Evaluation of pulse oximetry in anaesthetised foals using multiple combinations of transducer type and transducer attachment site.

A commercially available pulse oximeter was evaluated in anaesthetised foals to determine its accuracy for estimating arterial haemoglobin saturation (SaO2). Five different transducer/transducer attachment site (TTAS) combinations were evaluated; 1-3) a fingertip transmission transducer attached to the foal's ear, lip and tongue, 4) an adhesive transmission transducer positioned on the foal's ear and 5) a forehead reflectance transducer placed on the ventral aspect of the foal's tail-base. Eight normal, Quarter Horse foals (age 5-10 days) were studied while under general anaesthesia. Alterations in arterial carbon dioxide tension (PaCO2) were produced by changing the level of ventilation. At each level of ventilation, alterations of arterial haemoglobin saturation (SaO2) were produced by varying the inspired fraction of oxygen (FIO2). At each level of ventilation and each level of FIO2, arterial blood samples were obtained for blood gas analysis while pulse oximeter readings were recorded simultaneously for each TTAS combination. Arterial blood oxygen saturation (SaO2) was calculated from arterial blood gas values and the equine blood oxygen dissociation curve. Pulse oximeter readings from each TTAS combination were compared with SaO2 values with linear regression analyses. Bias and precision values were determined and the sensitivity and specificity of each TTAS combination for detecting desaturation (SaO2 < 90%) were determined. Linear regression analyses indicated significant (P < 0.05) linear correlation between oxygen saturation determined by pulse oximeter (SpO2) and SaO2 for each of the 5 TTAS combinations. The combinations TTAS-1, TTAS-3 and TTAS-4 tended slightly to underestimate SaO2 at high SaO2 ranges, but overestimated SaO2 at low ranges of SaO2. Combination TTAS-2 overestimated SaO2 over all ranges of SaO2. Combination TTAS-5 consistently underestimated SaO2 at all ranges of SaO2. In general, accuracy and precision of each TTAS combination decreased at lower SaO2 ranges. All TTAS combinations, except TTAS-2 and TTAS-4, had good sensitivity for detecting SaO2 less than 90%. All TTAS combinations except TTAS-5 demonstrated good specificity. We concluded that pulse oximetry appears to be a valuable method for assessing SaO2 and detecting desaturation in anaesthetised foals. Clinicians should be aware that the type of transducer used and the anatomical site to which it is attached can have marked effects on the accuracy of pulse oximetry; and that different TTAS combinations may behave differently over various ranges of SaO2.