Patients Recovered From COVID-19 Show Ongoing Subclinical Myocarditis as Revealed by Cardiac Magnetic Resonance Imaging

The cardiovascular complications of coronavirus disease-2019 (COVID-19) are still being established (1). Expert guidelines recommend the use of cardiac imaging in the management of patients with COVID-19 (2), and cardiac magnetic resonance (CMR) has shown utility in the noninvasive detection of myocardial inflammation (3). We present a case series of 16 patients who recovered from COVID-19 who underwent CMR to assess for evidence of myocardial involvement or ongoing myocarditis.

Ethics approval was obtained from the Hong Kong West Cluster (UW20-359) Institutional Review Board for this retrospective study. Inclusion criteria were COVID-19 patients admitted as inpatients to Queen Mary Hospital, referred for outpatient CMR post-recovery for raised troponin levels or electrocardiogram changes during the acute illness. Exclusion criteria were poor-quality CMR preventing assessment of ventricular function and late gadolinium enhancement (LGE). COVID-19 was diagnosed based on reverse transcription polymerase chain reaction test results of nasopharyngeal and throat swabs. Recovered COVID-19 status was based on: 1) 2 negative nasopharyngeal swab reverse transcription polymerase chain reaction results >24 h apart; and 2) absence of fever and improvement in respiratory symptoms. COVID-19 disease severity was defined according to World Health Organization criteria (4). CMR performed at 1.5-T (GE Healthcare Systems, Chicago, Illinois) included cine, native T1-mapping (SMART₁), T2-mapping, and LGE. T1/T2–mapping were analyzed in the mid-ventricular slice for an average value per patient. Images were reviewed independently by 3 cardiac radiologists.

Sixteen patients were identified (median age 68 years [interquartile range: 53 to 69 years]; 7 female subjects). Fifteen (94%) of the 16 patients had mild/moderate World Health Organization–defined disease severity. On admission, 14 (88%) had electrocardiogram changes, and 7 (44%) had raised troponin levels. At ≥2 weeks’ post-discharge, 11 (69%) patients were asymptomatic. Five (31%) had symptoms such as cough, shortness of breath, and mild chest pain.

CMR was performed at a median of 56 days’ post-recovery. Three (19%) patients had nonischemic LGE with elevated global T2-mapping values (57 to 62 ms), fulfilling the Lake Louise criteria for myocardial inflammation (3): 1 had chest discomfort with mildly elevated C-reactive protein (CRP) levels; 1 was asymptomatic but with elevated troponin levels (Figure 1 ); and 1 was asymptomatic with no blood biomarkers of inflammation. The fourth patient with LGE had a known history of non–ST-segment elevation myocardial infarction with circumflex artery stenting, showing a lateral wall infarct but no myocarditic changes. In the remainder (all 12 without LGE), 4 patients had elevated T1 only, 1 had elevated T2 only, and 1 had both elevated T1 and T2. Of these, 4 of 6 had blood biomarkers of inflammation (high white blood cell count, CRP, or troponin), and 3 of 6 had ongoing symptoms (1 cough, 1 cough/shortness of breath, and 1 shortness of breath/chest discomfort). The remaining 6 had normal T1 and T2 and no LGE; 5 of 6 were asymptomatic. Two of these 5 patients still had elevated troponin levels, 1 of 5 had elevated CRP levels, and 2 of 5 had normal blood test results. None had pericardial thickening or effusion.

Figure 1.

CMR Images From a Recovered Asymptomatic COVID-19 Patient With Myocardial Inflammation

(A and B) High global T1- and T2-mapping values. (C) Short-axis cine. (D) Small, subepicardial, basal anterolateral wall late gadolinium enhancement (arrow). (E) Cardiac magnetic resonance (CMR) results. ¶ = 1 patient had borderline dilated right ventricle (RV) and dilated main pulmonary artery (37 mm), with no initial suspicion of pulmonary embolus, and a ventilation/perfusion scan post-CMR was normal. #p < 0.02 when compared to 15 healthy volunteers with a mean T1 of 1,158 25 ms (2SD range 1,190 to 1,208 ms). ##p < 0.01 when compared to 15 healthy volunteers with a mean T2 of 48.2 ± 3.4 ms (2SD range 41.5 to 54.8 ms). COVID-19 = coronavirus disease-2019; IQR = interquartile range; LV = left ventricular; V/Q = ventilation/perfusion.

Our study describes subclinical ongoing or resolving myocardial inflammation in patients recovered from COVID-19, as revealed by CMR. A study from Wuhan, China, reported that 58% of patients who recovered from COVID-19 had abnormal CMR findings but all had cardiac symptoms (5). In contrast, our study extends that although 69% (11 of 16) of patients who recovered from COVID-19 were asymptomatic, a majority (56% [9 of 16]) exhibited abnormal CMR findings (high T1 and/or T2, ± nonischemic LGE), 67% (6 of 9) of whom had accompanying blood biomarkers of ongoing inflammation, even if asymptomatic (3 of 6). In asymptomatic patients, 45% (5 of 11) had abnormal CMR findings; 27% (3 of 11) of asymptomatic patients also had corroborating serological evidence of inflammation. In symptomatic patients, 80% (4 of 5) had abnormal CMR findings (high T1 and/or T2), 75% (3 of 4) of whom had corroborating serological evidence of ongoing inflammation. Overall, 6 of 16 (38%) patients had both imaging and serological evidence of myocardial inflammation, and may need follow-up within their individual clinical context. Three (19%) patients had either high T1 and/or T2 on CMR but without blood biomarkers of inflammation; the abnormal T1 or T2 signals may represent residual or resolving myocardial inflammation. Thus, in patients with COVID-19 deemed to have recovered, there remains a high index of suspicion of initial and ongoing myocardial inflammation, and CMR has demonstrable utility in identifying subclinical myocardial involvement post–COVID-19.