Most impactful predictors for hyperoxaemia in exacerbation of chronic obstructive pulmonary disease managed by Emergency Medical Services and Emergency Department

Hyperoxemia in acute exacerbation of chronic obstructive pulmonary disease (AECOPD) leads to adverse outcomes. It remains prevalent in the pre‐hospital Emergency Medical Services (EMS) and Emergency Department (ED). Objective To determine the key predictors for hyperoxemia in AECOPD in EMS and ED. Methods This was a prospective observational study of AECOPD patients in EMS and two EDs. Hyperoxemia was defined as PaO 2 > 65 mm Hg (corresponds to SpO 2 > 92%). We determined apriori candidate factors in Patient, Organization and Staff domains. Primary outcomes were the key predictors for hyperoxemia. Secondary outcomes were in‐hospital mortality and mechanical ventilation rates in hyperoxemic versus non‐hyperoxemic groups. We generated a logistic regression model for each domain. We reported the adjusted odds ratios (AORs), 95% CIs and p values. We selected the output factors using AOR ≥2.0 and ≥2.5 for modifiable and non‐modifiable factors, respectively. These selected factors were fed into a final model with eventual factors selected based on: threshold AORs as stated above and/or 95% CIs including these AORs. Results Three hundred and twenty‐six patients were analysed; 60.7% had hyperoxemia. We found three eventual modifiable factors; first, ED SpO 2 > 95% [AOR 2.62 (95% CIs: 1.61‐4.33); P  < 0.001], EMS non‐rebreathing mask [AOR 2.01 (95% CIs: 1.06‐3.97); P  = 0.04]; and ED nasal cannula [AOR 1.69 (95% CIs: 1.05‐2.72); P  = 0.03]. Secondary outcomes did not differ between groups. Conclusion We identified three key modifiable predictors. We intend to conduct an interventional study using them to reduce hyperoxemia rate in AECOPD.