Patients with epilepsy often experience cognitive dysfunction. Multiple factors can adversely affect cognition in epilepsy, including the etiology of the seizures, cerebral lesions acquired before the onset of seizures, seizure type, age at onset of epilepsy, seizure frequency, duration, and severity, intraictal and interictal physiologic dysfunction, structural cerebral damage caused by repetitive or prolonged seizures, hereditary factors, psychosocial factors, and sequelae of treatment for epilepsy, including antiepileptic drugs (AEDs) and epilepsy surgery.1-6 All these interrelated factors make complex contributions to cognitive deficits.5 AEDs affect cognition by suppressing neuronal excitability or enhancing inhibitory neurotransmission. Patients, and even some clinicians, tend to blame cognitive problems on AEDs because they are more identifiable than other factors. However, AED effects should not be overrated. Psychosocial problems, which are common, can be overlooked as a source of cognitive impairment. The stigma of epilepsy and the fear of having seizures in public can lead to low self-esteem, social isolation, and depression, all of which can negative affect cognitive function. Similarly, subclinical epileptiform activity is another important contributor to cognitive dysfunction that can go unrecognized, especially in patients with infrequent seizures.

Whilst AEDs are rarely the sole source of cognitive deficits, they do have the potential to affect cognition significantly in certain patients. AEDs exert dosage-dependent effects on cognitive functioning, which can be exacerbated by AED polytherapy. The main cognitive effects of AEDs involve attention/vigilance, psychomotor speed, and secondary involvement of other cognitive functions (e.g., memory).4 The effects of AEDs on cognition are especially significant since AEDs are often be selected based on both traditional measures of treatment effectiveness such as efficacy and tolerability, and their negative neuropsychological side effects. The presence of AED-induced cognitive side effects is an important concern of epilepsy patients taking medications.7 Consequently, the neuropsychologist should attempt to determine the potential effects of AEDs on cognitive performance, as well as the patient's subjective perception of performance, which can also be mediated by the mood and affective state.6

The magnitude of AED-related cognitive dysfunction is generally modest in monotherapy and when the AED is present at therapeutic serum concentrations. However, there are circumstances in which decreased cognitive function assumes greater importance, such as learning in school children,5,8 when driving or operating machinery, and when cognitive skills might be especially vulnerable, such as in the elderly.4 It is especially important to identify and minimize the cognitive effects of AEDs in children, whose developing nervous systems can be more vulnerable to the long-term consequences of AED-induced cognitive impairment.9 Individuals older than 65 years are more susceptible to the cognitive effects of AEDs due to both pharmacodynamic and pharmacokinetic factors.10 Nevertheless, very few studies have specifically examined the cognitive effects of AEDs in this population.

Attempts to identify and quantify the cognitive effects of AEDs have had limited success. Although hundreds of studies have been performed, their methodological flaws, differences in study designs, and contradictory results have resulted in no clear picture emerging. There are three primary factors to consider when assessing the cognitive risk of AEDs, which is a difficult area in which to perform studies in a largely clinical setting.6 The first results from failure to randomize the treatment regimen, since patients with epilepsy that is more difficult to control are more likely to try newer drugs, and hence have cognitive skills prior to treatment that are not representative of the population. The second factor is the difficulty of establishing comparable drug levels, since patients are treated clinically across a wide range of effective dosages, which might have varying effects on cognitive function. Third, the absence of blinding is a concern in many studies. Other methodological issues might also be present. Some studies have employed healthy volunteers rather than patients in order to control for the confounding effects of seizures and pre-existing brain abnormalities. The duration of treatment used in some crossover studies might be inadequate to determine the effects of long-term therapy on neuropsychological status, a concern that is especially relevant to pediatric AED studies since the cognitive effects occur against the backdrop of normal cognitive development. Here we review the current literature on the cognitive effects of AEDs, drawing general conclusions whenever possible whilst taking into account the many underlying methodological problems.

AEDs are often categorized into "older"(developed before the 1990s) and "newer" (introduced during in the 1990s or later) agents. Since numerous reviews and clinical data are available for older AEDs, here only a brief overview of their cognitive effects is provided. Phenytoin (PHT), carbamazepine (CBZ), valproate (VPA), phenobarbital (PB), and benzodiazepines have been used in epilepsy treatment for many years, and are categorized as older AEDs. The large Veterans Administration (VA) Cooperative Study comparing the cognitive effects of CBZ, PB, PHT, and primidone in a parallel study of patients with new-onset epilepsy found no consistent pattern across AEDs and little change in cognition after AED treatment.11 However, the study design did not control adequately for test-retest effects. The second VA Cooperative Study found no cognitive differences between CBZ and VPA.12 Another study comparing CBZ and PHT in patients with epilepsy also found modest negative effects on cognition for both drugs, and few differential effects.13 Randomized, double-blind crossover studies in healthy volunteers directly comparing AEDs have demonstrated that CBZ, PHT, and VPA have similar cognitive side effects.14,15 In contrast, PB scored significantly worse than PHT or VPA on 32% of the variables. All of these older AEDs affected cognition relative to the nondrug conditions, with approximately half of the variables being significantly worse for CBZ, PHT, and VPA. It was recently reported that total Intelligence Quotient (IQ) (especially performance items) improved in children who discontinued PB compared to those who continued to take PB.16 Thus, among the older AEDs, CBZ, PHT, and VPA exert similar effects on cognitive function, with PB producing greater cognitive impairment.

Many newer AEDs have been introduced since the early 1990s, including felbamate (FBM), gabapentin (GBP), lamotrigine (LTG), oxcarbazepine (OXC), topiramate (TPM), tiagabine (TGB), vigabatrin (VGB), zonisamide (ZNS), pregabalin (PGB), and levetiracetam (LEV). Because of serious safety issues, including aplastic anemia and hepatotoxicity, FBM therapy is recommended only for epilepsy patients who are refractory to other AEDs, and its cognitive effects have not been systematically studied.

Despite the large number of newer agents, most neuropsychological studies have compared newer drugs to older AEDs with a higher risk of cognitive impairment, or against newer drugs at dosages that do not reflect current prescribing patterns, making the comparative effects of newer agents unclear.6 Since these methodologies might be unduly influenced by the pharmaceutical companies that financially support cognitive studies, direct head-to-head comparative studies of newer drugs will probably require government support.

The ideal AED would reduce neuronal irritability without affecting neuronal excitability and cognitive function. The following techniques are used to reduce the cognitive side effects of AEDs: treat underlying disease processes, titrate slowly when initiating AEDs, use the lowest AED dosage possible, use AED monotherapy if possible, avoid AEDs with greater adverse events (e.g., PB), avoid adverse pharmacokinetic interactions, balance all factors with best seizure control, and confirm the seizure diagnosis if the patient is refractory to AEDs. Some patients are best controlled (with the least side effects) on low-dosage, addictive polytherapy.

Patients with epilepsy often experience cognitive dysfunction. An AED treatment can affect cognition either positively or negatively. Since greater seizure frequency, duration, and severity increase the likelihood of impaired cognition, AED intake can reduce cognitive dysfunction by controlling seizures. However, even though the cognitive effects of AEDs are usually modest when these agents are used as a monotherapy at blood concentrations within the standard therapeutic ranges, significant effects such as decreased quality of life or neuropsychological impairment can occur. Furthermore, intolerance of long-term AED-induced cognitive dysfunction will have longer lasting harmful effects on the quality of life.

Our epilepsy clinic investigated cognition using neuropsychological tests in patients with well-controlled epilepsy treated with AEDs for at least 1 year to determine whether cognitive deficits were induced by the long-term use of AEDs. We selected 48 patients who had been seizure-free for more than 1 year, normal EEG in neuropsychological tests, and normal MRI in order to exclude cognitive effects of interictal epileptiform discharges or structural brain lesions. Our patients were exposed to AEDs for a mean of 5.8 years (range: 1-30 years) and had a mean seizure-free period of 3 years (range: 1-9 years). Ninety percent of them received monotherapy at the usual dosages, being treated with older or newer AEDs as follows: VPA (n=16), LTG (n=12), OXC (n=7), TPM (n=6), CBZ (n=4), ZNS (n=4), PHT (n=1), Clonazopam (n=1), and PGN (n=1). We compared cognitive measures in these patients with those of 48 age- and educationmatched healthy controls. As indicated in Table 1, the IQ and mood status were normal in both of the epilepsy patients and controls, but the working memory, executive function, and verbal fluency were worse in the patients. The performances of learning and memory tests did not differ between the two groups. These cognitive deficits were not correlated with patient characteristics, epilepsy variables, or type of AEDs. Interestingly, the verbal-fluency performance was well correlated with the AED treatment duration (r=-0.304, p=0.035). Executive dysfunction is a common cognitive side effect of AEDs, and we concluded that any AED can elicit cognitive impairment when it is used for a sufficiently long time.

It is clear that the long-term use of AEDs can elicit cognitive dysfunction in epilepsy patients, but the short-term cognitive effects of AEDs (for periods of up to a year) are not conclusive from recent studies. We attempted to summarize the current literature and draw general conclusions despite previous cognitive studies exhibiting many methodological problems. Table 2 summarizes the cognitive effects of AEDs obtained by searching MEDLINE up to January 2008. In general, comparisons of mean data across groups revealed that cognitive performances are worse for older AEDs than for placebo or a nondrug condition. However, comparisons of older AEDs reveal few differences, with only PB showing some consistency in a trend for inferior performance compared to other AEDs. Data availability is worse for newer AEDs than for older AEDs. In general, it appears that newer AEDs performed more favorably compared with placebo or nondrug than have the older AEDs and also performed as well or better than the older AEDs in comparative studies. Especially, there appear to be fewer adverse cognitive effects for GBP, LTG, and LEV than for CBZ. TPM is the newer AED that is associated with the greatest risk of cognitive impairment irrespective of the comparator group, although this risk decreases with the use of slow titration and low target dosages. Although ZNS and PGB are tentatively associated with some cognitive risks, the current data are insufficient for drawing definitive conclusions.

Various factors can affect cognition in patients with epilepsy, and clinicians require considerable skill to elucidate these factors and mitigate those that can be influenced. The cognitive impact of AEDs can potentially be alleviated, and is therefore important to address. This should be seriously considered in those who are at an age vulnerable to cognitive function, e.g., children and the elderly, and who require maximal cognitive efficiency for their job, school, or daily activities. We therefore recommend that clinicians consider evaluating cognitive function prior to adding or switching an AED, with this evaluation including neuropsychological screening. Since neuropsychological tests can objectively discriminate subtle changes in cognitive function, they might assist when deciding the appropriate AED for improving cognition.

Since findings from studies usually only provide information on the "average" patient, individualization of AED therapy is essential. If a certain AED impairs cognition, trying another AED with a differing pharmacologic profile might be successful. Most of all, it is critical to guard against overmedication. Slow titration to the lowest effective dosage and avoiding polytherapy when possible are keys to success. Refractoriness to AEDs should be determined as rapidly as possible. In refractory patients, clinicians should consider video EEG monitoring to confirm the seizure diagnosis and determine if the patient is a candidate for epilepsy surgery.