慢性乙型肝炎免疫耐受期患者显著肝损伤的列线图模型及其预测价值分析

Value of a nomogram model in predicting significant liver injury in patients with immune-tolerant phase chronic hepatitis B

目的分析慢性HBV感染免疫耐受期(IT-CHB)患者显著肝损伤的高危因素并建立列线图预测模型。方法回顾性分析2002年8月—2017年12月在解放军总医院第五医学中心接受肝活检的382例慢性HBV感染者的资料,按照肝组织是否存在显著肝损伤分为2组,显著肝损伤组(≥G2或S2,n=82)和非显著肝损伤组(n=300)。正态分布的计量数据2组间比较采用独立样本t检验;非正态分布数据2组间比较采用Mann-Whitney U检验;多组比较采用Kruskal-Wallis H检验;计数资料2组间比较采用χ2检验。相关性分析采用Spearman秩相关。采用单/多因素logistic回归方法筛选高危因素,并建立列线图模型,用C-指数、ROC曲线、校准曲线以及Bootstrap法来评价列线图的区分度及校准度。结果2组年龄、HBV DNA载量、ALT、AST、PLT比较差异均有统计学意义(t=-7.071,Z值别为-4.924、-3.693、-6.945、-0.585、-5.723,P值均<0.001)。Logistic回归分析显示年龄(OR=1.074,95%CI:1.043-1.107,P<0.001),HBV DNA载量(OR=0.442,95%CI:0.314-0.624,P<0.001),AST(OR=1.096,95%CI:1.051-1.142,P<0.001),PLT(OR=0.992,95%CI:0.986-0.998,P=0.006)是显著肝损伤的高危因素。基于以上因素建立列线图模型,预测显著肝损伤的C-指数为0.845,并且有拟合度高的校正曲线,其ROC曲线下面积(AUC)为0.845(95%CI:0.795-0.895),显著优于APRI(AUC=0.781,95%CI:0.723-0.840)以及FIB-4(AUC=0.802,95%CI:0.746-0.859)。结论免疫耐受期具有显著肝损伤的患者比例并不少见,基于年龄、HBV DNA、AST、PLT构建的列线图模型具有良好的预测准确性,可用于个体化预测IT-CHB患者的显著肝损伤,减少肝活检,为抗病毒的精准治疗提供参考。

Objective To investigate the high-risk factors for significant liver injury in patients with immune-tolerant phase chronic hepatitis B( IT-CHB),and to establish a nomogram predictive model. Methods A retrospective analysis was performed for the data of 382 patients with chronic HBV infection who underwent liver biopsy in The Fifth Medical Center of Chinese PLA General Hospital from August 2002 to December 2017,and according to the presence or absence of significant liver injury,the patients were divided into significant liver injury group( ≥G2/S2) with 82 patients and non-significant liver injury group with 300 patients. The independent samples t-test was used for comparison of normally distributed continuous data between groups,and the Mann-Whitney U test was used for comparison of non-normally distributed continuous data between groups;the Kruskal-Wallis H test was used for comparison between multiple groups;the chi-square test was used for comparison of categorical data between groups. The Spearman rank correlation test was used to investigate correlation. Univariate and multivariate logistic regression analyses were used to screen out high-risk factors and establish a nomogram model. Concordance index( C-index),the receiver operating characteristic( ROC) curve,calibration curve,and the bootstrap method were used to evaluate the discrimination and calibration abilities of the nomogram. Results There were significant differences between the two groups in age,HBV DNA load,alanine aminotransferase,aspartate aminotransferase( AST),and platelet count( PLT)( t =-7. 071,Z =-4. 924,-3. 693,-6. 945,-0. 585 and-5. 723,all P < 0. 001). The logistic regression analysis showed that age( odds ratio [OR]= 1. 074,95% confidence interval [CI]:1. 043-1. 107,P < 0. 001),HBV DNA load( OR = 0. 442,95% CI: 0. 314-0. 624,P < 0. 001),AST( OR = 1. 096,95% CI: 1. 051-1. 142,P < 0. 001),and PLT( OR = 0. 992,95% CI: 0. 986-0. 998,P = 0. 006) were high-risk factors for significant liver injury. The nomogram model established based on the above factors had a C-index of 0. 845 in predicting significant liver injury and had a well-fitted calibration curve,with an area under the ROC curve( AUC) of 0. 845( 95% CI: 0. 795-0. 895),which was significantly better than aspartate aminotransferase-to-platelet ratio index( AUC = 0. 781,95% CI: 0. 723-0. 840) and fibrosis-4( AUC = 0. 802,95% CI:0. 746-0. 859). Conclusion There is a high proportion of IT-CHB patients with significant liver injury. The nomogram model established based on age,HBV DNA,AST,and PLT has a good predictive accuracy and can be used to predict significant liver injury in IT-CHB patients individually,reduce the need for liver biopsy,and provide a reference for precise antiviral treatment.