原发性肝癌危险因素的巢式病例对照研究

目的　应用巢式病例对照研究方法 ,探讨原发性肝癌 (HCC)的危险因素。方法　在江苏省海门市收集以队列为基础的病例对照资料 ,应用SAS软件进行两分类logistic回归分析和多分类logistic回归分析。 结果　经两分类logistic回归分析 ,发现HBsAg( )携带者和饮河沟水者患肝癌的风险较大 ,其OR值分别为 19.5 6(P <0 .0 0 1)和 4 .95 6 (P <0 .0 0 1)。交互作用分析表明 ,饮河沟水、表面抗原 乙肝、表面抗原 家族史三项进入多因素logistic回归模型 ,OR分别为 2 .4 9(P =0 .0 0 3)、7.0 92 (P <0 .0 0 1)和 4 .0 0 4 (P <0 .0 0 1)。经多分类logis tic回归分析 ,与对照组比较 ,发现河沟水对HBsAg(- )肝癌人群和HBsAg( )肝癌人群均是一危险因素 ,其在不同模型中的OR值分别为 5 .2 6 3(P =0 .0 0 0 3)、2 .174 (P =0 .0 377) ;而肝癌家族史在HBsAg( )肝癌人群中是一危险因素 ,OR =3.185 ,P <0 .0 0 1。当HBsAg( )组与HBsAg(- )组比较 ,提示饮河沟水仍然是一危险因素 ,OR =2 .373(P <0 .0 5 )。家族史未能引入这一模型。结论　HCC受多因素影响 ,尤其是HBsAg和饮河沟水是两个重要的危险因素 ,另外也提示存在一定的家族聚集现象

Nested Case-Control Study on Risk Factors of Hepatocellular Carcinoma

Purpose To explore the risk factors of hepatocellular carcinoma (HCC) with the aid of nested case control study. Methods Case control data based on cohort were collected in Haimen,Jiangsu Province,and binomial logistic regression analysis and multinomial logistic regression analysis were conducted with SAS software. Results With binary logistic regression analysis,it is found that there were excessive risks to HCC in those who were HBsAg carriers and those who drank ditch water,and those OR were 19.56 ( P <0.001) and 4.956 ( P <0.001),respectively.The result of interaction analysis showed that drinking ditch water,HBsAg*hepatitis B and HBsAg*family history could all be entered into the multinomial logistic regression model,and those OR were 2.49 ( P =0.003),7.092 ( P <0.001) and 4.004 ( P <0.001),respectively.With multinomial logistic regression analysis,ditch water,relative to the control group,was a risk factor among HCC cases with the HBsAg(-) and HCC cases with the HBsAg( ),OR were 5.263 ( P =0.000 3) and 2.174 ( P =0.037 7) in various models,while HCC family history was a risk factor only in HBsAg( ) cases,with its OR being equal to 3.185 ( P <0.001).Compared HBsAg( ) group with HBsAg(-) group,it was suggested that drinking ditch water was still a risk factor,its OR of which was 2.373 ( P <0.05),but family history was not fitted into the model. Conclusions The results suggested that HCC was influenced by various risk factors,among which HBsAg and drinking ditch water were expecially important,as well as certain familial aggression also existed.