Skip to main content
NCBI home page
VA Author Manuscripts logoLink to VA Author Manuscripts
. Author manuscript; available in PMC: 2026 Mar 14.
Published in final edited form as: Mov Disord. 2023 Jun 13;38(8):1483–1492. doi: 10.1002/mds.29457

Lifelong Association of Disorders Related to Military Trauma with Subsequent Parkinson’s Disease

Gregory D Scott 1,2,*, Lee E Neilson 3,4, Randy Woltjer 1, Joseph F Quinn 3,4, Miranda M Lim 3,4,5,6
PMCID: PMC12983286  NIHMSID: NIHMS2144142  PMID: 37309872

Abstract

Background:

Trauma-related disorders such as traumatic brain injury (TBI) and posttraumatic stress disorder (PTSD) are emerging as risk factors for Parkinson’s disease (PD), but their association with development of PD and independence from comorbid disorders remains unknown.

Objective:

To examine TBI and PTSD related to early trauma in military veterans using a case-control study.

Methods:

PD was identified by International Classification of Diseases (ICD) code, recurrent PD-specific prescriptions, and availability of 5+ years of earlier records. Validation was performed by chart review by a movement disorder–trained neurologist. Control subjects were matched 4:1 by age, duration of preceding health care, race, ethnicity, birth year, and sex. TBI and PTSD were identified by ICD code and onset based on active duty. Association and interaction were measured for TBI and PTSD with PD going back 60 years. Interaction was measured for comorbid disorders.

Results:

A total of 71,933 cases and 287,732 controls were identified. TBI and PTSD increased odds of subsequent PD at all preceding 5-year intervals back to year −60 (odds ratio range: 1.5 [1.4, 1.7] to 2.1 [2.0, 2.1]). TBI and PTSD showed synergism (synergy index range: 1.14 [1.09, 1.29] to 1.28 [1.09, 1.51]) and additive association (odds ratio range: 2.2 [1.6, 2.8] to 2.7 [2.5, 2.8]). Chronic pain and migraine showed greatest synergy with PTSD and TBI. Effect sizes for trauma-related disorders were comparable with established prodromal disorders.

Conclusions:

TBI and PTSD are associated with later PD and are synergistic with chronic pain and migraine. These findings provide evidence for TBI and PTSD as risk factors preceding PD by decades and could aid in prognostic calculation and earlier intervention.

Keywords: Parkinson’s disease, traumatic brain injury, posttraumatic stress disorder, military, chronic pain

Introduction

Identification of risk factors for Parkinson’s disease (PD) is important for understanding the pathogenesis, for prognostic calculation, and for earlier diagnosis of PD. One category of risk factors emerging in recent literature relates to war-related illness, such as traumatic brain injury (TBI) and posttraumatic stress disorder (PTSD). Both disorders have been recently implicated as PD risk factors, but their association with development of PD, timing relative to PD onset, and independence from other risk factors remains unknown.

TBI is the better studied of these two risk factors, and a subset of human studies have implicated TBI as an acquired PD risk factor.1,2 Mechanistic evidence in animal model systems and humans has supported this link by showing accumulation of pathologic proteins, including α-synuclein, after brain injury.3,4 Human studies have also linked brain injury related to repetitive head injury to Lewy body disease.5 However, some studies have suggested a reverse causation model where TBIs are the result of undiagnosed early PD6 (eg, due to postural instability). Although there is still lingering debate, recent meta-analyses and those studies with a longer lag time support TBI to be a true risk factor for PD.1,2 Moreover, association studies suggest the impact of TBI on PD risk is far greater in those who sustain early-in-life TBI, considered before the age of 30 years.7 These data highlight the importance of TBI as a PD risk factor but also the need for increased study duration and details on timing to disentangle associations. The lifelong time course of TBI-conferred risk and its interaction with comorbid disorders remains unknown.

The link between PTSD and PD has only recently been identified. Mechanistically, animal studies have linked chronic stress to aggregation of toxic proteins (eg, synuclein) and neurodegeneration.8 Epidemiological studies by our group and others have found an association between PTSD and REM sleep behavior disorder (RBD),9,10 and two larger studies (including one prospective study) reported direct associations between PTSD and PD.11,12 However, the reported associations of PTSD with PD were not controlled for important confounders (eg, smoking) and had shorter lag times susceptible to potential reverse causation.

Interactions between trauma-related disorders are unknown and could potentially cause more-than-additive (ie, synergistic) effects on increasing PD risk. This is particularly important to study for co-occurring disorders with similar causes that could confer any manner of positive or negative impact on PD risk. TBI and PTSD are two such disorders that often coexist as hallmarks of chronic neurobehavioral symptomatology along with other comorbid sleep, pain, mood/anxiety, and cognitive disturbances. Because some of these comorbid disorders are potentially treatable, it is important to define their interaction with TBI, PTSD, and later PD.

Overall, we hypothesized that prevalence of TBI and PTSD would be greater in individuals who later developed PD, compared with matched control cases who did not develop PD. Furthermore, we hypothesized that TBI and PTSD would interact in a synergistic manner to increase PD risk. To address these questions, we performed a retrospective case–control study within the U.S. Veterans Health Administration, a population enriched for early trauma, TBI, PTSD, and PD.13

Subjects and Methods

Research Design, Data Source, and Institutional Approval

A retrospective case–control study design was used to measure the association of TBI and PTSD with later development of PD. Data were obtained from the Veterans Health Administration Corporate Data Warehouse (CDW) under institution review board approval (MIRB #04744) using a waiver of participant consent. All time periods available within the CDW were included in the study. Dated International Classification of Diseases (ICD) codes in the CDW began in 1999 as ICD digitization was implemented at the U.S. Department of Veterans Affairs (VA) hospitals between 1999 and 2005.

Case Definition and Validation

PD was defined by (1) age older than 40 years, (2) the presence of one or more PD ICD-9 (332.0) or ICD-10 (G20) codes, (3) having at least 5 years of ICD-coded medical records before the first PD ICD code, and (4) having two PD medication prescriptions filled. PD medication was defined by the VA nationwide common data model for operations and research (Observational Medical Outcomes Partnership)14 (Supporting Information Table S1 in Data S1). To increase specificity, we created a secondary definition. In addition to meeting all previous criteria, these cases required affirmative mention of PD in at least two VA neurology notes, separated by a minimum of 1 year in time and starting at least 1 year after the first PD ICD code. Thus, each case would have a minimum of 2 years of free-text records referring to PD as a chief complaint. Regular expressions and iterative training were used to eliminate negations and ambiguous statements. Both definitions were validated against manual chart review as a gold standard in 150 randomly selected cases. The index date was defined as the date of the first PD ICD code in cases.

Control Selection

Four control subjects were chosen for each PD case by randomized selection and matching by sex, ethnicity, race, smoking status,15 and birth year. Controls were also selected for those with at least the same number of years of historic medical records before the index date as their matched case. The index date for controls was the same age (month and year) as the time of first coded PD diagnosis in their matched case.16,17 A total of 21,882,307 veterans were potential controls at the time of writing this manuscript.

Variables Definition

Extraction of race, ethnicity, sex, and birth year was from their own distinct fields in the CDW (see Supporting Information Methods for field names), and smoking information was categorized using a crosswalk method validated previously at the VA15(p20). Missing data were coded as unknown rather than performing imputation, and assuming missingness was random.

Comorbid disorders were defined by diagnostic codes (ICD-9, ICD-10) that were collected for each participant within 20 years of the index date. ICD codes for each disorder category are listed in Supporting Information Table S2 in Data S1. For TBI and PTSD in this military veteran cohort, person-specific onset was defined as the midpoint of active duty. Only TBI and PTSD with onset 5+ years before the index date is included to mitigate reverse causation. We do not use dates of ICD codes because VA records begin years to decades after active duty and would severely underestimate onset time of trauma exposures. To parallel PD case ascertainment, we created a secondary definition for TBI and PTSD to improve specificity, further enrich for true military-related TBI and PTSD, and further eliminate cases caused by nonmilitary factors in the veteran population. This required that these diagnoses are “service connected,” that is, explicitly indicated in the record to be connected to the participant’s active military service. We suspect usage and data completeness of the “service connected” field was low. For other disorders, the first instance of a relevant ICD code defined its onset. The possibility of multiple distinct episodes of the same disorder is not captured because the majority of repeated codes are related to the initial diagnosis.18

Statistical Analyses

For TBI, PTSD, and other clinical disorders defined by ICD codes, yearly prevalence was calculated before the index date. Association of each disorder with PD was calculated using conditional logistic regression adjusted for race, sex, smoking status, ethnicity, and birth year. Odds ratios (ORs) and 95% confidence intervals (CIs) are shown in the figures at 5-year and yearly epochs. The P values were adjusted for multiple comparisons using Bonferroni correction, and significant results were defined as P value less than 5 × 10−2. Analysis was conducted using RStudio (RStudio Team, 2021; RStudio: Integrated Development for R. Boston: RStudio, PBC; http://www.rstudio.com/).

Interaction Analysis

Interaction was measured using synergy index (SI), a recommended metric for case–control studies.19 Indices and CIs were quantified using the interactionR package in R studio.20 An SI greater than 1 indicates a positive interaction or greater than additivity.

Data Sharing

The VA legally restricts access to data that include sensitive and identifying patient information. The data used for this study are not permitted to leave the VA firewall without a Data Use Agreement. Behind the firewall, data are made available with an approved study protocol.

Results

Cohort Characteristics

A total of 71,933 patients fulfilled the primary PD case definition, and 287,732 controls (matched 4:1) were identified (Fig. 1, left side). The cohort was predominantly male (98.4%), white (81.4%), non-Hispanic (88.3%), and elderly (mean age at index date: 74.8 ± 8.9 years old) (Table 1). For smoking status, the cohorts’ proportions were split between former smokers (39.9%) and never smokers (37.1%) with a lower proportion of current smokers (20%) and unknown (2.9%). Manual chart review was performed for 150 randomly selected subjects (Fig. 1, right side), and 118 true positives were identified for a positive predictive value (PPV) of 78.6%. The most common specific types of false positives were drug-induced parkinsonism (n = 7) and essential tremor (n = 4).

FIG. 1.

FIG. 1.

Open in a new tab

Parkinson’s disease (PD) case definitions. Flowchart (left) shows stepwise criteria and case counts for primary and secondary PD case definitions. Definitions were validated (right) by manual review of medical records. FP, false positive; ICD, International Classification of Diseases; PPV, positive predictive value; TP, true positive.

TABLE 1.

Cohort characteristics used for matching

Characteristics Parkinson’s Disease Controls
N 71,933 287,732
Age, years 74.8 (±8.9) 74.8 (±8.9)
Birth year 1938.3 (±10.4) 1938.3 (±10.4)
Sex, n (%)
 Female 1118 (1.6%) 4472 (1.6%)
 Male 70,815 (98.4%) 283,260 (98.4%)
Duration of VA health care usage 11.1 (±4.5) 11.3 (±4.5)
Smoking status, n (%)
 Current 14,414 (20%) 57,656 (20%)
 Former 28,724 (39.9%) 114,896 (39.9%)
 Never 26,712 (37.1%) 106,848 (37.1%)
 Unknown 2083 (2.9%) 8332 (2.9%)
Race, n (%)
 African American/Black 4795 (6.7%) 19,180 (6.7%)
 Asian, Native American, Alaskan Native, Native Hawaiian, Pacific Islander 1276 (1.6%) 5104 (1.6%)
 White 58,569 (81.4%) 234,276 (81.4%)
 Unknown 7293 (10.1%) 29,172 (10.1%)
Ethnicity, n (%)
 Hispanic, Latino 3371 (4.7%) 13,484 (4.7%)
 Not Hispanic, not Latino 63,551 (88.3%) 254,204 (88.3%)
 Unknown 5011 (6.97%) 20,044 (6.97%)
History of traumatic brain injury, n (%total with data)
 Years before index date
  10 years 4171 (5.9%) 8403 (2.9%)
  20 years 4045 (5.8%) 8159 (2.9%)
  30 years 3825 (5.7%) 7747 (2.9%)
  40 years 3216 (5.3%) 6501 (2.7%)
  50 years 2044 (4.7%) 4284 (2.5%)
  60 years 645 (3.5%) 1320 (1.8%)
History of posttraumatic stress disorder, n (% total with data)
 Years before index date
  10 years 14,907 (21.1%) 37,330 (13.2%)
  20 years 14,506 (20.9%) 36,315 (13.1%)
  30 years 13,726 (20.5%) 34,373 (12.8%)
  40 years 11,441 (19.0%) 28,671 (11.9%)
  50 years 6361 (14.8%) 15,851 (9.2%)
  60 years 1333 (7.3%) 3350 (4.6%)
Open in a new tab

Abbreviation: VA, U.S. Department of Veterans Affairs.

In an attempt to improve specificity for secondary analyses, we next identified 13,237 veterans fulfilling the secondary PD definition, requiring affirmative mention of PD in at least two VA neurology notes. Validation by chart review identified 136 true positives and 14 false positives, yielding an improved PPV of 90.0%.

TBI and PTSD

Prevalence of TBI and PTSD was calculated at 5-year time epochs across a 60-year duration before the index date of PD diagnosis (Fig. 2A,B). Raw numbers of cases and controls with TBI and PTSD are also shown at 10-year epochs in Figure 2A,B and Table 1. TBI had an average prevalence of 2.6% ± 0.3% in controls and 5.3% ± 0.7% in cases who were later diagnosed with PD. PTSD prevalence was, on average, 11% ± 2.9% in controls and 17% ± 4.7% in cases. Association analysis was performed, and both TBI and PTSD had significantly increased odds of PD at each time epoch extending back 60 years. The association between TBI and PD was calculated at 5-year time epochs and ranged from OR of 2.0 [1.8, 2.1] to 2.1 [2.0, 2.1]. The association between PTSD and PD across the same time period ranged from OR of 1.5 [1.4, 1.7] to 1.9 [1.9, 2.0].

FIG. 2.

FIG. 2.

Open in a new tab

Association of trauma-related disorders with Parkinson’s disease (PD). (A) Association of traumatic brain injury (TBI) with subsequent PD. (B) Association of posttraumatic stress disorder (PTSD) with subsequent PD. Bottom panels: interaction of TBI and PTSD in association with subsequent PD. (C) Interaction of TBI and PTSD in influencing odds of subsequent PD. (D) Additive association of TBI plus PTSD on later PD diagnosis. For association analyses, odds ratios (black circles) and 95% confidence intervals (gray) shown at 5-year intervals beginning at midpoint of active military duty and extending up to 5 years before the index date (year 0, date of PD diagnosis). Tables show number of cases and controls with (+) and without (−) trauma-related disorders. For interaction analysis, synergy index (small black circles) and 95% confidence intervals (light gray) shown at 5-year intervals beginning at midpoint of active military duty and extending up to 5 years before the index date. Ctrl, control.

Analysis was repeated using the secondary PD definition (Supporting Information Fig. S1 in Data S1), with similar patterns emerging: mean TBI prevalence was 2.4% ± 0.3% in controls and 4.3% ± 0.4% in cases; mean PTSD prevalence was 13% ± 4.2% in controls and 22% ± 0.6% in cases. Moreover, association analyses revealed significantly increased odds of PD at each epoch extending back 60 years for both TBI (range of OR: 1.7 [1.6, 1.9] to 2.4 [1.5,3.9]) and PTSD (range of OR: 1.9 [1.9, 2.0] to 2.1 [1.7, 2.6]).

Analysis was again repeated on the primary PD definition after limiting TBI and PTSD diagnoses to those where a “service connected” field had been checked (Supporting Information Fig. S2 in Data S1), to improve specificity of military-related trauma diagnoses. Notably, case counts and prevalence for TBI dropped two orders of magnitude. Nevertheless, the same significant pattern emerged: TBI had an average prevalence of 0.05% ± 0.01% in controls compared with 0.09% ± 0.03% in cases, and PTSD prevalence was 4.5% ± 1.2% in controls versus 7.5% ± 2.0% in cases. Significant association for TBI and PD began at year −45 and ranged from OR of 1.5 [1.03, 2.1] to 2.0 [1.5, 2.6]. The association between PTSD and PD was significant across 60 years and ranged from OR of 1.7 [1.5, 1.9] to 1.8 [1.7, 1.8].

Interaction and Combined PD Association of TBI and PTSD

Because TBI and PTSD are commonly co-occurring conditions, we next aimed to explore whether their combination influenced PD risk beyond additivity. Interaction analysis was performed between TBI and PTSD (Fig. 2C). SI and 95% CIs were measured at 5-year intervals across 60 years before PD diagnosis. SI values were significantly above additivity up to year −50 at all but one time epoch. Values significantly above additivity ranged from SI of 1.14 [1.09, 1.29] to 1.28 [1.09, 1.51]. The one epoch that did not reach statistical significance was at year −30 with SI of 1.11 [0.98, 1.27].

Prevalence of individuals with both TBI and PTSD was calculated at 5-year time epochs across a 60-year duration before the index date/PD diagnosis (Fig. 2D). Raw numbers of cases and controls with both TBI and PTSD are shown at 10-year epochs in tables of Figure 2D. TBI + PTSD had a greater numerical difference between PD and controls when combined versus when compared with either disorder separately. TBI + PTSD had an average prevalence of 0.8% ± 0.2% in controls and 2.1% ± 0.5% in cases who were later diagnosed with PD. Association analysis was performed and having both TBI and PTSD significantly increased odds of subsequent PD at each time epoch extending back 60 years. This association was calculated at 5-year time epochs and ranged from OR of 2.2 [1.6, 2.8] to 2.7 [2.5, 2.8].

Interaction of TBI and PTSD with Other Disorders

Because TBI and PTSD are comorbid with other pain, sleep, cognitive, mood, and prodromal disorders, we performed interaction analysis between combinations of TBI and PTSD with these disorders (Table 2). SI was calculated at 1-year time intervals across 10 years from year −1 to year −11. The disorders with the largest significant positive interaction for both PTSD and TBI were migraine with average SI = 1.61 ± 0.33, chronic pain with average SI = 1.55 ± 0.34, and sleep apnea with average SI = 1.40 ± 0.19.

TABLE 2.

Disorders synergistic with PTSD and TBI in increasing odds of later PD

Disorders Synergy Index, Mean (SD)
Positive interaction with both PTSD and TBI
 Migraine 1.61 (±0.33)
 Chronic pain 1.55 (±0.34)
 Sleep apnea 1.40 (±0.19)
Positive interaction with just PTSD
 Cognitive impairment 1.55 (±0.07)
 Constipation 1.47 (±0.54)
 Orthostatic hypotension 1.47 (±0.42)
Positive interaction with just TBI
 Depression 1.46 (±0.05)
 Anxiety disorder 1.34 (±0.15)
Open in a new tab

Note: Greater than additivity is indicated by values >1 for synergy index. Values shown are significant, P < 0.05.

Abbreviations: PTSD, posttraumatic stress disorder; TBI, traumatic brain injury; PD, Parkinson’s disease; SD, standard deviation.

PD association analysis was performed for individuals with TBI, PTSD, and either chronic pain or migraine (Fig. 3). The yearly prevalence for participants with chronic pain, TBI, and PTSD was, on average, 0.14% ± 0.08% in controls and 0.45% ± 0.24% in cases, and PD association ranged from OR of 2.9 [2.6, 3.4] to 3.3 [2.2, 5.0]. The yearly prevalence of migraine, TBI, and PTSD was, on average, 0.14% ± 0.07% in controls and 0.40% ± 0.03% in cases, and PD association ranged from OR of 2.7 [2.1, 3.4] to 3.1 [2.7, 3.6]. Results were compared with International Parkinson and Movement Disorder Society (MDS) prodromal disorders (Fig. 3).

FIG. 3.

FIG. 3.

Open in a new tab

Synergistic disorders related to military trauma and also Movement Disorder Society (MDS) prodromal disorders associated with Parkinson’s disease (PD). Significant odds ratios (ORs; P < 0.05, Bonferroni’s correction) are shown for trauma-related disorders (circles) and MDS prodromal disorders (triangles) at 1-year intervals beginning at the first appearance of the relevant International Classification of Diseases code and extending up to 1 year before the index date (year 0, date of PD diagnosis). REM behavior sleep disorder is not shown because it demonstrated a larger effect size ranging from ORs of 5.4 [3.0, 9.8] to 8.5 [7.3, 9.9]. PTSD, posttraumatic stress disorder; TBI, traumatic stress disorder.

Discussion

We leveraged a nationwide electronic record of patients enriched for military trauma and show that TBI and PTSD are associated with later PD. Strengths of this study include improving the specificity of PD diagnosis, using active duty timing information for early-in-life neurotrauma, and testing comorbid disorder combinations. We also showed that TBI and PTSD are synergistic with chronic pain and migraine, and that synergy of TBI/PTSD with chronic pain and migraine further increased odds of PD. These findings provide evidence for TBI and PTSD as risk factors preceding PD by decades and could aid in prognostic calculation, earlier diagnosis, and earlier intervention for PD.

Association Between TBI and PD

Data confirm TBI as a significant early-in-life risk factor for PD. Notably, the average age during active duty is 28 years old.21 The current effect sizes and ages match those of a smaller nonmilitary study that reported an effect size of 2.3 (CIs not available) in those reporting head injury in their teens to twenties approximately 40 to 50 years before development of PD.7 The effect size is also similar to shorter-duration studies of VA participants (1.7 [1.5, 1.9])2 and slightly greater than pooled results of 15 other studies with shorter lag times (1.4 [1.1, 1.7]).7,22,23 In addition to extending on human epidemiologic data, this study corroborates chronic pathologies reported by animal studies of neurotrauma.24–28 These studies have catalogued long-term phenomena, including progressive dopaminergic cell loss, decreased tyrosine hydroxylase expression, and increased α-synuclein accumulation in surviving dopaminergic neurons. Interestingly, these studies also found nonneuronal and acellular aberrancies, including increased microglial inflammation and dysregulation of soluble α-synuclein, in some cases persisting into the chronic phase of recovery. Additional studies are needed to search in human tissues for pathologies of lifelong trauma-related PD risk.

Association Between PTSD and PD

Results also show PTSD is associated with increased odds of subsequent PD and is not just a surrogate for TBI-associated PD risk. Our data are similar to the association between PTSD and PD previously reported in a large retrospective study11 and multiple other studies, within and outside the VA.2,11,12,29 Mechanistically, animal studies have found a link between chronic stress, aggregation of toxic proteins (eg, synuclein), and neurodegeneration.8 Given these findings, translational studies of PTSD pathology in human biopsy and autopsy tissues are needed to pair with longitudinal clinical data.

Interaction between TBI, PTSD, and PD

The combination of TBI and PTSD further increased odds of PD beyond either factor alone. These results are not altogether surprising, because we and others have shown that the combination of TBI and PTSD in veterans synergizes to produce higher neuropsychiatric symptom burden,10,30,31 more profound disability,31 and higher rates of RBD.10 Potential mechanisms underlying the additive effects of comorbid TBI and PTSD are still unclear, but animal models suggest that neurodegeneration is accelerated by chronic stress and sleep disruption, two disorders increased in TBI and PTSD.1 Importantly, these data do not support the view that PTSD and TBI confer redundant risk and are not worth considering independently in PD subtyping and neuroprotection efforts.

Synergy Between TBI, PTSD, and Other Comorbid Disorders

A novel approach in this study was to explicitly examine comorbid disorders and screen for high-yield interactions. Interestingly, greater than additive association of developing PD in those with TBI and PTSD was strongest for chronic pain and migraine. The triad of TBI, PTSD, and chronic pain has been termed the “polytrauma clinical triad,” a syndrome recognized for its association with disability and poor outcomes.31 Our results suggest that veterans with the polytrauma clinical triad may also be at higher risk for neurodegenerative disorders. Additional human studies with more detailed phenotyping of pain and migraine will help identify what makes these potent mixtures with TBI and PTSD.

Limitations

Limitations of this study include those associated with retrospective case–control studies reliant on diagnostic coding. In general, using ICD codes for other disorders typically represents a “possible” or “probable” level of certainty depending on usage and classification criteria. Previous studies measuring the test performance of ICD codes at the VA report 75.6% PPV for PD at a single VA movement disorder specialty center,18 82% PPV and 84% sensitivity for PTSD across the VA,32 and 84.6% PPV for TBI at the VA, but with low sensitivity of 55.4%.33 Note that the PPV result for PD ICD codes is inflated by restricting to one specialty center, and in studies using all medical providers, PPV is 46% and sensitivity is 80%.34 Our PD case definitions were applied across the entire VA system and not just a specialty center, yet yielded higher PPVs of 78.6% and 90.0%, depending on our case definition. We suspect miscoding would only have lowered the observed effect sizes. For example, miscoding PD as drug-induced parkinsonism, the most common false positive in our study, is less likely to have the same prediagnosis disorders 10 years earlier. Capturing additional certainty in this study would have required a prohibitive amount of extraction of nonstandardized data from terabytes of free-text medical notes. We also suspect ICD insensitivity underlies low case ascertainment of entities such as RBD in this study. The prevalence of RBD in this study is far lower than previous studies in both PD and control groups, and we suspect it underpowers interaction analysis with PTSD and TBI. A final limitation of ICD codes is misdiagnosis, which can skew results. Misdiagnosis will be mitigated in future studies by newer tools, such as the validated questionnaire and the Comprehensive TBI Evaluation program.2

An additional limitation includes the use of birth year as an adjustment variable, which could be skewed toward Vietnam-era veterans who may have had exposure to Agent Orange. Although officially recognized by the VA in service connection to PD, at the time of this writing, Agent Orange has still not been proven as an established risk factor for PD and therefore was not included as an adjustment variable.

Another important limitation for generalizing findings is underrepresentation of non-White races, Hispanic ethnicity, and female sex. Underrepresentation of female sex, for example, may skew disorders that covary by sex, such as PTSD, depression, anxiety, headache, and migraine.35–38 Underrepresentation of race/ethnicity may falsely inflate sensitivity of diagnoses such as TBI where VA studies have shown underdiagnosis among Black, Hispanic, and Asian service members compared with White service members.39

A final limitation of a retrospective case–control design is disentangling the PD risk contribution of two co-occurring disorders. It is suggestive, albeit inconclusive, when two disorders (eg, PTSD and cognitive impairment) have synergistic effects on developing PD. This study cannot resolve independent risk and relies on studies with prospective design and more detailed patient-specific onset timing to determine independent risk of, for example, PTSD versus cognitive impairment. Accurate risk estimation is one reason MDS research criteria require prospective studies when designing a prodromal risk calculator.40 This limitation applies in general to other co-occurring diseases where, for example, TBI and PTSD and PD each may directly cause problems of sleep, cognition, pain, and anxiety/mood.

Conclusions

Overall, these data found a lifelong association of trauma-related disorders with subsequent PD. Interaction analysis uncovered synergism with comorbid chronic pain and migraine. These effects should help guide intervention trials and studies into distinct PD subtypes. We hope our results expand the search for putative acquired and preventable insults41 and jump-start earlier prodromal screening tools,16 which will be particularly impactful once effective neuroprotective therapy is available. Answering the questions of when and what disorders make PD more likely will be critical to understanding pathogenesis and spur innovative thinking about novel targets and optimal timing of potential interventions.

Supplementary Material

1

Additional Supporting Information may be found in the online version of this article at the publisher’s web-site.

Acknowledgments:

We thank Priya Srikanth, MPH, Biostatistics program at Oregon Health and Science University, for biostatistics help. We also thank the OHSU School of Medicine and Department of Pathology and the Parkinson’s Center of Oregon. This work was also supported by VA Biomedical Laboratory Research and Development Career Development Award 1 IK2 BX005760–01A1 (to G.D.S.), VA Clinical Science Research and Development Career Development Award IK2 CX00253 (to L.E.N.), and VA Clinical Science Research and Development Merit Award I01 CX002022 (to M.M.L.).

Funding agencies:

This work was supported by VA Biomedical Laboratory Research and Development Career Development Award 1 IK2 BX005760–01A1 (to G.D.S.), VA Clinical Science Research and Development Career Development Award IK2 CX00253 (to L.E.N.), and VA Clinical Science Research and Development Merit Award I01 CX002022 (to M.M.L.). This work was also supported by the Office of Academic Affiliations, Advanced Fellowship Program in Mental Illness Research and Treatment, Department of Veterans Affairs (L.E.N.). L.E.N. and J.F.Q. are also supported by the VA Parkinson’s Disease Research, Education and Clinical Centers, J.F.Q., M.M.L. and R.W. by the Oregon Alzheimer’s Disease Research Center (Grant P30 AG066518), and M.M.L. and L.E.N. by the VA VISN20 Northwest Mental Illness Research Education and Clinical Center. All authors report no disclosures relevant to this manuscript.

Financial Disclosures (for past 12 Months)

GS receives grant support from VA Biomedical Laboratory Research and Development Career Development Award Number 1 IK2 BX005760–01A1, John and Tami Marick Foundation, the Oregon Medical Research Foundation, the Oregon Health and Science University Parkinsons Center of Oregon, and the Collins Medical Trust. LEN receives grant support from the VA Clinical Science Research and Development Career Development Award IK2 CX00253. MML receives support from VA Clinical Science Research and Development Merit Award I01 CX002022. In addition, this work was supported by the Office of Academic Affiliations, Advanced Fellowship Program in Mental Illness Research and Treatment, Department of Veteran Affairs to LEN. LEN and JFQ are also supported by the VA Parkinson’s Disease Research, Education and Clinical Centers (PADRECC), JFQ and MML by the Oregon Alzheimer’s Disease Research Center P30 AG066518, MML and LEN by the VA VISN20 Northwest MIRECC. RW has received support from the following grants, NIH (RO1 AG056712, R01 DA047237, P30 AG08017, R01 AG0380651, R56 AG064016, TL4GM118965). JQ also reports one-time advisory board membership for Biogen, clinical trial reimbursement from Alkahest, Sanofi, Abbvie, Ascend, Neuraly, Sun-Pharma, and served on data-service monitoring board for VTV pharmaceuticals, Retrophin pharmaceuticals, Alzheimer’s Disease Cooperative Study, Alzheimer’s Therapeutic Research Institute.

Footnotes

Relevant conflicts of interest/financial disclosures: Nothing to report.

Data Availability Statement

The United States Veterans Affairs legally restricts access to Veteran health care data, which includes sensitive and identifying patient information. The data used for this study are not permitted to leave the Veterans Affairs firewall without a Data Use Agreement. However, behind the firewall, data are made available with an approved study protocol. This limitation is consistent with similar studies from the Veterans Affairs. For more information, please visit https://www.virec.research.va.gov.

References

  • 1.Jafari S, Etminan M, Aminzadeh F, Samii A. Head injury and risk of Parkinson disease: a systematic review and meta-analysis. Mov Disord 2013;28(9):1222–1229. 10.1002/mds.25458 [DOI] [PubMed] [Google Scholar]
  • 2.Gardner RC, Byers AL, Barnes DE, Li Y, Boscardin J, Yaffe K. Mild TBI and risk of Parkinson disease: a chronic effects of neurotrauma consortium study. Neurology 2018;90(20):e1771–e1779. 10.1212/WNL.0000000000005522 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Uryu K, Giasson BI, Longhi L, et al. Age-dependent synuclein pathology following traumatic brain injury in mice. Exp Neurol 2003; 184(1):214–224. 10.1016/s0014-4886(03)00245-0 [DOI] [PubMed] [Google Scholar]
  • 4.Uryu K, Chen XH, Martinez D, et al. Multiple proteins implicated in neurodegenerative diseases accumulate in axons after brain trauma in humans. Exp Neurol. 2007;208(2):185–192. 10.1016/j.expneurol.2007.06.018 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Adams JW, Alosco ML, Mez J, et al. Association of probable REM sleep behavior disorder with pathology and years of contact sports play in chronic traumatic encephalopathy. Acta Neuropathol 2020; 140(6):851–862. 10.1007/s00401-020-02206-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Ascherio A, Schwarzschild MA. The epidemiology of Parkinson’s disease: risk factors and prevention. Lancet Neurol. 2016;15(12): 1257–1272. 10.1016/S1474-4422(16)30230-7 [DOI] [PubMed] [Google Scholar]
  • 7.Taylor KM, Saint-Hilaire MH, Sudarsky L, et al. Head injury at early ages is associated with risk of Parkinson’s disease. Parkinsonism Relat Disord. 2016;23:57–61. 10.1016/j.parkreldis.2015.12.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Hemmerle AM, Herman JP, Seroogy KB. Stress, depression and Parkinson’s disease. Exp Neurol. 2012;233(1):79–86. 10.1016/j.expneurol.2011.09.035 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Yao C, Fereshtehnejad SM, Keezer MR, Wolfson C, Pelletier A, Postuma RB. Risk factors for possible REM sleep behavior disorder: a CLSA population-based cohort study. Neurology 2018;92(5): e475–e485. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Elliott JE, Opel RA, Pleshakov D, et al. Posttraumatic stress disorder increases the odds of REM sleep behavior disorder and other parasomnias in veterans with and without comorbid traumatic brain injury. Sleep. 2020;43(3):zsz237. 10.1093/sleep/zsz237 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.White DL, Kunik ME, Yu H, et al. Post-traumatic stress disorder is associated with further increased Parkinson’s disease risk in veterans with traumatic brain injury. Ann Neurol. 2020;88(1):33–41. 10.1002/ana.25726 [DOI] [PubMed] [Google Scholar]
  • 12.Chan YLE, Bai YM, Hsu JW, et al. Post-traumatic stress disorder and risk of Parkinson disease: a Nationwide longitudinal study. Am J Geriatr Psychiatry 2017;25(8):917–923. 10.1016/j.jagp.2017.03.012 [DOI] [PubMed] [Google Scholar]
  • 13.Ning WL, Ping TJ, Ge XH, et al. A cross-sectional study of neurological disease in the veterans of military communities in Beijing. Zhonghua Nei Ke Za Zhi. 2010;49(6):463–468. [PubMed] [Google Scholar]
  • 14.Lynch KE, Deppen SA, DuVall SL, et al. Incrementally transforming electronic medical records into the observational medical outcomes partnership common data model: a multidimensional quality assurance approach. Appl Clin Inform. 2019;10(5):794–803. 10.1055/s-0039-1,697,598 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Golden SE, Hooker ER, Shull S, et al. Validity of veterans health administration structured data to determine accurate smoking status. Health Informatics J. 2020;26(3):1507–1515. 10.1177/1460458219882259 [DOI] [PubMed] [Google Scholar]
  • 16.Schrag A, Anastasiou Z, Ambler G, Noyce A, Walters K. Predicting diagnosis of Parkinson’s disease: a risk algorithm based on primary care presentations. Mov Disord 2019;34(4):480–486. 10.1002/mds.27616 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Scott GD, Lim MM, Drake MG, Woltjer R, Quinn JF. Onset of skin, gut, and genitourinary prodromal Parkinson’s disease: a study of 1.5 million veterans. Mov Disord 2021;36(9):2094–2103. 10.1002/mds.28636 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Szumski NR, Cheng EM. Optimizing algorithms to identify Parkinson’s disease cases within an administrative database. Mov Disord 2009;24(1):51–56. 10.1002/mds.22283 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Skrondal A Interaction as departure from additivity in case–control studies: a cautionary note. Am J Epidemiol. 2003;158(3):251–258. 10.1093/aje/kwg113 [DOI] [PubMed] [Google Scholar]
  • 20.Alli BY. InteractionR: an R package for full reporting of effect modification and interaction. Softw Impacts. 2021;10:100147. 10.1016/j.simpa.2021.100147 [DOI] [Google Scholar]
  • 21.Department of Defense. 2020 Demographics Profile of the Military Community; 2020. Accessed January 2, 2023. https://download.militaryonesource.mil/12038/MOS/Reports/2020-demographics-report.pdf.
  • 22.Perry DC, Sturm VE, Peterson MJ, et al. Association of traumatic brain injury with subsequent neurological and psychiatric disease: a meta-analysis. J Neurosurg. 2016;124(2):511–526. 10.3171/2015.2.JNS14503 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Gao J, Liu R, Zhao E, et al. Head injury, potential interaction with genes, and risk for Parkinson’s disease. Parkinsonism Relat Disord. 2015;21(3):292–296. 10.1016/j.parkreldis.2014.12.033 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Acosta SA, Tajiri N, de la Pena I, et al. Alpha-synuclein as a pathological link between chronic traumatic brain injury and Parkinson’s disease. J Cell Physiol. 2015;230(5):1024–1032. 10.1002/jcp.24830 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Carlson SW, Henchir J, Dixon CE. Lateral fluid percussion injury impairs hippocampal synaptic soluble N-ethylmaleimide sensitive factor attachment protein receptor complex formation. Front Neurol. 2017;8:532. 10.3389/fneur.2017.00532 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Delic V, Beck KD, Pang KCH, Citron BA. Biological links between traumatic brain injury and Parkinson’s disease. Acta Neuropathol Commun. 2020;8(1):1–16. 10.1186/s40478-020-00924-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Hutson CB, Lazo CR, Mortazavi F, Giza CC, Hovda D, Chesselet MF. Traumatic brain injury in adult rats causes progressive nigrostriatal dopaminergic cell loss and enhanced vulnerability to the pesticide paraquat. J Neurotrauma. 2011;28(9):1783–1801. 10.1089/neu.2010.1723 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Mallah K, Quanico J, Raffo-Romero A, et al. Mapping spatiotemporal microproteomics landscape in experimental model of traumatic brain injury unveils a link to Parkinson’s disease*. Mol Cell Proteomics. 2019;18(8):1669–1682. 10.1074/mcp.RA119.001604 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Gardner RC, Burke JF, Nettiksimmons J, Goldman S, Tanner CM, Yaffe K. Traumatic brain injury in later life increases risk for Parkinson disease. Ann Neurol. 2015;77(6):987–995. 10.1002/ana.24396 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Balba NM, Elliott JE, Weymann KB, et al. Increased sleep disturbances and pain in veterans with comorbid traumatic brain injury and posttraumatic stress disorder. J Clin Sleep Med 2018;14(11): 1865–1878. 10.5664/jcsm.7482 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Lew HL, Otis JD, Tun C, Kerns RD, Clark ME, Cifu DX. Prevalence of chronic pain, posttraumatic stress disorder, and persistent postconcussive symptoms in OIF/OEF veterans: polytrauma clinical triad. J Rehabil Res Dev. 2009;46(6):697–702. 10.1682/jrrd.2009.01.0006 [DOI] [PubMed] [Google Scholar]
  • 32.Harrington KM, Quaden R, Stein MB, et al. Validation of an electronic medical record–based algorithm for identifying posttraumatic stress disorder in U.S. Veterans. J Trauma Stress 2019;32(2):226–237. 10.1002/jts.22399 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Carlson KF, Barnes JE, Hagel EM, Taylor BC, Cifu DX, Sayer NA. Sensitivity and specificity of traumatic brain injury diagnosis codes in United States Department of Veterans Affairs administrative data. Brain Inj. 2013;27(6):640–650. 10.3109/02699052.2013.771795 [DOI] [PubMed] [Google Scholar]
  • 34.Peterson BJ, Rocca WA, Bower JH, Savica R, Mielke MM. Identifying incident Parkinson’s disease using administrative diagnostic codes: a validation study. Clin Park Relat Disord. 2020;3:100061. 10.1016/j.prdoa.2020.100061 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Scher AI, Ross GW, Sigurdsson S, et al. Midlife migraine and late-life parkinsonism: AGES-Reykjavik study. Neurology. 2014;83(14): 1246–1252. 10.1212/WNL.0000000000000840 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Stein MB, Walker JR, Forde DR. Gender differences in susceptibility to posttraumatic stress disorder. Behav Res Ther. 2000;38(6):619–628. 10.1016/s0005-7967(99)00098-4 [DOI] [PubMed] [Google Scholar]
  • 37.Miller IN, Cronin-Golomb A. Gender differences in Parkinson’s disease: clinical characteristics and cognition. Mov Disord 2010; 25(16):2695–2703. 10.1002/mds.23388 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Lorentz IT. A survey of headache in Parkinson’s disease. Cephalalgia 1989;9(2):83–86. 10.1046/j.1468-2982.1989.0902083.x [DOI] [PubMed] [Google Scholar]
  • 39.Kysar-Moon A, Mustillo S. Race/ethnicity and traumatic brain injury: are there disparities in positive screenings and diagnoses among service members returning from Afghanistan and Iraq? Armed Forces Soc. 2019;45(1):155–176. 10.1177/0095327X18761852 [DOI] [Google Scholar]
  • 40.Berg D, Postuma RB, Adler CH, et al. MDS research criteria for prodromal Parkinson’s disease. Mov Disord 2015;30(12):1600–1611. 10.1002/mds.26431 [DOI] [PubMed] [Google Scholar]
  • 41.Langston JW, Forno LS, Tetrud J, Reeves AG, Kaplan JA, Karluk D. Evidence of active nerve cell degeneration in the substantia nigra of humans years after 1-methyl-4-phenyl-1,−2,3,6-tetrahydropyridine exposure. Ann Neurol. 1999;46(4):598–605. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS2144142-supplement-1.docx (453.5KB, docx)

Data Availability Statement

The United States Veterans Affairs legally restricts access to Veteran health care data, which includes sensitive and identifying patient information. The data used for this study are not permitted to leave the Veterans Affairs firewall without a Data Use Agreement. However, behind the firewall, data are made available with an approved study protocol. This limitation is consistent with similar studies from the Veterans Affairs. For more information, please visit https://www.virec.research.va.gov.

RESOURCES