Memory impairments in posttraumatic stress disorder are related to depression

Journal of Anxiety Disorders 22 (2008) 464–474

Memory impairments in posttraumatic stress disorder are related to depression Grethe E. Johnsen *, Pushpa Kanagaratnam, Arve E. Asbjørnsen University of Bergen, Bergen, Norway Received 31 October 2006; received in revised form 13 April 2007; accepted 20 April 2007

Abstract The present study focuses on verbal learning and memory alterations in refugees with posttraumatic stress disorder, and whether the alterations are related to attention, acquisition, storage, or retrieval. Twenty-one refugees exposed to war and political violence with chronic PTSD, were compared to an exposed control sample of 21 refugees without PTSD. No differences were found in attention span, but tests of verbal memory showed less efficient learning in the PTSD sample. Group differences in delayed recall could be explained by learning efficiency. No differences were seen in recognition memory. These results indicate that memory alterations in PTSD are related to impaired acquisition and less effective encoding of the memory material and not to impaired attention span and/or impaired retrieval. Controlling for specific PTSD symptom clusters and self-reported depression showed that the intrusion subscale and depressive reactions are the most important symptoms in understanding the memory alterations in PTSD. # 2007 Elsevier Ltd. All rights reserved. Keywords: CVLT; Depression; Memory; PTSD; Refugees; Verbal learning

The present study focuses on verbal learning and memory alterations in refugees with chronic posttraumatic stress disorder (PTSD) after exposure to war and political violence. PTSD is a disorder triggered by one or more traumatic events and the three main clusters of symptoms include intrusive reexperiencing symptoms, avoidance behaviors, and hyperarousal (DSM-IV; American Psychiatric Association, 1994). In addition, problems with memory and lack of concentration characterize most of the patients treated for PTSD. The reports of the extent and nature of the cognitive alterations in patients with PTSD have been inconsistent, and they range from global cognitive deficits to * Corresponding author at: Department of Biological and Medical Psychology, University of Bergen, Jonas Liesvei 91, N-5009 Bergen, Norway. Tel.: +47 55 58 25 61; fax: +47 55 58 98 72. E-mail address: [email protected] (G.E. Johnsen). 0887-6185/$ – see front matter # 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.janxdis.2007.04.007

specific memory disturbances (e.g., Bremner et al., 1993; Gilbertson, Gurvits, Lasko, Orr, & Pitman, 2001; Uddo, Vasterling, Brailey, & Sutker, 1993). The most consistent findings have been impaired performance in attention, verbal learning, and short-term memory (e.g., Bremner et al., 1995; Vasterling, Brailey, Constans, & Sutker, 1998; Yehuda, Golier, Halligan, & Harvey, 2004). Although the relationship between the experience of traumatic stress and verbal memory dysfunction is well documented, few of the previous studies have systematically analyzed the relationship between attention, encoding, storage, and retrieval. When testing the relationship between attention and memory, it is necessary to take into account that these two processes are closely related, and that the memory function is dependent on the effectivity of the attention process. Many studies do not report how deficits in attention are

G.E. Johnsen et al. / Journal of Anxiety Disorders 22 (2008) 464–474

related to the impairments in memory function (e.g., Jenkins, Langlais, Delis, & Cohen, 1998). In most of the previous studies where delayed recall was impaired, immediate recall was also impaired (e.g., Bremner et al., 1993), and few of the surveyed studies report control for differences in initial acquisition between PTSD patients and control groups. Some studies have attempted to give an answer to the relation between stage of processing (encoding, storage and retrieval) and memory impairments. Uddo et al. (1993) found no group difference on the first two trials on a list learning test, but on the last trials the PTSD group showed impaired acquisition. Yehuda and coworkers (2004) found impaired ability to learn new verbal information in Holocaust survivors with PTSD. More recently, Yehuda, Golier, Tischler, Stavitsky, and Harvey (2005) found more rapid forgetting in aging combat veterans. When comparing the PTSD group with non-exposed controls, these authors found a difference in total learning, reflecting problems with encoding of new information. Several of the former studies of cognitive dysfunctions were performed on treatment-seeking Vietnam War veterans. Studies on war veterans have methodological challenges, and uncontrolled factors like secondary gain from a diagnosis may cause unintended and artificial results (McNally, 2003). In addition, war veterans were specially trained for challenging situations, and this may affect their coping and later development of PTSD symptoms (Hytten, 1989; Johnsen, Eid, Løvstad, & Michelsen, 1997). One study on refugees found that PTSD was associated with altered fluid intelligence and episodic memory (Kivling-Bode`n & Sundbom, 2003). However, this study did not include more specific verbal memory tests, so memory functions in these samples require further exploration. Both the elevated arousal levels and the intrusive memories related to PTSD have been hypothesized to interfere with ongoing cognitive processing and thus produce the memory impairments (Kolb, 1987). However, few of the previous studies have analyzed specific hypotheses regarding how the different symptom clusters influence the performance on memory tasks. Other clinical states and premorbid cognitive abilities may contribute to the cognitive problems in patients with PTSD (Barrett, Green, Morris, Giles, & Croft, 1996; Macklin et al., 1998). In addition to PTSD, exposure to a traumatic event can also result in depression or anxiety, and may lead to substance abuse (Kessler, Sonnega, Bromet, & Hughes, 1995;

465

McFarlane & Papay, 1992). Major depression is found to be related to memory impairments, including problems with recall and effortful information processing (Burt, Zembar, & Niederehe, 1995; Egeland et al., 2003; Weingartner, 1981). Since PTSD often occurs comorbidly with other psychiatric disorders with known cognitive impairments, it is a challenge to assess the unique contribution of PTSD on memory functions. In addition, other psychiatric problems have frequently been discussed as comorbid conditions as if they occurred independently from the PTSD symptoms. There is now a growing awareness that complex and prolonged trauma may result in many posttraumatic outcomes (Bremner, 2002; McFarlane & Papay, 1992; Van der Kolk, Roth, Pelcovitz, Sunday, & Spinazzola, 2005). The aim of the present study is to analyze the different processing components of verbal memory and attention. The study will use a battery of different tests that distinguish more precisely among the memory components. In addition we will control for attention by covarying for a measure of attention. This will enable us to find the focal points at which the attention and memory processes become dysfunctional after trauma exposure. In addition the findings from trained samples will be replicated in a sample of war-exposed civilians. PTSD was measured with The Impact of Event Scale— Revised and the construct and its specific symptom clusters is used in reference to this scale (Weiss & Marmar, 1997). We will examine for the effect of these different clusters on verbal learning and memory. In order to examine other possible explanations for memory impairments, we would control for concurrent depressive reactions and general distress. Collecting data on the cognitive and clinical aspects of trauma in different ethnic minorities entail several challenges. The test materials are not always available in the different minorities’ native language. There is a lack of norms, and the use of interpreters increase possible sources of errors. This could be controlled for by balancing for ethnicity in the target group and the control group, and in having a design that focuses on the between-groups comparisons. In this study we will test a set of hypotheses regarding impairments in memory functions in refugees/immigrants with PTSD. The first question addressed is whether there are differences in verbal learning and memory as seen on a standardized memory test, between a group of war exposed refugees with PTSD and a control group of exposed refugees without PTSD. If the memory impairments in patients with PTSD are related to intrusive and arousal symptoms, it

466

G.E. Johnsen et al. / Journal of Anxiety Disorders 22 (2008) 464–474

1. Methods

hospitalizations and no psychiatric diagnosis. The distribution of gender in the PTSD group was 14:7 and in the non-PTSD group 18:3. Diagnoses were made according to DSM-IV criteria based on The Mini-International Neuropsychiatric Interview (MINI), Version 5.0.0 (Sheehan et al., 1998) and Clinician Administered PTSD Scale for DSM-IV Revised (CAPS; Blake et al., 1998). The CAPS assesses all the diagnostic criteria for PTSD, including criterion A, both frequency and intensity of criteria B–D, criterion E, and criterion F. The Life Event Check List in the CAPS, which identifies the main trauma of the participants, was used to confirm that their PTSD syndromes were related to political violence and ˚ sberg war events. IES-R and Montgomery and A depression rating scale were used to screen for PTSD and depressive symptomatology in the control group ˚ sberg, 1979). When participants in (Montgomery & A the non-PTSD group had elevated IES-R scores, they were measured with CAPS in order to rule out the possibility for a PTSD diagnosis. All participants in the PTSD group satisfied the DSM-IV criteria for chronic PTSD, opposed to none in the non-PTSD group. Participants were excluded from the study if they had organic brain damage or other neurological diseases. Those who had been involved in events that gave reason to expect brain injuries, such as blows to the head or loss of consciousness exceeding 30 min, were also excluded from the study (Vasterling et al., 1998), as were subjects with a history of alcohol/substance addiction and/or psychotic disorders. Among the PTSD participants, the diagnosis of major depression disorder was not an exclusion criterion. Participants on tricyclic antidepressant, benzodiazepine, and neuroleptic medication were excluded.

1.1. Participants

1.2. Procedure

Forty-two refugees/immigrants, 32 men and 10 women, mainly from the former Yugoslavia, Chile, and the Middle-East, were voluntary participants in the study. Twenty-one were referred from the primary health care on the basis of complaints of posttraumatic stress symptoms following experiences with actions of war and political violence (the PTSD group). They were tested as part of a clinical psychological screening as they were offered short-term cognitive behavioral therapy for PTSD symptoms. The control group, also consisting of 21 refugees/immigrants had similar age and ethnicity profile (the non-PTSD group). Both groups were homogenous in regard to their trauma exposure. The control group had no previous psychiatric

The participants gave written informed consent prior to participation and were informed that participation was voluntary. The study was approved by the Regional Committee for Medical Research Ethics. Information about the study was given to health personnel in outpatient clinics and at health and social services for refugees/immigrants. Written information was posted at local public health services. Participants who had been exposed to war or political violence, and who had a tentative PTSD diagnosis, were either referred by counselors or recruited themselves. Exposed participants for the non-PTSD group were recruited through the same formal contacts and procedures. All patients went through a diagnostic assessment administered by a

is expected that the symptoms will affect processing capacity or the number of items the refugees are able to process on the California Verbal Learning Test (CVLT; Delis, Kramer, Kaplan, & Ober, 1987). In addition, intrusive behavior could be seen as lack of monitoring of input and would then increase the effect of interference tasks. The second question is whether the learning and memory impairments are related to impaired attention or memory span. If this hypothesis is supported, then the participants with PTSD will show reduced performance on Digit Span forward and the first trial on the CVLT (Wechsler, 1981). Third, if the verbal learning and memory impairments are related to encoding and processing capacity, the differences between the groups will be expected to reveal themselves in the initial trials of CVLT. If the PTSD symptoms affect retention and retrieval, then there will be reduced performance on delayed recall. The fourth hypothesis tested is whether the intrusive symptoms in PTSD represent a generalized impairment in executive control. If this hypothesis is supported, there will be decreased performance on working memory or the ability to hold and manipulate information on Digit Span backward (Wechsler, 1981) and Paced Auditory Serial Addition Test (Gronwall, 1977). Further, we examine whether the verbal learning and memory impairments are more related to one of the PTSD symptom clusters. Earlier research has also shown that premorbid intelligence and comorbid conditions to some extent can explain the memory impairments in PTSD. If these factors are important contributors to the impaired verbal learning and memory, there will be reduced group effects.

G.E. Johnsen et al. / Journal of Anxiety Disorders 22 (2008) 464–474

licensed clinical psychologist with experience in the field of traumatology. All the participants in the study were interviewed about their traumatic stress exposure, and in this situation there may be a risk of the participants being retraumatized. The participants were asked to give a brief description of their experiences in order to judge the eligibility criteria for a diagnosis of PTSD. To further minimize the risk of participants being retraumatized, confidentiality, safety, and predictability were ensured. Most of the participants in the PTSD group were offered cognitive behavior therapy following their assessment, and they were not asked to disclose their experiences of exposure to political violence in detail unless they were to be included in the psychotherapy. All tests included in the test battery were formally translated to Norwegian, and for this study all tests and self-report scales were translated into the participants’ native languages. The diagnostic assessment and testing was done in collaboration with an authorized interpreter. The tests and self-report scales were administered in the same order for all the participants. 1.3. Clinical assessments The Impact of Event Scale—Revised (IES-R), a 22item self-report measure was used to examine the severity of PTSD symptoms the week before testing (Weiss & Marmar, 1997). Current depressive symptoms were ˚ sberg depression rating assessed with Montgomery and A scale (MADRS), a 10-item clinician administered questionnaire measuring depressive symptoms in the last 3 ˚ sberg, 1979). MADRS was used days (Montgomery & A as a structured interview to confirm depressive symptoms. General distress was measured with Symptom Checklist-90—revised (SCL-90-R; Derogatis, 1983). The subtests Picture Completion and Similarities from Wechsler Adult Intelligence Scale—Revised were used to estimate current IQ level (Wechsler, 1981). Picture Completion is seen as a ‘‘hold’’ test and is described as relatively stable and not so easily influenced by external factors, and is expected to be an indictor of premorbid IQ. Both subscales load high on the g-factor (Lezak, 1995). The War Exposure Questionnaire (WEQ), a 27-item self-report measure assessing threat, loss, injury, deprivation, grotesque impressions, and other aspects of war and political violence was administered to assess exposure to war and political violence. The WEQ is a converted version of the Childhood War Exposure Questionnaire, caretaker version (Netland & Kanaaneh, 1989).

467

1.4. Neuropsychological testing California Verbal Learning Test (CVLT; Delis et al., 1987) was used in the assessment of verbal learning and memory. The CVLT is a multi-trial serial learning test which measures immediate and delayed verbal learning. In addition, the test also allows for analysis of organizing skills and recognition memory. A list of 16 words (list A) is presented five times, and the participants are instructed to recall as many words as possible after each presentation of the list. After the five presentations of list A, a new list of words (list B) is presented to the participants, and they are instructed to recall as many as they remember from list B. Participants are then asked to recall list A again in short delay free recall (SDFR), and after a 20 min interval they are asked to recall list A in long delay free recall (LDFR). The standard procedure includes the use of semantic cues to enhance recall (cued recall). Cued recall was measured both as short delay cued recall (SDCR) after SDFR and long delay cued recall (LDCR) after the LDFR conditions. Using derived CVLT scores the participants were tested for proactive interference (list B minus list A trial 1), retroactive interference (SDFR minus list A trial 5) and retention (LDFR minus SDFR). Digit Span forward and backward from the Wechsler Adult Intelligence Scale—Revised (WAIS-R; Wechsler, 1981) were used to assess attention or memory span. Digit span backward measures the ability to hold and manipulate information and was used to assess working memory. Paced Auditory Serial Addition Test (PASAT; Gronwall, 1977) was used to assess working memory in addition to immediate memory and verbal learning ability as seen on CVLT. The PASAT consists of a random series of 61 digits from 1 to 9 and the test was presented to the participants via audiotape. The task is to add continuously the two last digits heard. The procedure requires that the last digit be moved to an instrumental string (the phonological loop) when used in one calculation and recalled again for the next. The stimuli were recorded on an audiocassette and presented in two conditions, with 3 and 2 s inter stimulus interval. 1.5. Statistical methods Data were statistically analyzed using t-tests to evaluate group differences in demographic, clinical and cognitive variables. Differences in learning and memory between the two groups were analyzed using a mixed analysis of variance (ANOVA), where group

468

G.E. Johnsen et al. / Journal of Anxiety Disorders 22 (2008) 464–474

2.2. Attention, learning and memory performances

(PTSD and non-PTSD) was treated as a between factor, and trials on the CVLT were treated as a within group factor. Analysis of covariance (ANCOVA) was used to control for initial attention on total learning, and to control for total learning on delayed recall. Differences on CVLT and Digit Span backwards were analyzed covarying for age and education. In addition, an analysis of covariance was used to evaluate the influence of estimated IQ, PTSD symptoms (IES-R), symptoms of depression (MADRS), and general distress (SCL-90-R) on CVLT performance. Multiple regression analysis, stepwise regression was used to analyze the relative impact of self-reported clinical variables and estimated intelligence on CVLT performance.

The performance of the two groups on the CVLT trials 1–5 is presented in Fig. 1. Initially we performed a two-way ANOVA with trials as a repeated variable and group as a between factor. The analysis yielded a significant main effect for trials (F (4,144) = 51.46, p < .001) in addition to a significant group effect (F (1,36) = 9.49, p = .003). Analysis of pair-wise contrasts revealed significant group differences on trials 2– 5, but not on trial 1 and list B performance (see Table 2). No other sources of variance yielded significant effects. Second, we analyzed the recall performance data with t-tests and revealed significant group differences in SDFR, LDFR, SDCR and LDCR (see Table 2). Further individual CVLT components revealed no increase from SDFR to LDFR for either of the groups (tPTSD group 0.15, tnon-PTSD group (1,33) = 1.12) This was (1,38) = also tested for cued recall, but neither of the groups had a significant increase from SDCR to LDCR (tPTSD group 0.10, tnon-PTSD group (1,33) = 0.58). (1,38) = Since performance on short and long delayed recall is dependent on the amount of information learned on total learning (CVLT 1–5), differences in delayed recall (SDFR, LDFR, SDCR and LDCR) were analyzed with an analysis of covariance (ANCOVA), controlling for total learning. When controlling for total learning the groups did not differ on either of the recall measures (F (1,34) = 0.49) (see Fig. 2). Since total learning is dependent on initial attention or memory span, the analysis was repeated using trial 1 performance as a

2. Results 2.1. Sample characteristics Demographic, estimated IQ, and clinical variables are presented by group in Table 1. There were no significant differences between the PTSD group and the non-PTSD group in age and years of education. The two groups were also equal regarding duration of exposure to war and political violence, time passed since their exposure, and time passed since they left their native country. As would be expected, the PTSD group had higher scores on IES-R, MADRS and SCL-90-R, and the effect sizes were large for these significant comparisons (Cohen, 1969).

Table 1 Demographic and clinical characteristics and estimated intelligence for the PTSD group and the non-PTSD group

Age Education (years) Exposure range (months) Time since exposure (months) MADRSa IES-Rb avoidance IES-R intrusion IES-R hyperarousal IES-R total-score IES-R GSI d SCL-90-Rc SCL-90-R GSId Estimated IQ e a b c d e

PTSD group (n = 21)

Non-PTSD group (n = 21)

M

(S.D.)

M

(S.D.)

36.90 12.12 83.14 107.43 26.20 18.29 20.81 20.29 59.38 2.70 190.95 2.12 82.19

(8.77) (2.86) (66.87) (96.00) (6.93) (5.83) (4.34) (3.42) (9.89) (0.45) (56.55) (0.63) (13.00)

39.24 14.05 100.81 166.27 5.85 8.10 8.15 6.60 22.85 1.09 76.75 0.85 91.90

(10.06) (4.12) (71.62) (109.66) (4.21) (8.36) (7.41) (6.54) (21.22) (0.94) (52.35) (0.58) (17.90)

˚ sberg depression rating scale. Montgomery and A The impact of event scale—revised. Symptom checklist-90—revised. GSI: general symptom index. Estimated IQ from the WAIS-R subtests picture completion and similarities.

t-Value

Effect Size (d)

0.80 1.76 0.83 1.71 11.23 4.54 6.71 8.45 7.12 7.03 6.28 6.27 2.01

0.25 0.56 0.26 0.59 3.64 1.45 2.15 2.71 2.28 2.25 2.12 2.12 0.64

p

n.s. n.s. n.s. n.s. .000 .000 .000 .000 .000 .000 .000 .000 n.s.

G.E. Johnsen et al. / Journal of Anxiety Disorders 22 (2008) 464–474

469

Fig. 1. Performance on the five successive learning trials on the CVLT, for the PTSD group and non-PTSD group (small bars depict the 95% confidence interval).

covariate. The group difference in total learning persisted when controlling for memory span (F (1,35) = 9.69, p = .004). Further, t-tests of retroactive interference, as defined by a decline in performance from learning trial 5 to short-delay free recall, and proactive interference, as defined by a decline in performance from learning trial 1 to learning list B (Yehuda et al., 1995), showed no differences between the groups. On recognition hits the two groups did not differ. The two groups differed on Digit Span backward, but not on Digit Span forward. No

significant group differences were seen on PASAT (see Table 2). Since the group differences on both age and education were insignificant, but relatively large, differences in repeated measures of CVLT and Digit Span backwards were analyzed controlling for age and education (ANCOVA). When age and education were introduced as covariates, the group differences on repeated measures of CVLT persisted (F (1,35) = 10.32, p = .003 and F (1,35) = 9.19, p = .005, respectively). After controlling for age and education on Digit Span

Table 2 Performance on the cognitive measures for the PTSD group and non-PTSD group

Digit span forward Digit span backward CVLT trial 1 CVLT trial 2 CVLT trial 3 CVLT trial 4 CVLT trial 5 CVLT1-5 (total learning) CVLT list B CVLT short delay free recall (SDFR) CVLT long delay free recall (LDFR) CVLT short delay cued recall (SDCR) CVLT long delay cued recall (LDCR) CVLT proactive interference CVLT retroactive interference CVLT retention (LDFR-SDFR) CVLT recognition hits PASAT 3 s PASAT 2 s

PTSD group (n = 21)

Non-PTSD (n = 21)

M

(S.D.)

M

(S.D.)

5.33 4.78 6.05 7.75 8.20 9.40 10.35 41.75 5.00 9.00 9.15 8.95 9.05 1.00 1.60 0.20 13.10 39.53 32.58

(2.93) (2.10) (2.46) (2.81) (3.66) (2.76) (3.36) (13.61) (2.53) (3.31) (3.23) (3.24) (3.17) (2.10) (1.70) (0.89) (3.14) (16.90) (14.60)

6.85 6.69 7.17 9.72 11.33 12.11 12.50 52.83 5.47 11.39 12.24 11.56 12.06 1.38 0.94 0.94 14.76 41.64 38.50

(0.90) (1.25) (2.07) (2.32) (2.40) (1.88) (1.89) (7.26) (1.77) (2.12) (2.33) (2.57) (2.56) (1.71) (1.43) (1.48) (1.64) (16.80) (16.69)

t-Value

1.79 2.92 1.51 2.34 3.08 3.50 2.39 3.08 0.64 2.62 3.27 2.73 3.14 0.58 1.28 1.85 1.96 0.36 1.08

Effect Size (d)

0.66 1.08 0.50 0.78 1.03 1.17 0.80 1.03 0.22 0.87 1.11 0.91 1.06 0.20 0.43 0.63 0.66 0.13 0.39

p

n.s. .007 n.s. .025 .004 .001 .022 .004 n.s. .013 .002 .010 .003 n.s. n.s. n.s. n.s. n.s. n.s.

470

G.E. Johnsen et al. / Journal of Anxiety Disorders 22 (2008) 464–474

Fig. 2. Performance on the two free recall tasks of the CVLT, short delay free recall (SDFR) and long delay free recall (LDFR) for the PTSD group and non-PTSD group controlling for total learning (small bars depict the 95% confidence interval).

backwards, the group differences persisted, indicating that this group difference could not be explained by individual differences in age and education (F (1,28) = 9.74, p = .004 and F (1,28) = 6.47, p = .02, respectively). 2.3. Relationship between estimated IQ, clinical measures and memory In order to analyze to the relative importance of the three IES-R subscales on the CVLT performance, a covariate analysis was done with the subscales of intrusion, hyperarousal and avoidance from the IES-R scale. When the intrusion scale was introduced as covariate the differences between the groups in total learning disappeared (F (1,35) = 2.61). When the hyperarousal and the avoidance scales were introduced, the overall group differences persisted (F (1,35) = 4.79, p = .035 and F (1,35) = 7.27, p = .01, respectively). The relation between PTSD and verbal learning performance may be influenced by a number of intervening variables. Estimated IQ, current depressive symptoms (MADRS) and general distress (SCL-90-R), in addition to the severity of PTSD symptoms (IES-R), may have contributed to the memory differences on CVLT. In order to analyze the potential contribution of these variables on repeated measures of CVLT, we performed both covariance analysis and multiple regression analysis. When controlling for MADRS in a covariance analysis, the group difference in repeated measures of CVLT disappeared (F (1,34) = 0.24). When controlling for estimated IQ and IES-R, the difference between the groups persisted (F (1,35 = 7.98, p = .007

and F (1,35) = 4.24, p = .046, respectively), indicating that the group differences could not be explained by individual differences in IQ or symptoms of traumatization, when the three subscales from IES-R were controlled for simultaneously. When controlling for SCL-90-R, the group effect was reduced to below the level of acceptance (F (1,32) = 4.10, p = .051). Pearson product moment correlations revealed significant correlations between the clinical measures. IES-R correlated with both MADRS (r = .76, p < .05) and SCL-90-R (r = .83, p < .05). MADRS correlated with SCL-90-R (r = .85, p < .05). Because of high intercorrelations between the clinical variables, a multiple regression analysis was performed to analyze which of the clinical variables best-predicted learning and memory alteration on the CVLT (CVLT 1–5). Since there is insufficient theoretic and empirical support for which of the clinical variables assumed to explain most of the variance, an exploratory stepwise multiple regression was chosen. In the stepwise regression, the following variables were entered, with criterion for inclusion set to p < .05; MADRS, IES-R, SCL-90-R and estimated IQ. Only depressive symptoms (MADRS) explained a significant part of the variance of CVLT and were therefore included in the model (R = .52, R2 = .27, b = 0.83, t = 2.63, p < .01). 3. Discussion The main finding of the present study is impaired verbal learning and memory problems associated with PTSD symptoms. The PTSD diagnosed refugees showed slower and less effective learning, and they

G.E. Johnsen et al. / Journal of Anxiety Disorders 22 (2008) 464–474

had impaired effect of rehearsal on the memory acquisition. The difference between the PTSD group and the non-PTSD group was not evident in the first trial of CVLT, but the learning differences between the two groups emerged over the following trials, suggesting decreased ability to elaborate the stimuli for a more efficient encoding. The lack of group differences on the first trial on CVLT and on Digit Span forward indicates that the learning problems cannot be explained by impaired memory span. In addition, the PTSD group reported fewer words in both short and long delay free recall, but when controlling for total learning these differences disappeared, indicating that the memory problems are connected to encoding and acquisition and not to retrieval. The impaired score on Digit Span backward indicates impaired working memory in the PTSD group, but this was not reflected by the performance on the PASAT. We had hypothesized that symptoms of intrusion and arousal would interfere with working memory in the PTSD group, in particular as evident on the PASAT. However, since we found no differences between the groups on the PASAT, the results from the working memory tests are inconclusive. Motivational aspects and stress in the test situation may explain why the two groups did not differ on this particular test. Our findings support the hypothesis of specific verbal memory disturbances associated with PTSD, and the effect is related to impaired encoding, but not retrieval. The memory impairment does not appear to be related to storage capacity, but is more likely to be a result of impaired processing and organization of the information. This interpretation is further supported by the finding that our two groups did not differ on recognition memory, which is more passive recollection of the memory material and to a lesser extent is dependent on organizational and retrieval strategies. This pattern of performance has been associated with frontal lobe dysfunction (Gershberg & Shimamura, 1995). Jenkins, Langlais, Delis, and Cohen (2000) also found digit span forward to be intact and backward to be impaired in patients with PTSD. Our findings are consistent with the Yehuda et al.’s (2004) study, in which Holocaust survivors with PTSD were compared to survivors with no PTSD diagnosis and non-exposed controls. This study showed significant differences on trials 1–5 on the CVLT, indicating problems related to the ability to encode and consolidate information. Compared to the most recent study performed by Yehuda et al. (2005) on aging combat veterans, our group effects are more pronounced than in their PTSD+ and PTSD groups. In this study Yehuda et al. does not

471

find differences in encoding between their two exposed groups, but between the PTSD+ group and the nonexposed controls. However, our findings are not consistent with the study by Yehuda et al. (1995) on cognitive dysfunction in a war veteran population. The veterans did not have problems on the encoding stage. Both groups showed similar learning curves on the first five trials on the CVLT, but the veterans showed problems at the retrieval stage, and the authors interpret this as problems due to retroactive interference. Jenkins et al. (1998) found differences in CVLT performance on number of words learned and on long-delayed recall. But in contrast to our findings, the PTSD group in that study improved their recall performance on the test in the cued recall condition. This difference in results could be explained by factors such as differences in extent of trauma exposure, but also by the fact that the veterans in the Yehuda et al.’s study are trained and selected for duty. In our study the covariance analysis showed that the intrusion subscale was the IES-R subscale which was most strongly related to CVLT performance, supporting the hypothesis that intrusive memories engage taskirrelevant processes and leave fewer resources available to complete the ongoing memory task. Vasterling et al. (1998) also found cognitive impairment in PTSD patients to be related to symptoms of reexperiencing rather than hyperarousal. The majority of our participants with PTSD had moderate self-reported depressive symptoms, as indicated by their elevated MADRS scores. Memory performance in the PTSD group appears to be related to depression, another common outcome of traumatic stress. PTSD is commonly associated with additional psychiatric diagnoses, in particular depressive disorder (Mollica, Wyshak, & Lavelle, 1987). Considerable evidence supports the hypothesis that depression interferes with tasks requiring effortful processing (Weingartner, 1986). In addition, Otto, Bruder, Fava, and Delis (1994) found that depressed outpatients performed between one-half to one standard deviation below normative standards on the CVLT. Poor use of encoding strategies and reduced levels of organization have also been found (Veiel, 1997). Studies on PTSD samples that have included clinical comparison samples have found less clear patterns of PTSD-specific impairment (Gil, Calev, Greenberg, Kugelmass, & Lerer, 1990; Zalewski, Thompson, & Gottesman, 1994). The interacting effects of co-existing PTSD and depression on cognitive functions need to be better understood. Most of the participants in the PTSD group

472

G.E. Johnsen et al. / Journal of Anxiety Disorders 22 (2008) 464–474

had relatively high scores on current depressive symptoms (MADRS), and this is assumed to be related to the exposure to multiple, sequential traumatic events involving threat to life as well as loss. In witnessing the killing of family members and friends, both loss and threat events occur simultaneously. Recruiting refugees without co-existing depressive symptoms may be difficult and would limit the generalizability of a study, and such individuals are not highly representative of PTSD-diagnosed populations (Kessler et al., 1995; McFarlane & Papay, 1992). It is possible that psychopathology in general is associated with poor performance on the CVLT or overlapping symptoms in PTSD and depression contributes to the memory problems. The inclusion of a clinical comparison group with a depressive disorder could have given valuable information concerning the specificity of the PTSDrelated impairments. We were not able to recruit a sufficient number of matched refugees with a major depression disorder to include such a control group in our study. Estimated IQ was not related to memory alterations in our samples, and this finding is consistent with other studies controlling for estimated IQ (Vasterling et al., 1998, 2002). However, our measures of general intellectual function did not predate the traumatic exposure. Thus we cannot exclude premorbid IQ as a cognitive risk factor for developing learning and memory dysfunctions in PTSD. However, care should also be taken when using a full scale IQ as a covariate, since IQ tests commonly also include measures of both short and long term memory. Our sample is relatively small, and future studies are needed to replicate the current findings in a larger sample. Our study shows some overlap on IES-R scores between the two groups. Some participants in the nonPTSD group report elevated IES-R scores, and some individuals in the PTSD group have low scores. Thus the comparison group reported similar stress reactions following the same traumatic events as the PTSD group, but they did not satisfy the DSM-IV criteria for a diagnosis of PTSD. The elevated scores in the control group may also be linked to situational factors. Low scores in the PTSD group may be related to different aspects or denial. These findings may question the validity of IES-R and self-report measures in differentiating between a PTSD population and an exposed non-PTSD population. Translating the different assessment scales and neuropsychological tests and the use of an interpreter in the clinical screening and test situations may have weakened the reliability and validity of the tests and

diagnostic scores. However, this kind of research would otherwise have been impossible to carry out, and valuable information would not have been collected. Efforts were made to minimize these effects by using the same interpreter within the same language groups. Also, since this study builds on a between-group comparison, and ethnicity was equally distributed in the two groups, these conditions affected both groups equally, and the impact of these factors is considered to be minimal. 4. Conclusions The present findings are important in suggesting impaired processing capacity, in particular evident in impaired encoding during the initial learning trials in refugees/immigrants with PTSD. The findings have implications for refugees’/immigrants’ adaptation, rehabilitation and education during resettlement. The observed impairments emphasize the importance of studying cognitive problems among refugees and war affected populations. Epidemiological studies have documented high rates of PTSD in these groups (Carlson & Rosser-Hogan, 1991), and they constitute groups in considerable need of diagnostic screening and effective treatment programs. Memory impairments may affect the developmental course of the disorder and the response to treatment. Acknowledgments This study was financially supported from the Research Council of Norway, L. Meltzer Høyskolefond, and the University of Bergen. References American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders: DSM-IV (4th ed.). Washington, DC (Author). Barrett, D. H., Green, M. L., Morris, R., Giles, W. H., & Croft, J. B. (1996). Cognitive functioning and posttraumatic stress disorder. American Journal of Psychiatry, 153(11), 1492–1494. Blake, D. D., Weathers, F. W., Nagy, L. M., Kaloupek, D. G., Charney, D. S., & Keane, T. M. (1998). Clinican-Administered PTSD scale for DSM-IV, Revised (CAPS). National Center for Posttraumatic Stress Disorder, Behavioral Science Division, Boston VA Medical Center & Neuroscience Division, West Haven VA, Medical Center. Bremner, J. D. (2002). Does stress damage the brain? Understanding trauma-related disorders from a mind-body perspective New York: WW Norton & Company Ltd. Bremner, J. D., Randall, P., Scott, T. M., Capelli, S., Delaney, R., McCarthy, G., et al. (1995). Deficits in short-term memory in adult survivors of childhood abuse. Psychiatry Research, 59(1–2), 97– 107.

G.E. Johnsen et al. / Journal of Anxiety Disorders 22 (2008) 464–474 Bremner, J. D., Scott, T. M., Delaney, R. C., Southwick, S. M., Mason, J. W., Johnson, D. R., et al. (1993). Deficits in short-term-memory in posttraumatic-stress-disorder. American Journal of Psychiatry, 150(7), 1015–1019. Burt, D. B., Zembar, M. J., & Niederehe, G. (1995). Depression and memory impairment: A meta-analysis of the association, its pattern, and specificity. Psychological Bulletin, 117(2), 285–305. Carlson, E. B., & Rosser-Hogan, R. (1991). Trauma experiences, posttraumatic stress, dissociation, and depression in Cambodian refugees. American Journal of Psychiatry, 148(11), 1548–1551. Cohen, J. (1969). Statistical power analysis for the behavioral sciences. New York: Academic Press. Delis, D. C., Kramer, J. H., Kaplan, E., & Ober, B. A. (1987). The California verbal learning test: Research edition. New York: Psychological Corporation. Derogatis, L. R. (1983). SCL-90-R: Administration, scoring, and procedures manual-II (2nd ed.). Towson, MD: Clinical Psychometric Research. Egeland, J., Sundet, K., Rund, B. R., Asbjørnsen, A., Hugdahl, K., Landrø, N. I., et al. (2003). Sensitivity and specificity of memory dysfunction in schizophrenia: A comparison with major depression. Journal of Clinical and Experimental Neuropsychology, 25(1), 79–93. Gershberg, F. B., & Shimamura, A. P. (1995). Impaired use of organizational strategies in free recall following frontal lobe damage. Neuropsychologia, 33(10), 1305–1333. Gil, T., Calev, A., Greenberg, D., Kugelmass, S., & Lerer, B. (1990). Cognitive functioning in post-traumatic stress disorder. Journal of Traumatic Stress, 3(1), 29–45. Gilbertson, M. W., Gurvits, T. V., Lasko, N. B., Orr, S. P., & Pitman, R. K. (2001). Multivariate assessment of explicit memory function in combat veterans with posttraumatic stress disorder. Journal of Traumatic Stress, 14(2), 413–432. Gronwall, D. M. (1977). Paced auditory serial-addition task: A measure of recovery from concussion. Perceptual and Motor Skills, 44(2), 367–373. Hytten, K. (1989). Helicopter crash in water: Effects of simulator escape training. Acta Psychiatrica Scandinavica, 80(355, Suppl), 73–78. Jenkins, M. A., Langlais, P. J., Delis, D., & Cohen, R. (1998). Learning and memory in rape victims with posttraumatic stress disorder. The American Journal of Psychiatry, 155(2), 278–279. Jenkins, M. A., Langlais, P. J., Delis, D. A., & Cohen, R. A. (2000). Attentional dysfunction associated with posttraumatic stress disorder among rape survivors. The Clinical Neuropsychologist, 14(1), 7–12. Johnsen, B. H., Eid, J., Løvstad, T., & Michelsen, L. T. (1997). Posttraumatic stress symptoms in nonexposed, victims, spontaneous rescuers after an avalanche. Journal of Traumatic Stress, 10(1), 133–140. Kessler, R. C., Sonnega, A., Bromet, E., & Hughes, M. (1995). Posttraumatic stress disorder in the national comorbidity survey. Archives of General Psychiatry, 52(12), 1048–1060. Kivling-Bode`n, G., & Sundbom, E. (2003). Cognitive abilities related to post-traumatic symptoms among refugees from the former Yugoslavia in psychiatric treatment. Nordic Journal of Psychiatry, 57(3), 191–198. Kolb, L. C. (1987). A neuropsychological hypothesis explaining posttraumatic stress disorders. American Journal of Psychiatry, 144(8), 989–995. Lezak, M. D. (1995). Neuropsychological assessment (3 ed.). New York: Oxford University Press.

473

Macklin, M. L., Metzger, L. J., Litz, B. T., McNally, R. J., Lasko, N. B., Orr, S. P., et al. (1998). Lower precombat intelligence is a risk factor for posttraumatic stress disorder. Journal of Consulting and Clinical Psychology, 66(2), 323–326. McFarlane, A. C., & Papay, P. (1992). Multiple diagnoses in posttraumatic stress disorder in the victims of a natural disaster. Journal of Nervous and Mental Disease, 180(8), 498–504. McNally, R. J. (2003). Progress and controversy in the study of posttraumatic stress disorder. Annual Review of Psychology, 54(1), 229–252. Mollica, R. F., Wyshak, G., & Lavelle, J. (1987). The psychosocial impact of war trauma and torture on Southeast Asian refugees. American Journal of Psychiatry, 144(12), 1567–1572. ˚ sberg, M. (1979). A new depression scale Montgomery, S. A., & A designed to be sensitive to change. British Journal of Psychiatry, 134, 382–389. Netland, M., & Kanaaneh, M. (1989). Tests and questionnaires for the research project ‘‘Children and political violence’’. Unpublished manuscript, Bergen, Norway: University of Bergen. Otto, M. W., Bruder, G. E., Fava, M., & Delis, D. C. (1994). Norms for depressed patients for the California Verbal Learning Test: Associations with depression severity and self-report of cognitive difficulties. Archives of Clinical Neuropsychology, 9(1), 81–88. Sheehan, D. V., Lecrubier, Y., Sheehan, K., Amorim, P., Janavs, J., Weiller, E., et al. (1998). The Mini-International Neuropsychiatric Interview (MINI): The development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. Journal of Clinical Psychiatry, 59(Suppl 20), 22–33. Uddo, M., Vasterling, J. J., Brailey, K., & Sutker, P. B. (1993). Memory and attention in combat-related post-traumatic stress disorder (PTSD). Journal of Psychopathology and Behavioral Assessment, 15(1), 43–52. Van der Kolk, B. A., Roth, S., Pelcovitz, D., Sunday, S., & Spinazzola, J. (2005). Disorders of extreme stress: The empirical foundation of a complex adaptation to trauma. Journal of Traumatic Stress, 18(5), 389–399. Vasterling, J. J., Brailey, K., Constans, J. I., & Sutker, P. B. (1998). Attention and memory dysfunction in posttraumatic stress disorder. Neuropsychology, 12(1), 125–133. Vasterling, J. J., Duke, L. M., Brailey, K., Constans, J. I., Allain, A. N., & Sutker, P. B. (2002). Attention, learning, and memory performances and intellectual resources in Vietnam veterans: PTSD and no disorder comparisons. Neuropsychology, 16(1), 5–14. Veiel, H. O. F. (1997). A preliminary profile of neuropsychological deficits associated with major depression. Journal of Clinical and Experimental Neuropsychology, 19(4), 587–603. Wechsler, D. (1981). Wechsler adult intelligence scale—revised. New York: The Psychological Corporation. Weingartner, H. (1981). Cognitive processes in depression. Archives of General Psychiatry, 38(1), 42–47. Weingartner, H. (1986). Automatic and effort-demanding cognitive processes in depression. In: L. W. Poon (Ed.), Handbook for clinical memory assessment of older adults (pp. 218–225). Washington, DC: American Psychological Association. Weiss, D. S., & Marmar, C. R. (1997). The impact of event scale— revised. In: J. P. Wilson & T. M. Keane (Eds.), Assessing psychological trauma and PTSD: A practitioner’s handbook (pp. 399–411). New York: Guildford Press. Yehuda, R., Golier, J., Tischler, L., Stavitsky, K., & Harvey, P. D. (2005). Learning and memory in aging combat veterans with PTSD. Journal of Clinical and Experimental Neuropsychology, 27, 504–515.

474

G.E. Johnsen et al. / Journal of Anxiety Disorders 22 (2008) 464–474

Yehuda, R., Golier, J. A., Halligan, S. L., & Harvey, P. D. (2004). Learning and memory in holocaust survivors with posttraumatic stress disorder. Biological Psychiatry, 55(3), 291–295. Yehuda, R., Keefe, R. S. E., Harvey, P. D., Levengood, R. A., Gerber, D. K., Geni, J., et al. (1995). Learning and memory in combat

veterans with posttraumatic stress disorder. American Journal of Psychiatry, 152(1), 137–139. Zalewski, C., Thompson, W., & Gottesman, I. (1994). Comparison of neuropsychological test performance in PTSD, generalized anxiety disorder, and control Vietnam veterans. Assessment, 1(2), 133–142.