Disorders of Extreme Stress Following Warzone Military Trauma: Associated Features of Post-traumatic Stress Disorder (PTSD) or Comorbid but Distinct Syndromes?

Julian D. Ford

As published in the February 1999 issue of Journal of Consulting and Clinical Psychology, 67(1), 3-12. Note that this online version may have minor differences from the published version.

Department of Veterans Affairs National Center for PTSD, Executive Division
Department of Psychiatry, Dartmouth Medical School
Please address correspondence to:
Julian D. Ford, Ph.D.
Director of Behavioral Healthcare Outcomes Research
Director of Outpatient Services
Department of Psychiatry 6410
University of Connecticut School of Medicine
University of Connecticut Health Center
10 Talcott Notch Rd.
Farmington, CT 06032
e-mail: Ford@Psychiatry.uchc.edu
Telephone: 860-679-6709/6732
FAX: FAX 860-679-6736

Acknowledgments: I appreciate the thoughtful comments of Drs. Paula Schnurr, Matthew Friedman, and John Nemiah, as well as the associate editor and anonymous reviewers, on earlier drafts of this manuscript. My thanks to Kathryn Thomas, M.D. and Patricia Cason, M.D. for providing independent diagnoses, to Patricia Chandler, Sandra Hardin, Michael Stevens, Barbara Thacker, and Jamie Zabukovec for clinician ratings, Phyllis Kidd for providing independent DESNOS ratings and collecting childhood trauma data, to Philip Fisher and Lynn Larson for helping to coordinate the overall project, and to Bud Halthen, Curtis Betters, Devon Wardwell, Mary Camacho, Kell Rudestam, Bill Folks, Earl Givens, and Dan Delapine for ensuring the collection and recording of the data.

Abstract

Disorders of Extreme Stress Not Otherwise Specified (DESNOS) and Post-traumatic stress disorder (PTSD) were found to be comorbid but in several fundamental ways distinct in a sample of military veterans seeking inpatient PTSD treatment: 31% qualified for both conditions, 29% were diagnosed only PTSD, 27% were classified only DESNOS, and 13% met criteria for neither. PTSD diagnosis was associated with elevated levels of warzone trauma exposure and witnessing atrocities, and with post-traumatic impairment on the Mississippi and Penn PTSD measures. DESNOS classification (but not PTSD) was associated with: (a) early childhood trauma and participation in warzone atrocities, (b) extreme levels of intrusive trauma reexperiencing symptoms, (c) impaired characterological functioning (object relations), and (d) elevated risk of utilization of intensive psychiatric services. The findings suggest that PTSD and DESNOS are comorbid but distinct post-traumatic syndromes. Implications for assessment, classification, and treatment of chronically impaired trauma survivors, and for research extending or revising the findings with different subpopulations of trauma survivors, are discussed.

Disorders of Extreme Stress Following Warzone Military Trauma:
Associated Features of Post-traumatic Stress Disorder (PTSD)
or Comorbid but Distinct Syndromes?

The longitudinal course following traumatization ranges from transient stress reactivity and acute stress disorders to episodic or enduring post-traumatic stress disorder (PTSD) and a range of associated psychosocial impairments (Yehuda & McFarlane, 1995; Friedman & Rosenheck, 1966). Complex PTSD or "Disorders of Extreme Stress Not Otherwise Specified" (DESNOS; Herman, 1992) was proposed as an alternative to Axis II personality disorder diagnosis when extreme trauma compromises the fundamental sense of self and relational trust at critical developmental periods (e.g., childhood sexual abuse; Roth, Newman, Pelcovitz, van der Kolk, & Mandel, 1997). Although not codified as a formal diagnosis, the DESNOS syndrome offers a conceptual framework for understanding and a clinical framework for assessing several sequelae often experienced by survivors of extreme trauma: (a) extreme affect and impulse dysregulation (e.g., rage, suicidality, self-destructiveness, unmodulated sexual activity), (b) pathological dissociation, (c) somatization (including alexithymia), and (d) fundamentally altered beliefs concerning self and relationships.

The National Institute of Mental Health DSM-IV PTSD Field Trial Study developed a structured interview to assess clinical and community respondents for DESNOS (Pelcovitz, van der Kolk, Roth, Mandel, Kaplan, & Resick, 1997). An early report found substantial intercorrelation among structured clinical interview measures of PTSD and three components of DESNOS (i.e., dissociation, somatization, and affect dysregulation) in adult survivors of childhood abuse, violence or abuse in adulthood, or disaster (van der Kolk, Pelcovitz, Roth, Mandel, McFarlane, & Herman, 1996). Most (92%) participants classified as DESNOS also met criteria for PTSD diagnosis, and DESNOS was associated with a history of childhood traumatic abuse (versus disaster; van der Kolk et al., 1996). Dissociation, affect dysregulation, and somatization each was problematic for 70-90% of community or clinical participants diagnosed with lifetime PTSD. Van der Kolk et al. (1996, pp. 89-90) concluded that DESNOS were "associated features of PTSD [that are] not likely to constitute separate 'double diagnoses' but represent the complex somatic, cognitive, affective, and behavioral effects of psychological trauma."

However, Field Trial results did not show dissociation, somatization, and affect dysregulation to be isomorphic with PTSD, but that each shared between 26-28% common variance PTSD (van der Kolk et al., 1996, Table 1). Despite their strong correlation with PTSD, DESNOS features thus may occur independently of PTSD. The large degree of nonoverlap between PTSD diagnosis and the features of DESNOS appears largely due to the fact that a substantial proportion of trauma survivors who did not meet criteria for lifetime PTSD were classified as exhibiting pathological dissociation (61%), somatization (47%), or affect dysregulation (34-37%). Thus, trauma survivors not meeting PTSD diagnostic often display substantial DESNOS symptoms.

Descriptions of sequelae of extreme traumatization often involve symptoms and impairment better characterized by DESNOS-like alterations of consciousness, affect and impulse regulation, and self-organization than by the fear, avoidance, numbing, hyperarousal, and hypervigilance cardinal to PTSD. Battered children and spouses, incest survivors, POWs, survivors of the holocaust or other incidents of genocide, survivors of kidnapping, torture, or terrorism, and military veterans exposed to war atrocities have been characterized as suffering from post-traumatic self disorder (Parson, 1988), serial selves (Laufer, 1988), or a death imprint (Lifton, 1979). Despite their substantial substantial comorbidity (van der Kolk et al., 1996; Roth et al., 1997), it is not clear that PTSD precedes nor even necessarily accompanies the cardinal DESNOS symptoms. PTSD and DESNOS involve several comparable symptom features (e.g., PTSDis emotional numbing and anger may overlap with DESNOSis affect dysregulation; PTSDis flashbacks or psychogenic amnesia may involve DESNOS-like dissociation; DESNOSis alterations in beliefs and relationships may overlap with PTSDis sense of foreshortened future and social detachment). However, PTSD does not specify self-destructive, impulsive, risk taking, or suicidal behavior (DESNOS affect dysregulation), pathological dissociation, or somatization. PTSDis diagnostic criteria also do not include a fundamentally impaired view of the self (e.g., guilt, shame, ineffectiveness, damage, hopelessness; DESNOS criteria III, VII). Further, trauma survivors not diagnosed with PTSDoalthough possibly suffering from undetected PTSDooften are best characterized by a range of Axis I and II disorders involving DESNOS-like impairments in affect, self-regulation, and consciousness (Zlotnick, Zakriski, Shea, Costello, Begin, Perlstein, & Simpson, 1996). Thus, DESNOS may be associated with traumatization yet distinct from PTSD.

One preliminary study of PTSD and DESNOS reported high rates (75+%) of DESNOS symptoms in veterans seeking treatment for chronic warzone-related PTSD, including problems with modulation of affect and anger, dissociative "spacing out," amnesia, inability to trust, despair, hopelessness, feeling permanently damaged misunderstood, and a loss of sustaining beliefs (Newman, Orsillo, Herman, Niles, & Litz, 1995). Other DESNOS symptoms reported by at least half the sample included dissociative derealization, shame, guilt, or fear of loss of control due to anger or violent impulses. This pilot study was not able to distinguish between DESNOS and PTSD due to the small sample size (N=10) and the absence of a control group of treatment-seeking veterans not diagnosed with PTSD. Morever, the study could not disentangle the roles of warzone and childhood traumatization because almost all participants (80%) had experienced childhood abuse. Although high rates of childhood trauma are the norm in samples of treatment-seeking veterans with PTSD (Bremner, Southwick, Johnson, Yehuda, & Charney, 1993), many report no history of childhood traumatization and could therefore comprise a comparison sample.

In the present study, PTSD, DESNOS, and history of both early childhood and warzone trauma were assessed in a sample of chronically impaired military veterans admitted for inpatient care for military-related PTSD (Ford, Fisher, & Larson, 1997). Participants were characterized by severe psychosocial impairment and psychiatric symptomatology, and concomitantly high levels of utilization of intensive psychiatric care, as well as documented warzone military trauma exposure. The primary issue this study addressed is whether DESNOS is best conceptualized as associated features of PTSD, as opposed to as a distinct but comorbid post-traumatic syndrome. Independent structured assessments of PTSD and DESNOS were obtained in an attempt to determine if the comorbidity pattern found in the DSM-IV Field Trial study (van der Kolk et al., 1996) could be replicated. Despite receiving treatment for PTSD, not all participants met research diagnostic criteria for PTSD, and thus it was possible to identify individuals with or without both PTSD and DESNOS. If DESNOS is found primarily only when PTSD also is diagnosed, an associated feature model is supported. If DESNOS occurs often without PTSD, then DESNOS may be a distinct syndrome even if there is evidence of some DESNOS-PTSD comorbidity.

Simply examining the rates of separate occurrence and comorbidity of PTSD and DESNOS is only a preliminary step, because any apparent independence or overlap may be an artifact of the specific diagnostic cutoffs used for each diagnosis or of the comorbidity percentage considered sufficient to justify defining the syndromes as distinct or associated. Several additional hypotheses offer a basis for contrasting the views of DESNOS as an associated feature or comorbid syndrome. Given the DSM-IV Field Trialis findings (Roth et al., 1997; van der Kolk et al., 1996), DESNOS was presumed to be an associated feature of PTSD, and the following hypotheses were tested:

1. DESNOS and PTSD should differ etiologically only in DESNOSis association of severe early childhood trauma (Roth et al., 1997; Zlotnick et al., 1998) and PTSDis association with severe combat trauma (Kulka et al., 1990)othe latter because PTSD was assessed with a focus on military trauma. To test the associated feature versus comorbid syndrome models of DESNOS, we also assessed exposure to war atrocities, because this is an extreme form of military-related trauma with profound potential impact on late adolescent/early adult development (Laufer, 1988). If DESNOS is an associated feature of PTSD, atrocity exposure should be associated with both syndromesoalthough possibly more strongly with DESNOS because of the link observed between war atrocity exposure and increased symptomatic severity (Yehuda, Southwick, & Giller, 1992). However, if atrocity exposure is associated differentially with DESNOS and PTSD, this would suggest that the syndromes may have distinct post-childhood etiologies or courses.

2. DESNOS and PTSD should differ quantitatively but not qualitatively in their association with symptomatic or impairment severity (Newman, Riggs, & Roth, 1997). DESNOS is defined by symptoms (e.g., somatization, dissociation) and psychosocial impairments (e.g., rage, guilt and shame, impulsiveness) that differ from the cardinal intrusive reexperiencing, avoidance, emotional numbing, and hyperarousal symptoms of PTSD. However, each of the DESNOS features has been empirically found to be correlated with PTSD: e.g., dissociation (Bremner et al., 1992); somatic dysregulation (Boscarino, 1997); rage (Chemtob et al., 1996); guilt and shame (Henning & Frueh, 1997); and, impulsivity (Joseph, Dagliesh, & Yule, 1997). Thus DESNOS and PTSD should either exert separate additive main effects on the reported levels of the same post-traumatic, psychiatric, and psychosocial problems, or there should be a DESNOS x PTSD interaction effect such that that symptom or impairment is most severe when DESNOS accompanies PTSD. If DESNOS is a separate comorbid syndrome, its effect on symptoms and impairment should not be in interaction with PTSD and its main effects should be for different aspects of symptomatic distress or functional impairment than those particularly associated with PTSD.

In addition to self-report measures, a clinician rating method for assessing object relations was utilized to contrast PTSD and DESNOS. In a prior study, PTSD was positively correlated with clinician object relations, suggesting that veterans seeking treatment for PTSD who do not fully manifest the PTSD syndrome were characterized by poor core psychosocial functioning (i.e., low object relation levels) while those who met criteria for PTSD had higher (moderate) object relations levels (Ford et al., 1997). This finding, along with the conceptual proposition advanced by Herman (1992) that DESNOS reflects trauma-specific developmental psychopathologyowhich should involve impaired object relations (Cicchetti & Toth, 1995)oraises the possibility that post-traumatically impaired individuals who do not clearly manifest PTSD may instead suffer from DESNOS. If DESNOS is an associated feature of PTSD then DESNOS and PTSD should interact, such that PTSDis positive effect should be present only when DESNOS is absent. If DESNOS and PTSD have separate opposite main effects a comorbid model is supported.

3. Both DESNOS (Herman, 1992) and PTSD (Friedman & Rosenheck, 1996) should be associated with an elevated likelihood of utilization of the most costly and intensive psychiatric treatment services, compared to similar PTSD-treatment-seeking individuals who do not meet criteria for DESNOS or PTSD. The increased severity in symptomatic distress and functional impairment attributed to DESNOS (Roth et al., 1997) may lead to a particularly high risk of costly intensive treatment for these individuals. However, if only DESNOS or only PTSD is associated with risk of intensive psychiatric treatment, this would suggest that the syndromes have differential rather than parallel or cumulative effects, and would support the comorbid model.

To provide the clearest possible test of these hypotheses with the present data set, two variables potentially associated with DESNOS and chronic PTSD will be controlled in case either should be a "third variable" accounting for what otherwise could be mistakenly interpreted as an effect of DESNOS and/or PTSD. Chronic PTSD often is comorbid with major depression (Kulka et al., 1990) and personality disorders (Southwick, Yehuda, & Giller, 1993). DESNOS has many features in common with both major depression and personality disorder (Herman, 1992). Therefore, major depression and personality disorder were used as covariates in multivariate and univariate analyses.

Method

Participants

Eighty-four of 85 consecutive admissions to a Department of Veterans Affairs specialized inpatient PTSD Residential Rehabilitation Program (PRRP) consented to participate in the study. All participants were male, ranging in age from 28 to 67 (M = 46, SD = 5.7) and education from 10th grade to Masters level (M = 12.1 years, SD = 1.6). Most (86%) were Caucasian, with a subgroup of Native American (12%) and Latino (2%) veterans. Most (94%) served in Vietnam. Seven (8%) served in the Korean war or World War II and one in the UN Somalia Peacekeeping Operation. Thirty-three per cent were currently married, 36% never married, and 30% divorced or widowed. Most of those currently married (i.e., 20 of 28) were remarried following one to four divorces. Most were currently unemployed (69%) or retired (4%). Few were employed full-time (7%) or part-time (20%). Almost all had extensive problems with employment, either having held at least 10 different jobs since military service (80%) or having served extensive (M = 12.5 years) prison terms (6%). Residential instability and homelessness were typical in the sample: 60% were homeless or had severe residential instability during the year prior to PRRP admission. Participants met the following inclusion criteria: (a) no severe organic or psychotic impairment or imminent suicidality, (b) no criminal charges pending or in adjudication, (c) no alcohol or drug use or prescribed opiate or benzodiazepine medications for at least one month, (d) in ongoing outpatient psychiatric or psychological treatment, and (e) documented exposure to warzone military trauma.

Extensive past histories of inpatient psychiatric treatment were described by patient self-report, with corroboration from medical records and referring clinicians. The mean number of prior psychiatric and substance abuse inpatient admissions were 1.68 and 1.01, with a range of 0-12 and 0-7, respectively. Most (77%) had at least one prior psychiatric or substance abuse inpatient admission and, of the 18 who had not, 8 (44%) had received residential care in a VA Homeless Domiciliary. Only 10 (12%) had never received any inpatient or care. All reported serious suicidal ideation at some time since military service, more than half currently.

Diagnostic and DESNOS Classification

The Structured Clinical Interview for DSM-III-R Axis I, Patient Version (SCID-P; Spitzer, Gibbon, & First, 1990b) and the Structured Interview for Disorders of Extreme Stress (SIDES; Pelcovitz et al., 1997) were administered by the author. The order of PTSD and DESNOS interview administration was randomly varied for each participant to control for possible order effects. Independent ratings were conducted in a conjoint interview by one of two psychiatrists for 20 randomly selected participants. Interrater agreement was adequate for a current PTSD diagnosis (60% prevalence, Kappa =0.91, p < .001). Consistent with Pelcovitz et al.is (1997) psychometric findings, scoring the SIDES for symptoms present/absent in the past six months yielded acceptable subscale internal consistency levels (Cronbachis alphas: Scale I, .93; Scale II, .80; Scale III, .74; Scale V, .88; Scale VII, .86. Interrater reliability was established by independent SIDES ratings by a post-doctoral fellow for 15 randomly selected participants yielded 13 agreements (Kappa = 0.79, p < .001), with the two disagreements classified by the primary assessoris data. Consistent with Pelcovitz et al. (1997, Table 1), subscale VI (altered perceptions of perpetrator) was excluded because more than half of the participants had no trauma for which there was a specific perpetrator. Subscale IV, somatization, was assessed but not counted toward DESNOS because complex medical histories and diagnoses in this sample resulted in almost universally-reported high levels of somatic symptoms which for which organic causation could not be ruled out. Participants meeting all other DESNOS criteria also reported the requisite two (of five) somatization symptom clusters, and no participant was not classified as DESNOS due to failing to meet somatization criteria.

SCID-P (Spitzer et al., 1990b) and SCID-II (Spitzer et al., 1990a) modules were coded by the author, with independent corroboration by one of two psychiatrist raters for a randomly selected sample of twenty cases for current major depression (39% prevalence, Kappa = 0.81, p < .001) and presence of any personality disorder (36% prevalence; Kappa = .67; p < .001).

Measures of Trauma History

All participants reported exposure to warzone trauma sufficient to qualify for DSM-III-R Criterion A. Warzone trauma varied from single or infrequent instances of rocket, mortar or "sapper" attacks, to many consecutive months of being under threat of intense fire from the enemy, to observing or participating in grotesque or abusive violence. War trauma was corroborated in every case by military records (e.g., DD214 discharge form) documenting duty sites and military operational specialties associated with combat or combat support. A dichotomously scored version of the Keane, Fairbank, Caddell, Zimering, Taylor and Mora (1989) Combat Exposure Scale was utilized to quantify severity of warzone trauma exposure based upon: hazardous duty, subject to enemy fire, being surrounded by the enemy, more than 25% of the soldiers in one's unit KIA/WIA/MIA, seeing others hit by incoming rounds, and being in danger of being severely injured or killed. The 0-7 point composite scale had good internal consistency (Alpha = .88; M item-total score r = .68) and retest stability over a 1-3 month period prior to PRRP admission (N = 14, r = .95, p < .001). War trauma scores ranged from 1 to 7 (M = 5.5, SD = 2.1). Independent clinician ratings of war trauma in detailed military history interviews with 24 participants showed good correspondence with war trauma self-ratings (r = .82, p < .001). Two one-item indices also assessed participants' (a) witnessing, or (b) participating in, abusive warzone violence ("atrocities").

In a detailed developmental history interview, PRRP clinicians inquired about the presence or absence between the ages of 0-6 of each of 10 discrete types of trauma exposure derived from lifetime trauma assessments developed by Resnick, Kilpatrick, Dansky, Saunders and Best (1993) and Weaver and Clum (1993): sexual abuse; physical abuse; witnessing or participating in intentional violence/killing; receiving a violent life-threat; directly experiencing a life-threatening natural or humanmade disaster or accident; witnessing other severe or violent injury or death; experiencing the murder or DUI death of a primary caregiver. Each trauma event type was defined in behavioral terms (Resnick et al., 1993) and specifically inquired about using a structured interview format similar to that validated by Bremner et al. (1993). If any traumatic event had occurred, the participant was classified as positive for early childhood trauma. Two clinicians made independent ratings of 32 randomly selected participants, agreeing on 29 cases (Kappa = .85). Disagreements were coded positive based upon definite information of early childhood trauma recorded by one of the two independent raters. Potential trauma events from other developmental periods in later childhood, adolescence, and adulthood were assessed but are not reported here.

Early childhood trauma primarily involved three reliably rated types of events: severe physical abuse (N = 32; kappa = .89), sexual abuse (N = 4; kappa = 1.00), or witnessing family violence or deaths (N = 14; kappa = .77). Family violence usually (i.e., 8 of 10 cases) co-occurred with physical abuse. Sexual abuse co-occurred with physical abuse and family violence in two of four cases. Caregiver death almost always occurred separately from physical abuse, family violence, or sexual abuse). Caregiver death was included only after it was determined that excluding the five cases in which only caregiver death occurred led to no change in the pattern or level of results.

Self-Report Measures of Symptomatic Severity and Impairment

Widely used and psychometrically sound questionnaires assessed PTSD symptomatic and impairment severity: the Penn PTSD Scale (Hammarberg, 1992), the Mississippi Scale for Combat-Related PTSD (Keane, Caddell, & Taylor, 1988), and the Impact of Event Scale (Weiss & Marmar, 1997) subscales (Intrusive reexperiencing, IES-I; Avoidance, IES-A). Widely used self-report measures with documented psychometrics assessed psychiatric symptom severity: the Global Severity Index (GSI) of the Hopkins Symptom Checklist-90-Revised (Derogatis, 1992); the trait subscale of the State-Trait Anxiety Inventory (STAI; Spielberger, Gorsuch, & Lushene, 1970); the Beck Depression Inventory (BDI; Beck, Steer, & Garbin, 1988); and the Dissociative Experiences Scale (Bernstein & Putnam, 1986).

Object Relations Clinician Rating (OR-C)a

Independent ratings of Westenis Social Cognition Object Relations Scale (Barends, Westen, Leigh, Silbert, & Byers, 1990) were done for each participant by trios of trained clinicians using the SCID-P Overview and Life Chart format to review the participant's major relationships and life experiences from birth to present (Ford et al., 1997). Unlike other study measures, object relations was scored in a positive direction, with higher scores representing lesser impairment and greater functionality. OR-C involves four 5-point scale ratings of the individualis general ability to (a) understand each person as having a coherent unique personality; (b) experience emotions as sources of helpful information, rather than as malevolent threats; (c) invest emotionally in relationships and moral standards or values; and (d) understand that psychological agency has a positive potential in determining the course of social events (rather than feeling helpless, dependent, or overcontrolling).

One rating was obtained for each participant on each of the four OR-C categories by averaging the 3 raters' ratings. Interrater reliability for each OR-C category was calculated by the Spearman-Brown version of the intraclass correlation coefficient, yielding the following estimates of the dependability of the average scores for the four OR-C categories: .75, .63, .69, .72 (all p < .01). Ratings of the four object relations categories were highly intercorrelated (rs = .78-.91, p < .001), so we summed them and divided by four to produce a single 5-point object relations composite score (OR-C) for each participant. The OR-C score was internally consistent (alpha = .95; mean item-total score r = .89). Retest reliability for a sample of 20 participants rated at a 15-30 day interval was r = .88. Evidence of convergent validity also was obtained by correlating OR-C ratings with independent object relations ratings from a Thematic Apperception Test (r = .84, p < .001).

Measures of Psychiatric Services Utilization

Medical Center database records were accessed to document utilization of inpatient psychiatric hospitalization in VA Medical Centers within the year prior to PRRP admission and lifetime. Pre-admission application records completed by each veteran and by a referring clinician, and an intake psychosocial history provided information confirming VA inpatient psychiatric hospitalizations, and also confirmed that no participant utilized inpatient psychiatric care outside VA that year.

Results

Participants diagnosed with PTSD did not differ from those not meeting criteria for PTSD on demographic factors (Table 1) including age and education level (t[82] = 0.76, 0.39, p=.39, .69, respectively), and current marital status and ethnic background c2 (1, N = 85) = 0.19, 0.07, p =.66, .77, respectively). Participants meeting criteria for DESNOS did not differ from others in age or education level (t(82) = 0.22, 1.71, p =.88, .09, respectively), or current marital status c2 (1, N = 85) = 0.29, p =.58). More non-caucasian than caucasian participants were classified with DESNOS (7% vs, 26%, c2 (1, N = 85) = 4.55, Phi = 0.24, p =.03). The four subgroups defined by crossing PTSD with DESNOS did not differ significantly on any demographic indicator.

TABLE 1. Demographic, Trauma History, and Inpatient Treatment Utilization Characteristics of Participants Classified by PTSD Diagnosis and DESNOS

PTSD Diagnosis / DESNOS Classification
Measure PTSD +
DESNOS +
PTSD +
DESNOS -
PTSD -
DESNOS +
PTSD -
DESNOS -
 
  N = 26 (31%) N = 25 (29%) N = 23 (27%) N = 11 (13%) c 2
(3,N=85)
Age 47 [5] 49 [6] 49 [6] 47 [6] NS
Education 12 [1] 13 [1] 12 [1] 13 [2] NS
% Married 35 40 30 36 NS
% Ethnic Minority 22 5 20 9 NS
% Early Child Trauma 71 36 71 27 13.8 **
% Severe Combat 84 84 41 36 17.3 ***
% Witness Atrocities 75 52 40 30 9.4 **
% Participate Atrocities 46 5 40 0 18.2 **
Past Year Inpatient Psych 67 35 62 46 6.0
Lifetime Inpatient Psych 82 39 72 30 13.8 **

Note: Data reported as Mean [Standard Deviation] or percentages.
PTSD + = Diagnosed PTSD; PTSD - = Not Diagnosed PTSD;
DESNOS + = Classified DESNOS; DESNOS - = Not Classified DESNOS;
% Married = Percent Currently Married; NS = not statistically sigificant.
* p < .05 ** p < .01 *** p < .001

Classification and Comorbidity of PTSD and DESNOS

All participants met PTSD Criterion A, having experienced warzone events in which they witnessed or were directly at risk for death or severe injury, and which were beyond the range of ordinary human experience. All participants met Criterion C, experiencing three or more of the seven symptoms of avoidance, emotional numbing, and social detachment in the past month. More than half of the sample (60%) met criteria for current military-related PTSD, but 34 participants (40%) did not, because: (a) no distinct military-related intrusive reexperiencing (N = 20), (b) fewer than two hyperarousal and hypervigilance symptoms (N = 7), or (c) both (N = 7). The No PTSD participants reported a morbid ruminative preoccupation with war in general, with violence, with death, with injustice, or with feeling psychically and physically "damaged."

Forty-nine participants (58%) met criteria for DESNOS. Prevalences of the five applicable DESNOS features were: altered regulation of affect and impulses (94%), altered consciousness or attention (46%), altered self-perceptions (e.g., alienation, guilt, shame; 79%), altered relationships (e.g., profound distrust, revictimization, victmizing; 60%), and altered systems of meaning (e.g., despair, hopelessness; 66%). The most prevalent and severe DESNOS symptoms were problems with rage, overwhelming distress, amnesia, a sense of being damaged, guilt and shame, the belief that no one understands, distrust, relational conflict or avoidance, hopelessness, and loss of meaning in life. Other DESNOS symptoms rated as at least moderately problematic for at least 67% of the sample included: extreme risk taking, constricted affect due to anger, carelessness about safety, felt personal ineffectiveness, altered ethics or religious beliefs, and suicidality.

PTSD and DESNOS diagnoses were highly comorbid but often distinct (Table 1). PTSD was accompanied by comorbid DESNOS in 51% of cases. DESNOS was accompanied by PTSD in 54% of cases. One participant in three (31%) qualified for both PTSD and DESNOS; 13% for neither. Almost half the DESNOS participants (23 of 49) were not diagnosed with PTSD. Thus, DESNOS occurred almost as frequently separate from PTSD as in conjunction with PTSD.

Major depression often was comorbid with DESNOS or PTSD. DESNOS was associated with an increased likelihood of major depression (52% vs. 21%), c2(1, N = 84) = 7.85, p = .005, but PTSD was not (44% vs. 30%), c2(1, N = 84) = 2.03, p = .15. DESNOS also was associated with the presence of an Axis II personality disorder (58% vs. 20%), c2(1, N = 84) = 12.18, p = .001, while PTSD diagnosis was not (34% vs. 53%), c2(1, N = 84) = 2.99, p = .083). All PTSD diagnosed participants who had a comorbid personality disorder also had DESNOS, suggesting a linkage between DESNOS, but not chronic PTSD, with Axis II personality disorders.

PTSD, DESNOS, and Trauma History

The four PTSD x DESNOS groups differed in likelihood having a history of trauma for each of the four trauma variables (Table 1). As predicted, DESNOS participants had particularly high likelihoods of having a history of early trauma regardless of PTSD diagnosis. PTSD participants had especially high likelihoods of a history of severe combat trauma regardless of DESNOS.

Comparing just participants with PTSD-only versus those with DESNOS-only, DESNOS-only participants were more likely to report early trauma than PTSD-only participants, c2(1, N = 47) = 6.3, p =.01, while PTSD-only participants were more likely to report severe combat trauma than DESNOS-only participants, c2(1, N = 47) = 9.4, p =.002. The PTSD-only and DESNOS-only groups were equally likely to report witnessing war atrocities, but the DESNOS-only group was more likely to have participated in atrocities, c2(1, N = 47) = 8.2, p < .01. DESNOS-only participants also were more likely than the no-PTSD/no-DESNOS group to have participated, c2(1, N = 33) = 3.7, p =.05 (but not to have witnessed), war atrocities. PTSD-only participants did not differ from no-PTSD/no-DESNOS participants on atrocity exposure. The DESNOS-only group was as likely as the the comorbid PTSD+DESNOS group to have participated in atrocities, but was less likely to have witnessed atrocities, c2(1, N = 48) = 6.0, p = .01. The PTSD-only group was as likely as the comorbid PTSD+DESNOS group to have witnessed atrocities, but less likely to have participated in war atrocities, c2(1, N = 51) = 11.2, p = .001.

Multivariate logistic regression analyses were conducted to examine the contribution of the four trauma variables, controlling for major depression or personality disorder, to the likelihood of (a) PTSD diagnosis or (b) DESNOS. The only predictor of PTSD diagnosis was severe combat, Odds Ratio (OR) = 7.3, 95% Confidence Interval (CI) = 2.4-19.9, p = .001. Significant predictors of DESNOS were early childhood trauma, OR = 4.5, 95% CI = 1.9-8.9, p < .01, participation in atrocities, OR = 13.5, 95% CI = 1.5-29.3, p < .05, personality disorder, OR = 4.4, 95% CI = 1.3-13.8, p < .05, and major depression, OR = 3.9, 95% CI = 1.2-7.9, p < .05. Thus, accounting for the effect of major depression and personality disorder diagnosis, traditional combat trauma was the trauma risk factor for PTSD, and childhood trauma and participation in atrocities for DESNOS.

PTSD, DESNOS, and Severity of Symptoms and Impairment

A two-way multivariate analysis of covariance (MANCOVA) with PTSD diagnosis and DESNOS as independent variables and two covariates (major depression diagnosis, any personality disorder diagnosis) was conducted to examine main and interaction effects on 9 dependent variables measuring the severity of PTSD symptoms and impairment (Mississippi, Penn, IES-I, IES-A), GSI, dissociation, anxiety, depression, and object relations. Consistent with a view of DESNOS as a severity marker for post-traumatic impairment, DESNOS was significantly associated with more distress on a multivariate basis, Hotellingis F (9, 64) = 3.1, p < .01, with univariate main effects for (increased) intrusive reexperiencing (IES-I) and (decreased) object relations, Fs(1, 72) = 12.5, 4.5, p < .001 and .05, respectively. PTSD also was significantly associated with increased distress on a multivariate basis, Hotellingis F (9, 64) = 7.0, p < .001, with univariate main effects for (increased) Mississippi and Penn, and (increased) object relations levels, Fs(1, 72) = 17.4, 7.2, and 46.6, respectively, p < .001. The PTSD x DESNOS interaction was nonsignificant, Hotellingis F (9, 64) = 0.8, p > .50. Thus, PTSD and DESNOS had distinctly different patterns of univariate effects, as displayed graphically for the Mississippi, Penn, IES-I, and OR-C scales in Figure 1.

Figure 1. Mississippi PTSD Scale, Penn PTSD Inventory, IES Intrusion subscale, and OR-C levels associated with groups classified by PTSD diagnosis (+ = present; - = absent) and DESNOS (+ = present; - = absent).

[Figure 1]

Next, participants having a PTSD diagnosis but not DESNOS (PTSD-only) were compared to those with DESNOS but not PTSD (DESNOS-only), via a oneway MANCOVA with the same two covariates and dependent variables. The multivariate effect for group was significant, Hotellings F(9, 28) = 6.7, p < .001. The only univariate differences were that DESNOS-only participants had higher IES-I intrusive reexperiencing scores and lower object relations scores than PTSD-only participants, Fs(1, 36 = 5.2, 40.2, p < .05 and .001, respectively).

Treatment Utilization Associated with PTSD and DESNOS

Dichotomous scores for admission for inpatient psychiatric hospitalization (a) in the past year and (b) ever as an adult, served as dependent variables in multivariate logistic regression analyses, with major depression, any personality disorder, PTSD diagnosis, and DESNOS as predictors. DESNOS was the sole risk factor for past year psychiatric hospital admission, OR = 2.72, 95% CI = 1.00-5.81, p = .05. DESNOS and major depression were independent risk factors for lifetime inpatient psychiatric admission, ORs = 3.67, 7.21, 95% CIs = 1.85 - 8.66, 2.93 - 12.11, p < .05 and .01, respectively. On an unadjusted basis, the 4 groups defined by DESNOS and PTSD diagnosis differed significantly on likelihood of lifetime inpatient admission, and showed a similar pattern (p = .10) for past yearis inpatient admission. Regardless of PTSD diagnosis, participants with DESNOS had higher likelihoods of past year or lifetime inpatient psychiatric hospitalization than those without DESNOS, c2(1, N = 84) = 5.46 and 10.53, p < .05 and .01, respectively (Table 1).

Discussion

The evidence from this study seems less consistent with the DSM-IV Field Trial view of DESNOS as an associated feature of PTSD than with a formulation of DESNOS and PTSD as separate comorbid post-traumatic syndromes. Concerning hypothesis 1, results for trauma history indicate that DESNOS involves a course or etiology that is more separate from than intertwined with PTSD. Findings regarding hypothesis 2 failed to show a DESNOS x PTSD interaction on indices of post-traumatic and psychiatric impairment, but instead that DESNOS and PTSD had separate main effects for different measures. Hypothesis 3 also was not supported, because only DESNOS was an independent risk factor for utilization of inpatient psychiatric treatment. These hypotheses and findings are limited to one sample of chronically impaired male military veterans receiving specialized PTSD treatment and to a limited set of measures assessed cross-sectionally, and thus cannot definitively rule out an associated features model or confirm a comorbid syndromes formulation of DESNOS and PTSD. However, the evidence indicates that DESNOS can be distinguished from PTSD (as well as from major depression and personality disorder) by: (a) clinical differential diagnosis, (b) a history of traumatization in early childhood and war atrocity participation in young adulthood, (c) extreme trauma-related intrusive reexperiencing symptoms and impaired object relations, and (e) high risk for inpatient psychiatric treatment.

Clinical Assessment of DESNOS and PTSD

PTSD diagnosis and DESNOS were highly comorbid but neither isomorphic nor inseparable. PTSD and DESNOS each occurred often (aproximately 60% prevalence), and were comorbid in almost one in three participants. However, aproximately half of those with DESNOS were not diagnosed with PTSD, suggesting substantial nonoverlap of the two syndromes. DESNOS also occurred as frequently separate from PTSD as in conjunction with PTSDoquite different from the DSM-IV PTSD Field Trial finding that only 8% of respondents classified as were not diagnosed with PTSD (van der Kolk et al., 1996; Roth et al., 1997). Although both major depression and personality disorder often were comorbid with both DESNOS and PTSD, only DESNOS had a more than chance association with major depression or the presence of a personality disorder.

Although requiring replication with populations other than chronically impaired male military veterans in inpatient PTSD care, these results suggest that DESNOS often occurs separately from PTSD. Differential assessment for both PTSD and DESNOS (as well as for other frequently comorbid Axis I or II disorders) appears advisable with individuals seeking treatment for chronic PTSD, in order to not prematurely attribute DESNOS-like symptoms to PTSD, depression, or a personality disorder. As discerned by the originators of the DESNOS construct (Herman, 1992; van der Kolk et al., 1996), some patients with extensive trauma histories may be better understood and treated with a focus on DESNOS as well as (or instead of) PTSD, personality disorder, or major depression. This is not an either-or issue, because DESNOS often co-occurred with not only PTSD but also with major depression and personality disorderoand may also be comorbid with other disorders not assessed in this study. Sensitive and accurate approaches to screening, assessing, and treating for trauma and its character-altering effects (e.g., traumatic grief, Prigerson et al., 1997) may make therapeutic gains possible with what otherwise may be considered either relatively intractable affective and/or characterological impairment. DESNOS does not replace any Axis I or II diagnosis but provides a complementary perspective for diagnosis and treatment planning.

The finding of greater independence of PTSD and DESNOS in this study than in the DSM-IV Field Trial warrants further investigation. The present sample included only male treatment-seeking war veterans. Gender thus may play a role in the studiesi discrepant findings: women were overrepresented in the Field Trial, especially in the group of violence victims (82%) that was at highest risk for DESNOSocompared to their more gender-balanced (51% female) group of disaster victims. DESNOS and PTSD also may be more difficult to distinguish in non-treatment-seeking or less chronically impaired samples due to their low prevalences of PTSD and DESNOS (van der Kolk et al., 1996). The chronicity of post-traumatic impairment in the present sample also may have made PTSD and DESNOS more discriminable by increasing both the frequency and severity of risk taking, violence, injury, self-destructive actions, and chronic illness (Friedman & Rosenheck, 1996). DESNOSis effects in this study appear robust because they were found despite a possible ceiling effect due to frequent endorsement of DESNOS symptoms by all groups.

The finding that non-caucasian participants more often than their caucasian counterparts were classified as DESNOS suggests a need for replication with other non-caucasian ethnicities beyond the primarily Native American subsample in this study, as well as for evaluation of the DESNOS criteria for both cultural sensitivity and bias. Native American Vietnam in-country veterans have very high prevalences of PTSD and severe psychosocial impairment as well as high levels of both combat and early childhood trauma (Friedman et al., 1998), which may also lead to DESNOS.

PTSD, DESNOS, and Trauma History

The trauma history data showed thatoconsistent with prior research on military-related PTSD (Bremner, Southwick, & Charney, 1995) and on DESNOS (Roth et al., 1997)otraditional combat trauma was the key trauma risk factor for PTSD and early childhood trauma was an independent risk factor for DESNOS, regardless of whether the syndromes occurred separately or comorbidly. However, the war atrocity data suggest still greater complexity in the etiology of DESNOS and PTSD. When the witnessing of atrocities was distinguished from participation in atrocities (Unger et al., 1998), DESNOS and PTSD were differentially related to atrocity exposure. Atrocity participation was a risk factor for DESNOS but not PTSD. Witnessing atrocities was related to (although not an independent risk factor for) PTSD. These findings suggest that PTSD and DESNOS may have distinct trajectories, rather than that DESNOS simply compounds PTSD.

These findings could be viewed as extending, rather than challenging, the associated features model. DESNOS could be viewed as a feature of PTSD caused by developmental adversities not only in childhood but also in later formative periods (e.g., the transition from adolescence to adulthood). However, there was no evidence that PTSD was in any way intrinsic to the association of DESNOS with early childhood or atrocity participation trauma: regardless of whether PTSD was present, both trauma types were independent risk factors for DESNOS. Thus, it may be more parsimonious to view DESNOS as a sequelae of developmentally adverse trauma independent of PTSD. Nevertheless, distinct post-traumatic etiologies cannot be certainly inferred for PTSD and DESNOS without longitudinal study with a sample sufficiently large and representiative to permit the reliable deployment of multivariate analyses (e.g., confirmatory factor and path analyses) with comprehensive assessment of the full range of types of trauma and of potentially mediating or moderating vulnerability and resiliency factors (L. King, King, Fairbank, Keane, & Adams, 1998).

Prior studies have reported an association between history of childhood trauma with PTSD in military veterans, perhaps due to methodological differences with the present study. Bremner et al.is (1993) comparison group included nonpsychiatric combat controls who rarely reported past child abuse, in contrast to this studyis no-PTSD psychiatric controls (who had moderate rates of child trauma). Zaidi and Foy (1994) and Engel et al. (1993) assessed PTSD symptom severity rather than diagnosis, and the latter study found PTSD symptoms associated with history of childhood trauma only for female veterans (replicating the present studyis null finding for male veterans). Of course, no prior study assessed DESNOS separately from PTSD. Thus, while many veterans with war trauma exposure may also have experienced childhood trauma, PTSD is of particular concern due to the war trauma, while childhood trauma confers additional risk for DESNOS. It should be noted that the low rate of childhood sexual abuse in this sample, while consistent with other studies with male veterans (Southwick et al., 1993), may be a substantive reason for the difference between this study and the Field Trial study. Childhood sexual abuse has profound effects (Finkelhor, 1995) that may lead to much higher rates of comorbid PTSD and DESNOS than other early traumas.

War atrocity participation, although clearly associated with early childhood trauma, r(84) = .45, p < .01, was independently predictive of DESNOS. Atrocity exposure has been shown to be an independent risk factor with a direct causal pathway to PTSD, its effect on PTSD was relatively small (L. King et al., 1998). The present studyis findings suggest that war atrocitiesi effect on PTSD may be due largely to indirect exposure (i.e., witnessing), while a separate pathogenic process may link direct atrocity exposure (participation) to the more extreme post-traumatic sequelae characterizing DESNOS. Participation in war atrocities as a young adult may constitute either (a) a further aspect of DESNOSis traumatic etiology beyond early childhood trauma, (b) a preclinical manifestation of DESNOS (i.e., atrocity participation itself often involves extreme affect dysregulation, dissociation, impulsivity, and disavowal of fundamental moral beliefs and commitments; Laufer, 1988), or (c) some combination of both.

PTSD, DESNOS, and Severity of Symptoms and Impairment

Hypothesis 2 proposed that DESNOSis effect on symptomatic or functional impairment occurs largely in interaction with PTSD or as parallel main effects. However, there was no interaction of DESNOS with PTSD in the MANCOVA (nor on a univariate basis), andDESNOS and PTSD had separate main effects on different variables. Similar results obtained in a direct test comparing participants with PTSD-only versus those with DESNOS-only. Specifically, DESNOS (but not PTSD) was independently related to extreme post-traumatic intrusive reexperiencing symptoms on the Impact of Events scale (IES-I), while PTSD diagnosis was associated with especially severe global post-traumatic impairment on the Mississippi and Penn PTSD measures. PTSD and DESNOS had opposite main effects for object relations, respectively positive and negative.

These apparently incompatible findings raise the question of why DESNOS would be a severity marker for one PTSD symptom scale (IES-I) and PTSD diagnosis for two other measures (Penn, Mississippi) that include many similar post-traumatic symptoms. One possibility is methodologic, in that the Mississippi and Penn measures are phrased in trait-like terms of general level of distress or difficulty, whereas the IES uses a state-like measurement frame (i.e., "past 7 days"). However, the GSI employs a state-like measurement frame similar to that for the IES, and the BDI and STAI use a trait-like methodology similar to that for the Mississippi and Penn, and these other scales showed no main effect for either PTSD or DESNOS. While the absence of a PTSD or DESNOS effect on any psychiatric symptom measure (also including the DES for dissociation) may reflect a ceiling effect for these more general psychiatric symptoms among chronically post-traumatically impaired individuals (Friedman & Rosenheck, 1996; Yehuda & McFarlane, 1995), the presence of differential effects for DESNOS and PTSD on the trauma-related measures suggests a substantive difference .

DESNOS may best be viewed as a severity marker (Newman et al., 1997) for post-traumatic impairment specifically associated with its cardinal features, i.e., extreme dysregulation of affect and consciousness. DESNOS appears to involve a variety of post-traumatic alterations in bodily and cognitive processing of emotionally-charged information, which may be due to neurobiological sensitization (Friedman, 1994), fragmented information processing (van der Kolk & Fisler, 1995), or unmodulated activation of cognitive-affective associational networks (Foa, Riggs, & Gershuny, 1994). Living with what often is described by sufferers as an unpredictable and overwhelming flood or firestorm of terrifying and humiliating emotions, thoughts, and impulses may be understood as an undampened positive feedback cycle in which unmodulated distress is associated with reliving the out of control experience of trauma, with escalating sensitization and a sense of being entrapped and re-victimized by oneis own body and mind. This is a terrible but precise recipe for intrusive trauma reexperiencing that is not just disturbing but terrifying, uncontrollable, and self-damaging. Thus, an association of DESNOS with extreme intrusive reexperiencing is plausible and of potential value to clinicians in understanding and treating trauma survivors who seem to have "more" than PTSD.

Chronic PTSD, by contrast, may be associated with the severity of more global post-traumatic psychosocial impairment (Friedman & Rosenheck, 1996). PTSDis cardinal features of are trauma-related hyperarousal that triggers intrusive reexperiencing, and the attempt to cope via hypervigilant or emotionally numbed avoidance (Yehuda & McFarlane, 1995). In this regard, it is interesting that, unlike the IES, the Penn includes several items assessing problems with personal and spiritual goal setting and attainment, and with self-knowledge, self-control, and self-efficacy (i.e., #2, 10, 14, 15, 18, 20, 23), and the Mississippi includes five items which have been shown factor analytically to comprise a distinct factor representing suicidal ideation and guilt (#2, 8,10, 12, 15; D. King & King, 1994)oas well as items loading on a factor for diminished self-efficacy in relational and work contexts (e.g., #5, 19). While it is not known whether a PTSD diagnosis correlated most strongly with these specific Penn and Mississippi items, it is possible that prolonged PTSD is distinguished by particularly severe emotional numbing and demoralization (Foa et al., 1994)oi.e., a chronic state of down-regulation, in contrast to DESNOSis apparent inability to down-regulate intense affects.

The directionally opposite relationship found for DESNOS versus PTSD with object relations may provide a clue to understanding why DESNOS but not PTSD is associated with particularly extreme intrusive reexperiencing. The extreme affective and impulsive lability, self-fragmentation, pathological dissociation, existential confusion, and interpersonal conflict characterizing DESNOS (Herman, 1992) seem likely to compromise an individualis core object relations. And without the coping capacities and social connectivity afforded by mature object relations, trauma reexperiencing may go largely unchecked and rarely be soothed or grounded by either self-regulation or social support. Prior studies have reported that object relational capacities positively predict (Ford et al., 1997) and DESNOS negatively predicts (Ford & Kidd, in press) PTSD treatment . Low object relations often accompanied DESNOS, and may explain why in DESNOS trauma reexperiencing might escalate unchecked because the individual is unable to successfully modulate intense affects.

It is important to note that PTSD may also be uncorrelated or negatively correlated with object relations. The positive correlation between PTSD and object relations most likely is attributable to an idiosyncracy of this sample, namely that PTSD-diagnosed participants had a restricted range of scores in the midrange on the object relations scale, and participants not meeting criteria for PTSD had more variability in object relations while tending to score at a lower level. Three participant clusters were identified empirically: PTSD/moderate object relations; no PTSD/moderate object relations; and no PTSD, low object relations (Ford et al., 1997). Why the fourth logical group, PTSD/low object relations, was not present is unknown, but may be due a propensity of veterans with chronic PTSD and poor object relations to eschew formal treatment due to alienation and isolation. PTSD often will be independent of object relations, because most traumas are severe but limited stressors that do not fundamentally impair psychological development (Yehuda & McFarlane, 1995). PTSDis relationship to object relations is more likely to be inverse rather than positive for those traumas that have high PTSD prevalences and adversely affect core object relational capacities (e.g., prolonged sexual molestation or physical abuse in early childhood; Cicchetti & Toth, 1995).

Treatment Utilization Associated with PTSD and DESNOS

DESNOS, but not PTSD, was an independent risk factor for recent or lifetime inpatient psychiatric treatment (controlling for the effect of major depression and personality disorder). DESNOS without PTSD was associated with as high a likelihood of recent or lifetime psychiatric hospitalization, as was DESNOS with comorbid PTSD. Although all participants were admitted to an inpatient PTSD treatment program at the time of the study, this admission was distinct from other forms of psychiatric inpatient hospitalization in being a nonacute residential milieu program rather than the more costly (due to high nursing and medical staffing intensity) and crisis-oriented (due to typically unscheduled acute admissions) nature of other inpatient psychiatric episodes. As Ronis, Garfein, Buit, Falcon and Liberzon (1996) showed in the larger VHA healthcare system, the military veteran trauma survivors in this study tended to be episodic high utilizers of psychiatric and PTSD services. Recognizing DESNOS as a particular risk for inpatient psychiatric episodes may help to reduce the stigma too often attached to chronic PTSD and increase clinicianis ability to distinguish veterans with chronic PTSD who are not at high risk for crises from those who may have full or only partial PTSD but whose care should include crisis prevention and management.

Conclusion

Limitations of the study include a sample restricted to chronic PTSD-treatment-seeking war veterans (although representative of this subpopulation, Fontana & Rosenheck, 1997), and reliance on Department of Veterans Affairs records for treatment utilization (although recent findings suggest that these sites represent the primary source of intensive psychiatric care for chronic military-related PTSD; Ronis et al., 1996). PTSD and DESNOS diagnoses were done by the same clinician rather than by independent clinicians, but with blind independent reliability ratings. Trauma history data were subject to retrospective biases (e.g., memory decay, motivated under- or overreporting, state dependency), but were collected via detailed historical interviewing with standard protocols, reliability checks, and a sampling the full range of childhood and war traumata. The study design was cross-sectional and neither longitudinal nor prospective, making causal inferences only tentative.

DESNOS itself is a new and complicated classification that warrants careful construct validity research (e.g., via a multitrait multimethod approach). The important question of the integrity and specificity of DESNOS classification within the broader nomological network of post-traumatic, psychiatric, and psychosocial constructs deserves further study. DESNOS assessed dichotomously appears discriminable from a dichotomously diagnosed PTSD, however, in its specific trauma correlates (i.e., early childhood trauma and war atrocity participation), extreme intrusive trauma reexperiencing symptoms and object relational impairment, and elevated risk of inpatient psychiatric treatment. When DESNOSis self-regulatory problems are present, trauma-focused treatment may require particularly careful grounding with ongoing community-based case management (Friedman & Rosenheck, 1996), therapeutic titration of the intensity of affect and reexperiencing (Ford & Kidd, in press), and strategic episodes of milieu care integrating affect regulation skills with trauma processing (Ford et al., 1997). The sequelae of early childhood and atrocity trauma may involve"traumagenic dynamics" (e.g., a sense of betrayal and stigmatization; Finkelhor, 1995) that go beyond PTSD and may best be understood in terms of DESNOS.

References

American Psychiatric Association. (1994). Diagnostic and statistical manual. (Fourth Revision). Washington, DC: Author.

Barends, A., Westen, D., Leigh, J., Silbert, D., & Byers, S. (1990). Assessing affect-tone of relationship paradigms in TAT and interview data. Psychological Assessment, 2, 329-332.

Beck, A., Steer, R., & Garbin, M. (1988). Psychometric properties of the Beck Depression Inventory. Clinical Psychology Review, 42, 841-865.

Bernstein, E., & Putnam, F. (1986). Development, reliability, and validity of a dissociation scale. Journal of Nervous and Mental Disease,1 74, 727-735.

Boscarino, J. A. (1997). Diseases among men 20 years after exposure to severe stress. Psychosomatic Medicine, 59, 605-614.

Bremner, J., Southwick, S., & Charney, D. (1995). Etiology of post-traumatic stress disorder. In M. Mazure (Ed.), Does stress cause mental illness? (pp. 68-104). Philadelphia: Saunders.

Bremner, J.D., Southwick, S., Brett, E., Fontana, A., Rosenheck, R., & Charney, D. (1992). Dissociation and post-traumatic stress disorder in Vietnam combat veterans. American Journal of Psychiatry, 149, 328-332.

Bremner, J., Southwick, S., Johnson, D., Yehuda, R., & Charney, D. (1993). Childhood physical abuse and combat-related post-traumatic stress disorder in Vietnam veterans. American Journal of Psychiatry, 150, 235-239.

Chemtob, C. M., Novaco, R., Hamada, R., Gross, D., & Smith, G. (1997). Anger regulation deficits in combat-related post-traumatic stress disorder. Journal of Traumatic Stress, 10, 17-36.

Cicchetti, D., & Toth, S., (1995). A developmental perspective on child abuse and neglect. Journal of the American Academy of Child and Adolescent Psychiatry, 34, 541-565.

Derogatis, L. (1992). SCL-90: Administration, scoring and procedures manual--II. Baltimore: Clinical Psychometric Research.

Engel, C., Engel, A., Campbell, S., McFall, M., Russo, J., & Katon, W. (1993). Post-traumatic stress disorder symptoms and precombat sexual and physical abuse in Desert Storm Veterans. Journal of Nervous and Mental Disease, 181, 683-688.

Finkelhor, D. (1995). The victimization of children: a developmental perspective. American Journal of Orthopsychiatry, 65,177-193.

Foa, E., Riggs, D., & Gershuny, B. (1994). Arousal, numbing, and intrusion. American Journal of Psychiatry, 152, 116-120.

Follette, V., Polusny, M., Bechtle, A., & Naugle, A. (1996). Cumulative trauma. Journal of Traumatic Stress,9, 25-36.

Fontana A., & Rosenheck R. (1997). Effectiveness and cost of the inpatient treatment of post-traumatic stress disorder. American Journal of Psychiatry,154, 758-765.

Ford, J. D., Fisher, P., & Larson, L. (1997). Object relations as a predictors of treatment outcome with chronic PTSD. Journal of Consulting and Clinical Psychology, 65, 547-559.

Ford, J. D., & Kidd, P. (in press). Early childhood trauma and Disorders of Extreme Stress as predictors of treatment outcome with chronic PTSD. Journal of Traumatic Stress.

Friedman, M. J. (1994). Neurobiological sensitization models of post-traumatic stress disorder: their possible relevance to multiple chemical sensitivity syndrome. Toxicology and Industrial Health, 10, 449-462.

Friedman, M. J., & the Matsunaga Study Investigators. Matsunaga Vietnam veterans project. Washington, DC: Department of Veterans Affairs.

Friedman, M. J., & Rosenheck, R. (1966). PTSD as a chronic disorder. In S. Soreff (Ed.), Handbook for the treatment of the seriously mentally ill (pp. 369-389). Seattle, WA: Hogrefe & Huber.

Friedman, M.J., & Schnurr, P.P. (1995). The relationship between trauma, post-traumatic stress disorder, and physical health. In M. Friedman, D. Charney & A. Deutch. (Eds.), Neurobiological and clinical consequences of stress. (pp. 507-534). Philadelphia: Raven

Hammarberg, M. (1992). Penn Inventory for post-traumatic stress disorder. Psychological Assessment, 4, 67-76.

Henning, K., & Frueh, B. C. (1997). Combat guilt and its relationship to PTSD symptoms. Journal of Clinical Psychology, 53, 801-808.

Herman, J. (1992). Complex PTSD. Journal of Traumatic Stress, 5, 377-391.

Horowitz, M. J. (1986). Stress response syndromes. (2nd Ed.) Northvale, NJ: Aronson.

Joseph, S., Dalgleish, T., Thrasher, S., & Yule, W. (1997). Impulsivity and post-traumatic stress. Personality and Individual Differences, 22, 279-281.

Keane, T., Caddell, J., & Taylor, K. (1988). Mississippi Scale for Combat-Related Post-traumatic Stress Disorder. Journal of Consulting and Clinical Psychology, 56, 85-90.

Keane, T., Fairbank, J., Caddell, J., Zimering, C., Taylor, K., & Mora, C. (1989). Clinical evaluation of a measure to assess combat exposure. Psychological Assessment, 1, 53-55.

King, D., & King, L. (1994). Latent structure of the Mississippi Scale for Combat-Related Post-Traumatic Stress Disorder. Assessment, 1, 275-291.

King, L., King, D., Fairbank, J., Keane, T., & Adams, G. (1998). Resilience-recovery factors in post-traumatic stress disorder among female and male Vietnam veterans. Journal of Personality and Social Psychology, 74, 420-434.

King, D., King, L., & Gudanowski, D., Vreven, D. (1995). Alternative representations of war zone stressors. Journal of Abnormal Psychology, 109, 184-196.

Kulka, R., Schlenger, W., Fairbank, J., Hough, R., Jordon, B., & Marmar, C. (1990). Trauma and the Vietnam generation. New York: Brunner/Mazel.

Laufer, R. (1988). The serial self. In J. Wilson, Z. Harel, & B. Kahana. (Eds.) Human adaptation to extreme stress (pp. 33-53). New York: Plenum.

Lifton, R.J. (1979). The broken connection. New York: Simon & Schuster.

Newman, E., Orsillo, S., Herman, D., Niles, B., & Litz, B. (1995). Clinical presentation of disorders of extreme stress in combat veterans. Journal of Nervous and Mental Disease, 183, 628-632.

Parson, E.R. (1988). Post-traumatic self disorders (PTSD). In J. Wilson, Z. Harel, B. Kahana (Eds.), Human adaptation to extreme stress. (pp. 245-283). New York: Plenum.

Pelcovitz, D., van der Kolk, B., Roth, S., Mandel, F., Kaplan, S., & Resick, P. (1997). Development of a criteria set and a structured interview for disorders of extreme stress (DESNOS). Journal of Traumatic Stress, 10, 3-16 .

Prigerson, H., Bierhals, A., Kasl, S., Reynolds, C., Shear, M. K., Day, N, Beery, L, Newsom, J., & Jacobs, S. (1997). Traumatic grief as a risk factor for mental and physical disability. American Journal of Psychiatry, 154, 616-623.

Resnick, H., Kilpatrick, D., Dansky, B., Saunders, B., & Best, C. (1993). Prevalence of civilian trauma and post-traumatic stress disorder in a representative national sample of women. Journal of Consulting and Clinical Psychology, 61, 984-991.

Ronis, D., Bates, E., Garfein, A., Buit, B., Falcon, S., & Liberzon, I. (1996). Longitudinal patterns of care for patients with post-traumatic stress disorder. Journal of Traumatic Stress, 9, 763-781.

Roth, S., Newman, E., Pelcovitz, D., van der Kolk, B., & Mandel, F. (1997). Complex PTSD in victims exposed to sexual and physical abuse. Journal of Traumatic Stress, 10, 539-555.

Southwick, S.M., Yehuda, R., & Giller, E. (1993). Personality disorders in treatment-seeking combat veterans with post-traumatic stress disorder. American Journal of Psychiatry, 150, 1020-1023.

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

Spitzer, R., Williams, J., Gibbon, M., & First, M. (1990). Structured Clinical Interview for the DSM-III-R, personality disorders (SCID-II). Washington, DC: American Psychiatric Press.

Spitzer, R., Williams, J., Gibbon, M., & First, M. (1990a). Structured Clinical Interview for the DSM-III-R, patient version (SCID-P). Washington, DC: American Psychiatric Press.

Unger, W., Gould, R., & Babick, W. (1998). The development of a scale to assess war-time atrocities: the War Events Scale. Journal of Traumatic Stress, 11, 375-383.

van der Kolk, B., & Fisler, R. (1995). Dissociation and the fragmentary nature of traumatic memories: overview and an exploratory study. Journal of Traumatic Stress, 8, 505-525.

van der Kolk, B., Pelcovitz, D., Roth, S., Mandel, F., McFarlane, A., & Herman, J. (1996). Dissociation, somatization, and affect dysregulation: the complexity of adaptation to trauma. American Journal of Psychiatry,153 (7 Festschrift Supplement), 83-93.

Weaver, T., & Clum, G. (1993). Early family environments and traumatic experiences associated with borderline personality disorder. Journal of Consulting and Clinical Psychology, 61, 1068-1075.

Weiss, D., & Marmar, C. (1997). Impact of Events Scale-Revised. In J. Wilson & T. Keane (Eds.), Assessment of PTSD (pp. 399-411). New York: Guildford.

Westen, D., Barends, A., Leigh, J., Mendel, M., & Silbert, D. (1990). Social cognition and object relations scale: Manual for coding interview data. Ann Arbor, MI: University of Michgan.

Yehuda, R., & McFarlane, A. (1995). The conflict between current knowledge about post-traumatic stress disorder and its original conceptual basis. American Journal of Psychiatry, 152, 1705-1713.

Yehuda, R., Southwick, S., & Giller, E. (1992). Exposure to atrocities and severity of chronic post-traumatic stress disorder in Vietnam combat veterans. American Journal of Psychiatry 149, 333-336.

Zaidi, L., & Foy, D. (1994). Childhood abuse experiences and combat-related PTSD. Journal of Trauma Stress, 7, 33-42.

Zlotnick, C., Zakriski, A., Shea, M.T., Costello, E., Begin, A., Perlstein, T., & Simpson, E. (1996). The long-term sequelae of sexual abuse. Journal of Traumatic Stress, 9, 195-205.