Amato 1998.

Study characteristics
Methods	Randomised controlled study conducted at 2 centres in Brazil Time period of study: December 1990 through July 1995
Participants	53 participants aged > 14 and < 70 (2 centres) Included: ARDS, LIS ≥ 2.5, Pwedge < 16 mmHg Conditions excluded: previous lung or neuromuscular disease, MV > 1 week, uncontrolled terminal disease, previous barotrauma, previous lung biopsy or resection, uncontrollable and progressive acidosis, signs of intracranial hypertension, documented coronary insufficiency Sample size was estimated from a previous study (Amato 1995), considering a maximum sample of 58 patients and assuming a type I error of 5%, statistical power of 85%, and a survival rate in the protective‐ventilation group that would be 2.4 times that in the conventional‐ventilation group
Interventions	Control (24): MV:TV 12 mL/kg; PEEP to optimise FIO₂ < 0.6 with adequate systemic oxygen delivery; mean PEEP 8.7 ± 0.4 during first 36 hours Intervention (29): MV:TV ≤ 6 mL/kg; recruiting manoeuvres; driving pressure < 20 cmH₂O; PEEP 2 cmH₂O above Pflex or 16 cmH₂O if no Pflex; mean PEEP 16.4 ± 0.4 during first 36 hours
Outcomes	Primary: mortality at Day 28 Secondary: mortality before hospital discharge, barotrauma, weaning rate adjusted for APACHE II score Other outcomes: mortality in intensive care unit (ICU), death after weaning from MV, nosocomial pneumonia, use of paralysing agents > 24 hours, neuropathy after extubation, dialysis required, packed red cells infused
Notes	Discontinued during fifth interim analysis because of a significant survival difference between groups Study authors do not declare a conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information about the sequence generation process
Allocation concealment (selection bias)	Low risk	Opaque sealed envelopes with a 1:1 assignment scheme
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Incomplete blinding (blinding of participants but not of personnel) but outcomes not influenced
Blinding of outcome assessment (detection bias) All outcomes	Low risk	No blinding but outcomes not influenced
Incomplete outcome data (attrition bias) All outcomes	Low risk	Analysis on the basis of the intention‐to‐treat principle. Minor protocol violations in both groups: 4 out of 29 participants from intervention group and 1 out of 24 participants from control group
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