Abstract
BACKGROUND: Oxygen supply is a common procedure performed by emergency medical services (EMS); however, whether prehospital oxygen or fraction of inspired oxygen (FiO(2)) therapy predict mortality has not been studied to date. This study aims to determine mortality associated with oxygen therapy in unselected patients with acute disease who underwent prehospital care. METHODS: This was a prospective, observational, cohort, multicenter, EMS-delivery, ambulance-based study. Adults with unselected acute disease who were managed by EMS and evacuated with high priority to the emergency department of four hospitals in three Spanish provinces. Epidemiological variables, on-scene vital signs, and prehospital blood analysis data were collected. The primary outcome was short- (2-, 7-, and 30-day), medium- (90- and 180-day), and long-term (365-day) all-cause cumulative mortality. The samples were a priori split according to thresholds of their received FiO(2)(FiO(2)=0.21, that is without oxygen therapy; FiO(2) between 0.22 and 0.49; and FiO(2)≥0.5). The categorical variables FiO(2), epidemiological variables, vital signs, prehospital point-of-care testing (POCT) and patient outcomes were fitted with a logistic regression model. Additionally, a propensity score matching and a survival analysis were used. RESULTS: The final sample included 7,494 patients, 70.3% of whom did not receive oxygen therapy, 15% with a FiO(2) between 0.22 and 0.49, and 14.7% with a FiO(2)≥0.5. The 2-day mortality was 0.4%, 5.3%, and 22.9% respectively (P<0.001). The 365-day mortality was 9.9%, 33.1%, and 50.5% respectively (P<0.001). Finally, the FiO(2) predictive capacities 2-,30-, and 365-day mortality were AUC=0.870 (95%CI: 0.840-0.899), 0.810 (95%CI: 0.784-0.837), 0.704 (95%CI: 0.679-0.728), respectively. CONCLUSION: Prehospital oxygen therapy by thresholds of FiO(2) was linked to death and allowed mortality prediction. This finding could provide an aid for EMS providers, allowing to assess more individualized patient risk.