Background: This study aimed to evaluate the prognostic value of key admission biomarkers in predicting mortality among hospitalized COVID-19 patients and to establish optimal cut-off thresholds for clinical decision-making.
Methods: Retrospective cohort study included 269 COVID-19 patients treated at Thu Duc City Hospital, Vietnam, during the peak of the fourth pandemic wave in 2021. Logistic regression identified independent predictors of mortality, and receiver operating characteristic (ROC) curve analysis assessed the diagnostic performance of biomarkers. The area under the ROC curve (AUROC), Sensitivity, Specificity and Accuracy Index were used to determine optimal cut-off values.
Results: Among the 269 patients, 53 (19.7%) died and 216 (80.3%) survived. Non-survivors exhibited elevated D-dimer (4.48 μg/mL vs 0.93 μg/mL, p < 0.0001), neutrophil counts (6.8 × 10⁹/L vs 3.5 × 10⁹/L, p < 0.0001) and white blood cell counts (11.68 × 10⁹/L vs. 7.87 × 10⁹/L, p < 0.0001). Lymphocyte counts and fibrinogen levels were significantly lower in non-survivors (p < 0.05). Logistic regression identified D-dimer (OR = 1.05, 95% CI: 1.02–1.09, p = 0.001), neutrophil counts (OR = 1.32, 95% CI: 1.10–1.63, p = 0.005) and lymphocyte counts (OR = 0.51, 95% CI: 0.26–0.92, p = 0.033) as significant predictors of mortality. ROC analysis revealed that D-dimer (AUROC = 0.809) and neutrophil counts (AUROC = 0.726) demonstrated strong discriminatory power, with cut-off values of ≥1.126 μg/mL (sensitivity = 90.57%, specificity = 60.19%) and ≥6.715 × 10⁹/L (sensitivity = 52.83%, specificity = 82.87%), respectively.
Conclusion: These findings support the use of admission biomarkers to guide early interventions and improve patient outcomes in severe COVID-19 cases. Further studies are warranted to validate these results and explore their applicability in other settings.