Whether or not antiepileptic drugs should be withdrawn after a patient has been seizure-free for several years is a complex issue. Some studies suggest the overall risk of seizure recurrence is approximately 30% if treatment is withdrawn. Clinical factors associated with a greater chance of successful withdrawal include childhood onset epilepsy, a normal electroencephalogram prior to drug withdrawal, being seizure-free for more than two years, monotherapy, normal neuroimaging and normal intellect. If antiepileptic drugs are withdrawn, they should be withdrawn slowly, ideally over several months.


Longitudinal and community-based studies suggest that antiepileptic drugs will result in approximately 70% of adults diagnosed with epilepsy becoming seizure-free.1 Whether or not antiepileptic drugs should be withdrawn after a patient has been seizure-free for several years is a complex issue. Discontinuing antiepileptic drugs implies the seizure tendency is no longer present. A number of clinical factors can help to predict the risk of seizure recurrence. This can assist when making a decision as to whether or not to discontinue antiepileptic drug treatment. The decision must include weighing up the harms and benefits of continuing and discontinuing therapy, and the risk and consequences of seizure recurrence.

Importance of remaining seizure-free

For many adults a recurrence of their seizures would have significant implications. Seizures impact on the ability to drive, possibly the ability to work. For many patients there is a significant negative impact on their psychological state and well-being. Seizures may also be associated with injury, and sudden unexpected death may occur as a direct result of seizures, particularly tonic-clonic seizures. Although the risk of seizure recurrence may be low, some patients choose to remain on treatment as they feel more secure taking an antiepileptic drug.

Reasons for discontinuing antiepileptic drugs

Patients may want to stop their treatment because antiepileptic drugs are commonly associated with mild adverse effects and have the potential for serious adverse effects. The risk of teratogenicity in women of childbearing years may prompt drug withdrawal earlier than might otherwise occur, but the risk of seizures recurring in pregnancy needs to be balanced against the possible harm of continuing treatment. Other problems include interactions with other drugs, for example the interaction between carbamazepine (liver enzyme inducer) and the oral contraceptive pill. For some patients, taking daily medication is a constant reminder of their chronic condition and has a negative impact on their well-being.

Relapse following withdrawal of antiepileptic drugs

A meta-analysis of 25 studies estimated the overall rate of seizure relapse following the withdrawal of antiepileptic drugs.2 After discontinuing therapy 25% of patients relapsed within one year and 29% by two years. A randomised study of continued treatment versus slow withdrawal in 1013 patients who had been seizure-free for at least two years, reported that 78% of patients who continued treatment and 59% of those who stopped treatment were still seizure-free two years later.3 These figures are overall estimates only and several factors need to be considered in individual patients when determining risk of relapse.

Factors influencing relapse

A number of studies have assessed clinical factors which might predict seizure recurrence following antiepileptic drug withdrawal (see Table 1).

Table 1
Factors predicting seizure recurrence following antiepileptic drug withdrawal

Associated with increased risk Associated with reduced risk

Juvenile myoclonic epilepsy
Partial seizures with secondary generalisation
Abnormal electroencephalogram
Epileptogenic lesion on neuroimaging
Childhood absence epilepsy
Benign rolandic epilepsy
Normal electroencephalogram
Normal neuroimaging
Onset in childhood
No seizures for more than two years prior to antiepileptic
drug withdrawal
No seizures following introduction of antiepileptic drug
Normal intellect

Age of onset of epilepsy

The risk of seizure recurrence following drug withdrawal is higher in adolescents than children.2,3 This probably reflects the prognosis of age-related syndromes. Childhood absence epilepsy and benign rolandic epilepsy have a good prognosis for antiepileptic drug withdrawal while juvenile myoclonic epilepsy is more likely to recur.

Seizure type and epilepsy syndrome

Epilepsy is not one disease, but comprises a number of syndromes each with a different prognosis (see Box 1). Juvenile myoclonic epilepsy, probably the commonest type of primary generalised epilepsy in adults, has a high risk of seizure recurrence on drug withdrawal.

Studies assessing seizure type report myoclonic seizures, tonic-clonic seizures and partial seizures as associated with an increased risk of seizure recurrence following antiepileptic drug withdrawal.3 One study reported 63% of patients with partial seizures relapsed on drug withdrawal.4 Remote symptomatic seizures (defined as seizures occurring in patients with a prior neurologic insult such as head injury, stroke or history of intellectual disability) are associated with a high risk of seizure relapse following discontinuation of therapy.2

Box 1
Features of common epilepsy syndromes

Childhood absence epilepsy
This is a form of primary generalised epilepsy which has a genetic predisposition and begins in childhood. It is characterised by frequent absence seizures, however it is estimated that 50% of patients will have a generalised tonic-clonic seizure. Seizures usually respond well to treatment and the condition often remits in adult life.

Juvenile myoclonic epilepsy
This is a common form of primary generalised epilepsy which usually begins in adolescence. It is characterised by tonic-clonic seizures and myoclonic jerks. Seizures may be precipitated by sleep deprivation and excess alcohol. The condition responds well to valproate, but there is a high chance of recurrence after drug withdrawal.

Benign rolandic epilepsy
This is a common form of idiopathic partial epilepsy which occurs in otherwise normal children and usually remits in adolescence. The seizures are often sleep-related and characterised by orofacial or oropharyngeal involvement which frequently evolve into secondarily generalised tonic-clonic seizures. The interictal EEG demonstrates a spike focus in the centrotemporal region. Seizures usually respond to carbamazepine.

Partial epilepsy (focal epilepsy)
The most common form is temporal lobe epilepsy characterised by frequent complex partial seizures and occasional secondarily generalised tonic-clonic seizures. Partial epilepsy is often resistant to antiepileptic drug treatment.

Seizure frequency, timing and antiepileptic drug usage

The UK Medical Research Council (MRC) trial found that an increased risk of seizure recurrence was associated with:

  • a shorter duration of seizure-free period prior to study entry
  • seizures after starting antiepileptic drug treatment
  • patients taking multiple antiepileptic drugs at the time of study entry.3


There is limited evidence to support the assumption that an interictal electroencephalogram (EEG) is predictive of seizure recurrence following antiepileptic drug withdrawal. The MRC study found that patients with only tonic-clonic seizures and generalised spike wave on EEG had a higher recurrence rate. Patients with tonic-clonic seizures and focal features or normal EEG had no increased risk of recurrence.3

The meta-analysis of 25 studies noted that an abnormal EEG was associated with an increased risk of recurrence, but there was considerable variability in the results and in most studies the epileptiform activity was not differentiated.2 In some studies patients with an abnormal EEG, particularly the presence of epileptiform activity, were excluded, biasing the results.

Another study assessed the role of the EEG in predicting seizure recurrence in partial epilepsies. It found that although the interictal EEG at time of antiepileptic drug withdrawal did not predict recurrence, a worsening of the EEG after withdrawal was predictive of seizure recurrence.4

Role of video-EEG monitoring

The value of video-EEG monitoring in assessing the chance of remaining seizure-free following antiepileptic drug withdrawal has not been systematically studied. However, it is not uncommon for patients with generalised epilepsy to report no seizures and yet continuous monitoring reveals frequent sub-clinical seizure activity.


Although studies are lacking, it seems likely that patients with an epileptogenic lesion on computed tomography or magnetic resonance imaging of the brain are more likely to have recurrences on drug withdrawal. This is consistent with the data on recurrences in a population of patients presenting with their first seizure - patients with an epileptogenic lesion on neuroimaging have a higher risk of seizure recurrence.5

Provoked seizure

It is likely that provoked seizures (such as seizures occurring with excess alcohol, sleep deprivation or induced by drugs) are less likely to recur if these factors are avoided.

Certainty of diagnosis

Occasionally patients are commenced on an antiepileptic drug and in retrospect there is some doubt about the diagnosis of epilepsy. In these patients drug withdrawal is a reasonable option.

Withdrawal after surgery for temporal lobe epilepsy

A reduction in antiepileptic drug therapy following successful surgery on the temporal lobe should be considered. However, the timing of withdrawal and whether or not all medications should be withdrawn is controversial. Many experts would agree that treatment should continue for two years. If the patient is then seizure-free and on several drugs, a reduction in medication could be commenced, but at least one antiepileptic drug should be continued for the long term.6

When to start antiepileptic drug withdrawal

No randomised-controlled trials have studied how long patients should be seizure-free before withdrawing their treatment can be considered. Studies in children suggest two years, but the MRC study of adults suggests a longer period of at least three years is desirable.3

Precautions during and after antiepileptic drug withdrawal

Patients need to be advised there is no guarantee they will remain seizure-free. The risk of relapse is greatest during the first 12 months.

'Assessing fitness to drive', produced by Austroads and the National Road Transport Commission, advises that the patient 'should not drive for the full period of withdrawal and for three months thereafter' unless withdrawal is advised by an experienced consultant on the basis that the risk of seizure recurrence is low.7 Safety advice such as not swimming alone, avoiding heights and having a shower rather than a bath, should be reinforced.

How to withdraw antiepileptic drugs

Withdrawal should be gradual and take place over approximately six months. Rapid withdrawal, particularly of barbiturates and benzodiazepines, can precipitate seizures. The withdrawal protocol for adults in the MRC study decreased doses every four weeks (see Box 2).3 This approach may need to be modified in patients on low doses of antiepileptic drugs. For patients taking multiple drugs, the withdrawal should be sequential.3

Box 2
Suggested protocol for antiepileptic drug withdrawal

The following decrements are recommended every four weeks:

30 mg

Management of seizure recurrence

If seizures recur the previous medication that controlled seizure should be restarted.


For adults who are seizure-free, there are limited data to help them make an informed decision regarding drug withdrawal. The decision is an individual one and in general should be patient driven, as it is very important that the patient supports the decision, is aware that seizures may recur and understands the associated risks.


  1. Kwan P, Brodie MJ. Early identification of refractory epilepsy. N Engl J Med 2000;342:314-9.
  2. Berg AT, Shinnar S. Relapse following discontinuation of antiepileptic drugs: a meta-analysis. Neurology 1994;44:601-8.
  3. Randomised study of antiepileptic drug withdrawal in patients in remission. Medical Research Council Antiepileptic Drug Withdrawal Study Group. Lancet 1991;337:1175-80.
  4. Tinuper P, Avoni P, Riva R, Provini F, Lugaresi E, Baruzzi A. The prognostic value of the electroencephalogram in antiepileptic drug withdrawal in partial epilepsies. Neurology 1996;47:76-8.
  5. Berg AT, Shinnar S. The risk of seizure recurrence following a first unprovoked seizure: a quantitative review. Neurology 1991;41:965-72.
  6. Lowe A, David E, Kilpatrick CJ, Matkovic Z, Cook MJ, Kaye A, et al. Epilepsy surgery for pathologically proven hippocampal sclerosis provides long-term seizure control and improved quality of life. Epilepsia 2004;45:237-42.
  7. Assessing fitness to drive, for commercial and private vehicle drivers. Medical standards for licensing and clinical management guidelines. Austroads and National Road Transport Commission. Sydney: Austroads; 2003. p. 59.