Dang 2013

Methods	Randomised controlled trial conducted in the First Hospital of Jilin University, China. Study period 21 May 2012 to 30 March 2013
Participants	Inclusion criteria: PMA ≤ 34 weeks; postnatal age < 14 days; echocardiographic diagnosis of haemodynamically significant PDA Exclusion criteria: congenital heart disease which required PDA to maintain blood flow; life‐threatening infection; recent (within the previous 24 h) intraventricular haemorrhage, Grade 3–4; urine output < 1 mL/kg/h during the preceding 8 h; serum creatinine > 88.4 mmol/L; platelet count of < 50 × 10⁹/L; hyperbilirubinaemia requiring exchange transfusion; active necrotizing NEC and/or intestinal perforation; liver dysfunction
Interventions	Eighty infants received oral paracetamol at the dose of 15 mg/kg every 6 h for 3 days, and 80 infants received oral ibuprofen at the initial dose of 10 mg/kg followed by 5 mg/kg after 24 and 48 h. Between doses of oral ibuprofen, infants of the ibuprofen group received the same volume of dextrose 5% in water (D5W) as that given for drug administration in the paracetamol group. Whether a subject received a second course of treatment depended on echocardiography evaluation after the first course. If only minor ductal shunting was present after 2 courses without the need for respiratory support, no further treatment was given.
Outcomes	Failure of PDA closure, all‐cause mortality, re‐opening of the ductus arteriosus, BPD (according to NICHD criteria: Jobe 2001), IVH (Grade I to IV; Grade I to II, Grade III to IV), PVL (diagnosed by cranial MRI), NEC (Bell staging criteria ‒ Grade IIa and above), gastrointestinal bleed, ROP (any stage), oliguria (< 1 mL/kg/h), sepsis (positive blood culture), hyperbilirubinaemia according to Maisels 2003 ‒ a serum bilirubin level higher than the exchange transfusion level according to the postnatal age and body weight), serum creatinine (µmol/L) following treatment.
Notes	In the report, although references were provided for BPD and hyperbilirubinaemia, it was not possible to ascertain exactly what criteria the authors applied. Sepsis was not defined. We wrote to the authors requesting clarification. We received a response and their definitions are included for the outcomes listed above. Funded by Jilin Department of Health
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation table (according to the published protocol)
Allocation concealment (selection bias)	Low risk	Quote: "The participants were randomly assigned at a 1:1 ratio between oral paracetamol and ibuprofen groups by using cards in sealed opaque envelopes"
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "...doctors and nurses were not blind"
Blinding of outcome assessment (detection bias) All outcomes	High risk	Quote: "...doctors and nurses were not blind"
Incomplete outcome data (attrition bias) All outcomes	Low risk	Results reported on an intention‐to‐treat basis which included patients who did not receive the complete course of treatment
Selective reporting (reporting bias)	Low risk	The study was entered in the Chinese Clinical Trial Register (http://www.chictr.org/cn). Registration number ChiCTR‐TRC‐12002177 and approved by the Hospital Ethics Committee of the First Hospital of Jilin University. There do not appear to be any deviations in study conduct between the study protocol and the publication
Other bias	Low risk	Appears free of other sources of bias