Effect of lingual gauze swab placement and light exclusion on pulse oximeter agreement with arterial haemoglobin-oxygen saturation derived from blood gas analysis in healthy anaesthetised horses.

Objective To compare peripheral arterial haemoglobin-oxygen saturation values obtained using pulse oximetry (SpO 2 ) with and without a lingual gauze swab under ambient light and light-exclusion conditions, with arterial haemoglobin-oxygen saturation derived from arterial blood gas analysis (SaO 2 ). Study design Prospective method-comparison study. Animals Sixty healthy adult horses undergoing elective general anaesthesia. Methods In American Society of Anesthesiologists (ASA) physical status classification I-II horses, SpO 2 readings were obtained under six conditions during anaesthesia: no swab, no swab with light exclusion, dry swab, dry swab with light exclusion, wet swab and wet swab with light exclusion. Agreement with SaO 2 values was assessed using Bland-Altman method for repeated measurements. Bias (defined as SpO 2 minus SaO 2 ) and limits of agreement (LoA) were estimated from mixed-effects models with anaesthetic episode as the primary clustering unit. Confidence intervals (95% CI) were obtained using cluster bootstrapping. Results A total of 154 paired SpO 2 –SaO 2 observations from 62 anaesthetic episodes with repeated measurements were analysed per experimental condition. SaO 2 ranged from 94.4% to 100%, and PaO 2 from 62.6 to 595.8 mmHg (8.3 to 79.4 kPa). SpO 2 underestimated SaO 2 across all conditions. Biases ranged from -1.6 to -2.7%, with least bias observed using a dry swab under ambient light (-1.6 %, 95% CI -2.1 to -1.1). LoA width ranged from 10.3 to 13.3 % across configurations. Light exclusion produced more negative bias and wider LoA width than the ambient light configurations; however, these differences were small relative to the overall LoA width. Conclusions and clinical relevance In healthy adult horses undergoing elective anaesthesia lingual gauze swab placement was not detrimental but did not meaningfully improve agreement between SpO 2 and SaO 2 . Individual SpO 2 readings were imprecise across all configurations. Arterial blood gas analysis remains necessary when accurate assessment of arterial oxygenation is required.