The Caerphilly Prospective Ischaemic Heart Disease (IHD) Study is based on a sample of 2512 men aged 45-59 years when first seen. Nutrient intakes, estimated using a self-administered semi-quantitative food frequency questionnaire, are available for 2423 men (96%). Amongst these, 148 major IHD events occurred during the first 5 years of follow-up. Associations were examined between these events and baseline diet. Incident IHD (new events) was negatively associated with total energy intake: men who went on to experience an IHD event had consumed 560 kJ (134 kcal)/d (6%) less at baseline than men who experienced no event (P = 0.01). The relative odds of an IHD event was 1.5 among men in the lowest fifth of energy intake, compared with 1.3, 1.2, 0.9 and 1.0 respectively for the other four fifths (P < 0.05). The difference in energy intake was reflected in lower intakes of every nutrient examined. When expressed as a percentage of total energy, mean intakes of men who experienced an IHD event were virtually identical to those of men who did not. There was some evidence suggesting a positive association between total fat intake and IHD risk, but the trend was not consistent and not statistically significant. There was no association for animal fat. Alcohol consumption was negatively associated with subsequent IHD, but only in men who already had evidence of IHD at baseline (P < 0.05). Dietary fibre, particularly from fruit and vegetables, was 7% lower in men who had an incident IHD event (P < 0.05), but the difference was not independent of total energy. There was a trend of increasing IHD risk with decreasing vitamin C intake, the relative odds of an IHD event being 1.6 among men in the lowest one-fifth of the vitamin C distribution, but this was not statistically significant.