Utility of a Proteomic Surrogate for Cardiovascular Outcomes to Predict and Monitor COVID-19 Induced Myocarditis

Abstract

Introduction: Poorer cardiovascular (CV) health is a risk factor for adverse outcomes from COVID-19, and a proportion of COVID-19 survivors exhibit CV dysfunction including myocardial inflammation after initial recovery. However, biomarker-driven predictors of CV outcomes in COVID-19 are lacking. We recently demonstrated that a proteomic surrogate of CV outcomes (MI, stroke, hospitalization for heart failure, or death within 4 years) is highly predictive of acute COVID-19 severity. The current study expands on this to determine if a proteomic CV risk test associates with myocarditis likelihood in COVID-19 survivors. Methods: The SomaScan® platform was used for plasma proteomic phenotyping in CISCO-19 trial participants (n=154, mean age 55y, 43% Female) at hospital discharge following COVID-19, and at follow-up within 6-months later when comprehensive CV assessment including cardiac MRI for myocarditis adjudication was performed. N=20 (13%) were classified as very likely to have myocarditis at follow-up. We used linear mixed models to test proteomic CV scores at both visits against myocarditis likelihood across COVID-19 recovery and compared the predictive performance of the CV risk score to discriminate myocarditis likelihood to performance of Troponin I (TnI) and NT-proBNP. Results: A 36% relative CV risk reduction was observed during COVID-19 recovery. CV risk scores were significantly higher among those very likely myocarditis positive at follow up (p < 0.05 for effects) (Figure 1). Additionally, the performance of the CV risk test in classifying myocarditis likelihood at follow-up was superior to that of TnI or NT-proBNP (AUC=0.64 vs 0.52 and 0.51, respectively). Conclusions: A proteomic predictor of CV risk has potential to support cardiac screening and monitoring CV health across the course of COVID-19 recovery, with diagnostic specificity superior to that of specific single cardiac biomarkers. Future studies with this approach seem warranted.

Rights

This content is not covered by the Open Government Licence. Please see source record or item for information on rights and permissions.

Cite as

Downloadable citations

Identifiers

DOI

https://doi.org/10.1161/circ.146.suppl_1.12132

Repository URI

https://eprints.gla.ac.uk/321233/