Utility of proteomic trajectories of cardiovascular risk and cardiorespiratory fitness to monitor adverse health states throughout post-COVID-19 illness

Background

• Cardiovascular involvement is a prominent observation in patients during the acute phase of COVID-19 infection, as well as in convalescence. However, the etiology, trajectory, and underlying biology of cardiac dysfunction across the spectrum of COVID-19 illness is not fully understood.
• To address this, the CISCO-19 study (NCT04403607) was formed to investigate the multisystem effects of COVID-19 from hospitalized patients using blood-based biomarkers, imaging, and patient-reported outcome measures¹.
• The CISCO-19 study identified a significant proportion of COVID-19 patients showed evidence for myocarditis in the short term (3 months) after COVID-19 hospitalization, and that this associates with persisting impairments in health status over longer-term follow-up².
• These results indicate the potential ongoing substantial demands on healthcare services following COVID-19 illness, and highlight the need for high-throughput tools for evaluation and monitoring cardiac involvement across the course of illness and recovery.
• The Cardiovascular Risk Secondary SomaSignal™ Test (CV Risk SST) is a proteomic machine-learning model comprised of 27 proteins encompassing 10 biological systems that predicts the risk of myocardial infarction, heart failure, stroke or all-cause death within four years from a plasma sample; the Cardiorespiratory Fitness SST is comprised of 52 proteins and predicts VO2 max measurements without the need for exercise stress-testing.

Objectives

1. Evaluate the ability of proteomic surrogates for CV Risk and Cardiorespiratory Fitness to monitor changes from acute COVID-19 illness into convalescence.
2. Determine the differences in predicted CV Risk and Cardiometabolic Fitness scores between myocarditis likelihood groups post-COVID-19, and between those that are rehospitalized.
3. Assess and compare the performance of the proteomic CV Risk SST to classify myocarditis likelihood to individual traditional circulating cardiac biomarkers.

Methods

• Paired EDTA plasma samples from the CISCO-19 trial collected from 154 patients hospitalized with COVID-19 at hospital discharge (Discharge) and again at 28-60 days post-discharge (Follow-up) were assayed using modified–aptamer proteomics technology, SomaScan Assay v4.1
• CV Risk and Cardiorespiratory Fitness SST scores were generated based on assayed proteomic measurements for each sample and compared between myocarditis likelihood groups which were adjudicated by consensus, as well as health record follow-up for rehospitalization.
• Circulating levels of high sensitivity cardiac troponin I (TnI) and NT-proBNP were measured in plasma using i1000SR ARCHITECHT (Abbott Diagnostics).
• Change in CV Risk scores and cardiorespiratory fitness between discharge and follow-up samples were assessed using paired t-tests.
• A linear-mixed model was designed using timepoint and myocarditis status (“very likely myocarditis” vs. others) or rehospitalization as fixed effects, and individual subject random effects. Effects were evaluated for significance using ANOVA.
• Classification performance of CV Risk was assessed by calculating area under the curve (AUC), and compared to TnI and NT-proBNP for 149 subjects that had all measurements available.

Results

• Myocarditis likelihood status and time after discharge were significant effects (ANOVA p<0.05) associated with predicted CV risk and cardiorespiratory fitness in COVID-19 recovery.
• Rehospitalization was also a significant effect associated with CV risk, but not cardiorespiratory fitness.

Conclusions

• Together our findings demonstrate the ability of proteomic CV risk and cardiorespiratory fitness surrogates to do the following:
  • Monitor risk trajectories following COVID-19 infection
  • Detect elevated risk of rehospitalization and consensus-based adjudicated myocarditis
  • Classify patients’ myocarditis likelihood with superior performance to that of traditional CV biomarkers.
• These findings suggest that cardiovascular-related proteomic SST models may have utility through use of clinicians to guide risk stratification of COVID-19 patients for precision-based medical management, monitoring and rehabilitation throughout illness and convalescence.

Authors

Clare Paterson¹
Michael A. Hinterberg¹
Andrew J. Morrow^2,3
Robert Sykes^2,3
Kenneth Mangion^2,3
Paul Welsh²
Alex McConnachie^2,4
Naveed Sattar²
Colin Berry^2,3,5
Stephen A. Williams¹

¹SomaLogic Operating Co., Inc, Boulder CO, USA
²British Heart Foundation Glasgow Cardiovascular Research Centre, University of Glasgow, Glasgow, UK
³Department of Cardiology, Queen Elizabeth University Hospital, Glasgow, UK
⁴Robertson Centre for Biostatistics, Institute of Health and Wellbeing, University of Glasgow, Glasgow, UK
⁵West of Scotland Heart and Lung Centre, NHS Golden Jubilee, Clydebank, UK