Faerk 2001.

Study characteristics
Methods	Parallel randomised controlled trial, multiple centres
Participants	Inclusion criteria: preterm infants with gestational age of < 32 weeks and no major congenital malformations Exclusion criteria: not stated Setting: NICU at Rigshospitalet and Hvidovre Hospital in Copenhagen. Some infants were discharged to local neonatal units at Glostrup and Hilleroed Hospitals, but continued in the study. Timing: not stated
Interventions	Human milk (mother + donor) supplemented with protein, calcium, and phosphate (Eoprotin, 0.4 g protein, 35 mg Ca and 17 mg P/100 mL; N = 51) versus human milk + phosphate (10 mg P/100 mL; N = 52) All infants received 800 IU of vitamin D daily. The intervention ceased at the end of the 36th gestational week or when infants were fully breast fed, whichever came first.
Outcomes	Primary: total body mineral content at term by DEXA scan Secondary: body weight, crown‐heel length, head circumference, body composition, and bone area at term.
Notes	Conflicts of interest: not stated Source of funding: not stated This study had three arms ‐ unsupplemented human milk or formula versus human milk supplemented with phosphate versus human milk supplemented with protein, calcium and phosphate. Infants in the unsupplemented arm were fed only their own mother's milk. However, if the mother's breast milk was insufficient, her infant was fed formula as well. Data on the unsupplemented human milk or formula arm were not included in this review.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No mention of how the random sequence was generated.
Allocation concealment (selection bias)	Unclear risk	The randomisation was done in an independent centre in Copenhagen (a human milk bank). They used sealed envelopes which gave some assurance. However, it was not stated if the envelopes were opaque.
Blinding (performance bias and detection bias) All outcomes	Low risk	“The study was kept blinded for the PI, the parents, all the attending physicians, and at Hvidovre Hospital, Rigshospitalet, all the attending staff”.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details
Incomplete outcome data (attrition bias) All outcomes	Low risk	A total of 217 infants fulfilled the inclusion criteria, parents of 52 infants refused to participate. 165 were randomised, 15 withdrew from the fortifier group due to unacceptable or no DEXA scans and other reasons, while 12 withdrew from the unfortified group also for unacceptable or no DEXA scans and other reasons. Other reasons included: abdominal discomfort in three cases, non‐acceptance of blinding principle in one case, on request of the attending physician in one case, and in two cases no reasons were stated. The remaining 11 withdrawals were from the formula group, which is not part of this review.
Selective reporting (reporting bias)	Low risk	The study protocol was available and all of the study’s prespecified (primary and secondary) outcomes were reported in the prespecified way.
Other bias	Unclear risk	No demographic tables were provided to make the assessment. The sample included sick preterm infants