A scoring system for differentiating multisystem inflammatory syndrome in children (MIS-C) after COVID-19 infection from Kawasaki disease

Abstract

Background: Multisystem inflammatory syndrome in children (MIS-C) is a post–COVID-19 systemic inflammatory condition with clinical features overlapping those of Kawasaki disease (KD), but with differences in severity, treatment strategies, and prognosis.

Objectives: To compare clinical characteristics of MIS-C and Kawasaki disease and to develop and evaluate a practical scoring system to assist in distinguishing between the two conditions.

Patients and Methods: A retrospective study was conducted in children younger than 18 years diagnosed with MIS-C or Kawasaki disease and admitted to Nakornping Hospital between January 2017 and December 2024. Clinical features, laboratory findings, echocardiographic results, and treatment outcomes were analyzed. Multivariable logistic regression was used to develop a diagnostic scoring system. The diagnostic performance of the model was internally assessed using receiver operating characteristic (ROC) curve analysis.

Results: A total of 220 patients were included: 41 with MIS-C and 179 with Kawasaki disease. Patients with MIS-C were older, had higher maximum body temperature, longer hospitalization, and more frequent need for vasoactive drugs and intensive care. Gastrointestinal, respiratory, and neurologic symptoms, as well as shock, were more common in MIS-C, while mucocutaneous features and higher erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) levels were more common in Kawasaki disease. Coronary artery aneurysms and left ventricular ejection fraction did not differ significantly between groups. The final scoring system included age >3 years, weight-for-height >120%, maximum temperature >39.3 °C, platelet count <500,000 cells/mm³, hemoglobin <9 g/dL, CRP <13 mg/dL, and ESR <50 mm/h. The area under the ROC curve (AuROC) was 0.87 (95% CI: 0.80–0.93). A simplified model excluding CRP demonstrated comparable performance (AuROC 0.86; 95% CI: 0.78–0.92). At the proposed cut-off, the model demonstrated high negative predictive

value but modest positive predictive value.

Conclusion: This simple and practical scoring system, based on routinely available clinical and laboratory parameters, demonstrates good discriminative ability and may be useful as a screening tool to support early clinical decision-making, particularly in resource-limited settings. However, external validation is required before clinical implementation.