Coronary risks after high-dose γ-globulin in children with Kawasaki disease

• Published in: Pediatrics international Vol. 42; no. 5; pp. 464 - 469
• Main Authors: Morikawa, Yoshiyuki, Ohashi, Yasuo, Harada, Kensuke, Asai, Toshio, Okawa, Sumio, Nagashima, Masami, Katoh, Toshiyuki, Baba, Kunizo, Furusho, Kenshi, Okuni, Masahiko, Osano, Mitsuru
• Format: Journal Article
• Language: English
• Published: Melbourne, Australia Blackwell Science Pty 01.10.2000
• Subjects: Child; Coronary Disease - etiology; cross validation; Humans; Immunoglobulin G - blood; Immunoglobulins, Intravenous - administration & dosage; intravenous γ-globulin therapy; Kawasaki disease; Logistic Models; mucocutaneous lymph node syndrome; Mucocutaneous Lymph Node Syndrome - complications; Mucocutaneous Lymph Node Syndrome - drug therapy; Mucocutaneous Lymph Node Syndrome - immunology; Multivariate Analysis; Predictive Value of Tests; Risk Assessment; risk factor
• ISSN: 1328-8067; 1442-200X
• DOI: 10.1046/j.1442-200x.2000.01288.x

Objectives: The goals of the present study were to develop a predictive coronary risk scoring system after intravenous γ‐globulin (IVGG) therapy of any dose for the different preparations currently used in the treatment of children with Kawasaki disease and to determine the predictive value of the system. The previously reported scoring systems were based on treatment with high‐dose IVGG therapy at limited doses and were determined using investigative methods. Methods : Four hundred and fifty‐one patients were randomized into one of three groups and received either i.v. polyethylene glycol‐treated human immunoglobulin at a dose of either 200 (n=147) or 400 mg/kg per day (n=152) or freeze‐dried sulfonated human immunoglobulin at 200 mg/kg per day (n=152) for 5 consecutive days. We documented 31 cases of coronary abnormalities (CA). Univariate and multivariate logistic regression was performed using 49 clinical variables and the resulting predictive model was validated. Results : The duration of fever (odds (1 day)/odds (≥ 5 days)=0.158; 95% confidence interval (CI) 0.0385–0.648), hemoglobin (odds (Q1=10.3)/odds (Q3=11.6) = 3.97; 95% CI 1.92–8.20), IgG (odds (Q1=1900)/odds (Q3=2658)=2.72, 95% CI 1.18–6.25) and IgA (odds (Q1=72)/odds (Q3=160) =0.415; 95% CI 0.253–0.680) levels after completion of γ‐globulin infusion were independent predictors. The model is quasi‐cross validated and has acceptable sensitivity and selectivity. The estimated risk and observed occurence of CA coincide. Conclusions : Determinants of the risk of CA after IVGG therapy are a longer duration of fever, a lower IgG level, a higher IgA level and a lower hemoglobin level after IVGG infusion. This model is applicable for IVGG doses from 1 to 2 g/kg and for at least two different γ‐globulin preparations.