Plasma levels of insulin-like growth factor I (IGF-I) have been associated with risk of several cancers. Although protein-calorie malnutrition is known to decrease IGF-I levels, few published studies have related diet to IGF-I levels in well-nourished humans. We examined the cross-sectional association of plasma IGF-I and IGF-binding protein 3 (IGFBP-3) levels with intakes of alcohol, energy, macronutrients, micronutrients, and specific foods in 1037 healthy women. Adjusted mean hormone levels across categories of dietary variables were calculated by linear regression. Results were adjusted for non-dietary factors found to be associated with IGF levels. Total energy intake was positively associated with IGF-I levels when adjusted for covariates. Adjusted mean levels of IGF-I (ng/ml) across increasing quintiles of energy intake were 181, 185, 191, 199, and 195 (P for the linear trend = 0.006). In other multivariate analyses, energy-adjusted fat and carbohydrate intake had no association with IGF-I levels. The most consistent finding was a positive association between protein intake with circulating IGF-I concentration (174, 188, 201, 192, and 196 ng/ml across quintiles of protein intake; P = 0.002), which was largely attributable to milk intake. Adjusted mean levels of IGF-I (ng/ml) across increasing quartiles of milk intake were 183, 189, 188, and 200 (P = 0.01). Higher fat intake, in particular saturated fat, was associated with lower levels of IGFBP-3. We conclude that higher energy, protein, and milk intakes were associated with higher levels of IGF-I. These associations raise the possibility that diet could affect cancer risk through influencing IGF-I level.