传染病信息

目的 探讨血清缺氧诱导因子1α(hypoxia-inducible factor-1α,HIF-1α)、T细胞免疫球蛋白黏蛋白分子-3 (T-cell immunoglobulin and mucin-domain-3,Tim-3)、壳多糖酶3样蛋白1 (chitinase-3-like protein1,CHI3L1)在慢性乙型肝炎(chronic hepatitis B,CHB)进展至乙型肝炎肝硬化(hepatitis B-related liver cirrhosis,HBLC)的表达变化情况,并建立预测模型。方法 选取邯郸市第一医院2021年8月至2024年1月收治的210例CHB患者作为研究对象,随访12～40个月,中位随访时间为28个月,根据是否发生HBLC分为HBLC组与非HBLC组,比较2组基线资料、血清HIF-1α、Tim-3、CHI3L1水平,分析血清HIF-1α、Tim-3、CHI3L1水平与病情指标的相关性,分析血清HIF-1α、Tim-3、CHI3L1对CHB进展至HBLC的影响,采用受试者工作特征(receiver operating characteristic,ROC)曲线、曲线下面积(area under the curve,AUC)、校正曲线评价血清HIF-1α、Tim-3、CHI3L1对CHB进展至HBLC的预测价值,并进行泛化能力的外部验证评价。结果 210例CHB患者随访期间失访5例,共205例完成随访,其中168例未进展至HBLC,37例进展至HBLC。2组CHB病程、肝纤维化分期、HBV DNA载量、丙氨酸氨基转移酶(alanine aminotransferase,ALT)、天门冬氨酸氨基转移酶(aspartate aminotransferase,AST)水平、抗病毒治疗依从性比较,差异有统计学意义(均P<0.05);HBLC组血清HIF-1α、Tim-3、CHI3L1水平高于非HBLC组(均P<0.05);血清HIF-1α、Tim-3、CHI3L1与肝纤维化分期、HBV DNA载量、ALT、AST水平呈正相关(均P<0.05);COX回归分析显示,在调整其他因素前后,血清HIF-1α、Tim-3、CHI3L1均是CHB进展至HBLC的独立影响因素(均P<0.05);ROC曲线分析显示,血清HIF-1α、Tim-3、CHI3L1预测CHB进展至HBLC的AUC为0.754、0.741、0.771,最佳截断值为4.05 pg/mL、16.22%、116.20 ng/mL,敏感度为86.49%、81.08%、78.38%,特异度为60.71%、64.88%、67.26%。建立血清HIF-1α、Tim-3、CHI3L1联合预测模型,公式为h(t)=-0.276+0.350×HIF-1α+0.374×Tim-3+0.430×CHI3L1,ROC曲线分析显示,该模型预测CHB进展至HBLC的AUC为0.904,最佳截断值为1.22,敏感度为83.78%,特异度为79.76%,预测价值较各指标单一预测价值显著提升(Z=3.157、3.557、3.346,P=0.002、0.000、0.001);且经由1 000次Bootstrap法抽样绘制校正曲线显示,血清HIF-1α、Tim-3、CHI3L1联合预测模型校准度为0.857,C-index为0.829,提示该模型预测能力较好;另选取同期90例CHB患者作为外部验证数据集,随访12～40个月,中位随访时间为28个月,失访1例,发生HBLC 12例,未发生HBLC 77例。ROC曲线分析显示,血清HIF-1α、Tim-3、CHI3L1联合预测模型在外部验证数据集中的AUC为0.900,敏感度为83.33%,特异度为89.61%,校正曲线分析显示,血清HIF-1α、Tim-3、CHI3L1联合预测模型在外部验证数据集中的AUC校准度为0.858,C-index为0.827,提示该预测模型具有较好的泛化能力。结论 血清HIF-1α、Tim-3、CHI3L1在CHB进展至HBLC中呈异常高表达,与HBLC发生风险密切相关,可作为HBLC独立预测因子,且联合预测模型的预测价值和泛化能力较好。

Objective To investigate the changes in serum levels of hypoxia-inducible factor-1α(HIF-1α), T-cell immunoglobulin and mucin-domain-3(Tim-3), and chitinase 3-like protein 1(CHI3L1) during the progression from chronic hepatitis B(CHB) to hepatitis B-related liver cirrhosis(HBLC), and to establish a predictive model. Methods A total of 210 CHB patients admitted to Handan First Hospital from August 2021 to January 2024 were enrolled as study participants. All patients were followed up for 12～40 months with a median follow-up duration of 28 months. Patients were divided into the HBLC group and the non-HBLC group based on whether HBLC occurred during follow-up. Baseline data and serum levels of HIF-1α, Tim-3 and CHI3L1 were compared between the two groups. We analyzed the correlations between serum levels of the three indicators and liver disease parameters, and explored the effects of serum HIF-1α, Tim-3 and CHI3L1 on the progression of CHB to HBLC. Receiver operating characteristic(ROC) curves, area under the curve(AUC) and calibration curves were used to evaluate the predictive value of serum HIF-1α, Tim-3 and CHI3L1 for CHB progressing to HBLC, and external validation was performed to evaluate model generalizability. Results Among the 210 CHB patients, 5 were lost to follow-up, and 205 patients completed follow-up. Of these 205 patients, 168 did not progress to HBLC and 37 progressed to HBLC. Statistically significant differences were found between the two groups in CHB disease course, liver fibrosis stage, HBV DNA load, alanine aminotransferase(ALT), aspartate aminotransferase(AST) levels and antiviral treatment compliance(P<0.05). Serum levels of HIF-1α, Tim-3 and CHI3L1 in the HBLC group were higher than those in the non-HBLC group(P<0.05). Serum HIF-1α, Tim-3 and CHI3L1 were positively correlated with liver fibrosis stage, HBV DNA load, ALT and AST levels(P<0.05). Cox regression analysis showed that serum HIF-1α, Tim-3 and CHI3L1 were all independent influencing factors for CHB progressing to HBLC before and after adjustment of other factors(P<0.05). ROC curve analysis showed that the AUC values of serum HIF-1α, Tim-3 and CHI3L1 for predicting CHB progression to HBLC were 0.754, 0.741 and 0.771, respectively; their optimal cutoff values were 4.05 pg/mL, 16.22% and 116.20 ng/mL; the sensitivities were 86.49%, 81.08% and 78.38%, and specificities were 60.71%, 64.88% and 67.26%. A combined predictive model incorporating serum HIF-1α, Tim-3 and CHI3L1 was established, with the formula: h(t)=-0.276 + 0.350 × HIF-1α + 0.374 × Tim-3 + 0.430 × CHI3L1. ROC analysis of this combined predictive model yielded an AUC of 0.904 for predicting CHB progression to HBLC, with an optimal cutoff value of 1.22, a sensitivity of 83.78%, and a specificity of 79.76%. Its predictive value was significantly higher than that of each single indicator(Z=3.157, 3.557, 3.346, P=0.002, 0.000, 0.001). Calibration curves plotted via 1,000 Bootstrap resampling showed that the combined predictive model had a calibration degree of 0.857 and a C-index of 0.829, indicating favorable predictive capacity of the model. In addition, 90 CHB patients treated during the same period were enrolled as the external validation dataset, with a follow-up period of 12～40 months, and a median followup time of 28 months. One patient was lost to follow-up; 12 patients developed HBLC, and 77 did not. ROC curve analysis showed that the combined predictive model achieved an AUC of 0.900 in the external validation dataset, with a sensitivity of 83.33%, and a specificity of 89.61%. Calibration curve analysis showed that the combined predictive model had a calibration degree of 0.858 and a C-index of 0.827 in the external validation dataset, which suggested that the predictive model possessed good generalizability. Conclusion Serum HIF-1α, Tim-3 and CHI3L1 are abnormally elevated during the progression from CHB to HBLC, and their levels are closely related to the risk of HBLC. The three indicators can serve as independent predictors for HBLC, and the combined predictive model exhibits satisfactory predictive value and good generalizability.