The urinary estrogen profile was studied in the midfollicular phase twice, and diet four times during 1 yr in 10 premenopausal breast cancer (BC) patients consuming an omnivorous normal Finnish diet and in two control groups, one consuming an omnivorous (n = 12) and the other a lactovegetarian (n = 11) diet. Total fat intake in relation to caloric intake was almost identical in all three groups. Only with regard to grain fiber intake did the BC patients differ significantly from both other groups. No differences were found between the groups with regard to urinary excretion of 13 individual estrogens and total estrogens, with the exception of 4-hydroxyestrone (4-OH-E1), which was significantly lower (P less than 0.05) in the BC group than in the vegetarians. A high carbohydrate to protein ratio in the diet had a negative correlation with the excretion of 2-hydroxyestrogens and 2-hydroxyesterone (2-OH-E1) to 4-OH-E1 ratio. The BC group had significantly higher urinary 2-OH-E1 to E1 ratio (P less than 0.05) compared to the vegetarians. The 2-OH-E1 to 4-OH-E1 ratio was highest in the BC group (= 7.1) and differed significantly from that of the omnivores (= 4.3; P less than 0.02) and vegetarians (= 3.6; P less than 0.005). This ratio showed a negative correlation with intake of carbohydrates, starch, total and grain fiber. Urinary excretion of 4-OH-E1 correlated positively with total and grain fiber intake and plasma SHBG. Protein intake correlated positively with urinary 2-methoxy-E1 excretion, and retinol intake positively with catechol estrogen, E1 and E2 excretion. It is concluded that estrogen production and urinary estrogen profile in premenopausal breast cancer patients is normal with the exception of a low 4-OH-E1 excretion and high urinary 2-OH-E1 to 4-OH-E1 ratio. This ratio, which seems to depend on diet, is the only urinary estrogen parameter separating premenopausal BC patients from the control omnivorous and lactovegetarian women.