ILAE seizure classification, spectrum-of-activity pharmacology, and individualized antiepileptic drug (AED) selection — вибір протиепілептичного препарату за типом нападу
Antiepileptic drug (AED) selection is not a single lookup table — it is a two-step decision tree that begins with accurate seizure classification. The International League Against Epilepsy (ILAE) 2017 operational classification splits seizures by onset: focal, generalized, or unknown. This single distinction determines the entire downstream pharmacologic strategy, because focal-onset seizures and generalized-onset seizures respond to fundamentally different mechanisms of drug action.
Focal-onset seizures: abnormal electrical activity begins in a network limited to one cerebral hemisphere. May remain focal-aware (no loss of consciousness) or evolve to focal-impaired-awareness, and can secondarily generalize into a bilateral tonic-clonic seizure.
Generalized-onset seizures: abnormal activity engages bilateral cortical and subcortical networks rapidly from the start. Subtypes include tonic-clonic, absence, myoclonic, atonic, and tonic. These are typically the electroclinical signature of genetic (idiopathic) generalized epilepsy (IGE) syndromes such as juvenile myoclonic epilepsy (JME) and childhood absence epilepsy (CAE).
Unknown-onset seizures: onset obscured by limited data (e.g., seizure occurring during sleep with no witness, incomplete EEG capture) — managed empirically with broad-spectrum agents until reclassified.
Why it matters pharmacologically: sodium-channel blocking drugs interrupt the fast, repetitive focal firing typical of focal epilepsy, but can paradoxically worsen the thalamocortical oscillatory circuits that generate absence and myoclonic seizures.
A patient misclassified as "focal" when they actually have juvenile myoclonic epilepsy and started on carbamazepine is one of the most common and consequential errors in epilepsy care — seizure frequency can increase within weeks of starting the wrong drug class.
Every major guideline (NICE, AAN/AES, ILAE treatment guidelines) structures its AED recommendations as a branching algorithm keyed to seizure/epilepsy type, not a flat list. The practical consequence: the same drug can be first-line for one seizure type and relatively contraindicated for another.
Clinical work-up feeding classification: • History — semiology, triggers, family history of epilepsy • EEG — interictal epileptiform discharges (focal spikes vs. generalized 3Hz spike-wave) • MRI — structural lesion suggests focal epilepsy • Age of onset and seizure pattern — strongly suggests IGE syndromes in children/adolescents
Getting this step right is the single highest-leverage decision in AED selection — it determines the entire menu of appropriate options before drug-specific factors are even considered.
Focal epilepsy responds well to drugs that dampen the fast, repetitive sodium-channel-dependent firing characteristic of a discrete cortical focus. Four agents anchor first-line focal therapy: lamotrigine, levetiracetam, carbamazepine, and oxcarbazepine — each with a distinct mechanism, efficacy profile, and tolerability trade-off.
Lamotrigine: blocks voltage-gated sodium channels, stabilizing the presynaptic membrane and reducing glutamate release. Broad-spectrum, well-tolerated, but requires slow titration to reduce rash risk.
Levetiracetam: binds synaptic vesicle protein SV2A, modulating neurotransmitter release presynaptically — a mechanism distinct from classic sodium-channel blockade. Rapid titration, minimal drug interactions, renally cleared.
Carbamazepine: classic voltage-gated sodium-channel blocker that stabilizes the inactivated channel state, reducing high-frequency repetitive firing. Strong focal efficacy but hepatic enzyme induction causes many interactions.
Oxcarbazepine: a keto-analog of carbamazepine with a similar sodium-channel mechanism but a cleaner interaction profile (weaker enzyme induction) and lower risk of certain idiosyncratic reactions.
The SANAD-II trial (Lancet 2021) found lamotrigine had superior 12-month remission and treatment retention compared to both levetiracetam and zonisamide for focal epilepsy, reinforcing its position as a preferred first-line choice.
All four drugs are considered appropriate for focal-onset seizures — selection among them is driven by patient-specific factors rather than seizure type:
• Rapid symptom control needed → levetiracetam (fast titration) • Concern for pregnancy / childbearing potential → lamotrigine (better fetal safety data than carbamazepine) • Cost-sensitive settings → carbamazepine (inexpensive, long track record) • Drug-interaction-sensitive regimens (oncology, HIV, transplant) → levetiracetam or lamotrigine (minimal enzyme interaction) over carbamazepine • Mood comorbidity → avoid levetiracetam if irritability is a concern; lamotrigine has mood-stabilizing properties
Generalized epilepsy syndromes need drugs that dampen thalamocortical oscillatory circuits broadly, not just cortical hyperexcitability at a single focus. Valproate remains the most broadly and reliably effective agent across all generalized seizure subtypes — but its teratogenicity has reshaped prescribing, elevating levetiracetam and lamotrigine as key alternatives, with an important caveat around lamotrigine and myoclonic seizures.
Valproate acts through multiple mechanisms: sodium-channel blockade, T-type calcium channel modulation (relevant to absence seizures), and enhancement of GABAergic transmission via GABA transaminase inhibition. This multi-mechanism profile is why it covers tonic-clonic, absence, AND myoclonic seizures effectively — a rare combination.
However, valproate carries the highest teratogenic risk of any commonly used AED: major congenital malformations in ~10% of pregnancies (vs. 2–3% background rate), plus dose-dependent neurodevelopmental effects (lower IQ, higher autism spectrum risk) documented in large registry studies. Regulatory agencies (MHRA, EMA, FDA) now require pregnancy prevention program enrollment for valproate use in anyone of childbearing potential.
Guidelines now explicitly state: valproate should not be prescribed to girls and women of childbearing potential unless other options have failed and a pregnancy prevention program is in place — even though it remains the single most effective drug for many generalized epilepsy syndromes.
Levetiracetam: effective across tonic-clonic, absence, and myoclonic seizures via its SV2A mechanism; increasingly first-line for JME specifically because it reliably controls the myoclonic component without the sodium-channel-related worsening seen with narrow-spectrum agents.
Lamotrigine: broad-spectrum and generally effective for tonic-clonic and absence seizures, and a reasonable choice for pregnancy-planning patients given a comparatively favorable fetal safety profile. The important caveat: a well-documented subset of patients with myoclonic seizures (particularly JME) experience aggravation of myoclonic jerks on lamotrigine, even though the drug controls their tonic-clonic seizures. This is why lamotrigine is flagged "caution" rather than "appropriate" for myoclonic seizures — effective for some, seizure-aggravating for others, and requires close monitoring after initiation.
The starkest teaching point in AED pharmacology: sodium-channel-blocking, narrow-spectrum drugs designed for focal epilepsy — carbamazepine, oxcarbazepine, phenytoin, and to a lesser extent gabapentin/pregabalin — can dramatically worsen absence and myoclonic seizures when mistakenly prescribed for generalized epilepsy syndromes. This is not a theoretical risk; it is one of the best-documented iatrogenic aggravation phenomena in neurology.
Absence and myoclonic seizures arise from thalamocortical circuits driven substantially by T-type calcium channels and GABA-B-mediated rebound bursting — not primarily by the fast sodium-channel-dependent repetitive firing that drives focal seizures. Sodium-channel blockers such as carbamazepine, oxcarbazepine, and phenytoin do not address this circuit and, in a substantial fraction of patients, actively worsen it — increasing spike-wave discharge frequency and myoclonic jerk frequency, and in severe cases precipitating absence status epilepticus.
This is the pharmacologic reason spectrum-of-activity classification (broad vs. narrow) is as important as knowing individual drug names: a narrow-spectrum agent is a rational, effective, first-line choice for focal epilepsy and simultaneously a high-risk mismatch for generalized epilepsy.
Clinical teaching mnemonic: "Carbamazepine, oxcarbazepine, phenytoin, gabapentin, and pregabalin can worsen absence and myoclonic seizures." Any patient whose seizures worsen shortly after starting one of these agents should prompt urgent reconsideration of the underlying seizure classification.
Safeguards used in real prescribing workflows:
• Never start a narrow-spectrum sodium-channel blocker without first confirming seizure/epilepsy classification via history and EEG • Treat "unknown onset" seizures with a broad-spectrum agent (levetiracetam, lamotrigine, valproate) until classification is clarified • Any suggestion of myoclonic jerks, absence staring spells, or a family history of generalized epilepsy syndromes should raise suspicion for IGE and steer away from narrow-spectrum agents • If seizures worsen after starting a new AED, reconsider the diagnosis before assuming inadequate dosing
This stage of the workflow is the pharmacologic safety check that sits between "what drugs work for this seizure type" (Stages 2–3) and "which specific drug fits this specific patient" (Stage 5).
Seizure type narrows the field to a set of pharmacologically appropriate drugs — but the final choice is individualized medicine. Comorbidities, side-effect tolerance, drug-drug interactions, cost and access, and patient-specific factors like pregnancy potential all shape which "appropriate" drug is actually the right drug for a given person.
Comorbidities: migraine favors valproate or topiramate; mood disorders favor lamotrigine (mood-stabilizing) over levetiracetam (irritability risk) or drugs with depressive side effects; neuropathic pain favors carbamazepine, oxcarbazepine, or gabapentin.
Side-effect tolerance: cognitive slowing (topiramate), weight gain (valproate), weight loss (topiramate, zonisamide), rash risk requiring slow titration (lamotrigine, carbamazepine) — must be weighed against patient priorities and lifestyle.
Drug-drug interactions: enzyme-inducing AEDs (carbamazepine, phenytoin) reduce efficacy of oral contraceptives, anticoagulants, and many oncology/HIV regimens — a critical consideration in polypharmacy patients.
Cost and access: generic older agents (carbamazepine, valproate, phenytoin) are far cheaper than newer branded options in many health systems, a major real-world constraint.
Patient-specific factors: pregnancy potential strongly disfavors valproate; renal impairment favors hepatically-cleared agents over renally-cleared levetiracetam; hepatic impairment favors the reverse.
Individualized selection is why two patients with the identical seizure classification can reasonably be started on two different first-line drugs — the "correct" answer is the intersection of pharmacologic appropriateness (Stages 1–4) and the individual patient in front of you.
1. Classify the seizure/epilepsy syndrome (ILAE framework, Stage 1) 2. Identify the pharmacologically appropriate drug set for that classification (Stages 2–3) 3. Exclude any narrow-spectrum agent that is a contraindicated mismatch (Stage 4) 4. Filter the remaining appropriate options against comorbidities, side-effect tolerability, interaction risk, cost, and patient-specific factors (this stage) 5. Initiate monotherapy at the lowest effective dose, titrate to seizure freedom or dose-limiting side effects, and reassess classification promptly if seizures worsen rather than improve
This five-step structure mirrors how epilepsy specialists actually reason through AED selection in clinic — matching pharmacology to seizure type is necessary but never sufficient on its own.