Abstract 14551: Biomarkers Predicting the Treatment-Resistant Kawasaki Disease

Abstract

Introduction: Kawasaki disease (KD), which is the most common multisystem vasculitis with unknown causes in childhood, causes coronary artery lesions (CALs). Treatment with a high dose of intravenous immunoglobulin (IVIG), plus steroids if needed, is the most effective therapy for the acute phase of KD. However, there are some very severe cases who need several times additional treatments and are at risk for CALs. In Japan, there are some scoring systems that initially predict IVIG-resistant patients. However, the problem is that these scoring systems fail in multiethnic populations. The aim of this study is to find universal biomarkers that predict treatment-resistant cases of KD. Methods: The subject was 276 KD patients, including Group 1 (n=214) who needed only 1 st line treatment, Group 2 (n=48) who needed 2 nd line treatment, Group 3 (n=14) who needed 3 rd line treatment or more. Tenascin C (TN-C), Pentraxin 3 (PTX3) and Procalcitonin (PCT) values, which were selected by systematic review, were measured before initial treatment in each group. Results: TN-C; 99.8±41.05 ng/ml in Group 1, 118.0±71.4 ng/ml in Group 2 and 183.0±25.0 ng/ml in Group 3. The TN-C level of Group 3 was significantly higher than that of all the others (p<0.01). The cutoff value for distinguishing Group 3 was 142.0 ng/ml (Area under the Curve (AUC)=0.81). PTX3; 16.2±9.0 ng/ml in Group 1, 31.4±19.7 ng/ml in Group 2 and 58.0±33.0 ng/ml in Group 3. The PTX3 level of Group 3 was significantly higher than that of all the others (p<0.01). The cutoff value for distinguishing Group 3 was 35.1 ng/ml (AUC=0.86). PCT; 0.79±0.77 ng/ml in Group 1, 2.55±3.01 ng/ml in Group 2 and 4.15±4.49 ng/ml in Group 3. The PCT level of Group 3 was significantly higher than that of all the others (p<0.01). The cutoff value for distinguishing Group 3 was 2.55 ng/ml (AUC=0.88). When those three biomarkers combined, Group 3 can be predicted with the sensitivity 79%, the specificity 96%, the positive predictive value 50% and the negative predictive value 99%. Conclusions: It may be possible to predict treatment-resistant KD cases with high sensitivity and specificity by combining the measurement from the universal biomarkers, TN-C, PTX3 and PCT, before initial treatment.