The impact of reproductive factors on breast cancer risk--the feasibility of using Swedish population-based registers to account for the effect of confounding in cohort studies.

OBJECTIVE: The aim of this study was to analyze the impact of reproductive factors on breast cancer risk among Swedish women by using nationwide population registers.
METHODS: A cohort including all Swedish women born between 1920 and 1959 was followed up to 1997 by record linkage to different population-based registers. More than 4 million children were linked to the women in the cohort and 60,328 women were diagnosed with breast cancer. Poisson regression was used to model the breast cancer incidence by risk groups. In a sub-cohort of 18,164 women irradiated for skin hemangioma in infancy, the breast cancer risk was analyzed in relation to radiation dose and accounting for reproductive factors.
RESULTS: The relative breast cancer risk (RR) for the reproductive factors was 0.35 [ 95% confidence interval (CI) 0.30--0.42] for women with 6 or more children and the first child before the age of 20 years, and RR was 1.11 (95% CI 1.06--1.18) for uniparous women with the first child at age 35 years or older, compared to nulliparous women. The discrepancies of reproductive factors in the hemangioma cohort compared to Swedish women had a minor effect on RR, with a reduction from 1.13 (95%CI 1.00--1.26) to 1.11 (95% CI 0.99--1.25).
CONCLUSIONS: This study shows the feasibility of using population-based registers to retrieve reliable information on reproductive factors to eliminate its confounding effect when analyzing other risk factors.