Comparative Effectiveness of Recombinant Human Follicle-Stimulating Hormone:Recombinant Human Luteinizing Hormone in a 2:1 Ratio versus Highly Purified Human Menopausal Gonadotropin Alone in Ovarian Stimulation for Medically Assisted Reproduction Treatment Using in vitro Fertilization/Intracytoplasmic Sperm Injection: A Systematic Review and Meta-Analysis

This systematic review and meta-analysis (PROSPERO: CRD42024549520) considered articles published in English, German, Italian, or Spanish between January 2000 and March 2024, regardless of the study design. Studies comparing OS with either r-hFSH:r-hLH in a 2:1 ratio or HP-hMG alone, both administered from day 1, and reporting the number of oocytes retrieved were included. We conducted the literature search according to PRISMA guidelines ( https://www.prisma-statement.org/ ) using Medline (NLM MEDLINE file), EMBASE (Excerpta Medica), Biosis (BIOSIS Previews/RN Database), SciSearch (Science Citation Index), and TOXCENTER (Toxicology Center Database) on STN. Duplicates were eliminated using an automated procedure. The search was performed as broadly as possible, covering various synonyms for the search terms (online suppl. Table 1; for all online suppl. material, see https://doi.org/10.1159/000550420 ). In Medline and EMBASE, the controlled terms were included. To optimize the results, the free-text terms were also searched in abstracts and titles with proximity operators. The titles and abstracts of the studies retrieved were screened for eligibility by J.-E.S. and B.H.; if there was disagreement about eligibility, a third reviewer adjudicated. Data on study information (reference, design, demographics/baseline characteristics), treatment information (downregulation protocol, stimulation dosing, days of stimulation, trigger used, insemination method), and outcome information (number of patients, oocytes recovered for each treatment group, MII oocytes, embryo transfers, implantation rate, clinical pregnancy rate, ongoing pregnancy rate, and live birth) were extracted from the full-text versions of eligible articles, independently, by B.H. and J.-E.S. The potential risk of bias from observational studies was assessed using the ROBINS-I tool (“Risk Of Bias In Non-randomized Studies – of Interventions”) [ 44 ] and visualized using the Risk-Of-Bias VISualization (robvis) program [ 45 ]. The primary outcome was the number of oocytes recovered per cycle. Secondary outcomes were the number of mature (MII) oocytes per cycle, clinical pregnancy rate, ongoing pregnancy rate, and live birth rate per cycle. The mean (standard deviation) number of oocytes recovered and number of MII oocytes per cycle were analyzed using a random-effects model and are reported as mean difference (95% confidence interval [CI]) for r-hFSH:r-hLH 2:1 versus HP-hMG alone. Clinical pregnancy, ongoing pregnancy, and live birth rates per cycle were also analyzed using a random-effects model and are reported as rate ratios (RRs) and 95% CI for r-hFSH:r-hLH 2:1 versus HP-hMG. To assess variance, the results for the primary and secondary outcomes were weighted using the inverse-variance method; heterogeneity was estimated using the I 2 statistic and tested using Cochrane’s Q chi-square test. Sensitivity analyses were performed to assess the impact of excluding studies with high risk of bias. The extent to which each study affected the overall estimate was assessed using the leave-one-out analysis. Additional sensitivity analyses were performed by excluding the studies that used a GnRH antagonist [ 39 , 41 ] and by excluding studies in women with diminished ovarian reserve [ 46 ], respectively. Publication bias was assessed with Egger’s linear regression test to evaluate asymmetry in funnel plots for the outcomes of number of oocytes and clinical pregnancy. All statistical analyses were performed using R (The R Foundation [ http://www.r-project.org ]).

Eighteen potential studies were identified (shown in Fig. 1 ). Seven of these studies compared r-hFSH:r-hLH with HP-hMG alone [ 38 – 41 , 46 – 48 ]. One study [ 48 ] was excluded, as it could not be confirmed that r-hFSH:r-hLH was used in the 2:1 ratio, leaving six studies eligible: four non-interventional [ 38 , 41 , 46 , 47 ], one non-interventional cross-over study [ 39 ], and one RCT [ 40 ] ( Table 1 ). The five non-interventional studies included analysis of previously collected data. The one RCT included prospective data collection and study-assigned intervention. Details of the OS protocols for each of the studies are reported in Table 1 . PRISMA diagram of the assessment of studies to be included in the analysis. hMG, human menopausal gonadotropin; IVF, in vitro fertilization; r-hFSH, recombinant human follicle-stimulating hormone; r-hLH, recombinant human luteinizing hormone. Summary of studies included in the systematic review and meta-analysis AFC, antral follicle count; AMH, anti-Müllerian hormone; GnRH, gonadotropin-releasing hormone; hCG, human chorionic gonadotropin; HP-hMG, highly purified menopausal gonadotropin; ICSI, intracytoplasmic sperm injection; IVF, in vitro fertilization; MAR, medically assisted reproduction; NIS, non-interventional study; RCT, randomized controlled trial; r-hFSH, recombinant human follicle-stimulating hormone; r-hLH, recombinant human luteinizing hormone. The number of oocytes per cycle was significantly higher in women treated with r-hFSH:r-hLH 2:1 compared with those treated with HP-hMG alone (mean difference 1.43, 95% CI: 0.18–2.69; p = 0.025; I 2 = 98%) (shown in Fig. 2 ). No study was considered to have critical bias (i.e., too problematic to provide useful evidence). Four studies [ 38 , 39 , 41 , 47 ] were considered to have low or moderate bias in the seven domains assessed (shown in Fig. 3 ). Two studies [ 40 , 46 ] were judged to have serious bias. Specifically, in Pacchiarotti et al. [ 40 ] the reported results were different from the pre-specified analysis (bias domain 7) and in Mignini Renzini et al. [ 46 ] there was a lack of adjustment for confounding variables (bias domain 1). Forest plot of the number of oocytes retrieved per cycle. CI, confidence interval; hMG, human menopausal gonadotropin; r-hFSH, recombinant human follicle-stimulating hormone; r-hLH, recombinant human luteinizing hormone; SD, standard deviation. Risk of bias assessment for number of oocytes retrieved per cycle. D, domain; D1, bias due to confounding; D2, bias due to the selection of participants; D3, bias in the classification of interventions; D4, bias due to deviations from the intended interventions; D5, bias due to missing data; D6, bias in measurement of outcomes; D7, bias in the selection of the reported result. The sequential removal of one study (leave-one-out) produced results consistent with the main analysis (shown in online suppl. Fig. 1). Similarly, when the studies [ 39 , 41 ] using a GnRH antagonist protocol for pituitary downregulation or the study in women with only diminished ovarian reserve [ 46 ] were excluded, respectively, the results were consistent with the main analysis (shown in online suppl. Fig. 2, 3). No reduction in heterogeneity was seen in any of the sensitivity analyses (range 91–99%). The Egger’s test of the intercept showed no evidence of publication bias (T = 1.22, p = 0.288). The number of MII oocytes retrieved did not differ significantly between the two treatment cohorts, with a mean difference of 0.72 (95% CI: −0.32 to 1.76; p = 0.177). The I 2 statistic was 57%, indicating moderate heterogeneity among the studies (shown in Fig. 4 ). However, this outcome was only reported in three studies [ 38 , 39 , 41 ]: one study did not report this outcome [ 47 ], and two studies only reported the rate and it was not possible to estimate the mean and standard deviation [ 40 , 46 ]. Forest plots of secondary outcomes: MII oocytes ( a ), clinical pregnancy rate ( b ). CI, confidence interval; CPR, clinical pregnancy rate; hMG, human menopausal gonadotropin; MII, metaphase II oocytes; r-hFSH, recombinant human follicle-stimulating hormone; r-hLH, recombinant human luteinizing hormone; SD, standard deviation. Clinical pregnancy rates were reported in five studies [ 38 – 40 , 46 , 47 ]. In two studies [ 39 , 40 ], clinical pregnancy rates were not reported “per started cycle”; however, we were able to calculate the rate per started cycle using data on the number of cycles reported in the articles. One article (Revelli et al. [ 41 ]) was not included in this outcome, as the clinical pregnancy rate per started cycle was not reported and insufficient data were provided to calculate this outcome. The Egger’s test of the intercept showed no evidence of publication bias (T = −1.04, p = 0.374). No standard definition of clinical pregnancy was cited across studies. No definition was specified in three studies [ 39 , 40 , 47 ], and different definitions of clinical pregnancy were used in two studies. In Fábregues et al. [ 38 ] clinical pregnancy was defined by increasing serum concentrations of hCG after embryo transfer and the subsequent demonstration of an intrauterine gestational sac by ultrasonography; in Mignini Renzini et al. [ 46 ] it was defined as ultrasonographic observation of a gestational sac, with or without fetal heartbeat, 25–30 days following embryos transfer. Owing to the absence of a standard definition of clinical pregnancy, we deferred to the definitions used in the contributing articles. Clinical pregnancy rate per started cycle was significantly higher with r-hFSH:r-hLH 2:1 compared with HP-hMG (RR 1.18, 95% CI: 1.06 to 1.33; p = 0.004; I 2 = 10%) (shown in Fig. 4 ). Ongoing pregnancy rate was reported in only one study [ 46 ]. No specific definition was included, and the results for ongoing pregnancy rate were not presented directly; therefore, we calculated the ongoing pregnancy rate using data from the reported number of clinical pregnancies minus the reported number of miscarriages. The calculated ongoing pregnancy rate was significantly higher with r-hFSH:r-hLH 2:1 compared with HP-hMG (RR 1.69, 95% CI: 1.06–2.71; p = 0.028; I 2 = N/A) (shown in online suppl. Fig. 4). Only two studies reported the live birth rate per started cycle [ 38 , 46 ], although no definition was specified in either. Live birth rate was not significantly different between r-hFSH:r-hLH 2:1 and HP-hMG (RR 1.40, 95% CI: 0.90–2.17; p = 0.137; I 2 = 29%) (shown in online suppl. Fig. 4).

In the analysis presented here, OS with r-hFSH:r-hLH in a 2:1 ratio resulted in a significantly higher number of oocytes retrieved compared with HP-hMG. This finding is important, as both the quantity and quality of oocytes influence the outcomes of OS. Furthermore, there is a positive correlation between the number of oocytes retrieved and the number of top-quality embryos, as well as the number of euploid embryos available for fresh transfer or cryopreservation [ 10 , 49 – 58 ]. The clinical pregnancy rate reported was significantly higher after r-hFSH:r-hLH 2:1 treatment in a meta-analysis of five studies, consistent with several studies in specific patient populations. In women with low ovarian reserve [ 46 ], the clinical pregnancy per started cycle was higher in the r-hFSH:r-hLH 2:1 group than in the HP-hMG group (12.5 vs. 8.1%, p < 0.02). And using data from the German DIR database in an observational matched case-control study, treatment with r-hFSH:r-hLH 2:1 compared with urinary menopausal gonadotropin with hCG-derived LH-like activity (u-hMG) resulted in higher pregnancy rates per cycle ( p = 0.006) and per embryo transfer ( p = 0.025) and implantation rate per embryo transferred ( p < 0.001) in the r-hFSH:r-hLH 2:1 cohort [ 47 ]. We were unable to draw any firm conclusions for live birth outcomes owing to the lack of data. When comparing the number of oocytes retrieved, it is important to consider inherent differences in the formulations of the respective gonadotropins used for OS. Firstly, the recombinant products have a higher level of purity and batch-to-batch consistency than urinary products [ 59 – 62 ] and also have a consistent glycoform profile, which, in the case of FSH, may influence bioactivity (or potency), pharmacokinetic/pharmacodynamic profiles, and clinical efficacy [ 62 ]. Secondly, HP-hMG also exhibits a high degree of oxidized samples, reaching up to 20% for the alpha subunit [ 9 ]. Oxidized gonadotropin samples exhibit lower biopotency in both in vivo and in vitro assays [ 63 ], potentially impacting clinical effectiveness. Thirdly, differences in the mode of action of the respective components should also be considered. In vitro studies demonstrate that although LH and hCG both bind to the same LHCGR receptor on antral follicles, they activate distinct intracellular signaling pathways, consistent with their different physiological roles – LH primarily supporting folliculogenesis and hCG predominantly driving steroidogenesis [ 64 ]. The action of LH is preferentially exerted via phosphorylated extracellular-regulated kinase 1/2 (pERK1/2) and phosphorylated AKT (also known as protein kinase B), resulting in proliferative/anti-apoptotic signals and partial agonism of progesterone production in vitro. This may have a positive effect on oocyte maturation, potentiated by r-hFSH, as well as having an anti-apoptotic effect on cumulus cells and promoting the paracrine signaling required for cell expansion and oocyte signaling during folliculogenesis [ 65 , 66 ]. Conversely, hCG displays notable cAMP/protein kinase A-mediated steroidogenic and pro-apoptotic potential [ 64 ], which may have a detrimental effect on oocyte development by activating the steroidogenic and pro-apoptotic pathways. Collectively, the chemical differences between r-hFSH:r-hLH 2:1 treatment and urinary FSH treatment combined with urinary hCG-derived LH-like activity of placental origin [ 9 ], and the differential biological effects outlined above, may explain the higher number of oocytes observed after treatment with rFSH:rLH 2:1 in our study. Our data are in line with results from other studies that consistently showed a higher number of oocytes after OS with r-hFSH, compared with OS with urinary FSH alone [ 53 , 67 ], after OS with hMG preparations [ 50 , 52 , 58 , 67 – 73 ], or treatment with r-hFSH:r-hLH 2:1 compared with hMG in combination with r-hFSH [ 74 ]. Several parameters that have a bearing on OS outcomes could not be accounted for in our analysis, owing to the lack of data available in the source material. Firstly, the starting dose and total gonadotropin dose are important determinants of the number of oocytes recovered after OS, and there is good evidence that a higher r-hFSH starting dose results in significantly more oocytes recovered in high, normal and low responders compared with a lower r-hFSH starting dose [ 75 ]. Dose increments of r-hFSH are the only reported significant predictor of the number of oocytes retrieved in the subsequent OS cycle, with an increase of 50 IU over the initial r-hFSH dose leading to an increase of one more oocyte [ 22 ], and an increase in the daily gonadotropin dose above 300 IU was also shown to result in a higher mature oocyte yield in women undergoing elective embryo freezing [ 76 ]. However, we were not able to establish any relationship between dosing and the number of oocytes retrieved (online suppl. Table 2), as there was a high degree of variability in dosing among the studies included, and it was unclear whether the reported dosing values represented only the FSH component of the treatment or the activity for the combined components. Secondly, we did not include or exclude studies based on the protocol used to prevent premature LH surge. Only two of the studies used GnRH antagonist protocols [ 39 , 41 ], which is the current prevailing protocol used for pituitary downregulation [ 77 ]. We do not envisage this to be of concern, as there is no reported difference in outcomes between the two protocols [ 78 , 79 ], although we acknowledge this may have implications when generalizing the results presented here to the general population and current practice. Thirdly, owing to a lack of data in the source publications, we were unable to account for the use of oral contraceptive pill pretreatment, which may be employed for several reasons, including the synchronization of follicular development and scheduling of IVF cycles for the convenience of clinicians and patients [ 80 , 81 ]. Nevertheless, current evidence for the effect of pretreatment hormonal contraception on OS is inconsistent [ 77 ], and the effect may be dependent on the type of hormonal contraception used [ 81 ]. Finally, we were unable to account for other factors, such as the use of nutraceutics or other compounds, sperm source or quality, or oocyte developmental competence. A key strength of this meta-analysis is that the greater number of oocytes retrieved with r-hFSH:r-hLH 2:1 compared with HP-hMG is based on a large number of cycles (2,639 for the recombinant and 2,648 for the urinary products) using observational and RCT data obtained across four countries. Furthermore, the high degree of similarity in patient baseline characteristics between the two treatment groups within each included study (online suppl. Table 3) strengthens the validity of this meta-analysis, and increases the credibility of any differences in outcomes reported. We also acknowledge several limitations. Firstly, only six studies were deemed eligible for inclusion in the analysis, and only one of these was an RCT. Secondly, the heterogeneity calculation identified that some factors, including patient populations, treatment protocols and dosing patterns, differed among the studies, but important features (i.e., the standard deviations between treatment groups) did not differ within each study. Nevertheless, while the differences across studies warranted sensitivity analyses (including leave-one-out analysis and analyses by downregulation protocol and by ovarian reserve), the results of these were consistent with those for the overall analysis. Thirdly, two studies [ 40 , 46 ] were judged to have serious reporting risk of bias. In Pacchiarotti et al. [ 40 ] the authors planned to correct for the multiple comparisons made on the numerous outcomes by adjusting the significance level that the p values had to reach. However, as the authors presented multiple outcomes as being significant without making the p value adjustment, it is unclear whether the multiple analyses were performed and selected results were presented, rather than all analyses or the pre-specified analyses. Since the most significant result (the number of oocytes retrieved) was higher in the r-hFSH:r-hLH 2:1 treatment group, we cannot rule out the possibility that there is bias in the results of this RCT that favors the r-hFSH:r-hLH 2:1 treatment group. In Mignini Renzini et al. [ 46 ] patients in the HP-hMG cohort had a higher baseline antral follicle count value compared with the r-hFSH:r-hLH 2:1 cohort (5.9 [2.4] vs. 4.9 [2.1]), but this confounding factor was not adjusted for when the study outcomes were reported, which could have biased the outcome towards a higher number of oocytes in the HP-hMG cohort compared with the r-hFSH:r-hLH 2:1 cohort. Fourthly, there was no standard definition of clinical pregnancy rate in the constituent studies, and three studies did not provide any definition; for this endpoint, we have deferred to the definitions used by the authors in the source material, and were not able to use the pregnancy definitions according to the ICMART Glossary [ 82 ]. The use of varying definitions for, and the timing of, clinical pregnancy is a recognized limitation in the literature [ 83 ], and we acknowledge that this may have the potential to influence the comparison of pregnancy rates across studies; for example, the use of a broad definition of clinical pregnancy (e.g., the presence of a gestational sac) may lead to higher pregnancy rates compared with studies that adopt a narrower definition (e.g., confirmed fetal heartbeat). However, our analysis is a summary across all studies of the difference in pregnancy rates between treatment groups within the same study. We are confident that a consistent definition of pregnancy was applied to both treatment groups within each study, enabling valid calculation of differences in pregnancy rates within studies and ensuring that the results were not affected by variability in definitions across studies. Additionally, when the study by Pacchiarotti et al. [ 40 ] was conducted, Italy permitted insemination of three oocytes per IVF cycle only. This may have resulted in a bias towards the selection of good-quality embryos and may have subsequently impacted the clinical pregnancy rate. Finally, quality, stage, and number of transferred embryos were not reported consistently (or were not reported at all), so the potential impact of embryo quality and/or embryo development state on pregnancy rates could not be assessed. Data on several outcomes were limited: only one study reported ongoing pregnancy rate [ 46 ] but did not define this outcome or report the result directly, only two studies reported live birth rate [ 38 , 46 ], and no studies reported on cumulative live birth rate. Egger’s test did not reveal evidence of publication bias in the study outcomes; however, the limited number of included studies may have reduced the statistical power to detect such bias.

We identified six studies that utilized r-hFSH:r-hLH in a 2:1 ratio, which were included in this meta-analysis. Despite a high level of heterogeneity across the six studies, the results indicated that the number of oocytes retrieved and the clinical pregnancy rate were significantly higher in patients treated with r-hFSH:r-hLH at a 2:1 ratio compared with those treated with HP-hMG. Our analysis has also highlighted both the lack of data for several outcomes (number of MII oocytes, ongoing pregnancy, and live birth rates) as well as inconsistency or absence in the reporting of outcomes such as clinical pregnancy, ongoing pregnancy, and live birth. In response to our findings, we call for more high-quality studies, both RCTs and observational studies, to address this gap.

Exogenous gonadotropins are used during ovarian stimulation (OS) to obtain a supraphysiological number of oocytes. During OS, residual circulating luteinizing hormone (LH) is adequate to support steroidogenesis in the follicles in many cases; thus, exogenous follicle-stimulating hormone (FSH) alone is considered sufficient for OS [ 1 , 2 ]. However, in the case of patients with LH and FSH deficiency due to reduced gonadotropin production or action [ 3 ], residual circulating LH is not sufficient, and exogenous LH in addition to FSH is required for optimal outcomes [ 3 ]. Different preparations containing FSH and LH or LH-like activity, including recombinant- and urinary-derived products, are used in OS. Originator recombinant human FSH (r-hFSH) and recombinant human LH (r-hLH) are produced using recombinant cDNA technology [ 4 ] and are available as separate products (originator r-hFSH; Gonal-f ® , Merck Europe BV, Amsterdam, The Netherlands; r-hLH; Luveris ® , Merck Europe BV, Amsterdam, Netherlands [ 5 , 6 ]) or as a 2:1 fixed-ratio combination product (r-hFSH:r-hLH 2:1; Pergoveris ® , Merck Europe BV, Amsterdam, The Netherlands) [ 7 ]. Based on the cumulative number of treatment cycles between June 2007 (when marketing authorization was granted) and the end of September 2024 (2,530,049) and using a conservative estimate of the live birth rate (20.3% [ 2 ]), an estimated 513,599 babies have been born following co-treatment with r-hFSH:r-hLH 2:1. The highly purified human menopausal gonadotropin HP-hMG (Menopur ® , Ferring Pharmaceuticals, West Drayton, UK) [ 8 ]) is obtained from the urine of postmenopausal women and comprises urinary FSH- and LH-like activity, which is mainly derived from placental human chorionic gonadotropin (hCG) purified from pregnant women [ 9 ]. Identifying the optimum OS protocol is essential for enhancing patient outcomes, as oocyte number is positively associated with the number of good-quality embryos obtained [ 10 ], the number of euploid blastocysts [ 11 , 12 ], the likelihood of pregnancy [ 13 , 14 ], live birth rates [ 15 – 21 ], and cumulative live birth rates [ 22 – 27 ]. There is published evidence that the combination of FSH and LH may be beneficial in several populations of women, especially those at risk of severe LH and FSH deficiency, which includes women of advanced reproductive age (ARA). A meta-analysis of 12 randomized controlled trials (RCTs) reported that women aged 35–40 years treated with r-hFSH:r-hLH had higher implantation and clinical pregnancy rates versus r-hFSH alone, suggesting that r-hFSH + r-hLH treatment may improve these outcomes in women of ARA [ 28 ]. In a non-interventional study based on real-world data derived from the Deutsches IVF Register, treatment with r-hFSH:r-hLH resulted in higher clinical pregnancy and live birth rates compared with r-hFSH alone in women of ARA and in women of ARA with predicted normal ovarian reserve (5–14 oocytes retrieved) [ 2 ]. Other populations that may benefit from co-administration of r-hFSH and r-hLH compared with r-hFSH alone include women with hypogonadotropic hypogonadism due to either reduced gonadotropin levels or reduced bioactivity or women with profoundly suppressed LH levels due to gonadotropin-releasing hormone (GnRH) agonist or antagonist treatment during ART, in terms of oocyte quality and implantation rate [ 29 – 32 ], and women who are hyporesponsive to FSH alone [ 33 , 34 ]. Limited data suggest that polymorphisms in LHB / LHCGR can influence OS outcomes [ 35 , 36 ]. However, while accumulating data suggest that ovarian response to OS may be mediated by various polymorphisms, further studies are needed to investigate the predictive value of these as markers of response in subgroups of women who may require supplementation with LH during OS [ 36 , 37 ]. In the general population, OS with r-hFSH:r-hLH 2:1 has been reported to yield a higher number of oocytes versus HP-hMG [ 38 – 41 ]. Furthermore, in egg donor cycles [ 42 ], more oocytes were retrieved after treatment with r-hFSH:r-hLH 2:1 compared with HP-hMG (16.5 vs. 11.8; p = 0.049). In isolation, these few studies do not constitute sufficient evidence for the benefit of r-hFSH:r-hLH 2:1 treatment over HP-hMG with respect to number of oocytes or other reproductive outcomes. A previous systematic review of studies found no statistically significant differences between r-hFSH plus r-hLH compared with hMG in clinical pregnancy or live birth rates [ 43 ]. However, the studies included in this review [ 43 ] reported on heterogenous gonadotropin preparation combinations administered at variable r-hFSH:r-hLH ratios, which makes it difficult to draw firm conclusions from this analysis. Given the preparations currently available, most OS cycles will use r-hFSH and r-hLH in a 2:1 ratio. Therefore, to provide a more precise answer, we conducted a systematic literature review and meta-analysis to assess the difference in the mean number of oocytes retrieved after OS with either r-hFSH:r-hLH in a fixed 2:1 ratio or HP-HMG.

M.H.D. has received fees Merck Healthcare KGaA, Ferring Inc., and Organon Inc. for consultations and scientific presentations. M.H.D. is the Editor and Chief of the Journal Clinical and Experimental Obstetrics and Gynecology and has published several textbooks related to infertility with diverse publishers. B.H. has received fees from Merck Healthcare KGaA as a statistical consultant. R.F. has received honoraria for lectures from Merck Healthcare KGaA, Medea, and Event Planet. S.C.E. has received unrestricted research grants from Merck KGaA; consulting fees from Merck, Medea, and Event Planet; payment or honoraria for lecture fees from Merck, Sanitanova, Medea, and Event Planet; is an advisory board member for Nature Reviews Urology (unpaid); and is the Head of the Andrology Committee, Brazilian Society of Human Reproduction (unpaid), and Topic Leader of the Male Infertility Group, WHO Infertility Guidelines (unpaid). MC, IF, and K.M.S. declare no conflicts of interest. L.G. has received honoraria for external consultant from CooperSurgical. D.S. has received honoraria for lectures from Merck Healthcare KGaA, Medea, and Event Planet. C.A. declares receipt of personal fees from Merck, Medea, Event Planet, IBSA, Ely Lilly, S&R Farmaceutici, and Biogen; and payment or honoraria from Merck, Medea, Event Planet, IBSA, and Ferring. J.-E.S., S.S.G., and T.M.D. are employees of Merck Healthcare KGaA, Darmstadt, Germany. J.-E.S. and T.M.D. were members of the journal’s Editorial Board at the time of submission.

Medical writing support was provided by Steven Goodrick of inScience Communications, Springer Health + , and was funded by Merck.

The work was funded by the healthcare business of Merck, Darmstadt, Germany (CrossRef Funder ID: 10.13039/100009945). The sponsor was involved in the study design, execution and analysis, manuscript conception, planning, writing, and decision to publish.

This study is not applicable, as this systematic review and meta-analysis used data from previously published literature.

M.H.D., R.F., S.C.E., M.-J.C., K.M.S., L.G., D.S., I.F., and C.A.: Interpretation of the data, drafting the article and critically revising it for important intellectual content, final approval for the article to be published, and accountability for all aspects of the work. J.-E.S. and S.S.G.: conception and design of the study, data acquisition, drafting the article and critically revising it for important intellectual content, final approval for the article to be published, and accountability for all aspects of the work. B.H.: conception and design of the study, data acquisition, data analysis, drafting the article and critically revising it for important intellectual content, final approval for the article to be published, accountability for all aspects of the work. T.M.D.: conception and design of the study, interpretation of the data, drafting the article and critically revising it for important intellectual content, final approval for the article to be published, and accountability for all aspects of the work.