Liu 2011.

Study characteristics
Methods	Randomised trial Single centre Randomisation method: random number table Blinding of randomisation: unclear Blinding of intervention: no Complete follow‐up: yes Blinding of outcome measurement: not clear
Participants	GA < 35 weeks Mechanical ventilation RDS Age < 12 hours old Arterial blood gas pH < 7.25; PaO₂ < 50mmHg; PaCO₂ > 50 mmHg PaO₂/FiO₂ ≤ 250 mmHg; a/APO₂ ≤ 0.22 Exclusion criteria: congenital lung abnormalities, pulmonary haemorrhage, pneumothorax, congenital pneumonia, meconium aspiration, wet lung, complex congenital heart disease, grade 3/4 intracranial haemorrhage Sample size: 84 SIPPV + VG: 31 CMV: 30 HFOV: 23
Interventions	SIPPV +VG vs CMV vs HFOV
Outcomes	Primary: Duration of mechanical ventilation Oxygenation status Secondary: Death Air leak Ventilation associated pneumonia Intraventricular haemorrhage (Grade 3/4)
Notes	Ventilator types: Babylog 8000 plus, Sensormedics 3100 A
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Comment: random number table used
Allocation concealment (selection bias)	High risk	Comment: probably not used
Blinding (performance bias and detection bias) All outcomes	High risk	Comment: probably not done
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: probably not done
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: probably not done
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: outcomes of all participants reported.