Central Type Benzodiazepine Receptors in Gulf War Veterans with Posttraumatic Stress Disorder

Masahiro Fujita, Steven M. Southwick, Christopher C. Denucci, Sami S. Zoghbi, Martha S. Dillon, Ronald M. Baldwin, Ali Bozkurt, Akira Kugaya, N. Paul L.G. Verhoeff, John P. Seibyl, and Robert B. Innis
Background: A previous single photon emission computed tomography study showed decreased central type benzodiazepine receptors in the prefrontal cortex in Vietnam War veterans with posttraumatic stress disorder. To assess the generalizability of this finding to patients with more recent history, we studied central type benzodiazepine receptors in Gulf War veterans with posttraumatic stress disorder. Methods: Nineteen Gulf War veterans with posttraumatic stress disorder and 19 age-matched, healthy, nondeployed veterans participated in a single photon emission computed tomography study using [123I]iomazenil. Regional total distribution volume (VT) was compared between two groups using Statistical Parametric Mapping 99 (Wellcome Department of Imaging Neuroscience, London, United Kingdom) and volumes of interest analysis. Results: Benzodiazepine receptor levels did not show regional differences between the two groups, either with or without global normalization. Average difference in VT was 2% across brain areas; however, by applying global normalization, VT in the patient group showed significant negative correlation with childhood trauma scores in the right superior temporal gyrus. Conclusions: Less severe symptoms and shorter duration of the illness in the current group than the prior one may be the source of the difference in the results of the two studies. Key Words: Single photon emission computed tomography, iomazenil, reproducibility, childhood trauma, combat, statistical parametric mapping osttraumatic stress disorder (PTSD) is characterized by exposure to psychologically distressing events, such as combat, rape, and child abuse, and symptoms of reexperiencing, avoidance, and hyperarousal. Animal and human studies have shown involvement of a number of neuroendocrine and neurotransmission systems, such as hypothalamic-pituitary-adrenal axis, norepinephrine, serotonin, opiates, and gamma-aminobutyric acid (GABA) systems. Animals exposed to acute inescapable stress such as forced swim or foot shock showed a decrease in benzodiazepine receptor densities in cerebral cortices and hippocampus (Lippa et al 1978; Medina et al 1983; Weizman et al 1989); however, chronic changes in these receptors are unknown, although patients suffer from symptoms for many years. Further, in most animal studies, old generation ligands were used that bind to both central and peripheral-type benzodiazepine receptors, and one study showed a decrease in peripheral-type receptors in the cerebral cortex (Drugan et al 1986). (Subsequently, the central type benzodiazepine receptor is described as BZR.) In patients with PTSD, benzodiazepine administration has been shown to be effective to treat anxiety but not core symptoms of PTSD (Braun et al 1990). These findings generally indicate a decrease of BZRs in the cerebral cortices and hippocampus in PTSD patients; however, until recently, changes in patients were unknown.
From the Departments of Psychiatry (MF, SMS, MSD, RMB, AB, AK, NPLGV, RBI) and Radiology (SSZ), Yale University and Veterans Affairs Connecticut Healthcare System, West Haven, Connecticut; Molecular Imaging Branch (MF, CCD, SSZ, RBI), National Institute of Mental Health, Bethesda, Maryland; and Institute for Neurodegenerative Disorders (JPS), New Haven, Connecticut. Address reprint requests to Masahiro Fujita, M.D., Ph.D., Molecular Imaging Branch, National Institute of Mental Health, Building 1, Room B3-10, 1 Center Drive, MSC-0135, Bethesda, MD 20892-0135. Received December 30, 2003; revised February 25, 2004; accepted March 17, 2004.

We studied patients with Vietnam War combat-related PTSD using a selective ligand for BZRs, [123I]iomazenil, and single photon emission computed tomography (SPECT) and found a decrease in the prefrontal cortex (Bremner et al 2000a). To our knowledge, this is the only study to date on BZRs in patients with PTSD. Other brain imaging techniques have been applied to PTSD patients. Using positron emission tomography (PET) and functional magnetic resonance imaging (MRI), several research groups studied changes in cerebral blood flow caused by personalized trauma scripts or trauma-related sounds or pictures (reviewed by Bremner 2002; Hull 2002). Replicated abnormal functional findings are located in the amygdala, Brocas area, prefrontal cortex, hippocampus, and cingulate. Volumetric analyses have been performed for MRI and several studies consistently found smaller hippocampi in PTSD patients (also reviewed by Bremner 2002; Hull 2002), although one study did not find atrophy 6 months after the traumatic event in subjects who developed PTSD at the time of the MRI scan (Bonne et al 2001). One study found larger right superior temporal gyrus, particularly the posterior half, in maltreated children and adolescents with PTSD (De Bellis et al 2002a). To assess the generalizability of the previous finding on BZRs in Vietnam War combat-related PTSD patients, we measured BZRs in Gulf War veterans with PTSD, who had shorter history of the disease. We used the same imaging techniques as in the previous study, which have been established (Abi-Dargham et al 1994) and validated by studying test-retest reproducibility in a small number (n 4) of healthy subjects (Abi-Dargham et al 1999). To further study the reproducibility of the measurement, we performed a test-retest study using a separate group of nine healthy subjects.

Methods and Materials

Subjects The study was approved by Yale University School of Medicine and VA Connecticut Healthcare System West Haven review boards. After complete description of the study to the subjects, written informed consent was obtained. The study was performed between February 2000 and August 2001. BIOL PSYCHIATRY 2004;56:95100 2004 Society of Biological Psychiatry

0006-3223/04/$30.00 doi:10.1016/j.biopsych.2004.03.010

96 BIOL PSYCHIATRY 2004;56:95100

Table 1. Sample Demographics Characteristics N Age (years) Gender (M/F) Ethnicity (AA/C/H) Height (cm) Weight (kg) Handedness (R/L/A) Education (years) WAIS-Ra PTSD 19 36 8 15/4 3/14/2 174 8 87 11b 16/1/2 15 2 102 12 Healthy 19 38 7 14/5 2/15/2 169 15 79 19 15/3/1 14 3 106 15

M. Fujita et al comparison subjects who had first degree relatives with psychiatric illnesses were excluded. A test-retest study was performed on a separate group of nine healthy subjects (ages: 32 9; with these and subsequent data expressed as mean SD; nine males; Caucasian: six, Hispanic: two, African Americans: one). These subjects were recruited via newspaper advertisement. History and Clinical Ratings Posttraumatic stress disorder patients were evaluated by the Clinician-Administered PTSD Scale for DSM-IV (CAPS) (Blake et al 1990), Mississippi Scale for Combat-Related PTSD (MISS) (Keane et al 1988), Combat Exposure Scale (CES) (Lund et al 1984), Early Trauma Inventory (ETI) (Bremner et al 2000b), Wechsler Memory Scales (WMS) (Russell 1975), Hamilton Anxiety Scale (HAM-A) (Hamilton 1959), Clinician Administered Dissociative States Scale (CADSS) (Bremner et al 1998), and Dissociative Experiences Scale (DES) (Bernstein and Putnam 1986). In the ETI, eight questions irrelevant to childhood trauma were not used. Wechsler Memory Scales, ETI, HAM-A, CADSS, and DES were also applied to the comparison subjects. SPECT and MRI [123I]Iomazenil was synthesized as described previously with specific activity greater than 185 MBq/nmol (Zoghbi et al 1992). Equilibrium SPECT with a blood sampling and MRI were performed as described previously (Verhoeff et al 1999) using a PRISM 3000 X P (Philips Medical Systems, The Netherlands). The dose of [123I]iomazenil for a priming bolus was 72.2 3.2 MBq for patients and 71.6 3.2 MBq for controls (no significant difference between groups [NS]), and the dose of continuous infusion was 18.8 .5 MBq/h for patients and 18.7 .8 MBq/h for controls (NS). Because our previous study showed that equilibrium was achieved by 5 hours 30 minutes (Abi-Dargham et al 1994), SPECT imaging was started at this time point. Subjects participating in the test-retest study received a priming bolus of 70.5 4.0 MBq, followed by continuous infusion at a rate of 18.4 1.0 MBq/h. Single photon emission computed tomography imaging was started at 6 hours. The subjects had two SPECT studies with an interval of 24 21 days. Plasma Metabolite Analysis Plasma samples were analyzed as previously described to measure [123I]iomazenil levels (Zoghbi et al 1992). The free fraction (f1) measured in the subject sample was corrected for interassay variability as described previously (Abi-Dargham et al 1999). In eight subjects among the study comparing PTSD patients and nondeployed veterans, f1 was not measured using a standard sample. Clearance was calculated by dividing infusion rate expressed as kBq/h by plasma total [123I]iomazenil measured under equilibrium conditions at 5 hours 48 minutes (PTSD patients and nondeployed veterans) or 6 hours 18 minutes (test-retest study). Image Analysis Single photon emission computed tomography projection data were processed and spatially normalized to the Montreal Neurologic Institute (MNI) space (Collins et al 1994) using Statistical Parametric Mapping 99 (SPM 99; Wellcome Department of Imaging Neuroscience, London, United Kingdom) as described previously (Verhoeff et al 1999). These normalized SPECT images were decay corrected and converted to parametric images of VT (brain activities/plasma free [123I]iomazenil) and VT (brain activities/plasma total [free protein-bound]

Mean SD values are shown on the table. Handedness was determined by the Edinburgh Handedness Inventory (Oldeld 1971). PTSD, posttraumatic stress disorder; M, male; F, female; AA, African American; C, Caucasian; H, Hispanic; R, right; L, left; A, ambidextrous; WAIS-R, Wechsler Adult Intelligence Scale-Revised. a Wechsler Adult Intelligence Scale-Revised, full scale score. b p .036.

Patients and healthy comparison subjects were recruited via direct mail to veterans who served in the period of the Gulf War or referred from the VA Connecticut Healthcare System Anxiety Disorders Clinic. Sample demographics are shown in Table 1. Nineteen Gulf War veterans with PTSD, who met criteria for PTSD as determined by the Structured Clinical Interview for DSM-IV (SCID), participated in the study. All participants were outpatients. At study entry, patients had not taken any psychotropic medications for at least 4 weeks. In addition, patients had no history of benzodiazepine usage in the preceding 6 months. Urinary toxicology screens were performed, and patients in whom the presence of benzodiazepines or other substances were detected were excluded from the study. Those who had a history of other brain disorders, psychiatric illnesses other than mood and anxiety disorders, major medical illnesses, and foreign bodies were excluded. Posttraumatic stress disorder patients were evaluated with SCID to determine the presence of psychiatric diagnoses. None of the 19 patients had a history of a psychiatric disorder that preceded the onset of PTSD, and in all cases, PTSD was the primary diagnosis. Comorbid conditions detected in 19 PTSD patients included major depression (current: n 7; lifetime: n 14), generalized anxiety disorder (current: n 7; lifetime: n 7), specific phobia (current: n 4; lifetime: n 5), social phobia (current: n 3; lifetime: n 4), panic disorder (current: n 3; lifetime: n 4), bipolar disorder (current: n 1; lifetime: n 3), obsessive-compulsive disorder (current: n 1; lifetime: n 1), and dysthymia (current: n 1; lifetime: n 1). Four patients met criteria for past alcohol dependence, and five patients met criteria for past alcohol abuse. The minimum interval between the SPECT study and the dependence or the abuse was 2 years and 8 months, respectively. Comparison subjects were healthy, nondeployed veterans who were in the service in the same period as the patients and were selected to be matched by age. There was no significant difference in age between the two groups. Those who had a history of psychiatric disorder as determined by the nonpatient version of SCID were excluded; however, comparison subjects with a history of alcohol dependence (n 3) and abuse (n 1) were included to minimize effects of alcohol use on the comparison between PTSD patients and nondeployed veterans. The minimum interval between the SPECT study and the dependence or the abuse was 2 years and 5 years, respectively. In addition to the exclusion criteria outlined above for PTSD patients, those www.elsevier.com/locate/biopsych

M. Fujita et al [123I]iomazenil) using the plasma metabolite data and a calibration factor of 33.7 Bq/cpm, which was obtained by imaging an anthropomorphic brain phantom (RS-900T, Radiology Support Device Inc., Long Beach, California). In the test-retest study, VT showed better reproducibility than VT. For this reason as well as the fact that f1 was not measured using a standard sample in eight subjects of the comparison study, parametric images of VT were created and analyzed to compare PTSD patients and nondeployed veterans. Test-retest reproducibility was assessed by obtaining volume of interest (VOI) data from spatially normalized parametric images using a template with 116 VOIs (Tzourio-Mazoyer et al 2002) developed for the MNI space. SPM Analysis Statistical Parametric Mapping 99 was used to compare distribution volume in each voxel between PTSD patients and healthy subjects and to study correlations between distribution volume in each voxel and history and clinical ratings. Subjects age was used as a nuisance variable in the correlation analyses. To analyze all brain regions, including low-density regions, gray matter threshold was set at 50%. In all SPM analyses, parametric images were analyzed in two ways, with and without global normalization using proportional scaling. Images were smoothed with 10-mm full width at half maximum (FWHM) and height threshold was set at uncorrected p .001. Corrected cluster level p .05 was considered as significant for the comparison between PTSD patients and nondeployed veterans. In correlation analyses with rating scales, corrected cluster level of p .003 (.05/16) was considered as significant, because 16 correlation analyses were performed. When clusters with uncorrected p .001 were detected in brain areas where previous studies reported abnormalities, small volume correction was applied using a corresponding VOI of the template developed for the MNI space (Tzourio-Mazoyer et al 2002). Partial Volume Correction In addition to voxel-based analyses using SPM 99, VOI data were analyzed in the following way to correct possible effects of gray matter atrophy. This correction was also applied to the test-retest data. Tissue outside the brain was removed from the original magnetic resonance (MR) images using the automatic brain extraction tool (Smith 2002) implemented in MEDx 3.42 (Sensor Systems, Inc., Sterling, Virginia). The extracted MR images were segmented to gray matter, white matter, and cerebral spinal fluid space using the adaptive fuzzy segmentation (Pham and Prince 1999). Gray matter images, which represented a fraction of gray matter in each voxel, were smoothed with the Gaussian function to adjust to the FWHM values of the SPECT scanner. Smoothed gray matter images were spatially normalized to the MNI space using the parameters obtained from the original MR images. Volume of interest data were obtained in the following two ways from both spatially normalized parametric SPECT and smoothed gray matter images. To compare distribution volumes of [123I]iomazenil between PTSD patients and nondeployed veterans in all brain regions, VOI data were obtained using the template developed for the MNI space (Tzourio-Mazoyer et al 2002). To examine the effects of brain atrophy in correlations between distribution volumes and ratings, VOI data were obtained from a cluster of the voxels where SPM detected significant correlation. In each VOI, partial volume corrected SPECT data were calculated by dividing SPECT data by the average

BIOL PSYCHIATRY 2004;56:95100 97

Table 2. Test-Retest Reproducibility of [123I]Iomazenil Equilibrium Measurement Cerebral Cortices Volume of VOIs (cm3) Number of VOIs VT Variability (%) ICC Mean SD VT Values (mL/cm3) VT Variability (%) ICC Mean SD VT Values (mL/cm3) 16.2 9.6 70 13.2 1.0 .63 .09 274 40 9.4 1.1 .58 .12 85 12 Cerebellar Lobes 8.6 6.8 18 13.8 2.1 .75 .16 256 78 11.1 2.8 .67 .21 79 24 Other Areas 6.0 4.5 28 15.0 2.9 .55 .15 215 42 12.1 2.9 .47 .22 67 13

Partial volume correction was applied. VOI, volume of interest; ICC, intraclass correlation coefcient.

fraction of gray matter. Partial volume correction was not performed by subtracting counts in white matter from total counts in gray and white matter because it is difficult to identify white matter in SPECT images of widely distributed BZRs due to the limited spatial resolution and scattered radiation, and the binding of [123I]iomazenil in white matter is only 10% of the total binding in gray and white matter (Innis et al 1991). Partial volume correction was not performed on each voxel because division of SPECT data with a small fraction of gray matter can create significant errors. In fact, although voxel-based partial volume correction, such as the method by Mu ller-Ga tner et al (1992), has been widely used in PET, such methods were applied in only a few SPECT studies. Statistical Analysis Differences were studied using two sample t tests. Correlation between VT values obtained from VOIs and history and clinical rating scores was tested by simple correlation analysis. The difference between simple correlation coefficient, r, was tested by comparing r values after applying Fisher z transformation (Zar 1996). Two-tailed hypothesis was applied in the analyses described above. Test-retest reproducibility of distribution volume measurements was evaluated by variability (absolute difference between two studies/mean of two studies) and intraclass correlation coefficient (ICC). SPSS 11.0 (SPSS, Inc., Chicago, Illinois) was used for statistical analysis.

Results
Test-Retest Reproducibility of SPECT Measurement Results of the test-retest study are shown in Table 2 by three groups of VOIs with high, medium, and low levels of distribution volume reported in our previous study (Abi-Dargham et al 1994). In each of the three groups of VOIs, VT showed significantly smaller variability than VT (p .005); however, VT showed significantly greater ICC (p .005). Therefore, although the variability was greater, VT may be more sensitive than VT to detect interindividual differences. PTSD Clinical Ratings. Compared with comparison subjects, PTSD patients showed greater scores in ETI, HAM-A, CADSS, and DES (Table 3). Each of the WMS was evaluated in terms of immediate and delayed recall and percent retention. Posttraumatic stress disorder patients and control subjects showed similar scores in all WMS tests with a mean difference of 4.2% (NS). www.elsevier.com/locate/biopsych

98 BIOL PSYCHIATRY 2004;56:95100

Table 3. History and Clinical Rating Scores Measure Clinician-Administered PTSD Scale for DSM-IV (CAPS) Current Clinician-Administered PTSD Scale for DSM-IV (CAPS) Lifetime Mississippi Scale for Combat-Related PTSD (MISS) Combat Exposure Scale (CES) Early Trauma Inventory (ETI)a Hamilton Anxiety Scale (HAM-A) Clinician-Administered Dissociative States Scale (CADSS) Dissociative Experiences Scale (DES) PTSD, posttraumatic stress disorder; NA, not available. Total scores of all age ranges are shown. b p .046. c p .0005. d p .001.
a

M. Fujita et al

PTSD 76 25 75 16 106 20 12 9 718 1004b 18 6c 25 10c .19 .16d

Healthy NA NA NA NA 192 318 2.7 3.6 6.3 7.1 .05 .04

SPECT The two groups showed almost identical clearance (PTSD: 115 23 L/h; healthy: 113 30 L/h, NS), indicating that equilibrium was achieved in a similar way in these two groups. The free fraction (f1) measured in 15 patients and 15 comparison subjects was also similar with 34 5% and 36 3% in these groups, respectively (NS). SPM Analysis Statistical Parametric Mapping did not detect differences in VT between PTSD patients and healthy subjects, either with or without global normalization. Among rating scales, only ETI showed significant correlation with VT after global normalization was applied. A cluster composed of 228 voxels was detected in the posterior portion of the right superior temporal gyrus (Figure 1). The greatest t-score in this cluster was 5.57 and the voxel was located at x 48, y 16, and z 6 mm. By applying small volume correction for the right superior temporal gyrus using the VOI for this area (Tzourio-Mazoyer et al 2002) and global normalization, SPM detected significant correlation (corrected p .000) between VT and total scores of ETI.

VOI Analysis with Partial Volume Correction Partial volume corrected VT values were also almost identical between two groups with differences of 2.1 1.2%, 2.1 1.7%, and 2.1 1.8% in cerebral cortices, cerebellar lobes, and other areas, respectively. The maximum difference between two groups among the 116 VOIs was 6.6% (NS even without correction for multiple comparisons) in the left superior parietal gyrus. These results from partial volume corrected VOI data support the negative finding of SPM where each voxel was analyzed without partial volume correction. VT values obtained from the cluster in the right superior temporal gyrus, where SPM showed significant negative correlation with ETI, were further analyzed with and without partial volume correction. Without partial volume correction, correlation was significant with r .72 and p .001. With partial volume correction, the correlation was still significant with r .57 and p .013. These two r values were not significantly different (p .24). There was no correlation between ETI scores and gray matter volume (r .037). These results indicate that VT levels per gray matter volume played a principal role in this correlation.

Discussion
In the present study, we compared BZRs in Gulf War veterans with PTSD and closely matched comparison subjects by analyzing each voxel and also partial volume corrected VOI data. There was no difference in BZRs in any region, including the prefrontal cortex where we detected a decrease in the previous study (Bremner et al 2000a). Partial volume corrected VOI data were almost identical between PTSD and comparison subjects in 116 areas across the brain with a mean difference of 2%, which was much smaller than the test-retest variability. If there had been changes in BZRs that were relevant to psychiatric symptoms or history of childhood trauma, there would have been significant correlation between the scores and BZRs. Therefore, no correlation between clinical rating or childhood trauma scores and absolute VT values (i.e., SPM analysis without global normalization) further supports the finding of no difference in BZRs between PTSD patients and healthy subjects. There may be several sources of the discrepancy between our two studies. Both patients and comparison subjects were different between the previous study and the current study. In the previous study, healthy subjects were not veterans. In the current study, comparison subjects were also veterans who served in the same time period but were not deployed in the Gulf region. The

Figure 1. Distribution volume of [123I]iomazenil relative to the entire brain showed signicant negative correlation with the Early Trauma Inventory in the right superior temporal cortex. t values in each voxel are shown by color.

www.elsevier.com/locate/biopsych

M. Fujita et al comparison subjects showed normal ranges in the three clinical ratings applied to them (Table 3); however, current normal scores do not ensure that they were not exposed to stress and anxiety in the past. Although the comparison subjects were not deployed, they might have been exposed to stress and anxiety. Therefore, one possible explanation is that the decrease of BZRs in the prefrontal cortex was caused by combinations of exposure to stress before traumatic experiences and pathologic changes caused by the development of PTSD. The interval between traumatic experiences and the SPECT studies was also different between the two studies. The interval was about 25 years in the previous study, while it was about 10 years in the current study. There is no clear consensus on the relationship between the duration of the disease and symptom severity, with some studies showing gradual decrease of symptoms (Green et al 1990; Port et al 2001) and others showing late onset (Herrmann and Eryavec 1994; Van Dyke et al 1985). Hamilton Anxiety Scale scores in the current study (18 6) were similar to those in the previous study (17 8). ClinicianAdministered PTSD Scale for DSM-IV scores in the current study were within a range of other studies on PTSD patients, slightly on the lower side. Mississippi Scale for Combat-Related PTSD and CES scores in the current study were lower than those in another study on the Vietnam War veterans with PTSD performed by our group (Grillon et al 1998). In addition, although our previous study on the Vietnam War veterans with PTSD showed memory deficits (Bremner et al 1993), the present study did not. Taken together, patients in the present study may have had less severe symptoms than those in the previous study imaged with [123I]iomazenil, and this may be a reason of not detecting a significant difference. There are a couple of other factors that may explain the discrepancy between the two studies. In the current study, each patient was closely matched with a comparison subject. The difference in age in each pair was only 2.2 1.2 in the current study, while the difference was greater in the previous study (PTSD: 49 2 years; healthy: 44 9 years). The greater difference in age in the previous study could have been a factor to cause the decreased receptors in the prefrontal cortex. For example, age-related decline of brain volume could have played a role in the decrease. In conjunction with the comparison study of patients and healthy subjects, we studied test-retest reproducibility of [123I]iomazenil distribution volume measurements. Before the current study, we performed test-retest studies twice, one with a small group of four subjects (Abi-Dargham et al 1999) and the other with seven subjects (Fujita et al 1999). Although the first study showed fairly good reproducibility of 13% variability for VT, the second study showed poor reproducibility of 20%. In these two studies, plasma metabolite analysis was performed in different laboratories by different chemists. In the present study, the chemist involved in the first test-retest study performed all metabolite analysis and recovered the reproducibility to a similar level as in the first study. The differences in VOI data between PTSD and healthy subjects were much smaller than the test-retest variability. It should be noted that the present study does not exclude possibilities of changes in GABAA receptor subunits. Affinity of iomazenil to individual subunits is not known; however, because a close analog of iomazenil, flumazenil, has high affinity to several subunits, including 1, 2, 3, and 5 (Sieghart 1995), it is likely that iomazenil lacks selectivity among subunits. There-

BIOL PSYCHIATRY 2004;56:95100 99

fore, the results in the present study do not indicate that there is no alteration in GABAA receptor subunits in PTSD patients. For example, an increase of one subunit configuration could be compensated by a decrease of another subunit configuration and not be detected by this SPECT ligand. Victims of child abuse are at increased risk for developing PTSD (Widom 1999), and a recent study showed greater volume in the right superior temporal gyrus gray matter, particularly in the posterior half, in such patients (De Bellis et al 2002b). With global normalization, in the exact same area, the current study detected a significant negative correlation between iomazenil distribution volumes and ETI scores (Figure 1). The ETI is a well-validated instrument to assess childhood trauma (Bremner et al 2000b), used in several studies. Because global normalization was applied, this is not correlation between ETI and absolute values of VT but correlation with VT relative to the average in the entire brain. Although it is difficult to interpret the meaning of such changes relative to the average value, the previous study on Vietnam War veterans with PTSD detected a significant decrease of iomazenil distribution volume in the prefrontal cortex after applying global normalization (Bremner et al 2000a). Because eight patients showed relatively low ETI scores of less than 100, it would be interesting to study correlation with BZRs and childhood trauma in a larger number of PTSD patients with significant history of childhood trauma. Posttraumatic stress disorder patients generally show high comorbidity with other anxiety and affective disorders. Both in the current and the previous studies (Bremner et al 2000a), the majority of patients had a history of major depression (78% in the current study and 85% in the previous study). In the current study, 37% of patients also had a history of generalized anxiety disorder. Although it is difficult to assess effects of the comorbidity in a small sample, it is unlikely that there were significant effects on the finding in the current study for the following reasons. First, we recently reported no change in BZRs in depressed patients using the same imaging techniques (Kugaya et al 2003), and second, the difference in VT between patients with and without a history of generalized anxiety disorder was only 5.3 3.0%, which is less than a half of the test-retest reproducibility of the measurement. In summary, the present study did not detect a difference in absolute value of [123I]iomazenil distribution volume in any brain region. The correlation between relative distribution volumes and childhood trauma in the right superior temporal gyrus may need to be examined in a larger sample with such a history. This work was supported by Veterans Affairs Research Service and Department of Defense DoD-90. We thank M.S. Al-Tikriti, L. Amici, S. Giddings, G. Morano, A. Perez, Q. Ramsby, N. Seneca, N. Sheung, and E.O. Smith for technical assistance and Jeih-San Liow, Ph.D., for assisting in image data processing.
Abi-Dargham A, Laruelle M, Krystal J, DSouza C, Zoghbi S, Baldwin RM, et al (1999): No evidence of altered in vivo benzodiazepine receptor binding in schizophrenia. Neuropsychopharmacology 20:650 661. Abi-Dargham A, Laruelle M, Seibyl J, Rattner Z, Baldwin RM, Zoghbi SS, et al (1994): SPECT measurement of benzodiazepine receptors in human brain with [123I]iomazenil: Kinetic and equilibrium paradigms. J Nucl Med 35:228 238. Bernstein EM, Putnam FW (1986): Development, reliability, and validity of a dissociation scale. J Nerv Ment Dis 174:727735.

www.elsevier.com/locate/biopsych

100 BIOL PSYCHIATRY 2004;56:95100

Blake DD, Weathers FW, Nagy LM, Kaloupek DG, Klauminzer G, Charney DS, et al (1990): A clinician rating scale for assessing current and lifetime PTSD: The CAPS-1. Behav Ther 13:187188. Bonne O, Brandes D, Gilboa A, Gomori JM, Shenton ME, Pitman RK, et al (2001): Longitudinal MRI study of hippocampal volume in trauma survivors with PTSD. Am J Psychiatry 158:1248 1251. Braun P, Greenberg D, Dasberg H, Lerer B (1990): Core symptoms of posttraumatic stress disorder unimproved by alprazolam treatment. J Clin Psychiatry 51:236 238. Bremner C, Scott TM, Delaney RC, Southwick SM, Mason JW, Johnson DR, et al (1993): Decits in short-term memory in post-traumatic stress disorder. Am J Psychiatry 150:10151019. Bremner JD (2002): Neuroimaging studies in post-traumatic stress disorder. Curr Psychiatry Rep 4:254 263. Bremner JD, Innis RB, Southwick SM, Staib L, Zoghbi S, Charney DS (2000a): Decreased benzodiazepine receptor binding in prefrontal cortex in combat-related posttraumatic stress disorder. Am J Psychiatry 157:1120 1126. Bremner JD, Krystal JH, Putnam FW, Southwick SM, Marmar C, Charney DS, et al (1998): Measurement of dissociative states with the Clinician-Administered Dissociative States Scale (CADSS). J Trauma Stress 11:125 136. Bremner JD, Vermetten E, Mazure CM (2000b): Development and preliminary psychometric properties of an instrument for the measurement of childhood trauma: The Early Trauma Inventory. Depress Anxiety 12:112. Collins DL, Neelin P, Peters TM, Evans AC (1994): Automatic 3D intersubject registration of MR volumetric data in standardized Talairach space. J Comput Assist Tomogr 18:192205. De Bellis MD, Keshavan MS, Frustaci K, Shifett H, Iyengar S, Beers SR, et al (2002a): Superior temporal gyrus volumes in maltreated children and adolescents with PTSD. Biol Psychiatry 51:544 552. De Bellis MD, Keshavan MS, Shifett H, Iyengar S, Dahl RE, Axelson DA, et al (2002b): Superior temporal gyrus volumes in pediatric generalized anxiety disorder. Biol Psychiatry 51:553562. Drugan RC, Basile AS, Crawley JN, Paul SM, Skolnick P (1986): Inescapable shock reduces [3H]Ro 5-4864 binding to peripheral-type benzodiazepine receptors in the rat. Pharmacol Biochem Behav 24:16731677. Fujita M, Woods SW, Verhoeff NPLG, Abi-Dargham A, Baldwin RM, Zoghbi SS, et al (1999): Changes of benzodiazepine receptors during chronic benzodiazepine administration in humans. Eur J Pharmacol 368:161172. Green BL, Lindy JD, Grace MC, Gleser GC, Leonard AC, Korol M, et al (1990): Buffalo Creek survivors in the second decade: Stability of stress symptoms. Am J Orthopsychiatry 60:4354. Grillon C, Morgan CA III, Davis M, Southwick SM (1998): Effects of experimental context and explicit threat cues on acoustic startle in Vietnam veterans with posttraumatic stress disorder. Biol Psychiatry 44:10271036. Hamilton M (1959): The assessment of anxiety states by rating. Br J Med Psychol 32:50 55. Herrmann N, Eryavec G (1994): Delayed onset post-traumatic stress disorder in World War II veterans. Can J Psychiatry 39:439 441. Hull AM (2002): Neuroimaging ndings in post-traumatic stress disorder. Systematic review. Br J Psychiatry 181:102110. Innis RB, Al-Tikriti MS, Zoghbi SS, Baldwin RM, Sybirska EH, Laruelle MA, et al (1991): SPECT imaging of the benzodiazepine receptor: Feasibility of in

M. Fujita et al
vivo potency measurements from stepwise displacements curves. J Nucl Med 32:1754 1761. Keane TM, Caddell JM, Taylor KL (1988): Mississippi Scale for Combat-Related Posttraumatic Stress Disorder: Three studies in reliability and validity. J Consult Clin Psychol 56:8590. Kugaya A, Sanacora G, Verhoeff NPLG, Fujita M, Mason GF, Seneca NM, et al (2003): Cerebral benzodiazepine receptors in depressed patients measured with [123I]iomazenil SPECT. Biol Psychiatry 54:792799. Lippa AS, Klepner CA, Yunger L, Sano MC, Smith WV, Beer B (1978): Relationship between benzodiazepine receptors and experimental anxiety in rats. Pharmacol Biochem Behav 9:853856. Lund M, Foy D, Sipprelle C, Strachan A (1984): The Combat Exposure Scale: A systematic assessment of trauma in the Vietnam War. J Clin Psychol 40:13231328. Medina JH, Novas ML, Wolfman CN, Levi de Stein M, De Robertis E (1983): Benzodiazepine receptors in rat cerebral cortex and hippocampus undergo rapid and reversible changes after acute stress. Neuroscience 9:331335. Mu ller-Ga tner HW, Links JM, Prince JL, Bryan RN, McVeigh E, Leal JP, et al (1992): Measurement of radiotracer concentration in brain gray matter using positron emission tomography: MRI-based correction for partial volume effects. J Cereb Blood Flow Metab 12:571583. Oldeld RC (1971): The assessment and analysis of handedness: The Edinburgh inventory. Neuropsychologia 9:97113. Pham DL, Prince JL (1999): Adaptive fuzzy segmentation of magnetic resonance images. IEEE Trans Med Imaging 18:737752. Port CL, Engdahl B, Frazier P (2001): A longitudinal and retrospective study of PTSD among older prisoners of war. Am J Psychiatry 158:1474 1479. Russell E (1975): A multiple scoring method for the assessment of complex memory functions. J Consult Clin Psychol 43:800 809. Sieghart W (1995): Structure and pharmacology of gamma-aminobutyric acidA receptor subtypes. Pharmaco Rev 47:181234. Smith SM (2002): Fast robust automated brain extraction. Hum Brain Mapp 17:143155. Tzourio-Mazoyer N, Landeau B, Papathanassiou D, Crivello F, Etard O, Delcroix N, et al (2002): Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI singlesubject brain. Neuroimage 15:273289. Van Dyke C, Zilberg NJ, McKinnon JA (1985): Posttraumatic stress disorder: A thirty-year delay in a World War II veteran. Am J Psychiatry 142:1070 1073. Verhoeff NP, Soares JC, DSouza CD, Gil R, Degen K, Abi-Dargham A, et al (1999): [123I]Iomazenil SPECT benzodiazepine receptor imaging in schizophrenia. Psychiatry Res 91:163173. Weizman R, Weizman A, Kook KA, Vocci F, Deutsch SI, Paul SM (1989): Repeated swim stress alters brain benzodiazepine receptors measured in vivo. J Pharmacol Exp Ther 249:701707. Widom CS (1999): Posttraumatic stress disorder in abused and neglected children grown up. Am J Psychiatry 156:12231229. Zar JH (1996): Biostatistical Analysis. Upper Saddle River, NJ: Prentice-Hall. Zoghbi SS, Baldwin RM, Seibyl JP, Al-Tikriti MS, Zea-Ponce Y, Laruelle M, et al (1992): Pharmacokinetics of the SPECT benzodiazepine receptor radioligand [123I]iomazenil in human and nonhuman primates. Int J Rad Appl Instrum B 19:881888.

www.elsevier.com/locate/biopsych