OBJECTIVE: Our purpose was to study the effects of long-term treatment with magnesium sulfate on hypoxic-ischemic brain damage in newborn rats. STUDY DESIGN: Seven-day-old rat pups (n = 120) were exposed to unilateral carotid artery ligation and 2 hours of hypoxia (8% oxygen in 92% nitrogen). Neuronal loss was evaluated in 4 groups. The loading dose group (n = 25) received a bolus injection of 270 mg/kg magnesium sulfate intraperitoneally. The maintenance group (n = 26) received 3 days of continuous infusion of magnesium sulfate with a micro-osmotic pump at a rate of 72 mg/kg per hour. The loading dose-plus-maintenance group (n = 23) received both. The control group (n = 46) did not receive magnesium. Seven days after the injury the pups were killed and the brains were removed for analysis. Severity of neuronal loss was evaluated in the cerebral cortex and hippocampus and compared among groups by chi2 test. RESULTS: Cyst formation resulting from necrosis was significantly decreased in the maintenance group (15%) and loading dose-plus-maintenance group (9%) but not in the loading dose group (52%) relative to the control group (46%). Severity of neuronal loss in the cerebral cortex and the hippocampus was also significantly improved in the maintenance group and the loading dose-plus-maintenance group but not in the loading dose group. Mortality rates were not different among the groups. Magnesium ion concentrations at 24 hours were significantly decreased from the preinfusion value of 0.49 +/- 0.01 mmol/L in the control group (0.34 +/- 0.03 mmol/L) but remained within normal ranges in the maintenance group (0.46 +/- 0.02 mmol/L with a pump infusing 72 mg/kg per hour and 0.56 +/- 0.05 mmol/L with a pump infusing 24 mg/kg hour). There was an inverse relationship between the maintenance dose and neuronal loss but not between the maintenance dose and mortality rate. CONCLUSION: Long-term magnesium administration had neuroprotective effects against hypoxia-ischemia in newborn rats.