Evaluation of pulse oximetry as a surrogate for PaO2 in awake dogs breathing room air and anesthetized dogs on mechanical ventilation. Issue 6 (18th October 2019)

• Record Type: Journal Article
• Authors: Farrell, Kate S.; Hopper, Kate; Cagle, Laura A.; Epstein, Steven E.
• Is Part Of: Journal of veterinary emergency and critical care. Volume 29:Issue 6(2019)
• Page Start: 622
• Page End: 629
• Publication Date: 2019-10-18
• Subjects: blood gas analysis -- hemoglobin oxygen saturation -- hypoxemia -- monitoring; Veterinary emergencies -- Periodicals; Veterinary critical care -- Periodicals; 636.089
• DOI: 10.1111/vec.12898
• Languages: English
• ISSNs: 1479-3261

Abstract

Abstract: Objective: To evaluate the ability of arterial hemoglobin oxygen saturation measurement via pulse oximetry (SpO2 ) to serve as a surrogate for PaO2 in dogs. Design: Two‐part study: prospective observational and retrospective components. Setting: University teaching hospital. Animals: Ninety‐two dogs breathing room air prospectively enrolled on a convenience basis. Retrospective evaluation of 1, 033 paired SpO2 and PaO2 measurements from 62 dogs on mechanical ventilation. Interventions: Dogs with concurrent SpO2 and PaO2 measured on room air had a data sheet completed with blood gas analysis. SpO2, PaO2, and FiO2 values were collected from medical records of dogs on mechanical ventilation. Measurements and Main Results: Predicted PaO2 was calculated from SpO2 using the dog oxyhemoglobin dissociation curve. The correlation coefficient between measured and predicted PaO2 was 0.49 ( P < 0.0001) in room air dogs and 0.74 ( P < 0.0001) in ventilated dogs. In room air dogs, Bland–Altman analysis between measured minus predicted PaO2 versus the average showed a mean bias of −6.0 mm Hg (95% limit of agreement, −35 to 23 mm Hg). The correlation coefficient between PaO2 /FiO2 and SpO2 /FiO2 ratios was 0.76 ( P < 0.0001). After combining data sets, receiver operating characteristic curve analysis showed the optimal cutoff value for detecting hypoxemia (PaO2 < 80 mm Hg) was an SpO2 of 95%, with sensitivity and specificity of 77.8% and 89.5%, respectively. Using this cutoff, 6.9% of SpO2 readings failed to detect hypoxemia, whereas 7.2% predicted hypoxemia that was not present. Conclusions: The SpO2 was not clinically suitable as a surrogate for PaO2, though it performed better in mechanically ventilated dogs. As sensitivity for the detection of hypoxemia was poor, pulse oximetry does not appear to be an acceptable screening test. The SpO2 /FiO2 ratio may have value for evaluation of anesthetized dogs on supplemental oxygen. Arterial blood gas analysis remains ideal for assessment of oxygenation.