Background: To investigate the diagnostic accuracy of CXCL10/IP10 for left ventricular (LV) dysfunction in multisystemic inflammatory syndrome (MIS-C). Methods: This cross-sectional, longitudinal study included 36 patients with MIS-C. Patients were classified as follows: (1) patients presenting with Kawasaki-like features (group I = 11); (2) patients presenting with LV systolic dysfunction (group II = 9); and (3) other presentations (group III = 3). CXCL10/IP10 levels were measured upon admission and on days 3 and 7 of treatment. Results: Twenty patients were male and 16 were female. The median age of patients at diagnosis was 7.5 (1.5–17) years. All patients had a fever lasting for a median of 4 (2–7) days. Ten patients had LV systolic dysfunction. The duration of hospitalization was longer in group II. Lymphocyte and platelet counts were lower, whereas NT-pro-BNP, troponin-I, D-dimer, and CXCL10/IP10 levels were higher in group II. Baseline levels of CXCL10/IP10 were weakly negatively correlated with ejection fraction (r = −0.387, p = 0.022). Receiver operator characteristic curve analysis yielded a cutoff value of CXCL10/IP10 to discriminate patients with LV dysfunction was 1839 pg/mL with sensitivity 88% and specificity 68% (Area under curve (AUC) = 0.827, 95% CI 0.682–0.972, p = 0.003). Conclusion: Having a good correlation with cardiac function, CXCL10/IP10 is a potential biomarker to predict LV dysfunction in MIS-C patients.
Read full abstract