BACKGROUND: The effect of positive end expiratory pressure (PEEP) on postoperative pulmonary complications (PPCs) in obese patients remains controversial.

OBJECTIVE: To test, whether intra-operative PEEP or PPCs are associated with plasma levels of biomarkers of lung injury.

DESIGN: A prospective substudy of a multicentre randomised controlled trial (PROBESE).

SETTING: Operating rooms of six tertiary care centres in the United States and Europe.

PATIENTS: Obese patients at risk for PPCs undergoing abdominal surgery.

INTERVENTION: Intra-operative low tidal volume ventilation with high PEEP (12 cmH 2 O) and recruitment manoeuvres, or low PEEP (4 cmH 2 O).

MAIN OUTCOME MEASURES: The primary endpoint was the association between absolute postoperative plasma levels of receptor for advanced glycation end-products (RAGE) and intra-operative PEEP; secondary endpoints included pre and postoperative plasma concentrations as well as the relative changes of interleukin-6, IL-8, tumour necrosis factor-α, surfactant protein D, mucin-1, clara cell protein-16, intercellular adhesion molecule-1 and vascular cell adhesion molecule. PPCs were assessed as a 'collapsed composite' of adverse pulmonary events. The predictive ability of biomarkers for PPCs was assessed with the receiver operating curve-area under the curve (ROC-AUC).

RESULTS: A total of 96 patients received low PEEP, and 95 patients high PEEP. Postoperative plasma concentrations of RAGE and other biomarkers did not differ between groups. The relative increase of RAGE during surgery was more pronounced with low than high PEEP; median [IQR], 1.2 [1.0 to 1.6] vs. 1.1 [0.9 to 1.3], P  = 0.012. Patients who developed PPCs showed higher postoperative plasma levels and relative increase of IL-6; 26.3 [12.6 to 139.5] vs. 15.1 [3.7 to 38.7] fold change. The ROC-AUC was less than 0.7 for all biomarkers.

CONCLUSIONS: In this subgroup, choice of PEEP did not affect postoperative biomarkers of lung injury. Irrespective of PEEP, PPCs were associated with an increase in plasma levels of these biomarkers, but their predictive capability was poor.

TRIAL REGISTRATION: Clinicaltrials.gov, identifier: NCT02148692.