Randomised trials of cholesterol reduction (26 trials, 50,000 patients, net cholesterol reduction approximately 10%) have demonstrated a clear reduction in the incidence of coronary heart disease (CHD) after just a few years of treatment. Overall, the reduction in CHD death was only half as large as the reduction in non-fatal myocardial infarction (MI), although both were statistically significant (2P < 0.005). In these trials, 60% of all deaths were from CHD, and since treatment reduced these by about 10%, the expected reduction in total deaths was about 6%. This expected reduction falls within the 95% confidence interval of the observed effect of cholesterol reduction on total mortality in these trials. There were small excesses of deaths from cancer and deaths from trauma among patients allocated active treatment. However, in no single trial, nor in the trials collectively, were these increases statistically significant. Furthermore, the increases did not appear to be specific to any one agent nor were the increases consistent between trials of the same agent. These observations are consistent with the hypothesis that the small excesses of cancer and trauma deaths observed in the cholesterol reduction trials occurred by chance.