To provide quantitative information on the role of age at any birth for breast cancer risk, we analyzed data from a cooperative Italian case-control study conducted between 1991 and 1994 on 2,569 incident, histologically confirmed breast cancer cases and 2,588 controls in hospital for acute, non-neoplastic, non-gynecological conditions. A single logistic model was fitted, including terms for number of births, age at each birth and at menarche, plus age and center. Age at first birth was the strongest reproductive determinant of subsequent breast cancer risk, with an estimated increase of 4.6% per year of delay of first birth. This was similar to the influence of age at menarche (4.7% decrease in risk per year of delay of menarche). Ages at subsequent births had an independent effect on breast carcinogenesis, with an estimated 0.7% increase in risk per year of delay. Multiparity showed also an independent protection on breast cancer risk, and a protective effect of parity > or = 3 was evident in all strata of age at first birth: the odds ratio was 0.81 for 3 births and 0.70 for > or = 4 births. However, the effect of parity was determined by the age of occurrence of various births.