Key Takeaways
- “Drug-resistant epilepsy” has a specific meaning: failure of adequate trials of two tolerated, appropriately chosen and correctly used antiseizure medication schedules to achieve sustained seizure freedom.[1]
- Referral for surgical evaluation is not consent to surgery. It can correct the diagnosis, optimise medication and identify non-surgical options even when no operation follows.
- The central question is whether seizures consistently arise from a brain region that can be treated without causing an unacceptable loss of language, memory, vision, movement or other essential function.
- A normal routine EEG does not exclude epilepsy. Presurgical work may require seizure videos, prolonged video-EEG, epilepsy-protocol MRI, neuropsychology and selected functional tests.[3][4]
- Resection or ablation aims at a defined epileptogenic region; VNS, DBS and responsive neurostimulation generally modulate a network and have different goals. They are not interchangeable “minimally invasive cures.”
- After treatment abroad, continue antiseizure medication unless the epilepsy team changes it. Take home the captured-event data, imaging, electrode map, operative report, pathology, device details and a seizure-safety plan.
Content
Epilepsy surgery begins with doubt, not an operating-room booking. The team must first check whether the events are epileptic seizures, classify them, understand why medicines failed, locate the seizure-generating network and estimate what treatment could damage as well as what it might improve.
That distinction matters for international patients. A centre that quotes a procedure after reviewing only a short MRI report has skipped the work that makes epilepsy surgery rational. The pathway may end in resection, laser or radiofrequency ablation, disconnection, an implanted stimulator, another medication plan—or a conclusion that current evidence is not strong enough to intervene.
Start by verifying the events and the epilepsy type
A useful seizure history is concrete. For each event, record the first symptom, awareness, speech, eye or head turning, limb movement, colour change, breathing, duration, injury and recovery. A witness video, when safely obtained, can be valuable. Note time of day, sleep loss, missed doses, fever, alcohol and menstrual association.
Syncope, sleep disorders, migraine, movement disorders and functional nonepileptic seizures can resemble epilepsy. Some patients have more than one event type. NICE states that EEG can support diagnosis but should not be used to exclude epilepsy; a 12-lead ECG and assessment for metabolic causes can also be relevant after a suspected first seizure.[3]
The centre should state whether the epilepsy is focal, generalised, combined or still uncertain, and name a syndrome or structural cause when justified. Operating on a presumed “focus” before this classification is unsafe.
Check whether two medication failures were genuinely informative
The ILAE definition of drug resistance requires failure of two tolerated and appropriately chosen, adequately used medication schedules—not simply a list of two drug names.[1] For every trial, document:
- generic medicine name and whether it was used alone or in combination;
- highest dose, blood level when relevant and duration at a therapeutic dose;
- seizure type and frequency before and during the trial;
- adherence, missed doses and access problems;
- adverse effects and reason for stopping; and
- interactions, pregnancy considerations and whether the medicine suited the seizure type.
Apparent resistance can reflect a wrong diagnosis, an unsuitable drug, inadequate dose, intolerable toxicity or inconsistent access. Clarifying this is not a delay tactic. It may prevent an unnecessary procedure.
Once true drug resistance is established, referral should not wait through an indefinite sequence of new medicines. ILAE expert consensus recommends offering surgical evaluation promptly to people with drug-resistant epilepsy, with individual judgement for age, contraindications and cooperation with care.[2] NICE similarly recommends tertiary referral for drug-resistant seizures and access to video-EEG, specialised imaging, neuropsychology and surgical or VNS assessment when indicated.[4]
Build a timeline that an epilepsy conference can trust
Before travel, assemble a compact but complete file:
- onset age, event types, current frequency and longest seizure-free periods;
- emergency visits, status epilepticus, injuries and rescue-medication use;
- every antiseizure medication trial and current dosing times;
- original EEG data where export is possible, not only the typed conclusion;
- all MRI, CT, PET or SPECT images in DICOM format;
- prior surgery, pathology, implanted device and programming records;
- developmental, psychiatric, sleep and cognitive history; and
- the patient's priorities: seizure freedom, fewer injuries, pregnancy planning, work, driving, school or reduced adverse effects.
Keep a bilingual event description, but do not replace the original medical documents with a translation. Exact dates and generic drug names prevent avoidable ambiguity.
Phase one asks whether non-invasive findings converge
Most centres begin with non-invasive evaluation. The exact combination varies, but the core often includes specialist history, prolonged scalp video-EEG to record typical events, high-resolution MRI using an epilepsy protocol and neuropsychological assessment. NICE lists video-EEG telemetry, specialised neuroimaging, neuropsychology and neuropsychiatry among tertiary epilepsy services.[4]
The team compares independent lines of evidence:
- Semiology: what the seizure looks and feels like at onset;
- Ictal EEG: where electrical seizure activity appears to begin;
- Interictal EEG: spikes or other abnormalities between seizures;
- Structural imaging: lesion, scar, malformation, tumour or hippocampal change;
- Neuropsychology: pattern of memory, language and executive strengths and weaknesses; and
- Functional anatomy: which areas support language, movement, vision and memory.
Concordance is stronger when these findings point to the same clinically plausible network. “The MRI has a spot” is not enough, and a normal MRI does not automatically end evaluation; NICE specifically supports referral of people with drug-resistant epilepsy even when no MRI lesion is identified.[5]
Understand what each additional test is trying to resolve
PET, ictal SPECT, MEG, source imaging, functional MRI and other studies are not a standard shopping list. Each should answer a remaining question. Functional MRI or other mapping may help estimate language or motor organisation; neuropsychology helps anticipate cognitive risk and establishes a baseline for postoperative comparison.
Ask the team to state, before ordering a test:
- the unresolved hypothesis;
- how the result could change electrode placement or treatment;
- what a negative or discordant result would mean; and
- whether the data can be exported for review at home.
More tests do not necessarily create more certainty. Conflicting results should be discussed explicitly at a multidisciplinary epilepsy surgery conference.
Intracranial EEG is a second operation with its own question
Stereo-EEG or subdural electrodes may be considered when non-invasive findings suggest a treatable hypothesis but cannot localise the seizure network or its relationship to essential cortex well enough. This is not merely a more accurate routine EEG. Electrodes are surgically implanted according to a pre-defined sampling plan.
Before consent, the team should explain:
- the competing seizure-onset hypotheses;
- why each electrode target was selected;
- infection, bleeding, neurological and anaesthetic risks;
- how medicines may be adjusted while seizures are recorded;
- emergency arrangements for prolonged or clustered seizures;
- whether stimulation mapping is planned; and
- which findings would lead to resection, ablation, stimulation or no further procedure.
Intracranial monitoring samples only the areas reached by electrodes. A negative study can mean the hypothesis was wrong, the seizures were not captured or the relevant network was not sampled; it is not automatically proof that no epileptogenic region exists.
Compare treatment families by purpose, not by incision size
When the evidence converges, the conference should present realistic alternatives.
Resection removes a defined region. Ablation destroys a selected target through a narrower trajectory. Disconnection interrupts pathways that spread seizures while leaving some tissue in place. In carefully selected focal epilepsy, these approaches may offer a chance of seizure freedom, but probability depends on diagnosis, location, completeness of treatment and follow-up.[5]
Neuromodulation—such as vagus nerve stimulation, deep brain stimulation or responsive stimulation—aims to reduce seizures by changing network activity. It is often considered when a safely removable single focus cannot be defined, when seizures arise from more than one region or when resection carries unacceptable functional risk. Outcomes should be discussed as device- and patient-specific seizure reduction, not a universal cure.
The comparison should include:
- chance of freedom from disabling seizures and expected time horizon;
- possibility of meaningful reduction without freedom;
- language, memory, mood, visual-field and motor risks;
- haemorrhage, infection and device-related risks;
- hospital stay, staged procedures and recovery restrictions;
- continued medication and device programming needs; and
- what remains possible if the first strategy fails.
“Minimally invasive” describes an access route, not the neurological consequence of the tissue treated.
Protect language, memory, mood and the person's own priorities
The epileptogenic region may overlap or interact with brain systems the patient relies on. Neuropsychological testing is not an intelligence exam to pass. It maps current strengths and weaknesses, estimates risk and gives a baseline for later comparison. Results must be interpreted in the patient's strongest language, with education and cultural background considered.
Professional interpretation is preferable for consent and testing. Family members can describe seizures and goals but should not translate complex risk discussions alone. Mood, anxiety, psychosis history and functional seizures deserve assessment; psychiatric comorbidity may need treatment but does not by itself make a person undeserving of specialist evaluation.[2]
Consent should describe the functional trade-off in the patient's terms: remembering names at work, reading, seeing on one side, using the dominant hand, studying or living independently.
Treat the operation as one step in long-term epilepsy care
Immediately after surgery, the team monitors neurological change, seizures, wound problems, infection, bleeding, mood and medication effects. A postoperative seizure does not have one universal meaning; timing, provoking factors and relation to the patient's previous events matter.
Do not stop antiseizure medication because an operation was completed or early seizures disappeared. Any reduction should be individualised by the epilepsy team. Continue seizure precautions for bathing, heights, cooking, machinery and water until local clinicians revise them. Driving rules depend on the jurisdiction where the patient will drive, not the country where treatment occurred.
Discuss SUDEP and injury risk without using fear. NICE recommends individual assessment and discussion of epilepsy-related mortality risks, including modifiable factors such as medication adherence and uncontrolled seizures.[6] The plan should include rescue medicine when prescribed, instructions for prolonged or repeated seizures and when to call emergency services.
Make cross-border follow-up possible before travelling
Device programming, medication adjustment, pathology review and seizure-outcome assessment may continue for years. Identify the home neurologist and, for an implant, a compatible programming centre before treatment. Confirm device model, regional technical support, MRI conditions, battery plan and what happens if a lead or generator problem occurs abroad.
CDC advises medical travellers to arrange follow-up and obtain complete records in English.[8] Take home:
- the epilepsy conference conclusion and its evidence;
- representative video-EEG events plus the formal report;
- pre- and postoperative DICOM imaging;
- intracranial electrode coordinates, labels and stimulation map;
- operative report, target or resection map and pathology;
- implant manufacturer, model, serial identifiers and programmed settings;
- discharge neurological and neuropsychological baseline;
- generic medication list, taper instructions and rescue plan; and
- wound review, seizure diary and follow-up timetable.
A successful operation without a usable longitudinal record is incomplete care.
Medical disclaimer: This article is general education. It cannot diagnose epilepsy, determine drug resistance, localise seizures, estimate individual surgical benefit or replace an epilepsy surgery conference. A prolonged seizure, repeated seizures without recovery, serious injury, breathing difficulty or new neurological deficit requires urgent local care.
FAQ
Does failure of two epilepsy medicines mean surgery is definitely needed?
No. It means a specialist evaluation should be considered promptly when both trials were appropriate, tolerated and adequately used.[1][2] The evaluation may lead to surgery, stimulation, medication optimisation, a revised diagnosis or no procedure.
Can a normal EEG rule out epilepsy?
No. NICE advises that EEG can support diagnosis but should not be used to exclude epilepsy.[3] The question may require repeat, sleep-deprived, ambulatory or prolonged video-EEG selected by an epilepsy specialist.
Can surgery still be assessed if the MRI is normal?
Yes. Drug-resistant epilepsy can be referred to a tertiary centre even without an identified MRI lesion.[5] Video-EEG, expert MRI review, neuropsychology and selected functional studies may build a localisation hypothesis, although not every case becomes operable.
Is an implanted stimulator the same as removing the seizure focus?
No. Resection or ablation targets defined tissue and may offer seizure freedom in selected focal epilepsy. VNS, DBS and responsive stimulation modulate networks and usually have different goals, follow-up and device obligations.
Can medicines be stopped immediately after successful surgery?
Usually not. Medication changes depend on the epilepsy type, procedure, postoperative events, EEG and individual risk. Continue the prescribed regimen until the treating epilepsy team provides a documented taper or change.
Sources
- International League Against Epilepsy — Definition of Drug-Resistant Epilepsy
- ILAE Surgical Therapies Commission — Timing of Referral to Evaluate for Epilepsy Surgery
- National Institute for Health and Care Excellence — Epilepsy Diagnosis and Assessment (NG217)
- National Institute for Health and Care Excellence — Referral to Tertiary Epilepsy Services
- National Institute for Health and Care Excellence — Rationale for Referral and Surgical Interventions
- National Institute for Health and Care Excellence — Epilepsies in Children, Young People and Adults (NG217 PDF)
- World Health Organization — Epilepsy Fact Sheet
- US Centers for Disease Control and Prevention — Medical Tourism, Yellow Book 2026
Image Review
- Decision: Approved and retained as hero-reviewed.png.
- Editorial note: The brain graphic and consultation among a patient, neurologist and surgical team member support the article's emphasis on evaluation and shared decisions. The image does not depict a specific device, invasive electrode or guaranteed seizure outcome.