Abstract
Sleep terrors (ST) and sleepwalking (SW) are partial arousal disorders, which occur in slow-wave sleep, usually during the first third of the sleep period [1,2]. Although the two sleep disorders have different clinical manifestations [3], they have a similar pathophysiology [4] and are related in terms of genetic [4,5] and developmental factors [6]. Sleep terrors and SW can be precipitated by fever [7], sleep deprivation, urinary bladder distention, a noisy environment or some central nervous system depressants [8]. Patients with ST or SW do not recall the attack in the morning upon awakening. In a study using the Minnesota Multiphasic Personality Inventory (MMPI) and the Symptom Check List (SCL-90), Kales et al. [9] reported that patients with ST had inhibited outward expressions of aggression and a predominance of anxiety, depression, obsessive-compulsive and phobic tendencies. In addition, Fisher et al. [10,11] reported that adult patients with ST were emotionally disturbed.
The frequency of ST and SW is greatest in childhood [3,12], decreases significantly in adolescence, and is lowest in adulthood. Jacobson et al. [3,13] found that of children aged 5–12 years, more than 15% had SW and 2–3% had ST. Our sleep habit questionnaire (SHQ) survey showed that of 930 junior high school students aged 13–15 years, 3.2% had had ST and 0.7% had had SW during the previous 6 months [14].
Based on the studies of Kales et al. [9] and Fisher et al. [10,11], who showed that those with ST and/or SW also suffer from emotional and personality problems, we proposed to examine this association in a community-based sample. Hence, this study was designed to investigate the psychiatric comorbidity and personality characteristics of those with ST and/or SW in a non-referred community sample of adolescents.
Methods
Study sample
A case-control design was used. In December 1993, among 70 junior high schools in Taipei City, two schools were selected randomly for the first-stage study. In each school, four classes of each grade level were selected as subjects. A total of 965 students, aged 13–15 years, and their parents completed a self-administered questionnaire (SHQ) about the students' sleep habits [14]. The response rate was 96.4% (930) for students and 88.6% (855) for parents. For detailed information on sampling and SHQ, please refer to the report of the first-stage study [14]. Of the 930 students surveyed, 32 had ST and/or SW during the previous 6 months. Of these 32, one student moved from Taiwan and another refused to be interviewed. This led to 30 subjects in the initial case group. The control group comprised the 30 students whose seat numbers were next to the case group students in the same classes.
Data collection
In November 1994, a child psychiatrist (S.F. Gau), who was blind to the subjects' SHQ responses, conducted an audiotaped psychiatric interview in the following sequence: the Chinese-version of the Kiddie-SADS-E (Schedule for Affective Disorders and Schizophrenia for Children—Epidemiology Version) [15] to determine psychiatric diagnoses and problems during the previous year; demographic data including sleep habits; and sleep disorders including insomnia, sleeptalking, nightmares, enuresis, bruxism, snoring and sleep paralysis before and during the previous year. The sleep disorders of immediate relatives were assessed by interviewing parents and siblings based on the diagnostic criteria of DSM-III-R. In addition, each subject completed a self-administered Junior Eysenck Personality Inventory (JEPI) [16] 1 day prior to the interview. Students were interviewed at school unless they had graduated or transferred, in which case they were interviewed at home. Parents and siblings were interviewed at home or by telephone in order to assess the personality characteristics and sleep disorders of the subjects and sleep disorders of the nuclear family members.
The presence or absence of ST and/or SW was determined by the subject's response to questions at the end of the psychiatric interview and confirmed by interviewing their parents and siblings. Subjects were classified as having ST and/or SW if they had had at least one episode within the previous year. Subjects were not divided into separate ST and SW groups because of high co-occurrence (only two case subjects had SW without ST). Due to recovery from ST and/or SW, six subjects were withdrawn from the case group. Because the diagnosis of ST and SW was based on identification of the symptoms and exclusion of organic disorders, another three subjects were excluded because they suffered from a physical illness: one had prolactinoma, one had hyperthyroidism and another had frequent asthmatic attacks during the previous year. There were 21 subjects remaining in the final case group and 30 subjects in the control group.
Study variables
Psychiatric abnormalities, personality characteristics, family history of ST and SW, and demographic data derived from the interviews and the questionnaires were used to contrast the case and control groups. To obtain more reliable information on the family history of ST and/or SW, only data on first-degree relatives of the subjects (i.e. parents and siblings) were collected and analysed.
Data analysis
The collected data were analysed statistically by SPSS/PC++ v6.0 (SPSS Inc., Chicago, IL, USA) and SAS v6.12 (SAS Institute Inc., Cary, NC, USA) with Wilcoxon rank-sum test for numeric variables. The Odds ratio (OR) statistic with 95% confidence interval (CI) and Chi-squared test or two-tailed Fisher's exact test was used for analysing categorical variables. For those two-by-two contingency tables with one empty cell, we estimated the Odds ratios with amendment [17]. We added 0.5 to each cell of those with one empty cell. The critical statistical confidence level selected for all statistical analyses was p < 0.05.
Results
Demographic data
There were eight boys and 13 girls in the case group with a mean age of 14 years and 4.1 months (SD = 11.0 months). There were 14 boys and 16 girls in the control group with a mean age of 14 years and 6.4 months (SD = 12.4 months). There were no significant differences between the case and the control group in terms of sex, age, grade, socioeconomic status [18] and birth order.
Sleep-related problems and physical problems
There were more complaints of fatigue (χ2 = 6.04, df = 1, p = 0.014) and low mood (Fisher's exact test, p = 0.023) in the morning in the case group. More girls in the case group complained of dysmenorrhea (Fisher's exact test, p < 0.01) than did those in the control group. There were no significant differences, at the p < 0.05 level, between the two groups regarding sleep setting, the main reason for going to bed, complaints of insufficient sleep and difficulty in awakening, or history of menstruation.
Psychiatric diagnosis and problems
Based on the criteria of DSM-III-R and ICD-10 for most diagnoses and only DSM-III-R for overanxious disorder, the case group had significantly (p < 0.05) more overanxious disorder, panic disorder, simple phobia, suicidal ideation, and cigarette or alcohol use. All of the diagnoses shown in Table 1 were based on both the diagnostic criteria of DSM-III-R and ICD-10 except overanxious disorder and panic disorders. Of the three case subjects with panic attacks, one had them at a frequency of more than once per week, reaching the criteria of panic disorder by DSM-III-R. The other two with frequencies of once per 2 weeks reached the criteria of mild panic disorder by ICD-10. Only one of the three had experienced sleep panic attack once.
Comorbidity of psychiatric diagnoses or problems for the past year
A total of 18 (85.7%) subjects from the case group had at least one disorder, compared to only eight (26.7%) subjects of the control group. Furthermore, there was a significant difference in the number of psychiatric diagnosis or problems per subject between the two groups. The case group averaged 2.5 (SD = 1.8) disorders per subject, whereas the control group averaged only 0.4 (SD = 0.8) disorder per subject (z = −4.68, p = 0.00).
Other sleep problems
Significantly more subjects from the case group suffered from sleeptalking (OR = 6.50, 95% CI = 1.93–21.94; χ2 = 9.28, df = 1, p < 0.001) and nightmares (OR = 5.33, 95% CI = 1.59–17.94; χ2 = 5.33, df = 1, p = 0.01) during the previous year and had a history of enuresis (OR = 14.50, 95% CI = 2.34-90.01; Fisher's exact test, p < 0.006) at primary school age. There were no statistically significant differences between the two groups on early insomnia, midnight waking, snoring, bruxium and sleep paralysis. All of the subjects in the case group and 20 in the control group had at least one of the other sleep problems in addition to ST and SW (OR = 22.02, 95% CI = 2.47-196.63; Fisher's exact test, p < 0.001). Subjects in the case group averaged 3.0 (SD = 1.3) additional sleep disorders, while subjects in the control group averaged only 1.6 (SD = 1.4) sleep disorders (z = −3.28, p = 0.001).
Personality characteristics from parental perspective
According to the parents' statements, more subjects in the case group were nervous, bad-tempered and pessimistic than were those of the control group (Table 2).
Self-administered scale for personality trait
Based on the JEPI, the case group had a significantly higher mean score on the Neuroticism Scale and a lower mean score on the Lie Scale (Table 2).
Personality traits from parental report and self-administered Junior Eysenck Personality Inventory (JEPI)
Family history of sleep terrors and sleepwalking
Eight (38.1%) of the case group and 4 (13.3%) of the control group had at least one immediate relative having ever experienced ST and/or SW. The frequency of ST and/or SW was statistically higher among nuclear family members of the case group than those of the control group (OR = 4.00, 95% CI = 1.05–15.24; Fisher's exact test, p = 0.05).
Discussion
Although the association of certain personality characteristics [12,19] and emotional problems [13,14,20] with ST and SW has been reported, this article is the first to report on the psychiatric comorbidity of adolescents with STand SW. In this study, the case group showed high psychiatric comorbidity, especially anxiety and phobic disorders, and a greater tendency for suicidal ideation. The adolescent subjects averaged a higher Neuroticism Score on the JEPI, and parental interview confirmed their neurotic and pessimistic tendencies. These findings were in keeping with the report of Kales et al. that clinical patients with ST had high rate of anxiety and depression [12]. Although the mood changes and personality characteristics of those who have perimenstrual complaints are controversial, it has been accepted that women with premenstrual syndrome and dysmenorrhea have high rates of depression and neuroticism [21,22] and adolescents with dysmenorrhea have been found to suffer from low self-esteem [23]. The finding of higher rate of dysmenorrhea among our female case subjects is consistent with previous studies suggesting an association between dysmenorrhea and anxiety-related symptoms.
Sleep terrrors and SW should be differentiated from other neuropsychiatric disorders such as brain tumour [12], nocturnal epilepsy [12,24,25], dissociation [12,19] and panic disorder [25–27]. Three subjects of the original case group had prolactinoma, hyperthyroidism and frequent asthma, respectively. Although these three cases were excluded from the final analysis, those who have problems of ST and/or SW may have physical illness or problems, and ST and SW may be clinical manifestations of physical illness or caused by medication. The greater number of sleep problems in addition to ST and SW among ST/SW subjects made us consider the interaction of different presentations of sleep-related disorders. When one sleep problem is observed, investigation into the possibility of other sleep problems is necessary. We should differentiate ST from sleep-related panic attacks [27,28], which can be associated with sudden awakenings from sleep, intense fear of impending doom, difficulty falling back to sleep, recall of the event in the morning, and absence of scream. Sleep terrors [1,3,9,28], however, are partial arousals from slow-wave sleep, are accompanied by a heralding scream, and are not associated with sleep onset difficulties, agoraphobia or daytime symptoms. Patients with panic disorder [29] often have sleep disturbances. They suffer from insomnia, fear of going to bed and midnight awakening from panic attacks, which often occur in light, stage 2 non-rapid eye movement sleep [30,31]. Nevertheless, previous studies did not reveal that patients with panic disorder also suffer from more sleep terrors or sleepwalking than the general population. In this study, comorbidity of panic disorders in three subjects (14.3%) of the case group encourages us to reconsider the relationship between these two diagnoses.
An anxious child may be more likely to describe all kinds of sleep and psychiatric problems than a non-anxious child, when we see the finding that 85.7% of the case subjects had at least one psychiatric disorder or problem. The fact is that we did psychiatric diagnosis first, then assessed sleep problems and finally ST and SW. The order of interview and the interviewer's blinding to subjects' status on ST and/or SW removed the possibility of interviewer's assessment bias. Furthermore, the presence or absence of sleep problems was confirmed by parents and siblings. Hence, the case subjects did not overestimate their problems but actually they suffered from more psychiatric and sleep problems at the same time. For clinical implications, however, it is important that those with adolescent onset of ST or SW have their complete history taken [32], receive a mental status examination and a physical check-up to reveal any psychopathology and/or organic problems.
Of the 21 adolescents with ST and/or SW, eight (38.1%) had at least one first-degree relative affected by ST and/or SW. This rate was similar to those of Kale et al. [7] and higher than those of Edwards [33]. The results of this community study and previous clinical studies support an increased familial occurrence of ST and SW. Based on the results of marginal statistical significance, small sample size and the limitation of this study design, we cannot draw any firm conclusion on familial aggregation of ST and SW.
In conclusion, even though the small sample size decreases the power to detect the true association, in this study, we still have some positive findings. This community-based study reveals that adolescents with ST and/or SW have more psychiatric comorbidity and other sleep disorders, are more neurotic and pessimistic, and have a higher frequency of ST and/or SW among their first-degree relatives. Regarding the limitation of the cross-sectional nature of this study design, the only conclusion we ultimately draw is that there is an association between higher rate of psychiatric problems and ST/SW. At this point, we do not infer a causal relationship between psychiatric disorders and ST/SW.
Footnotes
Acknowledgements
This study was supported by funds from a National Taiwan University Hospital Award, NTUH 84262-B26. We thank Kathleen R. Merikangas for permission to translate and use the Kiddie-SADS in our study and her comments on this article. We also thank the students and families who received the interviews, and their teachers.