Pregnancy related risk factors for breast cancer

Abstract: Pregnancies offer a life-long reduction in breast cancer risk. It has repeatedly been shown that both number of pregnancies and age at first birth affect the future breast cancer risk. The mechanisms for this protection are still not fully investigated. Promising results from animal experiments indicate that the placental hormone hCG could be involved, but data from human populations are lacking. Likewise, it is still unclear to what extent pregnancy characteristics influence the protection. These are the questions addressed by this thesis. The first study describes the risk of breast cancer after molar pregnancies. Molar pregnancies involve exposure to high levels of hCG, without other characteristics of a normal pregnancy. All women with a diagnosis of molar pregnancy in the Swedish Cancer Registry (n=3371) between 1958 and 1993 were followed up for subsequent diagnoses of breast cancer. The observed number (59) was then compared to the expected number based on year of birth and follow-up period (46), yielding a SIR of 1.3 (95% confidence interval 1.0-1.7). The second study uses another proxy variable for increased hCG exposure, namely hyperemesis. Breast cancer cases from the Swedish Cancer registry were compared to matched controls regarding diagnoses of hyperemesis in the Swedish In-Patient Care Registry. There was practically no difference between cases and controls (adjusted OR=1.05 for any versus no diagnosis of hyperemesis, 95% CI 0.86-1.27). The third study focuses on duration of molar pregnancies and breast cancer risk. Women diagnosed with hydatidiform mole and a subsequent breast cancer diagnosis were identified from the Swedish Cancer Registry. They were compared to matched controls regarding the duration of molar pregnancy before evacuation. A slightly shorter period was noted for cases than for controls. The fourth study investigates the possible relationship between abortions and breast cancer. Exposure information was abstracted from maternal care and birth records for 1759 cases and an equal number of matched controls. A history of at least one abortion was noted for 383 cases and 473 controls yielding an adjusted odds ratio of 0.84 (95% CI 0.720.99). Both induced and spontaneous abortions were associated with odds ratios below unity. The last study investigates the impact of weight change during pregnancy, child weight and placental weight on maternal breast cancer risk, in the same material as study IV. A slightly increased risk was found with increasing child weight and placental weight, whereas no association was found with maternal weight change. In conclusion, the presented studies do not support the theory of a protective effect of hCG on breast cancer risk. Further, abortions do not seem to increase breast cancer risk, at least not if followed by a childbirth. Finally, maternal risk of breast cancer seems to be modestly associated to offspring birth weight and placental weight, but not to maternal weight change during pregnancy.