Alcohol and Marijuana: Effects on Epilepsy and Use by Patients with Epilepsy


Address correspondence and reprint requests to Dr. O. Devinsky at NYU-Mount Sinai Comprehensive Epilepsy Center, 403 E 34th St, New York, NY 10016, U.S.A.


Summary: We review the safety of alcohol or marijuana use by patients with epilepsy. Alcohol intake in small amounts (one to two drinks per day) usually does not increase seizure frequency or significantly affect serum levels of antiepileptic drugs (AEDs). Adult patients with epilepsy should therefore be allowed to consume alcohol in limited amounts. However, exceptions may include patients with a history of alcohol or substance abuse, or those with a history of alcohol-related seizures. The most serious risk of seizures in connection with alcohol use is withdrawal. Alcohol withdrawal lowers the seizure threshold, an effect that may be related to alcohol dose, rapidity of withdrawal, and chronicity of exposure. Individuals who chronically abuse alcohol are at significantly increased risk of developing seizures, which can occur during withdrawal or intoxication. Alcohol abuse predisposes to medical and metabolic disorders that can lower the seizure threshold or cause symptoms that mimic seizures. Therefore, in evaluating a seizure in a patient who is inebriated or has abused alcohol, one must carefully investigate to determine the cause.

Animal and human research on the effects of marijuana on seizure activity are inconclusive. There are currently insufficient data to determine whether occasional or chronic marijuana use influences seizure frequency. Some evidence suggests that marijuana and its active cannabinoids have antiepileptic effects, but these may be specific to partial or tonic–clonic seizures. In some animal models, marijuana or its constituents can lower the seizure threshold. Preliminary, uncontrolled clinical studies suggest that cannabidiol may have antiepileptic effects in humans. Marijuana use can transiently impair short-term memory, and like alcohol use, may increase noncompliance with AEDs. Marijuana use or withdrawal could potentially trigger seizures in susceptible patients.


Central nervous system effects of alcohol


Alcohol (ethanol) is a CNS depressant with pharmacologic properties similar to those of general anesthetics. Depending on dose, rate of consumption, tolerance, and other factors, acute intoxication can cause cognitive impairment, emotional lability, nystagmus, dysarthia, ataxia, stupor, coma, and death. Jackson (1884) (1) recognized that the highest brain centers are most susceptible to alcohol's effects. Thus early depression of inhibitory centers can transiently stimulate and disinhibit behavior. The behavioral actions of alcohol may result from increased γ-aminobutyric acid (GABA)ergic activity with stimulation of the GABA receptor–mediated Cl uptake (2). However, alcohol affects multiple ion channels, and the relationship between effects on specific ion channels and different behavioral effects (e.g., intoxication, tolerance, dependence, seizure threshold) remains uncertain (3).

The most dangerous complication of intoxication is respiratory depression. A blood level of 500 mg/dl is lethal in ∼50% of patients. When used with other CNS depressants, much lower blood levels can be fatal (4,5). Seizures can occur during acute intoxication (6) and withdrawal (7). However, animal and human studies suggest that the long-term and immediate effects of alcohol on the CNS are different and often opposite (8) and that with the short-term administration, alcohol has antiepileptic properties (9). Therefore, in evaluating a seizure in an inebriated patient, one must carefully investigate to determine the cause.


Chronic alcoholism is associated with many neurologic disorders (e.g., cerebral atrophy, head injury). Seizures usually occur in individuals who have abused alcohol for a period of ≥10 years (6). Thus some long-term change in brain function (e.g., kindling) (8) or structure (e.g., loss of inhibitory neurons from metabolic deficiencies; cortical scars from head trauma) may be required (4,10). Chronic alcoholism can cause a cerebellar syndrome with nonepileptic action myoclonus, worse during abstinence and relieved during indulgence (11).


Withdrawal reactions are most severe in chronic alcoholics who rapidly taper or discontinue alcohol. Signs and symptoms include fine tremor, irritability, and insomnia in mild cases; in more severely affected subjects, lateral nystagmus (in the opposite direction of the nystagmus during intoxication), sensory illusions and hallucinations, tremulousness, anorexia, nausea, vomiting, anxiety, tachycardia, and diaphoresis may develop.

In chronic alcoholics, withdrawal seizures typically occur between 7 and 48 h after the cessation of drinking. Withdrawal seizures can occasionally occur in patients treated for alcohol withdrawal with benzodiazepines (BZDs; e.g., oxazepam), when the drug is discontinued (11,12). Patients with partial or generalized epilepsies may develop breakthrough seizures from alcohol withdrawal. Those with idiopathic generalized epilepsies may be most susceptible (13).

Delirium tremens (DT) is the most serious complication of alcohol withdrawal. DT occurs in 5% of hospitalized alcoholics and continues to have a mortality of ∼10% despite aggressive management. DT is a specific syndrome, diagnosed only if the following complex of clinical features is present: autonomic hyperactivity (fever, tachycardia, diaphoresis), severe agitated delirium, tremor, illusions, and hallucinations. DT usually begins 2–4 days after the last drink, has an acute or subacute onset, and lasts an average of 56 h (7).

Alcoholism and seizures

Evaluation of seizures in alcoholics

Alcohol abuse predisposes to many medical and metabolic disorders that mimic seizures. For example, syncope with mild convulsive movements may result from disorders associated with alcohol abuse [e.g., cardiomyopathy, arrhythmias, gastrointestinal (GI) hemorrhage, anemia, hypoglycemia, and dehydration]. First, ensure that the episode was a seizure.

Patients who abuse alcohol are at increased risk for seizures from different causes. The history, general medical, and neurologic examinations often point toward a specific etiology. Laboratory studies should include metabolic and hematologic blood tests, chest radiograph, toxicology screens (if there is suspicion of drug abuse or atypical features), lumbar puncture (based on clinical indications and contraindications), and computed tomography (CT) or magnetic resonance imaging (MRI) scans of the head should be considered in all patients with alcohol-related seizures. These neuroimaging tests also should be obtained in patients withdrawing from alcohol in whom the history or neurologic examination suggests new focal neurologic findings. All patients with a history of recent alcohol consumption and a first seizure (or initial seizure for which they sought medical attention) within 24 h of presentation, as well as those at risk of alcohol withdrawal should be considered for admission to the hospital for evaluation and treatment. The decision whether to admit patients with a history of recurrent alcohol-related seizures must be made on an individual basis.

Status epilepticus in alcoholics suggests the possibility of withdrawal from alcohol and short-term sedative/hypnotic agents or AEDs, or a coexisting, treatable CNS insult such as subdural hematoma or meningitis (7,14).

Ethanol: pharmacokinetics and effects on antiepileptic drugs

Alcohol elimination is zero-order (nonlinear) when blood levels are high, but becomes first-order as the levels decrease. Alcohol is metabolized by two enzyme systems into acetaldehyde: alcohol dehydrogenase (ADH; the major pathway, noninducible) and the microsomal ethanol oxidizing system (MEOS; the minor pathway, inducible). Then acetaldehyde is broken down into acetate by aldehyde dehydrogenase (ALDH). The ADH and MEOS systems are the rate-limiting steps. Approximately 50% of Asian people have a genetic variant in ALDH that limits the conversion rate of acetaldehyde and leads to alcohol intolerance. The half-life of ethanol elimination after high doses is ∼4.25 h (15). However, elimination also depends on sex, age, time of intake (ethanol elimination is fastest in late morning, afternoon, and early evening) (16), and presence of food (17) and other drugs within the system.

The hepatic mixed-function oxygenase system metabolizes alcohol and many AEDs. This system converts lipophilic compounds into hydrophilic compounds excreted in the urine. After acute dosing, ethanol inhibits the MEOS system. The hepatic metabolism of other drugs is slowed, partly because of competitive inhibition and enzyme saturation. This prolongs the half-life and enhances the pharmacodynamic effects of these drugs. This is especially true for carbamazepine (CBZ), phenytoin (PHT), primidone (PRM), valproate (VPA), and phenobarbital (PB). PB peak levels occur earlier and are higher when administered with alcohol than when PB is administered alone (18). Effects on respiratory depression also are synergistic. This has important clinical implications; patients can die after ethanol and PB ingestion, with mean ethanol levels found to be ∼175 mg/dl (19).

Acetaldehyde can bind to albumin, potentially becoming a competitive inhibitor for the attachment of AEDs to the serum proteins and increasing free drug levels. However, because acetaldehyde enhances the binding capacity of serum proteins for diazepam (DZP), high acetaldehyde levels (e.g., in chronic alcoholics), may increase DZP binding and decrease free levels (20). The clinical significance of this finding is uncertain (21).

After prolonged administration, enzyme induction leads to large increases in MEOS (2–3 times the normal rate) and relatively small increases in other monooxygenase systems. The induction of the hepatic systems can increase metabolism of both ethanol and other drugs including CBZ, PB, PHT, and PRM (4,5). The combined effects of long-term alcohol use and an AED regimen can increase clearance rates of AEDs, decreasing AED levels. One study found almost a 50% reduction in total PHT serum concentrations in alcoholics as compared with abstinent controls when tested 24 h after the last dose (22). The increased clearance rates may last 1 to 2 months after the cessation of drinking (23).

Liver disease due to chronic alcoholism can cause hypoalbuminemia, thereby reducing the protein binding of most AEDs, as well as decrease liver blood flow. The reduction in protein binding can lower total steady-state serum concentrations (Table 1), complicate interpretation of drug serum concentrations, and increase hepatic clearance rates of hepatically metabolized AEDs such as BZDs, CBZ, PHT, topiramate (TPM), VPA, zonisamide (ZNS), and, to a lesser extent, lamotrigine (LTG) and PB (24,25). With hypoalbuminemia, PHT free fraction may be slightly increased (26), and increased PHT dose can saturate metabolic pathways, increasing levels and producing excessive side effects. Decreased liver blood flow can diminish first-pass effect for drugs with a high intrinsic clearance such as oral chlordiazepoxide or DZP (27). This effect does not appear relevant to other hepatically metabolized drugs such as PHT (28).

Table 1.  Alcohol and antiepileptic drugs
Serum levels with
alcohol use
AcuteChronicSide effects of AED similar
to alcohol intoxication
BenzodiazapinesAtaxia, confusion, dizziness
   Nystagmus, sedation
CarbamazepineAtaxia, blurred vision, dizziness, drowsiness, vertigo, nystagmus, nausea, vomiting
EthosuximideAgitation, dizziness, nausea, vomiting
PrimidoneAtaxia, nystagmus, sedation, vertigo, nausea, vomiting
PhenobarbitalAtaxia, nystagmus, sedation, vertigo, nausea, vomiting
PhenytoinAtaxia, confusion, tiredness, dizziness, nystagmus, nausea, vomiting
Valproic acidAtaxia, sedation, dizziness, tremor, nausea, vomiting

Chronic alcoholics may require higher AED doses because of the enhanced clearance rates and more frequent, smaller AED doses to maintain therapeutic serum concentrations. For example, with VPA, both enzyme induction and protein binding must be considered when serum concentrations are interpreted because VPA has variable protein binding and, like other drugs, its clearance is related to free rather than total levels (19,29). Similarly, with CBZ, autoinduction of the enzyme system and changes in protein binding must be considered when interpreting serum concentrations (19,23). An accurate history of a patient's past or present alcohol use or abuse is essential when deciding on an AED.

Is alcohol abuse a risk factor for epilepsy?

Among patients who abuse alcohol, there is an increased prevalence of seizures and epilepsy. However, among patients with epilepsy, there is a decreased risk of alcohol use and abuse as compared with controls, probably because of warnings from physicians and pharmacists regarding increased risk of seizures or medication toxicity. Patients with recurrent episodes of only alcohol-withdrawal seizures do not have epilepsy.

There are currently insufficient data to determine if chronic alcohol abuse predisposes to recurrent seizures separate from head trauma, metabolic and medical disorders, and withdrawal (4,13). In some cases, isolated or multiple episodes of head trauma probably cause posttraumatic epilepsy (30). However, whether or not long-standing alcohol abuse causes physiologic (e.g., kindling) or pathologic (e.g., selective loss of inhibitory neurons) changes that result in recurrent seizures is unknown. In animals, repetitive alcohol withdrawal can kindle the amygdala and facilitate electrical kindling (31). Notably, acute alcohol intake blocks the effects of glutamate on the N-methyl-d-aspartate (NMDA) receptor. Prolonged exposure may lead to receptor supersensitivity, and withdrawal may increase glutaminergic activity and contribute to neurotoxicity and possibly seizure activity (32,33).

For those epilepsy patients who have chronically abused alcohol, the risk of seizures may increase on tapering or cessation of drinking because of an increase in AED metabolization and elimination caused by the relative lack of competing substrate. The physician should monitor drug levels during and after a patient stops chronic alcohol use.

Alcohol use in patients with epilepsy

Several studies showed that small to modest alcohol intake does not increase seizure frequency or significantly alter blood levels of AEDs (9,34). These studies (9,34) demonstrated that social drinking (one to two drinks/occasion) has no significant effect on blood levels of CBZ, ESM, and PHT. And although a marginal significant difference was found in the blood levels of VPA and PB in the same studies, the differences could be attributed to random variation in blood levels (9,34). Regardless, patients should be informed that intake of ethanol with CNS depressants such as PB can be dangerous.

The ability of epilepsy patients to consume limited amounts of alcohol can remove an unnecessary social restriction. Patients with epilepsy should therefore be allowed to use alcohol in small amounts (one to two drinks/occasion, no more than three to six drinks/week). However, exceptions may include patients with a history of noncompliance to AEDs, prior history of alcohol or other substance abuse, or those with alcohol-related seizures. Further exceptions may be adolescents and young adults who have difficulty limiting their alcohol intake. AED side effects can be similar to those of alcohol, leading to increased toxicity (Table 1).

Patients who drink moderate (three to four drinks/occasion) or heavy amounts (more than four drinks/occasion) of alcohol should be warned that they are at increased risk of seizures, with the greatest risk occurring 7–48 h after the last drink (35). Concurrence of withdrawal from moderate to heavy alcohol use and sleep deprivation or missed AED doses can be especially dangerous for epilepsy patients.


Marijuana and the CNS


Marijuana (Cannabis sativa) is a naturally growing plant with many chemical constituents present in varying levels in the different varieties. Approximately 60 cannabinoids and 260 noncannabinoid constituents have been identified (36,37). The main constituents of marijuana are delta-9-tetrahydrocannibinol (THC), the primary psychoactive constituent, and cannabidiol (CBD) the primary nonpsychoactive constituent.

Marijuana is generally smoked, but also may be ingested. Acute marijuana administration produces diverse cognitive and perceptual effects, the psychoactive “high.” The drug also produces cardiovascular and autonomic effects, including tachycardia, hypotension, dry mouth, and conjunctival injection (38). Other side effects include dizziness, dysphoria, and memory disturbances; infrequent effects include hallucinations, ataxia, tremulousness, and subjective muscle weakness (39). Acutely, positron emission tomography (PET) studies reveal that marijuana can increase blood flow in limbic areas, possibly mediating mood effects, and decrease blood flow in temporal neocortex, possibly mediating cognitive impairment (40).

In the 19th century, marijuana was used to treat epilepsy (41,42). Gowers reported, Cannabis indica (marijuana), which was first recommended in epilepsy by Dr. Reynolds, is sometimes, though not very frequently, useful. It is of small value as an adjunct to the bromide, but is sometimes of considerable service given separately … I have administered [it] in many cases, and with the effect of delaying the paroxysms and mitigating their severity in some individuals (42).

However, little medical attention was subsequently given to its possible antiepileptic effects. Recent animal and human research on the effects of marijuana or its constituents on seizure activity has led to mixed results (see later). The limited evidence suggests that marijuana and its active cannabinoids may have antiepileptic effects in some individuals.


With continued, frequent use of marijuana tolerance develops to some subjective, psychoactive effects and cardiovascular effects (e.g., tachycardia and hypotension) (43). The cardiovascular effects are potentially dangerous to those with pre-existing heart disease (38). Long-term pulmonary side effects may occur with the daily smoking of a few marijuana cigarettes (44). Endocrine effects may delay puberty in boys or disrupt hormonal and reproductive function in women (38). Marijuana can cause psychomotor slowing (45,46). Habitual use of marijuana may give rise to an “amotivational syndrome” marked by apathy, and decreased productivity and motivation. However, the controversial syndrome may not be a result of marijuana use, but instead may be the impetus for use of the drug.

Little is known about the extended effects of marijuana or its constituents on the brain. Short-term use of marijuana can decrease alpha amplitude and frequency (47), sleep duration, and rapid-eye-motion (REM) sleep (48). However, within an average of 10 days of continued administration, these functions returned to normal (48).


Limited data suggest that after repeated THC administration, withdrawal effects can occur. When 120 subjects were given moderate doses of oral THC for 5–21 days, self-reported withdrawal symptoms included irritability, restlessness, insomnia, sweating, salivation, tremor, nausea, and anorexia (48). Although the initial changes in intraocular eye pressure and sleep and waking EEG normalized with continued use, there was a rebound effect above the pre-THC levels after discontinuation. The withdrawal symptoms were greater for those on higher doses and were most severe 8–12 h after stopping the THC (48).

Marijuana use and seizures

Pharmacokinetics and effects on seizure threshold and AEDs

Cannabinoid receptors are found in the brainstem, limbic, and neocortical areas that modulate seizure activity (49). Receptor density is greatest in the substantia nigra pars reticulata, corpus striatum, cerebellum, hippocampus, and dentate gyrus. The mechanism by which marijuana and the cannabinoids alter seizure threshold are not well defined. There may be a functional connection between cannabinoid and NMDA receptors because the two receptors are co-localized in many brain areas (50). Cannabinoids influence levels of major catecholaminergic transmitters such as dopamine and norepinephrine (51). Cannabinoids also influence thalamocortical projections, which could alter seizure threshold by increasing synchronicity (52). Further, there is a dose-dependent effect of cannabinoids on CNS excitability, with low doses producing activation and high doses reducing electrical activity (51). The studies assessing the effects of THC, CBD, and related derivatives on seizure threshold and kindling reveal (a) marked variability for the different derivatives in different models and in different species; (b) the mechanisms of cannabinoid anti- or pro-convulsive effects are not well-defined; and (c) short term effects may be followed by a rebound effect (e.g., decreasing seizure threshold) after discontinuation (39,49,53).

The effects of THC on seizure threshold vary widely between studies that provide conflicting results (49,53). THC has both proconvulsant and anticonvulsant effects depending on dose, seizure model, and factors of seizure initiation versus seizure spread (39). THC has an anticonvulsant effect in models with rapidly evoked tonic discharges, utilizing post-tetanic potential for recruitment. In animal models of epilepsy, THC is effective against some forms of partial and generalized convulsive seizures. However, in animal models of genetic and absence epilepsy and other models of partial epilepsy, THC has proconvulsant effects. Animal studies also document a “rebound” effect to THC (54–56). After a single exposure to THC, the withdrawal phase enhanced CNS excitability, resulting in increased susceptibility to electrical-induced convulsions (55). This withdrawal hyperexcitability suggests that in susceptible patients, marijuana use may provoke withdrawal seizures (56).

CBD and its derivative analogs have anticonvulsant properties without stimulatory or convulsant properties. In animal models these cannabinoids (a) block or reduce the spread of generalized seizures induced by maximal electroshock or γ-aminobutyric acid (GABA)-inhibiting drugs, (b) block simple partial seizures induced by the topical application of convulsant metals on the cortex, and (c) increase the seizure threshold for electrical kindling (57). Further, these compounds increased the potency of AEDs in animal models of partial and generalized motor seizures, but inhibit the action of AEDs in animal models of absence seizures (58).

Several clinical studies examined the effects of CBD on seizure frequency, but these small, double-blinded studies found either some reduction of seizure frequency (59–61) or no significant reduction of seizures comparing CBD and placebo (61,62). However, most studies used low doses of CBD, and data suggest further study of CBD's antiepileptic and AED-potentiating effects at higher doses is warranted (59,61).

Is marijuana use a risk factor for seizures in epilepsy patients?

Few clinical studies have examined the effects of cannabinoids on seizure frequency or severity in epilepsy patients. No large double-blind or controlled studies have evaluated marijuana, or the cannabinoids, in treating epilepsy patients. One epidemiologic study of illicit drug use and new-onset seizures found that marijuana use appeared to be a protective factor against first seizures in men (63). The adjusted odds ratio was 0.42 for every marijuana use and 0.36 for marijuana use within 90 days of hospitalization.

Clinical anecdotes and single case reports suggest that marijuana may reduce seizure frequency or, conversely, may provoke seizure activity in selected cases. However, in most instances, it probably does not affect seizure activity. In our informal interviews with >215 patients with active epilepsy (seizures within the past 5 years or current use of AEDs to control seizures) who have used marijuana intermittently or regularly, the majority (194 patients; 90.2%) of patients have failed to identify a relationship between marijuana use and seizure frequency or severity. Sixteen (7.4%) patients believed seizures were less frequent around the time of marijuana consumption; five (2.3%) believed seizures were more frequent. However, these interviews represent retrospective recollections in a population with frequent short-term memory impairments. Several colleagues have informally reported to us that patients with idiopathic generalized epilepsy, particularly juvenile myoclonic epilepsy, have had seizures associated with marijuana use. Interpretation of these data is complicated by the fact that many of these patients may have consumed alcohol, may have missed doses of their AEDs, or may have been subjected to sleep deprivation near the time of marijuana use.

Two case reports suggest that marijuana use reduced seizure frequency (64,65). However, another young man who had been seizure free for 6 months smoked marijuana 7 times in 3 weeks and had three tonic–clonic seizures during that time (66), though none of these recurrent seizures occurred during or immediately after consumption. The difficulty with individual case reports is highlighted by a patient in whom marijuana use was implicated as the cause of new-onset seizures, but an infectious etiology was subsequently documented (67).

Recommendations for marijuana use in patients with epilepsy

In contrast to that of alcohol, most marijuana consumption in the United States is illegal. Therefore, although physicians may be asked about the safety of marijuana use by epilepsy patients, any discussion must be prefaced by the fact that illegal use of a substance cannot be recommended. There are insufficient data to determine whether occasional or chronic marijuana use influences seizure frequency.

Given the mixed evidence of marijuana's effect on seizure threshold and the current illegality of marijuana possession and consumption, epilepsy patients should be cautioned not to use marijuana. However, further study of the pharmacologic effects and potential anticonvulsant or proconvulsant effects is warranted, given the preliminary studies and anecdotal reports.