Changes in dietary fat intake preceding the diagnosis of cancer

We studied changes in dietary fat intake preceding the diagnosis of cancer in 96 men diagnosed with cancer during the Coronary Primary Prevention Trial and Post-trial Follow-up. Diet was assessed semiannually by 24-hour recall for the duration of the trial. Dietary fat intake was elevated in the interval 12-24 months before diagnosis compared with the interval 24-36 months before diagnosis [6.9 gm per day; 95% confidence interval (CI) = 0.04-13.7]. The percentage of calories as fat was also elevated (1.9%; 95% CI = 0.34-3.4). Fat consumption decreased in the 0- to 12-month prediagnostic interval. The basis for the elevation is unclear, but it may have been due to the metabolic effects of cancer before its diagnosis. Case-control studies of long-term diet may overestimate positive associations between dietary fat and cancer risk if the cases' recall includes the prediagnostic period, during which fat intake is elevated. In analyses of prospective studies of dietary fat and cancer, data from within the first 2-3 years of baseline should be examined separately to evaluate the potential for bias introduced by the inclusion of measurements representing transient elevations in dietary fat intake.