Diabetes and hyper-insulinemia as predictors of colorectal cancer risk in a prospective cohort of women

B93

Insulin, which is chronically elevated in the pre-diabetic condition and in diabetes before pancreatic exhaustion, stimulates growth of both normal colonic and carcinoma cells, and perhaps more importantly, modulates IGF-I and its binding proteins to create a pro-mitotic environment for colonic epithelial cells. We analyzed data from a prospective cohort of women selected from participants in a breast cancer screening program conducted between 1973-81 at 29 centers throughout the United States to determine if diabetes was a risk factor for colorectal cancer. Since hyper-insulinemia might be misclassified if we only considered subjects with diagnosed diabetes at baseline as exposed, we did additional analyses in which we analyzed risk of colorectal cancer among those who were either diabetics at baseline or likely pre-diabetic at baseline (identified based on subsequent self-report of diabetes at the 1993-1995 follow-up). Between 1987 and 1989, 45,519 women without a history of colorectal cancer or suspected Type I diabetes (onset before age 25) successfully completed a series of baseline questionnaires assessing dietary and other health and lifestyle risk factors for colorectal cancer. Mean age at time of interview was 61.9 years. Women were followed for an average of 8.5 years between 1987-89 and 1995-98, and during this time 489 women developed colorectal cancer. We used Cox proportional hazards models, with age as the underlying time metric, to estimate risk of colorectal cancer. Models of diabetes diagnosed at or before baseline controlled for age, physical activity, energy intake, alcohol, menopausal hormone therapy, smoking, multivitamin use, education, ethnicity, NSAID use, calcium from supplements, and intake of calcium from diet. The multivariate-adjusted hazard ratio (and 95% confidence interval) for prevalent diabetes at baseline was 1.50 (1.09-2.07). After reclassifying the exposed category to include those who were likely pre-diabetic at baseline, the multivariate-adjusted hazard ratio remained statistically significant, but was of lesser magnitude: 1.36 (1.01-1.83) for self-report of diabetes at baseline or at 1993-1995 follow-up. These results suggest that prior diagnosis of diabetes increased risk of colorectal cancer, and therefore support the hyper-insulinemia hypothesis. However, our observation that the risk estimate was attenuated if we included in the exposed category subjects who were likely pre-diabetic (and therefore hyper-insulinemic) at baseline suggests that either the degree of hyper-insulinemia in such subjects was not of sufficient magnitude or of sufficient duration to increase risk or that some factor related to diabetes but independent of hyper-insulinemia was responsible for the increased risk we observed for diabetic subjects.