Abstract
Background Pulse oximetry is widely used in the clinical setting. The purpose of this validation study was to investigate the level of agreement between oxygen saturations measured by pulse oximeter (SpO(2)) and arterial blood gas (SaO(2)) in a range of oximeters in clinical use in Australia and New Zealand. Methods Paired SpO(2) and SaO(2) measurements were collected from 400 patients in one Australian and two New Zealand hospitals. The ages of the patients ranged from 18 to 95 years. Bias and limits of agreement were estimated. Sensitivity and specificity for detecting hypoxaemia, defined as SaO(2) < 90%, were also estimated. Results The majority of participants were recruited from the Outpatient, Ward or High Dependency Unit setting. Bias, oximeter-measured minus arterial blood gas-measured oxygen saturation, was - 1.2%, with limits of agreement - 4.4 to 2.0%. SpO(2) was at least 4% lower than SaO(2) for 10 (2.5%) of the participants and SpO(2) was at least 4% higher than the SaO(2) in 3 (0.8%) of the participants. None of the participants with a SpO(2) >= 92% were hypoxaemic, defined as SaO(2) < 90%. There were no clinically significant differences in oximetry accuracy in relation to clinical characteristics or oximeter brand. Conclusions In the majority of the participants, pulse oximetry was an accurate method to assess SaO(2) and had good performance in detecting hypoxaemia. However, in a small proportion of participants, differences between SaO(2) and SpO(2) could have clinical relevance in terms of patient monitoring and management. A SpO(2) >= 92% indicates that hypoxaemia, defined as a SaO(2) < 90%, is not present.