AimsTo investigate the characteristics and clinical significance of myocardial injury in patients with severe coronavirus disease 2019 (COVID-19).Methods and resultsWe enrolled 671 eligible hospitalized patients with severe COVID-19 from 1 January to 23 February 2020, with a median age of 63 years. Clinical, laboratory, and treatment data were collected and compared between patients who died and survivors. Risk factors of death and myocardial injury were analysed using multivariable regression models. A total of 62 patients (9.2%) died, who more often had myocardial injury (75.8% vs. 9.7%; P < 0.001) than survivors. The area under the receiver operating characteristic curve of initial cardiac troponin I (cTnI) for predicting in-hospital mortality was 0.92 [95% confidence interval (CI), 0.87–0.96; sensitivity, 0.86; specificity, 0.86; P < 0.001]. The single cut-off point and high level of cTnI predicted risk of in-hospital death, hazard ratio (HR) was 4.56 (95% CI, 1.28–16.28; P = 0.019) and 1.25 (95% CI, 1.07–1.46; P = 0.004), respectively. In multivariable logistic regression, senior age, comorbidities (e.g. hypertension, coronary heart disease, chronic renal failure, and chronic obstructive pulmonary disease), and high level of C-reactive protein were predictors of myocardial injury.ConclusionThe risk of in-hospital death among patients with severe COVID-19 can be predicted by markers of myocardial injury, and was significantly associated with senior age, inflammatory response, and cardiovascular comorbidities.