Akdag 2014.

Study characteristics
Methods	Prospective, double‐blind, controlled study, conducted in the neonatal intensive care unit of Zekai Tahir Burak Maternity Teaching Hospital, Ankara, Turkey Period of study: August 2009 to October 2010 4‐arm study comparing pentoxifylline, IgM‐enriched IVIG, and both against placebo
Participants	Newborn infants with sepsis Exclusion criteria were major congenital abnormalities, intraventricular haemorrhage (grade 3 or 4), symptoms of a congenital infection, and inborn errors of metabolism Total number of participants: 204 in 4 groups Placebo group: (n = 51, median and range of bwt 1410 g (620 g to 4300 g), gestational age 31 (25 to 40) weeks) Pentoxifylline group (n = 51, median and range of bwt 1490 g (620 g to 4580 g), gestational age 31 (24 to 42) weeks) Pentaglobin group (n = 51, median and range of bwt 1320 g (620 g to 3860 g), gestational age 30 (24 to 41) weeks) Pentoxifylline + pentaglobin group (n = 51, median and range of bwt 1330 g (540 g to 4100 g), gestational age 30 (24 to 40) weeks)  89/204 had positive blood cultures, and 13 had positive urine cultures.
Interventions	Pentoxifylline (6 mg/kg IV over 4 hours for 3 consecutive days), pentoxifylline + IgM‐enriched IVIG (pentaglobin, 250 mg/kg over 4 hours for 3 consecutive days), IgM‐enriched IVIG (pentaglobin for 3 consecutive days), or placebo (normal saline for 3 consecutive days)
Outcomes	Mortality, NEC, oliguria/anuria, hepatic failure, disseminated intravascular coagulation, pulmonary haemorrhage, and lab parameters (white blood cell count, C‐reactive protein, interleukin‐6, TNF‐α, and neutrophil CD64)
Notes	The authors state that there were no funding or conflicts of interest. Length of hospital stay was reported in days (mean), but did not have standard deviations to use in meta‐analysis.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	None reported
Allocation concealment (selection bias)	Low risk	Sealed, opaque envelopes containing the randomisation arm.
Blinding (performance bias and detection bias) All outcomes	Low risk	The infusions were covered with plastic covers, and infusion vials were identical for the different arms.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The infusions were covered with plastic covers, and infusion vials were identical for the different arms.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up reported.
Selective reporting (reporting bias)	Low risk	All outcomes reported
Other bias	Low risk	None noted