Puberty represents a period of significant growth, hormonal change and the attainment of reproductive capacity. Its onset is marked by a significant increase in the amplitude of pulsatile release of gonadotrophin-releasing hormone (GnRH) by the hypothalamus (Mauras et al 1996). This usually occurs between the ages of 8 and 13.5 years in girls, and stimulates an increase in pituitary release of luteinizing hormone (LH) and follicle-stimulating hormone (FSH), which initially occurs as high-amplitude nocturnal pulses, although eventually both daytime and nocturnal pulsatile release are established (Delemarre et al 1991). FSH and LH, in turn, act upon the ovary to promote follicular development and sex steroid synthesis. As endocrine activity is initially nocturnal, there is no point in measuring gonadotrophin or sex steroid levels during the day.

In the female, this period is characterized clinically by accelerated linear growth, the development of breasts and pubic hair, and the eventual onset of menstruation (menarche) which occurs between the ages of 11 and 16 years in most women in the UK (Stark et al 1989). Menarche is often used as a marker of pubertal development as it is an easily identifiable event which can usually be dated with some accuracy. Adrenarche, the growth of pubic hair, is due to the secretion of adrenal androgens and precedes gonadarche by about 3 years. Thus, prepubertal children often have pubic hair, although, if pronounced, this should be investigated to exclude pathological causes (see below). Whilst androgen secretion is essential, oestrogen secretion facilitates pubic hair growth.

In tandem with the onset of pulsatile GnRH secretion, there is an increase in the amplitude of GH pulses released by the pituitary. Evidence suggests that this amplification of GH secretion may be regulated by the pubertal increase in levels of both androgenic and oestrogenic hormones (Mauras et al 1996, Caufriez 1997). In addition to this action, sex steroids have been shown to stimulate skeletal growth directly and thus augment the role of GH in promoting somatic growth and development (Giustina et al 1997). During the adolescent growth spurt, there is a greater gain in sitting height (mean 13.5 cm for girls, 15.4 cm for boys) than leg length (mean 11.5 cm for girls, 12.1 cm for boys) as the spurt for sitting height lasts longer (and is longer in boys). At this time, there is a minimum rate of fat gain and maximum attainment of muscle bulk, with differential and opposite changes in boys and girls. Ossification of the skeleton, as measured in terms of ‘bone age’, is helpful in determining a child’s progress through puberty and as a determinant of final stature.

Rising concentrations of GH are also believed to exert some effects on circulating insulin levels, although precise mechanisms are unclear. Some authors believe that GH induces peripheral insulin resistance which leads to compensatory increases in insulin secretion (Smith et al 1988, Nobels and Dewailly 1992). Increased levels of insulin during puberty may directly stimulate protein anabolism (Amiel et al 1991). Insulin also acts as a regulator of insulin-like growth factor-1 (IGF-1) through its effects on insulin-like growth factor binding protein-1 (IGFBP-1). IGF-1 is produced by hepatic cells under the influence of GH, and has actions which stimulate cellular growth and maturation. IGFBP-1 competes with IGF receptors to bind IGF-1, thus inhibiting cellular action of IGF-1. Studies have indicated that insulin acts to suppress production of IGFBP-1 and therefore increases circulating IGF-1 bioavailability (Holly et al 1989, Nobels and Dewailly 1992). Insulin has also been shown to be a regulator of free sex steroids through the control of sex-hormone-binding globulin (SHBG) production by the liver. Studies have shown that levels of SHBG decrease during puberty and that this fall parallels the rising levels of insulin (Holly et al 1989, Nobels and Dewailly 1992). The endocrine interactions involved in normal pubertal development are represented in a simplified diagram in Figure 14.1.

Figure 14.1 Schematic representation of the endocrine changes which regulate pubertal development. GnRH, gonadotrophin-releasing hormone; GHRH, growth-hormone-releasing hormone; LH, luteinizing hormone; FSH, follicle-stimulating hormone; SHBG, sex-hormone-binding globulin; IGFBP-1, insulin-like growth factor binding protein 1; IGF-1, insulin-like growth factor-1.

Tempo of pubertal development

Although the endocrine changes during puberty have become better understood, the exact ‘trigger’ which determines the onset of pulsatile GnRH release, and thus the initiation of puberty, is still unclear. The question of which factors might regulate the onset of puberty and control the rate of pubertal development has prompted extensive research. The majority of studies examining young women have used age at menarche as the key marker of sexual maturity, and have attempted to elucidate which factors are involved in the timing of this event.

The first physical sign of puberty in girls is breast development, which occurs at an average age of 10–12.5 years (Figure 14.2). Age at onset of puberty is influenced by family history, race (earlier in Afro-Caribbeans) and nutrition (earlier in obese girls, with a secular trend to younger ages). Fifty percent of girls show signs of breast development by the age of 11 years, and investigations should be performed if there is no sign of breast development by the age of 14 years. In girls, the adolescent growth spurt occurs around 1 year after pubertal onset, while menarche occurs in the later stages of puberty at an average age of 12–15 years, at breast and pubic hair stage 4 (Figures 14.3 and 14.4; Marshall and Tanner 1969). Oestradiol concentration affects both the uterus and skeletal maturity. Most girls begin to menstruate when their bone age is between 13 and 14 years, and fewer than 20% are outside these limits. The appearance of axillary hair may antedate the rest of puberty by a few years and does not bear a constant relationship to the development of breasts as, again, it is androgens and not oestrogens that are most significant. The apocrine glands of the axilla and vulva begin to function at the time when axillary and pubic hair is appearing.

Figure 14.2 Stages of breast development at puberty. Stage 1: the infantile stage, which persists from the time the effect of maternal oestrogen on the breasts disappears, shortly after birth, until the pubertal changes begin. Stage 2: the bud stage. The breasts and papillae are elevated as a small mound and there is an increase in the diameter of the areola. This stage represents the first indication of pubertal change in the breast. Stage 3: the breasts and areola are further enlarged to create an appearance similar to that of a small adult breast, with a continuous rounded contour. Stage 4: the areola and papilla enlarge further to form a secondary mound projecting above the contour of the remainder of the breast. Stage 5: the typical adult breast with a smooth rounded contour. The secondary mound present in stage 4 has disappeared.

Figure 14.3 Stages of pubic hair development at puberty. Stage 2: sparse growth of slightly pigmented hairs on either the labia or the mons pubis. Stage 3: the hair is darker and coarser, and spreads sparsely over and on either side of the midline of the mons pubis. Stage 4: the hair is adult in character but covers a smaller area than in most adults and has not spread to the medial surface of the thighs. Stage 5: hair distributed as an inverse triangle and spreading to the medial surfaces of the thighs. It does not spread to the linea alba or elsewhere above the base of the triangle.

Figure 14.4 Timing of pubic hair and breast changes and menarche. Horizontal line represents mean ± 2 standard deviations.

It has long been recognized that early maturing women have a tendency to be heavier for their height than their later maturing counterparts and, in contrast, that malnutrition and anorexia nervosa are associated with delayed menarche and amenorrhoea. These observed relationships between body weight and menarche led to the ‘critical weight hypothesis’ of Frisch and Revelle (1970), which hypothesized that attainment of a critical body weight led to metabolic changes which, in turn, triggered menarche. In further work, the investigators demonstrated that the ratio of lean body weight to fat decreased during the adolescent growth spurt in females from 5 : 1 to 3 : 1 at menarche, when the proportion of body fat was approximately 22% (Frisch 1990). They suggested that adipose tissue specifically was responsible for the development and maintenance of reproductive function, through its action as an extragonadal site for the conversion of androgens to oestrogens.

While few authors will dispute that nutrition and body weight play a role in pubertal development, the ‘critical weight hypothesis’ for menarche has not been supported by other studies. Cameron (1976) studied 36 British girls longitudinally, measuring weight and skinfold thickness at 3-month intervals for the 2 years before menarche to 2 years post menarche. He postulated that if a ‘critical body fatness’ was the true explanation for onset of menarche, he should expect to detect a reduction in the variability of his measurements at menarche compared with measurements taken before and after menarche, and this was not demonstrated in the study. In a much larger study, Garn et al (1983) analysed the triceps skinfold distributions of 2251 girls collected during three national surveys in the USA. Comparing premenarcheal and postmenarcheal girls, they confirmed that those who had attained menarche were, on average, fatter than premenarcheal girls. However, there was a marked overlap in skinfold thickness for both groups, and there was no evidence of a threshold level of fatness below which menarche did not occur. In a more recent study, de Ridder et al (1992) performed a longitudinal assessment of 68 premenarcheal school girls, including pubertal staging, skinfold measurements, waist–hip ratios and blood samples for the monitoring of gonadotrophins. They did not demonstrate a relationship between body fat mass and the age at onset of puberty or age at menarche, but did find that a greater body fat mass was related to a faster rate of pubertal development.

Although the studies above dispute the association of age at menarche with the achievement of a threshold weight or body fat percentage, the association of earlier maturation in heavier girls still exists. Stark et al (1989) reported on data from 4427 girls who were part of the National Child Development Study cohort. Data relating to birth weight, age at menarche, and weights and heights measured at 7, 11 and 16 years were available. The authors reported that a larger proportion of girls with early menarche (before the age of 11 years) were heavier for their height at all ages when compared with those with late menarche (after 14 years). Interestingly, however, changes in relative weight in the years preceding menarche (ages 7–11 years) were not strongly associated with age at menarche, whereas being overweight at the age of 7 years was much more strongly associated with an early age at menarche. Birth weight was not related to age at menarche. The authors concluded that the increase in weight for height associated with early maturation actually begins well before the onset of puberty. These findings were replicated by Cooper et al (1996) who examined 1471 girls in the Medical Research Council National Survey of Health and Development. They found that girls who were heavier at 7 years of age experienced an earlier menarche; however, contrary to Stark et al, they found that birth weight was related to age at menarche and that girls with heavier birth weights experienced menarche at a later age.

Intense exercise, such as long-distance running, ballet, rowing, long-distance cycling and gymnastics, is associated with delayed menarche in young girls and with amenorrhoea in older women (Malina et al 1978, 1979, Frisch et al 1981, Baxter-Jones et al 1994). These ‘endurance’ sports are associated with lower body weight and percentage fat. The extent to which menarche is ‘delayed’ has been shown to be related to the age at which participation in the sport begins and to the intensity of training (Frisch et al 1981). In view of the association between body weight and menarche described above, it is perhaps not surprising that girls participating in intense sporting activity experience a later age of menarche.

The influence of genetic factors on age at menarche is evident in the correlation demonstrated between mother–daughter and sister–sister pairs. Garn and Bailey (1978) analysed data from 550 mother–daughter pairs and reported a correlation of 0.25, which was adjusted to 0.23 once a correction was made for fatness. Similar correlation coefficients have been reported in other studies of mother–daughter pairs from European, African and American populations (Flug et al 1984, Benso et al 1989, Malina et al 1994, Cameron and Nagdee 1996). The strength of this ‘genetic effect’, however, must be interpreted carefully as the mothers’ age at menarche was based on recall in most of these studies, and the accuracy of the estimation is therefore questionable. In addition, one must consider the effect of similarity in socio-economic status (and hence nutrition and fatness) that is likely to exist between mothers and daughters, and this in itself may contribute to similarity in menarcheal age.

The association between body weight and age at menarche has formed part of the basis for the explanation of the secular trend towards earlier menarche noted in the UK and other industrialized countries over the last century (Roberts et al 1971, Ostersehlt and Danker Hopf 1991). It has been generally accepted that this trend has reflected improvements in nutrition, health and environmental conditions. However, a recent plateau in this trend has been observed, and even reversal in some countries, which is at present unexplained (Dann and Roberts 1993, Tryggvadottir et al 1994, Veronesi and Gueresi 1994). It has been suggested that the recent decrease in menarcheal age is related to changes in social conditioning in developed countries. Promotion of very thin women as ‘ideal’ role models through media and fashion has contributed to an increase in dieting in adolescent girls. Increasing participation of women in endurance sports which have intense training regimens may also be a factor related to lower body weight in young girls.

These studies all indicate that pubertal development and body weight are intrinsically linked. However, the mechanism for this relationship has not been defined conclusively. Insulin has been suggested as a modulator of the tempo of pubertal development through the regulation of IGFBP-1 and SHBG (Smith et al 1988, Holly et al 1989, Nobels and Dewailly 1992). States of overnutrition and obesity are associated with increased serum concentrations of insulin (Parra et al 1971, Rosenbaum et al 1997). Therefore, if excessive nutritional intake persists during childhood, it is possible that hyperinsulinaemia may lead to lower levels of IGFBP-1 and reduced SHBG concentrations, thus enhancing IGF-1 and sex steroid bioavailability. The converse would be true in states of malnutrition, where low levels of insulin would allow for the development of increased IGFBP-1 and SHBG levels.

However, it is still unclear whether hyperinsulinaemia in childhood is a result of obesity or if it is the cause of obesity. The roles of genetic factors, which may determine insulin production and obesity risk in childhood, are also yet to be clearly explained. Recently, there has been much interest in the actions of the newly identified hormone leptin which is produced by adipose tissue (Sorensen et al 1996). Serum concentrations of leptin, a 167 amino acid product of white fat, have been shown to be related to body fat mass, and it is believed that leptin exerts its action on the hypothalamus to control calorie intake, decrease thermogenesis, increase levels of serum insulin and increase pulsatility of GnRH (Considine et al 1996, Sorensen et al 1996). With these actions, leptin may potentially have a role in the hormonal control of pubertal development, and some studies have shown that leptin levels do increase before the onset of puberty (Clayton et al 1997, Mantzoros et al 1997). It has been postulated that changes in circulating leptin levels may act as an initiator for the onset of puberty, and that this may explain the relationships between body fat and maturation observed by Frisch and Revelle (1970). This hypothesis is supported by the recent work of Clement et al (1998), who identified a homozygous mutation of the leptin receptor gene which results in early-onset morbid obesity and the absence of pubertal development in association with reduced GH secretion. These findings suggest that increased leptin levels associated with gains in fat mass may signal the hypothalamus to act as an important regulator of sexual maturation. It is possible that future research may clarify the role of leptin and leptin receptors in pubertal development, and may explain the variation in the timing of onset of puberty between individuals.

Ovarian and uterine development

Development of primordial follicles and secretion of oestrogens may be regarded as the final common path in the hormonal activity of puberty, and there is a close correlation between oestrogen levels and sexual maturation. The uterus grows in concordance with somatic growth, with differential increased growth of the corpus starting from approximately 7 years of age. However, the main differential increase of uterine size compared with somatic growth is only obvious after oestradiol secretion is measurably increased, and tends to occur between breast stages 3 and 4. Awareness of this comparatively late relative increase in uterine size has clinical relevance in the differentiation of arrested puberty and Müllerian agenesis, and also in the diagnosis of causes of precocious puberty. Blood oestrogen levels rise before and during puberty, and continue to do so for approximately 3 years after menarche before the normal adult follicular levels of oestrogen are seen. This gradual maturation is a reflection of the finding that many of the early cycles are anovulatory, and of the observation that the early postmenarcheal years are relatively infertile.

Ultrasound is the most suitable imaging technique for examination of the internal genitalia of girls as it is free from the risks of radiation and involves a quick, quiet and non-invasive procedure. Whilst higher resolution images can be obtained of the ovaries and uterus using transvaginal ultrasonography, only the transabdominal approach can be used in young girls and it may not always be possible to visualize both ovaries; furthermore, a full bladder is required. Nonetheless, transabdominal ultrasonography is an invaluable tool for the delineation of normal changes before and through puberty, and also for the evaluation of paediatric disorders, such as pelvic masses and abnormal pubertal development. Ultrasonography can also be used in postmenarcheal girls to detect abnormal ovarian morphology and hence increase our understanding of common conditions such as polycystic ovary syndrome.

There are relatively few normal data on the maturation of the internal genitalia in girls, and a paper by Holm et al (1995) provides cross-sectional data of a large cohort of 166 school girls and medical students. It was demonstrated that uterine and ovarian volumes increase in size prepubertally, from as early as 6 years of age. Griffin et al (1995) demonstrated similar changes from the age of 4 years and, in their study of 153 girls, included subjects as young as 3 days. It was found that uterine shape changes constantly throughout childhood from a pear shape to a cylindrical shape and, finally, to the adult heart shape.

As girls go through puberty, the ovaries have been described as characteristically becoming multicystic due to low levels of FSH stimulating only partial folliculogenesis (Stanhope et al 1985). The multicystic appearance is also seen during resumption of ovarian activity after periods of quiescence; for example, in women who are recovering from weight-related amenorrhoea or those with hypogonadotrophic hypogonadism who are treated with pulsatile GnRH. The multicystic ovary differs from the polycystic ovary in that the cysts are larger (6–10 mm) and the stroma is of normal echogenicity. The morphological appearance of the polycystic ovary, on the other hand, requires the presence of at least 10 cysts (2–8 mm) per ovary in the presence of an echodense stroma (see Chapter 18).