The Relationship Between ROX Index and Inflammatory Markers In COVID-19 Patients

Background: The ROX index is a bedside measure used to assess respiratory status during noninvasive oxygen support, but its relationship with inflammatory burden in COVID-19 remains incompletely characterized. This study evaluated the association between early ROX values and inflammatory markers in hospitalized adults with COVID-19 receiving high-flow nasal cannula (HFNC) or bilevel positive airway pressure (BPAP).

Methods: This retrospective single-center cohort included 994 adults hospitalized with COVID-19 between 2020 and 2022. Primary analyses were limited to patients with documented ROX values at 2, 6, or 12 hours after HFNC or BPAP initiation. Spearman correlation was used to assess the association between ROX and the highest in-hospital values of C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), d-dimer, interleukin-6 (IL-6), interleukin-10 (IL-10), ferritin, lactate dehydrogenase (LDH), white blood cell count (WBC), fibrinogen, and neutrophil-to-lymphocyte ratio (NLR). Exploratory secondary analyses examined correlations with ICHIKADO score, APACHE score, SOFA score, hospital length of stay, invasive mechanical ventilation duration, and clinical outcome.

Results: During HFNC, lower ROX values were significantly associated with higher ferritin, ESR, d-dimer, LDH, and WBC at 2 hours; higher ESR, IL-6, LDH, NLR, and WBC at 6 hours; and higher IL-6, LDH, NLR, and WBC at 12 hours. During BPAP, lower ROX correlated with d-dimer at 2 hours and with ferritin and d-dimer at 6 hours, whereas no significant inflammatory-marker association was observed at 12 hours. CRP, fibrinogen, and IL-10 were not significantly associated with ROX at any time point. Lower HFNC ROX values also correlated with higher ICHIKADO, SOFA, and APACHE scores, and lower ROX values in both HFNC and BPAP groups were associated with worse clinical outcome.

Conclusions: In this retrospective cohort, lower early ROX values were associated with greater inflammatory burden and higher illness severity, particularly among patients receiving HFNC. These
findings should be interpreted as associative rather than predictive and warrant prospective validation
with serial biomarker sampling, standardized outcome definitions, and multivariable adjustment.