What is the negative predictive value of a completely unremarkable EMU admission?

By completely unremarkable, I mean zero clinical events AND zero interictal abnormalities despite full withdrawal of medications. Assume, for the sake of argument, between 5 and 10 days of monitoring.

2 Answers

Mednet Member

Neurology · USF Health

Answered on

Agree... A "negative" EMU evaluation does not, of course, rule out epilepsy.

"Unremarkable" or "inconclusive" means the event in question was not recorded and there were no interictal epileptiform abnormalities. In our own study [Benbadis et al., PMID 15329081], we found that represented 15% of EMU admissions. That has been consistent.

Options that can help are:

1) Using home (ambulatory) EEG-video where patients can be exposed to their usual triggers.

2) Repeating the EMU admission with more aggressive medication reduction.

3) Cellphone videos of the events.

4) Home detecting/diagnostic devices like Brain Sentinel (surface EMG), videos, etc.

Mednet MemberInvited Expert

Neurology · University of Rochester Medical Center

Answered on

There is no answer for this specific question, which I assume is meant to ask how certain we can be that the patient does not have epilepsy.

This paper (Robinson et al., PMID 21507729) reported on predictors of a non-diagnostic LTM stay, and found that absence of epilepsy risk factors and the presence of a prior normal EEG were the strongest predictors of a non-diagnostic EMU admission (ie typical events not captured). A minority of the patients in this retrospective study went on to have a repeat EMU admission (about 11%, which was 43 out of 380 patients). Of these, 18 patients (42%) had typical events recorded during the second EMU admission, and they were equally split between epileptic and non-epileptic events. However, 8 of the 9 patients with epileptic events had "clinically significant abnormalities" on their initial vEEG, whereas all 9 patients with non-epileptic events had a normal initial vEEG. I assume the "clinically significant abnormalities" on EEG referenced in the paper refers to epileptiform abnormalities, but the wording is a little unclear in the paper.

The whole clinical scenario needs to be taken into account when faced with a non-diagnostic EMU admission: the semiology of the events, the age at onset, epilepsy risk factors, prior EEG results, MRI results, etc. Many times, if the events are not captured, even if interictal EEG has been normal, the patient goes back on AEDs, especially if the event description is highly suspicious for an epileptic event. This also depends on patient preference, however.