Management and outcome of central precocious puberty

Abstract

Precocious puberty is a condition which has both important and diverse consequences for affected children and their families. The impact of precocious pubertal development on the patients is both physical and psychological. Because of the potential for a rapid, dynamic course of the disease it is necessary to make a correct diagnosis without delay in order to allow a judgement about the prospective course of the disease in terms of rate of pubertal progression, statural growth, bone age progression, development of reproductive functions and psychosocial adjustment and well-being. It is not surprising that there has been a controversial discussion about the indications for treatment, as precocious puberty is characterized by a great diversity in clinical presentation, in underlying causes for precocious development and thus, also, in its long-term prognosis. In this paper, we review some of the salient clinical, laboratory, and radiological features of the possible causes of precocious puberty and discuss the current treatment options as well as the long-term outcome. Puberty is the transitional period between childhood and adulthood during which human development progresses from the first appearance of secondary sexual characteristics to full sexual maturation with the capacity to reproduce. The onset of pubertal development and the progression to full maturity is controlled and regulated by several neuroendocrine factors and hormones (for review see Grumbach & Styne, 1998). The maturation of reproductive function depends on an increasing pulsatile secretion of GnRH into the hypophysial portal circulation, which in turn stimulates the pulsatile release of LH and FSH into the peripheral circulation (Boyar et al., 1974; Clarke & Cummins, 1982; Jakacki et al., 1982; Caraty & Locatelli, 1988; Wu et al., 1989, 1990; Oerter et al., 1990). LH stimulates Leydig cell secretion of testosterone in boys and ovarian gestagen secretion in girls after onset of ovulation. FSH has very little effect in boys until spermarche; in girls it stimulates follicle formation and oestrogen production. The physical changes of puberty are the result of the activation of the hypothalamic–pituitary-gonadal axis. In 95% of healthy girls puberty starts with the enlargement of the breasts (thelarche) between 8·5 and 13 years of age. This stage is generally followed by pubic and axillary hair development. The average time between first pubertal signs and menarche is 2·3 years. Most girls reach menarche at a mean (± SD) age of 13·5 ± 1·0 years (Marshall & Tanner, 1969). It has recently been suggested that the onset of puberty in U.S. girls seen in the paediatric office occurs earlier than previously thought (Herman-Giddens et al., 1997), however, several field studies have shown very little change in the mean age at menarche [Buckler, 1990 (13·35 ± 1·14 years), Engelhardt et al., 1995 (13·46 years = 50th centile), Fredriks et al. 2000 (13·15 years = 50th centile), Largo & Prader, 1983a (13·4 ± 1·1 years)]. Ethnic differences have also to be taken into account (Proos et al., 1991; Herman-Giddens et al., 1997). In 95% of boys pubertal development usually begins with testicular enlargement between the ages of 9·5–13·5 years (Largo & Prader, 1983b; Marshall & Tanner, 1970), followed by pubic hair development. The mean interval between initial genital development and mature external genitalia lasts 3 years (approximately 1 year per next genitalia stage). The pubertal growth spurt begins in girls in early puberty and in boys during mid-puberty (Marshall & Tanner, 1969, 1970; Largo & Prader, 1983a, 1983b). Precocious puberty is generally defined as the appearance of secondary sex characteristics before the age of 8 years in girls and 9 years in boys (Grumbach & Styne, 1998). In girls menarche before the ninth birthday may serve as an additional criterion. These diagnostic age limits have a statistical basis. However, they were chosen somewhat arbitrarily and may be subject to change over time (Kaplowitz & Oberfield, 1999). The arbitrary nature becomes obvious when one considers that the diagnostic age limit for thelarche corresponds to approximately –2·5 to –3·0 SD below the normal mean age [Largo & Prader, 1983a (10·9 ± 1·2 years); Marshall & Tanner, 1969 (11·2 ± 1·1 years)], and not the usual limit of –2·0 SD, while the cut-off age for menarche is in the range of –4 SD [Engelhardt et al., 1995 (11·32 years = 3rd centile); Fredriks et al., 2000 (11·77 = 10th centile)]. In addition, it is important to notice that two percent of healthy girls may show a pubertal stage B2 before their 8th birthday (Largo & Prader, 1983a). In boys the diagnostic age limit of 9 years for Tanner stage G2 corresponds to approximately –2·5 SD (Marshall & Tanner, 1970; Largo & Prader, 1983b; Willers et al., 1996). The aetiology of precocious puberty is variable and it is of utmost importance both for diagnostic strategy and treatment, and for distinguishing between central precocious puberty (CPP) which results from premature activation of the hypothalamic–pituitary-gonadal axis (GnRH-dependent) and GnRH-independent pseudoprecocious puberty (Fig. 1). Owing to the sequence of hormonal events, CPP is usually consonant although there may be marked variations in the effects of sex steroids on the development of secondary sex characteristics, height velocity and bone maturation. Furthermore, partial or incomplete forms of precocious puberty should be distinguished from complete forms, as should persistent from transient forms of CPP (Table 1). The recognition of transient forms is of particular importance in order to avoid unnecessary treatment in these patients (Palmert et al., 1999; Partsch et al., 1998) and so that outcome results of (unnecessary) treatment are not confused with the natural course of the disease (Partsch et al., 1999c). It is interesting to note that in some rare cases organic CPP may also be transient (Brauner et al., 1987a). Although it may be desirable for didactic reasons to think of CPP as one diagnostic category, it does not present as a homogeneous clinical picture. In contrast, CPP is much more a continuum of clinical presentation and rate of progression (Fig. 2), ranging from normal variants or incomplete forms of pubertal development to slowly progressive or transient forms and rapidly progressive forms (Pescovitz et al., 1986; Brauner et al., 1987b; Kreiter et al., 1993; Partsch et al., 1998; Palmert et al., 1999; Léger et al., 2000; Stanhope, 2000). Diagnostic scheme for the differentiation of gonadotropin-dependent from gonadotropin-independent precocious puberty (adapted from Partsch & Sippell, 2001). Gonadotropin-dependent or CPP is characterized by a pubertal LH response to a standard GnRH test (e.g. 60 µg/m2 i.v.). The hatched bar shows the reference range for stimulated LH in healthy prepubertal girls (Partsch et al., 1990). In addition, pulsatile spontaneous LH secretion is found at night (left panel). In contrast, in children with gonadotropin-independent or peripheral PPP no spontaneous LH pulses are present and the LH response to a GnRH challenge is absent or minimal (right panel). CAH = congenital adrenal hyperplasia. The spectrum of precocious pubertal development ranges from girls with the normal variant premature thelarche to those girls with the rapidly progressive forms of complete CPP. Persistent growth acceleration, early menarche, progressive bone age acceleration, and, thus, deterioration of final height potential is seen in the latter patients only. A progression from premature thelarche to overt progressive CPP is possible (Pasquino et al., 1995). In addition, a slowly progressive CPP may develop into a rapidly progressive form and vice versa (Pasquino et al., 1989; Schwarz et al., 1990; Kolmer et al., 1991; Partsch et al., 1998). Figure 3 gives an overview of the spectrum of the clinical, hormonal, and auxological findings seen in girls with progressive CPP. In rapidly progressive CPP, pubertal development is highly accelerated and proceeds much faster than in normal puberty (Schroor et al., 1995). Furthermore, the hormonal milieu is pathological: spontaneous LH pulses are of increased amplitude and plasma LH shows markedly increased levels after GnRH stimulation compared with the normal range for the respective pubertal stage (Partsch et al., 1989). Variance (ranges) of auxological and hormonal parameters in a cohort of 40 girls with progressive CPP at start of treatment, demonstrating the marked variability (Partsch et al., 1999c). Central precocious puberty is a rare disorder with an estimated incidence of 1 : 5000–1 : 10 000 (Gonzalez, 1982; Cutler, 1988). The sex distribution (girls to boys) is reported to be between 3 and 1 (Kappy & Ganong, 1994) and 23 and 1 (Bridges et al., 1994). There is a predisposition for the development of CPP in children with neurofibromatosis type 1 (2·4–5%: Habiby et al., 1995, 1997; Cnossen et al., 1997; Carmi et al., 1999; Virdis et al., 2000), in children with hydrocephalus (10–11%: Abdolvahabi et al., 2000; Brauner et al., 1987a; De Luca et al., 1985; Kaiser et al., 1989; Lopponen et al., 1996; Siddiqi et al., 1999; Tomono et al., 1983), in children with meningomyelocele (5–33%: Hunt, 1990; Meyer & Landau, 1984; Trollmann et al., 1996, 1998), in children with neonatal encephalopathy (4·3%: Robertson et al., 1990), and in children after low dose cranial irradiation (Brauner et al., 1984, 1999; Ogilvy-Stuart et al., 1994). An association of Williams–Beuren syndrome with an increased frequency of CPP was recently reported (Scothorn & Butler, 1997; Cherniske et al., 1999; Partsch et al., 1999b). Furthermore, children adopted from developing countries are prone to develop CPP (Bourguignon et al., 1992; Proos et al., 1991; Tuvemo & Proos, 1993; Virdis et al., 1998; Bureau et al., 1999; Tuvemo et al., 1999; De Monleon et al., 2000; Krsterska-Konstantinova et al., 2001; Mul et al., 2001). It has been postulated that CPP may be found in up to 25% of girls seen at one institution for precocious puberty (Bourguignon et al., 1992) and in up to 45% of adopted girls and in a much smaller percentage in boys (8·6%; Baron et al., 2000). While an underlying CNS pathology is found in a subset of CPP patients, this is not the case in many others even with the use of modern imaging techniques. The distribution of idiopathic vs. organic CPP varies widely: 69% to 98% idiopathic cases in girls and 0% to 60% in boys, respectively (Thamdrup, 1961; Pescovitz et al., 1986; Brauner & Rappaport, 1993; Bridges et al., 1994; Kappy & Ganong, 1994; Carel et al., 1999; Heger et al., 1999; Partsch et al., 1999b; Cisternino et al., 2000; De Sanctis et al., 2000; Mul et al., 2000). Thus, the risk for organic CPP is much higher in boys than in girls. In addition, the risk of a CNS lesion as the cause for CPP is higher in younger than in older patients (Cassio et al., 2000; Cisternino et al., 2000). An overview of the various aetiologies of CPP is given in Table 1. Hypothalamic hamartomas are congenital, non-neoplastic tumour-like lesions formed by heterotopic grey matter, neurones, glial cells and fibre bundles (Inoue et al., 1995). They are usually located at the base of the brain at the floor of the third ventricle, near the tuber cinereum or near the mamillary bodies. Owing to the nature of hypothalamic hamartomas as congenital malformations, they frequently cause precocious puberty at a relatively early age, in some cases even starting at birth (Albright & Lee, 1992; Guibaud et al., 1995; De Brito et al., 1999). Some hypothalamic hamartomas are associated with gelastic seizures which are often resistant to anticonvulsive treatment (Cascino et al., 1993; Marliani et al., 1991; Nishio et al., 1994; Fukuda et al., 1999; Regis et al., 2000; Rosenfeld et al., 2001). Highly suggestive signs for the diagnosis of a hypothalamic hamartoma are: (i) precocious onset of pubertal development at a very young age [usually < 4 years; often < 2 years of age (Cassio et al., 2000; Comite et al., 1984; Colaco et al., 1993; Mahachoklertwattana et al., 1993; Kornreich et al., 1995; de Brito et al., 1999; Jung et al., 1999)]; (ii) hormonal findings compatible with central activation of the GnRH pulse generator (i.e. pulsatile gonadotropin secretion and increased LH response in the GnRH stimulation test); (iii) the demonstration of an isointense tumour in typical location showing no gadolinium enhancement on magnetic resonance imaging (MRI) (Fig. 4); (iv) the absence of any sign of precocious pseudopuberty (i.e. absence of biochemical tumour markers, such as β-HCG and AFP). The younger the child, the higher the chance that the presence of a hypothalamic hamartoma is the cause of CPP (Cisternino et al., 2000). Hypothalamic hamartomas are responsible for CPP in 2–28% of patients [2·3% (Cacciari et al., 1990; Cisternino et al., 2000); 13·3% (De Sanctis et al., 2000); 14% (Hibi & Fujiwara, 1987); 12·9% (Kornreich et al., 1995); 28% (Lyon et al., 1985); 21% (Partsch et al., 1999c); 10% (Robben et al., 1995); 11·5% (Sharafuddin et al., 1994)]. Midsagittal T1-weighted MR image after gadolinium injection in an 11 month old boy with organic central precocious puberty (GnRH test: LH 1·2 → 12.7 IU/l, FSH 0·4 → 0·6 IU/l, testosterone 2.4 nmol/l, bone age acceleration to 2.5 years) showing hypothalamic hamartoma (arrow) without contrast enhancement. MR imaging is of particular importance in the diagnosis of hypothalamic hamartomas as these may be overlooked on cranial computerized tomography (CT) scans (Kornreich et al., 1995; Robben et al., 1995; Feilberg Jorgensen et al., 1998; De Sanctis et al., 2000) and histological examination will not be carried out in most patients. The typical MRI picture is that of an isointense structure on T1-weighted images which may be isointense or slightly hyperintense on T2-weighted images (Fig. 4). The question of the adequate and optimal treatment of children with hypothalamic hamartoma and CPP has been discussed controversially in the literature (Siegel-Witchel, 1995; Starceski et al., 1990). Over the past five decades neurosurgical treatment has been associated with considerable risk and high neurological morbidity (Zuniga et al., 1983; Breningstall, 1985; Sato et al., 1985; Berkovic et al., 1997). In general, the paediatric, and recently, also, neurosurgical recommendation is that long-acting GnRH agonists are the first choice of treatment in patients with precocious puberty due to hypothalamic hamartomas (Feuillan et al., 1999; Stewart et al., 1998; Partsch et al., 1999a). Successful suppression treatment has been reported by several groups for a duration of up to 8·4 years (Chamouilli et al., 1995; Comite et al., 1984; Mahachoklertwattana et al., 1993; Stewart et al., 1998; de Brito et al., 1999; Feuillan et al., 1999; Ishii et al., 1999). Long-term studies and outcome data after treatment with GnRH agonists are favourable and do not show negative sequelae (de Brito et al., 1999; Feuillan et al., 1999, 2000; Heger et al., 1999). In particular, depot preparations ensure an adult height within the genetic height potential with normal body proportions, bone density and reproductive function (Feuillan et al., 2000; Heger et al., 1999). Long-term exposure to sex steroids leads to an increased growth rate, to an accelerated bone age development, and to the maturation of hypothalamic centres that are important for the initiation of puberty. Efficient treatment of the primary disease causes a decrease in sex steroid levels and thereby activates the hypothalamic GnRH pulse generator via a feedback system that has matured prematurely during the period of exposure to sex steroids. Thus, secondary central precocious puberty develops. The most important examples in this category of precocious puberty are congenital adrenal hyperplasia (Wilkins & Cara, 1954; Penny et al., 1973; Pescovitz et al., 1984; Pouw et al., 1986; Dacou-Voutetakis & Karidis, 1993; Soliman et al., 1997; Frenzel & Doerr, 1998) and familial or sporadic male-limited precocious puberty (Bertelloni et al., 1997; Holland et al., 1987; Holland, 1991; Boepple et al., 1992; Laue et al., 1993; Gromoll et al., 1998; Leschek et al., 1999). Secondary central precocious puberty has also been described in a few patients with McCune Albright syndrome (Eugster et al., 1999; Foster et al., 1985; Kaufman et al., 1986; Lee et al., 1986; Feuillan et al., 1993; Schmidt & Kiess, 1998). After initiation of effective treatment of the primary, gonadotropin-independent puberty, in the majority of cases the course of these diseases is complicated by secondary central precocious puberty (Laue et al., 1989, 1993; Leschek et al., 1999). The occurrence of CPP depends on the maturational status of the children (Holland et al., 1987). In most patients secondary CPP occurs at bone ages between 10 and 13 years (Holland et al., 1987; Laue et al., 1989, 1993; Holland, 1991; Feuillan et al., 1993; Bertelloni et al., 1997; Gromoll et al., 1998; Leschek et al., 1999; Latronico et al., 2000b). However, some patients may experience CPP at higher bone ages (Boepple et al., 1992; Leschek et al., 1999) or, in some cases, not at all, despite older or even adult bone ages (Feuillan et al., 1993; Latronico et al., 2000a). The time course of development of secondary CPP is quite variable, but CPP may develop as early as one month after initiation of treatment of peripheral precocious puberty (PPP) (Holland et al., 1987). It is noteworthy that secondary CPP may also occur spontaneously without any treatment (Egli et al., 1985). In most cases secondary CPP is permanent, however, one case of ‘transient’ central precocious puberty secondary to nonclassic 21-hydroxylase deficiency has been reported (Speiser, 1995). The patient’s history should include the age at onset of pubertal development, the rate of progression of pubertal signs, the growth pattern during the last 6–12 months, the presence of secondary sexual characteristics and additional signs of puberty (acne, oily skin, erections, nocturnal emissions in boys and vaginal discharge and menstrual bleedings in girls). A positive family history of precocious puberty may suggest familial cases of CPP or familial male precocious puberty. The physical examination includes the pubertal stages according to Tanner (Marshall & Tanner, 1969; 1970) and the measurement of height, weight and body proportions. A growth curve including all available height data should be plotted (Fig. 5). Height velocity is calculated on the basis of the available height data: most patients with CPP have a height velocity > 75th centile. To determine the biological age of the child an X-ray of the nondominant (left) hand and wrist is taken. If the bone age (BA) is accelerated by more than 2 standard deviations for chronological age (CA) it is unlikely that the child has a normal variant of pubertal development. Whenever possible, the ratio of ΔBA/ΔCA should be calculated over a pretreatment observation period. This ratio is above 1·2 in the majority of patients with progressive CPP (Galluzzi et al., 1998; Partsch et al., 1999c). Growth curve of a girl with idiopathic central precocious puberty (GnRH test: LH 5·2 → 54·8 IU/l, oestradiol 130 pmol/l) of early onset at age 1·4 years. The child was first treated with daily Buserelin [(Bus) Hoechst Marion Roussel, Bad Soden, Germany] intranasally and subcutaneously between 2 and 4 years of age and was then switched to triptorelin depot by monthly intramuscular injection (Decapeptyl Depot, Ferring Arzneimittel, Kiel, Germany). Height for age (open circles), height for bone age (Greulich & Pyle, 1959; closed squares), and predicted adult height (Bayley & Pinneau, 1952; open diamonds) are shown until final height. TH = target height. The inset shows height SDS for bone age (BA; closed circles) during development. Note the rapid deterioration from age 2–4 years (during Buserelin treatment) and the subsequent improvement to a normal final height SDS (closed square). Bone age progression was effectively halted at 11 years between 5 and 8·8 years of chronological age. The initial hormonal evaluation includes plasma sex steroids and a standard GnRH test (e.g. GnRH 60 µg/m2 i.v.; LH and FSH measurement at 0 and 30 minutes; other protocols are available and may be used). In our opinion it is sufficient to take one plasma sample at 30 minutes after GnRH stimulation. There is no additional diagnostic value of the GnRH test with multiple samples (Cavallo et al., 1995). The GnRH test is important for the differentiation between central and gonadotropin-independent precocious puberty (Fig. 1; Brito et al., 1999). CPP is characterized by a pubertal response of LH to GnRH (cut-off levels for a pubertal LH level depend on the LH assay –Lee, 1994; Iughetti et al., 2000) with a predominant LH response compared to the FSH response (Partsch et al., 1989; Pescovitz et al., 1988). It is remarkable that the GnRH-stimulated LH concentration is not only higher than the prepubertal range but is also in excess of the normal range for the respective pubertal stage in 55% of girls with CPP (Fig. 6, Partsch et al., 1989). Basal and GnRH stimulated LH levels (mean ± SEM) are shown in 71 girls with CPP at diagnosis and for each half year of treatment with triptorelin depot until 48 months. Before treatment the mean LH response to GnRH was increased, not only compared to the prepubertal reference range but also compared to the reference range for late pubertal girls with breast stage five (hatched lines; Partsch et al., 1990). During treatment, the mean LH response to GnRH was suppressed. Six to 12 months after cessation of treatment stimulated plasma LH showed normal pubertal levels. Hyperstimulated LH levels were seen in a few girls only. With conventional radioimmunoassays or immunoradiometric assays, basal gonadotropin levels are of no great diagnostic value for CPP (see Fig. 6). With the new 3rd generation highly sensitive immunochemiluminometric (ICMA) assays it is possible to detect elevated basal serum gonadotropin levels (Neely et al., 1995a, 1995b). Elevated basal LH serum levels strongly correlate with an elevated GnRH-stimulated LH, and may therefore be a useful screening test for identifying patients with CPP (Neely et al., 1995b). However, the intravenous GnRH test still remains the single most important test in the diagnostic work-up of children suspected of having CPP (Lee, 1994). Other tests proposed in the literature, including the subcutaneous GnRH test with a single sample drawn after 40 minutes for stimulated LH (Eckert et al., 1996; Lawson & Cohen, 1999) and the leuprolide stimulation test (Garibaldi et al., 1993; Ibanez et al., 1994), may also be used but have not been tested in larger numbers of children and, in the case of the latter, include multiple blood sampling. In general, equivocal test results should prompt a reassessment of the patient after 3–6 months of follow-up. All the typical laboratory findings will not be present in every CPP patient. Thus, the diagnosis not only depends on the hormonal data but also on the clinical picture. Basal plasma testosterone levels are elevated in CPP boys. The diagnostic value of estradiol is limited as approximately half of the CPP girls show low prepubertal estradiol levels (Partsch et al., 1999c). Additional laboratory investigations including thyroid function tests, 17-hydroxyprogesterone and hCG determination, may be necessary to exclude other causes of precocious puberty. Pelvic ultrasound is a helpful tool in the work-up of girls with CPP (Stanhope et al., 1985, 1986; Salardi et al., 1988; Pelzer et al., 1990; King et al., 1993; Bridges et al., 1995; Griffin et al., 1995b; Haber et al., 1995; Buzi et al., 1998; Jensen et al., 1998). The volumes of ovaries and uterus can be compared to age-specific reference data (Bridges et al., 1993, 1996; Salardi et al., 1985; Griffin et al., 1995a; Haber & Mayer, 1994). Important diagnostic sonographic parameters for CPP are the volumes of ovaries (bilateral enlargement is a reliable indicator of CPP –Ambrosino et al., 1994; King et al., 1993; Griffin et al., 1995b) and uterus (Haber et al., 1995; Salardi et al., 1988). Similarly useful are uterine length (Ambrosino et al., 1994), shape (tubular in prepuberty, pear-like in puberty; Bridges et al., 1996), fundal/cervical ratio (Griffin et al., 1995b; Bridges et al., 1996; Cassio et al., 2000), as well as ovarian structure (homogeneous or microcystic in prepuberty and multicystic or macrocystic/follicular in CPP patients; Buzi et al., 1998; Stanhope et al., 1986; Salardi et al., 1988; Cassio et al., 2000), and thickness or absence/presence of the endometrium (Griffin et al., 1995b). However, most authors conclude from their studies that none of these parameters can reliably separate CPP from normal girls or from those with premature thelarche and, thus, should not be used as the only diagnostic tool for identifying CPP patients (Andolf, 1999). For uterine volume only a sensitivity and specificity for CPP of 100% has been reported (Haber et al., 1995). Abdominal and pelvic sonography is indicated in patients with precocious puberty to detect adrenal or ovarian tumours or ovarian cysts. Magnetic resonance imaging of the CNS is necessary in every child with proven CPP in order to exclude a central nervous system (CNS) lesion as the cause of CPP (Table 1). Recent studies have shown in large numbers of patients that CPP may be the only presenting symptom of an intracranial tumour or malformation (Cacciari et al., 1983; Robben et al., 1995; Cisternino et al., 2000; Chemaitilly et al., 2001). Occult intracranial lesions can be expected in 4·8–13·3% of girls and in 19·2% of boys with CPP. Restricting neuroradiological imaging to certain subgroups of CPP patients, i.e. only boys (Bridges et al., 1994) or only young girls is not adequate as intracranial lesions can be the cause of CPP in both sexes and in all age groups [up to 6·5 years in boys (Chemaitilly et al., 2001) and up to 8 years in girls (Cisternino et al., 2000: 0·0–3·0 year: 12·9%, 4·0–6·9 year: 9·1%, 7·0–7·9 year: 7·4%)]. The most common CNS lesion associated with CPP is hypothalamic hamartoma. Apart from CPP hamartomas are usually asymptomatic and may only be detected by MRI (Robben et al., 1995; Feilberg Jorgensen et al., 1998). Magnetic resonance imaging in all patients with proven CPP provides the chance to detect some CNS tumours early [i.e. optic glioma, astrocytoma, pinealoma (Chemaitilly et al., 2001; Cisternino et al., 2000)]. There is no general consensus in the literature about the indication for treatment in CPP children (Kaplan & Grumbach, 1990; Rosenfield, 1994; Oerter-Klein, 1999; Léger et al., 2000). However, there is consensus that not all patients with CPP need medical intervention. The indication for treatment may be either psychosocial/behavioural or auxological or a combination of both. Surprisingly, only few psychological problems and long-term psychological sequelae have been described in CPP patients (Ehrhardt & Meyer-Bahlburg, 1986, 1994; Ehrhardt et al., 1984; Galatzer et al., 1984; Sonis et al., 1985; Jackson & Ott, 1990; Mouridsen & Larsen, 1992) and sexual abuse and early pregnancy seem to be rare events (Herman-Giddens et al., 1988; Thamdrup, 1961; Money & Walker, 1971). However, it is clear that girls and boys with CPP have earlier sexual activity than age-matched healthy girls (Ehrhardt et al., 1984; Thamdrup, 1961; Money & Walker, 1971; Ehrhardt & Meyer-Bahlburg, 1986). Particularly in patients with mental retardation and/or specific character traits as in Williams–Beuren syndrome, this early sexual activity may pose severe problems for the families and may be of potential harm to the children. We therefore believe that the psychosocial indication for CPP treatment has to be a very individualized decision (Rosenfield, 1994). With respect to the auxological indication there is much uncertainty about adequate criteria (Oerter Klein, 1999). However, some objective criteria exist, i.e. complete and progressive CPP and abnormal height potential (Rosenfield, 1994) or deterioration of final height prediction (Leger et al., 2000). From these criteria it becomes obvious that many children need to have careful follow-up examination before a decision about the indication for treatment can be made (Oerter-Klein, 1999; Leger et al., 2000). In contrast to others (Kaplan & Grumbach, 1990; Rosenfield, 1994) our experience is that sex steroid levels are only helpful in the decision process for or against treatment if constantly elevated to the pubertal range (Oerter-Klein, 1999). Clear cut-off values for the criteria proposed in the literature are not available. Suggested indications for treatment are presented in Table 2. GnRH agonist treatment is not indicated in patients lacking evidence of pubertal gonadotropin secretion, with slowly progressing CPP, and with uncompromised height potential (Kaplan & Grumbach, 1990; Rosenfield, 1994; Oerter-Klein, 1999; Léger et al., 2000). Treatment goals are normalization of psychosocial well-being, prevention of early menarche and early sexual activity, avoidance of abnormal body proportions and preservation of normal height potential in comparison to target height. With the availability of agonistic analogues of GnRH in depot formulations licensed for the use in children, treatment of CPP with cyproterone acetate or medroxyprogesterone has become obsolete because of insufficient effects regarding hormonal suppression and auxological outcome (Sorgo et al., 1987) and their potential for side-effects (Savage & Swift, 1981). The first reports of successful short-term pituitary-gonadal suppression in CPP patients by GnRH agonists date back to 1981 (Comite et al., 1981; Crowley et al., 1981; Laron et al., 1981). Since then num