Elastance as a determinant of the effect of prone positioning on mortality in acute respiratory distress syndrome: a post hoc analysis of the PROSEVA trial

Abstract Background Patient factors determining the benefit of prone positioning remain uncertain, resulting in the maneuver being applied indiscriminately among those with moderate-severe ARDS. We aimed to assess if baseline respiratory system elastance (Ers), or “stiffness”, determines the treatment effect of prone positioning on mortality. Methods Bayesian logistic regression modeling of the PROSEVA Trial was used to estimate the posterior probability of prone positioning effect moderation by baseline Ers on 90-day mortality in patients with moderate-severe ARDS. As a secondary aim, we tested whether the absolute change in driving pressure of the respiratory system (∆DPrs ) in response to prone positioning predicted 90-day mortality, using logistic regression. Results The treatment effect of prone positioning on mortality did not meaningfully vary with baseline Ers (posterior probability of benefit OR < 0.95 = 52%; interaction OR 0.94, 90% credible interval, CrI, 0.74–1.20). Higher baseline Ers was associated with greater improvements in DPrs at the end of the first prone session (β= -3.3, 95% confidence interval (CI) -4.09, -2.49; p = < 0.001). However, this response was not associated with mortality benefit in adjusted models (OR 1.14, 95% CI 0.96, 1.37; p = 0.14). Conclusions The effect of prone positioning on mortality did not vary with Ers in the PROSEVA trial. Similarly, prone positioning-induced improvement in DPrs was not predictive of mortality in this cohort of passively ventilated ARDS patients.