Antiepileptics for Migraine Prevention
Topiramate and valproate/divalproex are the two antiepileptic drugs with Level A evidence for episodic migraine prevention. Topiramate is more widely used due to valproate's teratogenicity and weight gain profile. Both are effective, but their side effect burdens lead to high discontinuation rates — a key reason CGRP therapies have shifted treatment paradigms.
Bottom Line
- Topiramate (50-200 mg/day): Level A. Effective in EM and CM. Weight-neutral to weight loss. Limited by cognitive side effects ("brain fog"), paresthesias, and teratogenicity.
- Valproate/Divalproex (500-1500 mg/day): Level A. Effective but limited by weight gain, tremor, hair loss, and absolute teratogenicity contraindication in women of childbearing potential.
- Gabapentin and pregabalin: Insufficient evidence for routine migraine prevention (Level U/C). May be useful for comorbid neuropathic pain or anxiety.
- Both topiramate and valproate showed efficacy comparable to propranolol in head-to-head trials
- The ELEVATE trial showed atogepant 60 mg was more effective and better tolerated than topiramate in treatment-experienced patients
Topiramate
Evidence
MIGR-001 and MIGR-002 (Brandes 2004, Silberstein 2004): Two pivotal RCTs in episodic migraine, each randomizing ~480 patients to topiramate 50, 100, or 200 mg/day vs placebo for 6 months.
- Change in monthly migraine days: -1.4 to -2.1 days advantage over placebo (100 mg dose)
- ≥50% responder rate: 47-54% (100 mg) vs 23% placebo
- 100 mg/day was the optimal dose — 200 mg provided no additional benefit with more side effects
Chronic Migraine: Topiramate 100 mg/day reduced headache days by ~3.5/month vs placebo in CM trials (Silberstein 2007, Diener 2007). Less widely used for CM since onabotulinumtoxinA and CGRP therapies are generally preferred.
ELEVATE (Ailani 2024): Head-to-head trial of atogepant 60 mg vs topiramate (50-100 mg) in patients who had failed 2-4 prior preventives. Atogepant showed superior efficacy (58% vs 38% responder rate) and far lower discontinuation (6% vs 26%).
Dosing
Topiramate: Practical Prescribing
- Start: 25 mg at bedtime
- Titrate: Increase by 25 mg weekly. Slow titration minimizes cognitive side effects.
- Target: 50-100 mg/day (divided BID or once daily at bedtime)
- Maximum: 200 mg/day (rarely needed; more side effects without additional efficacy)
- Onset of benefit: 4-8 weeks at target dose; assess at 2-3 months
- Extended-release formulation (Trokendi XR, Qudexy XR) allows once-daily dosing and may have fewer peak-related cognitive effects
Side Effects
| Side Effect | Frequency | Notes |
|---|---|---|
| Cognitive dysfunction ("brain fog") | 15-25% | Word-finding difficulty, slowed processing. Most common reason for discontinuation. Dose-dependent. |
| Paresthesias (fingers, toes, perioral) | 30-50% | Related to carbonic anhydrase inhibition. Usually mild; resolves with continued use or dose reduction. |
| Weight loss / appetite suppression | 10-15% | Can be advantageous in overweight patients; problematic in underweight patients. |
| Taste alteration (carbonated drinks taste flat) | Common | Carbonic anhydrase inhibition |
| Kidney stones | 1-2% | Calcium phosphate stones. Counsel on hydration; avoid in patients with recurrent nephrolithiasis. |
| Metabolic acidosis | Dose-dependent | Check bicarbonate if symptoms (fatigue, dyspnea); more relevant at higher doses. |
| Acute myopia / angle-closure glaucoma | Rare | Typically within first month. Discontinue immediately if acute visual changes occur. |
Topiramate and Pregnancy
- FDA Category D (previously X for some indications) — associated with oral clefts (cleft lip/palate) with first-trimester exposure
- Risk of oral clefts: ~0.3% exposed vs ~0.07% general population (3-5x increase)
- Also associated with SGA (small for gestational age) infants
- Ensure reliable contraception in women of childbearing potential
- Topiramate reduces the efficacy of combined oral contraceptives at doses >200 mg/day (enzyme induction). Low-dose OCP may not be reliable.
- Use alternative preventive (beta-blocker, CGRP mAb with washout planning, or memantine) when pregnancy is planned
Valproate / Divalproex
Evidence
- Multiple RCTs (Hering 1992, Jensen 1994, Klapper 1997, Freitag 2002): Divalproex 500-1500 mg/day consistently reduces migraine frequency by ~40-45%
- ≥50% responder rate: approximately 44-48% vs 20-25% placebo
- FDA-approved for migraine prevention (divalproex sodium ER 500-1000 mg/day)
Dosing
- Start: Divalproex ER 250-500 mg at bedtime
- Titrate: Increase by 250 mg weekly
- Target: 500-1000 mg/day (ER formulation, once daily)
- Maximum: 1500 mg/day (higher doses rarely more effective for migraine)
- Serum levels are not routinely necessary for migraine dosing (unlike epilepsy), but can guide dosing if uncertain: target 50-100 mcg/mL
Side Effects
- Weight gain: Average 3-5 kg; significantly limits long-term use
- Tremor: Dose-dependent; can worsen essential tremor
- Hair loss / thinning: Common; may be mitigated with zinc and selenium supplementation (limited evidence)
- GI effects: Nausea, especially with immediate-release formulations; ER is better tolerated
- Hepatotoxicity: Rare in adults (>2 years old, monotherapy); check baseline LFTs, repeat at 1-2 months
- Thrombocytopenia: Dose-dependent; check CBC if bruising or at higher doses
- Pancreatitis: Rare but serious; counsel on abdominal pain symptoms
- PCOS-like syndrome: Weight gain, menstrual irregularities, hyperandrogenism reported with long-term use in women
Valproate and Pregnancy: Absolute Contraindication
- FDA Category X for migraine — the most teratogenic commonly used AED
- Neural tube defects (spina bifida): 1-2% risk (10-20x general population)
- Also causes cardiac malformations, craniofacial defects, limb anomalies, and neurodevelopmental impairment (lower IQ by 8-10 points, increased autism risk)
- Teratogenic effects are dose-dependent but no safe dose exists
- Never prescribe to women of childbearing potential for migraine without exhausting alternatives and ensuring iron-clad contraception
- International guidelines (EMA, FDA) recommend avoiding valproate in women <55 years unless no alternative and pregnancy prevention program is in place
Other Antiepileptics
| Agent | Evidence | Status |
|---|---|---|
| Gabapentin | Level U (conflicting data) | One positive trial at 2400 mg/day; subsequent trials negative. Not recommended as first-line. May help when comorbid anxiety or neuropathic pain. |
| Pregabalin | Limited data | No RCTs specifically for migraine prevention. Sometimes used off-label for comorbid fibromyalgia. |
| Lamotrigine | Negative for migraine frequency | Does not reduce migraine attack frequency. May reduce aura frequency specifically (small trials). Consider for migraine with frequent, disabling aura. |
| Levetiracetam | Insufficient evidence | Open-label data suggested benefit; RCTs were underpowered. Not recommended. |
| Zonisamide | Limited (Level C) | Similar mechanism to topiramate (carbonic anhydrase inhibition). Small studies suggest benefit at 200-400 mg/day. Alternative when topiramate causes intolerable cognitive effects. |
| Carbamazepine / Oxcarbazepine | Not effective | No evidence for migraine prevention. Should not be used. |
References
- Silberstein SD, et al. Evidence-based guideline update: pharmacologic treatment for episodic migraine prevention in adults. Neurology. 2012;78(17):1337-1345.
- Brandes JL, et al. Topiramate for migraine prevention: a randomized controlled trial. JAMA. 2004;291(8):965-973.
- Silberstein SD, et al. Topiramate in migraine prevention: results of a large controlled trial. Arch Neurol. 2004;61(4):490-495.
- Ailani J, et al. Atogepant versus oral standard of care for migraine prevention (ELEVATE). Lancet Neurol. 2024.
- Klapper J. Divalproex sodium in migraine prophylaxis: a dose-controlled study. Cephalalgia. 1997;17(2):103-108.
- Freitag FG, et al. A randomized trial of divalproex sodium extended-release tablets in migraine prophylaxis. Neurology. 2002;58(11):1652-1659.
- Meador KJ, et al. Fetal antiepileptic drug exposure and cognitive outcomes at age 6 years (NEAD study). Lancet Neurol. 2013;12(3):244-252.