Prediction of Coronary Artery Aneurysms in Children With Kawasaki Disease Before Starting Initial Treatment

Background: Precisely predicting coronary artery aneurysms (CAAs) remains a clinical challenge. We aimed to investigate whether coronary dimensions adjusted for body surface area (Z scores) on baseline echocardiography and clinical variables before primary treatment could predict the presence of late CAAs. Methods: We conducted a retrospective study including children hospitalized for Kawasaki disease and received intravenous immunoglobulin within 10 days of illness. We defined late CAAs as a maximum Z score (Zmax) ≥2.5 of the left main, right, or left anterior descending coronary artery at 11–60 days of illness. Associations between late CAAs and clinical parameters and baseline maximum Z scores were analyzed. Results: Among the 314 included children, 31 (9.9%) had late CAAs. Male, higher C-reactive protein, and higher baseline Zmax were risk factors of late CAAs. Late CAAs were significantly associated with baseline Zmax ≥2.0 vs. <2.0 (25 [32.5%] vs. 6 [2.5%], P < 0.001). The odds ratio for late CAAs among the patients with baseline Zmax ≥2.0 vs. <2.0 was 18.5 (95% confidence interval, 7.23 to 47.41, P < 0.001). The sensitivity, specificity, positive predictive value, and negative predictive value of baseline Zmax ≥2.0 for the presence of later CAAs were 80.6, 81.6, 32.5, and 97.5%, respectively. Conclusions: Findings from this study suggest that Zmax ≥2.0 of coronary arteries on baseline echocardiography may be used to predict children at a high risk of late CAAs and allow for targeted early intensification therapy.