How Using Different Oximeters May Affect Clinical Decision-Making : A Method Comparison Study in Patients Receiving CPAP or Noninvasive Ventilation

Background: Nocturnal pulsed oxygen saturation (SpO2 ) monitoring is recommended for detecting residual sleep-disordered breathing, including nocturnal hypoventilation, in patients treated by noninvasive ventilation (NIV) or CPAP. It is a general assumption that different pulse oximetry devices will provide similar results. However, this may not be correct. Research question: Does simultaneous nocturnal recording with different pulse oximetry devices lead to clinically meaningful differences? Study design and methods: In this prospective observational study, stable patients receiving CPAP or NIV with 3 different devices (a transcutaneous oxycapnograph with a probe on an earlobe and 2 pulse oximeters, one with a finger probe and the other with a ring probe) were monitored overnight. Comparison of results using Bland-Altman graphs, Cohen's κ values, and, interclass correlation coefficients (ICCs). Results: In all paired comparisons of devices, limits of agreement exceeded the predefined ± 2% limit for both mean and median SpO2 and for oxygen desaturation index (ODI). The ICC values ranged between 0.77 and 0.74 for mean and median SpO2 and between 0.33 and 0.75 for ODI. κ Values were moderate (0.46-0.73) for all paired comparisons. Interpretation: Use of different devices for monitoring stable patients receiving NIV or CPAP, all in current use in clinical practice and used by expert centers, may have an unpredictable impact on results and ensuing clinical decision-making. Clinical trial registration: ClinicalTrials.gov; No.: NCT04474574; URL: www. Clinicaltrials: gov.