Hypothalamic amenorrhea (HA) is a clinical disorder of unknown etiology. The diagnosis is made by exclusion of known abnormalities of pituitary and ovarian function. To determine if abnormalities of GnRH secretion could account for the anovulation and amenorrhea, we measured plasma gonadotropins every 20 min for 10- to 24-h periods in 19 women with HA. Ovarian steroids and gonadotropin responses to an iv bolus dose of GnRH (25 ng/kg) were also measured. The results were compared to those obtained during the early follicular (EF) and late luteal (LL) phases of ovulatory cycles in normal women. Plasma estradiol was lower (mean +/- SE, 52 +/- 5 pg/ml) than either cycle stage in normal women. Mean plasma LH was lower than EF values and FSH was higher than LL values. The amplitude of LH pulses in HA was similar to that in normal women. LH pulse frequency was the same as that present during the LL, but lower than that during the EF (HA, 4.7 pulses/12 h; EF, 7.7 pulses/12 h; P less than 0.05). In addition to the similar frequency, the patterns of LH secretion in HA resembled that of LL in that the amplitude of LH pulses was highly variable and pulses occurred at irregular intervals. Consistent changes in diurnal gonadotropin secretion were not found, and LH secretion was greater at night in 9 studies and during the day in 5 studies. Repeat studies in three patients (5-13 months later) revealed that LH pulse frequency was variable, being unchanged in 1, increased in 1, and decreased in the third patient. Thus, LH pulse frequency and, by inference, GnRH pulse frequency are similar in HA to those in the normal luteal phase despite a different steroid milieu. GnRH pulse frequency increases from the luteal to the follicular phases of normal cycles and may be important in the initiation of ovarian follicular maturation. These data suggest that the absence of cyclical gonadotropin secretion and anovulation in HA result from a decreased frequency and irregular amplitude of GnRH secretion and consequent absence of ovarian follicular maturation.

Pulsatile secretion of LH and FSH was examined to determine if the frequency of LH pulses, and by inference pulsatile GnRH secretion, varied during the human menstrual cycle. Blood samples were obtained at 10- or 20-min intervals for 12 or 24 h at 7-day intervals during the same ovulatory cycle in eight normal women. Ovarian steroids showed the expected cyclical changes and mean plasma FSH concentrations showed an inverse relationship to estradiol, being low when estradiol was greater than 150 pg/ml. Sampling every 10 min revealed a constant LH pulse amplitude but LH pulse frequency increased (from 11.8 to 14.3 pulses/12 h) during the follicular phase. LH pulse frequency was not further increased in two women sampled during the LH surge, but pulse amplitude was markedly higher. During the luteal phase LH pulse frequency was reduced to eight pulses/12 h but frequency was more variable between subjects than in the follicular phase. LH pulse amplitude showed striking variation (0.8-29.4 mIU/ml) during the luteal phase of the cycle and large LH secretory episodes which lasted 1-3 h were irregularly interspersed among periods of low amplitude LH secretion. These data show that the frequency of LH pulses, and by inference GnRH secretion, varies during the menstrual cycle but the degree of change is less than reported in previous studies. This observation may explain the reported efficacy of fixed frequency GnRH regimes in inducing ovulation and cyclical ovarian function.

The gonadotropin secretory patterns of 22 sexually immature children were analyzed in detail to determine whether pulsatile secretion occurs before the onset of puberty. Eight endocrinologically normal children, 13 children with isolated GH deficiency, and 1 girl with 45X gonadal dysgenesis were divided into 2 groups according to bone age. Group A children had bone ages less than 10 yr, and group B had bone ages between 10-11.5 yr. Blood samples were drawn every 20 min for periods of 3-11 h during both the day and night; in addition, 12-h urine collections were made for gonadotropin determinations. Mean nocturnal concentrations of LH and FSH were significantly greater than daytime values in 8 of 15 and 5 of 15 children in group A and in 6 of 7 and 1 of 7 in group B, respectively. Nocturnal urinary excretion of LH and FSH was significantly greater in group A children. Eight children in group A, including 4 whose bone ages were less than 5 yr, and 4 group B children had discernible LH pulses. LH pulses were detected during the day and night in both groups. LH pulse amplitude was greater during the night in both groups, but was greatest in group B (A, 1.9 +/- 0.2 mIU/ml; B, 3.0 +/- 0.3 mIU/ml). In children who demonstrated pulsatile secretion, LH pulse frequency appeared to be similar during the day and night and was slightly faster in the older children (A, every 3 h; B, every 2 h). These studies demonstrated that LH is secreted in a pulsatile manner well before the onset of puberty. Furthermore, the gonadotropin secretory pattern characteristic of early puberty results from the amplification of an already existing circadian pattern of gonadotropin secretion.

Pulsatile gonadotropin secretion was examined in seven women with hyperprolactinemia and amenorrhea by obtaining blood samples every 20 min for 24 h. When plasma PRL had returned to normal and menses had resumed during bromocriptine treatment, five women were restudied in an identical manner during the early to midfollicular stage of their cycles. Gonadotropin responses to a small dose of synthetic GnRH (25 ng/kg, iv) were measured after the initial 24-h study in each patient. In addition, low dose pulses of GnRH (25 ng/kg) were administered iv every 2 h for 88 h to three hyperprolactinemic women, and LH and FSH responses were determined. Before treatment with bromocriptine, mean +/- SE plasma gonadotropin concentrations (LH, 5.8 +/- 0.2 mIU/ml; FSH, 4.4 +/- 0.1 mIU/ml) were comparable to values during the follicular phase of normal menstrual cycles. LH pulse frequency during the pretreatment study in the hyperprolactinemic women (mean +/- SE, 7.6 +/- 1.2 pulses/24 h) was significantly less than that found during the early follicular stage of normal cycles (days 3-5; mean, 15.4 +/- 1.1 pulses/24 h). Mean +/- SE LH pulse amplitude before bromocriptine was 5.2 +/- 0.6 mIU/ml. The pattern of pulsatile LH secretion was highly variable before treatment and was characterized by prolonged periods (6-11 h) of low plasma LH concentrations. LH responses to GnRH were normal or increased (mean maximum increment in LH, 38.5 +/- 15.9; range, 4.3-125.2 mIU/ml), and no evidence of intermittent pituitary refractoriness was found during prolonged (88-h) administration of GnRH pulses. Treatment with bromocriptine was associated with the resumption of menses, and no significant change in mean gonadotropin concentrations. LH pulse frequency was increased (mean +/- SE = 10.2 +/- 1.0 pulses/24 h) and LH pulse amplitude was decreased (mean, 3.9 +/- 0.2 mIU/ml) in four of five patients receiving bromocriptine. Moreover, the pattern of pulsatile LH secretion was more uniform during treatment. We conclude that pituitary responsiveness to GnRH is not impaired in women with hyperprolactinemia and amenorrhea, and that periods of low LH secretion in these women are due to intermittent reductions in GnRH secretion. These observations suggest that the abnormal patterns of pulsatile gonadotropin secretion, and by inference GnRH secretion, are important factors in the etiology of amenorrhea associated with hyperprolactinemia.

Linear growth velocity in children with insulin-dependent diabetes mellitus is heavily related to metabolic control. Children who are prepubertal or in the early stages of puberty are the most vulnerable to growth suppression. Once puberty is well established, growth suppression does not occur until marked hyperglycemia (GHb greater than 16%) exists.