Prolactin and Seizures


“Doctor, I’m having a seizure!” says the patient shaking with his eyes open looking at you. Ah yes, the illustrious pseudo-seizure. What if the patient is not talking to you during the event? Is there any way to tell the difference between an epileptic seizure (ES) and a non-epileptic seizure (NES)? A fellow resident recommended I get a STAT prolactin level, (PRL), an idea I’d never heard of, so I briefly looked into it for your reading pleasure.


  • 1978 Trimble published the first paper that linked elevated PRL to generalized tonic-clonic seizures but not NES
  • …fast forward past articles that showed no increase in PRL with NES to more recent articles…
  • 2004 Vukmir wrote an article specific to PRL use in the ED and found that elevated prolactin was identified in 42% of seizures by hospital discharge diagnosis and 17.5% of patients without diagnosis of seizure (42% sensitive and 82% specific, 54% NPV). He basically calls it a reliable confirmatory but poor screening test. Interestingly, 57% of people diagnosed with seizure at hospital discharge had a normal PRL, and more people with an abnormal EEG during their hospitalization had a normal PRL in the ED than an elevated one (but this paper fails to talk about timing of lab draw relative to seizure activity and the primary endpoint is subjective since it’s a discharge diagnosis).
  • 2004 Willert et al looked at PRL, CK, and neuron specific enolase (NSE). To focus on their findings about PRL: they measured PRL at 0, 10, 20, 30, 60 minutes and 6, 12, and 24 hours during continuous EEG monitoring. The Δ PRL was much greater in the ES group than the NES group however both had significant elevations for the first 30 minutes. The sensitivity was greatest (88%) at 10 and 20 minutes but it had generally low specificity and NPV. There was a false positive rate of 42% at 10 minutes and 33% at 20 minutes for the NES group and a 60% false positive rate after vaso-vagal syncope.


  • Uptodate says that PRL should be drawn 10-20 minutes after the event and compared to a baseline level 6 hours later. The relative rise of twice or more from baseline is considered positive. This is not realistic for the ED so some studies talked about “pathologic elevation” from the normal lab range but didn’t exactly spell out what it meant to be “pathologically elevated”
  • PRL can be elevated for a multitude of reasons including syncope, neuroleptic drugs, stress, etc. Some syncopal episodes have seizure-like activity so where do you go from there?
  • The shorter the seizure free interval, the more likely PRL would be elevated

My Take

Does the result of a test change your management or did you just order it to order it? PRL is a test with a lot of false positives but the Vukmir paper also showed the possibility for false negatives with normal PRL and abnormal EEG. Both articles showed a low NPV so you cannot effectively rule out epilepsy with a normal PRL. Furthermore, if its not elevated, you have to decide if you feel comfortable sending the patient home without neurology follow-up / seizure precautions. Ultimately, I am not going to change my practice and order PRL to prove or disprove if someone is having a true seizure because it is not going to change my management.

Oh and by the way, the patient from the case was totally faking it, no PRL necessary.


  • “Serum neuron-specific enolase, prolactin, and creatine kinase after epileptic and psychogenic non-epileptic seizures.” Willert, et al. Acta Neurologica Scandinavica. (2004) 109: 318-323
  • “Does serum prolactin indicate the presence of seizure in the emergency department patient?” Vukmir, Rade. Journal of Neurology. (2004) 251: 736-739

Leave a Reply

Your email address will not be published. Required fields are marked *