Erythropoietin
Erythropoietin (EPO), epoetin alfa and darbepoetin alfa
Written by Aaron CuhaReviewed Sep 2026
Also known as: EPO, Epoetin alfa, Epoetin beta, Darbepoetin alfa, Epogen, Procrit, Aranesp, Retacrit, Mircera
The kidney hormone that drives red cell production, approved since 1989 for anaemia. Three randomised trials that pushed haemoglobin toward normal found harm rather than benefit, including a doubling of stroke in 4,038 people, and every product now carries a boxed warning.
Overview
Erythropoietin is a 166 amino acid glycoprotein hormone made mainly by cells in the kidney in response to low oxygen. It tells bone marrow progenitors to survive and multiply into red blood cells. Recombinant versions transformed the treatment of anaemia in kidney disease and cancer chemotherapy, and they also became one of the most notorious doping agents in endurance sport.
The part of this literature that matters most is what happened when trials tried to correct anaemia fully rather than partially. CHOIR randomised 1,432 people with chronic kidney disease to a haemoglobin target of 13.5 or 11.3 g/dL: the higher target produced more deaths, heart attacks, strokes and heart failure hospitalisations, hazard ratio 1.34, 95 percent confidence interval 1.03 to 1.74, with no improvement in quality of life, and the trial is listed as terminated. Sponsor Johnson and Johnson Pharmaceutical Research and Development. CREATE randomised 603 people to a normal or subnormal haemoglobin target and found no cardiovascular benefit from full correction, hazard ratio 0.78, 95 percent confidence interval 0.53 to 1.14, with more people in the normalisation group requiring dialysis, 127 versus 111, p = 0.03. Sponsor Hoffmann-La Roche. TREAT randomised 4,038 people with type 2 diabetes, chronic kidney disease and anaemia to darbepoetin alfa or placebo: neither primary composite improved, and fatal or non-fatal stroke occurred in 101 versus 53 participants, hazard ratio 1.92, 95 percent confidence interval 1.38 to 2.68, p < 0.001. Sponsor Amgen. A follow-up analysis could not explain the stroke excess by blood pressure, haemoglobin, platelet count or dose.
Those results are why the current label for these products opens with a boxed warning that erythropoiesis-stimulating agents increase the risk of death, myocardial infarction, stroke, venous thromboembolism, thrombosis of vascular access and tumour progression or recurrence, and states that no trial has identified a haemoglobin target, dose or dosing strategy that avoids those risks.
Regulatory position, verified 11 September 2026 through openFDA: EPOGEN and PROCRIT (epoetin alfa, BLA 103234, Amgen) approved 1 June 1989; ARANESP (darbepoetin alfa, BLA 103951, Amgen) approved 17 September 2001; MIRCERA (methoxy polyethylene glycol-epoetin beta, BLA 125164) approved 14 November 2007; RETACRIT (epoetin alfa-epbx, BLA 125545) approved 15 May 2018. All are listed Prescription.
Mechanism of action
In humans, erythropoietin binds the EPO receptor on erythroid progenitor cells, signalling through JAK2 and STAT5 to prevent apoptosis and drive proliferation, which raises red cell mass over days to weeks. The same signalling raises blood viscosity and platelet reactivity, which is the accepted explanation for the thrombotic and stroke risk seen in the normalisation trials. EPO receptors have been described outside the marrow, including in brain tissue, which is the rationale for the neuroprotection literature; that literature has not produced an approved indication in people.
Human evidence
Large, randomised, and largely cautionary. Three major trials totalling more than 6,000 participants tested pushing haemoglobin toward normal and found no benefit, plus a doubling of stroke in the largest of them. A separate literature tests EPO for neuroprotection and has not produced an approved indication.
- CHOIR (1,432 randomised): higher haemoglobin target produced a 34 percent higher rate of the composite of death, heart attack, heart failure hospitalisation and stroke, with no quality of life gain. The registry lists the trial as terminated.
- CREATE (603 randomised): full correction of anaemia gave no cardiovascular benefit and more people required dialysis (127 versus 111, p = 0.03), though general health and physical function scores improved.
- TREAT (4,038 randomised): neither primary composite improved; stroke occurred in 5.0 percent versus 2.6 percent, hazard ratio 1.92. Transfusions fell (297 versus 496 patients) and fatigue improved only modestly.
- TREAT stroke analysis: the doubling of stroke could not be explained by baseline features, blood pressure, achieved haemoglobin, platelet count or dose.
- Neuroprotection: a systematic review of 50 human studies in 4,351 patients found the literature heterogeneous, with 13 studies reporting adverse effects and 3 attributing serious events to EPO. No indication has been approved.
- Performance: a systematic review and meta-analysis of randomised trials in healthy adults found low to moderate quality evidence that EPO improves haematological measures, maximal power output and time to exhaustion, mostly at maximal exercise intensities.
What this does not tell you: The harm trials tested a strategy, targeting a high haemoglobin, rather than the hormone in isolation, so they do not show that EPO is harmful at doses that correct severe anaemia, which remains its approved use. They also enrolled people with chronic kidney disease and diabetes, the population where cardiovascular risk is already high. The performance literature is small trials in healthy volunteers measuring laboratory exercise tests, which is not the same as competition, and it says nothing about the doses or the thrombotic risk of covert use.
Reading the research record
Erythropoietin is the clearest example in this catalogue of a biomarker that looked like an outcome. Anaemia predicts death in kidney disease, EPO corrects anaemia, and the inference that correcting it fully would reduce death was tested three times in randomised trials and failed three times, once with a doubling of stroke. The regulatory response was a boxed warning that says in plain language that no target, dose or strategy has been found that avoids the risk. That sequence is worth holding onto whenever a compound is recommended because it moves a number in the right direction.
The doping history runs alongside. EPO transformed endurance sport in the 1990s, it is prohibited in and out of competition under the World Anti-Doping Agency's section S2, and the randomised evidence in healthy volunteers does show improvements in maximal power output and time to exhaustion, which is unusual: most doping agents have a thinner evidence base than their reputation suggests. The thrombotic risk that the kidney trials documented in patients is the same mechanism that makes covert use in athletes dangerous, and the dose and haematocrit involved in doping are higher than anything the trials tested.
One peptide on this site descends directly from this hormone: ARA-290, a non-erythropoietic EPO-derived peptide designed to keep the tissue-protective signalling without raising red cell mass. That design choice exists precisely because of the trials described above.
The evidence, charted
Fig. 1 · evidence composition
4of 7 citations (57%) are in people
Counted from the citation list on this page. The Human count is the same number shown in the badge at the top. Both understate any literature larger than the sources we cite.
Fig. 2 · evidence over time
Evidence spans 6 distinct years, 2006 to 2023, counted from the citation list on this page.
Fig. 3 · legal status at a glance
US
Approved
UK
Approved
AU
Approved
CA
Approved
Approved in all four jurisdictions shown. A jurisdiction's classification is a regulatory fact, not a verdict on the science; see Legal status below for the full text and any notes.
Key studies & citations
- Human2006
Correction of anemia with epoetin alfa in chronic kidney disease
CHOIR: 1,432 people with chronic kidney disease randomised open-label to a haemoglobin target of 13.5 or 11.3 g/dL, median 16 months. The composite of death, myocardial infarction, heart failure hospitalisation and stroke occurred 125 times in the high-target group and 97 times in the low-target group, hazard ratio 1.34, 95 percent confidence interval 1.03 to 1.74, p = 0.03, with no improvement in quality of life. Registry status TERMINATED; sponsor Johnson and Johnson Pharmaceutical Research and Development.
New England Journal of Medicine - Human2006
Normalization of hemoglobin level in patients with chronic kidney disease and anemia
CREATE: 603 people with estimated GFR 15 to 35 randomised to a normal (13.0 to 15.0 g/dL) or subnormal (10.5 to 11.5 g/dL) haemoglobin target using epoetin beta, 3 years. First cardiovascular events 58 versus 47, hazard ratio 0.78, 95 percent confidence interval 0.53 to 1.14, p = 0.20. More people in the normalisation group needed dialysis, 127 versus 111, p = 0.03. General health and physical function scores improved. Sponsor Hoffmann-La Roche.
New England Journal of Medicine - Human2009
A trial of darbepoetin alfa in type 2 diabetes and chronic kidney disease
TREAT: 4,038 people with type 2 diabetes, chronic kidney disease and anaemia randomised to darbepoetin alfa targeting 13 g/dL or to placebo with rescue below 9 g/dL. Death or a cardiovascular event, hazard ratio 1.05 (0.94 to 1.17); death or end-stage renal disease, hazard ratio 1.06 (0.95 to 1.19). Fatal or non-fatal stroke 101 versus 53, hazard ratio 1.92, 95 percent confidence interval 1.38 to 2.68, p < 0.001. Fewer transfusions (297 versus 496) and only a modest improvement in fatigue. Sponsor Amgen.
New England Journal of Medicine - Human2011
Stroke in patients with type 2 diabetes mellitus, chronic kidney disease, and anemia treated with Darbepoetin Alfa: the TREAT experience
Analysis of the 154 strokes in TREAT: 5.0 percent of the darbepoetin group and 2.6 percent of the placebo group. Assignment to darbepoetin was an independent predictor (odds ratio 2.1, 95 percent confidence interval 1.5 to 2.9). The doubling could not be attributed to baseline characteristics or to post-randomisation blood pressure, haemoglobin, platelet count or dose, so the risk could not be mitigated by monitoring those.
Circulation - Review2021
Systematic Review of Erythropoietin (EPO) for Neuroprotection in Human Studies
50 studies covering 4,351 patients, screened from 388 citations. 13 studies reported adverse effects and 3 attributed serious adverse events to EPO, 2 of those in combination with tissue plasminogen activator. Most work is in spinal cord injury, perioperative outcomes and central nervous system effects. No approved neuroprotection indication has resulted.
Neurochemical Research - Review2020
Effect of erythropoietin on athletic performance: a systematic review and meta-analysis
Systematic review of randomised trials of recombinant human EPO versus placebo in healthy adults aged 18 to 65. Low to moderate quality evidence that EPO improves haematological measures, maximal power output and time to exhaustion regardless of dose, with the improvements seen almost exclusively at maximal exercise intensities that may not reflect competition conditions.
BMJ Open Sport and Exercise Medicine - Review2023
Doping and sports endocrinology: growth hormone, IGF-1, insulin, and erythropoietin
Review of erythropoietin and other erythropoiesis-affecting agents among the substances prohibited by the World Anti-Doping Agency in section S2, peptide hormones, growth factors, related substances and mimetics, which is prohibited both in and out of competition. Covers prevalence, dosing, ergogenic effects, side effects and detection.
Revista Clinica Espanola
Frequently asked questions
Why does erythropoietin carry a boxed warning?
Because randomised trials that targeted higher haemoglobin found more harm, not less. CHOIR found a 34 percent higher rate of death, heart attack, heart failure hospitalisation and stroke at the higher target, and TREAT found stroke in 5.0 percent versus 2.6 percent on placebo. The label states that no haemoglobin target, dose or dosing strategy has been identified that avoids these risks.
Does EPO improve athletic performance?
A systematic review and meta-analysis of randomised trials in healthy adults found low to moderate quality evidence for improvements in haematological measures, maximal power output and time to exhaustion. The improvements appeared almost exclusively at maximal exercise intensities, which the authors note may not reflect real competition. EPO is prohibited in and out of competition.
Is EPO used for brain protection?
It has been studied for that, without an approved indication resulting. A systematic review of 50 human studies covering 4,351 patients found a heterogeneous literature, mostly in spinal cord injury and perioperative settings, with 13 studies reporting adverse effects and 3 attributing serious events to EPO, 2 of them when given with tissue plasminogen activator.
What is the difference between epoetin alfa and darbepoetin alfa?
Both act on the same receptor; darbepoetin is more heavily glycosylated so it lasts longer. The epoetin alfa label reports a plasma half-life of 4 to 13 hours after intravenous dosing in kidney disease, while darbepoetin has a mean terminal half-life of about 21 hours. Both carry the same boxed warning.