guideline on treating first seizure

The Am Acad of Neurology and Am Epilepsy Society just released an evidence-based guideline on the management of an unprovoked first seizure in adults (see seizure rx first seizure neurol 2015 in dropbox, or Neurology. 84;16: 1705-1713).

background: in the US, there are 150K adults annually who have a first unprovoked seizure, defined as a seizure (sz) of unknown etiology, or a seizure in relation to  a preexisting brain lesion or progressive CNS disorder.  they exclude seizures from acute symptomatic conditions (eg metabolic, traumatic, stroke).

general comments and guidelines:

--what is risk of recurrent seizures? studies were mostly done on tonic-clonic seizures, no randomization regarding using AEDs (anti-epileptic drug). but overall greatest risk of recurrent sz is in first year (approx 32%), which increases to 36% by 2 years, 40% by 3 years and then to 48% if >5years.  

--does immediate treatment with AEDs change the short-term (2 year) prognosis of sz recurrence? pretty clear that the risk is decreased in the first 2 years if give immediate AED, with absolute risk reduction of 20% , comparing immediate (recurrence rate of 23%) vs delayed AED (recurrence rate of 43%), but use of AEDs might not affect quality of life (only 1 study assessed this, not a great study, finding no difference)

--does immediate AED treatment influence prognosis (eg potential sz remission over the longer-term, >3 years), usually measured as being sz-free for 2-5 years? no (or, at least, unlikely)

--what adverse events (AEs) are there with AEDs and how common? data mostly available for older AEDs (phenytoin phenobarbital, carbamazepine, valproic acid, lamotrigine). range 7-31% (most studies in the lower range, around 7-15%). mostly mild, dose-dependent, and reversible. newer AEDs have fewer (though different) AEs.

so, implications:

--if someone presents with history of unprovoked sz several years ago, they are at significantly lower risk of a recurrent sz and this should play into their risk/benefit analysis. per this guideline: "immediate AED treatment at the time of the first unprovoked seizure is not well accepted and is debated".

--it seems to me that the use of AEDs with a first unprovoked sz should be decided collaboratively (a person working in construction of skyscrapers or someone severely psychologically traumatized by the sz may opt for more aggressive therapy.)

--there are some predictors to help with the decision analysis: greater sz risk if prior brain insult (relative recurrence rate at 1-5 years of 2.55, vs unprovoked first sz) or lesion associated with the sz, EEG has epileptiform abnormalities (rel recurrence rate at 1-5 years of 2.16), significant abnormality on brain imaging (rel recurrence risk at 1-4 years of 2.44), presence of nocturnal sz (rel recurrence risk at 1-4 yrs of 2.1, vs when patient is awake). and no consistent increased incidence associated with family history, sz type, age, sex, presentation with status epilepticus or multiple discrete sz in first 24 hours.

--those with 2nd or more sz are at much higher risk of further recurrences (57% by one year, 73% by 4 years)​

--lots of ripe areas for future research: better predictive models for risk of recurrence/risk stratification, newer data on AEs of the new drugs, how long should the AEDs be continued/risks of stopping AEDs (esp in the context of someone with only one sz. there are some data on those with various epilepsy or sz types who become sz-free and stopped their AEDs)