The extent to which diet influenced colon cancer survival was examined in 411 colon cancer cases identified in Utah between 1976 and 1981 using data from two population based case-control studies. After adjustment by proportional hazards regression models for the effects of tumour stage, age, sex, and religion, the hazard rate ratios (HRR) comparing highest to lowest quartile of intake for total calories, fat and protein were 0.60, 0.81 and 0.66 respectively, with the effect of calories being greatest for short-term survival (less than or equal to 24 months), HRR = 0.49. By contrast, the highest quartile of dietary fibre intake was associated with decreased survival (HRR = 1.53) when compared with the lowest quartile. More extensive studies are needed to verify these findings and to identify mechanisms underlying these associations.