[Reproductive variables and risk of breast cancer: a case-control study carried out in Brazil].

The purposes of the study were to identify reproductive risk factors for breast cancer and to test the hypothesis that reproductive variables operate through a common factor: the number of times that the mammary tissue has been exposed to the endocrine changes of the ovulatory cycle. The study was conducted in Campinas, and was based on interviews with 348 women with breast cancer first diagnosed between October 1979 and August 1984. The control group consisted of 348 women with healthy breasts. The data were obtained in interviews in the home, which were conducted using a structured questionnaire that had been pretested. The data were analyzed by calculation of the odds ratio, Mantel's statistic, to determine the linear trend; by Cornfield's method to calculate the confidence intervals; and by logistic regression adjusted for paired data. It was found that nulliparity, not having breast-fed, and a high number of menstrual cycles were significantly associated with the risk of presenting the disease. In the multivariate analysis, which included all the women, the only variable associated with the risk of mammary cancer was no history of breast-feeding. When the nulliparas were excluded, logistic regression showed that higher age of the woman at her first delivery was significantly associated with breast cancer.