Article Text

Download PDFPDF

Topiramate in clinical practice: first year’s postlicensing experience in a specialist epilepsy clinic
  1. Mark W Kellett,
  2. David F Smith,
  3. Paul A Stockton,
  4. David W Chadwick
  1. The Walton Centre for Neurology and Neurosurgery, Lower Lane, Liverpool L9 7LJ, UK
  1. Dr Mark W Kellett, The Walton Centre for Neurology and Neurosurgery, Lower Lane, Liverpool L9 7LJ, UK. Telephone 0044 151 529 4099; fax 0044 151 525 5512; email mark.kellett{at}


OBJECTIVE Topiramate became available for use in October 1995. Meta-analysis of its randomised controlled data suggested that it may be the most potent of the new antiepileptic drugs. The aim of this study was to assess the first year’s postlicensing experience in a specialist regional epilepsy clinic.

METHODS The case notes of 174 of 178 patients who were prescribed topiramate in the 12 months between November 1995 and October 1996 were retrospectively reviewed. Data were collected on seizure type, classification of epilepsy, presence or absence of learning difficulties, depression, or behavioural problems, co-medication, dosage escalation, efficacy, adverse events, whether or not the patient was still on topiramate and, if not, the reason for withdrawal. Kaplan-Meier survival analysis was used to estimate the overall retention rate and log rank tests were used to determine factors associated with stopping topiramate.

RESULTS Overall 90 of 174 patients had ceased taking topiramate at the end of the study. The median “survival time” was 427 days (95% CI 362.9–491.1). The cumulative probability for remaining on topiramate at 1 year was 0.549 (95% CI 0.475–0.623). The retention rate in patients in whom topiramate was substituted for another drug was significantly higher than in those in whom it was added to current therapy. Adverse events (CNS related) were the most common reason for stopping topiramate. Eight patients with partial and one patient with juvenile myoclonic epilepsy became seizure free.

CONCLUSIONS There is a significant (20–25%) chance of being intolerant to topiramate at relatively low doses. Substituting topiramate for another antiepileptic drug may reduce the chances of drug withdrawal. If topiramate is tolerated there is a good chance of worthwhile improvement in seizure control. These data, although not derived from randomised controlled trials, represent pragmatic use of novel antiepileptic drugs in “real life” and may be helpful to non-specialists when prescribing topiramate.

  • topiramate
  • epilepsy

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Four new drugs, vigabatrin, lamotrigine, gabapentin, and topiramate have recently received product licences for use as add on therapy in patients with refractory partial and tonic-clonic seizures. These licences have been granted on the basis of randomised controlled trials, meta-analysis of which suggests that topiramate may possess greater potency than the other compounds.1 2

Randomised control trials, however, are clinical experiments designed to provide sufficient evidence of efficacy and safety to satisfy regulatory authorities, but the relation between these results and the utility of these drugs in everyday clinical practice is poorly defined.3 Indeed a recent “audit” disclosed that, after 6 to 8 years, less than a quarter of patients still living continued taking these compounds.4 5 There are several possible interpretations of this finding. Firstly, the drugs may lack sustained efficacy or may not be as well tolerated as suggested by the trials. Secondly, their use in randomised control trials, which governs, at least initially, their postmarketing use, may have been suboptimal resulting in drugs being “’branded”’ as toxic or ineffective. Certainly recommendations for the use of lamotrigine (dosage escalation) and gabapentin (dosage range) have changed significantly with postmarketing surveillance. An alternative explanation of the data of Marson et al 1 2 is that topiramate may have been more thoroughly evaluated than other novel antiepileptic drugs and, accordingly, that trial results might accurately represent its efficacy and tolerability.

Although patients undoubtedly benefit from optimal use of the new drugs, inefficient use exacerbates the problems associated with similar usage of the conventional antiepileptic drugs (overrapid dosage escalation or failure to try maximally tolerated doses). It is incumbent on frequent prescribers to report their experience for the benefit of other physicians unfamiliar with these new drugs, to assist in the treatment and counselling of patients. We present the results of the first year’s experience of topiramate in specialist epilepsy clinics where experienced staff have adhered to recommendations for its use.


By searching data held on a computerised database of the Mersey Regional Epilepsy Clinic in Liverpool6 and hospital pharmacy records we identified 178 patients who were prescribed topiramate between November 1995 and October 1996 inclusive, and were able to review the case notes of 174 of these. Patients were seen in two settings; satellite epilepsy clinics in north Wales where all prescriptions were made by DFS and the Mersey Regional Epilepsy Clinic jointly run by DWC and DFS with the support of specialist registrars in neurology. For most of the period of interest the drug was administered to patients with partial epilepsies but, latterly, based on anecdotal reports of a broad range of efficacy, topiramate was given to some patients with idiopathic and symptomatic generalised epilepsies.

Data were collected on seizure type, classification of epilepsy, presence or absence of learning disability, depression or behavioural disturbance, comedication, dosage escalation, efficacy, toxicity, whether or not the patient was still on topiramate and, if not, the reasons for drug withdrawal. In all cases maximum and maintenance doses were documented for those still on topiramate. A logical hierarchy of efficacy was devised (table 1). Only reported side effects were recorded; specific inquiries about commonly described adverse events were not made. Details of admission to hospital due to adverse events were obtained.

Table 1

Characteristics of study population

For the purposes of survival analysis the end point was the date of discontinuation of topiramate; patients remaining on topiramate at the last clinical contact were censored. If the last clinical contact was less than 1 year after starting topiramate, attempts were made to contact the patient by telephone. Using Statview,7Kaplan-Meier survival analysis was used to estimate the overall retention rate and log rank tests were applied to subgroups to determine factors associated with stopping topiramate.


The characteristics of the study population are summarised in table 1. In the vast majority of patients topiramate was introduced according to (n=150) or more cautiously than (n=11) the data sheet recommendations of 50 mg/day for 2 weeks, with titration by 50 mg/day at 2 weekly intervals until 200 mg/day. Two thirds (119/174) of our patients were receiving two or more antiepileptic drugs for their refractory seizure disorders.

Overall 90 of 174 (51.7%) patients had ceased taking topiramate at the end of the study. The median period that patients remained on topiramate was 427 days (95% CI of median 362.9–491.1)(minimum 5 days, maximum 652 days). The cumulative probability of remaining on treatment for 12 months was 0.549 (95% CI 0.475–0.623, fig 1).

Figure 1

Kaplan-Meier cumulative survival plot for time to stopping topiramate.

The retention rate of patients in whom topiramate was substituted for another drug (n=49/88) was significantly higher than for those in whom it had been added to their current treatment (n=35/86) (fig 2 and table2). Furthermore, it seemed that patients who had received slower and faster dosage escalations than the data sheet recommendations had lower retention rates than those who adhered to the recommendations (fig 3and table 2). There was no association between seizure type or syndrome, presence of generalised tonic clonic seizures, history of depression, behavioural disturbance or learning difficulties, or number of drugs at time of starting topiramate, and an increased likelihood of stopping topiramate (table 2), although small numbers in some groups limited the ability to detect differences.

Figure 2

Kaplan-Meier cumulative survival plot for patients in whom topiramate was added on to or substituted for other drugs.

Table 2

Log rank (Mantel-Cox) test for time to stop topiramate

Figure 3

Kaplan-Meier cumulative survival plot for patients following standard, faster, or slower dose titrations.

Adverse events were the most common reason for stopping topiramate, especially within the first 4 months of treatment (table 3). Patients stopping topiramate in this period reached a significantly lower maximum dose (mean 203.6, 95% CI 178.1–229.2) than those stopping later (mean 349.0, 95% CI 305.1–392.8) or those continuing on topiramate (mean 447.6, 95% CI 409.5-485.7). Patients stopping topiramate after 4 months did so because of lack of efficacy in combination with adverse effects which precluded trying the effect of higher doses (table 3). Eleven patients (6.3%) reported a deterioration in seizure control with (n=7) or without (n=4) other adverse events. Patients with symptomatic generalised epilepsies were more likely to report a deterioration in seizure control (n=4/10) than those with partial (n=7/150) or idiopathic generalised epilepsies (n=0/14), although the numbers are small and must be interpreted with caution.

Table 3

Characteristics of patients stopping topiramate before and after 120 days

In the group as a whole there was a suggestion that adverse events were more frequent in the add on patients (n=65/86) than in substitution patients (n=54/88) (Fisher’s exact test, p=0.051), but this difference was not seen in the groups who stopped taking topiramate. Similarly, adverse events did not explain the different retention rates for different dose escalations (Fishers’ exact test, p=0.24) in the group as a whole.

Cognitive and behavioural problems were the most common adverse events in patients staying on and coming off topiramate (table 1). These included complaints of depression, mental slowing, reduced concentration, behavioural changes, and drowsiness. Four patients required admission to hospital; one patient developed depression with psychotic features (box 1), one patient developed increased frequency of blank spells different from her usual seizures but difficult to classify, one developed prolonged postictal confusion with marked sedation on 400 mg/day, but continues on 200 mg/ day with convincing improvement, another patient had marked cognitive/behavioural side effects.

Case 1

A 35 year old man had post-traumatic epilepsy manifest by weekly clusters of complex partial and secondary generalised tonic-clonic seizures, despite maximally tolerated doses of phenytoin. He had previously received thorough trials of all conventional antiepileptic drugs, lamotrigine, and vigabatrin. After the addition of topiramate he experienced only one cluster of complex-partial seizures in a 3 month period. However, he could not tolerate 150 mg twice daily because of depressive symptoms. Improved seizure control was maintained despite reducing the dose, in two stages, to 50 mg twice daily. He remained withdrawn and reported suicidal ideation. These symptoms resolved within 2 to 3 weeks of stopping topiramate. A second trial, with cautious escalation to 50 mg twice daily, was similarly effective but, unfortunately, the profound depression returned, necessitating withdrawal.

In those who tolerated topiramate there is an excellent chance of worthwhile improvement (table 1). The seizure free patients included eight with partial epilepsy and one with juvenile myoclonic epilepsy (case 2). One patient ceased having secondary generalised convulsions and a further eight showed a convincing improvement but had to come off topiramate because of adverse effects.

Case 2

A 32 year old woman had “refractory juvenile myoclonic epilepsy” manifest by daily matinal myoclonus and, on average one tonic-clonic seizure a week, despite sodium valproate (1 g in the morning plus 1.5 g in the evening) and a maximally tolerated dose of lamotrigine (100 mg twice daily). In the past combinations of valproate and benzodiazepines or phenobarbitone had failed to produce sustained remission. The addition of 100 mg topiramate twice daily resulted in cessation of tonic-clonic and, most impressively, myoclonic seizures. This remission has been sustained despite cautious tapering of lamotrigine to stop.


Data from a recent meta-analysis of randomised controlled trials involving new antiepileptic drugs suggested that topiramate was possibly one of the most potent new drugs.1 This audit was performed to assess whether the results of randomised controlled trials are borne out in clinical practice in a specialist regional epilepsy clinic. It is a retrospective study relying on patient reporting during outpatient clinics and, therefore, has limitations. Firstly, times to discontinuation, in some cases, had to be estimated relative to the clinic dates. Secondly, as we did not specifically enquire about adverse effects in all patients, it is likely that the figures reported are an underestimation of the true incidence of side effects, especially in patients continuing on topiramate. However, in most patients discontinuing topiramate, specific adverse effects were recorded. We have not relied on specific seizure counts but have used a logical hierarchy of improving seizure control which was easily applied to the information recorded during outpatient clinics. Whereas these limitations may have led to small inaccuracies in times to withdrawal, and underestimation of adverse effects, the study provides useful data on the pragmatic use of topiramate, which is relevant to everyday clinical practice.

We have found that at 12 months 55% of patients will still be taking topiramate. Sander, in a six month audit, found that 59% of patients remained on topiramate.8 Approximate retention rates for gabapentin (45%), vigabatrin (58%) and lamotrigine (60%) at 1 year are similar.5 However, after 6 years only one quarter of original patients will still be taking these drugs and about 3% will be seizure free.5 In a previous study from this centre, utilising a similar patient group, Schapel and Chadwick found that 57% of patients prescribed lamotrigine and 43% of those prescribed vigabatrin were still taking the drugs after 40 months, with 10% and 6% seizure free respectively.6 Pooled data from five double blind studies showed that withdrawal for any reason occurred in 20.7% of patients on topiramate and 7.4% on placebo.9Long term data regarding topiramate is limited, but evaluation of patients recruited into topiramate clinical trials in five centres in the United States found that of 214 patients, 64% were still receiving topiramate 3 months to 7 years (mean 2.5 years) after starting the drug.10

Five per cent of our patients became seizure free; these included eight patients with partial epilepsy and one with juvenile myoclonic epilepsy (case 2). Similar seizure free figures of 4% and 6% were reported in pooled data from five double blind studies,9 and in a 6 month audit performed by Sander.8 A further 15 patients ceased having generalised tonic-clonic seizures but continued having simple (n=2) or complex partial seizures (n=13) and 55 reported convincing improvement in their seizures. Despite these good responses, one of the patients ceased having generalised tonic-clonic seizures but developed severe depression requiring withdrawal (case 2), and nine patients with convincing improvements developed intolerable adverse effects. Four per cent of patients pooled from five double blind studies became seizure free, with 50% responder rates of 25% at 200 mg/day and 46% between 400 and 1000 mg.9

In animal seizure models, topiramate blocks voltage dependent sodium channels and enhances GABA evoked chloride channel currents. The second mechanism is thought to explain the ability to block spike wave discharges in the spontaneously epileptic rat, suggesting a broad range of action. This study, in agreement with preliminary data from patients with generalised epilepsies11 and the experience of other epileptologists,8 provides anecdotal evidence in support of this broad range of efficacy.

As in other studies, most patients who stopped taking topiramate did so because of adverse effects,12 13 with about 75% of patients who withdraw due to side effects alone, doing so in the first few months of treatment. Adverse events were the reason for stopping topiramate in 28% of our patients with a further 17% stopping due to inefficacy either alone or in combination with adverse events. Eleven patients (6%) reported a deterioration in seizure control in combination with adverse events as the reason for stopping the drug. There was a suggestion that patients with symptomatic generalised epilepsies were more likely to report a deterioration in seizure control than patients with partial or idiopathic generalised epilepsies, although small numbers limit any firm conclusions. In Sander’s audit, 19% stopped taking the drug due to adverse effects, 8% inadequate seizure control, and 15% due to a combination of the two.8

Common side effects of topiramate are CNS related—impaired concentration and memory, slowed thinking, and word finding difficulties.12 13 Cognitive or behavioural side effects were reported in almost 50% of patients in our study. Our limited classification makes it difficult to distinguish drowsiness and mental slowing from more severe psychiatric or behavioural changes. However, it seems that depressive symptoms, sometimes with psychotic features, can occur quite often, as reported by Betts et al.14 They found that 15% of their patients taking topiramate had severe psychiatric side effects, including unipolar depression, paranoid psychosis, acute confusional psychosis, or depression with paranoid ideation. These psychiatric symptoms were often insidious in onset but reversed in all patients seen.14 Nevertheless, these side effects may be severe and patients should be counselled as to their possibility. As we did not specifically ask about side effects, minor cognitive effects were probably not reported and are, therefore, underestimates of the true incidence. Similarly, the 10% of patients reporting anorexia or weight loss, is less than previously reported figures of about 15%. This is probably because these complaints are rarely the reason for topiramate withdrawal.

We found that patients who received topiramate as add on treatment were less likely to remain on topiramate than those in whom it was substituted for another drug. Intuitively this would seem most likely due to increased drug related adverse effects. Overall, there was a suggestion of increased adverse effects in the add on group of patients compared with substitution; however, this was not seen in the patients who withdrew from topiramate. The reason for higher retention rates when topiramate is substituted is therefore not clear; however the retrospective nature of the study and small numbers of patients undoubtedly limit ability to detect differences.

In controlled trials, upward dosage titration was rapid to satisfy dose finding objectives of the key double blind trials, which may have resulted in an artificially high incidence of adverse events. The vast majority of patients followed recommended titration regimes and, interestingly, both patients who followed faster and slower titration regimes had lower retention rates. It would seem likely that patients who followed a slower titration regime did so because the prescribing neurologist suspected that intolerance was likely, reflecting difficulties with previous antiepileptic drugs.

This retrospective audit reflects the pragmatic use of topiramate, by neurologists experienced in its use, in a population of patients with previously drug resistant epilepsies. Adverse effects seem relatively common and are the main reason for early withdrawal. Patients with history suggestive of antiepileptic drug intolerance seem to be most intolerant, although patients with symptomatic generalised epilepsies may be more likely to report deterioration in seizure control. Topiramate should be titrated slowly if it is tried in these patient groups, and it seems that substitution rather than addition of topiramate for another antiepileptic drug, may improve tolerability. Although side effects are common, patients who are able to tolerate topiramate are likely to report significant improvements in their seizures, with control of generalised tonic-clonic seizures in 9% and complete seizure freedom in 5% of patients. Evidence of a broad range of activity suggests that in patients with valproate or lamotrigine resistant generalised epilepsies, topiramate may be indicated, and indeed, it has recently received a product licence for this indication.