Ngan 2017.

Methods	Prospective randomised parallel controlled trial Setting: Delivery Room, NICU, Royal Alexandra Hospital (RAH), Edmonton, Canada Conducted: June 2013 to August 2014
Participants	Inclusion criteria: infants between 23+0 and 32+6 weeks of gestation who require respiratory support for resuscitation in the delivery room Exclusion criteria: congenital abnormality or condition that might have an adverse effect on breathing or ventilation; absence of parents' consent for inclusion in the study
Interventions	SLI group: 2 PIPs of 24 cmH₂O. Duration of first SLI was 20 seconds. Duration of second SLI was 20 or 10 seconds if ECO₂ value was < or > 20 mmHg, respectively. After SLIs, CPAP if breathing spontaneously or, if found to have apnoea or laboured breathing, mask IPPV at a rate of 40 to 60 bpm Control group: mask IPPV, ventilation rate of 40 to 60 inflations/min until spontaneous breathing, at which time CPAP will be provided
Outcomes	Primary outcome: BPD (rate of respiratory support or supplemental oxygen at corrected gestational age of 36 weeks) Secondary outcomes: rate of endotracheal intubation in the DR or the NICU, duration of MV and non‐invasive ventilation, neonatal death, air leak, PDA (medical or surgical), NEC, ROP, periventricular leukomalacia, abnormal cranial ultrasound (including IVH, parenchymal injury, and ventriculomegaly), surfactant administration, postnatal steroids, respiratory support or oxygen requirements at 28 days, neonatal death before discharge
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation scheme (1:1 ratio). Randomisation stratified according to gestational age (to infants 23+0 to 27+6 and 28+0 to 32+6 weeks). Twins and/or triplets were randomised as individuals
Allocation concealment (selection bias)	Unclear risk	A sequentially numbered, brown, sealed envelope contained a folded card box with treatment allocation opened by the clinical team immediately before delivery. Timing of randomisation resulted in many post‐randomisation exclusions with the potential of inadequate allocation concealment, as more post‐randomisation exclusions occurred in the SLI group than in the control group
Blinding (performance bias and detection bias) All outcomes	High risk	Assigned intervention could not be blinded to the resuscitation team
Blinding of outcome assessment (detection bias) All outcomes	Low risk	After admission into the NICU, the clinical team was not made aware of treatment allocation. In addition, both data collector and outcome assessor were unaware of group allocation. The research team was not involved in clinical care of the infants
Incomplete outcome data (attrition bias) All outcomes	High risk	Post‐randomisation exclusion (27%) resulted in fewer included infants in the SLI group; this discrepancy might have yielded different results
Selective reporting (reporting bias)	Low risk	Protocol was registered at ClinicalTrials.gov (NCT01739114)
Other bias	Unclear risk	Planned sample size of 93 infants in each group was not achieved. Moreover, incidence of the primary outcome in the control group was lower than assumed for the sample size calculation, further underpowering the trial to detect the desired effect size