Myoglobin Offers Higher Accuracy Than Other Cardiac-Specific Biomarkers for the Prognosis of COVID-19

Abstract

Although sporadic studies have shown that myoglobin may have better prognostic performance than other cardiac markers in COVID-19, a comprehensive comparative study is lacking. Herein, we retrospectively analyzed the clinical and laboratory data of COVID-19 patients admitted to the Guanggu Campus of Wuhan Tongji Hospital from February 9, 2020 to March 30, 2020, intending to compare the prognostic accuracy of three commonly used cardiac markers on COVID-19 mortality. Our results revealed that abnormal increases in myocardial biomarkers were associated with a significantly increased risk of in-hospital mortality with COVID-19. Interestingly, myoglobin, a non-cardiac-specific biomarker, also expressed in skeletal myocytes, had even higher prognostic accuracy than cardiac-specific biomarkers such as high-sensitivity troponin I (hs-TnI) and creatine kinase-MB (CK-MB). More importantly, multivariate Cox analysis showed that myoglobin, rather than hs-TnI or CK-MB, was independently prognostic for in-hospital mortality in COVID-19. These results were further confirmed by subgroup analyses of patients with severe and critical illnesses and those without a history of cardiovascular disease. Our findings suggest that myoglobin may be a reliable marker of illness reflecting general physiological disturbance and help to assess prognosis and treatment response in patients with COVID-19.

Keywords: COVID-19; in-hospital mortality; myocardial biomarkers; myoglobin; rhabdomyolysis.

Figure 1

Summary of myocardial biomarker levels in patients with COVID-19. (A–C) The early- and late-stage levels of myocardial biomarkers in patients stratified by mortality. (D–F) The early- and late-stage levels of myocardial biomarkers in patients stratified by clinical classifications.

Figure 2

Time-dependent ROC curve analysis based on the early-stage levels of myocardial biomarkers showing the prognostic performance of myocardial biomarkers. (A–C) ROC performance of myocardial biomarkers at different times during follow-up. (D,E) Comparison of the prognostic performance of each biomarker in predicting 14-day and 28-day mortality. (F) Time-dependent AUC curves of myocardial markers showed superior AUC values for myoglobin over hs-TnI and CK-MB throughout the follow-up period. ROC, receiver operating characteristic curve; AUC, area under curve.

Figure 3

Kaplan-Meier curves showing the cumulative survival rates of patient groups divided by cut-off and ULN. (A–C) Kaplan-Meier curve analysis based on the early-stage levels of myocardial biomarkers. (D–F) Kaplan-Meier curve analysis based on the late-stage levels of myocardial biomarkers. ULN, upper limit of normal.

Figure 4

Forest plots showing the results of multivariate Cox analysis. The potential prognostic factors included in the multivariate analysis were previously identified by LASSO regression. (A) LASSO coefficient profiles of the 30 variables including age, sex, comorbidities, physical examinations, and early-stage results of laboratory parameters. Dotted vertical lines were drawn at the optimal values by using the minimum criteria and the 1 SE of the minimum criteria (the 1-SE criteria). Five variables with non-zero coefficients (MYO, NEU, hs-CRP, IL-6, and D-dimer) were identified. (B) In multivariate analysis, all five variables had p-values < 0.05. (C) LASSO coefficient profiles of the 30 variables including age, sex, comorbidities, physical examinations, and late-stage results of laboratory parameters. Three variables (MYO, hs-CRP, and D-dimer) with non-zero coefficients were identified. (D) In multivariate analysis, they all had p-values less than 0.05. LASSO, least absolute shrinkage and selection operator; MYO, myoglobin; NEU, neutrophil; hs-CRP, high sensitivity C-reactive protein; IL-6, interleukin 6.

Figure 5

Subgroup analysis demonstrating the prognostic ability of MYO in COVID-19 patients with severe or critical conditions. The results confirmed that MYO provided a better prognostic performance than hs-TnI and CK-MB and had an independent prognostic effect on in-hospital mortality. (A) Standard ROC curve analysis based on the early-stage levels of biomarkers. (B) Time-dependent AUC curves based on the early levels of cardiac markers. (C) Standard ROC curve analysis based on the late-stage levels of biomarkers. (D–F) LASSO regression and multivariate COX analysis based on early levels of biomarkers. (G–I) LASSO regression and multivariate COX analysis based on late levels of biomarkers. LASSO, least absolute shrinkage and selection operator; MYO, myoglobin; NEU, neutrophil; hs-CRP, high sensitivity C-reactive protein; LYM, lymphocyte; ALB, albumin; N, number; AUC, area under curve.