Cognitive Effects of Antiepileptic Drugs

Kimford J. Meador

Epilepsy is not a single disease, so it is not surprising that there is considerable variability in cognitive abilities across patients with epilepsy. Nevertheless, cognitive problems are common in patients with epilepsy. Multiple factors can contribute to impairment of cognition in patients with epilepsy (1) (Table 6.1). Among these factors, antiepileptic drugs (AEDs) are of particular concern because they are the main therapeutic option for epilepsy. There is little evidence of differential efficacy across AEDs except for the lack of efficacy of some AEDs against primary generalized epilepsy. However, there is evidence of differential cognitive effects across AEDs (1). Given that most patients with epilepsy feel that the adverse cognitive effects of AEDs prevent them from achieving activities or goals (2), consideration of adverse AED cognitive effects should be important in the choice of AED.

The risk of cognitive side effects from AEDs rises with polypharmacy, increasing AED dosage, and increasing anticonvulsant blood levels (1). Reducing the number of AEDs can improve cognition and may even decrease seizure frequency (4). Nevertheless, an individual patient may obtain the best seizure control with tolerable cognitive side effects when treated with polytherapy. For each patient, the clinician must consider the overall risk-to-benefit ratio of the AED and the severity of the seizure disorder in question. When used in monotherapy with anticonvulsant blood levels within standard therapeutic ranges, the cognitive effects of AEDs are generally modest on formal neuropsychological testing (5) and may not be seen if appropriate study designs are not employed (6). However, these “modest” effects can be clinically significant. For example, several studies have consistently demonstrated that there is a highly significant inverse correlation between neurotoxicity symptoms and perception of quality of life (7). This effect is present despite the absence of overt toxicity on neurological examination. Another example of how the cognitive effects of AEDs can be clinically significant is the effect of AEDs on recent verbal memory. In a series of studies with healthy volunteers, delayed recall of memory for verbal paragraphs was reduced by approximately 15% by carbamazepine (CBZ), phenytoin, topiramate, and valproate compared with nondrug conditions (8,9). Differential effects
of AEDs can be seen for verbal memory because this effect is not present on the same task when gabapentin and lamotrigine are used (8).

TABLE 6.1 Factors that may Contribute to Cognitive Impairment in Patients with Epilepsy (3)

Etiology
Age of onset
Type of seizure
Seizure characteristics (i.e., frequency, duration, and severity)
Intraictal, interictal, or postictal dysfunction from seizures
Brain damage acquired before onset of seizures
Brain damage due to repetitive or prolonged seizures
Hereditary factors
Psychosocial factors (e.g., depression)
Adverse effects of treatments (e.g., surgery or antiepileptic drugs)

The literature examining the cognitive effects of AEDs must be viewed critically, because flaws in experimental design, analysis, and interpretation occur frequently (1,6,10). Several of these methodological issues are listed in Table 6.2 and discussed in detail elsewhere (1,6,10).

Older Antiepileptic Drugs

Many of the investigations of the older AEDs had limitations in design and analysis as listed in Table 6.2. The first large Veterans Administration (VA) Cooperative Study compared the efficacy and side effects of CBZ, phenobarbital, phenytoin, and primidone in patients with new-onset epilepsy; the study also examined the cognitive effects of these AEDs as a secondary goal (11). It reported “no consistent pattern” of cognitive effects across AEDs and few cognitive changes from pre- to post-AED treatment conditions. However, this study did not control for several factors such as test–retest effects, anticonvulsant blood levels, and seizure frequency. The cognitive effects of the older AEDs were readily apparent when randomized, double-blind, crossover designs in healthy volunteers were conducted, which controlled for confounding effects such as seizures, preexisting brain abnormalities, and anticonvulsant blood levels (12,13,14). These studies found no overall difference between CBZ, phenytoin, and valproate, but approximately 50% of the neuropsychological variables were significantly worse with AEDs compared with nondrug conditions (12,13,14). In contrast, 32% of the variables were significantly worse with phenobarbital than with phenytoin or valproate (14). Therefore, phenobarbital exhibits greater untoward cognitive effects than other older AEDs, but no clinically significant differences in cognitive effects were seen across CBZ, phenytoin, and valproate. The comparable cognitive effects of CBZ, phenytoin, and valproate are also supported by several studies in patients with epilepsy (15,16,17).

TABLE 6.2 Methodological Concerns in Antiepileptic Drug Cognition Studies

Experimental design
- Subject selection bias
- Nonequivalence of clinical variables
- Failure to control for anticonvulsant blood levels
- Failure to control for seizure frequency
- Nonequivalence of dependent variables
- Sample size
- Test–retest effects
- Characteristics of behavioral tests
Statistical analysis and interpretation
- Type I error
- Nonorthogonal contrasts
- Comparison of studies with nonequivalent designs/statistics
- Statistical vs. clinical significance

Newer Antiepileptic Drugs

Gabapentin

Several studies of gabapentin as add-on therapy in patients with epilepsy have reported subjective improvements in well-being (3). A double-blind, dose-ranging (1,200 to 2,400 mg per day), add-on, placebo-controlled, crossover study of patients with partial epilepsy found one positive effect and one negative effect (more subjective drowsiness) (18). A double-blind, randomized, crossover study of healthy volunteers found significantly better performance on 26% of the variables for gabapentin 2,400 mg per day versus CBZ (mean dose = 731 mg per day, mean blood level = 8.3 μg per mL) (19). In this same study, CBZ was worse on 48% of the variables and gabapentin worse on 19% of the variables compared with the nondrug state. Consistent with these results, gabapentin was tolerated better than CBZ in healthy elderly adults (20) and in elderly patients with new-onset epilepsy (21). In contrast, a randomized, double-blind, parallel-group study of healthy volunteers using quantitated electroencephalogram (EEG) and a cognitive battery reported no significant differences between CBZ (1,200 mg per day) and gabapentin (3,600 mg per day) (22).

Lamotrigine

Lamotrigine has demonstrated positive effects on perceived quality of life in patients with epilepsy compared with placebo or CBZ (23,24). One of these studies also employed a limited neuropsychological battery and found no adverse cognitive effects for lamotrigine (23). A double-blind, randomized, crossover study of healthy adults reported significantly better performance on over half of the neuropsychological variables for lamotrigine (150 mg per day) versus CBZ (mean dose = 696 mg per day; mean blood level = 7.6 μg per mL) (25). In the same study, 62% of the variables were worse with CBZ and 0% worse than lamotrigine compared with nondrug state. Consistent with the findings in the above study, several other studies of healthy volunteers have shown fewer cognitive side effects with lamotrigine compared with CBZ, diazepam, phenytoin, placebo, or valproate (26,27,28).

Levetiracetam

Although well tolerated in clinical trials, there are few published studies with formal neuropsychological data for levetiracetam. A double-blind, randomized, placebo-controlled crossover design study of healthy adults with 8-day treatment arms compared CBZ (800 mg per day), oxcarbazepine (1,200 mg per day), and levetiracetam (1,500 mg per day) (29). Although statistical power was limited, CBZ had the most adverse effects and levetiracetam had the least overall. In addition, preliminary results from a double-blind, randomized, placebo-controlled crossover study of healthy adults has reported significantly better profile on a neuropsychological battery for levetiracetam compared with CBZ (30).

Oxcarbazepine

There are also few investigations of oxcarbazepine using formal neuropsychological measures although it has been tolerated well in clinical trials. A small randomized, monotherapy, double-blind, parallel-group study of patients with new-onset epilepsy found no differences in cognitive effects for oxcarbazepine compared with phenytoin (31). A randomized, double-blind, placebo-controlled, crossover study of healthy volunteers treated for 2 weeks with low-dose (150 or 300 mg b.i.d.) oxcarbazepine revealed slowed reaction time, but slightly improved cancellation task performance and better subjective alertness (32). A double-blind, randomized, parallel design study of healthy volunteers comparing oxcarbazepine (19.9 μg per mL) with phenytoin (11.6 μg per mL) found no significant differences on neuropsychological and quantitative EEG measures (33).
Approximately half of the variables were worse for each AED compared with nondrug in this study.

Pregabalin

Recently approved by the U.S. Food and Drug Administration (FDA) and released in the US, pregabalin is an analog of gabapentin. There are no formal cognitive studies of pregabalin using formal neuropsychological measures.

Tiagabine

Despite the fact that tiagabine enhances the effect of γ-aminobutyric acid (GABA), the main inhibitory neurotransmitter in the brain, it has demonstrated no significant cognitive effects in a small, low-dose, add-on study (34), a small, double-blind, add-on, placebo-controlled, parallel-group, 3-month study (35), and a large, randomized, double-blind, add-on, placebo-controlled, parallel-group, dose–response study of patients with epilepsy (36).

Topiramate

Of all the newer AEDs, topiramate has raised the most concerns about cognitive effects. Somnolence, psychomotor slowing, difficulty with memory, and language problems (especially word-finding difficulty) were noted in the early clinical trials. A small, single-blind, randomized, parallel-group study of healthy volunteers reported greater adverse cognitive effects for topiramate compared with gabapentin and lamotrigine, although the titration rate for topiramate was faster than that recommended (37). Decline in verbal fluency, attention, processing speed, and working memory was found in a small study of patients with epilepsy tested on and off topiramate (38). A small, open, randomized trial comparing tiagabine and topiramate as add-on therapy in patients with epilepsy found significant deterioration in verbal fluency, language comprehension, working memory, and visual block tapping for topiramate but only a deterioration in verbal memory for tiagabine (39). Two multicenter, randomized, parallel-group, double-blind studies of topiramate versus valproate as adjunctive therapy to CBZ in patients with epilepsy showed less effects (40,41). Several significant neuropsychological differences were present at the end of titration, but only a few differences were present at the end of the 8-week maintenance phase. One study had only 1 of 17 variables (i.e., verbal memory) worse for topiramate (40), and the other study found 2 of 30 variables (i.e., verbal fluency and a graphomotor task) were worse for topiramate (41). A double-blind, randomized, placebo-controlled parallel study of healthy adults revealed significantly worse neuropsychological performance for topiramate (330 mg per day) compared with gabapentin (3,600 mg per day) on half of the variables (42). Topiramate was worse than nondrug state on 29% of the variables, but gabapentin did not differ from the nondrug state. A double-blind, randomized, crossover study of healthy adults comparing topiramate and lamotrigine at target doses of 300 mg per day found that 80% of the variables were significantly worse for topiramate (9). Topiramate was worse than nondrug conditions on 88% of the variables (no variables favored topiramate), and lamotrigine worse on 17% and better on 7% than nondrug state.

Only gold members can continue reading. Log In or Register to continue