Automated Oxygen Titration With Noninvasive Respiratory Support in Acute In-Patient Care

Background: When delivering high-flow nasal cannula (HFNC) therapy, automated oxygen titration increases time spent within a target S_pO2 range compared with manually adjusted titration. This trial explored if this benefit is also achieved when CPAP or noninvasive ventilation (NIV) is used. Methods: This open label exploratory study randomized participants to automated oxygen titration or manual oxygen titration using a single respiratory support device capable of delivering HFNC, CPAP, and NIV. Participants were hospital in-patients requiring supplemental oxygen and one or more of HFNC, CPAP, and NIV; and could interchange between the three modalities according to clinical need. The primary outcome was the proportion of time spent within a target S_pO2 range in participants who received ≥8 h of therapy. A secondary interaction analysis explored whether any difference between automated and manual titration differed by respiratory support modality. Results: A total of 68 participants received randomized therapy; 58 started on HFNC, 6 on CPAP, and 4 on NIV. A total of 62 had data for the primary end point, with median (interquartile range) proportion of time spent within the target S_pO2 range with automated oxygen (n = 32) of 91.4% (84.3 to 97.5) versus 75.9% (64.0 to 87.8) with manually adjusted oxygen (n = 30); difference 13.9% (95% confidence interval 6.2-21.2), P < .001. The difference between automated and manual titration did not depend on respiratory support modality, P-interaction = .94. Conclusions: In hospitalized subjects predominantly receiving HFNC, automated oxygen titration had the effect of increasing time spent within a target S_pO2 range compared with manual oxygen titration. Automated oxygen titration had a similar effect on a small subset of participants receiving CPAP and NIV.