Valproate (valproic acid) is one of the most pharmacologically versatile anti-seizure drugs (ASDs) in clinical use, with documented efficacy across virtually every seizure type and epilepsy syndrome. Its broad efficacy reflects multiple concurrent mechanisms of action rather than a single molecular target. These same pharmacological properties, combined with its substantial teratogenic risk and complex pharmacokinetics, require careful patient selection and longitudinal monitoring.
Valproate's antiseizure mechanism is genuinely multifactorial. At therapeutic concentrations, it produces use-dependent blockade of voltage-gated sodium channels in the inactivated state, an action mechanistically analogous to phenytoin and carbamazepine but weaker per molecule. It also enhances GABAergic inhibition through several routes: increased GABA synthesis (by stimulating glutamate decarboxylase), reduced GABA catabolism (by inhibiting GABA transaminase), and potentiation of postsynaptic GABA-A receptor responses. Its effect on absence seizures is mediated at least in part by inhibition of T-type calcium channels in thalamic neurons, suppressing the rhythmic low-threshold calcium spike that underlies thalamo-cortical synchronization in absence epilepsy. This mechanistic breadth is why valproate works across focal, generalized, and mixed epilepsy syndromes when many other agents with single mechanisms fail in some seizure types.1
Valproate's pharmacokinetics are governed by two important features that distinguish it from most ASDs. First, it is highly protein-bound (approximately 90–95% at low concentrations), primarily to albumin, and this binding saturates within the therapeutic range. As total valproate concentrations rise, the fraction of free (unbound, pharmacologically active) drug increases disproportionately, meaning that doubling the total concentration more than doubles the free concentration. Conventional TDM (therapeutic drug monitoring) measures total valproate, so free levels may be considerably higher than expected at elevated total concentrations, particularly in patients with hypoalbuminemia or those co-administered drugs that displace valproate from albumin.2 Second, valproate inhibits its own metabolism through inhibition of hepatic CYP2C9 and beta-oxidation enzymes, producing a non-linear relationship between dose and steady-state concentration that complicates titration.
Valproate's hepatotoxicity risk is age-stratified and enzyme-induction-dependent in a way that is critical for prescribers to understand. Fatal hepatotoxicity is rare in adults on valproate monotherapy (approximately 1 in 37,000) but dramatically more frequent in children under 2 years of age receiving polypharmacy, particularly those on enzyme-inducing agents, where the rate was reported as high as 1 in 500 in early studies. The mechanism involves accumulation of 4-en-valproic acid, a hepatotoxic metabolite produced by CYP2C9-mediated oxidation when the preferred mitochondrial beta-oxidation pathway is saturated. Enzyme inducers shunt more valproate through the CYP pathway, increasing toxic metabolite production. Patients with mitochondrial disease (especially POLG mutations, which impair mitochondrial DNA polymerase) are at substantially elevated hepatotoxicity risk and should generally not receive valproate. Liver function tests (LFTs) and ammonia levels should be obtained at baseline and monitored if symptoms of hepatotoxicity develop.3
The teratogenicity of valproate is the most serious prescribing consideration in women of reproductive potential. Valproate is a potent inhibitor of histone deacetylase (HDAC), an effect that disrupts neural tube closure by altering gene expression during weeks 2–4 of embryonic development. Neural tube defects (NTDs), principally spina bifida, occur in approximately 1–2% of exposed pregnancies – a rate 10–20 times the background risk. Beyond NTDs, valproate is associated with a syndrome of major congenital malformations (MCMs) affecting cardiac, urological, and craniofacial systems in 10% of first-trimester exposures. It is also the only ASD with documented dose-dependent reduction in child IQ when taken during pregnancy: children exposed in utero to valproate score approximately 6–9 IQ points below unexposed controls, an effect persisting into school age. Valproate should be avoided in women of childbearing potential unless no effective alternative exists, effective contraception is confirmed, and the patient has been fully counseled on these risks.4
The NEAD (Neurodevelopmental Effects of Antiepileptic Drugs) study demonstrated that children exposed in utero to valproate have a dose-dependent reduction in IQ, verbal abilities, and memory performance at age 6 compared with children exposed to other ASDs (lamotrigine, carbamazepine, or phenytoin). The risk was present even at doses below 800 mg/day, and even in pregnancies with no structural birth defects. Autism spectrum disorder (ASD) and attention deficit hyperactivity disorder (ADHD) are also more common in valproate-exposed children. Unlike structural teratogenic risk, which can be partially mitigated by folate supplementation, the cognitive risk cannot be eliminated by any known intervention once valproate exposure occurs in the critical developmental window.
Valproate inhibits carbamoyl phosphate synthetase I (CPS I), the first enzyme of the urea cycle, impairing the conversion of ammonium to carbamoyl phosphate. This causes mild asymptomatic hyperammonemia in up to 50% of patients on valproate, and symptomatic hyperammonemic encephalopathy in a smaller proportion. The syndrome presents with altered consciousness, confusion, vomiting, and asterixis, and can be mistaken for breakthrough seizures or other encephalopathies. Topiramate co-administration significantly amplifies valproate-induced hyperammonemia by an independent mechanism (topiramate also inhibits carbonic anhydrase in mitochondria, reducing urea cycle function). Serum ammonia should be measured when unexplained cognitive changes or altered consciousness develop in a valproate-treated patient.
Levetiracetam has achieved a central role in contemporary epilepsy management through a combination of a novel mechanism, an exceptionally favorable pharmacokinetic profile, broad efficacy, no significant drug interactions, and an IV formulation bioequivalent to the oral dose. Its psychiatric adverse effects, though occurring in a minority of patients, are sufficiently serious to require monitoring and patient counseling before initiation.
Levetiracetam's primary mechanism of action is binding to synaptic vesicle protein 2A (SV2A), a transmembrane glycoprotein present on all synaptic vesicles. SV2A modulates the priming and fusion of synaptic vesicles with the presynaptic membrane – the steps that immediately precede neurotransmitter release. By binding SV2A, levetiracetam slows the release of synaptic vesicle contents, reducing the rate of both excitatory (glutamate) and inhibitory (GABA) neurotransmitter release, but with a net anticonvulsant effect because normal inhibitory tone is less vesicle-dependent than the high-frequency sustained excitatory output characteristic of ictal firing. Levetiracetam also reduces high-voltage-activated calcium current and reverses the inhibition of GABA-A and GABA-B receptor function by zinc ions and beta-carbolines, contributing additional anticonvulsant activity. The SV2A mechanism has no equivalent among currently marketed sodium channel blockers, GABA potentiators, or calcium channel modulators, which is why levetiracetam and sodium channel blockers can be combined without mechanistic redundancy.5
The pharmacokinetic profile of levetiracetam is nearly ideal for clinical use. Oral bioavailability exceeds 95%, is not affected by food, and is dose-proportional across the full therapeutic range. It is not significantly protein-bound (less than 10%), minimizing displacement interactions. It is eliminated primarily by renal excretion of the parent compound (approximately 66%) and by hydrolysis to an inactive metabolite by non-hepatic esterases (approximately 24%). This hepatic-independent metabolism means it does not induce or inhibit CYP enzymes, and it is not affected by hepatic enzyme inducers or inhibitors. Its half-life of 6–8 hours allows twice-daily dosing. Renal dose adjustment is required in patients with creatinine clearance below 80 mL/min, since the drug is renally cleared. The IV formulation is bioequivalent to the oral form, allowing seamless conversion during acute hospitalization or when patients cannot take medications by mouth.5
Levetiracetam has established efficacy as adjunctive therapy in focal onset seizures, myoclonic seizures in juvenile myoclonic epilepsy (JME), and primary generalized tonic-clonic seizures in idiopathic generalized epilepsies (IGEs). It is notable for being one of the few ASDs that does not aggravate any seizure type – it does not provoke absence, myoclonic, or atonic seizures as carbamazepine and related sodium channel blockers do in generalized epilepsies. Brivaracetam, a second-generation SV2A-binding agent with 15–30 times higher SV2A affinity, provides an alternative for patients who cannot tolerate levetiracetam's behavioral adverse effects, with a similar spectrum of efficacy but fewer reported psychiatric effects at therapeutic doses.6
The psychiatric adverse effects of levetiracetam represent its primary limitation. Irritability, agitation, hostility, anxiety, and in severe cases, psychosis or suicidal ideation, occur in approximately 10–15% of patients. These effects appear to be mechanism-related rather than idiosyncratic, since they correlate with dose and occur more commonly in patients with a pre-existing psychiatric history, intellectual disability, or prior behavioral problems. Pyridoxine (Vitamin B6) has been used empirically to reduce levetiracetam-associated behavioral adverse effects, with some supportive evidence, though the mechanism is not fully established. Prior to initiating levetiracetam, prescribers should screen for psychiatric history and counsel patients and caregivers to monitor for mood or behavioral change, with a plan for dose reduction or substitution if these effects emerge.
The ESETT (Established Status Epilepticus Treatment Trial) randomized adult and pediatric patients with benzodiazepine-refractory convulsive SE to IV levetiracetam (60 mg/kg, maximum 4,500 mg), IV fosphenytoin (20 mg PE/kg, maximum 1,500 mg PE), or IV valproate (40 mg/kg, maximum 3,000 mg). Seizure cessation at 60 minutes without the need for additional anticonvulsants occurred in 47% (levetiracetam), 45% (fosphenytoin), and 46% (valproate) – no statistically significant difference among the three agents. This trial established that none of these agents is superior to the others for benzodiazepine-refractory SE, and that choice among them should be guided by individual patient factors including seizure type, comorbidities, pregnancy status, and drug interaction burden.14
Lamotrigine is among the most widely prescribed ASDs globally, valued for its broad seizure spectrum, low teratogenic risk relative to valproate, good cognitive tolerability, and mood-stabilizing properties in bipolar disorder. Its clinical pharmacology is dominated by its UGT1A4-mediated glucuronidation, which creates pharmacokinetic interactions of sufficient magnitude to require specific dosing adjustments across several clinically important drug combinations.
Lamotrigine blocks voltage-gated sodium channels by binding preferentially to the fast-inactivated state, stabilizing the channel and slowing recovery from inactivation – the same fundamental mechanism as phenytoin, carbamazepine, and lacosamide. Additionally, it inhibits the presynaptic release of glutamate and aspartate by a mechanism independent of sodium channel blockade, possibly through calcium channel modulation at the nerve terminal. This glutamate release inhibition may contribute to its efficacy in absence seizures and its mood-stabilizing properties, distinguishing lamotrigine mechanistically from pure sodium channel blockers like carbamazepine, which has minimal effect on glutamate release. Lamotrigine is therefore classified as a broad-spectrum ASD despite its primary sodium channel mechanism, because the glutamate-release inhibitory component provides activity against absence and myoclonic seizures that pure sodium channel blockers lack.7
Lamotrigine's primary metabolic pathway is glucuronidation by UGT1A4, producing an inactive N-2-glucuronide. Its half-life in monotherapy is 24–35 hours, allowing once- or twice-daily dosing. This half-life is dramatically altered by drug interactions. Valproate is a potent inhibitor of UGT1A4, reducing lamotrigine clearance by approximately 50% and doubling its plasma concentration and half-life. When lamotrigine is added to valproate, doses must be reduced by approximately 50% and the titration rate must be slowed significantly to avoid rash, which is concentration-dependent. Conversely, enzyme-inducing ASDs (carbamazepine, phenytoin, phenobarbital) upregulate UGT1A4 and other glucuronidating enzymes, increasing lamotrigine clearance by 40–50% and halving its steady-state concentration. When an enzyme inducer is discontinued, lamotrigine concentrations rise progressively over 2–4 weeks and may reach toxic levels unless the lamotrigine dose is proactively reduced.8
Oral contraceptives (OCs) containing ethinyl estradiol significantly increase lamotrigine glucuronidation through UGT1A4 induction, reducing lamotrigine plasma concentrations by 40–65% in most studies. This creates a bidirectional clinical problem: starting OCs in a woman on lamotrigine causes lamotrigine levels to fall over weeks, risking seizure breakthrough; stopping OCs causes lamotrigine levels to rise over weeks, risking toxicity. Progestin-only contraceptives do not appear to have this interaction. Women with epilepsy taking lamotrigine who use combined OCs should be counseled about this interaction, monitored for seizure breakthrough or toxicity around OC changes, and may require significant lamotrigine dose adjustments. Lamotrigine does not appear to reduce the efficacy of OCs, unlike enzyme-inducing ASDs.9
The major serious adverse effect of lamotrigine is Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN), life-threatening mucocutaneous reactions occurring at an overall rate of approximately 0.1% in adults and 0.3–0.8% in children. The risk is substantially higher when lamotrigine is initiated at high starting doses, titrated rapidly, or combined with valproate (which approximately doubles exposure). The prescribing protocol of slow initiation (starting at 25 mg every other day when combined with valproate, or 25 mg daily in monotherapy) and gradual 2-week titration steps was designed specifically to minimize SJS risk by preventing rapid concentration rises during immunological sensitization. Any rash appearing within the first 8 weeks of initiation should prompt consideration of drug discontinuation unless a clearly non-lamotrigine cause can be identified.7
Renal blood flow and glomerular filtration rate increase significantly during pregnancy, and the upregulation of UGT1A4 by gestational hormones further accelerates lamotrigine glucuronidation. Lamotrigine clearance increases by 40–65% over the course of pregnancy, causing a progressive fall in plasma concentrations that can lead to seizure breakthrough in the third trimester even when doses have not changed. Dose increases of 50–100% or more are frequently required during pregnancy to maintain seizure control. After delivery, clearance returns to pre-pregnancy levels over days to weeks, and the doses established during pregnancy can cause toxicity postpartum unless rapidly reduced. Lamotrigine monitoring during pregnancy and close postpartum follow-up are mandatory for women managed on lamotrigine.
Topiramate is a structurally unusual fructose-derived sulfamate ASD with four documented anticonvulsant mechanisms, a broad seizure spectrum overlapping with valproate, and a portfolio of non-epilepsy clinical applications. Its dose-dependent cognitive and language adverse effects – particularly word-finding difficulty and slowed information processing – are its principal clinical limitation, and they explain why many patients and clinicians prefer levetiracetam or lamotrigine when a broad-spectrum ASD is needed.
Topiramate's anticonvulsant mechanisms are mechanistically diverse. It blocks voltage-gated sodium channels in a use-dependent manner, reducing sustained high-frequency neuronal firing. It enhances GABA-A receptor-mediated chloride conductance at a site that is distinct from the benzodiazepine binding site, increasing inhibitory tone. It antagonizes AMPA (alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) and kainate subtypes of ionotropic glutamate receptors, reducing excitatory neurotransmission. It also inhibits carbonic anhydrase isoforms II and IV, though this is not considered the primary anticonvulsant mechanism and is instead responsible for several of its systemic adverse effects. These four concurrent mechanisms, none dominant alone, collectively produce broad-spectrum anticonvulsant activity comparable to valproate across generalized, focal, and mixed seizure types.10
The cognitive and language adverse effects of topiramate are the dominant clinical concern. Word-finding difficulty (anomia) and impaired verbal fluency are reported by 15–30% of patients, and slowed information processing affects a similar proportion. Memory impairment, psychomotor slowing, and confusion complete the cognitive profile. These effects are dose-dependent and partially reversible with dose reduction, but they do not resolve completely in some patients even after discontinuation. Slow titration (starting at 25–50 mg/day and increasing by 25–50 mg per week) significantly reduces but does not eliminate the cognitive burden. The negative cognitive profile has important occupational consequences and explains the common clinical scenario of patients requesting drug substitution because they cannot perform their work effectively despite satisfactory seizure control. These effects are more pronounced at doses above 200 mg/day and in patients with pre-existing cognitive vulnerability.11
Topiramate's carbonic anhydrase inhibition produces several systemic consequences. Inhibition of renal carbonic anhydrase reduces bicarbonate reabsorption, causing a dose-dependent non-anion-gap metabolic acidosis (hyperchloremic acidosis) in up to 30% of chronically treated patients. Serum bicarbonate should be measured at baseline and periodically during treatment; persistent acidosis below 17 mEq/L warrants dose reduction or discontinuation. Carbonic anhydrase inhibition also reduces urinary citrate excretion and increases urinary calcium, creating conditions favorable for nephrolithiasis (kidney stones). The risk of kidney stones is approximately 2–4 times higher on topiramate than in the general population. Adequate hydration and avoidance of the ketogenic diet (which worsens acidosis) are recommended for patients on long-term topiramate. Additionally, inhibition of ocular carbonic anhydrase can cause acute angle-closure glaucoma, a medical emergency presenting with acute visual blurring, ocular pain, and elevated intraocular pressure, typically within the first month of use.10
Topiramate produces dose-dependent weight loss, averaging 2–7 kg over 6–12 months of treatment, through mechanisms including anorexia, reduced appetite, and possibly altered fat metabolism. This effect distinguishes it sharply from valproate, which causes weight gain, and has driven its combined use with phentermine (Qsymia) as a weight-loss agent approved for obesity management. The same weight-loss property makes topiramate relatively preferred over valproate when an ASD is needed in an obese patient. Topiramate is also approved for migraine prophylaxis at doses of 100 mg/day, with efficacy independent of its antiepileptic action, likely reflecting its inhibition of cortical spreading depression through sodium channel and glutamate receptor mechanisms. Its teratogenic profile, though less severe than valproate's, includes an increased risk of oral clefts (cleft lip and/or palate) at approximately 1.4% versus 0.1–0.2% in the general population, requiring contraceptive counseling.12
Co-administration of topiramate and valproate produces hyperammonemia that is greater than either drug causes alone. Valproate inhibits carbamoyl phosphate synthetase I (CPS I), impairing the urea cycle at its first step. Topiramate inhibits carbonic anhydrase in hepatocyte mitochondria, which normally supplies CO2 for CPS I. Together, these effects cause a clinically significant reduction in urea cycle capacity and elevated serum ammonia. The presentation of encephalopathy in a patient on this combination should prompt urgent ammonia measurement. In patients who require both drugs (for instance, valproate-treated patients who need additional seizure control and tolerate topiramate), baseline and periodic ammonia monitoring is appropriate, and patients should be counseled to report symptoms of cognitive slowing, confusion, or unusual irritability.
The four broad-spectrum agents covered in this module have overlapping but distinct efficacy profiles, pharmacokinetic characteristics, interaction burdens, and adverse effect portfolios. No single agent is optimal for all patients. The clinical art in ASD selection lies in matching the agent's profile to the individual patient's seizure type, age, sex and reproductive status, comorbidities, co-medications, and priorities regarding cognitive function, body weight, and monitoring burden.
For idiopathic generalized epilepsies (IGEs) including JME, childhood absence epilepsy (CAE), and juvenile absence epilepsy (JAE), valproate remains the most efficacious single agent across all generalized seizure types (absence, myoclonic, and tonic-clonic), and is the treatment of choice in men and non-reproductive-potential women with JME or other complex IGEs. Levetiracetam and lamotrigine are alternatives for women of reproductive potential with IGE who cannot take valproate, though both have somewhat lower efficacy than valproate for the myoclonic component of JME. Lamotrigine has the important caveat that it can worsen myoclonic seizures in some patients with JME, particularly at higher doses; this apparent paradox reflects its sodium channel mechanism's potential to alter thalamocortical firing in ways that increase myoclonic frequency. Ethosuximide is preferred over valproate for pure CAE where tonic-clonic seizures are absent, as it has equivalent absence efficacy with a more favorable adverse effect profile.13
For focal onset epilepsies, all four broad-spectrum agents have documented efficacy as adjunctive therapy, and levetiracetam and lamotrigine have established monotherapy efficacy as well. In clinical practice, levetiracetam is frequently chosen as adjunctive or first-line therapy in focal epilepsy because of its lack of drug interactions, renal elimination without hepatic metabolism, and IV-to-oral convertibility. Lamotrigine is preferred when cognitive function is a priority and the patient can be carefully managed through the slow titration required for rash prevention. Valproate is less commonly chosen for focal epilepsy in adults due to its adverse effect profile and teratogenic risk, though it remains a rational choice in men and post-menopausal women when broad-spectrum coverage is needed. Topiramate is typically used as adjunctive therapy in focal epilepsy rather than monotherapy, due to its cognitive adverse effects limiting tolerability at doses required for optimal seizure control.13
Pregnancy and reproductive potential create the single most consequential selection constraint among these agents. Valproate should be avoided in all women of reproductive potential unless no alternative provides adequate seizure control, because no dose is known to be safe for neurodevelopment and the cognitive harm cannot be prevented. Lamotrigine has the lowest MCM rate of the four agents (approximately 2.3% at typical doses in the EURAP registry) and is the most commonly prescribed ASD in pregnant women in high-income countries.15 Levetiracetam appears to have low teratogenic risk based on current pregnancy registry data, though the database is smaller than for lamotrigine. Topiramate's oral cleft risk and the fact that oral clefts are among the more functionally significant MCMs makes it a second-choice agent in women of reproductive potential who require broad-spectrum coverage. All women with epilepsy who could become pregnant should receive 5 mg/day of folic acid supplementation regardless of which ASD they take, as this dose has been shown to reduce the baseline NTD risk, though it does not eliminate the valproate-associated risk.4
Male, any age, IGE with myoclonic component (JME): valproate first-line. High efficacy across all seizure types; no teratogenicity concern; cost-effective.
Female of reproductive potential, IGE: lamotrigine or levetiracetam preferred; avoid valproate. Counsel on lamotrigine-OC interaction. Monitor lamotrigine levels if OC is used.
Any patient, focal epilepsy, outpatient initiation: levetiracetam or lamotrigine. Levetiracetam if speed of initiation is important (no slow titration required); lamotrigine if cognitive tolerability is the priority.
Obese patient needing broad-spectrum coverage: topiramate (weight loss) preferred over valproate (weight gain).
Patient with migraine and epilepsy: topiramate addresses both (migraine prophylaxis + anticonvulsant) or valproate (also migraine prophylactic).
Patient with psychiatric comorbidity: avoid levetiracetam if behavioral history is concerning; lamotrigine has mood-stabilizing properties and may be beneficial.
Valproate + lamotrigine: valproate inhibits UGT1A4 → lamotrigine levels double. Use half the usual lamotrigine dose and titrate more slowly. Rash risk increased.
Valproate + topiramate: both impair urea cycle → additive hyperammonemia. Monitor ammonia; counsel on encephalopathy symptoms.
Valproate + carbamazepine: valproate inhibits epoxide hydrolase → carbamazepine-10,11-epoxide accumulates → toxicity at therapeutic carbamazepine levels.
Lamotrigine + enzyme inducers (CBZ, PHT, PB): inducers upregulate UGT1A4 → lamotrigine half-life halved. Dose must double; taper lamotrigine slowly when inducer stopped.
Lamotrigine + combined oral contraceptive: estrogen component induces UGT1A4 → lamotrigine levels fall 40–65%. Titrate up when OC started; reduce proactively when OC stopped.
Levetiracetam: no significant pharmacokinetic interactions. Dose-adjust for renal impairment only.
Loscher W. Basic pharmacology of valproate: a review after 35 years of clinical use for the treatment of epilepsy. CNS Drugs. 2002;16(10):669–694.
doi:10.2165/00023210-200216100-00003Patsalos 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.xDreifuss FE, Santilli N, Langer DH, Sweeney KP, Moline KA, Menander KB. Valproic acid hepatic fatalities: a retrospective review. Neurology. 1987;37(3):379–385.
doi:10.1212/wnl.37.3.379Meador 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-XLynch BA, Lambeng N, Nocka K, et al. The synaptic vesicle protein SV2A is the binding site for the antiepileptic drug levetiracetam. Proc Natl Acad Sci USA. 2004;101(26):9861–9866.
doi:10.1073/pnas.0308208101Ryvlin P, Werhahn KJ, Blaszczyk B, et al. Adjunctive brivaracetam in adults with uncontrolled focal epilepsy: results from a double-blind, randomized, placebo-controlled trial. Epilepsia. 2014;55(1):47–56.
doi:10.1111/epi.12432Messenheimer J. Lamotrigine. Epilepsia. 1995;36(Suppl 2):S87–S94.
doi:10.1111/j.1528-1157.1995.tb06001.xJohannessen SI, Landmark CJ. Antiepileptic drug interactions: principles and clinical implications. Curr Neuropharmacol. 2010;8(3):254–267.
doi:10.2174/157015910792246254Sabers A, Ohman I, Christensen J, Tomson T. Oral contraceptives reduce lamotrigine plasma levels. Neurology. 2003;61(4):570–571.
doi:10.1212/01.WNL.0000076485.09353.7AWhite HS. Molecular pharmacology of topiramate: managing seizures and preventing migraine. Headache. 2005;45(Suppl 1):S48–S56.
doi:10.1111/j.1526-4610.2005.4501006.xAldenkamp AP, Arends J. The relative influence of epileptic EEG discharges, short nonconvulsive seizures, and type of epilepsy on cognitive function. Epilepsia. 2004;45(1):54–63.
doi:10.1111/j.0013-9580.2004.33403.xHernandez-Diaz S, Smith CR, Shen A, et al. Comparative safety of antiepileptic drugs during pregnancy. Neurology. 2012;78(21):1692–1699.
doi:10.1212/WNL.0b013e3182574f39Marson AG, Al-Kharusi AM, Alwaidh M, et al. The SANAD study of effectiveness of valproate, lamotrigine, or topiramate for generalised and unclassifiable epilepsy: an unblinded randomised controlled trial. Lancet. 2007;369(9566):1016–1026.
doi:10.1016/S0140-6736(07)60461-9Kapur J, Elm J, Chamberlain JM, et al. Randomized trial of three anticonvulsant medications for status epilepticus. N Engl J Med. 2019;381(22):2103–2113.
doi:10.1056/NEJMoa1905795Tomson 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)30107-8