BACKGROUND: Extended criteria donor (ECD) and donation after circulatory death (DCD) kidneys are at increased risk of delayed graft function (DGF). Experimental evidence suggests that erythropoietin (EPO) attenuates renal damage in acute kidney injury. This study piloted the administration of high dose recombinant human EPO-beta at implantation of ECD and DCD kidneys, and evaluated biomarkers of kidney injury post-transplant. METHODS: Forty patients were randomly assigned to receive either rhEPO-b (100,000 iu) (n = 19 in the intervention group, as 1 patient was un-transplantable post randomisation), or placebo (n = 20) in this, double blind, placebo-controlled trial at Manchester Royal Infirmary from August 2007 to June 2009. Participants received either an ECD (n = 17) or DCD (n = 22) kidney. Adverse events, renal function, haematopoietic markers, and rejections were recorded out to 90 days post-transplant. Biomarkers of kidney injury (neutrophil gelatinase-associated lipocalin, Kidney Injury Molecule-1 and IL-18) were measured in blood and urine during the first post-operative week. RESULTS: The incidence of DGF (53% vs 55%) (RR = 1.0; CI = 0.5-1.6; p = 0.93) and slow graft function (SGF) (32% vs 25%) (RR = 1.1; CI = 0.5-1.9; p = 0.73) respectively, serum creatinine, eGFR, haemoglobin and haematocrit, blood pressure, and acute rejection were similar in the 2 study arms. High dose rhEPO-b had little effect on the temporal profiles of the biomarkers. CONCLUSIONS: High dose rhEPO-b appears to be safe and well tolerated in the early post- transplant period in this study, but has little effect on delayed or slow graft function in recipients of kidneys from DCD and ECD donors. Comparing the profiles of biomarkers of kidney injury (NGAL, IL-18 and KIM-1) showed little difference between the rhEPO-b treated and placebo groups. A meta-analysis of five trials yielded an overall estimate of the RR for DGF of 0.89 (CI = 0.73; 1.07), a modest effect favouring EPO but not a significant difference. A definitive trial based on this estimate would require 1000-2500 patients per arm for populations with base DGF rates of 50-30% and 90% power. Such a trial is clearly unfeasible. TRIAL REGISTRATION: EudraCT Number 2006-005373-22 ISRCTN ISRCTN85447324 registered 19/08/09.