Background: The effect of personalised mechanical ventilation on clinical outcomes in patients with acute respiratory distress syndrome (ARDS) remains uncertain and needs to be evaluated. We aimed to test whether a mechanical ventilation strategy that was personalised to individual patients' lung morphology would improve the survival of patients with ARDS when compared with standard of care.

Methods: We designed a multicentre, single-blind, stratified, parallel-group, randomised controlled trial enrolling patients with moderate-to-severe ARDS in 20 university or non-university intensive care units in France. Patients older than 18 years with early ARDS for less than 12 h were randomly assigned (1:1) to either the control group or the personalised group using a minimisation algorithm and stratified according to the study site, lung morphology, and duration of mechanical ventilation. Only the patients were masked to allocation. In the control group, patients received a tidal volume of 6 mL/kg per predicted bodyweight and positive end-expiratory pressure (PEEP) was selected according to a low PEEP and fraction of inspired oxygen table, and early prone position was encouraged. In the personalised group, the treatment approach was based on lung morphology; patients with focal ARDS received a tidal volume of 8 mL/kg, low PEEP, and prone position. Patients with non-focal ARDS received a tidal volume of 6 mL/kg, along with recruitment manoeuvres and high PEEP. The primary outcome was 90-day mortality as established by intention-to-treat analysis. This study is registered online with ClinicalTrials.gov, NCT02149589.

Findings: From June 12, 2014, to Feb 2, 2017, 420 patients were randomly assigned to treatment. 11 patients were excluded in the personalised group and nine patients were excluded in the control group; 196 patients in the personalised group and 204 in the control group were included in the analysis. In a multivariate analysis, there was no difference in 90-day mortality between the group treated with personalised ventilation and the control group in the intention-to-treat analysis (hazard ratio [HR] 1·01; 95% CI 0·61-1·66; p=0·98). However, misclassification of patients as having focal or non-focal ARDS by the investigators was observed in 85 (21%) of 400 patients. We found a significant interaction between misclassification and randomised group allocation with respect to the primary outcome (p<0·001). In the subgroup analysis, the 90-day mortality of the misclassified patients was higher in the personalised group (26 [65%] of 40 patients) than in the control group (18 [32%] of 57 patients; HR 2·8; 95% CI 1·5-5·1; p=0·012.

Interpretation: Personalisation of mechanical ventilation did not decrease mortality in patients with ARDS, possibly because of the misclassification of 21% of patients. A ventilator strategy misaligned with lung morphology substantially increases mortality. Whether improvement in ARDS phenotyping can decrease mortality should be assessed in a future clinical trial.

Funding: French Ministry of Health (Programme Hospitalier de Recherche Clinique InterRégional 2013).