Objective: Studies examining the association of dairy consumption with incident CHD have yielded inconsistent results. The current prospective study examined the association between dairy consumption and CHD in a population-based sample of older community-dwelling adults. Design: Baseline CHD risk factors were assessed and an FFQ was self-administered. Participants were followed for morbidity and mortality with periodic clinic visits and annual mailed questionnaires for an average of 16?2 years, with a 96 % follow-up rate for fatal and non-fatal CHD. Setting: Community. Subjects: Participants were 751 men and 1008 women aged 50-93 years who attended a clinic visit in [1984][1985][1986][1987]. Results: At baseline the mean age was 70?6 (SD 9?8) years for men and 70?1 (SD 9?3) years for women. Participants who developed CHD during follow-up were significantly older (P , 0?001), had higher BMI (P 5 0?035) and higher total cholesterol (P 5 0?050), and were more likely to be male (P , 0?001), diabetic (P 5 0?011) and hypertensive (P , 0?001), than those who did not develop CHD. Multivariate regression analyses adjusting for age, BMI, diabetes, hypertension, LDL-cholesterol and oestrogen use (in women) indicated that women who consumed low-fat cheese 'sometimes/often' and women who consumed non-fat milk 'sometimes/often' had an increased risk of incident CHD (hazard ratio 5 2?32; 95 % CI 1?57, 3?41) and CHD (hazard ratio 5 1?48; 95 % CI 1?02, 2?16) compared with women who 'never/rarely' ate these dairy products. Conclusions: Woman with higher intake of low-fat cheese and non-fat milk seem to have a higher risk of incident CHD. This needs further investigation considering recent evidence of cardiovascular benefits from certain dairy fat.