Does gas exchange response to prone position predict mortality in hypoxemic acute respiratory failure?

To determine whether gas exchange response to a first prone position session can predict patient outcome in hypoxemic acute respiratory failure.

Data from a previous multicenter randomized controlled trial were retrospectively analyzed for relationship between PaO(2)/FIO(2) ratio and PaCO(2) changes during the first 8-h prone position session to day 28 mortality rate; 370 prone position sessions were analyzed. Arterial blood gas was measured in supine position before proning and in prone position at the end of the session. Gas exchange improvement was defined as increase in the PaO(2)/FIO(2) ratio of more than 20% (PaO(2)R) or decrease in PaCO(2) of more than 1 mmHg (PaCO(2)R).

The 28-day mortality rate was 26.5% in PaO(2)R-PaCO(2)R, 31.7% in PaO(2)R-PaCO(2)NR, 38.9% in PaO(2)NR-PaCO(2)R, and 43% in PaO(2)NR-PaCO(2)NR (log-rank 14.02, p = 0.003). In a Cox proportional hazards model the gas exchange response was a significant predictor to patient outcome with a 82.5% increase in risk of death in the case of PaO(2)NR-PaCO(2)R or PaO(2)NR-PaCO(2)NR, relative to the gas exchange improvement response (odds ratio 1.825). However, after adjusting for the difference in oxygenation between day 2 and day 1 the gas exchange response does no longer reach significance.

In patients with hypoxemic acute respiratory failure initial improvement in gas exchange in the first PP session was associated with a better outcome, but this association disappeared when the change in oxygenation from day 1 to day 2 was taken into account, suggesting that underlying illness was the most important predictor of mortality in this patient population.