Study population and setting
This study explored cardiovascular sequelae following COVID-19 infection. The prospective cohort study enrolled 100 patients from the University Hospital Frankfurt COVID-19 Registry in Germany from April to June, 2020 following recovery from COVID-19 infection. Recovery was defined as resolution of respiratory symptoms and a negative swab for SARS-CoV-2 at the end of the isolation period (minimum of 2 weeks from diagnosis). Recovered patients with ongoing cardiac symptoms who were referred for cardiovascular magnetic resonance (CMR) imaging were not included. Two control groups sampled from the International T1 Multicenter Outcome Study were also included: 1) age- and sex- matched healthy controls and 2) risk-factor stratified controls, matched on age, sex, hypertension, diabetes, smoking, coronary artery disease, or comorbidities. CMR imaging was performed on clinical 3-T scanners, and cardiac parameters were measured using a combination of artificial-intelligence-based contour detection and standardized post-processing recommendations.
Summary of Main Findings
Of the 100 patients (median age 49 years, 53% female) who recently recovered from COVID-19, 18% were asymptomatic, 49% had mild-to-moderate presentation, and 33% required hospitalization during infection, including 2 who received invasive mechanical ventilation and 17 who received non-invasive mechanical ventilation with positive airway pressure. No patients had a history of heart failure or cardiomyopathy. CMRs were performed a median 71 days (IQR: 64-92) after testing positive for COVID-19, and at the time of the scan, 36% of patients had ongoing shortness of breath or general exhaustion, many of whom were never hospitalized. Troponin-T levels were above normal in 71% and significantly elevated in 5%. Overall, 78% of patients who had recently recovered from COVID-19 had abnormal CMR findings, most commonly myocardial inflammation, defined by raised native T1 (73%) and T2 (60%) measures, followed by regional scar (32%), pericardial enhancement (22%), and significant pericardial effusion (20%) all of which were higher than observed in risk factor matched controls (abnormal T1: 58%, abnormal T2: 26%, regional scar: 17%, pericardial enhancement: 15%, and pericardial effusion: 15%). Patients recovering in hospital had higher native T1 relative to those recovering at home.
The study characterizes a number of parameters describing cardiac involvement from CMR, and compares patients who recovered from COVID-19 infection to two control groups, including one that was matched on several clinical risk factors. The imaging sequences used in the study were validated and standardized and included state of the art T1 and T2 mapping.
The sample size of this study is small, including only 100 recovered COVID-19 patients. The selection process of recovered patients into the study is unclear. A significant number of patients were in fact symptomatic at the time of CMR and many had elevated troponin levels. The prevalence of cardiac abnormalities and troponin elevations were also high in the control groups, particularly the risk-factor matched cohort. Analyses of cardiac abnormalities did not thoroughly assess differences by COVID-19 severity. The authors attempt to minimize confounding by matching COVID-19 patients to controls based on age, sex, and other risk factors for cardiac involvement, but residual confounding likely exists, and adjusted analyses accounting for factors associated with testing positive for COVID-19 or CMR imaging would be helpful. Many patients had ongoing symptoms at the time of CMR, and it is unknown whether CMR findings would resolve alongside symptom resolution.
This is the first known prospective report of evaluation for cardiac involvement among patients recovered from COVID-19.
This review was posted on: 20 August 2020