Noise-induced hearing loss may result from excessive release of glutamate, nitrogen oxide and reactive oxygen species. The effects of these factors on the inner ear may potentially be prevented or reduced by erythropoietin (EPO), as indicated by previously demonstrated neuro-protective effects of EPO upon damage to the central nervous system and the retina. This paper reports three separate trials, conducted to investigate the hypothesis that noise-induced hearing loss is prevented or reduced by erythropoietin. The trials employed three different modes of drug application, different administration time windows and different rodent species. In trial 1, guinea pigs were exposed to 110dB SPL, 4-20kHz wide band noise (WBN) for 8h. EPO was administered to the round window membrane 24h after noise exposure, either sustained by pump for a week or by single dose middle ear instillation. In trial 2, rats were exposed to 105dB SPL, 4-20kHz WBN for 8h. EPO was administered by single dose middle ear instillation 1 or 14h after noise exposure. In trial 3, rats were exposed to 105dB SPL, 4-20kHz WBN for 8 or 3x8h. EPO was injected intraperitoneally 1h before noise exposure. Oto-acoustic emissions and auditory brainstem responses (at 16kHz) were recorded before and after noise exposure in all trials. The noise exposure induced a hearing loss in all animals. In trial 1, no recovery and no improvement of hearing occurred in any treatment group. In trial 2 and 3, a partial hearing recovery was seen. However, the hearing loss of the EPO treated animals was significantly worse than controls in trial 2. In trial 3, the hearing of the EPO treated animals exposed for 3x8h was significantly worse than controls. Thus, surprisingly, the results from 2 of the 3 present trials indicate that erythropoietin may in fact augment noise-induced hearing loss. This is contradictory to the beneficial effect of EPO reported by the vast majority of studies on stressed neural tissues. EPO administration may alter the blood flow dynamics of the cochlear vascular bed during or after noise exposure, by a potential induction of vasoconstriction. This may be the cause of the surprising findings.