Assess and manage normal puberty and its disorders, including precocious and delayed puberty and primary amenorrhoea

In one line

Puberty is the GnRH-driven reactivation of the hypothalamic–pituitary–gonadal axis, and almost every disorder of timing is read off two axes — whether the abnormal signal sits above the gonad (central/gonadotrophin-dependent) or at the gonad and beyond (peripheral/gonadotrophin-independent) — so the single decision that organises management is whether gonadotrophins are inappropriately high, low, or absent for the clinical stage in front of you.

Mechanism & pathophysiology

The pubertal cascade is the third act of a system that has run before. GnRH neurones, migrating from the olfactory placode into the hypothalamus in fetal life, fire in a pulsatile minipuberty in the first months after birth and then fall silent through childhood under central restraint. Puberty is the re-emergence of pulsatile GnRH secretion, not its first appearance — which is why a failure of GnRH neurone migration (Kallmann syndrome) and a failure of pubertal reactivation share the same final phenotype.

The brake on the prepubertal axis is now understood at the level of the kisspeptin–neurokinin B–dynorphin (KNDy) neurones of the arcuate nucleus, which sit upstream of the GnRH pulse generator. The permissive signal to release that brake integrates energy stores (leptin from adipose tissue), genetic timing and central maturation. Two mutational classes map directly onto the two clinical poles: loss-of-function mutations in KISS1/KISS1R or TAC3/TACR3 cause hypogonadotropic hypogonadism, while gain-of-function signalling — most importantly loss-of-function of MKRN3, an imprinted, paternally-expressed gene that normally restrains the pulse generator — releases the brake early and is now the commonest identified monogenic cause of familial central precocious puberty. MKRN3 is the molecular reason a precocious-puberty history can run down the paternal line.

Once GnRH pulses resume, the gonadotrophs translate frequency into the right mix of LH and FSH, and FSH drives the ovarian two-cell machinery — theca cells make androgen under LH, granulosa cells aromatise it to oestradiol under FSH — so that rising oestradiol produces the visible sequence. The order is itself diagnostic information: thelarche (breast budding, the first sign, oestrogen-driven) → pubarche (pubic then axillary hair, which is adrenal androgen-driven and can dissociate from the gonadal axis) → the growth spurt → menarche, typically about 2–2.5 years after thelarche at Tanner breast stage 4. Adrenarche — the rise in DHEAS from adrenal zona reticularis maturation — is a separate clock from gonadarche, and the dissociation matters: isolated premature pubarche reflects early adrenarche, not central puberty, and usually needs reassurance and exclusion of non-classic congenital adrenal hyperplasia rather than an MRI.

The growth spurt is the part most often mishandled in reasoning, because oestrogen is paradoxical at the growth plate. Low oestradiol concentrations promote chondrocyte proliferation and the pubertal height velocity peak; high concentrations accelerate epiphyseal fusion and close the plate. This single fact reconciles two clinical observations that otherwise look contradictory — that early sex-steroid exposure both makes a child taller than peers in the short term and shorter as an adult, because the same hormone that drives the spurt also spends the remaining growth potential prematurely. It is why bone age, not chronological age, is the currency of every height decision in this topic, and why a GnRH agonist preserves final height by protecting the growth plate from premature closure, buying back years of growth at a normal tempo.

Tanner staging quantifies this sequence (breast B1–B5, pubic hair PH1–PH5) and is the clinical currency of the assessment. The normal window matters as much as the sequence. The conventional thresholds — thelarche before 8 years as precocious, absent thelarche by 13 years or absent menarche by 15 years (or within 3 years of thelarche, or by 14 years with hirsutism/suspected outflow obstruction) as delayed/primary amenorrhoea — are not arbitrary cut-points but the tails of a population distribution, and a secular trend toward earlier thelarche (particularly in higher-BMI girls) means the lower boundary is genuinely contested rather than fixed.

The pathophysiology of the disorders falls out of this architecture. Precocious puberty is either central (premature reactivation of the whole GnRH axis — so gonadotrophins rise and the sequence is concordant) or peripheral (autonomous sex-steroid production downstream of a suppressed axis — so gonadotrophins are low and the sequence is often discordant, e.g. vaginal bleeding without progressive breast development, or virilisation). Within central precocious puberty the great majority of girls are idiopathic, but the differential that the imaging is designed to catch is the CNS lesion that has removed the central brake — a hypothalamic hamartoma (the classic ectopic GnRH-secreting lesion, often with gelastic seizures), optic-pathway glioma (consider neurofibromatosis type 1), other tumours, hydrocephalus, cranial irradiation, or post-inflammatory damage. The peripheral causes group by their steroid source: an autonomous ovarian follicular cyst or granulosa-cell tumour, an adrenal adenoma/carcinoma, congenital adrenal hyperplasia, McCune–Albright syndrome, exogenous oestrogen exposure, and — a frequently-missed mechanism — severe primary hypothyroidism, where very high TSH cross-reacts at the FSH receptor (the Van Wyk–Grumbach syndrome) to produce ovarian stimulation and even multicystic ovaries, the one peripheral precocity that comes with slow growth and a delayed bone age rather than acceleration, and that reverses completely on thyroxine. Delayed puberty and primary amenorrhoea are read the same way inverted: low gonadotrophins point above the gonad (constitutional delay, or hypogonadotropic hypogonadism), high gonadotrophins point to primary gonadal failure (hypergonadotropic hypogonadism — Turner syndrome, primary ovarian insufficiency), and a normally oestrogenised girl who does not menstruate points instead to an anatomical end-organ problem (outflow obstruction, or absent uterus in Müllerian agenesis and androgen insensitivity).

Assessment

The history and examination are designed to place the child on those two axes before any blood is drawn.

• History: age and sequence of each pubertal event (a documented timeline beats a single snapshot); growth pattern and parental heights (a target-height calculation and the child's own growth-chart trajectory); rate of progression (rapidly advancing signs are the ones that threaten final height); neurological symptoms (headache, visual change, seizures — a CNS driver of central precocious puberty); a careful drug/exposure history (oestrogen creams, anabolic supplements, endocrine-active products); and family history of pubertal timing (early menarche in mother/sisters; the paternal line for MKRN3-linked precocity). In a girl with absent or arrested puberty: anosmia or hyposmia (Kallmann), chronic illness, intense exercise or low weight (functional hypogonadotropism), and cyclical pelvic pain with primary amenorrhoea (obstructed outflow).
• Examination: accurate Tanner staging of breast and pubic hair separately (their dissociation is the clue); height, weight, BMI and height velocity plotted; signs of an underlying syndrome (Turner stigmata — short stature, webbed neck, widely-spaced nipples, cubitus valgus, lymphoedema; midline defects; visual fields and fundi); signs of androgen excess (clitoromegaly, acne, hirsutism → peripheral/adrenal cause); abdominal/pelvic mass; and, where indicated and consented, external genital inspection for a blind or absent vagina and an imperforate or bulging hymen.
• Hormonal first line: basal LH and FSH are the pivot. A basal LH already in the pubertal range (by an ultrasensitive assay) confirms central activation; a high FSH/LH signals primary gonadal failure; a low LH with low oestradiol is the hypogonadotropic pattern shared by constitutional delay and permanent GnRH deficiency. Add oestradiol, and — where androgen excess or peripheral precocity is in play — testosterone, DHEAS, 17-hydroxyprogesterone (non-classic CAH), and βhCG/AFP (a hormone-secreting germ-cell tumour).
• GnRH (or GnRH-agonist) stimulation test is the confirmatory step for central precocious puberty when the basal LH is equivocal but suspicion is high: a brisk LH rise (an LH-predominant response) confirms a primed pituitary; a flat, FSH-predominant response argues against central puberty. A low basal LH does not exclude central precocious puberty, which is exactly why the stimulation test still earns its place.
• Bone age (left wrist radiograph) is indispensable in both directions: advanced for chronological age in precocious puberty (and the substrate for predicting compromised final height), delayed in constitutional delay and in untreated hypogonadism. The gap between bone age and chronological age, read against the pace of progression, is what distinguishes a benign variant from a height-threatening one.
• Imaging: pelvic ultrasound to assess uterine size and endometrial thickness (an oestrogenised, enlarging uterus supports central activation) and ovarian morphology (a unilateral cyst/mass suggests a peripheral oestrogen source). Brain MRI is directed, not reflexive — it is mandatory in any child with neurological signs, in all boys with central precocious puberty (a far higher rate of intracranial pathology), and in younger girls; current guidance no longer demands routine MRI in an otherwise well girl in the 6–8-year window, where the yield of pathology is low.
• Karyotype is obtained in any hypergonadotropic picture (to diagnose Turner syndrome and its variants/mosaicism, and to detect Y-chromosome material that mandates gonadectomy), and where the uterus is absent (to separate 46,XX Müllerian agenesis from 46,XY androgen insensitivity).

The primary-amenorrhoea algorithm

Primary amenorrhoea is the one place in this topic where a disciplined algorithm out-performs pattern recognition, because the same presenting complaint hides causes at every level of the axis. The work-up is built on two examination findings read before any blood test — breast development (a biological assay of lifetime oestrogen exposure) and the uterus (present or absent on examination/ultrasound) — and then resolved with FSH and, where indicated, karyotype.

Two diagnoses are easy to confuse and worth holding apart. Müllerian agenesis and complete androgen insensitivity both present with breast development and an absent uterus, but the karyotype (46,XX vs 46,XY), the testosterone level (female- vs male-range), and the pubic-hair pattern (normal in MRKH, sparse/absent in CAIS because the receptor cannot respond to androgen) separate them — and the distinction carries a gonadectomy decision and a very different counselling conversation. And a normally oestrogenised girl with a uterus and cyclical pain is the algorithm's surgical exit: she has an obstructed outflow tract, and the next step is theatre, not endocrinology. This algorithm is the spine of the assessment; ambiguous external genitalia, gonadal dysgenesis with Y-chromosome material, and the broader spectrum of disorders of sex development extend it into the neonatal multidisciplinary setting and are managed there.

Management

The plan is built immediate → ongoing → long-term, and it diverges at the first fork: is this precocious, or delayed/absent?

Central precocious puberty — GnRH agonists are the treatment. The mechanism is counter-intuitive but decisive: continuous (rather than pulsatile) GnRH-receptor occupancy desensitises the pituitary gonadotrophs and switches off LH/FSH. A long-acting depot — leuprorelin or triptorelin, monthly or as 3-monthly preparations, with longer-acting implants where available — halts and often regresses pubertal progression and slows bone-age advance. The clinical aims are to preserve final adult height and to relieve the psychosocial burden of being years ahead of peers.

The indication is a judgement, not a reflex, and it turns on three things: a confirmed central process (rising or pubertal-range gonadotrophins, not a peripheral or benign-variant picture), rapid progression (advancing Tanner stages and bone age over months, an enlarging oestrogenised uterus on ultrasound), and a recoverable height deficit (a predicted adult height meaningfully compromised relative to target, read off the advanced bone age). The height benefit is real but selective: it is clearest in girls who start young — the gain in final height is substantial in those treated before about 6 years and progressively marginal as treatment starts later — so the older, slowly-progressing girl with a near-normal predicted height is one to observe with serial review, not reflexively treat. The honest framing of treatment in that older child is psychosocial rather than statural, and it should be offered as such.

Treatment is monitored by clinical regression of signs (breast development should stabilise or regress, growth velocity should slow toward a prepubertal rate), by demonstrating axis suppression (a blunted stimulated LH, or a low LH on a depot-timed sample), and by arrested bone-age advance on serial films. A child who continues to progress on a GnRH agonist needs the diagnosis re-examined — an undisclosed peripheral source, a missed McCune–Albright, or inadequate dosing — rather than simply more drug. Therapy is stopped at an age and bone age that let puberty resume in step with peers (commonly around 11 years of bone age), after which menarche typically follows within a couple of years. Suppression is reversible — the axis reactivates after withdrawal and long-term reproductive function and bone density are not impaired, the reassurance the foundational consensus established and longer follow-up has upheld.

Peripheral precocious puberty — treat the source, not the axis. GnRH agonists do nothing here because the axis is already suppressed by the autonomous steroid. Management is cause-directed: surgical removal of an oestrogen- or androgen-secreting ovarian or adrenal tumour; glucocorticoid replacement (and mineralocorticoid where needed) for congenital adrenal hyperplasia; aromatase inhibitors and/or oestrogen-receptor antagonists (e.g. letrozole/anastrozole, tamoxifen) for the autonomous ovarian oestrogen of McCune–Albright syndrome (the triad of peripheral precocity, café-au-lait macules with irregular "coast-of-Maine" borders, and polyostotic fibrous dysplasia, from a post-zygotic activating GNAS mutation); and removal of an exogenous steroid exposure. A peripheral picture can secondarily trip the central axis once bone age is sufficiently advanced — at which point a GnRH agonist is added on top of cause-directed therapy.

Delayed puberty and primary amenorrhoea — induce, replace, and treat the cause. The therapeutic decision follows the diagnostic axis:

• Constitutional delay of growth and puberty is a diagnosis of a self-limited variant, not a disease — it is the commonest cause of delayed puberty, runs in families, and resolves spontaneously. Most need explanation, reassurance and surveillance. A short course of low-dose sex steroid (low-dose oral oestradiol in girls; testosterone in boys) can be offered to relieve distress and "kick-start" the visible changes without compromising final height, but it is a comfort measure, not a cure.
• Permanent hypogonadism — pubertal induction is the core skill, and it must mimic physiology. Start low-dose oestradiol (transdermal where available, which avoids first-pass effects and is the preferred modality) and escalate gradually over 2–3 years to adult replacement, allowing the breast and uterus to develop and the growth potential to be used before the epiphyses fuse. Adding it too fast or too high produces tubular ("nipple-on-a-bump") breasts and premature fusion. Progestogen is added only once breakthrough bleeding occurs or after about 2 years of oestrogen, after which a combined regimen protects the endometrium and establishes withdrawal bleeds — but never lead with a progestogen or a standard combined oral contraceptive, which truncates breast and uterine development.
• Cause-directed management of primary amenorrhoea then depends on what the algorithm finds. Turner syndrome: growth hormone in childhood to optimise final height, oestrogen induction from around 11–12 years once spontaneous puberty has failed to begin (gonadotrophins rising), with lifelong cardiovascular, thyroid, hearing, hepatic and bone surveillance — and removal of any gonad containing Y-chromosome material because of dysgerminoma/gonadoblastoma risk. Primary ovarian insufficiency: in an adolescent this is gonadal failure before the menstrual life has properly begun — diagnosed on a high FSH (with the ESHRE threshold of >25 IU/L on repeated testing) and confirmed by karyotype, fragile-X premutation screening (FMR1) and autoantibody/adrenal screening, with a careful look for an iatrogenic cause (prior chemo- or radiotherapy for a childhood cancer is an increasingly common reason) and, in the SA setting, an awareness that infection and autoimmune disease contribute to the regional burden. Management is hormone therapy continued to the normal age of menopause — not as symptom relief but as primary prevention of bone, cardiovascular and cognitive morbidity, the point the current evidence-based guidance makes emphatically — together with bone-density surveillance, calcium/vitamin-D optimisation, and an early, honest fertility conversation (spontaneous conception is occasional but unpredictable; oocyte donation is the realistic route, with the access constraints of the public sector stated plainly). Müllerian agenesis (MRKH): a normally functioning ovary means no hormonal deficiency to replace; management is vaginal lengthening (graded dilators first-line, surgery reserved), psychological support, and fertility via surrogacy or uterine considerations. Complete androgen insensitivity (46,XY): oestrogen replacement after gonadectomy, careful gender-affirming and psychological care, vaginal dilatation as needed. Outflow obstruction (imperforate hymen, transverse vaginal septum) in an otherwise normally pubertal girl with cyclical pain and a haematocolpos is a surgical diagnosis — relieve the obstruction.

Psychological and growth considerations are part of the prescription, not an afterthought. Both ends of the spectrum carry real distress — the 7-year-old menstruating among unprepared peers, the 15-year-old without breasts among developed classmates — and the consultation must address body image, school, bullying and (where relevant) future fertility honestly. Growth is the time-limited variable: every plan is calibrated against bone age because the window to act on final height closes at epiphyseal fusion.

South African context. Paediatric and adolescent endocrinology is concentrated in academic centres, so the practical model is district recognition and timely referral of the abnormal-timing child to a tertiary paediatric-endocrine or specialist gynaecology service, rather than initiation of GnRH agonists or growth hormone at primary level. Depot GnRH agonists and growth hormone are costly and access is uneven across the public sector, which sharpens the case for confirming the diagnosis (and the likely height benefit) before committing — observation is a legitimate, resource-appropriate choice in the slowly-progressing older girl. Karyotyping and basic hormone panels are widely available through the NHLS; MRI access is the rate-limiting step in many settings and reinforces directed, rather than reflexive, imaging.

Guidelines compared

The major bodies agree on the architecture and differ mainly in emphasis and in how aggressively to investigate and treat.

The evidence & the controversy

The genuinely contested ground is the lower boundary of "normal". A documented secular trend toward earlier thelarche — entangled with rising childhood adiposity, and the subject of recurrent public anxiety about endocrine-disrupting chemicals in cosmetics, plastics and personal-care products — has reopened the question of whether the 8-year cut-off over-diagnoses. The defensible position holds the line on evaluation triggers while individualising treatment: a girl whose signs appear just before 8 but progress slowly, with a reassuring predicted height and no neurological feature, is a candidate for observation rather than reflexive suppression. The environmental-exposure argument is real enough to take seriously and weak enough not to act on as a diagnosis — there is association and biological plausibility, but the causal weight in any individual child is unestablished, and "it's the chemicals" is not a substitute for excluding a CNS lesion or an autonomous steroid source.

The treatment evidence has matured into a more selective practice. GnRH agonists unambiguously halt progression and rescue final height in the young, rapidly progressing child; the height case weakens steadily as onset moves past 6 and toward 8 years, where for many girls the predicted adult height is already near-normal and the honest indication becomes psychosocial rather than statural. The older feared harms have not held up — the foundational consensus explicitly declined to endorse weight-gain or bone-density concerns, and longer follow-up shows the axis reactivates and reproductive function is preserved. The live tension is therefore between over-treatment of the borderline child and under-recognition of the genuinely pathological one, which is why the modern guidance pushes basal LH and directed imaging rather than blanket testing and scanning.

At the delayed end, the persistent difficulty is that constitutional delay and congenital hypogonadotropic hypogonadism cannot be reliably separated at first presentation — both present prepubertal, and neither basal nor stimulated gonadotrophins discriminate cleanly in early adolescence. Anosmia points to Kallmann, very low inhibin B and absent prior minipuberty support permanent GnRH deficiency, and microphallus/cryptorchidism in a boy's history is a clue — but for many the answer only declares itself with time, so a period of low-dose induction with reassessment of spontaneous progression after a treatment-free interval is both diagnostic and therapeutic. Promising the family a definitive label too early is the error.

Landmark trials & key evidence

Exam traps & red flags

• Reading pubarche as puberty. Isolated early pubic hair is adrenarche (adrenal androgen), not central activation; it does not, on its own, warrant a GnRH agonist or a brain MRI — but it does warrant excluding non-classic CAH (17-OHP) and androgen-secreting tumours if it is rapid or virilising.
• Missing the discordant sequence of peripheral precocity. Vaginal bleeding or virilisation without progressive breast development, with low gonadotrophins, is peripheral until proven otherwise — look for an ovarian/adrenal mass, McCune–Albright (café-au-lait + fibrous dysplasia), CAH, or exogenous steroid. Giving a GnRH agonist here is useless because the axis is already suppressed.
• Skipping the brain MRI where it is mandatory. Every boy with central precocious puberty, and any child with neurological signs, needs imaging — the rate of intracranial pathology is high. "Idiopathic" is a diagnosis of exclusion, more safely applied to the well 6–8-year-old girl than to a boy.
• Treating every early girl. The older, slowly-progressing girl with a near-normal predicted height gains little statural benefit; the honest indication then is psychosocial, and observation is legitimate — especially where depot therapy is costly and access-limited.
• Leading pubertal induction with a progestogen or a combined pill. Oestrogen alone, started low and escalated slowly, allows breast and uterine development; an early progestogen or COC truncates it. Add the progestogen only after breakthrough bleeding or roughly 2 years of oestrogen.
• Forgetting the karyotype. In any hypergonadotropic picture, and whenever the uterus is absent: karyotype diagnoses Turner syndrome and finds Y-chromosome material (mandating gonadectomy for tumour risk), and separates MRKH from androgen insensitivity.
• Mistaking cyclical pain + primary amenorrhoea for a hormonal problem. A normally pubertal girl with monthly pain and a bulging/imperforate hymen or a palpable haematocolpos has outflow obstruction — a surgical, time-sensitive diagnosis, not a candidate for a hormone work-up.
• Promising a label in delayed puberty too soon. Constitutional delay and congenital hypogonadotropic hypogonadism look identical early; the distinction often needs time and a treatment-free reassessment, so commit to follow-up rather than a premature definitive diagnosis.

Evidence anchors

• Central precocious puberty — Endocrine Society clinical practice guideline (Latronico et al., J Clin Endocrinol Metab 2026)
• Use of GnRH analogs in children — international consortium update (Bangalore Krishna et al., Horm Res Paediatr 2019)
• Consensus statement on the use of GnRH analogs in children (Carel et al., Pediatrics 2009)
• Causes, diagnosis, and treatment of central precocious puberty (Latronico, Brito, Carel, Lancet Diabetes Endocrinol 2016)
• Clinical practice guidelines for the care of girls and women with Turner syndrome — 2023 Aarhus meeting (Eur J Endocrinol 2024)
• Evidence-based guideline: premature ovarian insufficiency (ESHRE/ASRM/IMS, Hum Reprod Open 2024)
• Current evaluation of amenorrhea: a committee opinion (ASRM Practice Committee, Fertil Steril 2024)
• Tanner staging of breast and pubic-hair development (Marshall & Tanner) — the standard clinical scale for pubertal stage; used throughout for assessment.
• South African paediatric and adolescent endocrinology referral model — abnormal-timing children investigated and treated at academic/tertiary centres; karyotyping and hormone panels via the NHLS; access to depot GnRH agonists and growth hormone uneven in the public sector.