Sleep Deprivation

Gender-Specific Associations of Short Sleep Duration With

Prevalent and Incident Hypertension
The Whitehall II Study
Francesco P. Cappuccio, Saverio Stranges, Ngianga-Bakwin Kandala, Michelle A. Miller,
Frances M. Taggart, Meena Kumari, Jane E. Ferrie, Martin J. Shipley,
Eric J. Brunner, Michael G. Marmot

AbstractSleep deprivation (5 hour per night) was associated with a higher risk of hypertension in middle-aged
American adults but not among older individuals. However, the outcome was based on self-reported diagnosis of
incident hypertension, and no gender-specific analyses were included. We examined cross-sectional and prospective
associations of sleep duration with prevalent and incident hypertension in a cohort of 10 308 British civil servants aged
35 to 55 years at baseline (phase 1: 19851988). Data were gathered from phase 5 (19971999) and phase 7
(20032004). Sleep duration and other covariates were assessed at phase 5. At both examinations, hypertension was
Downloaded from http://hyper.ahajournals.org/ by guest on April 15, 2017

defined as blood pressure 140/90 mm Hg or regular use of antihypertensive medications. In cross-sectional analyses
at phase 5 (n5766), short duration of sleep (5 hour per night) was associated with higher risk of hypertension
compared with the group sleeping 7 hours, among women (odds ratio: 2.01; 95% CI: 1.13 to 3.58), independent of
confounders, with an inverse linear trend across decreasing hours of sleep (P0.003). No association was detected in
men. In prospective analyses (mean follow-up: 5 years), the cumulative incidence of hypertension was 20.0% (n740)
among 3691 normotensive individuals at phase 5. In women, short duration of sleep was associated with a higher risk
of hypertension in a reduced model (age and employment) (6 hours per night: odds ratio: 1.56 [95% CI: 1.07 to 2.27];
5 hour per night: odds ratio: 1.94 [95% CI: 1.08 to 3.50] versus 7 hours). The associations were attenuated after
accounting for cardiovascular risk factors and psychiatric comorbidities (odds ratio: 1.42 [95% CI: 0.94 to 2.16]; odds
ratio: 1.31 [95% CI: 0.65 to 2.63], respectively). Sleep deprivation may produce detrimental cardiovascular effects
among women. (Hypertension. 2007;50:693-700.)
Key Words: sleep duration blood pressure hypertension gender differences confounders comorbidities

S leep-disordered breathing (SDB) has been linked to

elevated blood pressure and risk of hypertension in
several epidemiological observational studies.15 Growing
tension among individuals with usual sleep duration above
or below the median of 7 to 8 hours per night.14 The
association was stronger, ie, a 66% higher risk of hyperten-
evidence indicates that sleep deprivation is also associated sion, among short sleepers (6 hours per night). Although
with a number of health outcomes, including hyperten- this study attempted to account for a number of potential
sion.6 14 For example, in a recent longitudinal analysis of the confounders, including psychiatric and cardiovascular comor-
first National Health And Nutrition Examination Survey bidities, the cross-sectional design did not allow inference on
(NHANES-I), short sleep duration (5 hours per night) was the temporal relationship between sleep duration and
associated with a 60% higher risk of incident hypertension in hypertension.
middle-aged (32 to 59 years) American adults without appar- Several studies in humans indicate potential pathophysio-
ent sleep disorders during a mean follow-up of 8 to 10 logical mechanisms supporting the biological plausibility of
years.13 No association was found in individuals 60 years of the association between sleep deprivation and hypertension.
age. However, the outcome was based on self-reported For example, acute curtailments of sleep may induce an
diagnosis of incident hypertension, and no gender-specific overactivity of the sympathetic nervous system leading to
analyses were included. Furthermore, a cross-sectional anal- higher blood pressure in both normotensive and hypertensive
ysis from the Sleep Heart Health Study on a sample of 6000 individuals.1517 Other contributing mechanisms may include
US adults showed a significant higher prevalence of hyper- overactivity of the renin-angiotensin-aldosterone system,

Received May 30, 2007; first decision June 24, 2007; revision accepted August 3, 2007.
From the Clinical Sciences Research Institute (F.P.C., S.S., N-B.K., M.A.M., F.M.T.), Warwick Medical School, Coventry, United Kingdom; and the
International Centre for Health and Society (M.K., J.E.F., M.J.S., E.J.B., M.G.M.), University College London Medical School, London, United Kingdom.
Correspondence to Francesco P. Cappuccio, Clinical Sciences Research Institute, Warwick Medical School, UHCW Campus, Clifford Bridge Rd,
Coventry CV2 2DX, United Kingdom. E-mail [email protected]
2007 American Heart Association, Inc.
Hypertension is available at http://hyper.ahajournals.org DOI: 10.1161/HYPERTENSIONAHA.107.095471

693
 694 Hypertension October 2007

proinflammatory responses, endothelial dysfunction, and re- medication. At both phase 5 and 7 screening examinations, anthro-
nal impairment.18,19 On the other hand, intervention studies to pometric measures were recorded, including height, weight, and
waist circumference; body mass index (BMI) was calculated as
improve duration and quality of sleep have been effective in
weight in kilograms divided by height in meters squared. Blood
reducing both daytime and nighttime blood pressures.19,20 pressure was measured 3 times using a standard mercury manometer
However, there is concern that sleep habits may represent a by trained and certified technicians in both examinations. The onsets
marker of health status and quality of life rather than a casual of the first-phase (systolic) and fifth-phase (diastolic) Korotkoff
factor for hypertension and other health outcomes.21 sounds were recorded. The mean of the second and third measures
were used in the analyses. At both examinations, hypertension was
In the present analysis, we sought to examine both the
defined as blood pressure 140/90 mm Hg or regular use of
cross-sectional and prospective associations of sleep duration antihypertensive medications.
with prevalent and incident hypertension in the Whitehall II
Study, a prospective cohort of 10 308 white-collar British Ethical Approval
civil servants aged 35 to 55 at baseline (phase 1: 19851988). Ethical approval for the Whitehall II Study was obtained from the
Because reduced durations of sleep might be associated with University College London Medical School committee on the ethics
of human research.
more detrimental effects on cardiovascular outcomes among
women,6 9 unlike previous investigations, we conducted Statistical Analysis
gender-specific analyses with the inclusion of a number of For continuous and categorical variables, respectively, Kruskal
potential confounding variables. Wallis and 2 tests were used to determine the statistical significance
of any difference in the distribution of baseline variables at phase 5
Methods across categories of sleep duration. The statistical significance of the
Downloaded from http://hyper.ahajournals.org/ by guest on April 15, 2017

interaction between each baseline characteristic and sleep duration,

Study Population adjusted for other important baseline variables, was tested in multi-
The Whitehall II Cohort was recruited in 19851988 (phase 1) from variate logistic regression models that included this interaction and
20 London-based civil service departments. The rationale, design, the corresponding main effect terms. The interaction between sleep
and methods of the study have been described in detail elsewhere.22 duration and gender was significant (P0.05); therefore, all of the
Briefly, the initial response rate was 73%, and the final cohort analyses were stratified by gender. No significant interactions were
consisted of 10 308 participants; 3413 women and 6895 men. found between sleep duration and age or other relevant covariates.
Follow-up screening examinations took place in 19911993 (phase Univariate and multivariate regression analyses were conducted to
3), 19971999 (phase 5), and 20032004 (phase 7), whereas postal test the association between categories of sleep duration and preva-
questionnaires were sent to participants in 1989 (phase 2), 1995 lent hypertension at phase 5 (cross-sectional analyses), as well as the
(phase 4), and 2001 (phase 6). The participation rates of the original association between sleep duration at phase 5 and incident hyper-
cohort (n10 308) were 83%, 76%, and 68% at phases 3, 5, and 7, tension at phase 7 (prospective analyses). Covariates included the
respectively. In this report, we used data from phases 5 and 7. The following: baseline age, employment, alcohol consumption, smok-
total sample at phase 5 consisted of 7204 participants. The present ing, physical activity, BMI, cardiovascular disease (CVD) drugs
analyses were restricted to white individuals (n6592), given the (other than antihypertensive medications), the SF-36 mental and
low numbers of other ethnic groups (n612). For the cross-sectional physical health component scores, depression, and use of hypnotics.
analyses, only participants with a complete set of data at phase 5 In these analyses, 7 hours of sleep was selected as the reference
were included (n5766: 4199 men and 1567 women). Their char- category. All of the techniques were implemented using Stata 9.0.
acteristics were comparable to the overall sample (see Appendix).
For the longitudinal analyses, the incidence of hypertension at phase Results
7 was assessed among participants who were normotensive at phase
5 (n3691: 2686 men and 1005 women). Descriptives
Characteristics for both male and female participants (Table
Sleep Duration 1) at phase 5 (baseline) are reported by categories of sleep
At phases 5 and 7, sleep duration was elicited by the question, How duration. Among men, participants sleeping 9 hours were,
many hours of sleep do you have on an average week night?
on average, older than other participants; in addition, they
Response categories were 5 hours, 6 hours, 7 hours, 8 hours, and
9 hours. were less likely to be physically active and more likely to be
on medication for CVD drugs. On the other hand, male
Covariates participants sleeping 5 hours had a lower employment
For the present analyses, age and other covariates were derived from status, higher mean levels of BMI and waist circumference,
the questionnaires at phase 5. Employment grade was determined were more likely to be depressed and on medication for
from the participants last known civil service grade title (19% had
hypnotics, and reported lower scores for mental and physical
retired by phase 5) and divided into 3 categories in order of
decreasing salary: administrative, professional/executive, and cleri- health than other participants. No significant differences in
cal/support. Participants were allocated to 1 of 4 smoking categories: drinking and smoking habits were reported across categories
never, ex-smoker, pipe and/or cigar only, or current cigarette smoker of sleep duration. For blood pressures, no consistent pattern
(manufactured or hand-rolled cigarettes). Alcohol consumption in of association was noted; however, the mean levels of systolic
the previous week was recorded (units per week). Leisure-time
physical activity was categorized by energy use in 2 categories:
blood pressure (and pulse pressure) and the prevalence of
vigorous (subjects who reported 1.5 hour of vigorous activity per hypertension were significantly higher among participants
week) and no vigorous activity. General health status was assessed sleeping 8 hours than in other participants.
using the physical and mental health component summaries of the Among women, participants at the 2 extreme categories of
Short Form-36 (SF-36) health survey questionnaire23: low scores sleep duration (ie, 5 hours and 9 hours) were character-
indicate low functioning. Psychiatric morbidity, including depres-
sion, was assessed with a modified General Health Questionnaire ized, in general, by a poorer health status and lifestyle profile.
score. Participants taking sleep medication (hypnotics) or cardiovas- In particular, they had a lower employment status, were less
cular drugs were identified through a questionnaire item on current likely to be physically active, more likely to be on medication
 Cappuccio et al Short Sleep Duration and Hypertension 695

Table 1. Baseline Characteristics (Phase 5: 19971999) Across Categories of Sleep Duration: The
Whitehall II Study (n5766)
Sleep Duration

Characteristics 5 h 6h 7h 8h 9 h P*
Men (n4199)
No. of subjects 265 1383 1886 620 45
Age, y 54.6 (5.6) 54.6 (5.7) 55.5 (6.1) 57.5 (6.1) 57.7 (5.4) 0.001
DBP, mm Hg 79.2 (10.9) 78.2 (10.4) 78.6 (10.5) 79.0 (10.9) 78.1 (10.6) 0.57
SBP, mm Hg 124.0 (16.7) 122.3 (15.4) 123.9 (15.8) 125.1 (17.3) 122.5 (16.6) 0.009
Pulse pressure, mm Hg 44.7 (11.5) 44.1 (10.7) 45.3 (11.0) 46.1 (12.4) 44.4 (11.1) 0.003
Lowest employment, n (%) 26 (9.8) 66 (4.8) 62 (3.3) 18 (2.9) 3 (6.7) 0.001
BMI, kg/m2 27.0 (4.1) 26.4 (3.5) 25.9 (3.4) 25.4 (3.3) 25.7 (3.3) 0.001
Waist, cm 94.9 (11.6) 92.8 (10.2) 91.6 (9.7) 90.2 (9.9) 91.5 (8.5) 0.001
Weekly alcohol, units 18.0 (21.2) 16.9 (15.9) 16.5 (15.9) 16.3 (17.1) 17.7 (20) 0.58
CVD drugs, n (%) 49 (18.6) 185 (13.4) 259 (13.8) 111 (18.0) 9 (20.0) 0.013
Physical activity, n (%) 108 (40.8) 687 (49.7) 930 (49.3) 320 (51.6) 17 (37.8) 0.022
Downloaded from http://hyper.ahajournals.org/ by guest on April 15, 2017

SF-36 mental (score) 47.1 (12.2) 50.5 (9.2) 52.5 (8.2) 53.7 (7.5) 52.0 (9.9) 0.001
SF-36 physical (score) 50.2 (8.8) 52.1 (7.1) 52.4 (6.4) 51.9 (6.8) 50.9 (8.5) 0.019
Use of hypnotics, n (%) 7 (2.7) 8 (0.6) 4 (0.2) 2 (0.3) 0 (0.0) 0.001
Depression cases, n (%) 62 (23.8) 178 (13.1) 173 (9.3) 55 (9.0) 4 (8.9) 0.001
Current smoking, n (%) 30 (11.5) 123 (8.9) 154 (8.2) 49 (7.9) 4 (8.9) 0.45
Hypertensive, n (%) 84 (31.7) 358 (25.9) 533 (28.3) 205 (33.1) 14 (31.1) 0.014
No medication 45 (17.0) 218 (15.8) 323 (17.1) 115 (18.6) 9 (20.0)
Medication 38 (14.3) 140 (10.1) 208 (11.0) 90 (14.5) 5 (11.1)
Women (n1567)
No. of subjects 157 511 597 272 30
Age, y 56.4 (5.9) 56.1 (6.1) 55.7 (6.1) 56.8 (6.1) 54.7 (5.7) 0.07
DBP, mm Hg 75.7 (10.4) 74.9 (10.1) 74.4 (9.8) 74.2 (10.0) 74.0 (10.0) 0.50
SBP, mm Hg 124.2 (18.3) 121.2 (16.8) 120.3 (17.1) 121.4 (16.8) 118 (17.2) 0.07
Pulse pressure, mm Hg 48.5 (13.8) 46.3 (12.3) 45.9 (12.3) 47.2 (11.6) 44.0 (10.9) 0.07
Lowest employment, n (%) 52 (33.1) 143 (28.1) 162 (27.2) 67 (25.0) 11 (36.7) 0.003
BMI, kg/m2 27.1 (5.8) 26.2 (5.1) 25.9 (4.4) 26.3 (5.0) 25.8 (4.6) 0.31
Waist, cm 82.5 (14.3) 80.2 (11.8) 80.1 (11.2) 80.4 (11.9) 81.5 (12.0) 0.66
Weekly alcohol, units 7.3 (10.8) 8.0 (8.5) 8.2 (9.5) 8.4 (10.0) 9.8 (9.4) 0.10
CVD drugs, n (%) 36 (23.1) 58 (11.4) 94 (15.8) 46 (16.9) 4 (13.3) 0.007
Physical activity, n (%) 38 (24.2) 169 (33.1) 215 (36) 108 (39.7) 6 (20.0) 0.006
SF-36 mental (score) 42.9 (13.1) 49.0 (10.3) 50.8 (9.2) 52.6 (8.8) 46.9 (15.0) 0.001
SF-36 physical (score) 46.0 (12.5) 50.2 (8.4) 49.8 (9.0) 49.9 (8.8) 45.2 (10.1) 0.001
Use of hypnotics, n (%) 2 (1.3) 5 (1.0) 4 (0.7) 2 (0.7) 1 (3.3) 0.60
Depression cases, n (%) 50 (32.5) 72 (14.4) 66 (11.2) 20 (7.4) 5 (16.7) 0.001
Current smoking, n (%) 20 (12.9) 79 (15.5) 76 (12.8) 31 (11.4) 1 (3.3) 0.22
Hypertensive, n (%) 58 (36.9) 124 (24.3) 152 (25.5) 66 (24.3) 7 (23.3) 0.024
No medication 27 (17.2) 72 (14.1) 69 (11.6) 29 (10.7) 5 (16.7)
Medication 31 (19.7) 52 (10.2) 83 (13.9) 37 (13.6) 2 (6.7)
Data are expressed as the mean (SD) or as n (%). DBP indicates diastolic blood pressure; SBP, systolic blood pressure.
*P value for comparison across sleep duration groups using the 2 analysis for categorical variables and Kruskal-Wallis test for
continuous variables.
One value is missing.
Two values are missing.

for CVD drugs or hypnotics, more likely to be depressed, and duration. For blood pressures, there was a consistent pattern
reported lower scores for mental and physical health than of association among female participants sleeping 5 hours,
other categories. No significant differences in drinking and who reported higher mean levels of systolic blood pressure
smoking habits were reported across categories of sleep (and pulse pressure), as well as a significantly higher preva-
 696 Hypertension October 2007

Table 2. OR (95% CI) of Prevalent Hypertension Across Categories of Sleep Duration at Phase 5 (19971999): The Whitehall II Study
(n5766)
Sleep Duration P*

Sample Models 5 h 6h 7h 8h 9 h Linear Nonlinear

Men (n4199)
No. 265 1383 1886 620 45
Cases, n 84 358 533 205 14
Model 1 1.18 0.89 1.55 0.89 0.76 1.04 1 1.25 1.03 1.52 1.15 0.61 2.17 0.06 0.015
Model 2 1.23 0.92 1.63 0.93 0.79 1.09 1 1.11 0.91 1.36 0.99 0.52 1.90 0.70 0.09
Model 3 0.88 0.63 1.23 0.86 0.72 1.03 1 1.12 0.89 1.41 0.92 0.44 1.92 0.05 0.54
Women (n1567)
No. 157 511 597 272 30
Cases, n 58 124 152 66 7
Model 1 1.72 1.18 2.49 0.94 0.71 1.23 1 0.94 0.67 1.31 0.89 0.37 2.12 0.13 0.06
Model 2 1.64 1.11 2.42 0.90 0.68 1.19 1 0.83 0.58 1.17 0.98 0.40 2.40 0.12 0.15
Model 3 1.72 1.07 2.75 0.92 0.67 1.28 1 0.74 0.50 1.11 0.70 0.21 2.37 0.037 0.31
Downloaded from http://hyper.ahajournals.org/ by guest on April 15, 2017

*P value for test of linear and nonlinear trends.

P0.05.
Model 1: unadjusted.
Model 2: adjusted for age and employment.
Model 3: model 2alcohol consumption, smoking, physical activity, BMI, SF-36 mental, SF-36 physical, depression cases, hypnotics use, and CVD drugs.

lence of hypertension than other participants (in both treated higher risk of hypertension compared with the group sleeping
and untreated individuals). 7 hours (OR: 2.01; 95% CI: 1.13 to 3.58), independent of
several potential confounders, with a significant inverse
Cross-Sectional Analysis linear trend across decreasing hours of sleep (P0.003).
Table 2 displays the odds ratios (ORs) and 95% CIs of
prevalent hypertension across categories of sleep duration at Prospective Analysis
phase 5, using 7 hours of sleep as the reference category. Table 3 displays the ORs (and 95% CIs) of incident hyper-
Among men, no consistent pattern of association was noted. tension at phase 7 among participants who were normoten-
Among women, in fully adjusted analyses, short duration of sive at phase 5. During a mean follow-up of 5 years, the
sleep (5 hours per night) was associated with a significantly cumulative incidence of hypertension was 20.0% (740 of

Table 3. OR (95% CI) of Incident Hypertension at Phase 7 (20022003) Across Categories of Sleep Duration at Phase 5 (19971999):
The Whitehall II Study (n3691)
Sleep Duration P*

Sample Models 5 h 6h 7h 8h 9 h Linear Nonlinear

Men (n2686)
No. 160 900 1224 372 30
Cases, n 29 179 243 85 3
Model 1 0.89 0.58 1.37 1.00 0.81 1.24 1 1.20 0.90 1.58 0.45 0.13 1.49 0.60 0.53
Model 2 0.96 0.62 1.48 1.07 0.86 1.34 1 1.07 0.80 1.42 0.36 0.11 1.19 0.51 0.42
Model 3 0.89 0.55 1.43 1.02 0.80 1.31 1 1.11 0.80 1.53 0.13 0.02 0.98 0.55 0.18
Women (n1005)
No. 75 330 394 186 20
Cases, n 20 77 64 37 3
Model 1 1.88 1.05 3.34 1.56 1.08 2.27 1 1.28 0.82 2.01 0.91 0.26 3.20 0.029 0.38
Model 2 1.94 1.08 3.50 1.56 1.07 2.27 1 1.17 0.74 1.86 0.92 0.26 3.27 0.015 0.42
Model 3 1.31 0.65 2.63 1.42 0.93 2.16 1 0.99 0.59 1.69 1.07 0.29 3.94 0.12 0.61
*P value for test of linear and nonlinear trends.
P0.05.
Model 1: unadjusted.
Model 2: adjusted for age and employment.
Model 3: model 2alcohol consumption, smoking, physical activity, BMI, SF-36 mental, SF-36 physical, depression cases, hypnotics use, and CVD drugs.
 Cappuccio et al Short Sleep Duration and Hypertension 697

3691 total: 539 of 2686 in men and 201 of 1005 in women). Other Analyses
Among men, no consistent pattern of association was seen We also carried out linear regression analyses to test the
across categories of sleep duration. Among women, short association between sleep duration and blood pressures (sys-
duration of sleep was associated with significantly higher tolic, diastolic, and pulse pressures) at phase 5 (cross-
risks of hypertension compared with the group sleeping 7 sectional analyses), as well as the association between sleep
hours in unadjusted analyses, as well as in a reduced model duration at phase 5 and changes in blood pressures between
(age and employment; 6 hours per night: OR: 1.56 [95% CI: the 2 phases (prospective analyses), among participants not
1.07 to 2.27], 5 hours per night: OR: 1.94 [95% CI: 1.08 to taking antihypertensive medications. In cross-sectional anal-
3.50]). However, these associations were attenuated after yses, there were consistent, significant inverse associations
(P0.05) between duration of sleep and either systolic blood
accounting for cardiovascular risk factors and psychiatric
pressure (1.24 mm Hg per hour of sleep; 95% CI:
comorbidities (OR: 1.42 [95% CI: 0.94 to 2.16]; OR: 1.30
2.23 to 0.24 mm Hg per hour of sleep) or pulse pressure
[95% CI: 0.65 to 2.62], respectively). The proportion of the
(0.91 mm Hg per hour of sleep; 95% CI: 1.63 to
variance of the risk of developing hypertension explained by
0.20 mm Hg per hour of sleep), only among women, in
the age- and employment-adjusted model was 2.95%. In the fully adjusted models. In prospective analyses, no significant
fully adjusted model it was 6.14%. The major contributor in associations were found for any of the blood pressure
the full multivariate model was body mass index (explaining measures among either male or female participants (data not
2.23% of the added variance), whereas the remaining covari- shown).
ates all accounted for 1% of the remaining difference.
Downloaded from http://hyper.ahajournals.org/ by guest on April 15, 2017

Results were virtually unchanged after further adjustment for Discussion

the baseline values of systolic blood pressure (data not Findings from the Whitehall II cohort showed gender-specific
shown). associations between sleep duration and hypertension. Spe-

APPENDIX. Baseline Characteristics (Phase 5: 19971999): The Whitehall II Study

n6592* n5766 n612

Characteristics Men Women Men Women Men Women

No. of subjects 4724 1868 4199 1567 329 283
Age, y 55.6 (6.0) 56.3 (6.1) 55.4 (6.0) 56.1 (6.1) 57.7 (5.9) 56.2 (5.4)
DBP, mm Hg 78.6 (10.6) 74.6 (10.0) 78.6 (10.6) 74.6 (10.0) 79.2 (10.8) 78.5 (10.2)
SBP, mm Hg 123.5 (16.0) 121.1 (17.1) 123.6 (16.0) 121.1 (17.1) 125.6 (17.3) 127.5 (19.1)
Pulse pressure, mm Hg 44.9 (11.2) 46.5 (12.4) 45.0 (11.2) 46.5 (12.3) 46.4 (12.6) 49.0 (14.1)
Sleep duration, n (%)
5 h 296 (6.4) 190 (10.4) 265 (6.3) 157 (10.0) 43 (13.6) 34 (12.8)
6h 1521 (32.8) 602 (32.9) 1383 (32.9) 511 (32.6) 121 (38.2) 94 (35.3)
7h 2090 (45.0) 687 (37.6) 1886 (44.9) 597 (38.1) 95 (30.0) 101 (38.0)
8h 683 (14.7) 310 (17.0) 620 (14.8) 272 (17.4) 45 (14.2) 31 (11.7)
9 h 53 (1.1) 40 (2.2) 45 (1.1) 30 (1.9) 13 (4.1) 6 (2.3)
Lowest employment grade, n (%) 207 (4.4) 545 (29.5) 175 (4.2) 435 (27.9) 65 (19.8) 181 (64.4)
BMI, kg/m2 26.1 (3.5) 26.2 (4.9) 26.0 (3.5) 26.2 (4.9) 25.7 (3.2) 28.1 (5.1)
Waist, cm 92.1 (10.1) 80.4 (11.9) 92.0 (10.1) 80.5 (11.9) 90.6 (9.0) 85.1 (11.4)
Weekly alcohol, units 16.7 (16.6) 7.8 (9.5) 16.7 (16.5) 8.1 (9.6) 10.3 (18.3) 2.7 (5.4)
CVD drugs, n (%) 719 (15.3) 299 (16.1) 613 (14.6) 238 (15.2) 89 (27.2) 93 (32.9)
Physical activity, n (%) 2277 (48.2) 601 (32.2) 2062 (49.1) 536 (34.2) 99 (30.1) 47 (16.6)
SF-36 mental, score 51.6 (9.0) 49.7 (10.4) 51.7 (8.9) 49.7 (10.4) 50.0 (10.4) 49.1 (10.8)
SF-36 physical, score 60.0 (7.1) 49.1 (9.6) 52.1 (6.9) 49.5 (9.3) 49.2 (8.7) 44.5 (10.4)
Use of hypnotics, n (%) 26 (0.55) 17 (0.9) 21 (0.5) 14 (0.9) 2 (0.6) 0 (0.0)
Depression cases, n (%) 545 (11.8) 255 (14.1) 472 (11.4) 213 (13.8) 66 (21.4) 57 (22.2)
Current smoking, n (%) 436 (9.3) 266 (14.3) 360 (8.6) 207 (13.3) 40 (12.3) 13 (4.63)
Male hypertensive, n (%) 1194 (25.3) 407 (21.8) 1194 (28.4) 407 (26.0) 110 (33.4) 117 (41.3)
No medication 630 (13.3) 153 (8.2) 713 (17.0) 202 (12.9) 35 (10.6) 32 (11.3)
Medication 564 (11.9) 254 (13.6) 481 (11.5) 205 (13.1) 75 (22.8) 85 (30.0)
Data are expressed as the mean (SD) or as n (%).
*n6592 for all white participants.
n5766 for white participants with complete data on blood pressure.
n612 for ethnic minorities excluded from the analysis.
 698 Hypertension October 2007

cifically, short duration of sleep was associated with higher Strengths

risks of prevalent and incident hypertension only among Unlike these earlier investigations, our study examined both
women. We could not detect consistent associations among cross-sectional and prospective gender-specific associations
men or for long duration of sleep. For the latter finding, it between sleep duration and hypertension with the inclusion of
should be noted, however, that there were very few individ- a number of potential confounding variables. Our findings
uals (1.3% of the overall sample) in our study who reported suggest a potential role of sleep deprivation in the etiology of
sleeping 9 hours per night, thus yielding a limited statistical hypertension and other adverse health outcomes.
power to examine the association between sleep duration and First, the observation that reduced duration of sleep may be
hypertension in this subgroup. Furthermore, our findings associated with a higher risk of hypertension only among
point to the importance of a comprehensive scrutiny of potential women is a novel finding. Indeed, although previous inves-
confounders and mediators when examining the associations tigations have emphasized the potential impact of both short
between durations of sleep and health outcomes. In fact, the and long durations of sleep on chronic disease risk among
observed associations among women were attenuated after women,6 9 the mechanisms underlying the gender-specific
accounting for a number of cardiovascular risk factors, measures association between sleep deprivation and hypertension, ob-
of general health, and psychiatric comorbidities. served in our study, are unknown. Given the mean age of our
A substantial amount of evidence exists about the link female participants falling around the menopausal period
between SDB and hypertension.15 The epidemiological data (55 years), we can speculate that the periods marking shifts
have been corroborated by findings of mechanistic studies in the reproductive stages, such as menopause, are particu-
Downloaded from http://hyper.ahajournals.org/ by guest on April 15, 2017

emphasizing the critical role of sympathetic overactivity in larly vulnerable times for women, because they are associated
the etiology of SDB-related hypertension,24,25 although other with major hormonal turmoil and psychosocial stresses that
mechanisms are likely to be involved.19 More recently, sleep may, in turn, lead to adverse health outcomes.27 For example,
deprivation has been indicated as a risk factor for several in our sample of female participants, the prevalence of
chronic health outcomes in individuals without overt sleep depression cases was higher among women reporting short
disorders.6 12 However, we are aware of only 2 population- duration of sleep (5 hours per night) than in other sub-
based studies so far, both coming from the United States, that groups. In addition, as shown in our descriptive analyses, the
have attempted to examine the association between self- distribution of correlates of short sleep duration that have the
reported durations of sleep and risk of hypertension.13,14 potential to affect hypertension risk was different between
Specifically, in a longitudinal analysis of the NHANES-I genders and may have partially contributed to the observed
(n4810), short sleep duration (5 hours per night) was associations. Finally, we cannot rule out the possibility of
associated with a 60% higher risk of incident hypertension, in differential self-reporting of sleep habits between men and
fully adjusted models, among middle-aged (32 to 59 years) women, as suggested in a previous analysis from the Sleep
American adults without apparent sleep disorders.13 No Heart Health Study examining the relationship of gender to
association was found in individuals 60 years of age. subjective measures of sleepiness.28
However, in this study, the diagnosis of incident hypertension Second, although findings from cross-sectional analyses
was based on self-report with a potential of misclassification consistently showed a strong, significant association between
(underdiagnosis), as suggested by a lower cumulative inci- short sleep duration and risk of hypertension among women,
dence (647 of 4810 [13.5%]) as compared with that in prospective analyses, the risk estimates were attenuated
observed in our study (740 of 3691 [20.0%]), despite com- after accounting for cardiovascular risk factors, measures of
parable age ranges between the 2 studies and a longer general health, and psychiatric comorbidities. Thus, these
follow-up period in the NHANES-I (8 to 10 years versus 5 findings emphasize the importance of a comprehensive ex-
years, respectively). In addition, no gender-specific analyses amination of correlates that are likely to confound or may be
were included in the NHANES-I. Moreover, in a cross- on the causal pathway between sleep deprivation and adverse
sectional analysis of the large sample of the Sleep Heart health outcomes. Nevertheless, recent prospective analyses
Health Study (6000 US adults), a significantly higher from the Monitoring Trends Determinants in Cardiovascular
prevalence of hypertension was reported among either short Disease Augsburg survey indicate a modest but significant
(6 hours per night) or long sleepers (9 hours per night) as association between short sleep duration and incident myo-
compared with the median duration of sleep of 7 to 8 hours cardial infarction in middle-aged women, but not men, from
per night.14 However, the association was stronger among the general population.29
short sleepers than in long sleepers (ie, 66% versus 30% Third, our descriptive analyses clearly demonstrate that
higher risk of hypertension, respectively). Although this both short and long duration of sleep may indeed identify
study accounted for a number of potential covariates, includ- population subgroups with a distinct cluster of sociodemo-
ing psychiatric and cardiovascular comorbidities, the cross- graphic characteristics, lifestyle behaviors, and disease con-
sectional design does not allow us to exclude the possibility ditions that are likely to be affected by the cultural setting in
of residual confounding by unknown variables, as well as to which the research is being conducted.30 33 In this regard, the
exclude the potential of reverse causality. In addition, the study of health consequences related to curtailments of sleep
Sleep Heart Health Study sample cohort was, on average, seems to be epidemiologically relevant in the general popu-
older than those in both NHANES-I and Whitehall II, thus lation, given the downward trends in the average duration of
with a higher likelihood of geriatric comorbidities potentially sleep and the increasingly higher prevalence of short sleep-
affecting sleep patterns.26 ers in many Western countries.34,35 Conversely, it may be
 Cappuccio et al Short Sleep Duration and Hypertension 699

difficult for epidemiological studies to examine the health Sources of Funding

consequences of long durations of sleep in middle-aged, The Whitehall II Study has been supported by grants from the
healthy populations considering the relatively low preva- Medical Research Council; British Heart Foundation; Health and
lence of long sleepers in the absence of overt psychiatric Safety Executive; Department of Health; National Heart Lung and
Blood Institute (HL36310), National Institutes of Health; National
comorbidities.36 Institute on Aging (AG13196), National Institutes of Health; Agency
for Health Care Policy Research (HS06516); and the John D. and
Limitations Catherine T. MacArthur Foundation Research Networks on Success-
There are limitations in this study. First, the population under ful Midlife Development and Socioeconomic Status and Health.
J.E.F. is supported by the Medical Research Council (grant
investigation is an occupational cohort of white-collar work-
G8802774), M.J.S. by a grant from the British Heart Foundation, and
ers and limited to whites, which may reduce the generaliz- M.G.M. by a Medical Research Council research professorship.
ability of our findings to other populations. However, this
would not affect the internal validity of our results with Disclosures
respect to the prospective analyses. Second, information F.P.C. holds the Cephalon Chair, an endowed post at Warwick
about sleep duration was self-reported by the participants. Medical School, the result of a donation from the company. The
appointment to the chair was made entirely independent of the
Nevertheless, self-report assessments of sleep have been company, and the postholder is free to devise his own program of
shown to be valid measures compared with quantitative sleep research. Cephalon do not have any stake in intellectual property
assessments with actigraphy.37,38 Moreover, because the out- associated with the postholder, and the chair has complete academic
come was also assessed prospectively, any misclassification independence from the company. The remaining authors report no
conflicts.
Downloaded from http://hyper.ahajournals.org/ by guest on April 15, 2017

of sleep duration would be nondifferential with respect to

incident hypertension, thus resulting in underestimation of the
References
true effects. A further limitation of this study is the relatively 1. Young T, Peppard P, Palta M, Hla KM, Finn L, Morgan B, Skatrud J.
short time of follow-up (5 years), which may have precluded Population-based study of sleep-disordered breathing as a risk factor for
us from detecting larger and significant effects of sleep hypertension. Arch Intern Med. 1997;157:1746 1752.
2. Nieto FJ, Young TB, Lind BK, Shahar E, Samet JM, Redline S,
deprivation on subsequent hypertension incidence. The DAgostino RB, Newman AB, Lebowitz MD, Pickering TG. Association
strengths of this study include the simultaneous inclusion of of sleep-disordered breathing, sleep apnea, and hypertension in a large
a number of covariates known to be related to both sleep community-based study. JAMA. 2000;283:1829 1836.
3. Peppard PE, Young T, Palta M, Skatrud J. Prospective study of the
patterns and hypertension. A further strength is that the
association between sleep-disordered breathing and hypertension. N Engl
diagnosis of incident hypertension was also based on directly J Med. 2000;342:1378 1384.
measured blood pressures at both examinations, thus mini- 4. Duran J, Esnaola S, Rubio R, Iztueta A. Obstructive sleep apnea-
mizing the potential of misclassification that occurs when hypopnea and related clinical features in a population-based sample of
subjects aged 30 to 70 yr. Am J Respir Crit Care Med. 2001;163:
using self-report alone. 685 689.
5. Haas DC, Foster GL, Nieto J, Redline S, Resnick HE, Robbins JA, Young
Conclusions T, Pickering TG. Age-dependent associations between sleep-disordered
In summary, findings from the Whitehall II cohort suggest breathing and hypertension. Circulation. 2005;111:614 621.
6. Ayas NT, White DP, Manson JE, Stampfer MJ, Speizer FE, Malhotra A,
gender-specific associations between sleep duration and hy- Hu FB. A prospective study of sleep duration and coronary heart disease
pertension risk. Specifically, cross-sectional analyses showed in women. Arch Intern Med. 2003;163:205209.
a significant, consistent association between short sleep 7. Ayas NT, White DP, Al-Delaimy WK, Manson JE, Stampfer MJ, Speizer
FE, Patel S, Hu FB. A prospective study of self-reported sleep duration
duration (5 hours per night) and risk of hypertension only and incident diabetes in women. Diabetes Care. 2003;26:380 384.
among women, which was attenuated in prospective analyses 8. Patel SR, Ayas NT, Malhotra MR, White DP, Schernhammer ES, Speizer
after multivariate adjustment. FE, Stampfer MJ, Hu FB. A prospective study of sleep duration and
mortality risk in women. Sleep. 2004;27:440 444.
9. Patel SR, Malhotra A. White DP, Gottlieb DJ, Hu FB. Association
Perspectives between reduced sleep and weight gain in women. Am J Epidemiol.
Sustained sleep curtailment, ensuing excessive daytime sleep- 2006;164:947954.
iness, and the higher cardiovascular risk are causes for 10. Tamakoshi A, Ohno Y. Self-reported sleep duration as a predictor of
all-cause mortality: results from the J Am Coll Cardiol study, Japan.
concern. Emerging evidence also suggests a potential role for Sleep. 2004;27:5154.
sleep deprivation as a predictor or risk factor for conditions 11. Gangwisch JE, Malaspina D, Boden-Albala B, Heymsfield SB. Inad-
like obesity, diabetes, and metabolic syndrome not only in equate sleep as a risk factor for obesity: analyses of the NHANES I.
Sleep. 2005;28:1289 1296.
adults39 but also in children.40 Further prospective studies 12. Kohatsu ND, Tsai R, Young T, VanGilder R, Burmeister LF, Stromquist
with improved assessment of long-term exposure (repeated AM, Merchant JA. Sleep duration and body mass index in a rural
self-reported sleep duration or repeated actigraphy) and better population. Arch Intern Med. 2006;166:17011705.
13. Gangwisch JE, Heymsfield SB, Boden-Albala B, Buijs RM, Kreier F,
control for confounders are needed before causality can be
Pickering TG, Rundle AG, Zammit GK, Malaspina D. Short sleep
determined. duration as a risk factor for hypertension: analyses of the first National
Health and Nutrition Examination Survey. Hypertension. 2006;47:
833 839.
Acknowledgments 14. Gottlieb DJ, Redline S, Nieto FJ, Baldwin CM, Newman AB, Resnick
We thank all of the participating civil service departments and their HE, Punjabi NM. Association of usual sleep duration with hypertension:
welfare, personnel, and establishment officers; the Occupational the Sleep Heart Health Study. Sleep. 2006;29:1009 1014.
Health and Safety Agency; the Council of Civil Service Unions; all 15. Tochikubo O, Ikeda A, Miyajima E, Ishii M. Effects of insufficient
of the participating civil servants in the Whitehall II Study; and all sleep on blood pressure monitored by a new multibiomedical recorder.
members of the Whitehall II Study team. Hypertension. 1996;27:1318 1324.
 700 Hypertension October 2007

16. Lusardi P, Mugellini A, Preti P, Zoppi A, Derosa G, Fogari R. Effects of daytime somnolence in the Sleep Heart Health Study. Sleep. 2004;27:
a restricted sleep regimen on ambulatory blood pressure monitoring in 305311.
normotensive subjects. Am J Hypertens. 1996;9:503505. 29. Meisinger C, Heier M, Lowel H, Schneider A, Doring A. Sleep duration
17. Lusardi P, Zoppi A, Preti P, Pesce RM, Piazza E, Fogari R. Effects of and sleep complaints and risk of myocardial infarction in middle-aged
insufficient sleep on blood pressure in hypertensive patients: a 24-h study. men and women from the general population. The MONICA/KORA
Am J Hypertens. 1999;12:63 68. Augsburg Cohort Study. Sleep. In press.
18. Miller MA, Cappuccio FP. Inflammation, sleep, obesity and cardiovas- 30. Moore P, Adler N, Williams D, Jackson J. Socioeconomic status and
cular disease. Curr Vasc Pharmacol. 2007;5:95102. health: the role of sleep. Psychosom Med. 2002;64:337344.
19. Wolk R, Somers VK. Sleep and the metabolic syndrome. Exp Physiol. 31. Groeger JA, Zijlstra FR, Dijk DJ. Sleep quantity, sleep difficulties and
2007;92:6778. their perceived consequences in a representative sample of some 2000
20. Scheer FA, Van Montgrans GA, van Someren EJ, Mairuhu G, Buijs RM. British adults. J Sleep Res. 2004;13:359 371.
Daily night-time melatonin reduces blood pressure in male patients with 32. Adams J. Socioeconomic position and sleep quantity in UK adults.
essential hypertension. Hypertension. 2004;43:192197. J Epidemiol Community Health. 2006;60:267269.
21. Zee PC, Turek FW. Sleep and health: everywhere and in both directions.
33. Patel SR, Malhotra A, Gottlieb DJ, White DP, Hu FB. Correlates of long
Arch Intern Med. 2006;166:1686 1688.
sleep duration. Sleep. 2006;29:881 889.
22. Marmot MG, Davey Smith G, Stansfeld S, Patel C, North F, Head J,
34. National Sleep Foundation. Sleep in America Poll 2005: Summary of
White I, Brunner E, Feeney A. Health inequalities among British civil
Findings. Washington, DC: National Sleep Foundation; 2005.
servants: the Whitehall II study. Lancet. 1991;337:13871393.
35. Egan BM. Sleep and hypertension: burning the candle at both ends really
23. Brazier JE, Harper R, Jones NM, OCathain A, Thomas KJ, Usherwood
T, Westlake L. Validating the SF-36 health survey questionnaire: new is hazardous to your health. Hypertension. 2006;47:816 817.
outcome measure for primary care. BMJ. 1992;305:160 164. 36. Knutson KL, Turek FW. The U-shaped association between sleep and
24. Narkiewicz K, Somers VK. The sympathetic nervous system and health: the 2 peaks do not mean the same thing. Sleep. 2006;29:878 879.
obstructive sleep apnea: implications for hypertension. J Hypertens. 37. Hauri PJ, Wisbey J. Wrist actigraphy in insomnia. Sleep. 1992;15:
1997;15:16131616. 293301.
Downloaded from http://hyper.ahajournals.org/ by guest on April 15, 2017

25. Fletcher EC. Sympathetic overactivity in the etiology of hypertension of 38. Lockley SW, Skene DJ, Arendt J. Comparison between subjective and
obstructive sleep apnea. Sleep. 2003;26:1519. actigraphic measurement of sleep and sleep rhythms. J Sleep Res. 1999;
26. Quan SF, Zee P. Evaluating the effects of medical disorders on sleep in 8:175178.
the older patient. Geriatrics. 2004;59:37 42. 39. Knutson KL, Spiegel K, Penev P, Van Cauter E. The metabolic conse-
27. Parry BL, Newton RP. Chronobiological basis of female-specific mood quences of sleep deprivation. Sleep Medicine Rev. 2007;11:163178.
disorders. Neuropsychopharmacology. 2001;25:S102S108. 40. Currie A, Cappuccio FP. Sleep in children and adolescents: a worrying
28. Baldwin CM, Kapur VK, Holberg CJ, Rosen C, Nieto FJ. Sleep Heart scenario. Can we understand the sleep deprivation obesity epidemic?
Health Study Group. Associations between gender and measures of Nutr Metab Cardiovasc Dis. 2007;17:230 232.
 Gender-Specific Associations of Short Sleep Duration With Prevalent and Incident
Hypertension: The Whitehall II Study
Francesco P. Cappuccio, Saverio Stranges, Ngianga-Bakwin Kandala, Michelle A. Miller,
Frances M. Taggart, Meena Kumari, Jane E. Ferrie, Martin J. Shipley, Eric J. Brunner and
Michael G. Marmot
Downloaded from http://hyper.ahajournals.org/ by guest on April 15, 2017

Hypertension. 2007;50:693-700; originally published online September 4, 2007;

doi: 10.1161/HYPERTENSIONAHA.107.095471
Hypertension is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright 2007 American Heart Association, Inc. All rights reserved.
Print ISSN: 0194-911X. Online ISSN: 1524-4563

The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://hyper.ahajournals.org/content/50/4/693

An erratum has been published regarding this article. Please see the attached page for:
/content/50/5/e170.full.pdf

Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Hypertension can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial
Office. Once the online version of the published article for which permission is being requested is located,
click Request Permissions in the middle column of the Web page under Services. Further information about
this process is available in the Permissions and Rights Question and Answer document.

Reprints: Information about reprints can be found online at:

http://www.lww.com/reprints

Subscriptions: Information about subscribing to Hypertension is online at:

http://hyper.ahajournals.org//subscriptions/
 Correction

In the Hypertension article by Cappuccio et al (Cappuccio FP, Stranges S, Kandala N-B, Miller
MA, Taggart FM, Kumari M, Ferrie JE, Shipley MJ, Brunner EJ, Marmot MG. Gender-specific
associations of short sleep duration with prevalent and incident hypertension. The Whitehall II
study. Hypertension. 2007;50:693700), the OR, 95% CI, and P values reported in the Abstract
and Results as OR: 2.01; 95% CI: 1.13 to 3.58; P 0.003 should be OR: 1.72; 95% CI: 1.07 to
2.75; P 0.037. These values are listed correctly in Table 2. The authors regret the error.

(Hypertension. 2007;50:e170.)
2007 American Heart Association, Inc.
Hypertension is available at http://hyper.ahajournals.org DOI: 10.1161/HYPERTENSIONAHA.107.009592

e170