Estrogen/progestin steroid combinations adversely affect glucose tolerance and insulin resistance, but their effects in combined hormone replacement therapy (HRT) have rarely been evaluated. We studied 61 untreated symptomatic postmenopausal women randomized to receive oral (conjugated equine estrogens, 0.625 mg/d continuous + levonorgestrel, 0.075 mg/d for 12 days of each 28-day cycle) or transdermal therapy (estradiol 17 beta, 0.05 mg/d continuous + norethindrone acetate, 0.25 mg/d for 14 days of each 28-day cycle). An untreated control group of 30 postmenopausal women not seeking HRT was also studied. Intravenous glucose tolerance tests (IVGTT) were performed at baseline and 3, 6, and 18 months later. Mathematical modeling analysis of plasma glucose, insulin, and C-peptide concentration profiles provided measures of insulin resistance, secretion, and elimination. There were no changes in glucose or insulin concentrations with transdermal therapy. Oral therapy caused a deterioration of glucose tolerance and an increased overall plasma insulin response, apparently due to a reduction in the immediate plasma insulin response to glucose. This may have resulted from increased hepatic insulin uptake, uncompensated for by an increase in first-phase pancreatic insulin secretion. Neither treatment caused significant insulin resistance compared with baseline, but with the oral treatment insulin resistance was greater during the combined phase compared with the estrogen-only phase. Thus the oral regimen affected both insulin delivery and insulin resistance. The transdermal regimen had relatively few effects on insulin metabolism.