Abstract
Epidemiological studies of medical comorbidities and possible gender differences associated with traumatic brain injury (TBI) are limited, especially among military veterans. The purpose of this study was to examine relationships between TBI history and a wide range of medical conditions in a large, national sample of veterans, and to explore interactions with gender. Participants of this cross-sectional epidemiological study included 491,604 veterans (9.9% TBI cases; 8.3% women) who enrolled in the VA Million Veteran Program (MVP). Outcomes of interest were medical comorbidities (i.e., neurological, mental health, circulatory, and other medical conditions) assessed using the MVP Baseline Survey, a self-report questionnaire. Logistic regression models adjusting for age and gender showed that veterans with TBI history consistently had significantly higher rates of medical comorbidities than controls, with the greatest differences observed across mental health (odds ratios [ORs] = 2.10–3.61) and neurological (ORs = 1.57–6.08) conditions. Similar patterns were found when evaluating men and women separately. Additionally, significant TBI-by-gender interactions were observed, particularly for mental health and neurological comorbidities, such that men with a history of TBI had greater odds of having several of these conditions than women with a history of TBI. These findings highlight the array of medical comorbidities experienced by veterans with a history of TBI, and illustrate that clinical outcomes differ for men and women with TBI history. Although these results are clinically informative, more research is needed to better understand the role of gender on health conditions in the context of TBI and how gender interacts with other social and cultural factors to influence clinical trajectories following TBI. Ultimately, understanding the biological, psychological, and social mechanisms underlying these comorbidities may help with tailoring TBI treatment by gender and improve quality of life for veterans with TBI history.
Introduction
The high prevalence of traumatic brain injury (TBI) in military service members and veterans 1 has generated interest in studying clinical outcomes and recovery patterns following injury. Notably, prior research has established that veterans with a history of TBI are at increased risk for both acute and chronic health conditions, including mental health disorders, cognitive difficulties, and other medical comorbidities such as pain, as well as a variety of functional and psychosocial limitations. 2 –5 These sequelae are, in turn, associated with higher healthcare utilization, higher medical costs, greater disability, poorer quality of life, and increased mortality. 6 –9
Within the mental health domain, post-traumatic stress disorder (PTSD) and depression are among the most commonly studied comorbid conditions, co-occurring at a high frequency in veterans with TBI history. 2,10 Cognitive functioning has also frequently been studied in the context of military TBI, 11 –13 and a high-priority research area has been examining the increased risk for neurodegenerative conditions such as Alzheimer's disease (AD) and AD-related dementias (ADRD) 14 –16 in individuals with TBI history. Although the precise relationship between TBI and AD/ADRD is still debated, several studies have established that TBI is a risk factor for AD/ADRD and/or dementia. 3,15,17 –20 More recently, investigators have gone beyond psychiatric and cognitive conditions to evaluate other medical comorbidities such as hypertension, diabetes, and cardiovascular disease, highlighting an association between TBI history and these chronic health conditions. 21 –24
Evaluating individual differences as possible modifiers of TBI outcome and recovery is another area of exploration that has received increased attention recently, with biological sex and gender emerging as important variables of interest. 25 –27 However, sex and gender differences have been severely understudied within the context of military TBI, likely because of the disproportionately higher prevalence of men serving in the military relative to women. Although there is some evidence to suggest that women experience poorer clinical outcomes following TBI than men, 28 –30 this research is in its infancy, and well-powered studies are needed to better inform understanding of the extent to which men and women experience differential clinical outcomes post-injury. Learning more about sex and gender differences in the context of TBI will have important treatment and prognostic implications for veterans and their caregivers, especially women veterans who are currently among the fastest-growing veteran population. 31
Although prior research supports the link between TBI history and a wide range of medical comorbidities, questions remain about the nature of these associations and who may be most susceptible to experiencing chronic medical comorbidities following injury. The purpose of the present cross-sectional study was to examine associations between TBI history and medical comorbidities as well as the interaction of TBI with gender on medical comorbidities using a nationwide sample of military Veterans enrolled in the VA Million Veteran Program (MVP). Briefly, MVP is a nationwide research initiative that brings together electronic health record data and comprehensive surveys to examine how genes and environment influence health and behaviors in veterans. 32 MVP offers a unique opportunity to evaluate gender differences in the context of TBI because of the purposeful oversampling of women veterans. We hypothesized that veterans with a history of TBI would experience higher rates of medical condition comorbidities than controls (i.e., veterans without a history of TBI) and that gender would significantly moderate the association between TBI status and medical comorbidities.
Methods
Procedures and participants
MVP is a nationwide research initiative that received VA Central Institutional Review Board (IRB) approval in 2010. 32 The present study is part of the MVP project “MVP026,” which received VA Central IRB approval in 2019.
Any veteran is eligible to participate in MVP. MVP recruitment and enrollment procedures have been described previously by Gaziano and colleagues. 32 Data for the present study were solely derived from the MVP Baseline Survey (described subsequently). Eligible participants included MVP-enrolled veterans who completed the MVP Baseline Survey as of September 30, 2020 (n = 491,725); of these, 121 were excluded because of missing gender. Therefore, the total number of participants available for analysis was 491,604.
Data sources
MVP Baseline Survey
The MVP Baseline Survey 32 is a comprehensive self-report questionnaire administered to MVP participants upon enrollment that assesses sociodemographic data, military service history, activities and habits, health status, and medical history and healthcare usage. The current study evaluated sociodemographic and medical history data from this survey. Sociodemographic variables used in this study included gender, age, race, ethnicity, education, marital status, and annual household income. Gender was specifically assessed by the following question: “What is your gender?” with response options of “Male” and “Female.” As for medical history data, participants were provided with a wide range of medical conditions and were instructed to indicate whether they had ever been diagnosed with these conditions (i.e., lifetime history was evaluated, meaning that conditions could have occurred before, during, or after military service). The following medical condition domains were evaluated in this study: circulatory system (10 conditions), mental health (10 conditions), neurological (10 conditions), and other conditions (8 conditions).
TBI case/control definition
To assess TBI history, two items were evaluated from the medical history section of the MVP Baseline Survey: (1) “Concussion or loss of consciousness” and (2) “Traumatic brain injury.” Veterans were classified as a TBI case if they indicated having a history of either of these conditions (i.e., presence of concussion and/or TBI). Veterans were classified as a TBI control if there was no indication of both conditions (i.e., no indication of concussion and TBI). Note that TBI history was independent of military service and referred to lifetime history of TBI.
Data analysis
All analyses for this study were conducted in RStudio. Descriptive statistics (means and standard deviations (SDs) for continuous variables and counts and percentages for categorical variables) were conducted to characterize participant sociodemographic variables; data were examined for the overall sample and separately for men and women. Group differences in sociodemographic variables were evaluated using two-sample t tests for means and z tests for proportions (significance at the 5% level).
Primary analyses were performed using logistic regression adjusted for age and gender to evaluate the likelihood of experiencing various medical conditions in TBI cases versus controls for the overall sample and for men and women separately (these latter analyses were adjusted for age only). We also evaluated the interaction between TBI history and gender on the likelihood of reporting medical conditions with the addition of an interaction term in a similar model run on the whole cohort. Given the large number of medical conditions evaluated, the Benjamini–Hochberg procedure was applied, and false discovery rate (FDR)-adjusted p values were reported. Odds ratios (ORs) and 95% confidence intervals (CIs) were reported for all main effects (TBI case vs. control for overall sample; TBI case vs. control for men; and TBI case vs. control for women). We then analyzed the difference in likelihood of reporting each condition between men with TBI and women with TBI using the ratio of ORs and 95% CIs produced by the model, including the interaction term for TBI-by-gender.
Results
Sample characteristics
Participant sociodemographic characteristics are presented in Table 1 for the overall sample and separately for men and women, divided by TBI status (TBI case vs. control). Of the 491,604 veterans included in the study, 9.9% (n = 48,646) had a self-reported history of TBI and 8.3% (n = 40,979) identified as female. With regard to ethnicity, 7.0% of the overall sample self-identified as Hispanic. As for race, 13.3% identified as black, 82.3% identified as white, and 4.3% identified as another race. The average age of the sample was 65.1 years (SD = 12.5).
Participant Sociodemographic Characteristics
Percentages are calculated based on the available n for each variable.
Group differences in sociodemographic variables were evaluated using two-sample t tests for means and z tests for proportions; significant differences (p < 0.001) were observed across all variables evaluated.
TBI, traumatic brain injury; SD, standard deviation; GED, general education diploma.
Prevalence of medical comorbidities in overall sample and across men and women
Table 2 displays the prevalence of medical comorbidities for the overall sample (i.e., men and women together). Veterans with a history of TBI (i.e., the “TBI” group) consistently had significantly higher rates of medical comorbidities compared with controls (i.e., the “No TBI” group), with the most notable differences observed across mental health (ORadj = 2.10–3.61) and neurological (ORadj = 1.57–6.08) conditions. Among the overall sample, the largest ORs were observed for epilepsy/seizure (ORadj = 6.08; 95% CI [5.83, 6.34]), memory loss or impairment (ORadj = 6.06; 95% CI [5.92, 6.19]), other headaches (ORadj = 4.21, 95% CI [4.11, 4.31]), and migraine (ORadj = 4.05, 95% CI [3.95, 4.14]), such that veterans with a history of TBI had increased odds of having these conditions relative to controls. In contrast, medical conditions such as high blood pressure (ORadj = 1.13; 95% CI [1.11, 1.15]) and diabetes (ORadj = 1.03; 95% CI [1.01, 1.06]) had ORs close to 1, indicating that the odds of experiencing these conditions is similar for TBI cases and controls. Figure 1 displays all medical condition comorbidities evaluated, along with their adjusted ORs and 95% CIs.
Adjusted odds ratios with 95% confidence intervals of medical comorbidities for the overall sample: traumatic brain injury (TBI) versus no TBI.
Prevalence of Medical Comorbidities for the Overall Sample
Models adjusted for age and gender. bFalse discovery rate (FDR) correction for multiple testing.
TBI, traumatic brain injury; OR, odds ratio; CI, confidence interval; ADHD, attention-deficit/hyperactivity disorder; PTSD, post-traumatic stress disorder; CAD, coronary artery disease; CHD, coronary heart disease; DVT, deep vein thrombosis.
Similar patterns were generally observed when evaluating men and women separately (see Table 3 and Fig. 2). Across both genders, the most notable differences between TBI cases and controls were observed across mental health and neurological conditions. For men, the largest ORs were observed for epilepsy/seizure (ORadj = 6.16; 95% CI [5.89, 6.44]), memory loss or impairment (ORadj = 5.99; 95% CI [5.85, 6.13]), other headaches (OR = 4.38, 95% CI [4.27, 4.48]), and migraine (OR = 4.25; 95% CI [4.14, 4.36]). For women, the largest ORs were observed for memory loss or impairment (ORadj = 6.78; 95% CI [6.29, 7.30]), epilepsy/seizure (OR = 5.29, 95% CI [4.67, 5.99]), and dementia (ORadj = 4.37; 95% CI [3.21, 5.88]). As with the overall sample, both high blood pressure (men: ORadj = 1.13; 95% CI [1.11, 1.15]; women: ORadj = 1.05; 95% CI [0.98, 1.12]) and diabetes (men: ORadj = 1.03; 95% CI [1.01, 1.05]; women: ORadj = 1.03; 95% CI [0.95, 1.12]) had ORs close to 1 for men and women.
Adjusted odds ratios with 95% confidence intervals of medical comorbidities for men and women separately: traumatic brain injury (TBI) versus no TBI.
Prevalence of Medical Comorbidities by TBI Status and Gender and Interaction Results
Percentages are calculated based on the available n for each variable. The models were run separately for men (n = 450,625) and women (n = 40,979), and the final model was run on the overall sample (n = 491,604) with the addition of the interaction term.
Models adjusted for age. bFalse discovery rate (FDR) correction for multiple testing. cThis value represents the ratio of odds ratios and reflects the difference in likelihood of reporting each medical condition among men with TBI history versus women with TBI history, with the corresponding p value listed in the last column.
TBI, traumatic brain injury; OR, odds ratio; CI, confidence interval; ADHD, attention-deficit/hyperactivity disorder; PTSD, post-traumatic stress disorder; CAD, coronary artery disease; CHD, coronary heart disease; DVT, deep vein thrombosis.
TBI x gender interactions
Overlaying the adjusted ORs obtained by gender across the different medical condition comorbidities suggested that there may be a differential impact of TBI by gender (Fig. 3). To examine this hypothesis, we tested the significance of an interaction term for TBI and gender in the model (Table 3, Model 3). Gender significantly moderated the association between TBI status and the majority of mental health conditions, such that men with a history of TBI had greater odds of having these conditions than women with a history of TBI (ratios of ORadj = 1.16–1.47 for the significant interactions). Significant interactions were also observed for many of the neurological conditions evaluated. In particular, men with a history of TBI had increased odds of experiencing epilepsy/seizure, other headaches, migraine, and multiple sclerosis (ratios of ORadj = 1.18–1.45), but lower odds for memory loss or impairment and spinal cord injury/impairment relative to women with a history of TBI (ratios of ORadj = 0.85–0.89). Additional TBI-by-gender interactions were observed for some circulatory system conditions and other medical conditions.
Adjusted odds ratios with 95% confidence intervals of medical comorbidities for men and women overlaid: traumatic brain injury (TBI) versus no TBI.
Discussion
We evaluated rates of medical condition comorbidities in veterans with and without a history of TBI and found consistently higher rates of medical comorbidities in veterans with a history of TBI than in those without a history of TBI, even after adjusting for age and gender. Effect sizes were largest for mental health disorders and neurological conditions. Among these, epilepsy/seizure, memory loss or impairment, other headaches, and migraine had the largest effects in the overall sample. When examining men and women separately, similar patterns were observed (i.e., both male and female veterans with a history of TBI had higher rates of medical comorbidities than veterans without a TBI history), with the largest effects again found for mental health disorders and neurological conditions. Moreover, significant TBI-by-gender interactions were observed across several mental health and neurological conditions, as well as some circulatory conditions and other medical conditions. Our findings of increased rates of medical comorbidities and significant gender differences among veterans with TBI history are relevant for clinical care, and may be especially important when engaging in treatment and intervention planning.
The elevated rates of medical comorbidities in veterans with a history of TBI that were observed in this study—particularly for mental health and neurological conditions—are consistent with previously published findings in both the military and civilian TBI literature. 2 –4,22,23,33 –38 Additionally, the elevated rates of circulatory system conditions (e.g., stroke) in veterans with a history of TBI are similar to recent findings in Iraq/Afghanistan-era veterans. 22,24 However, unlike prior studies, 22,23,33 we did not observe elevated rates of hypertension or diabetes in veterans with a history of TBI relative to those without. In fact, both hypertension and diabetes were among the conditions with the smallest effects when comparing TBI cases with controls (see Fig. 1). Several factors have been previously shown to contribute to hypertension risk, including sociodemographic factors, mental health, combat exposure/combat injury, and lifestyle factors, 39 –42 and veterans tend to experience higher rates of hypertension than the general (civilian) population. 43 Although longitudinal research is needed to determine whether any causal associations exist between TBI and hypertension, our data fail to show an association between the two, suggesting that the relationship is more likely influenced by the many confounding factors described.
We also compared rates of medical condition comorbidities in veterans with and without a history of TBI among men and women separately, and similarly showed higher rates of medical comorbidities in veterans with a TBI history in both genders. Moreover, when evaluating TBI-by-gender interactions, our data showed that gender significantly moderated the association between TBI status and several medical comorbidities. It is noteworthy that significant interactions were observed for 8 of the 10 mental health conditions and 6 of the 10 neurological conditions. Remarkably, the interaction effect showed that relative to women with a history of TBI, men with a history of TBI had increased odds of having mental health conditions, as well as increased odds of having many of the neurological conditions including epilepsy/seizure, other headaches, migraine, and multiple sclerosis (this is depicted in Figure 3). In contrast, men with a history of TBI had lower odds of experiencing memory loss or impairment and spinal cord injury/impairment than women with a history of TBI. Despite the increased odds for many of the conditions evaluated for men with a history of TBI, it is important to still appreciate the generally high rates of mental health and neurological conditions experienced by both genders when TBI was also experienced (see Table 3).
Although prior research exploring gender differences in clinical outcomes following TBI in veterans is fairly limited, there are some data suggesting that women generally experience poorer mental health outcomes and greater neurobehavioral symptoms than men. 28 –30,44 Further, the broader veteran literature has reported gender differences across a wide range of medical conditions among Veterans Health Administration enrollees 45 and within MVP enrollees more specifically, 46 irrespective of TBI status. Taken together, these findings highlight the importance of taking gender into account when determining the treatment needs of veterans with a history of TBI, as men and women may experience different trajectories before and after injury. Ultimately, more research is needed to better contextualize these findings and understand mechanisms associated with the observed gender differences, but these initial MVP findings suggest that gender has an important moderating role on medical condition comorbidities in veterans with a TBI history.
The clinical implications of this study are far reaching, and further reinforce the consideration of individual differences (such as gender) when evaluating and treating veterans with a history of TBI. Results further support the need for tailoring treatment and intervention programs to the unique needs of men and women veterans. Currently, clinical practice guidelines for TBI have primarily been derived from male-dominated research, and many clinical interventions designed for the veteran population—particularly in the mental health sphere—have been assessed using predominantly male research participants. 47,48 However, recent efforts to tailor PTSD interventions to women veterans have shown success 49,50 and may provide a model for developing effective treatments for women with a history of TBI. Finally, it is worth mentioning the importance of integrated clinical care across the life span for men and women veterans with a history of TBI. 51
Although findings are clinically applicable, study limitations need to be considered. Our study utilized a cross-sectional design, so it is not possible to infer cause–effect associations. Also, the relationships between TBI and comorbid conditions are likely not uniform; although simultaneous onset of TBI and other conditions is possible (e.g., PTSD, depression), mechanisms by which TBI may increase risk of other conditions is less well documented and needs to be explored through longitudinal research. Further, all diagnoses (including TBI status and medical comorbidities) were based on self-report data and therefore are subject to recall bias. Another limitation is that we did not distinguish among TBI severity levels in our analyses or evaluate the impact of the number of TBIs, as we did not have this granularity of data available within the survey. Additionally, the MVP Baseline Survey assessed self-reported gender via a binary response option (using “Male” and “Female” as response options). Recent research by Giordano and colleagues 52 highlighted the need to incorporate gender beyond the binary, and others have emphasized the need to consider gender/gender identity versus biological sex and better integrate other aspects of intersectionality into this line of research. 26,53,54 Final study limitations relate to generalizability. Although the sample was extremely large and the study cohort consisted of veterans from across the nation, the majority self-identified as white and non-Hispanic. Moreover, even though our study represents one of the largest published studies on women veterans with a history of TBI, women made up only ∼10% of the overall sample, underscoring the need to increase representation of women in veteran TBI research.
Conclusion
Higher rates of medical condition comorbidities were observed in veterans with a history of TBI compared with those without a history of TBI, with the largest effect sizes found for mental health and neurological conditions. Further, our results suggest that clinical outcomes in the context of TBI differ as a function of gender, as significant TBI-by-gender interactions were observed across several medical conditions. Although these results are clinically informative, more research is needed to understand the role of gender on health conditions in the context of TBI and how gender interacts with other social and cultural factors to influence clinical trajectories following TBI. Future directions could also include latent class analysis to better characterize disease burden in veterans with a history of TBI and evaluate whether these classes vary by gender. MVP is well suited to examine these questions and explore the underlying genetic associations that may contribute to these disparate outcomes among men and women veterans.
Transparency, Rigor, and Reproducibility Summary
All data for this study came from the MVP and are subject to MVP rules and regulations. Per MVP, final data sets underlying this study cannot be shared outside the VA, except as required under the Freedom of Information Act (FOIA), per VA policy, because of the sensitive nature of medical health records of veterans. However, upon request through the formal mechanisms in place and pending approval from the VHA Office of Research Oversight (ORO), a de-identified, anonymized data set underlying this study may be created and shared. This study was not formally registered because of the MVP data restrictions described. The analysis plan was also not formally pre-registered. Given the sizeable nature of MVP, we were well powered to conduct all analyses. The final sample size of 491,604 subjects was determined based on the availability of relevant data in MVP. Given the large number of medical conditions evaluated, the Benjamini–Hochberg procedure was applied to adjust for multiple comparisons, and FDR-adjusted p values were reported.
Footnotes
Acknowledgments
The authors sincerely thank the veterans who volunteered to participate in the MVP. This research is based on data from the MVP (Project MVP026), Office of Research and Development, Veterans Health Administration. This publication does not represent the views of the Department of Veteran Affairs or the United States Government.
Authors' Contributions
V.C.M., M.G., J.Y., M.T.L., and C.C.C. were responsible for conceptualization; V.C.M., M.T.L., and M.G. were responsible for methodology; M.G. was responsible for formal analysis; M.G., and C.C.C. were responsible for data curation; V.C.M., and J.Y. were responsible for writing – original draft; V.C.M., M.G., J.Y., M.T.L., and C.C.C. were responsible for writing – review and editing; M.G. and J.Y. were responsible for visualization; V.C.M. was responsible for funding acquisition; and C.C.C. was responsible for supervision.
Funding Information
This work was supported by a VA Career Development Award (IK2 CX001952) from the United States Department of Veterans Affairs Clinical Science Research and Development Service awarded to Dr. Merritt. Dr. Gasperi is funded by a VA Career Development Award (1IK2 CX002107) from the United States Department of Veterans Affairs Clinical Science Research and Development Service.
Author Disclosure Statement
No competing financial interests exist.