Alan 2014.

Methods	Randomised controlled trial
Participants	Inclusion criteria Babies with gestational age ≤ 32 weeks and estimated BW ≤ 1500 g assessed by the obstetrics team. Multiple births included. 5 in each group. N = 48 babies randomised. Exclusions after randomisation left 44 and then 38 babies in the analyses (see below). Exclusion criteria Suspected twin‐to‐twin transfusion syndrome or discordant twins; major congenital anomalies or chromosomal anomalies; vaginal bleeding due to placenta previa or abruption or placental tear; haemolytic disease of the fetus and newborn like Rhesus sensitisation; IUGR; maternal gestational diabetes treated with insulin; hydrops fetalis; and refused parental consent.
Interventions	Intervention: UCM Infants were placed at the level of placenta in caesarean deliveries and below the level of placenta in vaginal deliveries in UCM group (group 1). The umbilical cord was held at 25 to 30 cm distance from the baby and milked vigorously toward the umbilicus for 3 times at a speed of approximately 5 cm/sec by the attending neonatologist before clamping N = 24 babies randomised but 2 were excluded because inappropriate milking – leaving N = 22. A further 3 babies were excluded in days 2‐7 for death or major bleeding. Comparator: ECC Infants in the control group (group 2) had immediate cord clamping (< 10 secs). N = 24 babies randomised but 2 excluded because: 1) tracheal bleeding during resuscitation in a preterm infant with 23 weeks of gestation, 2) 24 weeks' gestation infant who did not respond to resuscitation) leaving N = 22. A further 3 babies were excluded on days 2‐7 for death or major bleeding. Additional information Gestational subgroup: < 32‐34 weeks' gestation Resuscitation with cord intact: not available Access to NICU: yes Length of delay: n/a Baby placed: at the level of placenta at caesareans and below the placenta in vaginal births in UCM group Uterotonic: ‐ no information UCM: 3 x milking 25‐30 cm, cord intact at 5 cm Comparison 7 UCM vs ECC (subgroup by gestation) Subgroup 1: < 32‐34 weeks' gestation Comparison 8 UCM vs ECC (subgroup by type of intervention) Subgroup 1: cord intact during UCM
Outcomes	Primary Number and volume of PRBC transfusions received by the infant during the first 35 days of life. Secondary Hemodynamic variables during the first 24 hrs of life Number of infants undergoing PRBC transfusion within the first 3 days and the first 5 weeks of life, total volume of transfusions, and total phlebotomy losses during the first 5 weeks Haemodynamic parameters such as heart rate, respiratory rate, mean blood pressure on admission to the NICU, and at 6th, 12th, and 24th hour of hospitalisation, urine output, need for volume expanders (10 mL/kg normal saline), and inotrope drugs (dopamine, dobutamine, and adrenaline) during the first 24 hrs Haematological parameters such as Hb, Hct, white blood cell count, and Platelet counts at the first and 24th hour, day 7, and weekly thereafter; and Clinical outcomes such as percentage of nosocomial sepsis during the first 35 days of life, surfactant requirement for RDS, (PDA; without any treatment/with medical or surgical treatment), IVH (staging according to Papile13), NEC (NEC; staging according to Bell and colleagues 14), bronchopulmonary dysplasia (BPD; was defined by need for supplemental oxygen at 36 weeks postconceptional age), retinopathy of prematurity (RoP; was defined according to the International Classification15), hospital stay, and death. In addition, serum potassium, total bilirubin, blood urea nitrogen (BUN), creatinine, albumin, and total protein levels at 24th hour after birth and maximum serum total bilirubin and potassium levels within the first week of life were recorded for safety measures.
Notes	Setting: Ankara, Turkey Dates: April 2011 to February 2013 Declaration of interest: not reported Trial funding source: not reported Further information: Twins: in case of twin pregnancies, the first one was randomised and the second one was automatically assigned to the opposite arm without randomisation. Many outcomes were unable to be included in this review because they were reported as medians. Late sepsis chose data for 35 days. Data seems to have a high incidence in both groups. S Alan kindly provided further data (Tables 3,4,5) on the outcome measures in email of 14 April 2016, in particular the specific numerators and denominators for the various outcomes, including infant death which we report as UCM 2/22 vs ECC 3/24 (including the death on delivery suite).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Just reports 'randomly assigned'.
Allocation concealment (selection bias)	Unclear risk	Quote:"...Sequentially numbered sealed nontransparent envelopes...", however, it is not possible to have concealment of allocation if sequence generation is unclear.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote:"The intervention was unmasked for the attending neonatal and obstetric teams in the delivery room."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	There is no mention of whether the authors tried to blind outcome assessment.
Incomplete outcome data (attrition bias) All outcomes	High risk	48 randomised to 24 each group. UCM: 2 excluded in delivery room for inappropriate milking and 3 excluded later because of death or major bleeding. So N = 19 for analysis – loss of 5/24 = 21% ECC: 2 excluded in delivery room because of death or major bleeding and 3 excluded later because of death or major bleeding. So N = 19 for analysis ‐ loss of 5/24 = 21%
Selective reporting (reporting bias)	Low risk	Very comprehensive outcome measures listed in the methods section.
Other bias	Unclear risk	Baseline demographics were similar. Trial was small for assessing clinical outcomes, no other biases apparent.