Analysis Of IVIG Non-response Prediction Of Kawasaki Disease Based On Genetic And Laboratory Tests And Comparison Of Five Prediction Models

Objective: To collect clinical data from children with Kawasaki disease(KD)attending a tertiary general hospital in central and western Inner Mongolia and to analyze the laboratory test results associated with children with intravenous immunoglobulin(IVIG)nonresponsive KD.To investigate the correlation of IL1-B gene single nucleotide polymorphism(SNP)loci rs16944 and rs1143627,SMAD5 gene SNP loci rs10056474 and rs746994,and PLA2G7 gene SNP loci rs76863441 with IVIG non-responsive KD in the central and western regions of Inner Mongolia.To verify whether the Egami model,Kobayashi model,Capital Institute of Pediatrics2021 model,Wuhan Children’s Hospital-Beida Hospital model and Lin Xiaonan model are appropriate for children with KD in the central and western regions of Inner Mongolia and to assess their diagnostic efficacy.Methods: Using a case-control study,clinical data and blood samples of 130 children with KD were retrospectively analyzed.Children with KD were divided into IVIG non-responder group and IVIG responder group,peripheral blood was collected,and IL1-B gene rs16944,rs1143627;SMAD5 gene rs10056474,rs746994;PLA2G7 gene rs76863441 using polymerase chain reaction(PCR)technique and first-generation sequencing sanger method Sequencing was performed.Demographic characteristics,laboratory tests and the above SNP loci were analyzed separately for children with IVIG non-responder KD and IVIG responder KD using SPSS27.0 and R(4.2.2)software;five scoring systems were used for evaluation: Egami model,Kobayashi model,Capital Institute of Pediatrics 2021 model,Wuhan Children’s Hospital-Beida Hospital model and Lin Xiaonan model,ROC curve and DCA curve to determine the predictive effect.Reults: For age at month,white blood cell count(WBC),neutrophil count(NE),lymphocyte count(LY),neutrophil percentage(N%),C-reactive protein(CRP),erythrocyte sedimentation rate(ESR),hemoglobin(HB),platelets(PLT),mean red blood cell volume(MCV),serum Na+ concentration(Na),serum albumin(ALB),venous nutrition index(PNI),etc.were successively performed as the variables examined in univariate and multifactorial analyses.The number of fever days was finally determined to be an independent risk factor for IVIG nonresponsive KD(P=0.013).The IL1-B gene rs16944 was distributed TT,CC/CT according to IVIG non-responder genotype difference was not statistically significant(X 2=0.1,P=0.752).rs1143627 was distributed CC,TT/CT according to IVIG non-responder genotype difference was not statistically significant(X2= 0.1,P=0.752);SMAD5 gene rs10056474 according to IVIG non-responder genotype distribution GG,CC/CG differences were not statistically significant(X 2=0.031,P=0.861);rs746994 according to IVIG nonresponder genotype distribution GG,AG/AA differences were not statistically significant(X2=0.098,P=0.979).;PLA2G7 gene rs76863441 according to IVIG nonresponder genotype distribution GG,GT difference was not statistically significant(X2=1.022,P=0.507).The area under the ROC curve(AUC)for the Egami model was 0.591(95% CI:0.442-0.740)with a cut-off value of 0.5 at Yordon index = 0.227,sensitivity 65.6%and specificity 57.1% for prediction of IVIG non-responder KD using the above scoring system;the area under the ROC curve(AUC)for the Kobayashi model The area under the ROC curve(AUC)for the Kobayashi model was 0.619(95% CI:0.475-0.763),with a cut-off value of 0.5 and a Yordon index = 0.451,sensitivity77.6%,and specificity 50%;the area under the ROC curve(AUC)for the Capital Institute of Pediatrics 2021 model was 0.701(95% CI: 0.545-0.875),with a cut-off value of 6.75 and a Yordon index = 0.451,sensitivity 72.1%.0.451,with 72.4%sensitivity and 66.7% specificity;the area under the ROC curve(AUC)of Beida Hospital-Wuhan Hospital model was 0.592(95% CI: 0.463-0.720),and the Jorden index=0.185 when the cut-off value was 0.5,with 54.5% sensitivity and 64%specificity;the area under the ROC curve(AUC)of Lin Xiaonan model was 0.672(95% CI: 0.499-0.844),with a cut-off value of 2.75 and a Yordon index=0.303,sensitivity 58.3% and specificity 72%.Clinical decision analysis curves were plotted,and the net benefit of the Lin Xiaonan model was higher than other scoring systems in the Pt range of 0.2-0.7.Conclusions:1.days of fever are independent predictors of IVIG non-responsive KD.2.the current limited studies suggest no potential association of IL1-B gene SNP loci rs16944 and rs1143627,SMAD5 gene SNP loci rs10056474 and rs746994,and PLA2G7 gene SNP loci rs76863441 in IVIG nonresponsive KD in central and western Inner Mongolia.3.the Lin Xiaonan model was more suitable for children with IVIG non-responsive KD in central and western Inner Mongolia compared to the other 4 models.4.It was found that the above prediction models still had room for improvement in the prediction of IVIG nonresponse,and further search for specific IVIG nonresponse markers including genetic loci into the prediction models could help improve the prediction efficacy.