Post-Traumatic Epilepsy: Review of Risks, Pathophysiology, and Potential Biomarkers

The risk factors for PTE include advanced age, penetrating injuries, injury severity (e.g. neurosurgical procedure, intracranial hemorrhage, greater than 5 mm midline shift, duration of coma >24 hours, loss of consciousness >24 hours, prolonged length of post-traumatic amnesia), biparietal or multiple contusions, and frontal or temporal locations of the lesion.^7–10,27,28 In penetrating brain injury, a foreign body pierces the bony skull and passes into (and in some cases through) the substance of the brain. This leads to physical disruption of neurons, glial cells and fiber tracts. The estimated relative risk of seizures after penetrating war injuries is very high as compared with the risk in the general population.^8,10 Studies of the consequences of missile injuries to the head from World War I up through conflicts in the Middle East during the 1980s have shown a remarkably consistent pattern in terms of the development of epilepsy after this severe form of TBI. The Vietnam Head Injury Study, for example, found that incidence of PTE was 53% during the 15 years after the injury, with half still having seizures after 15 years.^29–31

Several population based studies have measured risk of developing PTE in groups (most or all civilians with nonpenetrating injuries) containing both inpatients and outpatients (Figure 2). A study from Denmark (1,605,216 subjects; 78,572 had at least one TBI) found that compared with no brain injury, the relative risk of developing epilepsy for people older than 15 years was 3.5 after mild TBI and 12.2 after severe TBI.² Risk was still elevated at 10 years after injury (1.5 for mild, 4.3 for severe). A U.S. study (4,541 subjects with a single TBI) reported 5-year cumulative PTE probability was 0.7% for mild, 1.2% for moderate, and 10% for severe TBI.¹ The 30-year cumulative PTE probability was 2.1% for mild, 4.2% for moderate, and 16.7% for severe TBI. A study in Taiwan (559,658 subjects age 15 years or older; 19,336 had at least one TBI) found that compared with no brain injury the hazard ratio (a different measure of relative risk) for PTE was 3.9 for mild TBI and 7.8 for severe TBI.³ Population-based studies that focused on development of PTE following hospitalization have reported relatively similar results (Figure 2). A U.S. study (2,118 patients with TBI age 15 years or older) reported 3 years cumulative incidence per 100 persons of PTE was 9.1 across injury severity.⁴ The cumulative incidence per 100 persons was 4.4 for mild TBI, 7.6 for moderate TBI, and 13.6 for severe TBI. Peak onset was in year two.⁴ A study from China (2,826 patients with TBI age 4 to 79 years) reported a 3 year cumulative incidence of PTE of 5% across injury severity.⁵ Cumulative incidence was 3.6% for mild, 6.9% for moderate and 17%, for severe TBI. Onset was within the first year in the majority of cases.⁵ Although there are many methodological differences across these studies, these results suggest that even mild TBI is associated with clearly increased risk for PTE. Variability in seizure occurrence may also be attributable to the unique genetic makeup that each patient brings to post-TBI recovery as not every patient with clinical risk factors develops PTE.^32–35

The development of PTE after a latent period is an example of human epileptogensis, whereby a nonepileptic brain undergoes molecular and cellular alterations after a brain insult, which increases its excitability.^7–9 This eventually leads to the occurrence of recurrent spontaneous seizures. Penetrating brain injury produces a cicatrix (scar) in the cortex and is associated with an increased risk of PTE of approximately 50%.²⁹ The initial disruption of tissue may be compounded by ischemia and hemorrhage. Nonpenetrating injuries, including focal contusions and intracranial hemorrhages, are associated with a risk of PTE of up to 30%. In this setting, the mechanism of epileptogenesis may be partly related to the toxic effects of hemoglobin breakdown products on neuronal function.^8,36 Closed head injuries produce diffuse axonal injury with shearing of axons, diffuse edema, and ischemia. The most common locations are the gray matter-white matter junction particularly in the frontal and temporal areas.³⁷ This process leads to the release of excitatory amino acids, cytokines, bioactive lipids, and other toxic mediators causing secondary cellular injury.^7–9 Epileptogensis is thought to be initiated by specific types of cell loss and neuronal reorganization, which results not only in enhanced excitation, but also in decreased inhibition, predisposing to hypersynchronization.³⁸ Brain injuries also lead to upregulation of proinflammatory cytokines and activated immune responses to further increase seizure susceptibility, promote neuronal excitability, and impair blood–brain barrier (BBB) integrity.^39,40 Early changes include the induction of immediate early genes and post-translational modifications of neurotransmitter receptor and ion channel/transporter proteins.⁴¹ Within days, neuronal death, initiation of an inflammatory cascade, and new gene transcription has been reported to occur.⁴² Later (days to weeks) anatomic changes include axonal sprouting and dendritic modifications. For example, mossy fiber sprouting can be observed in chronic epileptic brain.⁴³ Over time, seizure threshold is lowered by a growing increase in excitability, and the risk of a seizure increases.⁴⁴

The process by which trauma to the brain tissue leads to hyperexcitability, hypersynchronization, and development of recurrent seizures is relatively unknown, but there are several theories as to potential mechanisms based on experimental data, primarily from animal models of PTE. Animal models of TBI (mostly performed in rodents) are divided into focal brain injury, diffuse brain injury, mixed models of focal and diffuse brain injury, and models of repetitive concussive injury.^45,46 These models utilize a direct impact on the epidural space on the brain tissue.⁴⁵ A large number of studies have used lateral fluid-percussion or central fluid-percussion TBI models that produce both gray matter and white matter injury.⁴⁷ Late spontaneous seizures (PTE) have consistently been reported in the rat fluid-percussion TBI model, which is the most commonly used model of human closed head TBI.⁴⁷ The physiological effects from TBI are considered to consist of several phases including primary injury, evolution of primary injury, secondary injury, and regeneration.^45,47 Multiple mechanisms are likely activated simultaneously or sequentially by brain trauma. Animal studies demonstrate that a large number of molecular and cellular alterations occur during the latent period that could be related to epileptogenesis, including cell death, gliosis, neurogenesis, axonal and dendritic plasticity, rearrangement of the extracellular matrix, and angiogenesis.⁴⁷

Conventional neuroimaging methods including CT and MRI provide limited prognostic information for predicting development of PTE. Hippocampal atrophy associated with PTE is one of the most identified findings (Figure 1). There is promise with newer advanced MRI sequences that are more sensitive to microhemorrhages (e.g., susceptibility weighted imaging, SWI) and white matter injury (diffusion tensor imaging, DTI).⁴⁸ The search for neuroimaging features that correlate with presence of PTE and so might serve as biomarkers indicating increased risk of developing PTE and/or of epileptogenic processes is still in the early stages.^49,50

DTI shows potential for being more sensitive to the microstructural changes (e.g., gliosis, cell loss, axonal sprouting) thought to be involved in epileptogenesis than conventional MRI, as several studies have reported altered DTI metrics (lower fractional anisotropy [FA], higher mean diffusivity [MD]) in areas appearing normal on conventional MRI.^51–55 One study found that FA was lower and the volume of abnormal tissue was greater in patients with TBI who developed PTE than in patients with TBI who did not.⁵¹ The authors noted that their results suggest that DTI may have the potential to predict presence of epileptogenesis following TBI.

As noted above, trauma-induced hemorrhage is associated with higher risk for PTE, and animal models suggest multiple potentially seizure-promoting changes including generation of substances that can induce further injury, such as highly reactive free radical oxidants.⁵⁶ A longitudinal prospective study of patients with TBI (mild N=38, moderate N=26, severe N=71) used neuroimaging at multiple time after injury to assess whether the presence of gliosis and/or residual hemorrhage (hemosiderin) during the first 2 years increased risk of developing PTE.⁵⁷ As expected, patients with focal brain lesions requiring surgical treatment had the highest risk of developing PTE. Neuroimaging evidence of contusion in which residual hemosiderin was only partially encompassed by gliosis at earlier times after injury, particularly those that progressed to full enclosure at later times, was also associated with a higher risk of PTE. Presence of small areas of hemosiderin with no associated gliosis was not associated with higher risk of PTE. No evidence of medial temporal sclerosis was present. As noted by the authors of this study, this might be due to the relatively early phase (up to two years after injury) and/or the lower resolution of the neuroimaging (1.0 tesla, 5mm slice thickness).

Focal alterations in BBB permeability are common following TBI, with biphasic changes reported in animal models.^58,59 The early (hours) increase in permeability is likely a direct result of the injuring event, whereas the delayed (days) increase in permeability probably reflects secondary events. A study of patients with mild TBI (median 3 months post-injury) with (N=19) and without (N=18) PTE reported that BBB disruption (identified by contrast-enhanced MRI) was present more frequently in the PTE group (82.4% versus 25%) and the volume of cortex affected was larger.⁶⁰ BBB dysfunction is seizure-promoting and may contribute to epileptogenesis.^39,58,59 Activation of astrocytes has been reported following BBB disruption, perhaps a result of entry into the brain parenchyma of injurious serum constitutes (e.g. albumin), and is associated with epileptogenesis in animal models.^39,58,59

Traumatic brain injury is a heterogeneous disorder involving several different mechanisms including concussive forces, acceleration-deceleration forces, blast injury, and projectile missiles. Each of these mechanisms can result in numerous clinical sequelae including but not limited to PTE. Numerous studies have documented the risks associated with the development of PTE with moderate to severe brain injuries conferring the greatest risk. The characteristics of the seizure itself will also reflect the anatomic heterogeneity of TBI, with the majority of the seizure foci occurring in the temporal and frontal lobes. The process of epileptogenesis that occurs as a result of a brain injury is still not fully understood but it remains a prime target for the study and application of antiepileptogenic therapy. A large number of clinical trials have aimed at the prevention of PET. All failed to either prevent or alter the period of epileptogenesis associated with TBI.^7,8,10