Abstract
Objective: To determine whether APOE genotype explained variability in short-term recovery from predominantly mild traumatic brain injury (TBI).
Methods: A total of 87 adult patients presenting with mild or moderate TBI to a shock trauma center were enrolled prospectively. A battery of 13 neuropsychological tests was administered twice, at approximately 3 and 6 weeks after injury. Eighty of 87 patients were successfully genotyped for APOE using a buccal swab technique.
Results: Ninety percent of study patients had mild TBI (Glasgow Coma Scale score of 13 to 15); 18 (22.5%) had one APOE ε4 and none had two ε4 alleles. After adjusting for potential confounders, patients positive for the APOE ε4 allele had lower mean scores on 12 of 13 neuropsychological outcomes at visit 1 compared with APOE ε4-negative patients. Two of the differences were significant (grooved pegboard test, p = 0.005; paced auditory serial addition task 2.8-second trial, p = 0.004). At visit 2, APOE ε4-positive patients had lower adjusted mean scores on 11 of the 13 neuropsychological outcomes. None of the differences was significant.
Conclusions:APOE genotype may influence the severity of the acute injury. However, with no consistent pattern to the recovery curves, it is not clear if APOE genotype influences the rate of recovery.
Understanding variability in the persistence and severity of impairment following traumatic brain injury (TBI) has potentially important treatment implications. The force of impact and more generally the biomechanics of injury1-5⇓⇓⇓⇓ are known to influence outcomes. Other factors include the size, location, and nature of the brain lesion, duration of post-traumatic amnesia, history of brain injury,6 the psychological and emotional state of the individual,7 and selected traits (e.g., sex,8 level of education, and age at injury9). Evidence regarding the extent to which these prognostic factors can be generalized across the spectrum of TBI severity varies considerably. Most research has focused on severely brain-injured individuals10-12⇓⇓ or highly select patients,13,14⇓ even though mild and moderate TBI accounts for nearly 80% of all TBI. Moreover, a number of predictors are not likely to have direct clinical implications or influence the course of recovery, treatment decisions, or development of new treatments.
For a number of reasons, APOE genotype, a gene with three primary alleles (ε2, ε3, and ε4), may predict recovery from TBI. APOE ε4 phenotype is associated with poorer survival and function following hemorrhagic stroke,15,16⇓ with impaired cognitive function in retired professional boxers,17 with a lower probability of recovery from TBI-induced coma,18 and with a longer period of unconsciousness following TBI.19 Taken together, these studies support the notion that a genetic factor explains variability in short- and long-term recovery from CNS insults. To date, however, the influence of APOE on cognitive function has been limited to individuals who have had severe physical or vascular CNS trauma, findings that may not be generalizable to the majority of individuals who experience mild and, in some cases, moderate TBI. We conducted a study of patients with predominantly mild TBI to evaluate generalizability of findings concerning APOE genotype and recovery from severe TBI.
Materials and methods.
To evaluate the role of APOE in recovery from brain injury, we conducted a short-term longitudinal study of cognitive and neurobehavioral function in 87 patients presenting with mild or moderate TBI to a shock trauma center in Baltimore, MD, between June 29, 1996, and December 18, 1997. Eligible patients were selected as a consecutive case series. A battery of neuropsychological tests was administered twice, at approximately 3 and 6 weeks after injury. A buccal cell sample was collected during the first visit for APOE genotyping.
Recruitment of brain-injured patients.
Eligible patients were adults (18 years or older) admitted for a minimum of 24 hours to the trauma center with a mild or moderate TBI, defined by a Glasgow Coma Scale (GCS) score of no lower than 9 within the initial 24 hours following the injury. A CT scan was ordered on all eligible patients. Patients were excluded if there was no known mechanism for and no clear evidence of a TBI or if they had evidence of spinal cord injury, psychiatric disorder, neurologic illness, or a cerebrovascular event or had a positive toxicology screen for alcohol or drug use. Patients were also excluded if they were unable to speak or understand English, were participating in another research study, or resided outside the state of Maryland.
A recruiter monitored admissions, identifying patients who met initial enrollment criteria (i.e., age, GCS score, duration of admission, and absence of substance abuse), followed by a review of the patient’s clinical chart to determine final eligibility (e.g., CT scan results, history of neurologic or psychiatric illness). Recruitment was delayed until 2 weeks after injury to obtain consent of the attending physician and department chair and to ensure that the injured patient could provide informed consent. During an initial phone call to the head-injured patient, the recruiter described the research study, invited participation, and scheduled interested patients for their first study visit. Written informed consent was obtained at the first visit.
A total of 8,138 admission records of patients who presented to the trauma center with an alleged TBI were reviewed for recruitment. Of these, 4,445 were discharged within 24 hours, 850 did not receive a CT scan, and 273 were not diagnosed as brain injured. Of the remaining 2,570 patients with known or suspected TBI, 897 (34.9%) had evidence of substance abuse, 439 (17.1%) had a GCS score of 15 and no radiologic evidence of brain injury, 261 (10.2%) were younger than 18 years, 176 (6.8%) reported a history of psychiatric illness or seizure disorder, 161 (6.3%) reported a history of cerebrovascular accident or spinal cord injury, 140 (5.4%) lived outside the geographic area or were considered inappropriate, 119 (4.6%) presented with severe brain injury, 49 (1.9%) died before discharge from the trauma center, and 13 (0.5%) were transferred to the trauma center from another facility and >3 weeks had elapsed since the date of injury. A total of 315 (12.3%) patients were potentially eligible to participate in this study.
During enrollment, the study was temporarily suspended from July 6, 1996, to September 26, 1996, owing to a failure in the computerized data collection system. Of the 315 eligible admissions, 40 (12.7%) were admitted during this time period but were not invited to participate. Of the remaining 275 eligible admissions, 124 (45.1%) were lost to follow-up or could not schedule an appointment within the 3-week window, and 64 (23.3%) refused to participate. Eighty-seven (31.6%) patients agreed to participate and completed the study. The first examination was completed an average of 24.0 days (median 24 days, range 19 to 40 days) after the date of injury, and the second examination was completed an average of 42.4 days (median 42 days, range 32 to 60 days) after the date of injury.
Data collection.
During the first study visit, a baseline questionnaire was completed along with a questionnaire about postconcussive symptoms and a neuropsychological test battery. During the second visit, the baseline questionnaire was omitted. Questionnaires were interviewer administered.
The baseline questionnaire captured information on sex, race, highest level of education achieved, preinjury employment status, handedness, and history of head and neck injury. Questions about history of head and neck injury covered skull and neck fractures and injury requiring treatment by a doctor.
The symptom questionnaire captured information about emotional and psychological status and the presence of postconcussion symptoms including dizziness, headache, nausea/vomiting, irritability, sleep disturbance, drowsiness, poor concentration, post-traumatic amnesia, depression, anxiety, photophobia, phonophobia, ringing in the ears, blurred vision, and peripheral weakness.
The neuropsychological battery included tests of motor function, reaction time, memory, executive function, and attention/concentration. Tests selected for the battery were sensitive to recovery, challenging (i.e., a timed task, a task that required sustained attention, or a task that required multiple functions simultaneously), and had an adequate measurement scale to quantify change in status. Tests were also selected to have minimal practice effects or, if prone to practice effects, had several different forms to minimize the effect of memory on test performance. The test battery was administered in the following order: grooved pegboard test,20,21⇓ the Stroop test,22 the paced auditory serial addition task (PASAT),23,24⇓ a computerized number vigilance task with 450 stimulus–responses, and a computerized divided attention task, a 15-item verbal memory immediate word recall test with a delayed word recognition task, a visual memory picture presentation task with delayed recall, and a timed numeric memory task (memory scanning). The simple and choice reaction time, divided attention, word recall, word recognition, picture presentation, number vigilance, and memory scanning tests were administered in a computerized test battery provided by Cognitive Drug Research, Ltd.25,26⇓
APOE genotyping.
Buccal (cheek) cells were collected using a buccal swab-and-wash procedure.27 APOE genotype was completed using the method of Hixson and Vernier.28 Reactions were done in a total volume of 50 μL with 2.5 μM primer, 0.2 μM dNTP, 1.5 mM MgCl2, 0.25 μL (1.25 U) of Taq polymerase, and 10% dimethyl sulfoxide. Cycling was at 95°C for 30 seconds, 60°C for 30 seconds, and 71°C for 60 seconds for 40 rounds. The PCR product was digested by Hhal and run on 4% agarose gels for genotype determination. Primers were F4-ACAGAATTC-GCCCCGGCCTGGTAC and F6-TAAGCTTGGCACGGC-TGTCCAAGGA. The buccal swab sample was inadequate to determine APOE genotype on seven patients.
Analysis.
All data management and analysis tasks were performed using SAS statistical software (version 7; Cary, NC). The analysis is restricted to the 80 patients with a known APOE genotype. Mean differences in performance between APOE ε4-positive and ε4-negative patients were calculated for each of the 13 neuropsychological tests included in this analysis. Patients who were unable to complete the PASAT at visit 1 were excluded from the calculation of mean differences. Scores are scaled such that a negative difference is interpreted as poorer performance among the ε4-positive patients.
Multiple regression (SAS PROC REG) was used to estimate the mean difference in neurobehavioral test scores between APOE ε4-positive patients and APOE ε4-negative patients. Differences were adjusted for age (continuous), age squared, race (white vs other), sex, highest education achieved (less than or equal to high school vs more than high school), and history of concussive TBI. Indicators of injury severity were also included in the model, as measured by loss of consciousness (present at injury vs not present), GCS score (9 to 12 vs 13 to 15), and CT scan result (positive vs negative). Backward stepwise regression was used to develop the final linear regression model for each outcome variable (neuropsychology test measure). Variables that had a p value ≥0.20 were removed from the model.
Results.
Demographic and baseline health characteristics.
No significant differences in age and sex were observed between the 315 potentially eligible study patients and the 80 study participants. Education and employment status were not compared because these characteristics were not included in the electronic admission record or clinical chart for the 315 eligible patients. The 80 study patients were predominantly men (60.0%) and white (81.2%) and had at least a high school education (81.2%). Eighteen patients (22.5%) reported graduating from a 4-year college (table 1). Patients were 18 to 81 years of age (mean 41.4 years). Sixty (69.0%) patients worked for pay prior to the injury; 8 had part-time and 52 had full-time employment. Of the 52 who had full-time employment, only 9 (17.3%) had returned to full-time employment by the visit 1 date.
Distribution of demographic and injury characteristics among 80 study subjects, 62 subjects without ε4 allele, and 18 subjects with ε4 allele
Because of the small number of APOE ε4-positive patients, significant differences in demographic characteristics were not found by APOE genotype. However, APOE ε4-positive patients were more likely to be men (72.2 vs 56.5%), less likely to have graduated from high school (11.1 vs 21.0%), more likely to have graduated from college (44.5 vs 16.2%), and less likely to be employed full-time at the time of injury (38.9 vs 71.0%) (see table 1).
Injury characteristics.
Most (90%) of the 80 study patients presented with a mild TBI (i.e., a GCS score of 13 to 15) and a majority (55.0%) with a positive CT scan (see table 1). A total of 64 (80.0%) reported loss of consciousness at the time of injury, but only 12 could provide any information on duration of unconsciousness, and 62.5% reported post-traumatic retrograde amnesia. There were no significant differences in injury characteristics reported by the 315 potentially eligible patients and the 80 study participants.
Significant differences by APOE genotype were not identified, though this may be due to the small sample of APOE ε4-positive patients. However, APOE ε4-negative patients differed slightly from APOE ε4-positive patients on severity of brain injury (41.9 vs 33.3% presented with GCS score of 15), proportion with a positive CT scan (51.6 vs 66.7%), and proportion reporting post-traumatic loss of consciousness (82.3% vs 72.2) (see table 1).
APOE genotype and cognitive function.
At the first visit, APOE ε4-positive patients had lower mean scores on all 13 of the primary neuropsychological measures (see the figure); one difference was significant (PASAT 2.8-second trial total time: 22.0 vs 19.4; p = 0.03). Among the other 12 tests, the greatest differences between APOE ε4-positive and -negative patients were found in a slower mean response time on the word recognition task (1,304.1 vs 1,060.9 ms; p = 0.06) and the number vigilance task (489.5 vs 450.3 ms; p = 0.11). APOE ε4-positive patients also needed more time to complete the Stroop test (296.7 vs 257.2 seconds; p = 0.13). In addition, a higher proportion of the 18 APOE ε4-positive patients (50 vs 37.1%; p = 0.33) were unable to complete the PASAT.
Figure. Mean neuropsychology test scores at visits 1 and 2 for APOE ε4-positive and -negative subjects. Test scores for APOE ε4-positive subjects are represented by the solid lines, and test scores for APOE ε4-negative subjects are represented by the dashed lines. PASAT = paced auditory serial addition task; DH = dominant hand; RT = reaction time.
At visit 2, APOE ε4-positive patients had lower mean scores on all 13 of the primary neuropsychological outcomes. However, the mean difference in score for each test, with the exception of immediate word recall, was equal to or smaller than the differences observed at visit 1. None of the differences was significant.
After adjusting for other factors, APOE ε4-positive patients had lower mean scores on 12 of the 13 neuropsychological outcomes at visit 1 (table 2); 2 were significant (i.e., grooved pegboard, p = 0.005; PASAT, p = 0.004). At visit 2, APOE ε4-positive patients had lower adjusted mean scores on 11 of the 13 neuropsychological outcomes. None of the differences was significant.
Adjusted difference in means (SE) between APOE ε4 genotype groups for each primary outcome neuropsychological examination score at visits 1 and 2
Discussion.
The purpose of this study was to determine if APOE genotype explained variability in neurobehavioral function during the acute phase of recovery from mild to moderate brain injury. In fact, the data suggest that APOE genotype may influence the severity of the acute injury. Individuals who had an APOE ε4 allele had poorer performance on all but one test at visit 1. Test scores for almost all tests among APOE ε4-positive individuals were also lower at visit 2. However, as there is no consistent pattern to the recovery curves, it is not clear if APOE genotype influences the rate of recovery. These results suggest that any impact associated with genotype occurred prior to visit 1, approximately 3 weeks after injury.
It is noteworthy that the differences between those who were positive vs negative for the APOE ε4 allele were consistently smaller at visit 2 than visit 1, a finding that is consistent with delayed recovery among APOE ε4-positive individuals. However, the relatively few significant differences limit the certainty of the conclusions that can be drawn from these data. It is possible that the consistently lower scores at visit 1 and the relative improvement in scores by visit 2 among those who are positive for the APOE ε4 allele could be explained by chance and not delayed recovery. It is also possible that the poorer performance among those positive for the APOE ε4 allele could be explained by difference in premorbid status. We think the latter to be unlikely for two reasons: First, individuals who were positive for the APOE ε4 allele were considerably more likely to have completed at least 4 years of college (i.e., 44.5 vs 16.2%). Higher education is a known predictor of neurobehavioral test scores.8 If APOE genotype is actually associated with speed of recovery from mild TBI, the confounding of APOE genotype by education in our study would tend to mask differences between patients who were positive and negative for the APOE ε4 allele. Second, poorer premorbid test performance would be inconsistent with the relative improvement in test scores that we observed in APOE ε4-positive patients between the first and second visits.
The findings from our study are consistent with pathophysiologic evidence on the acute effects of CNS trauma and the role of APOE. In experimental studies, amyloid precursor protein and APOE up-regulate during the acute phase of injury.29-31⇓⇓ Different APOE phenotypes differentially bind to low-density lipoprotein receptors,32 which may affect lipid transport functions. More specifically, elevated levels of β-amyloid protein have been observed in humans during the acute phase of injury,33 deposits that appear to be more pronounced in those with the APOE ε4 allele.34 There may be secondary inflammatory reactions caused by the deposition of β-amyloid protein.33,34⇓ Increased levels of amyloid pathology associated with the APOE ε4 allele could induce immune-mediated events that slow the course of recovery from CNS insults. Moreover, APOE ε3 appears to increase neurite outgrowth, while APOE ε4 inhibits neurite outgrowth,35,36⇓ suggesting an effect on nerve regeneration. Finally, in vivo studies indicate that APOE ε4 can be neurotoxic37 and exhibits the lowest protective effect against oxidative stress of all three APOE forms,38 suggesting that APOE ε4 may not be as effective at preventing neurologic injury.
The findings from this study are also consistent with several previous studies in humans. Individuals with the APOE ε4 allele were at significantly higher risk for severe cognitive and functional impairment at 6 months after injury,19 risk of persistent coma following TBI,18 and poorer survival and risk of dementia following hemorrhagic stroke.15,16⇓ The consistency of findings with these previous studies of more severe trauma suggests that APOE may influence the course of recovery over a broad range of CNS insults from mild to very severe.
There are several important implications to confirming that APOE ε4 genotype is a significant predictor of slowed recovery from TBI. Individuals who have the APOE ε4 allele may be at greater risk of persistent problems from TBI, raising the need for more specific treatment options. Our understanding of the acute pathophysiology of transient changes in β-amyloid metabolism and related factors is not well established. Nonetheless, if β-amyloid production is a causal agent, new treatments that reduce production in the acute phase of injury could improve outcomes. While this factor may be more important for APOE ε4-positive individuals, β-amyloid may also play a more general role in CNS recovery.
A logical extension of our findings is that repeated subconcussive exposure to humans may be an important risk factor for CNS injury in individuals with the APOE ε4 allele. There is, as yet, no evidence of excess risk. Nonetheless, the possibility raises important questions. If this were to be confirmed, ≥20% of participants in sports like soccer, football, hockey, and boxing may be at risk of persistent CNS problems. Screening for the APOE ε4 could offer valuable information to those at risk. However, knowing that you are positive for the APOE ε4 allele can be traumatic, given that the allele is a strong risk factor for Alzheimer’s dementia and that there are currently no preventive agents for this disease.
Several factors may limit the generalizability of the results from our study. First, study participants represented only 27.6% of all eligible patients admitted to the trauma center with a TBI. No differences were found between participants and all eligible patients on age, sex, or head injury characteristics. Nonetheless, it is possible that participants and nonparticipants differ on other unmeasured factors (e.g., education, history of head injury) associated with recovery from head injury. Generalizability of our results may be limited to the extent that there are meaningful differences on other unmeasured factors. Second, generalizability may be limited because individuals with a positive toxicology screen (n = 897) were excluded from recruitment to ensure a more homogeneous study population. This was deemed essential, given the limited power of the study. Nonetheless, these individuals include casual or social drinkers as well as chronic substance abusers and represent a substantial proportion of all patients with TBI. Future studies should be less restrictive to determine if the results of this study are generalizable to all TBI patients. Finally, the study lacked a control group. Some of the improvement in performance on neurobehavioral tests that we observed between the first and second visits is likely to be due to practice. In this study, the practice effect could have been estimated with a noninjury control group. Without the control group, we are prone to overstating the rate at which patients with mild TBI recover. The absence of a control group, however, is not likely to limit interpretation of differences we observed by APOE genotype.
The results of our study require confirmation in studies adequately powered to detect recovery status from TBI during the acute phase, with a focus on the period immediately following injury. Also, the period of follow-up should be extended to allow for complete recovery.
- Received December 11, 2001.
- Accepted January 2, 2002.
References
Letters: Rapid online correspondence
REQUIREMENTS
You must ensure that your Disclosures have been updated within the previous six months. Please go to our Submission Site to add or update your Disclosure information.
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
