Drug-resistant epilepsy

Our approach to drug-resistant epilepsy

Drug-resistant epilepsy is not a single thing — it is a clinical situation defined by the failure of two appropriately chosen, dosed and tolerated antiseizure medicines to control seizures. My first job as a paediatric neurologist is to verify that definition before anything else. Were the previous trials truly representative — the right drug for the syndrome, at the right dose, for long enough, in a child who actually took it? Have the diagnoses (syndrome and aetiology) been re-examined to the best of current standards? In a substantial proportion of children referred as "drug-resistant," a sharper diagnosis or a corrected medication strategy changes the picture before we ever need to think beyond medicines.

For genuine drug-resistant epilepsy, the path forward depends on the underlying picture — the epilepsy syndrome (whether infantile spasms, Dravet, Lennox-Gastaut, focal epilepsy of structural origin, an mTORopathy, an autoimmune encephalitis, or another), the aetiology (genetic, structural, metabolic, immune), seizure semiology, EEG patterns, MRI findings, and the comorbidities that travel with it. We map all of these in one evaluation rather than across separate clinics, so that the next decision is not made in isolation.

The toolkit for drug-resistant epilepsy is broader than the medication shelf. In addition to newer antiseizure medicines, we weigh precision-medicine options when a treatable cause is identified, the ketogenic and modified Atkins diets, vagus nerve stimulation, surgical evaluation (including resective surgery, disconnection procedures and corpus callosotomy), responsive neurostimulation where appropriate, and — for selected children — regenerative and neuromodulation approaches in the framework of structured scientific studies. We choose what fits this child, not what fits a diagnosis label.

Drug-resistant epilepsy is also the area in which we carry the deepest day-to-day experience. We see a high volume of children with refractory seizures every week, which means we know the newer antiseizure medicines and their real-world side-effect and interaction profiles intimately — including those recently added to the paediatric toolkit such as cenobamate, fenfluramine, cannabidiol, ganaxolone, soticlestat and the precision-medicine agents for specific genetic epilepsies — and we know which combinations and sequences tend to work for which clinical pictures. Close, structured follow-up is part of how we use that experience: short interval reviews when a medicine is started, dose-adjusted or stopped, seizure-diary review at every contact, and a low threshold to bring families back online or in person when something does not look right.

What the in-person evaluation looks like

A first visit starts with a detailed seizure and medical history, with parents and — where possible — the previous treating team. From there I work through the picture in five overlapping passes.

The first pass re-examines the diagnosis. We read the EEG and the MRI together, alongside the seizure semiology and the medication history, and we ask whether the epilepsy syndrome and the aetiology have been correctly identified. Where the picture does not fit, we extend the work-up — epilepsy gene panel or whole-exome/whole-genome sequencing for unexplained epilepsies, targeted metabolic testing for treatable inborn errors (GLUT1, pyridoxine-dependent, biotinidase, and others), autoimmune antibody panels where the presentation suggests it, and advanced MRI or video-EEG where the resolution or detail of previous studies is insufficient.

The second pass characterises what travels with the epilepsy — development and cognition, behaviour and mood, sleep, learning, family functioning, and the medication side-effect burden. None of these can be assumed; each has to be measured before we change anything.

The third pass reviews the medication history honestly. Every antiseizure medicine the child has been on, the dose actually reached, the duration of the trial, the side effects, and the reason it was stopped or replaced. We do not call something a failed trial unless it really was one; and we do not move beyond medicines until the medicines have been given a real chance.

The fourth pass evaluates the next-step options against this particular child. Is there a precision-medicine match (for example, avoiding sodium-channel blockers in SCN1A-related Dravet syndrome, everolimus for tuberous sclerosis, vigabatrin for TSC-related infantile spasms, the ketogenic diet for GLUT1 deficiency or in any drug-resistant epilepsy where it fits)? Is the child a candidate for surgical evaluation? Would vagus nerve stimulation be a reasonable next step? Are regenerative and neuromodulation approaches — in the framework of scientific studies — worth discussing for this child?

Where the picture needs another set of eyes, I bring in the colleagues we work with regularly — clinical genetics and metabolism, paediatric neurosurgery for surgical candidacy review, a dietitian for ketogenic-diet planning, neuropsychology, child and adolescent psychiatry, and others as the case requires — for fast, focused consultations. Their findings come back to me, and the plan stays single and integrated rather than split across separate clinics.

The fifth pass is the plan itself — a written, family-facing plan that names the highest-yield next steps in order, sets explicit response and review timelines, and is honest about what we expect and what we do not.

Treatment options — beyond medicines

Precision medicine first, where it exists. When the aetiology work-up identifies a treatable underlying cause, the medication and lifestyle strategy changes accordingly — sodium-channel-blocker avoidance in SCN1A-related epilepsies, everolimus in TSC, pyridoxine in ALDH7A1 deficiency, the ketogenic diet in GLUT1 deficiency, immune therapy in autoimmune epilepsies, and so on. A single correct diagnostic step here often outperforms anything we add on top.

The ketogenic and modified Atkins diets are under-used outside specialist centres. For some syndromes (GLUT1 deficiency, infantile spasms, Doose, certain drug-resistant focal epilepsies) the diet is first-line or near first-line; for many others it is a serious option earlier than it is usually offered. Where the diet is the right step, we set it up with a paediatric ketogenic dietitian and a structured response review.

Surgical evaluation is the option families most often hear about last and most often deserve to hear about earlier. For children with structural drug-resistant focal epilepsy — particularly focal cortical dysplasia, mesial temporal sclerosis, tuberous sclerosis with a dominant tuber, and selected post-stroke or post-encephalitic epilepsies — surgery can be curative. We do not perform epilepsy surgery, but we identify candidacy, coordinate the work-up, and refer to the right surgical team with the right pre-surgical evaluation already in hand.

Neuromodulation has matured. Vagus nerve stimulation is established in selected children; deep brain stimulation and responsive neurostimulation have growing paediatric evidence in carefully chosen cases. We discuss them with families honestly — what they can and cannot deliver — and arrange the referral where they fit.

For selected children, regenerative and neuromodulation options in the framework of scientific studies become part of the conversation. Stem-cell therapies (from a number of different sources, each with its own biology and evidence profile) and exosome therapies (extracellular vesicles whose cellular source determines their molecular content, allowing patient-tailored selection) are an active research frontier. Their scientific efficacy in drug-resistant epilepsy is still in the research phase. Where we use them, it is with very explicit honesty about what is known and what is not, in carefully selected children, under appropriate authorisation, combined with photobiomodulation and other neuromodulation modalities as appropriate, and with intensive rehabilitation around the intervention — because the biology of brain plasticity argues for combining these together, not in isolation.

Follow-up is as much a part of the plan as the first decision. We set explicit response thresholds, measure progress against the same baseline, and re-evaluate at structured intervals. We do not call a treatment working on impression — we call it working when seizure frequency, cognition, behaviour and family functioning move together.