Topiramate: Independent Evidence on Epilepsy, Migraine & Risks

Table of contents

• Evidence summary
• What topiramate is
• How it works
• What it is prescribed for
• What works and what does not
• Benefits by claim
• Risks and all side effects
• All interactions
• Who should avoid topiramate
• Dosage and how to take it
• Follow the money: who makes it and who funded the evidence
• Related research
• Frequently asked questions
• Sources and funding notes

Evidence summary

Independent evidence and credibility scorecard

Independence tiers: 1 = regulator, government body or publicly funded research with no declared industry money; 2 = academic or Cochrane team with no declared commercial funding for the work, or a private foundation funder we did not trace further; 3 = academic work that also lists commercial grants, or analyses mostly manufacturer-run trials; 4 = manufacturer-funded or manufacturer-authored. Credibility runs from A (highest) to D.

Sources cluster in the UK, the Nordic countries and the US. Both main pregnancy datasets are Western registers, so rates for other populations are untested. The two groups with opposing financial interests, the maker's label and public registries, point in the same direction on pregnancy harm.

What topiramate is

Topiramate is an anti-seizure (antiepileptic) medicine taken by mouth. It is also called an antiseizure medication or ASM.

• Brand name: Topamax (tablets and "sprinkle" capsules). Original US approval was on 24 December 1996 for NDA 020505, sponsor Janssen, as a new molecular entity (Drugs@FDA). The 25, 50, 100 and 200 mg Topamax tablets are still listed as active.
• Generics: openFDA returns 27 generic (ANDA) application records for topiramate (openFDA query). The UK brand name is also Topamax (MHRA).
• Combination capsule: Qsymia contains phentermine plus extended-release topiramate. It was approved on 17 July 2012 under NDA 022580, sponsor VIVUS LLC (Drugs@FDA). It is a Schedule IV controlled substance (label heading "CIV") (Qsymia label).
• Prescription status: prescription-only (NHS).
• Forms in the UK: tablets, an oral liquid and capsules that can be swallowed whole or sprinkled on soft food (MHRA DSU).

How it works

Nobody knows exactly. The US label says the "precise mechanisms" for seizures and migraine are unknown. Laboratory studies show four actions that may contribute: blocking voltage-dependent sodium channels, boosting GABA-A receptor activity, blocking AMPA/kainate glutamate receptors, and inhibiting the enzyme carbonic anhydrase (FDA label 12.1). Those actions are shown in preclinical work. What is proven in humans is the clinical effect (fewer seizures and migraines) from the trials above. The carbonic anhydrase action is important for safety, because it explains kidney stones and acidosis.

For weight loss, the Qsymia label states that the mechanism of topiramate's effect is "not known". It suggests appetite suppression and increased fullness may be involved, and that phentermine, an amphetamine-like drug, is thought to act by releasing catecholamines in the hypothalamus (Qsymia label 12.1).

What it is prescribed for

• Epilepsy: US label: first single-drug treatment for partial-onset or primary generalised tonic-clonic seizures from age 2; add-on for those and for Lennox-Gastaut syndrome (FDA label 1).
• Migraine prevention: from age 12 in the US (FDA label 1.3). NICE says that in June 2025 topiramate for migraine prevention in children and young people was off-label in the UK (NICE CG150).
• Weight management (as Qsymia): adults and children from 12 with obesity, and adults with overweight plus a weight-related condition, alongside diet and activity (Qsymia label).
• Other uses: doctors also prescribe topiramate for other conditions. We did not review evidence for those uses.

What works and what does not

Benefits by claim

Seizure control

The Cochrane add-on review pooled 11 trials in 1,401 people whose partial seizures were not controlled. Compared with placebo, topiramate made a 50% or greater drop in seizure frequency almost three times as likely (RR 2.97, 95% CI 2.38 to 3.72). Seizure freedom was more likely too (RR 3.41, 95% CI 1.37 to 8.51), although the interval is wide because few people reached it. Doses above 300 to 400 mg a day added nothing. The same review found more dropouts on topiramate (RR 2.44), and that side effects were common (Pulman 2014).

As a first single drug, the picture is less rosy. SANAD randomised 1,721 people with partial epilepsy to carbamazepine, gabapentin, lamotrigine, oxcarbazepine or topiramate. Lamotrigine was better than topiramate on time to treatment failure (HR 0.64, 95% CI 0.52 to 0.79). Time to 12-month remission was not significantly different (carbamazepine versus topiramate HR 0.86, 0.72 to 1.03) (SANAD A). In 716 people with generalised or unclassifiable epilepsy, valproate beat topiramate on treatment failure (HR 1.57, 1.19 to 2.08), with no significant difference in remission (SANAD B). Valproate has its own severe pregnancy risks, which is why current NICE advice no longer makes it a first choice for women able to have children (NICE NG217).

Migraine prevention

Across 17 trials, topiramate cut migraine attacks by 1.2 per 28 days versus placebo (95% CI -1.59 to -0.80; 9 trials, 1,737 people). About twice as many people reached a 50% reduction (RR 2.02, 95% CI 1.57 to 2.60). That is a number needed to treat of 4 (95% CI 3 to 6): treat four people and one extra person gets at least a halving of attacks. At 100 mg the average gain was 1.15 fewer attacks, at 200 mg 0.94, and 200 mg was no better than 100 mg (Linde 2013). NICE says to consider propranolol, topiramate or amitriptyline after a full discussion of benefits and risks, and to avoid topiramate in pregnancy and in women able to have children unless the Pregnancy Prevention Programme is met (NICE CG150).

Weight loss

The Qsymia label reports two 56-week trials (NCT00554216 and NCT00553787). In Study 2 (people with overweight or obesity plus two or more conditions) average weight change was -1.2% on placebo, -7.8% on 7.5/46 mg and -9.8% on 15/92 mg. At least 5% weight loss was reached by 21%, 62% and 70% respectively (Qsymia label 14). The published CONQUER report of that trial lists "Funding Vivus" (Gadde 2011). The independent network meta-analysis of 28 trials (29,018 people) found phentermine-topiramate gave 8.8 kg more weight loss than placebo (95% CrI -10.20 to -7.42) and had the highest odds of 5% weight loss (OR 9.22, 6.63 to 12.85) among five approved drugs. But it noted high attrition (30% to 45%) lowers confidence in all estimates (Khera 2016). The label itself states the effect on heart disease and death has not been established.

Risks and all side effects

Common effects

The label lists the most common adverse reactions in epilepsy trials (at least 10% more frequent than placebo or low-dose topiramate) as paraesthesia, anorexia, weight loss, speech problems, fatigue, dizziness, somnolence, nervousness, psychomotor slowing, abnormal vision and fever. In migraine trials (at least 5% more frequent than placebo) they were paraesthesia, anorexia, weight loss, difficulty with memory, taste change, diarrhoea, hypoaesthesia, nausea, abdominal pain and upper respiratory tract infection (FDA label 6.1). The NHS lists feeling sick, dizziness, tiredness, diarrhoea, loss of appetite or weight loss, shaking and low mood, and says most ease after a couple of weeks (NHS).

Thinking and mood

The label groups cognitive effects into confusion, psychomotor slowing, trouble with concentration, memory, speech or word-finding; mood problems; and somnolence or fatigue. Fast titration and higher starting doses made cognitive problems more common. In adult epilepsy add-on trials with weekly increases of 100 to 200 mg, 56% at 800 to 1,000 mg a day, about 42% at 200 to 400 mg and 14% on placebo had cognitive-related dysfunction (FDA label 5.6). The Cochrane add-on review found concentration difficulties were nearly eight times more likely than on placebo (RR 7.81, 95% CI 2.08 to 29.29) and "thinking abnormally" nearly six times (RR 5.70, 99% CI 2.26 to 14.38) (Pulman 2014). Anti-seizure drugs as a class are linked with a slightly higher rate of suicidal thinking: 0.43% versus 0.24%, about one extra case per 530 people treated (FDA label 5.5).

Other label warnings

• Hyperammonaemia and encephalopathy: dose-related and more frequent with valproic acid; symptoms include lethargy and altered consciousness (FDA label 5.12).
• Hypothermia with valproate: body temperature below 35 °C has been reported (FDA label 5.14).
• Children: reduced bone mineral density (clinically important reductions in 21% of treated children in a one-year study) and slowed height and weight gain (FDA label 5.9 and 5.10).
• Stopping suddenly: may provoke seizures, even in people without epilepsy (FDA label 5.8).

Extra effects with Qsymia

The most common reactions in adults (at least 5% and 1.5 times placebo) were paraesthesia, dizziness, taste change, insomnia, constipation and dry mouth. On the top dose, 20% had paraesthesia and 19% dry mouth, versus 2% and 3% on placebo. Heart rate rose, and the label warns about mood and sleep disorders, cognitive impairment, and a reduction in renal function (Qsymia label 5 and 6).

Pregnancy and birth: the numbers

The 2024 UK review concluded that topiramate is among the anti-seizure medicines with the highest malformation risk (about 4 to 9 per 100 babies versus 1 to 3 in the general population), that risk rises with dose and no threshold below which risk is absent can be set, and that about 18 in 100 babies are born small for gestational age, compared with about 5 in 100 for women with epilepsy who take no anti-seizure medicine (MHRA DSU).

• Oral clefts: in the US pregnancy registry NAAED, 1.4% of topiramate-exposed infants had oral clefts versus 0.11% of infants of women without epilepsy or medicines (RR 12.5, 95% CI 5.9 to 26.37). The UK Epilepsy and Pregnancy Register found 3.2% on topiramate alone, 16 times the UK background rate of 0.2% (FDA label 8.1).
• Other malformations: excluding oral clefts, NAAED found 4.1% major malformations on topiramate versus 1.8% on a reference drug and 1.1% in unexposed women (FDA label 8.1).
• Versus lamotrigine: Nordic registers, adjusted RR 1.81 (95% CI 1.26 to 2.60) from 509 exposed pregnancies; risk rose with dose (Cohen 2023).
• Small babies: NAAED 19.7% SGA versus 7.9% on a reference drug and 5.4% unexposed; Norway 25% versus 9% (FDA label 8.1). A separate NAAED analysis put topiramate at 18.5% against 7.3% for lamotrigine (Hernandez-Diaz 2017).
• Autism and intellectual disability: in 4.49 million children, topiramate monotherapy had aHR 2.8 for autism and 3.5 for intellectual disability. Cumulative incidence by age 8 was 4.3% for autism and 3.1% for intellectual disability, and the authors reported that risks rose with dose (Bjørk 2022).

Observational studies cannot rule out confounding by the mother's epilepsy. But the registry results, the Nordic data and the animal data in the label point the same way, which is why regulators acted (FDA label 8.1).

All interactions

The tables below are drawn from the US label unless otherwise cited. Tell every prescriber and pharmacist you take topiramate.

Detailed UK advice on which contraceptive methods remain reliable with topiramate sits in FSRH guidance and the BNF, which we could not open. We therefore do not list specific methods.

Who should avoid topiramate

• Pregnancy: contraindicated for migraine prevention; for epilepsy only if there is no other suitable treatment (MHRA). Qsymia is contraindicated in pregnancy and is available through a restricted REMS programme (Qsymia label 4 and 5.1).
• Women and girls who could become pregnant: not to be prescribed unless the Pregnancy Prevention Programme is fulfilled: highly effective contraception, a negative pregnancy test before starting, and signing a Risk Awareness Form at start and yearly review. Contraception continues for at least 4 weeks after the last dose. The conditions also apply to girls who are not sexually active unless the prescriber sees compelling reasons that there is no risk (MHRA DSU).
• Breastfeeding: topiramate passes into breast milk; diarrhoea and sleepiness were reported in breastfed infants. For Qsymia, breastfeeding is not recommended (FDA label 8.2; Qsymia label 8.2).
• Glaucoma, kidney or liver problems: the NHS lists these as reasons topiramate may not suit you (NHS). Qsymia is contraindicated in glaucoma and hyperthyroidism (Qsymia label 4).
• Kidney impairment: clearance is reduced in moderate and severe impairment, and a dose adjustment is recommended (FDA label 8.6).
• Children: label approvals begin at age 2 for epilepsy and 12 for migraine; growth, bone density, heat and kidney stones need watching (FDA label 5.9 and 5.10).
• Anyone with prior serious rash or allergy to topiramate: do not take it again (NHS).

Dosage and how to take it

These are label doses and trial doses, not advice for any person.

• Epilepsy, single drug (US label): 400 mg a day in two divided doses for adults and children 10 and over, reached over six weeks of stepwise increases starting at 25 mg twice daily (FDA label 2.1).
• Epilepsy, add-on (US label): 200 to 400 mg a day in adults with partial seizures, 400 mg a day for primary generalised tonic-clonic seizures; start at 25 to 50 mg a day with weekly steps; doses above 400 mg have not been shown to help (FDA label 2.2).
• Migraine prevention (US label): 100 mg a day in two divided doses, built up over four weeks, starting with 25 mg in the evening (FDA label 2.3). The Cochrane review found 100 mg a day effective and 200 mg no better (Linde 2013).
• Qsymia (label): once daily in the morning; start at phentermine 3.75 mg / topiramate 23 mg for 14 days, then 7.5/46 mg. The label sets reassessment points and says to stop the top dose gradually to avoid seizure (Qsymia label 2).
• How it is taken: tablets are bitter and should not be broken; sprinkle capsules can be opened onto a teaspoon of soft food (FDA label 2.4). The NHS says to take it twice a day at the same times and not to stop suddenly (NHS).

Quality and product issues: we did not find a regulator recall or contamination finding for topiramate during this review. Absence of evidence from our searches is not proof that none exists.

Follow the money: who makes it and who funded the evidence

Who profits either way

Many generic makers sell topiramate; openFDA lists 27 generic applications (openFDA). Makers of alternative drugs (for example lamotrigine, levetiracetam, propranolol or amitriptyline) would gain if topiramate use fell, though most of those are also generics. For weight loss, topiramate sits inside a branded product with ten listed US patents, so its owner gains if the combination is prescribed (Qsymia label).

Who makes and sells it

• Originator: Topamax, NDA 020505, approved 24 December 1996, sponsor Janssen (Drugs@FDA). The label tells people to report problems to Janssen Pharmaceuticals, Inc. (FDA label). Johnson & Johnson's SEC filing lists its business address as New Brunswick, New Jersey (SEC EDGAR). We did not open a document stating Janssen's corporate parent in this review, so we make no claim about its corporate parent.
• Generics: companies named in the FDA's therapeutic-equivalents list for topiramate tablets include Accord, Aurobindo, Cipla, Glenmark, Unichem and Zydus (Drugs@FDA).
• Qsymia: sponsor VIVUS LLC, Campbell, California; made for VIVUS by Catalent Pharma Solutions in Winchester, Kentucky (Qsymia label; Drugs@FDA). Approved 17 July 2012. The last listed action is a REMS modification supplement in June 2026. We did not trace VIVUS's owners.

Who funded the evidence

• Public and charitable: SANAD (NIHR Health Technology Assessment programme and Wellcome Trust); the epilepsy add-on Cochrane review (UK Medical Research Council and NIHR grants); the Nordic malformation study (NordForsk, Research Council of Norway, UNSW); the obesity meta-analysis (US NIH grants). Source for each is the grant record in Europe PMC, cross-checked against the PubMed entries (SANAD A; Pulman; Cohen; Khera).
• Private foundation: the Nordic neurodevelopment study lists the Novo Nordisk Foundation and NordForsk (Bjørk 2022). We did not trace the foundation's holdings.
• Mixed: the NAAED pregnancy registry analysis lists a US NINDS grant plus Sunovion, AbbVie, Pfizer and Novartis (Hernandez-Diaz 2017). None of them is the Topamax sponsor, but we did not check their product lines or whether their support could matter.
• Manufacturer: CONQUER, the large Qsymia trial, lists "Funding Vivus" (Gadde 2011). The label's weight-loss and side-effect tables for Qsymia come from trials the company ran.

Documented integrity events

We could not open the US Department of Justice release on the 2010 Topamax marketing settlement, so we make no claim about it here. Check it directly if it matters to you.

Regulatory history of pregnancy warnings: the MHRA reports topiramate has been contraindicated in pregnancy for migraine prophylaxis since 2010 and tightened rules again in 2024 (MHRA DSU).

Where the sources come from

The UK (MHRA, NICE, MRC/NIHR, Wellcome), the Nordic countries and the US dominate. The manufacturers (Janssen and VIVUS) are US firms. No source we used came from a manufacturer of generic topiramate, from China or India (major generic producers), or from a low-income country.

What we could not trace

Funding statements for the 2013 migraine Cochrane review; the current ownership of VIVUS; the funding of the pooled epilepsy trials; and the full text of the DOJ settlement.

Related research

Related Pure City Research guides:

• Amitriptyline: independent evidence on pain, migraine, IBS and risks (another migraine-prevention option)
• Propranolol for anxiety: independent evidence and side effects (propranolol is also used for migraine prevention)
• Clonazepam: independent evidence on panic, epilepsy and withdrawal
• Lorazepam: independent evidence on anxiety, seizures and withdrawal
• Lithium: independent evidence on bipolar, suicide and kidney risk (interacts with topiramate)
• Bupropion: independent evidence on depression and side effects
• Semaglutide: independent evidence on weight loss, risks and money (a different weight-loss drug)
• Tirzepatide: independent evidence and risks (another weight-loss drug)
• Hypertension in pregnancy and pre-eclampsia: independent evidence
• Stress, anxiety and depression: evidence and care options

Frequently asked questions

Sources and funding notes

What we could not verify: the full text and funding section of the 2013 migraine Cochrane review; sponsorship of the individual pooled epilepsy trials; the March 2026 revision of the US Topamax label (search results indicated one; we read the May 2023 FDA-hosted version); FSRH guidance on which contraceptive methods remain reliable; the 2010 US Department of Justice Topamax settlement (the release could not be opened); the parent company of Janssen and current owners of VIVUS; the EMA's position on Qsymia; the Novo Nordisk Foundation's corporate holdings; and whether the Hernandez-Diaz registry sponsors influenced the work. Regulators' and NICE's own funding lines were not stated on the pages we opened except where noted.

1. FDA-hosted Topamax prescribing information (Janssen; revised May 2023) — US regulator-approved label; text drafted by the sponsor.
2. Qsymia prescribing information (VIVUS LLC; revised March 2026) — US regulator-approved label; sponsor-drafted; weight-loss data from sponsor trials.
3. Drugs@FDA, NDA 020505 (Topamax) — US FDA database; sponsor Janssen; approved 24 December 1996.
4. Drugs@FDA, NDA 022580 (Qsymia) — US FDA database; sponsor VIVUS LLC; approved 17 July 2012.
5. openFDA Drugs@FDA query (topiramate generics) — US government data; 27 ANDA records at time of checking.
6. MHRA Drug Safety Update, 20 June 2024 (web page) — UK government agency.
7. MHRA Drug Safety Update, June 2024 (PDF) — UK government agency; source of malformation, SGA and neurodevelopment figures.
8. MHRA news story: new pregnancy prevention measures for topiramate — UK government; gives the approximate number of affected female patients in England.
9. NICE CG150: headaches in over 12s — UK public body; migraine prevention recommendations.
10. NICE NG217: epilepsies in children, young people and adults — UK public body; add-on recommendations and topiramate restrictions.
11. NHS medicines page: topiramate (reviewed 29 July 2026) — UK health service; patient information.
12. Linde et al., Cochrane 2013: topiramate for migraine prophylaxis in adults — Cochrane (international); funding text not found.
13. Pulman et al., Cochrane 2014: topiramate add-on for drug-resistant partial epilepsy — Cochrane; MRC and NIHR grants.
14. Marson et al., Lancet 2007 (SANAD arm A) — NIHR HTA programme and Wellcome Trust.
15. Marson et al., Lancet 2007 (SANAD arm B) — NIHR HTA programme and Wellcome Trust.
16. Bjørk et al., JAMA Neurology 2022: prenatal antiseizure medication, autism and intellectual disability — Novo Nordisk Foundation and NordForsk.
17. Cohen et al., Annals of Neurology 2023: antiseizure monotherapy and major malformations — NordForsk, Research Council of Norway, UNSW.
18. Hernandez-Diaz et al., Annals of Neurology 2017: fetal growth and premature delivery — NINDS plus Sunovion, AbbVie, Pfizer and Novartis (Europe PMC grant list).
19. Gadde et al., Lancet 2011 (CONQUER) — Funding Vivus (the maker).
20. Khera et al., JAMA 2016: pharmacological treatments for obesity — US NIH grants.

Educational research reviewed 6 October 2026. This article is not medical advice; talk to a qualified clinician or pharmacist before starting, stopping or combining treatments.