Within the context of the pharmacologic treatment of epilepsy, the topic of drug interactions has received much attention. Because antiepileptic drugs (AEDs) have narrow therapeutic ranges, treatment is generally individualized, and unpredicted alterations in drug levels might require dose adjustments. Most of the older AEDs are prone to interactions with other AEDs, as well as with other concurrent medications such as anticoagulants, antidiabetic agents, and antidepressants. As a result, the literature associated with AED interactions is extensive. This chapter focuses on recent advances and the integration of pharmacokinetic and pharmacodynamic interactions.

The concept of pharmacodynamic interactions between AEDs is quite different from the concept of pharmacokinetic interactions. Whereas pharmacokinetic interactions determine changes in levels of a drug, pharmacodynamic interactions determine changes in pharmacologic effects when another drug is added or discontinued. These pharmacodynamic interactions, therefore, can determine whether it is meaningful for two particular drugs to be prescribed together. In contrast, pharmacokinetic interactions require only dose adjustments; they are unrelated to the qualitative aspects of drug combinations. The amount of information available about pharmacokinetic interactions between AEDs is considerably larger than that available about pharmacodynamic interactions, in part because it is easier to measure drug levels than to quantify various drug actions. In addition, for pharmacodynamic interactions, several pharmacologic actions of drugs can be involved, whereas with pharmacokinetic interactions there is always only one end point, namely, the drug concentration. To analyze pharmacodynamic interactions, it is necessary to understand what the possible interactions can be. Pharmacokinetic interactions can cause a change in drug absorption, metabolism, or elimination or a displacement from serum proteins. Pharmacodynamic interactions are distinguished by whether they are purely additive, supraadditive, or infraadditive. When the interaction is additive, the combined effect of the two drugs administered together for a given pharmacologic effect is equal to the expected sum of the corresponding activity of each drug used alone. For instance, if concentration x of drug A produces a certain effect, and concentration y of drug B produces the same effect, one-half of concentration x of A in the presence of one-half of concentration y of B will produce the same effect. One-half of x and one-half of y represent an equivalent “bolus,” and this concept forms the basis of the so-called “isobolographic” analysis, an established method for the quantitative assessment of pharmacodynamic interactions.²¹ When the interaction is supraadditive, or potentiated, the combined effect of the two drugs is greater than the expected sum of the individual effects of the two drugs. In the foregoing example, less than one-half the concentration of each drug would be required to achieve the same effect. Finally, if the interaction is infraadditive, or antagonistic, the combined effect is less than the expected sum of the individual effects, and, for instance, more than one-half the concentration of each drug would be required to achieve the same effect.

How does this apply to pharmacodynamic interactions between AEDs? AEDs form a heterogeneous group; their common denominator is the ability to prevent the occurrence of seizures. They also all tend to produce toxicity, in particular neurologic toxicity, at certain concentrations. The antiepileptic pharmacodynamic interaction between two drugs is irrelevant in itself. A supraadditive interaction is not necessarily beneficial. The concentration of a single drug could be increased indefinitely if it were not for the occurrence of toxicity. The same upper limit, however, will also apply to two drugs administered simultaneously. Therefore, for a combination of two AEDs to be advantageous, the seizure protection provided by the combination at a certain degree of toxicity (e.g., the threshold for overt toxicity) must be stronger than that with either drug alone at the same level of toxicity. Thus, the antiepileptic and the neurotoxic interactions must differ in favor of the antiepileptic interaction. In clinical reality, this issue is complicated by the fact that side effects of AEDs are not limited to neurotoxicity. In addition, a totally different situation in which a combination of two AEDs can be advantageous can arise when a patient has two different seizure types, each of which responds to a different drug.

Based on these considerations, it is understandable that the information on pharmacodynamic interactions between AEDs is limited. These interactions are difficult to quantify in patients. Even the available data from animal experiments are limited because certain studies considered only the antiepileptic interaction. In many other studies, the analysis was based on doses only. When drug effects are quantified on the basis of drug dose alone, pharmacokinetic interactions can falsify the analysis of pharmacodynamic interactions. For instance, earlier studies based on the analysis of doses in animals suggested an antiepileptic potentiation between phenytoin and phenobarbital.^28,172 This turned out to be a pharmacokinetic artifact due to an acute inhibition of phenytoin elimination in the presence of phenobarbital. Later studies revealed that, in the presence of phenobarbital, single doses of phenytoin produce higher phenytoin brain levels than when phenytoin is administered alone and that the antiepileptic interaction between phenobarbital and phenytoin is purely additive when brain levels are used for the analysis.^20,91

The results of the first series of experimental studies in which antiepileptic as well as neurotoxic interactions between AEDs were quantified are summarized in Table 1. These studies reveal that the majority of antiepileptic interactions are strictly additive and that antiepileptic potentiation is the exception
rather than the rule. In contrast, neurotoxic interactions are divided about equally between those that are infraadditive and those that are additive. These experimental data suggest that only a few combinations among the older AEDs are possibly superior to the corresponding monotherapies. These combinations include valproate with carbamazepine, valproate with ethosuximide, and valproate with phenytoin. Although an infraadditive neurotoxic interaction between phenytoin and phenobarbital was found, the therapeutic index of phenobarbital alone was very low, and the combination still had a lower therapeutic index than did phenytoin alone.²⁰

Interactions between low doses of clonazepam and valproate or ethosuximide were all found to be supraadditive for antiepileptic and for neurotoxic effects, but they result in a superior therapeutic index for both valproate and ethosuximide.²² Using a similar model, Gordon et al.⁶⁴ studied the pharmacodynamic interactions between felbamate and older AEDs. They found a potentiation of the antiepileptic activity of felbamate by phenytoin, carbamazepine, valproate, and phenobarbital. In contrast, the neurotoxicity was not potentiated, and the protective index of felbamate was raised by the addition of any one of these four drugs.

Many additional experimental studies of pharmacodynamic interactions have been carried out in recent years, involving mostly the newer AEDs (Table 1).³⁸ Overall, these studies again reveal that various combinations can have any possible type of association of antiepileptic and neurotoxic interactions. Accordingly, some drug combinations are more promising than others, at least on the basis of this experimental model.

In the end, whether a combination of two AEDs is beneficial for patients needs to be determined by careful clinical assessments. Although pharmacodynamic interactions are more difficult to study in patients than in experimental animals, valuable clinical data have accumulated. One of the first clinical studies addressing in a systematic manner the issue of the beneficial value of an AED combination was reported by Hakkarainen.⁷⁰ Among 100 newly diagnosed patients, 33 were refractory to carbamazepine alone and to phenytoin alone. Of those, five (15%) became seizure free on the combination. Rowan et al.¹⁴⁰ demonstrated that absence seizures could be fully controlled by valproate–ethosuximide combination therapy in a few
patients who had been refractory to either drug alone. Walker and Koon¹⁶⁹ found that some patients who had not responded to valproate alone and to carbamazepine alone became seizure free on the combination.

A positive synergism between valproate and lamotrigine has also been suggested. Among 347 patients refractory to monotherapy with valproate, carbamazepine, phenytoin, or phenobarbital, the seizure reduction was significantly greater when lamotrigine was added to valproate than when it was added to the other drugs.²⁴ In a rigorous systematic study of add-on valproate versus add-on lamotrigine, among 13 patients who had not responded to the addition of either one of the two drugs, 4 became seizure free when both drugs were added.¹²³ When 14 patients whose seizures had not been controlled by monotherapy with carbamazepine and with vigabatrin were given both medications, 5 (36%) became seizure free.¹⁵⁹

In addition to potentially favorable pharmacodynamic interactions, clinical observations have also revealed that adverse effects of one AED can be potentiated by those of another. An increase in tremor when lamotrigine was prescribed in combination with valproate was observed in two studies.^80,123 An increase in side effects characteristic of carbamazepine was noted in four patients in whom levetiracetam was added to carbamazepine in polytherapy.¹⁵⁶ An exacerbation of carbamazepine toxicity also was noted after the addition of lamotrigine, and this could not be attributed to a pharmacokinetic interaction.¹⁶ Chorea occurred in three patients on phenytoin and lamotri-gine in combination only, and it resolved when one medication was tapered.¹⁷⁵ Finally, it appears that valproate encephalopathy is more likely to occur in the presence of another AED, and it can resolve when either valproate or the other drug is discontinued.¹⁰⁴ More recently, two reports suggested that the addition of topiramate may enhance some side effects of valproate, in particular hyperammonemic encephalopathy.^63,101