Background and purpose: Numerous preclinical findings and a clinical pilot study suggest that recombinant human erythropoietin (EPO) provides neuroprotection that may be beneficial for the treatment of patients with ischemic stroke. Although EPO has been considered to be a safe and well-tolerated drug over 2 decades, recent studies have identified increased thromboembolic complications and/or mortality risks on EPO administration to patients with cancer or chronic kidney disease. Accordingly, the double-blind, placebo-controlled, randomized German Multicenter EPO Stroke Trial (Phase II/III; ClinicalTrials.gov Identifier: NCT00604630) was designed to evaluate efficacy and safety of EPO in stroke.

Methods: This clinical trial enrolled 522 patients with acute ischemic stroke in the middle cerebral artery territory (intent-to-treat population) with 460 patients treated as planned (per-protocol population). Within 6 hours of symptom onset, at 24 and 48 hours, EPO was infused intravenously (40,000 IU each). Systemic thrombolysis with recombinant tissue plasminogen activator was allowed and stratified for.

Results: Unexpectedly, a very high number of patients received recombinant tissue plasminogen activator (63.4%). On analysis of total intent-to-treat and per-protocol populations, neither primary outcome Barthel Index on Day 90 (P=0.45) nor any of the other outcome parameters showed favorable effects of EPO. There was an overall death rate of 16.4% (n=42 of 256) in the EPO and 9.0% (n=24 of 266) in the placebo group (OR, 1.98; 95% CI, 1.16 to 3.38; P=0.01) without any particular mechanism of death unexpected after stroke.

Conclusions: Based on analysis of total intent-to-treat and per-protocol populations only, this is a negative trial that also raises safety concerns, particularly in patients receiving systemic thrombolysis.