Diogo Faim et al. Cláudio Henriques Ana Brett Andreia Francisco … Fernanda Rodrigues António Pires
Abstract
Background:
Kawasaki disease (KD) is the leading cause of acquired cardiac disease in children, in developed countries.
Objectives:
To identify predictive factors for resistance to intravenous immunoglobulin (IVIG), calculate the effectiveness of Japanese predictive models and characterize cardiac complications.
Methods:
Retrospective analysis of KD cases admitted in a Portuguese paediatric hospital between january 2006 and july 2018. ROC curves were used to determine predictive factors for resistance and the multivariate logistic regression analysis was used to develop the predictive model. A significance level of 5% was used.
Results:
48 patients with a median age of 36 months were included. The IVIG resistance was 21%. Echocardiographic anomalies were noted in 46%, with coronary involvement in 25% of the sample population. As predictive variable of resistance, the C-reactive protein (CRP) presented an AUC ROC = 0.789, optimal cut-off value 15.1 mg/dL, sensitivity (Sn) 77.8% and specificity (Sp) 78.9%. The erythrocyte sedimentation rate (ESR) presented an AUC ROC = 0.781, optimal cut-off value 90.5 mm/h, Sn 66.7% and Sp 85.7%. The model with the two variables showed p = 0.042 and AUC ROC = 0.790. Predictive strength of Japanese models were: Kobayashi (Sn 63.6%, Sp 77.3%), Egami (Sn 66.7%, Sp 73.1%), Sano (Sn 28.6%, Sp 94.1%).
Conclusion:
CRP and ESR are independent variables that were related to IVIG resistance, with optimal cut-off points of 15.1 mg/dL and 90.5 mm/h, respectively. About half of the patients had some form of cardiac involvement. The Japanese models appeared to be inadequate in our population. (Arq Bras Cardiol. 2021; 116(3):485-491)