非吸烟女性肺癌危险因素的多分类logistic模型分析

目的 利用多分类logistic模型分析和评价非吸烟女性不同组织学类型肺癌的危险因素。方法 资料来源于非吸烟女性肺癌病例504例及人群对照601例，拟合多分类logistic回归模型估计各因素的比数比。结果 影响女性腺癌发病的危险因素是厨房小环境污染、经常食用菜油、BMI较低、一级亲属肺癌家族史，而保护因素是活产次数较多和胡罗卜素。女性鳞癌的危险因素则是厨房小环境污染和一级亲属肺癌家族史，保护因素是饮茶。女性其他类型肺癌危险因素主要是油炸烹饪次数较多等，而类型不明肺癌发病的危险因素类似于腺癌。结论 不同组织学类型的非吸烟女性肺癌病因可能存在一定的差异。

Polychotomous Logistic Regression Analysis for Lung Cancer among Nonsmoking Women in Shanghai, P. R. China

Objective To explore and evaluate the risk factors of lung cancer with different histology among nonsmoking women in Shanghai by using the polychotomous logistic regression model.Methods The data sets on 504 newly cases and 601 population controls were used to estimate the odds ratios by the polychotomous logistic regression model.Results The risk factors of adenocarcinoma were air pollution in cooking room,frequent intake of rapeseed oil,lower body mass index and family history of lung cancer among their first-degree relatives.The protective factors were increasing number of live births and intake of carotene.The main risk factors for squamous cell carcinoma of the lung were air pollution in cooking room and family history of lung cancer.Tea drinking was associated with the decreasing risk.The more activities with fried cooking,the higher risk with other histological types of female lung cancer.Regarding to those for the unknown histology of female lung cancer,the potential risk factors were similar with the adenocarcinoma.Conclusion Different histological types of female lung cancer among nonsmoking women may have different risk and protective factors in urban Shanghai.