Pre-existing Health Conditions and Repeat Traumatic Brain Injury

Article Outline

• Abstract
• Methods
  • Study Population
  • Definitions
  • Analysis
• Results
• Discussion
  • Study Limitations
• Conclusions
• References
• Copyright

Abstract 

Saunders LL, Selassie AW, Hill EG, Horner MD, Nicholas JS, Lackland DT, Corrigan JD. Pre-existing health conditions and repeat traumatic brain injury.

Objective

To assess and compare the effect of Pre-existing epilepsy/seizure disorder and drug/alcohol problem on the hazard of repeat traumatic brain injury (TBI) in persons with TBI who participated in a follow-up study.

Design

Retrospective cohort.

Setting

Acute care hospitals in South Carolina.

Participants

Participants were from the South Carolina Traumatic Brain Injury Follow-up Registry cohort of persons (N=2118) who were discharged from an acute care hospital in South Carolina and who participated in a year-1 follow-up interview.

Interventions

Not applicable.

Main Outcome Measures

Repeat TBI was defined by 2 isolated events of TBI in the same person at least 72 hours apart and recorded in hospital discharge or emergency department records from 1999 through 2005.

Results

A Cox proportional hazards model was used to assess the associations of Pre-existing epilepsy/seizure disorder and drug/alcohol problem with time to repeat TBI, controlling for other confounding factors. There were 2099 persons with information on both Pre-existing conditions. There were 147 (7%) persons who sustained repeat TBI after recruitment to the follow-up study, and 82 (3.9%) had a previous TBI before recruitment for which they were seen in the hospital discharge or emergency department since 1996. The hazard of repeat TBI for persons with Pre-existing epilepsy/seizure disorder was 2.3 times the hazard for those without (hazard ratio, 2.3; 95% confidence interval, 1.2–4.4; P=.011). Pre-existing drug/alcohol problem was not associated with repeat TBI. Other variables significantly associated with repeat TBI were having a prior TBI, being insured under Medicaid, and having no insurance.

Conclusions

Pre-existing epilepsy/seizure disorder predisposes to repeat TBI. Appropriate management of seizure control may be an important strategy to allay the occurrence of repeat TBI.

Key Words: Brain injuries, Epilepsy, Rehabilitation, Seizures

List of Abbreviations: AIS, Abbreviated Injury Scale, CI, confidence interval, CPH, Cox proportional hazard, ICD-9-CM, International Classification of Diseases, Ninth Revision, Clinical Modification, SCTBIFR, South Carolina Traumatic Brain Injury Follow-up Registry, TBI, traumatic brain injury

TRAUMATIC BRAIN INJURY is a significant public health problem in the United States, with 1.4 million Americans sustaining a TBI each year,¹ and 5.3 million living with a TBI-related disability.² More recently, Selassie et al³ estimated that 43.3% of TBI hospitalizations would result in long-term disability; however, their definition of disability included not only physical limitation but also cognitive complaints, postinjury symptoms, and mental health problems that are limiting to perform an activity. Persons who have sustained one TBI have been shown to be at increased risk to sustain another.⁴, ⁵ Research on repetitive TBI has been focused primarily on athletes, with very little information available on repeat TBI in the general population. However, evidence suggests a cumulative effect of TBIs on health outcome. Studies have found associations between the number of concussions and mild cognitive impairment⁶ and more permanent problems in memory and thinking among retired professional football players.⁷ Also, in men, a relationship has been found between the number of TBIs with unconsciousness, and symptoms such as dizziness, concentration difficulty, depression, visual problems, and overall health problems.⁸ Although these findings are informative of some athletes, they do not account for injuries occurring off the field, do not require the diagnosis of a physician for the classification of TBI, and do not include women. Further, the risk of repeat TBI as a function of Pre-existing conditions, such as epilepsy/seizure disorder or alcohol/drug problems, has not been explored.

Preinjury treatment for drug or alcohol problems has been associated with heavy drinking post-TBI.⁹ Limited previous research has been conducted evaluating the association between alcohol/drug abuse and repeat TBI. Walker et al¹⁰ reported that prisoners with 2 or more TBIs had more alcohol problems than prisoners with 1 or no TBIs. However, this was a cross-sectional study and therefore could not establish temporal sequence. Drubach et al¹¹ found that of persons admitted to a brain injury rehabilitation unit, those with prior drug and alcohol abuse were the most likely to have had a prior TBI. Lastly, in a review by Salcido and Costich,¹² alcohol/drug use was found to be associated with repeat TBI. These studies were conducted in very specific populations and therefore have limited generalizability.

Epilepsy/seizure disorder is another condition which might contribute to repeat TBI. Not only can seizures occur as a result of a TBI, but seizures can also contribute to TBIs.¹³ Wirrell¹⁴ found that persons with epilepsy had more injuries than the general population, and that a quarter of those injuries were seizure related, with head injuries being one of the more common injuries. The extent to which Pre-existing epilepsy contributes to the recurrence of TBI has not been quantified and described in a population-based setting.

The aim of this article is to assess and compare the effect of Pre-existing health conditions, specifically epilepsy/seizure disorder and drug/alcohol problems, on the risk of repeat TBI in a cohort of 2118 persons with TBI who participated in a follow-up study. We hypothesized that persons with Pre-existing epilepsy/seizure disorder and/or Pre-existing alcohol/drug treatment would have a larger hazard of repeat TBI than those without these Pre-existing conditions.

Back to Article Outline

Methods 

Study Population 

This study followed a retrospective cohort design in which a cohort of TBI hospital discharges was assessed for prior epilepsy/seizure disorder and alcohol/drug problems and then followed up for repeat TBI. Data from the SCTBIFR were used.¹⁵ In South Carolina, all nonfederal hospitals are required to provide Uniform Billing-92 abstracted data for all hospital discharges to the Office of Research and Statistics. Persons admitted to the hospital between January 1, 1999, and June 30, 2002, with a primary or secondary diagnosis of TBI made up the sampling frame. SCTBIFR excluded persons younger than 15 years at the time of injury, out-of-state residents, and persons who died in the hospital. Of 7613 persons in the sampling frame, 4519 persons were sampled using a 2-stage random sampling method.¹⁵ Medical records were abstracted on the sample to obtain more complete information about the circumstances of their injury.

Of 4519 persons sampled, 773 (17.1%) were found ineligible for recruitment for the following reasons: 382 (8.4%) had died after discharge from acute care, 80 (1.8%) were medically unable to participate and had no proxy, 92 (2.0%) were prisoners, 195 (4.3%) had moved out of state, and 24 (0.5%) were unable to speak English. Of those eligible, 2118 completed an in-depth telephone survey conducted by trained interviewers 1 year after their TBI. Lead questions related to events, dates, places, and how consistent these were to the records on file were used to assess the participants' cognitive competencies. Proxy respondents were used (n=285) when a person was deemed cognitively impaired or not able to follow commands or not able to respond to the telephone survey. Proxies were not asked all questions, especially pertaining to quality of life, but questions about alcohol and epilepsy/seizure were asked of proxies. The interview included questions on Pre-existing health conditions, including epilepsy/seizure disorder and drug/alcohol treatment. The survey instrument can be found at http://www.musc.edu/sctbifr. Eligible persons who did not complete the year-1 interview were more likely to be nonwhite males, have less severe TBI, and be without health insurance (χ² test of homogeneity, P<.001).¹⁵ However, persons who were interviewed were not significantly different from persons not interviewed with respect to their repeat TBI status (P=.932).

Definitions 

TBI was defined using the ICD-9-CM nature of injury codes (800.0–801.9, 803.0–804.9, 850.0–854.1, and 959.01) obtained through hospital discharge and emergency department records.¹⁶ Discharge diagnoses are 100% complete and have high accuracy. Both hospital discharge and emergency department records contain personal identifiers, patient demographics, admission and discharge dates, acute care charge, admission source, discharge disposition, external cause of injury codes, procedure codes, and 1 primary and up to 9 secondary diagnosis codes based on ICD-9-CM. The TBI that resulted in the selection of the person to the SCTBIFR study was defined as the index TBI.

The main outcome, repeat TBI, was defined by 2 isolated events of TBI in the same person at least 72 hours apart and recorded in hospital discharge or emergency department records from 1999 through 2005. Although 9.5% of the cohort had more than one repeated TBI, the response was made binary by lumping all repeats into one category. A person who was transferred to another hospital or readmitted for the same event because of complications, late effects, or worsening of symptoms was not identified as having repeat TBI. Hospital and emergency department records for persons included in the sample were linked across time using personal identifiers, enabling the identification of repeat TBI. Unique identifiers were available back through 1996, and therefore prior TBI for this period was assessed.

Epilepsy/seizure disorder was identified at the year-1 follow-up after the index TBI by the question, “Before your injury, did a doctor ever tell you that you had a seizure disorder or epilepsy?” In this analysis, the response reflected epilepsy/seizure disorder occurring before the TBI under reference. In addition, we identified persons with epilepsy or seizure diagnosis codes, ICD-9-CM 345 and 780.3, in their index TBI medical record. A 3-level categorical variable was created using these 2 sources: (1) persons with both self-reported epilepsy/seizure disorder and an epilepsy/seizure disorder diagnosis code; (2) persons with either, but not both; and (3) persons with neither. The purpose of this definition was to capture persons who have ever had epilepsy/seizure disorder (whether or not they have been told by a doctor they have epilepsy/seizure disorder) and to also identify persons who are having active seizures (through medical records).

Previous treatment for drug/alcohol problems was also identified at the year-1 interview (1 year after the index TBI) by the question, “Before your injury, had you ever been treated for a drug or alcohol problem?” In addition, drug and alcohol problems were assessed through medical record review. Trained abstractors extracted information on alcohol and drug use from the medical record using notes about drug or alcohol problems as well as previous referrals to drug or alcohol treatment.

Other explanatory variables were identified from hospital discharge records as noted at discharge from the index TBI. Because only 2.4% of persons in the sample were not white or black, race was categorized as white and nonwhite. Cause of injury was categorized as motor vehicle collision, fall, sports injury, violence injury, and all other using ICD-9-CM external cause of injury codes. Insurance was categorized as commercial, Medicaid, Medicare, and uninsured. Concomitant injury, other injury (eg, femoral fracture) co-occurring with the TBI at the time of the index injury, yes or no, was defined by counting the number of injuries other than TBI using the ICD-9-CM nature of injury codes (800–959). Comorbid conditions at the time of the index injury, yes or no, were defined using a modified Elixhauser's comorbidity index (codes 345 [.0–.9] and 780.39 were excluded because epilepsy/seizure disorder was a separate variable).¹⁷ While we used codes for epilepsy (345.0–345.9) and also for seizure (780.39), previous research has found that 73.1% of persons with seizure codes actually have a history of epilepsy.¹⁸ Lastly, county of residence was grouped according to Metropolitan Statistical Area designation into urban and rural.¹⁹

Severity of the index TBI was defined as mild, moderate, and severe using the AIS.²⁰ ICD-9-CM codes were translated into AIS using ICDMAP-90 software.^a Persons were assigned an AIS score for each of 9 body regions, and for this study, we used AIS of the head to define TBI severity. AIS of 2 represents a mild TBI, an AIS of 3 a moderate TBI, and an AIS of 4 to 6 a severe TBI. A mild TBI can be characterized by a period of disorientation or loss of consciousness lasting less than 30 minutes.²¹ The remaining variables considered for analysis were categorized as shown in table 1.

Table 1. Characteristics of Persons With and Without Repeat TBI

Characteristics	Repeat TBI		Total (n=2099)	P
	Yes (n=147)	No (n=1952)
Age (y)	45.8±24.3	44.8±22.2	44.9±22.4	.12
Sex
Male	59.9	60.7	60.6	.91
Female	40.1	39.3	39.4
Race
Nonwhite	20.4	23.9	23.6	.25
White	79.6	76.1	76.4
Pre-existing ESD
Self-report and in MR	7.5	2.9	3.2	<.001
Either positive	14.3	6.8	7.3
Neither positive	78.2	90.3	89.4
Pre-existing drug/alcohol
Self-report and in MR	11.0	4.9	5.3	.001
Either positive	27.6	22.8	23.2
Neither positive	61.4	72.3	71.6
Cause
Fall	37.4	29.4	29.9	.062
MVC	45.6	54.6	53.9
Violence	10.9	7.8	8.1
Other	6.1	8.3	8.1
TBI severity
Severe	55.8	45.7	46.4	.054
Moderate	19.7	22.6	22.4
Mild	24.5	31.7	31.2
Insurance
Medicaid	14.3	7.9	8.4	<.001
Medicare	25.2	25.7	25.7
Uninsured	28.6	17.7	18.5
Commercial	32.0	48.6	47.5
Residence
Rural	24.5	25.8	25.7	.63
Urban	75.5	74.2	74.3
Education
<12th grade	37.4	33.5	33.8	.072
12th grade or GED	38.1	32.8	33.2
College +	24.5	33.7	33.0
Concomitant injury
3+	12.9	17.8	17.5	.024
1–2	50.3	55.4	55.0
None	36.7	26.8	27.5
Comorbid conditions
Yes	48.3	45.1	45.4	.46
No	51.7	54.9	54.7
Prior history of TBI
Yes	8.8	3.5	3.9	.001
No	91.2	96.5	96.1

NOTE. Values are mean ± SD or percent.

Abbreviations: ESD, epilepsy/seizure disorder; GED, general educational development; MR, medical record; MVC, motor vehicle collisions.

Analysis 

All analyses were conducted by means of the SAS version 9.13 procedure applications.^b Our outcome was a repeat TBI event. We used survival analysis techniques to assess time to the outcome. Although some persons experienced more than one repeat TBI, there were not enough people to implement recurrent survival analysis. Kaplan-Meier curves were generated for the main exposure variable to assess for an association with time to repeat TBI. Variables with a chi-square significance of P less than .25 were considered for inclusion in the multivariable model (t test was used for age). Interactions of suspected variables were ruled out with the Breslow-Day test of homogeneity of the risk ratios. A CPH model was used to assess the associations between Pre-existing conditions and time to repeat TBI, controlling for other confounding factors. The proportional hazards assumption was tested by graphing the survival curves for each independent variable and looking for any intersection of the lines. Also, for each independent variable, a proportional hazards model was fit including an interaction term between the independent variable and time. If the interaction term was significant, this indicated a departure from the proportional hazard assumption. Lastly, we also assessed the proportional hazard assumption using a plot of the cumulative sums of martingale residuals.²²

Back to Article Outline

Results 

A total of 2118 persons completed the year-1 interview for SCTBIFR. There were 2099 persons for whom we had information on both Pre-existing epilepsy/seizure disorder and a Pre-existing alcohol/drug problem. The 19 persons with missing information were deleted.

Table 1 summarizes the characteristics by repetitiveness of TBI. There were 147 (7%) persons who sustained repeat TBI after the index TBI, and 82 (3.9%) had a previous TBI for which they were seen in the hospital discharge or emergency department between 1996 and the index visit. At the time of the index TBI, a significantly higher proportion of persons with repeat TBI had a Pre-existing epilepsy/seizure disorder or a drug/alcohol problem, had Medicaid or no health insurance, had fewer body region injuries, and had a prior history of TBI (see table 1). Of those with repeat TBI, 133 (90.5%) had 1 repeat event, 10 (6.8%) had 2, and 3 (2.7%) had 3 repeat events during the follow-up period. Also, of those who sustained repeat TBI, 11.6% occurred within 1 month, and 23.1% within 6 months (data not shown).

The distribution of the main hypothesized exposures by source of information and repeat TBI status showed that 8.2% had an epilepsy/seizure disorder diagnosis in their medical record, and 5.6% reported Pre-existing epilepsy. Also, 22.9% had alcohol use and 6.0% had drug use recorded in their medical record (table 2). Generally, measures of agreement beyond chance between self-report and medical record data expressed in kappa indicate fair agreement for alcohol and drug (κ=.21; 95% CI, .16–.25) and good agreement for epilepsy/seizure disorder (κ=.43; 95% CI, .36–.51) (table 3).

Table 2. ESD and Alcohol/Drug Use Characteristics for Subjects With and Without Repeat TBI

Characteristics	RTBI
	Yes (n=147)	No (n=1952)	Total (N=2099)
ESD
Diagnosis code
Yes	15.2	7.7	8.2
No	84.8	92.3	91.8
Self-report
Yes	13.8	5.0	5.6
No	86.2	95.0	94.4
Alcohol/drug
Alcohol record review
Yes	31.0	22.3	22.9
No	69.0	77.7	77.1
Drug record review
Yes	9.0	5.7	6.0
No	91.0	94.3	94.0
Self-report
Yes	18.6	8.9	9.6
No	81.4	91.1	90.4

NOTE. Values are percent.

Abbreviations: ESD, epilepsy/seizure disorder; RTBI, repeat TBI.

Table 3. Agreement of Self-Report and Medical Record Data for the Definition of Alcohol/Drug Problem and ESD

Medical Record
	Self-Report
	Yes	No
I. Alcohol/Drug
Yes	55.2	20.9
No	44.8	79.1
II. ESD
Yes	57.6	5.3
No	42.4	94.7

NOTE. Values are percent.

Abbreviation: ESD, epilepsy/seizure disorder.

Fig 1, Fig 2 show the Kaplan-Meier curves for pre-existing epilepsy/seizure disorder and alcohol/drug treatment. While 2 of the lines cross in figure 1, the cumulative sums of martingale residuals test found that epilepsy/seizure disorder did not violate the proportional hazard assumption. Results of the log-rank test showed that the survival curves were significantly different depending on epilepsy/seizure disorder and drug/alcohol status (P<.001 and P=.002, respectively). In both analyses, persons who both self-reported and also had medical record documentation were more likely to sustain a repeat TBI. The Kaplan-Meier survival estimates for those with both self-report and a medical record of epilepsy/seizure disorder decreased more rapidly than for persons with either self-report or medical record after the first year, and decreased more rapidly than persons with no epilepsy/seizure disorder over the entire follow-up (see fig 1). The survival estimates for persons with both self-report and a medical record of an alcohol/drug problem decreased more rapidly than the 2 other curves, either and neither (see fig 2).

• 
  • View Large Image
  • Download to PowerPoint

• Fig 1. 

  Kaplan-Meier estimates of survival for repeat TBI by pre-existing epilepsy/seizure disorder. Neither, no epilepsy/seizure disorder in medical record or self-report; Either, epilepsy/seizure disorder by either medical record or self-report; Both, epilepsy/seizure disorder by both medical record and self-report. Abbreviations: ESD, epilepsy/seizure disorder; RTBI, repeat TBI.

• 
  • View Large Image
  • Download to PowerPoint

• Fig 2. 

  Kaplan-Meier estimates of survival for repeat TBI by pre-existing alcohol/drug problem. Neither, no alcohol/drug problem in medical record or self-report; Either, alcohol/drug problem by either medical record or self-report; Both, alcohol/drug problem by both medical record and self-report. Abbreviations: ALC DRG, alcohol/drug problem; RTBI, repeat TBI.

Table 4 summarizes all variables that remained significant in the multivariable CPH model. All variables met the proportional hazard assumption. After adjusting for the covariates included in the analysis, epilepsy/seizure disorder remained significantly associated with repeat TBI, while alcohol/drug problems did not remain significant. The hazard of repeat TBI for persons who self-reported epilepsy/seizure disorder and who also had an epilepsy/seizure disorder diagnosis code in their medical record during the index TBI visit was 2.3 times the hazard of those with no epilepsy/seizure disorder (hazard ratio, 2.3; 95% CI, 1.2–4.4; P=.011). Further, the hazard for persons with either self-reported epilepsy/seizure disorder or an epilepsy/seizure disorder diagnosis code in their record was 1.8 times the hazard of those with no Pre-existing epilepsy/seizure disorder (hazard ratio, 1.8; 95% CI, 1.1–3.0; P=.017). The hazard of repeat TBI for persons with a record of prior TBI was 1.9 times the hazard of those with no prior TBI (hazard ratio, 1.9; 95% CI, 1.1–3.4; P=.031). Compared with participants with commercial insurance, those with Medicaid were 2.4 times as likely (hazard ratio, 2.4; 95% CI, 1.4–4.1; P=.002) and those with no insurance were 2 times as likely (hazard ratio, 2.0; 95% CI, 1.3–3.1; P=.002) to sustain a repeat TBI.

Table 4. Crude and Adjusted Hazard Ratios for RTBI

Characteristics	Crude HR	95% CI	Adjusted HR⁎	95% CI
Pre-existing ESD
Self-report and in MR	2.9	1.6,5.4	2.3	1.2,4.4
Either self-report or in MR	2.4	1.5,3.7	1.8	1.1,3.0
No self-report and not in MR	Reference
Pre-existing drug/alcohol problem
Self-report and in MR	2.5	1.5,4.3	1.5	0.8,2.6
Either self-report or in MR	1.3	0.9,2.9	1.2	0.8,1.7
No self-report and not in MR	Reference
Prior history of TBI
Yes	2.6	1.4,4.5	1.9	1.1,3.4
No	Reference
Insurance
Medicaid	2.7	1.6,4.4	2.4	1.3,3.1
Medicare	1.6	1.1,2.5	1.0	0.6,1.9
Uninsured	2.4	1.6,3.6	2.0	1.3,3.1
Commercial	Reference

Abbreviations: ESD, epilepsy/seizure disorder; HR, hazard ratio; MR, medical record; RTBI, repeat TBI.

Back to Article Outline

Discussion 

Findings from this study demonstrate the important relations between Pre-existing epilepsy/seizure disorder and repeat TBI after controlling for other influential variables. Epilepsy/seizure disorder that is confirmed through medical records and self-report showed a greater hazard ratio than epilepsy/seizure disorder identified by only one of those sources. This could suggest that persons with Pre-existing and active seizures at the time of the index TBI are the most likely to have repeat TBI.

We found that overall, 5.6% of the cohort self-reported epilepsy/seizure disorder. This is higher than what was seen in the study by Annegers et al,²³ who found that 98 (1.6%) of 5984 persons seen for TBI had epilepsy/seizure disorder. However, the authors were not clear as to how epilepsy/seizure disorder was determined. In our study, we asked persons discharged from the hospital for TBI about Pre-existing epilepsy and also examined their charts at that admission for a diagnosis of epilepsy/seizure disorder. We found that persons with epilepsy/seizure disorder were more likely to sustain repeat TBI during follow-up than persons without epilepsy/seizure disorder. Previous studies of injury resulting from epilepsy/seizure disorder support our results. Neufeld et al²⁴ found that among persons who sustained at least 1 injury, the average number of injuries sustained was 2. In addition, the head is the most commonly injured place in persons with epilepsy/seizure disorder.²⁵ However, no other studies have specifically examined repeat injuries to the head. It is possible that a TBI in persons with an epilepsy/seizure disorder could exacerbate their condition. One case series²⁶ found that of 5 persons with epilepsy/seizure disorder who sustained a TBI, all experienced a worsening of their seizure control, and only 1 person returned to their pre-TBI state 1 year after injury. If seizure control becomes worse, the person could be at an increased risk for a repeat TBI event. Seizure frequency has been noted to be associated with injury in persons with epilepsy.²⁷ In basic science, researchers are currently looking at the disruption of central nervous system synaptic circuits when cells do not die of injury.²⁸ These disruptions could lower the seizure threshold, possibly resulting in more seizures.

We found that 9.6% of our cohort reported having been treated for an alcohol/drug problem before their index TBI. A review by Parry-Jones et al²⁹ showed that there are varying reports of alcohol and substance abuse prevalence before TBI. Pre-existing alcohol abuse ranges from 37% to 51%, and other substance abuse from 10% to 44%.²⁹ The reported prevalence in our cohort is much lower. This could have resulted from the way the survey questions were framed. We asked about previous treatment for drug/alcohol problems, and therefore missed those with a problem who did not receive treatment. When including information from medical records, 22.9% had alcohol use and 6% had drug use recorded in their medical records. A study by Wong et al³⁰ found that 57.6% of persons with repeat TBI had a history of alcohol use in their medical records compared with 33.8% of those without repeat TBI. Problem drinking before TBI could be indicative of problem drinking after TBI.³¹ Thus, the lower rate of alcohol and drug abuse evident in our cohort (28%) may be due to unidentified cases, which would affect the results observed.

While we did not find an association between alcohol/drug use and repeat TBI, there is supporting literature for this relation. In a substance abuse treatment group, Walker et al³² found an association between having 2 or more lifetime TBIs with loss of consciousness and illicit drug use in the previous 12 months, but not with alcohol use in the previous 12 months. However, this study assessed lifetime TBIs, and thus a temporal sequence between alcohol/drug use and TBI cannot be established. These results do suggest that alcohol and drug use should be investigated separately, but the current study was not able to delineate between previous alcohol and drug abuse. In a study of substance-abusing prisoners, Walker¹⁰ found an association between the number of TBIs and alcohol use in the previous 12 months. These 2 studies were conducted in samples of prisoners and substance abusers, and therefore might not be comparable to our study findings. Lastly, a study by Drubach¹¹ found that persons with 2 or more TBIs were more likely to have either alcohol or drug abuse, or both. A similar finding across all of these studies is the lack of temporal sequence of TBI and alcohol or drug use. Therefore, it is possible that in these studies, alcohol and drug abuse did not commence until after repeat TBI.

We found that insurance status was significantly associated with repeat TBI, with persons who had Medicaid or who were uninsured having a higher rate of repeat TBI than persons with commercial insurance. This observation is likely due to the inverse association of injury rates with socioeconomic status, suggesting the increased risk of injury among persons with lower socioeconomic status.³³, ³⁴ Nonetheless, for this study, we were only able to identify TBI events where persons were seen in either the hospital or the emergency department. Many persons without insurance use the emergency department as their primary source of health care, whereas persons with insurance have the option to be seen in private physician offices.³⁵, ³⁶ However, the extent to which patients with insurance are evaluated for TBI in physicians' offices is not well established. In the current study, we did not capture persons who were seen in private physicians' offices, and there could have been underrepresentation of persons with insurance who sought care for TBI in physicians' offices.

Of our cohort, 3.9% had a record of a prior TBI. This percentage is slightly lower than that reported by Annegers,²³ who found that 397 (6.6%) of 5984 persons had a record of a previous TBI. However, Annegers was able to examine the lifetime history of prior TBI, while our data were limited to prior TBI in the previous 3 to 6 years. We found a significant relationship between having a prior TBI and repeat TBI. This finding is supported by previous work. One population-based study⁵ found that the risk for a subsequent TBI was 3 times greater in persons who had 1 TBI than those who had not, and 8 times greater among persons with 2 prior TBIs. The risk for TBI in collegiate players was related to their high school careers, with players who sustained a TBI in high school football almost twice as likely to have one during their college football career.⁴

Study Limitations 

While this study provides valuable information on Pre-existing conditions and repeat TBI, several limitations must be noted. First, we were able to identify episodes of repeat TBI only through emergency department visits and hospital discharges. It is possible that TBIs may have occurred outside the jurisdiction of the surveillance capture area, were not serious enough to warrant emergency department visits or hospital admissions, or that persons may not have sought medical attention. While severe TBIs are likely to have more of an impact on TBI-related outcomes than mild TBIs, it is possible that these undetected injuries could have swayed the results towards the null because those persons would only have been considered as having 1 TBI. Future studies should make an effort to quantify all TBIs, including those treated in private physicians' offices. However, it is expected that the number of South Carolina residents who sustain TBI while in another state will be very small, and those TBIs that are not recognized are likely to be minor injuries. Second, while personal identifiers allow linkage across data sets, these identifiers only go back to 1996, and therefore, no information on TBI before 1996 was available. Third, while much of the data were gathered from medical records, a portion was self-report, which can introduce recall bias. Specifically, the 2 Pre-existing conditions were determined partially by self-report, and in some cases epilepsy/seizure disorder might not have been recorded in the medical record. However, previous research has found that 23% of persons with confirmed epilepsy/seizure disorder will deny it in a survey.³⁷ Therefore, we could have an underestimate of epilepsy/seizure disorder. In addition, determination of previous alcohol/drug treatment was also by self-report, but previous research has shown that alcohol use in persons with TBI was concordant with their relatives' reports, suggesting that they are not underreporting their use.³⁸, ³⁹ Fourth, interviews were conducted on only 57% of eligible participants. Although nonwhite males, persons with less severe TBI, and persons without health insurance coverage were less likely to participate in the interview,¹⁵ we did not find differences in occurrence of repeat TBI between persons who participated and those who did not.

Although kappa statistics were reported for the 2 measurements of epilepsy/seizure disorder and drug/alcohol problems, we were not expecting them to be very high, especially for drug/alcohol problems. The self-report of drug/alcohol problems asked about previous treatment, which might not have been recorded in the medical record. In addition, persons who had not received treatment might have still had a drug or alcohol problem recorded in their medical record. Likewise, persons with self-reported epilepsy/seizure disorder, but who did not have active seizures, might not have had an epilepsy/seizure disorder code in the medical record. In addition, seizures occurring and recorded in the medical record might not have been established as epilepsy/seizure disorder by a physician previously.

This study has several notable strengths. It is the first large sample, population-based cohort study to examine the relation between Pre-existing conditions and repeat TBI. While other studies with smaller samples and inconsistent definitions of TBI have been published, this study consisted of 2118 persons with a clinical diagnosis of TBI. Through this study, we were able to use longitudinal information on the cohort to assess repeat TBI up to 6 years after the index TBI. Finally, we were able to establish the hazard of epilepsy and seizure disorders in predisposing persons to repeat TBI.

Back to Article Outline

Conclusions 

Our results showed that persons with TBI who have Pre-existing epilepsy/seizure disorders are more likely to sustain a subsequent TBI than persons without those conditions. TBI in a person with epilepsy/seizure disorder could exacerbate that condition, making that person susceptible to further TBIs. Future investigations should include measuring changes in Pre-existing conditions, identifying specific types of seizures that predispose to head trauma, and determining the influence of polypharmacy in the acquisition of repeat TBI.

Suppliers

Back to Article Outline

References 

1. Langlois JA, Thomas KE. Traumatic brain injury in the United States: emergency department visits, hospitalizations, and deaths. Atlanta: Centers for Disease Control and Prevention, National Center for Injury Prevention and Control; 2006;
   • View In Article
2. Thurman DJ, Alverson C, Dunn KA, Guerrero J, Sniezek JE. Traumatic brain injury in the United States: a public health perspective. J Head Trauma Rehabil. 1999;14:602–615
   • View In Article
   • MEDLINE
   • CrossRef
3. Selassie AW, Zaloshnja E, Langlois JA, Miller T, Jones P, Steiner C. Incidence of long-term disability following traumatic brain injury hospitalization, United States, 2003. J Head Trauma Rehabil. 2008;23:123–131
   • View In Article
   • CrossRef
4. Albright JP, McAuley E, Martin RK, Crowley ET, Foster DT. Head and neck injuries in college football: an eight-year analysis. Am J Sports Med. 1985;13:147–152
   • View In Article
   • MEDLINE
   • CrossRef
5. Annegers JF, Grabow JD, Kurland LT, Laws ER. The incidence, causes, and secular trends of head trauma in Olmsted County, Minnesota, 1935-1974. Neurology. 1980;30:912–919
   • View In Article
   • MEDLINE
6. Guskiewicz KM, Marshall SW, Bailes J, et al. Association between recurrent concussion and late-life cognitive impairment in retired professional football players. Neurosurgery. 2005;57:719–726
   • View In Article
7. Guskiewicz KM, Marshall SW, Bailes J, et al. Recurrent concussion and risk of depression in retired professional football players. Med Sci Sports Exerc. 2007;39:903–909
   • View In Article
   • MEDLINE
   • CrossRef
8. Carlsson GS, Svardsudd K, Welin L. Long-term effects of head injuries sustained during life in three male populations. J Neurosurg. 1987;67:197–205
   • View In Article
   • MEDLINE
   • CrossRef
9. Horner MD, Ferguson PL, Selassie AW, Labbate LA, Kniele K, Corrigan JD. Patterns of alcohol use 1 year after traumatic brain injury: a population-based, epidemiological study. J Int Neuropsychol Soc. 2005;11:322–330
   • View In Article
   • MEDLINE
10. Walker R, Hiller M, Staton M, Leukefeld CG. Head injury among drug abusers: an indicator of co-occurring problems. J Psychoactive Drugs. 2003;35:343–353
    • View In Article
    • MEDLINE
    • CrossRef
11. Drubach DA, Kelly MP, Winslow MM, Flynn JP. Substance abuse as a factor in the causality, severity, and recurrence rate of traumatic brain injury. Md Med J. 1993;42:989–993
    • View In Article
    • MEDLINE
12. Salcido R, Costich JF. Recurrent traumatic brain injury. Brain Inj. 1992;6:293–298
    • View In Article
    • MEDLINE
    • CrossRef
13. Annegers JF, Coan SP. The risks of epilepsy after traumatic brain injury. Seizure. 2000;9:453–457
    • View In Article
    • Abstract
    • Full-Text PDF (215 KB)
    • CrossRef
14. Wirrell EC. Epilepsy-related injuries. Epilepsia. 2006;(47 Suppl 1):79–86
    • View In Article
15. Pickelsimer EE, Selassie AW, Gu JK, Langlois JA. A population-based outcomes study of persons hospitalized with traumatic brain injury: operations of the South Carolina Traumatic Brain Injury Follow-up Registry. J Head Trauma Rehabil. 2006;21:491–504
    • View In Article
    • MEDLINE
    • CrossRef
16. Butler J, Langlois JA. Central nervous system injury surveillance: annual data submission standards-2000. Atlanta: US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Injury Prevention and Control; 2001;
    • View In Article
17. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36:8–27
    • View In Article
    • MEDLINE
    • CrossRef
18. Selassie AW, Ferguson PL, Wannamaker BB, Pickelsimer EE, Turner R, Smith GS. The South Carolina Epidemiological Studies of Epilepsy and Seizure Disorder. Charleston: Medical University of South Carolina; 2007;
    • View In Article
19. South Carolina Office of Research and Statistics. South Carolina population reports. 2006;http://www.ors2.state.sc.us/population/estimates/county1.aspAccessed August 4, 2009
    • View In Article
20. Association for the Advancement of Automotive Medicine. The Abbreviated Injury Scale. Des Plaines. 1990;
    • View In Article
21. National Center for Injury Prevention and Control. The report to Congress on mild traumatic brain injury in the United States: steps to prevent a serious public health problem. Atlanta: Centers for Disease Control and Prevention; 2003;
    • View In Article
22. Lin D, Wei L, Ying Z. The robust inference for the proportional hazards model. J Am Stat Assoc. 1993;84:1074–1078
    • View In Article
23. Annegers JF, Hauser WA, Coan SP, Rocca WA. A population-based study of seizures after traumatic brain injuries. N Engl J Med. 1998;338:20–24
    • View In Article
    • MEDLINE
    • CrossRef
24. Neufeld MY, Vishne T, Chistik V, Korczyn AD. Life-long history of injuries related to seizures. Epilepsy Res. 1999;34:123–127
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (61 KB)
    • CrossRef
25. Neufeld MY, Chistik V, Kogan E, Korczyn AD. Physically traumatic events in patients with epilepsy compared with non-epileptic subjects. Seizure. 2000;9:595–597
    • View In Article
    • Abstract
    • Full-Text PDF (48 KB)
    • CrossRef
26. Tai PC, Gross DW. Exacerbation of pre-existing epilepsy by mild head injury: a five patient series. Can J Neurol Sci. 2004;31:394–397
    • View In Article
    • MEDLINE
27. Lawn ND, Bamlet WR, Radhakrishnan K, O'Brien PC, So EL. Injuries due to seizures in persons with epilepsy: a population-based study. Neurology. 2004;63:1565–1570
    • View In Article
28. Cohen AS, Pfister BJ, Schwarzbach E, Grady MS, Goforth PB, Satin LS. Injury-induced alterations in CNS electrophysiology. Prog Brain Res. 2007;161:143–169
    • View In Article
    • CrossRef
29. Parry-Jones BL, Vaughan FL, Miles Cox W. Traumatic brain injury and substance misuse: a systematic review of prevalence and outcomes research (1994-2004). Neuropsychol Rehabil. 2006;16:537–560
    • View In Article
    • MEDLINE
    • CrossRef
30. Wong PP, Dornan J, Schentag CT, Ip R, Keating M. Statistical profile of traumatic brain injury: a Canadian rehabilitation population. Brain Inj. 1993;7:283–294
    • View In Article
    • MEDLINE
    • CrossRef
31. Ponsford J, Whelan-Goodinson R, Bahar-Fuchs A. Alcohol and drug use following traumatic brain injury: a prospective study. Brain Inj. 2007;21:1385–1392
    • View In Article
    • CrossRef
32. Walker R, Cole JE, Logan TK, Corrigan JD. Screening substance abuse treatment clients for traumatic brain injury: prevalence and characteristics. J Head Trauma Rehabil. 2007;22:360–367
    • View In Article
    • CrossRef
33. d'Errico A, Punnett L, Cifuentes M, et al. Hospital injury rates in relation to socioeconomic status and working conditions. Occup Environ Med. 2007;64:325–333
    • View In Article
    • CrossRef
34. Birken CS, Parkin PC, To T, Macarthur C. Trends in rates of death from unintentional injury among Canadian children in urban areas: influence of socioeconomic status. CMAJ. 2006;175:867
    • View In Article
    • CrossRef
35. Suruda A, Burns TJ, Knight S, Dean JM. Health insurance, neighborhood income, and emergency department usage by Utah children 1996-1998. BMC Health Serv Res. 2005;5:29
    • View In Article
    • MEDLINE
    • CrossRef
36. Lucas RH, Sanford SM. An analysis of frequent users of emergency care at an urban university hospital. Ann Emerg Med. 1998;32:563–568
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (60 KB)
    • CrossRef
37. Beran RG, Michelazzi J, Hall L, Tsimnadis P, Loh S. False-negative response rate in epidemiologic studies to define prevalence ratios of epilepsy. Neuroepidemiology. 1985;4:82–85
    • View In Article
    • MEDLINE
    • CrossRef
38. Sander AM, Witol AD, Kreutzer JS. Alcohol use after traumatic brain injury: concordance of patients' and relatives' reports. Arch Phys Med Rehabil. 1997;78:138–142
    • View In Article
    • Abstract
    • Full-Text PDF (689 KB)
    • CrossRef
39. Kruetzer JS, Witol AD, Sander AM, Cifu DX, Marwitz JH, Delmonico R. A prospective longitudinal multicenter analysis of alcohol use patterns among persons with traumatic brain injury. J Head Trauma Rehabil. 1996;11:58–69
    • View In Article
    • CrossRef

• a Johns Hopkins Bloomberg School of Public Health, 615 N Wolfe St, Baltimore, MD 21205; and Tri-Analytics, Inc, 134 Industry Ln, Ste C, Forest Hills, MD 21050.
• b SAS Institute Inc, 100 SAS Campus Dr, Cary, NC 27513-2414.