Counselling women with epilepsy about pregnancy has focused on malformations and neurodevelopment. Fetal growth has been a footnote. EURAP, the international registry of antiepileptic drugs and pregnancy, analysed 15,893 singleton livebirths exposed to antiseizure medication between 1999 and 2023, with birthweight centile as the primary outcome and adjustment for a wide range of clinical and demographic factors.
Polytherapy was worse than monotherapy: birthweight centile lower by 2.74 (95% CI -4.22 to -1.26), with about 50% higher odds of small for gestational age (adjusted OR 1.48, 1.29 to 1.70), severe small for gestational age (1.49, 1.22 to 1.83) and low birthweight (1.50, 1.15 to 1.95). The effect deepened with each added drug, reaching a 14.18 centile deficit on four medications, although only 48 pregnancies were in that group.
Among monotherapies, compared with lamotrigine, birthweight centile was lower with topiramate by 11.93 (95% CI -16.72 to -7.15), phenobarbital by 8.08, oxcarbazepine by 5.10, carbamazepine by 3.15, valproic acid by 2.54 and levetiracetam by 2.51. The combinations behaved unexpectedly: carbamazepine with levetiracetam was worse than lamotrigine monotherapy by 6.27 centiles, while lamotrigine combined with either levetiracetam or valproic acid was no different from lamotrigine alone.
Topiramate is the standout, and the size of its effect — roughly 12 centiles — is large enough to move a fetus from average to borderline. That has particular force in India, where maternal anaemia and undernutrition are common and the baseline birthweight distribution already sits low; a drug effect of this magnitude lands on a population with less margin. Phenobarbital matters here for the same reason, since it remains widely used where cost dominates prescribing.
The practical change is to add fetal growth to the preconception conversation and to the antenatal plan. Where a woman of childbearing age needs an antiseizure medication, this is another argument for lamotrigine or levetiracetam monotherapy and against topiramate. Where she is already established on something else and seizure control is fragile, the answer is not necessarily to switch — it is to plan serial growth scans.
- Add poor fetal growth to preconception counselling alongside malformation and neurodevelopmental risk.
- Prefer lamotrigine or levetiracetam monotherapy in women who may conceive; avoid topiramate.
- Where polytherapy is unavoidable, expect roughly 50% higher small-for-gestational-age risk and plan accordingly.
- Arrange serial growth scans in pregnancies exposed to topiramate, phenobarbital or polytherapy.
- Do not destabilise good seizure control to chase birthweight — monitor growth instead of switching reflexively.
The statistics, in plain English
This is registry data, not a trial. Women were not randomised to their medication, and the drug someone is on reflects their seizure type, severity and previous failures — all of which could independently affect fetal growth. The adjustment is extensive but cannot remove that entirely. Read the ranking as reliable in direction and approximate in size. The precision varies a lot with numbers: the topiramate estimate rests on 248 pregnancies and its interval is correspondingly wide at -16.72 to -7.15, while the four-drug polytherapy figure comes from just 48 pregnancies and should be treated as indicative only. The comparison of combinations is the least certain part, since some groups were small.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for neurology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free