肝纤维化指标对肾功能正常糖尿病患者经皮冠状动脉成形介入治疗术后造影剂肾病的预测价值

Predictive value of liver fibrosis index in contrast nephropathy of diabetes patients with normal renal function after percutaneous coronary intervention

目的 探讨肝纤维化指标对肾功能正常的糖尿病患者经皮冠状动脉成形介入治疗(percutaneous coronary intervention,PCI)术后造影剂肾病(contrast-induced nephropathy,CIN)的预测价值。方法 选取2020年1月至2022年6月在自贡市第一人民医院就诊的155例肾功能正常糖尿病合并冠心病患者作为研究对象,男109例(70.3%),女46例(29.7%),年龄(53.2±6.5)岁,年龄范围37~65岁,因病情需要均接受PCI,记录患者PCI术后72 h内CIN发生情况。根据CIN发生情况,将患者分为CIN组和N-CIN组,比较两组患者一般临床资料以及肝纤维化相关指标,多因素Logistic回归分析评估肝纤维化指标预测糖尿病患者PCI术后发生CIN的价值,绘制肝纤维化指标预测糖尿病患者PCI术后发生CIN的受试者工作特征曲线以及预测发生CIN净效益的决策曲线。结果 单因素分析显示,与N-CIN组比较,CIN组患者红细胞分布宽度[(13.3±1.6)%比(10.1±1.7)%,t=2.667,P=0.028]、血小板平均体积[(12.9±2.3)fL比(10.2±1.8)fL,t=2.458,P=0.034]、C反应蛋白[(7.4±1.4)mg/L比(5.7±1.2)mg/L,t=2.896,P=0.021]、Ⅳ型胶原(collagenⅣ,C-Ⅳ)[(65.5±7.4)μg/L比(52.3±6.9)μg/L,t=6.778,P<0.001]、透明质酸(hyaluronic acid,HA)[(89.4±7.7)μg/L比(78.2±7.2)μg/L,t=4.332,P<0.001]以及层黏蛋白(laminin,LN)[(93.3±8.2)μg/L比(81.2±8.9)μg/L, t=3.778, P<0.001]明显升高。多因素Logistic回归分析调整和校正混杂因素后,CIⅣ(OR=2.717,95%CI:1.237~4.347,P=0.004)、HA(OR=2.695,95%CI:1.932~4.989,P=0.008)以及LN(OR=2.588,95%CI:1.654~4.935,P=0.011)为糖尿病患者PCI术后发生CIN的独立预测因素(P<0.05)。受试者工作特征曲线分析结果显示,C-Ⅳ、HA以及LN对于糖尿病患者PCI术后发生CIN均具有一定的预测价值(P<0.05),其中C-Ⅳ预测糖尿病患者PCI术后发生CIN的曲线下面积(area under the curve,AUC)为0.932(0.887~0.966),最佳诊断截点为63.3μg/L,HA的AUC为0.846(0.774~0.889),最佳诊断截点为85.4μg/L,LN的AUC为0.812(0.745~0.857),最佳诊断截点为90.2μg/L,而3个指标联合预测的价值最高,AUC为0.963(0.921~0.992)。决策曲线分析结果显示,在大多数合理阈值概率范围内,3个独立预测指标预测糖尿病患者PCI术后是否发生CIN均具有良好的净获益,并且联合预测的净效益高于单一指标的净效益。结论 术前C-Ⅳ、HA以及LN为糖尿病患者PCI术后发生CIN的独立预测因素,并且三者联合预测的效能和净效益高于单一指标。

Objective To explore the predictive value of liver fibrosis index in contrast-induced nephropathy(CIN)of diabetics with normal renal function after percutaneous coronary angioplasty and interventional therapy(PCI).Methods From January 2020 to June 2022,155 diabetics of coronary heart disease(CHD)with normal renal function were selected as research subjects.There were 109 males(70.3%)and 46 females(29.7%)with an average age of(53.2±6.5)(37-65)years.PCI was performed as needed.The occurrence of CIN within 72h post-PCI was recorded.According to the presence or absence of CIN,they were assigned into two groups of CIN and N-CIN.General profiles and liver fibrosis related parameters of two groups were compared.Multivariate logistic regression analysis was employed for evaluating the value of liver fibrosis index in predicting CIN of diabetics post-PCI.Receiver operating characteristic(ROC)curve of liver fibrosis index was plotted in predicting CIN of diabetics post-PCI and decision curve for predicting the net benefit of CIN.Results Univariate analysis indicated that erythro-cyte distribution width[(13.3±1.6)%vs(10.1±1.7)%,t=2.667,P=0.028],average platelet volume[(12.9±2.3)vs(10.2±1.8)fL,t=2.458,P=0.034],C-reactive protein(CRP)[(7.4±1.4)mg/L vs(5.7±1.2)mg/L,t=2.896,P=0.021],typeⅣcollagen(C-IV)[(65.5±7.4)μg/L vs(52.3±6.9)μg/L,t=6.778,P<0.001],hyaluronic acid(HA)[(89.4±7.7)μg/L vs(78.2±7.2)μg/L,t=4.332,P<0.001]and laminin(LN)[(93.3±8.2)μg/L vs(81.2±8.9)μg/L,t=3.778,P<0.001]were significintly higher in CIN group than those in N-CIN group.After adjusting for confounding factors,multivariable logistic re-gression analysis revealed that C-Ⅳ(OR=2.717,95%CI:1.237-4.347,P=0.004),HA(OR=2.695,95%CI:1.932-4.989,P=0.008)and LN(OR=2.588,95%CI:1.654-4.935,P=0.011)were independent predictors of diabetics post-PCI(P<0.05).ROC analysis showed that C-Ⅳ,HA and LN had some predic-tive values for CIN of diabetics post-PCI(P<0.05).The area under the curve(AUC)of C-Ⅳpredicting CIN of diabetics post-PCI was 0.932(0.887-0.966)with an optimal diagnostic cut-off point of 63.3μg/L,AUC of HA 0.846(0.774-0.889)with an optimal diagnostic cut-off point of 85.4μg/L and AUC of LN 0.812(0.745-0.857)with an optimal diagnostic cut-off point of 90.2μg/L.The value of joint prediction of three indicators was the highest with AUC of 0.963(0.921-0.992).The results of decision curve analysis indicated that within the most reasonable range of threshold probability,these three independent predictors had an excellent net benefit in predicting the occurrence of CIN and the net benefit of joint prediction was higher than that of a single indicator.Conclusion Preoperative levels of C-Ⅳ,HA and LN are indepen-dent predictors of CIN in diabetics post-PCI.And the combined prediction efficiency and net benefit of three indicators are higher than that of a single index.