Sleep disturbance, circadian preference and symptoms of adult attention deficit hyperactivity disorder (ADHD)

Sleep disturbance, circadian preference and symptoms of adult attention deficit hyperactivity disorder (ADHD)
Bogdan Ioan Voinescu • Aurora Szentagotai •
Daniel David
Received: 1 February 2012 / Accepted: 30 June 2012 / Published online: 20 July 2012
Springer-Verlag 2012
hyperactivity disorder (ADHD) and one of the main
impairments resulting from insomnia disorders. These dis-
orders have also been reported to be linked with disturbances
in circadian rhythms and with increased eveningness. To
explore these associations, more than 550 adults were sur-
veyed across Romania. Using self-reported questionnaires,
the presence and severity of ADHD and insomnia symptoms
were determined, together with sleep and circadian typology
parameters. ADHD symptoms were more frequent and
severe among younger individuals. Subjects with probable
ADHD complained more frequently of sleep disturbance of
the insomnia type (more than 50 %) and reported shorter
sleep durations and longer sleep latencies and more frequent
unwanted awakenings. Individuals likely to suffer from
ADHD and/or insomnia disorder were significantly more
evening oriented than controls. Inattention was associated
with both insomnia and eveningness, while impulsivity was
associated with poor sleep. Hyperactivity and sleep timing
were associated with poor sleep only in probable insomnia
group. These findings highlight the reciprocal links between
ADHD symptoms, sleep and diurnal preference.
Keywords Insomnia Self-report Eveningness
5.29 % being affected worldwide (reviewed by Polanczyk
et al. 2007). Despite initial beliefs that ADHD is rare in
adulthood (Hill and Schoener 1996), more recent meta-
analyses revealed that at the age of 25 it persists in between
15 and 60 % of those with the childhood condition,
depending on the used criteria (Faraone et al. 2006), giving
an estimated prevalence of around 2.5 % (Simon et al.
2009). Insomnia is a common diagnostic criterion for many
psychiatric disorders and may cause several functional
impairments such as inattention, one of the core symptoms
of ADHD. Despite the possibility of an association
between sleep disturbance and ADHD symptoms, there is
little research on this relationship. In non-clinical samples,
inattention was associated with insomnia (Kass et al. 2003)
or difficulties falling asleep, longer times in bed and less
efficient sleep, while hyperactivity was associated with
decreased sleep duration (Gau et al. 2007). Decreased sleep
duration, increased sleep latency and more disturbed sleep
were also linked with hyperactivity in non-medicated
adults suffering from ADHD (Mahajan et al. 2010), while
longer sleep latencies, low sleep quality and daytime
sleepiness were reported by adult ADHD patients (Bijlenga
et al. 2011; Schredl et al. 2007; Surman et al. 2009).
Shorter and later sleep were found to be linked with
hyperactivity (Bijlenga et al. 2011). Polysomnography and
actigraphy studies in adults diagnosed with ADHD also
revealed reduced sleep efficiency, as well as alterations in
other sleep parameters (Boonstra et al. 2007; Kooij et al.
2001; Sobanski et al. 2008; Van Veen et al. 2010).
Eveningness, a psychometric parameter describing the
preference of an individual for later bed and wake times, as
well as patterns of alertness and activity associated with
later times during the day, has also been found to be linked
with inattention (Bae et al. 2010; Caci et al. 2009; Rybak
et al. 2007) and hyperactivity (Bae et al. 2010). Recent
B. I. Voinescu (&) A. Szentagotai D. David
Department of Clinical Psychology and Psychotherapy,
Babes-Bolyai University, Cluj-Napoca, Romania
DOI 10.1007/s00702-012-0862-3
preference are associated with changes in circadian
rhythms at both hormonal (Baird et al. 2011; Van Veen
et al. 2010) and molecular levels (Baird et al. 2011).
There are two main groups of adults with ADHD: those
with a known diagnosis in childhood and who still have
symptoms, and those who do not have a pre-existing diag-
nosis. As ADHD in adults has only been relatively recently
recognized in many European countries, many individuals
may have never been identified with the disorder as a child.
Moreover, as the comorbidity with other psychiatric disor-
der s is high, the risk of being misdiagnosed or for
remaining undiagnosed and consequently mis- or untreated
is considerable. To better understand the associations
between sleep disturbance, diurnal preference and ADHD
in adults, we conducted an exploratory research on young
adults. This population is at a higher risk to display ADHD
symptomatology, is generally physically healthy and may
represent an ideal population to examine relationships
between self-reported adult ADHD symptomatology, sleep
timings, circadian type, sleep quality and insomnia severity
(Sample 1). Then, we broadened our research by including
older adults and recruited participants from across Romania
(Sample 2). In our opinion, this gave us a glimpse on the
current situation of adult ADHD and chronic insomnia
symptoms in the general community, as well as the
opportunity to better understand the links between circadian
typology, sleep quality and timings on ADHD in adults.
Methods
only (Sample 1) and one in general population (Sample 2).
Along with providing demographic data, they all com-
pleted a battery of tests translated into Romanian that was
filled in via Surveygizmo platform. The studies had been
approved by the ‘‘Babes-Bolyai’’ University Ethics
Committee.
Participants
was 21.82 (standard deviation (SD) 2.52). No other
demographic data were collected.
were women (N = 176, 70.4 %). Mean age was 38.61 (SD
12.40). More than a half completed undergraduate educa-
tion (N = 134; 53.6 %) and 33.2 % postgraduate one
(N = 83). The o majority were employed (N = 218,
87.2 %) and 6.0 % (N = 15) were retired. Seven people
(2.8 %) declared they were unemployed. More than a half
(N = 144, 57.6 %) reported incomes of less than 1,500
RON (about €350) and being married (N = 140, 56.0 %).
The respondents were from all over Romania, but around a
third were from the cities of Baia Mare and Cluj-Napoca
only (33.2 %, N = 83).
The Adult Self-Report Scales (ASRS) Symptom Checklist
is the official instrument of the WHO (Kessler et al. 2005)
and is available in many languages.1 It has two forms: a
standard version, which contains eighteen DSM-IV items,
and a shorter one, the ASRS Screener, that has four inat-
tention items and two hyperactivity items. It is generally
accepted to detect self-report pathology in adults (Rosler
et al. 2010). The items can be scaled from 0 to 4. Internal
consistency reliability of the ASRS Screener was 0.678 and
of the whole scale 0.909, which is comparable to available
data (Kessler et al. 2007; Adler et al. 2006).
The Barkley Adult ADHD Rating Scale-IV (BAARS-
IV): Current Symptoms is a new behavior rating scale for
adult ADHD symptoms that consists of the 18 DSM-IV
symptoms (9 for inattention, 5 for hyperactivity and 4 for
impulsivity) together with 9 additional items for evaluating
the symptoms of sluggish cognitive tempo (SCT), a pos-
sible distinct type of ADHD (Barkley 2011). As in the case
of ASRS, it has also a shorter version: the Quick Screen.
The items are scaled from 1 to 4. Internal consistency was
0.777 for the screener, 0.914 for total score, 0.870 for
Inattention, 0.754 for the Hyperactivity, 0.807 for the
Impulsivity and 0.852 for SCT subscales, respectively.
Sleep disturbance
report categorical questionnaire, which evaluates the
presence of insomnia according to the DSM-IV and ICSD-
R criteria (Violani et al. 2004). It allows respondents’
classification into three categories: good sleep, chronic
insomnia and sub-threshold insomnia. In this paper, we
used the term ‘‘chronic insomnia’’ for the sleep disturbance
that was reported as being present three or more nights per
week, which lasted for more than 1 month and caused
significant impairment in social, occupational or cognitive
1 downloadable from http://www.hcp.med.harvard.edu/ncs/asrs.php.
daytime functioning. Sub-threshold insomnia was defined
as above in terms of duration and impairment, but had to be
reported for 2–3 nights per week instead of more than three
nights per week. The term ‘‘acute insomnia’’ was used
when insomnia had been present for less than a month, for
at least 3 days in a week, and had caused significant
impairment in at least one area of daytime functioning.
The Sleep Condition Indicator (SCI) is a new insomnia
self-report instrument (Espie 2011) that consists of eight
items, scored on a Likert scale from 0 (Not at all) to 4
(Very much) that gathers information on sleep pattern in
the last month (minutes needed for falling asleep, minutes
being awake during the night, number of problem nights in
a week), sleep quality, impact of poor sleep, level of
concern about poor sleep and history of poor sleep. It is
based on proposed DSM-5 criteria. Internal consistency
was 0.862.
Diurnal preference
used instrument for determining circadian typology (mor-
ningness/eveningness) and is available in several lan-
guages. It contains 13 questions, most of them having four
choices, with a Likert-type response format and total scores
range from 13 (extreme eveningness) to 55 (extreme
morningness) (Smith et al. 1989). We used authors’ rec-
ommendations and calculated the cut off scores for the
sample using the 10/90 percentile (higher than 42 for
morningness, lower than 24 for eveningness).
The Sleep Timing Questionnaire (STQ) is an alternative
to a sleep journal and contains 16 questions regarding sleep
timings and their variability for work and rest nights, as
well as 2 items for ‘‘unwanted wakefulness’’ (Monk et al.
2003).
Procedure
Psychology at ‘‘Alexandru Ioan Cuza’’ University in Iasi
and at ‘‘Babes-Bolyai’’ University in Cluj-Napoca,
Romania, to participate in an online survey. They were
given a web link with the description of the study and
offered extra credit for participating in the research. Col-
lection took place during the first and second semesters of
2011. Participants for Sample 2 were recruited from adults
from the general community via a dedicated website,
online adverts posted on health-related sites, as well as via
adverts in local mass media in Cluj-Napoca, Romania.
Participation was unpaid and all respondents received a
brief interpretation of their scores on the completion of the
questionnaires, as well as the possibility of a free basic
clinical assessment if it was the case. Data collection took
place during the second semester of 2011.
Data analysis
sub-total scores. T tests or ANOVAs were used to compare
the differences between means and the Chi-Square test for
differences between frequencies among various groups,
respectively, with p \ 0.05 considered significant. Pearson
R or Spearman q was used for calculating correlations
between different variables. FUZZY extension2 was used
to match participants. Data analysis was performed with
IBM SPSS Statistics 20.0.0.
We first examined differences in age and scores in ADHD
scales that could be related to gender and found that men
scored significantly higher in BAARS Hyperactivity sub-
scale in Sample 2. There were significant differences for
the total morningness and sluggish cognitive tempo scores
between Samples 1 and 2, As seen in Table 1, the differ-
ence in diurnal preference was obvious as about a third in
Sample 2 were morning types (compared to less than 10 %
in Sample 1), as might be expected due to the different age
profiles of the samples.
Using ASRS and BAARS-IV and the proposed cut-off
points (Barkley 2011; Gau et al. 2007), we identified 46
(8.3 %) participants (aged 26.93 (10.14), most of them
women (N = 32, 69.6 %)), who were likely to meet the
DSM-IV criteria for ADHD. The majority were from
Sample 1 (N = 30, 9.9 % of the total), while the remaining
third came from Sample 2 (N = 16, 6.4 % of the total).
Based on responses in SDQ, 73 (13.24 %) participants met
the criteria for chronic insomnia, while 327 (59.3 %)
reported sleeping soundly. A minority (N = 35, 6.4 %)
complained of acute symptoms, while the rest of the sleep
disturbance that did not meet the DSM or ICSD criteria for
acute or chronic insomnia (‘‘sub-threshold’’ insomnia).
Chronic insomnia had a prevalence of 11.3 % in Sample 1
and 15.6 % in Sample 2. From those likely to suffer from
ADHD, 21.7 % (N = 10) met the criteria for chronic
insomnia, another ten (21.7 %) for sub-threshold insomnia
and six more (13.0 %) for acute insomnia, meaning that
2 available for download at http://www.ibm.com/developerworks/
spssdevcentral.
Sleep disturbance, circadian preference and symptoms of adult ADHD 1197
54.5 %).
We eliminated those selected for being highly likely to
suffer from ADHD or from chronic insomnia and randomly
matched the ADHD-likely participants by age and sex with
46 ‘‘healthy controls’’ (mean age 27.00 ± 10.34, 30.4 %
men, N = 14). Evening typology was significantly more
frequent in the ADHD group (23.9 %, N = 11) than in the
control one (4.3 %; N = 2), while there were significantly
fewer morning types (10.9 %, N = 5 in ADHD group,
19.6 %, N = 9 in Non-ADHD one). Comparisons revealed
significant differences in sleep quality and morningness
score, with the ADHD-likely group sleeping worse, and
displaying more evening orientation, as indicated by
higher SCI scores and lower CSM scores, respectively
(Table 2).
We also compared these same parameters in those of the
remaining sixty-three participants complaining of chronic
insomnia (after exclusion of the ten ADHD-likely sub-
jects), as matching by age and sex could not be performed.
Significant differences were found in sleep quality score
and in all ADHD scales, with insomnia suspects scoring
higher in SCI and lower in ADHD scales (Table 2).
Compared to the control group, those likely to suffer from
chronic insomnia scored higher in all ADHD scales,
particularly in the SCT. Chronotype distribution did not
differ significantly between the groups. When analyzing
the differences in sleep timings (going to/going out of bed
times, both in workdays and rest days), we found that
ADHD-likely individuals went to bed later than the healthy
controls, particularly before a rest day, and also got out of
bed at a later time (Table 3). Regarding sleep duration and
latency, ADHD-likely respondents reported significantly
less time being asleep and more time needed to fall asleep
(19.6 %, N = 9 reported needing more than 45 min com-
paring to none among controls); when comparing these to
insomnia participants, they slept less, although the differ-
ences were not significant; nevertheless, 34.9 % (N = 22)
in the insomnia group reported needing more than 45 min
for falling asleep and 38.1 % (N = 24) being awake for
more than the same amount of time, compared to only
8.7 %, N = 4 in ADHD group.
In the ADHD-likely sample, we found that latest waking
up time during rest days was moderately correlated with
SCI (Spearman q = 0.348, p = 0.03) but not correlated
with ADHD symptoms. Also, SCT scores had a large
association with SCI (q = 0.659, p \ 0.001), and moder-
ately so with impulsivity (q = -0.410, p = 0.006) and
inattention (q = 0.488, p = 0.001). Morningness score
was moderately linked with SCI (q = -0.328, p = 0.026)
and inattention (q = -0.360, p = 0.018). Among the
insomnia group, habitual bed time (q = -0.304,
p = 0.028) and rising time at weekends (q = -0.329,
Table 1 Summary of the descriptive data for samples 1 and 2
Sample 1 Sample 2
Females Males Total Females Males Total
N (%) 255 (84.7) 46 (15.3) 301 176 (70.4) 74 (29.6) 250
Age (SD) 21.78 (2.55) 22.00 (2.39) 21.82 (2.52) 37.80 (12.10) 40.53 (12.95) 38.61 (12.4)
ASRS (SD)
Inattention (SD) 14.50 (6.18) 16.33 (5.54) 14.81 (6.10) 21.04 (3.85)a 22.00 (5.94)a 21.26 (4.28)
Hyperactivity/impulsivity (SD) 12.13 (5.88) 14.30 (6.43) 12.50 (6.01) 17.38 (5.85)a 14.86 (4.74)a 16.81 (5.64)
BAARS-IV
Inattention (SD) 15.26 (4.55) 16.37 (4.89) 15.43 (4.61) 14.46 (4.23) 15.09 (4.53) 14.64 (4.32)
Hyperactivity (SD) 7.42 (2.41) 7.94 (3.04) 7.49 (2.51) 6.82 (2.13) 7.90 (2.42) 7.12 (2.26)
Impulsivity (SD) 6.88 (2.46) 7.10 (3.04) 6.91 (2.54) 6.46 (2.14) 7.07 (2.25) 6.63 (2.18)
Sluggish cognitive tempo (SD) 17.24 (4.88) 17.58 (5.38) 17.29 (4.94) 15.50 (4.24) 16.41 (4.41) 15.76 (4.30)
CSM (SD) 33.24 (6.59) 32.32 (6.82) 33.1 (6.54) 37.42 (7.90) 36.53 (8.03) 37.15 (7.93)
SCI (SD) 11.96 (6.31) 11.30 (6.06) 11.86 (6.27) 11.85 (7.83) 11.21 (6.79) 11.66 (7.53)
ADHD likely N (%) 22 (73.3) 8 (26.6) 30 (9.9) 10 (62.5) 6 (37.5) 16 (6.4)
Chronic insomnia N (%) 31 (91.1) 3 (8.9) 34 (11.3) 27 (69.2) 12 (30.8) 39 (15.6)
Circadian typology
Morning (%) 26 (10.2) 2 (4.3) 28 (9.3) 64 (36.4) 18 (24.3) 82 (32.8)
Indifferent (%) 201 (78.8) 36 (78.3) 237 (78.7) 100 (56.8) 48 (64.9) 148 (59.2)
Evening (%) 28 (11.0) 8 (17.4) 36 (12.0) 12 (6.8) 8 (10.8) 20 (8.0)
a Scores were calculated only for those who screened positive in the ASRS screener (N = 31)
1198 B. I. Voinescu et al.
123
p = 0.021) were associated with SCI scores. Hyperactivity
scores were significantly, but weakly associated with sleep
duration (R = -0.177, p = 0.024) and sleep latency
(R = 0.176, p = 0.024) in the whole sample, but not in the
ADHD group.
during childhood significantly more frequently than con-
trols (N = 8, X = 6.03, p = 0.014), as well as more cur-
rent sleep problems (onset and maintenance insomnia, non-
refreshing sleep) and daytime sleepiness (Table 4).
Other analyses
quality, while hyperactivity showed a small magnitude
association. A small negative association of inattention
with total morningness score was detected in Sample 1 and
a larger one in Sample 2 (Table 5). When examining
differences between average scores on ADHD scales
according to circadian typology and presence of insomnia,
Table 2 Comparison of
BAARS-IV
ASRS
Circadian typology
Table 3 Sleep timings for
ADHD-likely, non ADHD and
Rest days
Good morning time (SD)
Rest days
Sleep duration (SD) 5.64 (1.39) 7.18 (1.23)** 6.17 (1.55)
Sleep latency (SD) 32.71 (24.77) 15.18 (7.36)** 41.79 (36.02)
Sleep disturbance, circadian preference and symptoms of adult ADHD 1199
123
we found significant differences (Table 6). As we had a
large number of morning types when using the authors’
criteria (Smith et al. 1989), we calculated the cut-off points
by three different age groups and 10/90 percentiles.
According to the new criteria, 54 (9.8 %) participants were
evening types (compared with 56 initially) and 58 (10.5 %)
morning types (110 initially). ANOVA revealed significant
differences across the different circadian preferences or
insomnia in Inattention, Hyperactivity, SCT and SCI
scores, but not in Impulsivity (Table 6). Post hoc tests
showed that the differences were significant when com-
paring insomnia subtypes to no insomnia (but not between
insomnia subtypes) and between chronotypes (evening
types scored higher).
ADHD and insomnia symptoms, sleep and circadian
typology. Our study protocol allowed us to gather and
compare data from two different populations: young adults
(students) and an older cohort from the general population.
We found that ADHD symptoms were more severe in
younger adults and that eveningness was more frequent
both among ADHD and chronic insomnia suspected indi-
viduals, while inattention was strongly linked with dis-
turbed sleep. Sleep disturbance, but not ADHD symptoms,
was linked with some of the sleep timings, and was
reported by more than a half of the ADHD-likely group
(summarized in Fig. 1).
the hyperactivity scale compared to women, across all
ages, and that this was more pronounced in older subjects.
Also, the young adult sample revealed significant lower
morningness scores compared to the older sample, with a
likely explanation being the effect of age on chronotype
(Paine et al. 2006; Park et al. 1998). ADHD-likely indi-
viduals were likely to be younger, probably as symptoms
reduce in severity during lifetime, while insomnia was
more prevalent in the general population sample, sleep
becoming more superficial and shorter in duration with
advancing age.
ing up too early or non-refreshing sleep, (the core symp-
toms of insomnia disorder) were reported by more than a
half of the probable-ADHD group, with slightly less than a
quarter of this group complaining of these symptoms…