Characterization of Myocardial Injury in Patients With COVID-19

Gennaro Giustino, MD; Lori B. Croft, MD; Giulio G. Stefanini, MD, PHD; Renato Bragato, MD; Jeffrey J. Silbiger, MD; Marco Vicenzi, MD; Tatyana Danilov, MD; Nina Kukar, MD; Nada Shaban, MD; Annapoorna Kini, MD; Anton Camaj, MD, MS; Solomon W. Bienstock, MD; Eman R. Rashed, MD; Karishma Rahman, MD, PHD; Connor P. Oates, MD; Samantha Buckley, BS; Lindsay S. Elbaum, MD; Derya Arkonac, MD; Ryan Fiter, MD; Ranbir Singh, MD; Emily Li, MD; Victor Razuk, MD; Sam E. Robinson, MD; Michael Miller, MS; Benjamin Bier, MD; Valeria Donghi, MD; Marco Pisaniello, MD; Riccardo Mantovani, MD; Giuseppe Pinto, MD; Irene Rota, MD; Sara Baggio, MD; Mauro Chiarito, MD; Fabio Fazzari, MD; Ignazio Cusmano, MD; Mirko Curzi, MD; Richard Ro, MD; Waqas Malick, MD; Mazullah Kamran, MD; Roopa Kohli-Seth, MD; Adel M. Bassily-Marcus, MD; Eric Neibart, MD; Gregory Serrao, MD; Gila Perk, MD; Donna Mancini, MD; Vivek Y. Reddy, MD; Sean P. Pinney, MD; George Dangas, MD, PHD; Francesco Blasi, MD, PHD; Samin K. Sharma, MD; Roxana Mehran, MD; Gianluigi Condorelli, MD; Gregg W. Stone, MD; Valentin Fuster, MD, PHD; Stamatios Lerakis, MD, PHD; Martin E. Goldman, MD


J Am Coll Cardiol. 2020;76(18):2043-2055. 

In This Article


Study Design

The Cardiac Injury Research in COVID-19 (CIRC-19) registry is an international, multicenter retrospective cohort study of hospitalized patients with confirmed severe acute respiratory syndrome-coronavirus-2 infection who underwent a transthoracic echocardiographic (TTE) evaluation during their index hospitalization at 7 clinical sites in New York City (United States) and Milan (Italy) between March 5, 2020, and May 2, 2020. Patients who did not have confirmed severe acute respiratory syndrome-coronavirus-2 infection (by polymerase chain reaction assay of nasal or pharyngeal swab specimens or serologic testing) and those who did not undergo a full TTE study were excluded. Patients who only had point-of-care cardiac ultrasound were not included. Approval for the study was obtained from each center's Institutional Review Board.

Data Collection and Endpoints

Data was collected from each center's electronic health record and included patient demographic information, presenting vital signs and symptoms, comorbidities, home medications, chest x-ray findings, ECG findings, laboratory values (reference values are reported in Supplemental Table 1), echocardiographic findings, inpatient treatments received, and in-hospital outcomes. Patients were then categorized according to the presence or absence of myocardial injury, defined as a serum cardiac troponin above the upper reference limit for the assay used at each participating site. Echocardiographic data examined included left ventricular (LV) ejection fraction, LV volumes, presence of regional wall motion abnormalities or global LV dysfunction, LV diastolic function, right ventricular (RV) size and function, and presence of pericardial effusions. Definitions of echocardiographic values are reported in Supplemental Tables 2 to 5. We defined "major echocardiographic abnormalities" as the composite of LV wall motion abnormalities, LV global dysfunction, LV grade II or III diastolic dysfunction, RV dysfunction or presence of a small or larger pericardial effusion. The primary clinical endpoint of interest was in-hospital all-cause mortality. Additional endpoints of interest included admission to an intensive care unit, need for mechanical ventilation, acute respiratory distress syndrome (ARDS), stroke, acute kidney injury (AKI), shock, and ventricular fibrillation or ventricular tachycardia. We defined ARDS according to the Berlin definition.[16] AKI was defined according to the Kidney Disease: Improving Global Outcomes definition.[17] All endpoints were site-reported.

Statistical Analysis

Continuous variables are reported as median (interquartile range) and were compared with the Wilcoxon rank sum test. Categorical variables are reported as percentages and were compared using the chi-square test. The Kaplan-Meier method was used to generate failure curves for descriptive purposes with censoring performed at either the date of discharge, date of last follow-up, or date of death. Multivariable logistic regression models were performed to evaluate the association between myocardial injury and mortality alone and with or without the presence of major echocardiographic abnormalities. The following covariates were included in the multivariable logistic regression model: age; sex; race; Hispanic ethnicity; history of heart failure; ARDS; AKI stage II or III; cardiocirculatory shock; myocardial injury (with or without major echocardiographic abnormalities), and center identifier. Results of the logistic regression models are reported as odds ratio (OR) and corresponding 95% confidence intervals (CIs). Multivariable Cox regression models for in-hospital death were also performed and the results were reported with hazard ratios and 95% CIs. Center identifiers were entered in the multivariable models to account for intercenter heterogeneity.

In separate analyses, we evaluate the characteristics and outcomes of subsets of patients according to the presence of major echocardiographic abnormalities. Also, we reported the clinical, echocardiographic characteristics and outcomes of those with confirmed ACS on coronary angiography defined as confirmed thrombotic lesion of a major epicardial coronary artery versus other types of myocardial injury. All analyses were performed with the use of Stata software version 14.2 (IBM Corp., Armonk, New York).