Hodgson 2011.

Study characteristics
Methods	Pilot randomised controlled parallel‐group study conducted in a UCI from Australia Time period of study: January 2008 through October 2009
Participants	20 participants aged > 15 (single centre) Included: patients with ARDS (AECC) and the presence of both an intra‐arterial line and a central venous catheter Excluded: chest trauma, intercostal catheter with air leak, pneumothorax on chest X‐ray, bronchospasm on auscultation, raised intracranial pressure, mean arterial pressure ≤ 60 mmHg, significant arrhythmias, ventilated longer than 72 hours
Interventions	Control (10): MV assist control, TV 6 mL/kg PBW, plateau pressure < 30 cmH₂O. Acidosis (pH < 7.3) was managed by increasing minute ventilation. Use of rescue therapies in participants receiving FIO₂ > 0.9 PEEP. PEEP/FIO₂ combination: mean PEEP 9.6 cmH₂O during first 72 hours Intervention (10): MV pressure control ventilation (PCV), TV 6 mL/kg PBW, plateau pressure < 30 cmH₂O. Target range for oxygenation: SpO₂ between 90% and 92%. Acidosis (pH < 7.15) was managed by increasing respiratory rate to maximum of 38 breaths per minute. PEEP: recruiting manoeuvres followed by decremental PEEP titration until decrease in SpO₂ ≥ 1% from maximum SpO₂ was observed. Mean PEEP: 13.5 cmH₂O during first 72 hours For calculation of sample size, study authors estimated that 10 patients per group would provide > 80% power to detect a difference of 1 standard deviation in cytokine levels, with a 2‐sided test for differences, P value of 0.01, whilst assuming an intraclass correlation of 0.2 between baseline level and Day 3 (pilot study)
Outcomes	Primary: measurement of plasma cytokines during first 7 days Secondary: PaO₂/FIO₂ ratio, static lung compliance, LOS in ICU, LOS in hospital, days of MV, mortality before hospital discharge, rescue therapies (number of patients), SOFA score (Day 7)
Notes	Rescue therapies (only control group): recruiting manoeuvres and inhaled nitric oxide Study authors do not declare a conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Study authors used computerised random block schedule
Allocation concealment (selection bias)	Low risk	Opaque sealed envelopes that were sequentially numbered
Blinding of participants and personnel (performance bias) All outcomes	High risk	Use of adjunctive therapy at the discretion of the attending physician ‐ not protocolised
Blinding of outcome assessment (detection bias) All outcomes	Low risk	No blinding but outcomes not influenced by lack of blinding
Incomplete outcome data (attrition bias) All outcomes	Low risk	Analysis on the basis of the intention‐to‐treat principle. No missing outcome data
Selective reporting (reporting bias)	Low risk	Published reports included all expected outcomes
Other bias	Low risk	Review author believed the study to be free of other sources of bias