Recently, a number of prospective studies showed evidence that the growth hormone/insulin-like growth factor I (IGF-I) axis may be important in the development of colorectal cancer. However, only a few studies have reported on the possible relationship of colorectal cancer risk with circulating levels of IGF-II, which are not growth hormone dependent and which do not vary with alterations in energy balance. In a case-control study of 102 cases and 200 matched controls nested within a cohort of 14,275 women in New York, we examined the relationship between colorectal cancer risk and prediagnostic serum levels of IGF-II. Conditional logistic regression analysis showed an odds ratio (OR) for colorectal cancer of 2.02 (95% confidence interval (CI): 0.83-4.93), comparing the upper to lower quintile of IGF-II. This association was slightly attenuated after excluding IGF-II measurements in serum samples taken within 1 year before case diagnosis (OR of 1.81; 95% CI: 0.71-4.64) and moderately attenuated after excluding IGF-II measurements in serum samples taken within 2 years before case diagnosis (OR of 1.47; 95% CI: 0.56-3.91). Adjustment for IGF-1, IGF binding protein (BP)-1, IGFBP-3, smoking, or body mass index did not substantially alter the association, whereas adjustment for IGFBP-2 moderately attenuated the relationship. Our results confirm those of three recent case-control studies, and collectively these results suggest a possible increase in colorectal cancer risk among subjects with comparatively elevated serum IGF-II. Mechanisms that might cause the increase in IGF-II levels are unknown but may include loss of parental imprinting of the IGF-II gene.