A nomogram for predicting ARDS progression and need for invasive ventilation in COVID-19 patients using plasma hyaluronic acid and clinical scores

Authors:
Open Access
Publisher website
Google Scholar
Abstract
Background: Acute respiratory distress syndrome (ARDS) is a major complication in hospitalized patients with coronavirus disease 2019 (COVID-19). Early identification of patients at increased risk of ARDS progression and invasive ventilation remains clinically important. This study aimed to investigate whether plasma hyaluronic acid (HA), alone or in combination with established clinical severity scores, could help identify ARDS, stratify severe disease, and assess the risk of endotracheal intubation in patients with COVID-19. Methods: This retrospective single-center cohort study included 502 adult patients with COVID-19 admitted between September 2022 and February 2023. Plasma HA levels were measured within 24 hours after admission. Demographic characteristics, comorbidities, laboratory findings, clinical severity scores, and outcome data were collected. Receiver operating characteristic analysis, multivariable logistic regression, and nomogram construction were used to evaluate the association of HA with ARDS diagnosis, severe ARDS stratification, and intubation risk. Results: This retrospective cohort analysis of 502 COVID-19 patients (361 in the non-ARDS group and 141 in the ARDS group) identified plasma hyaluronic acid (HA) as a potential biomarker for early ARDS diagnosis, severity stratification, and intubation risk assessment. HA levels were positively correlated with disease severity, with concentrations significantly increasing alongside ARDS severity, and showed good discrimination for ARDS (AUC = 0.904; sensitivity 81.6%, specificity 87.5% at a cut-off of 103 μg/L). HA also demonstrated excellent discrimination for severe ARDS, with an AUC of 0.953, comparable to that of the APACHE II and SOFA scores. It also showed good discrimination for endotracheal intubation risk in the overall cohort. The AUC for predicting intubation requirement reached 0.890 (sensitivity 76.6%, specificity 90.8%, NPV = 95.5% at a cut-off value of 140.1 μg/L). Integrating HA with clinical scores further improved model performance; incorporating age, lymphocyte count, CRP, CK, APACHE II, and SOFA into the baseline model increased the AUC for predicting severe ARDS from 0.960 to 0.973 (ΔAUC = 0.013). Furthermore, multivariable regression analysis showed that HA was independently associated with ARDS diagnosis (OR = 1.04, P = 0.007), severe ARDS (OR = 1.00, P < 0.001), and intubation risk (OR = 1.00, P = 0.011). Conclusion: Plasma HA was positively associated with ARDS severity and intubation risk in patients with COVID-19. HA may serve as a complementary biomarker for early risk stratification, particularly when used in combination with established clinical severity scores.

This publication has 26 references indexed in Scilit: