Onset of normal puberty in girls is somewhat variable, with disagreement in the literature concerning at what age to diagnose precocious puberty and at what age to suspect delayed puberty. The most generally agreed-on range of onset for female puberty is between 9 and 16 years. Signs of puberty include breast development, growth of pubic and axillary hair, and estrogen effect as seen in vaginal smears using the Papanicolaou stain. Menstruation also begins; if it does not, the possibility of primary amenorrhea arises. Primary amenorrhea may or may not be accompanied by evidence that suggests onset of puberty, depending on the etiology of the amenorrhea. Some of the causes of primary amenorrhea are listed in the box.

Some Etiologies of Female Delayed Puberty and Primary Amenorrhea
Mullerian dysgenesis (Mayer-Rokitansky syndrome): congenital absence of portions of the female genital tract, with absent or hypoplastic vagina. Ovarian function is usually normal. Female genotype and phenotype.
Male pseudohermaphroditism: genetic male (XY karotype) with female appearance due to deficiency of testosterone effect.
1. Testicular feminization syndrome: lack of testosterone effect due to defect in tissue testosterone receptors.
2. Congenital adrenal hyperplasia: defect in testosterone production pathway.
3. Male gonadal dysgenesis syndromes: defect in testis function.
a. Swyer syndrome (male pure gonadal dysgenesis): female organs present except for ovaries. Bilateral undifferentiated streak gonads. Presumably testes never functioned and did not prevent mьllerian duct development.
b. Vanishing testes syndrome (XY gonadal agenesis): phenotype varies from male pseudohermaphrodite to ambiguous genitalia. No testes present. Presumably testes functioned in very early embryologic life, sufficient to prevent mьllerian duct development, and then disappeared.
c. Congenital anorchia: male phenotype, but no testes. Presumably, testes functioned until male differentiation took place, then disappeared.
Female sex chromosome abnormalities (Turner’s syndrome—female gonadal dysgenesis— and Turner variants): XO karyotype in 75%-80% of patients, mosaic in others. Female phenotype. Short stature in most, web neck in 40%. Bilateral streak gonads.
Polycystic (Stein-Leventhal) or nonpolycystic ovaries, not responsive to gonadotropins.
Deficient gonadotropins due to hypothalamic or pituitary dysfunction.
Hyperprolactinemia: pituitary overproduction of prolactin alone.
Effects of severe chronic systemic disease: chronic renal disease, severe chronic GI disease, anorexia nervosa, etc.
Constitutional (idiopathic) delayed puberty: puberty eventually takes place, so this is a retrospective diagnosis.
Other: Cushing’s syndrome, hypothyroidism, and isolated GH deficiency can result in delayed puberty.

Physical examination is very important to detect inguinal hernia or masses that might suggest the testicular feminization type of male pseudohermaphroditism, to document the appearance of the genitalia, to see the pattern of secondary sex characteristics, and to note what evidence of puberty exists and to what extent. Pelvic examination is needed to ascertain if there are anatomical defects preventing menstruation, such as a nonpatent vagina, and to detect ovarian masses.

Laboratory tests

Basic laboratory tests begin with chromosome analysis, since a substantial minority of cases have a genetic component. Male genetic sex indicates male pseudohermaphroditism and leads to tests that differentiate the various etiologies. Turner’s syndrome and other sex chromosome disorders are also excluded or confirmed. If there is a normal female karyotype and no chromosome abnormalities are found, there is some divergence of opinion on how to evaluate the other important organs of puberty—the ovaries, pituitary, and hypothalamus. Some prefer to perform serum hormone assays as a group, including pituitary gonadotropins (FSH and LH), estrogen (estradiol), testosterone, prolactin, and thyroxine. Others perform these assays in a step-by-step fashion or algorithm, depending on the result of each test, which could be less expensive but could take much more time. Others begin with tests of organ function. In the algorithm approach, the first step is usually to determine if estrogen is present in adequate amount. Vaginal smears, serum estradiol level, endometrial stimulation with progesterone (for withdrawal bleeding), and possibly endometrial biopsy (for active proliferative or secretory endometrium, although biopsy is considered more often in secondary than in primary amenorrhea)—all are methods to detect ovarian estrogen production. If estrogen is present in adequate amount, this could mean intact hypothalamic-pituitary-ovarian feedback or could raise the question of excess androgen (congenital adrenal hyperplasia, Cushing’s syndrome, PCO disease, androgen-producing tumor) or excess estrogens (obesity, estrogen-producing tumor, iatrogenic, self-medication). If estrogen effect is absent or very low, some gynecologists then test the uterus with estrogen, followed by progesterone, to see if the uterus is capable of function. If no withdrawal bleeding occurs, this suggests testicular feminization, congenital absence or abnormality of the uterus, or the syndrome of intrauterine adhesions (Asherman’s syndrome). If uterine withdrawal bleeding takes place, the uterus can function, and when this finding is coupled with evidence of low estrogen levels, the tentative diagnosis is ovarian failure.

The next step is to differentiate primary ovarian failure from secondary failure caused by pituitary or hypothalamic disease. Hypothalamic dysfunction may be due to space-occupying lesions (tumor or granulomatous disease), infection, or (through uncertain mechanism) effects of severe systemic illness, severe malnutrition, and severe psychogenic disorder. X-ray films of the pituitary are often ordered to detect enlargement of the sella turcica from pituitary tumor or suprasellar calcifications due to craniopharyngioma. Polytomography is more sensitive for sellar abnormality than ordinary plain films. CT also has its advocates. Serum prolactin, thyroxine, FSH, and LH assays are done. It is necessary to wait 4-6 weeks after a progesterone or estrogen-progesterone test before the hormone assays are obtained to allow patient hormone secretion patterns to resume their pretest status. An elevated serum prolactin value raises the question of pituitary tumor (especially if the sella is enlarged) or idiopathic isolated hyperprolactinemia. However, patients with the empty sella syndrome and some patients with hypothyroidism may have an enlarged sella and elevated serum prolactin levels, and the other causes of elevated prolactin levels (see the box) must be considered. A decreased serum FSH or LH level confirms pituitary insufficiency or hypothalamic dysfunction. A pituitary stimulation test can be performed. If the pituitary can produce adequate amounts of LH, this suggests hypothalamic disease (either destructive lesion, effect of severe systemic illness, or malnutrition). However, failure of the pituitary to respond does not necessarily indicate primary pituitary disease, since severe long-term hypothalamic deficiency may result in temporary pituitary nonresponsiveness to a single episode of test stimulation. Hormone therapy within the preceding 4-6 weeks can also adversely affect test results.

The box lists some conditions associated with primary amenorrhea or delayed puberty and the differential diagnosis associated with various patterns of pituitary gonadotropin values. However, several cautionary statements must be made about these patterns. Clearly elevated FSH or LH values are much more significant than normal or mildly to moderately decreased levels. Current gonadotropin immunoassays are technically more reproducible and dependable at the upper end of usual values than at the lower end. Thus, two determinations on the same specimen could produce both a mildly decreased value and a value within the lower half of the reference range. In addition, blood levels of gonadotropins, especially LH, frequently vary throughout the day. The values must be compared with age- and sexmatched reference ranges. In girls, the levels also are influenced by the menstrual cycle, if menarche has begun. Second, the conditions listed—even the genetic ones, although to a lesser extent—are not homogeneous in regard to severity, clinical manifestations, or laboratory findings. Instead, each represents a spectrum of patients. The more classic and severe the clinical manifestations, the more likely that the patient will have “expected” laboratory findings, but even this rule is not invariable. Therefore, some patients having a condition typically associated with an abnormal laboratory test result may not show the expected abnormality. Laboratory error is another consideration. Also, the spectrum of patients represented by each condition makes it difficult to evaluate reports of laboratory findings due to differences in patient population, severity of illness, differences in applying diagnostic criteria to the patients, variance in specimen collection protocols, and technical differences in the assays used in different laboratories. In many cases, adequate data concerning frequency that some laboratory tests are abnormal are not available.

Gonadotropin Levels in Certain Conditions Associated With Primary Amenorrhea or Delayed Puberty
FSH and LH decreased*
Hypothalamic dysfunction
Constitutional delayed puberty
Some cases of primary hypothyroidism
Some cases of Cushing’s syndrome
Some cases of severe chronic illness
FSH and LH increased†
Some cases of congenital adrenal hyperplasia
Female gonadal dysgenesis
Male gonadal dysgenesis
Ovarian failure due to nonovarian agents
LH increased, FSH not increased‡
Testicular feminization
Some cases of PCO disease

Elevated FSH and LH levels or an elevated LH level alone suggests primary ovarian failure, whether from congenital absence of the ovaries or congenital inability to respond to gonadotropins, acquired abnormality such as damage to the ovaries after birth, or PCO disease.

An elevated estrogen or androgen level raises the question of hormone-secreting tumor, for which the ovary is the most common (but not the only) location. Nontumor androgen production may occur in PCO disease and Cushing’s syndrome (especially adrenal carcinoma).

As noted previously, there is no universally accepted single standard method to investigate female reproductive disorders. Tests or test sequences vary among medical centers and also according to findings in the individual patients.

When specimens for pituitary hormone assays are collected, the potential problems noted earlier in the chapter should be considered.