基于血液学指标经内镜逆行胰胆管造影术后胰腺炎及高淀粉酶血症列线图预测模型的构建

To construct a nomogram prediction model for postoperative pancreatitis and postoperative hyperamylasemia after endoscopic retrograde cholangiopancreatography based on hematological indicators

目的 回顾性研究行经内镜逆行胰胆管造影(ERCP)患者术后发生胰腺炎(PEP)及高淀粉酶血症(PEH)的危险因素,构建列线图预测模型并评估其预测效果。方法 根据ERCP术后淀粉酶水平将431例患者分为淀粉酶正常组216例和淀粉酶升高组215例。比较两组患者一般临床资料和实验室检查指标。采用单因素及多因素logistic回归分析评估ERCP术后PEP及PEH发生的独立危险因素。构建ERCP术后血清淀粉酶升高、PEP及PEH发生风险的列线图预测模型,采用校准曲线和受试者工作特征(ROC)曲线评估其预测性能。结果 PEP的发生率为9.28%(40/431)。淀粉酶升高组患者中性粒细胞(NE)计数、C反应蛋白(CRP)、直接胆红素(DBil)、AST、ALT水平及系统性免疫炎症指数(SII)、中性粒细胞/淋巴细胞比值(NLR)、单核细胞/淋巴细胞比值(MLR)、血小板/淋巴细胞比值(PLR)、中性粒细胞/单核细胞比值(NMR)均高于淀粉酶正常组,年龄、淋巴细胞(LY)计数、纤维蛋白原(FIB)水平均低于淀粉酶正常组(P<0.05)。多因素logistic回归分析结果显示,年龄、NE计数、SII及PLR均为ERCP术后淀粉酶升高的独立危险因素,NE计数、DBil及NLR均为ERCP术后PEP发生的独立危险因素,年龄及PLR均为ERCP术后PEH发生的独立危险因素(P<0.05)。预测ERCP术后淀粉酶升高的列线图模型内部验证的一致性指数为0.791(95%CI 0.752~0.832),ROC曲线下面积(AUC)为0.791(95%CI 0.749~0.833);预测ERCP术后PEP和PEH发生的列线图模型内部验证的一致性指数分别为0.815(95%CI 0.756~0.892)和0.701(95%CI 0.633~0.764),AUC分别为0.815(95%CI 0.750~0.880)和0.701(95%CI 0.652~0.750),均提示模型具有较高的预测效能。结论 基于血液学指标的列线图模型可有效预测ERCP术后患者PEP及PEH的发生,进而优化患者临床个体化治疗方案。

Objective To retrospectively study the risk factors of postoperative pancreatitis(PEP)and postoperative hyperamylasemia(PEH)in patients undergoing endoscopic retrograde cholangiopancreatography(ERCP),and to construct a nomogram prediction model and to evaluate its predictive effect.Methods According to the level of amylase after ERCP,431 patients were divided into normal amylase group(216 cases)and elevated amylase group(215 cases).General clinical data and laboratory examination indexes of two groups were compared.Univariate and multivariate logistic regression analysis were used to evaluate the independent risk factors for PEP and PEH after ERCP.A nomogram prediction model for the risk of serum amylase elevation,PEP,and PEH after ERCP was constructed,and the calibration curve and receiver operating characteristic(ROC)were respectively used to evaluate the predictive performance.Results The incidence of PEP was 9.28%(40/431).Levels of neutrophils(NE)count,c-reactive protein(CRP),direct bilirubin(DBil),AST,ALT and systemic immune inflammation index(SII),neutrophil/lymphocyte ratio(NLR),monocyte/lymphocyte ratio(MLR),platelet/lymphocyte ratio(PLR)and neutrophil/monocyte ratio(NMR)in elevated amylase group were higher than those in normal amylase group,and age,levels of lymphocyte(LY)count and fibrinogen(FIB)were lower than those in normal amylase group(P<0.05).Multivariate logistic regression analysis showed that age,NE count,SII and PLR were independent risk factors for postoperative amylase elevation after ERCP;NE count,DBil and NLR were independent risk factors for PEP after ERCP;age and PLR were independent influencing factors for PEH after ERCP(P<O.05).The internal consistency index of the nomogram model for predicting the increased amylase after ERCP was 0.791(95%CI 0.752-0.832),and the ROC area under the curve(AUC)was 0.791(95%CI 0.749-0.833).The internal consistency index of the nomogram model for predicting PEP and PEH after ERCP were 0.815(95%CI 0.756-0.892)and 0.701(95%CI 0.633-0.764)respectively,the AUC values were 0.815(95%CI 0.750-0.880)and 0.701(95%CI 0.652-0.750)respectively,which indicating that models had high prediction efficiency.Conclusion The nomogram model based on hematological indicators can effectively predict the occurrence of PEP and PEH in patients undergoing ERCP,and optimize the clinical individualized treatment plan for patients.