Hypertension (HTN) is a common complication of recombinant erythropoietin (EPO) therapy, but the mechanism of the EPO-associated HTN is uncertain. In the present study we examined the effects of EPO and the vehicle alone on rat caudal artery contractile response and basal and thrombin-stimulated platelet cytosolic Ca2+ concentration ([Ca2+]i) in vitro and on blood pressure (BP) and heart rate in vivo. At high concentrations (200 U/ml) EPO caused a small but consistent contraction in the caudal artery rings (P < 0.01) without affecting the response to either angiotensin II (ANG II) or the alpha 1-agonist methoxamine. Incubation with EPO significantly increased basal platelet [Ca2+]i (P < 0.01) and augmented the thrombin-induced rise of [Ca2+]i in Ca(2+)-free medium (P < 0.05). Long-term EPO administration led to a significant elevation of BP within 2 wk regardless of whether the hematocrit was allowed to rise or was kept constant by dietary iron deficiency. In contrast, single intravenous administration of high-dose EPO (400 and 5,000 U/kg), estimated to yield plasma concentrations comparable with those employed in vitro, failed to either alter BP or modify the BP response to ANG II during a 60-min observation period. This was associated with a significant rise in plasma guanosine 3',5'-cyclic monophosphate but no discernible change in plasma atrial natriuretic peptide, suggesting enhanced nitric oxide (NO) release. Thus, at high concentrations, EPO appears to possess a fast-acting pressor effect in vitro but not in vivo. The observed discrepancy may be due to enhanced NO release with EPO administration in vivo. However, HTN does occur with repeated EPO administration in a time-dependent and hematocrit-independent manner. This suggest that expression of the hypertensive effect of EPO in vivo involves a gradual conditioning process.