Development of a nomogram prediction model for early identification of persistent coronary artery aneurysms in kawasaki disease

Abstract

Abstract Background Coronary artery aneurysms (CAA) persistence prediction is critical in evaluating Kawasaki disease (KD). This study established a nomogram prediction system based on potential risk factors for assessing the risk of CAA persistence in a contemporary cohort of patients with KD. Methods This cohort comprised 105 patients with KD who had been diagnosed with CAA during the acute or subacute phase by echocardiography. The follow-up duration was at least 1 year. The clinical and laboratory parameters were compared between the CAA regression and persistence groups. Multivariable logistic regression analysis was used to identify the independent risk factors for CAA persistence, which were subsequently used to build the nomogram predictive model. Decision curve analysis was used to assess the net benefits of different nomogram scores. Results Of these patients with CAA, 27.6% of patients presented with persistent lesions. The incidences of CAA persistence were 14.1%, 81.3%, and 100.0% in patients with small, medium, and large aneurysms, respectively. The ratio of neutrophils to lymphocytes, γ-GT, and CAA size at diagnosis were considered as the independent risk factors for CAA persistence in patients with KD. The nomogram predictive models yielded a high capability in predicting CAA persistence, based on either univariable or multivariable analyses-identified parameters, compared with using CAA size as a single predictor. Conclusion The initial ratio of neutrophils to lymphocytes, γ-GT, and CAA size were the independent risk factors for CAA persistence in patients with KD. Nomogram scores could help elevate predictive efficacy in detecting CAA persistence.

Keywords

• Nomogram prediction
• Kawasaki disease
• Coronary artery aneurysms persistence
• Risk factor
• Decision curve analysis