AAN/AES/SMFM Guideline: ASM Teratogenesis, Perinatal & Neurodevelopmental Outcomes in Pregnancy (2024)
This topic summarizes the 2024 AAN/AES/SMFM practice guideline on anti-seizure medication (ASM) use during pregnancy, covering teratogenesis, perinatal outcomes, neurodevelopmental effects, and folic acid supplementation. Published in Neurology 2024;102:e209279, this guideline updates the 2009 AAN practice parameter based on a systematic review of 82 articles through August 2022.
🔹 Bottom Line
- Guideline: AAN/AES/SMFM 2024 practice guideline — 82 included articles (69 from 2009 review + 13 new); systematic review through August 2022
- Valproic acid has the highest MCM rate in monotherapy (9.7% prevalence per 1,000) — must be avoided in PWECP to reduce MCMs, NTDs, poor neurodevelopmental outcomes, and SGA risk (Level A)
- Safest ASMs (lowest MCM rates): Lamotrigine (3.1%), levetiracetam (3.5%), and oxcarbazepine (3.1%) in monotherapy → must consider these in PWECP (Level A)
- Valproic acid → lowest IQ: Full-scale IQ 93.9 (reference) vs lamotrigine 105.8, levetiracetam 99.0, carbamazepine 100.4 — likely associated with decreased IQ and increased ASD risk (Level A)
- Phenobarbital: Highest cardiac malformation rate (4.4%) — must avoid to reduce cardiac MCMs (Level A)
- Topiramate: Avoid in PWECP — associated with oral clefts (1.4%) and SGA risk (Level B)
- Folic acid ≥0.4 mg/d: Must prescribe preconceptionally and during pregnancy to reduce NTDs and possibly improve neurodevelopmental outcomes (Level A/B)
- Seizure control is paramount: Minimize convulsive seizures (GTC/FBTC) during pregnancy — mortality risk 5–12× greater in PWECP (Level A)
- Exercise caution in removing/replacing an ASM effective for GTC/FBTC once PWECP is already pregnant (Level B)
- Monitor ASM levels throughout pregnancy; lamotrigine and levetiracetam concentrations decrease during pregnancy (Level B)
Guideline Overview
Source & Scope
- Organizations: American Academy of Neurology (AAN), American Epilepsy Society (AES), Society for Maternal-Fetal Medicine (SMFM)
- Published: Neurology 2024;102:e209279 (June 11, 2024)
- Lead author: Alison M. Pack, MD, MPH (Columbia University)
- Updates: 2009 AAN practice parameter on women with epilepsy & pregnancy
- Evidence base: 82 total articles (50 initial + 13 update + 19 from 2009 guideline); Class IV studies excluded (n = 133)
- Population: People with epilepsy of childbearing potential (PWECP) — gender-neutral terminology adopted
- Process: AAN 2017 Clinical Practice Guideline Process Manual; modified GRADE methodology
Four Clinical Questions Addressed
- Prevalence of MCMs by specific ASMs (monotherapy vs polytherapy, high vs low doses)?
- Prevalence of adverse perinatal outcomes by specific ASMs?
- Prevalence of adverse neurodevelopmental outcomes by specific ASMs?
- Effect of folic acid on MCMs, perinatal outcomes, and neurodevelopmental outcomes?
Key Terminology
- PWECP: People with epilepsy of childbearing potential
- MCM: Major congenital malformation
- NTD: Neural tube defect
- SGA: Small for gestational age
- ASD: Autism spectrum disorder
- RMD: Raw mean difference (for IQ comparisons)
- PD: Prevalence difference; PR: Prevalence ratio
🔹 Clinical Pearl
The 2024 guideline uses PWECP (people with epilepsy of childbearing potential) instead of the 2009 term “women with epilepsy” — reflecting gender-inclusive language. The guideline emphasizes that >65% of pregnancies in PWECP are unintended, making preconceptional counseling critical at every visit.
Recommendation 1: General Principles
Recommendation 1 Statements
| Statement | Recommendation | Level |
|---|
| 1A | Clinicians should engage in joint decision-making with PWECP, taking individual preferences into account when selecting ASMs and monitoring dosing | B |
| 1B | When treating PWECP, clinicians should recommend ASMs and doses that optimize both seizure control and fetal outcomes at the earliest possible opportunity preconceptionally (e.g., at time of starting an ASM post-menarche) | B |
Recommendation 1 Rationale
- In utero ASM exposure may be associated with increased risk to fetus
- Discontinuing or changing ASMs also carries risks (seizure recurrence)
- Shared decision-making leads to more informed choices and better risk perception
- >65% of pregnancies in PWECP in the US are unintended
- The ASM regimen at time of conception is often the regimen used throughout pregnancy
Recommendation 2: Seizure Control During Pregnancy
Recommendation 2 Statements
| Statement | Recommendation | Level |
|---|
| 2A | Clinicians must minimize convulsive seizures (GTC and focal-to-bilateral tonic-clonic) in PWECP during pregnancy to minimize potential risks to the birth parent and fetus | A |
| 2B | Once PWECP is already pregnant, clinicians should exercise caution in removing or replacing an ASM effective for GTC/FBTC, even if not optimal for fetus (e.g., valproic acid) | B |
| 2C | Monitor ASM levels in PWECP throughout pregnancy as guided by individual ASM pharmacokinetics and patient clinical presentation | B |
| 2D | Adjust ASM dose during pregnancy in response to (1) decreasing serum ASM levels or (2) worsening seizure control | B |
| 2E | Counsel PWECP using understudied ASMs (acetazolamide, eslicarbazepine, ethosuximide, lacosamide, nitrazepam, perampanel, piracetam, pregabalin, rufinamide, stiripentol, tiagabine, vigabatrin) that limited data exist | B |
Recommendation 2 Rationale
- Mortality: Odds of mortality during pregnancy 5–12× greater in PWECP vs pregnant people without epilepsy
- Danish cohort (>2 million pregnancies) and US cohort (>20 million participants) confirm this elevated mortality risk
- 202 pregnancy-related deaths in epilepsy in the UK (2013–2015); most were from sudden unexpected death; 5 of 13 had stopped ASMs during pregnancy
- All 13 epilepsy-related deaths had pre-pregnancy uncontrolled seizures
- EURAP data: No statistical association between seizures and spontaneous abortion/stillbirth, but 1 stillbirth followed convulsive status epilepticus
- Frequency of GTC/FBTC seizures may be a risk factor for lower IQ in offspring
- Serum concentrations of lamotrigine and levetiracetam decrease during pregnancy → may need dose adjustment
- Valproic acid is one of the most effective ASMs for idiopathic generalized epilepsy; switching during pregnancy doubles seizure risk (EURAP data)
🔹 Clinical Pearl
Never abruptly stop valproic acid during pregnancy in a patient with GTC/FBTC seizures. EURAP data show that removing VPA and replacing with another ASM during pregnancy doubled the risk of GTC/FBTC seizures. Preconceptional optimization is key — once pregnant, exercise caution before changing effective therapy.
Recommendation 3: Major Congenital Malformations (MCMs)
Background
- Unadjusted birth prevalence of any MCM in the general population: approximately 2.4%–2.9%
- ASMs with sufficient data for reliable conclusions (>1,000 monotherapy exposures): lamotrigine, levetiracetam, oxcarbazepine, carbamazepine, valproic acid, phenobarbital, phenytoin, topiramate
- Lowest MCM prevalence (monotherapy): Levetiracetam (3.1%), oxcarbazepine (3.1%), lamotrigine (3.1%)
- Highest MCM prevalence (monotherapy): Valproic acid (9.7%)
Table: MCM Prevalence per 1,000 by ASM in Monotherapy
| ASM | Sample Size | MCM Prevalence per 1,000 (95% CI) | I² |
|---|
| Levetiracetam | 2,248 | 34.8 (19.5–54.3) | 77.8 |
| Oxcarbazepine | 1,036 | 31.3 (21.6–42.8) | 0.0 |
| Lamotrigine | 10,746 | 30.7 (25.4–36.4) | 49.4 |
| Gabapentin | 90 | 30.9 (5.5–76.1) | 0.0 |
| Clonazepam | 187 | 30.3 (7.4–67.8) | 26.5 |
| Clobazam | 64 | 31.3 (0.5–91.9) | 0 |
| Zonisamide | 116 | 39.2 (11.7–236.1) | 87.7 |
| Carbamazepine | 9,908 | 43.7 (35.7–52.6) | 69.6 |
| Topiramate | 748 | 44.5 (30.9–60.4) | 0.0 |
| Phenytoin | 1,604 | 51.3 (35.9–69.2) | 52.3 |
| Phenobarbital | 1,116 | 60.3 (47.1–75.0) | 0.0 |
| Valproic acid | 5,658 | 96.7 (80.4–114.2) | 67.0 |
| Primidone | 99 | 101.5 (50.4–167.7) | 0.0 |
MCM Prevalence Differences vs Reference (Valproic Acid)
| ASM (Monotherapy) | MCM per 1,000 | Difference vs VPA per 1,000 (95% CI) | Confidence |
|---|
| Carbamazepine | 43.7 | −53 (−71.9 to −34.1) significant | Moderate |
| Lamotrigine | 30.7 | −66 (−83.8 to −48.2) significant | Moderate |
| Levetiracetam | 34.8 | −61.9 (−86.2 to −37.6) significant | Moderate |
| Oxcarbazepine | 31.3 | −65.4 (−85.3 to −45.5) significant | Moderate |
| Phenobarbital | 60.3 | −36.4 (−58.3 to −14.5) significant | Low |
| Phenytoin | 51.3 | −45.4 (−69.1 to −21.7) significant | Low |
| Topiramate | 44.5 | −52.2 (−74.6 to −29.8) significant | Moderate |
Specific MCM Types by ASM (Monotherapy, Prevalence per 1,000)
| MCM Type | Highest-Risk ASM | Prevalence per 1,000 (95% CI) | Notes |
|---|
| Neural tube defects | Valproic acid | 14.3 (9.5–20.1) | Reference; highest NTD rate of all ASMs |
| Neural tube defects | Carbamazepine | 5.6 (2.6–9.7) | PD −8.7 (−15.1 to −2.3); moderate confidence |
| Neural tube defects | Lamotrigine | 3.4 (0.4–9.2) | PD −11.0 (−17.8 to −4.1); moderate confidence |
| Neural tube defects | Levetiracetam | 3.1 (0.2–9.3) | PD −11.3 (−18.3 to −4.2); moderate confidence |
| Cardiac malformations | Phenobarbital | 41.9 (25.1–62.7) | Reference; highest cardiac MCM rate |
| Cardiac malformations | Lamotrigine | 16.6 (7.8–28.5) | PD −25.3 (−46.8 to −3.8); moderate confidence |
| Cardiac malformations | Levetiracetam | 12.5 (0.1–53.4) | PD −29.4 (−62.0 to 3.2); low confidence |
| Cardiac malformations | Valproic acid | 25.1 (16.9–35.0) | PD −16.8 (−37.6 to 4.1); low confidence |
| Oral/cleft palate | Phenobarbital | 22.3 (7.1–45.6) | Reference; highest rate |
| Oral/cleft palate | Topiramate | 14.1 (7.3–23.1) | PD −8.2 (−29.0 to 12.6); low confidence |
| Oral/cleft palate | Valproic acid | 8.0 (4.6–12.2) | PD −14.3 (−34.0 to 5.3); low confidence |
| Urogenital | Valproic acid | 12.4 (7.4–18.8) | Reference; highest urogenital MCM rate |
| Renal | Valproic acid | 13.7 (8.6–19.9) | Reference; highest renal MCM rate |
Recommendation 3 Statements
| Statement | Recommendation | Level |
|---|
| 3A | Counsel PWECP that birth prevalence of any MCM in the general population is approximately 2.4%–2.9%, providing a comparison framework for individual risk | A |
| 3B | Must consider using lamotrigine, levetiracetam, or oxcarbazepine in PWECP when appropriate based on epilepsy syndrome, likelihood of seizure control, and comorbidities to minimize MCM risk | A |
| 3C | Must avoid valproic acid in PWECP to minimize risk of MCMs (composite) or NTDs, if clinically feasible | A |
| 3D | Must counsel PWECP treated with or considering valproic acid that the risk of any MCM is the highest with VPA compared with other studied ASMs | A |
| 3E | To reduce cardiac malformation risk, must avoid phenobarbital in PWECP, if clinically feasible | A |
| 3F | To reduce oral cleft risk, clinicians should avoid phenobarbital and topiramate in PWECP, if clinically feasible | B |
| 3G | To reduce urogenital and renal malformation risk, should avoid valproic acid in PWECP, if clinically feasible | B |
| 3H | Obstetricians should recommend fetal screening for MCMs (e.g., detailed anatomical ultrasound) for PWECP treated with any ASM during pregnancy | B |
| 3I | Obstetricians should recommend screening cardiac investigations of the fetus among PWECP treated with phenobarbital during pregnancy | B |
🔹 Clinical Pearl
For board exams, know the “Big Three Safe ASMs” in pregnancy: lamotrigine, levetiracetam, oxcarbazepine — all have MCM rates of ~3.1% (close to the general population baseline of 2.4–2.9%). Valproic acid at 9.7% is ~3× the background rate. Phenobarbital is the worst for cardiac malformations (4.4%). VPA is the worst for NTDs (1.4%), urogenital (1.2%), and renal (1.4%) malformations.
Dose-Dependent MCM Risk
Key Findings
- Pre-planned analyses from external comparisons did not reach sufficient evidence for a standalone dose recommendation
- The only Class I study addressing dose-response was EURAP, which demonstrated a dose effect for:
- Carbamazepine
- Lamotrigine
- Phenobarbital
- Valproic acid
- A statistically and clinically important difference in MCM prevalence was found for valproic acid and phenobarbital between high- and low-dose exposures
- Reasonable practice: Use the lowest appropriate dose of ASMs in PWECP to reduce MCM risk, if clinically feasible
Polytherapy vs Monotherapy
| ASM | Mono MCM/1,000 (95% CI) | Poly MCM/1,000 (95% CI) | PD Mono vs Poly (95% CI) |
|---|
| Carbamazepine | 43.7 (35.7–52.6) | 58.6 (38.8–82.1) | −14.9 (−38.1 to 8.3); low confidence |
| Lamotrigine | 30.7 (25.4–36.4) | 44.6 (34.1–56.5) | −13.9 (−26.4 to −1.4); low confidence |
| Levetiracetam | 34.8 (19.5–54.3) | 64.5 (30.1–110.8) | −29.7 (−73.7 to 14.2); low confidence |
| Oxcarbazepine | 31.3 (21.6–42.8) | 48.9 (26.2–78.2) | −17.6 (−45.7 to 10.5); low confidence |
| Phenobarbital | 60.3 (47.1–75.0) | 43.4 (24.4–67.5) | +16.9 (−8.8 to 42.6); low confidence |
| Phenytoin | 51.3 (35.9–69.2) | 38.0 (19.8–61.7) | +13.3 (−13.4 to 40.1); low confidence |
| Valproic acid | 96.7 (80.4–114.2) | 101.7 (81.0–124.5) | −5.1 (−32.6 to 22.5); low confidence |
🔹 Clinical Pearl
There is no clear evidence that polytherapy is worse than monotherapy for MCMs when comparing the same ASMs — most differences are small with wide confidence intervals. The key driver of MCM risk is the specific ASM used (especially valproic acid) rather than mono- vs polytherapy per se.
Recommendation 4: Perinatal Outcomes
Intrauterine Death
- Prevalence of intrauterine death is highly likely not to differ across ASMs in monotherapy
- Risk of intrauterine death is likely higher with polytherapy vs monotherapy
Prematurity
- Prevalence of prematurity is possibly no different across ASMs used in monotherapy
Small for Gestational Age (SGA) — Table 7 Data
| ASM (Monotherapy) | Sample Size | SGA Prevalence per 1,000 (95% CI) | Difference vs Reference (95% CI) |
|---|
| Phenytoin | 464 | 14.4 (2.7–35.1) | −65.8 (−206.3 to 74.8); low confidence |
| Zonisamide | 125 | 20.4 (3.1–52.4) | −59.7 (−201.6 to 82.1); low confidence |
| Levetiracetam | 835 | 52.9 (6.8–138.6) | −27.3 (−181.7 to 127.1); low confidence |
| Oxcarbazepine | 1,045 | 58.0 (6.8–154.2) | −22.2 (−180.1 to 135.7); low confidence |
| Gabapentin | 225 | 58.5 (0.1–214.2) | −21.7 (−197.7 to 154.3); low confidence |
| Carbamazepine | 3,033 | 75.7 (31.3–137.5) | −4.4 (−153.9 to 145.0); low confidence |
| Topiramate | 453 | 80.2 (0.3–279.6) | Reference |
| Lamotrigine | 2,597 | 85.1 (13.6–209.6) | 5.0 (−165.7 to 175.6); low confidence |
| Phenobarbital | 274 | 89.3 (0.3–310.0) | 9.1 (−199.4 to 217.6); low confidence |
| Primidone | 20 | 166.0 (40.7–352.9) | 85.8 (−123.6 to 295.2); very low confidence |
| Valproic acid | 1,829 | 147.1 (53.9–276.0) | 66.9 (−111.5 to 245.4); low confidence |
| Clobazam | 30 | 177.1 (64.6–329.9) | 96.9 (−95.7 to 289.6); very low confidence |
| Clonazepam | 276 | 165.4 (123.0–212.7) | 85.2 (−61.5 to 231.9); low confidence |
Recommendation 4 Statements
| Statement | Recommendation | Level |
|---|
| 4A | Counsel PWECP that prevalence of intrauterine death does not differ among different ASM exposures in monotherapy | B |
| 4B | Clinicians should avoid valproic acid or topiramate in PWECP to minimize risk of offspring being born SGA, if clinically feasible | B |
| 4C | To enable early identification of fetal growth restriction, obstetricians should recommend screening fetal growth throughout pregnancy among PWECP treated with valproic acid or topiramate | B |
- SGA prevalence is possibly greater after exposure to valproic acid or topiramate compared with lamotrigine
- Fetal growth restriction increases risk of perinatal morbidity and mortality
- Prenatal identification of SGA leads to improved perinatal outcomes by informing timely delivery
Recommendation 5: Neurodevelopmental Outcomes
Full-Scale IQ (Table 4 — Global IQ by ASM Monotherapy)
| ASM | Sample Size | Mean Global IQ (95% CI) | RMD vs VPA Reference (95% CI) | Confidence |
|---|
| Lamotrigine | 129 | 105.8 (100.9–110.6) | +11.85 (5.53–18.15) | Moderate (upgraded for magnitude) |
| Phenytoin | 76 | 103.2 (93.0–113.4) | +9.29 (−1.63 to 20.21) | Very low |
| Carbamazepine | 316 | 100.4 (95.8–105.1) | +6.53 (0.39–12.67) | Low |
| Topiramate | 27 | 100.5 (95.8–105.2) | +6.58 (0.37–12.80) | Very low |
| Levetiracetam | 42 | 99.0 (95.0–103.0) | +6.3 (0.9–11.7) | Very low |
| Valproic acid | 173 | 93.9 (89.1–97.9) | Reference | Reference |
Verbal IQ (Table 5)
| ASM | Sample Size | Mean Verbal IQ (95% CI) | RMD vs VPA (95% CI) | Confidence |
|---|
| Lamotrigine | 103 | 102.4 (96.5–108.2) | +10.3 (2.4–18.2) | Moderate (upgraded) |
| Phenytoin | 61 | 103.0 (95.8–110.2) | +10.9 (2.0–19.8) | Moderate (upgraded) |
| Levetiracetam | 42 | 101.0 (97.7–104.3) | +8.9 (2.7–15.1) | Very low |
| Carbamazepine | 283 | 98.4 (94.6–102.2) | +6.3 (−0.2 to 12.8) | Low |
| Topiramate | 27 | 99.2 (95.2–103.2) | +7.1 (0.5–13.7) | Very low |
| Valproic acid | 160 | 92.1 (86.9–97.4) | Reference | Reference |
Non-Verbal IQ (Table 5)
| ASM | Sample Size | Mean Non-Verbal IQ (95% CI) | RMD vs VPA (95% CI) | Confidence |
|---|
| Phenytoin | 40 | 106.0 (103.1–109.0) | +4.8 (0.1–8.7) | Very low |
| Lamotrigine | 103 | 105.8 (100.9–110.7) | +4.6 (−0.8 to 10.1) | Low |
| Carbamazepine | 197 | 104.7 (102.2–107.3) | +3.6 (0.0–7.1) | Low |
| Topiramate | 27 | 102.4 (97.1–107.7) | +1.2 (−4.6 to 7.1) | Very low |
| Levetiracetam | 42 | 99.6 (95.5–103.7) | −1.6 (−6.3 to 3.2) | Very low |
| Valproic acid | 96 | 101.2 (98.7–103.6) | Reference | Reference |
Autism Spectrum Disorder / Autistic Traits (Table 6)
| ASM (Monotherapy) | Sample Size | ASD/Autistic Traits per 1,000 (95% CI) | Difference vs VPA (95% CI) | Confidence |
|---|
| Levetiracetam | 1,226 | 11.3 (2.9–25.1) | −30.6 (−45.4 to −15.8) | Moderate (upgraded for magnitude) |
| Lamotrigine | 7,568 | 14.5 (8.6–22.2) | −27.4 (−39.3 to −15.6) | Moderate (upgraded for magnitude) |
| Carbamazepine | 4,493 | 17.1 (6.2–33.1) | −24.9 (−41.5 to −8.2) | Moderate (upgraded for magnitude) |
| Clonazepam | 587 | 20.8 (7.5–40.7) | −21.1 (−40.4 to −1.8) | Moderate (upgraded for magnitude) |
| Oxcarbazepine | 321 | 23.3 (9.7–42.6) | −18.6 (−37.8 to 0.5) | Low |
| Valproic acid | 3,399 | 41.9 (32.7–52.3) | Reference | Reference |
Recommendation 5 Statements
| Statement | Recommendation | Level |
|---|
| 5A | To reduce risk of poor neurodevelopmental outcomes, including ASD and lower IQ, clinicians must avoid valproic acid in PWECP, if clinically feasible | A |
| 5B | Must counsel PWECP treated with or considering valproic acid that in utero exposure is likely or possibly associated with a decrease in full-scale, verbal, and non-verbal IQ compared with other studied ASMs (CBZ, GBP, LTG, LEV, PHT, TPM) | A |
| 5C | Must counsel PWECP treated with or considering valproic acid that in utero exposure is possibly associated with an increased risk of ASD compared with other studied ASMs (CBZ, CZP, LEV, LTG) | A |
| 5D | Clinicians should implement age-appropriate developmental screening in children exposed to any ASM in utero born to PWECP | B |
Key Neurodevelopmental Summary
- Valproic acid: Lowest mean IQ (93.9) — likely associated with decreased IQ at age 6 compared with gabapentin and lamotrigine; possibly decreased vs carbamazepine, levetiracetam, and topiramate
- Lamotrigine: Highest mean IQ (105.8); RMD vs VPA = +11.85 points (moderate confidence, upgraded for large magnitude of effect)
- ASD risk: VPA has the highest ASD/autistic trait prevalence (41.9/1,000); levetiracetam (11.3/1,000), lamotrigine (14.5/1,000), carbamazepine (17.1/1,000), and clonazepam (20.8/1,000) all show statistically significantly lower rates
- Topiramate concern: Although pre-planned analysis was insufficient, the SCAN-AED study found even higher ASD and intellectual disability prevalence with topiramate than VPA (aHRs: 2.8 [95% CI 1.4–5.7] for ASD and 3.5 [95% CI 1.4–8.6] for intellectual disability)
🔹 Clinical Pearl
Valproic acid exposure in utero → ~12-point IQ deficit compared with lamotrigine (93.9 vs 105.8; RMD +11.85, moderate confidence). It is also associated with the highest ASD risk (41.9/1,000) — nearly 3× that of lamotrigine (14.5/1,000). These neurodevelopmental effects are the strongest argument for avoiding VPA in PWECP, even more than the MCM data.
Recommendation 6: Folic Acid Supplementation
Recommendation 6 Rationale
- Optimal dosing and timing of folic acid supplementation in PWECP are unknown
- No demonstrated benefit of folic acid supplementation (≥0.4 mg/d) specifically for MCM prevention in PWECP
- However, randomized trials in the general population (pre-US folic acid fortification) demonstrated that periconceptional multivitamin supplementation reduces NTDs
- Systematic review of 14 studies: folic acid supplementation of 0.2 mg/d (on top of US fortification levels) would reduce NTDs by 23%
- Protective effect greatest in those with initial low serum folate concentrations
- Neurodevelopmental benefits:
- Preconception folic acid → possibly associated with better neurodevelopmental outcomes in children born to PWECP
- Folic acid ≥0.4 mg/d → possibly associated with reduced autistic traits at 3 years (OR 7.9, 95% CI 2.5–24.9)
- Likely associated with higher global IQ (average 6 points) at age 6 in children born to PWECP exposed to ASMs in utero
- Caution with high-dose folic acid:
- Periconceptional folic acid >1 mg/d was associated with 0.9% absolute increase in childhood cancer before age 20 (HR 2.7, 95% CI 1.2–6.3)
- Sub-analysis restricted to maternal epilepsy + folic acid <3 mg/d was not significant (aHR 2.6, 95% CI 1.0–6.9)
- Very high maternal serum folate (≥60.3 nmol/L) at birth → 2.5× increased ASD risk (95% CI 1.3–4.6)
- Adherence to folic acid supplementation is generally poor among PWECP; ASM polytherapy further decreases adherence
- No high-dose folic acid formulation exists in the US → higher doses require multiple tablets, reducing adherence
Recommendation 6 Statements
| Statement | Recommendation | Level |
|---|
| 6A | Clinicians should prescribe at least 0.4 mg of folic acid daily preconceptionally and during pregnancy to any PWECP treated with an ASM to decrease the risk of NTDs in the offspring | B |
| 6B | Clinicians must prescribe at least 0.4 mg of folic acid daily preconceptionally and during pregnancy to any PWECP treated with an ASM to possibly improve neurodevelopmental outcomes such as ASD and global IQ in the offspring | A |
| 6C | Clinicians should counsel PWECP treated with an ASM that adherence to recommended folic acid supplementation preconceptionally and during pregnancy is important to minimize MCM and neurodevelopmental risk | B |
🔹 Clinical Pearl
The recommended folic acid dose is ≥0.4 mg/d (not the commonly cited 4–5 mg/d from older guidelines). This guideline found no demonstrated MCM reduction from folic acid in PWECP specifically, but Level A evidence supports it for neurodevelopmental benefit (higher IQ, reduced ASD risk). Paradoxically, very high folate levels may increase ASD risk — balance is key. Data do not support a dose recommendation beyond 0.4 mg/d.
Drug-by-Drug Quick Reference
Comprehensive ASM Pregnancy Risk Summary
| ASM | Any MCM /1,000 | Key Specific MCM Risks | IQ (Mean) | ASD /1,000 | SGA /1,000 | Guideline Action |
|---|
| Lamotrigine | 30.7 | NTD 3.4; Cardiac 16.6; Oral cleft 4.6 | 105.8 | 14.5 | 85.1 | Consider (Level A) |
| Levetiracetam | 34.8 | NTD 3.1; Cardiac 12.5; Urogenital 1.0 | 99.0 | 11.3 | 52.9 | Consider (Level A) |
| Oxcarbazepine | 31.3 | NTD 3.5; cardiac data limited | N/A | 23.3 | 58.0 | Consider (Level A) |
| Carbamazepine | 43.7 | NTD 5.6; Cardiac 8.5; Oral cleft 4.7 | 100.4 | 17.1 | 75.7 | Alternative option |
| Phenytoin | 51.3 | NTD 2.0; Cardiac 19.9; Oral cleft 9.7 | 103.2 | N/A | 14.4 | Use with caution |
| Phenobarbital | 60.3 | NTD 4.1; Cardiac 41.9; Oral cleft 22.3 | N/A | N/A | 89.3 | Must avoid for cardiac (Level A) |
| Topiramate | 44.5 | NTD 1.3; Oral cleft 14.1 | 100.5 | N/A | 80.2 | Avoid for oral clefts + SGA (Level B) |
| Valproic acid | 96.7 | NTD 14.3; Cardiac 25.1; Urogenital 12.4; Renal 13.7 | 93.9 | 41.9 | 147.1 | Must avoid (Level A) |
| Gabapentin | 30.9 | Limited data | N/A | N/A | 58.5 | Insufficient data |
| Zonisamide | 39.2 | Limited data | N/A | N/A | 20.4 | Insufficient data |
| Clobazam | 31.3 | Very limited data (n = 64) | N/A | N/A | 177.1 | Insufficient data |
| Clonazepam | 30.3 | Limited data | N/A | 20.8 | 165.4 | Insufficient data |
| Primidone | 101.5 | Oral cleft 16.6 | N/A | N/A | 166.0 | Very limited data |
Evidence Classification & GRADE Methodology
AAN Evidence Classification
- Class I: Lowest risk of bias (e.g., well-designed prospective registries with ≥80% follow-up)
- Class II: Moderate risk of bias
- Class III: Higher risk of bias (retrospective, incomplete follow-up, limited controls)
- Class IV: Highest risk of bias — excluded from this guideline (n = 133 articles)
Modified GRADE Confidence Levels
| Starting Confidence | Criteria |
|---|
| Low (starting point for most) | ≥2 Class III studies or ≥1 Class I/II study informing comparisons |
| Very Low | Single Class III study informing the estimate |
Upgrading/Downgrading Criteria
- Upgraded 1 level for large magnitude of effect: MCM/ASD PD >100/1,000 live births or PR >2 or <0.5; IQ RMD >10 points
- Upgraded 2 levels for very large magnitude: MCM/ASD PD >300/1,000 or PR >10 or <0.1; IQ RMD >20 points
- Downgraded for imprecision: 95% CI width >100/1,000 for MCMs or >300/1,000 for SGA
- Downgraded for indirectness: All comparisons are indirect (across different studies/registries)
- Perinatal outcomes (adjusted PR): Downgraded if CI width >2
Recommendation Levels (Actionable Statements)
| Level | Language | Meaning |
|---|
| A | “Must” / “Must not” | Strong recommendation; high certainty of benefit or harm |
| B | “Should” / “Should not” | Moderate recommendation; moderate certainty |
| C | “May” | Weak recommendation; low certainty |
Summary of All Recommendations by Evidence Level
Level A (“Must”) Recommendations
| # | Recommendation |
|---|
| 2A | Must minimize convulsive seizures (GTC/FBTC) in PWECP during pregnancy |
| 3A | Must counsel PWECP that general population MCM rate is 2.4%–2.9% as a comparison framework |
| 3B | Must consider lamotrigine, levetiracetam, or oxcarbazepine in PWECP to minimize MCM risk |
| 3C | Must avoid valproic acid in PWECP to minimize MCMs/NTDs |
| 3D | Must counsel that valproic acid has the highest MCM risk of all studied ASMs |
| 3E | Must avoid phenobarbital in PWECP to reduce cardiac malformation risk |
| 5A | Must avoid valproic acid in PWECP to reduce risk of poor neurodevelopmental outcomes (ASD, lower IQ) |
| 5B | Must counsel PWECP that VPA exposure is likely/possibly associated with decreased IQ vs other ASMs |
| 5C | Must counsel PWECP that VPA exposure is possibly associated with increased ASD risk vs other ASMs |
| 6B | Must prescribe ≥0.4 mg folic acid daily to possibly improve neurodevelopmental outcomes (ASD, IQ) |
Level B (“Should”) Recommendations
| # | Recommendation |
|---|
| 1A | Should engage in joint decision-making with PWECP |
| 1B | Should recommend optimal ASMs/doses preconceptionally at earliest opportunity |
| 2B | Should exercise caution in removing/replacing ASM effective for GTC/FBTC once pregnant |
| 2C | Should monitor ASM levels throughout pregnancy |
| 2D | Should adjust ASM dose in response to decreasing levels or worsening seizure control |
| 2E | Should counsel that limited data exist for understudied ASMs |
| 3F | Should avoid phenobarbital and topiramate to reduce oral cleft risk |
| 3G | Should avoid valproic acid to reduce urogenital and renal malformation risk |
| 3H | Should recommend fetal screening (anatomical ultrasound) for PWECP on any ASM |
| 3I | Should recommend cardiac screening of fetus for PWECP on phenobarbital |
| 4A | Should counsel that intrauterine death prevalence does not differ among ASMs in monotherapy |
| 4B | Should avoid valproic acid or topiramate to minimize SGA risk |
| 4C | Should recommend fetal growth screening for PWECP on VPA or topiramate |
| 5D | Should implement age-appropriate developmental screening in children exposed to any ASM in utero |
| 6A | Should prescribe ≥0.4 mg folic acid daily to decrease NTD risk |
| 6C | Should counsel on importance of folic acid adherence |
Limitations & Future Research
Key Limitations
- All comparisons between ASMs are indirect (different registries, populations, time periods) → inherent imprecision
- Confounders (genetics, seizure type, socioeconomic status, pregnancy conditions) could not always be adjusted
- Class I evidence was weighted more heavily, but most data comes from Class II–III observational registries
- Many newer ASMs have no or very limited data: acetazolamide, brivaracetam, eslicarbazepine, lacosamide, nitrazepam, perampanel, piracetam, pregabalin, rufinamide, stiripentol, tiagabine, vigabatrin
- Dose-response analyses could not generate standalone recommendations despite EURAP data showing dose effects
Future Research Priorities
- Pregnancy outcomes for newer ASMs (cenobamate, fenfluramine, lacosamide, zonisamide, clobazam, perampanel)
- Longitudinal neurodevelopmental outcomes beyond valproic acid comparisons
- Randomized controlled trials for optimal folic acid dose and timing
- Impact of socioeconomic status, ethnic/racial diversity on outcomes
- Pharmacokinetics and pharmacodynamics of ASMs during pregnancy and postpartum
- Uniform definitions for high vs low ASM doses
- Multi-variable analyses to untangle ASM effects from epilepsy syndrome effects
Key Registries Referenced
Major Pregnancy Registries in Evidence Base
| Registry | Details |
|---|
| EURAP | European Registry of Antiepileptic Drugs and Pregnancy — prospective, multinational; the only Class I study addressing dose-response for MCMs |
| UKIEPR | UK and Ireland Epilepsy and Pregnancy Register |
| NAAPR | North American AED Pregnancy Registry |
| KREP | Kerala Registry of Epilepsy and Pregnancy (India) |
| NEAD | Neurodevelopmental Effects of Antiepileptic Drugs study — Class I prospective observational study for IQ outcomes |
| MONEAD | Maternal Outcomes and Neurodevelopmental Effects of Antiepileptic Drugs study |
| SCAN-AED | Scandinavian study — found higher ASD/intellectual disability with topiramate than VPA |
