|Home | About | Journals | Submit | Contact Us | Français|
There is a complex relationship between posttraumatic stress disorder and traumatic brain injury. To understand and treat these conditions, it is necessary to apply an integrated physical and mental health care approach to postdeployment care.
Traumatic brain injury (TBI) is frequently referred to as the “signature wound” of the Iraq and Afghanistan wars. Previous studies estimate that 10% to 20% of U.S. veterans who served in these battles experienced mild to moderate TBI primarily related to blast exposure.1
Research studies conducted in the last decade reveal complex relationships between deployment-related factors and overlapping physical injury and psychiatric outcomes.2 To understand and treat these conditions, it is necessary to apply an integrated physical and mental health care approach to postdeployment care.
Under the direction of the VA Office of Public Health, the War Related Illness and Injury Center (WRIISC) clinical and research program in California focuses on comprehensive medical evaluations for veterans who have chronic, medically unexplained symptoms and on improving the health of deployed veterans by investigating innovative treatments.
In 2013 alone, 60% of the veterans that were seen at the WRIISC clinic listed TBI as one of the top 10 most common health issues.3 Additional complaints were also listed under mental health, including posttraumatic stress disorder (PTSD), depression, and anxiety. Cognitive complaints were the third most common concern.3
The WRIISC team is at a unique junction where clinical research can inform and aid clinical practice by evaluating the complex problems that are faced by returning veterans. Advanced neuroimaging measures, for example, have significantly aided the diagnosis and prognosis of TBI in veterans and civilians alike.1,4–15
In the VA system, conventional clinical neuroimaging, such as the computed tomography (CT) scan, is recommended in core and supplemental common data elements (CDE) initiatives for acute TBI. Recently, advanced techniques, such as diffusion tensor imaging (DTI), have been proposed to enhance knowledge of TBI's chronic stages.8
However, there are some major challenges in the application of new research techniques in a clinical setting, such as the technical capacity and availability of scanners, and obtaining the necessary equipment in the hospital (eg, servers, analysis software, etc). Despite these challenges, WRIISC in California has been able to incorporate a state-of-the-art clinical research DTI sequence to all veterans who visit the clinic.
In this article, we report on the method of integrating measures obtained from clinical and research neuroimaging. The WRIISC mission is to understand the long-term effects of TBI and its relation to other postdeployment health problems, particularly PTSD and cognitive impairments.
Diffusion tensor imaging is a type of magnetic resonance imaging (MRI) sequence that provides in vivo measures of directional water diffusion in brain white matter tracts. Common DTI metrics are fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD). These measures indicate the integrity of axonal membranes and myelin sheaths.7,9 When combat TBI exists, the integrity of white matter fiber is compromised as a result of blunt-force trauma or blast-related injury. Such damage can lead to bleeding, focal axonal shearing, and a degenerative condition called diffuse axonal injury (DAI).13,16
Impairments in attention, executive function and memory are frequently reported in TBI patients at all severity levels.13 It has been known for some time that susceptibility of frontal and temporal brain regions may be the basis for these symptoms.17
Studies of adult TBI patients show that an FA decrease in certain fiber tracts is often coupled with an increase in MD postinjury. Fractional anistropy and MD are also correlated with injury severity, functional outcome, as well as neurologic and cognitive ability.8
Traumatic brain injury can affect the major fiber bundles in the brain, including the corpus callosum, cingulum, the superior and inferior longitudinal fasciculus, the uncinate fasciculus, and brain stem tracts.5 Large fiber groups can potentially act as an index for the degree of damage sustained from a TBI. The superior longitudinal fasciculus (SLF) fiber group, including the SLF proper and arcuate fasciculus (AF), has shown a decreased FA even in mild TBI.9 In addition, performance on the first trial of a verbal learning task was associated with decreased FA in the SLF—and the uncinate fasciculus—in fully functioning patients with mild TBI.18
Previously, we presented a case where the shearing of fibers in the AF at the gray-white matter junction was associated with impaired language function.19 Based on these findings, we applied our integrated research and clinical method to the SLF group (SLF and AF) in veterans with mild and moderate TBI. As PTSD is the most common comorbidity of TBI, after cognitive impairment, we also included a measure of PTSD in our analysis.
In the WRIISC sample, 91 veterans received a clinical/research MRI scan, neuropsychologic evaluation, and clinical assessment (Controls: n = 16; PTSD only: n = 17; TBI only: n = 11; TBI/PTSD: n = 47 [Table]). Eighty-three veterans had been deployed to 1 of 4 conflicts (Vietnam: n = 14; Gulf War One: n = 44; Operation of Enduring Freedom (OEF): n = 11; Operation of Iraqi Freedom (OIF): n = 27). The Control, PTSD, and TBI + PTSD groups were compared.
Traumatic brain injury classification was based on the American Medical Association's (AMA) criteria that includes injury type, such as blast, contact, and mixed. Posttraumatic stress disorder was determined by clinical diagnosis, using the Clinician Administered PTSD Scale (CAPS). All participants completed a battery of neuropsychological tests, including Repeatable Battery for Assessment of Neuropsychological Status (RBANS), Mini-Mental State Examination (MMSE), Wechsler Adult Intelligence Scale (WAIS) Digit Span, Delis-Kaplan Executive Function System (DKEFS) Trails A & B, California Verbal Learning Test (CVLT), and Wechsler Test of Adult Reading (WTAR). These scores are shown in the Table.
The WRIISC patients were scanned with an ASSET (EPI) DTI sequence on a GE 3T MR scanner with an 8-channel head coil: 30 diffusion-encoding directions at b = 1000 s/mm2, 5 acquisitions at b = 0. Diffusion tensor imaging preprocessing was performed with mrDiffusion, freeware developed by the VISTA Lab (https://vistalab.stanford.edu). Using custom MAT-LAB programs, fiber tracts from whole brain tractography of each participant were automatically classified into 20 fiber structures as defined in the Johns Hopkins University white matter tractography atlas.14 Specifically, we manually defined reference region of interests (rROIs) describing 2 way-points for the major white matter tracts based on the literature.6 Fibers were retained if they passed through the paired rROIs.
Our findings indicate that patients with PTSD and TBI+PTSD have compromised fiber tracts compared to healthy controls.
To evaluate tract integrity, we divided the white matter tracts into 90 sections and retained FA values from each. This resulted in a profile consisting of a series of FA values running the length of each tract (n = 20) for every participant. Because TBI damage is often localized, we reasoned that averaging raw FA values across a tract could dilute potential effects. Instead, we compared patient profiles with a group of controls (WRIISC patients without neurologic or psychiatric conditions; n = 16) by calculating the number of standard deviations and FA values that fall below the control mean. The absolute values were then totaled across sections to produce scores for individual tracts as well as the total score for all tracts.
A similar method was applied to the controls, except that each individual control was compared with a group mean calculated without that participant's data. By summing z scores, this quantifies deviation from the normal population. We used these values in our correlations and group analyses. The larger scores indicate greater damage.
It was demonstrated that there were no significant differences in age or education between the 3 groups (P > .05).
T tests did not show significant differences between the control, PTSD, and TBI + PTSD groups on neuropsychological measures with Digit Span, CVLTII, and Trail Making Test B (TM-B). Correlations between Digit Span, CVLTII scores, and total FA were also nonsignificant. There was a significant positive correlation between the time (seconds) to complete the TM-B and total FA score (r = .34, P = .01). Data were distributed so that control participants had the lowest scores (least damage), PTSD patients had higher scores, and TBI + PTSD patients had the highest scores (most damage).
A one-way analysis of variance demonstrated there was a significant effect on the control, PTSD, and TBI + PTSD groups with total FA scores for all fibers (F (2, 68) = 3.21, P = .04). Posthoc t tests revealed total FA scores were significantly different between patients with PTSD (including PTSD and TBI + PTSD groups) and controls (t(40) = −3.27, P = .001 [one-tailed]). However, TBI + PTSD scores did not differ compared with those of the PTSD group, (t(35) = −1.32, P = .09 [one-tailed]).
We calculated Welch's t tests selectively on the SLF and the AF (Figure). We found FA scores in both the left and right SLF were higher for the TBI + PTSD group compared with controls (left, t(22) = −1.93, P = .03 [one-tailed]; right, t(24) = −2.12, P = .02 [one-tailed]) but not higher than the PTSD group (left, t(27) = −0.53, P = .30; right, t(28) = −0.74, P = .23).
Compared with controls, FA scores in the left and right AF were higher for the TBI + PTSD group (left, t(30) = −2.73, P = .005 [one-tailed]; right, t(27) = −1.89, P = 0.03 [one-tailed]) but not higher than the PTSD group (left, t(23) = 0.34, P = .36; right, t(26) = −0.64, P = .26). Though some P-values did not pass Bonferroni correction, effect sizes were medium to large for significant t tests (r = .45, r = .38, r = .39, r = .44, and r = .34, respectively), indicating that group differences are substantial.
Our findings indicate that patients with PTSD and TBI + PTSD have compromised fiber tracts compared with those of healthy controls, both in an aggregate measure of 20 tracts and in specific measures of the superior longitudinal and AF. However, we did not find differences in FA scores between the PTSD and TBI + PTSD groups. Previous research demonstrated that TBI and PTSD patients share similar symptoms6,8,20 Our initial findings indicate they may also share similar patterns of damage in white matter tracts. More research is necessary to disentangle any potential biomarkers for TBI and PTSD. Our findings here indicate damage to SLF and AF may primarily be due to PTSD, because damage to these tracts does not covary with TBI status. WRIISC is currently reviewing DTI data from a larger patient sample to determine whether any particular fascicle or combination is more predictive of TBI or PTSD. We hope to use this information to improve diagnosis and target areas for treatment with interventions such as rTMS, exercise and cognitive training, and language rehabilitation.19
Other forms of multimodal imaging studies are becoming more common, but DTI remains a promising tool in TBI research and clinical practice for several reasons: (1) it assists in clinical diagnosis—particularly in mild TBI and prognosis; (2) it provides metrics to understand the nature and time course of white matter changes in vivo; (3) it reveals target areas of neuroplastic changes due to injury or insult; and (4) it provides a platform for assessing intervention and rehabilitation.4,10–12,15,20
Further studies are necessary to validate the neuropathologic basis that underlies differences in various DTI-derived metrics.16 However, efforts are already underway to evaluate the utility of different commercially available software and public domain tools that may be very helpful in processing data. These tools may provide an excellent basis for data sharing and standardization in processing imaging data.8