The ability of Oxygen Reserve Index® to detect hyperoxia in critically ill patients

Hyperoxia is associated with increased morbidity and mortality in the intensive care unit. Classical noninvasive measurements of oxygen saturation with pulse oximeters are unable to detect hyperoxia. The Oxygen Reserve Index (ORI) is a continuous noninvasive parameter provided by a multi-wave pulse oximeter that can detect hyperoxia. Primary objective was to evaluate the diagnostic accuracy of the ORI for detecting arterial oxygen tension (PaO₂) > 100 mmHg in neurocritical care patients. Secondary objectives were to test the ability of ORI to detect PaO₂ > 120 mmHg and the ability of pulse oximetry (SpO₂) to detect PaO₂ > 100 mmHg and PaO₂ > 120 mmHg.

In this single-center study, we collected ORI and arterial blood samples every 6 h for 3 consecutive days. Diagnostic performance was estimated using the area under the receiver operating characteristic curve (AUROC).

There were 696 simultaneous measurements of ORI and PaO₂ in 62 patients. Considering the repeated measurements, the correlation between ORI and PaO₂ was r = 0.13. The area under the receiver operating characteristic curve (AUROC), obtained to test the ability of ORI to detect PaO₂ > 100 mmHg, was 0.567 (95% confidence interval = 0.566-0.569) with a sensitivity of 0.233 (95%CI = 0.230-0.235) and a specificity of 0.909 (95%CI = 0.907-0.910). The AUROC value obtained to test the ability of SpO₂ to detect a PaO₂ > 100 mmHg was 0.771 (95%CI = 0.770-0.773) with a sensitivity of 0.715 (95%CI = 0.712-0.718) and a specificity of 0.700 (95%CI = 0.697-0.703). The diagnostic performance of ORI and SpO₂ for detecting PaO₂ > 120 mmHg was AUROC = 0.584 (95%CI = 0.582-0.586) and 0.764 (95%CI = 0.762-0.766), respectively. The AUROC obtained for SpO₂ was significantly higher than that for ORI (p < 0.01). Diagnostic performance was not affected by sedation, norepinephrine infusion, arterial partial pressure of carbon dioxide, hemoglobin level and perfusion index.

In a specific population of brain-injured patients hospitalized in a neurointensive care unit, our results suggest that the ability of ORI to diagnose hyperoxia is relatively low and that SpO₂ provides better detection.