The relationship between oral contraceptives (OC), menopausal hormone replacement treatment (HRT) and breast cancer risk was analysed in a case-control study conducted in the Swiss canton of Vaud on 230 cases below the age of 75, linked with the Vaud cancer registry, and 507 controls in hospital for a wide spectrum of acute, non-neoplastic, non-hormone-related diseases. A total of 77 (37.4%) cases and 134 (31.6%) of controls below the age of 70 had ever used OC, corresponding to a multivariate odds ratio (OR) of 1.5, of borderline significance. The risk was related to duration of use, but this might be partly due to a recency effect. In fact, with reference to time since last OC use, the OR was above unity for up to 14 years, but declined to 1.0 for women who had stopped OC use for > or = 15 years. With reference to HRT, ever users were 64 (27.8%) of cases and 113 (23.3%) of controls, yielding a multivariate OR of 1.3. Also, for HRT the association was stronger for shorter time since last use; the OR was 1.5 for women who had stopped for < 10 years, but declined to 0.8 for those who had stopped for 10 years or longer. The association with HRT was stronger for women aged > or = 65. Thus, the present study confirms that breast cancer risk is moderately related to OC and HRT. The association, however, is essentially restricted to the 10-15 years after stopping use. This pattern of risk is consistent with a late-stage effect of steroid hormone preparations on the process of breast carcinogenesis, and has relevant implications for any risk/benefit assessment and public health evaluation.