Sleep Medicine 14 (2013) 1157–1163

Contents lists available at ScienceDirect

Sleep Medicine
journal homepage: www.elsevier.com/locate/sleep

Original Article

Poor sleep quality and resistant hypertension

Rosa Maria Bruno a,d, Laura Palagini b,⇑, Angelo Gemignani c, Agostino Virdis d, Alessia Di Giulio d,
Lorenzo Ghiadoni d, Dieter Riemann e, Stefano Taddei d
a
Institute of Clinical Physiology-CNR, Pisa, Italy
b
Department of Neuroscience, Sleep Center, University of Pisa, Via Roma 67, 56100 Pisa, Italy
c
Department of Surgical, Medical, and Molecular Pathology and of the Critical Area, University of Pisa, Pisa, Italy
d
Department of Clinical and Experimental Medicine, University of Pisa, Via Roma 67, 56126 Pisa, Italy
e
Department of Psychiatry and Psychotherapy, Freiburg University Medical Center, Freiburg, Germany

a r t i c l e i n f o a b s t r a c t

Article history: Objectives: We aimed to determine the relationship between sleep quality and treatment-resistant
Received 24 January 2013 hypertension (RH).
Received in revised form 18 April 2013 Methods: In our cross-sectional cohort study, 270 consecutive essential hypertensive patients were
Accepted 24 April 2013
recruited at the Outpatient Hypertension Unit, University of Pisa, Italy. The Pittsburgh Sleep Quality Index
Available online 28 August 2013
(PSQI), Beck Depression Inventory (BDI), and State-Trait Anxiety Inventory (STAI-Y2) were administered
to all subjects. RH was defined as office blood pressure (BP) >140/90 mmHg with three or more antihy-
Keywords:
pertensive drugs or controlled BP with four or more drugs. Poor sleep quality was defined as PSQI >5,
Sleep quality
Hypertension
depressive symptoms as BDI >10, and trait anxiety as STAI-Y2 >40. Patients with other sleep disorders
Resistant hypertension were excluded.
Gender differences Results: Complete data were available for 222 patients (50.9% men; mean age, 56.6 ± 12.5 y; RH, 14.9%).
Insomnia Poor sleep quality had a prevalence of 38.2% in the overall population. RH was associated with poor sleep
Sleep disorders quality, increased sleep latency and reduced sleep efficiency. No significant relationship was found
between RH and short sleep duration or depressive symptoms and trait anxiety. Poor sleep quality
was more prevalent in resistant vs nonresistant hypertensive women (70.6% vs 40.2%; P = .02) but not
in resistant vs nonresistant men (43.8% vs 29.2%; P = .24). In women poor sleep quality was an indepen-
dent predictor of RH, even after adjustment for cardiovascular and psychiatric comorbidities (odds ratio
[OR], 5.3 [confidence interval {CI}, 1.1–27.6), explaining 4.7% of its variance. In men age, diabetes mellitus
(DM), and obesity were the only variables associated with RH.
Conclusions: Poor sleep quality is significantly associated with resistance to treatment in hypertensive
women, independent of cardiovascular and psychiatric confounders.
Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction and cardiovascular events [4–7]. In a cross-sectional population-

based study, insomnia and short sleep duration were indepen-
There is growing evidence of a relationship between sleep dis- dently associated with a higher prevalence of hypertension, which
orders and cardiovascular disease. In particular, sleep disorders further increased in the presence of both conditions [8]. Prospec-
are linked through several pathophysiologic mechanisms to car- tive studies also provided evidence of increased incidence of
diovascular risk factors, such as hypertension, diabetes mellitus new-onset hypertension in subjects with insomnia or short sleep
(DM), and obesity, thus increasing the incidence of cardiovascular duration [5,6], though the relationship was attenuated or nonsig-
events [1,2]. This relationship is well-established for obstructive nificant after considering confounding factors [4]. The question of
sleep apnea syndrome (OSAS) [3], but it also has been suggested causality between disordered sleep and cardiovascular disease is
for other sleep disorders such as insomnia and short sleep duration obscured by the fact that a variety of cardiovascular risk factors
[4]. Indeed both insomnia and short sleep duration have been asso- and psychiatric disorders might simultaneously contribute to both
ciated with increased prevalence and incidence of hypertension insomnia and hypertension [9–11], making it difficult to unravel if
an independent cause-effect relationship that exists between these
two conditions. Furthermore, the relationship appears to be gen-
⇑ Corresponding author. Address: Department of Neuroscience, Sleep Center,
der-specific, being more prevalent in women than in men [12].
University of Pisa, Via Roma 67, 56126 Pisa, Italy. Tel.: +39 050 993165; fax: +39
050 992656.
The role of global sleep quality in hypertension severity is even
E-mail addresses: [email protected], [email protected] (L. Palagini). less established. In particular, the condition of resistant hyperten-

1389-9457/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.sleep.2013.04.020
1158 R.M. Bruno et al. / Sleep Medicine 14 (2013) 1157–1163

sion (RH) is of interest, as it identifies a nonnegligible percentage of Medical and pharmacologic history was recorded to identify
patients with a more severe form of hypertension, higher cardio- cardiovascular risk factors and disease. Habitual physical activity
vascular risk, and negative prognosis [13]. was defined as brisk walking or sports activities for at least 1 h
Thus we hypothesized that poor sleep quality or short sleep per week. Current use of cardiovascular and psychiatric drugs also
duration might be associated with increased hypertension severity. was recorded.
Accordingly, the aim of our study was to investigate the prevalence
and features of poor sleep quality in a population of patients 2.3. Questionnaires
accessing a Hypertension Outpatient Unit for the first time, to as-
sess if RH is associated with poor sleep quality or short sleep dura- 2.3.1. Pittsburgh Sleep Quality Index
tion. Moreover, the role of gender and cardiovascular and The PSQI is a widely used, self-rated, standardized question-
psychiatric comorbidities in the relationship between RH and sleep naire assessing sleep quality in the previous month [17]. The 19
quality also was evaluated. questions are grouped in seven component scores, each exploring
a different sleep feature; the sum yields a global PSQI score used
to define poor sleep quality when >5. On the basis of the sleep
2. Methods duration component of the PSQI score, self-reported short sleep
duration was defined as <6 h of sleep per night. The following
2.1. Patients PSQI-derived data also were analyzed: increased sleep latency
(>30 min), reduced sleep efficiency (<85%), sleep disturbance
Starting in May 2011, 270 consecutive patients attending their (sleep disturbances component score >1), and daytime dysfunction
first visit to the Hypertension Outpatient Unit of the University due to sleepiness (daytime dysfunction component score >1).
Hospital of Pisa, Italy were enrolled. Inclusion criteria were age
of >18 years, written informed consent, and diagnosis of essential
arterial hypertension according to current guidelines [14]. Criteria 2.3.2. Beck Depression Inventory
of exclusion were: known or suspected (according to clinical judg- To consider the frequent comorbidity of sleep disorders with
ment) secondary hypertension, including drug-induced hyperten- depression, the Beck Depression Inventory (BDI) was administered.
sion; hypertension in pregnancy; previous diagnosis of OSAS or The BDI is a 21-question inventory for self-assessment and is one
restless legs syndrome (RLS); severe chronic heart failure or hepa- of the most widely used instruments for measuring the severity
tic insufficiency; end-stage renal disease; and active neoplastic dis- of depression. The presence of depressive symptoms was defined
ease. On the basis of these inclusion and exclusion criteria, 436 with a BDI score >10 [18,19].
patients were screened to enroll 270 patients, with a rate of accep-
tance of 61.9%. Furthermore, patients with a score of 1 or more on 2.3.3. State-Trait Anxiety Inventory
item 10 of the Pittsburgh Sleep Quality Index (PSQI), regarding self- To consider the frequent comorbidity of sleep disorders with
reported symptoms or symptoms reported by the patient’s room- anxiety, trait anxiety was assessed using the State-Trait Anxiety
mate that were compatible with OSAS or RLS (i.e., snoring), were Inventory (STAI-Y2) and was defined as a score >40 [20].
excluded from the analysis after enrollment.
The study conformed to the Declaration of Helsinki and was ap-
2.4. Statistical analysis
proved by the local Ethical Committee. All patients provided writ-
ten informed consent prior to entering the study.
Statistical analysis was performed using NCSS 2008 (NCSS:
Kaysville, Utah, USA). For normally distributed data, results were
expressed as mean ± standard deviation (SD), whereas median va-
2.2. Measurements
lue and 25% to 75% interquartile range were used for abnormally
distributed data. Differences in means between groups were ana-
After enrollment, patients were asked to come to the Hyperten-
lyzed using analysis of variance for normally distributed variables
sion Outpatient Unit after an overnight fasting period. A blood
or the Kruskal–Wallis z test for abnormally distributed variables;
sample was drawn for routine examination (e.g., lipid profile, blood
categorical variables were analyzed using the v2 test. Question-
fasting glucose, serum creatinine, urinalysis), determined accord-
naire results and cardiovascular risk factors were expressed as bin-
ing to standard laboratory procedures. Hypercholesterolemia and
ary discrete variables to build multiple logistic regression models
DM were defined according to current guidelines [15,16] or in
including RH as a dependent variable. We analyzed the effects of
the presence of current lipid-lowering or glucose-lowering treat-
age, gender, and of a panel of cardiovascular (e.g., previous cardio-
ment, respectively. Glomerular filtration rate was estimated using
vascular events, DM, obesity) and psychiatric comorbidities (e.g.,
the Modification of Diet in Renal Disease formula, and chronic kid-
psychiatric medications use, depressive symptoms, anxiety) as
ney disease was defined as glomerular filtration rate <60 mL/min/
well as of short sleep duration, chosen according to a significant
1.73 m2.
association with RH in our population or in previous studies.
Office blood pressure (BP) was measured at the brachial level in
Chronic kidney disease was not inserted in the model, despite sig-
the sitting position by a trained physician, with the patients resting
nificant association with RH for high collinearity correlation with
for at least 10 min under quiet environmental conditions. BP mea-
DM. For the same reason, menopause status could not be consid-
surement was repeated at least three times at 2-min intervals
ered as a confounding factor (resistant hypertensive women were
using an automatic oscillometric device (OMRON-705IT, Omron
100% in postmenopause). The analysis was then stratified by
Corporation, Kyoto, Japan). Average BP was then calculated for
gender.
the last two measurements. On this basis of these measurements,
RH was defined as office BP >140/90 mmHg with three or more
antihypertensive drugs or controlled BP with four or more drugs, 3. Results
including a diuretic, according to current guidelines [13,14].
Weight and height were measured and a complete clinical exami- Among the 270 patients enrolled, those with self-reported (or
nation was performed; obesity was defined as a body mass index reported by roommate) sleep apneas, snoring, and leg restlessness
(BMI) >30 kg/m2. (n = 35) or with incomplete data (n = 13) were excluded. The final
 R.M. Bruno et al. / Sleep Medicine 14 (2013) 1157–1163 1159

Table 1
Clinical characteristics of the study population.

Overall (n = 222) Gender Resistant hypertension

Men (n = 113) Women (n = 109) No (n = 189) Yes (n = 33)
Age (y) 56.6 ± 12.5 55.4 ± 12.2 57.8 ± 12.8 54.9 ± 12.0 65.3 ± 11.9 
Men (%) 50.9 – – 48.5 51.5
Currently smoking (%) 15.1 16.2 14.0 16.2 9.1
Hypertension duration (y) 6 (2–12) 5 (2–11) 7 (3–12) 5 (2–10) 15 (8–30) 
Habitual physical activity (%) 37.0 38.8 35.0 38.1 30.8
Diabetes mellitus (%) 7.7 8.0 7.3 4.8 24.2 
BMI (kg/m2) 27.0 ± 4.4 27.7 ± 3.9 26.3 ± 4.9* 26.9 ± 4.5 27.7 ± 4.0
Obesity (%) 21.3 25.3 17.5 20.0 28.6
Hypercholesterolemia (%) 59.6 72.9 47.4* 55.9 65.1
Previous CV events (%) 11.7 13.3 10.0 9.1 12.2
Chronic kidney disease (%) 13.5 13.6 13.3 9.3 34.8 
Framingham Risk Score 19.3 25.4 13.4* 12.8 30.8 
Systolic BP (mmHg) 138.6 ± 17.4 141.3 ± 14.9 135.8 ± 19.4* 136.7 ± 15.9 149.8 ± 21.2 
Diastolic BP (mmHg) 81.9 ± 11.6 83.2 ± 11.8 80.4 ± 11.1* 81.7 ± 11.3 83.0 ± 13.0
Pulse pressure (mmHg) 56.7 ± 13.8 58.1 ± 12.0 55.3 ± 15.3 55 ± 12.0 66.9 ± 18.1 
Heart rate (bpm) 72.2 ± 13.4 71.5 ± 12.6 72.9 ± 14.3 73.0 ± 13.8 67.4 ± 9.2 
BP-lowering treatment (%) 84.2 82.3 86.2 – –
BP-lowering drugs (n) 2.0 ± 1.0 2.0 ± 1.1 2.0 ± 0.9 1.3 ± 1.0 3.6 ± 1.0 
Resistant hypertension (%) 14.9 14.2 15.6 – –

Abbreviations: y, years; BMI, body mass index; CV, cardiovascular; BP, blood pressure; bpm, beats per minute; n, number.
*
P < .05 vs men.
 
P < .05 vs nonresistant hypertensive patients.

analysis was performed on 222 patients. Clinical characteristics of mon in women than in men, whereas the use of psychiatric drugs
the study population are listed in Table 1. was similar for both sexes (11.1 vs 12.0%; P = .84).
The study population showed a median hypertension duration
of 6 years; 84.2% of the patients were on BP-lowering drugs, with
a mean number of two drugs; and 14.9% presented RH with no 3.1. Relationship between poor sleep quality and RH in the overall
gender differences. No gender differences in cardiovascular risk population
factors and disease prevalence were found, except for a higher
BMI and a greater prevalence of hypercholesterolemia in men, Patients with RH (n = 33) were older and had an increased prev-
who also showed higher BP values (Table 1). Among hypertensive alence of DM and chronic kidney disease compared with the non-
women, postmenopausal status was present in 73.5% of the overall resistant counterparts (Table 1). They also showed higher systolic
sample and in 100% of women with RH. and pulse pressure values, were treated by definition with a great-
Sleep features of the study population are listed in Table 2. Poor er number of antihypertensive drugs, and showed lower heart rate
sleep quality, defined as a global PSQI score >5, was present in in comparison to their counterparts (Table 1), possibly due to more
38.2% of the studied population; short sleep duration was present frequent b blocker use (48.5 vs 13.8%; P < .001).
in 31.3% with a mean sleep duration of 6.6 ± 1.6 h/night; depres- Treatment-resistant patients had a significantly higher preva-
sive symptoms were present in 13.3% of the population; and anx- lence of poor sleep quality (57.6 vs 34.6%; P = .01) and higher PSQI
iety was present in 43.8% of the sample. Women reported sleeping global score (5 [2–9] vs 3 [1–6]; P = .007) than nonresistant pa-
fewer hours per night than men (6.4 ± 1.6 vs 6.8 ± 1.6 h/night; tients. Analyzing components of PSQI score, resistant hypertensive
P = .05) and presented a higher prevalence of poor sleep quality patients showed increased sleep latency and daytime dysfunction
(46.1 vs 30.5%; P = .01) and higher PSQI global score (4 [1–8] vs due to sleepiness as well as a trend (P = .06) for reduced sleep effi-
2.5 [1–5]; P = .005). Depressive symptoms (20.0 vs 6.8%; P = .003) ciency but not reduced sleep duration (Table 2). Use of psychiatric
and anxiety (54.8 vs 33.1%; P < .001) were significantly more com- drugs and presence of anxiety and depressive symptoms were sim-
ilarly prevalent in patients with or without RH (Table 3). Binary lo-

Table 2
Sleep features of the study population.

Overall (n = 222) Gender Resistant hypertension

Men (n = 113) Women (n = 109) No (n = 189) Yes (n = 33)
Poor sleep quality (%) 38.2 30.5 46.1* 34.6 57.6 
Hours of sleep/night 6.6 ± 1.5 6.8 ± 1.4 6.4 ± 1.5* 6.6 ± 1.5 6.3 ± 1.6
Short sleep duration (%) 31.6 27.1 35.7 30.3 42.4
Sleep latency > 30 min (%) 19.7 16.1 22.6 16.5 33.3 
Sleep efficiency <85% (%) 37.6 31.4 43.5* 34.6 51.5
Sleep disturbance (%) 69.1 66.9 71.9 67.9 72.7
Day dysfunction due to sleepiness (%) 16.2 11.9 20.9* 13.8 27.3 
Psychiatric drugs users (%) 11.6 12.0 11.1 12.2 6.1
Depressive symptoms (%) 13.3 6.8 20.0* 11.2 18.2
Trait anxiety (%) 43.8 33.1 54.8* 44.1 42.4

Abbreviations: n, number of patients; min, minutes.

*
P < .05 vs men.
 
P < .05 vs nonresistant hypertensive patients.
1160 R.M. Bruno et al. / Sleep Medicine 14 (2013) 1157–1163

Table 3
Antihypertensive drug use and poor sleep quality.

Drug classes Prevalence of drug use (%) in hypertensive patients OR (95% CI) for PSQI >5
Overall PSQI <5 PSQI >5 Unadjusted Adjusted
Diuretics 32.7 25.7 44.0* 1.8 (1.1–3.7) 1.8 (0.8–3.8)
b blockers 19.1 19.9 17.9 0.8 (0.4–1.7) –
Calcium channel blockers 38.9 35.3 45.0 1.5 (0.9–2.6) –
RAS blockers 66.8 62.5 73.8 1.7 (0.9–3.1) –
a1 antagonists 9.1 11.0 6.0 0.7 (0.2–1.8) –
Others 2.7 2.2 3.6 1.6 (0.3–8.3) –

Abbreviations: OR, odds ratio; CI, confidence interval; PSQI, Pittsburgh Sleep Quality Index; RAS, renin–angiotensin system.
Odds ratio (OR) was obtained considering a Pittsburgh Sleep Quality Index (PSQI) >5 as a dependent variable and diuretic use as an independent variable. Adjusted OR was
obtained and also included age, gender, previous cardiovascular events, diabetes mellitus, obesity, depressive symptoms, trait anxiety, and psychiatric drug use in the model.
*
P < .05 vs patients with PSQI <5.

gistic regression analysis was performed in the overall population, 3.2. Relationship between poor sleep quality and resistant
including RH as a dependent variable and poor sleep quality as an hypertension: role of gender
independent variable. Poor sleep quality was associated with an in-
creased probability of RH in the unadjusted analysis (odds ratio Sleep characteristics in resistant and nonresistant hypertensive
[OR], 2.6 [confidence interval {CI}, 1.2–5.5]) and in a model that in- patients were analyzed according to gender. The prevalence of
cluded age, gender, sleep duration, and psychiatric comorbidities PSQI global score >5 was higher in resistant than in nonresistant
as confounding factors (OR, 2.7 [CI, 1.1–7.0]; Fig. 1). When cardio- hypertensive women (70.6% vs 40.2%; P = .02), while no significant
vascular comorbidities also were included in the model, poor sleep difference was found in men (43.8 vs 29.2%; P = .24).
quality was associated with RH with a borderline significance (OR, In women, binary logistic regression analysis revealed that poor
3.3 [CI, 1.0–10.5] explaining 2.4% of the variance), whereas age (OR, sleep quality was associated with an increased probability of hav-
4.5, [CI, 1.6–12.2]) and DM (OR, 5.3 [CI, 1.4–20]) remained signifi- ing RH in the unadjusted analysis (OR, 3.6 [CI, 1.2–11.0]) and in a
cant determinants (Fig. 1). In particular, DM and age were respon- model that included sleep duration and psychiatric and cardiovas-
sible for 3.4% and 4.9% of the variance of resistant hypertension, cular comorbidities as confounding factors (OR, 5.3 [CI, 1.1–27.6];
respectively. Fig. 1). In particular, poor sleep quality was responsible for 4.7%

Odds ratio (95% R2 (full model)

confidence
interval)
2.6 (1.2-5.5) 0.027

2.7 (1.1-7.0) 0.107

3.3 (1.0-10.5) 0.141

3.6 (1.2-11.0) 0.048

5.3 (1.3-21.0) 0.118

5.3 (1.1-27.6) 0.138

1.9 (0.6-5.6) 0.011

1.4 (0.3-5.8) 0.104

1.6 (0.3-9.8) 0.240

Fig. 1. Odds ratios (OR) and 95% confidence intervals (CI) in the multiple logistic regression analysis, considering resistant hypertension as a dependent variable and poor
sleep quality as an independent variable. ORs are shown for the overall population (black circles), women (white circles), and men (grey circles). Model 1 included age,
gender, short sleep duration, psychiatric medications, depressive symptoms, and trait anxiety as confounders. Model 2 included the same variables as Model 1, plus previous
cardiovascular events, diabetes mellitus, and obesity.
 R.M. Bruno et al. / Sleep Medicine 14 (2013) 1157–1163 1161

of the variance of RH in hypertensive women in the fully adjusted duration and hypertension appear to be robustly linked in current
model. On the contrary, poor sleep quality was not significantly literature, evidence regarding sleep quality is more elusive, possi-
associated with an increased probability of RH in men, either in bly due to the heterogeneity of diagnostic methods used [4]. Thus
the unadjusted (OR, 1.9 [CI, 0.6–5.6]) or in the fully adjusted model we used a standardized assessment of sleep quality by PSQI, which
(OR, 1.6 [CI, 0.3–9.8]). In the latter model, age, obesity, and DM is a validated and reliable tool [17]. Vgontzas et al. [8] showed a
were independent predictors of resistant hypertension, explaining lack of an association between short sleep duration and hyperten-
6.3%, 4.6%, and 5.9% of its variance, respectively (Fig. 1). sion in the absence of sleep complaints, suggesting that personal
To exclude the confounding role of menopause, we compared perception of sleep quality might have clinical consequences and
resistive hypertensive women with an age-matched group of 35 provide complementary information to objective sleep parameters
nonresistant, postmenopausal women, selected with a 1:2 ratio. [8]. In our population, short sleep duration and poor sleep quality
Resistant hypertensive women showed longer hypertension dura- largely overlapped (65.6% of patients with poor sleep quality slept
tion (14 [9–40] vs 8 [3–13]; P = .02], higher BP values and number less than 6 h per night); therefore, it was not possible to separately
of antihypertensive drugs. Global PSQI was significantly increased study the features of subjects with poor sleep quality but normal
in resistant compared to nonresistant hypertensive women (6 [3– sleep duration, due to small sample size. Future studies should ad-
11] vs 4 [2–8]; P = .03] mainly due to reduced sleep efficiency dress this crucial issue. It also should be noted that self-reported
(70.6% vs 31.4%; P = .08), while the remaining PSQI components, sleep duration systematically overestimates objective sleep dura-
BDI and STAI, were similar in the two groups. tion, even though the two variables are tightly correlated [23].
Our study suggests a gender-specific relationship between sleep
quality and resistant hypertension. This finding is in agreement with
3.3. Relationship between poor sleep quality and resistant
previous studies showing an association between sleep loss and
hypertension: role of antihypertensive drugs
hypertension incidence in women [12,24]. Our study confirms an in-
creased prevalence of poor sleep quality in women than in men;
All drug classes except for angiotensin-converting enzyme
however, the actual sources of this difference are unknown [25].
inhibitors (39.4 vs 30.9%; P = .33) were significantly more fre-
Interestingly, women showed similar bias in self-reported vs objec-
quently used in resistant than in nonresistant hypertensive pa-
tive sleep duration [23], and they reported similar symptoms sug-
tients (Table 3). In particular, 100% of treatment-resistant
gesting OSAS (i.e., apneas, snoring) compared with men [26]. One
patients were treated with diuretics of any kind as well as with re-
of the main criticisms of the role of insomnia in cardiovascular dis-
nin-angiotensin system blockers (Table 3). Resistant hypertensive
ease in our study was the difficulty in unraveling whether or not a
women and men were similarly treated (data not shown).
cause-effect relationship independently existed of common causal
To exclude the possibility that the relationship between RH and
factors [11]. Although this aspect cannot be completely clarified by
poor sleep quality was related to the adverse effects of antihyper-
our study, we showed that the relationship between RH and poor
tensive drugs, we tested each drug class (i.e., diuretics, b blockers,
sleep quality in women was significant using our cross-sectional de-
calcium channel blockers, renin–angiotensin system blockers, a1
sign, even after adjusting for psychiatric and cardiovascular comor-
antagonists) in a logistic regression model considering poor sleep
bidities and sleep duration. This finding highlights an important
quality as the dependent variable. Diuretic use was significantly
issue, given the greater prevalence of trait anxiety and depressive
more prevalent in patients with poor sleep quality than in those
symptoms in women compared to men in the presence of a similar
without (Table 3).
cardiovascular risk pattern. Our results are in agreement with those
After adjustment for a panel of cardiovascular and psychiatric
findings obtained in the National Health and Nutritional Examina-
comorbidities (e.g., age, sex, previous cardiovascular events, DM,
tion Survey (NHANES) cohort which demonstrated that the relation-
obesity, depressive symptoms, trait anxiety, psychiatric drug
ship between depression and hypertension incidence was mostly
use), diuretic use was no longer associated with poor sleep quality
driven by comorbidity with insomnia or short sleep duration [27].
(Table 3).
Possible mechanisms accounting for such a relationship were
beyond the aim of our study. However, evidence is accumulating
4. Discussion on the hypothesis that sleep deprivation might act as a neurobio-
logic stressor, leading to sympathetic nervous system and hypo-
The main finding of our study was the demonstration of an thalamus–pituitary–adrenal axis activation, proinflammatory
association between poor sleep quality and RH in a cohort of pa- responses, and endothelial dysfunction [28,29], thus favoring
tients on their first visit to a Hypertension Outpatient Unit. This hypertension development and maintenance. Experimental sleep
association is gender-specific, as it was more prevalent in women deprivation was shown to increase daytime BP [30] and to amplify
than in men and was independent of sleep duration, trait anxiety, systolic BP increases due to psychologic stress [31], possibly
or depression. Hypertensive women with PSQI >5 had 5.3-fold in- through sympathetic activation [32]. Acute psychologic stress can
creased probability of having RH compared to those with no sleep induce sodium retention and BP increase, mediated by renin–
complaints, even after correction for psychiatric and cardiovascu- angiotensin and sympathetic nervous system activation [33]; thus,
lar comorbidities. The results of our study suggest that mecha- the different sleep loss conditions acting as chronic stressors and
nisms involved in RH development might be gender-specific with therefore favoring allostatic load can conceivably induce and sus-
sleep quality being crucial in women, whereas traditional cardio- tain hypertension [34]. This mechanism could be gender-specific,
vascular risk factors play a major role in men. as postmenopausal women have an increased susceptibility to
Worldwide prevalence of both hypertension and sleep com- salt-induced rise in BP [35]. A number of pathophysiologic mecha-
plaints has increased over the past decades [21,22]. Increasing evi- nisms of hypertension development and maintenance have been
dence suggests that conditions of sleep loss are associated with suggested to show gender differences [36]; among others, muscle
increased prevalence and incidence of hypertension [4]. Our study sympathetic nerve activity had a steeper increase with age in wo-
showed an association between poor sleep quality and hyperten- men than in men [37], and women with elevated BP and metabolic
sion severity for the first time, identified by the definition of resis- syndrome showed a disturbed sympathetic firing pattern, which is
tant hypertension. proportional to trait anxiety and depressive symptoms scores [38].
Interestingly, sleep quality rather than sleep duration was Menopause status, which is associated with dramatic hormonal
found to be associated with resistant hypertension. Although sleep changes and psychosocial stress and with increased cardiovascular
1162 R.M. Bruno et al. / Sleep Medicine 14 (2013) 1157–1163

risk, also has been related to increased incidence of insomnia and Acknowledgments
other sleep disorders; however, discrepant results exist in the liter-
ature [39,40]. Thus sleep disorders might contribute to gender dif- We would like to acknowledge the contribution of Mario Gua-
ferences in the pathophysiology of hypertension, though this field zzelli, who prematurely passed away. We also would like to
is still largely unexplored. acknowledge Carolina Bernardo and Giuseppe Metallo for help in
The results of our study suggest that poor sleep quality may be a data collection.
mechanism involved in resistance to antihypertensive treatment in
women; however, other hypotheses should not be disregarded. We
cannot exclude an opposite cause-effect link. Living with a chronic References
disease may act as a chronic stressor, causing disruption of sleep
[1] Knutson KL. Sleep duration and cardiometabolic risk: a review of the
homeostasis, though this hypothesis is less established. Further- epidemiologic evidence. Best Pract Res Clin Endocrinol Metab
more, resistant hypertensive patients are more heavily treated 2010;24:731–43.
[2] Calhoun DA, Harding SM. Sleep and hypertension. Chest 2010;138:434–43.
with drugs that can interfere with sleep such as diuretics. How-
[3] Somers VK, White DP, Amin R, Abraham WT, Costa F, Culebras A, et al.
ever, our study tended to exclude the latter hypothesis, as none American Heart Association Council for High Blood Pressure Research
of the antihypertensive classes were independently associated Professional Education Committee, Council on Clinical Cardiology; American
with worse sleep quality. Antihypertensive treatment cannot ex- Heart Association Stroke Council; American Heart Association Council on
Cardiovascular Nursing; American College of Cardiology Foundation. Sleep
plain gender differences, as the men and women in our study apnea and cardiovascular disease: an American Heart Association/American
showed similar drug treatments. College of Cardiology Foundation Scientific Statement from the American
The frequent comorbidity of hypertension with OSAS [3] and Heart Association Council for High Blood Pressure Research Professional
Education Committee, Council on Clinical Cardiology, Stroke Council, and
RLS [41] is another possible confounder in the relationship be- Council on Cardiovascular Nursing. In collaboration with the National Heart,
tween RH and sleep quality. In a Brazilian cohort [42] of 125 resis- Lung, and Blood Institute National Center on Sleep Disorders Research
tant hypertensive patients with high prevalence of obesity, a (National Institutes of Health) [published online ahead of print August 25,
2008]. Circulation 2008;118:1080–111.
prevalence of 64% of moderate to severe OSAS was found. Although [4] Palagini L, Bruno RM, Gemignani A, Baglioni C, Ghiadoni L, Riemann D. Sleep
polysomnography was not performed in our study, we excluded all loss and hypertension: a systematic review. Curr Pharm Des
patients with known or suspected OSAS and RLS on the basis of 2013;19:2409–19.
[5] Fernandez-Mendoza J, Vgontzas AN, Liao D, Shaffer ML, Vela-Bueno A, Basta M,
self-reported (or reported by roommate) symptoms to evaluate et al. Insomnia with objective short sleep duration and incident hypertension:
the specific effect of poor sleep quality, independent of sleep-disor- the Penn State Cohort. Hypertension 2012;60:929–35.
dered breathing. It has been reported that snoring has high sensi- [6] 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
tivity, despite low specificity (87% and 57%, respectively), in
national health and nutrition examination survey. Hypertension
detecting OSAS in RH [42]; thus, the use of this selection criteria, 2006;47:833–9.
though possibly inadequate as screening in the hypertensive pop- [7] Cappuccio FP, Cooper D, D’Elia L, Strazzullo P, Miller MA. Sleep duration
ulation [43], makes the occurrence of type I error in our study predicts cardiovascular outcomes: a systematic review and meta-analysis of
prospective studies. Eur Heart J 2011;32:1484–92.
improbable. Our data also are supported by a number of studies [8] Vgontzas AN, Liao D, Bixler EO, Chrousos GP, Vela-Bueno A. Insomnia with
indicating that sleep quality is unrelated to apnea–hypopnea index objective short sleep duration is associated with a high risk for hypertension.
either in the general population [8,25] or in individuals with sleep- Sleep 2009;32:491–7.
[9] Parish JM. Sleep-related problems in common medical conditions. Chest
disordered breathing [44], and that the BP-raising effect of insom- 2009;135:563–72.
nia with short sleep duration is independent of apnea–hypopnea [10] Riemann D. Insomnia and comorbid psychiatric disorders. Sleep Med
index in the general population [5,8]. 2007;8(Suppl. 4):S15–20.
[11] Budhiraja R, Roth T, Hudgel DW, Budhiraja P, Drake CL. Prevalence and
Among limitations, we should consider the cross-sectional de- polysomnographic correlates of insomnia comorbid with medical disorders.
sign of our study and the use of self-reported evaluation of sleep, Sleep 2011;34:859–67.
as previously discussed. Furthermore, increasing evidence suggests [12] Cappuccio FP, Stranges S, Kandala NB, Miller MA, Taggart FM, Kumari
M, et al. Gender-specific associations of short sleep duration with
that diagnosis of true RH requires 24-h BP monitoring [45,46], prevalent and incident hypertension: the Whitehall II Study.
which was not included in our protocol. Another possible limita- Hypertension 2007;50:693–700 [published online ahead of print
tion is the small sample size of the RH group; however, RH preva- September 4, 2007].
[13] Calhoun DA, Jones D, Textor S, Goff DC, Murphy TP, Toto RD. American Heart
lence is in line with previous reports or may be even greater [47],
Association Professional Education Committee. Resistant hypertension:
possibly due to the fact that a Hypertension Unit only receives se- diagnosis, evaluation, and treatment: a scientific statement from the
vere cases from the general population. American Heart Association Professional Education Committee of the
In conclusion, our study showed that poor sleep quality was Council for High Blood Pressure Research. Circulation 2008;117:e510–26.
[14] Mancia G, De Backer G, Dominiczak A, Cifkova R, Fagard R, Germano G.
highly prevalent in resistant hypertensive patients and was associ- Management of Arterial Hypertension of the European Society of
ated with RH in women but not in men, even independent of sleep Hypertension; European Society of Cardiology. 2007 Guidelines for the
duration and cardiovascular and psychiatric comorbidities. Sleep Management of Arterial Hypertension: the Task Force for the Management
of Arterial Hypertension of the European Society of Hypertension (ESH) and
loss has been recently indicated as a new modifiable cardiovascu- of the European Society of Cardiology (ESC). J Hypertens 2007;2007:
lar risk factor [48]. Although our results need to be confirmed by 1105–87.
prospective studies demonstrating the cause-effect relationship [15] National Cholesterol Education Program (NCEP) expert panel on detection,
evaluation, and treatment of high blood cholesterol in adults (Adult Treatment
between poor sleep quality and RH and interventional studies Panel III). Third report of the National Cholesterol Education Program (NCEP)
investigating if treatment of insomnia might improve BP control, expert panel on detection, evaluation, and treatment of high blood cholesterol
they do suggest that sleep quality in treatment-resistant women in adults (Adult Treatment Panel III) final report. Circulation
2002;106:3143–421.
could be clinically relevant not only for quality of life but also for [16] Executive summary: standards of medical care in diabetes-2012. Diabetes
cardiovascular health. Care 2012;35:S4–S10.
[17] Buysse DJ, Reynolds 3rd CF, Monk TH, Berman SR, Kupfer DJ. The pittsburgh
sleep quality index: a new instrument for psychiatric practice and research.
Psychiatry Res 1989;28:193–213.
Conflict of interest [18] Richter P, Werner J, Heerlein A, Kraus A, Sauer H. On the validity of the beck
depression inventory. a review. Psychopathology 1998;31:160–8.
The ICMJE Uniform Disclosure Form for Potential Conflicts of [19] Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for
measuring depression. Arch Gen Psychiatry 1961;4:561–71.
Interest associated with this article can be viewed by clicking on [20] Spielberger C. State-trait anger expression inventory: professional
the following link: http://dx.doi.org/10.1016/j.sleep.2013.04.020. manual. Odessa: Psychological Assessment Resources; 1988.
 R.M. Bruno et al. / Sleep Medicine 14 (2013) 1157–1163 1163

[21] Rowshan Ravan A, Bengtsson C, Lissner L, Lapidus L, Bjorkelund C. Thirty-six- [35] Pechere-Bertschi A, Burnier M. Female sex hormones, salt, and blood pressure
year secular trends in sleep duration and sleep satisfaction, and associations regulation. Am J Hypertens 2004;17:994–1001.
with mental stress and socioeconomic factors-results of the population study [36] Fu Q, Vongpatanasin W, Levine BD. Neural and nonneural mechanisms for sex
of women in Gothenburg, Sweden. J Sleep Res 2010;19:496–503. differences in elderly hypertension: can exercise training help? Hypertension
[22] Kearney PM, Whelton M, Reynolds K, Muntner P, Whelton PK, He J. Global 2008;52:787–94.
burden of hypertension: analysis of worldwide data. Lancet 2005;365:217–23. [37] Matsukawa T, Sugiyama Y, Watanabe T, Kobayashi F, Mano T. Gender
[23] Lauderdale DS, Knutson KL, Yan LL, Liu K, Rathouz PJ. Self-reported and difference in age-related changes in muscle sympathetic nerve activity in
measured sleep duration: how similar are they? Epidemiology healthy subjects. Am J Physiol 1998;275:R1600–4.
2008;19:838–45. [38] Lambert E, Dawood T, Straznicky N, Sari C, Schlaich M, Esler M, et al.
[24] Dean E, Bloom A, Cirillo M, Hong Q, Jawl B, Jukes J, et al. Association between Association between the sympathetic firing pattern and anxiety level in
habitual sleep duration and blood pressure and clinical implications: patients with the metabolic syndrome and elevated blood pressure. J
asystematic review. Blood press 2012;21:45–57. Hypertens 2010;28:543–50.
[25] Buysse DJ, Hall ML, Strollo PJ, Kamarck TW, Owens J, Lee L, et al. Relationships [39] Krishnan V, Collop NA. Gender differences in sleep disorders. Curr Opin Pulm
between the Pittsburgh sleep quality index (psqi), epworth sleepiness scale Med 2006;12:383–9.
(ess), and clinical/polysomnographic measures in a community sample. J Clin [40] Young T, Rabago D, Zgierska A, Austin D, Laurel F. Objective and subjective
Sleep Med 2008;4:563–71. sleep quality in premenopausal, perimenopausal, and postmenopausal women
[26] Young T, Hutton R, Finn L, Badr S, Palta M. The gender bias in sleep apnea in the Wisconsin sleep cohort study. Sleep 2003;26:667–72.
diagnosis. Are women missed because they have different symptoms? Arch [41] Batool-Anwar S, Malhotra A, Forman J, Winkelman J, Li Y, Gao X. Restless legs
Intern Med 1996;156:2445–51. syndrome and hypertension in middle-aged women. Hypertension
[27] Gangwisch JE, Malaspina D, Posner K, Babiss LA, Heymsfield SB, Turner JB, et al. 2011;58:791–6.
Insomnia and sleep duration as mediators of the relationship between [42] Pedrosa RP, Drager LF, Gonzaga CC, Sousa MG, de Paula LK, Amaro AC, et al.
depression and hypertension incidence. Am J Hypertens 2010;23:62–9. Obstructive sleep apnea: the most common secondary cause of
[28] Bonnet MH, Arand DL. Hyperarousal and insomnia: state of the science. Sleep hypertension associated with resistant hypertension. Hypertension 2011;
Med Rev 2010;14:9–15. 58:811–7.
[29] Meerlo P, Sgoifo A, Suchecki D. Restricted and disrupted sleep: effects on [43] Drager LF, Genta PR, Pedrosa RP, Nerbass FB, Gonzaga CC, Krieger EM, et al.
autonomic function, neuroendocrine stress systems and stress responsivity. Characteristics and predictors of obstructive sleep apnea in patients with
Sleep Med Rev 2008;12:197–210 [published online ahead of print January 25, systemic hypertension. Am J Cardiol 2010;105:1135–9.
2008]. [44] Macey PM, Woo MA, Kumar R, Cross RL, Harper RM. Relationship between
[30] Zhong X, Hilton HJ, Gates GJ, Jelic S, Stern Y, Bartels MN, et al. Increased obstructive sleep apnea severity and sleep, depression and anxiety symptoms
sympathetic and decreased parasympathetic cardiovascular modulation in in newly-diagnosed patients. PLoS One 2010;5:e10211.
normal humans with acute sleep deprivation. J Appl Physiol 2005;98:2024–32. [45] de la Sierra A, Segura J, Banegas JR, Gorostidi M, de la Cruz JJ, Armario P, et al.
[31] Franzen PL, Gianaros PJ, Marsland AL, Hall MH, Siegle GJ, Dahl RE, et al. Clinical features of 8295 patients with resistant hypertension classified on the
Cardiovascular reactivity to acute psychological stress following sleep basis of ambulatory blood pressure monitoring. Hypertension
deprivation. Psychosom Med 2011;73:679–82. 2011;57:898–902.
[32] Ogawa Y, Kanbayashi T, Saito Y, Takahashi Y, Kitajima T, Takahashi K, et al. [46] Pierdomenico SD, Lapenna D, Bucci A, Di Tommaso R, Di Mascio R, Manente
Total sleep deprivation elevates blood pressure through arterial baroreflex BM, et al. Cardiovascular outcome in treated hypertensive patients with
resetting: a study with microneurographic technique. Sleep 2003;26:986–9. responder, masked, false resistant, and true resistant hypertension. Am J
[33] Light KC, Koepke JP, Obrist PA, Willis PW. Psychological stress induces sodium Hypertens 2005;18:1422–8.
and fluid retention in men at high risk for hypertension. Science [47] Persell SD. Prevalence of resistant hypertension in the United States, 2003–
1983;220:429–31. 2008. Hypertension 2011;57:1076–80.
[34] Harshfield GA, Dong Y, Kapuku GK, Zhu H, Hanevold CD. Stress-induced [48] Redline S, Foody J. Sleep disturbances: time to join the top 10
sodium retention and hypertension: a review and hypothesis. Curr Hypertens potentially modifiable cardiovascular risk factors? Circulation 2011;124:
Rep 2009;11:29–34. 2049–51.