Hyperandrogenism in female athletes with functional hypothalamic amenorrhea: a distinct phenotype

OBJECTIVE: To compare the reproductive, metabolic, and skeletal profiles of young athletic women with functional hypothalamic amenorrhea (FHA) as well as clinical or biochemical hyperandrogenism (FHA-EX+HA) with body mass index matched women with FHA due to exercise (FHA-EX) or anorexia nervosa (FHA-AN) alone. DESIGN: Retrospective cohort study. SETTING: Tertiary care teaching hospital. POPULATION: Adolescents and young women, 15–30 years of age, diagnosed with FHA along with concurrent signs of hyperandrogenism (n=22) and body mass index matched control groups consisting of 22 women in each group of FHA-EX and FHA-AN. MAIN OUTCOMES: 1) Reproductive hormone profile: luteinizing hormone (LH), follicle stimulating hormone (FSH), total testosterone, pelvic ultrasound features. 2) Metabolic function and skeletal health markers: fasting glucose, cholesterol, number of stress fractures and bone mineral density as assessed by spine dual-energy X-ray absorptiometry z scores. RESULTS: FHA-EX+HA group was older at diagnosis compared to the other groups with a median (interquartile range [IQR]) age of 22 (18.75–25.25) years versus (vs) 17.5 (15.75–19) for FHA-EX; (P<0.01) and 18 (16–22.25) years for FHA-AN (P=0.01). There were no differences among the groups based on number of hours of exercise per week, type of physical activity or duration of amenorrhea. Median (IQR) LH/FSH ratio was higher in FHA-EX+HA than both other groups, 1.44 (1.03–1.77) vs 0.50 (0.20–0.94) for FHA-EX and 0.67 (0.51–0.87) for FHA-AN (P<0.01 for both). Total testosterone concentrations were not different among the groups. Median (IQR) fasting serum glucose concentration was higher in FHA-EX+HA vs FHA-EX, 88.5 mg/dL (82.8–90 mg/dL) vs 83.5 mg/dL (78.8–86.3 mg/dL) (P=0.01) but not different from FHA-AN (P=0.31). Percentage of women with stress fractures was lower in FHA-EX+HA (4.5%) as compared to both FHA-EX (27.3%) and FHA-AN (50%); P=0.04 and 0.01 respectively. The LH/FSH ratio was weakly positively associated with serum glucose (adjusted r(2)=0.102; P=0.01) as well as with dual-energy X-ray absorptiometry spine score (adjusted r(2)=0.191; P=0.04) in the entire cohort. CONCLUSION: In a small cohort of female athletes with hyperandrogenism, a distinct reproductive hormone profile consisting of higher LH to FHS ratio may be associated with adverse metabolic health markers but improved skeletal health.