Relationship of time since childbirth and other pregnancy factors to premenopausal breast cancer prognosis.

To investigate the influence of time since childbirth and other pregnancy factors on the prognosis of premenopausal breast cancer. Women who delivered an infant in Nova Scotia, Canada, between 1980 and 2001 were identified from a provincial perinatal database and linked to the Nova Scotia Cancer Registry to determine primary breast cancer diagnoses among women aged younger than 50 years. Relative risks and Cox proportional hazards ratios were calculated to quantify the relationship of time from childbirth to diagnosis and other pregnancy factors to the extent of disease at diagnosis and on survival after breast cancer diagnosis. Of the 123,323 women who delivered an infant during the study period, 716 women were diagnosed with invasive breast cancer. Women with less than 5 years between their last delivery and diagnosis were more likely to be diagnosed with later-stage disease and had poorer survival even after adjusting for stage of disease (less than 2 years, adjusted hazards ratio 2.1, 95% confidence interval 1.2-3.9; 2-4 years, hazards ratio 1.6, 95% confidence interval 0.9-2.8) compared with women with 5 years or more. For every 13 women with less than 2 years between delivery and diagnosis, one excess death will occur, compared with women with 5 or more years between delivery and diagnosis. A time interval of less than 2 years (and 2-4 years) between childbirth and breast cancer diagnosis worsens the prognosis in a dose-response fashion. Clinicians should be aware of these findings when examining women in the first 5 years after a delivery. II.