Construction and verification of a risk prediction model for the patients with severe pneumonia complicated by disseminated intravascular coagulation

Lyu Wenyi; Chen Lingmin; Xu Yang

doi:10.3969/j.issn.1002-1949.2025.01.009

Abstract：Objective To investigate the influencing factors of severe pneumonia complicated by disseminated intravascular coagulation(DIC),thereby establishing and verifying risk prediction models based on independent influencing factors.Methods A total of 224 patients with severe pneumonia hospitalized in Yongkang Hospital of Traditional Chinese Medicine from January 2021 to January 2023 were prospectively selected and divided into coagulation group(n=55)and non-coagulation group(n=169)according to whether DIC occurred during hospitalization.Independent influencing factors were obtained by univariate analysis and multivariate Logistic regression analysis.The prediction model was constructed based on regression analysis,the corresponding nomogram was drawn with R language software,and the predictive efficiency of the model was tested by receiver operating characteristic(ROC)curve and calibration curve.In addition,96 patients admitted from February 2023 to December 2023 were selected as validation set,and ROC curve and calibration curve were drawn with validation set data to externally verify the predictive efficiency of the model.Results Univariate analysis and multivariate Logistic regression analysis showed that body temperature of the admission day,the type of pneumonia,acute physiology and chronic health evaluation Ⅱ(APACHE Ⅱ),platelet(PLT),white blood cell(WBC),prothrombin time(PT),D-dimer,fibrinogen,hypersensitive C-reactive protein(hs-CRP),tumor necrosis factor-α(TNF-α)and hyperlipidemia were independent influencing factors for the patients with severe pneumonia complicated by DIC(P＜0.05).The area under curve(AUC)of the risk prediction model based on the above 11 independent influencing factors was 0.936 and the optimal cut-off value was 0.122 with the sensitivity of 96.4％and the specificity of 85.7％,and the model has good discriminative ability.The calibration curve results showed that mean absolute error(MAE)was 0.049,and the calibration curve was close to the ideal curve,indicating that the model had good calibration performance and was relatively reliable and stable.The AUC of validation set was 0.978,the results of the calibration curve was good,indicating that the model has good external prediction efficiency.Conclusions DIC in the patients with severe pneumonia is affected by body temperature of the admission day,the type of pneumonia,coagulation function,inflammatory response and other factors.The risk prediction model based on the above 11 independent influencing factors has good predictive efficiency,and can provide the references for the prevention of DIC complications in clinic.