Associations between androgens and sexual function in premenopausal women: a cross-sectional study.

Background: Although clinicians often measure the serum concentration of androgens in premenopausal women presenting with sexual dysfunction, with some women given testosterone or dehydroepiandrosterone as treatment if their concentrations are low, whether androgens are determinants of sexual function in women of reproductive age is uncertain. We aimed to clarify the associations between androgens and sexual function in a community-based sample of non-health-care-seeking women.

Methods: This is a substudy of the Grollo-Ruzzene cross-sectional study, which recruited women aged 18-39 years from eastern states in Australia (QLD, NSW, VIC). After providing consent, women completed an online survey that included the Profile of Female Secual Function (PFSF) questionnaire, and those who were who were not pregnant, breastfeeding, or using systemic steroids were asked to provide a blood sample. At sampling, women were asked the dates of their last menstrual bleed. Serum androgens was measured by liquid chromatography and tandem mass spectrometry and sex hormone binding globulin (SHBG) by immunoassay. Associations between androgens and domains of sexual function, assessed by the PFSF, were examined in participants with regular menstrual cycles. After univariable linear regression (model 1), age, BMI, stage of menstrual cycle, and smoking status were added to the model (model 2), and then parity, partner status, and psychotropic medication use (model 3).

Findings: Of 6986 women who completed the online survey (surveys completed between Nov 11, 2016, and July 21, 2017), 3698 were eligible and 761 (20·6%) provided blood samples by Sept 30, 2017. Of those who provided a blood sample, 588 (77·3%) had regular menstrual cycles and were included in the analysis. Adjusting for age, BMI, cycle stage, smoking, parity, partner status, and psychoactive medication, sexual desire was positively associated with serum dehydroepiandrosterone (β-coefficient 3·39, 95% CI 0·65 to 6·03) and androstenedione (4·81, 0·16 to 9·12), and negatively with SHBG (-5.74, -9.54 to -1·90), each model explaining less than 4% of the variation in desire. Testosterone (6·00, 1·29 to 10·94) and androstenedione (6·05, 0·70 to 11·51) were significantly associated with orgasm, with the final models explaining less than 1% of the variation in orgasm. Significant associations were found between androstenedione (7·32, 0·93 to 13·08) and dehydroepiandrosterone (4·44, 0·86 to 7·95) and pleasure, and between testosterone and sexual self-image 5·87 (1·27 to 10·61), with inclusion of parity, partners status, and psychotropic drug use increasing the proportion of variation explained by each model to approximately 10%. There were no statistically significant associations between 11-oxygenated steroids and any PFSF domain, or between arousal or responsiveness and any hormone. No associations were seen between 11-oxygenated steroids and any sexual domain, or between arousal or responsiveness and any hormone.

Interpretation: Associations between androgens and sexual function in premenopausal women are small, and their measurement offers no diagnostic use in this context. Further research to determine whether 11-ketoandrostenedione or 11-ketotestosterone are of clinical significance is warranted.

Funding: The Grollo-Ruzzene Foundation.