[A study on treatment timing selection and short-term efficacy prediction with changes in cytokine levels before and after non-biological artificial liver treatment in acute-on-chronic liver failure].

Objective: To investigate the efficacy and diagnostic accuracy of changes in cytokine levels before and after non-biological artificial liver (referred to as ABL) treatment in patients with acute-on-chronic liver failure (ACLF) in order to establish a basis for treatment timing selection and short-term (28d) prognosis. Methods: 90 cases diagnosed with ACLF were selected and divided into a group receiving artificial liver treatment (45 cases) and a group not receiving artificial liver treatment (45 cases). Age, gender, first routine blood test after admission, liver and kidney function, and procalcitonin (PCT) of the two groups were collected. The 28-day survival of the two groups was followed-up for survival analysis. The 45 cases who received artificial liver therapy were further divided into an improvement group and a deterioration group according to the clinical manifestations before discharge and the last laboratory examination results as the efficacy evaluation indicators. Routine blood test, coagulation function, liver and kidney function, PCT, alpha fetoprotein (AFP), β-defensin-1 (HBD-1), 12 cytokines and other indicators were analyzed and compared. A receiver operating characteristic curve (ROC curve) was used to analyze the diagnostic efficacy of the short-term (28 d) prognosis and an independent risk factors affecting the prognosis of ACLF patients. According to different data, Kaplan-Meier method, log-rant test, t-test, Mann-Whitney U test, Wilcoxon rank-sum test, χ2 test, Spearman rank correlation analysis and logistic regression analysis were used for statistical analysis. Results: The 28-day survival rate was significantly higher in ACLF patients who received artificial liver therapy than that of those who did not receive artificial liver therapy (82.2% vs. 61.0%, P<0.05). The levels of serum HBD-1, alpha interferon (IFN-α) and interleukin-5 (IL-5) after artificial liver treatment were significantly lower in ACLF patients than those before treatment (P<0.05), while liver and coagulation function were significantly improved compared with those before treatment (P<0.05), and there was no statistically significant difference in other serological indexes before and after treatment (P＞0.05). Before artificial liver treatment, serum HBD-1 and INF-α levels were significantly lower in the ACLF improvement group than in the deterioration group (P<0.05) and were positively correlated with the patients' prognosis (deteriorating) (r=0.591, 0.427, P<0.001, 0.008). The level of AFP was significantly higher in the improved ACLF group than that in the deterioration group (P<0.05), and was negatively correlated with the prognosis (deteriorating) of the patients (r=-0.557, P<0.001). Univariate logistic regression analysis showed that HBD-1, IFN-α and AFP were independent risk factors for the prognosis of ACLF patients (P=0.001, 0.043, and 0.036, respectively), and that higher HBD-1 and IFN-α levels were associated with lower AFP level and a deteriorating prognosis. The area under the curve (AUC) of HBD-1, IFN-α, and AFP for short-term (28d) prognostic and diagnostic efficacy of ACLF patients was 0.883, 0.763, and 0.843, respectively, and the sensitivity and specificty was 0.75, 0.75, and 0.72, and 0.84, 0.80, and 0.83, respectively. The combination of HBD-1 and AFP had further improved the diagnostic efficiency of short-term prognosis of ACLF patients (AUC=0.960, sensitivity and specificity: 0.909 and 0.880 respectively). The combination of HBD-1+IFN-α+AFP had the highest diagnostic performance, with an AUC of 0.989, sensitivity of 0.900, and specificity of 0.947. Conclusion: Artificial liver therapy can effectively improve the clinical symptoms and liver and coagulation function of patients with ACLF; remove cytokines such as HBD-1, IFN-α, and IL-5 in patients with liver failure; delay or reverse the progression of the disease; and improve the survival rate of patients. HBD-1, IFN-α, and AFP are independent risk factors affecting the prognosis of ACLF patients, which can be used as biological indicators for evaluating the short-term prognosis of ACLF patients. The higher the level of HBD-1 and/or IFN-α, the higher the risk of disease deterioration. Therefore, artificial liver therapy should be started as soon as possible after the exclusion of infection. In diagnosing the prognosis of ACLF, HBD-1 has higher sensitivity and specificity than IFN-α and AFP, and its diagnostic efficiency is greatest when combined with IFN-α and AFP.目的： 分析慢加急性肝衰竭（ACLF）患者非生物型人工肝（简称人工肝）治疗的疗效及其治疗前后细胞因子水平变化，探讨细胞因子水平在ACLF患者短期（28 d）预后中的诊断效能并为人工肝治疗时机的选择提供依据。 方法： 选择诊断为ACLF患者90例，分为接受人工肝治疗组（45例）与未接受人工肝治疗组（45例），收集2组患者的年龄、性别、入院后第1次血常规、肝肾功能、降钙素原（PCT），追踪2组患者28 d生存情况，进行生存分析；接受人工肝治疗的45例患者根据出院前临床表现及最后一次实验室检查结果作为疗效判断指标，进一步分为好转组（24例）与恶化组（21例），收集患者第1次人工肝治疗前后血常规、凝血功能、肝肾功能、PCT、甲胎蛋白（AFP）、人β防御素-1（HBD-1）、12项细胞因子等指标，进行分析和比较；确定影响ACLF患者预后的独立危险因素，采用受试者操作特征曲线（ROC曲线）分析其对患者短期（28 d）预后的诊断效能。据资料不同用Kaplan-Meier法、log-rant检验、t检验、Wilcoxon秩和检验、Mann-Whitney U秩和检验、χ2检验、Spearman秩相关分析、logistic回归分析进行统计学分析。 结果： 接受人工肝治疗的ACLF患者28 d生存率明显高于未接受人工肝治疗组（82.2%对比61.0%，P<0.05）。经人工肝治疗后，ACLF患者的血清HBD-1、α干扰素（IFN-α）和白细胞介素-5（IL-5）水平较治疗前明显下降（P<0.05），肝功能及凝血功能较治疗前明显改善（P<0.05）；其余血清学指标治疗前后差异无统计学意义（P>0.05）。人工肝治疗前，ACLF好转组患者血清HBD-1和INF-α水平明显低于恶化组（P<0.05），与患者预后（恶化）呈正相关（r值分别为0.591、0.427，P值分别为<0.001、0.008）；ACLF好转组AFP水平明显高于恶化组（P<0.05），与患者预后（恶化）呈负相关（r=-0.557，P<0.001）。单因素logistic回归分析提示：HBD-1、IFN-α和AFP为ACLF患者预后的独立危险因素（P值分别为0.001、0.043、0.036），HBD-1、IFN-α水平越高，AFP水平越低，预后越差。ROC曲线评估HBD-1、IFN-α及AFP对ACLF患者短期（28 d）预后诊断效能的曲线下面积（AUC）分别为0.883、0.763、0.843，灵敏度分别为0.75、0.75、0.72，特异度分别为0.84、0.80、0.83；HBD-1与AFP联合时对ACLF患者短期预后的诊断效能进一步提高（AUC为0.960，灵敏度为0.909，特异度为0.880）；HBD-1+IFN-α+AFP三者联合的诊断效能最大，AUC为0.989，灵敏度为0.900，特异度为0.947。 结论： 人工肝治疗可有效改善ACLF患者临床症状、肝功能及凝血功能，清除肝衰竭患者体内HBD-1、IFN-α和IL-5等细胞因子，延缓/逆转疾病进展，提高患者生存率；HBD-1、IFN-α和AFP为影响ACLF患者预后的独立危险因素，可作为ACLF患者判断短期预后的生物学指标，HBD-1和/或IFN-α水平越高，疾病恶化风险越高，在排除感染后应尽早启动人工肝治疗；HBD-1诊断ACLF预后的灵敏度、特异度优于IFN-α和AFP，当HBD-1与IFN-α及AFP 3个指标联合应用时，诊断效能最高。.