EPO STROKE STUDIES
Responsible
Prof. Dr. Dr. Hannelore Ehrenreich, MD, DVM
Division Clinical Neuroscience
Max Planck Institute of Experimental Medicine
  Hermann-Rein-Str. 3
D-37075 Göttingen
Tel: +49-551-3899-628
Fax: +49-551-3899-670
Email: ehrenreich@em.mpg.de
STROKE IS AN EMERGENCY
Stroke, caused by sudden breakdown of perfusion in a certain brain area, constitutes, together with myocardial infarction (heart attack) one of the most frequent causes of serious disability or death in industrialized countries. Prominent risk factors for stroke apart from advanced age are high blood pressure, diabetes, smoking, high blood levels of cholesterol, obesity, contraceptive medication and excessive consumption of alcohol. But a stroke can also occur without any immediately apparent risk factors and even amongst young people. The most frequent signs of a stroke are a sudden paresis of an arm or a leg or even of a whole side of the body including the face. Often an additional speech disorder (aphasia or dysarthria) accompanies the paresis. Occasionally, the aphasia can appear as an isolated phenomenon. The speech disorder can render the affected person unable to talk at all or induce the production of apparently senseless words. In many cases this speech disorder is misinterpreted as a state of disorientation. If one or more of these most important symptoms of a stroke occur, the patient should immediately be brought to the hospital by ambulance. Even though the therapeutic possibilities are still limited, the supportive measures taken in the hospital, preferentially an institution with a special stroke unit will help to improve the outcome of stroke victims.
Therapeutic possibilities for stroke are limited -
EPO as an option
The introduction of thrombolysis therapy using recombinant tissue plasminogen activator (rtPA) in acute ischemic stroke has improved clinical outcome of patients who present early after symptom onset. With a narrow time-to-treatment window and multiple contraindications, however, thrombolytic therapy applies to a very restricted number of patients, leaving >90% without specific treatment. Therefore, there is an urgent need for new, safer therapies, accessible for more patients, in particular those non-qualifying for rtPA.

Apart from thrombolytic strategies, many agents targeting other aspects of stroke pathology were effective in animal models but unsuccessfully translated to man.1 Recombinant human erythropoietin (EPO) may be an exception. In fact, EPO represents an ideal compound for neuroprotection in brain disease, acting in an anti-apoptotic, anti-oxidant, anti-inflammatory, neurotrophic, neural stem cell modulating and neuroplasticity enhancing fashion. Independent of its hematopoietic effects, EPO proved neuroprotective/neuroregenerative in animal models of ischemia, hypoxia or combinations thereof.2-4 Despite a molecular weight of >30000 Dalton, EPO crosses the blood-brain-barrier to exert neuroprotection.5-7 In a proof-of-concept study, the "Göttingen EPO Stroke Study", EPO was safe and beneficial on clinical outcome in patients with ischemic stroke of the middle cerebral artery (MCA) territory.8
The Göttingen EPO Stroke Study (1997-2001)
Figure 2: Recruitment reminder for the Göttingen EPO Stroke Study (1997-2001; left side). Course of NIHSS throughout the complete study observation period of 30 days post-stroke (right side). EPO-treated patients showed significantly better neurological recovery compared to placebo-treated patients. NIHSS: National Institutes of Health Stroke Scale.
These promising results of the proof-of-concept study8 encouraged a larger trial on EPO for the treatment of acute ischemic stroke.9 However, since conclusion of the Göttingen EPO Stroke Study in 2001, the stroke landscape has changed due to approval and increasing use of rtPA in Germany.10 We present in the following a short synopsis of the results of the double-blind, placebo-controlled, randomized German Multicenter EPO Stroke Trial, designed as phase II/III trial to further investigate efficacy and safety of EPO with/without systemic thrombolysis in patients with ischemic stroke in the MCA territory (ClinicalTrials.gov Identifier: NCT00604630).9
German Multicenter EPO Stroke Trial (2003-2008)
PATIENTS AND METHODS
Multicenter trial
Figure 3: Map with all 11 participating study centers of the German Multicenter EPO Stroke Trial (2003-2008) and recruitment information.
Inclusion criteria / exclusion criteria
Figure 4: Inclusion and exclusion criteria of the German Multicenter EPO Stroke Trial. MCA: middle cerebral artery; cMRI: cranial magnetic resonance imaging; rtPA: recombinant tissue plasminogen activator.
Study intervention and clinical assessment
Figure 5: Study design of the double-blind, placebo-controlled, randomized German Multicenter EPO Stroke Trial. cCT: cranial computed tomography; cMRI: cranial magnetic resonance imaging; NIHSS: National Institutes of Health Stroke Scale; ICU: intensive care unit.
Patient flow and baseline characteristics
Figure 6: Overview of patient flow. In red, the ultimate target population of per-protocol (PP) non-rtPA treated patients is highlighted (comparable to the patients of the Göttingen EPO Stroke Study). rtPA: recombinant tissue plasminogen activator.
Upon inclusion, characteristics of all included patients (intent-to-treat population; ITT) seem to be well balanced between EPO and placebo groups regarding a selection of typical stroke variables (Table 1).
RESULTS
Clinical outcomes
For more detailed information, the reader is kindly referred to our publications in Stroke.9, 11
(1) Intent-to-treat analysis of major functional outcome measures:
      Barthel Index (primary outcome) and modified Rankin Scale:
      The German Multicenter EPO Stroke Trial is a formally negative study.
Figure 7: Neither primary outcome Barthel Index on day 90 nor modified Rankin Scale on day 90 (secondary outcome) revealed significant treatment effects (Chi2-test). Distribution of scores and proportion of patients (%) in the respective score range are presented. According to the ITT analysis of primary outcome Barthel Index on Day 90, the German Multicenter EPO Stroke Trial has to be seen officially as a negative study.
(2) Exploratory per-protocol subgroup analyses (post hoc):
a) Historical comparison of neurological recovery in the Göttingen EPO Stroke Study and the German Multicenter EPO Stroke Trial:
    Beneficial effect of EPO on neurological outcome reproduced
Figure 8: The historical comparison of the Göttingen EPO Stroke Study (all patients treated per protocol, no rtPA treatment yet available at this time-point) and the German Multicenter EPO Stroke Trial (subgroup of PP non-rtPA patients as only valid comparator for the Göttingen Study) showed that results on neurological recovery could be reproduced. Mean±SEM of NIHSS scores, expressed as % individual baseline over time, shown. Statistical strategy: repeated-measures ANOVA. rtPA: recombinant tissue plasminogen activator; NIHSS: National Institutes of Health Stroke Scale.
b) Comparison of neurological recovery in patient groups receiving EPO versus rtPA versus placebo mono-treatment:
    EPO may be an alternative for patients non-qualifying for rtPA.
Figure 9: EPO as well as rtPA mono-treatment led to similar beneficial effects on neurological recovery and were superior to placebo treatment. Estimated marginal mean±SEM of NIHSS scores, expressed as % individual baseline over time, shown. Statistical strategy: repeated-measures ANCOVA (age as covariate due to significant age differences between rtPA (66.7±13.6 years) and EPO (71.1±11.5 years; P=0.02) as well as between rtPA (66.7±13.6 years) and placebo groups (71.5±11.0 years; P=0.005). rtPA: recombinant tissue plasminogen activator; NIHSS: National Institutes of Health Stroke Scale.
Safety analysis
For more detailed information, the reader is kindly referred to our publications in Stroke.9, 11

Safety analysis of all included patients (intent-to-treat population; ITT) yielded an overall higher death rate in EPO as compared to placebo patients (16.4% versus 9.0%; odds ratio [OR] = 1.98, 95% confidence interval = 1.16-3.38; p=0.01; Table 2). This increased mortality was obvious in patients pre-treated with thrombolysis in both ITT as well as per-protocol (PP) population. The highest death rate occurred within the first week (29/66=43.9%), mainly attributable to intracerebral hemorrhage (13/66=19.7%), brain edema (10/66=15.2%) and thrombembolic events (10/66=15.2%). In the PP non-rtPA subpopulation, no safety risk was observed (Table 2, green labeled part). Remarkable is the low death rate of placebo-treated patients (as compared to comparable stroke populations of previous prominent stroke trials), pointing to an "under-mortality" in this group.
Possible reasons for the observed increased mortality:
(1) Unfavourable distribution of prognostically relevant characteristics upon inclusion
Subpopulation analysis of all patients who died revealed that several relevant baseline characteristics (i.e. data obtained before application of any study medication) were significantly different between groups, always in disadvantage of EPO-treated patients (Table 3, red labeled parts). This pattern was absent in surviving patients (no differences between EPO and placebo patients).
Most importantly, upon inclusion (before any study drug application), ITT non-rtPA EPO patients who died suffered from much more severe strokes as compared to placebo patients (NIHSS day 1: 20.4±5.4 versus 13.3±4.9; p=0.003). This highly significant prediction of a worse outcome explains the twofold higher very early death rate in the EPO group.
(2) Pretreatment with thrombolysis despite contraindications
Detailed additional safety analysis also uncovered that around 50% of the ITT rtPA-treated patients in this trial had contraindications to rtPA. Among the rtPA-treated patients of the ITT population with fatal outcome, 44.4% in the EPO but only 28.6% in the placebo group had one rtPA contraindication (p=0.05; Table 3). Figure 10 illustrates that EPO treatment on top of lege artis stroke treatment (i.e. no rtPA treatment in the presence of contraindications or rtPA treatment in the absence of contraindications) is free of any safety risk and even tends to be beneficial in patients without thrombolysis. Negative effects of EPO are restricted to patients who receive rtPA despite contraindications as demonstrated by their worse neurological recovery as compared to placebo.
Figure 10: ITT patients with lege artis stroke treatment show either slightly superior (non-rtPA patients; upper row) or similar neurological recovery post-stroke (rtPA patients without contraindications; upper row). Negative effects of EPO (i.e. worse neurological recovery) are restricted to patients who receive rtPA despite contraindications (lower row). Mean±SEM of NIHSS scores, expressed as % individual baseline over time, shown. ITT: Intent-to-treat population; rtPA: recombinant tissue plasminogen activator; NIHSS: National Institutes of Health Stroke Scale.
Figure 11: Additional subgroup analysis: ITT patients who received rtPA despite one contraindication (left lower panel of Figure 10) are further subdivided by type of contraindication. Owing to the small numbers, all conclusions can only be preliminary: Among the patients with one rtPA contraindication, negative effects of EPO (i.e. worse neurological recovery) are apparently restricted to patients on anticoagulants and patients with very large infarct size upon inclusion. Neither time-to-rtPA treatment nor old age seems to play a role. Mean±SEM of NIHSS scores, expressed as % individual baseline over time, shown. ITT: Intent-to-treat population; rtPA: recombinant tissue plasminogen activator; NIHSS: National Institutes of Health Stroke Scale.
Taken together, speculations on the increased mortality in the EPO arm comprise (1) a negative interaction of rtPA and EPO; (2) an unfavourable imbalance of relevant stroke characteristics, especially stroke severity upon inclusion in the dead patients; (3) an outcome by chance, considering the obvious under-mortality in the placebo group; (4) a negative influence of the escalating violations of thrombolysis contraindications which amounted to 50% of the rtPA treatments. These assumptions are supported by the lack of safety concerns in the pilot study and in two interim looks of the present trial.
Further subgroup analysis: Circulating damage marker profiles support a neuroprotective effect of EPO in ischemic stroke patients
The German Multicenter EPO Stroke Trial, investigating safety and efficacy of EPO treatment in ischemic stroke, had formally to be declared a negative study. Exploratory subgroup analysis, however, revealed that patients not receiving thrombolysis most likely benefited from EPO regarding clinical recovery - a result reproducing findings of the Göttingen EPO Stroke Study. We therefore investigated whether the positive signal on clinical outcome in this patient subgroup is mirrored by respective post-stroke biomarker profiles. All patients of the German Multicenter EPO Stroke Trial non-qualifying for thrombolysis were included, if they (I) were treated per protocol and (II) had at least 2 out of 5 follow-up blood samples for circulating damage markers drawn (n=163). The glial markers S100B and GFAP and the neuronal marker UCH-L1 were measured by ELISA in serum of days 1, 2, 3, 4, and 7 post-stroke. All biomarkers increased post-stroke. Overall, EPO treated patients had significantly lower concentrations (area under the curve) over 7 days of observation as reflected by the composite score of all 3 markers (Cronbach's a=.811) and by UCH-L1. S100B and GFAP showed a similar tendency. To conclude, serum biomarker profiles, as outcome measure of brain damage, corroborate an advantageous effect of EPO in ischemic stroke. In particular, reduction in the neuronal damage marker UCH-L1 may reflect neuroprotection by EPO.
Figure 12:Biomarkers substantiate the positive signal on clinical outcome of EPO compared to placebo patients. (A) Non-rtPA treated EPO patients show improved clinical outcome (NIHSS) after stroke as compared to the placebo group. (B) AUC mean±SEM values and (C) AUC z-standardized values demonstrate lower levels of circulating damage markers post-stroke upon EPO as compared to placebo. For more details see 12.
SUMMARY
In the German Multicenter EPO Stroke Trial (phase II/III), the results of the Göttingen EPO Stroke Study (phase IIb) were essentially reproduced: EPO is safe and beneficial with respect to neurological recovery in patients non-qualifying for rtPA. This conclusion is further supported by circulating biomarker profiles.

In rtPA treated patients, EPO did not improve outcome but rather increased the risk of serious complications: Death, bleeding, edema, and thrombembolic events. This result precludes the use of EPO in rtPA patients. More basic research is required to understand potentially detrimental mechanisms of interaction between rtPA and EPO.
References
1. O'Collins VE, Macleod MR, Donnan GA, Horky LL, van der Worp BH, Howells DW. 1,026 experimental treatments in acute stroke. Ann Neurol. 2006;59:467-477
2. Brines M, Cerami A. Emerging biological roles for erythropoietin in the nervous system. Nat Rev Neurosci 2005;6:484-494
3. Jelkmann W. Effects of erythropoietin on brain function. Curr Pharm Biotechnol. 2005;6:65-79
4. Juul S. Recombinant erythropoietin as a neuroprotective treatment: In vitro and in vivo models. Clin Perinatol. 2004;31:129-142
5. Brines ML, Ghezzi P, Keenan S, Agnello D, de Lanerolle NC, Cerami C, Itri LM, Cerami A. Erythropoietin crosses the blood-brain barrier to protect against experimental brain injury. Proc Natl Acad Sci U S A. 2000;97:10526-10531
6. Ehrenreich H, Degner D, Meller J, Brines M, Béhé M, Hasselblatt M, Woldt H, Falkai P, Knerlich F, Jacob S, Von Ahsen N, Maier W, Brück W, Rüther E, Cerami A, Becker W, Sirén A-L. Erythropoietin: A candidate compound for neuroprotection in schizophrenia. Molecular Psychiatry. 2004;9:42-54
7. Banks WA, Jumbe NL, Farrell CL, Niehoff ML, Heatherington AC. Passage of erythropoietic agents across the blood-brain barrier: A comparison of human and murine erythropoietin and the analog darbepoetin alfa. Eur J Pharmacol. 2004;505:93-101
8. Ehrenreich H, Hasselblatt M, Dembowski C, Cepek L, Lewczuk P, Stiefel M, Rustenbeck H-H, Breiter N, Jacob S, Knerlich F, Bohn M, Poser W, Rüther E, Kochen M, Gefeller O, Gleiter C, Wessel TC, De Ryck M, Itri L, Prange H, Cerami A, Brines M, Sirén A-L. Erythropoietin therapy for acute stroke is both safe and beneficial. Molecular Medicine. 2002;8:495-505
9. Ehrenreich H, Weissenborn K, Prange H, Schneider D, Weimar C, Wartenberg K, Schellinger PD, Bohn M, Becker H, Wegrzyn M, Jähnig P, Herrmann M, Knauth M, Bähr M, Heide W, Wagner A, Schwab S, Reichmann H, Schwendemann G, Dengler R, Kastrup A, Bartels C. Recombinant human erythropoietin in the treatment of acute ischemic stroke. Stroke. 2009;40:e647-656
10. Weimar C, Kraywinkel K, Maschke M, Diener HC. Intravenous thrombolysis in german stroke units before and after regulatory approval of recombinant tissue plasminogen activator. Cerebrovasc Dis. 2006;22:429-431
11. Ehrenreich H, Bartels C, Weissenborn K, Diener HC. NOT THE END OF THE EPO STORY … A safety risk of recombinant human erythropoietin (EPO) in ischemic stroke is restricted to patients receiving thrombolysis despite contraindications. Stroke. 2010:In press
12. Ehrenreich H, Kästner A, Weissenborn K, Streeter J, Sperling S, Wang KK, Worthmann H, Hayes RL, Von Ahsen N, Kastrup A, Jeromin A, Herrmann M. Circulating damage marker profiles support a neuroprotective effect of erythropoietin in ischemic stroke patients. Mol Med. 2011, Sep 2.
13. Worthmann H, Martens-Lobenhoffer J, Joumaah M, Li N, Lichtinghagen R, Hecker H, Kielstein JT, Ehrenreich H, Bode-Böger SM, Weissenborn K. Asymmetric dimethylarginine in response to recombinant tissue-type plasminogen activator and erythropoietin in acute stroke. Stroke. 2013;44:2128-33
Appendix
THE EPO STROKE TRIAL GROUP
Coordinating Investigator: H Ehrenreich
Coordinating study center: Division of Clinical Neuroscience, Max Planck Institute of Experimental Medicine, Göttingen.
Co-Coordinating Investigator: C Bartels
Study personnel: C Aust, K Hannke, C Norra, S Sperling, N Stender, W Timner, M Wegrzyn.


The following centers participated in the German Multicenter EPO Stroke Trial:
Hannover1, Göttingen2, Bremen3, Celle4, Erlangen5, Leipzig6, Dresden7, Essen8, Braunschweig9, Berlin10, Aachen11.

Local Principal Investigators (local PIs): K Weissenborn / R Dengler1, H Prange / M Bähr / A Kastrup2, M Herrmann / G Schwendemann3, W Heide4, PD Schellinger / S Schwab5, D Schneider / A Wagner6, K Wartenberg / H Reichmann7, C Weimar8, K Wessel9, P Marx10, J Noth11.

Enrolling investigators (in alphabetical order): B Ahl1, U Becker7, M Bigalke3, M Ebke3, M Feldmann3, G Gahn7, A Goldbecker1, K Gröschel2, Z Gumienny4, J Hanssen4, C Hobohm6, U Johansson8, M Köhrmann5, D Michalski6, G Moldrich2, T Nowe5, J Schaumberg3, S Schnaudigel2, S Schwarting2, B Sommer3, J Stewen4, J Thomsen3, A Tountopoulou1, A Tryc1, H Worthmann1, G Wortmann3. Neuroradiologists: H Becker1, M Bester2, A Dörfler5, F Donnerstag1, M Forsting8, S Jacob2, M Knauth2, R von Kummer7, H Lanfermann1, T Mitrovics3, HJ Roth6, M Schlamann8, M Stiefel2, U Thiemann3, BF Tomandl3. Study pharmacists: M Bohn (head of the department), K Linke, J Müller. Members of the scientific data and safety monitoring board: HC Diener, Essen, O Gefeller, Erlangen, D Neubert, Berlin, P Rieckmann, Vancouver, Canada. Consultant: H Becker, Applied Science and Technology, Zwingenberg. Clinical Research Organization: P Jähnig, PAREXEL International GmbH, Berlin.