1. On-admission laboratory predictors for developing critical COVID-19 during hospitalization - a multivariable logistic regression model.
- Author
-
Schmidt W, Jóźwiak B, Czabajska Z, Pawlak-Buś K, and Leszczynski P
- Subjects
- Hospitalization, Humans, Logistic Models, Retrospective Studies, Risk Factors, SARS-CoV-2, Troponin I, COVID-19
- Abstract
Introduction and Objective: Recognition of patients with COVID-19 who will progress clinically and need respiratory support remains challenging. The aim of the study was to identify abnormalities in on-admission laboratory results that can precede progression from moderate or severe to critical COVID-19., Material and Methods: Laboratory data analyzed of 190 patients admitted with moderate or severe COVID-19 to our ward. Laboratory results taken into analysis were obtained during the first 48 hours of hospitalization. Multivariate logistic regression was performed using risk factors obtained in the univariate analysis as dependent variables., Results: 42 patients were identified who developed critical COVID-19. In univariate analysis, 22 laboratory risk factors were detected that were used in logistic regression and in building model with following predictors: high-sensitive troponin I concentration (hs-TnI) >26 ng/mL (OR 13.45; 95%CI 3.28-55.11; P 15 (OR 5.67; 95%CI 1.97-16.36, P 50 pg/mL (OR 5.52; 95%CI 1.86-16.37; P = 0.001), fasting glycaemia >6.8 mmol/L (OR 4.74; 95%CI 1.65-13.66; P = 0.002), immature neutrophils count >0.06/µL (OR 4.06; 95%CI 1.35-12.2; P = 0.012) and urine protein concentration >500 mg/L (OR 2.94; 95%CI 1.04-8.31; P = 0.043)., Conclusions: The most significant risk factors of developing critical COVID-19 during hospitalization are: elevated hs-TnI, IL-6, and glucose serum concentrations, increased immature neutrophil count, neutrophils to monocytes ratio, and proteinuria during the first 48 hours after admission. The model built with these predictors achieved better predictive performance than any other univariately analysed laboratory markers in predicting the critical development COVID-19.
- Published
- 2022
- Full Text
- View/download PDF