Self-reported dietary intake was estimated from 3-day prospective food diaries completed by Type 2 diabetic patients in the UK Prospective Diabetes Study. All patients had received individual dietary advice and had been randomly allocated to diet, sulphonylurea or insulin therapy 3 months after diagnosis. A total of 132 patients (120 white Caucasian, 12 Asian) stratified for gender, obesity and allocated therapy with mean age 55 years (SD 8), body mass index 28 kg m-2 (SD 4), and with a diabetes duration of 3 to 6 years were selected at random from 5 of 23 clinical centres. Patients reported a similar proportion of their energy intake as carbohydrate (43%) to the general population and had not increased to the recommended 50-55%. Their protein intake (21%) was higher than the advised 10-15%. Estimated energy intake from fat (37%) was close to that recommended for diabetic patients (30-35%) and was lower than that reported for the UK population (40%). The estimated polyunsaturated/saturated fat intake ratio (0.48) was higher than that reported for the UK population (0.35) compared with the recommended 1.0. Mean fibre intake at 22 g day-1 was less than the recommended 30 g day-1. The 8 male Asian patients took a higher proportion of their dietary intake as fat (46% vs 37%) and lower as protein (14% vs 21%) than the male white Caucasian patients. No significant differences were seen in estimated nutrient constituents between patients allocated to diet, sulphonylurea or insulin therapy as part of the UK Prospective Diabetes Study and followed for mean 4.2 years (SD1.6). This suggests that dietary factors will not confound UK Prospective Diabetes Study treatment related analyses.