Insulin can inhibit dehydroepiandrosterone (DHEA) biosynthesis in humans, as suggested by several studies performed in induced or spontaneous hyperinsulinemia. The increased insulin resistance documented throughout aging, with its accompanying hyperinsulinemia, may contribute to the age-related decline in DHEA synthesis. The aim of this study was to assess if the aging-related differences in DHEA sulfate (DHEA-S) serum levels can be associated with differences in fasting insulin levels, as well as body composition. Two hundred fifty-two healthy subjects of both sexes aged 19 to 90 years with a body mass index (BMI) less than 30 (mean +/- SD, 23.5 +/- 2.4) were studied DHEA-S and insulin serum levels were determined by a radioimmunologic procedure; body composition was assessed by anthropometry (fat mass percentage [FM%] estimated from four skinfold thicknesses by Durnin and Womersley and Siri equations [FM%-SKF]) and by bioimpedance analysis (BIA) (FM% estimated by equations developed by Segal et al and Deurenberg et al for subjects < and > 62 years, respectively [FM%-BIA]). DHEA-S levels were significantly and inversely related to age in both sexes. No significant aging-related differences were found in fasting insulin levels, although a trend toward an increase was apparent in the women on simple regression analysis. No significant associations were found between DHEA-S and insulin levels. As for body composition, a positive relationship to age was apparent for FM%-SKF, FM%-BIA, and waist to hip ratio (WHR), whereas BMI and phase angle ([PA] a bioelectric parameter considered an index of the ratio between intracellular and extracellular water) were inversely related to age. Fasting insulin levels were positively related to FM% as estimated by both BIA and anthropometry, independently of age in both sexes; in addition, a positive correlation with WHR and with the subscapular to triceps skinfold thickness ratio (SS/TS) was found in men and women, respectively. No significant correlation was apparent between DHEA-S and body composition indices in men, whereas in women a slight negative correlation between DHEA-S and WHR was documented, and was still significant after adjustment for age and fasting insulin. Stepwise multiple regression analysis confirmed that DHEA-S levels are not related to fasting insulin, but are independently related to age and, in women only, to WHR. Our study suggests that the DHEA-S decline due to aging is independent of fasting insulin, at least in healthy, non-obese people. In addition, it is not related to the aging-dependent changes in body composition in terms of FM% and fat-free mass (FFM) percentage (FFM%). Only in women could changes in fat distribution be slightly associated with DHEA-S decline, although such a relation cannot be accounted for by changes in insulin levels.