慢性重型肝炎近期预后的多因素分析

Multivariate analysis of prognostic factors in chronic severe viral hepatitis with Cox Model

目的研究影响慢性重型肝炎(慢重肝)近期预后的主要危险因素及其与生存状况的综合性量化关系,建立一个有实用价值的危险指数(PI)方程。方法应用多因素Cox模型生存分析统计方法,收集190例慢重肝病例106项指标,对其入院后1,2及6个月的近期预后进行分析。结果16项指标与慢重肝近期预后有关;其中主要危险因素为肝性脑病程度(HE)、凝血酶原活动度(PTA)、血尿素氮(BUN)、血清总胆红素(TB)、血清钾(K+)、白细胞总数(WBC);根据主要危险因素对近期预后影响的大小,得出的最佳危险指数方程为1个月期的方程,表达如下:PI=(71.84HE+0.29TB+10.38BUN+124.5K++5.99WBC-9.01PTA) 1/100[方程中血清钾以正常血钾或高血钾为1,低血钾为0,其他全部为原测值];由上方程计算出的危险指数与生存时间及生存状况密切相关(偏相关系数和为0.88),明显高于单因素方程。结论慢重肝近期预后影响因素众多,有6项主要危险因素。据此建立的危险指数方程对病情及预后的判断明显优于单因素分析,并易于计算。可望用于本病的病情分析、预后判断及疗效考核。

Objective To study the risk factors in determining the extent of liver damage and the prognosis of the patients with chronic severe hepatitis （CSH） and to estabhsh an equation in estimating the risk factor. Methods The clinical and laboratory data from 190 cases of CSH were evaluated by Cox model analysis. 1, 2 and 6 months survivals were determined. Resalts There were 16 factors influencing the prognosis of CSH. The main risk factors and the best equation PI a derived from 1 month was expressed as following： PI = （71.84HE ＋ 0.29TB ＋ 10.38BUN ＋ 124.53K^＋ ＋ 5.99WBC- 9.01PTA- 3.13APB） × 1/100（In the equation, HE presented hepatic encephalopathy, TB presented total serum biliruhin, BUN presented blood urine nitro, K^＋ presented potassium of serum, PTA presented prothromtin time activity, and WBC presented total white blood cell count. Excluding potassium of serum, the value of all factors in the equation was measuring value. K^ ＋ was 1 when potassium of serum was normal or higher than normal, K^＋ .was 0 when potassium of serum was lower than normal）. The PI had better correlation with survival and survival status （with the largest sum of partial correlation coefficient 0.88） than monovariate analysis. Conclusion There are many risk factors affecting the prognosis of CSH. The equation derived from 6 major risk factors can discriminate the extent of the disease better than monovariate analysis, and may be useful to divide the extent liver damage, to predict the prognosis and to evaluate the effectiveness of medical treatment.