Introduction: The ability to study chronic stress in humans is complicated due to measurement error of questionnaires and the inability of short-term measures of stress hormones to reflect the chronic state. Therefore considerable controversy remains about whether chronic stress influences cardiovascular disease or not. The cortisol/testosterone (C/T) ratio was suggested to be a better predictor of heart disease in men than cortisol alone, as gonadotropin and cortisol are derived from the same biochemical precursor. This ratio has never been studied with metabolic syndrome (MetS) in a U.S. epidemiologic study, especially in women. Study question: Are C/T and C/sex hormone binding globulin (SHBG) ratio associated with MetS in women? Methods: In the Coronary Artery Risk Development in Young Adults (CARDIA) study, 367 women (age range 32 to 51, mean age 40 years old) who had both cortisol from year (Y) 15 and sex hormones from serum specimen Y16 measured are included. Metabolic syndrome (MetS) from Y15, 20, 25 were assessed. Due to the instability and diurnal characteristics of cortisol, area under the curve (AUC) of six samples and slope of 1st (or 3rd) and 6th (or 5th when 6th is not available) sample of salivary cortisol collected over one day were calculated. Ratios of AUC and of slope of cortisol to total testosterone (TT), free testosterone (FT), and sex hormone binding globulin (SHBG) were computed: AUC/TT, AUC/FT, AUC/SHBG, Slope/TT, Slope/FT, and Slope/SHBG. The associations of these variables categorized into tertiles with MetS were assessed cross-sectionally by logistic regression analysis. Model I controlled for age and race. Model II controlled for model I variables plus menopause, oral contraceptive usage, diabetes mellitus, alcohol consumption, and cigarette smoking. Results: MetS was present in 53, 69, and 74 participants at Y15, Y20, and Y25, respectively. The highest tertile of AUC/SHBG ratio was associated with Y15, Y20, and Y25 MetS prevalence in model I (OR 2.17, 3.77, 2.65, 95% CI 1.02-4.61, 1.64-8.44, 1.75-9.20, respectively). This association was slightly stronger in model II (OR 3.72, 4.76, 3.26, 95% CI 1.49-9.30, 2.00-11.34, 1.55-6.85, respectively). The highest tertile of slope/FT ratio was associated with Y20 and Y25 MetS prevalence in model I (OR 2.10, 1.68, 95% CI 1.03-4.26, 0.86-3.31 respectively). This association was slightly stronger in model II (OR 2.32, 2.13, 95% CI 1.09-4.95, 1.03-4.41 respectively). Conclusions: Findings suggest that some indicators of chronic stress are cross-sectionally associated with MetS in women.