Risk factors for breast cancer in women under 40 years.

The relation between hormonal and lifestyle factors and breast cancer risk in women younger than 40 years was investigated using data from two case-control studies conducted in Italy between 1983 and 1994. Cases were 579 women with histologically confirmed, incident breast cancer and controls were 668 women admitted to hospital for acute, non-neoplastic, non-hormone-related diseases. Breast cancer risk was inversely related to age at menarche with a multivariate odds ratio (OR) of 0.53 (95% confidence interval, CI, 0.31-0.89) for women reporting menarche at the age of > or = 15 years compared with < 12 years. Breast cancer risk was significantly higher in parae than in nulliparae (OR 1.58), and was directly associated with age at first birth (OR 5.31 among women aged > or = 30 years at first birth compared with those aged < 20), and inversely with time since last birth (OR 3.80 for < 3 years compared with > or = 12). Compared with women reporting no abortion, the OR were 1.10 for any spontaneous, 0.87 for any induced and 0.90 for > or = 2 abortions. With reference to oral contraceptive use, the OR was 1.05 for ever users compared with never users, and no material association was evident with duration, time since first and last use. The OR was 1.79 for more than 13 years of education compared with < 9, 1.85 for a family history of breast cancer and 1.85 for a history of benign breast disease. Breast cancer risk was inversely related to body mass index with an OR of 0.51 (95% CI 0.26-0.97) for > or = 30 kg/m2 compared to < 20. Total energy and alcohol intake were directly related to the risk (OR 1.38 and 1.27 for the highest tertiles of intake compared with the lowest), although the estimates were not significant, whilst raw vegetable and beta-carotene consumption were inversely related to breast cancer risk (OR 0.57 and 0.67 for the highest tertile of intake compared with the lowest). Thus, most risk factors in this large dataset of women aged less than 40 years were similar to those described in breast cancer epidemiology at any age. Of interest are the inverse associations with body mass index, age at menarche and time since last birth, the direct ones with age at first and last birth, and the higher risk of parous women compared with nulliparae.