Although certain risk factors for breast cancer incidence may also effect survival, findings have been inconsistent and the long-term role of childbirth is unknown. We studied the influence of number and timing of births on breast cancer prognosis prospectively. From 1958 to 1997, altogether 32,003 women, born 1932 or later, were notified to the Swedish Cancer Registry due to a primary invasive breast cancer. We obtained information on dates of all childbirths and achieved complete follow-up through 1997 by means of linkage to other nation-wide databases. Proportional hazards analyses were used to compute crude and multivariate hazard ratios (HR) with 95% confidence intervals (CI) of dying from breast cancer. We found a successively worse prognosis for women with a shorter delay between their last birth and breast cancer diagnosis (p for trend <0.0001). Compared to women with their last birth more than 10 years before diagnosis, the multivariate HR of breast cancer death was 1.39 (95% CI 1.17-1.67) for those with breast cancer diagnosis in the 3rd year after last birth and 1.72 (CI 95% 1.42-2.09) for those with diagnosis within 1 year after last birth. This adverse effect on prognosis of childbirth persisted beyond 10 years among women with a first birth before the age of 20 years. A pregnancy has marked adverse effects on the prognosis of a breast cancer diagnosed within 10 years after delivery. These findings suggest that pregnancy influences tumor biology.
Copyright 2004 Wiley-Liss, Inc.