Evaluation of fingertip pulse oximeters for monitoring haemoglobin oxygen saturation in arterial blood and pulse rate in isoflurane-anaesthetised horses breathing greater than 90 percent oxygen

Background Pulse oximetry is a noninvasive method for continuous monitoring of oxygen saturation (SpO(2)) and pulse rate (PR) used in equine anaesthesia. Despite species differences in the oxyhaemoglobin dissociation curve, pulse oximeters validated for humans are commonly used in veterinary medicine. These devices are expensive and at an increased risk for damage when used for equine anaesthesia. Human fingertip pulse oximeters (FPO) may be useful as an inexpensive alternative to standard devices. Objectives Evaluate accuracy and repeatability of FPOs applied to tongues of anaesthetised horses breathing >90% oxygen. Study design Prospective clinical study. Methods Duplicate SpO(2) and PR measurements from eight pulse oximeters (PO; seven FPOs and reference PO) placed in random order on tongues of horses under inhalational anaesthesia were recorded. A standard portable PO (RPO) was used as reference SpO(2) and invasive blood pressure trace or manual pulse palpation as reference PR. Nonparametric Bland-Altman analysis was performed. Acceptable coefficient of repeatability (CR) value was <= 6% SpO(2). Agreements were accepted if limits of agreement were <= +/- 3% difference in SpO(2) and <= +/- 10% relative difference in PR. Only PR values between 25 and 50 beats per minute and their corresponding SpO(2) measurements were used for agreement analysis. Results SpO(2) repeatability and agreement were unacceptable for all devices. The agreement of PR measurements was acceptable for one FPO and the RPO. Main limitations Utilisation of devices designed for the human oxyhaemoglobin dissociation curve, high number of absent or erroneous readings, use of healthy horses, small sample size, unavailability of co-oximetry, lack of a desaturation protocol and difficulty with proper placement of FPOs on horse tongues are all considered limitations in this study. Conclusions Accuracy of FPOs used in anaesthetised horses was poor. Their use in equine anaesthesia for SpO(2) monitoring is not recommended, but some models may be considered for continuous PR monitoring in conjunction with intermittent manual confirmation.