by Pedro David Wendel-Garcia, Arantxa Mas, Cristina
González-Isern, Ricard Ferrer, Rafael Máñez, Joan-Ramon Masclans, Elena
Sandoval, Paula Vera, Josep Trenado, Rafael Fernández, Josep-Maria Sirvent,
Melcior Martínez, Mercedes Ibarz, Pau Garro, José Luis Lopera, María Bodí…
Critical Care volume 26,
Article number: 37 (2022)
Background
Non-invasive oxygenation strategies have a prominent role in
the treatment of acute hypoxemic respiratory failure during the coronavirus
disease 2019 (COVID-19). While the efficacy of these therapies has been studied
in hospitalized patients with COVID-19, the clinical outcomes associated with
oxygen masks, high-flow oxygen therapy by nasal cannula and non-invasive
mechanical ventilation in critically ill intensive care unit (ICU) patients
remain unclear.
Methods
In this retrospective study, we used the best of nine
covariate balancing algorithms on all baseline covariates in critically ill
COVID-19 patients supported with > 10 L of supplemental oxygen at one of the
26 participating ICUs in Catalonia, Spain, between March 14 and April 15, 2020.
Results
Of the 1093 non-invasively oxygenated patients at ICU
admission treated with one of the three stand-alone non-invasive oxygenation
strategies, 897 (82%) required endotracheal intubation and 310 (28%) died
during the ICU stay. High-flow oxygen therapy by nasal cannula (n = 439) and
non-invasive mechanical ventilation (n = 101) were associated with a lower rate
of endotracheal intubation (70% and 88%, respectively) than oxygen masks (n = 553
and 91% intubated), p < 0.001. Compared to oxygen masks, high-flow
oxygen therapy by nasal cannula was associated with lower ICU mortality (hazard
ratio 0.75 [95% CI 0.58–0.98), and the hazard ratio for ICU mortality was 1.21
[95% CI 0.80–1.83] for non-invasive mechanical ventilation.
Conclusion
In critically ill COVID-19 ICU patients and, in the absence
of conclusive data, high-flow oxygen therapy by nasal cannula may be the
approach of choice as the primary non-invasive oxygenation support strategy.
No comments:
Post a Comment