不同评分系统对失代偿期肝硬化患者短期和中期死亡风险预测能力的比较

Comparison of the abilities of five scoring systems to predict short-term and medium-term death risks in patients with decompensated cirrhosis

目的比较Child-Turcotte-Pugh(CTP)评分、MELD评分、MELD-Na评分、iMELD评分以及MESO指数对失代偿期肝硬化患者3、12个月死亡风险的预测价值。方法回顾性分析2014年1月1日-2014年12月31日于四川大学华西医院消化内科住院的269例失代偿期肝硬化患者的资料。计算患者入院48 h以内评估得到的上述评分分值,并对所有患者随访至1年以上。满足正态分布的计量资料组间比较若方差齐性采用t检验;不满足正态分布的计量资料组间比较采用Mann-Whitney U检验。计数资料组间比较采用χ~2检验;采用logistic回归分析法进行多因素分析、Hosmer-Lemeshow检验各模型拟合度、Kaplan-Meier生存分析法对生存率进行分析。结果 3个月内死亡患者25例,12个月内死亡患者38例;TBil、肌酐、尿素、Alb、WBC、PT、INR、血钠、入院时是否合并肝性脑病、腹水程度、入院时CTP评分、MELD评分、MELD-Na评分、iMELD评分、MESO评分、既往是否出现肝性脑病、出院后是否对因治疗等因素在12个月内死亡患者与存活患者间差异均有统计学意义(P值均<0.05);CTP评分、MELD评分、MELD-Na评分、iMELD评分以及MESO指数都是影响预后的独立危险因素(比值比分别为2.020、1.252、1.088、1.114、1.368,P值均<0.01)。对患者入院日起3个月内和12个月内死亡风险进行预测,CTP评分、MELD评分、MELD-Na评分、iMELD评分以及MESO指数中CTP评分的受试者工作特征曲线下面积(AUC)最高(3个月的AUC分别为0.823、0.789、0.775、0.770、0.796;12个月的AUC分别为0.834、0.798、0.777、0.801、0.804)。用各项评分的受试者工作特征曲线的判断值行12个月内生存分析,5种评分方法都能很好地区分患者12个月以内的累积生存率(P值均<0.05)。结论 CTP评分、MELD评分以及3个MELD相关评分都能很好地预测肝硬化患者短期和中期的死亡风险。对于中国人群,CTP评分对肝硬化患者死亡风险的预测能力可能相对较好。

Objective To compare the abilities of Child-Turcotte-Pugh（CTP） score,Model for End-Stage Liver Disease（MELD）,MELD combined with serum sodium concentration（MELD-Na）,integrated MELD（iMELD）,and MELD to SNa ratio（MESO） to predict short-term and medium-term death risks in patients with decompensated cirrhosis at months 3 and 12.Methods The records of 269 patients with decompensated cirrhosis in Department of Gastroenterology,West China Hospital,Sichuan University from January 1 to December31,2014 were retrospectively analyzed.The CTP,MELD,MELD-Na,iMELD,and MESO scores were evaluated for these patients within48 hours after admission and these patients were followed up for at least 12 months.The predictive abilities of these five scoring systems were evaluated by the area under the receiver operating characteristic curve（AUC） at months 3 and 12.Comparison of continuous data was conducted using t test（for data with normal distribution and homogeneity of variance） or Mann-Whitney U test（for data with non-normal distribution）,while comparison of categorical data was conducted using χ~2 test.Meanwhile,logistic regression analysis,Hosmer-Lemeshow goodness-of-fit test,and Kaplan-Meier survival analysis were performed.Results Twenty-five（9.29%） and thirty-eight（14.13%） of all patients died within 3 months and 12 months,respectively.There were significant differences between the patients who died and survived within 12 months in total bilirubin,creatinine,urea,albumin,white blood cell count,prothrombin time,international normalized ratio,serum sodium,the presence or absence of hepatic encephalopathy on admission,degree of ascites,the CTP,MELD,MELD-Na,iMELD,and MESO scores on admission,history of hepatic encephalopathy,and treatments for these factors after discharge（all P〈0.05）.The multivariate logistic regression analysis revealed that all five scores were independent prognostic factors for the patients（odds ratios：CTP=2.020,MELD=1.252,MELD-Na=1.088,iMELD=1.114,MESO=1.368;all P〈0.01）.At month 3,CTP score had the highest AUC（0.823）,followed by MESO（0.796）,MELD（0.789）,MELD-Na（0.775）,and iMELD（0.770） scores.At month 12,the AUC values for CTP,MELD,MELD-Na,iMELD,and MESO were 0.834,0.798,0777,0.801,and 0.804,respectively.Therefore,CTP score was best in predicting the death risks within 3 and 12 months.The Kaplan-Meier curves showed that these five scoring methods clearly distinguished 12-month cumulative survival rates for these patients（all P〈0.05）.Conclusion CTP score,MELD score,and three MELD-related scores can reliably predict both short-term and medium-term death risks in patients with decompensated cirrhosis.Moreover,CTP score may have a better ability to predict death risks in patients with decompensated cirrhosis in China.