Publication Date




Download Full Text (728 KB)


Introduction – Kawasaki disease (KD) and Multisystem Inflammatory Syndrome in Children (MIS-C) associated with COVID-19 show considerable clinical overlap making differentiation challenging, particularly in the absence of evidence of evidence of prior COVID-19. Hypothesis – Cardiac biomarkers can differentiate KD from MIS-C. Methods – The International KD Registry enrolled n=2903 contemporaneous KD, MIS-C and acute COVID-19 pediatric patients from 42 sites in 8 countries from January 2020 through July 2022. The study population was confined to 1489 MIS-C patients meeting CDC criteria with confirmed evidence of prior COVID-19 infection and 387 KD patients meeting AHA guideline criteria with confirmed evidence supporting no prior COVID-19 infection, limited to those with sufficient echocardiogram data and at least one measurement of NTproBNP or Troponin I (118 KD, 946 MIS-C). A normalizing logarithmic transformation was applied to biomarker levels and multiple imputation of missing values of factors was performed before multivariable regression analyses to determine associated factors. Regression models were adjusted for diagnosis, age and creatinine at presentation and then explored for associations of biomarkers with clinical presentation, laboratory variables and cardiac involvement. Receiver operating characteristic curves were used to determine biomarker cut points differentiating KD from MIS-C. Results – Of 118 patients with KD, 90 had NTproBNP (median 370 ng/L) and 91 had Troponin I (median < 10 ug/L) assessed at presentation, while of 946 patients with MIS-C, 580 had NTproBNP (median 1944 ng/L; p<0.001 vs KD) and 763 had Troponin I (median 35.0 ug/L; p < 0.001) assessed. Baseline NTproBNP and Troponin I were not significantly correlated (r=0.08; p=0.10). Higher Troponin I was correlated with older age (r=0.07; p=0.04) and higher creatinine (r=0.25; p < 0.001), whereas NTproBNP was only correlated with higher creatinine (r=0.29; p < 0.001). After adjusting for diagnosis, age and creatinine, both higher log(NTproBNP) and log(Troponin I) were association with shock at presentation (p <0.001, p < 0.001, respectively) and ICU admission (p=0.003, p < 0.001). Higher log(NTproBNP) but not log(Troponin I) was associated with longer length of hospital stay (p < 0.001, p=0.23). After similar adjustments, laboratory variables at presentation independently associated with cardiac biomarkers are shown in TABLE. LV ejection fraction (LVEF) was lower for MIS-C versus KD (median 56% vs 63%; p<0.001). After adjusting for diagnosis, age and creatinine, higher baseline log(NTproBNP) was associated with lower LVEF (p<0.001) and LVEF<55% (p<0.001). After similar adjustments, higher baseline log(Troponin I) was association with lower LVEF (p=0.03) and LVEF<55% (p=0.03). Maximum coronary artery Z score was greater for KD versus MIS-C (median 1.36 vs 1.23; p<0.05). After adjusting for age, creatinine and diagnosis, higher baseline log(NTproBNP) was not associated with maximum coronary artery Z score (p=0.36) but with max Z score>2 (p=0.02). After similar adjustments, baseline log(Troponin I) was not associated higher maximum coronary artery Z score (p=0.23) or max Z score > 2 (p=0.72). Baseline Troponin I > 10 ug/L (c-statistic 0.65) predicted MIS-C vs KD with a sensitivity of 58% and specificity of 77%, with > 20 ug/L 44% and 89%, respectively. Baseline NTproBNP > 500 ng/L (c-statistic 0.71) with a sensitivity of 74% and specificity of 54%, >1000 ng/L 61% and 72%, and >1 500 ng/L 56% and 77% (FIGURE). The c-statistic improved to 0.74 with both biomarkers together, and to 0.78 with both biomarkers at peak. Conclusions – Higher baseline Troponin I and NTproBNP levels are predictive of MIS-C versus KD with reasonable sensitivity and specificity, and are associated with an increased likelihood of shock and ICU admission. Lower LVEF, more common with MIS-C, is associated with higher NTproBNP and Troponin I levels, while coronary artery involvement, more comm


Cardiology | Pediatrics


Presented at the American Heart Association - Scientific Sessions; Chicago, IL; November 5-7, 2022.

Cardiac Biomarkers in Differentiating Kawasaki Disease and Multisystem Inflammatory Syndrome in Children Associated with COVID-19