Recombinant human erythropoietin (EPO) has been successfully tested as neuroprotectant in brain injury models. The first large clinical trial with stroke patients, however, revealed negative results. Reasons are manifold and may include side-effects such as thrombotic complications or interactions with other medication, EPO concentration, penetration of the blood-brain-barrier and/or route of application. The latter is restricted to systemic application. Here we hypothesize that EPO is neuroprotective in a rat model of acute subdural hemorrhage (ASDH) and that direct cortical application is a feasible route of application in this injury type. The subdural hematoma was surgically evacuated and EPO was applied directly onto the surface of the brain. We injected NaCl, 200, 2000 or 20,000IU EPO per rat i.v. at 15min post-ASDH (400mul autologous venous blood) or NaCl, 0.02, 0.2 or 2IU per rat onto the cortical surface after removal of the subdurally infused blood t at 70min post-ASDH. Arterial blood pressure (MAP), blood chemistry, intracranial pressure (ICP), cerebral blood flow (CBF) and brain tissue oxygen (ptiO2) were assessed during the first hour and lesion volume at 2days after ASDH. EPO 20,000IU/rat (i.v.) elevated ICP significantly. EPO at 200 and 2000IU reduced lesion volume from 38.2+/-0.6mm(3) (NaCl-treated group) to 28.5+/-0.9 and 22.2+/-1.3mm(3) (all p<0.05 vs. NaCl). Cortical application of 0.02IU EPO after ASDH evacuation reduced injury from 36.0+/-5.2 to 11.2+/-2.1mm(3) (p=0.007), whereas 0.2IU had no effect (38.0+/-9.0mm(3)). The highest dose of both application routes (i.v. 20,000IU; cortical 2IU) enlarged the ASDH-induced damage significantly to 46.5+/-1.7 and 67.9+/-10.4mm(3) (all p<0.05 vs. NaCl). In order to test whether Tween-20, a solvent of EPO formulation 'NeoRecomon(R)' was responsible for adverse effects two groups were treated with NaCl or Tween-20 after the evacuation of ASDH, but no difference in lesion volume was detected. In conclusion, EPO is neuroprotective in a model of ASDH in rats and was most efficacious at a very low dose in combination with subdural blood removal. High systemic and topically applied concentrations caused adverse effects on lesion size which were partially due to increased ICP. Thus, patients with traumatic ASDH could be treated with cortically applied EPO but with caution concerning concentration.