We investigated estrogen (estrone and estradiol) levels in serum and in nipple aspirates of breast fluid in relation to reproductive and menopausal characteristics in 104 normal women. In general, breast fluid and serum estrogen levels were not correlated and breast fluid estrogen levels were approximately 5 to 45 times higher than serum levels. Serum estrogen levels were lower in post-menopausal than in premenopausal women. In contrast, breast fluid estrogen levels were approximately the same in pre-and post-menopausal women. Breast fluid estrogen mean levels were lower in premenopausal parous women than in nulligravidous or nulliparous women whereas serum estrogen levels did not differ in these 3 groups. Breast fluid estrogen levels were positively correlated with months since last birth or since last breastfeeding. Estrogen levels were low in nipple aspirates of breast milk but gradually increased in breast fluid of non-lactating women over a period of several years after cessation of lactation. Serum estrogen levels did not increase with months since last breast-feeding. We were unable to evaluate the postpartum effect of pregnancy without lactation due to the small numbers of these subjects. The high concentrations of estrogen in breast fluid and the absence of a relationship to serum estrogen levels may explain why prior serum studies have failed to link variations in serum estrogens with breast cancer risk. The prolonged low levels of breast fluid estrogens following full-term birth and lactation may, in part, provide a mechanism by which parity reduces breast cancer risk.
' To whom reprint requests should be sent, at the Department of Epidemiology and International Health (Box 0560). School of Medicine, Uni-