Age-Dependent Biomarkers for Prediction of In-Hospital Mortality in COVID-19 Patients

Abstract

Background: Several biomarkers and models have been proposed to predict in-hospital mortality among COVID-19 patients. However, these studies have not examined the association in sub-populations. The present study aimed to identify the association between the two most common inflammatory biomarkers in the emergency department and in-hospital mortality in subgroups of patients. Methods: A historical cohort study of adult patients who were admitted to acute-care hospital between March and December 2020 and had a diagnosis of COVID-19 infection. Data on age, sex, Charlson comorbidity index, white blood cell (WBC) count, C-reactive protein (CRP), and in-hospital mortality were collected. Discrimination ability of each biomarker was observed and the CHAID method was used to identify the association in subgroups of patients. Results: Overall, 762 patients (median age 70.9 years, 59.7% males) were included in the study. Of them, 25.1% died during hospitalization. In-hospital mortality was associated with higher CRP (median 138 mg/L vs. 85 mg/L, p < 0.001), higher WBC count (median 8.5 vs. 6.6 K/µL, p < 0.001), and higher neutrophil-to-lymphocyte ratio (NLR) (median 9.2 vs. 5.4, p < 0.001). The area under the ROC curve was similar among all biomarkers (WBC 0.643, NLR 0.677, CRP 0.646, p > 0.1 for all comparisons). The CHAID method revealed that WBC count was associated with in-hospital mortality in patients aged 43.1–66.0 years (<11 K/µL: 10.1% vs. 11+ K/µL: 27.9%), NLR in patients aged 66.1–80 years (≤8: 15.7%, >8: 43.3%), and CRP in patients aged 80.1+ years (≤47 mg/L: 18.8%, 47.1–149 mg/L: 43.1%, and 149.1+: 71.7% mortality). Conclusions: WBC, NLR, and CRP present similar discrimination abilities. However, each biomarker should be considered as a predictor for in-hospital mortality in different age groups.