DPDoctor's PassportEducation
Educational draft · awaiting clinical reviewThe full textbook explains uncertainty but does not replace live national or local guidance, specialist advice, or current prescribing information.
Full textbookMLAMSRAFoundation

Choosing and monitoring antiseizure medicines

Choose antiseizure treatment from the electroclinical syndrome and the person’s wider risks, titrate one medicine transparently, monitor meaningful outcomes and toxicity, and recognise when specialist or tertiary review is preferable to serial empirical prescribing.

!
Time-critical presentation

New widespread rash with mucosal involvement, facial swelling, fever or systemic illness after an antiseizure medicine may indicate severe cutaneous or hypersensitivity disease and needs urgent drug and medical review. Suicidal crisis, acute liver failure, severe cytopenia, pancreatitis, pregnancy while taking valproate or topiramate, or seizures after abrupt withdrawal also requires prompt escalation.

Open the sections you need. The overview is shown first.
01Purpose and principlesWhat the treatment does and how it fits into care.

Medication choice is a shared optimisation problem, not a league table. Seizure and syndrome efficacy must be balanced with mood, cognition, weight, bone health, cardiovascular and renal or hepatic disease, other medicines, work, driving, ability to adhere, pregnancy potential and family plans. Levetiracetam is easy to titrate and has few pharmacokinetic interactions but can cause behavioural or mood symptoms. Lamotrigine is often cognitively well tolerated but needs slow titration because rapid exposure increases serious rash risk. Carbamazepine can be effective for focal seizures yet induces enzymes, causes hyponatraemia and can worsen some generalised seizure types.

Monitoring starts before prescribing. Record baseline seizure frequency and phenotype, current medicines, allergy, mental health, organ function, weight and reproductive context. Obtain laboratory tests relevant to the chosen medicine rather than applying an unexamined panel to every agent. During titration, ask about target-seizure change and specific adverse effects at a time when action remains possible. Where seizure freedom is achieved, continue to monitor adherence, long-term toxicity, lifestyle and changing interactions.

Stopping treatment is also a clinical intervention. After a sustained seizure-free period, NICE supports an individualised recurrence-risk assessment, commonly after two years, rather than automatic withdrawal. Most medicines are tapered gradually over at least several months and barbiturates or benzodiazepines more slowly. Decisions must follow the current BNF, MHRA safety pack, specialist plan and local shared-care and laboratory protocols.

Key points

  • Confirm that recurrent events are epileptic, classify each seizure using the 2025 ILAE classes, and formulate epilepsy type separately as focal, generalized, combined generalized and focal, or unknown before selecting a medicine; the wrong narrow-spectrum agent can aggravate absence or myoclonus.
  • NICE first-line monotherapy for focal seizures is lamotrigine or levetiracetam, with the other considered if the first is unsuccessful; comorbidity and reproductive safety often decide between them.
  • For generalised tonic-clonic seizures, NICE offers lamotrigine, levetiracetam or sodium valproate, but valproate is subject to stringent MHRA restrictions and lamotrigine can occasionally worsen myoclonus.
  • Ethosuximide is first line for isolated absence seizures; myoclonic seizures generally need a broad-spectrum choice and avoidance of carbamazepine, phenytoin, gabapentin and related aggravating agents.
  • Start low, titrate at the product- and interaction-specific rate, define the target seizure type and allow an adequate tolerated trial before declaring failure unless toxicity requires earlier change.
  • Routine antiseizure serum concentrations are not a substitute for clinical review; use levels selectively for pregnancy, suspected non-adherence, toxicity, interactions, organ dysfunction or unexplained loss of control.
  • From 2024, valproate should not be started in a new patient under 55 unless two specialists independently document that no other effective or tolerated option exists, or compelling reasons make reproductive risks inapplicable.
  • After failure of two appropriately chosen and tolerated regimens, consider drug-resistant epilepsy and refer for tertiary evaluation rather than assuming additional medicine combinations are the only option.
02Indications, selection and cautionsWho may benefit, who needs urgent treatment and important alternatives.
Seizure-specific mismatchRed flag

New or worsening myoclonus, absence or generalised convulsions after a sodium-channel medicine should prompt urgent review of the electroclinical classification and possible medicine-induced aggravation.

Severe cutaneous reactionRed flag

Painful or extensive rash, blistering, mucosal erosion, facial oedema, fever or organ symptoms after lamotrigine, carbamazepine or another culprit requires emergency assessment rather than watchful titration.

Behavioural toxicityRed flag

Irritability, aggression, depression or suicidal thinking can emerge with several agents and is particularly important to ask about after levetiracetam initiation or dose increase.

Dose-related neurotoxicity

Diplopia, nystagmus, ataxia, sedation, tremor and cognitive slowing may indicate excessive exposure, interactions, rapid titration or impaired clearance and should be distinguished from seizure activity.

Drug-resistant pattern

Continued verified seizures despite two suitable, tolerated and adequately used schedules raises drug-resistant epilepsy and the need for comprehensive tertiary assessment.

03Assessment before treatmentTests and checks that guide safe selection.
Investigation order

Read from the initial assessment onwards. Tests may run in parallel in urgent care; first-line, preferred, confirmatory, definitive and gold-standard labels appear only when the chapter explicitly states them.

  1. 01
    Baseline phenotype and seizure diaryFirst step
    Why
    Define what improvement or failure will mean before medication changes alter the clinical picture.
    Interpretation and limitations
    Separate each seizure type and record frequency, severity, recovery and injuries. A treatment cannot be judged reliably when absences, focal events and tonic-clonic seizures are combined into one count.
  2. 02
    Medicine-specific baseline blood tests
    Why
    Identify contraindications and create a reference for foreseeable hepatic, renal, electrolyte or haematological toxicity.
    Interpretation and limitations
    Select full blood count, liver profile, renal function, sodium, weight or pregnancy testing according to the proposed agent and current product guidance. Normal baseline tests do not remove the need for symptom-triggered review.
  3. 03
    Selective serum drug concentration
    Why
    Answer a defined question about adherence, toxicity, altered clearance, pregnancy pharmacokinetics or a significant interaction.
    Interpretation and limitations
    A personal pre-pregnancy or seizure-free concentration may be more useful than a population range. Treat the person and event control, because therapeutic and toxic concentrations overlap for some agents.
  4. 04
    EEG and imaging review
    Why
    Reconsider syndrome or structural cause when the treatment response is unexpected.
    Interpretation and limitations
    A new seizure type, apparent paradoxical worsening or persistent events despite suitable therapy should trigger electroclinical reassessment; simply increasing dose may entrench a wrong diagnosis.
  5. 05
    Adherence and interaction assessment
    Why
    Identify modifiable reasons for breakthrough seizures or adverse effects before labelling pharmacological failure.
    Interpretation and limitations
    Use a non-judgemental schedule review, dispensing history where appropriate and a complete list of prescription, contraceptive, over-the-counter and recreational exposures. Enzyme induction and inhibition can change both antiseizure and co-medication effectiveness.
04Treatment approachPreparation, options, escalation and aftercare.
01SelectChoose the first monotherapyFirst stepEpilepsy is confirmed or treatment after a first seizure is justified by the recurrence-risk assessment.
  1. 11. Define seizure type and possible syndrome, revisiting witness, EEG and MRI evidence when focal and generalised features conflict.
  2. 22. Compare recommended options against pregnancy potential, contraception, mood, cognition, weight, organ function, interactions, occupation and the person’s treatment priorities.
  3. 33. Explain intended benefit, common and serious adverse effects, missed-dose and sick-day actions, driving implications and how quickly benefit can reasonably be judged.
  4. 44. Prescribe one medicine with a written titration and review plan, using current BNF, MHRA and local formulary instructions rather than memory alone.
02OptimiseRespond to incomplete controlSeizures continue or adverse effects limit the first regimen.
  1. 11. Verify events, adherence, dose, duration, sleep and provoking factors, and check whether another medicine or hormonal change altered exposure.
  2. 22. If tolerated, complete an adequate titration; if ineffective or unsafe, agree gradual substitution so abrupt withdrawal and avoidable polytherapy are minimised.
  3. 33. Recheck classification before selecting the next syndrome-appropriate monotherapy or add-on and document why the previous schedule failed.
  4. 44. After two well-conducted failures, refer to tertiary epilepsy services for diagnostic review, surgery assessment and suitable dietary, device or research options.
03ReviewMonitor benefit, harm and future plansAntiseizure treatment is continuing beyond initial titration.
  1. 11. Review seizure freedom, injuries, adverse effects, mood, cognition, adherence, driving, occupation and quality of life at intervals matched to risk.
  2. 22. Obtain agent-specific laboratory or concentration monitoring when clinically indicated and act on trends rather than collecting results without ownership.
  3. 33. Revisit pregnancy and paternity plans, contraception and regulatory documents at least whenever circumstances or medicine changes make them relevant.
  4. 44. If withdrawal is considered after sustained control, estimate recurrence consequences, taper slowly one medicine at a time and restore the previous effective dose if seizures recur under specialist advice.
05Regimens, contraindications and interactionsTreatment details and the circumstances that modify them.
NICE first-line option for focal seizures and an option for generalised tonic-clonic or idiopathic generalised epilepsy when the syndrome and risk profile fit.

Lamotrigine

For monotherapy a common start is 25 mg once daily for 14 days, then 50 mg once daily for 14 days before gradual increases; valproate and enzyme inducers require substantially different schedules from the current BNF.

Never accelerate titration casually. Stop and seek urgent assessment for concerning rash or mucosal symptoms; consider interaction with oestrogen-containing contraception and monitor for possible worsening of myoclonus.

Broad-spectrum option used first line for focal seizures and as an option for several generalised seizure contexts.

Levetiracetam

A common adult initiation is 250–500 mg twice daily with titration to response; reduce maintenance dosing in renal impairment and follow the BNF and specialist plan.

Ask proactively about irritability, aggression, depression and suicidal ideation. Adjust for renal function, avoid abrupt cessation and do not assume lack of pharmacokinetic interactions means lack of clinical adverse effects.

NICE second-line monotherapy option for focal seizures when first-line lamotrigine and levetiracetam are unsuccessful.

Carbamazepine

Adults are often started at 100–200 mg once or twice daily and increased gradually, with formulation and target dose individualised under current formulary guidance.

Can aggravate absence and myoclonic seizures, induce hepatic enzymes and reduce hormonal contraceptive effectiveness. Monitor for rash, hyponatraemia, blood or liver abnormalities and numerous interactions.

Effective broad-spectrum medicine for several generalised epilepsies when safer effective alternatives are unavailable or unsuitable.

Sodium valproate

Dose is individualised and titrated to control using the lowest effective exposure, but initiation and continuation must follow current MHRA specialist, age, sex and reproductive-risk requirements.

Highly teratogenic with neurodevelopmental risk; apply the Pregnancy Prevention Programme where required. Consider hepatic, pancreatic, platelet, weight and fertility effects, counsel men on current contraception advice, and never stop abruptly without specialist review.

06Complications, monitoring and follow-upAdverse effects, response and longer-term review.
  • Measure seizure-type-specific frequency, recovery and injury against the documented pretreatment baseline, including events the person may overlook such as absence or morning myoclonus.
  • Review mood, behaviour, sleep, cognition, balance, weight and rash during titration, involving family observations when appropriate and consented.
  • Repeat renal, liver, sodium, blood count or serum concentrations only at intervals justified by the selected agent, symptoms, comorbidity and current local guidance.
  • Audit reproductive safeguards, contraception interactions, pregnancy or paternity intentions and completion of required valproate or topiramate documents.
  • At each apparent failure, record adherence, maximum tolerated dose, duration, reason for stopping and electroclinical appropriateness so drug resistance is assessed from reliable trials.
07Special situationsVariants, exceptions and circumstances that change the usual approach.

Personal baselines beat ranges

For selected medicines, a concentration recorded during seizure freedom before pregnancy or an interaction can provide a more meaningful target than a broad laboratory reference interval.

Failure has several meanings

A medicine may fail through inefficacy, intolerability, non-adherence, interaction, wrong seizure classification or because the recurrent events are not epileptic; each demands a different response.

Polytherapy needs a hypothesis

Combining agents is most defensible when mechanisms, interactions and target seizure types have been considered and every component still contributes measurable benefit.

Withdrawal changes legal risk

A planned taper may affect driving and occupation even before a recurrence, so current DVLA rules and the consequences of relapse belong in shared decision-making.

Valproate rules evolve

Regulatory materials have changed repeatedly; prescribers must check current MHRA requirements at each relevant review rather than relying on an old consent form.

08Common pitfallsFrequent interpretation and management errors.
  1. 01

    Selecting medicine from the final bilateral convulsion while ignoring preceding morning myoclonus can expose a generalised syndrome to an aggravating narrow-spectrum drug.

  2. 02

    Escalating a dose without checking missed tablets, sleep, interactions or event diagnosis misclassifies potentially correctable breakthrough seizures as pharmacoresistance.

  3. 03

    Using serum levels as a routine pass-or-fail test disregards clinical control, toxicity and wide individual variation.

  4. 04

    Starting valproate in a person under 55 without the mandated independent specialist consideration and reproductive safeguards breaches current UK safety requirements.

  5. 05

    Stopping regular benzodiazepine, barbiturate or another antiseizure medicine abruptly can provoke withdrawal seizures and status epilepticus.

Practice

Two practice questions

Question 1 of 20 correct
NeurologyOriginal SBA

First-line focal monotherapy

A 31-year-old is diagnosed with focal epilepsy and needs first-line monotherapy. There are no immediate reproductive, renal or psychiatric factors favouring one option. Which pair does NICE recommend considering first?

Sources and review status5 sources · checked 27 Aug 2026 · clinical review pending
Sources

Sources and review status

National guidance is shown before implementation-dependent detail. Typical adult dose examples remain subject to patient factors, contraindications and the live BNF or specialist protocol. Source check completed 27 Aug 2026; clinical approval remains outstanding.

Authoring stateComplete draftClinical stateAwaiting reviewJurisdictionUnited Kingdom