Educational scope notice: This is a study note for medical students, not medical advice, diagnosis, or treatment guidance. Clinical management should follow local protocols and current guidelines. Drug dosing below is intentionally omitted; exact regimens follow local protocols.
Treatment serves two goals at once: seizure control and a life the patient accepts living. Antiseizure drugs work by damping neuronal excitability. Sleepiness, mental slowing, and confusion follow directly from that mechanism. Patients who suffer these effects stop their tablets and relapse. The trade-off needs open discussion from the first prescription.
When a first seizure warrants treatment
Epilepsy is defined by two unprovoked seizures more than 24 hours apart, by one seizure with at least 60% recurrence risk over 10 years, or by an epilepsy syndrome. After a single unprovoked seizure the 2-year recurrence risk is 21 to 45%. The strongest predictors are a prior brain insult such as stroke or trauma, epileptiform abnormalities on EEG, a structural imaging abnormality, and a nocturnal seizure. Nocturnal recurrence may stay nocturnal-only. Family history and presentation as status epilepticus were not consistently predictive and are not recurrence predictors.
Immediate treatment likely lowers 2-year recurrence but does not improve long-term remission or quality of life. Starting treatment is therefore an individualized wager of recurrence probability against drug burden. Drug adverse events run 7 to 31% and are mostly mild and reversible. NICE supports considering treatment after a first unprovoked seizure with a neurological deficit, unequivocal epileptiform EEG, a structural imaging abnormality, or a patient-judged unacceptable risk of another seizure. A first suspected seizure needs urgent specialist assessment within 2 weeks with an individualized second-seizure evaluation.
Drug selection by seizure type
The network biology differs, so the first-line drug differs. For focal seizures, NICE first-line monotherapy is lamotrigine or levetiracetam. Carbamazepine and oxcarbazepine are second-line options. For generalized tonic-clonic seizures there is a three-way choice of lamotrigine, levetiracetam, or valproate. For absence seizures, ethosuximide is first line with valproate second line. Carbamazepine, oxcarbazepine, and phenytoin can exacerbate absence or myoclonic seizures including juvenile myoclonic epilepsy.
The mechanisms explain the match. Levetiracetam modulates transmitter release through synaptic vesicle protein SV2A. Carbamazepine and oxcarbazepine block voltage-gated sodium channels. Valproate combines sodium blockade with GABA potentiation. Ethosuximide blocks thalamic T-type calcium channels that generate the 3 Hz absence rhythm. Phenobarbital has largely retreated from paediatric use because it impairs school performance. Phenytoin is limited by liver toxicity and gingival hypertrophy and keeps its place in the status epilepticus protocol.
Valproate is a serious teratogen. Major congenital malformations occur in 6.7 to 10.3% of exposed pregnancies, including neural tube defects, with additional dose-dependent neurodevelopmental harm. Valproate must not be started in anyone under 55 without two specialists independently agreeing that no other effective and tolerated treatment exists. Use in anyone of childbearing potential is restricted to cases where alternatives are unsuitable, under the Pregnancy Prevention Programme with counselling and contraception.
Monitoring, switching, and stopping
Treatment starts with one drug. Monotherapy is used whenever possible. When seizures persist despite apparent adherence, the first steps are confirming compliance and checking plasma levels. Levels are also monitored with unexpected adverse events, in early infancy when toxicity is hard to spot, during polytherapy where interactions shift concentrations unpredictably, with liver or kidney dysfunction, at relapse, and in pregnancy. Numbers never rule alone, especially in polytherapy.
When the first drug fails or harms, therapy changes by overlap, never by abrupt stop. The new drug rises to therapeutic levels before the old one tapers down. Sudden withdrawal can precipitate status epilepticus. Combined drugs should work through different mechanisms with low metabolic and adverse-event interaction. After 2 seizure-free years, NICE requires an individualized recurrence-risk assessment before any stop decision. Tapering typically runs at least 3 months, longer for benzodiazepines and barbiturates, one drug at a time. Epileptiform EEG argues against stopping in children. In adults tapered after 2-year freedom, long-term recurrence may be higher than with continued treatment (15% versus 7% in one class I study). Genetic epilepsies such as juvenile myoclonic epilepsy tend to need lifelong treatment, kept as a syndrome-aware caution rather than an absolute rule.
Evidence anchors
- NICE. Epilepsies in children, young people and adults (NG217): https://www.nice.org.uk/guidance/ng217
- Krumholz A, et al. AAN/AES first unprovoked seizure guideline (2015): https://www.aan.com/Guidelines/home/GetGuidelineContent/688
- Gloss D, et al. AAN antiseizure medication withdrawal practice advisory update (2021): https://www.neurology.org/doi/10.1212/WNL.0000000000012944
- Fisher RS, et al. ILAE official report: a practical clinical definition of epilepsy: https://pubmed.ncbi.nlm.nih.gov/24730690/