The predictive value of cardiac biomarkers and echocardiography in critical COVID-19

Abstract

Abstract Background Coronavirus disease (COVID-19) continues to challenge healthcare systems worldwide. Adequate triage and divergence to an appropriate level of care are essential. Myocardial injury is a common finding in COVID-19 patients, while cardiac complications seem to occur only in a minority of patients. Purpose To evaluate cardiac biomarkers and echocardiographic findings in critical COVID-19 at time of ICU admission, and to assess their association with ICU mortality in comparison to other biomarkers and risk factors. Methods Prospective, single-center, cohort study in patients with PCR-confirmed COVID-19 admitted to the ICU. Laboratory assessment included high sensitive troponin T (hsTnT) and N-terminal pro-brain natriuretic peptide (NT-proBNP). Standard transthoracic echocardiographic evaluation was performed upon ICU admission. The primary outcome was ICU mortality. Statistical analysis was performed using SPSS statistics (Version 27.0, IBM Corp, Armonk, NY) with χ2 and Mann–Whitney U tests. Predictive markers for mortality were assessed by ROC analysis and cut-off values by the Youden Index. Results We included 86 patients. Transthoracic echocardiography was not feasible in 11 patients (12.8%) due to poor visualization or prone ventilation. Baseline characteristics are presented in the table. Left ventricular ejection fraction (LVEF) was below 50% in 18.7% of patients and 20% had an E/e' septal ratio above 14. Right ventricular function, evaluated with TAPSE, was normal (>14mm) in 94% of patients. Cardiac biomarkers were elevated in almost half of all patients (hsTnT ≥14 μg/L in 46.5%, and NT-proBNP ≥450 pg/mL in 44,2%), and in up to 30% of patients with normal LVEF. The level of these cardiac biomarkers was significantly higher in non-survivors, while other routinely used biomarkers (D-dimers, ferritin, C-reactive protein), and commonly used clinical scores indicating the severity of illness (SOFA score) were not. A cut-off value of 16.5 μg/L for hsTnT corresponded with sensitivity and specificity for mortality of resp. 75% and 74.2%. For NT-proBNP, a cut-off value of 760 pg/mL corresponded with sensitivity and specificity for mortality of resp. 60% and 83.3%. LVEF was significantly lower and E/e' ratio was significantly higher in non-survivors. Conclusion Echocardiographic abnormalities, especially left and not right ventricular dysfunction, were found in up to 20% of patients, whereas up to half of all patients had elevated cardiac biomarkers at time of ICU admission. This study highlights the strong predictive value of cardiac biomarkers for ICU mortality and their possible role for triage in critical COVID-19. They outperform routinely used biomarkers of inflammation as well as clinical indices of disease severity in ICU. Transthoracic echocardiography should only be considered if it is likely to impact therapeutic management given the potential hazard for sonographers and the scarcity of personal protective equipment. Funding Acknowledgement Type of funding sources: None.