To assess the mechanisms responsible for the insulin resistance associated with both normal human pregnancy and gestational-onset diabetes, we have measured exogenous glucose disposal using sequential insulin infusions with the euglycemic glucose clamp technique and erythrocyte insulin binding. Three groups of women were studied: nonpregnant women with normal glucose tolerance (N = 7, mean age 32.9 +/- 2.1 yr), pregnant women with normal glucose tolerance (N = 5, mean age 24.8 +/- 3.5 yr), and pregnant women with gestational-onset diabetes (N = 5, mean age 34.6 +/- 2.6 yr). Despite normal plasma glucose levels obtained during a 100-g oral glucose tolerance test, plasma insulin levels were significantly elevated in the pregnant women compared with the nonpregnant control subjects, suggesting a state of insulin resistance. Insulin binding to erythrocytes was similar in all three groups (maximum specific binding being 5.0 +/- 0.6%, 5.5 +/- 1.1%, and 6.0 +/- 0.7% in nonpregnant, nondiabetic pregnant, and gestational-onset diabetic women, respectively). In vivo peripheral insulin action was measured using the euglycemic glucose clamp technique during an insulin infusion of 40 mU/m2 X min, with blood glucose clamped at a concentration of 75 mg/dl using a variable glucose infusion. Glucose infusion rates were 213 +/- 11 mg/m2 X min, 143 +/- 23 mg/m2 X min, and 57 +/- 18 mg/m2 X min in nonpregnant, nondiabetic pregnant, and gestational-onset diabetic women, respectively. This demonstrates that pregnant subjects display a state of insulin resistance, and that this appears to be more marked in gestational-onset diabetic subjects. To further define the possible mechanism of insulin resistance during pregnancy, the insulin infusion rate was increased to 240 mU/m2 X min and further euglycemic clamp measurements performed. Glucose infusion rates were 372 +/- 11 mg/m2 X min, 270 +/- 31 mg/m2 X min, and 157 +/- 26 mg/m2 X min, in nonpregnant, nondiabetic pregnant, and gestational-onset diabetic women, respectively. This demonstrates a shift to the right of the dose-response curve of insulin action and suggests that the insulin resistance of pregnancy may include a decrease in presumed "maximum" insulin responsivity. In four subjects, studies were repeated in the postpartum period, and these demonstrated that the insulin resistance of pregnancy is ameliorated shortly after delivery. These studies suggest that the insulin resistance of pregnancy results from a target cell defect in insulin action beyond the initial step of insulin binding to cellular receptors, a postreceptor (or postbinding) defect in insulin action.