High DHEAS Is Associated With Earlier Pubertal Events in Girls But Not in Boys.

CONTEXT: Premature adrenarche (PA) has been associated with increased metabolic risk.
OBJECTIVE: To describe the risk of precocious thelarche (PT; <8 years), pubarche (PP; girls <8 years, boys <9 years), and gonadarche (PG; <9 years) in children with high dehydroepiandrosterone sulphate (DHEAS [HD]) vs those with normal DHEAS (ND). SETTING AND INTERVENTION: Longitudinal Chilean cohort (n = 1052, 49.9% girls). Annual clinical examination including secondary sex characteristics by Tanner staging. Logistic regression models were adjusted by age and BMI. MAIN OUTCOME: Assess the relationship between DHEAS and premature thelarche, gonadarche, and pubarche in both sexes.
RESULTS: At age of DHEAS determination, overweight/obesity was present in 44.3% of boys and 42.9% of girls. Incidences of any precocious event were observed in 17.2% of boys and in 25.4% of girls, presented as 8.7% of PG and 8.5% of PP in boys and as 21.3% of PT and 4.1% of PP in girls. In crude and adjusted models in boys, HD did not increase the risk of earlier pubertal events. Conversely, girls with HD had a 2.6 times greater risk of early thelarche and a three times greater risk of early pubarche compared with girls with ND concentrations.
CONCLUSION: In Chilean adolescents, precocious events of pubertal development were in line with the worldwide secular trend of earlier sexual maturation. HD was only associated with PT and PP in girls. Continuous follow-up of this cohort is a unique opportunity to prospectively address and analyze the interrelationships among HD, early growth, and adiposity as determinants of gonadarche, pubertal rate/sequence progression, and ovarian function.

Adrenarche is a gradual maturation process of the adrenal zona reticularis, the inner most zone of the adrenal cortex, beginning during the first 5 years of life [1, 2], which results in increased secretion of the adrenal androgen precursor dehydroepiandrosterone (DHEA) and its sulfate ester dehydroepiandrosterone sulphate (DHEAS). Adrenarche is clinically evident ~2 years before the onset of puberty. The initiation of adrenarche is independent of gonadal axis maturation [3], and the physiological control of this event, as well as the significance for human prepubertal development, remains unknown.

Premature adrenarche (PA), initially described in 1952 [4], is defined biochemically as an increase in adrenal androgen (DHEA and DHEAS) above the age and sex specific normal reference range, before the age of 8 years old in girls and 9 years in boys. Traditionally, elevated levels are indicated by a DHEAS level above 1086 nmol/L (above average for 6 to 8 years), although within normal limits for early puberty, and minimal if any elevation of other androgens [5]. This process is clinically recognized by the presence of signs of androgen action, including adult-type body odor, oily skin and axillary and pubic hair growth. The prevalence of PA has been reported in only a few studies, and it seems to vary among different ethnic populations [6]. Most reports have focused on pubarche, onset of pubic sexual hair growth, although other signs of androgen action are more common [5]. In a recent Finish population sample of prepubertal children aged younger than 9 years, the serum DHEAS concentrations did not differ between the sexes, and levels for biochemical adrenarche were observed in 16.6% of girls and 18.4% of boys; however, the prevalence of any clinical sign of androgen action was higher in girls than in boys (26.1% vs 10.0%; P < 0.001). Similarly, in a Chilean longitudinal study of obesity, we also observed no differences in DHEAS concentration by sex at age ~6.8 years [7].

Early infancy weight gain has also been associated with increased metabolic risk [8], earlier puberty [9, 10], and high DHEAS (HD) [6]. We described that obese children had twice the risk of HD and that IGF-I concentrations and not leptin were also associated with HD, independent of obesity [7].

A dimorphic effect of obesity in adrenal and gonadal function has been postulated. Obese prepubertal children (boys and girls) and only obese pubertal girls have increased testosterone and DHEAS levels, whereas androgen concentrations are not different between obese and normal weight pubertal boys [11]. In a study of 1066 children (52% obese), a significant interaction between sex and obesity for prediction of pubertal development was observed, but in the opposite direction [12]. Nevertheless, this latter observation has not been confirmed in other studies where obesity has been related to earlier pubertal events in both sexes [13, 14].

Therefore, our aim was to assess the relative contributions of DHEAS concentrations at 7 years and premature thelarche, gonadarche, and pubarche in both sexes. For this purpose, we took advantage of a previously described cohort of children enrolled in the Growth and Obesity Chilean Cohort Study, a longitudinal study of Chilean children with birth weight between 2500 and 4500 g.

1. Subjects and Methods

We performed a longitudinal study within the Growth and Obesity Chilean Cohort Study with the primary aim of assessing the association between early growth and development of adiposity and metabolic risk [15]. Briefly, the inclusion criteria were as follows: all singleton children between 3.0 and 4.9 years attending Chilean National Nursery School Council Program from the south east area of Santiago, Chile in 2006; gestational age 37 to 42 weeks; birth weight ≥2500 g and ≤4500 g (data retrieved from medical registries); and with no physical or psychological conditions that could severely affect growth. Eighty-five percent of the children agreed to participate (n = 1190). We did not observe substantial differences in age, sex, birth, and anthropometry at enrollment between participants and nonparticipants. Thereafter, annual evaluations have been conducted (anthropometry, body composition, skeletal and hormonal maturation, and metabolic/inflammatory markers) and, since 2009, twice yearly visits were initiated to assess secondary sex characteristics. Maternal age at menarche was self-reported by the mother.

In 2009, 1044 children of the original cohort were evaluated (~88%). For the current analyses, we excluded a total of 58 children, 36 for whom no blood sample was obtained due to difficulties and 22 children without sexual maturation data. Thus, our final sample size was 986 children. The study protocol was approved by the Institutional Review Board of the Institute of Nutrition and Food Technology of the University of Chile. Written informed consent was obtained from all parents or guardians of the children and assent from the children.

A. Clinical Adrenal and Pubertal Development

At age ~7 years, a single pediatric endocrinologist (VM) assessed breast and genital development by palpation and classified breast and testes according to Tanner stages [16, 17]. The same endocrinologist evaluated the presence of clinical signs of androgen action. Thereafter, every 6 months, secondary sex characteristics were evaluated by a single dietitian (same sex) trained specially for this purpose, with permanent supervision of a single pediatric endocrinologist (VM). In girls, breasts were evaluated by inspection and palpation according to the Tanner scale [16] and, in boys, genitalia were evaluated by palpation using the Prader orchidometer [17]. In both sexes, pubarche was assessed based on the Tanner scale. Concordance between the dietitian and pediatric endocrinologist was 0.9 for breast [18] and genitalia evaluation.

Age at thelarche, gonadarche, and pubarche onset was defined as the midpoint of two consecutive visits, the last visit without signs of sexual development (breast, genital, or pubic hair) and the first visit in which one of these signs were detected. Thus, we defined precocious thelarche (PT) when this event (midpoint between visits) appeared before the age of 8 years, precocious gonadarche (PG) as a testicular volume >3 mL at age before 9 years, and precocious pubarche (PP) when pubic hair appeared before the age of 8 years in girls or before 9 years in boys.

B. Anthropometric Measures

Weight and height were collected using standardized protocols (barefoot and light clothes) by two dietitians (one female, one male) with inter- and intrarater correlation coefficients over 0.80 for all measurements. Weight was measured with a portable electronic scale (Seca 770), with precision of 0.1 kg, and height was measured with a portable stadiometer (Harpenden 603) to the nearest 0.1 cm. A complete description of anthropometry methodology is found elsewhere [7].

C. Blood Sample at 7 Years

A trained nurse collected a fasting venous sample (8 to 12 hours) from the children upon arrival to the Institute of Nutrition and Food Technology outpatient clinic. Mothers were contacted the day before sample drawing to confirm the absence of fever (>37.5°C) or symptoms of acute infection in the children. Serum glucose concentrations were assessed by enzymatic colorimetric techniques (HUMAN, Gesellschaft fur Biochemica und Diagnostica, Wiesbaden, Germany) and serum insulin, using a radioimmunoassay kit (Linco Research, Inc., St. Charles, MO). Serum leptin and adiponectin were measured by commercial radioimmunoassay (Millipore, Merck, Darmstadt, Germany). Analyses were conducted at the Nutrition Laboratory of the Catholic University of Chile. This laboratory conducts daily assessments of the accuracy of the measurements by using UNITY quality control software (Bio-Rad Laboratories, Inc., Hercules, CA). DHEAS and IGF-I analyses were conducted at the Institute of Maternal and Child Research University of Chile. Serum DHEA-S was determined by competitive specific binding RIA supplied by Diagnostic System Laboratories (Webster, TX); intra- and interassay coefficient of variation (CVs) were 3.5% and 5.1%, respectively. Serum IGF-I was measured by using a standardized locally developed radioimmunoassay requiring sample extraction as a first step (sensitivity: 5 ng/mL; intra- and interassay CVs: 8.6% and 10.2%, respectively) [19].

D. Computed Indices

Body mass index (BMI) was estimated by dividing weight (kg) by height squared (m2). We estimated height-for-age and BMI-for-age based on the World Health Organization 2006 standards and the World Health Organization 2007 growth reference [20]. We defined obesity as BMI standard deviation score (SDS) ≥ 2. Biochemical indicator of HD defined as DHEAS concentration drawn at age 7 based on sample distribution (75th percentile); cutoffs were 42.0 µg/dL for girls and 45.1 µg/dL for boys [7]. The homeostasis model assessment-estimated insulin resistance (HOMA-IR) was calculated as fasting glucose (mmol/L) X fasting insulin (mU/mL)/22.5.

E. Statistical Analysis

Descriptive analysis (mean, SD, percentage) of anthropometric, hormonal characteristics, and sexual maturation data were performed by stratifying PA. Statistical differences among groups were assessed using χ2 and Student t test accordingly and were considered significant at a P value < 0.05.

We performed logistic regression models [odds ratio (OR) and 95% confidence interval (CI)] to assess the association between precocious adrenarche and PG, PT, and PP (separately by sex). The OR was adjusted in three different models. Model 1 was adjusted only by age at DHEAS sampling and birth weight. Model 2 was adjusted by birth weight, age, and BMI SDS at DHEAS sampling. Model 3 was adjusted by birth weight, age, BMI SDS, insulin, IGF-1, and leptin at DHEAS sampling.

As age at thelarche, gonadarche, and pubarche onset was defined as the midpoint of two consecutive visits, it could occur within that period of time. To assess this error, we performed a sensitivity analyses, moving the age of pubertal onset to the beginning and to the end of this period.

2. Results

We collected blood samples from 494 boys and 492 girls at 7 years old and characterized the participants as HD or normal DHEAS (ND). HD was present in 26.1% of the samples and we did not observe differences across sexes. At follow-up, mean age 12 years, 99% of the girls and 60% of the boys reached a pubertal stage ≥ Tanner II. A complete description of the anthropometric and hormonal-metabolic characteristics of the study population at 2009 (age at DHEAS sampling) and percentage of PP, PT, and PG are show in Table 1.

Table 1.

Anthropometric and Hormonal-Metabolic Description of Study Population at 2009 and Percentage of PP, PT, and PG

	Boys, n = 494	Girls, n = 492
	Mean (SD)	Mean (SD)
Age at 2009, y	6.8 (0.4)	6.8 (0.4)
Birth weight, kg	3.4 (0.4)	3.4 (0.4)
Birth weight z score	0.08 (1.9)	0.18 (1.06)
Height, cm	121.2 (5.5)	120.4 (5.4)
Height z score	0.1 (0.9)	0.2 (0.9)
Weight, kg	25.4 (4.9)	25.1 (4.7)
BMI z score	0.9 (1.3)	0.9 (1.1)
Obese (>2 SD), n (%)	106 (21.5%)	70 (14.2%)
Overweight (>1 SD), n (%)	219 (44.3%)	211 (42.9%)
Insulin, µIU/mL	5.4 (1.3)	5.6 (1.7)
HOMA	1.2 (0.3)	1.2 (0.4)
IGF-I SDS	−0.1 (0.7)	−0.5 (0.7)
IGF-I, nmol/L	24.6 (8.16)	23.1 (5.7)
Leptin, nmol/L	0.36 (0.24)	0.38 (0.26)
Adiponectin, µg/mL	18.4 (7.2)	17.5 (6.1)
Precocious Adrenarche, n (%)	126 (25.5%)	131 (26.6%)
PG, n (%)	43 (8.7%)	—
PT, n (%)		105 (21.3%)
PP, n (%)	42 (8.5%)	20 (4.1%)
PG and PP, n (%)	10 (2.0%)	—
PT and PP, n (%)		8 (1.6%)

A. Boys

Boys with HD showed no difference in the incidence of PG and PP compared with boys who had ND concentrations; however, they were more obese, overweight, and taller and had a higher IGF-I, and leptin (Table 2). Boys with PG had a higher birth weight, were taller, and had a higher BMI SDS at age ~7 compared with those who did not show earlier gonadal development (data not shown). Four out of 14 boys with PG and HD had PP (Table 2). In boys with PP, no differences in anthropometric, hormonal, or metabolic data were found at this age that could predict earlier development of pubic hair (data not shown in table).

Table 2.

Description of PG, PT, PP and Anthropometric and Hormonal Characteristics Stratified by Sex and With or Without Precocious Adrenarche

	Boys			Girls
	HD, n = 126	ND, n = 368	P Value	HD, n = 131	ND, n = 361	P Value
	Mean (SD)	Mean (SD)		Mean (SD)	Mean (SD)
PG, n (%)	14 (11.1)	29 (7.9)	0.3
PT, n (%)				45 (34.4)	60 (16.6)	<0.001
PP, n (%)	12 (9.5)	30 (8.2)	0.6	10 (7.6)	10 (2.8)	0.02
PG and PP, n (%)	4 (3.2)	6 (1.6)	0.28a
PT and PP, n (%)				4 (3.1)	4 (1.1)	0.07a
BMI, SDS	1.4(1.31)	0.79 (1.22)	<0.001	1.16 (1.09)	0.75 (1.03)	0.0002
Obesity, n (%)	42 (33.3)	64 (17.4)	<0.001	26 (19.8)	44 (12.2)	0.03
Overweight, n (%)	71 (56.3)	148 (40.2)	<0.001	69 (52.7)	142 (39.3)	<0.001
Birth weight, kg	3.39 (0.44)	3.45 (0.41)	0.1	3.31 (0.38)	3.39 (0.39)	0.05
Birth weight z score	−0.01 (1.14)	0.11 (1.08)	0.31	0.09 (1.04)	0.21 (1.06)	0.26
Height, SDS	0.35 (0.85)	0.06 (0.95)	0.001	0.3 (0.99)	0.15 (0.86)	0.11
Insulin, pM/L	39.8 (10.1)	37.2 (8.3)	0.2	39.0 (9.9)	38.8 (12.2)	0.81
HOMA-IR	1.25 (0.37)	1.19 (0.3)	0.1	1.24 (0.36)	1.23 (0.44)	0.76
IGF-I, nmol/L	26.20 (8.45)	24.04 (7.9)	0.01	24.4 (6.45)	22.55 (5.34)	0.002
IGF-I, SDS	0.07 (0.75)	−0.20 (0.66)	0.008	−0.32 (0.77)	−0.54 (0.69)	0.004
Leptin, nmol/L	0.4025 (0.26)	0.34 (0.23)	0.02	0.39 (0.26)	0.37 (0.26)	0.58
Adiponectin, ng/mL	17.91 (6.85)	18.52 (7.33)	0.4	17.52 (6.03)	17.48 (6.18)	0.96

Fisher test.

In crude and adjusted models in boys, HD did not increase the risk of earlier pubertal events (gonadarche and pubarche), even after adjusting for birth weight, BMI SDS, insulin, leptin, and IGF-I at age 7 (DHEAS sampling; Table 3). The results were similar after performing a sensitivity analysis, redefining age at puberty onset as the upper or lower limit of the interval of the two consecutive visits.

Table 3.

Crude and Adjusted OR and 95% CI Between HD and Precocious Adrenarche and PG, PT, and PP

	OR	95% CI	OR 1a	95% CI	OR 2b	95% CI	OR 3c	95% CI
Male PG	1.46	0.73–2.81	1.58	0.76–3.16	1.23	0.57–2.53	1.13	0.37–3.19
Male PP	1.18	0.57;2.34	1.39	0.63–2.93	1.22	0.54–2.6	1.5	0.46–4.53
Female PT	2.63	1.66;4.13	2.67	1.65–4.31	2.48	1.52–4.04	2.37	1.44–3.90
Female PP	2.9	1.16;7.23	2.87	1.08–7.53	2.58	0.94–6.98	2.47	0.90–6.69

Adjusted by age and birth weight.

Adjusted by model 1 + BMI SDS.

Adjusted by model 2 + insulin, leptin, and IGF-1.

B. Girls

In contrast, girls with HD had an approximately two times greater risk of early thelarche and approximately three times greater risk of early pubarche, compared with girls with ND concentrations. Similar to boys, girls were more obese and overweight and had a higher IGF-I concentrations at age 7 (Table 2).

Four out of the 45 girls who had PT and HD had PP (Table 2). Girls with PP were taller and had higher BMI SDS than those who did not present earlier development of pubic hair (data not shown). In contrast to boys, HD increased the risk of early thelarche significantly in both crude and adjusted models. Thirty-six girls had Tanner II breast at the moment of DHEAS determination [OR 1.83 crude model (1.07 to 3.07), model 1 OR 1.91 (1.08 to 3.30), model 2 OR 1.97 (1.11 to 3.45), and model 3 OR 1.87 (1.04 to 3.32)]. HD also determined earlier pubarche in girls, but the association was significant only in the crude and adjusted model by age and birth weight. However, the magnitude of the association remained similar after further adjustment (the results were similar after the sensitivity analysis).

3. Discussion

In this longitudinal population-based study, higher levels of DHEAS at age ~6.8 years were associated with a higher BMI in both boys and girls, but only in girls with an earlier appearance of secondary sex characteristics. This finding remained true even after adjusting for IGF-1 and other relevant metabolic hormones, suggesting that these adrenal hormones directly promote earlier thelarche. In addition to increased DHEAS, both girls and boys had increased IGF-I, and in boys a taller stature and a greater concentration of leptin was observed.

This sex dimorphism in the association between HD and earlier appearance of sexual characteristics points toward the peripheral conversion of adrenal hormones through aromatase conversion to estrogen precursors, which play a role in increasing gonadotropin-releasing hormone secretion, follicle-stimulating hormone, and luteinizing hormone, finally activating the ovary [21]. The timing of puberty is precisely controlled by a plethora of endogenous signals and environmental factors that impinge at different levels of the hypothalamic-pituitary-gonadal (HPG or reproductive) axis. Accordingly, puberty has been regarded not only as a fundamental developmental event but also as a sensor of the proper interplay between genes and the environment along development. Additionally, pubertal timing is highly polygenic and a number of identified loci underlie both pubertal timing and related traits such as height and BMI [22]. Studies undertaken in many species indicate that kisspeptin-Gpr54 signaling is essential for the activation of gonadotropin-releasing hormone neurons to initiate puberty. This has recently been shown to be dependent upon circulating estradiol concentrations [21], supporting our hypothesis that increasing peripheral estrogens could sensitize the HPG axis.

Conversely, the earlier appearance of secondary sex characteristics might be dissociated from HPG maturation and may only reflect the peripheral action of DHEAS and estrogen metabolites. Recent data in a longitudinal study of 252 peripubertal girls in the United States show that DHEAS concentrations increased 24 months before breast development and estrone between 12 to 18 months before breast development; whereas estradiol and testosterone increased (mostly from ovarian source) more closely to the time of breast development [23]. Another interesting fact is that children with HD have a higher percent of obesity and overweight, and in this aforementioned study, girls with a greater BMI had a lower estradiol concentrations at onset of breast development, as well as 6 months after thelarche, pointing toward a peripheral source of sex steroids more than to a centrally active HPG as a cause of thelarche [23]. In a recent paper on the effect of obesity in puberty onset [12], there was a lack of association between adiposity and ovarian volume in girls, suggesting that the advanced breast development observed among early pubertal obese girls may be primarily due to peripheral conversion of relatively inactive androgens to more bioactive estrogens by aromatase in adipose tissues. We recently published that higher levels of estradiol equivalents in prepubertal girls were associated with an earlier thelarche, which remained significant after adjusting for BMI or other adiposity markers, suggesting that these associations were not mediated through adiposity [24]. These observations are also supported by evidence suggesting that age at the gonadotropin and sex steroid surges have not changed in population studies where age at thelarche has declined [25]. It may also be possible that some of our PT girls have HPG activation, whereas, in others, thelarche is mainly peripheral. In our sample of girls with HD, 3.1% presented with PT and PP together, suggesting HPG activation, at least in this population. Data on menarche will help to elucidate whether the earlier appearance of sex characteristics is a reflection of HPG activation. Nonetheless, evidence from the Avon Longitudinal Study of Parents and Children study of 329 girls showed that higher DHEAS at age 8 predicted earlier menarche [26], which is more in line with a centrally activated HPG axis.

Leptin is an essential signal in the control of body weight homeostasis, and its levels in circulation are proportional to the size of fat stores. Our results showing that HD is associated with earlier pubertal events in girls were, nonetheless, adjusted for leptin and insulin and are in agreement with pharmacological studies in humans and rodents with leptin deficiency, documenting that, although leptin is indispensable for puberty to proceed, leptin alone cannot trigger early puberty [27].

Another interesting finding of the current study is the increase in the percent of earlier thelarche (21.3%). In a 2004 cross-sectional analysis performed in our country with 758 girls aged 5.8 to 16.1 years, 16% of girls had thelarche [28] at ages below 8 years old. Our data are concordant with recent Danish longitudinal data describing a recent decline in age at thelarche in a decade [25].

Less data on pubertal maturation are available in boys, in part because pubertal development in boys is more difficult to ascertain on a large scale. In the current study, an increased BMI, mainly overweight, but not HD, was the only factor associated with earlier gonadal maturation. Interestingly, at high concentrations, as expected in obesity, leptin can directly inhibit gonadal function [29, 30]. Although available, data are mixed, a U-shaped relationship could help understand the controversy on the effect of BMI in male puberty [12, 31–33]. Although this was initially considered counterintuitive, the inhibitory actions of leptin may explain part of the hypogonadal state frequently observed in morbidly obese patients [12].

Pubic hair development is mediated by androgen exposure modulated by androgen receptor response. Among girls, HD more than doubled the risk of PP. Overall, in our sample the frequency of PP was 4.1% in girls and 8.5% in boys, which is higher than in previous cross-sectional Chilean studies in girls [28] and boys [34]. Pubic hair is highly dependent on ethnicity [35], and these differences appear to be at least in part mediated by the androgen receptor gene methylation pattern, as well as the presence of shorter CAG trinucleotide repeat [36]. Interestingly, in boys HD did not increase the risk of PP. DHEAS is a weak androgen and together with a lower sensitivity to androgens may not be enough to promote sexual hair growth in boys. Indeed, in a clinical evaluation of body hair in our country, the authors found that women in our country were more likely to be hairless than European or North American women and proposed a lower Ferriman and Gallwey score cutoff for hirsutism diagnosis [37].On the other hand, we cannot exclude that girls with PP alone were starting HPG axis maturation through this pathway and not through thelarche.

Our study is not exempt from limitations: (1) lack of assessment of luteinizing hormone and follicle-stimulating hormone at 2009, at age when DHEAS were obtained; (2) we could have misclassified age at sexual appearance because we use the midpoint of two consecutive visits; to overcome this problem, we carried out sensitivity analysis modifying the cutoff point in the interval of the two visits; and (3) in 36 girls, DHEAS determination occurred at the moment of thelarche, raising the possibility of reverse causality. Nevertheless, the association remained significant after excluding these girls. The study had the following strengths as well: (1) longitudinal follow-up since age 3 to 4 years; (2) visits every 6 months to assess sexual maturation; and (3) a highly trained evaluator to assess sexual maturation data (κ > 0.8).

4. Conclusions

In Chilean adolescents, precocious events of pubertal development were in line with worldwide secular trend of earlier sexual maturation. HD was only associated with PT and PP in girls. Continuous follow-up of this cohort is a unique opportunity to prospectively address the interrelationships among PA, early growth, adiposity as determinants of gonadarche, pubertal rate, and sequence progression and ovarian function.