The results of studies in databases and references concerning serum prolactin levels (PRL) in patients with suspected seizures were rated for quality and analyzed by members of the Therapeutics Subcommittee of the American Academy of Neurology. Eight prospective, controlled studies showed that an elevated PRL, measured at 10 to 20 minutes after a suspected event, was highly predictive of generalized tonic-clonic (GTCS) or complex partial seizures (CPS), and differentiated these epileptic from psychogenic nonepileptic seizures in adults and older children. Most studies used a PRL of at least twice baseline as abnormal (upper limits of normal in most laboratories was 18 to 23 ng/mL). Sensitivity was 60% for GTCS, and 46% for CPS, while specificity was 96% for both seizure types. Two studies showed elevated PRL levels after tilt-test-induced syncope, and PRL did not distinguish epileptic seizures from syncope. PRL data obtained after simple partial seizures, status epilepticus, repetitive seizures, and neonatal seizures were inconclusive. 
COMMENT. An elevated serum PRL is specific for GTCS and CPS, and may be used to differentiate these epileptic from psychogenic nonepileptic seizures. A positive test may be useful as a substitute for video-EEG monitoring when this is unavailable. The test has a low sensitivity, and a negative result cannot be considered diagnostic of a psychogenic seizure. In addition to epileptic seizures, other events, including syncope, pregnancy, hypothyroidism, and various drugs may be associated with elevated PRL. PRL sampling should be obtained within 10 and 20 minutes after a suspected seizure, and a return to a baseline level is reached after an interval of 6 hours. Further studies will be necessary to establish the value of the test in young children and in neonates.