It has been hypothesized that the androgens testosterone and dehydroepiandrosterone (DHEA) may have opposing actions on insulin sensitivity. To test this hypothesis, we selected patients with polycystic ovary syndrome (PCO) and hypertestosteronemia and a group of individuals with adrenal hyperplasia (AH) and elevated DHEA and studied their 1) insulin and glucose responses to a 75-g oral glucose tolerance test, 2) insulin resistance by hypoglycemic responses to a standard dose of intravenous (IV) insulin, and 3) insulin binding and pyruvate dehydrogenase (PDH) responsiveness to insulin in phytohemagglutinin (PHA)-activated T lymphocytes. PCO patients exhibited elevated basal and glucose-challenged insulin levels and had blunted hypoglycemic responses to IV insulin. Conversely, AH patients had hypoglycemic responses to IV insulin significantly > and basal and glucose-challenged insulin levels lower than the PCO patients and weight-matched control subjects. In vitro, T-lymphocyte insulin binding of the PCO patients was 40–60% below control values; in AH patients, insulin binding and PDH insulin sensitivity were above those of the control subjects. Testosterone levels in all study subjects were negatively correlated to T-lymphocyte insulin binding and positively correlated to basal insulin, insulin area under the curve (AUC), and insulin-glucose indices. DHEA levels were positively correlated to insulin binding and inversely related to basal insulin, insulin AUC, and insulin-glucose indices. In all instances, the parameters of insulin sensitivity were more strongly correlated to individuals' ratios of DHEA to testosterone than to either of these androgens alone. From the above in vivo and in vitro studies, we conclude that, in hyperandrogenic women, the ratio of DHEA to testosterone may be an important modulator of insulin sensitivity.