Introduction

Premature pubarche (PP) is defined as the development of pubic and/or axillary hair before the age of 8 years in girls and 9 years in boys, while puberty is otherwise normal. Pathological puberty resulting in PP is mainly due to non-classical congenital adrenal hyperplasia. PP is considered to be the result of premature adrenarche provided all other disorders have been ruled out, and it is followed by a normal central hypothalamic-pituitary puberty. The causes underlying the onset of physiological adrenarche are unknown, and the pathophysiology of premature adrenarche is still not clear. Although premature adrenarche was considered to be a benign disorder, recent studies have pointed out that there may be an association between intrauterine growth retardation (IUGR) and premature adrenarche, hyperinsulinaemia and functional ovarian hyperandrogenism in girls [7].

This aim of this investigation was to obtain a clearer picture of the physiology of adrenarche by evaluating the clinical-laboratory features of PP.

Subjects and methods

Subjects

A total of 216 consecutive children seen for PP by the same physician (RB) between September 1992 and September 2001 were studied. Pathological puberty resulting in PP was excluded by physical examination and normal basal prepubertal plasma concentrations of adrenal hormones. There was no sign of central precocious puberty as none of the girls had breast development and none of the boys had any increase in testicle size. None of the children had signs of systemic virilisation (clitoromegaly or phallic enlargement) or of hypercortisolism (normal blood pressure). The children were classified according to their sex and age of onset of PP: PP occurred before 2 years in 21 girls and one boy, between 4 and 7.9 years in 162 girls and between 5 and 8.9 years in 28 boys. It occurred between 2 and 4 years in only four girls.

Methods

Height was expressed as standard deviation (SD) and body mass index (BMI) as z-score compared with the data for French children of the same chronological sex and age [15, 16]. Birth weight and gestational age were routinely recorded on the health records of 204 children and expressed as percentiles [10]. The pubic hair development was assessed according to Tanner by the same physician [11], as were the bone ages for all children [5]. Pathological puberty was ruled out by normal basal prepubertal plasma concentrations of testosterone (n=210), Δ-4-androstenedione (n=136), 17-hydroxyprogesterone (n=213), 11-deoxycortisol (n=153) and adrenocorticotropin (ACTH, n=66). The 70 children seen most recently also underwent an ACTH test (0.25 mg intramuscularly), which showed normal increases in 17-hydroxyprogesterone and 11-deoxycortisol [12]. Plasma dehydroepiandrosterone sulphate (S-DHEA) was measured in 162 children. Basal blood samples were collected at any time of day, while the ACTH test was performed at 8:00 am after an overnight fast.

The plasma concentrations of the hormones were measured by commercial radioimmunoassays: 17-hydroxyprogesterone (OHP-CT, Cis Bio, Gif sur Yvette, France), Δ-4-androstenedione after extraction with isooctane (Immunotech, Marseille, France), testosterone after extraction with ether (Testo-CT2, Cis Bio), ACTH (IRMA, ACTH, Immunotech) and S-DHEA (Immunotech). S-DHEA was compared to normal concentrations for age and Tanner stage [4].

Results are expressed as means ± SD. Data were compared between groups using the Student’s t test. Correlations between variables were analysed using the Pearson correlation coefficient. The observed birth weights were compared to the theoretical distributions using a Chi-squared test. A value of P <0.05 was considered statistically significant.

Results

Clinical features

PP began between the ages of 7 and 7.9 years in 75 girls aged >4 years (46%), and between the ages of 8 and 8.9 years in 13 boys aged >5 years (45%). The BMI (available for all except two girls) was ≥+2SD in two children aged <2 years (9.1%), in 52 girls aged >4 years (32.5%) and in 13 boys aged >5 years (46.4%). The children <2 years had significantly lower BMI z-scores than the older ones (P=0.03) and fewer were obese (P=0.02, Table 1). The obese children aged >4 years were taller than the non obese ones (P=0.004) and their bone age was more advanced (P=0.0009). Of the 216 children, 42 (19%) had increased body hair. The extent and distribution of both body and pubic hair did not differ with sex or age. Pubic hair was stage 1 in five children (2%), who all had axillary hair alone; it was stage 2 in 121 children (56%), stage 3 in 81 (38%) and stage 4 or 5 in nine (4%, Table 1). Axillary hair developed before pubic hair in 14 children (6.5%), who were all girls aged >4 years. Of the children aged >4 years, 14% suffered from acne, while none of those under 2 years did.

Table 1 Clinical and laboratory features of 216 children with PP
Full size table

Laboratory features

The children aged <2 years had above-normal plasma Δ-4-androstenedione and S-DHEA concentrations for their sex and age, but these concentrations were lower than those expected in Tanner stage 2 (Fig. 1). They were lower than in the children >4 years (p<0.001). The S-DHEA concentrations in children >4 years were slightly higher than expected for their Tanner stage. They were higher in obese girls than in lean ones (P=0.009), with a positive correlation between S-DHEA and BMI z-score (P=0.004) (Fig. 2). This correlation persisted after adjustment for the time from the onset of PP to consultation and for the Tanner stage. The Δ-4-androstenedione concentrations were higher in obese girls (0.52±0.33 ng/ml) than in the non-obese ones (0.44±0.27 ng/ml), but the difference was not significant. The obese and non-obese boys had similar plasma hormone concentrations and there was no correlation between the S-DHEA and the BMI z-score. There was no correlation between the BMI z-score and the other hormones.

Fig. 1
figure 1

S-DHEA levels according to the sex and Tanner stage of the children. Open circles (children <2 years); filled circles (children >4 years). The straight line shows the mean S-DHEA concentration for the Tanner stage [4]

Full size image
Fig. 2
figure 2

Relationship between the plasma S-DHEA concentration and BMI for 115 girls aged 4 to 7.9 years. Open circles for lean children; filled circles for obese children

Full size image

Birth weight

The distribution of the birth weights by percentiles was compared to the expected distribution (Table 2). The prevalence of IUGR, defined as a birth weight ≤10th percentile, was 13.6% in children <2 years, 18.7% in girls >4 years and 18.5% in boys >5 years. This percentage was significantly higher in girls >4 years (P=0.02) than the 10% expected for the general population. There was a tendency towards a low birth weight in girls >4 years, with 67.7% of the children having a birth weight below the median for normal children (P<0.01 for the global comparison of the observed to the expected distributions).

Table 2 Distribution of birth weights (%) for the PP children and the expected distribution
Full size table

Discussion

This study was performed on a large population of children who were seen consecutively by the same physician. It shows that the cause of PP occurring in children aged less than 2 years is probably different from that occurring in older ones. It also shows that many of those who were older than 4 years at the onset of PP were obese, with a positive correlation between their plasma S-DHEA concentration and the BMI z-score in the girls.

We have examined the evaluation needed to exclude pathological puberty expressing as PP to determine if a normal physical examination and basal prepubertal plasma concentrations of adrenal hormones is sufficient, or whether an ACTH test is necessary. We performed an ACTH test on the 70 children seen more recently because one boy whose plasma 17-hydroxyprogesterone concentrations were normal (below 1.5 ng/ml) at 8 and 9 years showed a rapid bone age progression when he was 11 years old due to non-classical congenital adrenal hyperplasia. All the children in the present study had 17-hydroxyprogesterone concentrations lower than 1 ng/ml, except for 9, in whom it was 1 to 1.2 ng/ml. Siegel et al. [17] concluded that only the response to an ACTH test allowed the accurate classification of children with PP. A normal response to an ACTH test excludes congenital adrenal hyperplasia and allows the follow-up to be continued by the family physician. But girls aged 7 to 7.9 years are more likely to have PP than a pathological puberty, so that normal measurements of basal plasma 17-hydroxyprogesterone, Δ-4-androstenedione and testosterone concentrations may be sufficient.

As expected, a large proportion of the children with PP were girls aged 4 to 7.9 years; PP was rare in boys (13.4%). Balducci et al. [1] found that 28/135 (20.7%) of children with PP were boys. The PP of children aged less than 2 years (10.2%) had different clinical-laboratory features than those of older ones. The young children were not obese and their plasma concentrations of S-DHEA and Δ-4-androstenedione were lower than those expected in Tanner stage 2. Their PP therefore may not be simply a “premature adrenarche”, and the gonads may be involved in their development of pubic hair.

The children aged >4 years had the clinical features that have already been described, a height above the mean and an advanced bone age [6]. These children tended to be obese (32.5% of the girls and 46.4% of the boys). Some publications have reported frequent obesity in small groups of girls with PP [3, 18, 21]. Ibanez et al. [8] reported that girls with PP had excess total body and central fat mass throughout all pubertal stages. A study of PP in boys reported that all the children had a normal BMI [13]. This is therefore the first report of obesity being common in a large population with no recruitment bias because it concerns 216 children seen consecutively over a period of 9 years by the same physician. The obese and non-obese girls and boys had different clinical features; the obese children were significantly taller and had greater bone age advance than the non-obese children [19].

Our study confirms the biological features of PP, which are a moderate increase in the plasma concentrations of adrenal hormones, comparable to or slightly higher than those found in normal early pubertal children [9, 20]. The obese girls >4 years of age had significantly higher plasma S-DHEA concentrations, and a tendency towards higher Δ-4-androstenedione concentrations than the non-obese girls. The plasma S-DHEA concentration was correlated with the BMI z-score, but this was not true for the other adrenal hormones. This correlation is not particularly strong (r=0.27), but it is very significant (P=0.004). Clearly, the BMI cannot completely account for the difference in the S-DHEA concentration. S-DHEA is the most abundant circulating adrenal androgen and it does not show the circadian variation that is seen for Δ-4-androstenedione. This could explain why we found no significant association between the BMI and the concentration of the Δ-4-androstenedione. The correlation between S-DHEA and BMI z-score suggests that being overweight influences the onset of adrenarche. This has already been suspected in physiological adrenarche. A longitudinal study of normal children showed that the greatest increase in urinary S-DHEA occurred during the period of the greatest rise in the BMI [14]. Pathophysiological studies have emphasised the role of hyperinsulinaemia in the onset of premature adrenarche, possibly by decreasing the concentrations of insulin-like growth factor binding protein 1 and so increasing free insulin-like growth factor 1 [18, 21]. This is in agreement with our findings because obesity may be the first event of this cascade causing an increase in insulin. Obese children do indeed have significantly higher plasma insulin concentrations than controls [19]. The increased BMI is also associated with increased secretion of leptin, and this may interfere with the onset of the adrenarche [3]. Leptin could be involved in the onset of adrenarche by controlling the CYP17 phosphorylation [2]. Unfortunately, the concentrations of leptin, insulin and insulin-like growth factor binding protein 1 could not be measured over the whole period of the study. This relationship between S-DHEA and the BMI z-score was not found in boys, even after adjustment for the delay before consultation.

Intrauterine malnutrition has been incriminated in premature adrenarche. We compared the birth weights, expressed in percentiles, to the expected distribution. There was a shift towards a lower birth weight in girls aged 4 to 7.9 years, with IUGR being significantly more common than in the general population. This is in agreement with findings for other populations, showing that IUGR, or even a less pronounced prenatal growth restriction, can lead to PP [7]. This must now be confirmed by examining larger populations. The distribution of birth weights in the children aged less than 2 years is similar to that for the general population. The small number of boys aged 5 to 8.9 years prevented the percentage of IUGR appearing significantly different from the 10% seen in the general population, although it was 18.5%. A previous study of boys also showed no relationship between PP and reduced fetal growth [13], pointing to a sexual dimorphism in this disorder, whose physiological basis is unclear.

We therefore conclude that the BMI seems to influence the onset of adrenarche.