Of the newer agents vigabatrin has been the most studied, perhaps on account of its being the first to be tested and licensed. Shortly after the introduction of viga-
batrin, a London group had published an incidence of significant psychiatric complications in 7% of treated patients (Sander et al., 1991). Thomas et al. (1996) have analysed case records of psychiatric complications, episodes of psychoses or major depression, reported to the manufacturer of vigabatrin. With respect to psychoses the authors identified three patterns: of a total of 28 psychotic patients, eleven had become seizure free with vigabatrin, six had a postictal psychosis following a cluster of seizures after initial seizure control, possibly related to tolerance, and two psychoses occurred after withdrawal of vigabatrin.
In children, particularly in association with learning disabilities, the most common psychiatric side effects are agitation and excitation, hyperkinesia and aggression, a behaviour syndrome similar to that seen with barbiturates. In an early French study the incidence of behavioural disturbances was as high as 26% (Dulac et al., 1991).
Since these early reports, the clinical significance of vigabatrin-associated behavioural problems has been a matter of controversy, prompting Ferrie to perform a meta-analysis of psychoses and severe behavioural reactions leading to drug discontinuation, in seven placebo-controlled European studies (Ferrie et al., 1996). The overall incidence of these complications was 3.4% in the vigabatrin group and 0.6% in the placebo group. Although this is a significant difference in statistical terms, the authors conclude that the risk for psychiatric complications is not increased with vigabatrin referring to published incidence rates of psychiatric disorders in epilepsy patients in general. Remarkably, the incidence rates seem to be rather different in different studies, ranging from 1% to 12% suggesting that either the risk is not the same for all patient groups or that the threshold to report psychiatric side effects is not the same among different investigators.
Another meta-analysis on the risk for psychosis has only recently been published (Levinson and Devinsky, 1999). These authors used a uniform dictionary and translated psychopathological symptoms described in the investigator forms into standardized psychiatric terminology, which was then summarized into a syndro-matic diagnosis. This analysis of American and non-American double-blind studies demonstrated that there is a significantly increased risk for psychosis and particularly for depression. Psychoses occurred in 2.5% of patients treated with vigabatrin compared to an incidence of 0.3% in the placebo group (P < 0.05), and depression occurred in 12.1% of patients treated with vigabatrin in contrast to only 3.5% in the placebo group (P < 0.001). Less than 2% of the patients were taken off vigabatrin suggesting that these were relatively mild manifestations of depression in the majority of patients. It is however important to recognize that the duration of the analysed studies was 3-4 months, later complications are therefore not included. It is a pity that this useful analysis only came out in 1999, since most epi-leptologists have almost stopped using vigabatrin because of the risk for visual field defects since 1998.
Psychiatric side effects are not restricted to patients with complicated epilepsies who receive vigabatrin as an add-on treatment. The monotherapy trials showed a significantly increased incidence of depressive disorders in 5% of vigabatrin-treated patients as compared to only 1% in a carbamazepine-treated group (Chadwick, 1999).
In contrast to vigabatrin, lamotrigine has early on gained a reputation of having positive psychotropic properties, improving both mood and cognitive functions. Severe psychiatric complications seem to be uncommon with lamotrigine, and psychosis and depression occurred only in very few cases in the trials (Fitton and Goa, 1995).
Insomnia, which may be associated with irritability, anxiety or even hypomania, is the only significant psychiatric side effect, occurring in 6% of patients treated with lamotrigine in monotherapy, compared to 2% in patients treated with carba-mazepine and 3% in patients treated with phenytoin (Brodie et al., 1995).
When there were reports that carers complained that handicapped patients became more alert and demanding, this was interpreted as reflecting inadequate rehabilitation facilities rather than being a negative side effect (Binnie, 1997). Besag refers to this as a 'release phenomenon' (Besag, 2001). There are now, however, a number of reports that children with learning difficulties and adults with mental handicap develop behavioural problems such as aggression with lamotrigine (Beran and Gibson, 1998; Ettinger et al., 1998). More recently there have been reports on the induction of a reversible Gilles de la Tourette syndrome, which in some cases was accompanied by obsessive-compulsive symptoms (Lombroso, 1999).
Felbamate is at present only used in a minority of patients, particularly with Lennox-Gastaut syndrome, due to its haematologic and hepatic toxicity. According to the manufacturer psychoses are rare, reported as serious adverse events in 0.02% of all patients treated in 1996 (Essex Pharma, personal communication). Felbamate may lead to increased alertness, inducing sleep problems and behavioural problems related to agitation in some patients, again, particularly in children with learning disabilities (McConnell et al., 1996).
Ketter et al. (1999) have specifically investigated psychotropic effects of felba-mate. They concluded from their study of 30 refractory epilepsy patients that the stimulating effects of felbamate may be beneficial or negative depending on preexisting psychopathology. Patients with baseline insomnia or anxiety experienced a deterioration in their psychic state, while other children improved.
Beyond somnolence, negative psychotropic effects have not been demonstrated in the controlled studies of gabapentin, which is a generally well-tolerated but also a relatively weak anticonvulsant. However, there are a number of studies suggesting that gabapentin may induce behavioural problems such as aggression in children with learning disabilities and adults with mental handicap (Lee et al., 1996; Tallian et al., 1996; Wolf et al., 1995). It is not clear whether this could be related to rapid titration.
A specific problem with tiagabine is the paradoxical provocation of de novo non-convulsive status epilepticus due to a relatively narrow therapeutic window (Schapel and Chadwick, 1996). Therefore, EEG investigations are necessary when behavioural problems arise. Unfortunately, this complication was discovered following the initial trials and it is therefore not possible to know whether psychiatric complications in the trials were related to underlying epileptic activity or not.
In the placebo-controlled add-on studies nervousness and depressed mood were both increased in the tiagabine group (Leppik, 1995) (12% vs. 3%, 5% vs. 1%). The incidence for serious adverse events presenting as psychosis was not significantly increased in the tiagabine group (2% vs. 1%). A total of 84 psychoses had been reported to the manufacturer in 1996. In 30 patients tiagabine was withdrawn, in 38 patients tiagabine was reduced, and in 16 cases tiagabine was continued at an unchanged dosage.
The records of 19 patients with psychoses classified as serious adverse events have been further analysed (Schmitz, unpublished data). Psychoses occurred after a variable duration of treatment with tiagabine, with a mean of 267 days (range 13-606 days). The mean dosage was 48 mg/day (range 8-80 mg). Seven psychoses occurred postictally. Only one patient had an alternative psychosis following seizure control and one patient became psychotic following tiagabine withdrawal. There was no systematic EEG monitoring in these cases, so nonconvulsive status cannot be excluded and may have gone unrecognized in some cases. The psychoses were of a paranoid-hallucinatory type in 12 cases and duration was less than a week in 10, and less than a month in 4 cases. In summary, for psychoses with tiagabine there are no distinguishable patterns, as described for vigabatrin.
Topiramate, promising 'polytherapy with a single drug' because of three different modes of action, is a highly effective anticonvulsant, working against a broad spectrum of seizure types. Topiramate also has a high rate of reported side effects. These might in part relate to rapid titration schemes in the earlier studies. However, there is no evidence so far that severe psychiatric complications can be avoided by slow titration. An unusual idiosyncratic side effect of topiramate is amnesic or motor aphasia, and in the controlled trials 17-28 % of patients developed symptoms classified as 'abnormal thinking' (Janssen-Cilag, 1996).
The rate of affective symptoms is clearly dose-dependent with an incidence of 9% and 19% with a daily dose of 200 mg and 1000 mg respectively in one clinical study (Janssen-Cilag, 1996). The high incidence of depressive syndromes with topiramate may be related to cognitive side effects, which are particularly common with this drug. But this relationship has not been studied.
Trimble et al. (2000) have looked at the clinical data from a series of patients with psychoses and depression related to topiramate. In this analysis there was no specific pattern for the precipitation of psychoses. Interestingly, depressive syndromes were related to a complete control of seizures in five cases. They also suggested that, in contrast to vigabatrin, at least some of these psychoses were associated with high serum levels, and possible intoxication.
Levetiracetam is the latest drug, which only recently has been launched in Europe. Data on psychiatric effects of levetiracetam are limited but a preliminary analysis suggests that there are no significant psychiatric risks associated with this drug (Trimble, 2000). Affective disorders were reported in 0.02% and psychosis in 0.007% of patients treated with levetiracetam in clinical trials.
Was this article helpful?