Serologic precursors of cancer. I. Prediagnostic serum nutrients and colon cancer risk

A study was undertaken to determine whether prediagnostic serum levels of retinol, beta-carotene, vitamin E, and selenium are lower in colon cancer cases compared with matched, population-based controls. Sera were available from 25,802 participants of a serum collection campaign conducted in Washington County, Maryland in 1974. The authors identified from these participants 72 white colon cancer cases, who were first diagnosed with colon cancer during 1975-1983, and 143 white, living, cancer-free controls, matched to cases on the basis of age, sex, month of serum collection, and enumeration in a 1975 private census of Washington County. The mean values of serum nutrients in cases and controls, respectively, were 59.1 micrograms/dl and 61.8 micrograms/dl for retinol (p = 0.22), 32.9 micrograms/dl and 34.4 micrograms/dl for beta carotene (p = 0.52), 1.17 mg/dl and 1.27 mg/dl for vitamin E (p = 0.10), and 11.0 micrograms/dl and 11.5 micrograms/dl for selenium (p = 0.07). There were no consistent trends in the relative odds of colon cancer by quintiles of serum levels for any of the nutrients; however, a relative odds of 3.2 (95% confidence interval = 1.1-8.7) was found when persons in the four lowest quintiles of retinol were compared with those in the highest. No interactions with matching factors or between serum nutrients and no confounding effects of covariables were identified through conditional logistic regression analysis. The findings of this study do not support a strong association of low serum levels of retinol, beta-carotene, vitamin E, and selenium with an increased risk of subsequent colon cancer.