New predictors to identify severe cases during intravenous immunoglobulin therapy in Kawasaki disease

Background The purpose of this study was to predict which patients would require plasma exchange therapy (PEX) using laboratory findings during intravenous immunoglobulin (IVIG) therapy in Kawasaki disease. Methods A retrospective, single‐center, cohort study was conducted. Of the 621 IVIG‐treated patients, 166 patients who received a second IVIG dose on the second day after initiation of IVIG were included. The endpoint was PEX. Participants were divided into two groups: 10 patients who received PEX and 156 patients who did not. The variables were C‐reactive protein (CRP), white blood cell count (WBC), and neutrophil fraction (NEUT) at three time points: pre‐initial IVIG, pre‐second IVIG, and post‐second IVIG. The change ratio (post‐IVIG value/pre‐IVIG value) of each variable, with each IVIG dose, was calculated. Receiver operating characteristic analysis determined the area under the curve (AUC) and cut‐off values. Results The variables with an AUC > 0.9 were CRP (post‐second IVIG), change ratio of CRP (CRP value post‐second IVIG/CRP value pre‐second IVIG), NEUT (pre‐second IVIG), and NEUT (post‐second IVIG). Among these, the variables with high sensitivity were CRP (post‐second IVIG) and the change ratio of CRP (second IVIG), with cut‐off values of 9.52 mg/dL and 0.99, respectively. The sensitivity and specificity of these variables were 100% and 91%, and 100% and 80%, respectively. The combined sensitivity and specificity (95% confidence intervals) of these two variables were 100% (59–100) and 94% (89–97). Conclusion High CRP levels and the change ratio of CRP after the second IVIG dose were associated with PEX.