SLEEP, Vol. 34, No. 12, 2011
Heritability of Insomnia—Gehrman et al
were done with adults and relied on individual self-report items
rather than validated measures. As a result, findings from these
studies may reflect the more general concept of “sleep prob-
lems” rather than insomnia per se.
The lack of research in this area may reflect the traditional
notion of insomnia as a symptom of another disorder rather than
an independent clinical entity worthy of attention. For example,
insomnia often occurs in conjunction with psychiatric disor-
ders, most notably depression and generalized anxiety disorder,
both of which have insomnia as a diagnostic criterion.8 As a
criterion for each of these conditions, it is commonly felt that
insomnia is simply a symptom (secondary to primary medical
and/or psychiatric conditions). This said, it is now generally ac-
cepted within the sleep research community that insomnia can
occur as an independent condition and that it is best construed
as comorbid condition, one that has a bi-directional relationship
with disorders such as depression and anxiety.9-11
There are a number of longitudinal studies demonstrating
an increased risk of new-onset psychopathology over time in
individuals with insomnia at baseline.12-15 In cases of remitted
depression, insomnia often is a residual problem that confers
increased risk for relapse.16,17 In cases where insomnia does re-
solve it often occurs as a prodromal symptom prior to relapse.18
Lastly, the neurobiology of sleep/wake regulation overlaps con-
siderably with that for mood and anxiety,19 suggesting common
mechanisms. These intricate interrelationships raise the ques-
tion of the degree to which genetic effects are insomnia-specific
vs. shared with these disorders. Several studies have now dem-
onstrated that depression and generalized anxiety are different
Insomnia is a highly prevalent sleep disorder, affecting
10% of the U.S. population on a chronic basis and 30% epi-
sodically.1 There is a large body of research on pharmacologic
and cognitive behavioral treatments for insomnia, but inves-
tigations of the underlying mechanisms are relatively sparse.
In particular, studies of genetic factors are lacking. While the
heritability of most psychiatric disorders has been studied for
over 20 years, our understanding of the role of genetics in
insomnia remains limited.
To date, several twin studies in adults have examined wheth-
er insomnia symptoms, broadly defined, are heritable traits. In
one study, Heath and colleagues at the Australian Twin Registry
collected sleep data from almost 4000 pairs of twins.2 The pro-
portion of variance related to heritable factors (h2) was fairly
consistent across questions ranging from sleep quality (h2 =
32%) to “anxious insomnia” (h2 = 36%). Subsequent studies
generally yielded comparable estimates in the range of 20%
to 40%,3-6 with the exception of one study that found a herita-
bility of 64% for a single insomnia item.7 All of these studies
HERITABILITY OF INSOMNIA SYMPTOMS IN YOUTH
Heritability of Insomnia Symptoms in Youth and Their Relationship to
Depression and Anxiety
Philip R. Gehrman, PhD1,3; Lisa J. Meltzer, PhD2; Melisa Moore, PhD3; Allan I. Pack, MBChB, PhD3,4; Michael L. Perlis, PhD1; Lindon J. Eaves, PhD5;
Judy L. Silberg, PhD5
1Department of Psychiatry, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA; 2Department of Pediatrics, Perelman School of
Medicine, University of Pennsylvania and The Children’s Hospital of Philadelphia, Philadelphia, PA; 3Center for Sleep and Respiratory Neurobiology,
Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA; 4Division of Sleep Medicine, Perelman School of Medicine, University of
Pennsylvania, Philadelphia, PA; 5Virginia Commonwealth University School of Medicine, Richmond, VA
Submitted for publication July, 2010
Submitted in final revised form May, 2011
Accepted for publication June, 2011
Address correspondence to: Philip Gehrman, PhD, CBSM, Department of
Psychiatry & Penn Sleep Center, University of Pennsylvania, 3535 Market
Street, Suite 670, Philadelphia, PA 19104; Tel: (215) 746-3578; Fax: (215)
573-0759; E-mail: firstname.lastname@example.org
Study Objectives: Insomnia is a highly prevalent sleep disorder yet little is known about the role of genetic factors in its pathophysiology. The aim
of this study was to examine the relative contributions of genetic and environmental factors in explaining variability in insomnia symptoms.
Design: Traditional twin design.
Setting: Academic medical center.
Participants: 1412 twin pairs aged 8-16 years (48.8% MZ, 47.2% DZ, 4.0% indeterminate).
Measurements and Results: Ratings of insomnia symptoms, depression, and overanxious disorder were made by trained interviewers based
on DSM-III-R criteria. ACE models were conducted using Mx statistical software. Insomnia symptoms were prevalent in this sample based both
on parental (6.6%) and youth (19.5%) reports. The overall heritability of insomnia symptoms was modest (30.7%), with the remaining variance at-
tributed to unique environmental effects. There was no evidence of sex differences in the prevalence of insomnia symptoms or in the contribution
of genetic and environmental effects. In multivariate models, there was support for insomnia-specific unique environmental effects over and above
overlapping effects with depression and overanxious disorder, but no evidence for insomnia-specific genetic effects.
Conclusions: Genetic factors play a modest role in the etiology of insomnia symptoms in 8-16 year-olds. These effects overlap with the genetics
of depression and overanxious disorder. Further work is needed to determine which genes confer risk for all three disorders.
Keywords: Insomnia, genetics, twin study, depression, anxiety
Citation: Gehrman PR; Meltzer LJ; Moore M; Pack AI; Perlis ML; Eaves LJ; Silberg JL. Heritability of insomnia symptoms in youth and their relation-
ship to depression and anxiety. SLEEP 2011;34(12):1641-1646.
SLEEP, Vol. 34, No. 12, 2011
Heritability of Insomnia—Gehrman et al
youth interviews and the interview from mothers (or fathers if
mothers were not available) were used. All interviews were au-
diotaped throughout the study and rated by a team of monitors
for fidelity. For all questions, a rating of 0 was used if it was
determined that a symptom or disorder was not present. A rat-
ing of 2 indicated that the symptom or disorder was present at
least at the minimum level of severity, and a rating of 3 that the
symptom or disorder was present at a higher level of severity.
A rating of 1 was discouraged because it indicated that the rater
was not able to determine whether criteria were met, in which
the rater was supposed to continue to query the respondent until
a determination could be made.
The sleep module of the CAPA interview includes a series
of questions about the child’s/adolescent’s current sleep pat-
terns, including whether the child has difficulty falling asleep
or waking up too early in the morning, and then makes a clini-
cal judgment of whether or not clinically significant insomnia
symptoms are present. For these analyses, ratings of 2 and 3
were combined to create a dichotomous (yes/no) insomnia rat-
ing. For descriptive purposes, additional ratings of the timing of
insomnia during the night (initial, middle, or late), and medica-
tion for insomnia were examined.
The CAPA also contains modules for DSM-III-R depression
and overanxious disorder that were used for this study. For each
disorder clinical ratings are made for the presence/absence of
each of the criteria, allowing a determination to be made for
whether the criteria are met. No sleep related criteria were used
for either disorder. Youth and parent reports were combined us-
ing the standard “or” rule, in which a symptom is rated as pres-
ent if it is endorsed by either the child or parent.24 For insomnia,
univariate models were also computed for youth and parent re-
ports separately. For overanxious disorder a dichotomous yes/
no variable was computed based on whether criteria were met.
Depression was rare in this cohort (1.3%), so subjects were di-
chotomized based on the presence or absence of ≥ 3 symptoms,
with 3.4% meeting this threshold.
Zygosity was inferred using an algorithm that incorporates
data from parental responses to a questionnaire and ratings of
photographs, additional details of which have been published
previously.22 Parental questions asked (1) how often the twins
were mistaken for each other by strangers, (2) how similar the
twins were, and (3) the perceived likelihood that the twins were
identical. Photographs of the twins taken separately and to-
gether were scored by 2 independent raters as “definitely MZ,”
“probably MZ,” “indeterminate,” “probably DZ,” or “definitely
DZ.” In cases of disagreement the rating was assigned as inde-
terminate. The algorithm was validated in a subset of 231 twin
pairs for whom zygosity was confirmed by typing for blood
groups or DNA polymorphisms.22
Descriptive statistics for insomnia symptoms and correla-
tions (tetrachoric for bivariate variables) with depression and
overanxious disorder were computed. The correlations between
insomnia symptoms and age, gender, and pubertal status, as
determined by Tanner staging, were computed to explore the
influence of development on insomnia reporting.
manifestations of the same genetic factors,20 and it is possible
that insomnia is part of the same cluster.
While studies on the heritability of insomnia in children and
adolescents are lacking, the literature provides evidence for ge-
netic influences on more broadly defined sleep problems. For
example, a recent longitudinal study by Gregory6 found that
most of the variance in sleep problems could be explained by
genetic and nonshared environmental influences at both 8 years
(63% additive genetic factors, 32% nonshared environment)
and at 10 years (66% genetic factors, 27% nonshared environ-
ment, 7% shared environment). There appears to be a genetic
component to broadly defined sleep problems, but further stud-
ies are needed in different age groups to better understand how
genetic and environmental influences interact to confer risk for
insomnia over the course of development.
Thus, the primary goal of this study was to determine the
broad heritability of insomnia symptoms based on clinical rat-
ings in a large sample of twins aged 8-16. A second goal was
to determine the extent to which any genetic effects are shared
with depression and anxiety. It was hypothesized that there
would be unique genetic effects specific to insomnia symptoms
over and above those shared with depression and anxiety.
The data for the present analysis were drawn from the Vir-
ginia Twin Study of Adolescent Behavioral Development (VT-
SABD), a longitudinal sequential cohort study of twins aged 8
through 16 years.21,22 The current analyses focus on information
from the first wave of data collection that consisted of 1412
families. Twin pairs were identified through the state school
system and participating private schools in the state of Virginia
and then contacted by mail. Interested families were scheduled
for detailed assessments of behavioral development and psy-
chopathology. To maximize available sources of information,
all interviews and questionnaires were completed by each child/
adolescent as well as one or both parents or primary caregivers.
Caregivers had to have been living with the youth for at least 6
months to be included.
Child and Adolescent Psychiatric Assessment
The primary assessment utilized the Child and Adolescent
Psychiatric Assessment (CAPA), a semi-structured interview
with separate parent and child/adolescent formats designed
to assess a wide range of behavioral and psychological symp-
toms based on DSM-III-R.23 For the purposes of the current
analysis, the modules for sleep problems, depression, and
anxious affect and worries were utilized. The DSM-III-R
criteria for insomnia are very similar to those in the DSM-
IV-TR, and consist of (1) difficulty initiating or maintaining
sleep, or non-restorative sleep; and (2) occurs ≥ 3 times per
week for at least one month, and causes clinically significant
distress or impairment. In the DSM-IV-TR the frequency cri-
terion was dropped. In each area, the presence of symptoms
over the past 3 months was ascertained, along with the fre-
quency of occurrence, duration, and earliest age of onset (if
symptoms were present).
A pair of trained interviewers was assigned to each family
with each interviewer assessing one twin and one parent in the
family and with the twin interviews always occurring first. The
SLEEP, Vol. 34, No. 12, 2011
Heritability of Insomnia—Gehrman et al
to the anxiety and insomnia. The third model comparison de-
termines whether there is an “insomnia-specific” factor. This
type of model favors allocation of variance to depression and
anxiety as the first 2 variables in the model; this was intention-
ally done in order to determine the genetic variance that an in-
somnia-specific factor contributes over and above these shared
influences. Figure 1A provides a diagram of the model. In Fig-
ure 1A, there are 3 factors representing depression, overanx-
ious disorder, and insomnia. There are 3 additive genetic factors
(A1-A3) and 3 unique environmental factors (E1-E3) specified.
Note that common environmental factors were omitted from
the figure for simplicity. The A1 and E1 factors influence all 3
phenotypes, whereas the A3 and E3 factors only influence in-
somnia and represent the unique effects for this phenotype. This
approach extends previous studies in which a common factor
accounted for both depression and anxiety,27 with insomnia the
common factor in the current study. This analysis was also re-
peated with depression treated as the third factor, as depicted in
Figure 1B. Univariate models for depression and overanxious
disorder have been reported elsewhere.22,28
A total of 1412 twin pairs participated, and the sample was
split between males (46.3%) and females (53.7%). The mean
age was 12.0 years (SD 2.6) with a range of 8-16. There were
689 monozygotic (MZ) twin pairs (48.8%), 666 dizygotic (DZ)
twin pairs (47.2%), and 57 of indeterminate zygosity (4.0%).
This last group was excluded from further analyses. Age and
gender were not significantly correlated with any insomnia
variables (all P > 0.05).
The prevalence of insomnia symptoms is shown in Table 1.
The prevalence of insomnia symptoms in males and females
was very similar at 18.3% and 20.5%, respectively. Parents
consistently reported lower rates of insomnia symptoms (6.6%)
A series of twin models were conducted with the Mx statisti-
cal program for structural equation modeling.25 A full model
was fit that allows for additive genetic (A), common environ-
mental (C), and unique environmental (E) effects. This ACE
model was then compared with models in which one or more of
these effects were removed (AE, CE, and E), yielding a total of
4 models per analysis. The goal was to determine the model that
provided the best combination of fit to the data and parsimony.
Models were compared using likelihood ratio tests in which the
negative log likelihood (-2LL) for a restricted model was sub-
tracted from that for a more saturated model. The result follows
a χ2 distribution with df equal to the difference in the number
of parameters between the models. When the χ2 test yielded a
P-value > 0.05 the restricted model was deemed to provide a fit
that was not significantly worse than the more saturated model.
Models that provided the best combination of parsimony and fit
were characterized by the lowest values on the Akaike Informa-
tion Criterion (AIC).
The first ACE model was applied to data on the presence
or absence of insomnia symptoms. The second model allowed
separate parameters for boys and girls to determine if there were
gender differences in the relative contributions of genes and en-
vironment (i.e., sex-linked models). Separate models were then
fit based on pubertal status. Lastly, a multivariate ACE model
was fit to the data allowing for genetic and environmental ef-
fects for depression, anxiety, and insomnia symptoms based on
a Cholesky decomposition.26 This method allows the model to
contain multiple ACE factors up to the number of independent
variables, in this case 3. In these analyses the “or” rule was
used for insomnia symptoms. The first factor accounts for ge-
netic and environmental contributions that are common to all 3
variables (depression, anxiety, and insomnia). The second fac-
tor and resulting model comparison determines whether there is
second set of genetic and environmental contributions specific
Figure 1—Path diagram representing the Cholesky decomposition of the twin data. The model consists of three additive genetic (A1-3) and three unique
environmental (E1-3) factors, with each successive factor loading on fewer variables. In Figure 1A, A3 and E3 represent insomnia-specific factors. In Figure
1B, the phenotypes are presented in reverse order and A3 and E3 are depression-specific effects. Common environmental factors (C) have been omitted
because they were not statistically significant in any models. Coefficients are shown for all statistically significant paths. Note that, regardless of order of entry
for the phenotypes, the best fitting model included only a single genetic factor.
0.01 0.92 0.02
SLEEP, Vol. 34, No. 12, 2011
Heritability of Insomnia—Gehrman et al
the correlation between MZ twin pairs for insomnia was 0.18,
whereas for DZ twin pairs the correlation was 0.05, indicating
a greater degree of concordance for MZ compared to DZ twins.
On the other hand, the correlations between ratings of insomnia
in MZ and DZ twin pairs were both 0.10, indicating a similar
pattern of association regardless of zygosity.
For ratings of insomnia symptoms based on combined youth
and parental reports, an AE model provided the best fit to the
data (χ2 (4) = 5.7, P = 0.2, AIC = −2.31) with genetic effects
accounting for 37.4% of the variance and unique environmen-
tal effects accounting for 62.6% of the variance. For ratings of
insomnia symptoms based on youth reports, an AE model pro-
vided the best fit to the data (χ2 (4) = 7.4, P = 0.11, AIC = −0.56)
with genetic effects accounting for 30.7% of the variance and
unique environmental effects accounting for 69.3% of the vari-
ance. For ratings of insomnia symptoms based on parental re-
ports, again an AE model provided the best fit to the data (χ2 (4)
= 8.4, P = 0.08, AIC = 0.4) with genetic effects accounting for
51.4% of the variance, and unique environmental effects ac-
counting for 48.6% of the variance. There was no evidence for
significant common environmental effects in any of these mod-
els. In sex-linked analyses, models were examined for all ACE
effect combinations both with the effects for boys and girls con-
strained to be equal and with the effects allowed to differ. The
model with the lowest AIC, both when using youth and parent
ratings, was an AE model in which effects were constrained to
be equal across gender. For youth ratings, the AE model yielded
a χ2(10) = 11.6 (P = 0.31) and an AIC = −8.4. For parental rat-
ings, the AE model yielded a χ2(10) = 9.6 (P = 0.31) and an
AIC = −10.4. In analyses in which the sample was separated
by pubertal status, an AE model provided the best fit for both
pre-pubertal (χ2(4) = 7.5, P = 0.11, AIC = −0.5) and post-pu-
bertal (χ2(4) = 6.4, P = 0.17, AIC = −1.6) subjects compared to
the saturated ACE model. Additive genetic factors for pre- and
post-pubertal subjects were 27.8% and 31.5%, and unique en-
vironmental effects were 72.2% and 68.5%, respectively. Thus
the relative contribution of genetic and environmental factors
did not appreciably change with the onset of puberty.
Lastly, for the multivariate models an AE model that includ-
ed all 3 Cholesky factors provided a better fit to the data than
the saturated ACE model. The AE with 3 factors model was
then compared to one in which the third insomnia-specific ge-
netic factor was removed, and there was not a significant loss
of model fit (Δχ2(1) = 0.004, P = 0.95, ΔAIC = -2.0). The model
fit also did not significantly change when the second factor was
removed (Δχ2(3) = 0.001, P = 0.99, ΔAIC = -6.0). The final
model contained a single ge-
netic factor and 3 trait specific
unique environmental factors.
Standardized path coefficients
for each of the significant
pathways can be found in Fig-
ure 1A. When this analysis
was repeated using depression
as the last phenotype, the final
model again contained a single
genetic factor and three envi-
ronmental factors as depicted
in Figure 1B.
than their children (19.5%; Table 2). Parents and youth were in
agreement about the presence or absence of insomnia symptoms
79.8% of the time. Discrepancies consisted almost entirely of
youth reporting insomnia symptoms but parents not reporting
any symptoms. The parent-youth correlations between symp-
toms counts for depression and overanxious disorder previous-
ly reported21 were low (r ranging from 0.12 to 0.29), suggesting
that a discrepancy exists across disorders. The mean (SD) age
of onset was 9.9 (3.4) years from youth reports and 9.0 (4.4)
from parental reports. Depression was relatively rare in this co-
hort, with criteria met by only 1.3% of subjects and 3.4% of
subjects having ≥ 3 symptoms. Overanxious disorder was more
common, with 18.0% of subjects meeting criteria. The preva-
lence of insomnia symptoms in those with depression and/or
overanxious disorder was 32.4% compared to 16.5% in those
who did not meet criteria for either diagnosis. Table 3 displays
the cross-twin cross-trait tetrachoric correlations among dis-
orders. These correlations show the associations among traits
within a twin pair for both MZ and DZ twins. For example,
Table 1—Frequency of insomnia symptoms
Rating of clinically
% Positive From
% Positive From
Middle insomnia 19.6%6.3%
Table 2—Rates of youth and parent agreement for insomnia symptoms
Absent n = 2086
n = 98
n = 450
n = 74
Table 3—Average cross twin- cross trait tetrachoric correlations between insomnia, depression, and overanxious
disorder in MZ and DZ twins
MZ TwinsDZ Twins
*P < 0.05
SLEEP, Vol. 34, No. 12, 2011
Heritability of Insomnia—Gehrman et al
between the genetics of insomnia with that for depression and
anxiety. The same pattern was found when the order of the
phenotypes was reversed, indicating that the rarity of depres-
sion did not skew the model. These results are in contrast to
previous studies using different analytic approaches. Gregory
and colleagues approached this overlap from a different per-
spective by fitting ACE models separately to sleep problems
and depression in children and then computing correlations
between these factors.6 The correlation between the additive
genetic factors was r = 0.64 indicating substantial, but not
complete, overlap. Using this approach in the current data, the
correlation was 0.96, suggesting a stronger degree of overlap
than in their study. In adults, Kendler and colleagues conduct-
ed a factor analysis of the items from a scale of depression and
anxiety in a large sample of twins from the Australian Twin
Registry.42 The analysis led to a three-factor solution, with the
third factor loading only on two sleep-related items that ac-
counted for ~6% of the overall variance. When examined to-
gether, these studies provide mixed evidence for the existence
of unique insomnia-specific genetic factors that are distinct
from other symptoms of depression and anxiety.
This study had several limitations. The presence or absence
of insomnia, although an aggregation of multiple questions, ul-
timately was rated as a single item. A number of factors, such
as the schedule imposed by school start times, could have influ-
enced reporting of insomnia symptoms but are not specifically
addressed by the CAPA. The use of a validated instrument that
yields a score indicative of insomnia severity may have better
captured the wider range of insomnia manifestations. Further,
the wording of the items on the CAPA may have influenced
results. For example, the list of follow up questions in the sleep
module contains a reference to fear of the dark, which may have
fit better in the overanxious disorder module. Similarly, the
anxiety module contains the question of whether worry keeps
them up at night. To the extent that these items influenced the
orthogonality of insomnia, depression, and anxiety, there may
have been some impact on the genetic models. Objective as-
sessment of sleep, either via polysomnography or actigraphy,
may also have yielded different results because of their insen-
sitivity to the biases inherent in self-report and interview meth-
ods. The assessment of sleep also did not include a measure of
circadian tendency which, may have played an etiologic role,
or a measure of other sleep disorders. Lastly, the small num-
ber of youth with depression or overanxious disorder preclude
the ability to make subtle distinctions among the factors, but
the results establish a prima facie case for an overall common
genetic basis to the three traits. This genetic finding does not
preclude clinically significant effects specific to insomnia, but
raises the possibility that these might be environmental rather
than genetic in origin.
In summary, these results indicate that clinician ratings of
insomnia symptoms have moderate heritability but are not in-
fluenced by shared environmental factors. In this 8- to 16-year-
old sample, this pattern of results was consistent across gender.
Genetic factors related to the etiology of insomnia overlap with
those related to depression and anxiety but there are distinct
insomnia-specific environmental effects. Treatments for de-
pression and anxiety may also improve sleep, and vice versa,
although the distinct environmental effects suggest that insom-
This study examined the genetic and environmental contri-
butions to insomnia symptoms, both in isolation and in relation
to depression and anxiety, in youth ages 8-16 years. Insomnia
symptoms were prevalent, particularly when relying on youth
rather than parental reports. Almost one in five subjects report-
ed insomnia symptoms, which is higher than prior studies that
diagnostic criteria for an insomnia disorder,29-32 but comparable
to the rate of insomnia symptoms in another study.33 Given that
poor sleep has been associated with lower school performance
and other negative outcomes for youth,34-37 insomnia could be
associated with a substantial public health burden for this seg-
ment of the population.
Rates of insomnia symptoms were considerably lower when
based on parental reports. This is consistent with other stud-
ies that have found higher rates of internalizing symptoms by
youth report compared to parent report.38,39 Gregory and col-
leagues also found this pattern in a study of 300 8-year-old
twins, in which rates of sleep problems were higher on child
reports (45%) than parental reports (17%) of sleep latency,
as assessed with the Children’s Sleep Habits Questionnaire,
although other sleep variables were less discrepant.40 Unless
youth alert parents about their difficulties sleeping, parents may
be less likely to be aware of insomnia since the disturbance oc-
curs when parents are likely asleep. This pattern of higher heri-
tability for parental estimates was also reported in the Gregory
and colleagues study.40
The overall heritability of insomnia symptoms based on
combined youth and parental reports is moderate, with genetic
factors accounting for about 37% of the variance. By means of
comparison, using youth reports from these data, the heritabil-
ity of overanxious disorder was 30% for males and 46% for
females, and the heritability for depression was 11% for males
and 19% for females (note the presence of sex-linked effects
for these disorders).22 The rate for insomnia symptoms is highly
consistent with previous twin studies of insomnia symptoms in
adults that relied on self-report methods.2-5,7 Thus, self-report
vs. clinician ratings may not be a significant factor for deter-
mining the heritability of insomnia, which is important because
of the number of large-scale genetic studies that have only in-
cluded measures with sleep related items (e.g., items found in
most depression rating scales such as 41). The phenotypic data
from these studies may be sufficient for initial analyses of the
genetics of insomnia.
In adults, the prevalence of insomnia in women is higher
than in men but the reasons for the gender differences are not
known. In this younger sample there were no phenotypic differ-
ences between males and females. The genetic analyses further
suggest that there are no gender differences in the relative con-
tributions of genes and environment in this age group. Further
research is needed to determine the age at which gender differ-
ences in insomnia prevalence begin to emerge as well as the
factors underlying this transition.
The most significant finding from these analyses was that
there was not evidence for insomnia-specific additive genetic
factors, but insomnia-specific unique environmental factors
were supported. This is visually depicted in Figure 1A, for
which the only significant pathways are those with coefficients
provided. This suggests that there may be substantial overlap
SLEEP, Vol. 34, No. 12, 2011 Download full-text
Heritability of Insomnia—Gehrman et al
18. Perlis ML, Giles DE, Buysse DJ, Tu X, Kupfer DJ. Self-reported sleep
disturbance as a prodromal symptom in recurrent depression. J Affect
19. Saper CB, Cano G, Scammell TE. Homeostatic, circadian, and emotional
regulation of sleep. J Comp Neurol 2005;493:92-8.
20. Middeldorp CM, Cath DC, Van Dyck R, Boomsma DI. The co-morbidity
of anxiety and depression in the perspective of genetic epidemiology. A
review of twin and family studies. Psychol Med 2005;35:611-24.
21. Hewitt JK, Silberg JL, Rutter M, et al. Genetics and developmental
psychopathology: 1. Phenotypic assessment in the Virginia Twin Study
of Adolescent Behavioral Development. J Child Psychol Psychiatry
22. Eaves LJ, Silberg JL, Meyer JM, et al. Genetics and developmental psy-
chopathology: 2. The main effects of genes and environment on behav-
ioral problems in the Virginia Twin Study of Adolescent Behavioral De-
velopment. J Child Psychol Psychiatry 1997;38:965-80.
23. Angold A, Prendergast M, Cox A, Harrington R, Simonoff E, Rutter M.
The Child and Adolescent Psychiatric Assessment (CAPA). Psychol Med
24. Costello EJ, Angold A, Burns BJ, et al. The Great Smoky Mountains
Study of Youth. Goals, design, methods, and the prevalence of DSM-III-
R disorders. Arch Gen Psychiatry 1996;53:1129-36.
25. Neale MC. Mx: Statistical Modeling (2nd Ed). Richmond, VA: Virginia
Commonwealth University Department of Psychiatry, 1994.
26. Neale MC, Cardon LR. Methodology for genetic studies of twins and
families. Dordrecht: Kluwer Academic Publishers, 1992.
27. Kendler KS, Gardner CO, Gatz M, Pedersen NL. The sources of co-mor-
bidity between major depression and generalized anxiety disorder in a
Swedish national twin sample. Psychol Med 2007;37:453-62.
28. Silberg JL, Rutter M, Eaves L. Genetic and environmental influences on
the temporal association between earlier anxiety and later depression in
girls. Biol Psychiatry 2001;49:1040-9.
29. Canals J, Domenech E, Carbajo G, Blade J. Prevalence of DSM-III-R
and ICD-10 psychiatric disorders in a Spanish population of 18-year-olds.
Acta Psychiatr Scand 1997;96:287-94.
30. Johnson EO, Roth T, Schultz L, Breslau N. Epidemiology of DSM-IV
insomnia in adolescence: lifetime prevalence, chronicity, and an emergent
gender difference. Pediatrics 2006;117:e247-56.
31. Roberts RE, Roberts CR, Chan W. Ethnic differences in symptoms of
insomnia among adolescents. Sleep 2006;29:359-65.
32. Chung KF, Cheung MM. Sleep-wake patterns and sleep disturbance
among Hong Kong Chinese adolescents. Sleep 2008;31:185-94.
33. Ohayon MM, Roberts RE, Zulley J, Smirne S, Priest RG. Prevalence and
patterns of problematic sleep among older adolescents. J Am Acad Child
Adolesc Psychiatry 2000;39:1549-56.
34. Wolfson AR, Spaulding NL, Dandrow C, Baroni EM. Middle school start
times: the importance of a good night’s sleep for young adolescents. Be-
hav Sleep Med 2007;5:194-209.
35. Wolfson AR, Carskadon MA. Sleep schedules and daytime functioning in
adolescents. Child Dev 1998;69:875-87.
36. Morrison DN, McGee R, Stanton WR. Sleep problems in adolescence. J
Am Acad Child Adolesc Psychiatry 1992;31:94-9.
37. Sadeh A, Gruber R, Raviv A. Sleep, neurobehavioral functioning, and be-
havior problems in school-age children. Child Dev 2002;73:405-17.
38. Angold A, Weissman MM, John K, et al. Parent and child reports of de-
pressive symptoms in children at low and high risk of depression. J Child
Psychol Psychiatry 1987;28:901-15.
39. Waters E, Stewart-Brown S, Fitzpatrick R. Agreement between adoles-
cent self-report and parent reports of health and well-being: results of an
epidemiological study. Child Care Health Dev 2003;29:501-9.
40. Gregory AM, Rijsdijk FV, Eley TC. A twin-study of sleep difficulties in
school-aged children. Child Dev 2006;77:1668-79.
41. Boomsma DI, Willemsen G, Sullivan PF, et al. Genome-wide associa-
tion of major depression: description of samples for the GAIN Major De-
pressive Disorder Study: NTR and NESDA biobank projects. Eur J Hum
42. Kendler KS, Heath AC, Martin NG, Eaves LJ. Symptoms of anxiety and
symptoms of depression. Same genes, different environments? Arch Gen
nia-specific treatment components will also be necessary. Treat-
ment of insomnia in youth may also reduce risk for later onset
of depression and anxiety disorders. Further research into the
genetic and environmental bases for insomnia needs to more
comprehensively assess sleep and circadian rhythms in order to
better understand the nature of the phenotype. It is likely that
insomnia is a heterogeneous phenotype and a thorough assess-
ment will help to identify subgroups that may be more appro-
priate for genetic studies. This line of research may shed light
not only on the mechanisms of disturbed sleep but also on the
pathophysiology of mood and anxiety disorders.
Supported by: 1R21MH079187.
This was not an industry supported study. Dr. Pack is the
John Miclot Professor of Medicine. Funds for this endowment
are provided by Phillips Respironics. The other authors have
indicated no financial conflicts of interest.
1. Ohayon MM. Epidemiology of insomnia: what we know and what we still
need to learn. Sleep Med Rev 2002;6:97-111.
2. Heath A, Kendler K, Eaves L, Martin N. Evidence for genetic influences
on sleep disturbance and sleep pattern in twins. Sleep 1990;13:318-35.
3. Heath A, Eaves L, Kirk K, Martin N. Effects of lifestyle, personality,
symptoms of anxiety and depression, and genetic predisposition on sub-
jective sleep disturbance and sleep pattern. Twin Res 1998;1:176-88.
4. McCarren M, Goldberg J, Ramakrishnan V, Fabsitz R. Insomnia in Viet-
nam era veteran twins: influence of genes and combat experience. Sleep
5. Boomsma DI, van Someren EJ, Beem AL, de Geus EJ, Willemsen G.
Sleep during a regular week night: a twin-sibling study. Twin Res Hum
6. Gregory AM, Rijsdijk FV, Lau JY, Dahl RE, Eley TC. The direction of
longitudinal associations between sleep problems and depression symp-
toms: a study of twins aged 8 and 10 years. Sleep 2009;32:189-99.
7. Watson NF, Goldberg J, Arguelles L, Buchwald D. Genetic and environ-
mental influences on insomnia, daytime sleepiness, and obesity in twins.
8. American Psychiatric Association. Diagnostic and Statistical Manual of
Mental Disorders (4th ed.). Washington, DC: American Psychiatric As-
9. Stepanski EJ, Rybarczyk B. Emerging research on the treatment and etiol-
ogy of secondary or comorbid insomnia. Sleep Med Rev 2006;10:7-18.
10. Lichstein KL. Secondary insomnia: a myth dismissed. Sleep Med Rev
11. McCrae CS, Lichstein KL. Secondary insomnia: diagnostic challenges
and intervention opportunities. Sleep Med Rev 2001;5:47-61.
12. Riemann D, Voderholzer U. Primary insomnia: a risk factor to develop
depression? J Affect Disord 2003;76:255-9.
13. Pigeon WR, Perlis ML. Insomnia and depression: birds of a feather? Int J
Sleep Disord 2007;1:82-91.
14. Turek FW. Insomnia and depression: if it looks and walks like a duck.
15. Morawetz D. Insomnia and depression: which comes first? Sleep Res On-
16. Karp JF, Buysse DJ, Houck PR, Cherry C, Kupfer DJ, Frank E. Rela-
tionship of variability in residual symptoms with recurrence of major
depressive disorder during maintenance treatment. Am J Psychiatry
17. Paykel ES, Ramana R, Cooper Z, Hayhurst H, Kerr J, Barocka A. Residu-
al symptoms after partial remission: an important outcome in depression.
Psychol Med 1995;25:1171-80.