Aims of the study: Awake prone positioning (aPP) in non-intubated patients with severe SARS-CoV-2-related pneumonia improves oxygenation and reduces the intubation rate, but no early predictors for success or failure of the strategy have been described. The main objective of this study was to assess whether response to the first aPP in terms of PaO2/FiO2, alveolar-arterial gradient (Aa-O2), respiratory rate and PaCO2 could predict the need for intubation. As secondary objective, we assessed the effects of aPP on the same parameters for all the sessions considered together.
Methods: Retrospective analysis of consecutive SARS-CoV-2 pneumonia patients suffering from acute respiratory failure with moderate to severe hypoxaemia for whom aPP was performed for at least 45 minutes based on the prescription of the clinician in charge according to predefined criteria. Respiratory rate, blood gases and oxygenation parameters (PaO2/FiO2 and Aa-O2), before and after the first aPP were compared between patients who were subsequently intubated or not. Effects of all the aPP sessions together were also analysed.
Results: One hundred and sixty-six patients were admitted for SARS-CoV-2 pneumonia during the study period. Among them, 50 received aPP lasting at least 45 minutes. Because 17 denied consent for data analysis and 2 were excluded because of a "do not intubate order", 31 patients (for a total of 116 aPP sessions without any severe adverse events reported) were included. Among them, 10 (32.3%) were intubated. Mean age ± standard deviation (SD) was 60 ± 12 years. At ICU admission, respiratory rate was 26 ± 7/minute, median PaO2/FiO2 94 (interquartile range [IQR] 74-116) mm Hg and median Aa-O2 412 (IQR 286-427) mm Hg (markedly increased). Baseline characteristics did not statistically differ between patients who subsequently needed intubation or not. During the first aPP, PaO2/FiO2 increased and Aa-O2 decreased. When comparing patients who later where intubated or not, we observed, in the non intubated group only, a clinically significant decrease in median Aa-O2, from 294 (280-414) to 204 (107-281) mm Hg, corresponding to a 40% (26-56%) reduction, and a PaO2/FiO2 increase, from 103 (84-116) to 162 (138-195), corresponding to an increase of 48% (11-93%). The p value is <0.005 for both. When all the aPP sessions (n = 80) were considered together, aPP was associated with a significant increase in PaO2/FiO2 from 112 (80-132) to 156 (86-183) mm Hg (p <0.001) and Aa-O2 decrease from 304 (244-418) to 224 (148-361) mm Hg (p = 0.001).
Conclusions: Awake pronation in spontaneously breathing patients is feasible, and improves PaO2/FiO2 and Aa-O2. Response to the first session seems to be associated with lower intubation rate.