Predictive value of combined prediction model of CD64 index PCT and CD4+ T lymphocyte subsets for early infection in hypertriglyceridemic acute pancreatitis

Wei Guofeng; Zhang Hong

doi:10.3969/j.issn.1002-1949.2023.12.009

Abstract：Objective To explore the clinical value of multi-index combined prediction model in early secondary infection in the patients with hypertriglyceridemic acute pancreatitis(HTG-AP).Methods 248 cases with HTG-AP admitted to the First Affiliated Hospital of Anhui Medical University,the Hefei Second People's Hospital from January 2021 to May 2023 were divided into infection group and non-infection group according to infection.General data and infection indicators in two groups were compared and analyzed.The independent factors of secondary infection in HTG-AP patients were screened by univariate analysis and binary Logistic regression analysis in order to establish a multi-index combined prediction model.The receiver operating characteristic(ROC)curve and Hosmer-Lemeshow goodness-of-fit test were used to evaluate the discrimination and calibration of the model.The decision curve analysis was used to evaluate the model's clinical practicality.Results The 248 HTG-AP patients were divided into infected、group(70 cases)and non-infected group(178 cases).There were significant differences in fasting blood glucose and infection-related indexes such as neutrophil percentage,CRP,CD4+T lymphocyte subsets,PCT and CD64 index between the two groups(P＜0.05).CD64 index,PCT and CD4+T lymphocyte subsets were independent predictors of HTG-AP early secondary infection.The value of the combined prediction model(AUC = 0.865)in early diagnosis of HTG-AP secondary infection was better than that of a single index,with higher sensitivity(78.5%)and specificity(87.7%),positive predictive value(71.83%)and negative predictive value(91.08%).The calibration chart showed that the prediction results of the joint prediction model were in high agreement with the clinical observation results,and the AUC value of the combined prediction model was 0.865 by the internal verification of Bootstrap,and its prediction efficiency curve was in good agreement with the actual clinical curve.When the threshold probability of DCA analysis was 12%-88%,the secondary infection of HTG-AP can be predicted early.Conclusions The multi-index combined prediction model based on CD64 index,PCT and CD4+ T lymphocyte subsets can early diagnose HTG-AP secondary infection.