Song 2016

Methods	Randomised controlled trial Study location: 2 NICUs in China: Third Affiliated Hospital of Zhengzhaou University and Zhengzhou Children’s Hospital Study period: January 2009 to June 2013
Participants	Preterm infants, PMA ≤ 32 weeks' gestation, and < 72 hours of age
Interventions	EPO group (n = 366) received EPO at 500 IU/kg IV every other day for 2 weeks. Cumulative dose of 3500 IU/kg. First dose within 72 hours after birth. Placebo group (n = 377) received an equivalent volume of normal saline IV.
Outcomes	Head U/S within 3 days after birth, then weekly until discharge MRI at 49 weeks' PMA ICH PVL BPD NEC Sepsis ROP graded according to the international classification of ROP At 18 months' corrected age, neurological exam and Mental Developmental Index (Bayley Scales – Second Edition) Hearing test,‐ Deafness defined as a hearing disability that required amplification. Blindness defined as corrected visual acuity < 20/200 Moderate or severe disability defined as survival with at least 1 of the following complications: cerebral palsy, MDI < 70, deafness, or blindness
Notes	ROP grades are not reported. We wrote in March 2017 to the corresponding author, but we had not received a response as of 11 June 2017.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐based random number generator
Allocation concealment (selection bias)	Low risk	Group assignment for each consecutive participant was concealed in a sealed envelope before participants were included.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Doctors and nurses responsible for treatment were not blinded according to the rules of medical procedure in China.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	According to trial authors, "The investigators performing the short‐term and long‐term outcome assessments and the parents were blinded to patients' group allocation". It is difficult to understand this statement, as doctors and nurses were not blinded to treatments. Final evaluation at 18 months' corrected age was performed by doctors from the Child Growth and Development Department, who were blinded to the treatment protocol and were not allowed to have access to the treatment history of infants.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	All 743 randomised infants were accounted for in short‐term outcomes. For outcomes at 18 months, 309 children in the EPO group (36 infants were lost to follow‐up) and 304 in the placebo group (39 children were lost to follow‐up) were assessed (90% in both groups).
Selective reporting (reporting bias)	High risk	Study was registered as NCT02036073 but was registered in December 2013, after recruitment had been completed. Study started to recruit patients in January 2009. Therefore we are unable to tell whether any deviations occurred from the study protocol that was established before the study start. In the protocol, the primary outcome measure was: Incidence of MDI < 70 at corrected age of 18 months, and secondary outcome measures were Incidence of ROP at corrected age 42 weeks. In the full report, primary outcomes are listed as death, disability, or death + disability at 18 months' corrected age. ROP is listed as a neonatal complication.
Other bias	Low risk	Appears free of other bias