Construction and Analysis of Logistic Regression Model of Adverse Prognostic Factors in Children with Kawasaki Disease

WANG Yan; XIONG Yajuan; LIU Weiwei

doi:10.3969/j.issn.1004-437X.2025.01.009

Abstract：Objective To investigate the adverse prognostic factors in children with Kawasaki disease,establish a logistic regression prediction model and verify it,and provide guidance for early clinical intervention.Methods A total of 160 Kawasaki disease children admitted to the 7 th People's Hospital of Zhengzhou from February 2021 to February 2023 were retrospectively selected as the study objects,and were divided into a poor prognosis group(55 cases)and a good prognosis group(105 cases)according to the prognosis 1 month after treatment.The clinical data of the two groups were compared and the influencing factors of poor prognosis were analyzed.A logistic regression prediction model was established based on the influencing factors,and the predictive efficacy of the model was evaluated by area under the curve(AUC),net reclassification index(NRI)and integrated discrimination improvement(IDI).Results Delay of intravenous immunoglobulin(IVIG)treatment,duration of fever ≥10 days,non-response to IVIG treatment,erythrocyte distribution width(RDW),serum C-reactive protein(CRP),cardiac troponin Ⅰ(cTnⅠ),intercellular adhesion molecule-1(ICAM-1),interleukin-6(IL-6)levels were independent risk factors for poor prognosis.25-hydroxy vitamin D3[25-(OH)D3]level was an independent protective factor for poor prognosis(P＜0.05).Duration of fever ≥10 days,delay in IVIG treatment,no response to IVIG treatment,RDW,CRP,cTnⅠ were used as model 1,duration of fever ≥10 days,delay in IVIG treatment,no response to IVIG treatment,RDW,CRP,cTnⅠ,IL-6,ICAM-1,25-(OH)D3 were used as model 2.The AUC of model 1 and model 2 predicting poor prognosis was 0.813 and 0.915,respectively.The AUC of model 2 was greater than that of model 1,and the prediction effect of model 2 was improved compared with model 1(P＜0.05).Conclusion Delay of IVIG treatment,duration of fever ≥10 days,non-response to IVIG treatment,RDW and serum CRP,cTnⅠ,ICAM-1 and IL-6 levels are independent risk factors for poor prognosis in children with Kawasaki disease,and 25-(OH)D3 level is independent protective factor for poor prognosis.Based on the above influencing factors,a logistic regression prediction model was established,which has certain predictive value for the occurrence of poor prognosis,and can provide reference for clinical screening of high-risk children,and help to formulate individualized prevention and treatment plans to reduce the risk of poor prognosis.