To estimate genetic and environmental influences on parasomnias and dyssomnias and their association in 8-year-olds.
Parents of twins completed the Children's Sleep Habits Questionnaire.
Families were primarily tested at the Institute of Psychiatry,
A total of 300 pairs of 8-year-old twins and their parents participated in the study.
Main Outcome Measures
Sleep difficulties in children.
Individual differences in parasomnias and dyssomnias were largely explained by genes (accounting for 50% and 71% of the variances, respectively).
The rest of the variances were mainly due to nonshared environmental influences. A moderate association was found between parasomnias and dyssomnias (r = 0.42), which was mainly accounted for by genetic influences (87%). The genetic correlation between parasomnias and dyssomnias was moderate (r = 0.61). In contrast, the nonshared environmental correlation was small (r = 0.10).
The decomposition of the association between parasomnias and dyssomnias suggests that there may be different manifestations of shared underlying genetic risks for sleep problems partly dependent on nonshared environmental influence.
The Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition)1 distinguishes 2 main categories of primary sleep disorders, parasomnias and dyssomnias,
both of which may be experienced by children.2 Previous genetic research focusing on children has highlighted both genetic and environmental influences on specific sleep problems assessed individually (eg, sleepwalking and nightmares3) and sleep problems defined as a single category without distinguishing parasomnias and dyssomnias.4,5 Certain parasomnias and dyssomnias are known to co-occur in childhood,6,7 although little is known about the reasons for these associations. To address this issue, this article examines phenotypic and genetic associations between parasomnias and dyssomnias in a sample of twins aged 8 years.
Parasomnias are characterized by atypical behavior or physiologic events that occur in specific periods of sleep and include a range of difficulties, such as sleepwalking and bedwetting. Information is accumulating about both genetic and environmental risk factors for these disorders, with research showing that both types of influence are important in explaining individual differences in these sleep problems.6,8 Previous research has also examined the association among different types of parasomnias, showing that they co-occur in both children and adults and that this co-occurrence in adults can be partly explained by genes.3
Dyssomnias are characterized by an unusual amount, quality,
or timing of sleep and among other things include both insomnia and breathing-related sleep disorder. As with parasomnias, twin studies have revealed that both genetic and environmental factors have a role in the etiology of these difficulties in both children6 and adults.9- 11
Genetic analyses can provide information concerning reasons for the associations between disorders.12 Such analyses are potentially useful in the hunt for specific genetic or environmental influences on difficulties. For example, a high genetic correlation between different disorders (such as that found between anxiety and depression13,14) suggests that once we have identified specific genes associated with one disorder these may be worth examining in relation to the other disorder. Information about genetic influences on sleep difficulties is potentially useful in identifying individuals at risk for difficulties.
It is then possible to intervene to reduce the likelihood of difficulties developing. For these reasons, this study examined the extent to which parasomnias and dyssomnias co-occur during middle childhood and estimated genetic and environmental influences on this association.
All participants are members of the Emotions, Cognitions, Heredity and Outcome (ECHO) study, which consists of 300 twin pairs aged 8
years (range, 8 years 2 months to 8 years 11 months) at wave 1 of data collection. ECHO is a spin-off study from the longitudinal Twins Early Development Study (TEDS15),
focusing on twins born in England and Wales during 1994-1996. Data collection took place at the Institute of Psychiatry, London, England,
with the exception of a few families who were visited in their homes.
The ECHO study received ethical approval from the Maudsley Hospital Ethics Committee, London. Informed consent from parents was obtained through the mail before data collection.
Twins were selected for inclusion in the ECHO study sample based on their anxiety at 7 years. Anxiety was focused on when selecting participants because the current data were collected as part of a study investigating anxiety and links with other difficulties (eg,
sleep problems). This extremes design was used to maximize power.
Selecting participants by focusing on anxiety scores at the 7-year-old assessment is not a concern because the selected nature of the sample is corrected in all analyses.
Of the 5745 families in TEDS, parental data were available for 5343 twin-pairs, and 1378 of these families contained at least 1 child (or proband) scoring in the top 15% for parent-reported anxiety at 7 years. Of these families, 967 were excluded for various reasons (eg, 1 twin had a major medical condition). This left 411 potential proband pairs. An additional 30 families had moved, leaving 381 proband families to be invited to participate, of whom 247 pairs took part (65%). For the control subjects (n = 3965 pairs), the same criteria were applied, excluding 2794 pairs. From the remaining 1171
potential control families, a random 92 pairs were invited to participate,
of whom 53 participated (58%). The total number of families participating when the twins were 8 years of age was therefore 300, comprising 43
monozygotic male pairs, 57 monozygotic female pairs, 29 dizygotic male pairs, 53 dizygotic female pairs, 117 dizygotic opposite sex pairs, and 1 pair of unknown zygosity (who refused to provide DNA).
After testing, data from 11 twin pairs (4%) were considered unusable because at least 1 of the twins had neurologic impairments, autistic spectrum disorders, severe receptive language impairments, or persistent difficulties with attention. An additional 10 families (3%) did not have scores for the sleep measure reported herein because of missing data. Fifty-seven percent of the initial sample was female and 43%
male. Most families who participated in the ECHO study were white (n = 256 [87%]). Further information about the demographics of the sample can be found elsewhere.6,16
A parent-rated instrument was used to assign twin zygosity in the TEDS sample when the twins were 18 months old. This measure assigned zygosity accurately in 95% of cases when validated against DNA (for further discussion of zygosity ascertainment in the TEDS sample, see the article by Price et al17).
Anxiety was examined in the TEDS sample when the participants were aged 7 years using the parent-report Anxiety Related Behaviors Questionnaire.18 Twins with at least 1 member scoring in the top 15% of this measure were selected as the high-anxiety pairs in the ECHO study, whereas control pairs scored below this cutoff. The entire sample was analyzed as 1 group, taking this selection process into account.
Parents reported on their children's sleep problems using a 33-item version of the Children's Sleep Habits Questionnaire (CSHQ7). Each item was rated on a 3-point scale (1 indicating rarely to 3 indicating usually). The CSHQ consists of 8 subscales, focusing on bedtime resistance, sleep-onset delay, sleep duration, sleep anxiety, night waking, parasomnias, sleep-disordered breathing, and daytime sleepiness. When completing the CSHQ, parents were asked to report on the most recent typical week. The reliability and validity of this measure are demonstrated elsewhere.7,19
The focus of this article is on the parasomnias scale of the CSHQ (comprising 7 items including “child talks during sleep”
and “child grinds teeth during sleep”) and a measure of dyssomnias (comprising 10 items taken from 4 subscales of the CSHQ:
sleep-onset delay, sleep duration, night waking, and sleep-disordered breathing). The internal consistencies for parasomnias (α = .58)
and dyssomnias (α = .62) were moderate, reflecting the various sleep problems within each general category.
Both sleep problem variables were skewed (skew >1) and therefore were log transformed (reducing the skew to <1) for analyses. Log-transformed data are reported unless otherwise specified. Of note, all analyses were rerun on raw (untransformed) data, and substantively identical results were obtained.
All analyses were conducted in Mx,20 which is one of the most popular statistical programs designed to deal with genetically sensitive data. The selected sample (details reported previously) was used because we could only see a small proportion of the total TEDS participants and selecting from the extremes increases not only the power but also the likelihood of including children with clinically significant difficulties. To correct for ascertainment,
analyses on all variables, including descriptive statistics, and genetic analyses were conducted jointly with the 7-year screening variable from the TEDS sample. This technique is somewhat similar to using a weight (for further information, see other articles reporting data from the ECHO study6,16).
Twin studies compare within-pair similarity for groups of monozygotic twins, who are genetically identical, and dizygotic twins, who share half their segregating genes.12 This information can be used to estimate additive genetic influences, environmental influences that make family members similar (shared environmental influences), and environmental influences that make family members different (nonshared environmental influences) on single traits and correlations between traits. The bivariate correlated factors model reported herein allows influences on parasomnias to correlate with the same factors on dyssomnias. Consequently, the correlations between parasomnias and dyssomnias are mediated via genetic, shared, and nonshared environmental routes (Figure). As with other analyses, the selection variable was included in these analyses to control for the selected nature of the sample.
Correlated factor solution path diagram. Model shown for 1 twin. A indicates additive genetic influence; C, shared environmental influence; and E, nonshared environmental influence.
The fit statistic provided by Mx for raw data modeling is minus twice the log likelihood (−2LL) of the observations. This provides a relative measure of fit because the differences between −2LLs between models are distributed as χ2. To examine the overall fit of the genetic model, it is necessary to compare its −2LL to that of a saturated model. In a saturated model, the maximum number of parameters is estimated to describe the variance-covariance and means of the tested variables and to provide a perfect fit to the data. The Akaike Information Criterion, which was calculated as χ2 − 2 × df, considers the number of parameters being estimated and provides a further fit index (lower, negative values indicate a better fit).
The raw (untransformed) mean scores for parasomnias and dyssomnias at 8 years of age are given in Table 1. No significant differences were found in levels of the sleep problems between boys and girls. The phenotypic correlation between parasomnias and dyssomnias was moderate (r = 0.42). Table 1 also gives the cross-twin correlations for parasomnias and dyssomnias. For both sets of cross-twin correlations,
the monozygotic correlation is greater than the dizygotic correlation,
suggesting genetic influence. Although the monozygotic correlations are greater than twice the magnitude of the dizygotic twin correlations,
suggesting possible genetic dominance (where there is interaction of different alleles at a locus), the sample size was too small to differentiate additive genetic influence and dominance. Furthermore,
preliminary (unreported) analyses showed that for each phenotype,
models that included additive genetic, shared, and nonshared environmental influences provided a better fit to the data compared with models that included additive genetic, dominance, and nonshared environmental influences. The cross-twin cross-trait correlations (eg, the associations between parasomnias in twin 1 and dyssomnias in twin 2) were r = 0.39 for monozygotic twins and r = 0.10 for dizygotic twins.
Consistent with estimates derived from twin correlations, the results of model-fitting univariate data (reported in Table 2) revealed significant genetic influences on both parasomnias and dyssomnias (eg, genetic influences explained 50% of the variance of parasomnias). Nonshared environmental influences were also significant and accounted for 48% and 26% of the variance of parasomnias and dyssomnias, respectively. The confidence intervals (which included zero) suggest that shared environmental influences did not have a significant role.
Table 2 also presents the results from the bivariate model. The genetic, shared, and nonshared environmental estimates on parasomnias and dyssomnias were similar to those obtained from the univariate models. The bivariate genetic correlation (the overlap between genes that influence parasomnias and dyssomnias) suggests that there is overlap between genes that influence both difficulties (r = 0.61).
Overall, genes are most important in explaining the association between parasomnias and dyssomnias (accounting for 87% of the covariance between the two). Shared environmental influences on the sleep problems are negligible, but these influences appear to be identical for both parasomnias and dyssomnias (r = 1), although overall they still account for a small proportion of the association between parasomnias and dyssomnias (5% of the covariance). A small overlap was found between the nonshared environmental influences for parasomnias and dyssomnias (r = 0.10),
and nonshared environment accounted for little of the association between the 2 difficulties (explaining 8% of the covariance). Each model appeared to fit the data adequately as suggested by the negative Akaike Information Criterion values and nonsignificant P values, suggesting that the models do not fit the data significantly less well than the perfect models.
This study found a moderate overlap between parasomnias and dyssomnias, and individual differences in each category of difficulty were explained largely by genetic and nonshared environmental influences.
The novel finding here is that there was substantial genetic overlap between these 2 types of sleep problems, suggesting that genes that influence parasomnias influence dyssomnias and vice versa. Overall,
the moderate association between parasomnias and dyssomnias was largely explained by genes with smaller environmental influence.
The finding that parasomnias and dyssomnias co-occur agrees with previous literature demonstrating associations between these difficulties.7 Clinical knowledge concerning specific sleep difficulties supports links between these 2 categories of sleep problems. For example, it is known that excessive sleepiness (a possible feature of dyssomnias) may increase the chances of sleepwalking (a parasomnia).
In line with previous research, a genetic influence was seen on both parasomnias and dyssomnias.3,9,10 The strong genetic correlation between the 2 types of difficulties agrees with previous studies that highlight genetic overlap between diverse sleep problems, including different pairs of parasomnias3 and insomnia and sleepiness.21 This genetic overlap leads to speculation as to which genes are associated with both sets of difficulties. As with other complex phenotypes, there is a lack of replicated associations between specific genes and sleep difficulties. Nevertheless, clock genes have been linked to sleep difficulties,
including insomnia,22 and more specifically, Period genes are important in sleep timing.23 Furthermore, a range of sleep problems,
including insomnia and parasomnias, have been linked to symptoms of depression,16,24 and genetic overlap occurs between symptoms of sleep problems and depression.
This suggests that polymorphisms identified with regard to depression,
including genes related to monoaminergic neurotransmission, neurotoxic and neurotrophic processes, and activation of the hypothalamic-pituitary axis, may be candidates for further exploration with regard to sleep problems.25,26
Besides genetic influence, nonshared environmental influence also played a central role in explaining individual differences in parasomnias and dyssomnias. Many environmental factors have now been considered in association with sleep problems in children, and links have been made between sleep difficulties and family disorganization,
maternal depression, school demands, suffering bullying, and a host of life events.4,27- 30 Although influences such as these are often considered environmental,
it is widely accepted that genetic propensities can moderate both exposure and sensitivity to such environmental influences. The correlation between nonshared environmental influences on parasomnias and dyssomnias was small, suggesting that nonshared environmental influences on these difficulties were largely distinct. It is easy to see how environmental influences could lead to one type of difficulty and not another. For example, high levels of exposure to television could be associated with certain sleep difficulties (eg, sleep-wake transition) but not others (eg, sleep-disordered breathing31). Overall, the results of this study suggest that there may be a differential manifestation of shared underlying genetic risks for sleep problems partly based on nonshared environmental influences.
The finding of negligible shared environmental influence on parasomnias and dyssomnias is unsurprising given the rarity of shared environmental influences on traits in adults and to a lesser extent children.12 This supports the idea that environmental influences tend to work in a way that makes children within a family different and not similar.32
The results of this study should be considered alongside the limitations. The first limitation concerns the sample size. The thorough assessment of variables meant that it was infeasible to include a larger number of twin pairs. The modest sample size meant that there was no power to look at the associations between individual sleep problems (as opposed to combining difficulties into parasomnias and dyssomnias). This point is noteworthy because previous research examining the overlap between different pairs of parasomnias has revealed differing levels of genetic overlap (eg, the genetic covariation between sleepwalking and sleep talking was greater than that between sleep talking and nightmares3). Nevertheless, despite the modest sample size, there was power to highlight genetic influences that explain the association between parasomnias and dyssomnias. The second limitation concerns the examination of sleep problems in the full range. There was no power to investigate whether the magnitude and type of genetic and environmental influences on sleep problems are the same at the extremes as in the full range, and this is an issue that may be worthwhile addressing in larger twin samples. Although appropriate given the research question, the third limitation concerns the use of the twin method, for which concerns have been raised.33 The final limitation to be discussed concerns the employment of a parent-rated questionnaire to assess childhood sleep problems. Different phenotypic and genetic estimates have been obtained using different methods to assess sleep problems (eg, parent vs child self-report).6 Nevertheless,
this method of assessing sleep problems is appropriate given the size and nature of the study.
Despite limitations, this work has at least 3 sets of implications.
The first concerns clinical assessment of sleep problems. As highlighted elsewhere,7 we demonstrated that different types of sleep problems co-vary. This emphasizes the likely benefit of obtaining detailed information about both parasomnias and dyssomnias in children who present with a sleep problem. The familiality of the sleep problems reported herein also suggests that it may be worth inquiring about the sleep of family members when assessing patients.
The second set of implications concerns future research. Previous research has highlighted genetic overlap among different types of parasomnia,3 suggesting that it is worth examining whether genes found to influence one type of parasomnia also influence another. The results reported herein go further to suggest that genes that influence certain parasomnias may also influence certain dyssomnias and vice versa.
The final set of implications concern nosology. The finding that parasomnias and dyssomnias are phenotypically and genetically overlapping could have implications for the conceptualization of these problems.
This article focuses on 8-year-olds, and the findings reported herein are not necessarily applicable to other age groups. Indeed,
there is clear evidence that traits can become more or less heritable throughout life (eg, a recent report34 highlighted age as a moderator of the association between a polymorphism in the PER3 clock gene and diurnal preference).
In addition to paying close attention to developmental issues, researchers also need to examine ways in which genetic and environmental factors interact. Groundbreaking research in this area has demonstrated that sensitivity to environmental insults may be moderated by specific genes in relation to a host of difficulties, and a recent report revealed that a PER3 polymorphism is associated with individual differences in decrement in cognitive performance after sleep loss.35 Finally, to have a clear understanding of the mechanisms by which genes and the environmental factors have their effects, it is now essential to identify further specific genetic and environmental influences on sleep patterns and problems and to elucidate the ways in which these influence the brain and behavior.
Correspondence: Alice M. Gregory,
PhD, Department of Psychology, Goldsmiths College, University of London,
Lewisham Way, New Cross, London SE14 6NW, England (firstname.lastname@example.org).
Accepted for Publication: October 8,
Author Contributions: Dr Gregory had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Financial Disclosure: None reported.
Funding/Support: The ECHO study was supported by a Career Development Award from the UK Medical Research Council to Thalia C. Eley, PhD (Institute of Psychiatry).
Additional Contributions: Thalia Eley,
PhD, collected data, obtained funding, oversaw the conduction of the ECHO Study, and commented on an earlier draft of the manuscript; Simon Archer, PhD, Kay Jones, MSc, and Tom Willis, PhD, provided comments on an earlier draft of the manuscript; David Clark, PhD, and Peter McGuffin, PhD, provided scientific input about the ECHO study; Megan Crawford proofread an earlier draft of the manuscript; Georgina Hosang,
Jenn Lau, PhD, Fiona Mcleod, Jasmine Singh, and Lucy Stirling, DClinPsy,
collected data for the ECHO study; and Maria Napolitano, MPhil, was project coordinator for the ECHO study.
Country-Specific Mortality and Growth Failure in Infancy and Yound Children and Association With Material Stature
Use interactive graphics and maps to view and sort country-specific infant and early dhildhood mortality and growth failure data and their association with maternal
Thank you for submitting a comment on this article. It will be reviewed by JAMA Pediatrics editors. You will be notified when your comment has been published. Comments should not exceed 500 words of text and 10 references.
Do not submit personal medical questions or information that could identify a specific patient, questions about a particular case, or general inquiries to an author. Only content that has not been published, posted, or submitted elsewhere should be submitted. By submitting this Comment, you and any coauthors transfer copyright to the journal if your Comment is posted.
* = Required Field
Disclosure of Any Conflicts of Interest*
Indicate all relevant conflicts of interest of each author below, including all relevant financial interests, activities, and relationships within the past 3 years including, but not limited to, employment, affiliation, grants or funding, consultancies, honoraria or payment, speakers’ bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued. If all authors have none, check "No potential conflicts or relevant financial interests" in the box below. Please also indicate any funding received in support of this work. The information will be posted with your response.
Register and get free email Table of Contents alerts, saved searches, PowerPoint downloads, CME quizzes, and more
Subscribe for full-text access to content from 1998 forward and a host of useful features
Activate your current subscription (AMA members and current subscribers)
Purchase Online Access to this article for 24 hours
Some tools below are only available to our subscribers or users with an online account.
Download citation file:
Web of Science® Times Cited: 6
Customize your page view by dragging & repositioning the boxes below.
and access these and other features:
Enter your username and email address. We'll send you a link to reset your password.
Enter your username and email address. We'll send instructions on how to reset your password to the email address we have on record.
Athens and Shibboleth are access management services that provide single sign-on to protected resources. They replace the multiple user names and passwords necessary to access subscription-based content with a single user name and password that can be entered once per session. It operates independently of a user's location or IP address. If your institution uses Athens or Shibboleth authentication, please contact your site administrator to receive your user name and password.