ReviewA meta-analysis of structural brain abnormalities in PTSD
Introduction
Exposure to trauma can precipitate the development of posttraumatic stress disorder (PTSD), a complex syndrome comprising re-experiencing symptoms (e.g., nightmares, flashbacks) hyperarousal symptoms (e.g., insomnia), numbing symptoms (e.g., restricted affect, anhedonia), and avoidance symptoms (e.g., avoiding trauma-related stimuli) (DSM-IV, American Psychiatric Association, 1994) in addition to poor concentration and difficulty explicitly recalling aspects of the traumatic event (DSM-IV, American Psychiatric Association, 1994). PTSD may be accompanied by other types of mild cognitive impairment, such as relatively impoverished autobiographic memory for positive events (Harvey et al., 1998; McNally et al., 1995) as well as problems with attention, working memory (Vasterling et al., 1998, Vasterling et al., 2002), and learning novel word associations (Golier et al., 2002). Studies of electro-encephalographic activity (Karl et al., 2006) have found that PTSD is associated with enhanced processing of trauma-related stimuli and reduced processing of neutral stimuli. Converging evidence from neuroimaging research suggests that this altered information processing is associated with differential functional neuroanatomical activity in PTSD (Bremner et al., 1999b, Bremner et al., 2003b; Clark et al., 2003; Matsuo et al., 2003; Rauch et al., 1996; Shaw et al., 2002; Shin et al., 2004a, Shin et al., 2004b).
Studies of structural brain abnormalities in PTSD have focused in particular on the hippocampus, a grey matter structure in the limbic system that is critically involved in explicit (declarative) memory, working memory (O’Keefe and Nadel, 1978; Squire, 1992), and memory for episodic events (Eldridge et al., 2000; Tulving, 1985; Wheeler and Buckner, 2004). The hippocampus also has an important role in the regulation of stress (Jacobson and Sapolsky, 1991), and findings from animal research suggest that chronic stress may affect the hippocampus through excess release of glucocorticoids (Sapolsky et al., 1990), corticotropin-releasing hormone (Brunson et al., 2001), and glutamate (Moghaddam, 2002; Moghaddam and Bolinao, 1994), inhibition of neurogenesis (Gould et al., 1997); impaired long-term potentiation induction (Li et al., 2005); inhibition of brain-derived neurotrophic factor (BDNF, Duric and McCarson, 2005) and altered serotonergic receptor function (Harvey et al., 2003).
Because of its critical role in learning and memory as well as stress regulation, alterations in the hippocampus have been proposed as contributing to the etiology of PTSD (Bremner, 2001; Sapolsky, 2000). However, findings from PTSD neuroimaging research are equivocal (Jelicic and Merckelbach, 2004). Some cross-sectional studies find reduced hippocampal volumes (e.g., Bremner et al., 1995; Gurvits et al., 1996; Stein et al., 1997) in PTSD but others do not (e.g., Pederson et al., 2004; Schuff et al., 2001). Right-sided (Bremner et al., 1995), left-sided (Gurvits et al., 1996) as well as bilateral (Bremner et al., 2003a) volumetric reductions have been reported. One longitudinal study failed to find reduced hippocampal volume at 6 months post-trauma (Bonne et al., 2001), but the sample in this study experienced only a single incident trauma rather than chronic trauma exposure. Smaller hippocampal volumes have been associated with longer time since trauma (Villarreal et al., 2002) as well as trauma severity (Bremner et al., 1997; Gurvits et al., 1996; Winter and Irle, 2004) but there are negative findings as well (Stein et al., 1997). Winter and colleagues (Winter and Irle, 2004) found reduced hippocampal volumes in burn survivors with and without PTSD, compared to non-exposed healthy controls, which suggests that trauma exposure may produce reductions in hippocampal volumes in the absence of a PTSD diagnosis. In contrast, in Gilbertson et al.'s (2002) twin study, smaller hippocampal volumes were only found in combat veterans with more severe PTSD compared to non-exposed controls, with no significant differences when veterans with less severe PTSD were included in the analyses. Perhaps most critically, they found no significant difference in hippocampal volumes between monozygotic twin pairs with and without PTSD, and concluded that smaller hippocampal volume is a premorbid risk factor for severe and chronic PTSD, rather than a consequence of PTSD or trauma exposure.
In their critical review, Jelicic and Merckelbach (2004) noted that PTSD hippocampus volumetric studies are beset by a number of limitations, including small study sample sizes and low statistical power, methodological heterogeneity (e.g., neuroimaging measurements, type of control sample), and sample heterogeneity (e.g., type and severity of trauma exposure, comorbid psychopathology, medication use). Meta-analysis is a technique that can address some of these limitations, and the results of two recent meta-analyses have provided further evidence of hippocampal volumentric reduction in PTSD. Smith (2005) meta-analyzed 13 studies of adult patients with PTSD and found that persons with PTSD had left and right hippocampal volumes that were 7.2% and 7.0% smaller, respectively, than those of non-exposed controls, and 4.3% and 4.5% smaller, respectively, than those of trauma-exposed controls. Kitayama and associates (2005) also found smaller bilateral hippocampal volume in PTSD compared to both trauma-exposed and non-exposed controls in a meta-analysis of nine studies of adult patients, the majority of whom had chronic trauma exposure (combat veterans and adult survivors of childhood abuse).
The objective of the research that we present in this paper was to quantitatively integrate the literature through a comprehensive series of meta-analyses of structural abnormalities in PTSD. We expanded upon the results of the two previous meta-analyses (Kitayama et al., 2005; Smith, 2005) in the following ways. As recommended by Glass et al. (1981) we did not restrict the study sample to only those studies with the best methodology, which yielded a larger and more inclusive sample of studies. We then used empirical methods to identify sample heterogeneity and to construct homogenous groups for analyses. To examine whether volumetric reductions were specific to PTSD, we also meta-analyzed comparisons of trauma-exposed samples without PTSD versus healthy controls. To address method and sample variance, we conducted an extensive series of analyses examining the effects of moderator variables, including MRI methodology, gender, age and age of trauma exposure, PTSD severity and duration, comorbid disorders, and medication. To examine whether volumetric reductions were restricted to the hippocampus, we meta-analyzed PTSD volumetric studies of other brain regions. For ease of apprehension, we have organized the series of meta-analyses into separate sections punctuated by summaries. In the discussion we summarize the overall results and explicate their implications for the formulation of comprehensive neurobiological models of PTSD.
Section snippets
Studies/samples
Fifty English language candidate studies (23 hippocampus studies; 27 studies of other brain areas) were located through electronic indexes (Medline, PsychInfo; keywords: PTSD and MRI, hippocampal volume, amygdala volume, ACC, corpus callosum) and through perusing relevant journals from 1990 to 2005 (e.g., Neuroimage, Nature Neuroscience, Hippocampus, Biological Psychiatry, Biological Psychology). To address the “file drawer problem” (Hunter and Schmidt, 1990),1
Analyses 1: hippocampal volumetric studies
Studies were grouped according to type of control group: trauma-exposed (non-PTSD) or non-trauma exposed healthy controls (HC); and by hippocampal hemisphere.
PTSD vs. HC: The meta-analysis included 15 studies (Studies # 1–4, 8, 9, 11, 12, 14–17, 25, 26, 35 in Appendix A), . Persons with PTSD had significantly smaller bilateral hippocampal volume; see Table 1 and Fig. 1.
PTSD vs. non-PTSD: The meta-analysis included 12 studies (# 2, 4–7, 10, 13–15, 18, 25, 35 Appendix A), . The
Discussion
The main findings of the meta-analyses were as follows. Compared to control groups with no trauma exposure, samples with PTSD and trauma-exposed samples without PTSD showed significantly smaller hippocampal volume bilaterally. Compared to trauma-exposed controls, persons with PTSD reliably exhibited significantly smaller hippocampal volumes bilaterally only in samples with severe PTSD. MRI methodology differentially moderated results, depending upon the type of method and type of control group.
Acknowledgements
Our current research on PTSD is supported by the Deutsche Forschungsgemeinschaft (KA1476/3). Denise Dörfel was awarded a scholarship by the G.A. Lienert Foundation.
References (121)
- et al.
Association between lower P300 amplitude and smaller anterior cingulate cortex volume in patients with posttraumatic stress disorder: a study of victims of Tokyo subway sarin attack
Neuroimage
(2005) - et al.
Voxel-based morphometry—the methods
Neuroimage
(2000) - et al.
Reliability of in vivo volume measures of hippocampus and other brain structures using MRI
Magnetic Resonance Imaging
(1993) - et al.
Reliability of medial temporal lobe volume measurements using reformatted 3D images
Psychiatry Research
(1998) - et al.
Magnetic resonance imaging-based measurement of hippocampal volume in posttraumatic stress disorder related to childhood physical and sexual abuse-a preliminary report
Biological Psychiatry
(1997) - et al.
Neural correlates of exposure to traumatic pictures and sound in Vietnam combat veterans with and without posttraumatic stress disorder: a positron emission tomography study
Biological Psychiatry
(1999) - et al.
Neural correlates of declarative memory for emotionally valenced words in women with posttraumatic stress disorder related to early childhood sexual abuse
Biological Psychiatry
(2003) - et al.
Neural correlates of the classic color and emotional stroop in women with abuse-related posttraumatic stress disorder
Biological Psychiatry
(2004) - et al.
A unified approach for morphometric and functional data analysis in young, old, and demented adults using automated atlas-based head size normalization: reliability and validation against manual measurement of total intracranial volume
Neuroimage
(2004) - et al.
Attenuation of frontal asymmetry in pediatric posttraumatic stress disorder
Biological Psychiatry
(2001)