Increased Risk of Late Posttraumatic Seizures Associated With Inheritance of APOE ϵ4 Allele FREE

POSTTRAUMATIC epilepsy (PTE) is a major source of morbidity after traumatic brain injury (TBI), which complicates 25% to 30% of cases of severe head injury and 5% to 10% of cases of mild to moderate injury.¹ It is estimated that 5000 to 30 000 new cases of epilepsy each year result from TBI.² Posttraumatic epilepsy is the most common cause of new-onset epilepsy in young adults,³ is often refractory to medical therapy, and is the cause of medically refractory seizures in approximately 5% of patients referred to specialized epilepsy centers.^{4 - 5} Attempts to develop prophylactic antiepileptogenic therapy have so far been unsuccessful.^{2 ,6} Understanding the molecular mechanisms underlying epileptogenesis may allow the development of effective prophylactic therapies.

Primarily on the basis of studies on animal models, a large number of molecular events have been postulated to contribute to epileptogenesis after traumatic injury.^{7 - 8} These include release of excitatory amino acids, production of cytokines, bioactive lipids, oxygen free radicals, nitric oxide, lipid peroxides, activation of proteinases, and apoptosis. Genes regulating the function of these systems are potential targets for therapeutic intervention. However, since animal models are an imperfect model for human brain trauma, it is unclear whether such factors are important in human epileptogenesis. Therapies that have been validated as antiepileptogenic in animal models have failed in human trials.⁶ Recent progress in the Human Genome Project offers an alternate approach for determining which molecular factors play important roles in epileptogenesis after brain injury. Since most human genes are polymorphic, determining that inheritance of a particular variant of a certain gene is associated with increased risk of PTE would provide independent support for a role of that gene in epileptogenesis.

Apolipoprotein E (apoE) is the major lipid carrier molecule in the central nervous system^{9 - 10} and has been implicated in the neural response to injury.^{11 - 14} Three allelic variants of apoE exist, APOE ϵ2, APOE ϵ3, and APOE ϵ4, which have allelic frequencies of 0.075, 0.774, and 0.151, respectively, in most human populations. Inheritance of the APOE ϵ4 allele is associated with increased risk of Alzheimer disease,^{15 - 16} poor outcome after severe traumatic brain injury,^{17 - 18} ischemic cerebral infarction,¹⁹ and intracerebral hemorrhage,^{20 - 21} and is associated with faster progression of disability in multiple sclerosis.^{22 - 23} This study was undertaken to determine whether inheritance of APOE ϵ4 was associated with increased likelihood of developing late posttraumatic seizures in patients who suffered moderate to severe brain injury.

Subjects were recruited from patients admitted to the neurosurgical service at an urban level I trauma center. Subjects were enrolled during a 2-year period, from January 1, 1999, to December 31, 2001. To enrich for patients likely to develop late posttraumatic seizures, subjects were enrolled in the study if one or more of the following criteria were fulfilled: (1) cerebral contusion noted on computed tomographic scan, (2) any intracranial hematoma (epidural, subtotal, or intracerebral) noted on computed tomographic scan, (3) depressed skull fracture, (4) penetrating brain injury, or (5) early posttraumatic seizure (within 7 days of injury). Patients with preexisting epilepsy were excluded, as were patients with preexisting neurologic conditions (such as a brain tumor, nontraumatic hemorrhage, or major cortical infarction) that are commonly associated with epilepsy.

After informed consent was obtained, information was collected regarding severity of the injury, findings in the initial cranial computed tomographic scan, length of stay in the intensive care unit (ICU), length of stay in the hospital, and discharge disposition (whether to home, a rehabilitation unit, or a nursing home). Six months after injury, subjects were contacted by mail or telephone and a structured questionnaire was administered to assess functional outcome according to the Glasgow Outcome Scale–Expanded (GOS-E).²⁴ In addition, the structured interview asked about involuntary movements, transient alterations of consciousness, and abnormal motor, sensory, or psychic phenomena. These questions were adapted from a questionnaire used for more than 10 years at the University of Washington in PTE studies,^{2 ,6} kindly provided by Nancy Temkin, PhD. Information about each suspected event was reviewed by 3 collaborating epileptologists (R.D.-A., M.A.A., and P.C.V.N.) who were blinded as to the APOE ϵ4 genotype, and whether late posttraumatic seizures occurred was determined only when all concurred that the episode consisted of an epileptic seizure. In several instances, subjects were contacted again by telephone to obtain more details about their episodes. An abnormal electroencephalogram was not required for the diagnosis of late posttraumatic seizures. Although a single posttraumatic seizure does not fulfill the definition of PTE, 70% to 90% of patients with late posttraumatic seizures have 2 or more within 6 months.⁶

The institutional review board at the The University of Texas Southwestern Medical Center, Dallas, reviewed and approved of this project. All individuals provided written informed consent.

DNA was extracted from hair follicles by the method of Thomson et al.²⁵ After digestion with proteinase K (100 µg/mL), DNA was stored at 4°C until use. Genotyping at the APOE locus was carried out by a modification of the method of Chapman et al.²⁶ The upstream primer was 5′-TCCAAGGAGCTGCAGGCGGCGCCA-3′ and the downstream primer was 5′-ACAGAATTCGCCCCGGCCTGGTACACTGCCCA-3′. Polymerase chain reaction was carried out at 150nM concentration for each primer, 50µM for each for each dNTP (deoxynucleotide triphosphate), 2mM magnesium chloride, 10% dimethyl sulfoxide, and 2 units of Taq polymerase (New England Biolabs, Inc, Beverly, Mass). Reaction was carried out for 40 cycles using a melting temperature of 94°C, a reaction temperature of 72°C, and an annealing temperature of 65°C. The 227–base pair (bp) polymerase chain reaction product was separately digested by AflIII (New England Biolabs), which digests ϵ2 and ϵ3 into 177- and 50-bp fragments, but does not hydrolyze ϵ4, and by HaeII (New England Biolabs), which digests ϵ3 and ϵ4 into 195- and 32-bp fragments, but does not hydrolyze ϵ2. Polymerase chain reaction products were resolved by electrophoresis in 2% agarose slab gels (Agarose 1000; Invitrogen Corporation, Carlsbad, Calif), and bands visualized by staining with ethidium bromide.

Parametric data were analyzed by the 2-tailed Mann-Whitney test, and categorical data were analyzed by Fisher exact test or χ² test. Statistical analysis was performed with the InStat program version 3.0 (GraphPad Software, Inc, San Diego, Calif).

Two hundred four subjects met criteria for enrollment in the study. Outcome information was obtained from 106 subjects. The follow-up rate was 52%, which reflects the highly transient nature of our study population. There were no differences in age, sex, ethnicity, initial Glasgow Coma Scale (GCS) score, computed tomographic scan findings, or requirement for intracranial surgery between the 98 subjects lost to follow-up and those who completed the outcome questionnaire (Table 1). There was a trend suggesting that subjects lost to follow-up may have had less severe injury, such as a slightly shorter length of stay in the ICU and in the hospital, and lower need for inpatient rehabilitation therapy. However, only the slightly lower ICU stay reached statistical significance. Thus, although the follow-up rate is suboptimal, we believe these patients are a reasonable representation of the overall group.

Twenty percent (21/106) experienced at least 1 late posttraumatic seizure. This fraction is comparable but somewhat higher than the 13% to 15% incidence found at 6 months by Temkin et al.² Although our entry criteria were modeled on that earlier study, it is possible that unrecognized differences in injury severity may account for the difference. Table 2 compares the group who developed late posttraumatic seizures with those who did not in terms of demographic information, severity of injury, need for an intracranial surgical procedure, the presence of early posttraumatic seizures, length of stay, disposition, and outcome. There was no difference between the 2 groups as far as age, sex, and ethnicity. There was a trend suggesting that the group developing late posttraumatic seizures suffered somewhat more severe injury, as witnessed by the lower admission GCS score, greater likelihood of craniotomy, longer ICU stay, and longer total hospital stay. However, none of these trends reached statistical significance. There was no difference between the 2 groups in the fraction of patients who required inpatient rehabilitation services on discharge from the acute care hospital. Finally, 6 months after injury, there was no significant difference in functional outcome between the 2 groups, at least as measured by the GOS-E.

As anticipated, subjects who experienced early posttraumatic seizures (within 7 days of injury) were at significantly greater risk of developing late posttraumatic seizures (relative risk, 3.33; 95% confidence interval, 1.54-7.22; P = .03). We relied on medical and nursing staff to identify early posttraumatic seizures, and electroencephalographic studies were obtained in only a few cases at the request of the attending staff. It is likely that subclinical or subtle early seizures occurred in our patients but were not identified as such.²⁷ We are unable to determine whether these subclinical or nonconvulsive early seizures were associated with a greater risk of late seizures, or whether they were more prevalent in APOE ϵ4 carriers.

Subjects who developed late posttraumatic seizures were more likely to carry at least 1 copy of APOE ϵ4. The relative risk for ϵ4 carriers developing PTE was 2.41 (95% confidence interval, 1.15-5.07; P = .03). Only one individual in our cohort was APOE ϵ4/ϵ4 homozygous, and that patient developed late posttraumatic seizures. Since only 2.3% of the population is homozygous for ϵ4, our sample size was too small to determine whether inheritance of 2 copies of the ϵ4 allele confers a greater risk than inheritance of a single allele. Most subjects experienced more than 1 late posttraumatic seizure. There was no difference in the strength of the association with APOE ϵ4 when only data from subjects with multiple late posttraumatic seizures were analyzed.

Table 3 summarizes the injury severity and outcome results by APOE ϵ4 carrier status. We did not find an association between inheritance APOE ϵ4 and initial GCS score, hospital course, or 6-month outcome, as measured by the GOS-E. Stratifying GCS and GOS-E scores and analyzing them as categorical variables also did not result in a statistically significant association with ϵ4 status.

Apolipoprotein E is the major lipid carrier protein in the central nervous system,⁹ where it is produced primarily by astrocytes,¹⁰ although neurons and microglia may also contribute to apoE synthesis. In vivo, apoE expression is upregulated after injury (for review see Poirier¹¹ ), and recent work with transgenic mice indicates that apoE has a role modulating learning and memory,²⁸ structural plasticity during development and aging, and cell death after ischemic^{12 ,14} or convulsive¹³ brain injury.

In humans, apoE exists in 3 allelic forms: APOE ϵ2 (Cys112-Cys158), APOE ϵ3 (Cys112-Arg158), and APOE ϵ4 (Arg112-Arg158).²⁹ In mixed white populations, the allele frequencies for ϵ2, ϵ3, and ϵ4 are approximately 0.08, 0.77, and 0.15, respectively.^{30 - 31} There are some indications that the allelic frequency of APOE ϵ4 is higher in African Americans,³² a finding confirmed in our study.

Apolipoprotein E first came to the attention of neurologists as a result of the genetic association between APOE ϵ4 and late-onset Alzheimer disease (AD), which was first reported by Roses and collaborators in 1993¹⁵ and has since been confirmed by many independent workers.¹⁶ Pathologic studies in brains of patients with AD demonstrate that individuals with APOE ϵ4 have increased reactive gliosis³³ and lower levels of endogenous antioxidants³⁴ than those with ϵ2 or ϵ3. In addition, several groups have found that APOE ϵ4 is a risk factor for poor outcome after severe traumatic brain injury^{17 - 18} and for dementia after chronic concussive injury in boxers.³⁵ Furthermore, APOE ϵ4 is associated with increased risk of ischemic cerebral infarction,¹⁹ predicts poor outcome after intracerebral hemorrhage,^{20 - 21} and is associated with faster progression of disability in multiple sclerosis.^{22 - 23}

Others have shown that inheritance of APOE ϵ4 is associated with increased risk of death and poor functional outcome after TBI,¹⁷ a finding that we were unable to reproduce.^{18 ,36 - 37} It is likely, however, that the GOS-E is not sufficiently sensitive to subtle functional deficits to reliably detect APOE ϵ4 effects in all cohorts. Only 1 of the previous studies¹⁷ demonstrated a difference in GOS (the precursor measure to GOS-E) scores between ϵ4 carriers and noncarriers. That study analyzed data from all patients who were admitted with TBI at their center and included subjects with mild injury as well as those who died. In our study, we enrolled only subjects with a certain severity of injury, and individuals with very mild TBI were excluded. Patients who died within 6 months of injury were also excluded. It is likely that these differences in selection criteria explain our failure to detect a difference on GOS-E as a function of APOE ϵ4 status. It is also possible that use of a more sensitive outcome measure in our cohort would allow detection of an effect of APOE ϵ4 in outcome.

Only 1 of the previous studies of APOE ϵ4 in TBI¹⁸ looked at posttraumatic epilepsy as an outcome measure. These investigators did not find increased risk of PTE in APOE ϵ4–positive individuals. This finding was most likely a result of the relatively small size of their study, which included only 69 subjects. Furthermore, it is unclear from the report by Friedman et al¹⁸ what methods were used to identify whether patients had late posttraumatic seizures. We used a structured questionnaire that was developed and validated at the University of Washington and used for more than 10 years in their studies of posttraumatic epilepsy.^{2 ,6} The overall incidence of PTE in our study was similar to that reported by Temkin et al.^{2 ,6} Thus, while, in the absence of electroencephalographic studies and direct observation of the seizures, we cannot be certain that all subjects we scored as having late posttraumatic seizures in fact had epilepsy, we are confident that most of them did.

Since neuronal loss and aberrant synaptic reorganization are characteristic features of most partial epilepsies, several investigators have looked at the association between the APOE ϵ4 polymorphism and medically refractory epilepsy. All have studied patients who underwent temporal lobectomies for treatment of their seizures. Gouras et al³⁹ determined the APOE genotype of 28 patients with temporal lobe epilepsy who underwent temporal lobe resections at their center. They found that APOE ϵ4 was significantly more common in patients who had senile plaques identified histopathologically, but there were no differences between patients with and without senile plaques as far as severity of epilepsy, medication history, history of head trauma, or cognitive deterioration. Studying a similar group of 125 patients, most of whom had cryptogenic mesial temporal lobe epilepsy with Ammon horn sclerosis, Blumcke et al⁴⁰ found that the frequency of APOE ϵ4 was not different from that in the general population. More recently, Briellmann et al⁴¹ studied 43 patients with refractory epilepsy who had undergone temporal lobectomies, and while the overall distribution of APOE ϵ4 alleles was normal, patients with ϵ4 had an earlier onset of habitual epilepsy (5 ± 5 years) than those without ϵ4 (15 ± 10 years). Patients with symptomatic partial epilepsy, whose neural injury resulted from infarcts, traumatic or infectious insults, tumors, or other foreign tissue lesions, were not significantly represented in these 3 series. Since the mechanism of epileptogenesis in symptomatic partial epilepsy may be different from that in cryptogenic cases, failure to find an association between APOE ϵ4 and cryptogenic epilepsy did not discourage us from looking for such an association in PTE.

There was a nonsignificant trend (P = .40) suggesting that APOE ϵ4–positive subjects were at greater risk of developing early posttraumatic seizures. Early posttraumatic seizures are an indicator of severity of injury, as well as an important risk factor for late posttraumatic seizures.^{1 ,42 - 43} Our results are consistent with most studies of posttraumatic epilepsy in finding that incidence of late posttraumatic seizures is related to the severity of injury. Subjects who developed PTE tended to have lower GCS score on admission, to require longer ICU care, to stay longer in the hospital, and to have a lower GOS-E score 6 months later. That these trends did not reach statistical significance is likely a function of the relatively small size of our study and the insensitivity of the GOS-E as an outcome measure, as discussed above. Our findings are best interpreted as an indication that posttraumatic epilepsy is one of several adverse outcomes resulting from TBI that are influenced by inheritance of the APOE ϵ4 allele.

Posttraumatic epilepsy is a common and frequently disabling complication of TBI, for which there is no effective prophylactic therapy. Not all patients who experience severe TBI develop PTE, and inherited genetic factors may influence the likelihood of developing epilepsy after trauma. Our finding that APOE ϵ4 is associated with increased risk of PTE raises the possibility that therapeutic manipulation of lipid or lipoprotein metabolism in the brain may be useful as antiepileptogenic therapy.

Corresponding author and reprints: Ramon Diaz-Arrastia, MD, PhD, Department of Neurology, The University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390 (e-mail: Ramon.Diaz-Arrastia@UTSouthwestern.edu).

Accepted for publication January 21, 2003.

Author contributions: Study concept and design (Dr Diaz-Arrastia); acquisition of data (Drs Diaz-Arrastia, Gong, Garcia, and Carlile and Mss Fair and Scott); analysis and interpretation of data (Drs Diaz-Arrastia, Gong, Agostini, and Van Ness and Ms Scott); drafting of the manuscript (Dr Diaz-Arrastia); critical revision of the manuscript for important intellectual content (Drs Diaz-Arrastia, Gong, Garcia, Carlile, Agostini, and Van Ness and Mss Fair and Scott); statistical expertise (Dr Diaz-Arrastia); obtained funding (Dr Diaz-Arrastia); administrative, technical, and material support (Drs Diaz-Arrastia, Gong, Garcia, Agostini, and Van Ness and Mss Fair and Scott); study supervision (Drs Diaz-Arrastia and Carlile).

Dr Diaz-Arrastia was supported by grants RO1 AG1786RO3, MH64889, and R24 MH59656 from the National Institutes of Health, Bethesda, Md.

We thank Nancy Temkin, PhD, and Sureyya Dikmen, PhD, for the seizure adjudication questionnaire used in this study and for helpful suggestions regarding outcome measures.