For a long time it has been postulated that diet may influence estrogen metabolism and in this way affect breast cancer risk. In order to investigate possible effects of variations of dietary fiber intake on estrogen metabolism, the urinary estrogen profile (13 estrogens), including the catecholestrogens, was determined in one 72-h summer and one winter sample collected in the midfollicular phase of the menstrual cycle by 11 lactovegetarian and 12 omnivorous young Finnish women. Urinary estrogens were purified by ion-exchange chromatography and the quantitative determination was carried out by capillary gas chromatography-mass spectrometry. Detailed records of the subjects' diet during one 5-day period in summer and one in winter were obtained and dietary fiber intake calculated. The mean difference with regard to intake of total fiber in the two dietary groups was 3 g/day in the summer (not significant) and 5 g/day in the winter (P less than 0.05), the mean (geometric) consumption being 23 and 19 g/day by the vegetarian and omnivorous women, respectively. Within the groups we found seasonal variation in fiber intake only for the omnivorous women. During winter, compared to summer, the omnivorous women consumed significantly less grain (P less than 0.001), vegetable (P less than 0.02) and total fiber (P less than 0.02). The excretion of 13 estrogens was remarkably constant in the omnivoric group but a significant seasonal variation of total and individual catecholestrogens and of estrone was observed in the vegetarians (P less than 0.05-0.005). The quantitatively most important estrogen was 2-hydroxyestrone, followed by estrone, estriol, 2-hydroxyestradiol, 4-hydroxyestrone and estradiol, the three latter being excreted in similar amounts. Between the dietary groups there were no significant differences in excretion of total or individual urinary estrogens in any season or between the mean values for both seasons. However, numerous significant (P less than 0.05-0.01) negative correlations between dietary intake of total or grain fiber/kg body weight and the excretion of individual estrogens were found. These correlations disappeared if the fiber intake was not related to body weight. We conclude that dietary fiber intake significantly affects estrogen metabolism by reducing estrogen excretion in urine and that grain fiber seems to be most important in that respect. One of the mechanisms involved is a partial interruption of the enterohepatic circulation of the estrogens, due to alterations of the intestinal metabolism and reabsorption of these steroids, caused by the fiber.