHO 1999 , where the trial was commenced in 2013 or 2016 after the introduction of WHO 2010 ( Haje and Naoom, 2015 ; Hosseini et al. , 2016 ; Tehrani et al. , 2019 ). Although these trials may initially appear to use an outdated version of the WHO semen analysis manual, their design may have occurred prior to the publication of an updated WHO criteria. Deviation from the initial analysis plan, potentially using two different WHO criteria or favouring one criterion in particular, may have been considered a violation of the trial protocol. To achieve consistent outcome reporting, should these trials have used new WHO criteria, they may no longer be comparable to older trials evaluating similar interventions for male infertility.

A.P.—chairman of external scientific advisory committee of Cryos International Denmark ApS, member of the scientific advisory board for Cytoswim LDT and ExSeed Health. Guest lecture at the ‘Insights for Fertility Conference’. M.v.W.—holds a ZON-MW research grant.

A protocol was developed prior to commencing the review and included clearly defined objectives, including search criteria, study selection criteria and extraction of data ( Supplementary Data ). We followed the reporting guidelines for systematic reviews of RCTs, as outlined by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement ( Minebois et al. , 2017 ). The objective of our review was to characterize outcome reporting across RCTs evaluating interventions for male infertility. Our main outcome of interest was primary outcome reporting in these trails and the definition of this outcome. RCTs were identified by searching the Cochrane Register of Controlled Trials (CENTRAL) for RCTs published between 1 January 2010 and 24 July 2021. CENTRAL is populated by the Cochrane Collaboration by regularly searching the Cumulative Index to Nursing and Allied Health Literature (CINAHL), EMBASE, MEDLINE and PsycINFO ( Supplementary Data ). Two authors (M.P.R. and R.A.H.) independently performed the screening of each potentially relevant record, based on the title and abstract and reviewed the full text of each selected study to assess eligibility. Where data were reported twice, such as a conference abstract and peer reviewed paper published at a later date, extracted data from the peer reviewed paper was used. Discrepancies between the authors were resolved through discussion and a consensus being agreed. We included all RCTs which evaluated potential treatments for male factor infertility. We excluded systematic reviews and non-randomized trials. We limited our search to publications written in English. The largest 100 RCTs based on the number of participants were included in our analysis. Using a standardized data extraction form, two authors (M.P.R. and R.A.H.) independently extracted study characteristics, nature of the intervention and both the primary and secondary outcomes reported. We reported the definitions used for commonly reported outcomes, including semen quality, pregnancy rate and live birth, to illustrate how these definitions varied. An outcome was considered to be a primary outcome only if this was clearly started in the method section. Discrepancies between authors were resolved through discussion and a consensus being achieved. A comprehensive inventory of outcomes was developed. We used descriptive statistics to characterize outcome reporting across included RCTs. No risk of bias was undertaken as the scope of this review was to report outcome reporting across RCTs and not to assess the quality of the trials.

M.P.R., R.A.H., V.S. and J.M.N.D. undertook the searches, data extraction and drafted the manuscript. M.P.R., R.A.H., V.S., R.A.A., Y.B., R.P.B., S.K.S., A.P., B.P., R.T.M., A.P., M.v.W., C.M.F., C.N. and J.M.N.D. participated in data analysis and interpretation, preparation of the manuscript and critically revising the paper. C.M.F., C.N. and J.M.N.D. conceived the idea of the manuscript. All authors approved the final version of the manuscript.

Randomized trials reporting on interventions for male factor infertility frequently omit a primary outcome and often report these outcomes differently. This hinders the utility of these trials in how their results can be combined to inform health care professionals’ clinical decision-making and improve patient outcomes. Developing a core outcome set for male infertility trials will help inform how primary outcome measures are selected and reported on and translate into meaningful improvements in patient care.

This systematic evaluation of the literature of RCTs in male factor infertility identified a range of primary and secondary outcomes relevant to male, maternal and neonatal participants. Many trials omitted important information about the primary outcome of the trial and how this was defined. Of the 100 randomized trials included in our review, only 47 clearly stated a primary outcome in their methodology. This lack of clear outcome reporting is not uncommon and has been identified as a problem in other areas, including in IVF, neonatal and endometriosis trials ( Hirsch et al. , 2016 ; Wilkinson et al. , 2016 ; Webbe et al. , 2020 ). RCTs can be challenging to undertake and expensive to conduct; as such there is an ethical imperative to conduct them to the highest possible standards ( Macleod et al. , 2014 ). Less than half of the 100 largest trials included in our review reported a clearly defined primary outcome, which represents a lost opportunity to obtain further robust data to inform clinical decision-making. Where trials reported the same primary outcome, often different measurement tools and endpoints were used to define these, which precludes pooling data from these trials. Even trials with seemingly consistent primary outcome reporting and definitions are not without their limitations. We identified 57 trials using WHO semen analysis methods to report, with primary or secondary outcomes. Although WHO semen analysis is a robustly developed standard, there have now been six different editions, of which three were used in our identified trials, although the most up to date edition was used in the majority of trials at the time these they were conducted ( World Health Organization, 1992 , 1999 , 2010 ). Furthermore, semen is highly variable, even within individuals, which furthers the argument that semen quality may, in itself, not be an informative primary endpoint ( Oshio et al. , 2004 ; Castilla et al. , 2006 ). This is demonstrated in some of the trials included in our review, which showed that improved semen quality did not correlate with improved pregnancy outcomes. This finding challenges the assumption that improved pregnancy outcomes are always associated with improved semen quality and not achieved through other factors including the population under study and the intervention used. Once such example is Huang et al. (2020) , who demonstrated that folic acid supplementation was only effective at improving semen quality and pregnancy outcomes in a subgroup of patients with the homozygous polymorphism of the MTHFR 677 gene, while all other MTHFR polymorphisms studied showed no effect. Variable response in semen quality and pregnancy outcome was demonstrated by Hajizadeh Maleki et al. (2020) who investigated high-intensity interval training, reporting both semen parameters and live births. Patients categorized as asthenozoospermic, asthenoteratozoospermic, oligospermic or oligoasthenozoospermic demonstrated significantly improved semen quality following their exercise regime. Analysis of pregnancy outcomes in these cohorts, however, did not reveal a significant increase in live births. Another trial included in this review ( Haje and Naoom, 2015 ) reported the impact of tamoxifen and l -carnitine on semen parameters and pregnancy outcomes. Although semen parameters, including sperm count, sperm motility and sperm morphology, were found to be improved in men receiving tamoxifen or tamoxifen with l -carnitine compared to placebo or l -carnitine only, these improvements did not translate into a significant increase in pregnancy rate. In addition to outcome selection, inconsistent outcome reporting may result from a lack of validated instruments or poorly defined endpoints. One example is the assessment of sperm DNA fragmentation, for which at least eight different methods are available, with variable results obtained based on the test used and the laboratory undertaking the assessment ( Agarwal et al. , 2016a , b ; Pacey, 2018 ). Despite the large number of trials published on male factor infertility and the range of primary and secondary outcomes reported on, this inconsistency fundamentally limits their clinical utility and value to inform decision-making and patient care. In addition to difficulties in pooling results of trials, a lack of agreed core outcomes presents challenges for researchers designing future trials when selecting the outcomes to report, further compounded when considering factors such as sample size, cost and time. Our systematic review is the first to report on the primary and secondary outcomes reported in male factor infertility trials and the definitions used for the primary outcome. It builds on work undertaken in other areas of reproductive health to identify causes of subfertility and harmonize the way these data are reported ( Duffy et al. , 2017a,b ; Lee et al. , 2020 ; Turner et al. , 2020 ; Rimmer et al. , 2021 ). At present, there is no consensus on definitions to be used for outcomes relevant to male factor infertility. To address inconsistencies in outcome reporting across male and female infertility trials, an international working group of healthcare professionals and researchers have developed the Core Outcome Measures for Infertility Trials (COMMIT) initiative ( Core Outcomes in Women's and Newborn Health Initiative, 2014 ). This initiative will develop stakeholder-driven development of core outcome sets relevant to clinicians, researchers, and patients and has developed a consensus strategy for reporting core outcomes and standardizing their definitions ( Duffy et al. , 2020 , 2021 ). As no core outcome set for male fertility trials has been developed, it is therefore not surprising that we identified little consistency between outcome reporting and definitions used. This is further compounded by the nature of male infertility trials and the interventions studied. Trials reporting on exercise or dietary supplements to improve semen quality may not report the same outcomes as techniques to select sperm to be used in ART to achieve a pregnancy. Researchers planning future male infertility trials should consider either using a core outcome set for these trials or, in the absence of this, consider reporting outcomes and outcome measures previously used in the literature to improve pooling of data across trials. This review highlights inconsistencies in outcome reporting across male infertility trials but can also be used to identify commonly reported outcomes, when designing future trials. Using previously reported outcomes in new trials when evaluating interventions for male infertility may allow data from these trials to be pooled and meta-analysed. The outcomes identified in this review can be used to develop a core outcome set following discussion by a group of multinational, multiprofessional stakeholders as has been done for general infertility trials ( Duffy et al. , 2021 ). Development of this core outcome set would guide researchers in which core outcomes to report, allow data from several trials to be pooled better inform patient care and reduce research waste ( Duffy et al. , 2017a , b ). We plan to develop a core outcome set for future male infertility research using outcomes reported in this review, using a modified Delphi method and modified Nominal Group Technique to identify relevant outcomes, their measurement and definitions. Our review has several strengths. The comprehensive search strategy and methodological design gives us confidence in the results we have identified. Collecting outcomes reported by 100 trials means that the outcomes and definitions identified reflect a significant body of work and are representative of the field of male infertility trials. To avoid bias, abstract screening and data extraction were undertaken by two independent reviewers, utilizing a third to resolve any queries and reach a consensus. However, our review is not without its limitations. For example, owing to the large number of RCTs published in male factor infertility, we only included the 100 largest trials. This means smaller trials were excluded, and inclusion of the data in these trials may have altered the results obtained and the conclusions drawn. Many trials reported on outcomes but did not clearly state it was the primary outcome or base a sample size calculation on this; as such, they were not included as a primary outcome in our review, despite being reported on by the authors. We did not contact the authors to clarify the primary outcomes where it was not clearly stated or if the researchers extracting data were unsure. We were also unable to validate the quality of the outcomes reported, as there is no validated tool to do this. As our review assessed outcome reporting and the definitions of these outcomes, we did not undertake an integrity check of the trials included in our review. However, we did identify that the conduct of some of the included trials could have been improved upon. These include one trial which was not registered ( Karimi et al. , 2020 ) and four that were retrospectively registered ( Moslemi Mehni et al. , 2014 ; Youssef and Abdalla, 2015 ; Taiyeb et al. , 2017 ; Ketabchi et al. , 2018 ). One trial obtained ethics committee approval after trial registration and the proposed recruitment period (according to the clinical trial registration), although specific recruitment dates were not included in the manuscript ( Ketabchi et al. , 2018 ). One study had an enrolment period that ended 8 months prior to submission of the manuscript but also reported live birth as an outcome ( Majumdar and Majumdar, 2013 ). The conduct and integrity of RCTs are central to their ability to produce robust high-quality evidence ( Li et al. , 2020 ). This would be improved by the implementation of a core outcome set for future male infertility research and assessment of trial integrity when undertaking systematic reviews and meta-analysis.

Supplementary data are available at Human Reproduction Open online.

The data underlying this article will be shared on reasonable request to the corresponding author.

Click here for additional data file.