Embed Size (px)
Citation preview
Sleep and its Association with Cardiometabolic Risk Factors
in Young Adults
This thesis is presented for the degree of
Doctor of Philosophy
by
Anahita Hamidi
MD, MSc
Medical School
University of Western Australia
March 2018
“Home may be perilous and the destination out of reach
But there are no paths without an end, do not grieve”
Hafiz, Persian Poet
Poem, “The lost Joseph”
iv
Abstract
Cardiovascular disease, encompassing heart disease, stroke and vascular disease, is the
leading cause of death in Australia and globally, with the underlying cardiometabolic
risk factors including overweight and obesity, hypertension, dyslipidaemia, insulin
resistance and inflammation. Known lifestyle factors that influence cardiovascular
disease include physical inactivity, diet, alcohol consumption and smoking. Recent
evidence suggests that insufficient sleep and/or poor sleep quality are associated with
cardiometabolic risk factors. Therefore, inadequate sleep quality and/or duration is
potentially a further modifiable lifestyle factor. In particular, studies have shown a
positive relationship between insufficient sleep and obesity in children and adolescents.
However, evidence in other age groups is inconsistent. In addition to insufficient sleep,
a positive association between sleep disordered breathing and cardiometabolic risk
factors has been reported in adult studies, however, the results are inconsistent. The
influence of sleep and sleep-breathing disorders on cardiometabolic risk factors during
young adulthood, prior to establishment of chronic diseases, is still unknown.
This thesis investigated data from young adults that participated in the 22-year follow-
up of the Western Australian Pregnancy Cohort (Raine) Study. The aim was to
investigate the associations between: (i) objective measurements of sleep and
cardiometabolic risk factors; (ii) sleep quality and cardiometabolic factors; and (iii)
sleep-disordered breathing (high risk for obstructive sleep apnoea) and cardiometabolic
risk factors.
The Raine Study is a population-based pregnancy cohort study established between
1989 and 1991 that included 2900 pregnant women recruited at 18 weeks of gestation.
The Raine Study aimed to ascertain the effects of intrauterine, familial, environmental
and lifestyle factors on the health of the 2868 offspring. The study has detailed maternal
and paternal information during pregnancy, as well as sociodemographic,
developmental, anthropometric, clinical and biochemical data collected on the offspring
at birth and during follow-up assessments at 1, 2, 3, 5, 8, 10, 14, 17, 18, 20 and 22 years
of age. There were 1234 participants at the 22-year follow-up which was conducted
from 2012-2014. Sociodemographic and lifestyle behaviours (smoking, alcohol intake,
physical activity and depression and anxiety scores) were evaluated from
questionnaires. Anthropometric, blood pressure and fasting blood samples (glucose,
v
insulin, high sensitivity C - reactive protein [hs-CRP] and lipids) were assessed in 975
of the participants. Sleep characteristics were measured objectively by actigraphy with a
wrist accelerometer worn for 7 days in the home environment. Accelerometer data were
collected from 581 participants and sleep variables including sleep duration, sleep
efficiency, sleep latency and Wake after Sleep Onset (WASO) were analysed. Sleep
quality was measured subjectively using the Pittsburgh Sleep Quality Index (PSQI)
Questionnaire and high risk for sleep apnoea (OSA) was evaluated using the Berlin
Questionnaire. Data were analysed using linear and logistic regression models adjusting
for gender and lifestyle factors.
Mean sleep duration was 6.60 hours and the prevalence of short sleepers (< 6
hours/night) was 25.5% in this population. There were no associations between any of
the actigraphy-derived sleep variables and body mass index, blood pressure, or fasting
blood glucose, insulin, lipids and hs-CRP levels before or after adjustment for gender
and lifestyle factors.
Based on the PSQI Questionnaire 28% of the individuals had poor sleep quality. Poor
sleep quality was not associated with any cardiometabolic risk factors measured
although participants with poor sleep quality were more likely to be smokers and have
higher depression and anxiety scores.
Using the Berlin Questionnaire, the prevalence of high risk for OSA was 14.75% with
similar gender distribution. Smoking was the only lifestyle variable associated with high
risk for OSA. Those at high risk for OSA were more likely to be overweight/obese or
have central obesity. High risk for OSA was positively associated with systolic blood
pressure, triglycerides, hs-CRP and inversely associated with high density lipoprotein
cholesterol (HDL-C), before and after adjustment for gender and lifestyle factors.
However, these associations were no longer apparent when sensitivity analyses were
performed that accounted for BMI.
In summary, this study has shown that sleep characteristics are not related to
cardiometabolic risk factors in a young adult population. Further, while being at high
risk for OSA associated with obesity and related cardiometabolic risk factors, the
relationship is likely due to the underlying presence of obesity in these individuals. In
conclusion, data from the Raine Study do not show any evidence for a relationship
between sleep characteristics, high risk for OSA and cardiometabolic risk factors (other
than obesity) in a young adults.
vi
Acknowledgements
I gratefully acknowledge the University of Western Australia and the Western
Australian Pregnancy Cohort (Raine) Study for providing the scholarship and academic
support. I would also like to express my gratitude to all of the Raine Study participants,
their families and the Raine Study team.
I would like to express my gratitude to my main supervisors, Prof. Trevor Mori and Prof
Peter Eastwood for their invaluable advice, tremendous encouragement and support. It
has been a great honour to study and work with them. I would also like to thank my co-
supervisor A/Prof Rae-Chi Huang and E/Prof Lawrence Beilin for all of their valuable
suggestions and support. I am deeply thankful to Mrs. Sally Burrows for her statistical
advice and guidance. I would like to thank all of my colleagues at Royal Perth Hospital,
the Raine Study House and the Centre for Sleep Sciences at the University of Western
Australia.
To my dearest parents and my lovely sister, thank you for your constant support and
prayers, you always have given me even though you are thousands miles away from me.
Super special thanks go to my precious daughter, Vania for all of her understanding and
patience. Thank you for all the cheers and happy moments you created for your sole
parent during her PhD study. By completion of this PhD, I hope to model you that you
can pursue your dream with your determination and efforts despite the obstacles and
challenges. A big special thanks to Ms. Minoo Zand, thank you for all of your support,
tremendous encouragement and invaluable advice you have given to me. This thesis is
dedicated to my beloved grandmother who passed away while I was doing my PhD.
vii
List of Awards and Presentations
Awards
1. University Postgraduate Award-International student (UPAIS), University
Western Australia
2. Scholarship International Research Fee (SIRF), University Western Australia
3. Raine Study Top-Up Scholarship, The Raine Foundation
4. Foundation for High Blood Pressure Research Council Australia Travel Award,
Joint Annual Scientific Meeting of the Australian Atherosclerosis Society, the
Australian Vascular Biology Society and the High Blood Pressure Research
Council of Australia, Hobart, Tasmania, December 2016
Presentations
1. Relationship between sleep and cardiometabolic factors in young adults.
Anahita Hamidi, The Western Pregnancy Cohort (Raine) Study Annual Scientific
Meeting, Perth, Western Australia, September 2016
2. Relationship between sleep and cardiometabolic factors in young adults.
Anahita Hamidi, Royal Perth Hospital Medical Research Foundation Young
Investigators Day, Perth, Western Australia, October 2016
3. Relationship between sleep and cardiometabolic factors in young adults: the
RAINE Study.
Anahita Hamidi, Joint Annual Scientific Meeting of the Australian Atherosclerosis
Society, the Australian Vascular Biology Society and the High Blood Pressure
Research Council of Australia, Hobart, TAS, December 2016
viii
Table of Contents
Declaration ...................................................................................................................... iii
Abstract ........................................................................................................................... iv
Acknowledgements ......................................................................................................... vi
List of Awards and Presentations ................................................................................ vii
Table of Contents ......................................................................................................... viii
List of Tables ................................................................................................................... x
List of Figures ................................................................................................................ xii
Abbreviations and Acronyms ..................................................................................... xiii
Chapter 1: Literature Review & Study Aims ............................................................... 1 1.1 Cardiometabolic Risk Factors and Role of Sleep ................................................... 2
1.2 Human Sleep ........................................................................................................... 3 1.3 Objective Sleep Measurements ............................................................................... 4 1.4 Self-report Sleep Measurement ............................................................................... 6
1.4.1 Pittsburgh Sleep Quality Index Questionnaire ................................................. 6
1.4.2 Berlin Questionnaire ........................................................................................ 7 1.5 Association Between Short Sleep Duration and Weight Gain ................................ 8
1.5.1 Cross-sectional and Longitudinal Studies in Children and Adolescents ......... 8
1.5.2 Cross-sectional and Longitudinal Studies in Adults ...................................... 12 1.5.3 Physiological Basis for an Association Between Sleep Restriction and
Obesity ........................................................................................................... 15
1.6 Sleep Duration and Glucose Metabolism .............................................................. 17
1.6.1 Sleep Duration and Glucose Metabolism ....................................................... 17 1.6.2 Sleep Duration, Glucose Dysregulation and Risk of Diabetes ...................... 17
1.7 Sleep Duration and Other Cardiometabolic Risk Factors ..................................... 19 1.8 Sleep Quality and Cardiometabolic Risk Factors ................................................. 21 1.9 Sleep-disordered Breathing and Cardiometabolic Risk Factors in Adults ............ 22
1.9.1 Sleep-disordered Breathing in the General Population .................................. 22 1.9.2 Sleep-disordered Breathing and Obesity ........................................................ 23
1.9.3 Sleep-disordered Breathing and Glucose Metabolism ................................... 26 1.9.4 Sleep-disordered Breathing and Hypertension .............................................. 29 1.9.5 Sleep-disordered Breathing and Inflammation .............................................. 30
1.9.6 Sleep-disordered Breathing and Dyslipidaemia ............................................. 31 1.10 Summary ............................................................................................................. 32
1.11 The Western Australia Pregnancy Cohort (Raine) Study ................................... 32 1.12 Study Aims .......................................................................................................... 34
Chapter 2: Methods ...................................................................................................... 35 2.1 Background of the Raine Study ............................................................................ 36 2.2 Follow-up of the Raine Study Participants at 22 Years ........................................ 40 2.3 Anthropometric Measurements ............................................................................. 40 2.4 Blood Pressure Measurement ................................................................................ 41
2.5 Blood Collection ................................................................................................... 41 2.6 Questionnaires ....................................................................................................... 42 2.7 Selected Risk Taking and Lifestyle Variables ...................................................... 43 2.8 Subjective Sleep Measurements ............................................................................ 44
2.8.1 Sleep Diary ..................................................................................................... 44
ix
2.8.2 Sleep Questionnaires ...................................................................................... 44 2.8.2.1 Pittsburgh Sleep Quality Index Questionnaire ....................................... 44
2.8.2.2 Berlin Questionnaire ............................................................................... 45 2.8.2.3 Epworth Sleepiness Scale Questionnaire ................................................ 47 2.8.2.4 Short Functional Outcome of Sleep Questionnaire ................................ 48
2.9 Objective Sleep Measurement............................................................................... 48 2.9.1 Wrist Accelerometer ...................................................................................... 48 2.9.2 Wrist Accelerometry (Actigraphy) Data Collection ...................................... 49 2.9.3 Actigraphy Software &Analysing Sleep Actigraphy Data ............................ 50
2.10 Statistical Methods .............................................................................................. 52
Chapter 3: Results ......................................................................................................... 54 3.1 Follow-up Response of the Raine Study Participants at 22 Years of Age ............ 55 3.2 Socioeconomic and Lifestyle Characteristics of the Raine Study Participants
at 22 Years ............................................................................................................. 58
3.3 Anthropometric and Biochemical Characteristics of the Raine Study
Participants at 22 Years ......................................................................................... 59 3.4 Actigraphy Data of the Raine Study Participants at 22 Years .............................. 62
3.4.1 Actigraphy-derived Sleep Variables .............................................................. 65 3.4.2 Association between Actigraphy Sleep Variables and Cardiometabolic
Profile ............................................................................................................ 65 3.4.2.1 BMI and Sleep Variables ........................................................................ 65
3.4.2.2 Waist Circumference and Sleep Variables .............................................. 69 3.4.2.3 Blood Pressure, Biochemical Variables and Sleep Variables ................ 73
3.4.3 Sleep Duration Categorization ....................................................................... 77
3.5 Associations between the Pittsburgh Sleep Quality Index (PSQI) and
Socioeconomic, Risk Taking Behaviour and Cardiometabolic Factors in the
Raine Study Participants at 22 Years .................................................................... 79
3.6 Relationships Between Obstructive Sleep Apnoea Risk and Cardiometabolic
Risk in the Raine Study Participants at 22 Years .................................................. 85
Chapter 4: Discussion and Conclusions ...................................................................... 92
4.1 Sleep Duration ....................................................................................................... 93 4.2 Association between Actigraphy-derived Sleep Variables and
Cardiometabolic Risk Factors ............................................................................... 94
4.2.1 Association between Sleep Duration and Efficiency and Body Weight ........ 95 4.2.2 Association Between Sleep Duration and Glucose and Insulin Resistance ... 96
4.2.3 Association between Sleep Duration and Blood Pressure ............................. 97 4.2.4 Association Between Sleep Duration and Inflammatory Markers and
Lipids ............................................................................................................. 98
4.2.5 Summary of Evidence of Association Between Sleep Duration and
Cardiometabolic Risk Factors ..................................................................... 100 4.3 Sleep Quality (PSQI Questionnaire) and Socioeconomic, Lifestyle and
Cardiometabolic Risk Factors ............................................................................. 101
4.4 Relationships Between Risk of Obstructive Sleep Apnoea and
Socioeconomic, Lifestyle and Cardiometabolic Risk Factors ............................. 102 4.5 Strengths and Limitations ................................................................................... 106 4.6 Conclusions and Future Directions ..................................................................... 107
References .................................................................................................................... 109
Appendix 1 Berlin Questionnaire .............................................................................. 126
x
List of Tables
Table 1.1. Cross-sectional Studies of Sleep Duration and Weight Gain in Children
and Adolescents ............................................................................................ 10
Table 1.2. Cross-sectional Studies of Associations Between Sleep Duration and
Weight in Adults ........................................................................................... 14
Table 2.1. Data recorded during each of the cohort recalls in the Raine Study .... .. ..... 39
Table 3.1. Univariate Regression Analyses Evaluating the Factors Predicting the
Likelihood of Participating at the 22 Year Follow Up ............................ ...... 57
Table 3.2. Socioeconomic Characteristics of the Participants at 22 Years ..................... 59
Table 3.3. Characteristics of the Raine Participants at 22 Years Stratified by Gender... 61
Table 3.4. Characteristics of females stratified by hormonal contraceptive use. ............ 62
Table 3.5. Comparison of Actigraphy Sleep Parameters Stratified by Total Number
of Nights ......................................................................................................... 63
Table 3.6. Comparative Characteristics of the Raine Study Participants Stratified by
Attendance in the Actigraphy Study .............................................................. 64
Table 3.7. Actigraphy Sleep Characteristics of the Raine Participants........................... 65
Table 3.8. Linear Regression Models of BMI and Sleep Indices ................................... 68
Table 3.9. Linear Regression Models of Waist Circumference and Sleep Indices ......... 72
Table3.10. Linear Regression Models of Total Sleep Time and Blood Pressure and
Biochemical Variables ................................................................................... 74
Table 3.11. Linear Regression Models of Sleep Efficiency and Blood Pressure and
biochemical variables ..................................................................................... 75
Table 3.12. Linear Regression Models of Sleep Latency and Blood Pressure and
Biochemical Variables ................................................................................... 76
Table 3.13. General and Cardiometabolic Characteristics According to Sleep
Categories ....................................................................................................... 78
Table 3.14. Socioeconomic Status According to Sleep Categories ................................ 79
Table 3.15. Characteristics of the Raine Study Participants According to PSQI
Sleep Quality .................................................................................................. 81
Table 3.16. Linear Regression Analyses: Association Between Sleep Quality and
Waist Circumference, and Systolic and Diastolic Blood Pressure ................ 83
Table 3.17. Characteristics of the Raine Study Participants Stratified by Sex and
PSQI Sleep Quality ........................................................................................ 84
xi
Table 3.18. Number and Proportion of Individuals with Positive Berlin Scores for
(i) Overall Score, (ii) Each Category and (iii) Combinations of Individual
Categories. ...................................................................................................... 85
Table 3.19. Characteristics of the Raine Study Participants According to High Risk
for Obstructive Sleep Apnoea ........................................................................ 86
Table 3.20. Metabolic Variables of the Raine Study Participants According to High
Risk for Obstructive Sleep Apnoea (OSA) .................................................... 87
Table 3.21 Linear Regression Analyses Between High Risk for OSA (Total Berlin
Questionnaire) and Systolic and Diastolic Blood Pressure and Metabolic
Variables ........................................................................................................ 89
Table 3.22. Linear Regression Analyses Between One Positive Categories (Snoring
or Sleepiness) and Systolic Blood Pressure and Metabolic Variables ........... 90
Table 3.23. Linear Regression Analyses Between Both Positive Categories (Snoring
or Sleepiness) and Systolic Blood Pressure and Metabolic Variables ........... 91
xii
List of Figures
Figure 1.1. Normal sleep architecture during 8 hours of sleep . ....................................... 4
Figure 1 .2. Cycle of obesity and OSA and its potential mechanisms . ......................... 24
Figure 1.3. Obstructive sleep apnoea and cardiometabolic risk factors. ......................... 28
Figure 2.1. The Raine Study: Retention of the cohort from 1 to 22 years of age.. ......... 37
Figure 2.2. PSQI Questionnaire . .................................................................................... 45
Figure 2.3. Berlin Questionnaire . ................................................................................... 47
Figure 2.4. The Wrist Accelometer used in the Raine Study participants. ..................... 49
Figure 3.1. The Raine Cohort Study Retention from birth to 22 years. .......................... 56
Figure 3.2. Flowchart of sleep actigraphy data. .............................................................. 63
Figure 3.3. Scatter plot association between BMI and total sleep time (min). ............... 66
Figure 3.4. Scatter plot association between BMI and sleep efficiency (%). ................. 66
Figure 3.5 Scatter plot association between BMI and sleep latency (min). .................... 67
Figure 3.6. Scatter plot association between BMI and Wake After Sleep Onset
[(WASO); (min)] ............................................................................................ 67
Figure 3.7. Scatter plot association between waist circumference (cm) and total
sleep time (min).............................................................................................. 69
Figure 3.8. Scatter plot association between waist circumference (cm) and sleep
efficiency (%). ................................................................................................ 70
Figure 3.9. Scatter plot association between waist circumference (cm) and sleep
latency (min). ................................................................................................. 70
Figure 3.10 Scatter plot association between waist circumference (cm) and Wake
After Sleep Onset (WASO); (min). ................................................................ 71
Figure 3.11. BMI distributions according to sleep duration categories. ......................... 79
xiii
Abbreviations and Acronyms
AASM American Academy of Sleep Medicine
AHI apnoea-hypopnoea index
ALP alkaline phosphatase
ALT alanine aminotransferase
AST aspartate aminotransferase
AusDiab Australian Diabetes, Obesity and Lifestyle
BMI body mass index
BP blood pressure
CARDIA Coronary Artery Risk Development in Young Adults
CIH chronic intermittent hypoxia
CPAP continuous positive airway pressure
CRP C-Reactive Protein
DASS depression anxiety score system
DBP diastolic blood pressure
DEXA dual-energy X-ray absorptiometry
DoHAD Development Origins of Health and Disease
ECG electrocardiogram
EEG electroencephalographic
EMG electromyogram
EOG electrooculogram
FOSQ functional outcome of sleep questionnaire
xiv
GGT gamma-glutamyl transferase
GH growth hormone
g/wk gram/week
HbA1c haemoglobin A1c
HDL-C high density lipoprotein cholesterol
HELIUS multi-ethnic Healthy Life in an Urban Setting
HOMA-IR homoeostasis model assessment for insulin resistance
hs-CRP high sensitivity CRP
IL-6 interleukin 6
IPAQ international physical activity questionnaire
JNC-7 The Seventh Report of the Joint National Committee on the Prevention,
Detection, Evaluation and Treatment of High Blood Pressure
LDL-C low density lipoprotein cholesterol
METS metabolic equivalences
min minute
MMAS Massachusetts Male Aging Study
NHANES National Health and Nutrition Examination Survey
NHIS National Health Interview Survey
NO nitric oxide
NPY/AgRP neuropeptide Y/agouti-related protein
NREM non-rapid eye movement
OGTT Oral Glucose Tolerance Test
OSA obstructive sleep apnoea
xv
PA physical activity
POMC/CART propiomelanocortin/cocaine and amphetamine–regulated transcript
PSG polysomnography
PSQI Pittsburgh Sleep Quality Index
Raine Study The Western Australian Pregnancy Cohort Study
REM rapid-eye movement
ROS reactive oxygen species
SBP systolic blood pressure
SBSM Society of Behavioural Sleep Medicine
SDB sleep-disordered breathing
SE sleep efficiency
SOL Sleep Onset Latency
TG triglycerides
TNF-α tumour necrosis factor alpha
TST total sleep time
UA uric acid
WASO wake after sleep onset
WHO World Health Organisation
1
Chapter 1:
Literature Review & Study Aims
2
1.1 Cardiometabolic Risk Factors and Role of Sleep
Cardiometabolic risk factors encompass the related conditions of obesity, diabetes,
hypertension and cardiovascular disease (1). These conditions share common risk
factors and are associated with reduced quality of life, decreased life expectancy and
increased economic burden (1). Over the last few decades, the prevalence of these
cardiometabolic risk factors has increased worldwide in both developed and developing
countries (1). Amongst developed countries, Australia has one of the highest rates of
overweight and obesity; the prevalence of overweight and obesity has increased from
56% in 1995 to 63% in 2012 (2). Over the same time period, the proportion of obese
Australians increased from 19% to 28% (2). Obesity has been reported to shorten life
expectancy by nine years and is associated with chronic diseases and disabilities which
bring about significant economic burden (2). Based on the data from the Australian
Diabetes, Obesity and Lifestyle (AusDiab) study, the direct and indirect costs of obesity
is estimated at $56.6 billion per year (2).
The metabolic syndrome refers to the co-occurrence of several known cardiovascular
risk factors including hypertension, dyslipidaemia (elevated triglycerides and lowered
high-density lipoprotein cholesterol), insulin resistance, and central obesity (3). These
conditions are interrelated and share underlying mediators, mechanisms and pathways.
Although there are several definitions of the metabolic syndrome including those by the
International Diabetes Federation (4), the World Health Organization (WHO) (5)and the
National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATP III)
(5), a central feature of each is the inclusion of insulin resistance, visceral adiposity,
atherogenic dyslipidaemia and high blood pressure. The metabolic syndrome also
associates with proinflammatory and prothrombotic states, as well as increased
oxidative stress (6). Studies have shown that the metabolic syndrome is associated with
a 2-fold increase in cardiovascular mortality and a 1.5-fold increase in all cause
mortality, however, it is still unclear whether the prognostic significant of metabolic
syndrome exceeds the risk associated with the sum of its individual components
(7).There is strong evidence to support the hypothesis that many chronic diseases
including cardiovascular disease, begin a long time before their clinical manifestations
(8). Autopsy studies of young adults following death from trauma show a significant
link between cardiometabolic risk factors in childhood and fatty streaks in coronary
arteries (8). Many risk factors related to chronic disease (including cardiometabolic risk
factors) are also known to accumulate gradually over time thus contributing to the
3
development and progression of chronic disease (9). During adolescence and young
adulthood a number of behaviour and lifestyle factors can adversely impact on well-
being (10). Thus early identification and prevention of these factors may play a
significant role in health care policies designed to prevent and diminish the burden of
cardiometabolic risk factors.
Recently, the role of sleep in cardiometabolic risk factors has gained attention and the
increase in obesity is paralleled with a decrease in nightly sleep duration (11, 12).
According to National Health And Nutrition Examination Survey (NHANES) results,
adults aged 20-59 have less sleep each night than adults 60 years and over and 37% of
adults have less than six hours of sleep per night (13). Online surveys have shown that
the average sleep duration in Australia has dropped from 8 hours in 2000 to 7 hours in
2010 (14). Although these data do not provide evidence of a causal relationship between
sleep and cardiometabolic risk factors, they highlight the need for studies to better
understand the mechanisms underlying the associations between sleep and cardio
metabolic risk factors and how these may differ with age.
1.2 Human Sleep
Sleep is defined by the physiological stages of reduced consciousness, lack of activity
and reduction in sensory activity (15). The precise role of sleep remains a mystery,
although its role in maintaining normal cognitive, motor and metabolic function has
been clearly established (16). According to the American Academy of Sleep Medicine
(AASM), sleep involves two separate electroencephalograph (EEG) stages: non-rapid
eye movement (NREM) and rapid-eye movement (REM) (15). During a sleep episode
NREM and REM stages alternate cyclically (17). In a normal night of sleep, there are 3-
5 sleep cycles and each sleep cycle (NREM and REM) has a period of 90-120 minutes
duration (17). Sleep starts with NREM in all age groups except newborns (17). NREM
is further divided into three deeper stages defined as N1, N2 and N3 (Figure 1.1).
Increasing depth of NREM sleep is characterized by increases in EEG wave amplitude
and decreases in EEG wave frequency. Deeper stages of NREM occur before the first
episode of REM. In the average adult, 75-80% of total sleep is spent in the NREM
stage. During NREM sleep there is a down-regulation of cardiovascular activity
including a decrease in heart rate, stroke volume and arterial blood pressure (17). As a
result of the prevalence of parasympathetic over sympathetic activity, metabolic heat
4
production and body temperature drop at this stage of sleep. These changes in NREM
sleep contribute to postural and motor quiescence (17).
Figure 1.1. Normal sleep architecture during 8 hours of sleep (17).
In a normal young adult, NREM sleep is followed by REM sleep (17). During the REM
stage, low voltage EEG waves occur in addition to rapid movements of the eyes as
detected by electrooculogram (EOG). There is also a very low level of muscle tone
(except diaphragm, extraocular and sphincter muscles), as detected by electromyogram
(EMG);(17) .The most important features of REM are variability in heart rate, arterial
blood pressure and irregularities in breathing.
1.3 Objective Sleep Measurements
Studies examining sleep have used both objective and subjective methods to identify
sleep patterns and disturbances. Polysomnography and actigraphy are objective
methods, most commmonly used to assess sleep (18). Self-reported sleep characteritics
and validated sleep questionnaires are the most commonly used subjective measures of
sleep (18).
Polysomnography (PSG) is the gold standard method used to measure sleep patterns
and disturbances (18). It involves the simultaneous gathering, analyzing and interpreting
of physiologic signals during sleep, including the electroencephalograph (EEG),
electromyogram (EMG), electrooculogram (EOG), electrocardiogram (ECG) and
respiratory signals (19). Individuals undergoing PSG are typically required to spend a
night in a sleep laboratory, although over the last decade advances in technology have
made it possible to monitor sleep by PSG at home (18). Irrespective of whether PSG is
5
measured in the laboratory or at home, PSG is costly and the recording process may
disturb routine sleep patterns (18).
Polysomnography (PSG) is also the gold standard method used to assess obstructive
sleep apnoea (20). Obstructive Sleep Apnoea (OSA) is characterised by recurrent
decrease and cessation of respiratory airflow during sleep caused by upper airway
narrowing and collapse. The apnoea-hypopnoea Index (AHI) is commonly used for
estimation of the severity of OSA (20). The AHI is the combined average number of
apnoea episodes (cessation of airflow for at least 10 seconds) and hypopnoea episodes
(reduction in airflow of at least 50% with a decrease in arterial saturation of at least 4%
due to partial airway obstruction) occurring within each hour of sleep (20). The AHI is
used to define OSA as mild (5 < AHI < 15 events/hr), moderate (15 ≤ AHI ≤ 30
events/hr) or severe (AHI > 30 events/hr) (20).
Actigraphy is another common objective method of sleep assessment and the American
Academy of Sleep Medicine (AASM) considers actigraphy to be a useful and non-
invasive method (21). Actigraphy utilises one of the characteristic features of sleep;
relative immobility compared to wakefulness. By assessing mobility, actigraphy records
periods when an individual is asleep or awake (22). Most simply actigraphy assumes
that the wearer is asleep during periods of minimal activity and awake when there is
activity. Actigraphy measures limb movement, usually at the wrist, using a small
accelerometer embedded in a watch-like device. This device can also be placed on other
body parts such as ankles and hips (21).
The clinical diagnostic value of actigraphy is limited as it cannot detect and distinguish
sleep disorders such as restless leg syndrome (19) . However, actigraphy has a number
of advantages over PSG. For example, actigraphy can monitor and save information
about an individual's sleep over several days or even weeks (18). Moreover, it is
inexpensive and is able to collect sleep pattern information in the participant’s natural
sleep environment (18). This is particularly useful for those who do not tolerate sleeping
in a laboratory setting, such as some insomnia patients and small children in whom it is
often not possible to obtain self-reported behaviour (23). By providing an opportunity
for individuals to adhere more closely to their customary sleep and wake time,
actigraphy can more accurately estimate typical sleep patterns. Actigraphy is also a
practical method for sleep assessment in large-scale population studies (22). High
correlations have been reported between PSG and actigraphy-defined sleep, with an
6
overall agreement of over 90% in normal participants (24, 25). However, actigraphy is
known to overestimate the sleep period and underestimate wake time when compared to
PSG (18). This discrepancy is due to the different method used to measure sleep onset
in PSG and actigraphy. Specifically, sleep onset in PSG is signalled by changes in brain
electrical activity patterns, while sleep onset in actigraphy is signalled by immobility of
the participant (22) In actigraphy, periods of immobile wakefulness would be scored as
sleep, thus overestimating sleep time.
1.4 Self-report Sleep Measurement
Self-report of sleep by questionnaires is a subjective method of sleep measurement
often used in epidemiological studies. Previous studies estimating the agreement
between subjective and objective measures of sleep duration have reported correlations
of 0.31 to 0.63 between the two (26-28). A study of 2080 adults aged 18-74 years found
the correlation between self-reported sleep and actigraphy sleep duration was 0.43 (29).
Using self-report, the average sleep duration was 7.85 ± 1.12 hours versus 6.74 ± 1.02
hours using actigraphy, with lower correlations in males and those younger than 45
years (29).
A limitation of many previous studies is that correlations are not adequate for estimating
agreement, as a correlation only measures the strength of a relationship between two
variables and not the agreement between them (30). The few studies that have evaluated
sleep duration assessed by self-report and actigraphy using statistical methods other
than correlation, have reported a poor agreement between subjective and objective
methods of sleep measurement (31, 32).
1.4.1 Pittsburgh Sleep Quality Index Questionnaire
The Pittsburgh Sleep Quality Index (PSQI) Questionnaire is a standardized and
validated questionnaire used to assess the quality of sleep (33). The PSQI questionnaire
(see Section 2.8.2.1) can reliably categorise individuals as either “good” or “poor”
sleepers (33). The questionnaire includes 19 questions which subjectively evaluate
sleep. Sleep is then scored for the following categories: sleep quality (individual sleep
quality rating), sleep latency (length of time it takes to fully fall asleep), sleep duration,
habitual sleep efficiency, sleep disturbance, use of sleeping medication and daytime
dysfunction such as having trouble staying awake while driving. Participants are
required to complete the PSQI Questionnaire retrospectively by considering and
7
recalling their sleep in the past month (33). By adding each single score, the overall
score is determined if greater than five the respondent is categorized as being a poor
sleeper (33). The PSQI Questionnaire is recognised as a convenient and practical
method of assessing sleep quality (34).
Although the PSQI Questionnaire has been suggested as a valid tool for sleep quality
assessment, it has poor correlation with the sleep quality index as derived by PSG or by
actigraphy. For example, in a sample of 112 adults (53 young adults with a mean age of
23 years and 59 older adults with a mean age of 53 years), Grandner et al (35) showed a
poor correlation between PSQI score and actigraphy derived sleep variables. The poor
agreement between PSQI and PSG is likely due to the retrospective nature of the PSQI
Questionnaire (33). It is also possible that the average of multi-night recordings of PSG,
taken over the same period of time as PSQI, might improve the relationship between the
two measures (35).
1.4.2 Berlin Questionnaire
The Berlin Questionnaire was designed to identify individuals at risk for OSA (36). The
questionnaire has ten questions organised into three categories: snoring, sleepiness
/tiredness and obesity/hypertension (see Section 2.8.2.2). Each category is evaluated
and scored indiviually. If at least two of the categories are positive, an individual is
considered to be high risk for OSA (36). The Berlin Questionnaire has been validated in
adult populations (37, 38).
Previous studies compared the agreement rate between the Berlin Questionnaire score
and AHI using PSG. Comparing Berlin scores to mild OSA measured by PSG, the
questionnaire was found to have 83% sensitivity and 22% specificity (39). However,
when different AHI cut-offs were considered, the questionnaire was found to be limited
for mild and moderate OSA (39). Thus it has been suggested that the Berlin
Questionnaire may be most effective for predicting severe OSA rather than mild or
moderate OSA (40). Despite this, previous studies have recommended the Berlin
Questionnaire as a screening tool in the general population in view of its good
sensitivity and specificity. In a study of 240 adults with mean age of 48 years (50%
females),the sensitivity and specificity of the Questionnaire were 76.9% and 77.6% to
predict AHI ≥ 15 (40).
8
1.5 Association Between Short Sleep Duration and Weight Gain
1.5.1 Cross-sectional and Longitudinal Studies in Children and
Adolescents
A number of cross-sectional paediatric population studies (Table 1.1) have shown an
inverse association between sleep duration and increased risk of obesity (41-51). All of
the studies used age-adjusted thresholds of BMI for obesity but they used different cut-
offs for average sleep. The cut-off used for average sleep varied by age group according
to the sleep requirements of the age group. For example the average sleep duration
selected for children aged 5 years and 5-6 years were 11 hours and 11.5 hours
respectively. For children aged 7 and 9 years, it was 11 hours and 10 hours.
Although the definition of average sleep duration differed, the findings were similar.
For example, Locard et al (41) assessed 1031 five-year olds and found the odds ratio for
obesity was 1.4 for sleep duration less than 11 hours, compared to those with longer
sleep duration. Von Kries et al (42) showed in a study of 68625 children aged 5-6 years,
that the prevalence of obesity was 5.4% (95% CI: 4.1, 7.0) in those sleeping equal or
less than 10 hours, while this rate was 2.1% (95% CI: 1.5, 2.9) in those sleeping equal
or more than 11.5 hours. By using the wide range of confounders, the study concluded
that the effects of sleep duration on obesity are not influenced by confounders. In a
sample of 8274 children aged 6-7, Sekine et al (46) reported a 2.89 fold (95% CI: 1.61,
5.05) higher propensity for obesity for those sleeping less than 8 hours after adjusting
for confounders including age, gender and parental obesity. Padez et al (47) reported a
decreased risk of overweight (OR 0.44, 95%CI: 0.38, 0.49) and obesity (OR 0.39,
95%CI: 0.35, 0.42) with an increase in the number of hour of sleep in 4511 children
aged 7-9 years. It is noteworthy that most of the studies used one parental question to
estimate the sleep duration of the children and as such the estimated average sleep
duration might not be accurate. Additionally, the parental sleep question varied among
the studies; while some studies asked about sleep both during weekdays/weekend (42)
(47),other studies collected sleep duration irrespective of weekdays/weekends (41).
Lastly very few of these studies took into account important potential confounding
factors such as physical activity and some did not adjust for age and gender.
9
A number of cross-sectional studies have evaluated the association between sleep
duration and weight in adolescence. Two of these studies used either wrist actigraphy or
hip accelometer. By using wrist actigraphy in 383 adolescents, Gupta et al (48) reported
that for every hour reduction in sleep duration, the odds of obesity increased by 5-fold.
While other studies in adolescents reported an association between short sleep duration
and obesity risk (49, 50), one study on 4486 American teens reported no association in
girls but a significant association between short sleep duration and the incidence of
being overweight in boys (51). Other adolescent studies used a single self-report
question, which is a poor measure of sleep duration. The confounding effect of puberty,
which may influence obesity risk was not evaluated in any of the adolescent studies
(52).
10
Table 1.1. Cross-sectional Studies of Sleep Duration and Weight Gain in Children and Adolescents [modified (53)]
First Author (Year) Sample Size Age range or Mean Subjective Sleep Measure/Potential
confounders
Findings
Locard (1992)(41) 1,031 5 One question to parent/age, ethnicity,
breastfeeding, TV viewing, maternal
age, socioeconomic, single parent,
parental obesity
Obesity odds of 1.4 for sleep duration
< 11 hrs).
Von Kries (2002) (42) 6,862 5-6 Parental questions/age, breastfeeding,
diet, snacking, TV viewing,
socioeconomic, single parent, PA
Obesity odds of 2.2 for sleep duration
< 10.5 hrs.
Chaput (2006)(43) 422 5-10 One question to parent/ age, sex, TV
viewing breastfeeding, PA, regular
breakfast, parental obesity
Obesity odds of 3.4 for sleep duration
≤ 10 hrs).
Ben Slama (2002)(44) 167 6-10 One question to parent/ None Positive association between short
sleep duration and obesity risk (58%
obese participants had sleep duration
less than 8 hours).
Giugliano (2004)(45) 165 6-10 One question to parent/None Obese children had 30 minutes
shorter sleep compared to normal-
weight children.
Sekine (2002)(46) 8274 6-7 One question to parent/age, sex, TV
viewing, PA
Obesity ORs, 1.49 and 1,89 for sleep
durations 9-10 hrs and 8-9 hrs
respectively.
Padez (2005)(47) 4,511 7-9 Parental questions/age, sex Positive association between short
sleep duration and obesity risk.
11
First Author (Year) Sample Size Age range or Mean Subjective Sleep Measure/Potential
confounders
Findings
Chen (2006)(50) 656 13-18 One question to child/age, sex Inverse association between obesity
risk and having sleep duration of at
least 6-8 hours.
Knutson (2005)(51) 4,486 17 One question to child/age, sex,
ethnicity, PA, socioeconomic
Association between short sleep
duration and higher BMI z-score and
overweight risk in boys only.
First Author (Year) Sample Size Age range or Mean Objective Sleep Measure/Potential
confounders
Findings
Gupta (2002)(48) 383 11-16 Actigraphy (24 hours)/ age, sex,
puberty status
5-fold increase in obesity OR for
every hour reduction in sleep
duration.
Benefice Senegal (2004)(49) 40 13-14 Accelometer (72-96 hours)/ age, sex,
PA, height
Inverse association between short
sleep duration and BMI.
PA: Physical Activity
12
Consistent with cross-sectional studies, longitudinal studies have also found an inverse
association between sleep duration and obesity in children (54-57). Taveras et al (54)
found that at age 3, the odds of being overweight were 2.04 (95%CI: 1.07, 3.91) in
those sleeping less than 12 hours during infancy. Their study adjusted for a wide range
of infant confounders (birth weight, breast feeding duration and daily physical activity)
and maternal confounders (education, income, pre pregnancy BMI). However, a major
limitation of the study was that sleep duration was estimated using a maternal question
(asking about infant’s sleep duration during the last month) instead of using objective
method of sleep measurement or diaries (54). In a study of 1441 children aged 3-12,
Snell et al found that the association between sleep and overweight was moderated by
age (55). According to this study, at 3-8 years, an additional hour of sleep was
associated with reduced probability of being overweight (ẞ = - 0.061, P < 0.01) but this
association did not continue in older children (55). The use of parental reports and lack
of adjustment for confounders are both limitations of most of the previous longitudinal
studies undertaken in children and adolescents and some did not assess weight at the
time of sleep measurement (56, 57). As growth rate is not constant during childhood, it
is important to measure weight at the time of sleep assessment in order to have a better
estimation of the role of sleep on body weight.
1.5.2 Cross-sectional and Longitudinal Studies in Adults
In contrast to the link observed between sleep and obesity in children, the results of
cross-sectional studies of sleep and weight gain in adults are varied (Table 1.2). Studies
that have used questionnaires to ascertain sleep have reported a positive association
between short sleep duration and weight gain, a U-shaped relationship or no association.
A limitation of all of these large studies is that they were designed to evaluate the
effects of different behaviours on health outcome and not to only assess sleep.
Importantly, while these studies adjusted for gender, the effects of other important sleep
confounders were generally ignored. Lastly, these studies have used different cut offs to
define sleep duration categories.
One of the largest cross-sectional studies was conducted by the American Cancer
Society (58). The study included over 1 million adults with cancer, ranging in age from
33-102 years. They found that women had a BMI that was 1.39 kg/m2 higher if they had
4 hours of sleep compared to 7 hours (58). In men the BMI was 0.57 kg/m2 higher if
13
they had 4 hours of sleep compared to 7 hours (58). Similarly, in a study of 13742
adults aged ≥ 20 years, short sleepers (≤ 6 hours) were more likely to be obese (OR
1.10, 95%CI: 1.03, 1.16) compared to normal/long sleepers (≥ 7 hours) after adjusting
for a wide range of confounders (including age, gender, employment, smoking and
alcohol use) (59). In contrast, in a study of 104,000 adults aged 40-79 years, Tamakoshi
et al (60) reported a significant association between less sleep duration and reduced
BMI. For those with a sleep duration of ≤ 4, 5 and 7 hours the mean BMI were 22.2,
22.6 and 22.7 kg/m2 respectively (60). This is the only study reporting an association
between short sleep duration and reduced BMI.
There are several reports of a U-shaped association between sleep duration and BMI in
adults. The FIN-D2D Study of 2770 adults aged 45-74 reported higher prevalence of
obesity in men with either short sleep duration (≤ 6 hours) or long sleep duration ( ≥ 8
hours)(61). Similarly, Liu et al found a higher rate of obesity (OR 1.32, 95%CI: 1.21,
1.43) in short sleeper (≤ 6 hours) and those sleeping more than ten hours (OR 1.60,
95%CI: 1.34, 1.93) (62). Interestingly, in a study of 1486 adults with mean age of 70
years, Gottlieb et al (63) showed no association between sleep duration and BMI. In a
study of 191 women aged 80-92 years using actigraphy for sleep estimation, no
association was found between sleep duration and BMI (64)
14
Table 1.2. Cross-sectional Studies of Associations Between Sleep Duration and
Weight in Adults [modified (65)]
Author (Year) Sample
Size
Age range
or mean
Sleep measure Association between sleep duration
and BMI
Kripke et
al(2002)(58)
1,116,936
33-102
One sleep
question
Inverse association
Ford et al
(2014) (59)
13742 ≥ 20 One
sleep question
Inverse association
Gottlieb et al
(2005)(63)
1486 53-93 One sleep
question
No association
Chaput et al
(2007)(66)
90 50-70 One sleep
question
No association
Tuomilehto et
al (2008) (61)
2770 45-74 One sleep
question
U-shaped association
Liu et al (2013)
(62)
54269 ≥ 45 One sleep
question
U-shaped
Kim et al (2015)
(64)
669 35-49 5 days actigraphy No
The results of longitudinal adult studies have also provided inconsistent data, with
reports of negative associations, no associations or U-shaped associations between sleep
duration and weight gain. According to the Nurses’ Health Study of 68,183 women, the
relative risk of a 1.5 kg weight gain over 16 years in those sleeping ≤ 5 hours and 6
hours was 1.28 (95%CI: 1.15, 1.42) and 1.10 (95%CI: 1.04, 1.17) respectively (67). In
another longitudinal study of 3803 male adults aged 40-59 years, a significant increase
in BMI was reported in those sleeping less than 5 hours (ẞ = 0.015 kg/m2) over 4 years
(68).
There have also been reports of longitudinal studies showing no association between
sleep duration and BMI (69, 70). In NHANES I, 3,208 adults aged 32-49 years showed
no significant association (ẞ =- 0.053, P = 0.27) after 10 years of follow-up (69). The
only longitudinal study that has used actigraphy to monitor sleep included 612 adults
aged 33-45 years and reported no association between sleep and BMI after 5 years (70).
Similar to the findings in cross-sectional studies, some longitudinal studies have shown
a U–shaped association between sleep duration and BMI (71) and the association varies
15
by sex and age (72, 73). The Quebec family study of 276 adults aged 21-64 years over 6
years showed those who slept < 6 hours or > 9 hours gained 1.84 and 1.49 kg more
weight respectively, however other studies with U-shaped associations found this
association to be evident only in females (72) or those aged < 40 (73).
To summarize, there is no consistent evidence of an association between short sleep
duration and weight gain in adults. This may be due, in part, to lack of a standard
definition for short and long sleepers. Previous research studies have also used a single
self-reported question or sleep questionnaire to assess sleep duration estimation instead
of using actigraphy over several days. Another reason may be related to differences in
study populations. For example, some studies recruited participants from the general
population, whilst others recruited nurses or cancer survivors to evaluate the association
between sleep and obesity. Other potentially important confounders including
racial/ethnic influences which are often not taken into consideration.
1.5.3 Physiological Basis for an Association Between Sleep
Restriction and Obesity
Several mechanisms could underlie a possible association between sleep restriction,
body weight and obesity. Experimental physiological studies suggest the changes in
appetite-regulating hormones (leptin and ghrelin), sympathetic nervous activity and
energy intake that occur as a result of sleep restriction may induce obesity (74).
Leptin and ghrelin are appetite-regulating hormones which can be influenced by sleep
restriction (75). Leptin is secreted by adipose tissue and leptin levels increase following
a meal and during the night. Ghrelin is produced in the stomach. Both hormones exert
their effects in the arcuate nucleus of the hypothalamus through
propiomelanocortin/cocaine and amphetamine–regulated transcript (POMC/CART) and
Neuropeptide Y/agouti-related protein (NPY/AgRP) neurons. Leptin decreases hunger
and induces satiety (75). In contrast, ghrelin stimulates appetite, prolonging postprandial
glucose responses while stimulating growth hormone (GH) release.
Experimental sleep studies have shown changes in these hormones following sleep
restriction (75, 76). Following a short period of sleep restriction (2 consecutive nights of
4 hours in bed) in 12 healthy young men, leptin decreased by 18% while ghrelin
increased by 28% with an increase in appetite with a preference for salty or high
carbohydrate foods (74). Another study of eight young adult males undergoing sleep
16
restriction of 4 hours sleep per night for seven days showed a reduction in leptin level
by 33% after sleep restriction (76). In contrast, other studies have shown no change in
leptin and ghrelin levels (77, 78). A study of 11 healthy young adults (six males with a
mean age of 39 ± 5 years) that completed two 14-day periods (with three months
interval) in sleep laboratory found no change in leptin or ghrelin levels following sleep
restriction (5.5 hours) period compared to normal sleep duration (8.5 hours) period (77).
Additionally, no changes were found in the levels of leptin and ghrelin in a sample of 17
healthy young adults following sleep restriction of two-thirds of normal sleep time for
8 nights (78). All of these studies are limited by a small sample size and the inclusion of
a young healthy adult population. As the body metabolism varies through different
stages of life, it is unclear if similar associations exist in other age groups. Another
possible factor is differences in the duration of sleep restriction (short period of sleep
restriction/ long period of sleep restriction), as it has an impact on leptin and ghrelin
levels and weight gain (77).
Another possible mechanism linking sleep restriction to weight gain is through the
sympathetic nervous activity (79). The sympathetic system produces epinephrine and
norepinephrine. During sleep, vagal tone increases and sympathetic nervous activity
decreases. Experimental sleep studies show an increase in urinary epinephrine and
norepinephrine indicating increased sympathetic activity following a 24-h period of
continuous wakefulness in 14 adult males (79).
Sleep restriction can also potentially contribute to weight gain by influencing energy
intake and/or food preference. Previous studies have shown an increase in energy intake
following experimental sleep restriction from 180 to 559 kcal/d (78, 80-82). A similar
association between habitual sleep duration and energy intake has been reported in
observational studies (83). The results of the NANES Study of 5587 adults with a mean
age of 46 years showed the highest energy intake (2201 kcal) in short sleepers (5-6
hours) with higher intake of sugars and total fat compared to normal sleepers ( all P <
0.005)(83). Grander et al reported an inverse association between actigraphy sleep
duration and total energy intake and fat intake in 459 post-menopausal women
monitoring their sleep duration using one week of actigraphy(84). Overall, if the
observed dietary changes, particularly an increase in total energy and fat intake, are
sustained over long- term and are not accompanied by an equivalent increase in energy
expenditure, it could explain the association between short sleep duration and obesity.
17
1.6 Sleep Duration and Glucose Metabolism
1.6.1 Sleep Duration and Glucose Metabolism
During normal sleep, physiological changes occur in glucose metabolism in order to
maintain glucose level at the same level as wakefulness. A number of studies have
reported changes in sleep duration can alter glucose and insulin levels which may lead
to increased risk of diabetes (76, 77).
Blood glucose levels must be tightly regulated in the human body to prevent
hypoglycaemia and hyperglycaemia. During nocturnal sleep and overnight fasting, a
number of mechanisms keep glucose levels stable. During the first half of the sleep
period, glucose metabolism slows, as cerebral glucose uptake rates decrease by 40% and
peripheral glucose utilization rates decrease (85, 86). This reduction in glucose
metabolism during sleep occurs in brain regions such as the frontal, temporal and
occipital cortex and the anterior/dorsomedial thalamus (87).
Two hormones that regulate glucose during sleep are growth hormone and cortisol.
Growth hormone levels begin to rise at the onset of sleep and peak during slow wave
sleep (88). In contrast, cortisol levels rise during the second half of the sleep,
predominantly in REM sleep.
During the NREM phase of sleep, a reduction in brain glucose metabolism, reduced
muscle tone and anti-insulin effects of growth hormone are responsible for a drop in
glucose utilisation (75). During REM sleep, glucose utilization increases although it is
lower compared to the wakeful state.
1.6.2 Sleep Duration, Glucose Dysregulation and Risk of Diabetes
While a large number of studies have assessed the impact of sleep on glucose regulation
and metabolic dysregulation, the focus of these studies has shifted recently. With the
increase in the prevalence of obesity, studies have shifted towards examining the impact
of sleep duration, particularly the impact of sleep restriction on metabolic regulation
(12). The results of experimental, cross-sectional and longitudinal studies suggest that
sleep restriction plays a role in altering glucose metabolism (89-91).
18
Most of the recent experimental studies are focused on changes in glucose metabolism
in response to acute sleep restriction (92, 93). Generally these studies have shown
increased insulin levels and insulin resistance (lowered levels of response to insulin in
the body tissue) following acute sleep restriction (89, 92, 93). A small intervention trial
of 7 healthy adults (6 males, 1 female) with a mean age of 23 years showed sleep
restriction of four hours of sleep per night for four nights led to insulin resistance and
increased insulin concentration (measured by an intravenous glucose tolerance test and
cellular insulin sensitivity index in adipocyte subcutaneous fat biopsy) of up to 300%
(92). Another study of 21 healthy male adolescents with a mean age of 16 years showed
sleep restriction of four hours of sleep per night for three nights led to 59% increase in
fasting insulin without any changes in glucose (94).
Although the majority of these acute sleep restriction studies reported changes in insulin
sensitivity, these changes might be due to a physiologic response to acute intensive
temporary sleep restriction (four hours of sleep for 3-4 nights). Therefore further sleep
studies of chronic sleep loss that are more clinically relevant are required. Furthermore,
the majority of these studies have included only young male participants that is another
potential limitation of previous studies.
A number of cross-sectional and longitudinal studies have evaluated the association
between sleep duration and risk of diabetes. A study of 740 adults aged 21-64 years
who undertook 75 g Oral Glucose Tolerance Test (OGTT) found a 2-fold increased
odds ratio for diabetes in those who slept 5-6 hours, after adjusting for potential
confounders including BMI, waist circumference and physical activity (95). The
NHANES longitudinal study of 8,992 adults aged 32-86 years showed those adults who
only slept five hours a night had a 1.4 greater odds ratio (95%CI: 1.03, 2.09) for
diabetes (estimated by self-reported or physician-diagnosed diabetes) over a 10 year
follow-up period (96). This study also reported that 9 hours or more of sleep increased
the odds ratio to 1.5 (95%CI: 1.06, 2.18)(96). Similarly, a cohort study of 70,000 nurses
indicated an association between both short (≤ 5 hours) and long (≥ 9 hours) sleep
duration with type 2 diabetes. The relative risk of type 2 diabetes (assessed through self-
questionnaire) was 1.34 (95%CI: 1.04, 1.72) for short sleepers and 1.35 for long
sleepers (95%CI: 1.04, 1.75) (97). According to the Massachusetts Male Aging Study
(MMAS) which monitored men with no history of diabetes for 15 years, the risk of
diabetes (assessed through self-questionnaire) increased 2-fold (95%CI, 1.05, 3.48) in
those sleeping less than 6 hours while this risk increased 3-fold in those sleeping more
19
than 8 hours (95%CI: 1.83, 7.46) (98). A Swedish study following 2,000 people for over
10 years showed a positive association between short duration of sleep (< 5 hours) and
higher incidence of diabetes in men (99). All of these studies used a sleep question for
assessing sleep. Additionally, study participants limited to overweight participants (95)
or females only (97). The confounding effects of abdominal obesity is not considered in
some of these studies (97).
There are several meta-analyses evaluating the association between short sleep duration
and diabetes (100). The meta-analysis of Cappuccio et al (100) was based on seven
studies and showed a relative risk of 1.28 in short sleepers with larger effect in male
short sleepers. In contrast, the meta-analysis based of Holliday et al (101) including 10
studies showed an increased risk of diabetes in short sleepers irrespective of gender.
1.7 Sleep Duration and Other Cardiometabolic Risk Factors
Although most studies have investigated the role of sleep duration on obesity and
glucose metabolism, some studies have also evaluated the role of sleep duration on
other cardiovascular factors such as hypertension, dyslipidaemia and C-Reactive Protein
(CRP).
Studies examining short sleep duration and hypertension suggest the two are linked. A
study of 3,068 men and women aged 50 years evaluated sleep duration by self-report
questionnaire over four years showed that increased risk of incident hypertension in
those having short sleep duration (≤ 5 hours) (102). The participants had no previous
history of hypertension and were not taking any antihypertensive medication. The risk
of hypertension was 1.73 (95%CI: 1.08, 2.76) in men and 1.44 (95%CI: 1.00, 2.07) in
women after adjusting for potential confounders including age, BMI, physical activity
and smoking (102). The US National Health Interview Survey (NHIS) of 56,507 adults
aged 18-85 years showed both short (< 7 hours) and long sleep (> 8 hours) associated
with increased risk of hypertension (103). In a study of 578 adults aged 33-45 years,
Knutson et al (104) reported the odds of hypertension increased by 1.37 (95%CI: 1.05,
1.75) in those with shorter sleep duration after adjusting for age, race and sex
confounders. This is the only study that has evaluated the association between sleep
duration and hypertension in the later period of young adulthood to middle age. All of
these studies are limited as sleep duration was assessed using self-reported responses.
Furthermore, sleep-disordered breathing was not assessed in these studies. As sleep-
disordered breathing might contribute to hypertension better than sleep duration, it is
20
necessary to consider the confounding role of sleep-disordered breathing on the
association between sleep duration and hypertension.
It has been suggested that increased levels of CRP, IL-6 and TNF-α are important risk
factors for atherosclerosis, stroke and cardiovascular diseases. CRP a 120 kDa pentamer
is an important inflammatory marker predominately synthetised in the liver (105). CRP
has a significant role in the pathogenesis of atherosclerosis and a moderate elevation of
CRP in healthy adults is a strong predictor of cardiovascular risk (105). The studies that
have evaluated sleep duration and CRP have shown inconsistent results. The Wisconsin
Study of 907 males with a mean age of 52 years showed no association between sleep
duration and CRP levels after adjustment for potential confounders including age, sex
and BMI (106). In contrast, a study of 4,642 adults with a mean age of 50 years by
Miller et al (107) found a U shape association between sleep duration and CRP levels in
women only. Women with sleep duration < 5 hours or sleep duration > 9 hours had
higher CRP levels compared to those sleeping 7 hours (107).
Dyslipidaemia is an increase in total cholesterol or triglyceride levels with or without a
reduction in high density lipoprotein cholesterol (HDL-C) levels (108). Very limited
studies have evaluated the association between sleep duration and dyslipidaemia with
inconsistent results (109-111). One of these studies was the large epidemiological
Korean National Health and Nutrition Examination Survey (109). This study of 13,609
adults aged 20 and over that adjusted for covariates including age and sex found no
association between short sleep duration (≤ 5 hours) and high low density lipoprotein
cholesterol (LDL-C) or high triglycerides (109). However, a significant association was
found between long sleep duration (≥ 9 hours) and low HDL-C. In a study of 1600
women and 2,300 men aged 20 and over that adjusted for a wide range of confounders,
both short (< 5hours) and long (> 8 hours) sleep was associated with both high
triglyceride and low HDL-C levels in women. In men, a link between long sleep
duration and high LDL -C was the only observed association (110).
To summarize, the results of previous studies suggest the role of sleep duration on
cardiometabolic risk factors is inconsistent. It is noteworthy that the majority of these
studies have been carried out in middle-aged individuals and have used self-reported
sleep duration questionnaires instead of an objective sleep measure. Residual
confounding due to unmeasured variables may also contribute to the inconsistent
findings of previous studies.
21
1.8 Sleep Quality and Cardiometabolic Risk Factors
The majority of previous studies have analyzed sleep by duration and its effect on
cardiometabolic risk factors. However, as sleep has both quantitative and qualitative
aspects, to better understand the link between sleep and cardiometabolic risk factors the
impact of both quantity and quality of sleep needs to be investigated (112). In this
regard, a number of studies have evaluated the impact of sleep quality on
cardiometabolic risk factors particularly on body weight and hypertension (110-119).
However, according to our current knowledge there is little epidemiological data on
sleep quality and other cardiometabolic risk factors.
Studies examining the association between sleep quality and increased body weight
show varied results (112-114). The BiDirect Study of 753 adults aged 35-65 years
found an association between poor sleep quality (assessed by PSQI Questionnaire),
obesity and fat mass after adjustment for sociodemographic and lifestyle factors (115).
Rahe et al (115) also showed an association between poor sleep quality and general
obesity and fat mass after adjustment for sociodemographic and lifestyle factors
including education, job status,smoking and physical activity. However, further
adjustment for depression and comorbidities (including hypertension,diabetes and
myocardial infarction) attenuated the asociation (115). Despite the association between
poor sleep quality and obesity, the results of other studies suggest there is no
relationship between sleep quality and other cardiometabolic risk factors (10, 116, 118).
A study of 927 women of reproductive age (16-40 years) found no relationship between
poor sleep quality (assessed by PSQI Questionnaire) and BMI after adjusting for age,
race and depressive symtoms (117). Another study of 660 adults aged over 90 years
found 22% of adults had poor sleep quality (119). Using the World Health Organization
(WHO) recommendation for Asian populations and classifying particpants into
underweight (BMI < 18.5), normal weight (18.5 ≤ BMI < 23.0), overweight (23.0 ≤
BMI ≤ 27.5) and obese(BMI > 27.5 kg/m2), this study found no significant difference
among BMI categories and prevalence of poor sleep quality (using PSQI Questionnaire)
after adjusting for age,gender and lifestyle factors including smoking,alcohol drinking
and exercise habits (119). The discrepancies between studies may be due to differences
in study populations including age, limitation of criteria used to define poor sleep
22
quality (PSQI Questionnaire instead of using objective method of sleep quality
assessment) or use of different covariates.
Most of the studies have investigated the role of sleep quality on hypertension. In a
study of 133 adults with a mean age of 43 years, Erden et al (120) showed reducing
blood pressure at night was inversely correlated with sleep quality. Moreover, previous
studies reported an association between poor sleep quality and hypertension (121, 122).
In a study of 250 normal weight adults, a significant difference in the prevalence of
hypertension among poor and good sleep quality groups (87.1% vs 35.1%) was found
(121). Similarly, a study of 9,404 adults with a mean age of 52 years showed a 2.3-fold
greater risk of hypertension in those with poor sleep quality (122).
1.9 Sleep-disordered Breathing and Cardiometabolic Risk Factors in
Adults
In addition to sleep duration and quality, sleep disorders such as Sleep-Disordered
Breathing (SDB) are also thought to impact on cardiometabolic health (123). It is
important to note that the causes of sleep-disordered breathing in children are different
compared to adults. Adenotonsillar hypertrophy and anatomical or neuromuscular
factors play a significant role in SDB in children in contrast to adults (124).
Neurocognitive dysfunction and abnormal behaviours are the most common and
important complications of SDB, meanwhile metabolic complications of SDB are less
common in SDB children compared to adults (125). Therefore, the focus of the current
section is restricted to associations between SDB and cardiometabolic risk factors in
adults.
1.9.1 Sleep-disordered Breathing in the General Population
SDB refers to respiratory disorders characterized by abnormalities in respiration during
sleep (126). This comprises a wide spectrum of sleep-related breathing disorders
ranging from the mild form of snoring to the most serve form of obstructive sleep
apnoea (OSA) where upper airway respiratory collapse results in decreased oxygen
saturation in the blood (126, 127). The Apnoea Hyponea Index (AHI, see Section 1.3) is
used to indicate the severity of obstruction. These repetitive hypoxic events result in
23
arousal from sleep, sleep fragmentation and cause excessive daytime sleepiness and
fatigue.
SDB has been recognised as a common problem in the general population, with a
prevalence of 56% in the United States, 31% in Western Europe and 23% in Japan
(128). A study of 741 adult men suggests the prevalence of SDB increases with age
from 3.2% in participants aged 20-44 to 18.1% in those aged >61 years (129).
By using the minimal diagnostic criteria for the sleep apnoea syndrome (an apnoea-
hypopnea score of 5 or higher and daytime hypersomnolence), 4% of middle-aged work
force males and 2% of females were estimated to have sleep apnoea in the developed
world in 1993 (130). Recent data have shown the prevalence rate over the following two
decades has increased to 23% in females and 50% in males in the middle-aged to
elderly (131). In males and females aged 30-49 years, the prevalence of OSA is
estimated at 10% and 3% respectively (132). Although OSA is known to be the disorder
of middle and old-age groups, recent studies suggest OSA is also evident in young
populations (133, 134). A study of 916 Chilean college students reported that
approximately 8% had OSA (135). The prevalence of OSA in reproductive aged women
is 0.6-15%, with most undiagnosed (136). A study of a Chinese population with mean
age of 20 years reported OSA at a rate of 13% (137). In an Australian longitudinal
cohort study on 13423 adult men (aged 10-55 years), the prevalence of self-reported
professional-diagnosed OSA was 2.2% in those aged 18-25 years (138).
1.9.2 Sleep-disordered Breathing and Obesity
Obesity is an important risk factor for OSA (139). Obesity induces fat deposits in upper
airway muscles and lumen and reduces tracheal traction contributing to air flow
obstruction and subsequent OSA (139). While most of the previous studies suggest a
one-way cause and effect relationship, with obesity contributing to the pathogenesis of
OSA, current evidence indicates that OSA itself can contribute to or exacerbate obesity,
thus creating a bidirectional link between OSA and adiposity (139, 140;Wosu, 2014
#175)
There are two main mechanisms responsible for increased body weight in OSA. One of
these is sleep restriction (139). In OSA, disturbed respiratory gas exchange due to the
upper airway collapse contributes to sleep fragmentation and eventually sleep
restriction. This results in daytime tiredness and reduced physical activity leading to
24
weight gain (Figure 1.2). The other mechanism responsible for obesity in OSA is
variation in appetite regulatory hormones including leptin and ghrelin (139). Previous
studies have shown increased leptin levels in OSA participants, but others have reported
a reduction in leptin and increase in ghrelin levels (141-143). The changes in leptin and
ghrelin contribute to an increase in appetite and caloric intake and exacerbate obesity in
OSA participants (139). As oxidative stress and Reactive Oxygen Species (ROS) can
change metabolism in adipose tissue, they also can contribute to obesity in OSA (144).
Figure 1 .2. Cycle of obesity and OSA and its potential mechanisms
UAW, Upper Air Way (139).
A number of studies have shown a positive association between obesity and OSA (139).
The Wisconsin Sleep Cohort Study of 690 US adults (mean age of 46 years,56% males)
followed up during 4 years showed weight gain was associated with increased
prevalence of OSA, with a 10% weight gain as the severity of OSA increased to 32%
(95%CI, 20, 45) after adjustment for sex, baseline age and BMI and smoking
habit(145). As 27% of the participants in this study had incomplete follow-up, the
results can be due to participant bias. Furthermore, the average weight change was
within ± 20%, thus the effect of large weight change in OSA cannot be determined.
Lastly as the participants were initially overweight, the results of the study cannot be
generalised to normal-weight participants. In a study of 704 participants attending
primary health care clinic (aged 14 to 81 years, 58% males) Mahboub et al (146)
showed 70% of those at high risk of OSA were obese. Additionally, a cross-sectional
study of 916 Chilean college students showed a 9.9-fold (95%CI: 4.42, 22.45) higher
risk of obesity in those with OSA (135). A major limitation of the latter study is it used
25
the Berlin questionnaire instead of performing an overnight sleep study to assess OSA.
An Australian longitudinal cohort study of 13423 adult men (aged 10-55 years) also
found a significant association between OSA and higher BMI; the risk of obesity and
overweight was 3.6-fold and 1.7-fold higher respectively, in OSA participants (138). A
recent meta-analysis study of 2966 participants from 19 studies showed that weight and
BMI has no correlation with OSA after checking for publication bias (by removing
small studies)(147).
Continuous Positive Airway Pressure (CPAP) is the standard treatment for OSA.
Therefore, it theoretically could reduce BMI (148). This is supported according to the
results of two non-randomised studies (149, 150). In a study of 31 obese participants
(29 males, mean age of 52.2 ±3.3 years), visceral fat accumulation and subcutaneous fat
accumulation decreased significantly following six months treatment with CPAP
without any change in body weight, fasting insulin and cortisol levels(149). Another
retrospective cohort study of 32 overweight/obese participants (70% males, mean age of
59 years) showed a facilitating role of CPAP on weight loss (BMI change = −1.6 ± 0.7,
P < 0.05) in the CPAP group. (150). However, both of these studies have limitations
including a non-randomised design and small sample size. Additionally, it is unclear
whether CPAP can contribute to weight loss in severe obese or extreme obese
participants. Interestingly, these findings were not replicated by more recent studies
(151-153). The results of a retrospective study of 228 middle-aged adults (64% males)
showed the BMI of the CPAP and control groups did not differ following one year of
CPAP treatment (P = 0.32) (151). The limitations of the study were small sample size
and its retrospective design. In a study of 20 obese males with a mean age of 60, Garcia
et al(153) showed an increase in body weight (109.6 ± 5.4 0.04 vs. 108 ± 5.3 kg, P =
0.04), BMI (37.1 ± 1.80 vs. 36.5 ± 1.8, P = 0.06) and HOMA-IR index (5.9 ± 1 vs 7.5 ±
1.2, P = 0.04) following CPAP treatment. However, the study was initially designed to
evaluate the effect of CPAP on insulin and not BMI. The study is also limited in
comparing the BMI changes in the CPAP group and the non-interventional control
group due to the lack of a control group. In a recent meta-analysis of 25 randomised
trials of 3181 middle-aged participants (83.9% males) using CPAP at least for a month
(median three months), there was an increase in BMI with CPAP treatment group (154).
The calculated BMI post-intervention minus pre-intervention in the control group was –
0.018 ± 0.243 kg/m2 and 0.134 ± 0.27 kg/m2 in CPAP group (154). The increased BMI
26
in the CPAP treatment group was not influenced by age, gender, baseline BMI, OSA
severity and CPAP compliance.
In summary, the results of previous studies were inconsistent regarding the role of
CPAP on body weight. This is due to the study design (retrospective), variation in
CPAP therapy duration or the lack of having control group.
1.9.3 Sleep-disordered Breathing and Glucose Metabolism
A number of studies have shown the presence and/or severity of OSA is associated with
glucose metabolism and increased risk of diabetes or poor diabetic control (155, 156).
Studies of snoring and impaired glucose tolerance suggest the two are linked (157, 158).
The Nurses' Health Study cohort of 69,852 healthy female nurses aged 40-65 years with
a history of snoring, found the risk of diabetes was increased 2.02-fold (95%CI: 1.91,
2.66) over a 10 year follow up following adjustment for age and BMI (157). There are
several limitations of this study including the fact it only included females and it did not
account for the confounding role of abdominal obesity in the analyses. Similarly,
Elmasry et al in a study of 2504 men aged 30-69 showed that 5.4 % of snorers
developed diabetes over 10 years compared to 2.4% of those without snoring (P <
0.001) (158). Limitations of this study were that information regarding snoring and
diabetes was obtained through self- questionnaires and the study did not account for the
confounding role of abdominal obesity. As abdominal obesity is a component of the
metabolic syndrome and plays a vital role in the pathogenesis of insulin resistance, it is
important to consider the role of abdominal obesity when examining the associations
between diabetes and OSA (159).
As obesity plays an important role in both SDB and glucose metabolism, it is important
to consider obesity as a confounder when evaluating the association (155). The studies
controlling for obesity also reported an association between snoring and risk of diabetes
(155, 157, 158).
A number of studies have shown a positive association between OSA and diabetes (155,
156, 160). A study of 544 non-diabetic adults (85% males with a mean age of 60 years)
showed an independent association between OSA and incident diabetes (hazard ratio
per quartile 1.43; 95%CI, 1.10, 1.86) after adjusting for age, gender, baseline glucose
level, BMI and waist circumference over a 2.7 year follow up (160). The Sleep Heart
Health Study of 2,656 with a mean age of 68 years showed that the odds of glucose
27
intolerance (using oral glucose tolerance test) were 1.27 (95%CI, 0.98, 1.64) in
moderate OSA and 1.46 (95%CI, 1.09, 1.97) in severe OSA (155). Marshal et al
showed higher incidence of diabetes in 399 adults (aged 40-65 years) with moderate to
severe OSA (OR 13.45, 95%CI: 1.59, 114.11) after adjustment for confounders (age,
sex, BMI and BP) and within four years of follow-up (161). However, the results of
these studies are limited by several factors. Firstly, the populations comprised mainly
obese males. Secondly, the effect of residual confounding such as, for example, the role
of family history of diabetes has not been considered in all of these studies.
It has been suggested that insulin resistance can be influenced by SDB (155, 156, 162).
In this regard, a study of 270 non-diabetic subjects attending sleep clinic for OSA
evaluation showed a higher incidence of insulin resistance in OSA participants,
suggesting an independent role of OSA on insulin resistance (162, 163). A study of 400
non-diabetic Swedish women aged 20-70 also showed a relationship between OSA and
insulin resistance suggesting OSA severity can influence insulin resistance (163). The
study showed an association between AHI and increased fasting and 2-h insulin levels
(95% CI: 0.14, 0.99 and 95% CI: 0.28, 6.47) respectively) after adjusting for
confounders (163). The study used insulin sensitivity index derived from oral glucose
tolerance test instead of using gold standard method of euglycaemic insulin clamp
technique for insulin sensitivity evaluation.
Although the exact physiological mechanisms of the association between SDB and
glucose metabolism remain unknown, previous studies has suggested the roles of sleep
fragmentation, recurrent hypoxia, oxidative stress and sympathetic over activation
(Figure 1.3) on insulin resistance (144, 163). Insulin sensitivity is reduced during
hypoxia as a consequence of oxidative stress and cytokines release (164, 165).
Additionally, sympathetic overactivation can increase glycogen breakdown and higher
levels of gluconeogenesis. These changes contribute to impairement in glucose
homeostasis, resulting in higher glucose levels (166).
Despite the relationship between SDB and glucose metabolism, treatment of OSA with
CPAP has shown inconsistent results in improving insulin sensitivity and glucose
metabolism. A study of 50 diabetic participants (60% males with a mean age of 61
years) showed a 0.4% reduction in HbA1c levels (95%CI: 20.7, 20.04, P < 0.05) after
adjustment for age, sex and AHI following 6-months open-label CPAP intervention
(167). Loachimescu et al showed a decreasing trend in the level of HbA1c only in
28
individuals with excellent CPAP compliance (using CPAP for ≥ 90% of nights and ≥ 8
h per night) in a non-randomized cohort study of 928 overweight/obese participants
(86% males, with mean age of 54 years) (168). Conversely, the results of a study of 298
type 2 diabetic patients (60% male, mean age of 60 years) randomised to CPAP
treatment during a 6 months period showed no significant difference in HbA1c among
the groups (169). The CPAP studies to date had several limitations including small
sample size, non randomized design /open label designs and heterogeneity in patient
selection (male predominance, lack of normal weight group, lack of control group). The
posibility of the effects of residual confounders (including lack of control for family
history of diabetes, abdominal/visceral obesity and change in physical activity and
dietary pattern) remain in all of these studies.
Figure 1.3. Obstructive sleep apnoea and cardiometabolic risk factors.
29
1.9.4 Sleep-disordered Breathing and Hypertension
A number of studies have investigated the relationship between hypertension and SDB,
particularly OSA and reported an association between OSA and hypertension (170,
171).
Several mechanisms may be responsible for the relationship between hypertension in
SDB, though hypoxia and sympathetic activation are considered as the main factors
(128, 171-173). Generally blood pressure and heart rate fall by 25% or more during
uninterrupted sleep, due to diminished sympathetic overflow and amplification of vagal
tone (171). However, in patients with OSA, nocturnal dipping in blood pressure is
absent or diminished due to hypoxia and sympathetic over activation (171). Gradually
the cumulative effects of sympathetic over activation together with the release of
reactive oxygen species and reduced levels of nitric oxide (NO), an endothelial
molecule controlling vascular tone, induce constant daytime hypertension (171).
Large epidemiological studies have shown an association between OSA and
hypertension based on measures of office blood pressure ≥ 140/90 or using
antihypertensive medication (173, 174). The results of a cross-sectional study of 11911
participants (70% male, mean age of 52) suspected of having OSA underwent overnight
sleep studies showed that OSA was independently associated with hypertension after
adjustment for various confounders including age, sex, obesity and diabetes (173).
Additionally the severity of OSA can positively influence the odds of hypertension
(175). A study of 2677 adults aged 20-85 years referred to a sleep clinic with suspected
OSA, showed that OSA associated with hypertension after adjusting for potential
confounders including age, sex and BMI (175). This study also showed that the
incidence of hypertension increased by OSA severity; 46% in moderate and 53% in
severe OSA participants, suggesting for every unit increase in AHI (severity of OSA),
the odds of hypertension increased by 1% (95%CI: 0.75, 1.5) (175). Similarly in a study
of 1,741 participants (42% males, with mean age of 47 years), Bixler et al (129) showed
a positive association between OSA and hypertension which was stronger as the
severity of OSA increased. Another notable community based study was the Wisconsin
Sleep Cohort Study that evaluated 1,060 participants aged 30-60 years and showed a
linear increase in blood pressure with increasing AHI index after adjusting for age, sex
and BMI (176). According to the study, the odds of hypertension for mild OSA was
1.42 (95%CI, 1.13, 1.78), meanwhile in moderate OSA and severe OSA these figures
30
were 2.03 (95%CI: 1.29, 3.17) and 2.89 (95%CI: 1.46, 5.64) respectively. The majority
of these studies included middle-aged adults selected from sleep clinics and thus were
not representative of general population. The effects of other confounders including
abdominal obesity were not evaluated in these studies.
Despite the role of SDB in hypertension, there are studies that have suggested a role of
age on this relationship. These studies have shown inconsistent results, with some
showing the association in young and middle-aged while others reporting an association
in the elderly. The Sleep Heart Health Study of 6,120 adults stratified into middle-aged
(< 60 years) and older aged (≥ 60 years) showed a relationship between SDB and
hypertension only in middle-aged adults (177). However, another longitudinal study of
2,148 adults aged 30-70 years, found no association after adjusting for age, sex , BMI
and lifestyle factors including smoking and alcohol intake (178).
Despite the current evidence of an association between SDB and hypertension mainly in
middle-aged and older participants, the relationship between SDB and hypertension has
not completely established in other age groups.
1.9.5 Sleep-disordered Breathing and Inflammation
A number of studies have shown an association between OSA and inflammatory
markers such as CRP. A cross-sectional study of 245 adults with a mean age of 48 years
found increased levels of high sensitivity CRP (hs-CRP) in severe OSA participants (β
= 0.53, P = 0.005) after adjusting for BMI and metabolic syndrome (179). In a case-
control study of 76 middle-aged adults (mean age 52, mainly males), Guven et al (180)
reported higher hs-CRP levels in participants with OSA. The study also showed a
relationship between hs-CRP and AHI independent of obesity (180). In 1,835 middle-
aged Korean adults those with moderate-to-serve OSA had 1.78 higher odds of being in
the highest tertile of CRP after BMI adjustment (181). Despite the positive association
between SDB and increased CRP levels, the results of other studies have shown no
association between OSA and CRP levels (106, 182). In 231 habitual snores who
underwent PSG, the association between OSA severity (AHI) with hs-CRP was not
significant (r = 0.10, P = 0.33) after adjusting for age and BMI (182). In Wisconsin
Sleep study of 907 adults (50% males. Mean age of 52) Taheri et al (106) found no
independent association between OSA and CRP, suggesting the postive association
between OSA and CRP is driven by obesity.
31
Although the association between inflammation in OSA is not fully understood,
hypoxia, oxidative stress and synthesis of reactive oxygen species (ROS) are in part
responsible for the endothelial dysfunction that associates with OSA (183).
To summarize, the results of previous studies that have examined the association
between OSA and CRP are conflicting. Limitations of those studies that have shown an
association between OSA and CRP include small study sample size and lack of control
for abdominal obesity.
1.9.6 Sleep-disordered Breathing and Dyslipidaemia
The role of SDB on dyslipidaemia has been investigated mainly in clinical studies and
suggests a possible association between dyslipidaemia and OSA (184). In this regard, a
study of 2,081 patients (638 women) undergoing nocturnal recording for clinical
suspicion of OSA found higher triglyceride and lower HDL-C levels in OSA
participants (185). A systematic review study based on 64 studies also showed a
positive correlation between dyslipidaemia and OSA (186). Other studies have shown
an improvement in dyslipidaemia with CPAP therapy (187) and a 6% increase in HDL-
C levels after 6 months treatment with CPAP (187). A study of 6,440 females aged 65
and over with moderate to severe OSA showed similar findings (188). However, all
randomised controlled trials have shown no improvements in dyslipidaemia with CPAP
therapy (148, 189). A randomised study of 24 adults with a mean age of 46 years
reported no significant changes in lipid levels after CPAP treatment for four months
despite a significant reduction in carotid intima thickness (148). Another crossover
study on 34 adults aged 49 years found no difference in lipid levels following CPAP
treatment for six weeks (189).
Some studies have shown OSA associates with changes in lipoproteins (190, 191). For
example, higher levels of oxidized LDL-C, which is more atherogenic, have been
reported in patients with OSA (190). Furthermore, HDL-C from OSA participants is
less able to prevent oxidation of LDL -C leading to HDL-C dysfunction in OSA patients
(160, 161).
Overall, current data has suggested chronic Intermittent Hypoxia (CIH) is likely one of
the main factors linking endothelia dysfunction with OSA (180).
32
1.10 Summary
The prevalence of overweight and obesity has increased dramatically over the past few
decades, escalating the burden of cardiovascular disease. While an imbalance between
food intake and physical activity is known to be the key factor in increasing weight
gain, recent evidence suggests other factors, such as insufficient sleep may also be
important in weight gain and its cardiometabolic complications (10). Insufficient sleep
includes both sleeping for less than the recommended period or having poor sleep
quality (10). Previously published reviews suggest there is a relationship between
insufficient sleep, obesity and cardiometabolic risk factors; however, this relationship is
not seen in all age groups. While this association has been shown to be stronger in
children and adolescents, the association is inconsistent in other age groups.
To plan cardiometabolic prevention strategies to minimize the negative impact of
obesity and its cardiometabolic consequences, it is important to clarify potential risk
factors (other than physical inactivity and increased caloric intake) in different age
groups including young adults. Few studies have explored the association between
insufficient sleep, obesity and cardiometabolic risk factors in young adults, and rarely
have they used objective methods of sleep measurement. Similarly, only a few studies
have examined sleep-disordered breathing and cardiometabolic risk factors in a young
adult population. In a study of 558 adolescents and young adults (aged 14-28), those
with a high risk for OSA had higher BMI, waist circumference and neck circumference
(P = 0.001), and were more likely to have features of the metabolic syndrome (38.9% vs
7.0%, P = 0.001) (137). Young adults access the health service less often, the under-
representation of this age group in clinical setting contributes to under-representation of
this age group in clinical research studies (192, 193) . Moreover, in non-clinical
research settings, random sampling is rarely used in this age group, thus the validity and
generalizability of the findings are limited (194).
1.11 The Western Australia Pregnancy Cohort (Raine) Study
The Western Australian Pregnancy Cohort (Raine) Study (see further details in the
Methods Chapter) commenced in 1989-1992 (195). This unique longitudinal study has
gathered an extensive range of general health, clinical, biochemical, behavioural and
genetic data of the parents during pregnancy, and the offspring at birth and at ages
33
1,2,3,5,14,17,18,20 and 22 years. Some of the key cardiometabolic findings from the
Raine study are summarised below.
Based on the results of cohort at age 8, overweight and obese children had adverse
cardiometabolic risk factors such as lower HDL-C and higher triglycerides in addition
to higher blood pressure compared to their peers (196). BMI at age 8 was predicted
positively by birth weight, maternal BMI and paternal BMI (196). There was a U- shape
association between birth weight and a high-risk cardiometabolic cluster that included
higher body mass index, high blood pressure, an adverse lipid profile and high serum
glucose levels (197).
At 14 years of age, a two-step cluster analysis was used to derive a distinct high-risk
group with features consistent with the metabolic syndrome (197). The results showed
29% of participants were classified as being in the high-risk cardiometabolic cluster.
Compared to those in the low-risk cluster, these individuals had higher BMI, (95% CI:
24.5, 25.4 vs 19.5, 19.8), waist circumference (95% CI: 83.4, 85.8 vs 71.0-71.8), insulin
(95% CI: 3.5, 3.9 vs 1.7, 1.8 vs.), systolic blood pressure (95% CI, 116.7 - 118.9 vs
110.8, 112.1) and triglycerides (95% CI: 1.25, 1.35 vs 0.780.80.), and lower HDL
cholesterol (95% CI: 1.20, 1.26 vs 1.44, 1.48) (197). Those in the high-risk cluster also
had increased levels of CRP, ALT, GGT and UA irrespective of gender (198) and
central obesity appeared to play a central role (199).
Results from Raine participants at the age 17 years showed the prevalence of
prehypertension or hypertension in overweight and obese individuals was 34% and 38%
respectively (200). The study also examined the association between smoking status and
hs-CRP using a three-level variable, girls not using oral contraceptives, girls using oral
contraceptives, and boys (201). Smoking was significantly associated with higher hs-
CRP levels in girls not using oral contraceptives, but not in girls using oral
contraceptives or in boys. Oral contraceptives use in non-smoking girls was the
strongest factor associated with higher hs-CRP levels (P < 0.001). Further, girls taking
oral contraceptives had 3.27 and 1.74 mmHg higher systolic and diastolic blood
pressure compared with non-users (200).
At the age of 20, the prevalence of prehypertension was 33.4% while for hypertension;
the figure was 3.6% (202). There was a significant sex difference in mean systolic blood
pressure: males had on average 11.3 mmHg higher systolic blood pressure compared to
females (202). Longitudinal analysis of the association between pregnancy life stress
34
events (e.g. pregnancy problems ,death of close relatives or friends, marital problems,
problems with children, loss of job in the family, money problem and residential move)
and systolic blood pressure from age 14 through age 20, showed an inverse relationship
which was highly significant (202).
1.12 Study Aims
Considering the current evidence of the Raine Study findings suggesting the presence of
cardiometabolic risk factors at young age (variation of cardiometabolic factors among
genders) and knowledge from the literature of a possible association between sleep
characteristics/disorder (including sleep duration , quality and high risk of sleep apnoea
disorder) and cardiometabolic risk factors, the current research hypothesized that sleep
insufficiency /sleep disorder would associate with adverse cardiometabolic risk factors
in a young adult population.
Therefore, the objectives of this thesis were to investigate the association between:
1. Objective measurements of sleep and cardiometabolic risk factors;
2. Sleep quality and cardiometabolic risk factors;
3. Sleep-disordered breathing (high risk for obstructive sleep apnoea) and
cardiometabolic risk factors; and
4. Whether the associations differed between males and females
in the Raine Study participants at 22 years.
35
Chapter 2:
Methods
36
2.1 Background of the Raine Study
A complete description of the Western Australian Pregnancy Cohort (Raine) Study has
been previously published (195). In brief, the study was commenced as a randomised
controlled trial that aimed to evaluate the effects of ultrasound scans on pregnancy
outcomes. The study was conducted between May 1989 and November 1991 (195).
Pregnant women attending the public antenatal clinic at King Edward Memorial
Hospital were enrolled in the project that formed the basis of the Raine Study. The
inclusion criteria were gestational age between 16 and 20 weeks, an intention to remain
in Western Australia in the coming years, an expectation to deliver at the King Edward
Hospital and sufficient proficiency in English (195). As this hospital was at the time the
sole tertiary referral centre for obstetrics in Western Australia, these mothers could be
considered as representative of the antenatal population of the Perth metropolitan area.
The initial written informed consent sought approval for long term follow up of the
cohort with the view that the study would have a significant potential for investigating
the developmental origins of health and disease.
A total of 2,900 pregnant women were recruited into the study. They were asked to
complete questionnaires that asked about their socio-economic status, lifestyle factors,
environmental exposure and past medical history. The partners were also asked to
complete a questionnaire containing information regarding their height, weight,
occupation, education and environmental exposures. The women were then allocated to
either an intensive group who had 5 ultrasounds during their pregnancy (n = 1,415) at
18, 24, 28, 34 and 38 weeks gestation or a ‘regular’ group who had one ultrasound
examination at 18 weeks. Any further ultrasound examinations were conducted only if
requested by clinician (n = 1,419). Allocation was done using computer-generated
random numbers. There were 66 enrolled women with multiple pregnancies who were
not allocated to either intensive or regular groups.
Of those recruited to the study there were 2868 live births from 2,826 pregnant mothers.
There were 2,801 (99%) women with single pregnancies, while 66 women had multiple
pregnancies during the recruitment phase. The outcomes of the initial study indicated a
higher intrauterine growth retardation (birthweight < 10th centile and birthweight <3 rd
centile) in the intensive ultrasound group, suggesting that frequent exposure to
ultrasound may influence fetal growth (195). However, by the age of 1 and thereafter no
significant difference observed in physical size among the ultrasound groups (203).
37
Subsequent recalls of the cohort have examined and documented a wide range of health
and behavioural data variables at 1, 2, 3, 5, 14, 17, 18, 20 and 22 years of age (Figure
2.1). Currently over 85,000 phenotypic and behavioural variables have been collected in
addition to establishment of a genetic database (Table 2.1). Biological samples
including blood and urine have been collected at many of these follow-up assessments
(204).
Figure 2.1. The Raine Study: Retention of the cohort from 1 to 22 years of age.
The Raine Study currently incorporates 25 broad areas of research including asthma &
allergy, anaesthetic, Development Origins of Health and Disease (DoHAD),
cardiovascular, cognitive neuroscience, dental health, nutrition, eating disorders,
endocrinology, infectious diseases, language & social development, mental health, risky
behaviour, musculoskeletal, ophthalmology, physical activity, reproductive health
(204).
The overarching aim of the Raine study has been to evaluate the interrelationship
between health, diet, lifestyle, behavioural, psychosocial and genetic factors from
infancy through childhood into adolescents and young adulthood. The adolescent/young
adulthood period is considered a critical period during which lifestyle behaviours that
are likely to substantially influence adult cardiovascular risk tend to become established.
At all recalls there have been core data collected including questionnaires relating to
psychosocial, lifestyle and medical information, in addition to anthropometry (height,
38
weight, waist circumference, skinfolds) and clinical (blood pressure, vascular function)
measures (Table 2.1).
Furthermore, blood samples have been collected for biochemical variables including
fasting lipids, glucose, insulin, liver function and iron metabolism. The cohort recall at
22 years of age had a specific focus on aspects of sleep, cardiovascular measures, spinal
pain and work productivity (204).
39
Table 2.1. Data recorded during each of the cohort recalls in the Raine Study (204)
Doppler
Ultrasound,
Maternal
blood, cord
blood
Family,
Social,
Economic/
Medical
Data
Mental
Health
Nutrition Asthma
/Allergy
Hearing
Screening
Tests/Milk
Teeth
Risky
Lifestyle
Physical
Assessment
Blood Fitness
Assessment
/ DEXA
Liver
Ultrasound
DNA
Sampling
Sleep
Antenatal /
Neonatal
1-3 yrs
5 yrs
8 yrs
10 yrs
14 yrs
17 & 18 yrs
20 yrs
22 yrs
40
2.2 Follow-up of the Raine Study Participants at 22 Years
Participants were contacted using an established confidential contact database of the
Raine study by telephone around the time of their 22nd birthday. The follow-up
procedure was explained to them by the recruitment team members and an appointment
was arranged. All of the participant information, consent forms and core questionnaires
(general and medical questionnaires) were mailed to those who accepted to participate
in the follow-up study at 22 years of age. On arrival, participants provided written
informed consent. The trained research assistants explained protocols to the participants.
The physical assessments were completed over a 3 hour appointment and included
asthma, skin, muscloskeletal and tissue sensitivity. The asthma assessments included
spirometry, mannitol challenge test and induced sputum. The skin assessments included
skin prick testing and mole counting. The muscloskeletal assessments included back
muscle endurance (in which the participants completed prone trunk hold) and hip
accelometer (which used to capture movement during wakefulness). Tissue sensitivity
examinations included pressure pain threshold and cold pain threshold (204).
The participants were asked to complete food frequency, sleep specific and work
attendance questionnaires. Once these questionnaires were completed they were asked
to stay overnight at the Centre for Sleep Science, at the University of Western Australia
for an overnight sleep study (including polysomnography and wrist accelometer which
used to capture movement data during sleep period). Participants went home wearing
the hip and wrist accelometers for the following 7 days (see section Objective Sleep
Measurement for details). The recruitment for the study was commenced in March 2012
and completed in July 2014. The total number of participants in this follow up was
1,234.
All aspects of this study were approved by the University of Western Australia Human
Research Ethics Committee (RA/4/1/5002) (204). The participants provided written
informed consent for data collection and the study was conducted in accordance with
the Declaration of Helsinki (204).
2.3 Anthropometric Measurements
Body weight was measured to the nearest 100 g using a Wedderburn Chair Scale
(Wedderburn, AUS). Height was measured without shoes to the nearest 0.1 cm using a
41
Holtain Stadiometer (Crosswell, UK). All anthropometric measurements were
performed with participants wearing light clothing. Body Mass Index (BMI) was
calculated as weight/height2 (kg/m2) (205). The BMI was categorised as normal-weight
(< 25 kg/m2), over-weight (> 25 < 30 kg/m2) and obese (≥ 30 kg/m2) based on WHO
recommendation (205). As the majority of the Raine Study population (90%) is
Caucasian, the Caucasian-based BMI cutoffs were used in the current study (206).
Waist circumference was measured using a measuring tape at the halfway point of the
lowest rib and iliac crest. Hip circumference was measured at the level of maximum
extension of the buttocks, to the nearest 0.1 cm and hip-to-waist ratio calculated.
2.4 Blood Pressure Measurement
Systolic and diastolic blood pressures (BP) were recorded using an oscillometric
sphygmomanometer (Dinamap ProCare 100; Soma Technology, Bloomfield,
Connecticut, USA) with an appropriate cuff size in a sitting position. Following resting
for 5 minutes, BP and heart rate recordings were repeated every 2 minutes until at least
5 separate measurements were obtained. BP was calculated from the average of the last
5 readings. Systolic and diastolic prehypertension were defined as SBP 120-139 mmHg
and DBP 80-89 mmHg. Systolic BP ≥ 140 and diastolic BP ≥ 90 were defined as
systolic and diastolic hypertension using the adult The Seventh Report of the Joint
National Committee on the Prevention, Detection, Evaluation and Treatment of High
Blood Pressure (JNC-7) criteria (207).
2.5 Blood Collection
Participants were advised not to eat or drink after 10 pm in order to fulfil eight hours of
fasting. Blood samples were collected in the morning after an overnight fast by standard
phlebotomy from an antecubital vein. The total volume of collected blood was 100 ml.
Blood samples were processed within 2 hours of collection by centrifugation at 3,500
rpm for 5 minutes at 4 C. Serum and plasma were stored at -80°C until analyses were
performed. All blood biochemical measurements were carried out at the Path West
Laboratory at Royal Perth Hospital.
Fasting blood glucose was measured by a standard spectrophotometric assay (Abbott
Diagnostics, Abbott Laboratories, USA). Serum insulin was assayed by the
immunoassay technique (Abbott Diagnostics, Abbott Laboratories, USA). Inter-assay
42
coefficients of variation for glucose and insulin were recorded as 0.54%-1.76% and
1.78%-2.59%, respectively. The homoeostasis model assessment for insulin resistance
(HOMA-IR) score was calculated based on fasting insulin (μU/ml) × fasting glucose
(mmol/l) /22·5 (208).
Total cholesterol and triglycerides (TG) were measured with enzymatic colorimetric
assays (Abbott Diagnostics, Abbott Laboratories, USA). The inter-assay CVs of
cholesterol and TG ranged from 0.51% to 0.87% and 1.02%to 1.92%, respectively. The
LDL cholesterol (LDL-C) was calculated according to the Friedewald equation (209).
The HDL cholesterol (HDL-C) was measured with an enzymatic, colorimetric assay
(Abbott Diagnostics, Abbott Laboratories, USA). The inter-assay CV of HDL-C ranged
between 1.94%-2%.
High sensitivity CRP (hs-CRP) was measured by CRP Vario Reagent (Abbott
Diagnostics, Abbott Laboratories, USA); the inter-assay CV was 2.07%-2.08%.
Liver function tests included measurements of alanine aminotransferase (ALT),
aspartate aminotransferase (AST), gamma-glutamyl transpeptidase (GGT) and alkaline
phosphatise. ALT was measured using an activated alanine aminotransferase reagent
(Abbott Diagnostics, Abbott Laboratories, USA); the inter-assay CV ranged between
1.37%-3.68%. AST was measured using an activated aspartate aminotransferase reagent
(Abbott Diagnostics, Abbott Laboratories, USA); the inter-assay CV ranged between
0.77% to 2.87%. GGT was measured using a GGT reagent (Abbott Diagnostics, Abbott
Laboratories, USA); the inter-assay CV ranged between 0.64% to 1.44%. ALP was
assayed using an Alkaline Phosphatase Reagent (Abbott Diagnostics, Abbott
Laboratories, USA) with an inter-assay CV of 1.96% to 2.18%.
2.6 Questionnaires
The questionnaires included general, medical, food frequency and sleep questionnaires.
All of the questionnaires were delivered in a paper-based format. The general
questionnaire focused on 4 main categories including:
a. Socio-economic factors (e.g. income, employment and education).
b. Risk taking behaviours (e.g. smoking and alcohol).
c. Family life (marital status, having child).
d. Lifestyle (e.g. physical activity, hormonal contraceptive use in females only).
43
A medical questionnaire evaluated the presence of any medical condition in addition to
past medical history and prescription medication of individuals. The sleep
questionnaires included Pittsburgh Sleep Quality Index, Berlin and Epworth Sleepiness
Scale Questionnaires. The details of all sleep questionnaires are provided in the next
sections.
2.7 Selected Risk Taking and Lifestyle Variables
Smoking and alcohol consumption were used as the main indicators of risk taking
behaviour in the final analyses. The dichotomous response (yes/no) to the questions
about smoking was used in this study. Alcohol intake was estimated based on the self-
report questionnaire. For each type of alcohol beverage the number of drinks consumed
on one occasion and the frequency of drinking were asked. Based on the average
alcohol content in different types of beverages, the average of alcohol consumption per
week (gram/week) was calculated and used in subsequent analyses.
In females, current use of any hormonal contraceptive pill, implant, intrauterine device
and injection in females was classified as hormonal contraceptive usage.
The physical activity level of participants was estimated from self-report questionnaires
based on the International Physical Activity Questionnaire (IPAQ). The IPAQ
questionnaires assessed 3 different types of activity including walking, moderate
intensity activities and vigorous intensity (210). For each of the activities the frequency
(as days per week) and duration (time per day) were measured. Time spent in each
activity was converted to Metabolic Equivalences (METS) defined by weighting each
type of activity by its energy requirements. Met-minutes were calculated by multiplying
the MET score by the minutes performed
The median MET -minutes were used as a continuous variable for assessing physical
activity in participants based on the following formula:
Walking MET-minutes/week = 3.3 * walking minutes * walking .days.
Moderate MET-minutes/week = 4.0 * moderate-intensity activity minutes *
moderate days.
Vigorous MET-minutes/week = 8.0 * vigorous-intensity activity minutes *
vigorous-intensity.
44
Total physical activity in MET-min/week was calculated as the sum of walking
+moderate+ vigorous MET-min/week scores.
2.8 Subjective Sleep Measurements
2.8.1 Sleep Diary
During the study at the Centre for Sleep Sciences, University of Western Australia, a
sleep diary was given to each participant in order to collect sleep hours and time to bed
and time to rise. All of the participants received information on how to fill in their sleep
diary in the evening and morning. They were required to note the time they went to bed
and the time they got up each day for 7 days. The sleep diary was given to the
participants in the morning after the overnight sleep study at the Centre for Sleep
Sciences. Sleep diary reminders were added to the calendar system and telephone calls,
text messages or Facebook messages were used to remind participants to fill in their
diaries.
2.8.2 Sleep Questionnaires
A number of sleep-related questionnaires were given to the participants in the evening
of overnight sleep study at the Centre for Sleep Sciences. These included: the Pittsburgh
Sleep Quality Index, the Berlin questionnaire (to assess high risk for OSA), the Epworth
Sleepiness Questionnaire and the Short Functional Outcome of Sleep Questionnaire
(FOSQ).
2.8.2.1 Pittsburgh Sleep Quality Index Questionnaire
The Pittsburgh Sleep Quality Index (PSQI) questionnaire is a self-administrated
questionnaire that measures sleep quality and sleep disturbance during the previous
month (33). It has seven components including subjective sleep quality, sleep latency
(the length of time it takes to achieve transmission from full wakefulness to sleep), sleep
duration, habitual sleep efficiency (time spent in bed divided by actual asleep time),
sleep disturbance, usage of sleep medication and daytime dysfunction. The total score is
calculated by adding the seven components scores together. A high total score (≥ 5)
indicates of poor sleep quality (Figure 2.1) (33).
45
Figure 2.2. PSQI Questionnaire (33).
2.8.2.2 Berlin Questionnaire
The Berlin questionnaire is a self-administrated questionnaire that identifies individuals
at risk of having sleep apnoea (36). It has 3 main categories. The first category consists
of 5 questions that aim to evaluate the presence and frequency of snoring. The second
category consists of 3 questions regarding daytime sleepiness and fatigue and the last
(third) category evaluates history of obesity or hypertension. The first and second
46
categories are considered a positive response when there are 2 or more positive
responses to the questions. The last category is considered positive in the presence of
BMI > 30 kg/m2 or a history of hypertension (Figure 2.2). Two or more positive
categories indicate a high likelihood of sleep apnoea (36). Previous studies have
recommended the Berlin Questionnaire as a screening tool in the general population for
its good sensitivity and specificity (40; Senaratna, 2017 #458). In the Raine Study the
last category of the Berlin questionnaire was omitted because these aspects were
addressed elsewhere during the 22-year assessment. The modified Berlin Questionnaire
used in the Raine Study is presented in Appendix 1.
47
Figure 2.3. Berlin Questionnaire (211).
2.8.2.3 Epworth Sleepiness Scale Questionnaire
This questionnaire aims to evaluate daytime sleepiness and is composed of 8 questions
which are scored equally. The questions ask the probability of falling sleep in different
conditions (212). A higher total score indicates higher probability of falling asleep
(212).
48
2.8.2.4 Short Functional Outcome of Sleep Questionnaire
This self-administrated questionnaire is designed to evaluate the impact of excessive
sleepiness on various activities relating to an individual’s daily living including
relationships, productivity, social and physical activity, and vigilance (213).
2.9 Objective Sleep Measurement
2.9.1 Wrist Accelerometer
Sleep actigraphy has been used as an objective measurement of sleep (214). This
method is based on monitoring the amount of movement a person makes. As minimal
movement occurs during sleep compared to wakefulness, sleep actigraphy can be used
to distinguish between sleep and wakefulness (214). By using sleep actigraphy overall
sleep measures can be obtained. The overall sleep measurements included:
1. Total Sleep Time (TST): duration of sleep which reported as hours or minutes
(215).
2. Sleep Onset Latency (SOL): time period between reported bed time and
actigraphy scored sleep onset time (215).
3. Wake After Sleep Onset (WASO): the minutes awake during the sleep period
after sleep onset (215).
4. Sleep Efficiency (SE): time spent in bed divided by actual asleep time (215).
In order to calculate all mentioned sleep measurements, time into bed (lights out) and
time out of bed (lights on) are also needed (215). According to the Society of
Behavioral Sleep Medicine (SBSM) actigraphy monitoring guideline, it is
recommended to identify lights out and lights on time by individual-recorded method
such as individual sleep diary or actigraphy log (215). According to the guideline,
participants are required to keep and fill out their actigraphy log or sleep diaries while
they are wearing a wrist accelerometer. This method uses a device called an
accelerometer that is equipped with a light sensor and is placed on the wrist. The
common location for placing an accelerometer is the non-dominant wrist although it can
also be placed on the leg, shoulder or hip (215, 216).
49
The reliability and validity of sleep actigraphy has been evaluated in many studies (22,
25, 217, 218). The range of agreement between actigraphy and Polysomnography (PSG)
has been shown to be between 91-93% (219)
In the Raine Study we used a GT3X+ activity monitor (Actigraph, FL, USA; Figure 2.3)
which is a small (4.6 x 3.3 x 1.5 cm), lightweight (19 g) tri-axial solid state water-
resistance accelerometer.
Figure 2.4. The Wrist Accelometer used in the Raine Study participants.
2.9.2 Wrist Accelerometry (Actigraphy) Data Collection
During the overnight sleep study at Centre for Sleep Science, at the University of
Western Australia, the participants received an information sheet and verbal instructions
on how to wear and use the device. They were required to wear the device for at least 8
consecutive nights (including the overnight study). They were advised to continue their
50
normal lifestyle and sleep-wake pattern in their home environment. They were provided
with an instruction sheet and a self-addressed Express post bag in order to return the
device when the measurements were completed.
2.9.3 Actigraphy Software &Analysing Sleep Actigraphy Data
The raw data were collected at 30 Hz and recorded in one -minute period (epoch). All
accelerometer files were downloaded on a computer and checked by a Raine study
research assistant for any technical problems. The Actilife software (Actigraph 2012,
Actilife 6.8) was used to download and analyse the data. As the device used in this
study was not equipped with an event marker button to record significant events such as
“lights-out” and “lights-on” time , these variables were extracted from the sleep diaries
as recommended by the SBSM Actigraphy Monitoring Guidelines (215).
The Actilife software uses the Sadeh algorithm (219) that is applied to a one minute
period (epoch) of vertical axis movement data. The algorithm then transforms that
epoch of movement data, based on an 11 minute window, with that epoch at centre, and
also considers the previous and the following five minutes of data (219). Any missing
epochs are considered as zero. The vertical axis data are placed in the Sadeh algorithm.
By using this algorithm the software considers an epoch as asleep if the result of the
algorithm is greater than -4. The software calculates all four overall sleep measurements
including total sleep time (TST), sleep latency, sleep efficiency and wake after sleep
onset (WASO) (see Wrist Accelometer Section in Chapter 2-Methods, page 48).
According to the Society of Behavioral Sleep Medicine (SBSM) actigraphy monitoring
guideline, participants are required to keep and fill out their actigraphy log or sleep
diaries while they are wearing a wrist accelerometer (215).
During the Raine study, actigraphy data were considered invalid if the wrist
accelerometer malfunctioned (220). The sleep diary data were also considered invalid if
the participant did not enter the diary variables correctly (wrong date and time) or if
there was more than one hour discrepancy between the sleep diary and actigraphy (220).
Sleep diaries were used to identify when the lights in the room were turned off (time of
going to bed with intention to sleep) and when they were turned on (time out of bed and
awake) in those participants that had sleep diary data. After checking the actigraphy
data, sleep times (lights out and lights on) were entered manually into the sleep analysis
51
tool of Actilife software by clicking on the “Add Bed Time “button. By confirming the
time, the selected time periods were highlighted on the graph window. The final sleep
report for each individual was exported to an excel sheet.
Prior to entering information relating to lights out and lights on, the activity counts were
evaluated in order to check for periods of low and moderate activity. If the actigraphy
data of low and moderate activity were in accordance with the sleep diary data, the sleep
times (lights out and lights on) were extracted from the sleep diary and entered into the
software. In total there were 497 participants in which their sleep diaries were valid and
therefore used to identify “lights on”’ and “lights out”.
Data loss from either the wrist accelerometer, actigraphy log or sleep diary is a common
issue in sleep studies (215). As a solution, an alternative method adapted from the
SBSM guide to actigraphy monitoring was applied to missing data in the Raine sleep
study. According to the SBSM actigraphy monitoring guideline, the following
alternative methods are recommended in order to identify lights out and lights on time
in those whose their sleep diaries are incomplete, inaccurate or disagreeing with
actigraphy data (215) . These included:
• Visual identification of time based on movement data.
• Excluding a night with unreliable or missing diary data.
• Using lights out and lights on time from previous night if having regular sleep
habits.
• Additional device feature such as light measure and non-wear detection.
• Proprietary device and software features such as automated bed period detection
algorithms.
Visual manual identification of sleep periods whenever sleep diary data are incomplete,
inaccurate or missing is the most common approach used in the literature (221). This
approach was used in the Raine study.
All actigraphy files with sleep diaries were analysed using The Actilife software. The
actigraphy files without sleep diaries were scored by visual manual identification, and
also by a second qualified sleep technician who was blinded to the first scorer’s values.
After recognising light out and light on time using the visual manual identification
method, results were compared. Following visual manual scoring all 84 files were
added to the rest of the data.
52
2.10 Statistical Methods
Quality and efficiency of sleep was assessed using the four actigraphy derived variables:
total sleep time, sleep efficiency, sleep latency and Wake after Sleep Onset (WASO) as
well as PSQI derived sleep quality. In addition, a classification of Obstructive Sleep
Apnoea (OSA) was determined from the Berlin Questionnaire.
As BMI was included in the Berlin Questionnaire, obesity-adjusted estimates may be
subject to over adjustment. By using the Berlin Questionnaire a modified classification
of OSA, excluding category three (in which BMI is located) was undertaken. In the
Berlin Questionnaire, OSA is defined as 2 positive classifications from any of the three
categories (snoring, fatigue/sleepiness and BMI/hypertension). A modified, 3 level
classification was generated using the number of categories in which a positive score
was obtained from only the snoring and fatigue/sleepiness categories. A resulting score
of 2 equated to a diagnosis of OSA on the full Berlin as the third category would be
unnecessary for a positive determination. A score of 1 remains dependent on the result
of the third omitted category to form a definite diagnosis so these were classified as
possible OSA. A score of 0 identified negatives as the result of the third category would
not alter the diagnosis.
All continuous variables were tested for normality and where departures were found,
log transformed. Subject characteristics were summarised using mean (95% CI) or
median (1st, 3rd quartiles) as appropriate for continuous variables and proportion (95%
CI) for categorical variables. For log transformed variables, data is presented as
geometric means (95% CI).
Comparisons of those who attended follow-up at age 22 with those who did not attend
were performed to identify potential bias in the participating sample. Similar
comparisons were made within the 22 year participants between those who participated
in the actigraphy sleep study and those who did not. Comparisons were also made
between male and female participants in the 22 year age cohort to identify gender
differences.
BMI, waist circumference, systolic and diastolic blood pressure, biochemical variables
and HOMA-IR as an index of insulin resistance (see Sections 2.3, 2.4 and 2.5 in
Chapter 2-Methods, pages 40-41) were analysed as continuous variables. Differences in
53
continuous variables were investigated using linear regression and logistic regression
for categorical variables.
The following method was used to assess the association between each of the
independent sleep variables and cardiometabolic risk factors. Firstly scatterplots,
Locally Weighted Scatter plot Smoothing (lowess), Multiple Variable Rank Sum
multivariable regression splines (mvrs) were used to assess the linearity of associations
with continuous independent variables (only for actigraphy derived sleep variables),
prior to analysis using univariate linear regression. Univariate regressions were
followed by models adjusted for gender then gender and BMI (excluding models where
BMI was outcome) and finally gender, BMI and lifestyle factors. The lifestyle
confounders included in each model were selected based on their associations with the
particular cardiometabolic risk factors identified through univariate regression models
(P < 0.2).
There was a large amount of missing data for physical activity variable. Therefore, it
was necessary to separate the effect on the sleep variable coefficient of accounting for
physical activity and the loss of sample its inclusion generated. The regression was run
using the sample with physical activity data, without including the physical activity
variable. The same model was then run again including the physical activity variable.
Comparison of the sleep coefficient from these two models allowed the effect of
physical activity to be assessed. If the coefficient was unchanged, physical activity was
excluded to retain the larger sample in the analysis.
Examination of scatterplots identified some outlier values of BMI whose influence
warranted further investigation. Where statistical significance was identified, a
sensitivity analysis was undertaken to assess the results for robustness to the removal of
these outliers. Both analyses, with and without the outliers, are reported.
The Raine cohort is known to contain siblings whose measures may not be independent
due to familial genetics or environmental background. A family identifier variable was
used to apply a per family cluster variance adjustment to all regressions to account for
the potential correlation between siblings.
Statistical significance was set at P < 0.05. The data was analysed using STATA
(StataCorp. 2013. Stata Statistical Software: Release 13. College Station, TX: StataCorp
LP).
54
Chapter 3:
Results
55
3.1 Follow-up Response of the Raine Study Participants at 22 Years of
Age
Figure 3.1 shows the retention rates of the Raine study cohort from birth to 22 years of
age. Of the original cohort of 2,868 live births there are 41 deceased, with an additional
561 that have withdrawn and 1,029 who are lost and /or have deferred. There were
1,234 participants from the original 2,868 live-born children who participated at the 22
year follow-up, representing a 43% response which is similar to previous cohort recalls.
Of the 1,234 that participated at 22 years, 975 healthy young men and women as
ascertained from medical history questionnaires, provided full anthropometry and
biochemical data.
A logistic regression model determined potential predictors for participating at age 22.
These included maternal (age at conception, BMI before pregnancy, education),
socioeconomic and demographic (family income) variables, in addition to
anthropometric variables collected at birth. Table 3.1 shows maternal age, maternal
higher educational level and high family income were significantly positively associated
with participation at age 22.
56
Number of Participants at birth N = 2868
Follow-up:
Year 1 N = 2441
Deferred=204 Lost=174 Total Withdrawn=21
Deceased=28
Follow-up:
Year 2
N = 1988
Deferred=381 Lost=418 Total Withdrawn=51
Deceased=30
Follow-up:
Year 3 N = 2280
Deferred=321 Lost=158 Total Withdrawn =79
Deceased=30
Follow-up:
Year 5 N = 2237
Deferred=339 Lost=135 Total Withdrawn =127
Deceased=30
Follow-up:
Year 8 N = 2140
Deferred=376 Lost=124 Total Withdrawn =198
Deceased=30
Follow-up:
Year 10 N = 2047
Deferred=281 Lost=162 Total Withdrawn =348
Deceased=30
Follow-up:
Year 17
N = 1771
Deferred=379 Lost=206 Total Withdrawn=447
Deceased=35
Follow-up:
Year 20 N = 1342
Deferred/lost =840 Total Withdrawn=485
Deceased=36
Follow-up:
Year 22 N = 1234
Deferred/lost =1029 Total Withdrawn=561
Deceased=41
Figure 3.1. The Raine Cohort Study Retention from birth to 22 years.
57
Table 3.1. Univariate Regression Analyses Evaluating Factors Predicting the Likelihood of Participating at the 22 Year Follow Up
Characteristics Participants attending follow
up (n=1234)
Participants not attending
follow up (n=1634)
OR (95%CI) P value
Maternal age (years) 29.45 (29.1, 29.7) 27.02 (26.7, 27.3) 1.03(1.02,1.05) 0.016
Maternal BMI (kg/m2) 22. 17 (22.0, 22.40) 22.50 (22.30, 22.70) 0.99 (0.95, 1.03) 0.73
Offspring birth weight (kg) 3.3 (3.2, 3.3) 3.2 (3.2, 3.3) 1.00 (0.99, 1.00) 0.19
Income*
Low Income %
High Income%
38 (35, 40)
61 (59, 64)
56 (53, 58)
43 (41, 46)
1.03 (0.89, 1.1)
< 0.001
Maternal post-school higher
education (%)
57.4 (54.5, 60.1) 41.5 (39.1, 44.0) 1.42 (1.2, 1.6) < 0.001
Data are presented as mean (95%CI) for continuous variables and proportion (95%CI) for categorical variables.
*Low and high income is defined by the 1989 poverty line cut-offs (AUD$ 325.8 /week).
58
3.2 Socioeconomic and Lifestyle Characteristics of the Raine Study
Participants at 22 Years
Socioeconomic and lifestyle characteristics are presented in Table 3.2. There was no
significant difference in the distribution of higher education (post school higher
education) between males and females (Table 3.2). Although there was no difference in
the proportion of females and males in full/part time employment (P = 0.23), a
significantly greater proportion of males earned more than $604 per week (P = 0.002).
Smoking, alcohol intake, usage of hormonal contraceptive, physical activity and
depression, anxiety and stress score (DASS) were evaluated as lifestyle behaviours of
the study population. There was a significant difference in the rate of smoking among
gender (P = 0.02, Table 3.2). The percentage of alcohol drinkers in females was 76.4%
(95%CI: 72.8, 79. 7) compared with 82 % (95%CI: 78.4, 85.2) in males (P = 0.024).
Males consumed more alcoholic drinks per week (g/wk) compared to females (P <
0.001). Hormonal contraceptives were used by 50.6% of females. Males undertook
almost twice as much physical activity per week than females (P < 0.001). Evaluating
the adverse emotional states of DASS score questionnaires showed that females had a
significantly higher total DASS score (P < 0.001).
59
Table 3.2. Socioeconomic Characteristics of the Participants at 22 Years
Parameters N (Females/Males) Females Males P value
Post-school higher
education, %
1101
(593/508)
52.6
(47.9, 57.2)
48.4
(43.6, 53.2)
0.22
Total usual pay/wk
after tax >$ 604
1026
(557/469)
41.8
(37.1, 46.7)
52.9
(47.9, 57.8)
0.002
Employed, % 1139
(604/535)
82.9
(79.6, 85.7)
80.5
(76.9, 83.6)
0.23
Current smokers,
%
1139
(603/536)
14.0
(11.5, 17.1)
19.6
(15.87, 22.6)
0.02
Total weekly
alcohol
consumption
(g/wk)*
1146
(607/539)
88
(73.64, 101.84)
194.4
(170.3, 218.6)
< 0.001
Physical activity
(MET-min/wk)*
962
(515/447)
2392.50
(2133.90, 2682.42)
4003.57
(3551.10, 4513.68)
< 0.001
Total DASS score* 1078
(583/495)
18.00
(16.48, 19.62)
12.70
(11.50, 14.00)
< 0.001
Data are presented as mean (95%CI) for continuous variables or proportion (95%CI) for categorical
variable.
*Not-normally distributed variables presented as geometric mean (95%CI).
Standard error adjusted for 20 families who had two children in the study.
3.3 Anthropometric and Biochemical Characteristics of the Raine
Study Participants at 22 Years
There was no significant difference in BMI between males and females (Table 3.3).
According to the adult criteria of BMI categorization, 61.5% of participants were
normal weight (BMI < 25), 24% were overweight (25 ≥ BMI < 30) and 15% were obese
(BMI ≥ 30). More males were overweight (28.6%) compared with 18.4 % of females
(OR 1.70, 95% CI: 1.2, 2.2, P < 0.001) but there was no significant difference in the
proportion of obesity between genders (OR 0.79, 95%CI: 0.5, 1.1, P = 0.18).
Using adult criteria of waist circumference, 37% of females and 20% of males (P <
0.001) had central obesity according to the criteria of the World Health Organization
(males ≥ 102 cm and female ≥ 88 cm) (205).
Mean Systolic blood pressure (SBP) was higher in males compared to females (123
mmHg vs 113.8 mmHg, P < 0.001). There was no significant difference in DBP
between genders. Using criteria for pre-hypertension as defined by 120 mmHg < SBP ≤
139 mmHg or 80 mmHg < DBP ≤ 89 mmHg, 125 (23%) females and 289 (53%) of
60
males were pre-hypertensive (222). Hypertension defined as SBP ≥ 140 mmHg or DBP
≥ 90mmHg mmHg, showed 5 females and 34 (6%) males were categorized as
hypertensive (222). Males had a higher prevalence of pre hypertension and hypertension
compared to females (P < 0.001).
Table 3.3 illustrates the anthropometric, blood pressure and biochemical characteristics
of female and male participants. Mean glucose levels were significantly higher in males
(P < 0.001), whereas females had higher levels of insulin (P < 0.001) and HOMA-IR (P
< 0.001). Using HOMA-IR ≥ 2.73, as a measure of insulin resistance, showed 15.8% of
females and 11.3% of males were insulin resistant (P = 0.04) (223).
Females had higher levels of total cholesterol (P = 0.001), HDL-C (P < 0.001) and hs-
CRP (P < 0.001) compared to males. In contrast males had higher levels of the liver
enzymes AST, ALT, ALP and GGT, compared with females (P < 0.001). Females using
hormonal contraceptive had lower BMI and waist circumference, while having higher
total cholesterol, HDL-C and triglycerides (Table 3.4).
61
Table 3.3. Characteristics of the Raine Participants at 22 Years Stratified by
Gender
Characteristics Total (975) Females (480) Males (495) P value
BMI
(kg/m2)
25.22
(24.94, 25.59)
25.25
(24.73, 25.78)
25.28
(24.89, 25.67)
0.85
Waist (▪)
(cm)
83.27
(82.45, 84.10)
80.54
(79.27, 81.81)
85.95
(84.96, 86.94)
< 0.001
SBP
(mmHg)
118.6
(117.9, 119.3)
113.8
(112.9, 114.7)
123.3
(122.3, 124.2)
< 0.001
DBP
(mmHg)
67.0
(66.5, 67.4)
67.2
(66.6, 67.8)
66.7
(66.1, 67.4)
0.27
Glucose
(mmol/L)
4.97
(4.95, 5.00)
4.86
(4.82, 4.89)
5.0
(5.05, 5.12)
< 0.001
Insulin (▪)
(mU/L)
7.36
(7.14, 7.59)
8.0
(7.72, 8.41)
6.74
(6.46, 7.04 )
< 0.001
HOMA-IR(▪) 1.62
(1.57, 1.67)
1.73
(1.65, 1.81)
1.52
(1.45, 1.59)
< 0.001
Total Cholesterol
(mmol/L)
4.62
(4.57, 4.67)
4.71
(4.64, 4.78)
4.53
(4.46, 4.60)
0.001
HDL-C
(mmol/L)
1.35
(1.33, 1.37)
1.47
(1.44, 1.50)
1.23
(1.21, 1.25)
< 0.001
LDL-C
(mmol/L)
2.76
(2.71, 2.80)
2.74
(2.69, 2.80)
2.77
(2.70, 2.83)
0.50
Triglycerides
(mmol/L)
1.09
(1.06, 1.13)
1.06
(1.02, 1.11)
1.12
(1.08, 1.17)
0.05
hs-CRP (▪)
(mg/L)
1.06
(0.98, 1.15)
0.75
(0.67, 0.84 )
0.45
(0.38, 0.53)
< 0.001
ALT (▪)
(U/L)
25.26
(24.49, 26.04 )
22.81
(21.65, 23.97)
34.60
(32.77, 36.43)
< 0.001
AST (▪)
(U/L)
26.48
(25.85, 27.12)
23.85
(23.0, 24.6)
29.0
(28.0, 30.0)
< 0.001
ALP
(U/L)
69.49
(68.22, 70.75)
61.4
(59.8, 62.9)
77.3
(75.64, 79.09)
< 0.001
GGT (▪)
(U/L)
17.62
(17.10, 18.15 )
15.27
(14.66, 15.91)
20.24
(19.47, 21.05 )
< 0.001
Data are presented as mean (95%CI). Linear regression models for continuous variables.
(▪) log transformed was used for in regression model.
62
Table 3.4. Characteristics of females stratified by hormonal contraceptive use.
Characteristics Females not using contraceptive Females using contraceptive P value
BMI (kg/m2) 26.38
(25.30, 27.46)
24.70
(24.06, 25.35)
0.009
Waist (cm) 83.77
(81.26, 86.28)
79.12
(77.50, 80.74)
0.002
Total Cholesterol
(mmol/L)
4.58
(4.46, 4.70)
4.82
(4.73, 4.92)
0.002
HDL-C
(mmol/L)
1.39
(1.33, 1.46)
1.51
(1.47, 1.55)
0.002
Triglycerides
(mmol/L)
0.98
(0.89, 1.03)
1.14
(1.08, 1.19)
< 0.001
hs-CRP (▪)
(mg/L)
1.08
(0.88, 1.33)
1.87
(1.61, 2.17)
< 0.001
ALP (U/L) 66.30
(63.52, 69.07)
58.45
(56.51, 60.39)
< 0.001
Data are presented as mean (95%CI).
(▪) log transformed was used for in regression model.
3.4 Actigraphy Data of the Raine Study Participants at 22 Years
Figure 3.2 summarises the sleep actigraphy data. In total, 587 participants had at least
one night of actigraphy data. Of these, 6 participants were excluded from further
analyses as their actigraphy variables were in outlier range [sleep efficiency < 50%,
Wake After Sleep Onset (WASO) >120 minutes, sleep latency is less than 5 minutes or
more than 60 minutes]. Therefore a total of 581 participants (303 females and 278
males) were used for data analysis of actigraphy data (ranging from one night to 8
nights per participant). There were 497 actigraphy files with sleep diaries and 84 files
without any sleep diary data.
A total of 111 participants had less than 3 nights of actigraphy data, while 470 had ≥ 3
nights. The mean values of the actigraphy sleep variables [total sleep time, sleep
efficiency, sleep latency and Wake After Sleep Onset (WASO)] were compared among
these groups. Because there were no significant differences among those having
actigraphy sleep variables for less than 3 nights compared to those with ≥ 3 nights for
any sleep variable, data from all participants were pooled (Table 3.5).
63
Figure 3.2. Flowchart of sleep actigraphy data.
Table 3.5. Comparison of Actigraphy Sleep Parameters Stratified by Total
Number of Nights
Sleep parameter Participants
having ≥3 nights
(Mean ± SD)
Participants having <3
nights data
(Mean ± SD)
Difference
among groups
(Mean ± SE)
P value
Total sleep time
(min)
398.3 ± 52.5 386.6 ± 79.9 11.6 ± 6.1 0.06
Sleep efficiency
(%)
81.6 ± 8.1 81.2 ± 8.1 0.36 ± 0.6 0.59
Sleep latency
(min)
11.6 ± 8.4 10.2 ± 10.8 1.3 ± 0.9 0.138
Wake After Sleep
Onset (min)
78.6 ± 32.1 78.1 ± 41.8 0.43 ± 3.3 0.89
64
Logistic regression was used to compare characteristics between those participating (n =
529) and those not participating (n = 415) in the actigraphy study. These characteristics
included variables collected at birth (birth weight and mother’s weight) and at 22 years
(age, BMI, waist circumference, smoking, self-reported sleep duration, Table 3.6).
There were no significant differences observed in any birth variables or gender
distribution among those who participated or did not participate in the actigraphy study,
although those who participated in the actigraphy study were 10 months younger, less
likely to smoke and less physically active. There was no significant difference in blood
pressure between the groups, although those that participated in the actigraphy study
tended towards being less obese (P = 0.07).
Table 3.6. Comparative Characteristics of the Raine Study Participants Stratified
by Attendance in the Actigraphy Study
Characteristics Total
Participants
with full
data (n)
Participants
who did the
sleep
actigraphy
study(n)
Participants who
did not do sleep
the actigraphy
study(n)
P value
Birthweight (g) 3319 (944) 3318.4 (529) 3320 (415) 0.954
Mother weight (kg) 59.52 (923) 59.52 (523) 59.51 (400) 0.99
Age 22.16 (945) 22.11 (530) 22.21 (415) 0.024
Gender (F/M) 43/56 (945) 47/53 41/59 0.054
BMI at age 22
(kg/m2 )
25.28 (945) 24.99 (530) 25.64 (415) 0.07
Waist circumference
at age 22 (cm)
83.5 (944) 82.8 (530) 84.5 (414) 0.059
Systolic BP (mmHg) 118.90 (945) 118.44 (530) 119.48 (415) 0.173
Diastolic BP
(mmHg)
67.04 (945) 66.76 (530) 66.96 (415) 0.74
Smoker at age 22, n
[ %]
134 (892)
[15.0]
66 (518)
[12.7]
69 (384)
[18.1]
0.023
Self-reported
duration of sleep
(mins)
480.3 (873) 481 (508) 478 (365) 0.599
Total weekly alcohol
consumption (g/wk)
137.20 (896) 126.68 (520) 151.73 (376) 0.08
Total DAS score 20.28 (845) 20.40 (495) 20.12 (350) 0.83
Physical Activity
(MET-min/wk)
5565 (759) 4700 (457) 6874 (302) < 0.001
Data are presented as mean. Linear regression models for continuous variables.
65
3.4.1 Actigraphy-derived Sleep Variables
There was no significant difference in the number of males and females with actigraphy
data. The average sleep duration (minutes) for males and females combined was 395.6
mins (SD 59.6) and the average sleep efficiency was 81 % (SD 6.2). Mean sleep latency
and wake after sleep onset (WASO) were 14.4 mins (SD 8.4 mins) and 71.2 mins (SD
23.7 mins) respectively. Males had a significantly lower sleep duration compared to
females (P = 0.023, Table 3.7). Sleep efficiency, latency and WASO were not different
between males and females.
Table 3.7. Actigraphy Sleep Characteristics of the Raine Participants
Parameters Total Females (n =
303)
Males (n = 278) P
value
Total Sleep Time
(min)
395.70 ± 59.51 402.29 ± 59.53 388.47 ± 58.75 0.02
Sleep efficiency (%) 81.62 ± 6.22 82.06 ± 6.05 81.13 ± 6.37 0.21
Sleep Latency (min) 14.40 ± 8.42 13.45 ± 7.6 9.19 ± 13.29 0.56
Wake After Sleep
Onset (min)
71.23 ± 23.70 70.52 ± 22.80 72.09 ± 22.09 0.96
Data are presented as mean ± SD. Linear regression models for continuous variables.
3.4.2 Association between Actigraphy Sleep Variables and
Cardiometabolic Profile
3.4.2.1 BMI and Sleep Variables
There was no evidence of any linear association between sleep variables and BMI as
determined by scatterplot and multivariable spline models (Figures 3.3 - 3.6). In
addition based on the results of multivariable spline models, there was no evidence of
any nonlinear relationship observed between BMI and total sleep time, sleep efficiency,
sleep latency and WASO.
Table 3.8 shows the results of linear regression models examining the association
between actigraphy sleep variables and BMI. There was no association between total
sleep time (min) in univariate analysis (model 1) and no gender interaction. The
association remained non-significant after adjusting for gender (Model 2) and additional
inclusion of potential confounders (gender, current smoker, physical activity, alcohol
68
Table 3.8. Linear Regression Models of BMI and Sleep Indices
BMI
Model 1* Model 2** Model 3***
COEFF 95% CI P value COEFF 95% CI P value COEFF 95% CI P value
Total Sleep Time (min) -.0002 -.007, .006 0.95 -.007 -.007, .007 0.95 -.002 -.007, .006 0.91
Sleep efficiency (%) -.011 -.07, .05 0.74 -.016 -.08, .05 0.74 -.016 -.090, .056 0.65
Sleep Latency (min) .003 -.04, .047 0.88 .003 -.04, .04 0.88 .0004 -.04, .04 0.98
Wake After Sleep Onset
(min)
.003 -.01, .01 0.62 .003 -.01, .01 0.62 .002 -.01, .01 0.52
* Model 1: Unadjusted ** Model 2: adjusted for gender.
* **Model 3: Model 2 + physical activity, current smoker, alcohol intake and total depression and anxiety score (DASS). COEFF: Beta correlation coefficient
71
Y axis: Waist circumference (cm). Lowess: Locally Weighted Scatter plot Smoothing. Fitted values: Ordinary
least square regression fit.
Figure 3.10 Scatter plot association between waist circumference (cm) and Wake After Sleep
Onset (WASO); (min).
72
Table 3.9. Linear Regression Models of Waist Circumference and Sleep Indices
Waist Circumference
Model 1* Model 2** Model 3***
COEFF 95% CI P value COEFF 95% CI P value COEFF 95% CI P value
Total Sleep Time (min) .0002 -.01, .02 0.98 .005 -.01, .02 0.61 .002 -.01, .02 0.80
Sleep efficiency (%) -.049 -.21, .11 0.56 -.019 -.18, .14 0.81 -.033 -.21, .14 0.71
Sleep Latency (min) .084 -1.2, 1.4 0.90 -.019 -1.31, 1.27 0.97 .16 -1.14, 1.46 0.80
WASO (min) .011 -.021, .043 0.49 .009 -.02, .04 0.56 .011 -.02, .04 0.53
* Model 1: Unadjusted.
** Model 2: adjusted for gender.
* **Model 3: Model 2 + physical activity, current smoker, alcohol intake and total depression and anxiety score (DASS).
73
3.4.2.3 Blood Pressure, Biochemical Variables and Sleep Variables
Scatterplot showed no evidence of linear relationship between sleep variables and systolic and
diastolic blood pressure or any of the biochemical variables. In addition multivariable spline modes
showed no evidence of any nonlinear relationship between sleep variables and blood pressure or
biochemical variables.
Regression analyses showed no evidence of an association between total sleep time and systolic or
diastolic blood pressure or any of the biochemical variables measured including glucose, insulin,
HOMA-IR, total cholesterol and hs-CRP (Table 3.10). Similarly there were no significant
associations between systolic or diastolic BP or biochemical variables and sleep efficiency (Table
3.11), sleep latency (Table 3.12) or WASO.
74
Table3.10. Linear Regression Models of Total Sleep Time and Blood Pressure and Biochemical Variables
Total Sleep Time (mins)
Variable Model 1* Model 2** Model 3***
COEFF 95% CI P value COEFF 95% CI P value COEFF 95% CI P value
Systolic BP(mmHg) 5.45e-06 -.01, .01 0.99 5.45e-06 -.01, .01 0.99 .009 -.004, .023 0.18
Diastolic BP (mmHg) .011 .002, .021 0. 12 .011 .002, .020 0. 16 .012 .001, .021 0. 1
Glucose (mmol/L) -.0004 -.001,.00007 0.08 -.0003 -.001, .0004 0.36 -.0009 -.0009, .0001 0.08
Insulin● (mU/L) -.00006 0007, .0006 0.85 -.0001 -.0008, .0005 0.67 .00002 -.0006, .0006 0.94
HOMA - IR● -.0002 -.0009, .0004 0.51 -.0002 -.001, .0004 0.44 .0002 -.0003, .0009 0.34
Total cholesterol (mmol/L) .0007 -.0004, .001 0.22 .0005 -.0005, .001 0.31 .0004 -.0007, .001 0.42
hs-CRP● (mg/L) .0001 -.001, .001 0.83 -.001 -.003, .001 0.25 .0002 -.001, .001 0.73
●Log transformed used * Model 1: Unadjusted ** Model 2: adjusted for gender
* **Model 3: Model 2 + physical activity, current smoker, alcohol intake and total depression and anxiety score (DASS)
75
Table 3.11. Linear Regression Models of Sleep Efficiency and Blood Pressure and biochemical variables
Sleep Efficiency (%)
Variable Model 1* Model 2** Model 3***
COEFF 95% CI P value COEFF 95% CI P value COEFF 95% CI P value
Systolic BP (mmHg) -.014 -.14, .11 0.82 .033 -.08, .15 0.59 .031 -.09, .15 0.62
Diastolic BP (mmHg) .072 -.014, .160 0.10 .069 -.018, .156 0.12 .064 -.030, .159 0.18
Glucose (mmol/L) -.005 -.014, .003 0.20 .004 -.013, .003 0.27 -.006 -.014, .002 0.17
Insulin● (mU/L) -.002 -.001, .004 0.47 -.003 -.010, .004 0.40 -.002 -.008, .004 0.50
HOMA - IR● -.002 -.009, .005 0.56 -.002 -.009, .005 0.52 -.002 -.008, .004 0.55
Total Cholesterol (mmol/L) -.007 -.01, .004 0.20 -.008 -.02, .003 0.15 -.01 -.02, .00001 0.5
hs-CRP● (mg/L) -.005 -.02, .01 0.49 -.008 -.02, .007 0.28 -.007 -.02, .008 0.33
●Log transformed used * Model 1: Unadjusted. ** Model 2: adjusted for gender.
* **Model 3: Model 2 + physical activity, current smoker, alcohol intake and total depression and anxiety score (DASS).
76
Table 3.12. Linear Regression Models of Sleep Latency and Blood Pressure and Biochemical Variables
Sleep Latency (mins)
Variable Model 1* Model 2** Model 3***
COEFF 95% CI P value COEFF 95% CI P value COEFF 95% CI P value
Systolic BP (mmHg) .07 -.02 ,.16 0.13 .04 -.04, .12 0.35 .04 -.04, .12 0.34
Diastolic BP (mmHg) .01 -.05, .07 0.75 .01 -.05, .07 0.70 -.002 -.06, .06 0.93
Glucose (mmol/L) .0006 -.004, .005 0.79 -.0001 -.004, .004 0.96 .0001 -.004, .005 0.96
Insulin● (mU/L) .003 -.001, .007 0.16 .003 -.0009, .007 0.12 .003 -.007, .007 0.11
HOMA- IR● .002 -.001, .007 0.22 .003 -.001, .007 0.20 .002 -.001, .007 0.16
Total Cholesterol (mmol/L) -.0007 -.008, .006 0.83 -.0002 -.007, .006 0.94 .001 -.005, .008 0.70
hs-CRP● (mg/L) -.001 -.01, .01 0.86 .0007 -.01, .01 0.91 -.001 -.01, .01 0.82
●Log transformed * Model 1: Unadjusted ** Model 2: adjusted for gender.
***Model 3: Model 2 + physical activity, current smoker, alcohol intake and total depression and anxiety score (DASS).
77
3.4.3 Sleep Duration Categorization
There were 447 participants with actigraphy data from both a weekday and a weekend.
For sleep duration, the difference between weekends and weekdays was 12.00 ± 3.57
minutes (P = 0.001). Adjusting analyses to compensate for weekend vs weekday data
had no effect on the associations between sleep variables, BMI, blood pressure other
biochemical variables.
The participants were divided according to whether they were short sleepers (< 6 hrs),
normal sleepers (≥ 6 hrs) or normal/long sleepers (≥ 8 hrs, Table 3.12) (224). The
percentages of short sleepers and normal/long sleepers were 25.6% (95%CI: 22.2, 29.3),
60.7% (95%CI: 62.8, 70.5) and 7.5% (95%CI: 5.6, 10.0), respectively. As there were
only few long sleepers in this population, normal and long sleepers were combined and
categorised as one group. Males were more likely to be short sleepers (P = 0.012).
Analyses examining associations between socioeconomic variables and sleep categories
showed no evidence of a gender interaction. Among sleep duration categories, there
were no significant differences in socioeconomic status such as post school education,
income or having job (Table 3.13). The odds of unemployment in those that sleeping ≥
6hrs was 0.56 compared to those that slept < 6 hrs (P = 0.06). Lifestyle behaviours
including smoking, physical activity and weekly alcohol consumption were no different
according to sleep duration categories across the two genders (Table 3.14). There were
no significant differences between sleep categories for BMI, waist circumference and
systolic and diastolic BP. There was also no significant difference in distribution of
overweight (OR 0.71, 95%CI: 0.45, 1.12, P = 0.15) and obese participants (OR 0.82, 95
%CI: 0.46, 1.45, P = 0.51) between the sleep categories (Figure 3.11).
78
Table 3.13. General and Cardiometabolic Characteristics According to Sleep
Categories
Sleep Duration Categories
< 6 hrs (149) ≥ 6 hrs (423)
Females Males Females Males
Number
Employed (%)
63
56 (89.5)
86
76
(89.0)
240
208
(87.0)
183
144
(78.7)
Post school
education (%)
30
(55.5)
36
(47.3)
121
(52.3)
90
(51.4)
High income
[>605/wk], (%)
17
(27.0)
40
(44.0)
91
(38)
73
(44)
Current smoker
(%)
9
(14.0)
8
(9.0)
30
(12.5)
30
(16.3)
Total weekly alcohol
consumption (g/wk)*
41.5
(11.1, 97.1)
79.0
(5.9, 206.8)
46.2
(8.3, 110.6)
95.6
(18.4, 212.2)
Physical activity
(MET-min/wk)*
1575
(480, 3279)
3791
(1539, 9036)
2063
(762, 4716)
2786
(1099, 5971)
DASS score* 20 (10, 44) 14 (6, 28) 18 (8, 34) 14 (4, 24)
BMI (kg/m2) 26.5 ± 7.2 24.7 ± 4.0 24.6 ± 5.3 25.0 ± 4.4
Waist circumference (cm) 82.4 ± 16.5 83.8 ± 10.3 79.5 ± 13.2 86.0 ± 11.5
BMI 25-29.9 (kg/m2) 13 (20.6) 28 (32.5) 43 (18.0) 48 (25.1)
BMI >30 )kg/2m( 13 (20.6) 8 (9.3) 34 (14.0) 24 (12.5)
Abdominal obesity, (%) 19 (74.6) 6 (7) 42 (17.5) 19 (10)
Systolic BP (mmHg) 113.5 ± 8.7 122.1 ± 10.2 114.3 ± 9.3 123.1 ± 10.2
Diastolic BP (mmHg) 65.8 ± 6.0 66.17 ± 6.3 68.0 ± 6.9 67.1 ± 7.5
Data presented as mean ± SD for continuous variables and number (proportion) for categorical variables.
*Not-normally distributed variables presented as median (Q1, Q3).
80
and 42.15% was males (P = 0.01). Within gender 31.5% of females had poor sleep
quality while within male gender 24.0% had poor sleep quality.
The socioeconomic, lifestyle, anthropometric and blood pressure characteristics of the
study population stratified by sleep quality are presented in Table 3.15. The
characteristics of study population stratified by sleep quality and sex are also presented
in Supplementary Table S1. Participants with good sleep quality were more likely to be
employed (P = 0.05) and to have a higher income (50.2% vs. 40.5%, P = 0.01). There
was no significant difference in post-school higher education between the high and poor
sleep quality groups (P = 0.43).
With respect to lifestyle behaviours, there was a significantly higher percentage of
smoking in the group with poor sleep quality (P = 0.002). Low sleep quality was also
associated with a higher DASS score compared to high sleep quality (P < 0.001). No
significant difference was observed in the level of alcohol consumption or physical
activity according to sleep quality. Low sleep quality was associated with higher regular
intake of sleeping pill (P < 0.001).
There was a significant difference in mean BMI among sleep quality subgroups (P =
0.01) with more individuals likely to be overweight/obese if they had poor sleep quality
(P = 0.02). There were 129 (22.3% ) overweight and 74 (12.8% ) obese participants in
the high sleep quality group, while in the poor sleep quality group these figures were 59
(26.4% ) and 39 (17.4% ), respectively. Participants with poor sleep quality were more
likely to have central obesity (P = 0.01). There were no significant differences in mean
systolic or diastolic blood pressure or the proportions of individuals with
prehypertension and hypertension between the groups.
Those differences in characteristics that differed between the poor sleep quality
compared with good sleep quality groups were evident in both females and males
(Supplementary Table S1).
81
Table 3.15. Characteristics of the Raine Study Participants According to PSQI
Sleep Quality
PSQI ≤ 5
Good Sleep Quality
(n = 578)
PSQI > 5
Poor Sleep Quality
(n =223)
P value
Females to males ratio, (Female
%)
280/298 (48.4) 129/94 (57.4) 0.01
Employed (%) 322 (85.8) 145 (80.3) 0.05
Post school education (%) 297 (51.7) 108 (48.6) 0.43
High income (%) 273 (50.2) 83 (40.5) 0.01
Current smoker (%) 67 (11.6) 45 (20.3) 0.002
Total weekly alcohol consumption
(g/wk)●
69.3 (11.1, 165.6) 70.3 (11.6, 187) 0.51
Physical activity (MET-
min/week)●
3093 (1455, 6792) 2796 (1142, 5973) 0.79
Weekly intake of sleeping pill (%) 2 (0.3) 26 (11.6) < 0.001
Total DASS score● 12 (4,20) 28 (16, 50) < 0.001
BMI
(kg/m2 )
23.8 (21.7, 26.6) 24.3 (21.4, 28.4) 0.01
Overweight and obese a (%) 203 (35.1) 98 (43.9) 0.02
Waist
(cm)
79.5 (74.0, 88.7) 80.1 (72.4, 90.2) 0.45
Central obesity b (%) 78 (13.5) 45 (20.3) 0.01
SBP
(mmHg)
117.4 (110.8, 125.0) 117.2 (109.6, 123.8) 0.48
DBP
(mmHg)
66.6 (62.2, 71.2) 67.2 (62.2, 71.8) 0.87
Prehypertension & hypertension c
(%)
242 (41.8) 88 (39.4) 0.53
Data are presented as number of participants (%) for categorical variables or median (1rd,
3rdquartiles) for continuous variables.
● Variable not normally distributed a Overweight was defined as (25 kg/m2 > BMI < 30 kg/m2) and obese as BMI > 30 kg/m2. b Central obesity was defined as waist circumference ≥102 cm in males and waist
circumference ≥ 84 cm in females. c Prehypertension was defined as 120 mmHg < SBP ≤ 130 mmHg or 80 mmHg < DBP ≤89
mmHg and hypertension was defined as SBP ≥ 140 mmHg or < DBP ≥ 90 mmHg.
82
Linear regression analyses showed the PSQI score was significantly associated with
BMI in univariate analysis (Coeff, 0.23; 95%CI: 0.05, 0.42; P = 0.012) and after
adjusting for gender (Coeff, 0.24; 95%CI: 0.06, 0.43; P = 0.008). The association
between PSQI score and BMI remained significant (Coeff, 0.19; 95%CI: 0.02, 0.37; P =
0.02) after adjusting for gender and socioeconomic status (i.e. being employed, income)
and lifestyle characteristics (i.e. smoking). There were ten morbidly obese participants
with a BMI in the range 41 to 55 kg/m2, of those four participants with a BMI > 47
kg/m2. While an accurate measure of each individual’s BMI, these 4 measurements
could be considered as outliers in terms of the statistical distribution of BMI in this
particular sample. For this reason, sensitivity analyses were undertaken to examine
whether the association between PSQI score and BMI was influenced by these four
outliers. Linear regression analysis (unadjusted model) was performed by consecutively
removing each measure of BMI in descending order (i.e. between 55 and 47 kg/m2).
By removing BMI > 55 kg/m2, the association between BMI and PSQI score became
weaker (Coeff, 0.17; 95%CI, 0.02, 0.32; P = 0.02). Sequentially removing each of the
four participants with BMI > 47 kg/m2, the association between PSQI score and BMI
was no longer evident (Coeff, 0.14; 95%CI: -0.004, 0.30; P = 0.05). Removing the four
participants with BMI > 47 kg/m2 revealed that these morbidly obese subjects were
responsible for the association between PSQI and BMI. As there were only ten
participants with morbid obesity, there was insufficient power to evaluate this group
separately.
Table 3.16 shows the results of unadjusted and adjusted models of linear regression
analyses between sleep quality and waist circumference and systolic and diastolic blood
pressure. There was no relationship between sleep quality and waist circumference,
systolic and diastolic blood pressure with and without adjustment for gender and BMI.
In summary, the prevalence of poor sleep quality according to PSQI in Raine Study
participants at 22 years of age was 27.8%, with higher prevalence estimates in females.
Participants in the poor sleep quality group were more likely to be smokers,
unemployed and with higher DASS scores. These differences were consistent in both
males and females. No associations were found between sleep quality and BMI, waist
circumference and blood pressure in this study sample.
83
Table 3.16. Linear Regression Analyses: Association Between Sleep Quality and
Waist Circumference, and Systolic and Diastolic Blood Pressure
Model Waist Circumference (cm)
Coeff (SD) P value
Systolic BP (mmHg)
Coeff (SD) P value
Diastolic BP(mmHg)
Coeff (SD) P value
Model 1:
Unadjusted
0.92 (1.13) 0.45 - 0.60
(0.85)
0.48 0.08 (0.54) 0.87
Model 2:
Adjusted for
gender
1.45 (1.11) 0.19 0.46
(0.77)
0.55 0.05 (0.55) 0.92
Model 3:
Adjusted for
gender + BMI
- 0.26 (0.39) 0.50 - 0.13
(0.71)
0.84 - 0.01
(0.54)
0.97
84
Table 3.17. Characteristics of the Raine Study Participants Stratified by Sex and
PSQI Sleep Quality
Females (n = 409) Males (n = 392)
PSQI ≤ 5
Good Sleep
Quality
(n = 280)
PSQI > 5
Poor Sleep
Quality
(n = 129)
PSQI ≤ 5
Good Sleep
Quality
(n = 298)
PSQI > 5
Poor Sleep
Quality
(n=94)
Employed, % 89.2 (85.0, 92.4) 80.6 (72.7, 86.6) 82.5 (77.7, 86.4) 79.7 (70.2, 86.8)
Post school
higher
education, %
54.1 (48.2, 60.0) 51.1 (42.4, 59.8) 49.4 (43.8, 55.1) 45.1 (35.2, 55.5)
High income, % 46.6 (41.4, 51.8) 34.5 (27.6, 42.0) 54.7 (49.2, 60.0) 44.6 (35.2, 54.5)
Current smoker,
%
10.0 (7.3, 13.5) 19.4 (14.2, 25.9) 15.8 (12.4, 20.1) 25.0 (17.9, 33.8)
Total weekly
alcohol
consumption
(g/wk)●
41.7 (6.5, 103.6) 52.7 (11.1, 122.1) 110.4 (27.6, 247) 120.2 (19.4,
225.6)
Physical
activity (MET-
min/wk)●
2453 (996, 5172) 2238 (924, 4644) 4341 (2022, 8693) 4645 (1848,
9918)
Weekly intake
of sleeping pill,
%
0
12.40 (7.6, 19.4) 0.6 (0.1, 2.6) 10.6 (5.7, 18.8)
Total DASS
score●
14 (6, 24) 35 (18, 54) 10 (2, 18) 23 (14, 34)
BMI (kg/m2 ) 24.6 (24.0, 25.1) 26.2 (25.0, 27.4) 25.1 (24.6, 25.6) 25.4 (24.5, 26.4)
Overweight and
obese, % a
31.9 (26.8, 37.4) 43.4 (35.6, 51.4) 38.2 (33.0, 43.7) 45 (35.3, 55.0)
Waist (cm) 79.3 (77.8, 80.8) 81.9 (79.2, 84.5) 85.80 (84.5, 87.0) 86.1 (83.6, 88.6)
Central
Obesity, % b
18.3 (14.3, 23.1) 25.5 (19.1, 33.2) 16.1 (14.3, 22.4) 25.0 (17.8, 33.8)
SBP
(mmHg)
113.6 (112.6,
114.6)
113.7 (112.1,
115.3)
122.7 (121.6,
123.8)
123.7 (121.5,
125.9)
DBP
(mmHg)
67.2 (66.4, 68.0) 66.9 (65.7, 68.0) 66.7 (65.9, 67.5) 67.2 (65.6, 68.8)
Prehypertension
& hypertension,
% c
22.6 (18.1, 27.6) 26.0 (19.5, 33.6) 59.7 (54.2, 65.0) 58 (48.0, 67.4)
Data are presented as mean (95%CI) for continuous variables and proportion (95%CI) for categorical
variables.
● Variable not normally distributed presented as median (1rd, 3rdquartiles).
a Overweight was defined as (25 kg/m2 > BMI < 30 kg/m2) and obese as BMI > 30 kg/m2.
b Central obesity was defined as waist circumference ≥102 cm in males and waist circumference ≥ 84
cm in females.
c Prehypertension was defined as 120mmHg < SBP ≤ 130 mmHg or 80mmHg < DBP ≤89 mmHg and
hypertension was defined as SBP ≥ 140 mmHg or DBP ≥ 90mmHg mmHg.
85
3.6 Relationships Between Obstructive Sleep Apnoea Risk and
Cardiometabolic Risk in the Raine Study Participants at 22 Years
The Berlin Questionnaire is a validated tool used to identify participants with high risk
for Obstructive Sleep Apnea (OSA) (189, 190). It contains three main categories:
Category 1 evaluates sleep and snoring behaviour; Category 2 determines the presence
of daytime sleepiness and Category 3 assesses the patient’s history of obesity (BMI >
30 kg/m2) and/or self-reported hypertension. High risk for OSA is defined when two or
more categories are positive (see Method Chapter).
Overall there were 850 study participants (48.5% females) who completed the Berlin
Questionnaire. According to the criteria defined by the Berlin Questionnaire, there were
125 (14.7%) participants with high risk for OSA (i.e. positive for two or three
categories, Table 3.18). There were 332 (39.1%) participants who were positive for
either Category 1 (snoring) or Category 2 (fatigue/sleepiness), and 67 (7.9%) were
positive for both categories (snoring and fatigue/sleepiness). Amongst those with high
risk for OSA, 45.6% were female and 54.4% were male (P = 0.46). Within each gender,
13.8% of females and 15.5% of males were identified as having high risk for OSA.
Table 3.18. Number and Proportion of Individuals with Positive Berlin Scores for
(i) Overall Score, (ii) Each Category and (iii) Combinations of Individual
Categories.
Categories n (%) 95% CI
Category 1 positive 186 (21.8) 19.2, 24.8
Category 2 positive 280 (32.9) 29.8, 36.1
Category 3 positive 160 (18.8) 16.3, 21.6
Categories 1and 2 positive 67 (7.9) 6.2, 9.9
Categories 1or 2 positive 332 (39.1) 35.8, 42.4
Total Berlin Score positive 125 (14.7) 12.5, 17.2
Data are presented as number of individuals, N (%) and 95% Confidence Interval (95% CI).
Table 3.19 shows the characteristics of study participants stratified by high risk for
OSA. There were no differences between participants at high- compared with low risk
for OSA with respect to gender, social factors (employment, education and income),
alcohol consumption or physical activity. However, high risk for OSA was associated
with a higher BMI, waist circumference and systolic blood pressure. Participants with
86
high risk for OSA were more likely to be smoker, overweight/obese, hypertensive and
having central obesity.
Table 3.19. Characteristics of the Raine Study Participants According to High Risk
for Obstructive Sleep Apnoea
High Risk for OSA
Total (n = 850) No (n = 725) Yes (n = 125) P value
Female to male ratio, (females %) 413/437
(48.5)
356/369
(49.1)
57/68
(45.6)
0.46
Currently employed, % 83.2
(80.4, 85.7)
84.5
(81.4, 87.1)
78.1
(68.5, 85.3)
0.32
Post school education, % 51
(47.4, 54.5)
52
(47.9, 55.8)
43.1
(33.0, 53.8)
0.11
High income, % 47.4
(43.8, 51.1)
47.2
(43.1, 51.3)
57.8
(46.7, 68.1)
0.05
Current smoker, % 17.5
(15.0, 20.3)
14.4
(11.8, 17.4)
28.8
(20.6, 38.8)
< 0.001
Total weekly alcohol consumption
(g/wk)●
71.4
(14.3, 185)
71.3
(16.6, 178)
73.6
(0, 230)
0.24
Physical activity
(MET-min/wk) ●
3279
(1386, 660)
3292
(1386, 6742)
3262
(1485, 6102)
0.25
BMI (kg/m2), mean 25.5
(25.1, 25.9)
24.5
(24.1, 24.8)
31.2
(30.0, 32.4)
< 0.001
Normal weight
[BMI <25 kg/m2], %
59.7
(56.3, 62.9)
66.5
(62.8, 70.0)
20.8
(14.4, 28.9)
< 0.001
Overweight & obese
[BMI≥25 kg/m2], %
40.2
(37.0, 43.6)
33.4
(29.9, 37.1)
79.2
(71.0, 85.5)
< 0.001
Waist circumference (cm) 83.8
(82.9, 84.7)
81.7
(80.8, 82.6)
96.9
(94.0, 99.8)
< 0.001
Central obesity a, % 17.4
(15.0, 20.1)
11.3
(9.1, 14.0)
53.2
(43.6, 62.6)
< 0.001
Systolic BP ( mmHg) 118.8
(118.0, 119.5)
117.9
(117.1, 118.7)
123.6
(121.2, 126.0)
< 0.001
Diastolic BP (mmHg) 67.0
(66.5, 67.5)
66.7
(66.1, 67.2)
68.3
(66.7, 70.0)
0.05
Prehypertension &
Hypertension b, %
43.8
(40.5, 47.2)
42.2
(38.5, 46.0)
55.5
(45.9, 64.7)
0.01
*Data are presented as mean (95%CI) for continuous variables or proportion (95%CI).
● Variable not normally distributed. Data presented as median (1rd, 3rdquartiles) for continuous variables.
a Central obesity was defined as waist circumference ≥102 cm in males and ≥ 84 cm in females.
b Prehypertension was defined as 120mmHg < SBP ≤ 130 mmHg or 80mmHg < DBP ≤ 89 mmHg and
hypertension was defined as SBP ≥ 140 mmHg or DBP ≥ 90mmHg.
87
Table 3.20. Metabolic Variables of the Raine Study Participants According to High
Risk for Obstructive Sleep Apnoea (OSA)
High Risk for OSA
Total
(n = 780)
No
(n = 669)
Yes
(n = 111)
P value
Total Cholesterol (mmol/L) 4.6
(0.8)
4.6
(0.81)
4.6
(0.8)
0.50
Triglycerides (mmol/L) 1.1
(0.5)
1.0
(0.5)
1.2
(0.5)
< 0.001
HDL-C
(mmol/L)
1.3
(0.3)
1.3
(0.3)
1.2
(0.2)
< 0.001
LDL-C
(mmol/L)
2.7
(0.7)
2.7
(0.7)
2.8
(0.6)
0.07
Glucose
(mmol/L)
5.0
(0.6)
5.0
(0.4)
5.1
(1.3)
0.23
Insulin ●
(mU/L)
7.0
(5.4, 9.6)
6.8
(5.2, 9.1)
9.4
(6.7, 13.4)
< 0.001
HOMA,IR ● 1.5
(1.1, 2.1)
1.5
(1.1, 2.0)
2.0
(1.4, 3.0)
< 0.001
hs-CRP ●
(mg/L)
1.0
(0.4, 2.3)
0.8
(0.4, 2.1)
1.9
(0.9, 4.2)
< 0.001
Data are presented as mean (SD).
● Variable not normally distributed. Data presented as median (1rd, 3rd quartiles).
Table 3.20 shows the metabolic variables of the participants stratified by high risk for
OSA. Univariate regression analyses showed that participants with high risk for OSA
had higher triglycerides (P < 0.001) and lower HDL-C levels (P < 0.001). High risk for
OSA was also associated with significantly higher insulin (P<0.001), HOMA-IR
(P<0.001), and hs-CRP levels (P < 0.001).
Using linear regression analyses, high risk for OSA was significantly positively
associated with systolic blood pressure (P < 0.001), triglycerides (P < 0.001), HOMA-
IR (P < 0.001) and hs-CRP (P < 0.001), and inversely associated with HDL-C (P <
0.001), before and after adjustment for gender and lifestyle factors (Table 3.21).
As blood pressure and metabolic variables can be influenced by BMI (i.e. Category 3
measures in the Berlin Questionnaire), univariate and multivariate analyses were further
adjusted for BMI and sensitivity analyses were undertaken.
Tables 3.22 and 3.23 show the results of sensitivity analyses (linear regression models
with and without adjustments for BMI) in participants with one positive category (either
88
snoring or fatigue/sleepiness) and both positive categories (both snoring and
fatigue/sleepiness). Such analyses were not performed for Category 3 due to the focus
of this category on blood pressure and obesity.
Those with only one positive category showed no association with systolic blood
pressure and metabolic variables before or after adjustment for BMI (Table 3.22). In
those with two positive categories, there was a significant inverse association with
HDL-C, and a positive association with HOMA-IR and hs-CRP levels, before
adjustment for BMI. These associations were no longer statistically significant after
adjustment for BMI (Table 3.23).
89
Table 3.21 Linear Regression Analyses Between High Risk for OSA (Total Berlin Questionnaire) and Systolic and Diastolic Blood Pressure
and Metabolic Variables
Model SBP
(mmHg)
DBP
(mmHg)
Triglycerides
(mmol/L)
HDL-C
(mmol/L)
HOMA-IR * hs-CRP*
(mg/L)
Coeff
(95% CI)
P value Coeff
(95% CI)
P value Coeff
(95% CI)
P value Coeff
(95% CI)
P value Coeff
(95% CI)
P value Coeff
(95% CI)
P value
Model 1:
unadjusted
5.3
(3.0, 7.7)
< 0.001 1.5
(0.01, 3.1)
0.05 0.2
(0.1, 0.3)
< 0.001 -0.1
( -0.2, -0.1)
< 0.001 0.3
(0.2, 0.4)
< 0.001 0.7
(0.5, 0.9)
< 0.001
Model 2:
adjusted for
gender
5.0
(2.9, 7.1)
< 0.001 1.5
(0.02, 3.1)
0.05 0.2
(0.1, 0.3)
< 0.001 -0.1
( -0.2, -0.1)
< 0.001 0.3
(0.2, 0.4)
< 0.001 0.7
(0.6, 1.0)
< 0.001
Model 3:
adjusted for
gender +
lifestyle
factors
4.9
(2.7, 7.7)
< 0.001 1.7
( -0.1, 3.6)
0.07 0.2
(0.1, 0.3)
< 0.001 -0.16
( -0.2, -10.0)
< 0.001 0.3
(0.2, 0.4)
< 0.001 0.8
(0.6, 1.0)
< 0.001
Data presented as Coeff (95% CI).
*log transformed.
90
Table 3.22. Linear Regression Analyses Between One Positive Category (Snoring or Sleepiness) and Systolic Blood Pressure and Metabolic
Variables
Model SBP
(mmHg)
Triglycerides
(mmol/L)
HDL-C
(mmol/L)
HOMA-IR * hs-CRP*
(mg/L)
Coeff
(95% CI)
P value Coeff
(95% CI)
P value Coeff
(95% CI)
P value Coeff
(95% CI)
P value Coeff
(95% CI)
P value
Model 1:
unadjusted
0.9
( -0.6, 2.5)
0.2 - 0.005
( -0.07, 0.06)
0.8 0.007
( -0.04, 0.06)
0.7 0.03
( -0.03, 0.1)
0.3 0.06
(-0.1, 0.2)
0.5
Model 2:
adjusted for
gender
0.38
( -1.0, 1.8)
0.6 - 0.01
( -0.07, 0.06)
0.8 0.02
( -0.02, 0.07)
0.4 0.04
( -0.03, 0.1)
0.2 0.01
( -0.08, 0.3)
0.3
Model 3:
adjusted for
gender +
BMI
0.08
( -1.2, 1.4)
0.9 - 0.02
( -0.08, 0.04)
0.5 0.02
( -0.02, 0.07)
0.2 0.02
( -0.04, 0.09)
0.4 0.05
( -0.1, 0.3)
0.5
Data presented as Coeff (95% CI). *Log transformed.
91
Table 3.23. Linear Regression Analyses Between Both Positive Categories (Snoring or Sleepiness) and Systolic Blood Pressure and Metabolic
Variables
Model SBP
(mmHg)
Triglycerides
(mmol/L)
HDL-C
(mmol/L)
HOMA-IR * hs-CRP*
(mg/L)
Coeff
(95% CI)
P value Coeff
(95% CI)
P value Coeff
(95% CI)
P value Coeff
(95% CI)
P value Coeff
(95% CI)
P value
Model 1:
unadjusted
2.1
(-0.7, 2.5)
0.2 0.07
(-0.07, 0.2)
0.3 - 0.11
(-0. 2, -0. 4)
0.002 0.2
(0.07, 0.4)
0.004 0.4
(0.1, 0.7)
0.004
Model 2:
adjusted for
gender
2.3
(-0.2, 5.0)
0.7 0.06
(-0.08, 0.2)
0.8 - 0.11
(-0.2, -0.04)
0.002 0.2
(0.08, 0.4)
0.003 0.4
(0.2, 0.7)
0.001
Model 3:
adjusted for
gender +
BMI
0.6
( 2.0, 3.1)
0.6 -0.04
(-0.1, 0.01)
0.5 - 0.05
(-0.1, 0.008)
0.08 0.09
(-0.03, 0.2)
0.15 0.1
(-0.1, 0.3)
0.32
Data presented as Coeff (95% CI). *Log transformed.
92
Chapter 4:
Discussion and Conclusions
93
The work presented in this thesis represents one of the largest studies evaluating the
associations between various aspects of sleep and cardiometabolic risk factors in young
adults from the general population. It addresses a gap in our understanding of these
relationships in young adults, as the majority of previous studies have focused on older
populations. All measurements were undertaken in the Western Australian Pregnancy
Cohort (Raine) Study, one of the largest successful prospective cohorts of pregnancy,
childhood, adolescence and now early adulthood to be carried out anywhere in the
world.
The Raine Study provided an excellent resource with which to study the associations
between sleep and cardiometabolic factors in young adults, as it has collected a wealth
of data on familial, socioeconomic, lifestyle, anthropometric, biochemical and clinical
factors when the participants were approximately 22 years old. These factors and other
variables were used as potential confounders, strengthening the analyses of the presence
or absence of any cardiometabolic associations by adjusting for many potential
covariates. A unique aspect of the study is the use of objective measures of sleep
through actigraphy in addition to subjective measures, including the PSQI and Berlin
Questionnaires.
4.1 Sleep Duration
Compared with previous epidemiological studies of young adults, the mean sleep
duration was lower in the current study. The average sleep duration in the Raine Study
participants was less than the National Sleep Foundation’s recommendation for 7–9
hours of sleep in those aged 18–25 years (225).
The mean sleep duration of participants was 6.6 hours (396.10 ± 59.5 minutes), with
25.5% classified as short sleepers (< 6 hours). In a study of 658 adults with a mean age
of 24 years, mean sleep duration was 7.7 hours (226). According to the US National
Longitudinal Study of Adolescent Health that evaluated 15,700 individuals aged 13–32
years, the average sleep duration was 8.5 hours (SD 2.6) in those aged 22 years (227).
This latter study also reported short sleep duration (< 6 hours) at a rate of 8.9% in those
aged 19–22 (227). Another study of 4714 participants with a mean age of 27 years
reported mean sleep duration of 6.9 hours (228).
94
There is strong evidence showing a decline in sleep duration in adolescents (participants
aged 13–18 years) caused by puberty onset sleep phase delay and social factors, such as
early school start time (229, 230).
However, it is not clear if sleep duration curtailment continues during early young
adulthood (19–22 years). Two studies have suggested that there is an increase in sleep
duration beginning at age 19, a result of the typical transition from high school,
followed by a decrease in sleep duration at age 22 (226, 227). It has also been suggested
that early young adulthood is a period when individuals possess a greater ability to
choose their own sleep schedule compared with other age periods (227).
Large epidemiological studies have shown variable rates of short sleep duration in
young adult populations. The National Health and Nutrition Examination Survey 2005–
2008 examined 2,391 young adults (aged 20–39) and reported that 14% of individuals
experienced short sleep duration (< 6 hours) (231). Although a study of 17,465
university students aged 17 to 30 years reported that short sleepers comprised 6% of the
cohort (232).
In the current study, the percentage of short sleepers was higher than in previous reports
(227, 229, 232). However, it should be emphasised that all the previous epidemiological
studies in young adults measured sleep using sleep questions and / or questionnaires,
which can be strongly influenced by recall bias.
4.2 Association between Actigraphy-derived Sleep Variables and
Cardiometabolic Risk Factors
The results of the current study showed no association between actigraphy-derived sleep
variables (sleep duration, sleep efficiency, sleep latency and Wake after Sleep Onset
[WASO]) and cardiometabolic risk factors in this young population.
This study is one of only a few to evaluate actigraphy-derived sleep variables in a
healthy young adult population. Generally previous association studies using actigraphy
have not reported sleep latency and WASO, as these measures are known to be
accompanied by large errors in their estimations (215, 216, 233). However, as sleep
latency and WASO are considered as two of the four main variables measured by
actigraphy, both are included in analyses in this study. For this reason and in consistent
95
with previous studies, the Discussion chapter of this thesis focuses mainly on the
association between sleep duration and cardiometabolic factors.
4.2.1 Association between Sleep Duration and Efficiency and Body
Weight
In this study, there was no association between sleep duration and body weight. Further,
there was no significant difference in the distribution of overweight and obese
participants among sleep categories (short vs normal/long sleepers).
Previous systematic reviews examining the association between sleep duration and body
weight in children and adolescents have consistently shown short sleep duration to be
associated with increased body weight (53, 234). In contrast, the association between
sleep duration and body weight in adults has been less consistent, and appears to differ
throughout the different stages of adulthood (53, 234, 235). For example, studies with
larger sample sizes drawn from general populations have shown a negative linear
association in young adults (235-237), a U-shape association in the middle-aged (235,
236, 238) and an absence of any association in the elderly (235, 239-241). The results of
the National Health and Examination Survey (NHANES) of 5607 adults aged 16 and
over showed a linear negative association between sleep duration and higher BMI only
in those aged 18–49, a U-shaped association in middle-aged adults and no evidence of a
linear relationship among older age groups (235). The results of the present study are
contrary to those previous studies reporting a negative linear association between sleep
duration and body weight in young adults.
A potential reason for these discrepant findings is the lack of objective measures of
sleep in most previous studies (234, 242). Only a few small studies have used objective
measures of sleep in a young population (243, 244). One such study of 132 young adults
(mean age of 23 years) measured sleep by actigraphy for 1–2 weeks and found no
association between sleep duration and body weight or adiposity (fat mass and fat-free
mass measured by bioelectrical impedance analysis) (243). Another study of 330 young
women (mean age of 20 years) measured sleep by actigraphy for 7 nights and reported
no association between sleep duration and weight or adiposity (244). Therefore, the
results of the current study, with a substantially larger sample size, agree with these
studies.
96
Several studies have reported that weight gain was less likely to occur if short sleepers
at the time of the study had reported that they had been short sleepers for many years
prior to study commencement (245, 246). Such findings have suggested that as the body
physiologically adjusts itself to the effects of chronically short sleep, such individuals
do not continue to gain weight linearly throughout the period of being short sleepers
(245, 246). Thus, it has been suggested that in order to detect any physiological change,
the start point of the study needs to be around the acute phase of having short sleep (i.e.,
start of the short sleep period). In the current study, this was not accounted for as initial
measurements were obtained when the participants were 22 years of age. In addition, it
has been suggested that having extra longer sleep (recovery) on non-working days can
modify sleep debt (sleep deficiency) and might be an important factor in body weight
regulation (247).
The present study found no relationship between sleep efficiency and body weight.
Currently, only a few studies have investigated the role of actigraphy-derived sleep
efficiency on body weight in young adult populations. The two above-mentioned
studies of young adults reported no association between sleep efficiency and body
weight; however, both studies found an inverse association between sleep efficiency and
fat mass (243, 244). No association between sleep efficiency and any anthropometric
measurement was found in the present study. Such inconsistent results may be due to
the way in which body composition is measured. Specifically, the present study used
BMI as an overall measure of body composition, whereas other studies have used air
displacement plethysmography (244) or bioelectrical impedance (243), both of which
can purportedly distinguish between fat and fat-free mass.
Considering the absence of any association between sleep and body weight in the
current study, and consistent with the findings of the two other studies that have used
objective sleep measures in this age range (243, 244), it seems unlikely that sleep
duration and efficiency are related to body weight in young adults.
4.2.2 Association Between Sleep Duration and Glucose and Insulin
Resistance
This study found no evidence of any associations between actigraphy-based sleep
duration and fasting glucose or insulin resistance in this young adult population.
97
These results contrast with cross-sectional studies that have shown higher levels of
fasting plasma glucose, insulin and insulin resistance (HOMA-IR) among short sleepers
in middle-aged adults (95, 248). In the Quebec Family Study of 740 adults aged 21–64
years, the odds of type 2 diabetes/impaired glucose tolerance was 2.09 (1.34–2.98) for
those with 5–6 hours of sleep, after adjustment for potential confounding variables (95).
In a study of 161 type 2 diabetic participants (119 women, 42 men) with a mean age of
57, sleep duration was a significant predictor of haemoglobin A1c (HbA1c), a key
marker of long term glycaemic control (249). The absence of any association between
sleep and glucose metabolism in this study contrasts with the studies of middle-aged
adults showing an association between sleep duration and increased risk of developing
type 2 diabetes. According to the results of a meta-analysis on seven studies involving
10307 adults, there was no significant difference in the degree of insulin resistance
measured by HOMA-IR between short and long sleepers in those participants without
Obstructive Sleep Apnoea (OSA) or diabetes (250). The majority of studies with
positive relationships between short sleep duration and poor glycaemic control were
carried out in either diabetic patients or individuals with OSA. In addition it has
suggested that sleep recovery during non-working days can improve insulin sensitivity
(251). Consistent with the results of this meta-analysis mentioned, our results also
indicated that there is no association between insulin resistance and sleep duration in a
healthy young population.
4.2.3 Association between Sleep Duration and Blood Pressure
Several studies reported an association between short sleep duration and increased
blood pressure in different age groups, including adolescents and young and middle-
aged adults (104, 252-255), while others report variable associations (256-259). The
present study undertaken in young healthy adults showed no association between sleep
duration and increased blood pressure.
In a study of 331 healthy young adults with a mean age of 22 years, Fujiyama et al.
(254) showed that short sleep duration (5 hours or less) was associated with elevated
24-hour blood pressure. Similarly, a study of 246 adolescents with a mean age of 16
years showed that actigraphy-assessed sleep duration was inversely related to 24-hour
systolic and diastolic blood pressure in adolescents (255). The results of the Coronary
Artery Risk Development in Young Adults (CARDIA) cohort study of 578 adults aged
33–45 years using 3 days of actigraphy-assessed sleep found that short sleep duration
98
predicted higher systolic and diastolic blood pressure levels cross-sectionally and over
five years follow-up after adjustment for a wide range of covariates (104).
Other studies have reported variable associations between short sleep duration and
blood pressure (256-258). For example, the NHANES study of 4810 adults from the
general population found that short sleep duration (≤ 5 hours) was associated with
increased blood pressure only in those aged between 32 and 59 years (257). In a study
of 1214 middle-aged adults (30–54 years), Hall et al. (258) was unable to find any
association between short sleep duration and increased blood pressure. Similarly, the
multi-ethnic Healthy Life in an Urban Setting (HELIUS) Study of 12805 adults aged
18–70 years showed no relationship between short sleep duration (< 7 hours) and
hypertension after adjusting for potential covariates in most ethnic groups, except Turks
(259).
The inconsistent findings between studies may be a consequence of the different
methods used to measure blood pressure (e.g. 24-hour monitoring vs office
measurement) or to the use of different categorisations for short, normal and long sleep
duration. In view of the results from the two largest studies to date—NHANES, which
showed an association only in those aged 32–59 years, and the HELIUS Study, which
showed no relationship—it appears that the participants in current study were too young
for any strong relationship between sleep duration and blood pressure to be evident.
4.2.4 Association Between Sleep Duration and Inflammatory
Markers and Lipids
In this study of Raine participants at 22 years, actigraphy-derived sleep duration showed
no association with hs-CRP and lipid levels.
Previous studies that have examined the association between sleep duration and
inflammatory markers such as CRP have reported inconsistent findings (259-263). For
example, the Cleveland Family Study of 614 adults, each extra hour of habitual sleep
duration (based on self-reported questionnaire) was significantly associated with
increased CRP levels, while each hour of reduction in sleep was associated with no
change in CRP levels (264). In contrast, in the Wisconsin Sleep Cohort Study of 907
Caucasian participants, which evaluated sleep by both questionnaires and PSG, no
association was detected between sleep duration and CRP levels (106). Similarly,
99
Suarez et al. (265) study on 127 healthy young males and females found no association
between sleep duration and CRP levels.
Several studies have shown that CRP levels can be influenced by age, BMI and
cardiometabolic risk factors, including hypertension and diabetes (266, 267). In this
regard, a study of 5,003 middle-aged adults showed that the association between short
sleep duration and CRP levels was attenuated, and disappeared after adjustment for
BMI, systolic blood pressure and cholesterol levels (267). The results of the current
study agree with previous reports that indicate no association between sleep and CRP
levels.
The current study also found no relationship between sleep duration and fasting lipids.
These results are consistent with data from the HELIUS Study of 12,805 adults aged
18–70 years, which reported no association between sleep duration and lipids (259).
Other epidemiological studies examining the association between sleep duration and
lipids have shown inconclusive results. For example, while some studies have reported
an association between short sleep duration and increased odds of high triglycerides
(110, 258), Aora et al (268) have reported an association between high triglycerides and
long sleep duration. Studies examining the relationship between sleep duration and
HDL cholesterol have produced similarly inconsistent results. In a longitudinal study of
8,766 healthy males aged 40–55 years, moderate (5–7hours) to long (> 7 hours) sleep
duration was reported to decrease the risk of having low HDL cholesterol after 6 years
follow-up (111). In contrast, the Korean National Health and Nutrition Examination
Survey of 13609 adults aged 20 and over found no association between short sleep
duration and HDL cholesterol (109).
These inconsistencies results are most likely caused by differences in age distribution of
study subjects, different cut-offs used to define short sleep duration or incomplete
adjustment for important confounding factors, including hypertension, diabetes and
stress level. In general, very few cross-sectional studies have been undertaken to
evaluate the association between sleep duration and lipid profile and those that have
been done have reported inconclusive findings. Further studies are required to better
understand the relationship, if any, between sleep duration and the lipid profile.
100
4.2.5 Summary of Evidence of Association Between Sleep Duration
and Cardiometabolic Risk Factors
The current study of 22-year old Raine participants showed no relationship between
actigraphy-derived sleep duration and cardiometabolic risk factors. Such a finding is
consistent with a study using objective sleep measures in 492 adults aged 35–64 years,
which also showed no association between sleep and cardiometabolic risk factors
between sleep duration and hypertension, obesity and diabetes (269). Another cross-
sectional study of 1524 employees (92% females with a mean age of 36 years)
evaluating sleep duration by using actigraphy, showed that different dimensions of
work-family conflict adversely affect sleep duration and cardiometabolic health (270).
Previous studies that have reported an association between sleep and cardiometabolic
factors have used subjective sleep measures. This is an important consideration as such
measures can be influenced by recall bias, which could confound any reported
associations (26).
Another significant source of variability between study findings is the definition of
short and long sleep duration among studies. For instance, some studies defined normal
sleepers as having sleep duration of > 6, ≥ 7, 7–8 and 7–9 hours (271) and others have
used different categorisations for short, normal and long sleepers (227, 231, 232). The
specific criteria used in the present study for short sleep duration (< 6 hours) or the short
period of monitoring (7 days) compared with other studies (2 weeks) may have
contributed to the lack of association between sleep duration and cardiometabolic risk
factors. Furthermore, the lack of association using sleep duration as a continuous
measure supports the lack of associations among categorical sleep groups.
It is notable that some studies have reported associations between cardiovascular
function and sleep in young adults. In general these studies have used more specific or
more targeted measurements of cardiovascular function. For example, several studies
have shown a relationship between sleep duration on coronary calcification and
circulation (272, 273). In a study of 26 healthy males with a mean age of 29 years,
Sekine et al. (272) showed lower coronary flow circulation after sleep restriction.
Similarly, a longitudinal study of 494 participants aged 35–47 years found that each
additional hour of sleep was associated with a 33% reduction in coronary calcification
(273).
101
Further, changes in 24-hour blood pressure and sympathetic activity have been reported
to occur early with insufficient sleep (274, 275). Thus, it is better to use more sensitive
assessment for vascular function, such as 24-hour blood pressure monitoring, to detect
blood pressure variation, sympathetic activity and heart rate variability. Alternatively,
Doppler ultrasound evaluation could track changes in vascular endothelium.
4.3 Sleep Quality (PSQI Questionnaire) and Socioeconomic, Lifestyle
and Cardiometabolic Risk Factors
Using the PSQI Questionnaire to assess sleep quality, our study showed that 28% of the
Raine participants experienced poor sleep quality, with females more likely to fall into
this category. Participants with poor sleep quality were more likely to be unemployed,
smoke and have higher DASS scores. No association was found between sleep quality
and anthropometric measures, including BMI and waist circumference.
The prevalence of poor sleep quality in our study is consistent with a study of
Mongolian university students that reported 27.8% of the students had poor sleep
quality (276). Consistent with our data, previous studies have shown that females have
poorer sleep quality than males (115, 277). A study of 26,851 members of the general
population living in rural areas (mean age of 38 years) and urban areas (mean age of 42
years) of Hunan, China, showed a higher PSQI score for females than males (277).
Similar to our findings, the results of the BiDirect Study, which included 753
participants with a mean age of 53, showed that poor sleepers (35%) were more likely to
be females (115).
An association between unemployment and poor sleep quality has been reported in
previous studies (278, 279). In a German community sample of 9,284 participants aged
18–80 years, poor sleep quality was twofold higher (OR, 2.11; 95% CI = 1.73 – 2.53;
P < 0.001) in unemployed participants (278). Additionally, Wong et al. (279) reported
poor sleep quality was three times higher in unemployed participants. The results of our
study are consistent with these previous studies that indicate unemployment is
associated with poor sleep quality.
An association between lifestyle factors and sleep quality has been reported in previous
studies; in particular, among those with poor sleep quality there is a higher prevalence
of cigarette smoking (118, 278, 280). Similar to the results of this study, Dugas et al.
(280) study of 405 young adults found that cigarette smokers were more likely to have
102
poor sleep quality. Our results also indicated a higher percentage of smokers in those
with poor sleep quality.
Consistent with previous studies, low sleep quality in our study was associated with
adverse psychological factors, including higher depressive and anxiety symptoms (115,
280). These results highlight that lifestyle factors in young adults associate with poor
sleep quality.
We found no association between poor sleep quality and obesity or a range of
cardiometabolic measures. The association between poor sleep quality and
overweight/obesity has been reported in studies of young populations (244, 281, 282).
However, a study of 702 young college students with a mean age of 21 years found that
PSQI scores were not correlated with anthropometric measures (283). The results of the
present study are in agreement with this; no association was found between poor sleep
quality and anthropometric measures, including BMI and waist.
Our study did not find any association between blood pressure and poor sleep quality.
These results contrast with previous studies that found a positive association between
blood pressure and poor sleep quality (269, 284, 285). However, these findings are
confined to studies with older participants (122, 269, 285). In a study of 5461 Chinese
adults stratified into two groups, with the age of 45 years as the cut-off, poor sleep
quality was associated with hypertension in male subjects of all ages and in female
subjects aged ≤ 45 years (286).
The lack of association in this study could be because of the effects of young age and
the lower rate of hypertension in this age group. Mechanistically, this may be caused by
sympathetic nervous activation accompanying aging, which could be a consequence of
sleep disorders and may contribute to hypertension (287).
4.4 Relationships Between Risk of Obstructive Sleep Apnoea and
Socioeconomic, Lifestyle and Cardiometabolic Risk Factors
This study aimed to evaluate high risk for OSA (by using the Berlin Questionnaire) and
its relationship with cardiometabolic risk factors. The Berlin Questionnaire has three
main categories: sleep/snoring behaviour, daytime sleepiness and obesity or
hypertension.
103
Our data show that 14.7% of the participants were at high risk of OSA, with a similar
distribution between genders. Most of the previous studies examining this relationship
were undertaken on middle-aged or elderly populations, although there are several
studies of young adults (133, 135, 288). The prevalence for high risk of OSA in our
study sample (14.7%) is higher than what has generally been reported for young adult
populations (133, 135, 289). For example, Wosu et al. (135) and Pensuksan et al. (290)
reported that the prevalence of high risk for OSA in young non-Western populations
was between 6% and 9%, respectively. Additionally, a National Sleep Foundation Sleep
in America 2005 poll with 1,506 respondents reported the rate of high risk for OSA was
18% among those aged 18–29 years (289). The likely explanation for the difference in
prevalence rates between studies are the differences in obesity distribution among
Western and non-Western studies. For example, in our study the percentage of high risk
for OSA decreased to 8% when obesity was excluded from the obesity/hypertension
category of the Berlin Questionnaire.
For Raine participants, there was no significant difference in gender distribution among
those with or without high risk for OSA. Similar to our results, a study of 916 college
students with a mean age of 21 found no gender difference among OSA categories
(135). However, current evidence indicates that OSA is more prevalent in males than
females (163), which is likely related to males having more fat deposits in the neck,
trunk and abdomen, making them more susceptible to OSA (290, 291). It is well known
that obesity and central adiposity are strong risk factors for sleep apnoea through
mechanisms such as mechanical effects and the effects of adipokines on airway
neuromuscular control through the central nervous system (291). In our study, the
prevalence of obesity was similar in both genders, and males were less likely to have
abdominal obesity compared with females (20% vs 37%, P < 0.001).
Evaluation of lifestyle variables in relation to risk of OSA showed that only smoking
was associated with high risk of OSA in this young adult population. This finding is
consistent with other studies of similar age populations (135, 290). Wosu et al. (135), in
a study of 916 college students with a mean age of 22, found an association between
high risk for OSA and age and cigarette smoking. Similarly, a study of 2,911 young
adults with a mean age of 20 years reported a relationship between high risk for OSA
and age, male sex and cigarette smoking (290).
104
Cigarette smoking can induce OSA or increase its severity through mechanisms such as
increased upper airway inflammation, alternation in sleep architecture and upper airway
neuromuscular function (292).
The current study found no associations between high risk for OSA and a range of
cardiometabolic risk factors, including blood pressure and biochemical variables. The
prevalence of overweight/obesity and central obesity in those at high risk for OSA was
high in the present study (79% and 53%, respectively). Such findings are consistent
with several previous studies showing a relationship between high risk for OSA and
obesity (135, 290, 293). In a study of 916 college students, Wosu et al. (135) reported
10-fold increased odds of general obesity and 2.7-fold increased odds of central obesity.
Similarly, Qu et al. (137) found in their study of 559 adolescents and young adults with
a mean age of 20 years that those at high risk for OSA had higher BMI and waist
circumferences. Mahboub et al. (146) reported that among 1214 adults with a mean age
of 36, nearly 70% of those at high risk for OSA had BMI ≥ 30 kg/m2, while 70% of
those with low risk for OSA had BMI < 30 kg/m2.
Obesity and increased fat deposits in the neck and abdomen can contribute to OSA
through mechanisms such as reduced pharyngeal lumen size (due to fat deposition in the
airway walls). In those with central obesity, due to increased abdominal pressure, end-
expiratory lung volumes reduce, which contributes to reducing tracheal traction and
increasing collapsibility of the upper airway (294). There is good evidence that OSA
can also cause obesity and metabolic dysfunction (140). Insufficient sleep, sleep
fragmentation and daytime sleepiness can increase food intake through mechanisms
such as alternation in appetite regulatory hormones, physical activity and energy
balance (80, 295-297).
We found no significant association between high risk for OSA and blood pressure
based on our sensitivity analyses that excluded the obesity/hypertension category from
the Berlin Questionnaire and used a classification based on the presence of snoring and
sleepiness. The association between OSA and hypertension could be caused by an
independent effect of OSA on increased blood pressure or obesity-induced hypertension
in OSA (222). However, data from the Sleep Heart Health Study of 6,132 participants
(aged ≥ 40 years) indicates that the association between OSA and hypertension was
only significant (OR, 1.54; 95% CI= 1.02 - 2.31) in obese participants (298). In support
of this hypothesis, longitudinal analyses of the Sleep Heart Health Study with 2,470
105
participants found no statistically significant association between OSA and hypertension
following adjustment for baseline BMI (299). Another study on 2,911 college students
found a significant association only in overweight/obese participants or those with
central obesity (290).
The current data from the Raine participants supports previous reports suggesting
obesity is the main factor contributing to hypertension in individuals with OSA.
Several mechanisms likely explain the association between OSA and increased blood
pressure in obesity. These mechanisms include increased sympathetic activity,
intermittent hypoxia, elevated oxidative stress and inflammatory markers, and increased
angiotensin level, in addition to endothelial and kidney dysfunction (127, 139, 300). As
the independence of the relationship between OSA and hypertension is still unclear,
further research is required to determine whether the association is dependent on
obesity.
The current analysis of the Raine participants at 22 years showed higher levels of
insulin and HOMA-IR in those at high risk for OSA. However, sensitivity analyses
showed the association was no longer present after accounting for obesity. Previous
studies have reported conflicting results regarding the relationship between measures of
insulin resistance and OSA. In a study of 52 young, healthy, men of normal weight aged
18–30 years, OSA was independently associated with a 27% decrease in insulin
sensitivity and a 37% increase in insulin levels (164). Conversely, Kapsimalis et al.
(301), in a study of 67 middle-aged men, showed that after adjustment for obesity there
were no significant differences in insulin and glucose levels or the HOMA-IR index
between participants with or without OSA. Additionally, a case-control study in which
OSA patients and a control group were matched for age, gender and BMI reported no
difference in insulin levels in OSA between the two groups (302). In adolescents, many
of the positive associations between OSA and insulin level and HOMA-IR disappear
after controlling for adiposity and BMI (303). It is likely that these inconsistent results
are due to variations in OSA assessment (e.g., using the PSG or Berlin Questionnaire)
or study populations (e.g., general population, participants having OSA or type 2
diabetic patients with OSA).
There was no association between high risk for OSA and hs-CRP after adjusting for
BMI. Such a finding suggests that the association between OSA and hs-CRP is likely to
be a consequence of the coexistence of obesity in high risk OSA participants.
106
Conversely, several previous reports found an independent relationship between OSA
and increased CRP level after controlling for BMI (181, 304, 305). According to a study
of 1835 adults with a mean age of 58 years, moderate and severe OSA participants had
1.7 and 2.0 times the risk of having higher hs-CRP after adjustment for obesity (181).
Similarly, in a study that compared 22 newly diagnosed OSA patients (mean age of 48
years) with 20 healthy adults (matched for age and BMI), Shamsuzzaman et al. (305)
reported higher CRP levels in OSA participants. However, in agreement with the
findings from the current study, Ryan et al. (306) showed no association between CRP
and OSA in 110 men with a mean age of 40 years. It is likely that any association
between CRP and OSA may be related to increased CRP levels in adipose tissue as a
consequence of obesity (295). Moreover, it should be noted that a positive relationship
between CRP and OSA has, to date, been reported exclusively in older participants.
4.5 Strengths and Limitations
The major strength of this study is the use of data from a large population-based cohort
of 22-year-olds. The study population is representative of the Western Australia
population, with all participants having the same age and similar gender distribution
(307). The large sample size and comprehensive anthropometric, clinical and
socioeconomic and lifestyle information allowed for adjustment of association analyses
for a wide range of confounders. Objective assessment of sleep using actigraphy is
another strength of the study, as most previous studies of large cohorts have used
subjective evaluation of sleep through questionnaires.
There are several limitations to the study. Firstly, its cross-sectional design means that
causality cannot be inferred. Comparing the Raine Study participants who attended the
sleep actigraphy study monitoring with those who did not, showed that those in the
sleep actigraphy study were less physically active and had a smaller waist
circumference. Thus, the presence of selection bias in the actigraphy study is possible.
Secondly, few participants were long sleepers or very short sleepers. A consequence of
this is that it was not possible to categorise by the extremes of sleep and independently
investigate such groups. Thirdly, as body fat was not directly measured in our study
(e.g., via a DEXA scan or impedance), we were unable to assess changes in fat mass
independently of BMI. Fourthly, 24-hr blood pressure was not measured in our study,
so any such variability over the course of day and night, particularly nocturnal blood
pressure dipping, could not be assessed.
107
The advantage of examining young adults is that these participants are less likely to
have co-morbidities that can confound the analyses. However, it is becoming increasing
apparent form previous studies that healthy young individuals might require more
sensitive measurements (than the standard measurements) to detect metabolic impacts.
For instance, Donga et al showed that by using the hyperinsulinaemic euglycemic clamp
in nine healthy lean volunteers, there was impairment in insulin sensitivity (increased
endogenous glucose production) following a night of sleep restriction (4 hours of sleep),
despite their normal fasting insulin and glucose levels (308). Similarly, Harsch et al
showed impaired insulin sensitivity by using the hyperinsulinaemic euglycaemic clamp
in forty OSA patients (309). Therefore, fasting insulin, glucose or HOMA
measurements as used in the current study, may not be sufficiently sensitive for
detecting early changes in glucose metabolism in young healthy adults.
Actigraphy data collected over seven days has been used to evaluate the predicting roles
of a number of parameters including circadian rhythm, diurnal activity, napping and
sleep fragmentation, on cardiometabolic risk factors (310) However, the present study
was unable to collect actigraphy data for seven consecutive days in all participants, thus
the effects of these sleep parameters could not be examined. Similarly, the present study
was unable to obtain complete sets of sleep diary data which limited our ability to
determine correlations between total sleep time derived by sleep diaries with total sleep
time derived actigraphy.
Lastly, there are possible risks of under-adjustment and over-adjustment when using the
Berlin questionnaire to identify OSA and its association with cardiometabolic risk
factors. The current study is limited in diagnosing OSA by the Berlin questionnaire
instead of objective OSA diagnosis. Considering the strong effect of obesity on
metabolism and OSA, it is likely that adjustment for obesity would remove any
potential effect of OSA in statistical models. Therefore, an objective OSA diagnosis is
essential for detecting potential interactions between OSA and obesity.
4.6 Conclusions and Future Directions
This study has shown that among a young adult population sleep characteristics
including objective (actigraphy) sleep measures and subjective sleep measures (sleep
quality questionnaire) and high risk for obstructive sleep disorder are not related to
cardiometabolic risk factors. Further, while being at high risk for OSA as a young adult
is associated with BMI, any subsequent association with cardiometabolic risk is likely
108
to be modulated by obesity and is not yet apparent in these 22-year-olds. Considering
the lack of association in the present study, it is more likely that mechanisms linking
sleep duration to cardiometabolic risk factors have varying effects at different stages of
life.
The results of this study lead to several recommendations for future studies. Firstly, we
suggest evaluating this association through longitudinal studies. As the exposure to
shorter sleep varies across days, months and years, a snapshot view may not be
indicative of sleep patterns over time. It has been suggested that metabolic
consequences of sleep loss can be resolved with recovery sleep (311). Therefore,
targeting habitual short sleepers and evaluating their cardiometabolic risk factors is
recommended.
The current study has few participants who are very short sleepers or long sleepers.
Targeting and recruiting participants who are long, short or very short sleepers would be
helpful to investigate the cardiometabolic effects of sleep duration.
Short sleepers are a heterogenous group, comprising those who choose to be short
sleepers to meet their personal demands, those who are natural short sleepers and those
with sleep impairment. As previous studies have not evaluated these characteristics of
short sleepers, evaluating such characteristics and their effects on cardiometabolic
health is recommended in future studies.
Future studies would also benefit by the inclusion of anthropometric assessments that
measure fat mass, such as DEXA scans. Additionally, the Berlin Questionnaire is
limited so far as it only estimates the risk of OSA. Future studies should consider using
polysomnography, which is the gold standard method for diagnosing OSA. Such
polysomnography based measures would provide a better estimate of the association
between OSA, its severity and its relation to cardiometabolic risk factors.
109
References
1. Rangaraj VR, Knutson KL. Association between sleep deficiency and cardiometabolic disease: implications for health disparities. Sleep Med. 2016;18:19-35.
2. Colagiuri S, Lee CM, Colagiuri R, Magliano D, Shaw JE, Zimmet PZ, et al. The cost of overweight and obesity in Australia. Med J Aust. 2010;192(5):260-4.
3. Dabass A, Talbott EO, Rager JR, Marsh GM, Venkat A, Holguin F, et al. Systemic inflammatory markers associated with cardiovascular disease and acute and chronic exposure to fine particulate matter air pollution (PM2.5) among US NHANES adults with metabolic syndrome. Environmental research. 2018;161:485-91.
4. Savadatti SS, Bell EM, Gates MA, Hosler AS, Yucel RM, Misra R. Metabolic Syndrome Among Asian Indians in the United States. Journal of public health management and practice : JPHMP. 2018.
5. Lone S, Lone K, Khan S, Pampori RA. Assessment of metabolic syndrome in Kashmiri population with type 2 diabetes employing the standard criteria's given by WHO, NCEPATP III and IDF. Journal of epidemiology and global health. 2017;7(4):235-9.
6. Cozma A, Orasan O, Sampelean D, Fodor A, Vlad C, Negrean V, et al. Endothelial dysfunction in metabolic syndrome. Romanian journal of internal medicine = Revue roumaine de medecine interne. 2009;47(2):133-40.
7. Mottillo S, Filion KB, Genest J, Joseph L, Pilote L, Poirier P, et al. The Metabolic Syndrome and Cardiovascular Risk: A Systematic Review and Meta-Analysis. Journal of the American College of Cardiology. 2010;56(14):1113-32.
8. Raitakari OT, Juonala M, Kahonen M, Taittonen L, Laitinen T, Maki-Torkko N, et al. Cardiovascular risk factors in childhood and carotid artery intima-media thickness in adulthood: the Cardiovascular Risk in Young Finns Study. JAMA. 2003;290(17):2277-83.
9. Rose G. Incubation period of coronary heart disease. Br Med J (Clin Res Ed). 1982;284(6329):1600-1.
10. Vargas PA. The Link Between Inadequate Sleep and Obesity in Young Adults. Curr Obes Rep. 2016;5(1):38-50.
11. Grandner MA, Alfonso-Miller P, Fernandez-Mendoza J, Shetty S, Shenoy S, Combs D. Sleep: important considerations for the prevention of cardiovascular disease. Curr Opin Cardiol. 2016;31(5):551-65.
12. Depner CM, Stothard ER, Wright KP, Jr. Metabolic consequences of sleep and circadian disorders. Curr Diab Rep. 2014;14(7):507.
13. Wheaton AG, Perry GS, Chapman DP, McKnight-Eily LR, Presley-Cantrell LR, Croft JB. Relationship between body mass index and perceived insufficient sleep among U.S. adults: an analysis of 2008 BRFSS data. BMC Public Health. 2011;11:295.
14. Bin YS, Marshall NS, Glozier N. Secular trends in adult sleep duration: a systematic review. Sleep medicine reviews. 2012;16(3):223-30.
15. Mesarwi O, Polak J, Jun J, Polotsky VY. Sleep disorders and the development of insulin resistance and obesity. Endocrinol Metab Clin North Am. 2013;42(3):617-34.
16. Siegel JM. Clues to the functions of mammalian sleep. Nature. 2005;437(7063):1264-71.
17. Carskadan MA, Decement WC. Normal human sleep. In: Kryger MH, Roth T, Decement WC., editors. Priciple and practice of sleep medicine. 4nd ed. Philadelphia: Saunders; 2005.
18. Blackwell T, Redline S, Ancoli-Israel S, Schneider JL, Surovec S, Johnson NL, et al. Comparison of sleep parameters from actigraphy and polysomnography in older women: the SOF study. Sleep. 2008;31(2):283-91.
19. Patil SP. What every clinician should know about polysomnography. Respir Care. 2010;55(9):1179-95.
110
20. Park JG, Ramar K, Olson EJ. Updates on definition, consequences, and management of obstructive sleep apnea. Mayo Clin Proc. 2011;86(6):549-54; quiz 54-5.
21. Practice parameters for the use of actigraphy in the clinical assessment of sleep disorders. American Sleep Disorders Association. Sleep. 1995;18(4):285-7.
22. Marino M, Li Y, Rueschman MN, Winkelman JW, Ellenbogen JM, Solet JM, et al. Measuring sleep: accuracy, sensitivity, and specificity of wrist actigraphy compared to polysomnography. Sleep. 2013;36(11):1747-55.
23. Martin JL, Hakim AD. Wrist actigraphy. Chest. 2011;139(6):1514-27. 24. Jean-Louis G, Kripke DF, Cole RJ, Assmus JD, Langer RD. Sleep detection with an
accelerometer actigraph: comparisons with polysomnography. Physiol Behav. 2001;72(1-2):21-8.
25. de Souza L, Benedito-Silva AA, Pires ML, Poyares D, Tufik S, Calil HM. Further validation of actigraphy for sleep studies. Sleep. 2003;26(1):81-5.
26. Lauderdale DS, Knutson KL, Yan LL, Liu K, Rathouz PJ. Self-reported and measured sleep duration: how similar are they? Epidemiology. 2008;19(6):838-45.
27. Grutsch JF, Wood PA, Du-Quiton J, Reynolds JL, Lis CG, Levin RD, et al. Validation of actigraphy to assess circadian organization and sleep quality in patients with advanced lung cancer. J Circadian Rhythms. 2011;9:4.
28. Lockley SW, Skene DJ, Arendt J. Comparison between subjective and actigraphic measurement of sleep and sleep rhythms. J Sleep Res. 1999;8(3):175-83.
29. Cespedes EM, Hu FB, Redline S, Rosner B, Alcantara C, Cai J, et al. Comparison of Self-Reported Sleep Duration With Actigraphy: Results From the Hispanic Community Health Study/Study of Latinos Sueno Ancillary Study. Am J Epidemiol. 2016;183(6):561-73.
30. Watson PF, Petrie A. Method agreement analysis: a review of correct methodology. Theriogenology. 2010;73(9):1167-79.
31. Girschik J, Fritschi L, Heyworth J, Waters F. Validation of self-reported sleep against actigraphy. J Epidemiol. 2012;22(5):462-8.
32. Van Den Berg JF, Van Rooij FJ, Vos H, Tulen JH, Hofman A, Miedema HM, et al. Disagreement between subjective and actigraphic measures of sleep duration in a population-based study of elderly persons. J Sleep Res. 2008;17(3):295-302.
33. Buysse DJ, Reynolds CF, 3rd, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res. 1989;28(2):193-213.
34. Mollayeva T, Thurairajah P, Burton K, Mollayeva S, Shapiro CM, Colantonio A. The Pittsburgh sleep quality index as a screening tool for sleep dysfunction in clinical and non-clinical samples: A systematic review and meta-analysis. Sleep medicine reviews. 2016;25:52-73.
35. Grandner MA, Kripke DF, Yoon IY, Youngstedt SD. Criterion validity of the Pittsburgh Sleep Quality Index: Investigation in a non-clinical sample. Sleep Biol Rhythms. 2006;4(2):129-39.
36. Netzer NC, Stoohs RA, Netzer CM, Clark K, Strohl KP. Using the Berlin Questionnaire to identify patients at risk for the sleep apnea syndrome. Ann Intern Med. 1999;131(7):485-91.
37. Chiu HY, Chen PY, Chuang LP, Chen NH, Tu YK, Hsieh YJ, et al. Diagnostic accuracy of the Berlin questionnaire, STOP-BANG, STOP, and Epworth sleepiness scale in detecting obstructive sleep apnea: A bivariate meta-analysis. Sleep medicine reviews. 2017;36:57-70.
38. Senaratna CV, Perret JL, Matheson MC, Lodge CJ, Lowe AJ, Cassim R, et al. Validity of the Berlin questionnaire in detecting obstructive sleep apnea: A systematic review and meta-analysis. Sleep medicine reviews. 2017;36:116-24.
39. Karakoc O, Akcam T, Genc H, Yetkin S, Piskin B, Gerek M. Use of the Berlin Questionnaire to screen at-risk patients for obstructive sleep apnea. B-ent. 2014;10(1):21-5.
111
40. Tan A, Yin JD, Tan LW, van Dam RM, Cheung YY, Lee CH. Using the Berlin Questionnaire to Predict Obstructive Sleep Apnea in the General Population. Journal of clinical sleep medicine : JCSM : official publication of the American Academy of Sleep Medicine. 2017;13(3):427-32.
41. Locard E, Mamelle N, Billette A, Miginiac M, Munoz F, Rey S. Risk factors of obesity in a five year old population. Parental versus environmental factors. Int J Obes Relat Metab Disord. 1992;16(10):721-9.
42. von Kries R, Toschke AM, Wurmser H, Sauerwald T, Koletzko B. Reduced risk for overweight and obesity in 5- and 6-y-old children by duration of sleep--a cross-sectional study. Int J Obes Relat Metab Disord. 2002;26(5):710-6.
43. Chaput JP, Brunet M, Tremblay A. Relationship between short sleeping hours and childhood overweight/obesity: results from the 'Quebec en Forme' Project. Int J Obes (Lond). 2006;30(7):1080-5.
44. Ben Slama F, Achour A, Belhadj O, Hsairi M, Oueslati M, Achour N. [Obesity and life style in a population of male school children aged 6 to 10 years in Ariana (Tunisia)]. Tunis Med. 2002;80(9):542-7.
45. Giugliano R, Carneiro EC. [Factors associated with obesity in school children]. J Pediatr (Rio J). 2004;80(1):17-22.
46. Sekine M, Yamagami T, Handa K, Saito T, Nanri S, Kawaminami K, et al. A dose-response relationship between short sleeping hours and childhood obesity: results of the Toyama Birth Cohort Study. Child Care Health Dev. 2002;28(2):163-70.
47. Padez C, Mourao I, Moreira P, Rosado V. Prevalence and risk factors for overweight and obesity in Portuguese children. Acta Paediatr. 2005;94(11):1550-7.
48. Gupta NK, Mueller WH, Chan W, Meininger JC. Is obesity associated with poor sleep quality in adolescents? Am J Hum Biol. 2002;14(6):762-8.
49. Benefice E, Garnier D, Ndiaye G. Nutritional status, growth and sleep habits among Senegalese adolescent girls. European journal of clinical nutrition. 2004;58(2):292-301.
50. Chen MY, Wang EK, Jeng YJ. Adequate sleep among adolescents is positively associated with health status and health-related behaviors. BMC Public Health. 2006;6:59.
51. Knutson KL. Sex differences in the association between sleep and body mass index in adolescents. The Journal of Pediatrics. 2005;147(6):830-4.
52. Cheng HL, Amatoury M, Steinbeck K. Energy expenditure and intake during puberty in healthy nonobese adolescents: a systematic review. The American Journal of Clinical Nutrition. 2016;104(4):1061-74.
53. Patel SR, Hu FB. Short sleep duration and weight gain: a systematic review. Obesity (Silver Spring). 2008;16(3):643-53.
54. Taveras EM, Rifas-Shiman SL, Oken E, Gunderson EP, Gillman MW. Short sleep duration in infancy and risk of childhood overweight. Archives of Pediatrics & Adolescent Medicine.. 2008;162(4):305-11.
55. Snell EK, Adam EK, Duncan GJ. Sleep and the body mass index and overweight status of children and adolescents. Child Development. 2007;78(1):309-23.
56. Reilly JJ, Armstrong J, Dorosty AR, Emmett PM, Ness A, Rogers I, et al. Early life risk factors for obesity in childhood: cohort study. BMJ. 2005;330(7504):1357.
57. Agras WS, Hammer LD, McNicholas F, Kraemer HC. Risk factors for childhood overweight: a prospective study from birth to 9.5 years. The Journal of Pediatrics. 2004;145(1):20-5.
58. Kripke DF, Garfinkel L, Wingard DL, Klauber MR, Marler MR. Mortality associated with sleep duration and insomnia. Archives of General Psychiatry. 2002;59(2):131-6.
59. Ford ES, Li C, Wheaton AG, Chapman DP, Perry GS, Croft JB. Sleep duration and body mass index and waist circumference among U.S. adults. Obesity (Silver Spring). 2014;22(2):598-607.
60. Tamakoshi A, Ohno Y, Group JS. Self-reported sleep duration as a predictor of all-cause mortality: results from the JACC study, Japan. Sleep. 2004;27(1):51-4.
112
61. Tuomilehto H, Peltonen M, Partinen M, Seppa J, Saaristo T, Korpi-Hyovalti E, et al. Sleep duration is associated with an increased risk for the prevalence of type 2 diabetes in middle-aged women - The FIN-D2D survey. Sleep Medicine. 2008;9(3):221-7.
62. Liu Y, Wheaton AG, Chapman DP, Croft JB. Sleep duration and chronic diseases among U.S. adults age 45 years and older: evidence from the 2010 Behavioral Risk Factor Surveillance System. Sleep. 2013;36(10):1421-7.
63. Gottlieb DJ, Punjabi NM, Newman AB, Resnick HE, Redline S, Baldwin CM, et al. Association of sleep time with diabetes mellitus and impaired glucose tolerance. Archives of Internal Medicine. 2005;165(8):863-7.
64. Kim M, Sasai H, Kojima N, Kim H. Objectively measured night-to-night sleep variations are associated with body composition in very elderly women. Journal of Sleep Research. 2015;24(6):639-47.
65. Norton MC, Eleuteri S, Cerolini S, Ballesio A, Conte SC, Falaschi P, et al. Is poor sleep associated with obesity in older adults? A narrative review of the literature. Eat and Weight Disorders. 2017.
66. Chaput JP, Lord C, Aubertin-Leheudre M, Dionne IJ, Khalil A, Tremblay A. Is overweight/obesity associated with short sleep duration in older women? Aging Clinical and Experimental Research. 2007;19(4):290-4.
67. Patel SR, Malhotra A, White DP, Gottlieb DJ, Hu FB. Association between reduced sleep and weight gain in women. American Journal of Epidemiology. 2006;164(10):947-54.
68. Nishiura C, Hashimoto H. A 4-year study of the association between short sleep duration and change in body mass index in Japanese male workers. Journal of Epidemiology. 2010;20(5):385-90.
69. Gangwisch JE, Malaspina D, Boden-Albala B, Heymsfield SB. Inadequate sleep as a risk factor for obesity: analyses of the NHANES I. Sleep. 2005;28(10):1289-96.
70. Lauderdale DS, Knutson KL, Rathouz PJ, Yan LL, Hulley SB, Liu K. Cross-sectional and longitudinal associations between objectively measured sleep duration and body mass index: the CARDIA Sleep Study. American Journal of Epidemiology. 2009;170(7):805-13.
71. Chaput JP, Despres JP, Bouchard C, Tremblay A. The association between sleep duration and weight gain in adults: a 6-year prospective study from the Quebec Family Study. Sleep. 2008;31(4):517-23.
72. Lopez-Garcia E, Faubel R, Leon-Munoz L, Zuluaga MC, Banegas JR, Rodriguez-Artalejo F. Sleep duration, general and abdominal obesity, and weight change among the older adult population of Spain. The American Journal of Clinical Nutrition. 2008;87(2):310-6.
73. Hairston KG, Bryer-Ash M, Norris JM, Haffner S, Bowden DW, Wagenknecht LE. Sleep duration and five-year abdominal fat accumulation in a minority cohort: the IRAS family study. Sleep. 2010;33(3):289-95.
74. Spiegel K, Tasali E, Penev P, Van Cauter E. Brief communication: Sleep curtailment in healthy young men is associated with decreased leptin levels, elevated ghrelin levels, and increased hunger and appetite. Annals of Internal Medicine. 2004;141(11):846-50.
75. Lucassen EA, Rother KI, Cizza G. Interacting epidemics? Sleep curtailment, insulin resistance, and obesity. Annals of the New York Academy of Sciences. 2012;1264:110-34.
76. Guilleminault C, Powell NB, Martinez S, Kushida C, Raffray T, Palombini L, et al. Preliminary observations on the effects of sleep time in a sleep restriction paradigm. Sleep Med. 2003;4(3):177-84.
77. Nedeltcheva AV, Kilkus JM, Imperial J, Kasza K, Schoeller DA, Penev PD. Sleep curtailment is accompanied by increased intake of calories from snacks. American Journal of Clinical Nutrition. 2009;89(1):126-33.
78. Calvin AD, Carter RE, Adachi T, Macedo PG, Albuquerque FN, van der Walt C, et al. Effects of experimental sleep restriction on caloric intake and activity energy expenditure. Chest. 2013;144(1):79-86.
113
79. Benedict C, Hallschmid M, Lassen A, Mahnke C, Schultes B, Schioth HB, et al. Acute sleep deprivation reduces energy expenditure in healthy men. American Journal of Clinical Nutrition. 2011;93(6):1229-36.
80. Markwald RR, Melanson EL, Smith MR, Higgins J, Perreault L, Eckel RH, et al. Impact of insufficient sleep on total daily energy expenditure, food intake, and weight gain. Proceedings of the National Academy of Sciences of the United States of America. 2013;110(14):5695-700.
81. Spaeth AM, Dinges DF, Goel N. Effects of Experimental Sleep Restriction on Weight Gain, Caloric Intake, and Meal Timing in Healthy Adults. Sleep. 2013;36(7):981-90.
82. St-Onge MP, Roberts AL, Chen J, Kelleman M, O'Keeffe M, RoyChoudhury A, et al. Short sleep duration increases energy intakes but does not change energy expenditure in normal-weight individuals. American Journal of Clinical Nutrition. 2011;94(2):410-6.
83. Grandner MA, Jackson N, Gerstner JR, Knutson KL. Dietary nutrients associated with short and long sleep duration. Data from a nationally representative sample. Appetite. 2013;64:71-80.
84. Grandner MA, Kripke DF, Naidoo N, Langer RD. Relationships among dietary nutrients and subjective sleep, objective sleep, and napping in women. Sleep Medicine. 2010;11(2):180-4.
85. Boyle PJ, Nagy RJ, O'Connor AM, Kempers SF, Yeo RA, Qualls C. Adaptation in brain glucose uptake following recurrent hypoglycemia. Proceedings of the National Academy of Sciences of the United States of America. 1994;91(20):9352-6.
86. Maquet P. Positron emission tomography studies of sleep and sleep disorders. Journal of Neurology. 1997;244(4 Suppl 1):S23-8.
87. Nofzinger EA, Buysse DJ, Miewald JM, Meltzer CC, Price JC, Sembrat RC, et al. Human regional cerebral glucose metabolism during non-rapid eye movement sleep in relation to waking. Brain. 2002;125(Pt 5):1105-15.
88. Holl RW, Hartman ML, Veldhuis JD, Taylor WM, Thorner MO. Thirty-second sampling of plasma growth hormone in man: correlation with sleep stages. The Journal of clinical Endocrinology and Metabolism. 1991;72(4):854-61.
89. Spiegel K, Leproult R, Van Cauter E. Impact of sleep debt on metabolic and endocrine function. Lancet. 1999;354(9188):1435-9.
90. Nuyujukian DS, Beals J, Huang H, Johnson A, Bullock A, Manson SM, et al. Sleep Duration and Diabetes Risk in American Indian and Alaska Native Participants of a Lifestyle Intervention Project. Sleep. 2016;39(11):1919-26.
91. Matthews KA, Dahl RE, Owens JF, Lee L, Hall M. Sleep duration and insulin resistance in healthy black and white adolescents. Sleep. 2012;35(10):1353-8.
92. Broussard JL, Ehrmann DA, Van Cauter E, Tasali E, Brady MJ. Impaired insulin signaling in human adipocytes after experimental sleep restriction: a randomized, crossover study. Annals of Internal Medicine. 2012;157(8):549-57.
93. Nedeltcheva AV, Kessler L, Imperial J, Penev PD. Exposure to recurrent sleep restriction in the setting of high caloric intake and physical inactivity results in increased insulin resistance and reduced glucose tolerance. The Journal of clinical Endocrinology and Metabolism. 2009;94(9):3242-50.
94. Klingenberg L, Chaput JP, Holmback U, Visby T, Jennum P, Nikolic M, et al. Acute Sleep Restriction Reduces Insulin Sensitivity in Adolescent Boys. Sleep. 2013;36(7):1085-90.
95. Chaput JP, Despres JP, Bouchard C, Tremblay A. Association of sleep duration with type 2 diabetes and impaired glucose tolerance. Diabetologia. 2007;50(11):2298-304.
96. Gangwisch JE, Heymsfield SB, Boden-Albala B, Buijs RM, Kreier F, Pickering TG, et al. Sleep duration as a risk factor for diabetes incidence in a large US sample. Sleep. 2007;30(12):1667-73.
97. Ayas NT, White DP, Al-Delaimy WK, Manson JE, Stampfer MJ, Speizer FE, et al. A prospective study of self-reported sleep duration and incident diabetes in women. Diabetes care. 2003;26(2):380-4.
114
98. Yaggi HK, Araujo AB, McKinlay JB. Sleep duration as a risk factor for the development of type 2 diabetes. Diabetes care. 2006;29(3):657-61.
99. St-Onge MP, Grandner MA, Brown D, Conroy MB, Jean-Louis G, Coons M, et al. Sleep Duration and Quality: Impact on Lifestyle Behaviors and Cardiometabolic Health: A Scientific Statement From the American Heart Association. Circulation. 2016;134(18):e367-e86.
100. Cappuccio FP, D'Elia L, Strazzullo P, Miller MA. Quantity and quality of sleep and incidence of type 2 diabetes: a systematic review and meta-analysis. Diabetes care. 2010;33(2):414-20.
101. Holliday EG, Magee CA, Kritharides L, Banks E, Attia J. Short sleep duration is associated with risk of future diabetes but not cardiovascular disease: a prospective study and meta-analysis. PLoS One. 2013;8(11):e82305.
102. Jackowska M, Steptoe A. Sleep and future cardiovascular risk: prospective analysis from the English Longitudinal Study of Ageing. Sleep Medicine. 2015;16(6):768-74.
103. Buxton OM, Marcelli E. Short and long sleep are positively associated with obesity, diabetes, hypertension, and cardiovascular disease among adults in the United States. Social Science & Medicine. 2010;71(5):1027-36.
104. Knutson KL, Van Cauter E, Rathouz PJ, Yan LL, Hulley SB, Liu K, et al. Association between sleep and blood pressure in midlife: the CARDIA sleep study. Archives of Internal Medicine. 2009;169(11):1055-61.
105. Ridker PM, Hennekens CH, Buring JE, Rifai N. C-reactive protein and other markers of inflammation in the prediction of cardiovascular disease in women. The New England journal of medicine. 2000;342(12):836-43.
106. Taheri S, Austin D, Lin L, Nieto FJ, Young T, Mignot E. Correlates of serum C-reactive protein (CRP)--no association with sleep duration or sleep disordered breathing. Sleep. 2007;30(8):991-6.
107. Miller MA, Kandala N-B, Kivimaki M, Kumari M, Brunner EJ, Lowe G, et al. Gender differences in the cross-sectional relationships between sleep duration and markers of inflammation: Whitehall II study. Sleep. 2009;32(7):857-64.
108. Fodor G. Primary prevention of CVD: treating dyslipidaemia. BMJ Clinical Evidence. 2008;2008.
109. Shin HY, Kang G, Kim SW, Kim JM, Yoon JS, Shin IS. Associations between sleep duration and abnormal serum lipid levels: data from the Korean National Health and Nutrition Examination Survey (KNHANES). Sleep Medicine. 2016;24:119-23.
110. Kaneita Y, Uchiyama M, Yoshiike N, Ohida T. Associations of usual sleep duration with serum lipid and lipoprotein levels. Sleep. 2008;31(5):645-52.
111. Kinuhata S, Hayashi T, Sato KK, Uehara S, Oue K, Endo G, et al. Sleep duration and the risk of future lipid profile abnormalities in middle-aged men: the Kansai Healthcare Study. Sleep Medicine.. 2014;15(11):1379-85.
112. Okubo N, Matsuzaka M, Takahashi I, Sawada K, Sato S, Akimoto N, et al. Relationship between self-reported sleep quality and metabolic syndrome in general population. BMC Public Health. 2014;14:562.
113. Hung HC, Yang YC, Ou HY, Wu JS, Lu FH, Chang CJ. The association between self-reported sleep quality and overweight in a Chinese population. Obesity (Silver Spring). 2013;21(3):486-92.
114. Bidulescu A, Din-Dzietham R, Coverson DL, Chen Z, Meng YX, Buxbaum SG, et al. Interaction of sleep quality and psychosocial stress on obesity in African Americans: the Cardiovascular Health Epidemiology Study (CHES). BMC Public Health. 2010;10:581.
115. Rahe C, Czira ME, Teismann H, Berger K. Associations between poor sleep quality and different measures of obesity. Sleep Medicine.. 2015;16(10):1225-8.
116. Gildner TE, Liebert MA, Kowal P, Chatterji S, Josh Snodgrass J. Sleep duration, sleep quality, and obesity risk among older adults from six middle-income countries: findings from the study on global AGEing and adult health (SAGE). American Journal of Human Biology. 2014;26(6):803-12.
115
117. Tom SE, Berenson AB. Associations between poor sleep quality and psychosocial stress with obesity in reproductive-age women of lower socioeconomic status. Women's Health Issues. 2013;23(5):e295-e300.
118. Fatima Y, Doi SA, Mamun AA. Sleep quality and obesity in young subjects: a meta-analysis. Obesity Reviews. 2016;17(11):1154-66.
119. Ji-Rong Y, Hui W, Chang-Quan H, Bi-Rong D. Association between sleep quality and arterial blood pressure among Chinese nonagenarians/centenarians. Medical Science Monitor. 2012;18(3):PH36-42.
120. Erden I, Erden EC, Ozhan H, Basar C, Aydin M, Dumlu T, et al. Poor-quality sleep score is an independent predictor of nondipping hypertension. Blood Pressure Monitoring. 2010;15(4):184-7.
121. Fiorentini A, Valente R, Perciaccante A, Tubani L. Sleep's quality disorders in patients with hypertension and type 2 diabetes mellitus. International Journal of Cardiology. 2007;114(2):E50-2.
122. Liu RQ, Qian Z, Trevathan E, Chang JJ, Zelicoff A, Hao YT, et al. Poor sleep quality associated with high risk of hypertension and elevated blood pressure in China: results from a large population-based study. Hypertension Research. 2016;39(1):54-9.
123. Jackson CL, Redline S, Emmons KM. Sleep as a potential fundamental contributor to disparities in cardiovascular health. Annual Review of Public Health. 2015;36:417-40.
124. Huang YS, Guilleminault C. Pediatric Obstructive Sleep Apnea: Where Do We Stand? Advances in oto-rhino-laryngology. 2017;80:136-44.
125. Amaddeo A, de Sanctis L, Olmo Arroyo J, Giordanella JP, Monteyrol PJ, Fauroux B. [Obesity and obstructive sleep apnea in children]. Archives de Pediatrie : Organe Officiel de la Societe Francaise de Pediatrie. 2017;24 Suppl 1:S34-s8.
126. Chervin RD, Guilleminault C. Obstructive sleep apnea and related disorders. Neurologic Clinics. 1996;14(3):583-609.
127. Wolk R, Shamsuzzaman AS, Somers VK. Obesity, sleep apnea, and hypertension. Hypertension. 2003;42(6):1067-74.
128. Sawatari H, Chishaki A, Ando SI. The Epidemiology of Sleep Disordered Breathing and Hypertension in Various Populations. Current Hypertension Reviews. 2016;12(1):12-7.
129. Bixler EO, Vgontzas AN, Lin HM, Ten Have T, Leiby BE, Vela-Bueno A, et al. Association of hypertension and sleep-disordered breathing. Archives of Internal Medicine. 2000;160(15):2289-95.
130. Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S. The occurrence of sleep-disordered breathing among middle-aged adults. The New England Journal of Medicine. 1993;328(17):1230-5.
131. Heinzer R, Vat S, Marques-Vidal P, Marti-Soler H, Andries D, Tobback N, et al. Prevalence of sleep-disordered breathing in the general population: the HypnoLaus study. The Lancet Respiratory Medicine. 2015;3(4):310-8.
132. Peppard PE, Young T, Barnet JH, Palta M, Hagen EW, Hla KM. Increased prevalence of sleep-disordered breathing in adults. American Journal of Epidemiology. 2013;177(9):1006-14.
133. Lee YC, Eun YG, Shin SY, Kim SW. Prevalence of snoring and high risk of obstructive sleep apnea syndrome in young male soldiers in Korea. Journal of Korean Medical Science. 2013;28(9):1373-7.
134. Yasin R, Muntham D, Chirakalwasan N. Uncovering the sleep disorders among young doctors. Sleep Breathing. 2016;20(4):1137-44.
135. Wosu AC, Velez JC, Barbosa C, Andrade A, Frye M, Chen X, et al. The Relationship between High Risk for Obstructive Sleep Apnea and General and Central Obesity: Findings from a Sample of Chilean College Students. ISRN Obes. 2014;2014:871681.
136. Cai XH, Xie YP, Li XC, Qu WL, Li T, Wang HX, et al. The prevalence and associated risk factors of sleep disorder-related symptoms in pregnant women in China. Sleep Breathing. 2013;17(3):951-6.
116
137. Qu XX, Esangbedo IC, Zhang XJ, Liu SJ, Li LX, Gao S, et al. Obstructive Sleep Apnea Syndrome is Associated with Metabolic Syndrome among Adolescents and Youth in Beijing: data from Beijing Child and Adolescent Metabolic Syndrome Study. Chinese Medical Journal. 2015;128(17):2278-83.
138. Bearpark H, Elliott L, Grunstein R, Hedner J, Cullen S, Schneider H, et al. Occurrence and correlates of sleep disordered breathing in the Australian town of Busselton: a preliminary analysis. Sleep. 1993;16(8 Suppl):S3-5.
139. Pillar G, Shehadeh N. Abdominal fat and sleep apnea: the chicken or the egg? Diabetes Care. 2008;31 Suppl 2:S303-9.
140. Phillips CL, Hoyos CM, Yee BJ, Grunstein RR. CrossTalk opposing view: Sleep apnoea causes metabolic syndrome. The Journal of Physiology. 2016;594(17):4691-4.
141. Ip MS, Lam KS, Ho C, Tsang KW, Lam W. Serum leptin and vascular risk factors in obstructive sleep apnea. Chest. 2000;118(3):580-6.
142. Ozturk L, Unal M, Tamer L, Celikoglu F. The association of the severity of obstructive sleep apnea with plasma leptin levels. Archives of Otolaryngology--Head & Neck Surgery. 2003;129(5):538-40.
143. Harsch IA, Konturek PC, Koebnick C, Kuehnlein PP, Fuchs FS, Pour Schahin S, et al. Leptin and ghrelin levels in patients with obstructive sleep apnoea: effect of CPAP treatment. The European Respiratory Journal. 2003;22(2):251-7.
144. Mesarwi OA, Sharma EV, Jun JC, Polotsky VY. Metabolic dysfunction in obstructive sleep apnea: A critical examination of underlying mechanisms. Sleep and Biological Rhythms. 2015;13(1):2-17.
145. Peppard PE, Young T, Palta M, Dempsey J, Skatrud J. Longitudinal study of moderate weight change and sleep-disordered breathing. JAMA. 2000;284(23):3015-21.
146. Mahboub B, Afzal S, Alhariri H, Alzaabi A, Vats M, Soans A. Prevalence of symptoms and risk of sleep apnea in Dubai, UAE. International Journal of General Medicine. 2013;6:109-14.
147. Cho JH, Choi JH, Suh JD, Ryu S, Cho SH. Comparison of Anthropometric Data Between Asian and Caucasian Patients With Obstructive Sleep Apnea: A Meta-Analysis. Clinical and Experimental Otorhinolaryngology. 2016;9(1):1-7.
148. Drager LF, Bortolotto LA, Figueiredo AC, Krieger EM, Lorenzi GF. Effects of continuous positive airway pressure on early signs of atherosclerosis in obstructive sleep apnea. American Journal of Respiratory and Critical Care Medicine. 2007;176(7):706-12.
149. Chin K, Shimizu K, Nakamura T, Narai N, Masuzaki H, Ogawa Y, et al. Changes in intra-abdominal visceral fat and serum leptin levels in patients with obstructive sleep apnea syndrome following nasal continuous positive airway pressure therapy. Circulation. 1999;100(7):706-12.
150. Loube DI, Loube AA, Erman MK. Continuous positive airway pressure treatment results in weight less in obese and overweight patients with obstructive sleep apnea. Journal of the American Dietetic Association. 1997;97(8):896-7.
151. Redenius R, Murphy C, O'Neill E, Al-Hamwi M, Zallek SN. Does CPAP lead to change in BMI? Journal of clinical sleep medicine : JCSM : official publication of the American Academy of Sleep Medicine. 2008;4(3):205-9.
152. Quan SF, Budhiraja R, Clarke DP, Goodwin JL, Gottlieb DJ, Nichols DA, et al. Impact of treatment with continuous positive airway pressure (CPAP) on weight in obstructive sleep apnea. Journal of clinical sleep medicine : JCSM : official publication of the American Academy of Sleep Medicine. 2013;9(10):989-93.
153. Garcia JM, Sharafkhaneh H, Hirshkowitz M, Elkhatib R, Sharafkhaneh A. Weight and metabolic effects of CPAP in obstructive sleep apnea patients with obesity. Respiratory Research. 2011;12(1):80.
154. Drager LF, Brunoni AR, Jenner R, Lorenzi-Filho G, Bensenor IM, Lotufo PA. Effects of CPAP on body weight in patients with obstructive sleep apnoea: a meta-analysis of randomised trials. Thorax. 2015;70(3):258-64.
117
155. Punjabi NM, Shahar E, Redline S, Gottlieb DJ, Givelber R, Resnick HE, et al. Sleep-disordered breathing, glucose intolerance, and insulin resistance: the Sleep Heart Health Study. American Journal of Epidemiology. 2004;160(6):521-30.
156. Punjabi NM, Sorkin JD, Katzel LI, Goldberg AP, Schwartz AR, Smith PL. Sleep-disordered breathing and insulin resistance in middle-aged and overweight men. American Journal of Respiratory and Critical Care Medicine. 2002;165(5):677-82.
157. Al-Delaimy WK, Manson JE, Willett WC, Stampfer MJ, Hu FB. Snoring as a risk factor for type II diabetes mellitus: a prospective study. American Journal of Epidemiology. 2002;155(5):387-93.
158. Elmasry A, Janson C, Lindberg E, Gislason T, Tageldin MA, Boman G. The role of habitual snoring and obesity in the development of diabetes: a 10-year follow-up study in a male population. Journal of Internal Medicine. 2000;248(1):13-20.
159. Owolabi EO, Ter Goon D, Adeniyi OV. Central obesity and normal-weight central obesity among adults attending healthcare facilities in Buffalo City Metropolitan Municipality, South Africa: a cross-sectional study. Journal of Health, Population, and Nutrition. 2017;36(1):54.
160. Botros N, Concato J, Mohsenin V, Selim B, Doctor K, Yaggi HK. Obstructive sleep apnea as a risk factor for type 2 diabetes. American Journal of Medicine. 2009;122(12):1122-7.
161. Marshall NS, Wong KKH, Phillips CL, Liu PY, Knuiman MW, Grunstein RR. Is Sleep Apnea an Independent Risk Factor for Prevalent and Incident Diabetes in the Busselton Health Study? Journal of Clinical Sleep Medicine : JCSM : official publication of the American Academy of Sleep Medicine. 2009;5(1):15-20.
162. Ip MS, Lam B, Ng MM, Lam WK, Tsang KW, Lam KS. Obstructive sleep apnea is independently associated with insulin resistance. American Journal of Respiratory and Critical Care Medicine. 2002;165(5):670-6.
163. Theorell-Haglow J, Berne C, Janson C, Lindberg E. Obstructive sleep apnoea is associated with decreased insulin sensitivity in females. European Respiratory Journal. 2008;31(5):1054-60.
164. Pamidi S, Wroblewski K, Broussard J, Day A, Hanlon EC, Abraham V, et al. Obstructive Sleep Apnea in Young Lean Men. Diabetes Care. 2012;35(11):2384-9.
165. Cain MA, Ricciuti J, Louis JM. Sleep-disordered breathing and future cardiovascular disease risk. Seminars in Perinatology. 2015;39(4):304-9.
166. Braun B, Rock PB, Zamudio S, Wolfel GE, Mazzeo RS, Muza SR, et al. Women at altitude: short-term exposure to hypoxia and/or alpha(1)-adrenergic blockade reduces insulin sensitivity. Journal of Applied Physiology. 2001;91(2):623-31.
167. Martinez-Ceron E, Barquiel B, Bezos AM, Casitas R, Galera R, Garcia-Benito C, et al. Effect of Continuous Positive Airway Pressure on Glycemic Control in Patients with Obstructive Sleep Apnea and Type 2 Diabetes. A Randomized Clinical Trial. American Journal of Respiratory and Critical Care Medicine. 2016;194(4):476-85.
168. Ioachimescu OC, Anthony J, Constantin T, Ciavatta M-M, McCarver K, Sweeney ME. VAMONOS (Veterans Affairs' Metabolism, Obstructed and Non-Obstructed Sleep) Study: Effects of CPAP Therapy on Glucose Metabolism in Patients with Obstructive Sleep Apnea. Journal of clinical sleep medicine : JCSM : official publication of the American Academy of Sleep Medicine. 2017;13(3):455-66.
169. Shaw JE, Punjabi NM, Naughton MT, Willes L, Bergenstal RM, Cistulli PA, et al. The Effect of Treatment of Obstructive Sleep Apnea on Glycemic Control in Type 2 Diabetes. American Journal of Respiratory and Critical Care Medicine.. 2016;194(4):486-92.
170. Phillips CL, O’Driscoll DM. Hypertension and obstructive sleep apnea. Nature and Science of Sleep. 2013;5:43-52.
171. Gonzaga C, Bertolami A, Bertolami M, Amodeo C, Calhoun D. Obstructive sleep apnea, hypertension and cardiovascular diseases. Journal of Human Hypertension. 2015;29(12):705-12.
172. Kasai T, Floras JS, Bradley TD. Sleep apnea and cardiovascular disease: a bidirectional relationship. Circulation. 2012;126(12):1495-510.
118
173. Tkacova R, McNicholas WT, Javorsky M, Fietze I, Sliwinski P, Parati G, et al. Nocturnal intermittent hypoxia predicts prevalent hypertension in the European Sleep Apnoea Database cohort study. The European Respiratory Journal. 2014;44(4):931-41.
174. Grote L, Ploch T, Heitmann J, Knaack L, Penzel T, Peter JH. Sleep-related breathing disorder is an independent risk factor for systemic hypertension. American Journal of Respiratory and Critical Care Medicine. 1999;160(6):1875-82.
175. Lavie P, Herer P, Hoffstein V. Obstructive sleep apnoea syndrome as a risk factor for hypertension: population study. BMJ. 2000;320(7233):479-82.
176. Peppard PE, Young T, Palta M, Skatrud J. Prospective study of the association between sleep-disordered breathing and hypertension. The New England Journal of Medicine. 2000;342(19):1378-84.
177. Haas DC, Foster GL, Nieto FJ, Redline S, Resnick HE, Robbins JA, et al. Age-dependent associations between sleep-disordered breathing and hypertension: importance of discriminating between systolic/diastolic hypertension and isolated systolic hypertension in the Sleep Heart Health Study. Circulation. 2005;111(5):614-21.
178. Cano-Pumarega I, Duran-Cantolla J, Aizpuru F, Miranda-Serrano E, Rubio R, Martinez-Null C, et al. Obstructive sleep apnea and systemic hypertension: longitudinal study in the general population: the Vitoria Sleep Cohort. American Journal of Respiratory and Critical Care Medicine. 2011;184(11):1299-304.
179. Wu WT, Tsai SS, Shih TS, Lin MH, Chou TC, Ting H, et al. The impact of obstructive sleep apnea on high-sensitivity C-reactive protein in subjects with or without metabolic syndrome. Sleep Breathing. 2015;19(4):1449-57.
180. Guven SF, Turkkani MH, Ciftci B, Ciftci TU, Erdogan Y. The relationship between high-sensitivity C-reactive protein levels and the severity of obstructive sleep apnea. Sleep Breathing. 2012;16(1):217-21.
181. Kim J, Lee SJ, Choi KM, Lee SK, Yoon DW, Lee SG, et al. Obstructive Sleep Apnea Is Associated with Elevated High Sensitivity C-Reactive Protein Levels Independent of Obesity: Korean Genome and Epidemiology Study. PLoS One. 2016;11(9):e0163017.
182. Sharma SK, Mishra HK, Sharma H, Goel A, Sreenivas V, Gulati V, et al. Obesity, and not obstructive sleep apnea, is responsible for increased serum hs-CRP levels in patients with sleep-disordered breathing in Delhi. Sleep Medine. 2008;9(2):149-56.
183. Ryan S, Taylor CT, McNicholas WT. Systemic inflammation: a key factor in the pathogenesis of cardiovascular complications in obstructive sleep apnoea syndrome? Thorax. 2009;64(7):631-6.
184. Adedayo AM, Olafiranye O, Smith D, Hill A, Zizi F, Brown C, et al. Obstructive sleep apnea and dyslipidemia: evidence and underlying mechanism. Sleep Breathing. 2014;18(1):13-8.
185. Trzepizur W, Le Vaillant M, Meslier N, Pigeanne T, Masson P, Humeau MP, et al. Independent association between nocturnal intermittent hypoxemia and metabolic dyslipidemia. Chest. 2013;143(6):1584-9.
186. Nadeem R, Singh M, Nida M, Waheed I, Khan A, Ahmed S, et al. Effect of obstructive sleep apnea hypopnea syndrome on lipid profile: a meta-regression analysis. Journal of clinical sleep medicine : JCSM : official publication of the American Academy of Sleep Medicine. 2014;10(5):475-89.
187. Borgel J, Sanner BM, Bittlinsky A, Keskin F, Bartels NK, Buechner N, et al. Obstructive sleep apnoea and its therapy influence high-density lipoprotein cholesterol serum levels. The European Respiratory Journal. 2006;27(1):121-7.
188. Newman AB, Nieto FJ, Guidry U, Lind BK, Redline S, Pickering TG, et al. Relation of sleep-disordered breathing to cardiovascular disease risk factors: the Sleep Heart Health Study. American Journal of Epidemiology. 2001;154(1):50-9.
189. Coughlin SR, Mawdsley L, Mugarza JA, Wilding JP, Calverley PM. Cardiovascular and metabolic effects of CPAP in obese males with OSA. The European Respiratory Journal. 2007;29(4):720-7.
119
190. Barcelo A, Miralles C, Barbe F, Vila M, Pons S, Agusti AG. Abnormal lipid peroxidation in patients with sleep apnoea. The European Respiratory Journal. 2000;16(4):644-7.
191. Tan KC, Chow WS, Lam JC, Lam B, Wong WK, Tam S, et al. HDL dysfunction in obstructive sleep apnea. Atherosclerosis. 2006;184(2):377-82.
192. Fortuna RJ, Robbins BW, Halterman JS. Ambulatory care among young adults in the United States. Annals of Internal Medicine. 2009;151(6):379-85.
193. Ozer EM, Urquhart JT, Brindis CD, Park MJ, Irwin CE, Jr. Young adult preventive health care guidelines: there but can't be found. Archives of Pediatrics & Adolescent Medicine. 2012;166(3):240-7.
194. Peterson RA, Merunka DR. Convenience samples of college students and research reproducibility. Journal of Business Research. 2014;67(5):1035-41.
195. Newnham JP, Evans SF, Michael CA, Stanley FJ, Landau LI. Effects of frequent ultrasound during pregnancy: a randomised controlled trial. The Lancet. 1993;342(8876):887-91.
196. Burke V, Beilin LJ, Simmer K, Oddy WH, Blake KV, Doherty D, et al. Predictors of body mass index and associations with cardiovascular risk factors in Australian children: a prospective cohort study. International Journal of Obesity. 2005;29(1):15-23.
197. Huang R-C, Burke V, Newnham JP, Stanley FJ, Kendall GE, Landau LI, et al. Perinatal and childhood origins of cardiovascular disease. International Journal of Obesity. 2007;31(2):236-44.
198. Huang R-C, Mori TA, Burke V, Newnham J, Stanley FJ, Landau LI, et al. Synergy between adiposity, insulin resistance, metabolic risk factors, and inflammation in adolescents. Diabetes Care. 2009;32(4):695-701.
199. Beilin L, Huang RC. Childhood obesity, hypertension, the metabolic syndrome and adult cardiovascular disease. Clinical and Experimental Pharmacology and Physiology. 2008;35(4):409-11.
200. Le-Ha C, Beilin LJ, Burrows S, Huang R-C, Oddy WH, Hands B, et al. Oral contraceptive use in girls and alcohol consumption in boys are associated with increased blood pressure in late adolescence. European Journal of Preventive Cardiology. 2013;20(6):947-55.
201. Le-Ha C, Beilin LJ, Burrows S, Oddy WH, Hands B, Mori TA. Gender and the active smoking and high-sensitivity C-reactive protein relation in late adolescence. Journal of Lipid Research. 2014;55(4):758-64.
202. Bhat SK, Beilin LJ, Robinson M, Burrows S, Mori TA. Contrasting effects of prenatal life stress on blood pressure and body mass index in young adults. Journal of Hypertension. 2015;33(4):711-9.
203. Newnham JP, Doherty DA, Kendall GE, Zubrick SR, Landau LL, Stanley FJ. Effects of repeated prenatal ultrasound examinations on childhood outcome up to 8 years of age: follow-up of a randomised controlled trial. Lancet. 2004;364(9450):2038-44.
204. Straker LM, Hall GL, Mountain J, Howie EK, White E, McArdle N, et al. Rationale, design and methods for the 22 year follow-up of the Western Australian Pregnancy Cohort (Raine) Study. BMC Public Health. 2015;15:663.
205. World Health Organization. Obesity : preventing and managing the global epidemic : report of a WHO consultation Geneva: Geneva : World Health Organization; 2000 [cited2016Aug].Availablefrom: http://www.who.int/nutrition/publications/obesity/WHO_TRS_894/en/.
206. Straker L, Mountain J, Jacques A, White S, Smith A, Landau L, et al. Cohort Profile: The Western Australian Pregnancy Cohort (Raine) Study-Generation 2. International Journal of Epidemiology. 2017;46(5):1384-5j.
207. Chaturvedi S. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7): is it really practical? The National Medical Journal of India. 2004;17(4):227.
208. Gayoso-Diz P, Otero-Gonzalez A, Rodriguez-Alvarez MX, Gude F, Garcia F, De Francisco A, et al. Insulin resistance (HOMA-IR) cut-off values and the metabolic syndrome in a
120
general adult population: effect of gender and age: EPIRCE cross-sectional study. BMC Endocrine Disorders. 2013;13:47.
209. Johnson R, McNutt P, MacMahon S, Robson R. Use of the Friedewald formula to estimate LDL-cholesterol in patients with chronic renal failure on dialysis. Clinical Chemistry. 1997;43(11):2183-4.
210. Ainsworth BE, Macera CA, Jones DA, Reis JP, Addy CL, Bowles HR, et al. Comparison of the 2001 BRFSS and the IPAQ Physical Activity Questionnaires. Medicine and Science in Sports and Exercise. 2006;38(9):1584-92.
211. Thurtell MJ, Bruce BB, Rye DB, Newman NJ, Biousse V. The Berlin questionnaire screens for obstructive sleep apnea in idiopathic intracranial hypertension. Journal of Neuro-Ophthalmology. 2011;31(4):316-9.
212. Johns MW. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep. 1991;14(6):540-5.
213. Chasens ER, Ratcliffe SJ, Weaver TE. Development of the FOSQ-10: a short version of the Functional Outcomes of Sleep Questionnaire. Sleep. 2009;32(7):915-9.
214. Tryon WW. Nocturnal activity and sleep assessment. Clinical Psychology Review. 1996;16(3):197-213.
215. Ancoli-Israel S, Martin JL, Blackwell T, Buenaver L, Liu L, Meltzer LJ, et al. The SBSM Guide to Actigraphy Monitoring: Clinical and Research Applications. Behavioral Sleep Medicine. 2015;13 Suppl 1:S4-S38.
216. Slater JA, Botsis T, Walsh J, King S, Straker LM, Eastwood PR. Assessing sleep using hip and wrist actigraphy. Sleep and Biological Rhythms. 2015;13(2):172-80.
217. Sadeh A. The role and validity of actigraphy in sleep medicine: an update. Sleep Medicine Reviews. 2011;15(4):259-67.
218. Meltzer LJ, Walsh CM, Traylor J, Westin AM. Direct comparison of two new actigraphs and polysomnography in children and adolescents. Sleep. 2012;35(1):159-66.
219. Sadeh A, Sharkey KM, Carskadon MA. Activity-based sleep-wake identification: an empirical test of methodological issues. Sleep. 1994;17(3):201-7.
220. Cole RJ, Kripke DF, Gruen W, Mullaney DJ, Gillin JC. Automatic sleep/wake identification from wrist activity. Sleep. 1992;15(5):461-9.
221. Boyne K, Sherry DD, Gallagher PR, Olsen M, Brooks LJ. Accuracy of computer algorithms and the human eye in scoring actigraphy. Sleep Breathing. 2013;17(1):411-7.
222. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL, Jr., et al. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 report. JAMA. 2003;289(19):2560-72.
223. Oliveira AC, Oliveira AM, Oliveira N, Oliveira A, Almeida M, Veneza LM, et al. Is triglyceride to high-density lipoprotein cholesterol ratio a surrogates for insulin resistance in youth? Health. 2013;5(3):481.
224. Knutson KL, Van Cauter E, Rathouz PJ, DeLeire T, Lauderdale DS. Trends in the Prevalence of Short Sleepers in the USA: 1975–2006. Sleep. 2010;33(1):37-45.
225. National Sleep Foundation. National Sleep Foundation recommends new sleep times [Internet]. Washington, D.C: National Sleep Foundation; 2015 [cited 2017 June]. Availablefromhttps://sleepfoundation.org/press-release/national-sleep-foundation-recommends-new-sleep-times
226. Kakinami L, O'Loughlin EK, Brunet J, Dugas EN, Constantin E, Sabiston CM, et al. Associations between physical activity and sedentary behavior with sleep quality and quantity in young adults. Sleep Health. 2017;3(1):56-61.
227. Maslowsky J, Ozer EJ. Developmental trends in sleep duration in adolescence and young adulthood: evidence from a national United States sample. The Journal of Adolescent Health. 2014;54(6):691-7.
228. Walsemann KM, Ailshire JA, Gee GC. Student loans and racial disparities in self-reported sleep duration: evidence from a nationally representative sample of US young adults. Journal of Epidemiology and Community Health.. 2016;70(1):42-8.
121
229. Leger D, Beck F, Richard JB, Godeau E. Total sleep time severely drops during adolescence. PLoS One. 2012;7(10):e45204.
230. Knutson KL. The association between pubertal status and sleep duration and quality among a nationally representative sample of U. S. adolescents. American Journal of Human Biology. 2005;17(4):418-24.
231. Chen X, Gelaye B, Williams MA. Sleep characteristics and health-related quality of life among a national sample of American young adults: assessment of possible health disparities. Quality of Lfe Research. 2014;23(2):613-25.
232. Steptoe A, Peacey V, Wardle J. Sleep duration and health in young adults. Archives of Internal Medicine. 2006;166(16):1689-92.
233. Tryon WW. Issues of validity in actigraphic sleep assessment. Sleep. 2004;27(1):158-65. 234. Magee CA, Huang XF, Iverson DC, Caputi P. Examining the pathways linking chronic
sleep restriction to obesity. Journal of Obesity. 2010;2010. 235. Grandner MA, Schopfer EA, Sands-Lincoln M, Jackson N, Malhotra A. Relationship
between sleep duration and body mass index depends on age. Obesity (Silver Spring). 2015;23(12):2491-8.
236. Moreno CR, Louzada FM, Teixeira LR, Borges F, Lorenzi-Filho G. Short sleep is associated with obesity among truck drivers. Chronobiology International.. 2006;23(6):1295-303.
237. Kohatsu ND, Tsai R, Young T, Vangilder R, Burmeister LF, Stromquist AM, et al. Sleep duration and body mass index in a rural population. Archives of Internal Medicine. 2006;166(16):1701-5.
238. Taheri S, Lin L, Austin D, Young T, Mignot E. Short sleep duration is associated with reduced leptin, elevated ghrelin, and increased body mass index. PLoS Medicine. 2004;1(3):e62.
239. Marshall NS, Glozier N, Grunstein RR. Is sleep duration related to obesity? A critical review of the epidemiological evidence. Sleep Medicine Reviews. 2008;12(4):289-98.
240. Amagai Y, Ishikawa S, Gotoh T, Doi Y, Kayaba K, Nakamura Y, et al. Sleep duration and mortality in Japan: the Jichi Medical School Cohort Study. Journal of Epidemiology. 2004;14(4):124-8.
241. Gortmaker SL, Dietz WH, Jr., Cheung LW. Inactivity, diet, and the fattening of America. Journal of the American Dietetic Association. 1990;90(9):1247-52, 55.
242. Knutson KL, Spiegel K, Penev P, Van Cauter E. The metabolic consequences of sleep deprivation. Sleep Medicine Reviews. 2007;11(3):163-78.
243. Kahlhofer J, Karschin J, Breusing N, Bosy-Westphal A. Relationship between actigraphy-assessed sleep quality and fat mass in college students. Obesity (Silver Spring). 2016;24(2):335-41.
244. Bailey BW, Allen MD, LeCheminant JD, Tucker LA, Errico WK, Christensen WF, et al. Objectively measured sleep patterns in young adult women and the relationship to adiposity. American Journal of Health Promotion. 2014;29(1):46-54.
245. Magee L, Hale L. Longitudinal associations between sleep duration and subsequent weight gain: a systematic review. Sleep medicine reviews. 2012;16(3):231-41.
246. Magee LL, Hale LE. Re: "Cross-Sectional and Longitudinal Associations between Objectively Measured Sleep Duration and Body Mass Index: The Cardia Sleep Study''. American Journal of Epidemiology. 2010;171(6):745-.
247. Roenneberg T, Allebrandt KV, Merrow M, Vetter C. Social jetlag and obesity. Current Biology. 2012;22(10):939-43.
248. Flint J, Kothare SV, Zihlif M, Suarez E, Adams R, Legido A, et al. Association between inadequate sleep and insulin resistance in obese children. The Journal of Pediatrics. 2007;150(4):364-9.
249. Knutson KL, Ryden AM, Mander BA, Van Cauter E. Role of sleep duration and quality in the risk and severity of type 2 diabetes mellitus. Archives of Internal Medicine. 2006;166(16):1768-74.
122
250. Upala S, Sanguankeo A, Congrete S, Romphothong K. Sleep duration and insulin resistance in individuals without diabetes mellitus: A systematic review and meta-analysis. Diabetes Research and Clinical Practice. 2015;109(3):e11-2.
251. Killick R, Hoyos CM, Melehan KL, Dungan GC, 2nd, Poh J, Liu PY. Metabolic and hormonal effects of 'catch-up' sleep in men with chronic, repetitive, lifestyle-driven sleep restriction. Clinical Endocrinology. 2015;83(4):498-507.
252. Kuciene R, Dulskiene V. Associations of short sleep duration with prehypertension and hypertension among Lithuanian children and adolescents: a cross-sectional study. BMC Public Health. 2014;14:255.
253. Sun X-m, Yao S, Hu S-j, Liu Z-y, Yang Y-j, Yuan Z-y, et al. Short sleep duration is associated with increased risk of pre-hypertension and hypertension in Chinese early middle-aged females. Sleep and Breathing. 2016;20(4):1355-62.
254. Fujikawa T, Tochikubo O, Kura N, Umemura S. Factors related to elevated 24-h blood pressure in young adults. Clin Exp Hypertens. 2009;31(8):705-12.
255. Mezick EJ, Hall M, Matthews KA. Sleep duration and ambulatory blood pressure in black and white adolescents. Hypertension. 2012;59(3):747-52.
256. Min H, Um YJ, Jang BS, Shin D, Choi E, Park SM, et al. Association between Sleep Duration and Measurable Cardiometabolic Risk Factors in Healthy Korean Women: The Fourth and Fifth Korean National Health and Nutrition Examination Surveys (KNHANES IV and V). International Journal of Endocrinology. 2016;2016:3784210.
257. Gangwisch JE, Heymsfield SB, Boden-Albala B, Buijs RM, Kreier F, Pickering TG, et al. Short sleep duration as a risk factor for hypertension: analyses of the first National Health and Nutrition Examination Survey. Hypertension. 2006;47(5):833-9.
258. Hall MH, Muldoon MF, Jennings JR, Buysse DJ, Flory JD, Manuck SB. Self-reported sleep duration is associated with the metabolic syndrome in midlife adults. Sleep. 2008;31(5):635-43.
259. Anujuo K, Stronks K, Snijder MB, Jean-Louis G, Rutters F, van den Born BJ, et al. Relationship between short sleep duration and cardiovascular risk factors in a multi-ethnic cohort - the helius study. Sleep Medicine. 2015;16(12):1482-8.
260. Grandner MA, Sands-Lincoln MR, Pak VM, Garland SN. Sleep duration, cardiovascular disease, and proinflammatory biomarkers. Nature and Science of Sleep. 2013;5:93-107.
261. Grandner MA, Chakravorty S, Perlis ML, Oliver L, Gurubhagavatula I. Habitual sleep duration associated with self-reported and objectively determined cardiometabolic risk factors. Sleep Medicine. 2014;15(1):42-50.
262. Altman NG, Izci-Balserak B, Schopfer E, Jackson N, Rattanaumpawan P, Gehrman PR, et al. Sleep duration versus sleep insufficiency as predictors of cardiometabolic health outcomes. Sleep Medicine. 2012;13(10):1261-70.
263. Chiang JK. Short duration of sleep is associated with elevated high-sensitivity C-reactive protein level in Taiwanese adults: a cross-sectional study. Journal of clinical sleep medicine : JCSM : Official Publication of the American Academy of Sleep Medicine. 2014;10(7):743-9.
264. Patel SR, Zhu X, Storfer-Isser A, Mehra R, Jenny NS, Tracy R, et al. Sleep duration and biomarkers of inflammation. Sleep. 2009;32(2):200-4.
265. Suarez EC. Self-reported symptoms of sleep disturbance and inflammation, coagulation, insulin resistance and psychosocial distress: evidence for gender disparity. Brain Behavior and Immunity. 2008;22(6):960-8.
266. Nadeem R, Molnar J, Madbouly EM, Nida M, Aggarwal S, Sajid H, et al. Serum inflammatory markers in obstructive sleep apnea: a meta-analysis. Journal of clinical sleep medicine : JCSM : official publication of the American Academy of Sleep Medicine. 2013;9(10):1003-12.
267. Ferrie JE, Kivimaki M, Akbaraly TN, Singh-Manoux A, Miller MA, Gimeno D, et al. Associations between change in sleep duration and inflammation: findings on C-reactive protein and interleukin 6 in the Whitehall II Study. American Journal of Epidemiology. 2013;178(6):956-61.
123
268. Arora T, Jiang CQ, Thomas GN, Lam KB, Zhang WS, Cheng KK, et al. Self-reported long total sleep duration is associated with metabolic syndrome: the Guangzhou Biobank Cohort Study. Diabetes care. 2011;34(10):2317-9.
269. Montag SE, Knutson KL, Zee PC, Goldberger JJ, Ng J, Kim KA, et al. Association of sleep characteristics with cardiovascular and metabolic risk factors in a population sample: the Chicago Area Sleep Study. Sleep Health. 2017;3(2):107-12.
270. Berkman LF, Liu SY, Hammer L, Moen P, Klein LC, Kelly E, et al. Work-family conflict, cardiometabolic risk, and sleep duration in nursing employees. Journal of Occupational Health Psychology. 2015;20(4):420-33.
271. Wang Q, Xi B, Liu M, Zhang Y, Fu M. Short sleep duration is associated with hypertension risk among adults: a systematic review and meta-analysis. Hypertension Research. 2012;35(10):1012-8.
272. Sekine T, Daimon M, Hasegawa R, Toyoda T, Kawata T, Funabashi N, et al. The impact of sleep deprivation on the coronary circulation. International Journal of Cardiology. 2010;144(2):266-7.
273. King CR, Knutson KL, Rathouz PJ, Sidney S, Liu K, Lauderdale DS. Short sleep duration and incident coronary artery calcification. JAMA. 2008;300(24):2859-66.
274. Neufeld EV, Carney JJ, Dolezal BA, Boland DM, Cooper CB. Exploratory Study of Heart Rate Variability and Sleep among Emergency Medical Services Shift Workers. Prehospital Emergency Care. 2017;21(1):18-23.
275. Lusardi P, Vanasia A, Mugellini A, Zoppi A, Preti P, Fogari R. Evaluation of nocturnal blood pressure by the Multi-P Analysis of 24-hour ambulatory monitoring. Zeitschrift Für Kardiologie. 1996;85 Suppl 3:121-3.
276. Wang L, Qin P, Zhao Y, Duan S, Zhang Q, Liu Y, et al. Prevalence and risk factors of poor sleep quality among Inner Mongolia Medical University students: A cross-sectional survey. Psychiatry Res. 2016;244:243-8.
277. Tang J, Liao Y, Kelly BC, Xie L, Xiang YT, Qi C, et al. Gender and Regional Differences in Sleep Quality and Insomnia: A General Population-based Study in Hunan Province of China. Scientific Reports. 2017;7:43690.
278. Hinz A, Glaesmer H, Brahler E, Loffler M, Engel C, Enzenbach C, et al. Sleep quality in the general population: psychometric properties of the Pittsburgh Sleep Quality Index, derived from a German community sample of 9284 people. Sleep Medicine. 2017;30:57-63.
279. Wong WS, Fielding R. Prevalence of insomnia among Chinese adults in Hong Kong: a population-based study. Journal of Sleep Research. 2011;20(1 Pt 1):117-26.
280. Dugas EN, Sylvestre MP, O'Loughlin EK, Brunet J, Kakinami L, Constantin E, et al. Nicotine dependence and sleep quality in young adults. Addictive Behaviors. 2017;65:154-60.
281. Vargas PA, Flores M, Robles E. Sleep quality and body mass index in college students: the role of sleep disturbances. Journal of American College Health.. 2014;62(8):534-41.
282. Quick V, Byrd-Bredbenner C, White AA, Brown O, Colby S, Shoff S, et al. Eat, sleep, work, play: associations of weight status and health-related behaviors among young adult college students. American Journal of Health Promotion. 2014;29(2):e64-72.
283. de Vasconcelos HC, Fragoso LV, Marinho NB, de Araujo MF, de Freitas RW, Zanetti ML, et al. [Correlation between anthropometric indicators and sleep quality among Brazilian university students]. Revista da Escola de Enfermagem da U S P. 2013;47(4):852-9.
284. Bruno RM, Palagini L, Gemignani A, Virdis A, Di Giulio A, Ghiadoni L, et al. Poor sleep quality and resistant hypertension. Sleep Medicne. 2013;14(11):1157-63.
285. Sforza E, Saint Martin M, Barthelemy JC, Roche F. Association of self-reported sleep and hypertension in non-insomniac elderly subjects. Journal of Clinical Sleep Medicine : JCSM : official publication of the American Academy of Sleep Medicine. 2014;10(9):965-71.
124
286. Lu Y, Lu M, Dai H, Yang P, Smith-Gagen J, Miao R, et al. Lifestyle and Risk of Hypertension: Follow-Up of a Young Pre-Hypertensive Cohort. International Journal of Medical Sciences.. 2015;12(7):605-12.
287. Esler M, Hastings J, Lambert G, Kaye D, Jennings G, Seals DR. The influence of aging on the human sympathetic nervous system and brain norepinephrine turnover. American Journal of Physiology Regulatory, Integrative and Comparative Physiology. 2002;282(3):R909-16.
288. Sharma SD, Kanona H, Kumar G, Kotecha B. Latest trends in the assessment and management of paediatric snoring and sleep apnoea.The Journal of Laryngology and Otology. 2016;130(5):482-9.
289. Hiestand DM, Britz P, Goldman M, Phillips B. Prevalence of symptoms and risk of sleep apnea in the US population: Results from the national sleep foundation sleep in America 2005 poll. Chest. 2006;130(3):780-6.
290. Pensuksan WC, Chen X, Lohsoonthorn V, Lertmaharit S, Gelaye B, Williams MA. High risk for obstructive sleep apnea in relation to hypertension among southeast Asian young adults: role of obesity as an effect modifier. American Journal of Hypertension. 2014;27(2):229-36.
291. Schwartz AR, Patil SP, Laffan AM, Polotsky V, Schneider H, Smith PL. Obesity and obstructive sleep apnea: pathogenic mechanisms and therapeutic approaches. Proceedings of the American Thoracic Society. 2008;5(2):185-92.
292. Krishnan V, Dixon-Williams S, Thornton JD. Where there is smoke...there is sleep apnea: exploring the relationship between smoking and sleep apnea. Chest. 2014;146(6):1673-80.
293. Kim NH, Lee SK, Eun CR, Seo JA, Kim SG, Choi KM, et al. Short sleep duration combined with obstructive sleep apnea is associated with visceral obesity in Korean adults. Sleep. 2013;36(5):723-9.
294. Franco I, Reis R, Ferreira D, Xara S, Ferreira W, Bettencourt N, et al. The impact of neck and abdominal fat accumulation on the pathogenesis of obstructive sleep apnea. Revista Portuguesa de Pneumologia. 2016;22(4):240-2.
295. Dempsey JA, Veasey SC, Morgan BJ, O'Donnell CP. Pathophysiology of sleep apnea. Pulmonary Medicine. 2010;90(1):47-112.
296. Nino G, Gutierrez MJ, Ravindra A, Nino CL, Rodriguez-Martinez CE. Abdominal adiposity correlates with adenotonsillectomy outcome in obese adolescents with severe obstructive sleep apnea. Pulmonary Medicine. 2012;2012:351037.
297. Shechter A, Grandner MA, St-Onge MP. The Role of Sleep in the Control of Food Intake. American Journal of Lifestyle Medicine. 2014;8(6):371-4.
298. Nieto FJ, Young TB, Lind BK, Shahar E, Samet JM, Redline S, et al. Association of sleep-disordered breathing, sleep apnea, and hypertension in a large community-based study. Sleep Heart Health Study. JAMA. 2000;283(14):1829-36.
299. O'Connor GT, Caffo B, Newman AB, Quan SF, Rapoport DM, Redline S, et al. Prospective study of sleep-disordered breathing and hypertension: the Sleep Heart Health Study. American Journal of Respiratory and Critical Care Medicine. 2009;179(12):1159-64.
300. Bradley TD, Floras JS. Obstructive sleep apnoea and its cardiovascular consequences. Lancet. 2009;373(9657):82-93.
301. Kapsimalis F, Varouchakis G, Manousaki A, Daskas S, Nikita D, Kryger M, et al. Association of sleep apnea severity and obesity with insulin resistance, C-reactive protein, and leptin levels in male patients with obstructive sleep apnea. Lung. 2008;186(4):209-17.
302. Davies RJO, Turner R, Crosby J, Stradling JR. Plasma-Insulin and Lipid-Levels in Untreated Obstructive Sleep-Apnea and Snoring - Their Comparison with Matched Controls and Response to Treatment. Journal of Sleep Research. 1994;3(3):180-5.
125
303. Stoohs RA, Facchini F, Guilleminault C. Insulin resistance and sleep-disordered breathing in healthy humans. American Journal of Respiratory and Critical Care Medicine. 1996;154(1):170-4.
304. Chien MY, Lee P, Tsai YF, Yang PC, Wu YT. C-reactive protein and heart rate recovery in middle-aged men with severe obstructive sleep apnea. Sleep Breathing. 2012;16(3):629-37.
305. Shamsuzzaman AS, Winnicki M, Lanfranchi P, Wolk R, Kara T, Accurso V, et al. Elevated C-reactive protein in patients with obstructive sleep apnea. Circulation. 2002;105(21):2462-4.
306. Ryan S, Nolan GM, Hannigan E, Cunningham S, Taylor C, McNicholas WT. Cardiovascular risk markers in obstructive sleep apnoea syndrome and correlation with obesity. Thorax. 2007;62(6):509-14.
307. Straker L, Mountain J, Jacques A, White S, Smith A, Landau L, et al. Cohort Profile: The Western Australian Pregnancy Cohort (Raine) Study–Generation 2. International Journal of Epidemiology. 2017:dyw308.
308. Donga E, van Dijk M, van Dijk JG, Biermasz NR, Lammers GJ, van Kralingen KW, et al. A single night of partial sleep deprivation induces insulin resistance in multiple metabolic pathways in healthy subjects. The Journal Clinical Endocrinology and Metabolism. 2010;95(6):2963-8.
309. Harsch IA, Schahin SP, Radespiel-Troger M, Weintz O, Jahreiss H, Fuchs FS, et al. Continuous positive airway pressure treatment rapidly improves insulin sensitivity in patients with obstructive sleep apnea syndrome. American Journal Respiratory and Critical Care Medicine. 2004;169(2):156-62.
310. Malone SK, Zemel B, Compher C, Souders M, Chittams J, Thompson AL, et al. Characteristics Associated with Sleep Duration, Chronotype, and Social Jet Lag in Adolescents. The Journal of School Nursing : the Official Publication of the National Association of School Nurses. 2016;32(2):120-31.
311. Grandner MA, Patel NP, Gehrman PR, Perlis ML, Pack AI. Problems associated with short sleep: bridging the gap between laboratory and epidemiological studies. Sleep Medicine reviews. 2010;14(4):239-47.
126
Appendix 1
Berlin Questionnaire
![The association between burning mouth syndrome and sleep ... final version.pdf · portunity for sleep [American Psychiatric Association (APA), 2013]. Estimates of the prevalence of](https://img.dokumen.tips/doc/110x75/5e640bf763ea8044f566a8b7/the-association-between-burning-mouth-syndrome-and-sleep-final-versionpdf.jpg)
