Epilepsy in pregnancy requires balancing two competing imperatives: controlling maternal seizures, which themselves pose fetal risk through hypoxia and trauma, and minimizing fetal exposure to teratogenic anti-seizure drugs (ASDs). Neither objective can be sacrificed for the other, and the goal is to use the most effective ASD at the lowest effective dose in a woman who has received preconception counseling.
Uncontrolled seizures during pregnancy carry independent risks to the fetus: generalized tonic-clonic seizures (GTCSs) cause maternal hypoxia, acidosis, and physical trauma that can result in fetal heart rate decelerations and, rarely, fetal demise. Status epilepticus (SE) in pregnancy carries maternal and fetal mortality risks that substantially exceed those of any teratogenic ASD. These considerations mean that abrupt discontinuation of ASDs at conception – a common but dangerous patient behavior – should be actively discouraged. Women with well-controlled epilepsy on a teratogenically acceptable ASD should continue that drug through pregnancy with appropriate monitoring rather than switching regimens in the first trimester, when organogenesis is already underway.1
The EURAP (European and International Registry of Antiepileptic Drugs and Pregnancy) registry is the largest prospective dataset on ASD teratogenicity, covering over 100,000 pregnancy exposures across multiple countries. Its most clinically significant finding is the dose-dependent, agent-specific major congenital malformation (MCM) rate hierarchy. Valproate carries the highest MCM risk of any commonly used ASD at standard doses (approximately 10% at doses above 1,500 mg/day), with neural tube defects (NTDs), cardiac defects, and hypospadias predominating. Phenobarbital (PB) and phenytoin (PHT) carry MCM rates of approximately 6–7%. Carbamazepine (CBZ) at standard doses has an MCM rate of approximately 5%, with a specific NTD risk of approximately 0.5–1%. Lamotrigine (LTG) and levetiracetam (LEV) have the most favorable profiles among commonly used ASDs, with MCM rates of approximately 2–3% at standard doses – close to the general population baseline of approximately 2%. The data for newer agents (brivaracetam, cenobamate, perampanel) remain limited, and these should generally be avoided in pregnancy unless alternatives are inadequate.1
Valproate's teratogenic profile warrants special emphasis because it is unique in causing both structural malformations and neurodevelopmental harm. The Neurodevelopmental Effects of Antiepileptic Drugs (NEAD) study demonstrated that children exposed to valproate in utero scored 6–9 points lower on IQ testing at age 6 compared to children exposed to other ASDs, with dose-dependent effects present even below 800 mg/day and even in pregnancies without structural defects. Autistic spectrum disorder and attention deficit hyperactivity disorder (ADHD) occur at significantly higher rates in valproate-exposed children. Because no dose of valproate is known to be neurologically safe for the developing brain, and because this neurodevelopmental harm cannot be mitigated by any intervention after exposure, valproate should be avoided in women of reproductive potential unless no alternative provides adequate seizure control.2
Folic acid supplementation is recommended for all women with epilepsy who could become pregnant. The dose recommended by most epilepsy guidelines is 5 mg/day, substantially higher than the 400 mcg/day recommended for the general obstetric population, because enzyme-inducing ASDs (carbamazepine, phenytoin, phenobarbital) reduce folate levels through induction of folate-metabolizing enzymes. Folic acid supplementation reduces the background risk of NTDs and is beneficial regardless of which ASD is used, but it does not eliminate the excess NTD risk associated with valproate or carbamazepine at high doses. Supplementation should ideally begin at least one month before conception, as neural tube closure occurs by week 4 of gestation – before many pregnancies are confirmed.3
Lamotrigine pharmacokinetics change dramatically during pregnancy because estrogen-driven upregulation of uridine diphosphate glucuronosyltransferase 1A4 (UGT1A4) – the primary enzyme responsible for lamotrigine glucuronidation – accelerates lamotrigine clearance. Lamotrigine clearance increases by 40–65% during pregnancy, causing plasma levels to fall progressively across all three trimesters. This pharmacokinetic change has direct clinical consequences: women whose seizures were well controlled on a stable lamotrigine dose before pregnancy may experience seizure breakthrough in the second or third trimester solely due to declining drug levels, not disease progression. Therapeutic drug monitoring (TDM) of lamotrigine levels is recommended monthly during pregnancy, with dose increases as needed to maintain the pre-pregnancy baseline level. After delivery, clearance returns to pre-pregnancy rates within days to weeks, requiring rapid dose reduction to avoid lamotrigine toxicity in the postpartum period.4
Valproate should not be prescribed to women of reproductive potential unless they are enrolled in a risk-minimization program (REMS in the U.S.; Prevent in Europe), have confirmed that they understand the teratogenic and neurodevelopmental risks, and are using effective contraception. If valproate must be used (because no alternative controls the patient's seizures), the lowest effective dose should be employed, 5 mg/day of folic acid prescribed, and pregnancy planning discussed at every visit. The neurodevelopmental harm from valproate – 6–9 IQ point reduction, increased autism and ADHD risk – is not preventable by folic acid supplementation and occurs even without structural malformations.
Epilepsy is the third most common neurological condition in adults over 65, and its incidence rises steeply with age as cerebrovascular disease, neurodegenerative conditions, and brain tumors become more prevalent. Prescribing ASDs in older adults is complicated by age-related pharmacokinetic changes, high baseline rates of polypharmacy, narrowed therapeutic windows, and heightened sensitivity to CNS adverse effects.
Multiple pharmacokinetic parameters change with aging in ways that alter ASD exposure and tolerability. Renal function declines progressively from approximately age 30 onward at a rate of roughly 1 mL/min/year, meaning that renally eliminated ASDs – levetiracetam (LEV), gabapentin, pregabalin, topiramate (partially) – accumulate in older adults unless doses are reduced. Hepatic blood flow and cytochrome P450 (CYP) enzyme activity decline with age, reducing the clearance of hepatically metabolized ASDs including carbamazepine, phenytoin, phenobarbital, and lamotrigine. Albumin levels fall with aging and illness, reducing protein binding of highly bound drugs (phenytoin is 90% protein-bound), which increases the free fraction and pharmacological activity even at unchanged total plasma concentrations – a phenomenon that causes phenytoin toxicity to appear at total concentrations previously tolerated. Body fat increases and total body water decreases with aging, altering the volume of distribution of lipophilic and hydrophilic drugs respectively. These combined changes require more conservative initial dosing, slower upward titration, and more frequent monitoring in elderly patients than in younger adults.5
Polypharmacy is nearly universal in elderly patients with epilepsy, and the interaction potential of the classic enzyme-inducing ASDs – carbamazepine, phenytoin, and phenobarbital – makes them particularly hazardous in this population. These drugs induce CYP3A4, CYP2C9, CYP2C19, and P-glycoprotein, reducing the efficacy of statins, anticoagulants (warfarin), calcium channel blockers, direct oral anticoagulants (DOACs), beta-blockers, and many other drugs commonly prescribed in older adults. Phenytoin additionally exhibits nonlinear (zero-order) pharmacokinetics at therapeutic concentrations, meaning that small dose increases can produce disproportionately large rises in plasma levels and rapid toxicity – particularly dangerous in elderly patients with reduced protein binding and altered renal and hepatic clearance. Carbamazepine can cause hyponatremia in elderly patients through syndrome of inappropriate antidiuretic hormone secretion (SIADH), which is more common in older adults and can be clinically severe.6
Falls and cognitive impairment are the two most clinically consequential adverse effects in elderly epilepsy patients, and ASD selection should actively minimize these risks. Sedating ASDs – particularly barbiturates (phenobarbital), benzodiazepines, and older agents like clonazepam – substantially increase fall risk in older adults, where falls cause hip fractures, subdural hematomas, and excess mortality. Phenytoin and carbamazepine cause dizziness, ataxia, and cognitive slowing at therapeutic levels. Lamotrigine, levetiracetam, and lacosamide have the most favorable cognitive and sedation profiles among effective ASDs, and one or more of these are preferred as first-line therapy in older adults with newly diagnosed focal epilepsy. The VA Cooperative Study confirmed that lamotrigine and gabapentin were better tolerated than carbamazepine in older adults with newly diagnosed epilepsy, with fewer treatment withdrawals due to adverse effects despite similar efficacy.7
For newly diagnosed focal epilepsy in older adults, current practice guidelines favor lamotrigine or levetiracetam as first-line agents. Lamotrigine has the advantage of established safety data, no renal dose adjustment requirement, and a favorable cognitive profile, but requires slow titration (8–12 weeks) and TDM during intercurrent illness or medication changes. Levetiracetam requires renal dose adjustment but has no pharmacokinetic drug interactions and can be initiated at therapeutic doses rapidly. Lacosamide is an increasingly preferred alternative, particularly in patients who cannot tolerate the slow lamotrigine titration or whose focal epilepsy requires IV administration. Enzyme-inducing ASDs (carbamazepine, phenytoin, phenobarbital) are generally avoided as first-line choices in elderly patients due to drug interaction burden, narrow therapeutic windows, cognitive adverse effects, and osteoporosis risk from accelerated vitamin D metabolism.5
Phenytoin is approximately 90% protein-bound, and total plasma phenytoin concentration reflects both bound and free drug. In elderly patients with hypoalbuminemia (common in illness, malnutrition, or chronic disease), the free fraction rises while total concentration may appear unchanged or even low. A patient with albumin of 2.0 g/dL can have phenytoin toxicity at a total level of 12 mcg/mL that would be well within the therapeutic range in a normal-albumin patient. The corrected phenytoin level should be calculated: Corrected phenytoin = measured total phenytoin ÷ [0.2 × albumin (g/dL) + 0.1]. Alternatively, free (unbound) phenytoin levels should be measured directly in hypoalbuminemic patients. This limitation, combined with nonlinear PK and high drug interaction burden, explains why phenytoin is rarely a rational first choice in older adults.
Pediatric epilepsy is not simply adult epilepsy in smaller patients. The dominant epilepsy syndromes of childhood are age-specific, genetically driven, and often have syndrome-specific treatment hierarchies that differ substantially from adult focal epilepsy management. Matching drug to syndrome is the central principle of pediatric epilepsy pharmacology.
Infantile spasms (IS), also called West syndrome, is an epileptic encephalopathy presenting in the first year of life with characteristic flexion-extension spasms, arrest of developmental milestones, and hypsarrhythmia on electroencephalogram (EEG). The two evidence-based first-line treatments are adrenocorticotropic hormone (ACTH) and vigabatrin. ACTH produces short-term spasm cessation in approximately 55–87% of patients; vigabatrin is the preferred first-line agent when the underlying etiology is tuberous sclerosis complex (TSC), where spasm cessation rates exceed 95%. The Child Neurology Society/American Epilepsy Society (AES) guidelines recommend ACTH or vigabatrin as first-line, with combination therapy for refractory IS. Vigabatrin causes irreversible visual field constriction (peripheral retinal toxicity) in approximately 30% of treated patients, making ophthalmologic monitoring every 3 months mandatory for all patients on vigabatrin. Standard ASDs such as carbamazepine and phenytoin are ineffective for infantile spasms and should not be used.8
Childhood absence epilepsy (CAE) is the most common pediatric epilepsy syndrome, characterized by frequent brief absence seizures typically beginning between ages 4 and 8. As established in the 2010 CAE randomized trial, ethosuximide is the preferred first-line agent for pure CAE without generalized tonic-clonic seizures (GTCSs), offering equivalent seizure control to valproate with superior attentional function in children. Valproate is preferred when GTCSs accompany absence seizures, because ethosuximide has no efficacy against tonic-clonic seizures. Lamotrigine is a second-line option for CAE when ethosuximide and valproate are not tolerated or are contraindicated, but produces lower seizure-freedom rates than either first-line agent. Most children with CAE achieve remission before or during adolescence, allowing ASD discontinuation in the majority of patients.9
Juvenile myoclonic epilepsy (JME) is a lifelong idiopathic generalized epilepsy (IGE) syndrome beginning in adolescence, characterized by myoclonic jerks on awakening, generalized tonic-clonic seizures, and, in many patients, absence seizures. Valproate is the most effective single agent for JME, controlling all three seizure types in the majority of patients; it is the treatment of choice in male patients and post-menopausal women. In females of reproductive potential, valproate is avoided due to teratogenicity, and lamotrigine or levetiracetam is used instead – accepting some reduction in efficacy, particularly against the myoclonic component. Lamotrigine has the important caveat that it can worsen myoclonic jerks in some JME patients, particularly at higher doses; this apparent paradox reflects its sodium channel mechanism's potential to alter thalamocortical firing in ways that aggravate myoclonus. Seizure remission is uncommon in JME and most patients require lifelong treatment, making initial ASD choice and tolerability particularly consequential.10
Lennox-Gastaut syndrome (LGS) is a severe childhood-onset epileptic encephalopathy characterized by multiple seizure types (tonic, atonic, atypical absence, myoclonic), intellectual disability, and a characteristic slow spike-and-wave pattern on EEG. LGS is refractory to most ASDs; no single agent produces sustained seizure freedom in the majority of patients. Valproate is the most frequently used backbone agent. Adjunctive agents with evidence in LGS include rufinamide (particularly for tonic and atonic seizures), clobazam, lamotrigine, and cannabidiol (CBD). Cannabidiol (Epidiolex) is FDA-approved for seizures associated with LGS and Dravet syndrome at 10–20 mg/kg/day and has demonstrated approximately 43% reduction in drop attacks (tonic and atonic seizures) versus placebo in pivotal trials.11 Fenfluramine is also approved for LGS as of 2022. Felbamate has efficacy in LGS but carries black box warnings for aplastic anemia and hepatic failure, limiting its use to refractory cases where benefit clearly outweighs risk.
Dravet syndrome is a severe developmental and epileptic encephalopathy caused in approximately 80% of cases by de novo pathogenic variants in SCN1A, the gene encoding the Nav1.1 sodium channel subunit expressed predominantly in GABAergic interneurons. Loss-of-function SCN1A variants impair inhibitory interneuron function, producing the characteristic treatment-resistant epilepsy with multiple seizure types triggered by fever, bathing, and infection. Sodium channel blockers – carbamazepine, phenytoin, lamotrigine, and oxcarbazepine – worsen seizures in Dravet syndrome by further reducing Nav1.1 activity and are strictly contraindicated. Valproate and clobazam are the primary treatment backbone. Stiripentol (a cytochrome P450 (CYP) inhibitor that also enhances GABA-A receptor function) is approved as adjunctive therapy in Europe and the U.S. for Dravet syndrome in patients already on valproate and clobazam. Cannabidiol (Epidiolex) is FDA-approved for Dravet syndrome and reduces seizure frequency by approximately 39% versus placebo. Fenfluramine (Fintepla) received FDA approval in 2020 for Dravet syndrome, reducing monthly convulsive seizures by approximately 63% versus placebo in pivotal trials, via a mechanism involving serotonin receptor modulation.12
Carbamazepine, phenytoin, lamotrigine, and oxcarbazepine are all contraindicated in Dravet syndrome. Because Nav1.1 loss-of-function impairs GABAergic inhibitory interneurons, further reduction in sodium channel activity by these drugs paradoxically worsens seizures and can precipitate status epilepticus. This is a critical prescribing safety rule: any child with an SCN1A-confirmed diagnosis or a clinical presentation consistent with Dravet syndrome must not receive sodium channel blocking ASDs. Genetic testing for SCN1A should be obtained early in any child with a febrile seizure pattern suggesting Dravet syndrome to guide ASD selection before sodium channel blockers are inadvertently prescribed.
Renal and hepatic impairment alter ASD pharmacokinetics through distinct and predictable mechanisms. Renal impairment reduces the elimination of drugs that are excreted unchanged in the urine; hepatic impairment reduces the metabolism of drugs that undergo first-pass or systemic hepatic biotransformation. Understanding which elimination pathway each ASD relies on is the foundation for rational dose adjustment in organ-impaired patients.
The ASDs that rely primarily or exclusively on renal elimination and therefore require dose reduction in renal impairment are levetiracetam, gabapentin, pregabalin, topiramate (approximately 70% renally excreted), and vigabatrin. Levetiracetam is approximately 66% eliminated by renal hydrolysis of its acetamide group (not by glomerular filtration of unchanged drug alone), making creatinine clearance (CrCl) the dosing parameter – manufacturers recommend dose reduction when CrCl falls below 80 mL/min, with substantial reductions below 50 mL/min and 30 mL/min thresholds. Gabapentin and pregabalin are eliminated entirely by renal filtration of unchanged drug, and their dose must be reduced proportionally with declining CrCl – both have established dosing tables based on CrCl ranges that should be followed precisely. Topiramate is approximately 70% renally excreted and requires dose reduction in moderate to severe renal impairment. Dialysis removes significant amounts of levetiracetam, gabapentin, and pregabalin, requiring supplemental dosing after hemodialysis sessions.13
Hepatically metabolized ASDs do not require routine dose adjustment in mild hepatic impairment but require caution in moderate impairment and are often contraindicated or require significant reduction in severe (Child-Pugh C) hepatic failure. Valproate is particularly hazardous in liver disease: it undergoes extensive hepatic metabolism through mitochondrial beta-oxidation and glucuronidation, and it is itself hepatotoxic – its use is contraindicated in patients with significant hepatic disease or a family history of severe hepatic dysfunction. Carbamazepine and phenytoin are extensively metabolized by cytochrome P450 (CYP) enzymes in the liver; in severe hepatic failure their clearance falls substantially, necessitating dose reduction and level monitoring. Lamotrigine is metabolized primarily by uridine diphosphate glucuronosyltransferase 1A4 (UGT1A4) glucuronidation in the liver – mild to moderate impairment reduces its clearance by approximately 25–50%, with more severe effects in severe impairment. Levetiracetam, though predominantly renally eliminated, undergoes partial hepatic hydrolysis and is generally the safest choice in patients with combined renal and hepatic impairment because dose adjustment is primarily driven by the well-characterized CrCl relationship.14
Practical monitoring principles in organ-impaired patients center on two strategies: therapeutic drug monitoring (TDM) of drug plasma levels, and clinical monitoring for toxicity and breakthrough seizures. TDM is particularly valuable for narrow-therapeutic-index drugs – phenytoin, carbamazepine, and valproate – and for any ASD in a patient whose organ function is changing (acute illness, progressive chronic kidney disease (CKD), post-transplant). For highly protein-bound drugs in patients with hypoalbuminemia (common in both hepatic and renal failure), total drug levels underestimate active drug exposure, and free drug levels should be measured where available. In renal failure, phenytoin protein binding is also reduced because uremic toxins compete for albumin binding sites, further increasing the free fraction. For renally eliminated ASDs in patients with CKD progressing to dialysis, dose schedules may need adjustment as residual renal function declines, and supplemental post-dialysis dosing protocols should be established proactively.14
Dose adjustment required with renal impairment (CrCl <60–80 mL/min): levetiracetam, gabapentin, pregabalin, topiramate (moderate–severe), vigabatrin, brivaracetam (minor via metabolites — monitor). Supplemental dosing after hemodialysis: levetiracetam, gabapentin, pregabalin.
No routine renal dose adjustment required: lamotrigine, carbamazepine, phenytoin, valproate, phenobarbital, lacosamide (mild impairment), perampanel. These rely on hepatic metabolism; renal function has minimal effect on clearance at typical CrCl levels.
Hepatic impairment — greatest caution: valproate (contraindicated in significant liver disease), carbamazepine, phenytoin, phenobarbital, lamotrigine. Levetiracetam is preferred when both renal and hepatic impairment are present.
The special-population considerations covered in this module converge on a core clinical discipline: the same seizure type in different patients may require different ASDs, because the optimal drug is determined as much by the patient's physiological context as by the seizure mechanism. The following frameworks synthesize the most consequential decision points.
In women with epilepsy of reproductive potential, the overriding constraint is teratogenic risk, and the decision tree begins before conception. Preconception counseling should cover: the teratogenic hierarchy (lamotrigine and levetiracetam lowest risk; valproate highest and should be avoided), the importance of 5 mg/day folic acid supplementation beginning before conception, the need to continue rather than abruptly stop ASDs at confirmation of pregnancy, and the pharmacokinetic changes that will necessitate more frequent monitoring during pregnancy. For women already on lamotrigine, establishing a baseline plasma level before conception and committing to monthly monitoring during pregnancy is mandatory. For women on valproate who wish to become pregnant, a proactive switch to a safer alternative ASD during a planned preconception period is strongly preferred over continuing valproate through pregnancy. Postpartum, dose reductions (particularly for lamotrigine) should be implemented proactively within days of delivery to match the rapid return of pre-pregnancy clearance rates.15
In elderly patients with new-onset epilepsy, the default position is to favor lamotrigine or levetiracetam as first-line agents and to avoid enzyme-inducing ASDs unless a specific clinical rationale favors them. Dosing should begin at 25–50% of the standard adult starting dose and be titrated more slowly than in younger adults, with the target plasma level range set at the lower end of the therapeutic window until individual response is established. Monitoring should include renal function at baseline and annually (to guide levetiracetam dosing), albumin (to interpret phenytoin levels if that drug must be used), and a medication reconciliation review at every visit to identify new pharmacokinetic interactions with newly added medications. Falls screening and gait assessment should accompany every neurology visit in elderly epilepsy patients, and any ASD that worsens gait, balance, or cognition should be reconsidered promptly.5
In pediatric patients, syndrome identification precedes drug selection, and the clinician must maintain awareness of the sodium channel blocker contraindication in Dravet syndrome, the vigabatrin visual toxicity monitoring requirement in infantile spasms, and the LGS drug landscape now including cannabidiol. Weight-based dosing is standard for all pediatric ASDs, and children's higher metabolic rates relative to body surface area often require higher mg/kg doses than adults to achieve equivalent plasma levels. ASDs that carry reproductive risk (valproate) should be discussed with parents well before a female child reaches reproductive age, with a plan to reassess the regimen at that transition. For idiopathic generalized epilepsy syndromes that typically remit (CAE), a trial of ASD discontinuation after 2–3 years of seizure freedom is appropriate; for syndromes that do not remit (JME, Dravet, LGS), this expectation should be communicated to families from the outset.10
In patients with organ impairment, the initial ASD selection should account for the elimination pathway: renally impaired patients should preferentially receive hepatically metabolized ASDs if seizure type permits, or renally cleared ASDs with precise CrCl-based dose adjustment. Levetiracetam is the most versatile choice in this context because its dose adjustment is transparent and predictable, it has no drug interactions, and it covers both focal and generalized seizure types. Valproate should be avoided in any patient with significant liver disease. In patients with both renal and hepatic impairment – as occurs in severely ill patients or those with hepatorenal syndrome – levetiracetam's dosing remains the most manageable, and phenytoin should be used with extreme caution, requiring free-level monitoring and frequent reassessment. For any ASD in an organ-impaired patient, the principle of starting low, titrating slowly, and monitoring frequently applies regardless of which drug is chosen.13
Pregnant women: lamotrigine or levetiracetam first; 5 mg/day folic acid; monthly lamotrigine TDM; avoid valproate; do not stop ASDs abruptly.
Elderly: lamotrigine or levetiracetam first; start low, titrate slowly; avoid enzyme inducers; monitor renal function, albumin, and falls risk.
Pediatrics: match drug to syndrome; ethosuximide for CAE; valproate for JME (males); vigabatrin for IS/TSC with ophthalmology monitoring; cannabidiol for LGS and Dravet; sodium channel blockers contraindicated in Dravet syndrome.
Renal impairment: dose-reduce levetiracetam, gabapentin, pregabalin, topiramate by CrCl; supplement after dialysis; prefer hepatically metabolized ASDs where possible.
Hepatic impairment: avoid valproate; reduce lamotrigine, carbamazepine, phenytoin doses; levetiracetam is the safest choice; monitor free drug levels in hypoalbuminemia.
Tomson T, Battino D, Bonizzoni E, et al. Comparative risk of major congenital malformations with eight different antiepileptic drugs: a prospective cohort study of the EURAP registry. Lancet Neurol. 2018;17(6):530–538.
doi:10.1016/S1474-4422(18)30107-8Meador KJ, Baker GA, Browning N, et al. Fetal antiepileptic drug exposure and cognitive outcomes at age 6 years (NEAD study). Lancet Neurol. 2013;12(3):244–252.
doi:10.1016/S1474-4422(12)70323-XMorrow JI, Hunt SJ, Russell AJ, et al. Folic acid use and major congenital malformations in offspring of women with epilepsy: a prospective study from the UK Epilepsy and Pregnancy Register. J Neurol Neurosurg Psychiatry. 2009;80(5):506–511.
doi:10.1136/jnnp.2008.156109Pennell PB, Newport DJ, Stowe ZN, Helmers SL, Montgomery JQ, Henry TR. The impact of pregnancy and childbirth on the metabolism of lamotrigine. Neurology. 2004;62(2):292–295.
doi:10.1212/01.WNL.0000103286.47129.F8Brodie MJ, Kwan P. Epilepsy in elderly people. BMJ. 2005;331(7528):1317–1322.
doi:10.1136/bmj.331.7528.1317Johannessen SI, Landmark CJ. Antiepileptic drug interactions: principles and clinical implications. Curr Neuropharmacol. 2010;8(3):254–267.
doi:10.2174/157015910792246254Rowan AJ, Ramsay RE, Collins JF, et al. New onset geriatric epilepsy: a randomized study of gabapentin, lamotrigine, and carbamazepine. Neurology. 2005;64(11):1868–1873.
doi:10.1212/01.WNL.0000167384.68207.3EGo CY, Mackay MT, Weiss SK, et al. Evidence-based guideline update: medical treatment of infantile spasms. Report of the Guideline Development Subcommittee of the American Academy of Neurology and the Practice Committee of the Child Neurology Society. Neurology. 2012;78(24):1974–1980.
doi:10.1212/WNL.0b013e318259e2cfGlauser TA, Cnaan A, Shinnar S, et al. Ethosuximide, valproic acid, and lamotrigine in childhood absence epilepsy. N Engl J Med. 2010;362(9):790–799.
doi:10.1056/NEJMoa0902014Marson AG, Al-Kharusi AM, Alwaidh M, et al. The SANAD study of effectiveness of valproate, lamotrigine, or topiramate for generalised and unclassifiable epilepsy. Lancet. 2007;369(9566):1016–1026.
doi:10.1016/S0140-6736(07)60461-9Devinsky O, Patel AD, Cross JH, et al. Effect of cannabidiol on drop seizures in the Lennox-Gastaut syndrome. N Engl J Med. 2018;378(20):1888–1897.
doi:10.1056/NEJMoa1714631Lagae L, Sullivan J, Knupp K, et al. Fenfluramine hydrochloride for the treatment of seizures in Dravet syndrome: a randomised, double-blind, placebo-controlled trial. Lancet. 2019;394(10216):2243–2254.
doi:10.1016/S0140-6736(19)32500-0Patsalos PN, Berry DJ, Bourgeois BF, et al. Antiepileptic drugs: best practice guidelines for therapeutic drug monitoring. Epilepsia. 2008;49(7):1239–1276.
doi:10.1111/j.1528-1167.2008.01561.xJohannessen Landmark C, Patsalos PN. Drug interactions involving the new second- and third-generation antiepileptic drugs. Expert Rev Neurother. 2010;10(1):119–140.
doi:10.1586/ern.09.136Harden CL, Meador KJ, Pennell PB, et al. Practice parameter update: management issues for women with epilepsy — focus on pregnancy. Neurology. 2009;73(2):142–149.
doi:10.1212/WNL.0b013e3181a6b325