| Summary: | 碩士 === 臺北醫學大學 === 公共衛生學系 === 93 === Background Female breast cancer was the second leading incidence of malignant tumors and of female cancers in Taiwan. Although there were a great deal of investigations focusing on the elucidation of risk factors of developing breast cancer, the distinguishing of environment factors from genetic factors still attracts great interests. In addition, as the peak of incidence rate in Taiwan is around 45 years of age, which is earlier compared with western countries. As familial effects have been recognized as increasing in younger women, the elucidation of risk factors such as reproductive factors for the risk of breast cancer in women with positive family history of breast cancer is therefore of paramount important. In the current thesis, we aimed to investigate the risk factors for breast cancer among women with family history from a hospital-based high-risk cohort with breast cancer screening with consideration of correlation in the cohort.
Materials and methods A total of 4867 women in the current study was included from a hospital-based screening program, Taiwan Multicentre Cancer Screening Project (TAMCAS), for women with female breast cancer within second-degree relatives underwent breast cancer screening between 1994 and 1998. Subjects from the same family were treated as in one family proband. Logistic regression models were firstly used to investigate the effect of different risk factors assuming independency among participants regardless family probands. Generalized estimating equation (GEE) model was used to consider correlation among participants from the same family proband.
Results The results of multi-variate analysis of logistic regression recognized risk factors of breast cancer included age, regularity of menses, smaller number of child birth or nullparity, higher level of education, and not employed. Since the correlation between age and period of lifetime CBMA, after excluding age the final model included lifetime BCMA, regularity of menses, smaller number of child birth or nullparity, menopause, and three or more first-degree relatives with breast cancer. The similar findings of univariate analysis from GEE model were found in both models with autoregressive and unstructured working correlation matrix. The GEE model included age, age at menarche, regularity of menses, and contraception as recognized risk factors. When excluding age from
the multi-variate model, the significant factors retained in the model was lifetime BCMA, regularity of menses, number of child birth, menoparuse, and three or more first-degree relatives with breast cancer.
Conclusion In the present study, we confirmed genetic factors and reproductive factors play respective important roles in the risk for developing breast cancer even in Taiwanese women with positive family history. The built-in predictive model may be helpful for prevention of women with family history.
|
|---|
