Dunn 1988.

Methods	Randomised controlled trial
Participants	VLBW infants with respiratory distress syndrome treated with mechanical ventilation and with an umbilical artery catheter in situ. Setting: Rainbow Babies and Children's Hospital, Cleveland, USA
Interventions	Minimal enteral nutrition (N = 19) vs. enteral fasting (N = 20) until 9 days after birth. Intervention group infants received minimal enteral feeds from 48 hours at 15 to 20 ml/kg/day of diluted preterm formula milk
Outcomes	Time to establish full enteral feeds Incidence of necrotising enterocolitis Growth: time to regain birth weight and growth throughout study period Duration of phototherapy Mortality Incidence of sepsis Duration of hospital stay
Notes	All infants received formula milk. Feeds were given by intermittent gavage nasogastric technique. Data enabling calculation of SD relating to duration of hospital stay were not provided. We have imputed this information from standard deviations provided by Meetze 1992, a trial with similar sample size, as recommended by the Cochrane Handbook
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Stratified into groups according to birth weight then randomised using cards in paired envelopes
Allocation concealment (selection bias)	High risk	Unclear if envelopes were sealed ‐ possibility that allocation groups could have been predicted
Blinding (performance bias and detection bias) All outcomes	Unclear risk	No blinding of care givers or investigators after allocation
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No reference to whether interpretation of radiographs was blind
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	9 infants were excluded from some of the outcome data: 5 deaths in the control group, 1 death in the intervention group and 3 infants removed from the minimal enteral nutrition group due to severe unrecognised aortic coarctation, systemic candidiasis and ileus precluding the introduction of feeds. These infants have been included in intention‐to‐treat analysis. Uncertainty exists about whether these infants went on to develop necrotising enterocolitis as this is not formally reported. We have assumed they did not