The coronavirus disease 19 (COVID-19) has become a global health event. Cardiac biomarkers like creatine kinase isoenzyme (CK-MB), myoglobin, and high-sensitivity troponin T were usually elevated in early stages. This study aimed to investigate whether the elevated cardiac biomarkers could become effective prognostic predictors for COVID-19 patients.

The present study involved 357 COVID-19 patients. The potential predictors for two study outcomes (in-hospital death and recovery status) in 28 days were selected by LASSO regression analysis. Prognostic values of cardiac biomarkers selected were evaluated using the receiver operating characteristic curve (ROC) and the area under ROC (AUC).

After 28-day follow-up, overall 357 patients were divided into death group (n = 25) and survival group (n = 332), or non-recovery group (n = 43) and recovery group (n = 314). The LASSO regression analysis showed elevated CK-MB and myoglobin were independent risk predictors for in-hospital death, and CK-MB and myoglobin were also independent risk predictors for non-recovery. The AUC of CK-MB and myoglobin for in-hospital death were 0.862 (95%CL: 0.804–0.920, p < 0.001) and 0.838 respectively (95%CL: 0.729–0.947, p < 0.001). The AUC of CK-MB and myoglobin for non-recovery were 0.839 (95%CL: 0.786–0.892, p < 0.001) and 0.841 (95%CL: 0.765–0.918, p < 0.001) respectively. We also found AUC of combined use of CK-MB and myoglobin for in-hospital death and non-recovery were 0.883 (95CL: 0.813–0.952, p < 0.001), and 0.873 (95%CL: 0.817–0.930, p < 0.001) respectively.

In patients with COVID-19, elevated CK-MB and myoglobin on admission may be effective predictors for adverse outcomes, and combined use of CK-MB and myoglobin had a better performance for prediction.