PM1021: Does Admission Echocardiogram Predict Clinical Illness Severity in MIS-C?

Background: Multisystem inflammatory syndrome in children (MIS-C) has variable clinical presentations with a subset requiring vasoactive support. Echocardiography is typically performed early to evaluate the extent of cardiac involvement. We hypothesized that measurements of ventricular systolic and diastolic function, including myocardial deformation, on the admission echocardiogram predicts the need for a higher acuity of care.

Methods: This single center retrospective cohort study included all consecutive patients diagnosed with MIS-C from 4/1/2020 to 12/31/2021. Patients with preexisting cardiomyopathy, congenital heart disease, or prior exposure to cardiotoxic therapies were excluded. Patients were divided into 3 groups based on illness severity, defined by vasoactive-inotrope score (VIS) which adjusts for cumulative dosing and differential units of measurement: VIS 0= no vasoactives/mild disease, VIS 2= lower intensity vasoactive need/moderate disease, VIS 3= higher intensity vasoactive need/severe disease. Admission labs and echocardiographic measurements were compared using Kruskal-Wallis test or Fisher’s exact test as appropriate. P-values of multiple pairwise comparisons between vasoactive-score subgroups were corrected by Bonferroni adjustment.

Results: Of 127 MIS-C patients, 9 were excluded. Median age was 7.8 years. 57 patients (48%) received vasoactive medication (moderate n=43, severe n=14). Admission B-type natriuretic peptide (120.0 vs 341.5 vs 932.5, p< 0.01) and troponin (0.01 vs 0.09 vs 0.13, p< 0.01) were higher for severe illness. Differences in left ventricular (LV) ejection fraction (EF), LV longitudinal strain (LS), and reservoir left atrial strain (LAS) between mild and severe subgroups were noted (Figure). Those with severe illness had a lower median LVEF (62% vs 59% vs 52%, p < 0.01), worse LVLS (-17.2% vs -17.0% vs -15.7%, p< 0.02), and worse LA reservoir strain (35.3% vs 30.6% vs 28.7%, p< 0.01). There were no differences between mild and moderate groups in LVEF, LVLS, and reservoir LAS (Figure). Right ventricular LS was similar in all 3 groups. Admission LVEF was within normal limits for 39 of the 57 patients who ultimately required vasoactive support.

Conclusion: Admission echocardiographic measurements of ventricular function in MIS-C patients did not reliably predict the need for vasoactive medications. Mildly depressed LVEF was identified as a risk factor, but LVEF was within normal limits in 68% of patients who later received vasoactive support. LVLS was mildly impaired at presentation in most patients, but the small differences between the lower- and higher-vasoactive-need groups limit its discriminatory ability. This stresses the need for serial echocardiography, as measures of ventricular dysfunction may lag symptoms and laboratory markers of inflammation.