Patient characteristics and prediction of CoViD-19 in-hospital mortality: a retrospective cohort study in Crete, Greece before and after the onset of a targeted vaccination strategy in 2021

Abstract

Background: CoviD-19 in-hospital mortality rates show variations across time, setting, patient populations and health-systems. Aim: The aim of the present study was to explore the association of patient characteristics with in-hospital mortality taking into consideration temporal changes of CoViD-19 prevention and treatment, focusing on the negative predictive value of specific biomarker thresholds for death and oxygen requirements. Methods: This retrospective study included a sample of 425 vaccine-naive patients hospitalized for CoViD-19 in Crete, Greece. Time-specific mortality rates and negative predictive values of biomarkers on admission were calculated. The association of patient characteristics with mortality was assessed using multivariate logistic regression models. Results: The overall in hospital case fatality rate (CFR) was 19.53%. The in-hospital CFR [95% CI] dropped from 22.74% [17.96-27.52] to 11.9% [6.17-17.63] before and after 31/03/2021, as did the mean age of hospitalized patients. LDH & D-dimers ≥240 or ≥0.5), CRP & D-dimers (≥3 or ≥0.5), CRP & Ferritin (≥3 or ≥205) had negative predictive values [95% CI] for mortality of 100% [86.7-100], 96.1% [80.3-99.9] and 95.4% [84.5-99.4] %, but the respective negative predictive values for oxygen requirements were 19.2% [6.5-39.3], 19.2% [6.5-39.3] and 27.2% [14.9-42.7] . Conclusion: A decreasing trend of in-hospital case fatality rates over the course of our study was associated with a decreasing age of hospitalized patients, which could be associated with community vaccination uptake in people over 60 in Greece through reduction of persons “susceptible” to severe disease. Mortality prediction formulas incorporating biomarkers should be used cautiously for the risk stratification of patients with CoViD-19 in the setting of the emergency department assessment.