A genetic study of sex hormone--binding globulin measured before and after a 20-week endurance exercise training program: the HERITAGE Family Study.

Familial aggregation and a major gene effect were assessed for baseline serum sex hormone-binding globulin (SHBG) levels and the response (post-training minus baseline) to a 20-week endurance training program in a selected sample of 428 non-obese nonhypertensive individuals from 99 white families who were sedentary at baseline in the HERITAGE Family Study. Baseline SHBG levels were not normally distributed, and were therefore logarithmically transformed prior to genetic analyses. In a sample without postmenopausal mothers, maximal (genetic and familial environmental) heritabilities were 50% averaged across sexes, 73% in men, 50% in women, and 31% in men versus women for the age-body mass index (BMI)-adjusted baseline. The estimate reached 64% when the baseline was further adjusted for the effects of estradiol, fasting insulin, and testosterone levels. For the response to training, no sex difference was found and the heritability reached about 25% to 32%. Segregation analysis was separately performed in the whole sample and in the sample without postmenopausal mothers. In addition to a multifactorial effect for both the baseline and the response to training, a major effect for the baseline appeared to be familial environmental in origin, whereas a major effect for the response to training was Mendelian in nature. The major gene effect for the response to training in the whole sample was undetectable in the sample without postmenopausal mothers, and it is therefore possible that the postmenopausal mothers, characterized by decreased sex hormones with or without estrogen replacement therapy for menopause, produced some confounding effects. In addition, the reduced sample size might also be a plausible candidate explanation. The novel finding in this study is that baseline SHBG levels and the response to training were influenced by a multifactorial effect with sex difference for the baseline. The response to training appeared to be additionally influenced by a single recessive locus that is independent of baseline SHBG levels.