Journal of Truurnutir Strrss. Vol. 12. No. 4.

1999

Evaluation of Dream Content in

Combat-Related PTSD
Karin Esposito,’ Amparo Benitez,’ Lydia Bana,’ and Thomas Mellman’.*

Nightmares that replicate traumatic events are among the criteria that define
posttraumatic stress disorder (PTSD);however, there has been limited systematic
assessment of the content of dreams in PTSD. A PTSD dream rating instrument
was developed to operationalize features attributed to dreams associated with
combat-related PTSD that incorporated dimensions from the dream content anal-
ysis literature, and its reliability was assessed. Then the instrument was used to
characterize dream reports in 18 Vietnam combat veterans diagnosed with PTSD.
Approximately half of the group’s target dreams contained features characteristic
of combat, and all but 3 ( 17%) were moderately to highly threatening. Fifry-three
percent were set at least partially in ihe present and 79% contained distorted el-
ements. We therefore conclude that target dreams of combat veterans with PTSD
vary with regard to replication of trauma and elements normally associated with
dreaming, but typically are threatening.
KEY WORDS: posttraumatic stress disorder; combat; dreams: nightmares.

Reexperiencing criteria for posttraumatic stress disorder (PTSD)in the

Diagnostic and Statistical Manual of Mental Disorder (DSM-IV; American Psy-
chiatric Association [APA], 1994) include “recurrent distressing dreams of the
event,” further described in the text as dreams “during which the event is re-
played.” The association between €TSD and d r e a m that replicate trauma has
received the most support from studies of combat veterans. Van der Kolk, Blitz,
Burr, Sherry, and Hartmann (1984) reported that Vietnam veterans who met cri-
teria for PTSD were more likely to state that their nightmares exactly or almost

‘Department of Psychiauy and Behavioral Science. University of Miami School of Medicine, and VA
Medical Center, Miami, Florida.
*Towhom correspondence should be addressed at Miami VAMC,Psychiatry Service (116A). 1201
N.W. 16fhStreet, Miami. Florida 33125-1693;e-mail: TMell~~medner.med.miami.edu.

681
0894-9867/99/1oM)-w81S16.wII0 1999 hsmtionrl Society for Traumatic Stress SNdiCi
a2 Esposito, Benitez, Bana, and Mellman

exactly replicated an actual event than lifelong nightmare sufferers without combat
experience. Mellman, Kulick-Bell, Ashlock, and Nolan ( 1995) found endorsement
of disturbing dreams about combat experience to be more specifically associated
with PTSD than disturbing dreams about other topics. In a recent secondary anal-
ysis of the National Vietnam Veterans Re-adjustment Study database, a measure
integrating the frequencies of nightmares related to military experiences and dis-
tress provoked by dreaming was found to be “virtually specific for PTSD” and to
represent the domain of sleep disturbance most related to exposure to war-zone
trauma (Neylan et al., 1998).
Although they support an integral relationship between combat-related PTSD
and nightmares, the aforementioned studies are limited by having assessed dreams
retrospectively and by use of restrictive or nonsystematic assessment of dream con-
tent. Dow, Kelsoe, and Gillin (1996) applied content ratings to dreams reported
from awakenings induced during rapid eye movement ( E M ) sleep. The main
finding was that the dreams of Vietnam veterans with depression and PTSD were
more likely to be set in the past in comparison to reports from non-ill but depressed
groups. Kramer, Schoen, and Kinney (1984) reported findings from a sleep lab-
oratory evaluation of Vietnam veterans with and without disturbed dreaming and
described dreams from spontaneous awakenings preceded by REM and non-REM
sleep stages. Although infrequent in the comparison group, approximately half
of the dreams of the study group contained military references. This finding was
viewed as supporting the continuing mental processing of military trauma as well
as other concerns during dreaming. Hartmann (1998) has suggested that the rep-
resentation of threat in nonreplicative settings in dreams following trauma serves
to contextualize the emotional concern of the dreamer in a manner that can be
adaptive.
Thus, information on dreaming can be informative regarding the status of
emotional processing in PTSD. Kramer et al. (1984) also have linked dream dis-
turbance to greater severity of psychological distress among combat veterans. The
occurrence of “repetitive replicas” of combat scenes in the dreams of veterans with
PTSD in contrast to dreams normally associated with REM sleep that have bizarre
and implausible elements led Ross, Ball, Sullivan, and Caroff (1989) to propose a
REM abnormality in the disorder.
The DSM-IV ( M A , 1994) criterion and these implications for pathogenesis
and prognosis underscore the need for additional systematic assessment of the con-
tent of dreams in PTSD subjects. The initial aim of our study was to operationalize
the construct of a PTSD dream in the context of combat trauma. A rating scale was
developed that incorporated the aforementioned features attributed to dreams in
PTSD and constructs from the empirically based dream content analysis literature
(Hall & van de Castle, 1966). After assessing the reliability of the instrument,
it was utilized to characterize target dreams from a treatment-seeking group of
combat veterans with chronic PTSD.
Dream Evaluation in Combat-Related PTSD 683

Methods

Participants

Veterans were recruited from consecutive appointments with one of the inves-
tigators (TAM) in the PTSD outpatient clinic (n = 4) or consecutive admissions to
a residential program for PTSD rehabilitation (n = 14) at the Department of Vet-
erans Affairs Medical Center, Miami, Florida. The mean age of participants was
50.21 (SD= 4.03, range = 41 to 61); 58% of participants were White, 32% were
Black, and 11% were Hispanic. All participants were Vietnam combat veterans
and had a current principal diagnosis of PTSD based on the Structured Clinical In-
terview for DSM-IV, PTSD section, and the Mississippi Scale for Combat-Related
PTSD. Military discharge forms (DD-2-14) had been reviewed to confirm combat
exposure. Subjects were further determined not to have a primary psychotic dis-
order and were monitored to exclude substance abuse disorders that were active
in the preceding 2 months. All were being treated with one or more psychotropic
medications, including antidepressants, mood stabilizers, benzodiazepines, and
antihistamines.

Procediires

A dream rating scale (DRS) for combat-related PTSD was developed that in-
corporated parameters previously utilized in dream content analysis (Hall & van de
Castle, 1966).These include the setting, characters, and objects in the dream; de-
gree of overt threat to the dreamer; contemporaneity; and degree of distortion
from actual or plausible events. “PTSD-like” dreams were conceptualized as fea-
turing sertings, charucters, and objecrs similar to traumatic experiences, along
with high threat, low conremporuneiry, and low distortion. Ratings were made
after hearing a detailed description of the dream with clarifications as needed.
For each parameter, anchoring descriptions for ratings from 0 to 4 were devel-
oped. For all parameters, 4 represented the extreme most consistent with PTSD
(i.e., setting of combat, weapon objects, high threat, set in the past, low distor-
tion), 0 represented the opposite extreme (i.e., noncombat settings, no overt threat,
etc.), and 2 represented intermediate representations or mixtures (see Tables 1
and 2).
The four investigators reviewed several dreams together and established con-
sensus ratings and then evaluated five dreams concurrently blind to each others’
ratings. Interrater reliability and internal consistency were evaluated by calculating
the intraclass correlation coefficient and Cronbach’s alpha, respectively.
For the descriptive portion of the study a spontaneously recalled dream was
rated in all 19 subjects. PTSD severity for the month preceding the target dream was
684 Esposito, Benitez, Bana, and Mellman

Table 1. Overview of DRS

Dimensions Anchor-Point Guidelines"

Setting 4-characteristic of combat

Character 2-military reference. noncombat; mixture of combailnoncombat
objects 0-civilian, no military references
Threat 4-catastrophic, death possible
2-discomfort
0-friendly,comfortable
Contemporaneity 4-past or future
2-mixture of pastlpresent
0-present
Distortion 4-replication of actual event
2-did not occur but plausible
0-bizarre, magical

"Anchor points are described for extreme (4.0) and middle (2) nting points; ntings of 1 or 3 are used
when proportion of elements in dream falls between middle and extreme points.

Table 2. Examoles of Dream Descriotion S u m r i e s with DRS Scores

Esurnplu I
A two-man team went out and came back with a dead VC, air-fire was called in. then we went into the
village with grenades; the nex! morning a woman came with a dead baby and blamed me.
Setting 4 Threat 3 Total score 12
Chancters 4 Contemporanrity 3
Objects 4 Distortion 4
E.roniplr 2
Fishing off a very high bridge; a young kid asks to borrow my net and I wam him not to lose it; he
throws i t in. jumps after it and drowns; I hear a firefight at the police station 10 miles away.
Setting 1 Threat 2 Total score 6
Chancters 0 Contemponneity 2
Obiects 0 Distonion 1

assessed in each subject using the Clinician-Administered FTSD Scale Severity

(CAPS;Blake et al., 1990). Functional impairment was assessed with the employ-
ment/housework, interpersonal relationship, and sexual activities subscales of the
Longitudinal Interval Follow-up Evaluation (LIFE; Keller et al.. 1987) adminis-
tered by a psychology technician independently of the dream evaluations.
To enhance recall and increase accuracy, participants were provided morning
diaries to record dream reports for each morning of the week preceding the as-
sessment. Dream reports elicited on awakening in the morning do vary somewhat
from, and tend to be more dramatic than, mentation elicited after experimen-
tally awakening subjects out of REM sleep (Domhoff & Kamiya, 1964). A recent
study indicates that morning reports have characteristics similar to reports from
Dream Evaluation in Combat-Related PTSD 685

awakenings induced from late-night REM periods, however (St. Onge et al., 1999).
Slightly more than half of the participants recorded more than one dream and were
asked to describe the one that was most typical of dreams they had been having.
This method balanced the number of dreams per participant and also corresponded
to the DSM-IV symptom criteria specification of dreams that are recumng.
The target dream was rated by an investigator who was blinded as to that
veteran's CAPS and LIFE scores. Frequencies of aream parameters were tabulated
and relationships within DRS subscales and between total DRS and CAPS and
LIFE scores were explored with Pearson correlations.

Results

DRS

Interrater reliability and internal consistency were high (intraclass correlation

coefficient = .99, Cronbach's alpha = .93, respectively). Correlations ofDRS sub-
scales were lowest for threat with the other parameters: characters (r(18) = -58,
p < .Ol), objects (r(l8) = .66, p < .01), setting (r(18) = .44.ns), contempo-
raneio ( r ( 18) = .22, ns), and distortion (r( 18) = .25, ns). All other correlations
between parameters exceeded r(18) = .60 ( p c .01).

Content Ratings

Frequencies of dream content ratings were calculated. In summary, slightly

less than one half of dreams reported contained settings (47%).characters (47%),
or objects (42%)characteristic o f combat (i.e., ratings = 3.4);only three dreams
(14%) were rated as low in threat ratings (0, I), 53% were set at least partially
in the present (contemporaneity = 0, 1, 2) and 79% demonstrated some degree
of distortion (ratings 5 3; i.e., only 21% of dreams were rated as having exactly
replicated an event).
CAPS andLIFE scores did not correlate with total DRS scores ( r (18) = -. 19,
ns). Total DRS did significantly correlate with the CAPS reexperiencing cluster
subscore (r(18) = .58, p < .05).

Conclusions

Our experience indicates that dream content can be rated reliably in combat
veterans with FTSD along dimensions referred to in the FTSD and dream content
analysis literatures (American Psychiatric Association, 1996; Dow et al., 1996;
Hall & van de Castle, 1966; Ross et al., 1989; Van der Kolk et al., 1984). Modifying
the DRS to apply to other populations with PTSD also might be of interest.
686 Esposito, Benitez, Barza, and Mellman

Despite sampling considerations that might have biased toward reporting of

ESD-like elements, dream descriptions varied with regard to dimensions that
reflect the degree to which traumatic combat experiences were replicated. This
observation is consistent with Kramer et al. (1984) reporting that 47% of dreams
elicited in the sleep laboratory from combat veterans with disturbed dreaming
contained military references. The data do support a relationship between the
construct of a PTSD-like dream and overall seventy of reexperiencing symptoms.
Although not consistently replicative or set in the past, all but three of the dreams
rated in the study were at least moderately threatening to the dreamer. Although
inferences regarding specificity to PTSD are limited here by the absence of a control
group, other findings suggest that threat is neither an unusual nor a predominate
theme of normative dreaming. Specifically, Hall and van de Castle (1966) found in
a large, nonclinical sample that aggression occurred in 46% of dreams and that the
dreamer was a victim of the aggression in 42% of these dreams (i.e.. 19% of dreams
involved aggression to the dreamer). Whereas Hartmann (1988) has pointed to an
adaptive function of recontextualizing a dominant emotion, in these symptomatic
combat veterans threatening dreams may contribute to generalization of a threat
response over time.
The lack of significant relationships between dream ratings and measures of
FTSD severity and psychosocial functioning could be a consequence of the limited
number of subjects and the homogenous nature of the study group. Factors in
addition to the focus on symptomatic combat veterans that limit the generalizability
of the study include the use of medications that often were prescribed to help
maintain sleep.
The study does support that reported dream content can be measured and that
recent, representative dreams reported by patients with chronic, combat-related
PTSD typically feature threat, some degree of reality distortion, and variably repli-
cate trauma. The characterization in the DSM-IV for dreams to replay a traumatic
event may be justified by its specificity to the diagnosis, but appear to capture only
a subset of distressing dreams experienced by patients with PTSD.

References

American Psychiatric Association. ( 1994). Diugnosfic und stufisficulrncmuul ojrnenfaf disorders (4th
ed.). Washington DC: Author.
Blake, D. D., Weathers, F. W., Nagy, L. M., Kaloupek, D. G., Klauminzer, G., Chmey. D. S., & Keane,
T. M. (1990). A clinician rating scale for assessing cumnt and lifetime PTSD: The CAPS-I.
Behavior Therupist. 13,181-188.
hmhoff, B.. & Kamiya, J. (1964). Problems in dream content study with objective indicators: 1. A
comparison of home and labontory dream reports. Archives ofCeneru1 Psychiutry, IJ, S 19-524.
Dow. B. M.. Kelsoe, J. R., Jr., & Gillin, J. C. (1996). Sleep and dreams in Vietnam FTSDand depression.
Biological Psychiatry, 39.42-SO.
Hall, C. S.. & van de Castle, R. L. (1966). The confent analysis of dreams. New York Appleton-
Century-Crofts.
Dream Evaluation in Combat-Related PTSD 687

H ~ ~ I I - I E.
~ I(1998).
M . Nightmre after t n u m as paradigm for all dreams: A new approach to the nature
and function of dreaming. Phyrhiufry,61,223-238.
Keller, M . B.. Lavori, P. W., Friedman, B., Nielsen, E., Endicott. J., McDonald-Scott, P.,& Andreasen,
N. ( 1987). The longitudinal interval follow-up evaluation. Archives of General Psyrhiutry, 44.
540-548.
Kramer. M.,Schoen, L. S., & Kinney, L. (1984). The dream experience in dream-disturbed Vietnam
veterans. In van der Kolk (Ed.), Posffraumaticsiress disorders: Psyrhological and biolo.qiru1
sequelue (pp. 82-95). Washington. DC: American Psychiatric Association.
Me1lman.T. A., Kulick-Bell, R., Ashlock, L. E.. & Nolan, B.(1995). Sleep events among veterans with
combat-related posttraumatic stress disorder. Amerirun Journal of P.vchiufry. 152, I 10-1 15.
Neylan. T. C., M m r , C. R.. Metzler. T. J.. Weiss, D. S., Zauick, D. F., Delucchi, K. L.. Wu. R. M.,&
Schoenfeld, F.B. (1998). Sleepdisturbances in the Vietnam generation: Findings frorna nationally
representative sample of male Vietnam vetenns. Amerirun Journal o/Pvrhiurry. I55.929-933.
ROSS,R. J.. Ball, W. A., Sullivan, K. A,, & Caroff, S. N. (1989). Sleep disturbance as the hallmark of
posttnumatic stress disorder. American Journul of Psyrhiurry, 146,697-707.
St. Onge. M. A.. Grenier, J.. Vinette, S., Vachon. J., Dupuis-Roy, N.. Roussy, F.. & de Koninck. J.
(1999). A comparison of visual imagery in labontory and home dreams. Sleep. 22 (Suppl. I),
170.
Van der Kolk, B.. Blitz, R., Burr, W., Sherry, S . , & Hart-.E. (1984). N i g h t m e s and tnum: A
comparison of nightmares after combat with lifelong nightmares in vetenns. Amerirun Journul
of Pqchiufry 141. 187-190.