慢性乙型重型肝炎预后影响因素Logistic回归分析

Logistic regression analysis of prognostic factors for chronic severe hepatitis B

目的探讨影响慢性乙型重型肝炎预后的因素,并建立Logistic回归方程预测预后。方法回顾性调查2006年1月—2008年12月5家医院收治的慢性乙型重型肝炎住院患者618例,采用非条件二分类反应变量Logistic回归分析,得出相应的危险因素并建立回归方程。结果 Logistic回归分析显示,高龄、TBIL升高、BUN升高、CRE升高、国际标准化比值(in-ternational normalized ratio,INR)增大、甘油三酯降低、腹水、肝性脑病、上消化道出血是影响患者预后的危险因素,而AFP升高提示预后较好(P<0.05)。回归方程预测总的正确率为78.3%。结论患者年龄、TBIL、BUN、Cr、INR、甘油三酯、AFP指标及腹水、肝性脑病、上消化道出血等并发症可作为判断慢性乙型重型肝炎预后的指标,由此建立的方程预测效果较好,对临床治疗有一定的参考价值。

Objective To investigate prognostic factors for chronic severe hepatitis B, and to establish logistic regression equation for the prediction of the prognosis. Methods A retrospective study was carried out on 618 patients with chronic severe hepatitis B treated in 5 hospitals from Jan. 2006 to Dec. 2008. Unconditioned binary response logistic regression model was used to determine the corresponding risk factors, and then a regression equation was established. Results Logistic regression analysis showed that older age, elevated levels of TBIL, BUN and CRE, increased international normalized ratio （INR）, reduced level of triglycerides, ascites, hepatic encephalopathy and upper gastrointestinal bleeding were the risk factors for the prognosis; increased AFP level indi- cated a better prognosis （P〈0.05）. The predictive accuracy rate of the logistic regression equation was 78.3%. Conclusions Age, TBIL, BUN, CRE, INR, triglycerides, AFP, ascites, hepatic encephalopathy and upper gastrointestinal bleeding can serve as prognostic indicators of chronic severe hepatitis B. The logistic regression equation based on the above indicators is preferable in predicting the prognosis of chronic severe hepatitis B, which may have reference value in clinical practice.