Accuracy of pulse oximetry in the intensive care unit

Objective: Pulse oximetry (SpO₂) is a standard monitoring device in intensive care units (ICUs), currently used to guide therapeutic interventions. Few studies have evaluated the accuracy of SpO₂ in critically ill patients. Our objective was to compare pulse oximetry with arterial oxygen saturation (SaO₂) in such patients, and to examine the effect of several factors on this relationship. Design: Observational prospective study. Setting: A 26-bed medical ICU in a university hospital. Patients: One hundred two consecutive patients admitted to the ICU in whom one or serial arterial blood gas analyses (ABGs) were performed and a reliable pulse oximeter signal was present. Interventions: For each ABG, we collected SaO2, SpO₂, the type of pulse oximeter, the mode of ventilation and requirement for vasoactive drugs. Measurements and results: Three hundred twenty-three data points were collected. The mean difference between SpO₂ and SaO₂ was -0.02% and standard deviation of the differences was 2.1%. From one sample to another, the fluctuations in SpO₂ to arterial saturation difference indicated that SaO₂ could not be reliably predicted from SpO₂ after a single ABG. Subgroup analysis showed that the accuracy of SpO₂ appeared to be influenced by the type of oximeter, the presence of hypoxemia and the requirement for vasoactive drugs. Finally, high SpO₂ thresholds were necessary to detect significant hypoxemia with good sensitivity. Conclusion: Large SpO₂ to SaO₂ differences may occur in critically ill patients with poor reproducibility of SpO₂. A SpO₂ above 94% appears necessary to ensure a SaO₂ of 90%.