Glucocorticoids induce insulin resistance in humans, whereas thiazolidinediones enhance insulin sensitivity. Although the effects of glucocorticoids and thiazolidinediones have been assessed in isolation, interaction between these drugs, which both act as ligands for nuclear receptors, has been less well studied. Therefore, we examined the metabolic effects of dexamethasone and troglitazone, alone and in combination, for the first time in humans. A total of 10 healthy individuals with normal glucose tolerance (age 40 ± 11 years, BMI 31 ± 6.1 kg/m2 were sequentially studied at baseline, after 4 days of dexamethasone (4 mg/day), after 4-6 weeks on troglitazone alone (400 mg/day), and again after 4 days of dexamethasone added to troglitazone. Key metabolic variables included glucose tolerance assessed by blood glucose and insulin responses to an oral glucose tolerance test (OGTT), insulin sensitivity evaluated via hyperinsulinemic-euglycemic clamp, free fatty acids (FFAs) and FFA suppressibility by insulin during the clamp study, and fasting serum leptin. Dexamethasone drastically impaired glucose tolerance, with fasting and 2-h OGTT insulin values increasing by 2.3-fold (P < 0.001) and 4.4-fold (P < 0.001) over baseline values, respectively. The glucocorticoid also induced a profound state of insulin resistance, with a 34% reduction in maximal glucose disposal rates (GDRs; P < 0.001). Troglitazone alone increased GDRs by 20% over baseline (P = 0.007) and completely prevented the deleterious effects of dexamethasone on glucose tolerance and insulin sensitivity, as illustrated by a return of OGTT glucose and insulin values and maximal GDR to near-baseline levels. Insulin-mediated FFA suppressibility (FFA decline at 30 min during clamp/FFA at time 0) was also markedly reduced by dexamethasone (P = 0.002). Troglitazone had no effect per se, but it was able to normalize FFA suppressibility in subjects coadministered dexamethasone. Futhermore, the magnitudes of response of FFA suppressibility and GDR to dexamethasone were proportionate. The same was true for the reversal of dexamethasone-induced insulin resistance by troglitazone, but not in response to troglitazone alone. Leptin levels were increased 2.2-fold above baseline by dexamethasone. Again, troglitazone had no effect per se but blocked the dexamethasone-induced increase in leptin. Subjects experienced a 1.7-kg weight gain while taking troglitazone but no other untoward effects. We conclude that in healthy humans, thiazolidinediones antagonize the action of dexamethasone with respect to multiple metabolic effects. Specifically, troglitazone reverses both glucocorticoid-induced insulin resistance and impairment of glucose tolerance, prevents dexamethasone from impairing the antilipolytic action of insulin, and blocks the increase in leptin levels induced by dexamethasone. Even though changes in FFA suppressibility were correlated with dexamethasone-induced insulin resistance and its reversal by troglitazone, a cause-and-effect relationship cannot be established. However, the data suggest that glucocorticoids and thiazolidinediones exert fundamentally antagonistic effects on human metabolism in both adipose and muscle tissues. By preventing or reversing insulin resistance, troglitazone may prove to be a valuable therapeutic agent in the difficult clinical task of controlling diabetes in patients receiving glucocorticoids.