NEW YORK (Reuters Health) – Recombinant human erythropoietin (EPO) provides no benefit after acute ischemic stroke, and may even be dangerous, particularly in patients receiving systemic thrombolysis.
These findings, the results of a phase II/III trial, are reported in the December issue of Stroke.
In fact, the study was “designed to reproduce the promising results” of an earlier trial that showed improved recovery with EPO treatment after ischemic stroke, according to lead author Dr. Hannelore Ehrenreich, from the Max Planck Institute of Experiment Medicine in Gottingen, Germany, and her colleagues in the German Multicenter EPO Stroke Trial Group.
Theoretically, the researchers point out, EPO would have neuroprotective properties, since it binds to neuronal EPO receptors and “acts in an antiapoptotic, antioxidant, anti-inflammatory, neurotrophic, neural stem cell-modulating and neuroplasticity-enhancing fashion.”
Given the increasing use of recombinant tissue plasminogen activator (rtPA) in Germany, the investigators revisited the efficacy and safety of EPO in what they describe as today’s changed “stroke landscape” in their country.
Their intent-to-treat population comprised 522 patients with ischemic stroke in the middle cerebral artery territory, who were enrolled in the study between 2003 and 2008 and randomized to IV infusion of EPO 40,000 IU within 6 hours of symptom onset, or to placebo. Treatments were repeated 24 and 48 hours later.
The primary outcome measure, Barthel Index on Day 90, did not differ between the EPO and placebo groups, nor did modified Rankin Scale scores and NIH Stroke Scale (NIHSS) scores on day 90. Lesion sizes, as measured by magnetic resonance imaging, did not change between baseline and day 7 in either group.
“Unexpectedly, a very high number of patients” – 63.4% — “received rtPA,” the researchers report.
Sixty-six patients died, including 16.4% of the EPO group and 9.0% of the placebo group (OR 1.98, p = 0.01). The significantly increased death rate on EPO was observed in rtPA-treated patients as well, on intent-to-treat and per protocol analyses. In patients not treated with rtPA, there was a trend toward a higher death rate with EPO treatment in the intent-to-treat analysis but not in the per protocol population.
EPO was also associated with increased risk of serious complications, including intracerebral hemorrhage, brain edema, and thromboembolism.
But the researchers also performed a post hoc analysis that raises the question “as to whether the higher number of deaths in the EPO arm may be related to some potential interaction with rtPA.”
As they point out, only the per protocol patients who did not receive rtPA represent an appropriate comparison group for the per protocol patients in the earlier study from the pre-rtPA era, the results of which they were now trying to reproduce.
In fact, when the rtPA patients from the current trial are excluded, “improvement from individual baseline to day 30 (predefined endpoint of the first EPO stroke study) looks identical,” the authors say.
The investigators conclude that “rtPA-treated patients should be excluded from acute poststroke EPO applications.”
In the future, they add, “Careful risk-benefit considerations will have to precede any such trials.”