Prospective electrocardiographic and cardiovascular magnetic resonance alterations in the UK Biobank coronavirus disease 2019 repeat imaging study

Sucharitha Chadalavada, Ahmed Salih, Hafiz Naderi, Elisa Rauseo, Jackie Cooper, Stefan van Duijvenboden, C. Anwar A. Chahal, Gaith S Dabbagh, Liliána Szabó, Mohammed Y Khanji, Jose D. Vargas, Mihir M. Sanghvi, Kenneth Fung, José Miguel Paiva, Stefan K. Piechnik, Betty Raman, Patricia B. Munroe, Aaron M. Lee, Alborz Amir-Khalili, Luca Biasiolli, John P. Greenwood, Paul M. Matthews, Wenjia Bai, Stefan Neubauer, Nay Aung, Nicholas C. Harvey, Zahra Raisi‐Estabragh, Steffen E. PetersenJournal of Cardiovascular Magnetic Resonance 2025 · DOI

TL;DR

This study leverages the UK Biobank's unique longitudinal data to investigate cardiovascular changes after COVID-19 infection. It finds that while pre-existing cardiovascular health influences COVID-19 risk, mild infection itself doesn't appear to cause new or altered cardiovascular changes detectable by MRI or ECG over a 3-year period.

What Problem Does This Paper Address?

The COVID-19 pandemic has raised significant concerns about its long-term cardiovascular consequences. While many studies have reported cardiovascular abnormalities (like myocarditis) in individuals after COVID-19 using imaging techniques like cardiovascular magnetic resonance (CMR) and electrocardiography (ECG), a major limitation has been the lack of pre-infection baseline data. This makes it difficult to determine if the observed abnormalities are truly caused by the infection or if they represent pre-existing conditions that also predispose individuals to severe COVID-19. This paper aims to address this by using a unique dataset with paired pre- and post-infection CMR and ECG scans to establish a clearer causal link.

Methods

The researchers utilized data from the UK Biobank, a large prospective cohort study. They identified participants who had undergone CMR imaging before the pandemic and subsequently contracted COVID-19. These individuals (cases) were then matched with uninfected controls based on age, sex, ethnicity, and location. Both groups underwent repeat CMR and ECG scans.

Key methodological aspects include:

  • Longitudinal Design: The study's strength lies in its pre- and post-infection imaging, allowing for the assessment of interval change rather than just cross-sectional abnormalities.
  • Automated Pipelines: Standardized automated pipelines were used for extracting ECG phenotypes and CMR measures (cardiac structure, function, aortic distensibility, flow, and native T1). This ensures consistency and allows for large-scale analysis.
  • Statistical Analysis:
    • Baseline characteristics of cases and controls were compared.
  • Logistic regression was used to assess the association between baseline CMR/ECG metrics and the risk of incident COVID-19.
  • Crucially, to assess interval change, a method involving standardized residuals was employed. This technique accounts for baseline values to represent the degree of change from the expected trajectory, providing a more robust measure of change than simple subtraction.

Datasets

The primary dataset used is the UK Biobank. This is a large, well-established prospective cohort study that includes detailed health information, genetic data, and, for a subset, advanced imaging like CMR. The study specifically utilized data from the UK Biobank COVID-19 Repeat Imaging Study, which provided the unique pre- and post-infection CMR and ECG data for both COVID-19 cases and matched controls.

Key Insights and Ideas

  • Pre-existing Health Matters: Poorer baseline cardiovascular health, specifically lower left ventricular ejection fraction (LVEF) and reduced LV strain (global longitudinal, circumferential, and radial), was associated with a higher risk of contracting COVID-19. This suggests that individuals with underlying cardiovascular vulnerabilities may be more susceptible.
  • Mild COVID-19 Doesn't Seem to Cause New Cardiac Damage: Despite concerns about long-term cardiac sequelae, this study found no significant differences in the interval change of CMR or ECG metrics between COVID-19 cases and controls. This implies that mild to moderate SARS-CoV-2 infection, in this cohort, did not lead to de novo cardiovascular abnormalities detectable by these sensitive measures over a 3-year follow-up.
  • Methodological Rigor is Key: The use of standardized imaging protocols and advanced statistical methods (like standardized residuals for change analysis) is highlighted as crucial for avoiding confounding and reverse causation, which have plagued previous studies.
  • Reassurance for Mild Cases: The findings offer reassurance that for individuals with mild COVID-19, significant, persistent cardiovascular damage detectable by CMR or ECG is unlikely.

Key Results

  • Baseline Associations: Lower baseline LVEF and poorer LV strain metrics were associated with an increased risk of incident COVID-19 infection.
  • Interval Change: No statistically significant differences were found in the degree of interval change for any of the analyzed CMR or ECG metrics between the COVID-19 cases and their matched uninfected controls. This applied to measures of cardiac structure, function, myocardial tissue character (native T1), aortic mechanics, and various ECG parameters.
  • Clinical Significance: While some statistically significant differences in LVEF and corrected QT interval (QTc) were observed at the repeat imaging visit, the authors note these were not clinically relevant in magnitude.

Limitations and Future Work

  • Mild Infection Cohort: The study primarily included individuals with mild or asymptomatic COVID-19. The findings may not be generalizable to those with severe COVID-19 requiring hospitalization or intensive care.
  • Asymptomatic Controls: There's a possibility that some individuals in the control group may have had undiagnosed, asymptomatic COVID-19 infections.
  • CMR Protocol Limitations: The UK Biobank CMR protocol is designed for population screening and has limitations, such as the absence of contrast-enhanced imaging and limited non-parametric tissue characterization sequences.
  • Follow-up Interval: The median follow-up of 1.3 years might be too short to detect very late-onset cardiovascular changes, although it is considered appropriate for assessing persistent changes.
  • Future Work: The authors suggest that future research could explore individuals with severe COVID-19 to understand potential differences in cardiovascular impact. Further investigation into the persistence of symptoms in "long COVID" and their correlation (or lack thereof) with objective cardiovascular findings remains an important area.

Summary updated

Abstract

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.
BibTeX
@article{chadalavada2025prospective,
  author = {Sucharitha Chadalavada and Ahmed Salih and Hafiz Naderi and Elisa Rauseo and Jackie Cooper and Stefan van Duijvenboden and C. Anwar A. Chahal and Gaith S Dabbagh and Liliána Szabó and Mohammed Y Khanji and Jose D. Vargas and Mihir M. Sanghvi and Kenneth Fung and José Miguel Paiva and Stefan K. Piechnik and Betty Raman and Patricia B. Munroe and Aaron M. Lee and Alborz Amir-Khalili and Luca Biasiolli and John P. Greenwood and Paul M. Matthews and Wenjia Bai and Stefan Neubauer and Nay Aung and Nicholas C. Harvey and Zahra Raisi‐Estabragh and Steffen E. Petersen},
  title = {{Prospective electrocardiographic and cardiovascular magnetic resonance alterations in the UK Biobank coronavirus disease 2019 repeat imaging study}},
  journal = {Journal of Cardiovascular Magnetic Resonance},
  year = {2025},
  doi = {10.1016/j.jocmr.2025.101957},
}
Figure