Background: Recent literature suggests that the respiratory rate oxygenation (ROX) index may be useful in predicting intubation indication in hypoxemic patients.
Objective: In this study, we evaluated the accuracy of the ROX index in predicting intubation, length of stay in the ICU, and mortality in ICU patients with hypoxemic respiratory failure with and without hypercapnia.
Methods: A single-centre retrospective cohort study of 290 patients with a preliminary diagnosis of respiratory failure were treated with low-flow oxygen systems. Demographics, medical history, clinical, laboratory, treatment, and prognostic data were obtained from the hospital's electronic records. The ROX index was calculated at the time of ICU admission.
Results: Thirty-seven percent of non-hypercapnic and 69% of hypercapnic patients were intubated (p:0.005). In hypercapnic patients, ROX: 6.9 had highest sensitivity (81%) and specificity (65%) values for intubation (p:0.005). In non-hypercapnic patients, ROX: 6.2 had the highest sensitivity (81%) and specificity (40%) values. While 11% of hypercapnic patients and 30% of non-hypercapnic patients died (p:0.05), 22% of hypercapnic patients and 33% of non-hypercapnic patients stayed in the ICU longer than 14 days (p:0.044). The highest sensitivity and specificity values were found for mortality in hypercapnic patients; for ROX value of 5.94 (sensitivity:86%, specificity:61%) and ICU stay longer than 14 days; for ROX value of 7.4 (sensitivity:71%, specificity:68%).
Conclusion: Results of our study suggest that the ROX index calculated during ICU admission can help predict intubation indication and length of ICU stay in patients with respiratory failure and hypercapnia may influence the cutoff values.
Keywords: ROX index, PaO2/FiO2, respiratory failure, intubation, intensive care unit