Pregnancy has been associated with increased risk for diabetes, but an association independent of obesity has been questioned. The hypothesis that the repeated demands of pregnancy lead to reduced sensitivity to insulin and hyperinsulinemia was tested by examining the relation of pregnancy history to fasting and postchallenge insulin levels and insulin sensitivity in 764 Caucasian, nondiabetic women aged 50-89 years from the Rancho Bernardo cohort. Between 1984 and 1987, an oral glucose tolerance test was administered, fasting and postchallenge levels of glucose and insulin were measured, and an index of insulin sensitivity was calculated. After adjustment for covariates, an increased number of pregnancies was associated with increased fasting insulin (p < 0.05) and decreased insulin sensitivity (p < 0.05). Postchallenge insulin was unrelated to pregnancy history. Obesity and fat distribution were associated with higher fasting and postchallenge insulin levels and lower insulin sensitivity (ps < 0.001), but did not explain the associations between pregnancy history and insulin or insulin sensitivity. Analyses restricted to women with normal glucose tolerance yielded similar results. Thus, there is a small, but significant increase in fasting insulin and a decrease in insulin sensitivity with increasing pregnancies many years after childbearing that is independent of obesity and present even in normoglycemic women. Confirmatory studies using insulin clamp techniques would be of interest.