Patients and methods

The UK Database of Ambiguous Genitalia and Intersex Disorder was set up to investigate AIS and consisted mainly of cases of male undermasculinisation where other causes had been excluded. The diagnosis was entered as reported by the clinician. Among the 706 entries in the register, there were 395 cases of 46XY male undermasculinisation: eight with 5α-reductase deficiency, three with a defect in testosterone biosynthesis, seven cases of Drash syndrome, 20 cases of abnormal testes, 210 cases of AIS, and 147 cases where the diagnosis was unclear. The diagnosis of abnormal testes was reached on the basis of abnormal histology in 13 of 20 cases (table 1); in the remainder, it was based on inadequate response to hCG stimulation. The diagnosis of 5α-reductase deficiency was confirmed by analysis of urinary steroids or detection of a mutation in the 5α-reductase type II gene. Cases of confirmed AIS (n = 210) have been reported previously4 and were excluded from this study. Cases where investigations had been performed only after gonadectomy were also excluded.

Data analysed included age at investigation, dose of hCG used in the stimulation test, and serum testosterone (nmol/l), DHT (nmol/l), luteinising hormone (LH) (U/l), and follicle stimulating hormone (FSH) (U/l) concentrations. The results of these tests were available in a variable number of cases (table 2). Hormonal analyses were performed by standard immunoassays in endocrine laboratories participating in a national external quality control service (UKNEQAS). Interassay variability for the assays ranged between 10% and 20% (data courtesy of UKNEQAS). Assay results are presented as medians and 90th centiles (P10, P90); intergroup comparison was performed by the Wilcoxon signed rank test. Testosterone, LH, and FSH values were compared with those described in normal children.3

Results

TESTOSTERONE, DHT, AND hCG STIMULATION

Basal serum testosterone reached peak concentrations at around 2 months of age, declining to low concentrations in childhood and rising again postpubertally. Twenty nine of 50 measurements performed in the first year of life were within the age related reference range (fig 1). Of the 66 cases where an hCG test was performed, 25 (38%) and 16 (24%) cases received hCG 1000 U and 1500 U on three consecutive days, respectively. Thirteen different protocols were used in the remaining 24 cases, ranging from 200 U on three consecutive days to 1500 U on five consecutive days. The median rise in testosterone with 1000 U hCG stimulation was 2.6 times the basal concentrations (P10, P90 = 1, 29.4). The median testosterone increment was 4.2 nmol/l (P10, P90 = 0, 14.5). In comparison, the median rise in testosterone with 1500 U hCG was 6.5 times the basal level (P10, P90 = 1, 28) and the median testosterone increment was 3 nmol/l (P10, P90 = 0, 19.2). There were no significant differences in the results obtained using the two regimens. In the two largest groups of patients (abnormal testes and unknown), the rise in testosterone and testosterone increment were not related to the age at which tests were performed. However, there was a significant relation between these two parameters and the basal concentration of testosterone in the group of patients where the diagnosis was unclear (r = 0.3, p < 0.05). The response to hCG stimulation was most notable in 5α-reductase deficiency with peak median testosterone of 15 nmol/l (range 3.1–38; fig 2). The basal and stimulated testosterone:DHT ratios were higher in those cases diagnosed as 5α-reductase deficiency (table 3). The highest value in the remainder of the cases was 16.8. There was no clear age related change in testosterone:DHT ratio before or after hCG stimulation (fig 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Basal plasma testosterone concentrations in cases of male undermasculinisation. Shaded area represents normal reference range in the first year of life.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Serum testosterone concentrations before and after hCG stimulation in different diagnostic groups.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Plasma testosterone:DHT ratio before and after hCG stimulation.

GONADOTROPHINS AND LHRH STIMULATION

Basal LH and FSH were observed to be highest during infancy. Over the first year of life, 10 of 20 LH measurements and 13 of 16 FSH measurements were within the normal reference range. Two of these 10 normal LH values and five of 13 normal FSH values were observed in cases with abnormal testes. In the latter group, the median basal concentrations of LH and FSH were 2 U/l (P10, P90 = 1.1, 7.3) and 6.4 U/l (P10, P90 = 2.6, 29.7), and peak stimulated LH and FSH were 21 U/l (P10, P90 = 4.5, 31.4) and 28 U/l (P10, P90 = 9.7, 79), respectively. An exaggerated gonadotrophin response to LHRH stimulation was observed during mid-childhood in those children where the diagnosis was unclear and in infants as well as children with abnormal testes (table 4).

Discussion

The results of this study indicate that while testosterone and gonadotrophin concentrations are often determined in the undermasculinised male, hCG and LHRH stimulation tests are performed in less than half the cases. Basal testosterone and gonadotrophin concentrations followed the pattern seen in normal children with an early peak at 1 to 2 months of age during infancy followed by low concentrations in childhood.3 6-8 The mechanisms underlying the transient postneonatal surge in testosterone are ill understood. The concurrent rise in gonadotrophin concentrations suggests that it is centrally mediated, which can be abolished in primates with a gonadotropin releasing hormone (GnRH) agonist.9 The decline in testosterone and LH in childhood is thought to be due to increasingly sensitive negative feedback in the pituitary–gonadal circuit as well as a central neuroinhibitory mechanism on release of GnRH.3 10 Our previous observations as well as those reported in this study show that the neonatal surge of testosterone and gonadotrophins is generally preserved in children with male undermasculinisation with or without evidence of androgen insensitivity. This suggests that this phenomenon may not be due to changes in sensitivity of testosterone feedback.4

The hCG stimulation test is a valid indicator of the presence of functional testicular tissue.11 There is minimal spontaneous testosterone secretion in young boys after the first few weeks of life, so that stimulation of Leydig cell activity with gonadotrophins is needed when investigating prepubertal subjects.12 A two to threefold rise in testosterone is generally considered to be normal.11 13 A positive testosterone response to a single hCG injection when used as a screening test in the evaluation of hypogonadism is conclusive.14 We have recently shown that the test can be negative in proven cases of AIS.4 There was considerable variation in the dosage and duration of hCG stimulation documented in our study, ranging from 200 to 5000 U and one to five days, respectively. The lack of a relation between the dose of hCG and the subsequent testosterone response suggests that the unit dose may not be critical. The magnitude of testosterone response appears to be related to basal testosterone concentrations and, possibly, Leydig cell mass.

Basal and post-hCG stimulated DHT measurements were not often performed in this series of cases. The ratio of testosterone to DHT following hCG stimulation is regarded as a useful test for 5α-reductase deficiency. The typical pattern is a normal or raised concentration of plasma testosterone with decreased DHT and increased testosterone:DHT ratio, particularly after hCG stimulation.15 However, the finding of a testosterone:DHT ratio of 12.5 following hCG stimulation in one infant with 5α-reductase deficiency suggests that the ratio may not be completely specific. This value is within the testosterone:DHT range of 1.5–17 reported in normal male infants.16 A prolonged hCG stimulation test may be a better discriminator than the standard test17 as well as perhaps urinary steroid analysis.18 Decreased 5α-reductase activity in genital skin may also help in the diagnosis,19 although the normal range can extend to very low concentrations.20 Phallic growth in this disorder should typically respond to DHT cream application which may be an indirect indicator of the diagnosis.18 Direct molecular genetic testing is an alternative approach that is likely to be diagnostic of 5α-reductase deficiency.17

The variable patterns of basal LH and FSH with age were similar to those observed in normal children and children with the syndrome of gonadal dysgenesis.8 10 In this study some cases were reported as having “abnormal testes” based solely on the results of an hCG stimulation test as well as their clinical phenotype and karyotype. Our analysis of hCG stimulation tests in AIS showed that some proven cases of this syndrome may have a poor testosterone response so that it is unwise to use this solely for the diagnosis of gonadal dysgenesis.4 The data in the group of children categorised as having abnormal testes did not show severely raised basal gonadotrophin concentrations, even during infancy. However, an exaggerated LH response following LHRH stimulation was observed. Basal and stimulated FSH concentrations were higher than LH, suggesting that FSH concentrations as an indirect measure of Sertoli cell function may be a more useful indicator of testicular dysfunction. Serum inhibin B concentrations may be decreased in clinical conditions associated with impaired Sertoli cell function and have a reciprocal relation with serum FSH concentrations.21 22 A recent study of serum anti-Mullerian hormone measurements in a large group of intersex patients has proposed this Sertoli cell product as a useful marker to distinguish gonadal dysgenesis from isolated defects in androgen production or action.23

A register based study has the problem of data reporting from multiple and variable sources. Nevertheless, this disadvantage was more than compensated for by analysing data from a large number of cases of male undermasculinisation where a number of appropriate tests had been performed and reported. The results of this study would have been even more meaningful if a greater number of cases had been thoroughly investigated. Hormone assays were performed in a number of laboratories, but each was a participant in a national external quality control scheme. Nevertheless, data from this study do help with interpreting the results of investigations performed in cases of male undermasculinisation which are not due to AIS. While a raised testosterone:DHT ratio following hCG stimulation is generally diagnostic, a normal ratio does not necessarily exclude 5α-reductase deficiency. An acute LHRH stimulation test may be useful to evaluate cases of male undermasculinisation presenting in mid-childhood. Most cases of undermasculinisation, however, currently remain undiagnosed.

As our understanding of the pathophysiology of male undermasculinisation improves, the practice of evaluating the pituitary–gonadal axis in the undermasculinised male using uniform protocols needs to become more thorough if the diagnostic yield in this condition is to be improved. We would recommend the protocol described in the paper by Viner et al.24These guidelines, however, are not based on any prospective study of the true value of the investigations detailed. Such a study would require the cooperation of general paediatricians and paediatric endocrinologists nationwide. We would also recommend establishing early contact with a paediatric endocrine centre so that appropriate investigations can be performed and interpreted as early as possible.