Abstract
BACKGROUND: Automated oxygen titration to maintain a stable S(pO(2)) has been developed for spontaneously breathing patients but has not been evaluated during CPAP and noninvasive ventilation (NIV). METHODS: We performed a randomized controlled crossover, double-blind study on 10 healthy subjects with induced hypoxemia during 3 situations: spontaneous breathing with oxygen support, CPAP (5 cm H(2)O), and NIV (7/3 cm H(2)O). We conducted in random order 3 dynamic hypoxic challenges of 5 min (F(IO(2)) 0.08 ± 0.02, 0.11± 0.02, and 0.14 ± 0.02). For each condition, we compared automated oxygen titration and manual oxygen titration by experienced respiratory therapists (RTs), with the aim to maintain the S(pO(2)) at 94 ± 2%. In addition, we included 2 subjects hospitalized for exacerbation of COPD under NIV and a subject managed after bariatric surgery with CPAP and automated oxygen titration. RESULTS: The percentage of time in the S(pO(2)) target was higher with automated compared with manual oxygen titration for all conditions, on average 59.6 ± 22.8% compared to 44.3 ± 23.9% (P = .004). Hyperoxemia (S(pO(2)) > 96%) was less frequent with automated titration for each mode of oxygen administration (24.0 ± 24.4% vs 39.1 ± 25.3%, P < .001). During the manual titration periods, the RT made several changes to oxygen flow (5.1 ± 3.3 interventions that lasted 122 ± 70 s/period) compared to none during the automated titration to maintain oxygenation in the targeted S(pO(2)) . Time in the S(pO(2)) target was higher with stable hospitalized subjects in comparison with healthy subjects under dynamic-induced hypoxemia. CONCLUSIONS: In this proof-of-concept study, automated oxygen titration was used during CPAP and NIV. The performances to maintain the S(pO(2)) target were significantly better compared to manual oxygen titration in the setting of this study protocol. This technology may allow decreasing the number of manual interventions for oxygen titration during CPAP and NIV.