In 1982 we started a series of pilot studies to examine the feasibility of dietary intervention with a low-fat, high-carbohydrate diet in women with extensive mammographic densities. The purpose of the present paper is to examine the long-term effects of participation in these studies by assessing nutrient intake and other variables several years after active participation had stopped. Two hundred sixteen women were eligible for the follow-up study and were invited to attend and interview with a dietician. Data were collected by food frequency questionnaire from 157 subjects (73%), and blood was obtained from 115 subjects. Total energy intake was slightly lower in the intervention group. Total fat and percent energy from fat were significantly lower in the intervention group. The intake of all types of fat (saturated fat, linoleic acid, and oleic acid) and dietary cholesterol was lower in the in the intervention group; however, the polyunsaturated/saturated fat ratio did not differ between the groups. Total cholesterol and apoprotein B levels were lower in the intervention group compared to the control group. Follicle-stimulating hormone was 29% higher in postmenopausal members of the intervention group than in controls, but there was no difference in levels of estradiol. A total of 19 women enrolled in pilot studies had developed breast cancer, 5.7 times the number expected, confirming that the selection of women with extensive mammographic densities does identify a high-risk group. These data suggest that even quite short periods of intensive dietary counselling may have prolonged effects on diet, and that once subjects have adopted new dietary habits, the habits may persist even in the absence of continued counselling.