Analysis Of Risk Factors And Establishment Of Clinical Prediction Model For IVIG Non-response In Kawasaki Disease

Research background and purpose:The incidence rate of Kawasaki disease(KD)is in Crsing every year,and now it has become the primary cause of acquired heart disease in children.The choice of treatment plan mainly depends on high-dose intravenous immunoglobulin(IVIG).Clinical data show that about 10-20% of children with KD are still insensitive to singledose IVIG treatment,and this part of children with KD have a significantly higher risk of coronary artery disease(CAL),which seriously affects the quality of life and longterm prognosis of children with KD.At the same time,the incidence of IVIG unresponsive KD is in Crsing.Many literatures have reported that early identification of IVIG non-reactive KD and intensification of initial treatment are essential to reduce the risk of CAL.However,current research has found that the clinical prediction effect of the existing prediction scoring system is not good,and its application has obvious limitations.No model can be effectively applied to different regions and different populations.The purpose of this study was to analyze the risk factors of IVIG resistance in children with KD in Sichuan area,and on this basis to build and verify a clinical prediction model for IVIG non-responsive KD suitable for Sichuan Province.In addition,we verified and evaluated the predictive efficiency of some published predictive scoring systems applied to children in Sichuan Province.Methods:This study is a multicenter retrospective study,collecting the continuous cases of KD diagnosed from January 1,2019 to December 31,2022 and receiving IVIG treatment for the first time in 4 hospitals in Sichuan(Sichuan Province Maternity and Child Health Care Hospital,Suining Central Hospital,Affiliated Hospital of North Sichuan Medical College,Deyang People’s Hospital),and collecting the demographic characteristics,clinical manifestations and physical signs,laboratory examination and other information of the children.The children were divided into modeling group(n=656)and validation group(n= 101)according to the length of hospitalization,and each group was divided into IVIG non-response group and IVIG response group according to whether IVIG non-response occurred.Single factor analysis and multiple logistic regression analysis were used to screen out the independent risk factors of KD children without response to IVIG,and the binary logistic regression clinical prediction model was further constructed.The consistency of the model was evaluated by the Hosmer-Lemeshow test,and the clinical effectiveness of the model was evaluated by the ROC curve.Result:1.Among the 757 children with KD included in the whole population,704(704/757,93.00%)were IVIG sensitive children and 53(53/757,7.00%)were IVIG unresponsive children.There are 289 children with KD in our hospital,of which 23children(23/289,7.96%)have no response to IVIG treatment,and 266 children(266/289,92.04%)are sensitive to IVIG treatment.2.The univariate analysis of KD children in our hospital found that lip change,NEUT%,heat duration(days)at the first use of IVIG,CRP,PCT,NLR,PLR,PLT,AST,ALT,S/T,GGT,TBA,TBIL,Cr,blood sodium,blood potassium,and urinary white blood cells were the risk factors for IVIG resistance.3.Based on the single-factor analysis and multi-factor analysis of the whole sample population,four variables including PLT,RDW-CV,Cr and blood potassium concentration were found to be independent risk factors for IVIG resistance,and a regression model was established.Prediction model AUC=0.910(95% CI: 0 870-0.949),with sensitivity of 77.3% and specificity of 82.0%.The clinical data of 101 patients in the validation group were brought into the new prediction model for internal validation,with sensitivity of 88.9% and specificity of 88.0%.Conclusion:1.The clinical prediction model based on PLT,RDW-CV,Cr and blood potassium concentration can better predict the occurrence of non-reaction of IVIG in children with KD in Sichuan region than the previous prediction model.2.The seven domestic and foreign predictive scoring systems included in this study are not fully applicable to the prediction of IVIG unresponsive Kawasaki disease in Sichuan children.The new scoring system established in this study has good predictive efficiency,but its predictive efficiency in other regions and populations needs to be further confirmed.