The degree of practice variation in continuance of phenobarbital treatment despite resolution of neonatal seizures was evaluated by national survey conducted at the University of Rochester Medical Center, New York. Surveys mailed to 609 randomly selected child neurologists and 579 neonatologists were completed by 20.7% and 23.1%, respectively. Practices varied widely, with little difference in response frequencies between child neurologists and neonatologists. For child neurologists, prophylactic phenobarbital was always used in 5%, sometimes used in 72%, rarely in 19%, and never in 3%. Responses of neonatologists were similar. Duration of treatment was <1 month in 8%, 1-3 months in 45%, 3-6 months in 37%, none longer than 6 months. Drug levels were monitored routinely by 34%, and only when indicated by 57%. Physicians were more likely to respond yes to continuation of treatment for a given scenario than would be predicted by their overall responses to questions. Since the survey of practices 15 years ago, physicians are reporting less frequent and shorter phenobarbital treatment after resolution of neonatal seizures. 
COMMENT. The relatively low response to this survey and surveys in general is explained by the length and complexity of questions, and the increasing number of similar requests. Possible late cognitive effects of long-term phenobarbital in the infant are one reason to limit duration of prophylactic treatment. A randomized trial is needed to determine benefits and adverse effects of continued therapy after discharge from the NICU.