Le décubitus ventral dans le syndrome de détresse respiratoire aiguë. Revue des essais contrôlés et méta-analyse

Mechanical ventilation in prone position has been reported to be less frequently used in patients with ARDS. This has been ascribed to the absence of evidence of a positive impact of proning on mortality. Critical appraisal of published randomised controlled trials on shows that most of these trials suffer methodological bias that threaten their internal validity: included patients of variable severity and disease stage (case-mix), non standardisation of the prone duration across studies, lack of lung protective ventilatory strategy in the earliest trials, premature inclusion cessation with reduced power of most studies, and inappropriate intention-to-treat analysis in the trials that allowed crossover. Most studies consistently reported a substantial increase in blood oxygenation with prone positioning, and a trend toward a lower incidence of ventilator associated pneumonia without an increase in the rate of adverse events. We did a meta-analysis of randomised controlled trials assessing proning, showing that the pooled effect of proning on mortality progressively shifted from a negative to a positive effect as a probable consequence of prolonged duration of proning, and generalised use of lung protective ventilatory strategies in the most recent studies. Our meta-analysis also emphasizes the fact that the study that accounts for the major effects of the whole meta-analysis has also major methodologic limitations. Definitive evidence on the role of prone positioning in ARDS will probably stem from RCTs that are currently enrolling patients. Meanwhile, proning should still be proposed to the most severe patients.