Prospective electrocardiographic and cardiovascular magnetic resonance alterations in the UK Biobank COVID-19 repeat imaging study

Abstract

Background: Cardiovascular magnetic resonance (CMR) and electrocardiographic (ECG) abnormalities after COVID-19 are widely reported. However, the absence of pre-infection assessments limits causal inference from these studies. This study aims to compare interval change in CMR and ECG measures in participants with incident COVID-19 and matched uninfected controls in UK Biobank.

Methods: UK Biobank participants with documented COVID-19 who had CMR and ECG performed prior to the pandemic were invited for repeat assessment, along with uninfected participants matched on age, sex, ethnicity, location, and date of baseline imaging. Automated pipelines were used to extract ECG phenotypes and CMR measures of cardiac structure and function, aortic distensibility, aortic flow, and myocardial native T1. Logistic regression was used to examine associations of baseline metrics with incident COVID-19. Standardized residual approach was used to compare the degree of interval change in CMR and ECG metrics between cases and controls.

Results: We analyzed 2,092 participants (1,079 cases, 1,013 controls) with average age of 60±7 years. 47% were male. There was 3.2±1.5 years between pre- and post-infection assessments. 4% of cases were hospitalized. Lower baseline left ventricular ejection fraction and worse longitudinal, circumferential, and radial strain were associated with higher risk of incident COVID-19. There were no significant differences in interval change of any CMR or ECG metric between cases and controls.

Conclusions: While pre-existing cardiovascular abnormalities are linked to higher risk of COVID-19, exposure to infection does not alter interval change of highly sensitive CMR and ECG indicators of cardiovascular health.

Keywords: Cardiac magnetic resonance; Cardiovascular disease; Electrocardiogram; Long Covid; Myocarditis; SARS-COV-2.