An observational method comparison study examined whether different pulse oximetry devices provide consistent nocturnal oxygen saturation measurements in stable adults treated with continuous positive airway pressure (CPAP) or long-term noninvasive ventilation (NIV). Although overnight pulse oximetry is widely recommended to detect residual sleep-disordered breathing (SDB), nocturnal hypoventilation, and hypoxaemia in these patients, clinicians often assume that different devices yield equivalent results. In this study, the authors questioned this assumption and investigated whether device variability might meaningfully influence clinical decision-making.
Pulse oximetry has been a cornerstone of respiratory and critical care monitoring since the 1970s and is often considered a ‘fifth vital sign’. In patients receiving CPAP or NIV, nocturnal SpO₂ trends are used to assess treatment effectiveness and guide interventions, such as adding supplemental oxygen or adjusting ventilator settings. Clinical thresholds, including spending more than 10% of the night with SpO₂ below 88–90% or having an elevated oxygen desaturation index (ODI), frequently trigger changes in therapy. Therefore, inaccurate or inconsistent readings may directly affect management decisions .
The study was conducted at a specialised pulmonary rehabilitation centre in Switzerland. Stable adult patients already receiving long-term CPAP or NIV were monitored simultaneously overnight with three commercially available, European Community-certified devices. Device 1 was a Sentec Digital Monitor combining transcutaneous capnography with an earlobe SpO₂ probe heated to 42°C. Device 2 was a ResMed system using a conventional finger pulse oximeter. Device 3 was a Viatom wrist-worn device with a ring probe. All were in routine clinical use. Inclusion criteria required clinical stability and awake SpO₂ above 90% on room air.
Fifty-one patients were enrolled; one recording shorter than four hours was excluded, leaving 50 participants (mean age 68 years; 40% female). Thirty-one were using NIV and 19 CPAP. The median recording duration was approximately 7.5 hours. For each device, the investigators extracted mean, median and minimum SpO₂, time spent below predefined saturation thresholds, ODI (≥ 3% desaturations), heart rate, and, for device 1, transcutaneous CO₂.
Across all paired comparisons, agreement proved poor. Limits of agreement for both mean and median SpO₂ exceeded the predefined ± 2% threshold, indicating clinically meaningful discrepancies. ICCs for mean and median SpO₂ ranged from 0.77 to 0.84, suggesting only moderate to good reliability rather than strong concordance. Bias was particularly evident between devices 1 and 3, with the Sentec monitor systematically overestimating saturation by more than 2% compared with the wrist device. These differences were large enough to change clinical interpretation .
Variability was even greater for ODI. Limits of agreement were wide, and ICCs ranged from 0.33 to 0.75, demonstrating weak to moderate reliability. Bias between some devices exceeded five events per hour. Consequently, the classification of patients as having residual SDB differed substantially depending on the device used. Depending on the comparison, 12–28% of patients were classified inconsistently for nocturnal hypoxaemia, and up to 30–43% were classified differently regarding residual SDB. Such discordance could readily lead to different treatment decisions, including adding oxygen, modifying pressure settings, or ordering further investigations .
Several explanations were proposed. The heated earlobe probe used by the Sentec device increases local perfusion and may artificially elevate readings. Probe location may also matter, as earlobe measurements reflect more central perfusion than fingertip sensors. Devices differed in averaging times, optical configurations (transmission vs reflectance technology), calibration methods, and proprietary signal-processing algorithms. These technical variations could substantially influence overnight averages and event detection .
Overall, the study demonstrates that commonly used pulse oximeters and combined capnography-oximetry devices do not provide interchangeable nocturnal SpO₂ or ODI measurements in stable patients receiving CPAP or NIV. Device choice alone can materially alter the identification of hypoxaemia or residual SDB and thus influence management. The authors advocate caution when interpreting results across different systems and suggest the need for formal independent validation and potential recalibration of devices to ensure reliable clinical decisions .
Source: CHEST
Image Credit: iStock
