Recombinant human erythropoietin (rEpo) is neuroprotective in neonatal models of hypoxic-ischemic brain injury. However, the optimal rEpo dose, dosing interval, and number of doses for reducing brain injury are still undetermined. We compared the neuroprotective efficacy of several subcutaneous rEpo treatment regimens. Seven-day-old rats underwent unilateral carotid ligation plus 90 min 8% hypoxia. Treatment began immediately after injury. Treatment regimens examined included 1, 3, or 7 daily subcutaneous injections of either 0 (vehicle), 2,500, 5,000, or 30,000 U/kg rEpo. Gross brain injury, neuronal apoptosis (TUNEL), and gliosis (glial fibrillary acidic protein) were assessed at 48 h or 1 wk post injury. Immunoreactive cells and brain injury were quantified for statistical comparison to vehicle controls. rEpo treatment reduced brain injury, apoptosis, and gliosis, in a dose-dependent U-shaped manner at both 48 h and 1 wk. Neither one injection of 2,500, seven injections of 5,000, or three injections of 30,000 U/kg rEpo were protective. Three doses of 5,000 and one dose of 30,000 U/kg rEpo were most protective at both time intervals. rEpo provides dose-dependent neuroprotection. Of the regimens tested, three doses of 5,000 U/kg was optimal because it provided maximal benefit with limited total exposure. (Pediatr Res 61: [451][452][453][454][455] 2007) N eonatal brain injury continues to affect the lives of thousands of infants in the United States and worldwide. In fact, perinatal asphyxia accounts for 23% of deaths globally (1). Hypothermia following hypoxia-ischemia offers promise of improved neurodevelopmental outcomes, however, thus far, the therapeutic benefit has been modest (2). When strict entry criteria are defined, up to 60% of untreated infants exposed to moderate to severe perinatal hypoxia-ischemia die or have significant neurodevelopmental sequelae (3,4). Erythropoietin (Epo), a hematopoietic cytokine known best for its erythropoietic effects, has neuroprotective effects in many animal models of brain injury, including hypoxia-ischemia, trauma, and excitotoxic injury (5-12). Although many neuroprotective strategies appeared promising in animal models, most have failed clinically (often due to toxicity). In contrast, rEpo is a safe treatment for anemia (13) and, more importantly, the neuroprotective effects of rEpo have translated well from bench to clinical trials for both amyotrophic lateral sclerosis (ALS) (14) and middle cerebral artery stroke (15).Initial studies provided no evidence that the 37 kD charged glycoprotein rEpo could cross the blood-brain barrier (BBB) (16 -18). Therefore, early rEpo neuroprotection experiments used either in vitro application (19,20) or direct intracerebral injection (21). More recently, robust neuroprotection was produced after systemic injection of high-dose rEpo (5,000 U/kg i.p.) (5). Although the neuroprotective rEpo dose is clearly above the anemia treatment range, there is no consensus as to the optimal neuroprotective dose, dose number, or in...