A practical nomogram for predicting coronary thrombosis for Kawasaki disease patients with medium or large coronary artery aneurysm

Kawasaki disease (KD) is the main cause of acquired heart disease in children. Coronary thrombosis is a serious cardiovascular complication of KD, which affects the long-term treatment effect. The purpose was to develop and validate a model for predicting coronary thrombosis in KD with medium or large coronary artery aneurysm (CAA). A total of 358 consecutive KD patients with medium or large CAA from Chongqing Children's Hospital were enrolled retrospectively. The demographic data, clinical characteristics, laboratory features before intravenous immunoglobulin (IVIG) treatment, and all radiological features during hospitalization and follow-up were collected. Eligible patients follow-up for > 2 years. Follow-up was weekly for the first 1 month, monthly for the next 11 months, and every 3-6 months after 1 year. The main examinations included echocardiogram and electrocardiogram. The primary endpoint was defined as coronary thrombosis during the follow-up. Coronary thrombosis was assessed by echocardiographic assessment of the presence of echoes in the lumen of the right coronary artery, left main coronary artery, left anterior descending artery, or left circumflex artery by echocardiologists. The independent risk factors were identified using univariate analyses and multivariate logistic regression analyses, and the nomogram was constructed for predicting coronary thrombosis. Tenfold cross-validation was used to perform internal validation. The area under the ROC curve (AUC), calibration curve, and decision curve analysis were used to evaluate the discrimination, calibration, and clinical utility of the nomogram, respectively. Multivariate logistic regression analysis revealed that male (odds ratio [OR] 3.491; 95% confidence interval [CI] 1.570-7.765), large CAA (OR 3.725; 95% CI 1.388-9.999), no use high-dose aspirin prior to IVIG (OR 3.114; 95% CI 1.291-7.510), two-vessel coronary artery involvement (OR 4.433; 95% CI 1.732-11.344), three-vessel coronary artery involvement (OR 5.417; 95% CI 2.048-14.328), four-vessel coronary artery involvement (OR 13.183; 95% CI 3.408-50.997), serum fibrinogen level > 5.325 g/L (OR 14.233; 95% CI 5.479-36.921), serum thrombin time level <= 15.15 s (OR 3.576; 95% CI 1.756-7.284) were significantly associated with coronary thrombosis. The nomogram was established based on these variables. The AUC of the nomogram were 0.920, and tenfold cross-validation (repeated 100 times) showed that the average AUC was 0.902. Moreover, the nomogram had a well-fitted calibration curve and also exhibited good clinical usage. The nomogram is based on six ready-made clinical variables, is easy to use, has excellent diagnostic performance, and can help clinicians make better clinical decisions on the management and treatment of KD patients with medium or large CAA.