Opening GPAtlas…

Epilepsy and valproate: both sexes, under fifty-five

So this one is a screen rather than a patient. It's Friday, I'm doing repeats, and there's a request for sodium valproate for a thirty-four year old man. One clinical scenario, worked through by Sarah and Ben. Every claim in this episode is taken from our own Ocean article "Epilepsy", last reviewed 15 November 2025.

Clinical Rounds, the GPAtlas podcast. Neurology. 20 minutes. Published 31 August 2026. Free to listen.

Clinical source: our own Ocean article "Epilepsy", last reviewed 15 November 2025.

Transcript

Sarah: Hello, and welcome to GPAtlas Clinical Rounds. I'm Sarah, and this is the podcast where we take one thing you'll genuinely see this week and work out what to do with it.

Ben: And I'm Ben. Today, epilepsy, from the point of view of a repeat prescribing list rather than a neurology clinic.

Sarah: So this one is a screen rather than a patient. It's Friday, I'm doing repeats, and there's a request for sodium valproate for a thirty-four year old man. He's been on it for years, he's stable, and there is absolutely nothing about that request that asks me to think. And that's the problem, because the rules on that drug changed and most of us still file it mentally under a women's issue.

Ben: Three things. What the valproate rules actually say now, including who they apply to. The contraception interactions, which run in both directions and catch people out. And the driving rules, which patients ask us about constantly and which we answer badly.

Sarah: So valproate. Our article's line, and read the words carefully: it must not be started in any new patient, male or female, under fifty-five years, unless two specialists independently consider and document that there is no other effective or tolerated treatment.

Ben: Male or female. That's the part most people miss, because valproate has been a teratogen story for a decade and it's filed in everybody's head under pregnancy. It's now a whole- list restriction with an age on it, and the bar is two independent specialists documenting that nothing else works.

Sarah: And for women and girls of childbearing potential, it's used only under the Pregnancy Prevention Programme, and the numbers behind that are worth being able to quote, because patients ask what the actual risk is. Around one in nine babies affected by birth defects, and thirty to forty per cent by neurodevelopmental disorders. Those are big numbers and they explain why the programme is as strict as it is.

Ben: And for men there's now advice too, which is newer and less well known. The MHRA advises that men and their female partner use effective contraception during treatment and for three months after stopping, given a possible small risk of neurodevelopmental disorders in offspring, and a risk of impaired male fertility. So my thirty-four year old man on the repeat list has a conversation owing to him that probably hasn't happened.

Sarah: And one that travels with it and gets forgotten: topiramate is likewise contraindicated in women and girls of childbearing potential unless the Pregnancy Prevention Programme conditions are met. Topiramate turns up for migraine as well as for epilepsy, often prescribed by somebody who isn't thinking about epilepsy rules at all.

Ben: Male or female, under fifty-five. Two specialists, documented.

Sarah: A prescribing detail that's pure primary care and easy to breach without noticing. Category one drugs must always be maintained on the same manufacturer's product: phenytoin, carbamazepine, phenobarbital and primidone. So a switch at the pharmacy for supply reasons is a clinical event for those four, not an administrative one.

Ben: And on monitoring, our article says something that saves you work: routine blood-level monitoring is not needed. You check levels only for specific reasons, so suspected non- adherence, toxicity, pregnancy or status. What it does suggest is considering vitamin D for the enzyme-inducing drugs, carbamazepine, phenytoin, phenobarbital and primidone, because of reduced bone density.

Sarah: No routine levels. But keep those four on the same brand.

Sarah: Now contraception, which is the interaction I'd most want a GP to have straight, because it runs in two directions and the two directions have opposite consequences.

Ben: Direction one. Enzyme-inducing antiseizure medicines, so carbamazepine, oxcarbazepine, eslicarbazepine, phenytoin, phenobarbital, primidone, perampanel and topiramate, reduce the effectiveness of combined hormonal contraception, the progestogen-only pill and the implant. So the drug undermines her contraception, and she gets pregnant on a teratogen.

Sarah: Direction two, and this is the one people don't know. Lamotrigine is not an enzyme inducer, but combined hormonal contraception roughly halves lamotrigine levels, risking breakthrough seizures, with a rebound rise in the pill-free week. So here the contraception undermines the epilepsy drug, and her seizure control falls apart in a pattern that follows her pill packet.

Ben: And there's one answer that sidesteps the whole thing, which our article calls the practical default: the copper coil or the levonorgestrel coil is reliable with virtually all antiseizure medication. So when in doubt, that's the recommendation, and you don't have to hold the whole interaction table in your head.

Sarah: And a prescription that should be near-automatic and often isn't: folic acid five milligrams daily for any woman of childbearing potential on an antiseizure medication. Not just those planning pregnancy. Any.

Ben: Inducers break her contraception. The pill breaks her lamotrigine.

Sarah: Emergencies. Status epilepticus is defined as a convulsive seizure lasting five minutes or more, or three or more seizures in an hour without recovery, and it's a medical emergency. Five minutes is shorter than people instinctively wait.

Ben: And in the community, for a convulsive seizure lasting five minutes or more, or repeated seizures, you give buccal midazolam first-line, or rectal diazepam, as set out in the patient's individualised emergency plan, and you call nine nine nine. Both. The rescue medication is not instead of the ambulance.

Sarah: Five minutes. Rescue medication and the ambulance, not one or the other.

Sarah: Driving, which is the question we get asked most and answer worst. After a first seizure, driving must cease for six months from the seizure, or twelve months if there's an underlying causative factor that may increase the risk of another.

Ben: In established epilepsy it's twelve months seizure-free. And the detail that catches people: the twelve-month period restarts after any seizure, including minor seizures, auras, absences and seizures during sleep, unless a permitted-seizure exception applies. So an aura is a seizure for licensing purposes, and patients absolutely do not know that.

Sarah: And one that surprises everybody, including me. For licensing, seizures due to sleep deprivation, stress or missed meals are not classed as provoked. So the patient who says it only happened because I'd been up all night with the baby has not given you a reason to be lenient. That explanation doesn't help him.

Ben: Two more. Stopping medication is a specialist decision, usually only after a sustained seizure-free period, often at least two years, and there's no driving during withdrawal and for six months after the last dose. And for a bus or lorry licence, epilepsy requires ten years seizure-free without antiseizure medication before licensing is even considered.

Sarah: And the duty question, which comes up when someone tells you they're still driving. Notifying the DVLA is the patient's legal duty. But if a patient continues to drive against advice and cannot or will not stop, GMC guidance permits the doctor to inform the DVLA. So you're not powerless, and it's worth knowing that before you're in the conversation.

Ben: An aura restarts the twelve months. And sleep deprivation is not an excuse.

Sarah: Two more things worth carrying. First, refer every first suspected seizure for specialist assessment within two weeks, to a first-seizure clinic or a neurologist. Two weeks, for a first event.

Ben: And second, a normal EEG proves nothing. Our article says a normal EEG occurs in up to half of people with epilepsy, and incidental epileptiform activity is seen in people who never have seizures. So it's a test that's wrong in both directions, and a patient waving a normal EEG at you as evidence they don't have epilepsy needs that explained gently.

Sarah: And SUDEP, sudden unexpected death in epilepsy, which people are frightened to raise. Our article's framing makes it actionable rather than only frightening: the single biggest modifiable risk factor is uncontrolled seizures, especially nocturnal generalised tonic- clonic seizures, and medication adherence is the most effective protection.

Ben: Which turns the adherence conversation into something other than nagging. Taking the tablets is the intervention with the largest effect on the thing they're most afraid of. That's a much better reason than because I said so, and it's true.

Sarah: Adherence is the SUDEP intervention. Say it that way round.

Sarah: I want to come back to that Friday afternoon repeat list, because that's where this actually bites. A request for valproate for a thirty-four year old man doesn't look like a decision. It looks like admin. And the whole design of a repeat system is to make it not require thought, which is exactly why a rule that changed quietly doesn't reach it.

Sarah: So the practical version of this episode isn't remembering the rule. It's running a search. Everyone on valproate under fifty-five, both sexes. And then looking at what's documented, because the requirement is two specialists independently considering and documenting that there's no other effective or tolerated treatment. For someone started years ago, that documentation almost certainly doesn't exist in the form the rule now expects.

Ben: With a caution about what you do next, because the wrong response to that search is worse than no search. You don't stop anybody's valproate. Stopping an antiseizure medicine is a specialist decision and abrupt withdrawal risks status. What the search generates is a list of people who need a conversation and a specialist review, not a list of prescriptions to cancel.

Sarah: And the conversation itself is one most of these men have never had, which is why it's worth doing properly rather than as a letter. He's been on this drug for years, nobody has ever mentioned fertility or contraception to him in relation to it, and now we're raising both. If that lands badly it sounds like we've been careless for a decade.

Sarah: So I'd be honest about the shape of it. The advice about men is genuinely new, it isn't that we knew and didn't tell him, and the recommendation is contraception during treatment and for three months after stopping, with a specialist review of whether valproate is still the right drug for him. That's a straightforward, honest sentence and it doesn't require anybody to have been at fault.

Ben: Search the list. Then refer. Do not stop anybody's valproate.

Sarah: And one more group that the repeat list hides, which is topiramate. It's contraindicated in women and girls of childbearing potential unless the Pregnancy Prevention Programme conditions are met, and it turns up as a migraine drug at least as often as an epilepsy one.

Sarah: So the person prescribing it may have been thinking entirely about headaches, and the person authorising the repeat may be thinking about neither. It's the same structural problem as the trimethoprim one: the rule lives in one specialty and the prescription lives in another.

Ben: Topiramate is a migraine drug too. Same rule, different clinic.

Ben: Let me add the piece that sits underneath all of this, which is what epilepsy care in primary care actually consists of. We don't diagnose it, we mostly don't change the drugs, and the seizures happen where we aren't. What we do own is the repeat list, the contraception, the folic acid, the vaccination-style annual checks, and the conversations nobody else has time for.

Ben: And the driving one is squarely ours, because it comes up opportunistically rather than in clinic. Someone mentions in passing that they had a funny turn last month and they're driving to work. That's not a neurology conversation, that's a Tuesday conversation, and it needs the twelve-month rule and the fact that auras count.

Ben: Same with the contraception. A woman on carbamazepine asking for the pill will ask us, not her neurologist. And a woman on lamotrigine who's been started on a combined pill by a different service is a seizure risk that only shows up if somebody looks at both lists together, which is a thing only general practice does.

Sarah: And that's genuinely the argument for why this episode is worth twenty minutes of your week, even though epilepsy is a specialist condition. The specialist decisions are theirs. Almost every one of the ways this goes wrong is ours.

Sarah: The diagnosis is theirs. Most of the failure modes are ours.

Ben: One practical addition on brand consistency, because it's the most likely thing to happen without anybody deciding it. The four category one drugs, phenytoin, carbamazepine, phenobarbital and primidone, must always be maintained on the same manufacturer's product.

Ben: And the way that breaks is a supply problem at the pharmacy, which is resolved helpfully and invisibly by substituting an equivalent. For most drugs that's fine. For those four it's a clinical change, and the person best placed to notice is whoever authorises the next repeat and sees the product name has changed.

Sarah: A helpful substitution at the pharmacy is a clinical event for four drugs.

Ben: Exam corner. A thirty-four year old man with generalised epilepsy has been stable on sodium valproate for eight years. He and his partner are planning a pregnancy. What is the most appropriate advice? A, no action is needed, as valproate restrictions apply only to women. B, stop the valproate immediately. C, discuss the MHRA advice on contraception during treatment and for three months after stopping, and refer for specialist review. Or D, switch him to topiramate.

Ben: It's C. The MHRA advises that men and their female partner use effective contraception during treatment and for three months after stopping, because of a possible small risk of neurodevelopmental disorders in offspring and a risk of impaired fertility, so this needs a conversation and a specialist review rather than nothing. A is the misconception this episode exists for. B is dangerous, because stopping an antiseizure medicine abruptly risks status and it is a specialist decision. And D swaps one restricted drug for another, since topiramate carries its own Pregnancy Prevention Programme contraindication.

Sarah: So. Three things for Monday. One. Valproate is not a women's issue any more. No new starts under fifty-five in either sex without two specialists documenting there's nothing else. Two. The contraception interaction runs both ways, and the coil sidesteps all of it. And three. An aura restarts the twelve-month driving clock, and sleep deprivation doesn't count as provoked.

Sarah: One thing to reflect on, if you're logging this. Search your list for men under fifty-five on valproate. Has anybody had the contraception conversation with them? The full transcript and the references are on the episode page. Ben and I are synthetic voices. The medicine isn't. See you next week.

More from Clinical Rounds

Every episode of Clinical Rounds, free to listen with a full transcript.

We use cookies to enhance your browsing experience, provide personalised content, and analyse our traffic. By clicking "Accept All", you consent to our use of cookies. Privacy policy