Tarnow‐Mordi 2017.

Methods	Randomised controlled trial With minimisation and stratification according to gestational age (< 27 weeks vs ≥ 27 weeks), by centre, and multiple birth status (singleton birth vs. multiple birth). Infants of multiple births underwent randomisation individually.
Participants	Inclusion criteria Women expected to give birth before 30 weeks' gestation. Babies were eligible if obstetricians or maternal–fetal medicine specialists considered that they might be born before 30 weeks of gestation. 24.8% multiple births: babies: 1176 singletons; 344 twins; 42 triplets and 4 quadruplets = 1566 babies. Mothers: 1176 singletons; 172 twins; 14 triplets and 1 quads = 1363 mothers in total. N = 1634 babies randomised. Mortality data on 1156 babies and primary analyses on 1497. 1363 mothers included. Exclusion criteria Exclusion criteria included fetal haemolytic disease, hydrops fetalis, twin–twin transfusion, genetic syndromes, and potentially lethal malformations.
Interventions	Intervention: DCC Clamping 60 secs or more after birth, with the infant held as low as possible below the introitus or placenta and without palpation of the cord. N = 818 babies with data on 748 (and on 784 for death) Comparator: ECC Cord clamping within 10 secs N = 816 babies with data on 749 (and on 782 for death) Additional information Gestational subgroup: < 32‐34 weeks' gestation Resuscitation with cord intact: not available Access to NICU: yes Length of delay: 60 secs Baby placed: low Uterotonic: recorded as an outcome UCM: n/a Comparison 1 DCC with neonatal resuscitation after cord clamping vs ECC (subgroup by gestation) Subgroup 1: < 32‐34 weeks' gestation Comparison 2 DCC with neonatal resuscitation after cord clamping vs ECC (subgroup by type intervention) Subgroup 4: DCC at 1‐2 mins with baby low (+ gravity)
Outcomes	Primary outcomes Composite of death or major morbidity ‐ initially defined as severe brain injury on postnatal ultrasonography, severe RoP, NEC, late‐onset sepsis, or chronic lung disease, each diagnosed by 36 completed weeks of postmenstrual age. The protocol was amended in July 2016 to reflect the updated primary outcome of death, severe brain injury, severe RoP, NEC, or late‐onset sepsis. Secondary outcomes Death by 36 completed weeks of postmenstrual age Death or severe brain injury on postnatal ultrasonography Severe brain injury Late cerebral abnormality on ultrasonography IVH (all grades, grade 3 or 4, and grade 4 only) Severe RoP NEC Late‐onset sepsis Treated PDA Chronic lung disease, defined as below Additional secondary outcomes of death, disability, and death or disability by 3 years are not reported here Tertiary outcomes (analyses of which were considered to be hypothesis generating) Birthweight, Number of red‐cell transfusions by 36 weeks, Temperature of the infant on admission, Peak bilirubin level in the first week, Peak Hct in the first week; Duration of hospital stay if the infant was discharged alive, Maternal blood transfusion for postpartum haemorrhage, The use of uterotonic drugs, Exchange transfusions by 36 weeks of gestation, Because rates of endotracheal intubation at delivery can vary considerably among centres and may not correlate with the rate of morbidity, they were not recorded, 5‐minute Apgar score of less than 4 was considered to be a better index of initial risk than endotracheal intubation, Apgar score at 1 minute and 5 mins and an Apgar score of less than 4 at 5 mins were prespecified as tertiary outcomes in the statistical analysis plan.
Notes	Setting: 25 centres in 7 countries: Australia; New Zealand; Canada; France; Northern Ireland; Pakistan and USA. Led by University of Sydney, Australia. Dates: December 2010 to January 2017 Declaration of interest: disclosure forms provided by the authors are available with the full text of this article at NEJM.org. Trial funding source: supported by the National Health and Medical Research Council (NHMRC) and by the NHMRC Clinical Trials Centre, University of Sydney. Further information: The trial registration in 2009 was for a 4‐arm trial: 1) DCC with baby held low; 2) UCM; 3) DCC with baby low + UCM; 4) ECC. However, the APTS trial undertaken was only on DCC with baby held low vs ECC. For the outcome of death, we have added the 5 babies who were stillborn in each group so we report all deaths after randomisation. D. Osborn kindly provided additional data on 'Maternal blood transfusions' for women with singleton births as we were unable to use the data in the publication because the randomisation was by baby. He also clarified the data on IVH and we included that from Supplementry Appendix Table 4 (previously Table 3).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: “…computer generated randomisation…”
Allocation concealment (selection bias)	Low risk	Quote: “The computer generated randomisation lists used with the interactive voice response system will be prepared by an independent statistician at the NHMRC Clinical Trials Centre, University of Sydney. The randomisation code will be stored securely by the statistical group at the centre.” (Trial registration form) “Randomization was performed centrally … with the use of an interactive voice‐response system with minimization and with stratification according to gestational age (<27 weeks vs. ≥27 weeks), center, and multiplebirth status (singleton birth vs. multiple birth).” (2017 publication)
Blinding of participants and personnel (performance bias) All outcomes	High risk	Clinicians at the birth cannot be blinded and it is unclear whether women knew or not but women knowing is unlikely to affect outcome assessment.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Qquote: “For practical reasons, no attempt was made to make staff who were diagnosing these morbidities unaware of the timing of cord clamping.”
Incomplete outcome data (attrition bias) All outcomes	Low risk	1634 infants randomised. 68 were excluded and 69 had missing data for ≥ 1 component of primary outcome. So overall loss of primary data were 8%.
Selective reporting (reporting bias)	Low risk	All outcomes from full trial protocol are reported in the 2017 paper and Supplementary Appendix.
Other bias	Low risk	No other biases apparent.