Executive functions in recently postmenopausal women: absence of strong association with serum gonadal steroids.

It is controversial to what extent endogenous gonadal hormone exposures influence executive functions in midlife women. Participants in the population-based Melbourne Women's Midlife Health Project were administered a battery of neuropsychological tests on two occasions 2 years apart. Tests of executive functions were the Trail Making Test (Part B), Tower of London (administered at baseline only), Symbol Digit Modalities Test, Digit Span Backward, and Letter-Number Sequencing. Estrone, free estradiol, and free testosterone levels were measured at the time of the first testing, and analyses were restricted to 147 women aged 56-64 years who had recently undergone natural menopause (mean age of menopause 53 years) and were not using hormone therapy. In multiple linear regression analyses, 2 of 20 baseline associations were significant at an alpha level of 0.05. Estrone concentrations were positively associated with Tower of London performance (p=0.02), and the ratio of free testosterone to free estradiol was positively associated with scores on the Symbol Digit Modalities Test score (p=0.04). No hormone measure was significantly predictive of 2-year change in executive functions performance. Significance levels in these exploratory analyses were unadjusted for multiple comparisons, and observed associations could be due to unique psychometric properties of these particular tasks or due to chance. Sex hormone binding globulin concentrations were unrelated to executive function scores. In recently postmenopausal women not receiving hormone therapy, serum concentrations of estrone, estradiol and testosterone, and the testosterone/estradiol ratio are not strongly associated with executive functions.