Sleep Pathophysiology in Posttraumatic Stress Disorder

and Idiopathic Nightmare Sufferers

Anne Germain and Tore A. Nielsen

Background: Nightmares are common in posttraumatic Introduction

stress disorder (PTSD), but they also frequently occur in
idiopathic form. Findings associated with sleep distur-
bances in these two groups have been inconsistent, and
sparse for idiopathic nightmares. The aim of the present
R apid eye movement (REM) sleep disturbances and
nightmares have been suggested to be the hallmark of
posttraumatic stress disorder (PTSD) (Ross et al 1989).
study was to investigate whether sleep anomalies in PTSD The presence of REM sleep disturbances remains equiv-
sufferers with frequent nightmares (P-NM) differ from ocal, however, and no profile of sleep disturbances unique
those observed in non-PTSD, idiopathic nightmare (I-NM) to PTSD has yet been established (e.g., see Pillar et al
sufferers and healthy individuals. 2000 for review). One study showed that the nature of
Methods: Sleep measures were obtained from nine P-NM posttraumatic nightmares may shape sleep profiles in
sufferers, 11 I-NM sufferers, and 13 healthy control PTSD inpatients (Woodward et al 2000). Nightmares are
subjects. All participants slept in the laboratory for two part of the intrusion symptom cluster of PTSD (American
consecutive nights where electroencephalogram, electro- Psychiatric Association 1994) and are reported by as many
oculogram, chin and leg electromyogram, electrocardio- as 75% of PTSD patients (e.g., Kilpatrick et al 1994).
gram, and respiration were recorded continuously. Although spontaneous awakenings following distressing
Results: Posttraumatic nightmare sufferers had signifi- dreams are rarely observed when PTSD patients sleep in a
cantly more nocturnal awakenings than did I-NM sufferers laboratory environment, chronic and frequent distressing
and control subjects. Elevated indices of periodic leg dreams (associated or not with a sudden awakening) may
movements (PLMs) during rapid eye movement (REM) nevertheless influence sleep profiles in PTSD patients in a
and non-REM sleep characterized both P-NM and I-NM
pervasive manner. In other words, we postulate that sleep
sufferers.
disturbances in PTSD may be in part due to nightmare
Conclusions: Posttraumatic nightmare sufferers exhibit pathophysiology, rather than being exclusively attributable
more nocturnal awakenings than do I-NM sufferers and
to PTSD-specific processes.
control subjects, which supports the hypothesis of hyper-
Two studies conducted with individuals reporting fre-
arousal in sleep in PTSD sufferers; however, elevated
PLM indices in both P-NM and I-NM sufferers suggest quent nightmares but who do not suffer from PTSD
that PLMs may not be a marker of hyperarousal in sleep observed sleep disturbances comparable to those observed
of PTSD sufferers. Rather, PLMs may be a correlate of in PTSD patients, including increased or unaltered phasic
processes contributing to intense negative dreaming. REM sleep activity, decreased total sleep time, increased
Biol Psychiatry 2003;54:1092–1098 © 2003 Society of number and duration of nocturnal awakenings, and de-
Biological Psychiatry creased slow-wave sleep (Fisher et al 1970; Newell et al
1992). Another study (van der Kolk et al 1984) compared
Key Words: Idiopathic and posttraumatic nightmares, sleep complaints in nightmare sufferers with PTSD (n ⫽
sleep disturbances, periodic leg movements in sleep 15) and individuals with idiopathic nightmares (n ⫽ 10).
Posttraumatic stress disorder–related nightmares tended to
occur earlier in the night, were more frequent, and were
more often associated with gross body movements than
were idiopathic nightmares; however, a lack of objective
From the Sleep Research Center (AG, TAN), Hôpital du Sacré-Cœur de Montréal;
sleep measures obviates the possibility of determining
and the Departments of Psychology (AG) and Psychiatry (TAN), Université de more specifically whether these two groups had different
Montréal, Montréal, Québec, Canada.
Address reprint requests to Anne Germain, Ph.D., University of Pittsburgh School
types of sleep disturbances. Together, these observations
of Medicine, Department of Psychiatry, 3811 O’Hara Street, Suite E-1116, suggest that the relationship between chronic frequent
Pittsburgh PA 15213.
Received January 17, 2002; revised July 10, 2002; revised November 27, 2002;
nightmares and sleep anomalies may not be exclusive to
accepted January 6, 2003. PTSD. Rather, these findings further support the notion

© 2003 Society of Biological Psychiatry 0006-3223/03/$30.00

doi:10.1016/S0006-3223(03)00071-4
Sleep Pathophysiology in Nightmare Sufferers BIOL PSYCHIATRY 1093
2003;54:1092–1098

that sleep disturbances in PTSD patients may be in part a from an advertisement in the same university newspaper. Be-
function of the nightmare psychopathology rather than cause of limited resources available for the present study,
other, more global PTSD processes. eligibility of control subjects was assessed during a telephone
In fact, no study has investigated whether PTSD pa- interview and through the completion of self-report measures
(see below). All reported being good sleepers, free of sleep and
tients with frequent nightmares differ from idiopathic
dreaming disturbances, and otherwise met the same inclusion and
nightmare sufferers on laboratory-recorded sleep parame-
exclusion criteria.
ters. Idiopathic nightmare sufferers offer a unique oppor- The Sacré-Coeur Hospital Ethics Committee approved the
tunity to determine whether the sleep disturbances ob- study. Written and verbal consent was obtained from all partic-
served in PTSD patients are due to intrinsic ipants.
pathophysiologic factors proper to PTSD, or whether some
of these anomalies are attributable to nightmare patho-
physiology. Thus, the goal of the present study was to Self-Report Measures
investigate whether the sleep attributes of PTSD patients At intake, all participants completed the Beck Depression Inven-
with frequent nightmares differ from those of idiopathic tory (BDI; Beck et al 1961), the Beck Anxiety Inventory (BAI;
nightmare sufferers and healthy participants matched for Beck et al 1988), the Nightmare Distress Questionnaire (NDQ;
Belicki 1992), and the Posttraumatic Symptom Scale, self-report
age and gender.
version (PSS-SR; Foa et al 1993). The BDI is a 21-item
self-report questionnaire that assesses severity of the behavioral,
Methods and Materials cognitive, emotional, and somatic symptoms associated with
depression. The BAI is a similar 21-item self-report checklist that
Participants assesses the severity of anxiety-related symptoms. Both the BDI
Self-referred nightmare sufferers were recruited mainly from and the BAI are scored by summing responses for each of the 21
advertisements in the University of Montreal’s campus newspa- items, with each item rated on a 0 –3 scale. The NDQ is a 13-item
per and following a short televised documentary on the study and self-report scale that measures the level of waking distress
treatment of nightmares that aired in the evening. To enter the associated with the experience of nightmares. This instrument
study, participants had to be at least 18 years of age and to report has been shown to be reliable; high scores are significantly
recalling more than one nightmare per week for a minimum of 6 correlated with interests in pursuing therapy for nightmares
months. They were excluded if 1) they were currently under (Belicki 1992). Finally, the PSS-SR is a measure of PTSD
medications known to influence sleep and dreams; 2) they were severity according to DSM-III-R criteria (American Psychiatric
currently suffering from a major psychiatric disorder other than Association 1987), and it evaluates the severity of intrusion,
PTSD; 3) they reported currently suffering from another sleep avoidance, and arousal symptoms in the preceding 2-week
problem; 4) they suffered from a neurologic disorder; 5) they period. On all measures, higher scores reflect greater symptom
reported irregular sleep–wake schedules or had undergone jet lag severity.
in the previous 3 months; 6) they reported using alcohol or drugs
on a regular basis; or 7) they were currently engaged in legal
Polysomnography
proceedings involving events related to their nightmares. Eligi-
bility was ascertained during extensive clinical interview con- All participants slept in the laboratory for two consecutive nights.
ducted by either one of the two authors. Nine individuals with Sleep recordings were performed with a 32-channel montage that
PTSD and nightmares (P-NM; four men and five women, age measured electroencephalogram (EEG) with the international
[mean ⫾ SD] ⫽ 39.0 ⫾ 12.1 years) and 11 individuals with 10 –20 electrode placement system (FP1, FP2, F3, F4, F7, F8,
idiopathic nightmares (I-NM; five men and six women; age ⫽ C3, C4, T3, T4, T5, T6, P3, P4, O1, O2, Fz, Cz, Pz), eye
28.2 ⫾ 5.3 years) met these criteria and participated in the study. movements (LOC-A2, ROC-A1), EMG (submental and right
For participants who reported that the onset of nightmares tibialis), oral-nasal airflow, and electrocardiogram. A referential
occurred after exposure to a traumatic event (i.e., P-NM suffer- montage (with linked-ears reference) was used to record the 19
ers), PTSD status was determined using the Clinician’s Assess- EEG channels. Recordings were performed using RHYTHM
ment of Posttraumatic Stress (Blake et al 1990). Three P-NM version 10.0 (Stellate Systems, Montréal, Québec, Canada) and
sufferers (patients 1, 2, and 9) met DSM-IV criteria for current were scored manually, according to Rechtschaffen and Kales
depressive episodes in the severe range; however, all three (1968) criteria using HARMONY version 4.1 (Stellate Systems),
sufferers attributed their current depressive symptoms to the by an experienced polysomnographic technician who had not
PTSD-related events (i.e., financial problems, social isolation, conducted the sleep recordings and who was blind to the purpose
and imminent prison release of the aggressor); they were thus of the study. Sleep onset latency (SOL) was computed as the
included. One I-NM patient (patient 1) reported a history of interval between lights out and the first episode of any sleep
substance abuse but had been abstinent for 6 years, whereas none stage. Periodic leg movements (PLMs) were scored according to
of the PTSD patient did. Coleman’s criteria (Coleman 1982), and the PLM indices were
Seven men and six women (age ⫽ 32.6 ⫾ 11.2 years) computed as the number of PLM ⫻ 60/number of minutes of
constituted the control group. These subjects were paired for age sleep. Rapid eye movement density was computed as the
and gender to the I-NM and P-NM sufferers. They were recruited absolute number of REMs (Tashibana et al 1994) during the last
1094 BIOL PSYCHIATRY A. Germain and T.A. Nielsen
2003;54:1092–1098

Table 1. Descriptive Information on P-NM and I-NM Patients

PTSD
NM PTSD Chronicity
Gender Age Chronicitya Trauma Severity (years)
P-NM
1 M 28 3 Car crash Severe 3
2 F 46 4 Rape Severe 4
3 F 58 53 Sexual abuseb Severe 14
4 M 35 25 Physical abuseb Moderate 25
5 F 40 10 Physical and sexual abusec Moderate 10
6 F 40 3 Rape Moderate 3
7 M 46 20 Parachute accident Moderate 20
8 F 22 5 25-foot fall Moderate 5
9 M 36 2.5 Physical assault Severe 2
I-NM
1 F 36 30 Kidnappedb None
2 M 19 13
3 M 30 10
4 F 27 9 Father was shot None
5 M 36 25
6 M 24 19 Witnessed armed robbery None
7 F 25 13
8 F 30 25 Intruder in the house None
9 M 29 24
10 F 31 2
11 F 23 2.5
P-NM, posttraumatic stress (PTSD) nightmare; I-NM, idiopathic nightmare; M, male; F, female
a
Nightmare chronicity provided in years, and did not differ between groups.
b
Trauma occurred in childhood.
c
Trauma occurred during adolescence.

5 minutes of each REM sleep episode and then averaged over all and Newman-Keuls post hoc comparisons were also conducted
REM sleep episodes. Micro-arousals were identified as abrupt to assess group differences on the self-report measures of
changes in EEG frequency with a minimal duration of 3 sec and psychological distress. The significance level was set at .05.
a maximal duration of 10 sec and could include alpha or theta
frequencies but not spindles. A minimal interval of continuous
sleep of 10 sec was necessary to score a second micro-arousal
(American Sleep Disorders Association 1992). The first night Results
was considered an adaptation night, and only results collected
from the second night are reported. Bedtime was between 10:00
Sample Characteristics and Self-Report Measures
PM and midnight, depending on each participant’s usual bedtime. Table 1 presents information on the age, gender, night-
The morning awakening was conducted between 6:00 and 8:00 mare chronicity, type of trauma, and PTSD severity (when
AM, again depending on each participant’s typical schedule. In applicable) of all nightmare sufferers. Four of the I-NM
the morning, electrodes were removed and participants were free sufferers also reported past traumatic events, but the onset
to go for the day. Before leaving and after the first recording
of nightmares preceded the trauma in all cases. None of
night, they were reminded to avoid caffeine consumption and
naps during that day. All participants received a monetary these met the criteria for past or current PTSD. Age tended
compensation of $20 per night slept in the laboratory. to differ across the three groups [F(2,32) ⫽ 3.11, p ⫽ .06];
P-NM sufferers were relatively older than I-NM sufferers
(p ⫽ .05). (Because the present study does not have
Statistical Analyses sufficient statistical power to further discriminate regard-
Statistica 5.1 software (StatsSoft, Tulsa, OK) was used. Square ing the contribution of age from that of nightmare etiol-
root transformations were applied to SOL, wake time after sleep ogy, age was not entered as a covariate in subsequent
onset, number of awakenings, REM latency, REM density, and
analyses.) Nightmare chronicity did not differ between the
all variables related to PLM. Logarithmic transformations were
applied data related to sleep and REM sleep latencies, sleep two groups of nightmare sufferers [F(1,19) ⫽ .09, ns].
stages, sleep efficiency, and REM sleep efficiency. One-way Control subjects did not report significant levels of psy-
analyses of variance (ANOVAs) were then computed and New- chological distress. One control subject reported a trau-
man-Keuls post hoc comparisons performed. One-way ANOVAs matic event (motor vehicle accident) 2 years before
Sleep Pathophysiology in Nightmare Sufferers BIOL PSYCHIATRY 1095
2003;54:1092–1098

Table 2. Scores on Measures of Anxiety, Depression, Nightmare Distress, and Posttraumatic Symptom Severity (when applicable),
for P-NM, I-NM, and Control Subjects
P-NM (n ⫽ 9) I-NM (n ⫽ 11) CTL (n ⫽ 13) F (df) p Post-hoc Comparisons
Anxiety 18.22 (9.58) 9.27 (10.55) 6.58 (5.07) 5.07 (2,31) .01 P-NM ⬎ CTL, p ⬎ .01
Depression 21.56 (11.10) 7.36 (6.18) 3.33 (3.47) 17.49 (2,31) ⬍.001 P-NM ⬎ I-NM, p ⬍ .001
P-NM ⬎ CTL, p ⫽ .001
Nightmare Distress 39.89 (7.37) 34.81 (6.75) 13.25 (11.44) 24.18 (2,31) ⬍.001 P-NM ⬎ CTL, p ⫽ .001
I-NM ⬎ CTL, p ⬍ .001
Posttraumatic Symptom 34.31 (6.93) 5.40 (4.06) 2.50 (2.84) 129.42 (2,29) ⬍.001 P-NM ⬎ I-NM, p ⬍ .001
Severity P-NM ⬎ CTL, p ⬍ .001
Values are mean (SD). P-NM, posttraumatic stress (PTSD) nightmare; I-NM, idiopathic nightmare; CTL, control subjects.

participating in the study but did not meet present past or exhibited more nocturnal awakenings than both I-NM and
current PTSD symptoms. control groups (p ⫽ .007 and p ⫽ .01, respectively). The
Table 2 presents mean scores on the self-report mea- latter two groups did not differ on these three measures.
sures, including prospective nightmare frequency for the The groups did not differ on any of the REM sleep
three study groups. Significant group differences were parameters.
observed on all measures of psychological distress. The Mean percent sleep stage scores are presented in Table
P-NM group endorsed significantly higher scores on the 4. None of these sleep stage variables, including REM
BDI, BAI, NDQ, and PSS-SR than did the control group sleep percent, differentiated the groups.
(all p ⬍ .05). The P-NM group also endorsed higher scores
on the BDI, PSS-SR, and NDQ than the I-NM group (all Periodic Leg Movements during Sleep
p ⬍ .05). The I-NM group endorsed higher scores than did
control group on the NDQ (p ⬍ .001), whereas the two Table 5 presents results for PLMs in REM and non-REM
groups were comparable on the BDI, BAI, and PSS-SR. sleep. Differences across the three groups were found for
all PLM indices, with and without associated micro-
arousals, in both REM and non-REM sleep. In all cases,
Polysomnography P-NM and I-NM sufferers exhibited elevated PLM indices
Mean polysomnography scores for the three groups are compared with control subjects, but did not differ from
presented in Table 3. Sleep efficiency significantly dif- one another.
fered across the three study groups [F(2,29) ⫽ 3.50, p ⫽
.04]. This was attributable to differences in the total
Discussion
number of nocturnal awakenings [F(2,29) ⫽ 5.61, p ⫽
.002], and total wake time after sleep onset [F(2,29) ⫽ Consistent with prior studies (Engdahl et al 2000; Kramer
4.13, p ⫽ .03] across the three groups. The P-NM group and Kinney 1988; Lavie et al 1979; Mellman et al 1995;

Table 3. Polycomnography Scores for the Three Study Groups: P-NM sufferers, I-NM sufferers, and Control Participants
P-NM (n ⫽ 9) I-NM (n ⫽ 11) CTL (n ⫽ 13) Fa Post-hoc Comparisons
b
SOL 11.7 (8.5) 14.0 (17.2) 11.2 (13.1) .12, ns
TST 376.0 (48.1) 328.5 (101.0) 403.0 (65.7) 3.90, ns
WASOc 70.6 (61.2) 30.9 (13.9) 31.2 (24.2) 4.13, p ⫽ .03 P-NM ⬎ CTL, p ⫽ .03
P-NM ⬎ I-NM, p ⫽ .02
No. Awakeningsc 44.8 (21.3) 25.4 (9.7) 26.5 (9.7) 5.61, p ⫽ .009 P-NM ⬎ CTL, p ⫽ .007
P-NM ⬎ I-NM, p ⫽ .01
REM Latencyb 83.4 (27.2) 82.3 (31.9) 99.0 (6.6) .32
Sleep Efficiencyd 85.1 (11.5) 92.1 (3.3) 92.7 (5.8) 3.50, p ⫽ .04 P-NM ⬍ CTL, p ⫽ .04
P-NM ⬍ I-NM, p ⫽ .06
REM Efficiencyd 79.1 (17.1) 90.8 (9.1) 87.7 (11.0) 1.93
REM Densityc 5.4 (4.7) 6.8 (7.2) 4.4 (3.8) .96
Micro-Arousalsc 8.6 (5.3) 6.6 (3.2) 8.3 (4.2) .46
Values are mean (SD). P-NM, posttraumatic stress (PTSD) nightmare; I-NM, idiopathic nightmare; CTL, control subjects; SOL, sleep onset latency; TST, total sleep
time; WASO, wake time after sleep onset, REM, rapid eye movement.
a
All df ⫽ 2,30.
b
Log transformation Ln performed before analyses. Mean values are presented in original units.
c
Square root transformation performed before analyses. Mean values are presented in original units.
d
Log transformation Ln (100-n-1) performed before analyses. Mean values are presented in original units.
1096 BIOL PSYCHIATRY A. Germain and T.A. Nielsen
2003;54:1092–1098

Table 4. Average Percent Sleep Stages for the Three Study support the hypothesis that a lowered arousal threshold
Groups: P-NM Sufferers, I-NM Sufferers, and Control characterizes sleep in PTSD (Brown and Boudewyns
Participants
1996; Mellman et al 1995; Ross et al 1989). The present
P-NM I-NM CTL results may not contradict findings from three previous
(n ⫽ 9) (n ⫽ 11) (n ⫽ 13) F
studies that have shown that PTSD patients exhibit higher
% S1a 12.2 (5.6) 8.2 (3.1) 8.4 (4.2) 2.36b awakening thresholds during sleep than do control sub-
% S2a 64.5 (7.3) 60.6 (8.6) 64.2 (4.4) 1.23b
jects (Dagan et al 1991; Lavie et al 1998; Schoen et al
% S3a 3.6 (3.7) 8.4 (4.9) 5.1 (3.9) 1.78c
% S4a .2 (.4) 1.7 (2.7) .8 (1.5) .03d 1984). It remains possible that emotional–attentional pro-
% REMa 19.5 (5.1) 21.2 (4.4) 21.5 (3.6) .81b cesses are shifted inward toward intensified negative sleep
Values are mean (SD). P-NM, posttraumatic stress (PTSD) nightmare; I-NM, mentation in PTSD sufferers, rather than outward toward
idiopathic nightmare; CTL, control; REM, rapid eye movement. external stimuli. Intensified oneiric processes may thus be
a
Log transformation Ln (n) performed before analyses. Mean values are
presented in original units. related to an abnormal emotional–attentional shift during
b
df ⫽ 2,30 sleep, increased number of nocturnal awakenings, and an
c
df ⫽ 2,28
d
df ⫽ 2,16 increase in body movements during sleep.
The finding that all PLM indices were elevated in both
van der Kolk et al 1984), P-NM sufferers exhibited lower groups of nightmare sufferers relative to control subjects
sleep efficiency than did I-NM sufferers and healthy does not support the hypothesis that PLMs are a specific
control participants. Poorer sleep efficiency in P-NM correlate of hyperarousal in P-NM patients (Brown and
sufferers was due to increases in the number and duration Boudewyns 1996; Ross et al 1989, 1994). Examination of
of nocturnal awakenings in P-NM sufferers compared with the PLM inter-movement intervals indicated that gross
I-NM and control subjects. Posttraumatic nightmare suf- body movements, as well as PLMs, may be more frequent
ferers did not differ from I-NM sufferers or control in both groups of nightmare sufferers. Despite the low
subjects on any of the REM sleep and sleep stage PLM indices and the questionable clinical significance of
measures. Both P-NM and I-NM sufferers exhibited ele- PLMs (e.g., Mahowald 2001; Monplaisir et al 2000), the
vated PLM indices compared with control subjects. De- present observations nevertheless bear empirical signifi-
spite the relatively small sample sizes attributable to cance in indicating that PLMs/gross body movement may
stringent selection criteria, and the consequent lack of be closely related to nightmare pathophysiology, or more
statistical power to control for the potential confound of globally, to abnormal oneiric processes. The direction of
age on these sleep parameters, the present study neverthe- this relationship, however, remains unclear. Abnormal
less provides preliminary support for the hypothesis that central motor activation patterns may be translated at
sleep anomalies in PTSD may in part be a function of higher cortical levels into vivid terrifying dream imagery
PTSD-specific processes, and partially attributable to per- (Ross et al 1994) in both P-NM and I-NM sufferers.
vasive effects of chronic frequent nightmares. Alternatively, intense negative dreaming may facilitate the
The findings that the P-NM group exhibited more release of motor inhibition during all sleep stages. This is
nocturnal awakenings than the other two groups further not to say that leg or body movements are necessarily

Table 5. Scores for Periodic Leg Movements in REM and non-REM Sleep,a with and without Micro-arousals for P-NM Sufferers,
I-NM Sufferers, and Control Participants
P-NM (n ⫽ 9) I-NM (n ⫽ 11) CTL (n ⫽ 13) Fb Post-hoc Comparisons
PLMs 7.0 (5.6) 8.8 (8.9) 1.9 (2.7) 8.93 (p ⬍ .001) P-NM ⬎ CTL, p ⫽ .004
I-NM ⬎ CTL, p ⫽ .002
PLMs ⫹ MA 2.1 (1.7) 1.9 (1.2) .6 (.9) 7.70 (p ⫽ .001) P-NM ⬎ CTL, p ⫽ .005
I-NM ⬎ CTL, p ⫽ .005
REM PLMs 14.2 (8.6) 18.61 (17.9) 5.2 (9.1) 7.96 (p ⫽ .002) P-NM ⬎ CTL, p ⫽ .005
I-NM ⬎ CTL, p ⫽ .004
REM PLMs ⫹ MA 2.7 (2.7) 1.8 (1.8) .5 (1.0) 1.77 (ns) P-NM ⬎ CTL, p ⫽ .008
Non-REM PLMs 5.2 (4.9) 5.7 (7.9) 1.0 (1.4) 5.94 (p ⫽ .007)
I-NM ⬎ CTL, p ⫽ .02
Non-REM PLMs ⫹ MA 1.9 (1.5) 1.7 (1.4) .6 (1.0) 6.03 (p ⫽ .006) P-NM ⬎ CTL, p ⫽ .02
I-NM ⬎ CTL, p ⫽ .008
Values are mean (SD). REM, rapid eye movement; P-NM, posttraumatic stress (PTSD) nightmare; I-NM, idiopathic nightmare; CTL, control subjects; PLMs, periodic
leg movements; MA, micro-arousals.
a
Square root transformation performed before analyses. Mean values are presented in original units.
b
All df ⫽ 2,30
Sleep Pathophysiology in Nightmare Sufferers BIOL PSYCHIATRY 1097
2003;54:1092–1098

correlates of nightmare content, in a manner analogous to References

the motor correlates of REM-sleep behavior disorder American Psychiatric Association (1987): Diagnostic and Sta-
(Mahowald and Schenck 2000), but rather to claim that tistical Manual of Mental Disorders, 3rd ed, revised. Wash-
body movements may reflect an underlying disposition of ington, DC: American Psychiatric Press.
system dysfunction that produces intense negative dream- American Psychiatric Association (1994): Diagnostic and Sta-
ing both in and out of PTSD. Neuroimaging developments tistical Manual of Mental Disorders, 4th ed. Washington, DC:
in the study of sleep (e.g., Nofzinger et al 1998) may American Psychiatric Press.
clarify the neuroanatomic substrates underlying the rela- American Sleep Disorders Association (1992): EEG arousals:
tionships between waking anxiety, disturbing dreams, and Scoring rules and examples. A preliminary report from the
increased motor activity during sleep in P-NM and I-NM Sleep Disorders Atlas Task Force of the American Sleep
Disorders Association. Sleep 15:174 –184.
sufferers.
The main limitations of the present study are inherent to Beck AT, Epstein N, Brown G, Steer R (1988): An inventory for
measuring clinical anxiety. J Consult Clin Psychol 56:893–
the sample sizes in both groups of nightmare sufferers due 897.
to the use of stringent selection criteria. Replications with
Beck AT, Ward C, Mendelson M, Mock J, Erbaugh J (1961): An
larger samples are required to further characterize the inventory for measuring depression. Arch Gen Psychiatry
specific contribution of nightmare etiology in shaping 4:561–571.
sleep disturbances both in and out of PTSD. Larger Belicki K (1992): Nightmare frequency versus nightmare dis-
samples will provide the statistical power necessary to tress: Relations to psychopathology and cognitive style. J
control for the potential confounding effects of age on Abnorm Psychol 101:592–597.
sleep, as well as to investigate which qualities of night- Blake DK, Weathers FW, Nagy LM, Kaloupek DG, Klauminzer
mares (e.g., frequency vs. distress) more directly influence G, Charney DS, Keane TM (1990): A clinician rating scale
sleep parameters in PTSD and non-PTSD patients. An- for assessing current and lifetime PTSD: The CAPS-1. Behav
other limitation concerns the absence of a group of PTSD Ther 13:187–188.
subjects without frequent nightmares; however, because as Brown TM, Boudewyns PA (1996): Periodic limb movements of
many as 75% of PTSD patients endorse frequent night- sleep in combat veterans with posttraumatic stress disorder.
J Trauma Stress 9:129 –135.
mares (e.g., Kilpatrick et al 1994), the feasibility of
Coleman RM (1982): Periodic limb movements in sleep (noc-
recruiting nonmedicated PTSD patients without night-
turnal myoclonus) and restless leg syndrome. In: Guillemi-
mares remains uncertain. Finally, eligibility of control nault C, editor. Sleeping and Waking Disorders: Indications
participants did not include the use of structured diagnos- and Techniques. Menlo Park, CA: Addison Wesley.
tic interview. Although the levels of psychopathology Dagan Y, Lavie P, Bleich A (1991): Elevated awakening
endorsed by the control participants were below clinically threshold in sleep stage 3– 4 in war-related post-traumatic
significant thresholds, the possibility that subsyndromal stress disorder. Biol Psychiatry 30:618 –622.
psychopathologies may have been present cannot not be Engdahl BE, Eberly RE, Hurwitz TD, Mahowald MW, Blake J
ruled out. (2000): Sleep in a community sample of elderly war veterans
Despite these limitations, this is the first study to with and without posttraumatic stress disorder. Biol Psychi-
atry 47:520 –525.
investigate the pathophysiology of sleep in nonveteran
P-NM sufferers in comparison with I-NM sufferers and Fisher C, Byrne J, Edwards A, Kahn E (1970): A psychophysi-
ological study of nightmares. J Am Psychoanal Assoc
healthy individuals. The results support and refine the
18:747–782.
sleep-related hyperarousal hypothesis during sleep in
Foa E, Riggs D, Dancu C, Rothbaum B (1993): Reliability and
PTSD but indicate that increased motor activity during
validity of a brief instrument for assessing posttraumatic
sleep is a correlate of chronic frequent nightmares both stress disorder. J Trauma Stress 6:459 –473.
inside and outside of PTSD. In addition, the inclusion of
Kilpatrick DG, Resnick HS, Freedy JR, Pelcovitz D, Resick P,
I-NM sufferers allowed us to provide new data regarding Roth S, van der Kolk B (1994): Posttraumatic stress disorder
the neurophysiologic correlates of abnormal dreaming field trial: Evaluation of PTSD construct criteria A through E.
processes during sleep in non-PTSD nightmare sufferers. In: Widiger TA, Frances AJ, Pincus HA, Ross R, First MB,
Davis W, Kline M, editors. DSM-IV Sourcebook, volume 4.
Washington, DC: American Psychiatric Press.
Kramer M, Kinney L (1988): Sleep patterns in trauma victims
with disturbed dreaming. Psychiatr J Univ Ottawa 13:12–16.
This research was sponsored by the Canadian Institutes of Health Lavie P, Hefez A, Halperin G, Enoch D (1979): Long-term
Research, the Fonds de la Recherche en Santé du Québec (FRSQ), and effects of traumatic war-related events on sleep. Am J
Fondation J.A. DeSève. Psychiatry 136:175–179.
The authors thank Benoit Adam, Sonia Frenette, Sébastien Saucier,
Lavie P, Katz N, Pillar G, Zinger Y (1998): Elevated arousal
and Stéphanie Barbier for their technical support in this project.
threshold during sleep: Characteristics of chronic war-related
1098 BIOL PSYCHIATRY A. Germain and T.A. Nielsen
2003;54:1092–1098

posttraumatic stress disorder patients. Biol Psychiatry Pillar G, Malhotra A, Lavie P (2000): Post-traumatic stress
44:1060 –1065. disorder and sleep—What a nightmare! Sleep Med Rev
4:183–200.
Mahowald MW (2001): Assessment of periodic leg movements
is not an essential component of an overnight study. Am J Rechtschaffen A, Kales A (1968): A Manual for Standardized
Respir Crit Care Med 164:1340 –1341. Terminology, Techniques and Scoring System for Sleep
Stages of Human Subjects. Los Angeles: UCLA Brain Infor-
Mahowald MW, Schenck CH (2000): REM sleep parasomnia. mation Service: Coleman.
In: Kryger MH, Roth T, Dement WC, editors. Principles and
Practice of Sleep Medicine, 3rd ed. Philadelphia: W.B. Ross RJ, Ball WA, Dinges DF, Kribbs NB, Morrison AR, Silver
Saunders. SM, Mulvaney FD (1994): Motor dysfunction during sleep in
posttraumatic stress disorder. Sleep 17:723–732.
Mellman TA, Kulick-Bell R, Ashlock LE, Nolan B (1995): Sleep Ross RJ, Ball WA, Sullivan KA, Caroff SN (1989): Sleep
events among veterans with combat-related posttraumatic disturbance as the hallmark of posttraumatic stress disorder.
stress disorder. Am J Psychiatry 152:110 –115. Am J Psychiatry 146:697–707.
Montplaisir J, Michaud M, Denesle R, Gosselin A (2000): Schoen L, Kramer M, Kinney L (1984): Auditory thresholds in
Periodic leg movements are not more prevalent in insomnia the dream disturbed. Sleep Res 13:102.
or hypersomnia but are specifically associated with sleep Tashibana N, Sugita Y, Terashima K, Teshima Y, Shimizu T,
disorders involving a dopaminergic impairment. Sleep Med Hishikawa Y (1994): Scoring REM density. Neurology
1:163–167. 44:987–988.
Newell SA, Padamadan H, Drake ME (1992): Neuropsychologic van der Kolk B, Blitz R, Burr W, Sherry S, Hartmann E (1984):
studies of nightmares sufferers. Clin Electroencephalogr Nightmares and trauma: A comparison of nightmares after
4:203–206. combat with life-long nightmares in veterans. Am J Psychia-
Nofzinger EA, Mintum MA, Price J, Meltzer CC, Townsend D, try 141:187–190.
Buysse DJ, et al (1998): A method for the assessment of Woodward SH, Arsenault NJ, Murray C, Bliwise DL (2000):
functional neuroanatomy of human sleep using FDG PET. Laboratory sleep correlates of nightmare complaint in PTSD
Brain Res Protocols 2:191–198. inpatients. Biol Psychiatry 48:1081–1087.