Prediction of unresponsiveness to second intravenous immunoglobulin treatment in patients with Kawasaki disease refractory to initial treatment

PURPOSE This study investigated predictors of unresponsiveness to second-line intravenous immunoglobulin (IVIG) treatment for Kawasaki disease (KD). METHODS This was a single-center analysis of the medical records of 588 patients with KD who had been admitted to Asan Medical Center between 2006 and 2014. Related clinical and laboratory data were analyzed by univariate and multivariate logistic regression analyses. RESULTS Eighty (13.6%) of the 588 patients with KD were unresponsive to the initial IVIG treatment and received a second dose. For these 80 patients, univariate analysis of the laboratory results obtained before administering the second-line IVIG treatment showed that white blood cell count, neutrophil percent, hemoglobin level, platelet count, serum protein level, albumin level, potassium level, and C-reactive protein level were significant predictors. The addition of methyl prednisolone to the second-line regimen was not associated with treatment response (odds ratio [OR], 0.871; 95% confidence interval [CI], 0.216-3.512; P=0.846). Multivariate analysis revealed serum protein level to be the only predictor of unresponsiveness to the second-line treatment (OR, 0.160; 95% CI, 0.028-0.911; P=0.039). Receiver operating characteristic curve analysis to determine predictors of unresponsiveness to the second dose of IVIG showed a sensitivity of 100% and specificity of 72% at a serum protein cutoff level of <7.15 g/dL. CONCLUSION The serum protein level of the patient prior to the second dose of IVIG is a significant predictor of unresponsiveness. The addition of methyl prednisolone to the second-line regimen produces no treatment benefit.