Definition and prognostic value of response to prone positioning in ARDS: a systematic review and meta-analysis

Prone positioning improves survival in patients with ARDS; however, no consensus exists on how to define a positive response to this intervention. We conducted a systematic review and meta-analysis to map existing definitions of response to prone position in invasively ventilated patients with ARDS and to quantify their pooled proportion across existing body of evidence. We also evaluated the association between responsiveness to prone position and mortality.

We surveyed PubMed, Embase, and Cochrane Central Register of Controlled Trials databases from inception to July 2025. For the primary outcome (proportion of responders), pooled estimates were calculated using a random-effects model with logit transformation of individual study proportions. For the secondary outcome, we performed a pairwise meta-analysis to estimate pooled odds ratios for mortality.

Oxygenation response, defined as a change of PaO₂/FiO₂ from the supine to the prone position, was adopted as definition by 53 non-randomized studies. The pooled proportion of responders was estimated as 68% (95% C.I. 63-72%). Twenty-one studies assessed responsiveness using physiological variables other than, or in addition to, oxygenation, including carbon dioxide clearance, respiratory mechanics, or ventilation-perfusion matching. Using these alternative definitions, the pooled proportion of responders ranged from 45% to 53%. Of the 26 studies providing unadjusted mortality data in responders and non-responders, 11 (42%) reported a reduced risk of mortality in the responder cohort. A sensitivity analysis restricted to the five studies at serious risk of bias showed a reduced unadjusted risk of mortality in responders (OR 0.41, 95% CI 0.25-0.68; I² = 78%).

Definitions of responsiveness to prone positioning are highly heterogeneous across the literature, and the reported proportion of responders varies widely depending on the definition adopted. High risk of bias, residual confounding and substantial between-study heterogeneity, limit robust conclusions regarding the association between physiological responsiveness to prone positioning and survival.

CRD420251104725.