Background: It has been postulated that COVID-19 lung injury may share some features with those observed in HALI. Thus, a correct target PaO₂ during oxygen supplementation may be crucial to protect the lung from further tissue damage.

Aims and objectives: 1) to evaluate the effects of conservative oxygen supplementation during Helmet CPAP therapy on mortality and ICU admission in patients with COVID-19 and respiratory failure; 2) to evaluate the effect of conservative oxygen supplementation on new-onset organ failure and secondary pulmonary infections.

Methods: This was a single-center, historically controlled study of patients with severe respiratory failure due to COVID-19 pneumonia, receiving either conservative or nonconservative oxygen supplementation during Helmet CPAP. A cohort receiving conservative oxygen supplementation was studied prospectively in which oxygen supplementation was administered with a target PaO₂ < 100 mmHg. Results of this cohort were compared with those of a cohort who had received liberal oxygen supplementation.

Results: Seventy-one patients were included in the conservative cohort and 75 in the nonconservative cohort. Mortality rate was lower in the conservative cohort (22.5% vs 62.7%, p<0.001). Rates of ICU admission and new-onset rate organ failure were lower in conservative cohort (14.1% vs 37.3%, p=0.001, and 9.9% vs 45.3% p<0.001, respectively).

Conclusions: In patients with COVID-19 and severe respiratory failure, conservative oxygen supplementation during Helmet CPAP was associated to improved survival, lower ICU admission rate and less new-onset organ failure.