Effect of oestrogen modulation on semen parameters in men with secondary hypogonadism: Systematic review and meta‐analysis

Selective oestrogen receptor modulators and aromatase inhibitors stimulate endogenous gonadotrophins and testosterone in men with hypogonadism. There are no systematic reviews/meta‐analyses assessing the effects of selective oestrogen receptor modulators/aromatase inhibitors on semen parameters in men with secondary hypogonadism.


INTRODUCTION
Reduced semen quality (male infertility) is one of the most common reasons a couple cannot conceive within 12 months of regular unprotected intercourse.Male infertility may result from secondary hypogonadism, in which gonadotropin levels are low or inappropriately normal.2][3][4] Accordingly, unlike primary hypogonadism, gonadotrophin therapy (GnT) has been proven to restore fertility as well as normal testosterone levels in men with secondary hypogonadism. 5Unfortunately, GnT is quite expensive and therefore unaffordable in most healthcare systems worldwide.
2][13] These agents have also been shown to be beneficial in improving testosterone deficiency in men with secondary hypogonadism, mostly data limited to men with dysmetabolic conditions with functional hypogonadism such as obesity, metabolic syndrome and functional hypogonadism. 8,14,15Nevertheless, it remains unresolved whether SERMs/AIs might have utility in men with infertility associated with milder cases of secondary hypogonadism, or cases of functional, secondary hypogonadism unresponsive to lifestyle intervention.In addition, these drugs are not approved for the treatment of male hypogonadism and their use must be considered an 'off-label' approach.Finally, SERM agonistic effect on venous vessels could facilitate the development of venous thromboembolic diseases in predisposed men. 16 SERMs/AIs could be used to improve semen quality in a subset of men with secondary hypogonadism-related infertility, this would enable many affected couples to access fertility treatment, otherwise unaffordable because of the high costs of GnT.There currently exists no consensus on the effectiveness of SERMs/AIs to treat men with secondary hypogonadism-related infertility.We conducted a systematic review and meta-analysis investigating the effect of monotherapy or a combination of SERMs/AIs on sperm parameters and/or fertility in men with secondary hypogonadism, and to identify predictors of successful treatment if observed.

MATERIALS AND METHODS
We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines in conducting this systematic review.The study was registered in PROSPERO under the registration number CRD42022306535.

Search strategy
We performed an electronic search in the Cochrane Library, MED-LINE and PubMed in January 2022.We searched using the search strategy ('hypogonadotropic hypogonadism' OR 'secondary hypog- To identify additional papers, we performed citation searching by manually screening references of selected articles.We performed an electronic search in ClinicalTrials.govfor registered studies of potential relevance, but without publications.
We contacted the principal investigator for data; studies were excluded if the results were not available or there was no response from the investigators.The detailed search strategy is available in Supporting Information S1.

Study selection
Titles and abstracts of all records were screened independently by two reviewers (N.L. de S. and H. D.) to identify publications eligible for full-text review.The same reviewers screened selected full-text articles independently to identify eligible studies to include in the review.When there were discrepancies in selected full-text articles, two reviewers discussed and reached a consensus.Conflicts were resolved by a third reviewer (C.J.).
All randomised-controlled trials (RCTs) and non-randomised studies of intervention (NRSIs; prospective or retrospective) reporting the effect of SERMs/AI as monotherapy or in combination on semen parameters in men with secondary hypogonadism of any aetiology (defined as hypogonadism with low or inappropriately normal gonadotrophins) 1 irrespective of fertility history were included in the analysis.We included any RCT reporting semen parameters (sperm concentration/motility/total motile sperm count/morphology/volume) compared to placebo or other treatment modalities for secondary hypogonadism.We also included NRSIs reporting semen parameters before and after the intervention with SERMs and/or AIs.Additionally, our inclusion criteria included RCTs and NRSIs reporting conception or live birth.We searched for full-text original articles published in English.We excluded case reports.
When available data in the full-text article were inadequate to define eligibility, corresponding authors were contacted for clarification.and because it is beyond the review's primary focus.

Risk of bias analysis and assessment of the quality of evidence
The risk of bias (ROB) in RCTs was assessed using a revised version of the Cochrane risk of bias (ROB2) tool. 17 interested in the effect of the assignment to the intervention at baseline (intention-to-treat effect) rather than the effect of adhering to the intervention (per-protocol effect).We reviewed the published articles and records from Clinicaltrials.gov to obtain information on reporting bias.
The ROBINS-I tool was used to assess the ROB in NRSIs. 18Potential confounders considered were as follows: severity and cause of hypogonadism; baseline sperm concentration; previous testosterone replacement therapy (TRT) stopped at the time of intervention; other interventions that could have improved the underlying secondary hypogonadism.
The quality of evidence was assessed according to the 'GRADE' approach using GRADEproGDT software.The certainty of evidence was graded based on five main considerations: ROB, consistency of effect, imprecision, indirectness and publication bias. 19

Data analysis
The planned meta-analysis of effect estimates from the available studies could not be performed for RCTs because outcome measures differed across studies and there was incomplete reporting of effect estimates.Therefore, we adapted the vote counting method to determine the direction of effect. 20Vote counting was performed without considering the statistical significance as per the standard method in data synthesis.The benefit or harm of the intervention was defined based on the direction of the effect.
Available NRSIs were pre-and post-intervention studies with critical ROB.The statistical heterogeneity was calculated using I 2 statistics.
Even when low heterogeneity was detected, a random-effect model was applied because the validity of tests of heterogeneity can be limited with a small number of component studies.
While acknowledging the limitation in synthesising data from NRSIs with a critical ROB, we decided to conduct a meta-analysis because available NRSIs and their data were directly relevant in answering the research question of the review.Meta-analysis was performed in Comprehensive Meta-Analysis V4, with single-group pre-and postintervention data calculation and random effect model.

Randomised controlled studies
Six RCTs including 747 participants were eligible for inclusion: among included studies, two compared the effects of SERMs versus placebo versus TRT (n = 377), 21,22 or just with placebo (n = 332). 23,24In addition, one study investigated the role of SERMs compared to TRT (n = 12), 25 or AIs (n = 26). 26The characteristics of the included studies are summarised in Table 1.All studies recruited participants with mean serum testosterone <300 ng/dL and mean serum luteinising hormone (LH) within the reference range; however, the minimum levels of serum LH were reported only in one study. 26Three of the six studies were restricted to men with overweight/obesity [22][23][24] ; the aetiology of secondary hypogonadism was not specified in the remaining studies. 21,25,26However, the mean BMI of participants in two other studies was >30 kg/m 2 , 21,26 and the mean weight was 105 kg in the other study. 25Normospermia was an inclusion criterion for two studies, 23,24 and another study excluded men with a sperm concentration <1 million/mL. 26All participants had previously received TRT in one study, 25 whereas in another study, 23% of the participants had received TRT. 22TRT within the last 3 26 or 6 months 23,24 was an exclusion criterion in three studies.Data on TRT were not available in one. 21men analyses were performed 3-6 months after starting treatment in all included studies.All included studies, except one, had a high ROB, as summarised in Figure 2.
The main RCT results related to semen outcome are summarised in Table 2. Outcomes and reporting differed across studies, so it was not possible to compute summary statistics.Therefore, we adapted the vote counting method to adjudicate the overall direction of effect.F I G U R E 2 Summary of risk of bias assessment of randomised controlled studies.Based on the overall assessment, the first five studies a have high risk of bias and the other study has some concerns on the risk of bias.clomiphene, respectively, conceiving during the study period.However, its intervention period was only 12 weeks.Two studies did not report safety data, 21,25 and one study reported that there were no serious adverse effects in participants. 23One study reported one fatal stroke in a participant on enclomiphene, although causality was not established. 22One serious adverse event with enclomiphene was reported in another study; however, the adverse effect was not specified. 24In the remaining study, anastrozole treatment was associated with skin rash in one patient, and deep vein thrombosis in another patient (who had a previous episode of deep vein thrombosis). 26

Non-randomised studies
8][29][30][31] Four studies were uncontrolled retrospective observational, while the study by Lima et al. was uncontrolled prospective.Although details of patients treated with human chorionic gonadotropin (hCG) are also reported in two studies 29,31 the authors did not consider this a comparator group.There were no eligible studies reporting data on AIs.Characteristics of included studies are summarised in Table 3.
All studies included men with infertility.Whitten et al. included men with idiopathic secondary hypogonadism (n = 3) or panhypopituitarism (n = 1). 29The aetiology of secondary hypogonadism was not reported in the other four studies.The mean/median BMI of participants in all four studies reporting BMI was >28 kg/m 2 . 27,28,30,31In two studies, previous TRT was an exclusion criterion, 27,28 whereas in one study, previous TRT was not reported, 30 and in the other, one patient had been on TRT. 29 In one study, patients who had been on TRT were given a washout period of 8 weeks. 31Except in the study by Patel et al., 28 mean/median sperm concentration at baseline was low or close to the lower limit of the normal.The approximate follow-up duration was 3 months for all the studies.
All five studies had a critical ROB because of the risk of confounding because all of them had a single pre-intervention outcome measurement and a single post-intervention outcome measurement.Other domains of bias were variable between the studies (Figure 3).The publication bias could not be assessed because less than 10 studies were included in the quantitative analysis.The main results of the studies in relation to semen parameters and the overall ROB judgment are summarised in Table 4.
All five studies (n = 105) reported changes in sperm concentration.
reported that there were no significant adverse effects. 30In the other study, three patients had a paradoxical drop in total testosterone and switched to hCG. 28One patient had fatigue and mood swings.number of studies, low total numbers of participants, and unavailability of disaggregated data from subgroups of interest.Safety data were not reported in many studies.Few major adverse events were observed, but causality could not be attributed to SERM/AI treatment.

DISCUSSION
Our systematic review was conducted to appraise for the first time whether SERMs/AIs might improve semen quality in men with secondary hypogonadism.Non-comparative cohort studies suggest that SERM treatment is associated with improved sperm concentration and total motile sperm count in men with low testosterone with low/normal gonadotrophins and infertility; however, RCTs show heterogeneous effects of SERMs compared with placebo.Although studies comparing SERMs to testosterone gel were also summarised in our review, we have not pursued to draw any conclusions using their data because of known suppressive effects of exogenous testosterone on spermatogenesis.
Although we aimed to review studies including men with secondary hypogonadism of diverse aetiology, there was an overrepresentation of men with low testosterone with low/normal gonadotrophins in the background of obesity.From a pathophysiological point of view, a greater response is expected with SERMs in men with obesity-related low testosterone because oestradiol produced through increased aromatase activity from adipose tissue is thought to mediate suppression of the hypothalamus and pituitary. 324][35] In con- a For randomised-controlled trials (RCTs)-summary according to vote counting; for non-randomised studies of intervention (NRSIs)-summary is pooled mean difference.b Studies with high risk of bias, imprecision because of small sample size.c Some concern in risk of bias, small sample size.d Very serious risk of bias because of critical risk of bias in all non-randomised studies of intervention.
hypogonadism in these men may be recommended. 36Additionally, hypogonadism associated with obesity or type 2 diabetes is considered an increasingly prevalent cause of male infertility. 37,38Considering these factors, we included these studies in our review acknowledging the limitation in generalising these data to men with organic secondary hypogonadism and we have identified the broad patient population in our review as having low testosterone with low/normal gonadotrophins.
There is a gap in the evidence between the androgenic and spermatogenic effects of oestrogen modulators among men considered in our review because the evidence on the improvement of testosterone seems to be more robust. 8,14It is not clear whether this is because of the true inefficacy of these agents to promote spermatogenesis or the lack of high-quality studies.This ambiguity is further intensified by some data that suggest that SERMs enhance Leydig cell function but not Sertoli cell action. 39Another important factor is the possible onadism') AND ('selective estrogen receptor modulator' OR SERM OR clomiphene OR tamoxifen OR enclomiphene OR aromatase inhibitor OR anastrozole OR letrozole) in PubMed and MEDLINE from inception to 11 January 2022 without additional filters.All search results were exported into EndNote, and duplicates were removed before screening.The search was updated on 11 November 2022.
searches yielded 1157 records.After removing duplicates, 1008 records remained for screening.After excluding 949 abstracts, 59 full-text articles were screened for eligibility.After excluding 51 studies, eight studies were selected for the systematic review.The updated search did not yield any additional eligible studies.One additional study was identified through citation searching.Two studies identified through ClinicalTrials.govwere included using the results published in the database.Attempts to obtain published full-text articles of these studies by contacting the investigators were unsuccessful.Other studies were excluded because they were either already included through database search (n = 4) or had no published results and could not be accessed even after contacting the principal investigators (n = 2).Details of studies excluded after full-text review are summarised in Supporting Information S3.PRISMA flow diagram summarising the search strategy is given in Figure1.

0. 38 ), I 2 =
0% and 10.52 (95% CI 1.46-19.59),I 2 = 0%, respectively.The certainty of evidence was very low because of the critical ROB in these studies.Surbone et al. reported fertility data, with three female partners becoming pregnant during the study period out of 57.In the study by Whitten et al., female partners of two study participants (out of four) conceived.The other three studies did not report fertility data.

Identification of studies via databases Identification of studies via other methods
Fifty-nine studies were selected to review full-text articles.Eight studies from the database search were included in the review.Three additional studies were included through citation searching and ClinicalTrials.gov.

Table 5
summarises findings, including the assessment of the certainty of evidence using the GRADE tool.Four RCTs with 591 participants compared SERMs against placebo.Two published studies showed evidence towards the benefit; however, two unpublished studies suggested that SERMs were inferior to placebo (possible overlap of patients not excluded).Overall, the certainty of evidence was very low because of the high ROB and imprecision because of the small total sample size.The study comparing anastrozole to clomiphene suggested better sperm concentration outcomes with clomiphene (mean difference -15.00 [-25.40,-4.60]). 26Only the study by Helo et al. mentioned fertility outcomes reporting the conception of one and two partners of men on anastrozole and TA B L E 1 Characteristics of selected randomised controlled studies.a Number of men with sperm concentration <15 million/mL out of the participants with semen parameters.

Study Outcome Overall risk of bias judgement Pre-intervention Post-intervention Mean difference (95% confidence interval) Significance
Characteristics of selected non-randomised studies of intervention with selective oestrogen receptor modulators.Summary of risk of bias assessment of non-randomised studies of intervention on selective oestrogen receptor modulators.All included studies have a critical risk of bias in the overall assessment with bias of confounding been critical in all the studies.Summary of results from non-randomised studies of intervention on selective oestrogen receptor modulators.Summary of findings for effects of selective oestrogen receptor modulators/aromatase inhibitors in men with hypogonadotropic hypogonadism.