To assess the effects of neonatal seizures on the hypothalamus and to test clinical use of prolactin as a neonatal seizure marker, we studied postictal and recovery baseline serum prolactin levels in 19 neonates whose seizures were classified according to their clinical and EEG features. Postictal prolactin levels were obtained 30 min after the seizure, and recovery levels were ascertained 2-4 days later. The ratio of postictal prolactin level to recovery baseline level (prolactin ratio) was used as an indicator of postictal prolactin increase. The specificity and sensitivity of a prolactin ratio of > 2 was compared with the current standard of diagnosis (seizure discharges recorded by ictal EEG). Infants with electroclinical seizures had significantly higher prolactin ratios than control infants or infants with seizures without EEG correlation. Marked prolactin increases were noted only in infants with focal tonic seizures and temporal electrode involvement. A prolactin ratio of > 2 had a specificity of 100% and a sensitivity of 40%. We conclude that neonatal seizures have variable effects on the hypothalamus and that the low sensitivity and the need to await recovery levels limit the clinical value of prolactin ratio as a neonatal seizure marker.