Mohammed 2015.

Study characteristics
Methods	Study design: double‐blind, pilot RCT Location: Egypt Setting: Neonatal Intensive Care Unit (NICU) of Mansoura University Children’s Hospital, Mansoura Duration: July 2011 to July 2012
Participants	Inclusion criteria: 120 infants < 32 weeks' gestational age who exhibited apnea of prematurity within first 10 days of life. Exclusion criteria: major congenital malformations and chromosomal anomalies.
Interventions	High dose: loading dose 40 mg/kg/day and maintenance dose 20 mg/kg/day Low dose: loading dose 20 mg/kg/day and maintenance dose 10 mg/kg/day
Outcomes	Outcomes reported that are considered for this review: mortality during first admission; chronic lung disease at 36 weeks of corrected age; severe IVH ≥ grade 3; periventricular leukomalacia (PVL); lesions indicative of brain injury (detected by US and MRI): IVH, any grade; PVL; apnea; retinopathy of prematurity, ≥ grade 3; extubation failure (defined as need for re‐intubation within 72 hours of extubation from mechanical ventilation); any tachycardia; need for treatment of PDA; necrotizing enterocolitis (NEC); and somatic growth.
Notes	Caffeine initiation time: postnatal mean (SD) age of 2.5 (2.6) days in the high‐dose group 2.7 (2.8) days in the low‐dose group. Baseline imbalances: none The trial protocol is available at www.clinicaltrials.gov/ct2/show/NCT02103777 (19 April 2018)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Enrolled infants were assigned randomly to treatment groups using Internet‐based random table technique. A designated pharmacist was responsible for the randomization of selected infants and the preparation of caffeine dose.
Allocation concealment (selection bias)	Low risk	Cards in opaque, sealed envelopes.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Nursing staff and family were blinded to patient's allocation.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Investigators were blinded to patient's allocation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Data available on all of the infants randomized. No exclusions or attrition after randomization.
Selective reporting (reporting bias)	Unclear risk	The protocol is available at www.clinicaltrials.gov/ct2/show/NCT02103777 (24 April 2018). The study did not report on hydrocephalus as pre‐specified in the trial protocol. Furthermore, the duration of CPAP and postnatal steroid therapy for BPD were reported as outcomes in the study but not pre‐specified in the protocol.
Other bias	Low risk	None known