Chapter 4 Long-Term Effects of Seizures on Brain Structure and Function
Perhaps one of the most common questions asked by patients with epilepsy or by parents of children with epilepsy is, “What are the seizures doing to the brain?” Implicit in the question is a concern that the seizures are having a negative effect. The number of questions that derive from this one are almost endless: “Are the seizures causing brain damage?” “Will my child’s cognitive development be impaired by having seizures?” “Will the seizures cause more seizures?” The answers to these and other related questions are as varied as the causes of epilepsy, the types of seizure, and the ages of the patient when the seizures occur. In short, there is no single, simple answer. The physiology associated with seizures varies from a short run of rhythmic spike and wave activity to prolonged tonic discharges. Some types of epilepsy are seemingly benign and self-limited, whereas others are inexorably progressive. Seizures can occur in a neonatal brain with years of development to come or in aging brains that have suffered the ravages of time and are hanging by their fingertips over the abyss of oblivion. In approaching these questions, we therefore have to consider, among other things, which kind of seizure, what cause, and what type of brain. Many of the answers are often colored by the physician’s interpretation of imperfect data, an interpretation that can be further clouded by attempts to extrapolate data across seizure types, developmental state, and species. In this chapter, we will provide an overview of some of the relevant data. However, we cannot offer clarity where none exists. Thus, readers will find answers that contain enough caveats to cast doubt on any attempt at certainty. Our goal is to provide a perspective on the complexities involved in answering your patients’ and their families’ legitimate questions.
Before we enter the discussion regarding the effect of seizures on the brain, it is essential that we differentiate among the different types of epilepsy and seizures because it is clear that the pathophysiology of the many syndromes and situations are quite varied in both cause and consequence. These are important distinctions to make, as one of the problems that we encounter in determining the effect of seizures on the brain is that many of the causes involve preexisting abnormalities or destructive injuries. Thus, claiming any change found as a potential consequence of seizure activity cannot be supported. The primary distinction to be made between seizure types is between status epilepticus (SE) and intermittent spontaneous seizures. Even though SE is often caused by events such as stroke, infection, or trauma, there is ample evidence, especially from the laboratory, that after a duration of as short as 30 minutes, the seizure activity itself can induce neuronal loss.1,2 However, it is likely that even in the general area of SE, there are types that cause injury and types that do not. Convulsive SE or SE associated with high-frequency discharges on electroencephalogram (EEG) clearly can cause damage, even when the high-intensity physical convulsions are controlled.3,4 Absence SE or spontaneous episodes of prolonged partial SE associated with lower frequency discharges (3 c/s spike and wave or rhythmic delta/theta activity) may not be, as there appear to be no long-term consequences to these episodes.
The other major grouping of seizures are intermittent spontaneous, which is the type that will be encountered most often in clinical practice. These seizures range from benign neonatal convulsions to infantile spasms (West syndrome) to absence seizures and partial seizures associated with well-defined focal pathology with many types and etiologies. Each seizure type has a distinctive physiology on EEG, a variety of underlying pathologies, and a defined natural history ranging from those that stop on their own with no identified poor outcomes, to intractable epilepsy associated with severe neurological and cognitive problems. The extent to which these different outcomes are the consequence of continued seizures and the contribution is from the underlying neuropathology, which is at times quite severe, is unclear. In our attempt to clarify the problem, we will draw on a variety of animal studies as well as a number of clinical observations. In each instance, keep in mind that no finding can be universally applicable to all seizure types and all forms of epilepsy. To answer the question for any given human syndrome, it will ultimately be necessary to extrapolate, often imperfectly, from what is known about the syndrome’s natural history and pathophysiology, as well as relevant experimental data, recognizing that not all laboratory observations are applicable to a particular (or in some cases any) human syndrome. Answering the questions for any one patient will require careful consideration and judgment. We will begin the discussion of these issues with some clinical examples, which will illustrate the complexity of differentiating cause from consequence as well as the inappropriateness of applying a universal rule to all epilepsies, which are a wondrously varied collection of syndromes.
The developing brain is predisposed to seizures,5,6 and seizures are common in the neonate with an estimated incidence of 1 to 6 per 1000 live births.7–11 The term neonatal seizures encompasses seizures of diverse etiologies ranging from acute symptomatic causes (e.g., hypoxia-ischemia, intracranial hemorrhage, infections, and metabolic derangements such as hypocalcemia) to remote symptomatic causes (e.g., cerebral dysgenesis and inborn errors of metabolism) to idiopathic, genetically determined epilepsies (e.g., early myoclonic encephalopathy of infancy and familial neonatal convulsions).
As a group, the outcome for children with neonatal seizures, although improving, is poor.12 The mortality rate for neonates with seizures is high, with values ranging from 15 to 30%. In survivors, approximately 30% will develop neurological sequelae that include epilepsy, mental retardation, and cerebral palsy. Those most at risk for mortality and neurological morbidity are those neonates with persistent, difficult-to-treat seizures.13–17 It has been proposed that these seizures produce neurological injury as the result of energy depletion and cerebral metabolism dysfunction.18 However, because these prolonged, persistent seizures tend to be observed in those neonates with the most severe CNS involvement and injury (e.g., severe hypoxic ischemic encephalopathy), it is difficult to distinguish between the secondary damage, if any, imposed by the seizures and the primary injury imposed on the brain by the underlying etiology.
There are good examples in which neonatal seizures tend to be associated with a more favorable outcome. One example is the genetically determined benign familial neonatal convulsions, an autosomal dominantly inherited epilepsy syndrome resulting from mutations in voltage-gated potassium channel. These convulsions are brief events commencing as tonic posturing with apnea and other autonomic features, which can evolve to clonic movements in otherwise healthy neonates. These convulsions typically commence during the first week of life, can be observed multiple times per day, and may be difficult to treat. In most cases, the seizures will remit by 16 months of age. Although there is a increased risk of febrile and afebrile seizures later in life, psychomotor development tends to be normal.19 For those with psychomotor delay, it is likely that this outcome is the result of genetic factors and not the seizures.20
In summary, caveats abound when attempting to answer the question of whether seizures damage the neonatal brain. Although all recognize that the answer to this question is important, clinical data are limited.21 It is assumed that few would argue against the concept that in some circumstances seizures in the neonatal brain impose additional injury, but not all would agree on the circumstances or the extent of the injury, especially as some neonatal seizures have a benign outcome.
For many children with epilepsy, the outcome is good. The seizures remit and may not recur, even after medication is discontinued. Although cognitive and behavioral disorders may be observed transiently during the time period when the epilepsy was most active and required treatment with an antiepileptic medication, these children have no apparent evidence of a significant, long-term disability resulting from their epilepsy. In contrast, a group of children with epilepsy will have an associated psychomotor slowing or regression that persists even in the absence of seizures. In some of these children, data suggest that the seizures or the interictal EEG abnormalities contributed to the regression. Most childhood epilepsies fall somewhere between these two extremes (Table 4-1).
|Benign familial neonatal convulsions|
|Benign myoclonic epilepsy of infancy|
|Childhood absence epilepsy|
|Juvenile absence epilepsy|
|Benign rolandic epilepsy|
|Juvenile myoclonic epilepsy|
|Early myoclonic encephalopathy|
The epilepsies in which the outcome tends to be good, the so-called benign epilepsies, include benign rolandic epilepsy, childhood absence epilepsy, and juvenile absence epilepsy. The epilepsies in which a potentially long-lasting psychomotor slowing or regression is observed include Ohtahara syndrome (early infantile epileptic encephalopathy with suppression burst), West syndrome (infantile spasms), Dravet syndrome (severe myoclonic epilepsy of infancy), Lennox-Gastaut syndrome, and Landau-Kleffner syndrome. For purposes of illustration, we will review two of these disorders, the evidence supporting the absence of a long-term effect of the epilepsy in one (benign rolandic epilepsy), and evidence supporting the long-term effect of the epilepsy in the other (West syndrome).
Benign rolandic epilepsy (BRE; also known as benign childhood epilepsy with centrotemporal spikes or BECT) is a common childhood epilepsy accounting for 15 to 25% of all childhood epilepsies. This syndrome is named for the sleep-activated interictal centrotemporal spikes with a tangential dipole observed on EEG and its characteristic partial seizures of the face and tongue that, especially during sleep, may secondarily generalize. It is most commonly observed in boys between the ages of 3 and 13 years with a peak at 9 to 10 years.
Given that the seizure frequency associated with this syndrome is low,22 and that the seizures most commonly occur during sleep, the majority of children with this syndrome do not require antiepileptic therapy.23 In addition, these children outgrow the epilepsy, and the interictal EEG abnormality resolves during their teenage years.24 That the seizures remit and the rate of epilepsy in these children in adulthood is no greater than that in the general population24 is in contrast to the adage that seizures beget seizures.
Although BRE is considered a benign epilepsy, cognitive and behavioral difficulties may be observed.25–28 However, for the most part, these difficulties are transient and appear to be related to the severity of the interictal discharges and not the result of a permanent injury to the brain.29 Indeed, as a group, the developmental outcome for these children, including those with cognitive and behavioral difficulties, whether treated or not, is good. A potential alternative explanation is that an underlying genetic predisposition underlies both the epilepsy and cognitive dysfunction.
West syndrome is a cryptogenic or symptomatic generalized epilepsy syndrome composed of infantile spasms, a hypsarrhythmic EEG, and mental retardation. The spasms, which tend to occur in clusters, are characterized by a brief, massive flexion, extension, or mixed contraction of the axial musculature coincident with an electrodecrement on the EEG. The incidence for this syndrome ranges from 3 to 4.5 per 10,000 live births, and the majority of children present during the first year of life.30–32 Causes of West syndrome are multiple and include the neurocutaneous syndromes (e.g., tuberous sclerosis complex), brain malformation and tumors, inborn disorders of metabolism, and genetic disorders, as well as acquired injuries from fetal infections, trauma, and hypoxic-ischemic injury.
As a group, the long-term prognosis is poor and includes medically refractory epilepsy, cognitive impairments ranging from learning disorders to mental retardation, and autism. To a large extent, like neonatal seizures, the prognosis is dependent on the etiology.33 However, medical and surgical reports demonstrating improved developmental outcome in those children with the shortest epilepsy duration and prompt response to therapy suggest that the seizure activity itself may be detrimental. But once again, caution in interpretation is required, as those with a short treatment lag or who responded quickly to therapy or potentially those most amenable to surgical resection may be those who would have had a better developmental outcome, even in the absence of effective treatment.34
One of the areas of great uncertainty regarding the long-term consequences of seizures is the potential relationship between febrile seizures as a young child and the development of mesial temporal lobe epilepsy (MTLE) with hippocampal sclerosis/atrophy as an adult. Patients with MTLE are much more likely to have a history of febrile seizures as a child, and this association has raised the issue of whether febrile seizures may in some way cause the limbic pathology that eventually leads to MTLE.
On the other hand, a number of epidemiological studies indicate that the risk for developing epilepsy (or any other negative consequence) following febrile seizures is not greater than for children who never experienced them. However, the studies have also revealed that febrile seizures are not a single entity. Some are classified as simple and last less than 5 minutes, and these do not carry any additional risk for a poor long-term outcome. Complex febrile seizures (focal ictal behavior or seizures lasting more than 30 minutes), especially those that are prolonged, have a much higher risk for poor long-term outcome and the development of chronic epilepsy.