Overview
- The multicenter SAVE-O2 AI trial of 300 adults was published in JAMA Internal Medicine and presented at the Military Health System Research Symposium after enrollment at four U.S. hospitals.
- Patients whose oxygen was controlled by the closed-loop O2matic PRO100 spent an adjusted 85% of time in the target SpO2 range (90–96%) versus 63% with usual care, an adjusted increase of 21 percentage points.
- The system cut time in hyperoxemia (SpO2 >96%) and reduced minutes spent below key hypoxemia thresholds, and it produced no increase in serious adverse events or 28-day in-hospital mortality.
- Autonomous control lowered total oxygen volume used per patient and required fewer manual flow adjustments, suggesting reduced oxygen waste and less clinician workload during the first 72 hours.
- Key limits include the open-label design, reliance on pulse oximetry with known accuracy issues, use of a single investigational device under an FDA IDE, and no demonstrated improvements in hospital stay, respiratory escalation, or survival, so further studies are planned in transport and prehospital settings.