On the basis of observation of the effects on glucose metabolism of estrogens and progestins administered alone or in combination, it is believed that deterioration of glucose tolerance in the 1st trimester of oral contraceptive (OC) use is due to the synthetic estrogen, while hyperinsulinism after the 6th month of use is due to the progestin component. Either the estrogen-induced deterioration of glucose tolerance or the progestin-induced insulin resistence may lead to poor glucose tolerance or diabetes in predisposed individuals. The early deterioration of glucose tolerance with synthetic estrogens is probably due to a direct action of estrogens on the pancreas, where specific estrogen receptors have been identified, but the regression of this effect has not been explained. The reduction of fasting glucose levels by synthetic estrogens is due to a reduction of glucogenolysis and of gluconeogenesis with an increase of hepatic gluconeogenesis. These 3 effects may be explained by the action of estrogens on insulin and glucagon: estrogens inhibit the secretion of glucagon induced by arginine, an effect not obtained by progestins. A daily injection of estradiol for 6 weeks in overiectomized rats results in a lowered level of insulin and glucagon with an elevated insulin/glucagon ratio in the portal vein; this results in a decline in gluconeogenesis with a particular decline in phospho-enol-pyruvate-carboxy-kinase activity. Progestins on the other hand augment insulin and glucagon secretion without modifying their ratio in the portal vein. Hyperinsulinism in the basal state and after stimulation is observed in women using combined OCs or progestins only, whether or not there is a deterioration of glucose tolerance. The effect appears after 3 months of use and continues as long as OCs are used. This hyperinsulinism is due to a hypersecretion and not to a diminution in degradation of insulin. The insulin/glucose ratio is elevated; the hypersecretion of insulin compensates for a lowered peripheral sensitivity to the hormone. Pill-induced insulin resistence is comparable to that of pregnancy. Although all progestins used in contraception induce insulin resistence with hyperinsulinism, the effect is more marked with 19 nortestosterone derivatives than with those of 17 alpha hydroxyprogesterone, because of the structural similarity of 19 norsteroids to anabolic androgens. 2 arguments suggest a potentialization of the effects of progestins by the estrogens in combined OCs: estrogens diminish the biliary flow and may therefore diminish progesterone secretion, and progestin catabolism is diminished by estrogens. Theories have been proposed to explain the diabetogenic effect of combined OCs through indirect mechanisms involving cortisol, growth hormone, or tryptophane metabolism.