Trauma-Induced Weight Loss and Cognitive Deficits among Former Prisoners of War.

Acknowledgement: This work was supported by a Veterans Administration Medical Research Award to Patricia B. Sutker.
The authors are grateful for the consultation and assistance of Nelson M. Butters, John E. Mendoza, and Franlynn Bugg.

It is well documented that exposure to severe and prolonged stress exacts demonstrable toll on psychosocial adjustment and biological functioning. Studies of Nazi concentration camp survivors and former prisoners of war (POWs) reveal marked psychological and medical residuals that do not appear to remit with passage of time. Pioneering research by Nefzger (1970), Beebe (1975), and Keehn (1980) indicated that relative to comparison samples of non-POW combat veterans, World War II (WWII) and Korean Conflict POWs evidenced significantly increased rates of physical disability, morbidity, and mortality including enhanced risk for infectious, degenerative, and cardiovascular diseases and psychiatric disorders up to 35 years postinternment. Shortly after WWII, Wolf and Ripley (1947) pointed to alterations in memory and learning among survivors of Japanese imprisonment, and more recently White (1983) and Arthur (1982) documented POW complaints of problems with memory, attention, and concentration.

Several early studies provide evidence of neuropsychological impairment among WWII POWs who suffered extreme mistreatment during imprisonment at the hands of the Japanese, and researchers suggested that central nervous system (CNS) damage may contribute to performance and adaptation difficulties found among some former POWs. Kral, Pazder, and Wigdor (1967) compared 20 Canadian POWs held by the Japanese for approximately 44 months with 20 of their male siblings seeing similar WWII service using the Wechsler-Bellevue Intelligence Scale I and other measures in combination with interview reports of psychiatric and neurological symptoms. Their findings showed increased rates of CNS and peripheral neurological impairments as well as more exaggerated presentation of affective, psychosocial, and anxiety-related complaints among former POWs compared with their brothers. Klonoff, McDougall, Clark, Kramer, and Horgan (1976) contrasted groups of veterans held as prisoners of the Japanese and Germans across seven physiological and psychiatric dimensions and documented neuropsychological impairments that could reflect cerebral or other CNS pathology. Both groups exhibited residual problems that correlated with internment length, and these deficits were significantly greater among the Japanese-held men.

More recently, Sutker, Allain, and Winstead (1987) described deficits on Wechsler Adult Intelligence Scale–Revised (WAIS–R) Digit Span, Similarities, Digit Symbol, and Object Assembly subtests compared with mean Verbal and Performance scores among 80 former POWs undergoing clinical assessment. Although replication and extension using suitable control subjects are needed, findings are consistent with POW self-reports of problems with attention, concentration, and memory and suggest that subtle acquired cortical dysfunction may contribute to the constellation of common residuals of severe psychological and biological stress despite an absence of gross intellectual compromise. Interestingly, Sutker, Allain, and Motsinger (1988) found that 78% and 93% of two Minnesota Multiphasic Personality Inventory (MMPI) defined subgroups of WWII and Korean former POWs cited problems with concentration and memory.

There have been few research attempts to describe the symptoms of cognitive dysfunction among former POWs in relation to appropriate control veteran samples and to determine the extent to which lasting intellectual impairments might be associated with the parameters of the stress experience. Complementing the work of Klonoff and his colleagues (1976), Beebe (1975) and Keehn (1980) showed that subsequent appearance of psychiatric symptomatology among WWII or Korean POWs was positively associated with site or nature of confinement, severity of physical disease, and estimated weight loss during captivity. Although Beebe (1975) failed to find evidence of CNS damage among former POWs relative to combat control subjects, Thygesen, Hermann, and Willanger (1970) pointed to symptoms of organic brain damage among Danish survivors of Nazi concentration camps who lost more than 35% of their preconfinement body weight. Results of this 23-year follow-up indicate that CNS residuals may reach maximum significance among veterans who experienced the greatest biological and psychological stress. Therefore, degree of self-reported individual weight loss during captivity, serving as a composite index of physical and mental hardship, may constitute a predictor of subsequent adjustment difficulties, including symptoms of intellectual and personality dysfunction.

Our study compared two subgroups of former POWs, differentiated by the weight-loss parameter suggested by Thygesen et al. (1970) as predictive of organic brain residuals, and combat veterans who were not captured and interned, on WAIS–R and Wechsler Memory Scale Logical Memory measures. It was hypothesized that those former prisoners subjected to the most severe psychological and biological stress, conceptualized by amount of trauma-induced weight loss, would show the least effective performances on cognitive tasks compared with their low weight-loss counterparts and combat veterans. It was also predicted that former POWs who reported less significant weight loss during confinement would demonstrate poorer efforts than combat veterans.

Participants were 173 former POWs and 50 veterans who served in combat during WWII or the Korean Conflict. Former POWs were selected from 202 men who responded to a Veterans Administration invitation for comprehensive medical and psychiatric evaluation extended to POW survivors by Public Law 97-37. Constituting more than half of POW survivors residing in the New Orleans, Louisiana Medical Center catchment area, veterans were not necessarily seeking VA health-care services, and approximately 40% of the participants customarily received medical care elsewhere. The sample comprised 180 WWII veterans, 21% of whom were interned by the Japanese, and 22 Korean Conflict POW survivors. Of 202 volunteers presenting for the psychological assessment component of the medical protocol, 29 were excluded from the study because of protocol noncompletion (n = 15) , self-report of severe head trauma, neurosurgery, or other CNS disease during confinement (n = 3) , current active alcoholism defined by Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev.; DSM-III–R,American Psychiatric Association, 1987) criteria (n = 3) , and postconfinement history of seizure disorder, cerebrovascular accident, or neurosurgery (n = 8) . Men reporting beatings with blows to the head from rifle butts and other weapons were not eliminated from the study sample unless they described more than momentary loss of consciousness, because these were common occurrences during POW confinement, regardless of weight-loss magnitude.

Men who served in combat during WWII and the Korean Conflict but were not taken captive served as control subjects. They were recruited through letters inviting participation in a study of psychological residuals to combat and POW experiences, which were mailed to individuals selected by computer-generated random sequence from currently served veteran patients on VA Medical Center rolls as well as to Commanders of local Veterans of Foreign Wars and American Legion Posts for display on bulletin boards. Approximately two thirds (66%) of volunteers constituted regular VA patients. The sample comprised 41 WWII veterans, 54% of whom saw combat in the Pacific Theater of Operations (PTO), and 21 Korean Conflict servicemen. Of the 62 veterans who volunteered to undergo assessment requirements, 12 were excluded from the study because of protocol noncompletion (n = 2) , history of cerebrovascular accident (n = 3) , current active alcoholism (n = 4) , and behavioral evidence of gross intellectual decline (n = 3) .

As might be expected in an aging sample, prescription use of medications was not uncommon, and exclusion of potential participants for medicinal use of drugs would have drastically reduced the numbers available for study. Therefore, prescribed medications were recorded and classified into the following categories: no current medications (30%), antianxiety agents or medications for sleep (33%), antidepressants (10%), antihypertensive/cardiac medications (39%), analgesic/anti-inflammatory agents (10%), medications for diabetes (6%), and neuroleptics (0%). POW and combat veteran subgroups did not differ significantly in medication use by types, including antidepressant drugs with anticholinergic effects, which, as dose increases, may be associated with sedation or result in problems with concentration and recall (Salzman, Hoffman, & Schoonover, 1983).

Most of the former POWs reported significant weight loss during POW captivity, and they provided weight estimate figures for precapture and lowest confinement weights. The sample was divided into high (n = 60) and low (n = 113) weight-loss subgroups using the criterion suggested by Thygesen et al. (1970) in their 23-year follow-up of Nazi concentration camp survivors (i.e., loss of preconfinement weight > 35% and ≤35%, respectively). Weight-loss categorization of POW survivor subgroups was judged to be the most appropriate test of Thygesen et al.’s (1970) hypothesis of a critical weight-loss threshold concept and of the notion that degree of hunger dystrophy or body wasting is a parsimonious reflection of the combined effects of psychological and biological trauma, an index quantifying combined and confounded effects of confinement length, physical brutality, psychological harassment, and biological insults of starvation and lack of medical care. Although length of confinement may be associated with semistarvation, captor brutality, and weight loss, Thygesen et al. (1970) found severe confinement weight loss to be the best predictor of organic brain residuals, which is the focusof this study. Mean reported percentage weight losses for the high and low subgroups were 42.87 ± 5.81 and 22.51 ± 8.29, respectively. Of those men composing the high weight-loss subgroup, 27% were prisoners of the Japanese, 50% of the Germans, and 23% of the North Koreans/Chinese. In contrast, 82% of those assigned to the lower weightloss subgroup were held by the Germans, 11% by the Japanese, and 7% by the North Koreans/Chinese.

Two examiners independently inspected military medical records housed at the VA Regional Office for evidence to confirm or disconfirm POW retrospective estimates of precapture and lowest confinement body weights. Corroborative materials included summaries of repatriation, induction, discharge, and inservice physical examinations as well as results of the medical protocol and other correspondence bearing on the POW experience. More than half (n = 76) of available files (n = 145) contained sufficient information to corroborate veteran reports of weight loss (i.e., included precapture or confinement weight-loss figures), and raters achieved 100% agreement for group assignments using these materials. In two cases, materials were found to influence weightloss categorization, and in both instances the discrepancy involved less than five percentage points in men whose weight loss was near the cutting point for group assignment. These data suggested that in most instances POW self-reports of weight loss appear to be accurate for defining weight-loss subgroups as well as estimating total number of pounds lost during confinement.

The majority of men in the three groups were White (82%), right-handed (97%), and of middle/lower-middle socioeconomic status (65%) defined by Hollingshead-Redlich (1958) criteria, with mean age and education of 63.85 ± 4.90 and 11.70 ± 3.64 years, respectively, at assessment. Former POW and combat veterans did not differ in formal schooling, racial distribution, or socioeconomic status (ps > .05) , but mean age differed across groups,F (2, 220) = 5.22, p < .01, with the low weight-loss subgroup (M = 64.84 ± 3.86 years) being older than their high weight-loss counterparts (M = 63.20 ± 5.51 years) and the combat veterans (M = 62.40 ± 5.76 years) . Among the POWs, high weight-loss men were confined longer, 22.18 ± 15.26 months versus 12.19 ± 10.75 months,F (1, 171) = 25.05, p < .01, and experienced more severe captivity mistreatment reflected by the POW Trauma Index (Sutker, Winstead, Goist, Malow, & Allain, 1986); 11.18 ± 3.24 versus 8.12 ± 3.21,F (1, 171) = 35.37, p < .01. Veteran groups differed in theater of war operations, χ2(4, N = 223) = 54.69, p < .01, and combat exposure,F (2, 220) = 63.79, p < .01. POW survivors who saw duty in the WWII Pacific Theater of Operations (PTO) or Korea were more likely to be assigned to the high weight-loss subgroup, whereas POWs who served in the WWII European Theater of Operations (ETO) were more numerous among the low weight-loss subgroup. Combat veterans were relatively evenly distributed across theaters of operations. Non-POW veterans reported longer combat exposure (16.46 ± 10.95 months) than low (6.02 ± 4.37 months) and high weight-loss (4.06 ± 2.95 months) POWs who did not differ on this dimension.

Participants were administered the WAIS–R and Logical Memory indices of the WMS under conditions of immediate and 30-min delayed recall. WAIS–R subtest data were transformed from raw to scale scores using age-appropriate norms, and the WAIS–R derived factors of Verbal–Comprehension, Attention–Concentration, and Perceptual–Organization described by Witkin and Goodenough (1981) were extracted. WMS scores were converted to scale scores using Russell’s (1988) normative adjustments for age, education, and intelligence, resulting in a range of values from above average (0) to profound impairment (6). Former POWs were also administered a structured background information questionnaire incorporating POW Trauma Index (Sutker et al., 1986) items, the MMPI including the Posttraumatic Stress Disorder (PTSD) scale of Keane, Malloy, and Fairbank (1984), the Beck Depression Inventory (BDI; Beck, Ward, Mendelson, Mock, & Erbaugh, 1961), and the A-Trait scale of the State-Trait Anxiety Inventory (Spielberger, Gorsuch, & Lushene, 1970).

Cognitive comparisons on WAIS–R and WMS derived scores were performed using multivariate analyses of variance followed by univariate analyses and Newman-Keuls tests as appropriate. Men assigned to the low weight-loss subgroup were older than their high weight-loss counterparts and veteran controls, but age was not included as a covariate in data analysis because both WAIS–R and WMS scores were adjusted for age using standardized norms. Theater of war operations and length of POW confinement were examined as possible covariates, but neither of these factors proved significant in multivariate analyses of covariance of WAIS–R variables, Witkin-Goodenough factors, or WMS indices. Chi-square analyses were used to compare POW subgroups on items composing the POW Trauma Index.

Canonical correlation analysis was conducted to evaluate possible relationships among cognitive performance variables and measures of general psychological distress or psychopathology, or predictability of scores on one set of measures from scores on the other, and to determine the number and nature of different dimensions of association between the two variable domains (Cliff, 1987). This analysis was performed with approximately 85% of the participant pool, or those men for whom the full set of psychopathology indices were available. The cognitive variables included WAIS–R and WMS subtest scores, and those reflecting psychopathology consisted of the BDI, A-Trait, Keane et al. (1984) PTSD scale, and Skinner and Lei’s (1980) MMPI-derived elevation parameter, the latter considered to reflect overall level of symptom severity or psychological disturbance.

The three veteran subgroups differed significantly across the sets of WAIS–R derived measures, including the summary indices,F (6, 436) = 3.06, p < .01; subtest scores,F (22, 420) = 1.77, p < .05; and Witkin-Goodenough factors,F (6, 436) = 2.89, p < .05. As can be seen in Table 1, former POWs who reported loss of more than 35% of their preconfinement weights showed significantly less proficient performances than did combat veterans on the WAIS–R Performance IQ summary score, Arithmetic, Similarities, and Picture Completion subtests, and Attention–Concentration Factor. They were deficient relative to their low weight-loss counterparts on Arithmetic and Attention–Concentration Factor. Former POWs classified in the low weight-loss subgroup did not differ significantly from the combat controls on any of the WAIS–R derived measures. Review of the data presented in Table 1 also reveals that high weight-loss POWs scored lower than both other veteran subgroups across the entire range of cognitive measures included in this investigation.

POW and combat control veterans differed on the Logical Memory indices of the WMS,F (4, 438) = 4.93, p < .01. Former POWs assigned to the high weight-loss subgroup performed more poorly on measures of immediate memory than did those in the low weight-loss subgroup, who were in turn less proficient than were combat veterans. High weight-loss POWs also performed more poorly than combat veterans after a period of delay. Application of Russell’s (1988) formula for differentiating normal and brain-damaged performances showed significant differences for both immediate and delayed conditions, χ2s(2, N = 223) = 9.81 and 13.62, respectively, ps < .01. Under immediate recall, classification as impaired or compromised was observed for 68% of the high weight-loss POW subgroup and 51% of the low weight-loss men compared with 38% of combat veterans. Corresponding percentages were 63%, 42%, and 28%, respectively, for delayed recall.

Results of canonical correlation analysis revealed a significant interrelation between scores on cognitive and personality measures. A significant chi-square was observed with all canonical correlations included, χ2(52, N = 190) = 72.99, p < .05, but subsequent chi-square tests were not statistically significant. Thus, the first canonical correlation of .47 accounted for the significant interrelations between the two sets of variables, or 22% of the variance. Structure correlations and percentages of variance accounted for by the canonical variates within their corresponding variable set and within the other variable set (i.e., redundancies; Cliff, 1987) are presented in Table 2. Application of Tabachnick and Fidell’s (1983) cutoff correlation of ≥ / .30/ for interpretation indicated that WMS Delayed Recall was the only variable not relevant to its respective canonical variate. In general, as scores increased within the personality domain, scores decreased on WAIS–R subtests and increased (i.e., became more impaired) on WMS indices. The canonical variate for the cognitive variables accounted for 31% of the variance in the cognitive domain and 7% of the variance in the personality measures, whereas the canonical variate for the personality indices accounted for 76% of the variance in the personality measures and 17% of the variance in the cognitive variables.

Self-reported confinement stress and hardship were endorsed as significantly greater among POW survivors who lost more than 35% of their body weight during captivity, reflected by the type and frequency of stressor events composing the POW Trauma Index. As may be seen in Table 3, higher percentages of men assigned to the high weight-loss subgroup described being wounded at capture, having contracted illnesses during POW confinement, and suffering beatings, and physical torture, as well as individual and group death threats. However, more than two thirds of both POW survivor groups were exposed to freezing temperatures without adequate protection, forced marches, intimidation, interrogations, and overcrowding, and almost half of the overall sample witnessed fellow servicemen tortured.

Present findings support the Thygesen et al. (1970) hypothesis that severity of confinement stress, reflected parsimoniously by an index of hunger dystrophy or trauma-induced weight loss, is associated with lessened intellectual efficiency among former POW captives. As predicted, high weight-loss POW survivors performed significantly more poorly than did veterans seeing similar military combat on tasks requiring attention, concentration, and memory as well as on measures demanding such higher order, problem-solving functions as abstraction and organization. The three veteran groups differed least on measures dependent upon verbal comprehension and expression, fund of knowledge, and vocabulary resources, tasks which are thought to be relatively impervious to the effects of subtle or generalized cortical dysfunction. Although the magnitude of differences among the three groups is not dramatic, significant findings and group mean trends are compatible with the hypothesis of lessened intellectual performance among men exposed to prolonged brutality and extreme, inescapable psychological and biological hardship.

Significant differences were found between the two POW survivor subgroups on attention–concentration and WMS immediate recall tasks; but contrary to prediction, POWs experiencing less significant weight loss performed similarly to combat veterans, differing only on WMS immediate recall. Application of Russell’s recommended cutoff scores for classifying memory impairment suggests deficits among both POW subgroups, but scores among the high weight-loss men were relatively more impaired. The finding of deficits on measures of immediate recall, combined with lowered Arithmetic and Attention–Concentration Factor performances, most apparent among the high weight-loss POW survivors, raises the possibility that memory problems cited frequently in POW self-reports and clinical descriptions may be more attributable to deficiencies in attention, concentration, and perhaps organizing functions than to storage or retrieval processes.

Norwegian (Eitinger, 1980) and Danish (Thygesen et al., 1970) investigators described a pattern of neurological findings and intellectual impairment among Nazi concentration camp survivors. Although not focusing on one aspect of prolonged traumatization to exclusion of another, these researchers pointed to a relation between physical trauma and clinical evidence of organic brain syndrome. For example, Eitinger (1980) identified a subgroup of Norwegian concentration camp survivors with circumscribed features suggestive of chronic brain syndrome judged to be residual to captivity stress, including chronically elevated anxiety and intellectual impairment exhibited as problems with memory, concentration, concreteness, and poverty of ideas found in association with a multiplicity of diseases affecting gastrointestinal, cardiovascular, and skeletal systems. Among the confinement factors hypothesized to contribute to intellectual impairment, Eitinger (1980) cited weight loss and longstanding undernourishment, as well as brutal torture experiences and beatings to the head and other body parts. Present findings are supportive of this conclusion.

Although the majority of POWs were subjected to exposure to extreme temperatures, forced marches, interrogations, intimidations, and overcrowding, those reporting more than 35% total body weight loss described more severe biological and psychological insult, including being wounded at capture, troubled by confinement illnesses, subjected to personal and generalized death threats, and punished by beatings. Almost half of the high weight-loss subgroup reported having endured torture, including mock executions, beatings to head and soles of the feet, solitary confinement, and forced standing. It might be argued that findings merely reflect head trauma; however, elimination of those participants who experienced more than momentary lapses in consciousness makes this explanation less compelling. Rather, it appears that the observed symptom constellation is multifaceted in origins, possibly encompassing insult to the organism derived from nutritional deprivation and avitaminosis, untreated medical diseases, and psychological torture. Indeed, Goldfeld, Mollica, Pesavento, and Faraone (1988) contended that psychological symptoms among victims of torture may be secondary to CNS dysfunction rather than to the psychological effects of the experiences themselves.

Beebe (1975) and Keehn (1980) showed that psychiatric symptomatology among WWII and Korean POWs was positively correlated with nature of confinement, severity of physical disease, and estimated captivity weight loss. More recently, Engdahl (1987) and Speed, Engdahl, Schwartz, and Eberly (1989) reported that POW captivity weight loss was correlated with increased risk for development of depression and symptoms of chronic PTSD. Such psychological factors as higher levels of emotional and behavioral disturbance may be associated with reduced cognitive efficiency among subsets of former POW survivors. However, present data show that, despite shared variance among measures of psychopathology and cognitive functioning, differences among groups cannot be attributed solely to gross psychological disturbance; and disrupted concentration, attention, memory, and higher order cognitive functions are probably most reasonably linked to an interdependence of both psychological and biological factors. Present findings, however, may not be interpreted to rule out the possibility that preexisting person variables, such as coping ability and style or intellectual sophistication, could influence response to captivity stress, such as degree of confinement weight loss, and thus postconfinement adaptive functioning.

Prolonged conditions of extreme hardship, semistarvation, untreated medical diseases, and physical brutalization may be expected to produce a complex symptom pattern characterized by an interaction of psychological and medical residuals traceable to biological insult and mental duress. It is not our intention to point to a single determinant of late-life POW sequelae or to presume to unravel by retrospective analysis the intricate etiology of chronic residuals to severe stress. However, our findings add to the body of literature suggesting that reduced intellectual efficiency, and perhaps specific cognitive deficits, may contribute to difficulties in psychosocial adaptation exhibited by that subset of POW survivors subjected to the most severe trauma. This work extends our earlier effort to characterize the problem-solving performances of former POWs (Sutker et al., 1987) by inclusion of appropriate control subjects, verification of self-reported weight loss using military records, and assessment of the interrelations between psychological distress/psychopathology and cognitive deficits. Additional research is required to describe more specifically memory, concentration, efficiency, and organization problems residual from POW confinement and to determine the extent to which functional and structural CNS changes may derive from biological and psychological insult. Brain studies using computerized tomography and magnetic resonance imaging may shed more light on the possibility of brain abnormalities associated with chronic malnutrition, disease, avitaminosis, and other factors. In any case, results underscore the need for comprehensive clinical assessment of the varied symptoms and physical findings among former POW subgroups and development of therapeutic strategies to address problems identified among these men who were exposed to gruesome trauma over many months.

American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders (, rev.) (3rd ed.). Washington, DC: Author.

Arthur, R. J. (1982). Psychiatric syndromes in prisoner of war and concentration camp survivors. In C. T. H.Friedman & R. A.Faguet (Eds.), Extraordinary disorders of human behavior (pp. 47–63). New York: Plenum Press.

Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory for measuring depression. Archives of General Psychiatry, 4, 561–571.

Beebe, G. W. (1975). Follow-up studies of World War II and Korean War prisoners: 2. Morbidity, disability, and maladjustments. American Journal of Epidemiology, 101, 400–422.

Cliff, N. (1987). Analyzing multivariate data. New York: Harcourt Brace Jovanovich.

Eitinger, L. (1980). The concentration camp syndrome and its late sequelae. In J. E.Dimsdale (Ed.), Survivors, victims, and perpetrators: Essays on the Nazi holocaust (pp. 127–162). Washington, DC: Hemisphere.

Engdahl, B. E. (1987, August). Psychological consequences of the World War II prisoner of war experience. Paper presented at the 95th Annual Convention of the American Psychological Association, New York.

Goldfeld, A. E., Mollica, R. F., Pesavento, B. H., & Faraone, S. V. (1988). The physical and psychological sequelae of torture: Symptomatology and diagnosis. Journal of the American Medical Association, 259, 2725–2729.

Hollingshead, A. B., & Redlich, F. C. (1958). Social class and mental illness. New York: Wiley.

Keane, T. M., Malloy, P. F., & Fairbank, J. A. (1984). Empirical development of an MMPI subscale for the assessment of combat related posttraumatic stress disorder. Journal of Consulting and Clinical Psychology, 52, 888–891.

Keehn, R. J. (1980). Follow-up studies of World War II and Korean Conflict prisoners: 3. Mortality to January 1, 1976. American Journal of Epidemiology, 111, 194–211.

Klonoff, H., McDougall, G., Clark, C., Kramer, P., & Horgan, J. (1976). The neuropsychological, psychiatric, and physical effects of prolonged and severe stress: 30 years later. Journal of Nervous and Mental Disease, 163, 246–252.

Kral, V. A., Pazder, L. H., & Wigdor, B. T. (1967). Long-term effects of a prolonged stress experience. Canadian Psychiatric Association Journal, 12, 175–181.

Nefzger, M. D. (1970). Follow-up studies of World War II and Korean War prisoners: 1. Study plan and mortality findings. American Journal of Epidemiology, 91, 123–138.

Russell, E. W. (1988). Renorming Russell’s version of the Wechsler Memory Scale. Journal of Clinical and Experimental Neuropsychology, 10, 235–249.

Salzman, C., Hoffman, S. A., & Schoonover, S. C. (1983). Geriatric psychopharmacology. In E. L.Bassuk, S. C.Schoonover, & A. J.Gelenberg (Eds.), The practitioner’s guide to psychoactive drugs (2nd ed., (pp. 293–311). New York: Plenum Press.

Skinner, H. A., & Lei, H. (1980). Modal profile analysis: A computer program for classification research. Educational and Psychological Measurement, 40, 769–772.

Speed, N., Engdahl, B., Schwartz, J., & Eberly, R. (1989). Posttraumatic stress disorder as a consequence of the POW experience. Journal of Nervous and Mental Disease, 177, 147–153.

Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E. (1970). STAI: Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press.

Sutker, P. B., Allain, A. N., & Motsinger, P. A. (1988). Minnesota Multiphasic Personality Inventory (MMPI) derived psychopathology subtypes among former prisoners of war (POWs): Replication and extension. Journal of Psychopathology and Behavioral Assessment, 10, 129–140.

Sutker, P. B., Allain, A. N., & Winstead, D. K. (1987). Cognitive performances in former WWII and Korean-Conflict POWs. VA Practitioner, 4(6), 77

Sutker, P. B., Winstead, D. K., Goist, K. C., Malow, R. M., & Allain, A. N. (1986). Psychopathology subtypes and symptom correlates among former prisoners of war. Journal of Psychopathology and Behavioral Assessment, 8, 89–101.

Tabachnick, B. G., & Fidell, L. S. (1983). Using multivariate statistics. New York: Harper & Row.

Thygesen, P., Hermann, K., & Willanger, R. (1970). Concentration camp survivors in Denmark: Persecution, disease, disability, compensation. Danish Medical Bulletin, 17, 65–108.

White, N. S. (1983). Posttraumatic stress disorder [Letter to the editor]. Hospital and Community Psychiatry, 34, 1061–1062.

Witkin, H., & Goodenough, D. R. (1981). Cognitive styles: Essence and origins: Field dependence and field independence. New York: International Universities Press.

Wolf, S., & Ripley, H. C. (1947). Reactions among Allied prisoners of war subjected to three years of imprisonment by the Japanese. American Journal of Psychiatry, 104, 180–193.

Submitted: June 9, 1989 Revised: November 15, 1989 Accepted: November 15, 1989