This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Ibáñez, L. Articles by Saenger, P. Search for Related Content PubMed PubMed Citation Articles by Ibáñez, L. Articles by Saenger, P.

Premature Adrenarche—Normal Variant or Forerunner of Adult Disease?¹

Endocrinology Unit, Hospital Sant Joan de Deu, University of Barcelona, Barcelona, Spain 08950 (L.I.); Division of Pediatric Endocrinology (J.D.-N., P.S.), Albert Einstein College of Medicine/Montefiore Medical Center, Bronx, New York 10467; and Hormonal Laboratory (N.P.), Hospital Materno-Infantil, Vall d’Hebron, Autonomous University of Barcelona, Barcelona, Spain 08035

	Abstract

Top Abstract I. Adrenarche II. Premature Adrenarche III. Conclusions References

 
Adrenarche is the puberty of the adrenal gland. The descriptive term pubarche indicates the appearance of pubic hair, which may be accompanied by axillary hair. This process is considered premature if it occurs before age 8 yr in girls and 9 yr in boys.

The chief hormonal product of adrenarche is dehydroepiandrosterone (DHEA) and its sulfated product DHEA-S. The well documented evolution of adrenarche in primates and man is incompatible with either a neutral or harmful role for DHEA and implies most likely a positive role for some aspect of young adult pubertal maturation and developmental maturation. Premature adrenarche has no adverse effects on the onset and progression of gonadarche in final height.

Both extra- and intraadrenal factors regulate adrenal androgen secretion. Recent studies have shown that premature adrenarche in childhood may have consequences such as functional ovarian hyperandrogenism, polycystic ovarian syndrome, and insulin resistance in later life, sometimes already recognizable in childhood or adolescence. Premature adrenarche may thus be a forerunner of syndrome X in some children. The association of these endocrine-metabolic abnormalities with reduced fetal growth and their genetic basis remain to be elucidated.

I. Adrenarche

A. Introduction

B. Definition

C. Hormonal basis of adrenarche and reference data for steroid hormone levels in adrenarche

D. Biological role of adrenarche

E. Control of adrenarche

F. Adrenarche and gonadarche

II. Premature adrenarche

A. Definition

B. Pathophysiological basis

C. Clinical features

D. Adrenal androgens in premature adrenarche

E. Differential diagnosis

F. Timing of puberty and final height

G. Postpubertal outcome

H. Patterns of insulin secretion

I. Lipid levels in premature adrenarche

J. Acanthosis nigricans

K. Future avenues of investigation

L. Premature adrenarche, hyperinsulinism, and ovarian dysfunction: possible relation to reduced fetal growth

III. Conclusions

	I. Adrenarche

Top Abstract I. Adrenarche II. Premature Adrenarche III. Conclusions References

 
A. Introduction
Adrenarche is the "puberty" of the adrenal gland. It is characterized by the activation of adrenal androgen production and by impressive increases in dehydroepiandrosterone (DHEA) and dehydroepiandrosterone sulfate (DHEAS), both products of the zona reticularis of the adrenal gland (see Fig. 1).

View larger version (92K): [in this window] [in a new window]   Figure 1. Sections through the adrenal glands of a 6-month old infant (left) and of an adult man (right). While the infant has no reticularis zone, the reticularis zone is present in the adult adrenal (see also Fig. 3). [Reprinted with permission from W. Bloom, D.W. Fawcett: Textbook of Histology, ed. 9. W.B. Saunders Co., Philadelphia, 1986, p 461 (323 )].

The descriptive clinical term pubarche indicates the appearance of pubic hair, which may be accompanied by axillary hair. This process is considered premature if it occurs before age 8 in girls and before 9 in boys (4, 5, 6).

C. Hormonal basis of adrenarche and reference data for steroid hormone levels in adrenarche
Studies by several investigators (1, 7, 8, 9, 10, 11, 12) showed that adrenarche is characterized by dramatic increases in urinary 17-ketosteroids and serum levels of dihydrotestosterone, DHEA, and DHEAS (13). Androstenedione, a zona fasciculata product, and 11-hydroxyandrostenedione (14, 15), a zona reticularis product, do not rise during adrenarche. These increases take place in girls and boys between 6 and 8 yr of age, approximately 2 yr before the onset of gonadal maturation and puberty (gonadarche) (10, 11, 12, 16, 17, 18). Cortisol concentration, production, and excretion remain constant (19).

Absence of normative data for adrenal steroidogenesis in children hampered characterization of the endocrine effects of adrenal androgens in the past. Over the past 10 yr, reference data for steroid hormones at baseline and after standard ACTH stimulation have been published and demonstrate a substantial breadth of normalcy as well as gender differences (20, 21). The most widely used protocol for the ACTH stimulation test (250 µg Cortrosyn iv or im) eschews dexamethasone suppression before ACTH testing. Responses to ACTH stimulation change throughout childhood, with definite age-, sex-, and pubertal stage-dependent differences in resulting steroid levels. Ethnic origin may also influence ACTH response pattern (22, 23). An enhanced adrenal sensitivity to ACTH and additional alterations in the metabolic clearance rates of 17- hydroxyprogesterone (17-OHP) or progesterone are characteristic of obese adults but have not been conclusively observed in children (24). In careful longitudinal studies the progressive increase in serum concentrations of DHEA and DHEAS in healthy boys and girls that begins at the age of 6 to 8 yr roughly parallels an increase in skeletal age (25, 26, 27, 28, 29, 30, 31, 32) (Fig. 2). Adrenal androgen levels rise steadily up to age 18–20 yr. During this period a 20-fold increase in DHEAS concentration is accompanied by an increase in the secretion of 17-ketosteroids, especially deoxy C19 steroids. Because androstenedione can be formed peripherally from DHEAS as well as directly by the gonads, circulating levels do not necessarily reflect adrenal production rates. A surrogate marker for adrenal production of androstenedione is 11-hydroxyandrostenedione (15). The enzyme necessary for its formation is expressed only in the adrenal gland; this steroid is therefore specific for the adrenal cortex.

View larger version (31K): [in this window] [in a new window]   Figure 2. Serum DHEAS concentrations in normal boys and girls. Each entry represents a simple value: serial values in children are connected by lines. The area between the two vertical lines encompasses the usual age of onset of puberty in girls, while the shaded area encompasses the usual age of onset of puberty in boys. [Reprinted with permission from S. Korth-Schutz et al.: J Clin Endocrinol Metab 42:1005–1013, 1976 (27 ) © The Endocrine Society.]

For clinical purposes, DHEA and particularly DHEAS are useful markers for adrenal androgen secretion. Based on several studies, levels of DHEAS above 40–50 µg/dl are considered to be consistent with the advent of adrenarche (34, 35).

Numerous studies suggest that the adrenal androgens, DHEA and DHEAS, emanate chiefly from the zona reticularis (36, 37, 38). Recent elegant studies by Endoh et al. (39), using dispersed adrenal cells, identify the zona reticularis, the innermost layer of the adrenal cortex, as the site of biosynthesis of DHEA and DHEAS. The zona reticularis is theorized to be the morphological equivalent of the fetal zone of the adrenal cortex. The fetal zone virtually disappears in the first few months after birth, and production of DHEA and DHEAS virtually ceases, only to resume some 6 yr later (40, 41, 42, 43, 44).

D. Biological role of adrenarche
Growth patterns have been investigated in normal children going through adrenarche. A small but significant growth spurt has been found by two independent investigators to occur between 6.5 and 8.5 yr of age—exactly the age when adrenarche occurs (45, 46). Others have been unable to demonstrate a midchildhood growth spurt (47). Treatment of a child with adrenal hyperplasia with oral DHEA at a dose sufficient to raise DHEAS into the normal range increased linear growth and caused growth of pubic hair, although puberty did not occur (48). In children with precocious puberty who are being treated with GnRH agonists for gonadotropin suppression, DHEA concentrations were found to correlate well with the rate of skeletal maturation (35).

Adrenal androgens also lower serum levels of sex hormone binding globulin (SHBG). This may represent an effect of adrenal androgens on the tempo of the pubertal process through the augmentation of biologically available free testosterone (49). Other investigators found, even with short-term elevation of plasma testosterone levels to 130 ng/dl, only a slight depression of SHBG (50). The well known gender-dependent differences in postnatal SHBG may also suggest a role for prenatal induction in SHBG levels (51).

The event of adrenarche occurs only in humans and higher primate species (chimpanzee, gorilla) that have a long childhood preceding the advent of puberty (52, 53, 54). While the Rhesus monkey, Cynomolgus monkey, and the crab-eating macaque have low levels of DHEA that are at their highest in the newborn and decline thereafter with no discernible adrenarcheal process (55, 56), their DHEAS appears to originate from a persistent fetal zone rather than from a zona reticularis arising at adrenarche (57). On the other hand, the increase of adrenal androgen levels with age in the chimpanzee closely resembles adrenarche in man. The rise in DHEA levels in the chimpanzee preceding gonadal maturation is also comparable to that in man: DHEA levels begin to rise by 5 yr of age, exactly 2 yr before testosterone levels begin to increase in that species.

Androgens of adrenal origin have been postulated to initiate activation of the hypothalamic/pituitary/gonadal axis in puberty; witness the fact that children untreated or poorly treated for congenital adrenal hyperplasia, who consequently have markedly increased androgens, enter central puberty at an earlier or even precocious age (38). The persistent high levels of adrenal androgens in the Rhesus monkey after birth may therefore play a contributory role in its early sexual maturation, whereas the low level of adrenal androgens before adrenarche in man and chimpanzee may be one of several factors in the relatively delayed onset of puberty in these species (53). The role of adrenal androgens in sexual maturation doesn’t apply broadly, however. The rat, for example, does not make DHEA or its sulfate and yet has an early puberty.

The well documented evolution of adrenarche in primates and man is incompatible with either a neutral or harmful role for DHEA but most probably implies a positive role for some aspect of young adult health and reproduction (57).

E. Control of adrenarche
1. The role of the zona reticularis in adrenal androgen production. The adrenal gland of the young child between 1 to 6 yr of age makes predominantly cortisol, a C21 steroid, but virtually no androgens (C19 steroids) (58, 59). The zona reticularis, not perceptible in children under 6, later recapitulates the secretory pattern of the fetal zone, forming DHEA and DHEAS (58, 59, 60, 61, 62, 63, 64, 65, 66, 67). Since the zona reticularis is the only adrenal zone with sulfotransferase activity, DHEAS is a good marker for functional activity of the zona reticularis (62, 63, 68). Both DHEA and DHEAS are products of the 5 pathway (see Fig. 3). The development of the zona reticularis correlates closely with the increasing DHEAS production (31), which is due to low expression of 3ß-hydroxysteroid dehydrogenase (3ß-HSD) activity (14). While the major source of sulfation of DHEA is obviously the adrenal gland, other tissues also have limited sulfotransferase activity (69). The production rate for DHEAS is about 31 mg/day in young men and 19 mg/day for young women, making this the most abundant steroid in humans. The half-life of DHEAS is between 9 and 11 h, whereas it is 30–60 min for the unconjugated DHEA (67). The plasma concentration of DHEAS exhibits a high correlation with urinary 17-ketosteroids and can be used to assess adrenal androgen production rates (67). DHEAS, the steroid hormone in the greatest concentration in the human circulation, can also be synthesized from other sulfated precursors, such as cholesterol sulfate and pregnenolone sulfate (69). Plasma DHEAS concentrations show only minor circadian fluctuations, while those of DHEA seem to follow a circadian pattern similar to that of cortisol (70).

View larger version (21K): [in this window] [in a new window]   Figure 3. Scheme of steroidogenic pathways. 17-Hydroxylase and 17,20-lyase are activities of cytochrome P450C17. 3HSD, 3ß-Hydroxysteroid dehydrogenase; DOC, 11-deoxycorticosterone; S, 11-deoxycortisol; C21, P450C21; c11, P450C11. Note that in the adrenal, little androstenedione is formed from 17OH pregnenolone. Most androstenedione is derived from DHEA and through peripheral conversion.

Careful histological studies by Dhom (73) suggest that the appearance of adrenarche is associated with an increase in the thickness of the zona reticularis (Fig. 4). The zona reticularis begins to develop in foci at age 3 to 5 yr, and by age 7 to 8 yr it is usually present as a continuous zone, as the medullary capsule of the adrenal disintegrates at the same time. Growth of the zona reticularis is directly related to rises in DHEAS levels (Fig. 5). After a peak of adrenal androgen production at age 20 to 25, DHEAS, particularly, begins a steep, continuous decline (57), while serum levels of aldosterone and cortisol undergo relatively little change with age (57, 74, 75, 76).

View larger version (24K): [in this window] [in a new window]   Figure 4. Age at which focal islands of reticular tissue or a continuous reticular zone were found in a series of patients with sudden deaths who had not had an antecedent illness. [Reprinted with permission from G. Dhom: Beitr Pathol 150:357–377, 1973 (73 ).]

View larger version (16K): [in this window] [in a new window]   Figure 5. Relation of plasma DHEAS to growth of the zona reticularis and increase in adrenal volume with age. [Reprinted with permission from M. M. Grumbach et al.: In V. H. T. James et al. (eds) The Endocrine Function of the Human Adrenal Cortex. Serono Symposia 18. Academic Press, London, pp 583–612 (16 ).]

2. The Regulation of adrenal androgen secretion.
a. Extraadrenal factors.
Numerous endocrine signals endogenous and exogenous to the adrenal gland (37) have been proposed as stimuli of adrenal androgen secretion. Among those proposed as exogenous to the adrenal gland were PRL (77, 78), estrogen (79, 80, 81, 82, 83), epidermal growth factor (84), prostaglandins (85), angiotensin (86), GH (87), gonadotropins (88, 89), ß-lipotropin, ß-endorphin, and CRF (90, 91, 92). The adrenal cortex has a high level of PRL receptors in several species (93). In women with PRL-secreting tumors there is a correlation between PRL levels and DHEAS (94). GH receptors are present in the adrenal cortex (95, 96). Recently, administration of MK 0677, a nonpeptidyl compound that restores pulsatile GH secretion by the pituitary, was found to increase DHEAS levels in adults (97). To date, none of these factors has been conclusively identified as a regulator of adrenal androgen secretion of biological significance.

Patients with familial glucocorticoid deficiency due to mutations in the coding region of the ACTH receptor show not only low cortisol levels but also low DHEAS and androstenedione levels and a complete lack of adrenarche. Despite adequate glucocorticoid replacement, ACTH levels remain elevated (98).

ACTH and CRH as a dual control mechanism (99, 100) probably have a permissive role in the modulation of adrenal androgen secretion but are thought not to be the sole stimuli for the rise in adrenal androgen secretion (101). Another pituitary factor, "ACTH-like," which might stimulate adrenal androgen secretion was postulated previously by Mills et al. (8) and by Grumbach and co-workers (16). Based on studies in hypophysectomized, ACTH-replaced chimpanzees, Cutler and colleagues (101) developed the hypothesis that ⁵ androgen secretion is dependent upon a non-ACTH pituitary factor or that different ACTH requirements exist for maintenance of normal cortisol and adrenal androgen secretion. This hypothesis was strengthened by the clinical finding that in pediatric patients with Cushing’s syndrome due to central ACTH overproduction, there is generally no increase in DHEAS and DHEA above normal levels for chronological and bone age, despite the marked increase in cortisol secretion. Hauffa et al. (102) interpreted this observation as lending further support for their theory that there is yet another adrenal androgen-stimulating factor that may indeed be central (18). POMC-related peptides are elevated in pituitary adenomas of patients with Cushing’s disease (103, 104), yet DHEA and DHEAS levels are not elevated. Known POMC-related peptides do not appear to be the adrenal androgen-stimulating factor (104). POMC was believed to be the leading candidate for a glucocorticoid-suppressible adrenarche-stimulating factor. A human pituitary fraction containing a 60,000-dalton glycopeptide that is capable of stimulating the zona reticularis selectively in the dog has been described by Parker et al. (105). This fraction, sharing amino acids POMC_79–96 with human POMC, stimulated DHEA secretion without affecting cortisol secretion in an in vitro dog adrenal bioassay. In subsequent studies using human pituitary fractions and cultured human adrenal cells, it was identified by Parker et al. (106, 107) as central androgen-stimulating hormone (CASH). The synthesized 18-amino acid peptide, CASH-18, stimulated production of DHEA from cultured adult adrenal cells but had no effect on cortisol secretion. When POMC_79–96 was studied by three additional groups (108, 109, 110) with and without ACTH in cultured human fetal and adult adrenal cells, it had no demonstrable effect. No specific binding to human adrenocortical cells could be measured (109). This does suggest strongly that POMC_79–96 is not the elusive central androgen-stimulatory hormone in man. The fact that in an adenoma tissue sample from human Cushing’s disease, increased POMC (111) was not associated with elevated DHEA or DHEAS casts further doubt on the relevance of CASH in the initiation of adrenarche.

A change in nutritional status, measurable in the form of body mass index (BMI) increases, also appears to be an important physiological regulator of adrenarche regardless of individual adrenal androgen secretion, age, and developmental stage (112).

b. Intraadrenal factors.
As adrenarche represents a change in the pattern of adrenal-secretory response to ACTH, another theory for its biochemical foundation is that it is dependent on intraadrenal factors that control growth and differentiation of the zona reticularis, with concomitant changes in the activity of steroidogenic enzymes. Anderson (113) has formulated an attractive hypothesis relating adrenarche to the maturation of the zona reticularis as observed in the histological substrate by Dhom (73). According to this hypothesis the reticularis is exposed to high cortisol concentrations from the adjacent zona fasciculata. Gradually the innermost cells of the fasciculata start to respond to the very high cortisol levels by undergoing morphological and functional changes. Zonal and developmental changes of steroid enzyme activities, as described by Winter and colleagues (66, 114, 115), particularly increased activity of 17,20-lyase, sulfokinase, and sulfatase and reduced activity of 3ß-HSD, especially in the reticularis zone, would lead to production of more DHEA, DHEAS, and androstenedione in response to ACTH.

Adrenarche, thus, is characterized by a profound change in the degree and in the pattern of the adrenal secretory response to ACTH. The levels of 17-OH pregnenolone (17-OH Preg), DHEA, and DHEAS increase strikingly. Maturational increases in 17-hydroxylase and 17,20-lyase are seen together with a low activity of 3ß-HSD (116, 117, 118, 119), particularly in the developing zona reticularis.

Using adult human fasciculata and reticularis cells L’Allemand et al. (120) demonstrated that both insulin-like growth factors I and II (IGF-I and IGF-II) enhance steroidogenic enzyme activity of 17ß-hydroxylase and 3ß-HSD. ACTH receptor mRNA was also slightly increased, while mRNA for cytochrome P450_scc remained unchanged. Thus, IGF-I and -II mimic some of the changes observed in adrenarche; other effects, such as the increase in 3ß-HSD activity, are opposite to those typically observed at the time of adrenarche. The possible role of transforming growth factor-ß1 (TGF-1) in adrenarche is less clear. TGF-1 stimulates 3ß-HSD activity in adult human adrenal cells. A local diminution of TGF-1 production might be involved in the steroid hormone changes observed at adrenarche. The factor responsible for this reduction in TGF-1 expression remains to be elucidated (121). T-cells within the adrenal gland have direct cell-to-cell contact with epithelial cells of the adrenal zona reticularis; this provides a mechanism for immune system-mediated stimulation of androgen secretion in vitro. This establishes evidence for a non-ACTH-mediated mechanism of adrenocortical androgen regulation (122).

Thus, adrenal mass, pattern of intraadrenal blood flow, intraadrenal steroid concentrations, and immune system-mediated stimulation, together with enzymatic changes and changes in ACTH response, affect adrenal androgen production as adrenarche begins.

c. P450C17 and adrenarche.
An especially intriguing new, molecular genetic approach was suggested by the observation that increasing the molar ratio of isolated, purified electron donors, such as P450 oxidoreductase (OR) or cytochrome b5, to porcine P450C17 would increase the ratio of 17,20-lyase-to-hydroxylase activity (123, 124). Recent experiments by Miller and co-workers (125) with transfected cos-1 cells confirm that the expression of vectors encoding human OR and human P450C17 results indeed in a substantial increase in 17,20-lyase activity. However, it seems unlikely that adrenarche could result from a large increase in the expression of an electron donor, as the activity of adrenal cytochrome P450C21 (steroid 21-hydroxylase), which uses the very same electron donors, is unchanged during adrenarche (58).

Human cytochrome b5 acts principally as an allosteric effector that interacts primarily with the P450C17 OR complex to further stimulate 17,20-lyase activity. Complete absence of cytochrome b5, as described in a splicing mutation, may lead to low levels of androgen synthesis and even male pseudohermaphroditism (126).

Since the regulation of 17-hydroxylase and 17,20-lyase determines the degree or amount of precursor steroids that are converted to sex steroids, regulation of these two enzymatic steps coded by a single human gene for P450C17 is extremely important. The 17-hydroxylation of pregnenolone and progesterone and the subsequent cleavage (17, 20-lyase activity) of 17-OH Preg and 17-OHP are catalyzed by a single enzyme, cytochrome P450C17. In the testes, all precursor steroids are converted to sex steroids; the ratio of lyase to hydroxylase activity is therefore 1. In the human adrenal cortex, however, activity of these enzymes, as well as other enzymes, is under closely regulated control during development, which may determine timing as well as tempo of adrenarche. Previous studies have shown that a specific amino acid sequence is required for maintenance of 17,20-lyase activity (126, 127). Since the amino acid sequence of P450C17 cannot change with adrenarche, Zhang and co-workers postulated that a posttranslational modification of P450C17 could alter the ratio of hydroxylase to lyase activity (64).

Consistent with their hypothesis, these authors have found in an in vitro system using African green monkey kidney cells that the serine phosphorylation of cytochrome P450C17 by a cAMP-dependent kinase accounts for a large increase in 17,20-lyase activity (64). This process differs from the regulation of 17ß-hydroxylase activity, which is needed to produce cortisol throughout life. The 17,20-lyase enzyme is controlled independently in an age-dependent pattern. Early activation of this process increases 17,20-lyase activity. P450C17 is phosphorylated on serine and threonine residues by a cAMP-dependent protein kinase; phosphorylation of P450C17 increases lyase activity, while dephosphorylation virtually eliminates this activity. Hormonally regulated serine phosphorylation of human P450C17 suggests a possible mechanism for human adrenarche that would unify all clinical findings.

These studies do require independent confirmation, and altered phosphorylation of P450C17 has yet to be demonstrated in children at adrenarche or, for that matter, in children with premature adrenarche (119). Thus, the role of serine phosphorylation remains only a hypothesis until such a time as the kinase is cloned and activating mutations are found in families with polycystic ovary syndrome (PCOS) or premature adrenarche.

Hornsby cautions that an absence in the change of adrenal production of androstenedione makes it unlikely that adrenarche involves changes in 17,20-lyase activity of CYP17, the gene encoding for P450C17 (57), although it should be pointed out that physiological responses of ⁴ to ACTH are modest (20, 21). A quite simple explanation for the absence of a rise in androstenedione may be that the preferentially used ⁵ pathway in the adrenal, of course, bypasses androstenedione altogether (57).

While the physiological trigger for adrenarche and/or altered P450C17 hydroxylase and lyase activity is currently unknown, Zhang et al. (64) speculate that IGF-I and possibly also insulin are good candidates. Both insulin and IGF-I transmit their signals by initiating tyrosine autophosphorylation of the insulin/IGF-I receptors, while the phosphorylation of serine and threonine residues markedly diminishes signal transduction (128, 129, 130).

F. Adrenarche and gonadarche
The increase in adrenal androgens is not associated with an increase in sensitivity of gonadotropins to GnRH or with sleep-associated LH secretion characteristic of puberty; rather it occurs at an age when the hypothalamic/pituitary/gonadal axis is functioning at a lower level of activity and gonadarche has not yet occurred. Adrenarche and gonadarche are thus two separate maturational events (131). Timing of adrenarche in girls with Turner syndrome who do not undergo gonadarche is perfectly normal (132). Similarly, children with isolated gonadotropin deficiency will undergo normal adrenarche while children with adrenal insufficiency will not. In true isosexual central precocious puberty occurring before the age of 6 yr, there is generally no adrenarche, whereas in precocious puberty occurring after age 6 yr, adrenarche may be present (133). Boys treated for primary adrenal insufficiency have been noted to enter puberty at a normal age (134). Thus, adrenarche and gonadarche are independent events controlled by separate mechanisms (131) (see Table 1).

	II. Premature Adrenarche

Top Abstract I. Adrenarche II. Premature Adrenarche III. Conclusions References

Premature development of pubic hair with and without axillary hair and without other signs of virilization or puberty was first described by Wilkins (135) and was descriptively named premature pubarche. Some years later, other investigators suggested that the adrenal glands could be involved in the development of this condition and they named it, therefore, "precocious adrenarche" (136).

An increased frequency of premature adrenarche has been reported in children with cerebral dysfunction with a sex ratio close to 1 (138, 139), although children with premature adrenarche do not have more developmental or behavioral problems (140). Weight gain may be a trigger for adrenarche (112), and obesity has also been associated with a higher incidence of premature adrenarche (141, 142, 143).

Recent data suggest that girls seen by primary care practitioners in the United States show pubic hair and/or breast development at younger ages than stated above (144). In a cross-sectional study involving 17,077 girls, striking differences were detected in pubic hair development between black and white girls. At 6 yr of age, 9.5% and, at 8 yr of age, 34.3% of black girls had at least Tanner stage 2 pubic hair, whereas 1.4% and 7.7% of white girls, at these ages, had pubic hair. Although these observations might suggest a revision of the current criteria for referral of premature adrenarche patients, the data should be interpreted with caution, as there may have been a bias in self-referral of these patients to the pediatric practices. Furthermore, as no endocrine evaluations were carried out in the study, it is not known whether some of the girls included had pathological conditions accounting for the early appearance of pubertal milestones (144).

Pubarche after age 7 yr is often slowly progressive. However, that does not mean that it is normal. Evidence is emerging that premature pubarche may on occasion be a risk factor for subsequent reproductive endocrine system dysfunction (144A ).

B. Pathophysiological basis
Generally, premature adrenarche is secondary to an early isolated maturation of the adrenal gland (26, 30, 145, 146, 147). Adrenal androgens, particularly DHEA, DHEAS, androstenedione, and testosterone, are in most cases moderately increased for chronological age but fall within the expected range according to the pubertal stage of pubic hair (59, 145, 147, 148). In some patients, the early development of pubic hair is associated with normal androgen levels for chronological age, suggesting increased peripheral sensitivity (59, 145, 149). Lee et al. (150) described a family in whom adrenal androgen hypersecretion was transmitted as a dominant non-HLA-linked trait (150).

The cause of the adrenal oversecretion in premature adrenarche is currently unclear (151). Gonadotropins do not play a role in the development of premature adrenarche (152, 153) just as in normal adrenarche.

C. Clinical features
In typical or isolated premature adrenarche the appearance of pubic hair, which is usually dark, straight or curly, and coarse, is mostly limited to the labia majora in girls and thus may elude detection on casual examination in an obese girl. The development of pubic hair is non- to slowly progressive and may spread throughout the pubic area (154). Axillary hair growth may also be noted (6, 141). A mild hypertrichosis with fine hair over the extremities and back is much less frequently observed (141). Increased body odor, oily skin, and acne, usually in the form of a few microcomedones, may be present. Clitoral or penile enlargement are usually absent, and testicular and breast size remain at the prepubertal stage (155). Growth velocity may be increased, and moderately advanced bone maturation (<±2 SD is often present, but is generally correlated with the height age) (4, 141, 155, 156, 157).

D. Adrenal androgens in premature adrenarche
Although DHEA and DHEAS are relatively weak androgens, they serve as a substrate for the synthesis of more potent androgens, such as androstenedione and testosterone (158, 159). In premature adrenarche, baseline serum levels of DHEA, and to a lesser degree, those of androstenedione and testosterone as well as their urinary metabolites, the 17-ketosteroids, are in the range of those found in early puberty (6, 26, 33, 59, 145, 147, 148, 157, 160, 161, 162, 163) (Fig. 6). However, DHEAS levels may exceed those of pubertal controls (148, 162). Serum DHEAS concentrations can be suppressed to a greater extent after dexamethasone treatment, although the degree of suppression is highly variable and seems to be related to bone age (23).

View larger version (31K): [in this window] [in a new window]   Figure 6. Baseline plasma androgen levels in girls with premature appearance of pubic hair compared with age-matched controls. The enclosed areas represent the mean ± 2 SD levels in controls. 17-OHP, 17-Hydroxyprogesterone; 5 , androstenedione. [Modified with permission from R. Virdis et al.: Riv Ital Pediatr (IJP) 19:569–579, 1993 (160 ).]

The responses of the steroid precursors 17-OHP, 17-OH Preg, and 11-deoxycortisol to ACTH are between the prepubertal and adult range in more than 90% of children with premature adrenarche (145).

E. Differential diagnosis
Premature adrenarche is a diagnosis of exclusion. In those patients in whom pubic hair is accompanied by testicular, breast, or clitoral enlargement (atypical premature adrenarche), the strong possibility of precocious puberty or a virilizing adrenal or gonadal tumor must be ruled out (4, 141, 155). The possibility of iatrogenic androgen administration must also be kept in mind. A careful history and physical examination can potentially rule out these entities. In some cases, measurement of gonadotropins and gonadal steroids may be necessary (Table 2).

The incidence of mild defects of steroidogenesis among premature adrenarche patients is not well defined and ranges from about 0 to 40% of cases (144, 148, 167, 168, 169, 170, 171, 172, 173, 174). This discrepancy may be due to several factors. The ethnic origin of the patients is important and can partially explain the high variability of incidences reported. For example, in Ashkenazi Jews, the prevalence of late-onset congenital adrenal hyperplasia due to P450C21 deficiency has been calculated to be as high as 3.7% with a disease frequency of 1 in 27 (175). Conversely, the incidence of this enzymatic defect among Spanish children presenting with premature adrenarche is 7% (148). The number of subjects studied and the diversity of criteria adopted for diagnosis may also account for the differences.

Most patients with premature adrenarche due to late-onset congenital adrenal hyperplasia have clinical features characteristic of atypical premature adrenarche and present with elevated baseline hormone (17-OH Preg, 17-OHP, androstenedione, testosterone) levels. However, there is still controversy as to whether all children with premature adrenarche should undergo an ACTH stimulation test (250 µg Cortrosyn iv or im), based on the fact that, in some cases, baseline hormonal levels may be normal in mild errors of steroidogenesis (148, 168). We recommend ACTH testing in those children with ratios of bone age to statural age greater than 1, and/or elevated basal androgen levels, and/or signs of atypical premature pubarche (4, 155) (Fig. 7). Atypical premature adrenarche is characterized by bone age advancement (155), cystic acne, and signs of systemic virilization.

View larger version (26K): [in this window] [in a new window]   Figure 7. Differential diagnosis of elevated androgens in premature adrenarche. Algorithm for the hormonal diagnosis of premature adrenarche. Baseline androgen levels are assessed. Markedly elevated plasma androgen levels point toward tumor or Cushing’s syndrome (if cortisol levels are concomitantly elevated). Moderately elevated plasma androgen levels other than DHEAS indicate the need for an ACTH test to rule out congenital adrenal hyperplasia. If only plasma DHEAS is moderately elevated, the diagnosis of premature adrenarche is made.

The hormonal criteria for mild 3ß-HSD deficiency are still very controversial. Some 1–13% of children presenting with premature adrenarche and 3–50% of older female patients with hirsutism and menstrual disorders have been reported to have this enzymatic defect based on published hormonal criteria (179, 180). The molecular analysis of the type I and type II 3ß-HSD genes in children with premature adrenarche and hyperandrogenic women has failed to demonstrate mutations in those patients with post-ACTH ⁵-steroid precursor levels between 5 and 10 SD above the normal mean levels (181). Therefore, the hormonal levels for ACTH-stimulated 5-steroids in patients with a mild variant of 3ß-HSD deficiency are predicted to be higher than 10 SD above the normal mean value (169).

2. Idiopathic functional adrenal hyperandrogenism. Exaggerated androgenic precursor responses to ACTH testing were first reported in adult hyperandrogenic women (182) and subsequently found to be a common finding in hyperandrogenic adolescents and children with premature adrenarche (183, 184). Typically, these patients show prompt and prominent hyperresponsiveness to ACTH (not only more than 2 SD above the mean for normal age- and sex-matched controls but also above Tanner stage-matched controls) of the ⁵-steroids DHEA and 17-OH Preg, with 50% of them also showing an excessive response of androstenedione (185) and a concomitant hyperresponse of 17-OHP (182, 186). In this group of patients, the post-ACTH ratios of plasma 17-ketosteroids to cortisol have been found compatible with increased 17,20-lyase activity (186). This pattern of adrenal secretion resembles an exaggeration of adrenarche and has conservatively been considered "idiopathic," as it cannot be assigned to any well established pathophysiological entity, such as late-onset congenital adrenal hyperplasia. Thus, the entity has been described as functional adrenal hyperandrogenism (186).

Recently Banerjee et al. (187) reported that many prepubertal African-American and Caribbean Hispanic girls with premature adrenarche can have an androgen response to standard ACTH testing that is different from that which has been reported for the early pubertal stages. Approximately one-third of the 72 patients tested were found to have ACTH-stimulated levels of 17-OH Preg that were more than 2 SD above the mean for normal early pubertal children. In contrast, ACTH-stimulated levels of DHEA, androstenedione, and 17-OHP all remained in the early pubertal range.

Although the cause of idiopathic functional adrenal hyperandrogenism is unknown, it does not imply an enzymatic abnormality. It may simply represent hyperplasia of the zona reticularis (59). Another possibility is that it may be due to abnormal regulation or dysregulation of androgen formation by 17-hydroxylase and 17,20-lyase involving adrenal P450C17 activity, most prominently expressed in the ⁵-pathway (186). Furthermore, dysregulation of ovarian cytochrome P450C17, although most prominently involving the 4 pathway, often seems to coexist in a significant proportion of these patients, as discussed below (59, 186, 188).

F. Timing of puberty and final height
Previous analysis of small groups of patients has suggested that isolated premature adrenarche is not usually associated with a marked alteration in the timing of the child’s subsequent pubertal development (141, 147). Recent follow-up studies from two larger European population of girls with premature adrenarche and similar ethnic characteristics (70 Northern Italian and 57 Northern Spanish) have shown that advanced bone age and tall stature are frequently seen during the first years of follow-up and subsequently wane (157). Times of initiation of gonadarche (Tanner breast stage 2) at age 9.7 + 0.9 yr and menarche at 12.0 + 1.0 yr were comparable to maternal and population data (157) (Table 3). Furthermore, adult heights correlated well with height prognosis at time of diagnosis and at onset of puberty (Fig. 8). Final adult heights were generally above midparental heights, following the secular trend still present in both populations (157). Thus, premature adrenarche appears to cause a transient acceleration in growth and bone maturation with negligible effects on the onset and progression of puberty and final height (157, 189).

View larger version (35K): [in this window] [in a new window]   Figure 8. Relationship between height prognosis at diagnosis of premature adrenarche at onset of puberty and final height in 38 postpubertal girls, expressed in percentiles. Each three-dotted line represents an individual subject. Reprinted with permission from L. Ibáñez et al.: J Clin Endocrinol Metab 74:254–257, 1992 (157 ). © The Endocrine Society.]

In a preliminary study we performed in a group of postpubertal girls with a history of premature adrenarche (193), 9 of 27 showed an increased score of hirsutism, using the Ferriman and Gallway scale (194) and elevated baseline androgen levels. Three of the girls also had oligomenorrhea and polycystic ovaries on ultrasound.

a. PCOS.
PCOS is the most common cause of hyperandrogenism in young females, with an incidence of approximately 3% in both adolescents and adults (59). Stein and Leventhal (195) were the first to define the association of polycystic ovaries and amenorrhea, and to recognize the high incidence of hirsutism and obesity in these patients (195). In its fully developed form, PCOS is characterized by menstrual abnormalities with anovulation, obesity, hyperandrogenemia, elevated plasma LH concentrations, and ultrasonographic evidence of polycystic ovaries (196, 197, 198). However, the subject remains controversial, in part due to the paucity of knowledge of its pathogenesis, and partly because endocrinological criteria for diagnosis are not well defined. Indeed, half of the women with the clinical syndrome lack the classic sonographic features of PCOS (199). Consequently, PCOS has come to be empirically defined on clinical grounds as chronic hyperandrogenic anovulation that is not secondary to underlying disease of the pituitary, ovaries, or adrenal glands (198, 200, 201).

GnRH agonists are potent and specific stimulators of the pituitary-gonadal axis. A single dose maximally stimulates gonadotropin secretion both in children and adults within 3–4 h and gonadal secretion within 18–24 h (202, 203, 204, 205, 206). It has proved to be a more effective stimulator of pituitary-gonadal function than a standard GnRH test (206). In adult women with well defined PCOS, the administration of the GnRH agonist nafarelin elicits pituitary-gonadal responses that are similar to those found in normal men and differ significantly from those elicited in normal women (204). This "masculinized" response is, according to Barnes et al. (204), characterized by an early increase in plasma LH levels 30–60 min after challenge and by androstenedione and a predominant 17-OHP hyperresponsiveness 16–24 h after GnRH agonist administration (204). This response pattern is unaffected by dexamethasone pretreatment (204). The secretory pattern seems to result from a generalized overactivity of steroidogenesis, which is particularly evident at the level of thecal 17-hydroxylase and 17,20-lyase activities of cytochrome P450C17 (199, 204). The exaggerated ovarian 17-OHP response does not appear to be mediated by increased secretion of LH in response to the agonist, as similar increases in 17-OHP can be elicited by the administration of a single and standardized (5000 IU, im) dose of human CG (hCG), both in normal and hyperandrogenic women (Fig. 9) (207). Therefore, it has been postulated that abnormal regulation of this androgen-forming enzyme within the ovary, rather than a steroidogenic block, underlies most PCOS cases (199).

View larger version (39K): [in this window] [in a new window]   Figure 9. Top, Baseline and peak 17-OHP responses to leuprolide acetate challenge (500 µg sc) in 16 oligomenorrheic premature adrenarche girls, 19 regularly-menstruating premature adrenache girls, and 12 age-matched controls. Values are mean ± SEM; a, P < 0.0001 vs. regularly menstruating patients and controls. [Modified with permission from L. Ibáñez et al.: J Clin Endocrinol Metab 76:1599–1603, 1993 (209 ). © The Endocrine Society.] Bottom, Basal and peak 17-OHP levels in response to GnRH agonist leuprolide acetate (open bars) and hCG (filled bars) administration in PCOS women (right) and controls (left). *, Each 17-OHP value is significantly higher in PCOS subjects than in controls. [Modified with permission from L. Ibáñez et al.: J Clin Endocrinol Metab 81:4103–4107, 1996 (207 ). © The Endocrine Society.]

We assessed the ovarian responses to challenge with the GnRH agonist leuprolide acetate in 35 adolescents with a history of premature adrenarche (age: 15.4 ± 1.5 yr) who were at least 3 yr beyond menarche (209). Sixteen of them showed hirsutism, oligomenorrhea, and elevated baseline testosterone and/or androstenedione levels. The remaining 19 were eumenorrheic, nonhirsute, and showed baseline androgen levels similar to those present in a group of 12 age- and BMI-matched controls. Subcutaneous administration of the agonist (500 µg) produced similar increases in gonadotropin levels in the three groups when tested at 6 h. However, 17-OHP and androstenedione levels 24 h after leuprolide acetate challenge were significantly higher in the oligomenorrheic girls than in the other two groups (Fig. 9). Specifically, only oligomenorrheic girls showed stimulated 17-OHP levels exceeding the mean ± 2 SD of the values found in controls (>160 ng/dl). The responses of the remaining androgens to the agonist were very similar among the three groups. In this cohort of 35 postpubertal girls with a history of premature adrenarche, almost half (45%) show a distinct response to GnRH agonist challenge, suggestive of functional ovarian hyperandrogenism, indicating the need for a continued postpubertal follow-up of these patients (209).

This pattern of ovarian steroidogenic response appears to be also particularly frequent in unselected hyperandrogenic women and adolescents (18, 185, 210). In our series, 58% of girls presenting with signs or symptoms of androgen excess showed abnormal 17-OHP responses to leuprolide acetate testing (183, 210). The sensitivity and specificity of the 17-OHP response to leuprolide acetate challenge in the diagnosis of functional ovarian hyperandrogenism compared with those after the dexamethasone suppression test, performed in the same patients, were 72.8% and 94.7%, respectively (183). Similar results have been reported in adult hyperandrogenic women after nafarelin testing, suggesting that the response of 17-OHP after GnRH agonist challenge can be used as a marker for the diagnosis of this type of ovarian dysfunction (183, 199).

c. Premature adrenarche and subsequent development of functional ovarian hyperandrogenism.
To identify possible biochemical markers for predicting the development of ovarian hyperandrogenism in girls with premature adrenarche, the relationship between adrenal androgen levels at premature adrenarche diagnosis and androgen responses to GnRH agonist challenge were examined. Baseline DHEAS and androstenedione levels at diagnosis of premature adrenarche correlated positively with 17-OHP values after leuprolide acetate challenge (Fig. 10), suggesting that functional ovarian hyperandrogenism is more frequent in those girls with pronounced premature adrenarche (209). Cytochrome P450C17 is encoded by the same gene in the adrenal and in the gonads, resulting in androgen synthesis in both glands (211). Thus, increased cytochrome P450C17 activity in both the adrenals and ovaries might first begin in the adrenal during childhood, causing premature adrenarche, and subsequently occur in the ovary, leading to signs and symptoms of functional ovarian hyperandrogenism.

View larger version (19K): [in this window] [in a new window]   Figure 10. Correlation between basal DHEAS and androstenedione (A) levels at diagnosis of premature adrenarche and 17-OHP values after GnRH agonist (leuprolide acetate) challenge. [Modified with permission from L. Ibáñez et al.: J Clin Endocrinol Metab 76:1599–1603, 1993 (209 ). © The Endocrine Society.]

View larger version (35K): [in this window] [in a new window]   Figure 11. Incremental rises of 17-hydroxypregnenolone (17-Preg), DHEA, 17-OHP, and androstenedione (A) after GnRH agonist (leuprolide acetate) challenge in premature adrenarche girls and Tanner stage- and bone age-matched controls in early puberty (B2), midpuberty (B3), late puberty (B4), and postmenarche (B5). Values are the mean ± SEM. Note: postmenarcheal girls (i.e., B5) included in this study were specifically selected from among those who did not develop signs of hirsutism (31 B2 patients, age 9.8 ± 0.1 yr; 15 B3, age 11.0 ± 0.2 yr; 12 B4, age 12.0 ± 0.4 yr; 18 B5, age 15.4 ± 0.6 yr. Controls: 11 B2, age 11.2 ± 0.3 yr; 10 B3, age 13.1 ± 0.4 yr; 9 B5, age 15.7 ± 0.3 yr). Patients were matched for bone age. For details, see Ref. 213. [Reprinted with permission from L. Ibáñez et al.: Fertil Steril 67:849–855, 1997 (213 ).]

2. Adrenal function. Reevaluation of the adrenal function in postpubertal girls with a history of premature adrenarche has shown an increased incidence of idiopathic functional adrenal hyperandrogenism, i.e., DHEA, 17-OH Preg, and androstenedione hypersecretion in response to ACTH stimulation (182, 183). Fifty-five percent of the adolescent girls tested showed post-ACTH DHEA or 17-OH Preg and androstenedione levels greater than 2 SD beyond the mean for controls (215). All patients with idiopathic functional adrenal hyperandrogenism were hyperinsulinemic and in 70% of them, exaggerated 17-OHP responses to GnRH agonist stimulation suggestive of functional ovarian hyperandrogenism were also found (215). Overactivity of both adrenal 17,20-lyase and ovarian 17-hydroxylase/17,20-lyase coexist in the majority of cases of functional ovarian hyperandrogenism (186).

H. Patterns of insulin secretion
1. Insulin resistance at puberty. Puberty has been associated with increased fasting and glucose-stimulated insulin concentrations and a decrease in insulin sensitivity (216, 217, 218, 219). The insulin resistance during puberty is restricted to peripheral glucose metabolism and is associated with concomitant increases in GH, IGF-I, and IGF binding protein-3 (IGFBP-3) levels and a decrease in insulin-like growth factor binding protein-1 (IGFBP-1) and SHBG concentrations (216, 220, 221, 222, 223).

Differences in insulin sensitivity throughout puberty appear to be sex dependent (224) and also to show racial differences (225, 226). For example, African-American adolescents have lower insulin sensitivity and higher insulin levels during a hyperglycemic clamp than Tanner stage- and weight-matched Caucasian adolescents (225).

2. Hyperinsulinemia and premature adrenarche.
a. Prepubertal period.
The hyperinsulinemia and increased IGF-I activity during puberty have been proposed as inducing factors in the development of PCOS (212). Both insulin and IGF-I are capable of stimulating androgen production by ovarian thecal-interstitial cells and to augment the steroidogenesis and ACTH responsiveness of human adrenocortical cells in culture (121, 227, 228). However, whether hyperinsulinemia and insulin resistance may be primary in the development of ovarian hyperandrogenism is still unclear (229).

Oppenheimer et al. (230) were the first to relate ACTH-stimulated steroid hormone data to insulin sensitivity obtained from the frequent sampling intravenous glucose tolerance test (FSIVGTT) (231, 232) in 21 prepubertal African-American and Hispanic girls with premature adrenarche. Eleven girls had normal insulin sensitivity and 10 girls had an insulin sensitivity more than 2 SD below the mean of normal prepubertal girls. Insulin sensitivity correlated inversely with the ACTH-stimulated levels of 17-OH Preg and the ratio of 17-OH Preg/17-OHP. IGFBP-3 levels were normal and IGFBP-1 levels were low normal.

Just as in many women with PCOS, the hyperandrogenism of prepubertal African-American and Caribbean Hispanic girls with premature adrenarche can be associated with hyperinsulinism. The increased insulin levels may result in decreased levels of IGFBP-1, which in turn, can increase the availability of IGF-I (233). IGF-I, together with insulin, may directly stimulate ovarian steroidogenesis (227, 233). In carefully conducted in vitro studies IGF-I and insulin synergize with LH to stimulate androgen production by normal ovarian theca-interstitial cells (228). More recent data from our group suggest a primary role of altered insulin sensitivity and IGFBP-1 activity as hyperandrogenism develops (234, 235).

Hyperinsulinemia after an oral glucose load is a common feature in lean premature adrenarche girls before and also during pubertal development (236) (Fig. 12). The hyperinsulinism is associated with an increased initial insulin response to glucose and a later rise in insulin sensitivity compared with bone age- and Tanner stage-matched girls used as controls. These patients also show increased free androgen indexes and lower serum SHBG and IGFBP-1 levels at most pubertal stages tested (236).

View larger version (30K): [in this window] [in a new window]   Figure 12. Mean serum insulin (MSI) in premature adrenarche girls and in Tanner stage-matched controls. Values are mean ± SEM. *, Significantly different from controls. P < 0.03, P = 0.03, P = 0.03, and P < 0.05 for B1, B2, B3, and B5, respectively. [Reprinted with permission from L. Ibáñez et al.: J Clin Endocrinol Metab 82:2283–2288, 1997 (236 ). © The Endocrine Society.]

More recently we documented the utility of fasting glucose insulin ratios as a simple measure evaluating insulin resistance in girls with premature adrenarche (242A 242B ).

b. Postpubertal period.
Hyperinsulinism and insulin resistance have been consistently reported in obese and lean women with functional ovarian hyperandrogenism (243), PCOS patients (244, 245, 246, 247, 248, 249, 250, 251, 252), and hyperandrogenic adolescents (210), although some reports have failed to find a linear relationship between hyperinsulinemia and hyperandrogenism in hirsute patients (253). PCOS women may already have impaired glucose tolerance (IGT) or frank type 2 diabetes by their third decade (244, 245, 254). Although PCOS and obesity have a synergistic deleterious effect on glucose homeostasis (244), insulin resistance has also been reported in lean PCOS patients (245, 246, 247, 251) and appears to be directly related to the degree of hyperandrogenism (246, 247).

Ethnicity has been proven to be an independent risk factor for insulin resistance development in PCOS women (255, 256, 257). For example, Caribbean-Hispanic PCOS women are significantly more insulin resistant than non-Hispanic women matched for age, weight, and body composition (255).

The mechanisms of insulin resistance in PCOS are still unclear (258). Although it has been hypothesized that androgens directly decrease insulin action (259, 260), studies in which the hormonal environment has been manipulated have yielded conflicting results. Suppression of androgen action with antiandrogenic drugs in hyperandrogenic women has been shown to either have no effect on insulin levels or result in significant improvement in insulin sensitivity (261, 262, 263, 264). On the other hand, it has been proposed that hyperinsulinemia per se causes hyperandrogenism (259, 265). Consistent with this hypothesis, in short-term studies of women with PCOS, insulin infusions have been shown to increase androgen levels, whereas lowering circulating insulin levels with diazoxide, troglitazone, or a somatostatin analog has also decreased androgen levels (266, 267, 268, 269). The report of Nestler and Jakubowicz (270) that hyperinsulinemia stimulates ovarian P450C17 activity in obese PCOS women suggests that, at least in some subsets of hyperandrogenic patients, hyperinsulinemia and dysregulation of ovarian androgen secretion are pathogenetically linked (270, 271). Postpubertal girls with a history of premature adrenarche and functional ovarian hyperandrogenism demonstrate more hyperinsulinism than normal adolescents after an oral glucose load (272). Furthermore, 27% of our cohort of postpubertal premature adrenarche girls without ovarian androgen excess also show mean serum insulin values well above the upper normal limit for controls. The hyperinsulinemia is directly related to the degree of ovarian hyperandrogenism (assessed by an abnormal 17-OHP response to GnRH agonist challenge) in functional ovarian hyperandrogenism patients and to the free androgen index (equivalent to free testosterone) levels in both girls with ovarian hyperandrogenism and subjects without ovarian hyperandrogenism (272) (Fig. 13).

View larger version (16K): [in this window] [in a new window]   Figure 13. Relationship between mean serum insulin (MSI) levels and the free androgen indexes (FAI) in postpubertal premature adrenarche girls with functional ovarian hyperandrogenism (FOH), premature adrenarche girls without functional ovarian hyperandrogenism (non-FOH), and controls. [Modified with permission from L. Ibáñez et al.: J Clin Endocrinol Metab 81:1237–1243, 1996 (272 ). © The Endocrine Society.]

c. Type 2 diabetes mellitus and adrenal/ovarian hyperandrogenism: a familial syndrome?
Type 2 diabetes mellitus and/or hirsutism with or without polycystic ovaries have been described in kindreds of probands with PCOS (273). However, these association studies may be weakened by the heterogeneity of the syndrome, and by the diverse etiologies of these phenotypic characteristics. Consequently, homogenous populations are needed for identifying biochemical intermediary phenotypes that are fixed defects in PCOS and functional ovarian hyperandrogenism and may be present in family members.

Preliminary studies performed by us suggested that the search for biochemical intermediary phenotypes may begin with a homogeneous population in which probands meet four strict criteria: premature adrenarche, clinical evidence of hyperandrogenism, exaggerated 17-OHP responses to GnRH agonist challenge suggestive of functional ovarian hyperandrogenism, and hyperinsulinemia. The study of 60 first-degree relatives belonging to nine families with two affected adolescents each has shown an increased prevalence of both type 2 diabetes mellitus and IGT compared with our normal age- and BMI-matched population (type 2 diabetes mellitus, 22.2% vs. 1.5%; IGT, 27.7% vs. 8.2%) (274).

Even more striking findings with regard to the prevalence of type 2 diabetes mellitus were obtained among first-degree relatives of a group consisting of African-American and Caribbean-Hispanic patients with premature adrenarche (235). In this study 25 of 35 children studied had at least one first- or second-degree relative with type 2 diabetes mellitus. Female first-degree relatives also had lower serum SHBG levels compared with age-matched population controls, possibly secondary to their hyperinsulinemia (274).

I. Lipid levels in premature menarche
Considerable epidemiological controversy exists as to whether hyperinsulinemia, both fasting and postprandial, is an independent risk factor for the development of cardiovascular disease (275, 276, 277). Preliminary data of the lipid patterns in premature adrenarche girls show that hyperinsulinemia is accompanied by increased triglyceride levels compared with a Tanner stage- and age-matched population (278) (Fig. 14) and support the proposal that the genesis of an atherogenic pattern of risk factors may start in childhood (279, 280). Insulin appears to be a major determinant of their lipid status with no additional effects of androgens or estrogens on serum lipid levels.

View larger version (17K): [in this window] [in a new window]   Figure 14. Serum triglyceride levels in premature adrenarche patients and controls throughout all stages of pubertal development. B1 to B5, Tanner breast stages 1 to 5. ap = 0.01; bp = 0.006, and cp = 0.001 vs. controls. [Modified with permission from L. Ibáñez et al.: Diabetologia 41:1057–1063, 1998 (278 ).]

View larger version (27K): [in this