It has now been established that antiepileptic drug treatment may be associated with a variety of side-effects. Some effects appear immediately after the start of drug exposure, such as nystagmus, but are relatively benign because they show habituation, or are reversible when they are dose dependent. Others may be of insidious onset, emerging only after extended periods of treatment (i.e. chronic side-effects). A multitude of such chronic side-effects have been documented, but the most frequently reported effects concern central nervous system effects. This post reviews some of our knowledge about a specific subgroup of such central nervous system-related chronic side-effects of antiepileptic drug treatment, that is, cognitive side effects: the adverse effects of drug treatment on information-processing systems.

Such effects are considered to be much more moderate than for example, some of the idiosyncratic reactions to drugs and normally do not lead to discontinuation of drug treatment. Nonetheless, a number of studies have claimed that the drug-induced cognitive impairments may have a much greater impact on daily life function than had hitherto been suspected, for example through the impact on critical functions, that is, learning in children or driving capacities in adults (often requiring milliseconds precision), or on vulnerable functions such as memory function in elderly. Moreover, as the cognitive side-effects represent the long-term outcome of antiepileptic drugs, the effects may increase with prolonged therapy, which may contribute to the impact on daily life functioning in patients with refractory epilepsies.

Review of psychometric studies

Clinical effects

Subjective patient complaints

Conclusion

Systematic analysis of subjective patients complaints about side-effects of antiepileptic drugs show that the impact of side-effects maybe larger than hitherto suspected both in number of patients involved (our community-based sample suggests that almost 60% of the patients with antiepileptic drug have complaints) and the frequency of the complaints. Especially the behavioral (and within this class the cognitive) side-effects occur frequently and require careful monitoring and possible interventions.

Still using subjective patient complaints it is clear that a switch from monotherapy to polytherapy entails a serious risk of increasing side-effects. This has been reported from clinical groups, in patients with refractory epilepsy and within the context of many drug trials (most recently for topiramate), but is now also confirmed in a community-based sample.

Table Differences per area of complaint between the four groups: valproate, carbamazepine, combination of phenytoin and polytherapy. Results from a community-based study in 346 patients

On the other hand, Bourgeois reports reduction of side effects when reducing polytherapy; this is considered proof for a partially cumulative toxic effect. When in clinical decision-making the option of polytherapy arises, the serious risk of an increase in side-effects should be taken into consideration carefully. This is especially important in the light of recent revivals of polytherapy, for example, within the context of rational polytherapy.

A similar effect is often observed when new drugs proceed from initial add-on studies to studies in monotherapy. Although the efficacy profile often remains unchanged, the tolerability profiles often have to be adjusted with much more moderate profiles in monotherapy.

Formal psychometric studies are much more difficult to interpret, especially when formal scientific standards in line with evidence-based medicine are applied.

Nonetheless, we may claim that a systematic review supports these conclusions. It may be considered conceivable that polytherapy increases the risk of behavioral and specifically cognitive impairments. We may therefore hypothesize a potentia-tion of tolerability problems leading to cognitive impairments due to interactions between antiepileptic drugs. This seems to be a general effect as it occurs in many combinations of drugs and so far no specific combination has been identified.

1   English-language report of original research, published in peer-reviewed journals in the period 1970-1994; studies after 1970 were all done at a time when most of the current antiepileptic drugs had become available and modern cognitive tests had come into widespread use.

2  Studies that report psycho metrically assessed cognitive functions (excluding for example clinical observations).

3  Only current antiepileptic drugs (excluding experimental drugs that have been removed from study programmes, such as zonisamide, felbamate or flunarizine).

4  Only studies on patients with epilepsy (excluding antiepileptic drug studies in for example psychiatric patients). The resulting meta-analysis has been published on the data concerning monotherapy. Here we focus on the results for combination therapy or polytherapy.

In the meta-analysis, studies were classified into the polytherapy category if subjects were treated with more than one drug at a time and no comparisons between individual drugs, or single drug vs. no antiepileptic drug, were possible. Studies that were identified through the aforementioned procedure and involving polytherapy are listed in Table Review summary of polytherapy polytherapy studies (1970-1994)

1   Treatments. This section shows the treatment conditions associated with assessment points. The nomenclature and abbreviations for individual antiepileptic drugs comply with the recommendations in Epilepsia, 1993,34,1151. In addition: P, polytherapy; SAD, single additional dose; Mono, monotherapy; plac, placebo; none, no antiepileptic drugs. Subscripts: AED_cr, controlled release formulation; AED_Dhi/Dme/Di₀, high, medium, low dosage; AED_shi/si₀, high vs. low serum (or saliva) levels; P_red,mod> P reduction [c.q.], other modification; Paed+/-> P ^{wmi vs}- without a particular antiepileptic drug; P_tox+/-> P with toxic vs. non-toxic serum levels. Slashes (/) indicate contrasts under study in a parallel group or post-test-only design. Crosses (X) indicate crossover elements. Arrows (—>) indicate change of one treatment to another. Plus signs ( +) indicate that medication is added to an existing regimen.

2  Number of subjects. The numbers are shown separately for each treatment condition and untreated controls; they indicate the number of subjects who completed the trial and for who test data were available. A range is given for n when not all subjects completed all tests. Occasionally, we were unable to determine these numbers for the separate treatments (e.g. when only an overall n was provided), or for one or more outcome measures. This is indicated by a question mark.

3  Drop-out rate. This gives a rough indication as to whether a selection artefact might have developed during the trial. An overall rate is given separately for subjects on antiepileptic drugs and untreated controls. About half the studies reviewed mention dropout losses and present sufficient data to compute a drop-out rate for each outcome measure. A range may be given here as well as due to incompletion in various degrees. A few studies explicitly state that no dropout losses occurred (-). In others, dropout losses are mentioned but insufficient data is provided to compute a loss rate (?). Often, dropout losses or their absence are not mentioned (n.m.)> which may or may not mean that no such losses occurred. Sometimes a minimum rate is quoted (5^s), more subjects may have been lost, but the data are unclear or ambiguous in this regard.

4  Design. (Table 22.2 gives an overview of the general design types encountered.) The term design as used here refers to the scheduling of treatments (i.e. antiepileptic drugs, placebo, no treatment) and outcome measurement sessions, and to the way subjects were assigned to treatment groups (i.e. on a random basis or not). Occasionally, we were unable to discover a consistent principle underlying

Table Review summary of polytherapy polytherapy studies (1970-1994)

Randomized treatment allocation, or treatment sequencing in a crossover and parallel design, is indicated by the suffix (R).

Table Design nomenclature and classification

Randomized treatment allocation, or treatment sequencing in a crossover design, is indicated by the suffix (R).

The scheduling of treatments and assessment points, or different schedules were employed for different subjects. In such cases the design was classified as unclear (?).

5  Number of cognitive variables. This gives an indication of the possible scope of the study with respect to cognitive functioning; also, this is a statistically relevant characteristic. Uncertainty as to the number of variables actually employed (?) in analysing the data may occur even if the tests used are mentioned; often multiple outcome variables may be derived from a single test (e.g. response speed, accuracy, subscales in intelligence tests).

6   Time on antiepileptic drug. This characteristic is important in judging the relevance of the results to chronic antiepileptic drug use. Its meaning depends on the particular design employed. In a post-test-only design the figures quoted relate to the duration of treatment prior to the assessment point. In repeated measurement designs with one or more groups (i.e. single and parallel) this refers to the duration of the experimentally changed antiepileptic drug-treatment or the continuous medication interval studied. With multiple assessment points during the trial, the maximum interval studied is given. In a crossover design where multiple antiepileptic drugs or dosages are given, this refers to the time on each antiepileptic drug [c.q.] dosage.

Methodological considerations

Closer inspection of the studies that we identified shows many methodological problems, most of which are inherent to polytherapy as such. These methodological problems must be taken into consideration carefully because they restrict the validity of the information from these studies.

Treatment reproducibility

Polytherapy is by nature a heterogeneous treatment category; thus, one finds treatment descriptions such as ‘various combinations of the three major antiepileptic drugs’ or ‘combination of phenytoin and one or more other antiepileptic drugs’ [c.q.], ‘drug regimens exclusive of combination of phenytoin’, or even ‘no attempt was made to standardize drug therapy as part of the study’. Obviously, widely different drug regimes would fit such descriptions, and results established with one regimen may not apply to another. Also, the polytherapy manipulations used in many studies are actually quite complex, making replication problematical. For example, all polytherapy reduction studies are done as part of individualized programs of therapy rationalization. That is, patients did not have their medications changed for research purposes, and different types of medication change were not subjected to randomization. Rather, changes were typically made ‘according to the individual needs of each patient’. The clinical considerations underlying the medication changes are a major ingredient of the treatment package, albeit one that may not be easily reproduced.

Drug interactions

Combinations of antiepileptic drugs may alter metabolism to produce changes in the level of active and/or toxic metabolites. Examples include the decrease in carbamazepine levels due to the increased elimination of the drug when given together with combination of phenytoin and/or phenobarbital. Such interactions can alter seizure control efficacy and maybe relevant to cognitive functioning. With multiple drugs, identifying the components of a treatment most responsible for any observed effects presents a difficult problem.

Serum concentration-effect relationships

Cognitive antiepileptic drug effects may be examined through an analysis of the relationship between test scores of subjects and their individual serum drug levels, and this approach seems to offer a way out of the problem mentioned above. In fact, a number of studies report such relationships suggesting that, generally, higher serum levels are associated with lower cognitive scores. However, in patients with epilepsy, higher serum concentrations maybe the reflection of higher antiepileptic drug doses prescribed for more severe epilepsy, perhaps with seizures not fully controlled. Also, antiepileptic drugs may interact on receptor sites (pharmacodynamics), which would not necessarily be reflected in the pharmacokinetics expressed in serum concentrations. Such factors greatly reduce the interpretability of relationships between serum concentrations and cognitive performance.

Seizure confound

Polytherapy is typically given to patients with refractory epilepsy, and separating seizure effects from antiepileptic drug effects may thus be very difficult, particularly in add-on studies, where the cognitive evaluation is usually made in connection with an efficacy trial. That is, adverse cognitive antiepileptic drug effects may be masked by beneficial effects of better seizure control. Also, patients with refractory seizures may not be representative of the general population with epilepsy.

Discussion of cognitive effects

Due to the validity threats described above, acting singly or simultaneously, drawing conclusions about the cognitive effects of polytherapy studies is not without complications. Moreover, the anecdotal-type of information is best illustrated by the large number of question marks both in the column expressing numbers of patients, for the dropout rates and even for design, number of cognitive variables, and time on antiepileptic drugs. Starting from the principles of evidence-based medicine we can therefore only proceed carefully. Conclusions will be drawn on a general level. Reviewing the literature, five types of studies can be distinguished:

•  The first type of study is the single measurement polytherapy study. Corbett et al. (1985) is an example of studies that analyze polytherapy in a single measurement design. Patients who received polytherapy are analysed for cognitive impairments. Although, without exception all these studies report severe cognitive impairment the design does not allowthe isolation of drug effects from the effects of the epilepsy.

•  The second type consists of studies comparing monotherapy with polytherapy. Brodie et al. (1987) showed no difference between monotherapy carbamazepine, valproate, combination of phenytoin and polytherapy at a single assessment study. Other studies did, however, show serious impairments for polytherapy. Bittencourt et al. (1993) used a complex add-on with a polytherapy at baseline (with either combination of phenytoin or carbamazepine added to an existing low-dose phenobarbital regime) and monotherapy (combination of phenytoin or carbamazepine) at endpoint. The study shows statistically significant improvements on measures of memory and attention after withdrawal from polytherapy. As for the former type of study, it is extremely difficult to avoid the seizure confound here as polytherapy is mostly given to different patients.

•  A convincing group of studies showed the effect of reduction of polytherapy. Durwen etal. (1989) showed that reduction of polytherapy resulted in improvements of verbal memory. Duncan etal. (1990) used a rather interesting design in which separate drugs (combination of phenytoin, carbamazepine, valproate) were removed from polytherapy regimes showing consistent improvements in cognitive function, irrespective of the type of drug that was discontinued. Thompson and Trimble (1980,1982) Ludgate etal.(1985) and Van Rijckervorsel etal. (1990) are other examples of studies that showed marked improvement after reduction of polytherapy.

•  In contrast, the fourth type of study does not show convincing effects of polytherapy. In these add-on studies a new drug is added to either monotherapy or to an existing polytherapy. Berents et al. (1987) showed impaired verbal learning when a new drug was added to an already existing polytherapy. Most other studies, however, showed no effects of newer drugs to an existing polytherapy. These are, however, all studies within the context of drug trials in refractory epilepsy, where the added effects of a new drug are difficult to entangle from the beneficial effects of improved seizure control.

•  Finally, the last type of study that we could distinguish analyzed the relationship between cognitive impairment in polytherapy with serum level. Dekaban and Lehman (1975) claim a relationship but the study does not control interfering factors such as dose and seizure confound and, hence, does not guarantee valid interpretation. The same situation occurs for other studies such as Reynolds and Travers (1974), Matthews and Harley (1975) and Thompson and Trimble (1983).

The existing evidence from especially the reduction studies, therefore suggests the possibility of potentiation of tolerability problems in polytherapy and specifically an increase of cognitive problems. It may therefore be hypothesized that drug interactions maybe responsible for this potentiation. This seems to be a general effect as it occurs in many combinations of drugs and so far not a specific combination has been identified.

Nonetheless, we may look at more anecdotal clinical information. In many drug trials a similar effect has been found as suggested by the psychometric studies: a higher incidence of side-effects in combination therapy when compared to the same drug in monotherapy. This is often observed in post-marketing studies. Many of the new drugs are first tested in add-on designs and monotherapy is only used later.

Table Percentage of side-effects for topiramate (topiramate) in polytherapy vs. monotherapy; potentiation of side-effects due to polytherapy?

Side-effects in combination therapy but also different types of dominant complaints when compared to monotherapy.

When inspecting this example we see cognitive impairments are reduced in monotherapy, but other impairments, especially paresthesia, are increased. This may lead to the hypothesis of potentiation of some tolerability problems and especially cognitive side-effects.

Of course, it is imperative to emphasize that the monotherapy studies typically include patients with other epilepsies compared to the initial add-on studies, which are done in refractory partial epilepsies with an associated risk of epilepsy-induced cognitive impairments. Moreover, in contrast with the former paragraph of this chapter, these side effects have not been established with formal psychometric or other objective measurements, but are based on clinician ratings of subjective patient complaints.

Table Subjective reported side effects in 346 patients in a community-based study

Taking advantage of the reliable databases of antiepileptic drug use in the pharmacies in the Netherlands we were able to establish a non-selected unbiased community-based study group of adult patients with epilepsy from a suburban area (100 000 inhabitants) with a prevalence equivalent of 0.4% (i.e. 346 patients). All patients finished a rating scale on side-effects of their treatment. Almost half the patients reached a 2-year seizure remission; about one-third considers the seizures unacceptable. About 80% of the patients are on monotherapy. Nonetheless, almost 60% of the patients report side-effects in at least three areas.

Table Subjective reported side effects in 346 patients in a community-based study shows that the two areas with clearly most reports are: (a) general central nervous system-related complaints (such as fatigue and dizziness) with 68% complaints; and (b) cognitive complaints (61.8%). If we combine cognitive and mood areas, then behavioural complaints are, with 84.1% complaints, the dominant complaint in our study group. Within the areas, two types of complaint have been reported by >20% of the patients: memory problems (21.4%) and fatigue (20.3%). Two other complaints are reported by between 15% and 20% of the patients: tiredness (18.8%) and concentration difficulties (16.1%). It is thus clear that cognitive complaints are a dominant complaint even in a group of patients with a well-controlled epilepsy, mainly using monotherapy.

Subsequently, differences in side-effect profile were tested per antiepileptic drug. This was only possible for four groups with >50 patients, that is, patients on monotherapy of valproate, carbamazepine or combination of phenytoin and patients on polytherapy (28.7%, 24.7%, 15.7% and 19.1% of the study group respectively). All remaining groups were too small to achieve sufficient statistical power. Table Differences per area of complaint between the four groups: valproate, carbamazepine, combination of phenytoin and polytherapy shows exclusively the 7 areas with statistical differences between the four groups. On all areas the differences were caused by the higher percentage of complaints in patients using polytherapy. The remaining differences show more concentration difficulties for combination of phenytoin compared to valproate, more weight gain for both valproate and combination of phenytoin compared to carbamazepine.

The relationship between serum concentration, seizure frequency and side effects gives the therapeutic range or target range (Table Antiepileptic drug therapeutic ranges). In this range, an antiepileptic drug can be considered to be associated with no dose-dependent side effects in the majority of the patients in whom it is effective. Thus the range provides guide values which give a more rapid identification of a patient’s individual therapeutic range, which reflects the patient’s clinical susceptibility to seizures, seizure type, etc. However, despite a good deal of anecdotal testimony, surprisingly little has been published demonstrating the benefits of anticonvulsant therapeutic drug monitoring in epileptic populations. In two randomized controlled trials on the clinical impact of therapeutic drug monitoring in patients with epilepsy, the implementation of serum antiepileptic drug level monitoring did not improve overall therapeutic outcome, and the majority of patients could be satisfactorily treated by adjusting dose on clinical grounds. In the study by Camfield et al. (1985), 82 newly diagnosed children were started on antiepileptic drug therapy and followed prospectively for 12-36 months. Serum concentrations were not observed to be different between children who had a relapse and children who continued to have seizure control; concentrations were within the therapeutic range in both groups. In a study of antiepileptic drug side effects in 197 patients, Laplane and Carydakis

Table Antiepileptic drug therapeutic ranges

“During oxcarbazepine treatment it is only necessary to determine the pharmacologically active monohydroxy derivative (MHD) metabolite. ^b During primidone treatment it is only necessary to determine the phenobarbital concentration.

(1985) concluded that the recording of side effects from a case history and a physical examination is far more useful than the determination of the serum concentration. By these results, on the one hand the value of a routine serum concentration determination is doubtful, on the other hand most epileptologists are convinced that monitoring improves the pharmacotherapy of the epilepsies even if there is a lack of statistically significant results. At present, the determination of antiepileptic drugs is considered to be indicated in the situations which are listed in Table Indications for the determination of antiepileptic drugs in serum. These indications are valid not only for the established drugs but also for the new antiepileptic drugs, with the exception of vigabatrin, which is associated with a somewhat unusual mechanism of action whereby its pharmacological effect long outlasts its serum concentration.

Indications for drug monitoring in antiepileptic combination therapy

Prerequisites of the determination of antiepileptic drugs in serum

The essential prerequisite for monitoring serum concentrations, both in monotherapy and in combination therapy, is the reliability (accuracy and precision) of the assay methodology. Another important factor is the consideration of the time interval between drug intake and the collection of blood. Thus if a drug has a short half-life value, the blood sampling of a patient in an outpatients setting often does not result in the determination of a trough concentration. In combination with an enzyme-inducing antiepileptic drug, the half-life value of tiagabine, for example, is reduced from 5-8 h (monotherapy) to 2-5 h. When phenobarbital and valproic acid are administered in combination, the half-life value of valproic acid is reduced while the half-life of phenobarbital is increased and its peak concentration is delayed.

Also, as in the case of monotherapy, one has to take into consideration whether or not the measured serum concentration is reflective of steady state. When the dosage is changed, the new serum concentration should not be determined until steady state has been achieved (typically this can be expected to occur after five half-life values of the affected drug).

Measurement of the free (non-protein bound) concentration of antiepileptic drugs

Some antiepileptic drugs are highly protein bound in blood to albumin. These include carbamazepine (80%), phenytoin (90%), tiagabine (96%) and valproic acid (95%). Some antiepileptic drugs may interact at the albumin protein-binding site and free concentration may increase; for example free carbamazepine-epoxide concentrations are significantly increased in patients taking carbamazepine plus valproic acid. Since only the free concentration can penetrate the blood-brain barrier, several attempts have been made to improve treatment by using free drug concentration measurements; for example by determining antiepileptic drugs in saliva or by measuring the free concentration in blood by ultrafiltration. The clinical relevance of the determination of the free antiepileptic drug concentrations is somewhat controversial, even for highly protein bound drugs. Nevertheless, overall there are clinical settings where patient management would best benefit from measurement of free serum concentrations. Indeed, when phenytoin and valproic acid are co-prescribed, measurement of total phenytoin concentrations would be misleading because of the protein-binding displacement interaction that occurs between these antiepileptic drugs.

Limits and dangers of the determination of serum concentrations of antiepileptic drugs

The effectiveness of serum concentration determination depends on the accuracy of the therapeutic ranges of the individual antiepileptic drugs. Therefore, for phenytoin with its narrow therapeutic range the indication for serum concentration monitoring is much clearer than, for example, phenobarbital, for which the upper limit of the therapeutic range is very blurred.

As for the new antiepileptic drugs, the clinical relevance of drug monitoring is limited by the fact that the target ranges are derived from the clinical trial data collected during their evaluation as add-on therapy. These studies enrol patients that are highly selected and as such are not reflective of patients seen in general clinical practice. Nevertheless, these ranges are valuable as they provide reference values, which can aid patient management. Serum monitoring of new antiepileptic drugs is particularly useful when interacting antiepileptic drug combinations are co-prescribed.

The appropriate procedure for adjusting the dosage is essentially the same irrespective of whether or not the serum concentration is determined. In both cases the dose is increased slowly until seizures cease or, if seizures persist, until the intoxication limit is reached. The upper limit of the serum concentration is merely a warning signal, not a barrier. The procedure must be similar because the individual therapeutic ranges are different. The lower limit of the target range or therapeutic range is even more indistinct than the upper limit.

It is imperative that therapeutic ranges are appreciated for what they are, that is statistical ranges whereby if a patient achieves a serum concentration within that range, the probability of achieving a desirable therapeutic response (with minimal side effects) is high. A serum concentration should be used in conjunction with the knowledge of the patient, the clinical information and also the pharmacokinetic characteristics of the drug. The therapeutic range should be used as a guide.

Special laboratory tests

Sometimes when there is an interaction associated with antiepileptic drug treatment, a surrogate physiological marker may be necessary other than the measurement of a serum drug concentration, so as to ascertain the interaction. An example of this is the combination of phenprocoumon or warfarin and an enzyme-inducing antiepileptic drug, whereby the serum concentration of phenprocoumone/warfarin will decrease. In this setting the careful monitoring of the prothrombin time (or internationalized normalized ratio, internationalized normalized ratio) is indicated. Another example is that of valproic acid-induced encephalopathy. Valproic acid occasionally induces stuporous or comatose states. This encephalopathy is frequently accompanied by hyperammonemia without signs of hepatic failure. Valproic acid encephalopathy may occur after initiation of valproic acid as monotherapy or, more often, in combination with other antiepileptic drugs (e.g. phenobarbital, phenytoin, topiramate). In this setting monitoring of plasma ammonia concentrations would be a valuable diagnostic tool.

Future of monitoring

Therapeutic drug monitoring has greatly enhanced the treatment of epilepsy in that it has allowed the individualization of treatment and thus maximized the desirable anticonvulsant effects of antiepileptic drugs while keeping to a very minimum the undesirable side effects that are associated with antiepileptic drugs. Monitoring is essential for drugs with a narrow therapeutic index and for those drugs with unacceptable incidence of toxicity. This is particularly exemplified by phenytoin. With regards to the new antiepileptic drugs, target ranges are at their infancy and will inevitably require fine tuning as our clinical experience with these drugs increases. The usefulness of drug monitoring during combination polytherapy is dependent on the propensity of the combination to interact and the extent and mechanism of the interaction. For pharmacodynamic interactions, drug monitoring is only of value in that it is used to exclude the possibility that an interaction is pharmacokinetic in nature. For many pharmacokinetic interactions, the direction and extent of the associated change in serum drug concentration cannot be reliably predicted. Furthermore, some interactions may be associated with either an increase or a reduction in serum concentration. The likelihood of an unpredictable interaction is much lower with some of the new antiepileptic drugs compared with the established drugs. Indeed gabapentin and levetiracetam have a particularly low propensity to interact. Therefore, the role of drug monitoring will not be important during combination therapy with these antiepileptic drugs.

Summary

Combination therapy is an important indication for antiepileptic drug therapeutic drug monitoring as it is an invaluable aid in avoiding underdosage and intoxication and also to confirm drug-related side effects.

If it is not possible to conclude from the case history which component of a combination is probably effective, the determination of the serum concentration helps identify which drug dosage should be increased or lowered. In this setting one has to take into account the possibility of drug interactions. In particular, consequent to a pharmacokinetic interaction, the addition or withdrawal of a drug to or from a combination may result in a considerable shift in serum concentration. The early detection and correction of such a shift may help to prevent, for example, underdosage with recurrences of seizures. With the combination of a medium dose of lamotrigine and a low dose of valproic acid, the withdrawal of valproic acid may induce seizures not because of the low dose of valproic acid but as a consequence of the decrease of the lamotrigine serum concentration. When combination therapy has led to intoxication symptoms the clinical picture and the electroencephalograph are often not helpful in recognizing the responsible drug. In sedated patients taking a combination of valproic acid and phenobarbital, drug monitoring can help to differentiate between valproic acid encephalopathy with a ‘normal’ serum concentration and phenobarbital intoxication as the consequence of an inhibitory interaction between phenobarbital and valproic acid with a consequent elevation of phenobarbital serum concentrations.

Table Indications for the determination of antiepileptic drugs in serum

Avoidance of underdosage

One of the most frequent prescription errors relates to treatment with two drugs that are below the effective dose. If it is not possible to tell from the case history as to which drug is effective, the determination of the serum concentration can provide invaluable information as to which of the two drugs should be adjusted in relation to its dosage. Thus if a patient continues to have seizures on drug A and a drug B, one will determine the serum concentration of both drugs. If the serum concentration of A is within the target range and the serum concentration of B is below, then a first therapeutic step would be to increase the dosage of drug B. In contrast, in a seizure-free patient with the same treatment regimen described above, one would reduce the dosage of drug B and keep the dosage of drug A the same. Undertaking such a step-wise approach also serves to take into account the possibility of pharmacokinetic interactions. Owing to the possibility of drug interactions, the addition or withdrawal of a drug to or from a drug combination may result in a considerable shift in serum concentration. The immediate detection and correction of such a shift may help to prevent underdosage with recurrences of seizures or intoxication. If a patient is seizure free with a combination of lamotrigine with a medium serum concentration and a low serum concentration of valproic acid, the withdrawal of valproic acid may induce seizures, not because of the high efficiency of the low dose of valproic acid but because of the drop of the lamotrigine serum concentration (valproic acid is an enzyme inhibitor, capable of reducing the rate of metabolism of the co-administered lamotrigine; Patsalos et al., 2002). Although valproic acid withdrawal may be undertaken without knowledge of lamotrigine serum concentrations, their knowledge would allow for a more gradual and predictive therapeutic response.

Another setting whereby underdosage can occur is when an antiepileptic drug, whose metabolism is susceptible to enzyme induction, is co-prescribed with an enzyme-inducing antiepileptic drug. If, for example, in a patient with focal epilepsy the occurrence of tonic-clonic seizures has been stopped by valproic acid monotherapy but complex focal seizures persist, one could add the enzyme-inducing drug carbamazepine. This might induce an increase of seizure frequency by an acceleration of the elimination of valproic acid. The cause of the deterioration of the seizure frequency will be explained quickly by measuring the serum concentration of valproic acid. In another patient, the same combination might not increase the seizure frequency but induce toxicity by an increase of the serum concentration of the carbamazepine-epoxide metabolite. The clinical significance of this interaction is particularly important in children, where high concentrations of carbamazepine-epoxide have been observed, along with severe side effects such as vomiting and tiredness. This example also demonstrates the value of monitoring a pharmacologically active metabolite such as carbamazepine-epoxide in special situations.

When an enzyme-inducing antiepileptic drug such as phenytoin is withdrawn from a treatment regimen, one needs to take into account the consequent de-induction. If phenytoin is discontinued abruptly rather than gradually, the valproic acid concentration will increase to its new steady-state level at about 1 week later. In contrast, a drug with a longer half-life (e.g. phenobarbital) would take longer to reach its new steady-state concentration. In these complex situations, monitoring of serum concentration provides invaluable information for the optimal management of patients.

Antiepileptic drug underdosage may also be induced by the addition of a drug which has been indicated for a non-epilepsy-related condition. If, for example, gabapentin is co-ingested with hydroxides of aluminium or magnesium (antacids), the absorption of gabapentin will be reduced. The extent of the reduced absorption is best ascertained by measuring the serum concentration.

Avoidance of intoxication and identification of suspected side effects

Clinically relevant examples of interactions between antiepileptic drugs with the consequence of an increase of the serum concentration of one of the drugs (or both drugs, e.g. when phenytoin and phenobarbital are co-prescribed) are the combination of the enzyme-inhibiting valproic acid and phenobarbital (increase of phenobarbital), valproic acid and lamotrigine (increase of lamotrigine), topiramate and phenytoin (increase of phenytoin), sulthiame and phenytoin (increase of phenytoin), and the aforementioned combination of valproic acid and carbamazepine (increase of carbamazepine-epoxide).

When combination therapy is associated with symptoms of intoxication, the clinical features and the electroencephalogram (electroencephalograph) are often not the right tools to ascertain which antiepileptic drug is responsible. When non-specific neurological symptoms such as tremor and ataxia appear, it is not always clear whether the symptoms are due to intoxication or to the underlying disease, particularly if the disease is progressive. The same is true for psychic symptoms. The difficulty of distinguishing between the symptoms of an underlying disease and intoxication is illustrated by the following case history of a patient on phenytoin monotherapy. We observed a 55-year old epileptic patient with a psychosyndrome, which had been interpreted as postcontusional by the doctor in attendance. The patient was disoriented and decelerated. He was treated with 300 mg phenytoin per day. The neurological examination was hindered because the patient did not cooperate. Antiepileptic drug side effects were not expected by the doctor in attendance. The phenytoin level was 53 µg/ml. In reality the psychosyndrome thought to be postcontusional was a pharmacoge-netically associated psychosis which disappeared after the withdrawal of phenytoin.

In ambiguous cases of intoxication in which the concentration of the administered drug is ‘normal’ and the drug has a pharmacologically active metabolite, the measurement of the metabolite, for example phenylethylmalonamide during primidone therapy or carbamazepine-epoxide during carbamazepine therapy, may be diagnostically helpful. Indeed it should not be forgotten that with antiepileptic drugs such as carbamazepine and primidone the alleged monotherapy in fact comprises of two or three components (carbamazepine, carbamazepine-epoxide; phenobarbital, phenylethylmalonamide, primidone). Another consideration in doubtful cases of intoxication is the possibility of the intake of drugs which have not been prescribed including over the counter supplements and herbal remedies. Such drugs can only be detected and quantitated by methodologies that are based on basic analytical principles such as high performance liquid chromatography. The use of reagent-based commercial analysis such as enzyme immunoassay may not be sufficient to identify the drug that is responsible for the clinical presentation as the following case history demonstrates. A 28-year old female patient with focal epilepsy was admitted to the hospital because of gait ataxia. The only prescribed drug was carbamazepine. Therefore, at admittance, only the serum concentration of this drug was determined. The carbamazepine value was in the lower part of the target range. With some delay we detected that the patient had continued to take phenytoin which she should have stopped 18 months before. The subsequently measured phenytoin concentration was 60 µg/ml, which was far above the target range.

When antiepileptic drugs are co-prescribed with drugs for other indications, intoxication can similarly occur. For example, when carbamazepine and erythromycin are co-prescribed, the measurement of carbamazepine serum concentration is indicated because this combination can be associated with a significant increase in carbamazepine concentrations; the extent of which cannot be predicted reliably.

Monitoring of concomitant medication

Another consideration arising from the combination of antiepileptic drugs and drugs used for other indications maybe an underdosage of the latter drugs. If one combines, for example, cyclosporine with an enzyme-inducing drug such as carbamazepine, phenobarbital, phenytoin or oxcarbazepine, careful monitoring of cyclosporine is necessary because serum cyclosporine concentrations can be expected to decrease. Thus in patients that have undergone an organ transplantation, this interaction would quickly result in a rejection of the transplanted organ.

Ever since they first appeared, antiepileptic drugs have not infrequently been used to treat patients with conditions other than epilepsy. Some antiepileptic drugs, e.g. phenytoin (combination of phenytoin), carbamazepine and valproic acid (valproate), have for long been indicated in a number of neurological and psychiatric disorders. The same is true for some of the new generation antiepileptic drugs, such as gabapentin, lamotrigine, levetiracetam,oxcarbazepine,tiagabine,topiramate and pregabalin. It seems that newer antiepileptic drugs – compared with the older ones – may be at least equally effective in non-epileptic disorders, but with fewer adverse events, and with minimal or no drug interactions. However, it should be stressed that evidence-based medicine varies broadly as far as the efficacy of particular drugs in given disorders or health conditions is concerned. Moreover, the number of reports and trials, and the extent of usage of these drugs vary greatly.

Epilepsy with its prevalence of about 1% is one of the most common neurological conditions. However, because antiepileptic drugs have been used in several other neurological and psychiatric conditions with a higher prevalence than epilepsy, altogether they present a large market for antiepileptic drug usage. For example, in the United States lamotrigine, topiramate and Gabapentin use – in terms of pharmaceutical market (IMS, 2000,2001,2002) -is greater for non-epileptic disorders than for epilepsy; moreover, there is an increasing trend for use of topiramate and lamotrigine from 2000 to 2002 (Table Antiepileptic drugs used in non-epileptic disorders in 2002 (first 9 months of the year)). In European countries (e.g. France, Germany, Italy, Spain and UK) Gabapentin use is also greater in other fields than epilepsy, and as in the USA this trend is increasing. Two other antiepileptic drugs (topiramate and lamotrigine) have been, also, showing an increasing use in non-epileptic disorders in Germany and Spain (IMS, 2000,2001,2002).

A national survey in the USA showed that approximately 10% of nursing home residents were taking antiepileptic drugs, usually with other drugs. In 18% of the residents receiving antiepileptic drugs, indications were other than epilepsy. In Poland in 2002, the number of valproate prescriptions for non-epileptic disorders was about 21%. Thus, the term antiepileptic does not reflect the whole spectrum of these drugs’ potential therapeutic effects.

Table Antiepileptic drugs used in non-epileptic disorders in 2002 (first 9 months of the year)

It is estimated that combination therapy occurs in about 10% of the general population, and in the elderly and in women the percentage is even higher. Patients over 65 years use 2-6 prescribed medications, and 1-3.4 over-the-counter drugs. Therefore, knowledge of possible drug interactions in non-epileptic patients taking antiepileptic drugs is just as important as it is in epileptic patients.

The purpose of this chapter is to emphasize the spectrum and scale of antiepileptic drug usage in medical disciplines other than epilepsy, and to increase awareness of unpredicted drug interactions when combination therapy with two or three drugs is used. As in the treatment of epilepsy, awareness of possible drug interactions with antiepileptic drugs is an important part of the treatment strategy.

Rationale for using antiepileptic drugs in disorders other than epilepsy

There are a number of pharmacological reasons why antiepileptic drugs have therapeutic effects in non-epileptic neurological and psychiatric conditions. With the exception of two antiepileptic drugs (tiagabine and vigabatrine ) that are thought to have a single mechanism of pharmacological action, all other antiepileptic drugs have multiple neurophysiological and neurochemical actions. Even when various mechanisms are involved, the primary mechanism of reducing high frequency firing in neurones is by enhancing sodium channel inactivation. Indeed, this mechanism may be one of the main reasons for the antineuralgic effects of antiepileptic drugs in various pain syndromes. However, antiepileptic drugs have several other ways of modifying abnormal neuronal activity, presumably involved in a number of neurological and psychiatric disorders, which at first glance are very different from those found in epilepsy. These mechanisms involve: inhibition of the sodium, L-,N-,T-calcium and chloride channels; blockage of the N-methyl-D-aspartic acid (NMDA) receptor, decrease of glutamate release, antagonism of alpha-amino-3-hydroxy-5-methyl-4-isoxasole propionate (AMPA) and adenosine receptors, increase in 5-HT release, increase in different modes of operation of gamma-amino butyric acid – one of the principal inhibitory neurotransmitters. Antiepileptic drugs have been shown to potentiate GABAergic synaptic transmission either by increasing gamma-amino butyric acid concentrations through the inhibition of gamma-amino butyric acid-transaminase (vigabatrin) or gamma-amino butyric acid up-take (tiagabine). Other drugs act directly at the synaptic gamma-amino butyric acid receptor complex (benzodiazepines, phenobarbital) or increase gamma-amino butyric acid concentration in several specific brain regions (valproate). The parallels of the neurochemistry and pathophysiology of epilepsy and chronic pain provide the basis for re-evaluating the use of antiepileptic drugs in various pain syndromes.

Recently, research into the mechanisms of migraine and the progressive recognition that cortical hyperexcitability and an imbalance between neuronal inhibition and excitation (mediated by gamma-amino butyric acid and amino acids, respectively) may play an important role in migraine pathophysiology, provided rationale for using antiepileptic drugs in prophylaxis and treatment of migraine and other headaches. Quality of evidence-based reviews and guidelines for the efficacy of various antiepileptic drugs in migraine prophylaxis and in migraine aura therapy have been provided by the American Academy of Neurology (2000) and D’Andrea et al. (2003).

There are also physiological reasons for using antiepileptic drugs in bipolar psychosis or other recurrent disorders. Ever since Kraepelin (1921), it has been postulated that bipolar affective disorders are of a progressive nature, just like seizures, and a kindling mechanism has been proposed.

Drug interactions in various health conditions

Unpredicted interactions when more than two drugs are used

Pindolol (beta-adrenergic blocking agent) does not increase serum combination of phenytoin concentrations when combination of phenytoin is administered in monotherapy. However, it does increase combination of phenytoin concentrations when combination of phenytoin is combined with other antiepileptic drugs. Gabapentin does not seem to have any pharmacokinetic interaction with combination of phenytoin; however, co-administration with vigabatrin and combination of phenytoin results in marked reduction in clearance of combination of phenytoin.

It has been found that antiepileptic drugs increase fenantyl requirement during anesthesia for craniostomy. However, this is a dose-effect relationship between the number of antiepileptic drugs received and the maintenance dose of fentanyl required for balanced anesthesia. Antiepileptic drugs have a similar dose-effect relationship with pipercuronium neuromuscular blockade (myorelaxants), also resulting in induction of a significant effect of antiepileptic drugs.

A case of retroperitoneal hematoma due to interaction between combination of phenytoin and aceno-coumarol, possibly potentiated by concomitant administration of paroxetine, has been reported.

The fact that over-the-counter drugs and nutritional supplements are increasingly being self-administered by patients creates the risk of drug interactions. Internet self-diagnosed and self-treated cases can also contribute to drug-drug interactions.

Interactions with folk medicine

In many countries folk medicine is frequently used for various reasons. Knowledge of active ingredients and possible interactions with antiepileptic drugs is usually poor. Widely used ginkgo preparations are a good example. Co-medication of ginkgo and antiepileptic drugs may result in decreased effectiveness of antiepileptic drugs due to the presence of seizure provoking contaminants in some ginkgo preparations. Ginkgo products may contain neurotoxin 4′-0-methylpyridoxine, which is a B₆ antivitamin. When seizures occur in patients for the first time, particularly in children, it is recommended that subjects be asked whether they have been taking ginkgo seeds or leaf extracts. In the probable mechanism of seizures, 4′-0-methylpyridoxine appears to inhibit pyridoxal kinase and when taken in a sufficient amount may result in convulsions. The amount of this neurotoxin in gingko leaves or seeds depends on the growing seasons during which the product was harvested.

Another example is primrose oil. Concurrent use of evening primrose oil and antiepileptic drugs may result in seizures. Combination of phenothiazines (for presumed schizophrenia) and evening primrose oil resulted in epileptic seizures. Withdrawal of primrose oil and carbamazepine administration resulted in seizure control. Evening primrose oil activated temporal lobe epilepsy in two patients with schizophrenia. For the same reason, evening primrose oil is contraindicated in patients with mania and epilepsy.

Antiepileptic drugs in non-epileptic disorders

Summary

Many drug interactions can be demonstrated but only a few of them are so clinically significant that they require adjustment of drug dosages. However, some drug combinations may produce unexpected changes of various extents and directions in different subjects and in different health conditions. The reasons for this variability include genetic control of the rate of drug metabolism as well as internal factors, such as serum changes, renal or hepatic disorders, gender and ageing. In this post, clinically and/or potentially significant drug interactions between antiepileptic drugs and non-antiepileptic drugs in health conditions other than epilepsy are discussed. Case reports of toxic effects due to drug interactions are presented as a warning signal calling for attention when polytherapy has to be used. In such cases, careful drug selection and dosage adjustment based on serum drug monitoring and clinical observation are the main rules for risk minimization. Awareness and knowledge of possible drug interactions is a good starting point before making treatment decisions.

It has also been reported that clearance of lamotrigine in pregnancy may be increased. Moreover, in one study lamotrigine plasma concentrations were slightly lower in females (13.7%) than in males. However, in another study this difference was not confirmed.

Combination of phenobarbital and theophylline results in increased theophylline clearance in children and adults but not in premature neonates.

Serum concentration of tirilazad mesylate (a membrane lipid peroxydation inhibitor), when given with phenobarbital in subarachnoid hemorrhage, may increase by 69%.

Drugs which compete for albumin bindings may increase the risk of kernicterus, e.g. combination of phenobarbital with aminophyline, cefotaxime and vancomycin shows that the bilirubin-displacing effect in the drug combinations cannot be predicted from each drug’s individual effect in premature infants.

Primidone withdrawal in a 14-year-old girl with congenital adrenal dys-plasia and epilepsy resulted in a three-fold increase of hypercorticolism and a reduction of plasma testosterone and 17 OH progesterone concentrations, thus showing the need to adjust the dose.

In patients with liver cirrhosis grade A, B and C (Child-Hugh classification), the median oral clearance was 0.31,0.24 and 0.10 ml/min/kg, respectively, in comparison with 0.34 ml/min/kg in normal healthy subjects. Correspondingly, median T_1/2 of LTGwas 36,60 and 110 h, whereas for patients with normal liver function it was 32 h. Lamotrigine dosage must be downwardly adjusted in patients with liver dysfunction.

Lamotrigine clearance in patients with hyperbilirubinemia (Gilbert’s syndrome) was 32% lower, and T_1/2 was 37% longer than in the healthy controls. Close clinical monitoring of such patients is needed when lamotrigine is administered; possible downward dosage adjustment should be considered.

Clearance of lamotrigine in Asians and non-Whites is lower than in Whites. This difference may have significant clinical relevance for non-Whites if lamotrigine is administered, particularly when it is combined with non-antiepileptic drugs, which are hepatic enzyme inhibitors.

Renal failure slows down the urinary excretion of prednisone and its metabolites, making dose reduction of corticosteroids necessary. However, combination of prednisone and phenobarbital increases excretion of prednisone without any clear change in 17 OH steroids and prednisolone urinary excretion . This drug interaction is associated with a decrease of graft tolerance in renal transplant patients.

It is interesting to note that serum valproate concentrations are higher in uremic serum than in normal serum, but there is no further displacement of valproate in the presence of mefenamic acid or fenoprofen. However, when uremic serum is treated with charcoal at pH 3.0, it removes the protecting effect of uremic serum, and valproate displacement from protein binding is higher.

Valproate displacement from albumin binding may depend on concentrations of non-antiepileptic drug, e.g. ketoconazole is an antifungal agent widely used in the management of patients with fungal infections, especially patients with acute acquired immunodeficiency syndrome. Ketoconazole is 99% bound to serum albumin and readily interacts with valproate. Statistically significant displacement of valproate has been observed at normal albumin level but only when ketoconazole concentrations were high (10-20 µg/ml). However, in patients with hypoalbuminemia, significant displacement of valproate was observed with lower ketoconazole concentrations. It is interesting that there is no displacement of valproate by ketoconazole in uremic serum. On the contrary, the free fraction of valproate decreases in the presence of ketoconazole in uremic serum.

Salicylate displaces carbamazepine from protein binding in normal sera but this effect is significantly reduced in uremic sera. On the other hand, significant displacement of carbamazepine from protein binding by tolmetin, ibuprofen and naproxen (non-steroidal inflammatory drugs) has been observed in uremic serum whereas in normal serum significant displacement has been found only with higher concentrations of naproxen.

Renal elimination plays only a minor role in overall elimination of lamotrigine. Thus, even in patients with moderate renal dysfunction and much lower clearance than in healthy people, the difference is not clinically relevant. However, in patients with more severe renal failure and particularly if hemodialy-sis is required, the daily dose of lamotrigine should be downwardly adjusted to the overall renal clearance.

Seizures are a relatively common occurrence in patients with human immuno virus infection. Seropositive patients are usually treated with antiepileptic drugs. In such patients antiepileptic drugs should be carefully chosen. The ideal antiepileptic drug:

1   should not stimulate viral replication,

2  has limited protein bindings,

3  has no effect on the cytochrome P450 system.

Gabapentin, topiramate, tiagabine and PRG meet these criteria. Valproate stimulates human immuno virus replication. Thus combination of antiepileptic drug with antiretrovirals should be carefully considered. Combination of carbamazepine and ritonavir in patients with human immuno virus infection may result in carbamazepine intoxication and antiretroviral therapy failure.

Carbamazepine is one of the most commonly used antiepileptic drugs in epilepsy and other neurological and psychiatric disorders. Carbamazepine mechanisms involve inhibitory action on sodium and on calcium (L- and N-type) channels, inhibitory effect on the release of somatostatin, increase of 5-HT release, effect on synaptic transmission and receptors, purine, monoamine, acetylcholine, adenosine and NMDA receptors. Its broad spectrum of pharmacological actions may explain the potent effect of carbamazepine in disorders other than epilepsy.

The analgesic effect is most comprehensively documented in neuralgias. However, it is also used in diabetic polyneuropathy, phantom limb pain syndrome, thalamic pain, cerebellar tremors and migraine.

There are reports that carbamazepine is also effective in hemifacial spasm, myotonia, restless legs syndrome and hyperactivity disorders in children. In patients with dementia, carbamazepine is given to alleviate agitation, aggressiveness or other behavioral abnormalities. There are many reports and a long history of carbamazepine use in alcoholism, psychiatric disorders such as acute mania, bipolar disorders and mood stabilizing in affective and aggressive disorders.

Gabapentin

Gabapentin is widely and more frequently used in fields other than epilepsy. This may be due to its multiple mechanisms of action and to the fact that it is not associated with any significant pharmacokinetic interaction with other drugs. Most frequently, Gabapentin is indicated for various pain syndromes and bipolar disorders. In these health conditions Gabapentin efficacy is well documented in clinical controlled trials.

In pain-resistant conditions, combination of Gabapentin with other analgesics is frequently used. In these cases Gabapentin is the optional drug in elderly patients because in this population polytherapy is frequently used and Gabapentin, usually, does not interact with other drugs. However, antacids (Maalox (R)) given concurrently with Gabapentin reduced Gabapentin bioavailability by 20% and when given 2h after Gabapentin, reduction was by 5% only.

Gabapentin in combination with antiretroviral medication showed some effects in neuropathic pain due to immunodeficiency syndrome. Amitriptyline or Gabapentin are alternatives for postherpetic neuralgia, and other antiepileptic drugs (Gabapentin, lamotrigine, tiagabine and topiramate) are alternatives for seizures since indinavir interaction with carbamazepine causes antiretroviral therapy failure. Adding Gabapentin to stable opioid medication in neuropathic cancer pain resulted in significant pain reduction without new adverse events. However, concurrent use of Gabapentin and morphine may result in an increase of Gabapentin plasma concentration, requiring Gabapentin dosage reduction in the elderly.

Gabapentin administration resulted in significant reduction of spontaneous or evoked pain (brush-induced allodynia, cold-induced allodynia and hyperalgesia). Good effects of Gabapentin were reported in newly diagnosed trigeminal neuralgia (Magnus, 1999), in diabetic neuropathy in randomized studies and in postherpetic neuropathy.

Gabapentin (1800-2400 mg/day) administered in patients with migraine resulted in significant prophylactic migraine attack reduction: in 36% of the patients, 50% reduction of migraine attacks was observed (in comparison with 14% in the placebo group).

In patients with bipolar or monopolar disorders and mild depression, moderate to marked response was reported using Gabapentin. Good effect was also observed in the treatment of acute mania using Gabapentin alone or in combination with other antimanic drugs. These positive effects were observed in open-label study, on rather small number of patients and short-treatment duration. However, placebo-controlled studies showed that Gabapentin does not have such beneficial effects in bipolar psychosis.

Other neurological disorders: Gabapentin administration resulted in significant beneficial effects in spasticity and paroxysmal symptoms associated with multiple sclerosis. In Parkinson’s disease Gabapentin in addition to dopaminergics showed a significant improvement in favour of Gabapentin over a short period of co-medication. In amyotrophic lateral sclerosis controversial results were reported during Gabapentin administration.

Gabapentin was administered in Huntington’s disease and other movement paroxysmal disorders. In open studies, long-term Gabapentin administration of 900 mg showed moderate to good beneficial effects, without adverse events in tardive diskinesia, facial diskinesia, blepharospasm, hemichorea or hemibalismus. On the other hand, various movement disorders appeared when Gabapentin was initiated; these disappeared when Gabapentin was withdrawn. In 48% of the patients with restless legs syndrome clinical improvement was observed during Gabapentin administration.

Paroxysmal dystonic movement, in both hands, occurred during combination of 900 mg Gabapentin with propranolol in the elderly. After reduction of propranolol to 40 mg/day the paroxysmal dystonia subsided immediately. A pharmacodynamic interaction effect was suggested.

Combination of Gabapentin with propranolol led to significant tremor improvement. Orthostatic tremor was reduced in the majority of patients with Gabapentin treatment. In another study, however, Gabapentin 1800 mg/day was added for 2 weeks to baseline anti-tremor treatment without any significant tremor reduction compared to the placebo. No drug interactions were reported.

Various rare neurological conditions: Gabapentin was administered in reflex sympathetic dystrophy, central pain, myokymia, cramp syndrome, idiopathic chronic hiccup and usually with clinical improvement.

Since Gabapentin is eliminated predominantly by renal excretion, it may be influenced or may affect pharmacokinetics of other drugs showing the same pattern of elimination at the renal site. In patients with renal dysfunction or in the elderly, the daily dose of Gabapentin should be downwardly adjusted according to creati-nine clearance decrease.

Lamotrigine

Lamotrigine has multiple mechanisms of action including decrease of glutamate release in addition to inhibition of sodium and calcium (L- and N-type) currents, and increase of gamma-amino butyric acid.

It has been suggested that lamotrigine possesses distinct psychotropic effects in addition to its antiepileptic action. Placebo-controlled trials in epilepsy treatment show some mood improvement (greater well-being) and there are theoretical reasons to suggest that lamotrigine, like other antiepileptic drugs, may possess mood-stabilizing properties. Polytherapy is usually used in bipolar psychoses, since there is no single mood stabilizer. Lamotrigine monotherapy administered in two groups of patients with bipolar-I depression showed that 250 mg of lamotrigine was significantly better than the placebo. Lamotrigine was effective in patients with rapid-cycling bipolar disorder and was useful in the treatment of bipolar-II disorder. Lamotrigine has not been shown to have clear efficacy in the treatment of mania or unipolar depression. Based on efficacy, adverse events and costs, it has been suggested that the use of lamotrigine in mood disorders should probably be on the basis of a second-line agent for bipolar depression.

Levetiracetam

Levetiracetam is an antiepileptic drug with unique profile of activity with potent broad-spectrum efficacy including effect on the high voltage N-type calcium channel, and at gamma-amino butyric acid and glycine-gated channels.

Information concerning levetiracetam usage in non-epileptic disorders is too limited to allow any firm conclusions to be made. However, a number of preliminary reports show that levetiracetam is well tolerated and effective in a wide variety of pain states (cervical and lumbar radiculopathy, traumatic peripheral nerve injury, neuropathic component in neoplastic pain, postherpetic neuralgia, allodynia, myelopathic pain and paresthesis in multiple sclerosis). levetiracetam is presently undergoing extensive evaluation for the treatment of various neuropathic pains and migraines; however, it is only registered for the treatment of epilepsy.

Migraine and various headaches are other disorders in which levetiracetam has been used with positive effects in reducing severity and frequency with modest side effects. In refractory migraines levetiracetam was given intravenously (i.v.) with good effect and was well tolerated.

There is also a suggestion that levetiracetam may by used as a mood stabilizer.

Levetiracetam is not associated with any pharmacokinetic interactions.

Oxcarbazepine

There are few publications concerning the use of O-carbamazepine in non-epilepsy conditions, although it has been used to treat acute mania. However, since O-carbamazepine is better-tolerated than carbamazepine (with the exception of more common hyponatremia), and has similar mechanisms of action, it may be used in indications similar to those for carbamazepine. O-carbamazepine is associated with far fewer pharmacokinetic interactions than carbamazepine.

Phenobarbital

Phenobarbital is the oldest antiepileptic drug in use and is still extensively used in developing countries. The mechanism of action of phenobarbital involves antagonism of AMPA receptor subtype and includes enhancement of GABAergic inhibition, enhancement of ionic currents by interactions with GABA_A receptor, decrease of excitatory amino acid release and post-synaptic response due to blocking of the excitatory glutamate response. A broad spectrum of pharmacological actions may contribute to potential therapeutic activity in neurological conditions other than epilepsy. However, cognitive impairment, morning sedation, potential for abuse, severe toxicity and withdrawal syndrome are contraindications for routine use of phenobarbital (strong inducer) in such disorders.

In the past, i.v. injections of phenobarbital were frequently used to prevent cerebral hemorrhage in preterm neonates. However, a critical review of the literature suggests that phenobarbital has no beneficial effect and in fact increased the incidence of intraventricular hemorrhage in infants with respiratory disease.

Phenobarbital has also been used in increased intracranial pressure to reduce the effect of cerebral blood flow and metabolism. However, it may impair cerebral perfusion pressure by inducing hypotension and therefore the benefit to risk ratio is low.

Neonatal hyperbilirubinemia can be controlled with a high single dose of phenobarbital (12 mg/kg) after birth. However, such a dose results in a prolonged sleep-state. In this condition, infants spend more time sleeping than they do with smaller doses.

Combination of chenodeoxycholic acid (750 mg/day) with phenobarbital (90-180 mg/day) was effective on the rate-limiting enzymes of liver cholesterol and bile acid synthesis. In patients with gallstones this effect was more pronounced than when each drug was used alone. Thus, an advantageous interaction was observed.

When asthmatic children were treated with phenobarbital, theophylline clearance increased by 42%, resulting in a 30% decrease in steady-state serum theophylline concentration. This drug combination requires theophylline dosage upward adjustment.

Reversible toxic encephalopathy was reported in a girl, possibly due to the toxic effect of ifosamide (cytostaticum) in combination with phenobarbital.

The rapidly fatal outcome of fulminant hepatitis caused by nilutamide, a non-steroidal antiandrogen derivative, was enhanced by co-administration with phenobarbital.

Phenytoin

Combination of phenytoin seems to be used much more frequently in the USA and the UK than in other European countries. In Poland, combination of phenytoin constitutes 6.7% of the pharmaceutical market. Combination of phenytoin is a strong inducer of hepatic enzymes and is involved in numerous drug interactions with antiepileptic drugs and non-antiepileptic drugs. Moreover, due to non-linear pharmacokinetics and side effects, combination of phenytoin is less frequently used today in non-epileptic disorders than it was before the introduction of the new generation of antiepileptic drugs.

Combination of phenytoin has a broad spectrum of pharmacological action on neurotransmitter receptors and ion channels and this may explain why combination of phenytoin is so effective in conditions other than epilepsy, such as: neuropathic pain, various pain syndromes, spasticity, myotonia and other disorders. However, evidence on the efficacy of combination of phenytoin from randomized clinical trials in these and other non-epileptic conditions is rather scant.

Neuropathic pain: It has been reported that combination of phenytoin may have a beneficial effect in trigeminal neuralgia, glossopharyngeal and superior laryngeal neuralgias, postherpetic and diabetic neuropathy, thalamic syndrome, phantom limb pain, diabetic pain and cancer pain. The efficacy of combination of phenytoin in other pain syndromes is at best mode. Unlike carbamazepine, evidence for efficacy of combination of phenytoin in trigeminal neuralgia and similar conditions is based on an uncontrolled study only. However, combination of phenytoin was more effective than aspirin in reducing pain in glycolipid lipidosis (Fabry disease).

In myotonic treatment, combination of phenytoin and carbamazepine were used interchangeably and their efficacy was comparable to the efficacy of procainamide. However, adverse events may be more pronounced. In a double-blind placebo-controlled study combination of phenytoin had a positive effect on motion sickness.

Pregabalin

Pregabalin ((S) —( + ) —3 isobutylgaba) is a gamma-amino butyric acid derivative, but does not interact with GABA_A or GABA_B receptors and does not influence gamma-amino butyric acid concentrations. Instead, pregabalin binds to sub-units (3₂, a₁ a₂-8 of the Ca²⁺ channel and this reduces the release of glutamate, noradrenaline and substance P. These mechanisms of action seem to be important in the treatment of epileptic seizures, pain and anxiety. Pregabalin is not associated with any pharmacokinetic interactions with carbamazepine, lamotrigine, phenobarbital, combination of phenytoin, tiagabine, topiramate or valproate.

Primidone

Primidone has been used in prospective, randomized clinical trials in essential tremor and is as effective as propranolol and more effective than phenobarbital.

Possible adverse events associated with primidone are similar to those with phenobarbital, which limits their use.

Tiagabine

Tiagabine is an inhibitor of gamma-amino butyric acid uptake. The drug was developed specifically for use as an antiepileptic drug based on the concept of the GABAergic mechanism of epileptic seizures. Since reduction in GABAergic neuronal activity has been proposed not only in epilepsy but also in various neuropsychological disorders, anxiety and pain, tiagabine may have a beneficial effect in these health conditions.

Tiagabine has been evaluated in various GABAergic mechanism-related disorders e.g. sleep disorders, pain (postherpetic and diabetic neuropathy), movement disorders (related to basal ganglia disorders, e.g. tardive diskinesia), spasticity, bipolar disorders, anxiety and neuroprotection against ischemia-induced cell loss. A moderate effect of tiagabine in migraine has been observed.

In casuistic observation tiagabine was administered in psychiatric patients (bipolar disorders) as add-on therapy to venlafaxine, lithium, flurazepam, bupropion, methylphenidate and paroxetine.

Dosages of tiagabine should be adjusted in patients with liver dysfunction.

In general, preliminary reports suggest that tiagabine use in non-epileptic conditions requires longer-term studies based on larger numbers of patients and on evidence-based medical principle.

Topiramate

The mechanisms of action of topiramate involve: sodium channel blockade, inhibition of AMPA glutamate receptors, potentiation of gamma-amino butyric acid-related neuroinhibition at GABA_A receptors; blocking of excitatory neurotransmission mediated by non-NMDA receptors. Topiramate is also a weak carbonic anhydrase inhibitor and may have an inhibitory effect on calcium channels.

Preliminary data suggest that topiramate, with multiple pharmacological properties, may have therapeutic effects in various chronic pain syndromes, migraine and cluster headache prophylaxis, tremor and certain psychiatric disorders.

Analgesic effect of topiramate combination with opioids was reported in neuropathic pain; and these effects were not the result of any drug interaction.

Topiramate was prophylactically effective in migraine and other headaches. topiramate and propranolol combination resulted in control of essential tremor, particularly in the hands compared to the head or voice. No drug interaction was reported.

In psychiatric disorders topiramate was combined with tricyclic antidepressants (Ortho McNeil Pharmacological) or with serotonin reuptake. Topiramate has been used as an alternative treatment for bipolar disorder, and was effective in 55% of initially manic patients after a mean of 312 days of treatment. There was no clinically significant effect of topiramate on haloperidol serum concentrations but a modest decrease in lithium serum concentration was observed though the interaction was without clinical relevance.

Nightmares and binge-eating responded well to topiramate in an open-label trial. Two patients with Tourette’s syndrome were successfully treated with topiramate while previous medications were discontinued. No interaction was reported.

In general, the reports on topiramate administration in non-epileptic disorders are based on short preliminary studies and/or small numbers of patients.

Valproic acid

Valproate is an antiepileptic drug with broad-spectrum efficacy against various forms of epileptic seizure. This is due to the combination of several neurochemical and neurophysi-ologic mechanisms, which may explain its effects in various neuronal dysfunctions. The mechanisms of valproate action include

1   increase of gamma-amino butyric acid turnover potentiating GABAergic functions in various specific brain regions,

2  inhibitory effect on voltage-sensitive sodium channels,

3  inhibitory effect on neuronal excitation mediated by the NMDA.

Several double-blind controlled trials have demonstrated the efficacy of valproate in migraine treatment and prophylaxis. Migraines with paroxysmal discharges in the electroencephalograph, mainly of the dysrhythmic type, were successfully treated with valproate. valproate is also effective in chronic headaches, and in cluster-form headaches. Valproate can occasionally be combined with other groups of medication for migraine treatment, including β-adrenergic channel blockers or anti-inflammatory drugs. In such cases potential drug interaction with valproate may occur.

In addition to its analgesic effect, valproate also shows efficacy in various psychiatric and neurotic disorders. It was reported that valproate is effective in patients with acute mania and its subtypes, depression and bipolar disorders. Moreover, valproate has been used in anxiety disorders, stress condition, aggressive behavior and tardive diskinesia.

Evidence-based medicine varies greatly but even so, valproate is widely used in fields other than epilepsy in the majority of countries.

Antidepressant drugs

Everybody is familiar with the tricyclic antidepressant drugs. However, in recent years a number of newer antidepressant drugs have been introduced into clinical practice (Table Classification of the psychotropic drugs currently in use). Essentially these are mainly non-tricyclic, earlier variants included mianserin, maprotiline and viloxazine.

The selective serotonin re-uptake inhibitors are represented by citalopram, fluoxetine, fluvoxamine, sertraline and paroxetine. Of these, citalopram is the most selective on serotonergic uptake, inhibiting serotonin uptake 3000 times more than noradrenaline uptake, and 22 000 times more than that of dopamine. In general, the selective serotonin re-uptake inhibitors are better tolerated and safer in overdose when compared with tricyclic drugs.

The latest generation of antidepressants has been developed to derive therapeutic benefits from tailor-made actions at specific monoamine receptor and re-uptake sites, in theory providing better efficacy and tolerability.

Reboxetine is a selective noradrenergic re-uptake inhibitor (NARI) with low affinity for histaminergic, cholinergic, dopaminergic and alpha-1 adrenergic receptors. It appears to be equally as effective as the tricyclics in treating depression, and there is a suggestion that it may be more effective than fluoxetine. Venlafaxine is a serotonin-noradrenergic re-uptake inhibitor, which is similar to the earlier generation of antidepressants, but it does not interact with histaminergic or cholinergic receptors, thus diminishing side effects due to those receptor systems. Several studies have indicated equipotentiality or superior effectiveness with this compound compared with tricyclics.

Nefadazone is a noradrenaline serotonin re-uptake inhibitor whose most potent action is blockade of 5HT2 post-synaptic receptors, leading to a dual mechanism of action on the serotonin system at 5HT1 and 5HT3 subsites. Noradrenaline re-uptake inhibition is only minimal, and there is no interaction with histamine or cholinergic receptors.

Table Classification of the psychotropic drugs currently in use

Mirtazapine, or noradrenaline-specific serotonergic antidepressant has a selective action at alpha-2 adrenoreceptors, and only at some serotonin receptor subtypes. Its actions increase noradrenergic and serotoninergic transmission by blocking the alpha-2 autoreceptors. However, because it also blocks 5HT2 and 5HT3 receptors, the increased serotonin turnover only stimulates 5HT1 receptors. Thus it enhances noradrenergic and 5HT1A mediated serotoninergic neurotrans-mission. It is free of muscarinic, alpha-1 adrenergic and 5HT2- and 5HT3-related side effects, but its effects on histamine receptors can cause sedation and increased appetite. Several studies have shown equal or superior efficacy of this compound compared with other antidepressants.

Antipsychotic drugs

As with the antidepressant drugs, in recent years there have been several newer agents introduced into clinical practice. These essentially, with some exceptions, fall into the class of atypical antipsychotics.

The classical neuroleptic drugs, such as chlorpromazine and haloperidol, antagonize dopamine D2 receptors. Their clinical efficacy has been shown to correlate with inhibitory activity at these receptor subtypes. However, these drugs block dopamine receptors in the striatum leading to catalepsy in animal models, and unwanted extrapyramidal side effects in patients.

The new generation of antipsychotic drugs essentially fall into two categories; those that are clozapine related, which included olanzapine and quetiapine, and others such as risperidone.

Although clozapine has been available for many years, it was initially not available for clinical use on account of its potential to produce agranulocytosis. However it has been reintroduced into clinical practice as a model of an atypical antipsychotic. The term essentially relates to the low potential of these compounds to cause extrapyramidal problems, and to have minimum effects on serum prolactin levels. The mechanism of atypicality seems to relate to activity at different receptor subtypes.

In general, the atypical antipsychotics occupy lower levels of D2 receptors than the classical antipsychotics, but one reason for their differing profile may be due to the rapid displacement of these agents from receptors by endogenous dopamine, then thus being more loosely bound to the receptor. The newer antipsychotic agents also have lower relative affinity for striatal D2 receptors as opposed to limbic D2 receptors (dorsal vs. ventral striatum).

Others

The minor tranquilizers mainly in use are the benzodiazepines, but their use in epilepsy is limited. Problems with dependency have led to caution with the use of these drugs, and in epilepsy withdrawal seizures can be a problem. Clobazam, a 1-5-benzodiazepine is used in the management of intractable seizures, and has effective anxiolytic properties.

The mood stabilizers include lithium, which is proconvulsant, and several antiepileptic drugs. Of the older generation, carbamazepine and valproic acid have been shown to have antimanic properties and they help in the prophylaxis of mood disorders. Topiramate and lamotrigine are under investigation at the present time for their mood regulating properties. The mode of action on mood is unclear, and it may not be directly related to their antiepileptic properties.

Stimulants include amphetamine and methylphenidate. These are used mainly to control attention-deficit-hyperactivity disorders, which are not uncommon in the learning disabled, many of which patients also have epilepsy.

On the use of psychotropic drugs in epilepsy

It seems accepted that many patients with epilepsy have psychiatric syndromes, and recent epidemiologic evidence from selective clinics suggests that over 50% of patients may have a recognizable psychiatric disorder. It is also known that many patients with epilepsy receive psychotropic drugs, sometimes but not always on account of psychiatric symptoms. However, the effect of these drugs on the seizure threshold is variable, some, such as the benzodiazepines, being anticonvulsant, others, including many antidepressant and antipsychotic drugs, in contrast, are pro-convulsant.

It has been known ever since their introduction that tricyclic drugs are proconvulsant, and lead to seizures, which, for example in overdose, is one method of fatality.

Of the non- tricyclic drugs, both maprotiline and mianserin seem to be at the more proconvulsant end of the spectrum. Of the newer generation of drugs, the selective serotonin re-uptake inhibitors are considered to provoke less in the way of seizures than tricyclics, and there is a possibility that the even newer, more selective drugs provoke even less in the way of seizures than the selective serotonin re-uptake inhibitors, but more data on these compounds are needed. The reporting of seizures with all of the new drugs in clinical trials is at very low levels, either similar to, or lower than the less convulsant tricyclics.

As with the newer antidepressants, there is little information about the effect of atypical neuroleptic drugs on the seizure threshold with the singular exception of clozapine. The latter was known to be proconvulsant from early studies, the seizures seemed to be a dose-related effect. The incidence of seizures rises to about 5% at doses of 600 mg, although electroencephalograph changes maybe recorded at lower doses, these results emerging from patients with schizophrenia, and not epilepsy. The seizures are often myoclonic, but can be generalized tonic/clonic, or partial depending on the individual patient.

It is perhaps of no coincidence, and of considerable interest that clozapine is probably the most effective antipsychotic drug available, reinforcing again a link between seizures and psychosis, and an integral part of neuropsychiatric practice.

Pharmacokinetic interactions between psychotropic drugs and anticonvulsants

Pharmacodynamic interactions between antiepileptic and psychotropic agents

Conclusions

Several factors must be considered when predicting the outcome of a potential interaction: patient-related (sex, age, ethnicity) and drug-related (the presence of active metabolites, the activity and potency at the enzyme site, the therapeutic window). Clinicians should be aware of these potential interactions especially if the patient shows no response to a psychotropic drug therapy or signs and symptoms of intoxication.

As far as antidepressant drugs are concerned, fluoxetine and nefazodone interactions are probably the most relevant in epilepsy from a clinical point of view. The former for its long half-life and the presence of different enantiomers with different kinetic properties, and the latter for its inhibitory properties on CYP3A4. antiepileptic drug inducers increase the clearance of all the antipsychotic drugs; therefore dosage adjustments are quite often required for antipsychotics.

Careful clinical monitoring and personalized dosages and titration time can usually lower the risk threshold of side effects due to pharmacologic interactions, which is one of the major factors for good clinical practice and patient compliance.