Date	Event	Description
25 March 2017	New search has been performed	This review was updated in March 2017.
25 March 2017	New citation required and conclusions have changed	For this update, we included erythropoietin (EPO)/darepoetin (erythropoiesis‐stimulating agents ‐ ESAs) as possible neuro protective agents and as agents protecting against necrotising enterocolitis. We provided additional information in the Background section to justify these inclusions. We changed the review title to "Early erythropoiesis‐stimulating agents in preterm or low birth weight infants". In the previous update, we included a post hoc analysis that examined results for retinopathy of prematurity (ROP) ≥ 3 for all available studies, regardless of the age of the infant at initiation of treatment with EPO. For this update, we excluded that analysis and corresponding references to studies that initiated EPO treatment beyond 7 days' postnatal age. We excluded post hoc analyses on the basis of study quality. No current evidence suggests that early EPO increases the risk of ROP. Promising evidence indicates that early ESAs may prevent necrotising enterocolitis (NEC), intraventricular haemorrhage (IVH), and periventricular leukomalacia (PVL), while improving long‐term neurodevelopmental outcomes. Ongoing trials will prove whether this is the case.
19 July 2016	Amended	We made minor edits.
9 September 2013	New citation required and conclusions have changed	We initiated this unscheduled update after receiving feedback from Dr. Robin Ohls. Dr. Ohls questioned our inclusion of the study by Dr. Romagnoli and coworkers (Romagnoli 2000) in this review (early erythropoietin (EPO) review) and suggested that this study should be included in the late EPO review (New Reference; Ohls 2013a). We contacted Dr. Romagnoli, who informed us that the mean (± SD) age of infants at the start of EPO treatment was 10 ± 1 days. We therefore moved the study to the late EPO review. In addition, Dr. Ohls informed us that Bierer 2006 was a report on a subgroup of the Ohls 2001A study. It could not be ascertained from the published report that it was a duplicate publication. As Ohls 2001A reported all outcomes, we excluded Bierer 2006. We conducted literature searches on 1 July 2013. We identified 2 new studies for inclusion (Yasmeen 2012; Ohls 2013), as well as 2 new studies for exclusion (Saeidi 2012; Costa 2013). As expected, when we excluded 2 studies (Romagnoli 2000; Bierer 2006) and added the results of 2 new studies (Yasmeen 2012; Ohls 2013), almost all point estimates and confidence intervals changed. The major difference was that the outcome of retinopathy of prematurity (ROP) stage ≥ 3 was no longer statistically significantly increased but remained a matter of concern as the typical risk ratio (RR) was 1.37 (95% confidence interval (CI) 0.87 to 2.17) with no heterogeneity (I2 = 0%); the typical risk difference (RD) was 0.03 (95% CI ‐0.01 to 0.06) with low heterogeneity (I2 = 29%). Our decision to divide the EPO studies into early and late studies was based on initiation of EPO treatment at the cutoff of ≤ 7 days of age for early and > 7 days for late treatment with EPO. Although arbitrary, we chose this cutoff on the basis of previously published meta‐analyses (Garcia 2002; Kotto‐Kome 2004) to allow us to compare results of our reviews with those of previously published reviews. Concerns about a possibly increased risk of ROP remain, and because of the arbitrary cutoff age for early versus late EPO, we decided post hoc to perform a secondary analysis of data from all studies that reported on ROP (stage ≥ 3), regardless of age at initiation of EPO treatment. We included 3 studies from the late EPO review: Shannon 1995; Al‐Kharfy 1996; and Romagnoli 2000. In these studies, investigators initiated treatment with EPO at 10 to 17 days; 10 ± 1 (SD) days; and 23 to 24 days, respectively. The outcome of ROP ≥ 3 was statistically significantly increased following EPO treatment initiated at any age during the neonatal period (typical RR 1.48, 95% CI 1.02 to 2.13, P = 0.04 with no heterogeneity (I2 = 0%); typical RD 0.03, 95% CI 0.00 to 0.06, P = 0.03 with moderate heterogeneity (I2 = 50%); number needed to treat for an additional harmful outcome 33, 95% CI 17 to infinity). When the Romagnoli 2000 study was moved to the late EPO review, the risk for ROP stage ≥ 3 was not statistically significantly increased in the late EPO review, but results showed a trend in the direction of increased risk (RR 1.73, 95% CI 0.92 to 3.24; RD 0.05, 95% CI ‐0.01 to 0.10; 3 trials enrolling 442 infants) (New Reference). In the latest study by the Romagnoli group (Costa 2013), trial authors compared the use of early intravenous EPO vs subcutaneous EPO and reported that the incidence of stage ≥ 3 was high in both groups, at 16% and 14%, respectively (overall 15%), similar to the incidence of 17% in the Romagnoli 2000 study. Thus our concerns about increased risk of ROP ≥ 3 following EPO treatment remain.
1 July 2013	New search has been performed	We conducted literature searches on 1 July 2013.
2 May 2012	New citation required but conclusions have not changed	This update identified several randomised controlled studies that were excluded, as they compared one EPO dosing regimen vs another, did not provide numbers randomised to EPO and placebo groups, or did not state the dose of EPO. Review authors identified cohort/case‐control studies reporting on possibly increased risk of ROP following treatment with EPO and ROP. Review authors did not change the review conclusions. Studies using EPO for neuroprotection will be reviewed separately (Yu 2010).
2 May 2012	New search has been performed	This updates the review "Early erythropoietin for preventing red blood cell transfusions in preterm and/or low birth weight infants", which was published in the Cochrane Database of Systematic Reviews, Issue 3, 2006 (Ohlsson 2006), and updated in August 2009.
21 August 2009	New search has been performed	This updates the review "Early erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants", which was published in the Cochrane Database of Systematic Reviews, Issue 3, 2006 (Ohlsson 2006). We have included 4 additional studies (adding 145 additional infants) in this review update. Studies using EPO for neuroprotection are emerging.