Kliethermes 1999.

Methods	Randomised controlled trial. Study duration ‐ 22 months
Participants	Single centre, Neonatal Intensive Care Unit, USA Inclusion criteria: birth weight 1000 g to 2500 g, < 1 week of age, no congenital or neurological abnormalities that interfered with cardiopulmonary status Gestational age at birth ‐ experimental: 32 weeks, SD not reported, range 26 to 35 weeks; control: 32 weeks, SD not reported, range 28 to 35 weeks; birth weight ‐ experimental: 1.73 kg, range 1.05 kg to 2.43 kg; control: 1.64 kg, range 1.0 kg to 2.35 kg; twins ‐ experimental: 8 (21%); control: 16 (35%) Sample size: 99 randomised (47 experimental/tube alone, 52 control/bottle); 84 included in analysis (38 experimental/tube alone, 46 control/bottle)
Interventions	Both groups of infants breast fed when mother was present. Experimental group: feeds given by indwelling size 3.5 FG nasogastric tube when mother not available, or top‐up after breast feed required. Tube was removed during last 24 to 48 hours of parent 'rooming‐in' period; a cup or syringe was used during this time if needed. Control group: fed by bottle when mother not available, or top‐up after breast feed required. Indwelling nasogastric tube was removed as directed by clinicians.
Outcomes	Breast feeding, exclusive and partial, at discharge home, and at 3 days, 3 months and 6 months post discharge. Length of hospital stay, apnoea/bradycardia, weight gain to discharge home, infection rate
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomization was achieved by using sealed envelopes, which were physically mixed and drawn in random sequence after enrolment of the dyad into the study".
Allocation concealment (selection bias)	Low risk	Quote: "...sealed envelopes"
Blinding (performance bias and detection bias) All outcomes	High risk	Comment: Blinding of intervention not possible. Blinding of outcome assessment not reported
Incomplete outcome data (attrition bias) On discharge home	High risk	Missing outcome data (n = 15, 15%) (experimental 9, control 6): Deaths 1: experimental Clinical conditions 4: experimental 2 (chronic lung disease, congenital heart defect); control 2 (NEC, subglottic stenosis) Transfer to another hospital 2: 1 in each group Protocol violation 5: experimental 3, control 2 Maternal conditions 3: experimental 2 (scleroderma, +ve cocaine screen), control 1 (+ve cocaine screen) Comment: high risk of bias due to incomplete outcome data. Difference in proportion of missing data across groups (19% experimental, 12% control). For 4 infants, valid reasons were given for missing outcome data (1 died, 2 were transferred to another hospital).
Incomplete outcome data (attrition bias) 3 months post discharge	High risk	Missing outcome data (n = 15, 15%) (experimental 9, control 6): Deaths 1: experimental Clinical conditions 4: experimental 2 (chronic lung disease, congenital heart defect); control 2 (NEC, subglottic stenosis) Transfer to another hospital 2: 1 in each group Protocol violation 5: experimental 3, control 2 Maternal conditions 3: experimental 2 (scleroderma, +ve cocaine screen), control 1 (+ve cocaine screen) Comment: high risk of bias due to incomplete outcome data
Incomplete outcome data (attrition bias) 6 months post discharge	High risk	Missing outcome data (n = 15, 15%) (experimental 9, control 6): Deaths 1: experimental Infant clinical conditions 4: experimental 2 (chronic lung disease, congenital heart defect); control 2 (NEC, subglottic stenosis) Transfer to another hospital 2: 1 in each group Protocol violation 5: experimental 3, control 2 Maternal conditions 3: experimental 2 (scleroderma, +ve cocaine screen), control 1 (+ve cocaine screen) Comment: high risk of bias due to incomplete outcome data
Selective reporting (reporting bias)	Low risk	Before clinical trial registration requirements, all expected outcomes reported
Other bias	Low risk	Nil noted