By Kimberly Pargeon, MD

Assistant Professor of Clinical Neurology, Weill Cornell Medical College

Dr. Pargeon reports no financial relationships relevant to this study.

SYNOPSIS: Using positive likelihood ratios, an adult and child with epileptiform discharges on electroencephalography were estimated to have a 77% and 66% probability, respectively, of recurrent seizures.

SOURCE: Bouma HK, Labos C, Gore GC, et al. The diagnostic accuracy of routine electroencephalography after a first unprovoked seizure. Eur J Neurol 2016;23:455-463.

Prior to 2014, the International League Against Epilepsy (ILAE) defined epilepsy as two or more unprovoked seizures separated by more than 24 hours; however, in 2014, the ILAE revised the definition and acknowledged circumstances in which a single unprovoked seizure may be diagnosed as epilepsy if the risk of seizure recurrence is 60% within the next 10 years.1 One way to assess risk is the routine electroencephalography (EEG), which per the last practice parameter, published in 2007, stated that epileptiform abnormalities after a first unprovoked seizure occur in about 23% of patients and are predictive of future recurrence.2 Bouma et al updated previous reviews and meta-analyses of the predictive accuracy of routine EEG after a first unprovoked seizure using more rigorous guidelines for conducting and reporting reviews published in 2007 and 2009.

The authors systematically searched for prospective or retrospective cohort studies featuring patients of any age, presenting with a first unprovoked seizure who underwent a routine EEG, defined as lasting up to 60 minutes, and were followed for recurrence for a minimum of one year. A “positive” test was defined as the presence of epileptiform discharges. Other information recorded included seizure type, epilepsy etiology, any treatments with anti-epileptic drugs, timing of EEG relative to the seizure occurrence, and activation procedures. Two authors independently screened all titles and abstracts identified by the initial search (n = 3,096), of which 180 full texts were reviewed, yielding 15 studies.

Despite their systematic process for study selection, there was some heterogeneity between the 15 selected studies, which included 1,799 total participants, with some characteristics variably reported. For instance, in nine studies participants were primarily pediatric, with the remainder mostly a mixture of adult and pediatric populations. In seven studies, the duration of the EEG was not defined (described as “routine”). Also, in seven studies the timing of the EEG relative to the first seizure was not defined, and for the remaining eight studies there was a range from < 48 hours to several months.

The reported pooled risk of seizure recurrence after a first unprovoked seizure was 44.2%, with the overall pooled sensitivity and specificity for routine EEG of 44.5% and 79.6%, respectively. The pooled sensitivity for routine EEG for adults as compared to children, however, was significantly lower at 17.3% vs. 57.8%, whereas the pooled specificity trended higher in adults (94.7% vs. 69.6%). Positive and negative likelihood ratios were calculated for adults and children, which were used with Fagan nomograms (assuming a pretest probability of 50%) to estimate post-test probabilities of epilepsy given a “positive” or “negative” test. They estimated an adult presenting with a first unprovoked seizure has a 77% post-test probability of recurrence if the routine EEG is positive and 47% if it is negative, whereas a child has a 66% post-test probability of recurrence with a positive test and 38% probability with a negative test. They were unable to conduct any other subgroup analyses given the heterogeneity of the data.


The 2014 revised ILAE definition of epilepsy, allowing for potential diagnosis after one unprovoked seizure in circumstances where there is at least a 60% chance of recurrence, emphasizes the need to know the diagnostic accuracy of tests, especially EEG. This was the primary goal of this study; however, as with all meta-analyses, it is sometimes difficult to answer specific questions when the source articles try answering others or are missing key pieces of information. For instance, a few studies included only “genetic” forms of epilepsy, whereas the majority were inclusive of those with either unknown causes or any etiology, which may affect the generalizability of these results. Additionally, for many of the studies, abnormal EEG was one of only many factors considered for future seizure recurrence, including presence of structural abnormalities, timing of first seizure, treatment with automated external defibrillators, and seizure type. Within a few individual studies, abnormal EEG was not a statistically significant variable related to recurrence. Further, while a majority of the studies reported either focal or generalized epileptiform discharges, several simply reported “abnormal” EEGs. Finally, the timing of EEG relative to the first seizure either was not reported or was highly variable across the included studies, but early EEG may be more likely to capture epileptiform discharges than a delayed study.3


  1. Fisher RS, Acevedo C, Arzimanoglou A, et al. ILAE official report: A practical clinical definition of epilepsy. Epilepsia 2014;55:475-482.
  2. Krumholz A, Wiebe S, Gronseth G, et al. Practice parameter: Evaluating an apparent unprovoked first seizure in adults (an evidence-based review), report of the Quality Standards Subcommittee of the American Academy of Neurology and the American Epilepsy Society. Neurology 2007;69:1996-2007.
  3. King MA, Newton MR, Jackson GD, et al. Epileptology of the first-seizure presentation: A clinical, electroencephalographic, and magnetic resonance imaging study of 300 consecutive patients. Lancet 1998;352:1007-1011.