Why Survival Goes Up with the Prone Position in the Most Severely Hypoxemic Patients?
As discussed previously, the physiologic beneficial effects of prone positioning should explain its effect on survival. However, the cause to effect relationship between them is not so clear in particular regarding oxygenation.
Even though there is no doubt that oxygenation was found better in the prone as compared to the supine group, the fact that better survival results from better oxygenation is not supported by the analysis of the data. In the PROSEVA trial [21], the patients in each group were split into four quartiles of PaO2/FiO2 ratio at the time of randomization. The benefit of proning was present at every quartile. Furthermore, the magnitude of change in oxygenation during the first proning session was not associated with a better survival [36]. This result is in line with previous studies [37, 38]. Therefore, even though prone positioning may prevent death from life-threatening hypoxemia, and in this way proning can be seen as a rescue therapy, this mechanism is not the main factor to explain survival improvement with this procedure.
VILI prevention is the second candidate for explaining survival improvement with prone. Even though it is highly likely, it has not been directly demonstrated in the trials because these did not include any assessment of VILI.
The hemodynamic effects may be a mechanism of better survival with prone that has not receive much attention. The use of proning in the early experience regularly showed lack of hemodynamic worsening contrary to the well-known deleterious effects of higher PEEP. In the PROSEVA trial, there was a greater number of days where patient was not in shock, i.e., 2 days greater for the prone group as compared to the supine group [21]. Recently, prone position was associated with an even increase in cardiac output in those patients who were preload dependent in the supine position [39]. By increasing abdominal pressure, prone position may push the stressed intravascular volume into the right atrium in patients with reserve of preload. Interestingly in this study, the improvement of oxygenation in prone was less important in those patients with the increased cardiac index than in those with no change (with no preload reserve). This may be explained by the increased intrapulmonary shunt resulting from the increased cardiac index as originally demonstrated by Dantzker [40]. As important, proning has a beneficial impact on the right ventricle function [41], and this may explain survival benefit with this strategy. The right ventricle function can be impaired as a result from an acute increase in afterload due to different contributing factors that are working in ARDS, like hypoxemia, hypercapnia, overdistension [42]. This culminates into the acute cor pulmonale (ACP) and acute right ventricle failure. ACP has been found to occur in 22% of ARDS patients and is a predictor of worst outcome [43]. Earlier, Vieillard-Baron by using transesophageal echocardiography found that prone position reverted acute cor pulmonale in ARDS patients [44]. Very recently a multicenter observational study in ICUs in France over a large cohort of ARDS patients confirmed these findings and found that ACP was an independent predictor of death [45]. Interestingly in this cohort, prone position was an independent factor associated with better survival. By allowing lower PEEP, and hence minimizing the risk of overdisetnsion as previously discussed, and by improving oxygenation, prone position could prevent ACP occurrence and could be a strategy for the clinicians facing ACP in association with inhaled nitric oxide and dobutamine [41, 42]. This real preventive and curative effect of ACP by using prone position should be tested prospectively in trials, and, as a whole, the hemodynamic effects of prone position deserve further studies.
Finally, prevention of ventilator-associated pneumonia (VAP) could be a mechanism by which prone position reduces mortality. Prone position may reduce VAP by enhancing respiratory secretion removal. However, the attributable mortality of VAP is low, in particular in ARDS patients [46]. Indeed, a post hoc analysis of the PROSEVA trial found that prone position was not associated with a reduction in VAP [47].
