Evaluation of four novel prognostic scores on admission for COVID-19 mortality. An experience from a Mediterranean tertiary center

Velissaris, Dimitrios; Paraskevas, Themistoklis; Oikonomou, Eleousa; Bizos, Aristeidis; Karamouzos, Vasileios; Marangos, Markos

AIM: To assess the performance of four novel prognostic scores on admission in predicting in-hospital mortality in patients with confirmed SARS-CoV-2 infection and compare it to NEWS2 and respiratory SOFA score. METHODS: A total of 85 adult patients admitted to a tertiary hospital in Western Greece with positive SARS-CoV-2 PCR test, were enrolled and divided into the non-survivor (n = 10) and survivor (n = 75) groups. Receiver Operating Characteristic (ROC) analysis was conducted to determine the predictive effect of the COVID-19 Mortality Score, COVID-19 Severity Index, 4 C Mortality Score and COVID-IRS NLR. Subsequently, they were compared to the respiratory component of the SOFA score and NEWS2. RESULTS: ROC curve analysis showed that the COVID-19 Mortality Score (score ≥4) had the highest combination of sensitivity and specificity values for predicting in-hospital mortality (Sensitivity = 0.8, Specificity = 0.853). The Area Under Curve (AUC) for predicting in hospital mortality for the COVID-19 Mortality Score, COVID-19 Severity Index, 4 C Mortality Score and COVID-IRS NLR were 0.846, 0.815, 0.789 and 0.787, respectively. Comparison between the AUC of the four novel COVID-19 scores, respiratory SOFA and NEWS2 showed no significant differences. CONCLUSION: All four novel prognostic scores had acceptable to excellent AUC values for predicting in hospital mortality. Out of the four novel prognostic scores for patients with COVID-19, the COVID-19 mortality score showed the best results in our cohort. Its prognostic ability was superior to that of the NEWS2 and respiratory SOFA score.