Risk Factors Analysis And Establishment Of Scoring Model For Immunoglobulin Resistant Kawasaki Disease

Objective:To investigate the risk factors of immunoglobulin resistance in children with Kawasaki disease(KD)and select effective predictive indicators and predictive system assessments before initial treatment.Methods:1.A retrospective analysis was performed on KD patients admitted to Children’s Hospital of Soochow University from April 2017 to March 2019.It was divided into two groups(IVIG-responsive and IVIG-resistant)according to their response to IVIG.IVIG resistant patients were divided into IVIG-partial responsive group and IVIG non-responsive group.2.Using multivariate analysis,we identified risk factors for demographic,clinical,and laboratory parameters of IVIG-responsive and IVIG-resistant patients.3.Logistic regression analysis was performed on the indicators with significant differences in the univariate analysis between the IVIG-responsive group,the IVIG-partial responsive group and the IVIG non-responsive group to construct a suitable risk prediction model.Result:1.A total of 540 KD patients were included in the study,and 66 patients were IVIG resistance.There was no statistical significance in gender,age,and season of onset between the IVIG-responsive and IVIG-resistant group.2.The time of onset and initial IVIG use in IVIG-resistant patients were significantly lower than those in IVIG-responsive patients(P<0.05).The incidence of coronary artery lesions(CALs)and digestive,respiratory,urinary,musculoskeletal,and central nervous system damage of IVIG-resistant patients was significantly higher than that of IVIG-responsive patients(P<0.05).3.IVIG-resistant patients had higher percentage of neutrophils(N%),C-reactive protein(CRP),procalcitonin(PCT),alanine aminotransferase(ALT),aspartate aminotransferase(AST),serum Bilirubin(TB),serum ferritin(SF),percentage of CD19+cell(P<0.05),lower platelet count(PLT),serum cholesterol(TC),serum albumin(ALB),serum globulin(GLB),serum sodium(Na+),serum potassium(K+),C3,C4,percentage of CD3+cell,percentage of CD3+CD4+cell,percentage of CD3+CD8+cell(P<0.05).4.IVIG-resistant patients had higher N%,TB and lower PLT and Na+when treated before day 5 of illness.5.Logistic multivariate regression analysis showed that PLT,N%,TB,and TC were independent risk factors for IVIG resistance in KD.These four variableswere used to construct a four-variable logistic regression model.The cut-off points of each variable are PLT≤300×109/L,N%≥75%,TB≥9 umol/L,TC≤3 mmol/L,the scores are taken as 2,2,1,and 2.Using a cut-off point of 4 and more with this prediction score,the area under the ROC curve is 0.847(95%CI:0.790-0.904),and the sensitivity and specificity are78.8%and 81.0%.6.In the IVIG resistance group,there were 37 IVIG-partial responsive patients and 29 IVIG non-responsive patients.The scoring model predicts that the area under the ROC curve of IVIG non-responsive patients in the IVIG resistance group is 0.652(95%CI:0.520-0.784),with specificity only was 54.1%.Comparing the clinical parameters of the two,54.1%of IVIG-partial responsive patients have at least 2 systemic impairments compared to 82.8%of IVIG non-responsive patients(P<0.05).Conclusion:1.Age,gender,season of onset may not predict IVIG resistance.IVIG-resistant patients havehigher risks of CALs and other systemic damage.2.Laboratory parameters such as PLT,N%,CRP,PCT,ALT,AST,TB,TC,ALB,GLB,Na+,K+,SF are of great significance in predicting IVIG resistance in KD patients.3.Logistic multivariate regression analysis showed that PLT,N%,TB,and TC were independent risk factors for IVIG resistance in KD.These four variableswere used to construct a simple predictive scoring model with high sensitivity and specificity.