Testicular Growth and Spermatogenesis: New Goals for Pubertal Hormone Replacement in Boys With Hypogonadotropic Hypogonadism?

A Multicentre Prospective Study of hCH/rFSH Treatment Outcomes During Adolescence

Julia Rohayem; Berthold P. Hauffa; Margaret Zacharin; Sabine Kliesch; Michael Zitzmann


Clin Endocrinol. 2017;86(1):75-87. 

In This Article


A total of sixty patients aged 14–22 years with HH were enrolled. Group A boys (n = 34), mean age 15·5 years, were prepubertal or had early arrested puberty. Group B adolescents (n = 26), mean age 18·8 years, had received previous full testosterone-enanthate replacement for 1·5–5·7 years (mean: 2·5 years).

Following monotherapy with low doses of hCG replacement, two (additional) previously prepubertal patients were recognised as having CDGP and not HH: rising LH serum levels and pubertal testicular growth were observed in these subjects. These patients were not included in the study, and hCG replacement was ceased.

In group A, three patients discontinued hCG/rFSH replacement; one patient with congenital multiple pituitary hormone deficiencies (MPHD), and 2 patients with congenital normosmic HH (CHH). Four patients had not yet reached the therapeutic end-points at the time of evaluation of the study, leaving 27 participants in group A. Twenty-three group A boys provided semen samples for final assessment. In group B, three patients withdrew from the study: one patient with MPHD after tumour surgery, one patient with congenital MPHD and one patient with CHH, leaving 23 group B participants. Nineteen group B adolescents provided semen samples for final assessment.

Puberty Induction

In all group A boys, pubertal virilization to Tanner stage V occurred without major adverse side effects. Mild gynaecomastia (Tanner B2–3) was observed in four subjects; severe acne did not occur. Pubertal growth from a mean pretreatment height of 168 ± 10 cm to a mean final height of 181 ± 8 cm, appropriate for mid-parental target height (180 ± 6 cm), was documented. Bone age matured from 14 ± 1·4 to 17 years. Adolescents in group B grew from 176 ± 9 cm (bone age pretreatment: 16·7 ± 0·7 years) to 178 ± 8 cm (parental target: 177 ± 4 cm). Pubertal T levels were reached after 6 ± 3 months of hCG treatment in group A and after 4 ± 3 months in group B.


Final Bi-testicular Volumes (BTVs): Gonadotropins were administered for 24 ± 7/22 ± 6 months in group A/B, respectively, until cessation of testicular growth.

BTVs rose from 5 ± 5 ml at baseline to 10 ± 8 ml on hCG alone and to 34 ± 3 ml after combined treatment with hCG and rFSH in group A and from 5 ± 3 to 8 ± 5 to 32 ± 3 ml in group B (Figure 1). Changes in testicular sizes in response to gonadotropin treatment in the different HH patient subsets of both groups (A/B) are detailed in Table 1.

Figure 1.

(a) Testicular growth over time in response to gonadotropin replacement with hCG and rFSH in prepubertal (group A) and testosterone-virilized (group B) adolescents with HH. (b) Final bi-testicular volumes (BTV) over time to final BTV from start of hCG therapy in group A and B. The dashed lines indicate the lower limit of normal BTVs (24 ml); the black lines indicate mean final BTVs and mean duration from start of hCG until attainment of final BTVs. A: Mean final BTVs: 34 ± 16 ml with 74% (20/27) of patients reaching a normal BTV ≥24 ml; mean duration from start of hCG therapy until final BTV: 24 ± 7 months. B: Mean final BTV: 32 ± 16 ml with 70% (16/23) of patients reaching a normal BTV≥24 ml; mean duration from start of hCG until final TV: 22 ± 6 months; all P > 0,05; n.s.

Sperm Concentrations (SCs) and Other Semen Parameters

Sperm were found in 91% (21/23) of group A vs 95% (18/19) of group B patients. Two group A patients (one with KS and one with CHH, both with initial BTVs of 6 ml) and one group B patient (with CHH, with initial BTVs of 2 ml) remained azoospermic. Only one of them had a history of bilateral cryptorchidism. Successful microscopic testicular sperm extraction (mTESE) was performed in the latter patient with KS, and mTESE samples were cryostored for potential future use in assisted reproduction. The other two azoospermic patients did not wish to undergo surgery for sperm retrieval.

SCs plateaued after 31 ± 6/30 ± 7 months from start with hCG and after 25 ± 9/25 ± 9 months of combined hCH/rFSH treatment, in A/B, respectively (Figure 2). Final SCs were normal (≥15mill/ml) in 61% (14/23) in group A and 32% (6/19) in group B, and mean SCs were non-significantly higher in A (40 ± 73 mill/ml) than in B (19 ± 38 mill/ml; P = 0·07).

Figure 2.

(a) Increase in sperm concentration over time in response to gonadotropin replacement with hCG and rFSH in previously prepubertal (group A) and previously testosterone-virilized (group B) adolescents with HH. Baseline azoospermia was assumed in all boys of group A, as pretreatment semen analysis was not possible due to psycho-sexual immaturity. (b) Final sperm concentration over time of rFSH treatment until a plateau in group A and B was reached. A: Mean final sperm conc.: 40 ± 73 mill/ml, with 61% (14/23) of patients reaching a normal sperm concentration ≥15 mill/ml; mean duration from start of rFSH until final sperm concentration: 25 ± 7 months. B: Mean final sperm concentration: 20 ± 9 mill/ml, with 32% (6/19) of patients reaching a normal sperm conc. ≥15 mill/ml (A vs B: P = 0·007); mean duration from start of rFSH therapy until final sperm concentration: 25 ± 9 months (A vs B: n.s.).

In group B, first sperm were found 15 ± 7 months after start of hCG administration and 11 ± 6 months after initiation of FSH treatment. The previously prepubertal boys required 2·0 ± 1 years of gonadotropin replacement before 'feeling mature enough' to provide a semen sample for laboratory analysis. In this group, first sperm were documented after 21 ± 10 months of hCG and 17 ± 7 months of combined hCG/rFSH administration.

Mean final ejaculate volume was slightly lower than the WHO normal value in group A (A: 1·3 ± 0·2 ml; B: 3·8 ± 0·8 ml; normal: ≥1·5 ml). Final total sperm counts were not significantly different (A: 60 ± 160 mill; B: 42 ± 55 mill; normal: ≥39 mill; P = 0·43), neither was progressive motility (A: 43 ± 18%; B: 42 ± 14%; normal: ≥32%) nor sperm morphology (A: 4 ± 3%; B: 3 ± 2%; normal: ≥4%).

Quality of Life (QoL) Before and After Gonadotropin Replacement

At baseline, health-related QoL scores (ILK) (Figure 3) were at the lower limit of the normal range in both groups (median: A: 74%; B: 75%). Health-related problem scores (ILK) (on a scale from 1 to 7) were comparable in A and B, but showed large intra-individual variations (median (range): A: 2·0 (0–7); B: 1·0 (0–5)). Group B adolescents had significantly higher baseline depression scores (DIKJ) than group A boys, with less variation (median (range): A: 34 (0–100); B: 50 (35–73); A/B pretreatment P = 0·03).

Figure 3.

Box and whisker plots showing medians, interquartile ranges (boxes) and ranges (whisters) for results of QoL questionnaires. These were filled in by n = 26 group A boys before gonadotropin treatment and again (n = 15) after puberty induction with gonadotropins, while n = 17 testosterone-virilized adolescents answered all questions prior to gonadotropin substitution and n = 13 of these again following gonadotropin replacement.

After gonadotropin treatment, QoL scores were significantly higher than pretreatment in both groups (A: 86%; pre/post P = 0·03; B: 82%; pre/post P = 0·03), accompanied by significantly lower post-treatment problem scores (A: 0; pre/post P = 0·04; B: 0; pre/post P = 0·01) and lower depression scores in group A (A: 19 (3–94); pre/post P = 0·05), while depression scores in group B had not significantly changed (B: 43 (35–66)). When comparing post-treatment scores between the two groups, depression scores of group B were significantly higher than those of A (post-treatment A/B P < 0·01), while post-treatment QOL and problem scores were not different.

There were no significant changes in response to gonadotropin treatment in both groups with respect to scores for adaptive and maladaptive strategies of emotional regulation (FEEL-KJ) (data not shown).

Self-reported satisfaction (on a scale from -2 to +2) concerning testis size was considerably higher in both groups after gonadotropin treatment (mean pretreatment scores A/B: −1·2/−1·2 vs +1·3/+1·1 post-treatment in A/B, respectively). Satisfaction concerning masculinity increased more in group A (A: from −0·5 to +0·9 vs B: from −0·2 to +0·2) (Figure S1).

Analysis of Baseline Variables Potentially Influencing Therapeutic Response to Gonadotropins

Causes of HH. With respect to adherence to treatment, which was better in group A (Figure S2), there was a trend towards higher final BTVs and higher final sperm concentrations achieved by patients with childhood-acquired causes of HH (MPHD after tumour surgery and CHH with pubertal arrest), compared to those with congenital causes (Kallmann syndrome, CHH with absent puberty, congenital MPHD (group A: final BTV HH acquired: 50 ± 21 ml, vs HH congenital: 34 ± 14; P = 0·07; final sperm concentration acquired HH: 94 ± 81mill/ml vs congenital HH 43 ± 17; P = 0·3) ( Table 1 ).

Undescended Testes. Final sperm concentrations of patients with bilateral cryptorchidism at birth were lower than for those with unilateral or no undescended testis (Figure 4a). Of the whole cohort of adolescents, 15 subjects (45% of group A and 32% of group B) ( Table 1 ) had a history of undescended testes, 4 unilateral and 11 bilateral at birth. All had orchidopexy before the age of six, most of them before the age of two.

Figure 4.

Predictors of response to hCG/rFSH treatment: presence of undescended testes at birth, baseline testicular volumes, baseline inhibin B and AMH serum levels. (a) Mean sperm concentrations for patients with bilateral/unilateral/no cryptorchid testes at birth were 4 ± 6/38 ± 46/44 ± 74 mill/ml, respectively. (b) Correlation of initial ultrasound bi-testicular volume (BTV) with final ultrasound BTV in both groups (A+B); Spearman r: 0·56/0·57; P < 0·001. (c) Correlation of baseline inhibin B serum levels with final BTV (Prader) in both groups; r: 0·51/0·57; P < 0·01. (d) Correlation of baseline AMH serum levels with final sperm quality (sperm concentration and total sperm count) (r: 0·42/0·41; P < 0·02) in both groups.

Initial Testicular Size. Initial BTVs (by ultrasound) correlated with final ultrasound BTV on gonadotropin replacement in both groups (r: A:0·56/B:0·57; P < 0·001) (Figure 4b). BTVs also correlated with final sperm concentrations (and final total counts) in group A (r: 0·51; P = 0·025), but not in group B.

Markers of Sertoli Cell Maturity (Inhibin B, AMH). Baseline inhibin B levels before gonadotropin replacement correlated with final BTVs in both groups (r: A: 0·51/B: 0·57; P < 0,01) (Figure 4c). A significant correlation with final SCs (r: 0·64; P = 0·002) and final total count (Spearman r: 0·73; P = 0·0002) was found only in group A.

There was a significant correlation of baseline AMH and final SCs and total sperm count (r: A: 0·42/B: 0·41; P < 0·02) in both groups (Figure 4d) and a significant correlation with final BTVs in group A (r: 0·40; P = 0·047).

Kinetics of Sertoli Cell Markers During Gonadotropin Replacement

Group A had baseline inhibin B levels of 39 ± 35 pg/ml, rising to 76 ± 61 pg/ml on hCG alone with maximum inhibin B levels on hCG/rFSH of 177 ± 118 pg/ml (normal adult range: 125–330 pg/ml) ( Table 1 ). Mean serum inhibin B levels were lower at baseline in group B with 27 ± 23 pg/ml (P = 0·02). HCG-stimulated levels (44 ± 28 pg/ml) and maximum inhibin B levels on hCG+rFSH (122 ± 73 pg/ml) were not significantly different from group A.

Mean baseline AMH levels were not significantly different between groups A/B (respectively, 31 ± 32 vs 20 ± 13 ng/ml), declined to 17 ± 19/16 ± 15 ng/ml on hCG, reaching minimum levels on hCG/rFSH of 5·8 ± 4·3/3·7 ± 2·7 ng/ml (normal adult range: 1·3–14·8 ng/ml) ( Table 1 ).