Adaptive Oxygen Delivery in Home Settings for Patients With Chronic Obstructive Pulmonary Disease on Long-Term Oxygen Therapy - A Pilot Randomized Crossover Feasibility Trial

Rationale In patients with COPD on long-term oxygen therapy (LTOT), oxygen doses are typically set at rest to maintain saturation above 90% while avoiding hyperoxemia, despite fluctuating needs throughout the day. Although emerging technology enables continuous, saturation-based adjustments, it is unclear whether such adjustments are feasible in a home setting. This pilot study assessed the feasibility and preliminary effect of saturation-responsive automated oxygen adjustments at home. Methods The intervention consisted of four days of continuous oxygen titration using an electronic device. Patients wore a wrist pulse oximeter that connected to the device via Bluetooth, creating a closed-loop system. In this setup, oxygen flow was adjusted from 1- 7 L/min based on real-time oxygen saturation (SpO₂), targeting a level of 90% to 94%. In the control period, the patients received their usual fixed oxygen, and they were not blinded to the intervention. Data were transmitted to a cloud-based solution. Feasibility was defined as successful time during which the patients were automatically titrated, patients´ willingness toward the intervention, and clinically relevance. The Clinical COPD Questionnaire (CCQ, 0-6 points, with 0.4 suggested as a minimal important difference) evaluated changes in the 24-hour health status. Results Twelve patients, four women and eight men with an average ±SD home oxygen dose of 2.0 ± 0.8 L/min were included. The patients' oxygen flow was automatically adjusted for a median (interquartile range) of 77 (68.0-84.3) hours per patient, reflecting 83% of the time they were compliant with wearing the equipment. More than 217,000 paired live data points on SpO2 and oxygen flow were received on average per patient. The time spent within target was statistically different between arms with 51.3 ± 12.2% when using fixed oxygen flow, and 83.4 ± 8.3% of the time when flow was automatically adjusted, p<0.001. Complaints from the patients were mainly regarding the noise from the 9 L oxygen concentrator. A change in the CCQ score of 0.74 ± 0.47 points favored the automated oxygen titration, p<0.001. Conclusion It is possible to titrate oxygen flow in a home setting and achieve almost perfect control of SpO2, minimizing time with hypoxemia. The patients in this pilot study accepted the equipment and reported a statistical and clinically significant reduction in COPD symptoms measured with CCQ. The clinical importance of controlling SpO2 needs to be examined in a larger study.