Context: The incidence of menstrual disorders declines during adolescence. The mechanism responsible is unknown.

Objective: To test the hypothesis, formulated a priori, that the dependence of LH (LH) pulsatility and ovarian function on energy availability declines with gynecological age (years since menarche).

Design: A controlled experiment repeated in two menstrual cycles; performed 2001-2004.

Setting: University laboratory and General Clinical Research Center.

Participants: Healthy, habitually sedentary, young women of normal body composition with 5-8 yr (adolescents, n = 9) and 14-18 yr (adults, n = 10) of gynecological age recruited by advertisement from 9,000 women aged 18-34 in a college community. Samples were similar in age of menarche, length of menstrual cycle and luteal phase, body size and composition, aerobic capacity, and dietary intake. None were withdrawn due to adverse effects.

Interventions: Energy availabilities of 45 and 10 kilocalories per kilogram of fat-free mass per day for five days in the early follicular phases of separate menstrual cycles in random order.

Main Outcome Measures: LH pulsatility, estradiol, and luteal phase length.

Results: Low energy availability reduced LH pulse frequency in adolescents (P < 0.01) but not adults (P = 0.39), did not increase LH pulse amplitude in either group (both P = 0.13), and suppressed 24 h mean LH in adolescents (P = 0.01) but not adults (P = 0.72). Estradiol was unaffected (both P = 0.48), but the subsequent luteal phase was shorter in adolescents (P < 0.01).

Conclusion: In women of normal body composition, the response of LH pulsatility and ovarian function to five days of low energy availability disappears by 14 yr of gynecological age.