Efficacy of Gonadotropin Therapy to Induce Spermatogenesis and Fertility in Men with Pathologic Gonadotropin Deficiency.

Gonadotropin-deficient men (n = 99) received 160 independent cycles of gonadotropin therapy. Baseline characteristics are presented in Table 1 . The median cycle duration was 16 ( 10 , 30 ) months. Beyond the 99 first cycles, 40 men were treated with second cycles, 19 third cycles, and 2 fourth cycles ( Fig. 1 ). For 40 men receiving multiple treatment cycles, cycles were separated by median 18 ( 10 , 27 ) months. The modal hCG dose required was 4500 IU (urinary) and 250 µg (recombinant) per week. The modal FSH dose required was 450 IU weekly. Swimmer plot displaying all 160 gonadotropin treatment cycles administered to 99 gonadotropin-deficient men. The end of a treatment cycle was recorded as the achievement of first pregnancy (circle) or when treatment was opted to be ceased by the patient and clinician. Baseline participant and treatment cycle characteristics Data are n (%), mean ± SEM, or median (interquartile range). Abbreviations: BMI, body mass index; CHH, congenital hypogonadotropic hypogonadism. The proportion achieving (%) and median time to achieve (months) sperm thresholds for the 160 treatment cycles was >0 (78%, 5), >2 (56%, 12), >5 (49%, 14), >10 (37%, 21), and >20 (24%, not determined) M sperm/mL ( Fig. 2 ). Subsequent cycles of gonadotropin treatment were marginally independently associated with earlier time to sperm concentrations >0 [hazard ratio (HR) 2.06, 95% confidence interval (CI) 1.09-3.89; P = .03) ( Table 2 ) and >2 (HR 1.96, 95% CI 0.95-4.06; P = .07) [Table S1 ( 18 )] M sperm/mL. Men with bilateral cryptorchidism had slower time to appearance of sperm compared to men without cryptorchidism (HR 0.06, 95% CI 0.01-0.60; P = .02). Higher pretreatment summed testes volume was marginally associated with earlier time to sperm concentrations >10 (HR 1.03, 95% CI 1.00-1.06; P = .02) and >20 (HR 1.05, 95% CI 1.01-1.08; P = .006) M sperm/mL on univariate analyses only. Postpubertal onset of HH and anthropometric measurements [Table S2 ( 18 )] were not significantly associated with time to any sperm threshold. Proportions achieving sperm concentrations >0 M/mL were similar in those with mean pretreatment testis volume 4 mL (81% vs 88%, respectively). Kaplan-Meier plots showing the cumulative incidence over time by which sperm density thresholds of more than 0 (red line), 2 (olive green line), 5 (green line), 10 (blue line), and 20 (purple line) million sperm per mL were achieved. The median is plotted as a horizontal dashed line and indicates when half the treatment cycles reached the corresponding sperm threshold. Analysis of the association between covariates and attainment of sperm concentrations >0 million sperm/mL, >5 million sperm/mL, and >10 million sperm/mL from the Cox regression analysis using all covariates Data is reported as HR (95% confidence interval). An HR of more than 1 indicates a faster rate to attain the respective threshold. An HR of less than 1 indicates a slower rate to attain the respective threshold. Corresponding analyses for sperm concentrations >2 million sperm/mL and >20 million sperm/mL are reported in Table S1 ( 18 ). Abbreviations: CHH, congenital hypogonadotropic hypogonadism; hCG, human chorionic gonadotropin; HR, hazard ratio. Of the 157 gonadotropin cycles conducted in men seeking fertility, 90 (57%) resulted in a pregnancy: 72 naturally and 18 via IVF. Pregnancy occurred after a median 27 months of treatment ( Fig. 3 ). Adverse female fertility factors, affecting 69/157 (44%) of treatment cycles, was the major determinant of successful pregnancy, with only 35% (24/69) of treatment cycles with female factors achieving pregnancy compared to 75% (66/88) of cycles without female factors present (HR 0.20, 95% CI 0.10-0.40; P < .001) ( Table 3 ). Median time to pregnancy was also shorter for men whose partners did not have adverse fertility factors (15 vs 43 months; P < .0001). IVF was more frequently required to achieve pregnancy when adverse female fertility factors were present (46% vs 11%). Subsequent cycles of gonadotropin treatment were associated with faster time to pregnancy relative to cycle 1 (HR 2.59, 95% CI 1.56-4.31; P < .001). Proportions achieving natural pregnancy were similar in those with mean pretreatment testis volume 4 mL (45% vs 45%, respectively). Kaplan - Meier plots showing the cumulative incidence over time by which conception is achieved, in the absence (blue line) and presence (red line) of adverse female factors. The median is plotted as a horizontal dashed line and indicates when half the treatment cycles have resulted in conception. Analysis of the association between covariates and attainment of pregnancy from the Cox regression analysis using all covariates Data is reported as HR (95% CI). An HR of more than 1 indicates a faster rate to attain the respective threshold. An HR of less than 1 indicates a slower rate to attain the respective threshold. Abbreviations: CHH, congenital hypogonadotropic hypogonadism; CI, confidence interval; hCG, human chorionic gonadotropin; HR, hazard ratio. In total, 20/160 (13%) gonadotropin cycles were conducted using hCG treatment alone, with no requirement for FSH as spermatogenesis occurred within 6 months of treatment initiation. The remaining 140 (88%) gonadotropin treatment cycles consisted of sequential combined hCG and FSH treatment. 19/20 (95%) hCG-only cycles occurred in men with mean pretreatment testis volume >4 mL. For the gonadotropin cycles conducted in men with mean pretreatment testis volume <4 mL, only 1/44 (2%) were conducted using hCG treatment alone. The remaining 43 (98%) consisted of sequential combined hCG and FSH treatment. Of the 20 hCG-only treatment cycles, 18 were for intended fertility, with 8/18 (44%) resulting in pregnancy. For all gonadotropin cycles which resulted in pregnancy, the median (interquartile range) sperm concentration and total sperm output at pregnancy was 5.4 (0.4, 13.6) M/mL and 13.4 (0.9, 49.5) M, respectively, conditional on the presence and absence of adverse female factors ( Table 4 ). Sperm concentration required for pregnancy was significantly lower in men whose partners did not have, compared to those who had, adverse female fertility factors (2.8 vs 35.4 M/mL; P = .001). Median sperm concentration and total sperm output at time of conception for the 90 gonadotropin cycles that resulted in pregnancy Data are median (interquartile range). Abbreviation: IVF, in vitro fertilization. a Data not available for all cycles. There was no difference in time to any sperm concentration threshold for recombinant (n = 38) hCG compared with urinary (n = 122) hCG, nor between recombinant (n = 117) FSH compared with urinary (n = 23) FSH [Fig. S1 ( 18 )].

The present time-dependent analysis of sperm output and pregnancy outcomes across 160 gonadotropin treatment cycles has sufficient power to perform multivariate survival-type analyses. Most (78%) gonadotropin treatment cycles resulted in the appearance of ejaculated sperm, usually within 5 months, and 57% of treatment cycles resulted in pregnancy, mostly within 27 months. These data reinforce previous findings ( 11 , 12 ) but additionally report novel relevant factors impacting pregnancy outcomes, notably the presence of adverse female fertility factors, which had a greater impact than many highly relevant male factors such as sperm concentration achieved, cryptorchidism, and testicular volume. Our data affirm that gonadotropin treatment for gonadotropin-deficient men is effective at inducing spermatogenesis ( 2 , 3 , 5 , 7 , 8 , 11 , 12 , 19-21 ). Nevertheless, most previous studies have insufficient power for time-to-event analyses, do not include multiple treatment cycles per man, and are often limited to the appearance of sperm ( 3 ), rather than higher sperm thresholds required for natural conception ( 5 , 22 ). In the present study, the median time for appearance of sperm was 5 months, consistent with some ( 12 ) but faster than other previous reports of 7 to 15 months ( 7 , 11 ). Furthermore, most cycles resulted in sperm concentrations of >5 M/mL within 14 months, also faster than previous reports of 18 to 24 months ( 7 , 11 ). Overall, the proportions of men attaining sperm-density thresholds of 0, 2, 5, 10, and 20 M/mL after gonadotropin treatment in our study were similar to those reported by a recent meta-analysis ( 2 ) [Fig. S2 ( 18 )]. An important novel finding of this study was the detrimental effect of adverse female fertility factors on pregnancy outcomes, not reported in previous studies focused primarily on male factors. Treatment cycles in couples with adverse female factors resulted in a slower time to conception (median 43 vs 15 months), slower rate of conception (0.16 vs 0.54 pregnancies per year), and lower chance of conception (35% vs 75%). When adverse female factors were present, almost half of pregnancies required IVF, with high median sperm concentration at conception (14.0 M/mL), suggesting adverse female factors were the predominant reason for accessing IVF. Comparatively, the median sperm concentration required for natural conception was 4.9 M/mL in couples without female factors, lower than 7 to 8 M/mL reported by large studies that did not account for adverse female factors ( 6 , 11 ). Therefore, the sperm concentration required for conception in HH patients treated with gonadotropins may have been previously overestimated due to the failure to account for the contribution of female subfertility. Similarly, the chance of fertility in men with HH, reported as 28% to 55% ( 2 , 5-8 , 11 , 12 , 23 ), may underestimate the chance of pregnancy in couples without female factors (75% in our study). Thus, assessment of female fertility should be considered promptly for all men with HH desiring fertility, particularly if conception is not achieved after sperm concentrations exceed 4 to 5 M/mL. The median sperm concentration required for IVF pregnancies in cycles without female factors was much lower (0.2 M/mL), supporting previous evidence of successful conception with low sperm concentrations using IVF and intracytoplasmic sperm injection (ICSI) ( 24 ). No difference in time to conception was observed in our study for natural vs IVF pregnancies, but most treatment cycles included in our analysis occurred before IVF and ICSI were readily available. Although median time to conception in couples without female factors was 15 months, the median time to sperm concentrations required for IVF was only 4 to 12 months. Therefore, it is probable that conception can be achieved earlier than 15 months with modern use of assisted reproductive technology (IVF and ICSI), although at increased cost and invasiveness to the couple. The present findings confirm that second or later cycles are faster than the initial cycle in time to pregnancy ( 11 ). This difference is not due to female factors (42% vs 44% of first and subsequent cycles, respectively) but may be due to differences in sperm output, given the marginal independent association of subsequent cycles with faster time to sperm concentration thresholds of 0 and 2 M/mL. Improved pregnancy outcomes in later cycles may also be due to participation bias, whereby couples with proven fertility may be more likely to pursue further treatment cycles, or that sperm output and testis volume may not return to pretreatment baseline prior to the next cycle commencing. Previous studies have shown that higher baseline testis volume is associated with improved spermatogenesis outcomes ( 6 , 7 , 11 , 12 , 19 , 20 , 23 ), an unsurprising finding given that the bulk (>90%) of testis volume is spermatogenic tissue ( 25 ). In the present study, this was only evident on univariate analyses for higher sperm concentration thresholds of >10 and >20 M/mL and was not evident on multivariate analysis for any sperm concentration threshold. Pretreatment testes volume was recorded prior to any gonadotropin treatment exposure per man rather than per cycle, which may have diminished the effect of this variable in subsequent treatment cycles. Similarly, other studies have linked postpubertal (vs prepubertal) onset of HH with improved spermatogenesis outcomes ( 2 , 5 ), but the present multivariate Cox regression analysis did not identify postpubertal HH as a significant predictor of sperm output when controlled for testes volume. The lack of independent association with spermatogenesis outcomes may indicate a collinear effect of diagnosis with testis volume, as men with prepubertal CHH usually had smaller testes than men with adult onset HH due to pituitary disease (12 vs 18 mL), consistent with other data ( 26 ). Bilateral cryptorchidism was marginally associated with slower appearance of sperm, consistent with some ( 6 , 27 , 28 ) but not all ( 11 , 12 , 29 ) previous studies. Although cryptorchidism has a well-known association with male infertility ( 30 ), it has been difficult to assess in gonadotropin-deficient men due to small sample sizes ( 12 ). Cryptorchidism may reflect failure of the masculinization programming window during fetal life and/or absent mini-puberty in men with CHH, leading to the reduced proliferation of immature Sertoli cells ( 31 , 32 ) that may blunt response to gonadotropin therapy in adulthood. The efficacy of urinary vs recombinant hCG and FSH for the induction of spermatogenesis has not been studied previously. The present study found no significant effect of urinary vs recombinant hCG and FSH, respectively. This supports previous evidence demonstrating similar pharmacokinetic and pharmacodynamic properties of urinary and recombinant hCG ( 33 , 34 ). However, these findings should be interpreted with caution as they are a nonrandomized comparison. While combined hCG/FSH treatment improves sperm output compared to hCG monotherapy ( 2 ), 8 pregnancies in 18 cycles were achieved with hCG treatment alone. These cycles, which achieved adequate spermatogenesis without requiring FSH, had higher baseline testes volume (mean 30 mL), with only 1 cycle occurring in a man with mean pretreatment testis volume <4 m, and only 1 cycle affected by bilateral cryptorchidism. Men with pre-pubertal testis volumes (4 mL) treated with combined hCG/FSH (47% vs 49%, respectively). The high prevalence of cryptorchidism in CHH (up to 36%) indicates a more severe reproductive phenotype ( 35 ), whereas the prevalence of cryptorchidism in men with acquired gonadotropin deficiency is likely <1% ( 36-38 ), comparable with the general adult male population. Together, these findings support previous suggestions that hCG monotherapy may be sufficient in some men where they have experienced at least partial spontaneous puberty ( 39 ), whereas combination therapy is more likely to be required in cases with absent puberty. However, further comparator studies are required to confirm this. While the pregnancy rate of 44% for hCG-only cycles was lower than the overall 57%, the difference is likely due to nonspermatogenesis factors, notably higher rates of female factors (56%) and fewer men who underwent multiple cycles (22%). Important strengths of our study are the large sample size, time-to event multivariate analyses across multiple treatment cycles, comprehensive covariate evaluation (notably the contribution of adverse female factors to fertility outcomes), and a consistent protocol and therapeutic team over 4 decades. Limitations include the study's observational nature, including nonrandomized comparisons between urinary and recombinant gonadotropins. Pregnancy outcome data were present for every gonadotropin cycle, producing reliable estimates of fecundity in our population of 99 infertile men. However, over a third of gonadotropin cycles were affected by missing sperm data, with risk of attrition bias, especially in less successful treatment cycles, which may artificially lower estimates of median time to sperm density thresholds. Proportions of men who can achieve higher sperm concentrations may be underestimated, as cycles usually ended at conception, and sperm concentrations <5 M/mL were sufficient for natural conception in many men in the absence of adverse female factors. Lastly, despite its relatively large sample size, the present analysis still lacks power in many areas, particularly to analyze the independent effect of closely related male factors (such as testis volume, disease etiology, and cryptorchidism) on spermatogenesis outcomes. As subsequent treatment cycles were associated with improved spermatogenesis and pregnancy outcomes, our study highlights the need for randomized studies to assess the efficacy and feasibility of gonadotropin therapy, rather than testosterone (which promotes virilization but not testicular growth), for induction of puberty in patients with prepubertal onset HH but no immediate desire for fertility ( 19 ). Additional research is also needed to evaluate the role of IVF in HH men on time to and rates of conception. In conclusion, our findings affirm that gonadotropin therapy for gonadotropin-deficient men is highly effective for inducing spermatogenesis and fertility, with most men able to achieve sperm output required for pregnancy within 1 year of commencing treatment in the absence of female fertility limitations. We also show that it is critical to evaluate the fertility status of the female partner, even in the setting of significant male infertility, as adverse female factors were the most important prognostic factor for successful fertility. Encouragingly, in the absence of adverse female factors, men with HH achieve fertility faster, at greater proportions, and at lower sperm concentrations than previously thought and similar to men without HH.

In this retrolective (data collected prospectively via a predetermined standard protocol and treatment regimen but analyzed in retrospect) ( 10 ) cohort study, infertile men with HH received treatment to induce spermatogenesis and fertility using a standardized protocol of hCG alone for 6 months followed by the addition of FSH; as described previously ( 11 ). Briefly, gonadotropin treatment consisted of urinary or recombinant hCG in 2 or 3 split doses per week, with dosing individually optimized by trough serum testosterone concentration for up to 6 months. If no sperm appeared in the ejaculate after 6 months of hCG treatment, urinary or recombinant FSH was added in 2 or 3 split doses per week. All men with pathologic gonadotropin deficiency (structural or genetic etiology) treated from 1983 until August 2024 were included, excluding men with functional or drug-induced disorders causing azoospermia (eg, opiate or androgen misuse or abuse). This report updates previous interim analyses of 29 men treated until 2002 ( 12 ) and 50 men treated until 2008 at this center ( 11 ). The diagnosis of HH was established by standard medical history; physical examination (including testicular volume and testing for anosmia); and low plasma testosterone, LH, and FSH levels with a plausible clinical diagnosis of gonadotropin deficiency ( 13 ). Kallmann syndrome was diagnosed in a subset of men with congenital hypogonadotropic hypogonadism (CHH) and anosmia ( 13 ). Men with hypopituitarism were treated with other hormones if required, including thyroxine, glucocorticoids, and vasopressin, but none were treated with GH. All female partners were assessed and referred for standard investigations via fertility specialists as required. During gonadotropin treatment, men were reviewed at 3-month intervals. At each visit, clinical features (androgenic effects, testis size) were monitored, and blood (plasma testosterone, LH, FSH, SHBG) and semen samples were obtained. Total testicular volume was estimated using a Prader orchidometer. Semen samples collected by masturbation were analyzed according to the contemporaneous World Health Organization method ( 14 ). Gonadotropin treatment was continued until completion of the first trimester of pregnancy, after which treatment was switched back to testosterone replacement therapy. Conception dates were estimated to the nearest week based on the date of positive urinary pregnancy tests, otherwise designated as the 15th day of the month. Semen analysis at conception was taken as that closest to and within 90 days of conception. Adverse female factor data encompassed a composite of adverse female fertility factors (age ≥40 years, fallopian tube disease, endometriosis, polycystic ovarian syndrome or other ovulatory, tubal, or uterine disorder) and relationship factors (breakdown or living apart). In vitro fertilization (IVF) pregnancies were patient-reported. Institutional ethics committee approval for this analysis of completed treatments was not required. No additional procedures or data beyond standard of care were sought, data were de-identified before analysis, and it was not feasible to obtain retrospective consent from all individuals treated over the last 40 years. The time to sperm density thresholds [0, 2, 5, 10, and 20 million (M)/mL] and time to pregnancy were the coprimary outcomes of this study, assessed using a time-to-event analysis per cycle of treatment. Covariables included in analyses were (1) type of gonadotropin treatment (urinary or recombinant hCG or FSH), (2) etiology of gonadotropin deficiency (congenital/prepubertal onset vs acquired/postpubertal onset), (3) treatment cycle number (cycle 1 vs cycle 2 or later), (4) cryptorchidism (unilateral or bilateral), (5) presence of adverse female fertility factors, (6) age at start of treatment cycle, (7) anthropometry (height, weight, body-mass index, body surface area), and (8) pretreatment total testicular volume. Data for each treatment cycle were collected until the end of treatment or after 5 years of treatment, whichever was sooner. Nonparametric Kaplan-Meier analyses were performed to assess the time required to achieve threshold sperm concentrations and pregnancy. Differences in survival plots were assessed with the log-rank (Mantel-Cox) test subject to Bonferroni adjustment for 5 sperm density threshold endpoints. Cox model univariate and multivariate regression analyses included prespecified categorical and continuous variables considered as independent cofactors. Data is expressed as mean ± SEM if normally distributed or median (interquartile range) otherwise. Two-tailed P -values of <.05 were considered statistically significant, except for spermatogenesis thresholds, where a Bonferroni-adjusted P -value of <.01 was used. Statistical analysis was conducted using R version 4.1.0 ( 15 ) through survival ( 16 ) and survminer ( 17 ) packages.