APOE genetic associations with seizure development after severe traumatic brain injury

Abstract

Primary objective: The purpose of the current study is to assess the role of the APOE genotype in post-traumatic seizure (PTS) development.

Research design: A retrospective study of 322 adult Caucasians with a severe TBI and APOE genotype.

Methods and procedures: Medical records were searched for PTS. Time to first seizure was categorized as early, late or delayed-onset PTS. Potential PTS associations by genotype, grouped genotype and allele were investigated.

Main outcome and results: No statistically significant associations were found. However, two out of the four individuals (50%) with the E4/E4 genotype had late/delayed-onset PTS. Furthermore, none with a E2/E2 or E2/E4 genotype seized in the late periods.

Conclusions: The results of this study may suggest 4/4 as a risk genotype for late/delayed onset PTS and a potential neuroprotective role of the E2 allele. However, this study did not definitively support a role for the APOE genotype in PTS susceptibility and indicates that larger populations are needed to fully evaluate the potential impact of APOE on PTS.

Introduction

Traumatic brain injury (TBI) is prevalent in industrialized societies and carries with it a significant amount of associated morbidity and disability. The World Health Organization states that TBI will surpass several diseases to become a major cause of disability and mortality by 2020 [1]. In the US alone, 1.4 million people currently live with TBI [2]. Typically, most TBI occurs with falls, vehicular collisions and violence. Furthermore, TBI is a common injury for those serving in the US military, with the current conflicts; as of 2008, over 25 000 soldiers have sustained blast injury [3].

Post-traumatic seizures (PTS) are a frequent complication from TBI. TBI accounts for 20% of symptomatic epilepsy and 5% of epilepsy overall [4]. Importantly, PTS rates in the military are reported to run as high as 50% [5]. Furthermore, there may be a higher rate of PTS than what is published due to sub-clinical seizures that can occur during intensive care [6]. Both early and late PTS can have a negative impact on outcome after TBI. Individuals who suffer early PTS (within the first week post-injury) are at a higher risk for developing late PTS and those with ongoing PTS must deal with limited mobility in the community and independence [7]. Asikainen et al. [8] found that individuals with late PTS had worse recovery, based on Glasgow Outcome Scale (GOS) scores, than those individuals who had a TBI with similar severity but without PTS. Also, individuals with late PTS (beyond 1 week) are more likely to die from all causes at a younger age [9].

Published recommendations for PTS management support a limited time window after TBI for PTS prophylaxis. Yet, the evidence suggests that the standard prophylactic regime with Dilantin and several other therapies fail to significantly reduce PTS, particularly late PTS [10–12]. However, long-term prophylaxis is still often clinically prescribed, which may adversely impact neurorecovery and cognition as well as result in negative side-effects [7], [13–15].

Based on the current evidence about PTS rates, as well as the efficacy and adverse effects of current treatments, more research is needed to better pinpoint which individuals are most susceptible to PTS and to identify new potential neurobiological mechanisms as treatment targets for PTS. In addition to several risk factors well documented in the literature like injury severity, depressed skull fractures and penetrating trauma [7], [16], [17], genetic variability may be another factor post-TBI that has significant bearing on PTS susceptibility and treatment. Some recent studies have explored genetic links to the occurrence of epilepsy after injury [18], [19], but less is known about how genetic variability influences PTS.

The apolipoprotein E (ApoE = protein, APOE = gene) gene codes for a protein that is mainly responsible for lipid transport has three common alleles (E2, E3 and E4) which correspond with three different isoforms. The E3 allele is the most common and the E2 and E4 alleles are less prevalent across populations. Overall 20–39% of the American population can be expected to carry an E4 allele, while 8–20% is expected to carry the E2 allele [20].

The APOE E4 allele has been repeatedly associated with increased susceptibility to Alzheimer's dementia (AD). As in AD, APOE may play an important role in TBI as well. After closed head injury, ApoE-deficient mice showed markedly impaired motor and cognitive abilities [21] which suggests ApoE's role in diminished neuronal repair. Poor outcome after TBI has also been repeatedly linked to the E4 allele [22]. TBI patients carrying the E4 allele have been found to have longer loss of consciousness periods [23] along with poorer functional outcomes [24], GOS scores [25], [26] and delayed recovery [27]. In one of the largest cohort studies, an interaction was found between the E4 allele and young age, with increased risk for poor outcome [28].

The association between PTS and APOE has also been explored. Specifically, Briellmann et al. [29] found an association between the E4 allele and a shortened latency period between temporal lobectomy and epilepsy. Previous work also linked the E4 allele to late PTS occurrence after TBI [30]. In this latter study of 106 adults with moderate-to-severe TBI, E4 allele carriers were more likely to have PTS develop after the first week post-TBI (late PTS) compared to individuals without the E4 allele. To date this has not been replicated in a small case-control study [31]. However, based on the role of APOE in TBI pathophysiology and outcomes and a previous positive association with E4 and PTS, the purpose of this study was to examine potential APOE associations with PTS in a larger independent cohort of adults with severe TBI.

Methods

Study population

This study was approved by the Institutional Review Board at the University of Pittsburgh. Three hundred and sixty-eight adults with severe TBI were enrolled between 1992–2008 and were genotyped for the APOE gene as a part of a study evaluating APOE genotype and outcome as well has having appropriate medical records. The subjects ranged from 18–75 years old and were enrolled if they had a severe TBI based on a GCS ≤ 8 with positive findings on head CT and required an extraventricular drainage catheter (EVD) for intracranial pressure (ICP) monitoring and management. Patients with penetrating head injury as well as prolonged cardiac or respiratory arrest at injury (greater than 30 minutes) were excluded from the study.

Critical care management

All subjects enrolled were admitted to the neurotrauma intensive care unit and received treatment consistent with The Guidelines for the Management of Severe Head Injury [32]. Patients had a CT scan performed to identify intracranial lesions or haemorrhages. Each patient had an EVD, venous catheter and arterial catheter placed upon arrival. When clinically necessary, surgical intervention for decompression of mass lesions was provided. Elevated ICP was treated in a stepwise fashion to maintain pressure within normal parameters (<20 mmHg) and cerebral perfusion pressure (CPP) was maintained at >60 mmHg. If CPP remained low, then mean arterial pressure (MAP) was supported with pressors or inotropes to maintain a MAP >90 mmHg. Standard EEG was ordered intermittently for patients for whom there was clinical suspicion of PTS activity. In general, patients with severe injuries received PTS prophylaxis for 1 week based on previously published studies [10].

Temperature was monitored regularly and a sub-set of subjects received moderate hypothermia as a part of clinical care or if they were enrolled in a randomized controlled clinical trial evaluating moderate hypothermia after severe TBI. As a part of the study patients were brought to a target temperature of 32.5–33.5°C with 48 hours of cooling with passive rewarming. There were 101 individuals that received hypothermia. Subjects not receiving hypothermia were treated to maintain a normothermic state.

Demographic and injury data

Information regarding the patient's age, gender and race were abstracted from the medical records. A history of pre-morbid seizures, length of hospital stay, cranial surgeries to treat a TBI lesion, mechanism of injury, hypothermia status and GCS scores were also collected. Initial hospital GCS scores were recorded for each patient after resuscitation and without the influence of paralytics. Information about administration of anti-epileptic drugs (AEDs) during their acute hospital stay was also noted. Common AEDs used at the centre for seizure prophylaxis or treatment were included in the analysis. Drugs with anti-convulsant properties that were explicitly designated in the medical record for a condition other than seizure treatment/prophylaxis were excluded from analysis.

Seizure assessment

Information about post-traumatic seizures (PTS) was collected through electronic medical records. Time to first seizure was the primary measure abstracted from the medical record. Inpatient notes used to determine time to first seizure included ambulance emergency room reports, progress notes, nursing notes, EEG reports, patient history and physical reports and discharge or transfer summaries. Subjects were included for analysis if there was at least one clinical note on record related to their injury. Notation in medical records referring to convulsions, seizures, status epilepticus or seizure disorder was taken as PTS occurrence for an individual. Last available records found for the cohort ranged from 6 months post-injury to 18 years post-injury. Time to first seizure was divided into three different time-based categories: early, late and delayed-onset late PTS. In order to meet early PTS criteria subjects must have had a documented seizure within the first week post-injury. Individuals categorized as having late PTS must have had first documented seizure outside of 1 week past injury. Subjects were designated as having delayed-onset of PTS if the time to first documented seizure occurred beyond 6 months post-injury.

Outcome measures

An outcome assessment battery was also collected for each subject at 6 months post-injury that included the Glasgow Outcome Scale (GOS) score. The GOS is commonly used as a global outcome measure in TBI [33]. To appropriately account for mortality in PTS analysis, death date was recorded and coded in the same way as PTS. Death was recorded as occurring (1) within the first week post-TBI, (2) beyond the first week after TBI or (3) beyond the first 6 months after TBI. Mortality was collected using death reports in medical records as well as the social security death index online (http://ssdi.rootsweb.ancestry.com/).

Genotyping

DNA was extracted from one of two sources for each subject, whole blood collected prior to transfusion or cerebrospinal fluid (CSF). Whole blood was collected into EDTA vacutainer tubes, processed to retrieve the buffy coat and DNA extracted using a simple salting out procedure [34]. CSF was collected by passive drainage as standard of care and DNA was extracted using the Qiamp DNA extraction protocol for extraction from body fluids (Qiagen Corporation, Valencia, CA).

Genotypes were obtained using polymerase chain reaction and restriction fragment length polymorphism techniques as previously described [35], [36]. Genotypes were assigned by two independent, blinded technicians and any discrepancies between the two assigned APOE genotypes were reconciled by additional genotyping.

Statistical analysis

To avoid confounding interactions, 46 individuals were excluded from the final statistical analysis because of race and/or history of pre-morbid seizures. Thirty-two individuals did not self-report European ancestry, which did not provide a large enough cohort to analyse separately. Nineteen individuals had pre-morbid seizures and were therefore removed from the population. The final cohort included for analysis was n = 322 subjects.

Summary statistics including means, standard error of the mean (SEM), medians and frequencies were computed for final population. Independent t-tests were used to assess differences among PTS groups for continuous variables. Chi-square analysis, using the Fisher's exact test when appropriate, was used to explore significant associations between APOE genotype and descriptive variables like GOS and mortality. Genotype associations with PTS were evaluated such that subjects within each ‘time to first PTS category’ were compared to those without PTS at any time point. Chi-square, with Fisher's exact test as appropriate, was also performed to compare individual APOE genotype to PTS. PTS occurrence was also analysed in relation to the presence vs absence of the E2, E3 or E4 allele.

To avoid the confounding factor of death on seizure incidence, subjects were removed from analysis who died within a week of injury and did not seize when comparing APOE genotypes and alleles to early PTS. All documented deceased subjects, except three subjects with late PTS, were removed to assess APOE associations with late and delayed-onset late PTS.

Population sizes were estimated to find appropriate sample size required to definitively analyse less frequent genotypes. Estimated sample sizes were created using Fisher's exact method and assuming a two-sided hypothesis with a confidence level (α) of 0.05 and a power greater than 80%.

Results

Population description

Three hundred and twenty-two adult subjects with severe TBI were included in this analysis and had both medical record information and APOE genotype. The APOE genotype distribution is depicted in Figure 1 Of the entire population, 79 (24.5%) subjects carried at least one E4 allele, while five (1.6%) subjects possessed the E4/E4 genotype (However, one E4/E4 individual died a week post-injury and was removed from analysis). The median GCS score for the population was 6. Seventy-seven per cent of the sample was male, while 23% were female. The most prevalent mechanism of injury was automobile collisions (44.7%) followed by injury from a fall or jump (14.6%) and motorcycle accidents (10.6%). Approximately 96% of the subjects in this cohort received AEDs during their acute care hospital stay and 18.6% of the population had documented evidence of PTS. Of those who seized, 16 patients seized within 1 week of injury, while 44 individuals seized beyond 1 week from injury. At the end of 2009, the cumulative mortality rate (based on the SSI) for the entire population was 30.7% (n = 99 subjects). Of this population, 44 subjects died within the first week of injury.

Figure 1.

APOE genotype frequency for the entire study population.

A comparison of demographic- and injury-related data for each PTS cohort (early, late and delayed onset) is provided in Table I. For most variables, there were no significant differences between those with no PTS, early PTS, late PTS and delayed onset PTS. For the early PTS cohort, individuals receiving hypothermia were significantly less likely to seize than those who were not (p = 0.041) (see Table I). Additionally, 140 individuals in the overall population had at least one cranial operation. Cranial surgery was significantly associated with increased seizure occurrence in both the late PTS (p = 0.0001) and the delayed-onset cohort (p = 0.001) (see Table I). As both hypothermic status and cranial surgery were significantly associated with PTS, further analyses were run to assess possible relationships between these variables and APOE genotype. There were no significant associations in any cohort with genotype (or presence or absence of the E2, E3 or E4 allele) when comparing hypothermic status or cranial surgery at all time points.

Table I. .

Descriptive comparisons of PTS cohorts

Variable	No PTS vs. Early PTSa		p-valuec	No PTS vs. Late PTSa		p-valuec	No PTS vs. Delayed-onset PTSb		p-valuec
Gender			1.00			0.841			1.00
Female	49 (94.2%)	3 (5.8%)		39 (78%)	11 (22%)		39 (88.6%)	5 (11.4%)
Male	170 (92.9%)	13 (7.1%)		133 (80.1%)	33 (19.9%)		133 (87.5%)	19 (12.5%)
AED treatment			1.00			0.364			0.599
Yes	10 (100.0%)	0 (0.0%)		8 (100.0%)	0 (0.0%)		8 (100.0%)	0 (0.0%)
No	209 (92.9%)	16 (7.1%)		164 (78.8%)	44 (21.2%)		164 (87.2%)	24 (12.8%)
GCS score			0.767			0.529			1.00
3–4	50 (92.5%)	4 (7.4%)		32 (76.2%)	10 (23.8%)		32 (88.9%)	4 (11.1%)
5–8	168 (93.3%)	12 (6.7%)		139 (80.3%)	34 (19.7%)		139 (87.4%)	20 (12.6%)
Hypothermia status			0.041*			0.363			0.813
Cooled	72 (98.6%)	1 (1.4%)		56 (76.7%)	17 (23.3%)		56 (87.5%)	8 (12.5%)
Normothermic	142 (91%)	14 (9%)		112 (82.3%)	24 (17.6%)		112 (88.8%)	14 (11.2%)
Depressed skull fracture			0.607			1.00			0.646
Yes	15 (100.0%)	0 (0.0%)		10 (83.3%)	2 (16.7%)		10 (83.3%)	2 (16.7%)
No	203 (92.7%)	16 (7.3%)		161 (79.3%)	42 (20.7%)		161 (88.0%)	22 (22.0%)
Injury mechanism			0.521			0.121			0.668
Motor vehicle	109 (94%)	7 (6%)		88 (84.6%)	16 (15.4%)		88 (88.9%)	11 (11.1%)
Fall	25 (96.2%)	1 (3.8%)		16 (72.7%)	6 (27.3%)		16 (80.0%)	4 (20.0%)
Motorcycle	21 (91.3%)	2 (8.75%)		17 (73.9%)	6 (32.4%)		17 (85.0%)	3 (15.0%)
Other	27 (100.0%)	0 (0.0%)		23 (67.6%)	11 (32.4%)		23 (88.5%)	3 (11.5%)
Cranial surgery			0.118			0.000*			0.001*
Yes	79 (95.2%)	9 (4.8%)		56 (65.1%)	30 (34.9%)		56 (76.7%)	17 (23.3%)
No	139 (89.8%)	7 (10.2%)		115 (79.3%)	30 (20.7%)		115 (94.3%)	7 (5.7%)
Age	33.2 ± 0.98	35.3 ± 3.5	0.589	31.1 ± 1.0	31.6 ± 1.7	0.789	31.1 ± 1.0	32.6 ± 2.4	0.588
Acute care LOS	24.2 ± 0.977	29.1 ± 7.1	0.510	24.6 ± 1.1	26.8 ± 2.0	0.351	24.6 ± 1.1	23.7 ± 1.9	0.766

^aAnalysis does not include subjects who died in during the first week after TBI and did not have PTS.

^bAnalysis does not include subjects who died in acute care or had documented death after 1 week post-injury and did not have PTS.

^cp-values based on chi-square analysis and independent t-test as appropriate.

*notes statistically significant values (p<0.05).

Early PTS

The early PTS cohort included a total of 235 subjects. Chi-square analysis showed no significant association between APOE genotype and early PTS (p = 0.479) (see Figure 2(a)). Table II shows that individuals carrying either the E2 allele or the E3 allele did not have significantly different early PTS rates compared to those without the respective allele. Notably, there also was no significant association between individuals carrying the E4 allele and early PTS (p = 0.377) (see Figure 3(a)).

Figure 2.

APOE genotype by (a) early (within 1 week of injury) PTS frequency, (b) late (beyond a week from injury) PTS frequency and (c) delayed onset (beyond 6 months) PTS frequency.

Table II. .

Chi-square analysis p-values for APOE genotype associations with PTS cohorts

	Early PTS:	Late PTS: Delayed onset PTS:
	within 1 week	Beyond first week	Beyond 6 months
Genotype	0.479	0.133	0.153
2 Allele selected	1.00	0.126	0.385
3 Allele selected	0.436	0.667	0.254
4 Allele selected	0.377	0.43	0.803

Figure 3.

APOE genotype selected for the E4 allele carriers by (a) early (within 1 week of injury) PTS frequency, (b) late (beyond a week from injury) PTS frequency and (c) delayed onset (beyond 6 months) PTS frequency.

Late PTS

Data for 216 individuals were analysed for APOE genetic associations with late PTS. There was no genotype difference in the proportion of subjects who experienced late PTS (see Figure 2(b)). None of the subjects with an E2/E2 and E2/E4 genotype developed late PTS and Table II shows that neither the individuals with the E2 allele (p = 0.126) nor the E3 allele (p = 0.667) were significantly more or less likely to have late PTS. Individuals carrying the E4 allele had similar late PTS rates than those without the E4 allele (see Figure 3(b)) (p = 0.430). However, 50% of subjects with the E4/E4 genotype had late PTS.

Delayed-onset late PTS

A total of 196 individuals were included for delayed-onset PTS analysis. Similar to that reported for late PTS, there was no significant difference found between APOE genotype and delayed-onset late PTS (see Figure 2(c)). Similar to that observed for late PTS, subjects with the E2/E2 and E2/E4 genotypes had no delayed onset PTS and there was no significant association found between individuals selected for the E2 or E3 allele and delayed onset PTS. Although the E4/E4 group had a 50% PTS rate, there were no differences in delayed onset PTS rates for individuals either with or without the E4 allele (p = 0.803) (see Figure 3(c)).

Sample size calculations for specific genotype analyses

Although not significant in genotype analysis, the population with E4/E4, E2/E2 and E2/E4 genotypes showed qualitative trends for late and delayed onset PTS rates. Two of the four subjects with the E4/E4 genotype (50%) had late seizures and delayed onset PTS. Sample size calculations showed that over 2000 individuals would be needed to have a cohort large enough to definitively evaluate the E4/E4 genotype in relation to late/delayed onset PTS. Additionally, none of the individuals with an E2/E2 genotype or an E2/E4 genotype seized in the late or delayed onset cohort. Further sample size calculations showed that over 1300 subjects would be needed to definitively evaluate the E2/E2 and E2/E4 genotype associations with late/delayed onset PTS.

Discussion

Overall this study does not show a significant association between APOE genotype or alleles and PTS after TBI. However, it is important to note that 50% (two out of four) of the individuals seized within the late/delayed-onset time point. Additionally, none of the individuals with the E2/E2 or E2/E4 genotype seized in the late periods.

ApoE is an essential protein in several CNS cellular mechanisms. It is also important in development (neurite growth) and CNS pathology. Several studies have linked ApoE and the development of AD. Specifically, the E4 allele is associated with an increased risk of developing AD, which features amyloid-β plaque formation, also found in some cases after human TBI [37]. Measurements of ApoE and amyloid-β in CSF following TBI are significantly lower than controls [38]. This is consistent with the idea that ApoE is taken into neurons after injury to influence repair and survival [38], [39] as well as clearing lipids created by injury [40]. Furthermore, reduced CSF amyloid-β levels also have been hypothesized to be reduced due to an increased deposition within the formulating amyloid plaques [38].

The propensity for amyloid-β plaque formation after TBI may be ApoE isoform-specific. Animal models demonstrate that ApoE isoforms may have specific effects on amyloid-β accumulation and clearance relative to the ApoE concentration present [41]. Specifically, past studies show that individuals with the E4 allele are more susceptible to amyloid-β plaque development following TBI [42]. One mechanistic hypothesis for this association includes the finding that E4 binds to amyloid β and facilitates its passage across the blood–brain barrier, while the E2 and E3 isoforms prevent blood–brain barrier transport into the brain [43]. Reciprocally, E4 is less efficient then the E2 and E3 alleles at binding and clearance of the brain [44].

Amyloid β plaques contribute to the development of temporal lobe epilepsy (TLE) [45], [46], potentially through disruption of normal neuron signalling. Given the development of amyloid β plaques in TBI and plaque associations with TLE, it is reasonable to hypothesize that amyloid β could contribute to PTS and that APOE genotype also contribute indirectly to this process. However, TLE studies have been unable to demonstrate a significant association between APOE genotype and earlier onset of TLE after focal lesions [47] or show an overall increased risk of TLE [48].

The overall PTS rate was 18.6% and is within the range of what other reports have offered for PTS incidence [49]. The results show that there was no significant link between individual APOE genotype and PTS at any time point. There were no associations between presence of the E4 allele and early PTS rates. Unlike previous work [30], there also were no associations between presence or absence of the E4 allele and late PTS. Moreover, the findings show no relation between presence of the E4 allele and delayed PTS. The study population is approximately three times larger than that reported with earlier work showing a positive relationship between carriage of the E4 allele and late PTS [30]. However, beyond the difference in sample size, there are comparative differences in cohort characteristics between studies that may have influenced the findings. For example, the cohort was comprised only of subjects with severe TBI, does not include subjects with penetrating TBI and used medical records instead of patient interview/report to generate phenotypes. Further, it was chosen to limit the analyses to Caucasians in an attempt to control for population stratification given the large differences in APOE allele frequencies from different ancestries [20].

However, analysis with this small group of subjects with the E4/E4 genotype showed a 50% PTS rate (n = 4 subjects) for late and delayed onset PTS. Although small numbers, these findings suggest that there may be a link between the APOE genotype and PTS development. The link between the E4 allele, post-traumatic amyloid β plaque development and epilepsy may be one viable explanation about why the E4/E4 individuals in this study had a 50% seizure rate for late and delayed onset PTS. Although this cohort of severely injured adults with TBI is larger than other published work on APOE genotype associations and PTS [30], [31], one limitation for this study still is the relatively small sample size, particularly for evaluating those with the E4/E4 genotype. Sample size calculations show that a definitive study would need at least 2000 subjects in order to have enough power to fully assess PTS associations with E4/4E individuals with severe TBI.

The APOE E2 allele may also have isoformic effects on amyloid β plaque formation after TBI. In the past, studies of AD have found evidence that the E2 allele is less frequent in AD subjects than what would have been predicted from normal population frequency studies [50], [51]. However, little research has been done in the protective effects of the E2 allele in TBI as it is a less frequent allele. Although not found to be significant, the data may add to the understanding of the protective effect. Of the four individuals carrying either the E2/E2 or E2/E4 genotype none of them had a late or delayed onset seizure. This 0% seizure rate may be indicative of a protective role of the E2 allele. Any potential protective role of the E2 allele, however, would again need to be evaluated in larger populations. In fact, the sample size calculations estimate that a cohort size of >1300 individuals is needed to definitively evaluate a potential protective role of the E2 allele and PTS in TBI patients. The role of APOE in TBI, AD, epilepsy pathology, combined with the data and data from other studies [30], may provide reasonable evidence to support further evaluating the presence of the E2 allele and the E4/E4 genotype in the context of PTS development in a large multi-centre study.

Out of the clinical variables which were collected, only two were significantly related to PTS: hypothermia and cranial surgery. Individuals who were randomized to hypothermia treatment were significantly less likely to seize. This finding supports previous hypothermia work in animal models [52], [53] in which hypothermia significantly reduced seizure risk. This is also supported through a trend seen in previous studies between adenosine receptor genes and PTS [54]. The neuroprotective role of hypothermia in early PTS may be due to decreased levels of glutamate after cooling [55]. In addition, late PTS and delayed onset PTS were significantly associated with cranial surgery. Recent findings show glial scars associated with cranial surgery increase the rate of PTS after TBI [56]. Both the hypothermia and cranial surgery variables were compared to genotype or presence or absence of the E2, E3 or E4 allele to determine if APOE influenced these associations and there were no significant genotype associations with these variables.

Although this study utilized a large cohort and rigorous approach to chart extraction and analyses, potential limitations for this negative finding study still include the fact that PTS status relied largely on documentation of clinical evidence of PTS within patient medical records. Routine diagnostic or continuous EEG has not been a part of the standard of care and, as such, the PTS rates may be lower than what actually occurred [6]. However, other study design formats like prospective patient interview may lead to significant patient recall bias about PTS in a cognitively impaired population. With the retrospective approach to phenotyping, time to first seizure was the primary measure of interest. As such, one cannot conclude anything about recurrent seizure or the development of post-traumatic epilepsy (PTE). Prospective studies that employ continuous EEG monitoring in the acute care setting and routine diagnostic EEG follow-up assessments in the outpatient setting not only may be required for accurate PTS rates, but also to assess how E4 allele/genotype influences the development of PTE as well as to characterize the evolution of how specific electrographic abnormalities may portend the eventual development of PTS or PTE.

Given the potential role of amyloid β plaques in TBI pathology as well as in the development of PTS, investigating how genetic variation for other key factors in AD like pathology after TBI may be reasonable. For example, Neprilysin has recently been implicated as modulating amyloid β plaque formation after TBI [57]. Further, recent studies implicate variation in the promoter of the Neprilysin gene as playing an important role in amyloid β plaque formation in humans after TBI [37]. Assessing PTS rates in the context of variability for genes coding for proteins associated with familial AD, like amyloid precursor protein (APP) and presenilin, as well proteins associated with sporadic late onset dementia [58] also may be warranted.

In summary, there are potential neurobiological mechanisms for the APOE genotype to contribute to the development of PTS. The current study shows no significant associations linking the presence of the E4 allele with higher early or late/delayed-onset PTS rates. However, an association between subjects with the E4/E4 genotype and late and delayed onset PTS was noted in the small sub-group of adults with severe TBI. Subjects with the E2/E2 or E2/E4 genotype may also have some degree of protection against the development of late/delayed onset PTS. These results highlight the need for replication in a larger multi-centre cohort.

Acknowledgements

We would like to thank the Conley Genetics Lab staff, especially Sandra Delouches, for their work in genotyping our patients. We would also like to thank the Brain Trauma Research Center at the University of Pittsburgh Medical Center for their support in recruitment and collecting patient outcome measures.

Declaration of Interest: The authors would like to acknowledge the NIH grant number NIHR01HD048162, NIHP50NS030318, and NIHR01NR008424 as well as DOD grant number W81XWH-07-1-0701.