Epilepsy, Cognition, and Behavior: The clinical picture

Abstract

Although epilepsy is defined by the occurrence of spontaneous epileptic seizures, a large body of evidence indicates that epilepsy is linked to a spectrum behavioral, psychiatric, and cognitive disorders as well as to sudden death. Explanations for these associations include: (1) The effects of structural lesions which may impair the functions subserved by the regions of the brain involved in the lesion. (2) The effects of seizure activity which may begin well before a clinical seizure occurs and may persist long after it is over raising questions about what truly constitutes “interictal.” In addition, encephalopathic effects of epilepsy in infancy during critical periods in development may be particularly severe and potentially irreversible. (3) Shared mechanisms underlying seizures as well as these other disorders in the absence of structural lesions or separate diseases of the CNS. Epidemiological and clinical studies demonstrate the elevated risk of cognitive, psychiatric, and behavioral disorders not just during but also prior to the onset of epilepsy (seizures) itself. These may outlast the active phase of epilepsy as well. The mounting evidence argues strongly for the recognition of epilepsy as part of a spectrum of disorders and against the notion that even uncomplicated epilepsy can a priori be considered benign.

Epilepsy is defined by its quintessential and obligatory symptom, spontaneous epileptic seizures. Most of the focus of clinical care including AED and surgical therapy is directed at controlling these seizures. A large literature repeatedly highlights the observation that there are many other disorders that co-segregate with epilepsy and occur in people with epilepsy at rates much higher than what would be expected in the general population. People, especially those with childhood onset epilepsy, do not do as well as expected with respect to school completion, employment, marriage, and parenthood (Camfield et al. 1993, Gaitatzis et al. 2004, Shackleton et al. 2003, Strine et al. 2005, Tellez-Zenteno et al. 2007). They are more likely than others to report symptoms of depression, sleep disturbances, and pain syndromes. Approximately a quarter of all children in the population with epilepsy have impaired intellectual function consistent with the designation “mental retardation.” (Berg et al. 2004, Camfield et al. 2002). Death rates are higher from all causes(Lhatoo et al. 2001)but of particular concern are sudden death and suicide. Sudden death may account for 10–20% of the mortality seen in epilepsy, may occur at any age, and is particularly of concern in patients with uncontrolled tonic-clonic seizures (Walczak et al. 2001, Nashef et al. 2007)although asytole, a potential mechanism for SUDEP, can be seen with nonconvulsive seizures as well (Rocamora et al. 2003, Zijlmans et al. 2002). Suicide risk is also increased by up to five-fold following the diagnosis of epilepsy (Christensen et al. 2007). Finally, there is a strong association between autism and epilepsy in childhood (Tuchman et al. 1991).

There are many explanations for these findings. This review focuses on (1) the impact of the underlying structural lesions and other conditions that cause epilepsy, (2) the effects of seizures and inter-ictal discharges on brain function, and (3) potential shared mechanisms that may produce epilepsy and independently cause some of the other disorders seen so frequently in people with epilepsy.

Structural lesions

Approximately a quarter of all childhood epilepsy occurs in association with identifiable structural brain lesions, presumed early insults as evidenced by cerebral palsy, or other metabolic-genetic encephalopathies. Pre-and perinatal hypoxic-ischemic insults and stroke as well as malformations of cortical development are the most common structural lesions seen in childhood onset epilepsy (Berg, et al. 2009). In adults, one would expect to see a different spectrum of causes with stroke, tumor, and trauma being most common. Most of our information regarding underlying causes is from developed countries with relatively low infant mortality rates and long life-expectancies. Ultimately, the causes of epilepsy reflect the epidemiology of neurological morbidity and can be expected to vary from one age group to another and from one region of the world to another. To the extent that the integrity of brain function is dependent on the integrity of brain structure, it is not surprising that people whose epilepsy is attributable to a structural or metabolic cause have high rates of a variety of difficulties. In one study, 64.5% children with abnormal brain imaging had intellectual impairments (IQ<80), and 9.4% died. By comparison, in those without structural-metabolic causes, 15.2% were intellectually impaired and <1% died (both p-values <0.0001) (Berg et al., 2009).

Effects of seizures and interictal discharges

Seizures and interictal activity may also contribute to cognitive and behavioral disturbances. These effects may be loosely partitioned into three types although the boundaries between these categories are increasingly blurred.

1. Episodic transient effects

In addition to the obvious disruptions due to seizures themselves, some evidence indicates that interictal discharges, brief enough not to produce overt, recognizable clinical seizures, may still have subtle but important effects on the quality of ongoing cognitive activity (Kasteleijn-Nolst Trenité et al. 1987). A recent review suggested that much of what was being called “interictal” epileptiform activity was in fact ictal activity associated with subtle seizures (Aldenkamp and Arends 2004). We are beginning to understand that there is likely a continuum in the manifestations and impacts of epileptiform discharges. This raises questions about where exactly to draw the line between ictal versus interictal and seizure versus non-seizure.

Post-ictal disruption of behavior and function is a well-documented phenomenon. Abundant clinical anecdote and now some evidence from both humans and animal models suggests that the postictal effects of some seizures may persist in subtle form for far longer than previously appreciated. In a study of patients with newly-presenting motor seizures, Badawy et al. demonstrated that 24 hours prior to a seizures, there were discernable increases in motor cortex excitability as measured by transcranial magnetic stimulation. During the 24 hours after, there was a detectable decease (depression) in excitability in the same areas (Badawy et al. 2009). These were motor seizures; however, potentially similar patterns may occur in association with virtually any type of seizure leaving the question of the impact of this subtle postictal suppressionon subtle cognitive functions. In adult rats, Lin demonstrated seizure-related declines in a spatial hidden goal task. Rats that had been trained in this task increasing lost proficiency over 11 days day during which a seizure was induced each day. During the 9-day post-seizure recovery period, experimental rats eventually returned to baseline levels of performance, similar to control rats (Lin et al. 2009). These types of studies extend the notion of postictal and point to prolonged subclinical postictal effects. Further, if these findings accurately reflect what happensin uncontrolled human epilepsy, it raises that possibility that patients with daily seizures may never have periods in which they are truly interictal but are always to some degree in an ictal or peri-ictal state.

2. Progressive effects

In adults, and particularly in the case of mesial temporal lobe epilepsy, there is good reason to believe that the structural hippocampal damage is progressive (Briellmann et al. 2002). Damage may extend beyond the hippocampus and surrounding limbic structures. A small but potentially important thinning of the neocortex has been documented to occur over time in patients with ongoing MTLE (Bernhardt et al. 2009). These changes may explain the progressive decline in cognitive function, particularly memory functions, seen in these patients. In one series, substantial proportions (25–55%) of patients with MTLE incurred significant losses on tests of memory, confrontational naming, and also fine motor function over the course of four years(Hermann et al. 2006). Fewer than 5% of controls (as expected by chance) experienced such losses.

3. Encephalopathic effects

The concept of epileptic encephalopathy is not new and is often used to refer to a group of very severe epilepsy syndromes with onset typically in infancy though early childhood (Dulac 2001). The recent report from the ILAE Commission on Classification and Terminology (Berg et al. 2010)formally defined the concept as meaning that “ The epileptic activity itself may contribute to severe cognitive and behavioral impairments above and beyond that expected from the underlying pathology alone (e.g. cortical malformation).” Evidence from laboratory models of long-term potentiation, the cellular basis of learning provides a plausible basis for seizure activity to disrupt learning, memory and other functions (Lamprecht and LeDoux 2004). A critical departure from other work in this area was the proposition that the encephalopathic effects of seizures might occur in any form of epilepsy and at any age. In fact, given the evidence of prolonged transient effects of individual seizures (e.g. (Badawy et al., 2009, Lin et al., 2009), some of the encephalopathic effects could represent sub-clinical postictal phenomenon which, if seizures are frequent, could be virtually continuous.

The work of Lin et al. cited above was in adult rats. The effects were reversible. The reason for the concept of epileptic encephalopathy having developed in childhood epilepsy first is because of the dramatic deficits seen in children with syndromes such as West, Dravet, and Lennox-Gastaut. Approximately three quarters of children with these syndromes have mental retardation or are developmentally devastated (Berg et al. 2008a). The risk of autism or autistic features is also high, particularly in West syndrome (Saemundsen et al. 2007)suggesting the epilepsy, mental retardation, and autism form a triangle of encephalopathy. In these and other syndromes, there is a strong impression that the epileptic activity itself interferes with normal processes of brain development during critical times in development. In essence, the seizures create an epileptic cacophony which derails the initial efforts to establish functional connections and networks.

The process underlying epileptic encephalopathy need not be limited to those electro-clinical syndromes traditionally identified as such but should apply to virtually any form of epilepsy. Young age at onset, however, should still confer a high vulnerability, and all evidence so far indicates that it does. For example, in a series of children referred for surgical evaluation, Vasconcellos studied the association of age at onset, seizure frequency, and underlying cause with cognitive test scores (Vasconcellos, et al. 2001). Their results demonstrated a clear strong correlation between younger age at onset and lower IQ (r=0.36, p<0.001). The combined impact of seizure frequency and young age at onset was striking. Of children with age at onset >2 years and with seizures occurring monthly to weekly, 3% had mental retardation. In that same age group, 17% of those with daily seizures had mental retardation. In children <2 years at onset, the comparable figures were, respectively, 22% and 65%. Specific underlying cause did not explain these findings. Thus two factors, young age at onset and daily seizures accounted for a nearly 22-fold difference in the proportion with mental retardation. Other studies have highlighted the relatively poor cognitive outcomes seen in association with younger age at onset. In a separate surgical referral series of children with intractable temporal lobe epilepsy, full-scale IQ was again strongly correlated with age at onset (r=0.39, p=0.001) although not with duration of epilepsy (r=−0.06, p=0.64) (Cormack et al. 2007). Of children with onset during the first year of life, 82% were intellectually disabled (including not assessable) compared to 32% of those with older age at onset (p<0.001). In a study of adults with chronic temporal lobe epilepsy, patients with younger age at onset (<7 years, reflecting the median of the case group) had lower cognitive scores than did cases with older age at onset (Hermann et al. 2002). Older onset cases were closer to healthy controls than they were to younger-onset cases. Finally, in a community based study of all forms of epilepsy, children with younger age at onset (<5 years) were more likely to have evidence of intellectual impairment compared to those with older onset regardless of the type of epilepsy and regardless of whether the epilepsy was pharmacoresistant or not (Berg et al., 2008a). In all, these studies converge on the conclusion that young age at onset is associated with greater cognitive consequences. The hypothesis of epileptic encephalopathy may explain at least part of this repeated finding.

The hypothesis of epileptic encephalopathy also comes with the prediction that effective intervention may reverse some of the cognitive losses and in the case of developmental encephalopathic effects, rescue development from what might otherwise be irreversible effects. This is very hard to study in the human situation; however, repeated observations, largely from the surgical literature provide some support for this notion. Two surgical studies of children with very young age at onset (average 1–2 years) found striking correlations between duration of epilepsy prior to surgery and post-surgical developmental quotients (DQ); the longer the delay to surgery, the lower the DQ (Freitag and Tuxhorn, I 2005, Jonas et al. 2004). Generally, improvements were confined to children who became seizure free after surgery. Although the preference is for early intervention, it is possible that even delayed intervention may still be of some benefit. In a series of children with Lennox-Gastaut syndrome who underwent resective surgery to treat their epilepsy, a 9-point gain in pre-to postsurgical DQ was found in those children who became completely seizure-free even though many of them had substantial delays prior to surgery (Lee, et al. 2010). Although additional evidence is needed, the current evidence is compelling enough and strongly supports a recent recommendation from the ILAE Subcommission on Pediatric Epilepsy Surgery that any child whose seizures have failed to come under complete control after the use of two or three AEDs should be evaluated at a comprehensive epilepsy center (Cross, et al. 2006).

Essential Co-morbidity

In addition to the obvious effects of demonstrable brain lesions and other disorders and apart from the effects of the seizures themselves, there is a growing body of literature demonstrating that various cognitive and psychiatric co-morbidities may precede the onset of epilepsy, that is these disorders appear to identify individuals at increased risk of developing epilepsy. This pattern of findings suggests the possibility that some of these various co-morbidities may share common mechanisms with epilepsy, the hypothesis of “essential co-morbidity.” This is of particular interest in the case of epilepsy in people with no identifiable structural lesion or metabolic disorder underlying their epilepsy who are sometimes referred to as having “just epilepsy.”

In the recent SANAD study (Taylor et al. 2009), patients with newly diagnosed but not yet treated epilepsy as well as healthy controls participated in a battery of neurocognitive tests. Even prior to initiating treatment, cases scored relatively worse compared to controls on 11 of 14 tests, in some cases substantially so. These findings suggest that subtle cognitive impairments likely preceded the onset of epilepsy as these patients had only had a few seizures at the time of testing.

In a population-based, prospective incident case-control study, children with newly diagnosed epilepsy were compared to age and sex-matched controls from the population. A standardized psychiatric interview was used. Cases were two to three times as likely as controls already to have met criteria for attention deficit disorder. This was true for boys and girls separately and particularly true for the inattentive type of the disorder.

Within that same study and extending the age range into adulthood, the authors found that cases were more likely than matched controls to have, in the past, suffered significant symptoms of depression. They were approximately five times as likely to have attempted suicide. Thus, the association goes both ways: people with epilepsy are at increased risk of committing suicide (Christensen et al., 2007)and people who are depressed or who have attempted suicide are more likely to develop epilepsy (Hesdorffer et al. 2006). In this context, it is worth noting that there is an increased risk of cardiac events (including sudden death) associated with both depression (Frasure and Lespérance, 2008)and with epilepsy (Nashef et al. 2007). Reviews by Dr. Kanner and by Dr. Richerson in this issue explore some of the mechanisms linking these three disorders.

Co-morbid conditions after epilepsy is resolved

Epilepsy often lasts for only a few years. At this point in time, we have very little information about whether the cognitive, behavioral, and psychiatric disorders seen in epilepsy resolve with the epilepsy itself or whether they persist. Certainly in the case of disorders directly attributable to an underlying lesion, we would not expect the disorder to resolve. In the case of developmental encephalopathic effects, we would also not expect a full recovery. In the case of shared mechanisms “essential co-morbidity,” it is less clear. Some studies have found poor educational and social outcomes in people with childhood onset epilepsy (Camfield et al., 1993, Shackleton et al., 2003)however, those studies do not entirely address the issue. In one prospective study of childhood onset epilepsy, cases with epilepsy of genetic or unknown cause were compared to their sibling controls (Berg et al. 2008b). Only cases and controls with full scales IQ ≥60 were included. Cases were more likely than their sibling controls to have IQ scores in the borderline to very mild MR range(60–79) (13% vs 5%, p=0.005). Cases had lower scores than their sibling controls on almost all cognitive and achievement measures. When comparisons were limited to only pairs in which both case and control had FSIQ≥80, a residual effect was still seen in processing speed. Adjustment for the case’s remission and drug status did not influence the findings. This could represent a subtle but potentially important form of residual cognitive difficulty. In light of studies reviewed above that demonstrate that depression, suicidality, attention deficit disorder, and subtle cognitive impairments preceded the onset of seizures, we are left with the question whether these co-morbidities of epilepsy may persist long after the seizures themselves have been controlled or, as in the case of some pediatric syndromes, completely resolved. Is there a post-clinical phase to otherwise seemingly uncomplicated epilepsy, and if so, how is it to be addressed?

Conclusions

Epilepsy is a complex set of brain disorders with epileptic seizures being the quintessential defining element. Abundant evidence makes it clear that there is a spectrum of cognitive, behavioral, and psychiatric disorders that appear to be part and parcel of many forms of epilepsy. Investigations in so many areas of the neurosciences, as discussed by others in this supplement, are beginning to elucidate the basis for these associations, and these discoveries may contribute to the development of therapies and management techniques that will better enable physicians to treat the full spectrum of disorders that epilepsy entails. In this context, it is worth noting that the ILAE recently recommended that the term “benign” no longer be used to describe epilepsy precisely because of the large number of disorders often seen in association with even relatively uncomplicated epilepsy (Berg et al., 2010). For the practicing clinician, perhaps the most important message to remember is that there is no such thing as “just epilepsy.” Any patient with epilepsy should be viewed as someone who is at risk of encountering a variety of consequences including cognitive difficulties, behavioral disorders, depression, suicide, and also sudden death. There is absolutely no reason to re-stigmatize epilepsy, but there is every reason to make sure physicians, patients, families, as well as educators and others are adequately prepared to recognize any such difficulties and to manage them appropriately as they arise.