Admission Predictors of Mortality in Hospitalized COVID-19 Patients-A Serbian Cohort Study.

ABSTRACT:

Background

Early prediction of COVID-19 patients' mortality risk may be beneficial in adequate triage and risk assessment. Therefore, we aimed to single out the independent morality predictors of hospitalized COVID-19 patients among parameters available on hospital admission.

Methods

An observational, retrospective-prospective cohort study was conducted on 703 consecutive COVID-19 patients hospitalized in the University Clinical Center Kragujevac between September and December 2021. Patients were followed during the hospitalization, and in-hospital mortality was observed as a primary end-point. Within 24 h of admission, patients were sampled for blood gas and laboratory analysis, including complete blood cell count, inflammation biomarkers and other biochemistry, coagulation parameters, and cardiac biomarkers. Socio-demographic and medical history data were obtained using patients' medical records.

Results

The overall prevalence of mortality was 28.4% (n = 199). After performing multiple regression analysis on 20 parameters, according to the initial univariate analysis, only four independent variables gave statistically significant contributions to the model: SaO2 < 88.5 % (aOR 3.075), IL-6 > 74.6 pg/mL (aOR 2.389), LDH > 804.5 U/L (aOR 2.069) and age > 69.5 years (aOR 1.786). The C-index of the predicted probability calculated using this multivariate logistic model was 0.740 (p < 0.001).

Conclusions

Parameters available on hospital admission can be beneficial in predicting COVID-19 mortality.

SUBMITTER: Poskurica M 

PROVIDER: S-EPMC9605560 | biostudies-literature | 2022 Oct

REPOSITORIES: biostudies-literature