Breast cancer epidemiology : influence of hormone-related factors

Abstract

The main purpose of this thesis was to explore how hormone-related factors, including use of hormone replacement therapy and oral contraceptives, body size in different periods of life and reproductive factors, influence the risk of breast cancer. Other aims were to assess how family history of breast cancer modifies the associations between such factors and risk of the disease and how hormone replacement therapy affects tumour characteristics.

The principal part of this work was based on a population-based case control study conducted in all of Sweden during October 1993 to March 1995. In this study 3,345 women aged 50-74 years with a diagnosis of primary breast cancer (84% of the eligible) and 3,454 age frequency-matched control women (82% of the selected) were included. Detailed information on hormone-related factors was collected through mailed questionnaires and telephone interviews. In another approach, a pathology register of 1,589 women with primary breast cancer diagnosed in Uppsala county from May 1977 to December 1991 was used. These women's exposure to hormone replacement therapy was determined through cross-linkage with a cohort of 23,234 women in the Uppsala health care region who had received prescriptions of these drugs.

We found a duration-dependent increase in breast cancer risk among current as well as past users of oestrogen and oestrogen-progestin replacement therapy (odds ratio 2.23, 95% confidence interval 1.71-2.93 for women who had been treated for at least ten years as compared to those never treated). Only non-obese women seemed to be susceptible to this adverse effect. Breast cancers diagnosed after hormone replacement therapy had apparently less malignant features than other cancers. We also demonstrated a relation between childhood build and breast cancer risk, and women with the leanest constitution at age 7 had an approximately threefold higher risk than those with the most obese (p for trend 0.0009). Weight gain between age 18 and recent ages, rather than obesity per se, was also a convincing predictor of risk. Yet, this association was only demonstrated in older post-menopausal women. Among women who were 20 years past their menopause, the risk for those who had gained 30 kg or more was doubled as compared to those who had maintained their weight.

We found a negative association between increasing age at menarche and breast cancer risk in women born before 1925 (p for trend 0.008) but not after. Augmenting number of births was strongly protective of breast cancer, while neither lactation, menopausal symptoms or past use of oral contraceptives were associated with risk. Hormone-related risk factors were similarly related to breast cancer risk among women with and without a family history of the disease.