Supplementary Table 1 The predictive values for the onset of KD with single SNP and multi-locus models

Sensitivity, % / Specificity, % / PPV, % / NPV, %
Single SNP (Gene)*
rs1801274 (FCGR2A) / 53.6 / 55.1 / 34.4 / 76.8
rs2857151 (HLA) / 58.1 / 49.7 / 33.7 / 73.0
rs2254546 (BLK) / 68.4 / 40.9 / 33.7 / 74.7
2-locus models*,#
rs1801274+ rs2857151 / 32.1 / 77.3 / 38.3 / 72.2
rs113420705+ rs2254546 / 36.0 / 73.5 / 37.4 / 72.3
3-locus models*,
rs113420705 + rs2857151+ rs4813003 / 12.6 / 93.7 / 46.9 / 70.9

PPV: positive predictive value; NPV: negative predictive value.

*The single SNP and multi-locus models observed significant effect on KD in regression analyses.

#The frequency comparison of subjects with high-risk genotypes at both loci (See Table 3) between 358 patients with KD and 815 controls.

The frequency comparison of subjects with high-risk genotypes at all three loci (See Table 3) between 358 patients with KD and 815 controls.

Supplementary Table 2Conditional multi-variable logistic regression analyses for the associations of the SNPswith CALs

SNP (Gene) / Allele / Genotype / OR (95% CI) / P valuea(α’= 0.007)
High-risk / Low-risk
rs1801274 (FCGR2A) / A/G / AA / AG/GG / 1.915 (0.986,3.722) / 0.055
rs113420705 (CASP3) / G/A / AA/AG / GG / 1.574 (0.809,3.063) / 0.182
rs2857151 (HLA) / A/G / GG / AG/AA / 1.702 (0.823,3.520) / 0.152
rs2254546 (BLK) / A/G / GG / AG/AA / 1.343 (0.706,2.555) / 0.368
rs4813003 (CD40) / C/T / CC / CT/TT / 1.408 (0.740,2.680) / 0.297

The risk alleles were underlined.

a Multiple regression analysis in 46 KD patients with CALs and 312 KD patients without CALs.TheP values were adjusted for gender and age. The significant level was corrected with the formula of α’=α/5=0.01(5 tests, 5 variables × 1 test/variable) according to the Bonferroni method.

Supplementary Table3Conditional multi-variable logistic regression analyses for the associations of the SNPswith IVIG unresponsiveness

SNP (Gene) / Allele / Genotype / OR (95% CI) / P valuea(α’= 0.007)
High-risk / Low-risk
rs1801274 (FCGR2A) / A/G / AA / AG/GG / 1.232 (0.666,2.280) / 0.506
rs113420705 (CASP3) / G/A / AA/AG / GG / 1.083 (0.583,2.012) / 0.802
rs2857151 (HLA) / A/G / GG / AG/AA / 1.340 (0.675,2.657) / 0.403
rs2254546 (BLK) / A/G / GG / AG/AA / 1.051 (0.568,1.946) / 0.874
rs4813003 (CD40) / C/T / CC / CT/TT / 1.052 (0.574,1.926) / 0.870

The risk alleles were underlined.

a Multiple regression analysis in 51 KD patients with IVIG unresponsiveness and 307 KD patients without IVIG unresponsiveness.TheP values were adjusted for gender and age. The significant level was corrected with the formula of α’=α/5 =0.01(5 tests, 5 variables × 1 test/variable)according to the Bonferroni method.

Supplementary Table 4 Association ofthe SNPs with general characteristics in patients withKD

General characteristics / High risk groupa (n) / Low risk groupb (n) / Pvaluec
Age at onset (years,mean ± SD)
rs1801274* / 1.57 ± 1.86 (192) / 1.63 ± 1.93 (166) / 0.765
rs2857151* / 1.62 ± 1.80 (208) / 1.57 ± 1.96 (150) / 0.803
rs2254546* / 1.53± 1.78 (245) / 1.72± 2.01 (113) / 0.369
Gender
rs1801274+ rs2857151*
Male (%) / 73 (63.5) / 149 (61.3) / 0.694
Female (%) / 42 (36.5) / 94 (38.7)
rs113420705+ rs2254546*
Male (%) / 84 (65.1) / 138 (60.3) / 0.364
Female (%) / 45 (34.9) / 91 (39.7)
rs113420705 + rs2857151+ rs4813003*
Male (%) / 29 (64.4) / 193(61.7) / 0.719
Female (%) / 16 (35.6) / 120 (38.3)

aSubjects with high-risk genotypes shown in Table 3 at all loci.

bSubjects with low-risk genotype shown in Table 3 at any locus.

cStudent’s t test for the comparison of age at onset,the significant level was corrected with the formula of α’=α/3=0.017(3 tests, 3 SNPs×1test/SNP) according to the Bonferroni method.Chi-Squared test for the comparison of gender, the significant level was corrected with the formula of α’=α/3=0.017(3 tests, 3multi-locus models× 1 test/ multi-locus models) according to the Bonferroni method.

*The single SNP and multi-locus models observed significant effect on Kawasaki disease in regression analyses.

1