Review

The impact of physical activity on mortality in

patients with high blood pressure: a systematic review
Amanda Rossi a,b,c, Anastasia Dikareva d, Simon L. Bacon a,b,c,e, and Stella S. Daskalopoulou d

See editorial comment on page 1310

Background: Physical activity has been shown to be

INTRODUCTION

H
beneficial for the prevention and management of ypertension represents the highest proportion
hypertension. In the general population, physical activity of attributable mortality amongst all global risk
has been shown to decrease mortality. factors and is a large burden to healthcare
Purpose: The purpose of this systematic review was to systems worldwide [1,2]. There exists a strong, direct
identify and synthesize the literature examining the impact relationship between blood pressure and risk of stroke
of physical activity on mortality in patients with high blood mortality, ischemic heart disease mortality, and all-cause
pressure (BP). mortality [3].
Current guidelines recommend regular physical activity
Methods: An extensive search was conducted by two
as a preventive measure and a first-line nonpharmacolog-
independent authors using Medline, Embase and Cochrane
ical treatment for hypertension [4–6]. Habitual leisure time
Library electronic databases (between 1985 and January
physical activity has been shown to reduce all-cause
2012) and manual search from the reference list of
mortality in both men and women [7]. A study of Harvard
relevant articles. Inclusion criteria were as follows:
alumni showed that those who engaged in regular physical
longitudinal design with minimum 1-year follow-up;
activity lived over a year longer than their sedentary
hypertensive status of the cohort was indicated; and BP,
counterparts [8]. Furthermore both leisure time physical
physical activity, and mortality were measured.
activity and occupational activity have shown similar results
Results: Six articles evaluating a combined total of 48 448 with respect to reducing risk of death from ischemic heart
men and 47 625 women satisfied the inclusion criteria. disease [9]. A review of 44 studies concluded that the
Cardiovascular and/or all-cause mortality were shown to be volume of physical activity and all-cause mortality are
inversely related to physical activity in all studies. For related in an inverse, linear dose-dependent manner [10].
example, patients with high BP who participated in any Researchers have also shown that cardiorespiratory fitness,
level of physical activity had a reduced risk (by 16–67%) measured by maximal exercise stress testing, is related to
of cardiovascular mortality, whereas a greater than two- mortality [11]. Of note, Blair et al. [12] have shown that
fold increase in risk of mortality was noted in nonactive this is consistent for both normotensive and hypertensive
individuals. However, activity classification and parameters, men, in that men with higher cardiorespiratory fitness have
such as frequency, duration, intensity, and volume, as well a decreased risk of mortality. Moreover, participation in
as BP status, were not consistent across studies. aerobic [13,14] or resistance [15,16] exercise can lead to
Conclusions: Regular physical activity is beneficial for modest reductions in blood pressure.
reducing mortality in patients with high BP. More research Despite the available literature to support the benefits
is needed to establish the impact of specific kinds of of physical activity on blood pressure and mortality in
physical activity and whether any differences exist between the general population, it is not clear whether these
sexes.
Keywords: blood pressure, hypertension, mortality,
physical activity, systematic review Journal of Hypertension 2012, 30:1277–1288
a
Department of Exercise Science, Concordia University, bMontreal Behavioural Medi-
Abbreviations: BP, blood pressure; CVD, cardiovascular cine Centre, cResearch Centre, Montreal Heart Institute – a University of Montreal
disease; D&B, Downs and Black; HPA, high physical Affiliated Hospital, dDepartment of Medicine, McGill University and eResearch Centre,
Hôpital du Sacré-Cœur de Montréal – a University of Montreal Affiliated Hospital,
activity; ICD, International Classification of Diseases; IL-6, Montréal, Quebec, Canada
interleukin-6; LIFE, Losartan Intervention for Endpoint; LPA, Correspondence to Stella S. Daskalopoulou MD, MSc, DIC, PhD, FRSQ, Chercheur-
low physical activity; Meds, medications; MONICA, Boursier Clinicien, Assistant Professor in Medicine, Department of Medicine, Division
Multinational Monitoring of trends and determinants in of Internal Medicine, McGill University, McGill University Health Centre, Montreal
General Hospital, 1650 Cedar Avenue, B2.101.4, Montreal, Quebec, Canada H3G
Cardiovascular disease; MPA, moderate physical activity; 1A4. Tel: +1 514 934 1934 x 42295; fax: +1 514 934 8564; e-mail: stella.
NHANES, National Health and Nutrition Examination [email protected]
Survey; PRISMA, Preferred Reporting Items for Systematic Received 4 October 2011 Revised 21 March 2012 Accepted 30 March 2012
Reviews and Meta-Analyses; RR, relative risk; TNF-a, tumor J Hypertens 30:1277–1288 ß 2012 Wolters Kluwer Health | Lippincott Williams &
necrosis factor-a Wilkins.
DOI:10.1097/HJH.0b013e3283544669

Journal of Hypertension www.jhypertension.com 1277

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Rossi et al.

benefits translate to decreases in cardiovascular or Retrieved records were retained if they fit all of
all-cause mortality specifically in patients with high blood the following criteria: longitudinal design with a minimum
pressure. Therefore, the purpose of this systematic review 1-year follow-up; adult participants (>18 years of age) had
is to present the results of prospective longitudinal studies high blood pressure or hypertensive status was indicated;
exploring the effect of physical activity on mortality and blood pressure, physical activity and cardiovascular or
(cardiovascular and all-cause) in patients with high blood all-cause mortality were measured.
pressure. Risk of bias was evaluated in the selected studies using a
modified version of the Downs and Black [18] tool so that
METHODS only questions pertinent to prospective cohort studies were
retained. This same method has been used previously
The present systematic review was conducted in accord- [19,20]. Thus 15 of the original 27 items (reporting: 1–4,
ance with the Preferred Reporting Items for Systematic 6, 7, 9, 10; external validity: 11–13; internal validity: 16–18,
Reviews and Meta-Analyses (PRISMA) guidelines [17]. and 20) were considered for a possible total score of 15, in
The literature search was conducted using the Medline, which a higher score indicates better quality publication.
Embase and Cochrane Library electronic databases and Additionally, funnel plots were used to evaluate publication
manual search from the reference list of relevant articles. bias. In cases when group sample sizes were not detailed in
Records were identified using standardized search terms. the original article, efforts were made to contact the authors
The Medline search strategy, as seen below, was adapted by telephone and E-mail; however, identifying current
according to the respective indexing systems for the contact information was not always possible.
Embase and Cochrane Library databases. No previously A meta-analysis would have allowed us to quantify the
established review protocol exists for this theme. English overall effect of physical activity on mortality in this popu-
language longitudinal studies collecting data from human lation. However, there was a substantial lack of consistency
samples, published between the beginning of January in the reporting of physical activity, whereby each study
1985 and the end of January 2012, were considered, was classified in a different manner according to varying
without any other limitations. The search and screening criteria in the self-report questionnaires, which made
phases were conducted independently by two authors formal statistical analysis impractical. For instance, some
(A.R. and A.D.) with the help of two medical librarians, studies classified physical activity groups according to the
one from McGill University (A.L.) and the other one from number of steps or city blocks walked each day, whereas
Concordia University (D.K.). Any discrepancies were other studies used minutes per day, metabolic equivalent
resolved through consensus. All the authors participated scales or kilocalories per day to categorize the participants.
in the final selection of the included studies. Data were
extracted by one author (A.R.) using an electronic form RESULTS
and checked for accuracy (A.D.). All authors have
reviewed the extracted data. Variables of interest A total of 3217 records were retrieved (see Fig. 1). Of the 26
included: study and participant characteristics (e.g. length full-text articles [7,8,11,12,21–42] evaluated for eligibility,
of follow-up, age, etc.), blood pressure and physical 20 were eliminated for the following reasons; in one article
activity measurement tools and classification schemes, [24] the analysis was based on a sub-sample of a larger trial
method of mortality and cause of death verification, [32] included in the systematic review, two articles [12,35]
cardiovascular and all-cause mortality hazard ratios as did not measure leisure time physical activity (only fitness
well as study-specific covariates. or work activity were evaluated), and the remaining 17
Medline search strategy: studies [7,8,11,21–23,25–29,33,34,36,38,40,41] reported
collecting data relating to blood pressure, physical activity,
1. Hypertension/or hypertens
.mp. or mortality, but they did not evaluate the relationship
2. blood pressure.mp. or Blood Pressure/ between the three variables. Usually, physical activity
3. Normotens
.mp. and blood pressure were considered covariates in these
4. Arterial pressure.mp. reports. Thus, six studies were identified. Table 1 describes
5. 1 or 2 or 3 or 4 the characteristics of these studies. Altogether these studies
6. Exercise/ or exercise.mp. evaluated 48 448 men and 47 625 women for a total of
7. physical active
.mp. 96 073 adults. Of the six studies, two considered only male
8. physical
active.mp. participants [30,39] and the remaining four included both
9. Motor Activity/ men and women [31,32,37,42]. Only three studies [32,37,42]
10. resistance training.mp. or Resistance Training/ reported results for men and women separately. The
11. exercise
.mp. cohorts originated from Northern Europe (Denmark,
12. 6 or 7 or 8 or 9 or 10 or 11 Sweden, Iceland, Norway, Finland and the UK) or USA.
13. Mortality/ Medication usage was only indicated for the Losartan
14. Death/ Intervention for Endpoint (LIFE) trial [32]. None of the other
15. Fatal Outcome/ studies reported the type of medications, apart from stating
16. 13 or 14 or 15 that participants using blood pressure-lowering drugs were
17. 5 and 12 and 16 included and classified as hypertensive. Vatten et al. [42]
18. Limit 17 to (English language and humans and stated that participants with specific comorbidities were
yr ¼ ‘1985-Current’) excluded, including patients using antihypertensive

1278 www.jhypertension.com Volume 30  Number 7  July 2012

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 Blood pressure, physical activity and mortality

No. of articles identified through database

searches No. of additional articles identified
Medline: 106 through manual search:
Embase: 2814 1
Cochrane clinical trials: 296

Total no. of records: 3217

No. of records screened: No. of duplicates removed:

3001 216

No. of articles excluded:

2975

No. of full text articles No. of articles excluded:

assessed for eligibility: Sub-analysis of larger study already
26 included in review: 1
Did not mesaure physical activity: 2
Did not evaluate relationship between
physical activity, blood pressure and
mortality: 17

No. of studies included in review: 6

FIGURE 1 Literature search results.

medications. Exclusions were also specified for the LIFE cohort. In contrast, Vatten et al. [42] excluded individuals
trial recruitment [32]. By design, the LIFE cohort also had who reported using blood pressure-lowering medications
left-ventricular hypertrophy [32]. The National Health prior to entering the study. The authors established
and Nutrition Examination Survey I (NHANES I) report by four categories for SBP (<120 mmHg, 120–139 mmHg,
Fang et al. [31] and LIFE trial, reported by Fossum et al. [32], 140–159 mmHg, >160 mmHg) and DBP (<80 mmHg,
were the only reports to include alcohol consumption and 80–89 mmHg, 90–99 mmHg, 100 mmHg) classification
race/ethnicity; additionally NHANES I considered diet and spanning normotensive and hypertensive values.
socioeconomic measures in their model for analysis. Paffenbarger et al. [39] stated that all participants were
Amongst the articles selected are several sub-analyses hypertensive; however, they did not describe what blood
of larger trials [31,32,37,39]. In cases when information pressure threshold level or criteria were used to define high
regarding the methods of blood pressure measurement, blood pressure.
physical activity assessment, or mortality was not available
in the text, the reference list or original publications were Measurement
consulted. Details for the measurement of blood pressure can be found
in Table 2. Three studies [30,31,42] reported measuring
blood pressure with a manual sphygmomanometer accord-
Blood pressure ing to a defined protocol [43]. Engström et al. [30], Fossum
et al. [32,44,45] and Fang et al. [31] specifically reported the
Classification patients to be in a seated position; however, Paffenbarger
According to the design of this review, each of the selected et al. [39] and Vatten et al. [42] did not describe the posture
publications evaluated patients with high blood pressure; of the participants. The LIFE trial reports indicate a stand-
however, the criteria used to diagnose hypertension varied ardized protocol was used to measure blood pressure
between studies. Engström et al. [30] used cut-off values of [44,45]. Hu et al. [37,46] described all but one of their
at least 160 mmHg or at least 95 mmHg for SBP and/or DBP, multiple sites to have measured blood pressure in the
respectively, or self-reported use of antihypertensive medi- seated position; this single site evaluated patients in a
cation. Fang et al. [31] and Hu et al. [37] established their recumbent position [46]. Although the method of blood
own respective classification schemes (see Table 1). pressure measurement was consistent within each partic-
Fossum et al. [32] selected participants based on their blood ipating site, the WHO Multinational Monitoring of trends
pressure following 2 weeks of placebo treatment. If SBP and determinants in Cardiovascular disease (MONICA)
ranged between 160–200 mmHg and/or 95–115 mmHg blood pressure assessment document [46] explained
they were classified as hypertensive and included in the that both random-zero sphygmomanometers and simple

Journal of Hypertension www.jhypertension.com 1279

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Rossi et al.

sphygmomanometers were used. Details regarding

score
D&B
measurement of blood pressure in the University of
11

12

10

13

13
8
Pennsylvania College Alumni cohort [39] were not
available.
Medsb

Yes
No

No

No

No

No
Physical activity assessment
Left ventricular
hypertrophy
Classification
conditionsa

participants
No exclusions

No exclusions

No exclusions
Comorbid

indicated

indicated

indicated
At baseline, the LIFE cohort [32] classified participants as

Excluded
No CVD

sedentary (never active), intermediate (30 min of activity/

week), or active (>30 min/week) based on responses to
their questionnaire. Engström et al. [30] conducted struc-
tured interviews and following an initial classification
22.8–27.04
24.5–24.5

27.4–29.1

26.1–28.6

(almost completely inactive, some, regular and regular hard

(kg/m2)

activity) they collapsed the groups into nonvigorous (i.e.

BMI

inactive and some activity) and vigorous (i.e. regular and

–

regular hard activity) groups. Fang et al. [31] asked their

participants two questions regarding physical activity: ‘Do
age (years)

you get much exercise in things you do for recreation, or

Baseline

hardly any exercise, or in between?’ and ‘In your usual day,

38–55

66–67

41–51

 20

aside from recreation, how active are you?’ The MONICA

55

cohort [37] qualified their occupational and leisure time

physical activity as low, moderate, or high based on
12244/14399

24409/24584

descriptors given for each respective type of physical

6011/3780

4961/4224
(female/

activity. Additionally, commuting physical activity was

male)
0/642

0/819
Sex

classified as motorized, walk/cycle less than 30 min or

walk/cycle at least 30 min. Based on information collected
about the number of city blocks walked per day, number of
26643

48993
9791

9185

stairs climbed daily and the type and frequency of sport or

642

819
N

recreational activities, Paffenbarger et al. [39] calculated a

physical activity index which estimated the amount of
11.8–17 (1971/1975–

kilocalories expended per week. However, for this analysis

23.8 (1968/69-death/

19.9 (1972–2003)

16 (1984/86-death/
December 1993)

December 2002)

they only used a classification scheme according to level

6.6–31.7; mean:

24 (1962–1985)

of sport participation. Vatten et al. [42] collected subjec-

June 1992)
Follow-up

tive information on the frequency of exercise, average

BMI, body mass index; CVD, cardiovascular disease; D&B, Downs and Black [18]; Meds, medications.
(years)

duration of each session, activity intensity and based on

responses graded participants into the following hier-
4.8

archy: no activity, low, medium and high. Engström

et al. [30] and Fang et al. [31] used descriptive categories,
College Alumni Health Study,
University of Pennsylvania

whereas the remaining studies attempted to quantify

Losartan Intervention for

physical activity by time [32,37,42] or by type and fre-

Nutrition Examination

Nord Trøndelag Health

Survey (NHANES I)

quency of exercise [39,42].

National Health and

Indicates whether or not the participants’ medication use was reported.

Endpoint (LIFE)

Indicates if patients with comorbid diseases were included in the study.

Study (HUNT)

Measurement
Cohort II

Information regarding physical activity was obtained

Study
TABLE 1. Study and sample characteristics

through self-report in all studies. Engström et al. [30] con-

–

ducted interviews with their participants, whereas all other

studies reported using questionnaires [31,32,37,39,42].
United Kingdom,
Iceland, Norway,

General and leisure time physical activity were the main

Denmark, Finland,

types of activity considered; however, Hu et al. [37] also

collected information specifically relating to occupational
Sweden,
Country
Sweden

Norway

and commuting physical activity (these results are not

Finland
USA
USA

USA

presented here). Each study established a classification

scheme for activity levels (see Table 2).
Fang (2005) [31]

Hu (2007) [37]
First author

Fossum (2007)

Follow-up and mortality

Vatten (2006)
(1999) [30]

(1991) [39]
Paffenbarger

The length of follow-up ranged from almost 5 years up

Engström
(year)

to 24 years. Cardiovascular mortality [30–32,37,42] and

[32]

[42]

all-cause mortality [30–32,39] were evaluated in the

selected studies. Most studies obtained a confirmation of
b
a

1280 www.jhypertension.com Volume 30  Number 7  July 2012

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 Journal of Hypertension
TABLE 2. Summary of methods and classification schemes for selected studies
Blood pressure Physical activity
First author Mortality and cause
(year) Classification Measurement Classification Measurement of death
Engström Hypertension SBP  160 mmHg or Morning, seated position, measured Non-vigorous: Inactive (group 1) þ Structured interview Mortality Register, Swedish
(1999) [30] DBP  95 mmHg or reported to nearest 5 mmHg with mercury some activity (group 2); vigorous: National Bureau of Statistics;
using antihypertensive medication sphygmomanometer and regularly active (group 3) þ regular ICD, 8th and 9th revision
12 x 16 rubber cuff hard activity (group 4)
Fang (2005) [31] Normal BP: no history of Seated position, with weekly calibrated Recreational activity: 1. low activity; Self-report questionnaire; Death certificate or proxy
hypertension and BP manometer and falling pressure 2. moderate activity; only considered respondent; ICD, 9th
<120/80 mmHg; prehypertension: at 2–3 mmHg, measured to nearest 3. high activity ‘recreational activity’ revision
no history of hypertension and 2 mmHg [43]
BP 120–139/80–89 mmHg;
hypertension: history of
hypertension, reported using
antihypertensive medication,
BP 140/90 mmHg
Fossum (2007) [32] Hypertension; SBP 160–200 mmHg Seated position following a standardized Sedentary: never active; intermediate: Self-report questionnaire Deaths were reported
and/or DBP 95–115 mmHg after protocol [44,45] 30min/week; Active: >30 min/week separately and directly to the
2 weeks of placebo treatment independent data and safety
monitoring board for
validation [47]
Hu (2007) [37] Hypertension: SBP 140 or Varied across sites (mainly seated position), Low: almost completely inactive; Self-report questionnaire; Statistics Finland; ICD, 8th,
DBP 90, using or approved simple or random-zero moderate: some physical activity leisure time physical 9th, and 10th revision
reimbursement for antihypertensive sphygmomanometer [46] >4 h/week; high: vigorous activity activity
medication; Moderate or severe >3 h/week
hypertension: SBP 160 or
DBP 95
Paffenbarger – – Sport participation: none, light only, Self-report questionnaire; –
(1991) [39] light and vigorous, vigorous only only sport participation
Vatten (2006) [42] Blood pressure groups: SBP: Calibrated mercury manometers, 1. no activity; and three equal Self-report questionnaire Cause of Death Registry,
<120, 120–139, 140–159, standard cuff size, measured to the activity groups; 2. low, 3. medium, Norway
160 mmHg; DBP: <80, 80–89, nearest 2 mmHg and 4. high
90–99, 100 mmHg

ICD, International Classification of Diseases.

www.jhypertension.com
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
1281
Blood pressure, physical activity and mortality
Rossi et al.

death through their respective national registries or official The results from each study (all based on multivariate
documentation [30,31,37,42,47]. No information was pro- analyses) as well as the variables included in the respective
vided regarding how mortality data was acquired for the statistical models are illustrated in Table 3. In hypertensive
study by Paffenbarger et al. [39]. patients who engaged in vigorous physical activity (i.e.

TABLE 3. Summary of cardiovascular and all-cause mortality results for selected studies
First author (year) Cardiovascular mortality All-cause mortality Multivariate model

Engström [30] (1999) Relative risk (95% CI): hypertensive/ Relative risk (95% CI): Normotensive: smoking; hypertensive:
vigorous physical activity: hypertensive/vigorous physical smoking, antihypertensive therapy
0.33 (0.11–0.94); hypertensive/ activity: 0.43 (0.22–0.82); and SBP
nonvigorous physical activity: 1.00; hypertensive/nonvigorous physical
normotensive/vigorous physical activity: 1.00; normotensive/
activity: 0.72 (0.39–1.35); vigorous physical activity:
normotensive/nonvigorous physical 0.89 (0.60–1.31); normotensive/
activity: 1.00 nonvigorous physical activity: 1.00
Fang [31] (2005) Hazard ratios (95% CI): Normotensive: Hazard ratios (95% CI): normotensive: Age, sex, race, BMI, education, diabetes,
LPA 1, MPA 0.76 (0.39–1.49), LPA 1, MPA 0.75(0.53–1.05), smoking, alcohol, dietary caloric,
HPA 0.65 (0.24–1.77); prehypertensive: HPA 0.71 (0.45–1.12); sodium, calcium and potassium intake,
LPA 1, MPA 0.79 (0.58–1.09), prehypertensive: LPA 1, MPA 0.79 SBP and serum cholesterol
HPA 0.89 (0.61–1.31); hypertensive: (0.65–0.97), HPA 0.93 (0.74–1.18);
LPA 1, MPA 0.84 (0.73–0.97), hypertensive: LPA 1, MPA 0.88
0.80 (0.66–0.96) (0.80–0.98), HPA 0.83 (0.72–0.95)
Fossum [32] (2007) Hazard ratios (95% CI) Hazard ratios (95% CI) Baseline current smoking, alcohol, sex,
age, race, left-ventricular hypertrophy,
Framingham risk scorey
Sedentary: reference; intermediate: Sedentary: reference; intermediate:
0.80 (0.63–1.01); active: 0.49 0.85 (0.71–1.02); active:
(0.39–0.62) 0.65(0.55–0.77)
Men Men
Sedentary: reference; intermediate: Sedentary: reference; intermediate:
0.65 (0.47–0.90); active: 0.45 0.77 (0.60–1.00); active:
(0.33–0.61) 0.60 (0.48–0.76)
Women Women
Sedentary: reference; intermediate: Sedentary: reference; intermediate:
1.029 (0.73–1.44); active: 0.55 0.95 (0.74–1.24); active:
(0.38–0.79) 0.72 (0.56–0.92)
Hu [37] (2007) Hazard ratios (95% CI) – men: Low 1, – Age, study year, education, alcohol,
Mod 0.84 (0.77–0.91), High 0.73 smoking, BMI, SBP, cholesterol,
(0.62–0.86); trend P < 0.001; antihypertensive drug use and
women: Low 1, Mod 0.78 (0.70–0.87), diabetes
High 0.74 (0.58–0.94); trend P < 0.001
Paffenbarger [39] (1991) – Relative risk: none 1.00, Adjusted for age
light only 1.00, light and
vigorous 0.73, vigorous only
0.63, trend P ¼ 0.1276
y
Vatten [42] (2006) Relative risk (95% CI): high, medium, – Age, BMI, marital status, education,
low, no activity alcohol and smoking
Men
<120 mmHg: 0.68 (0.43–1.07),
0.99 (0.70–1.39), 0.78 (0.51–1.20),
1.15 (0.72–1.85)
120–139 mmHg: 1.00 (Reference),
1.06 (0.86–1.32), 0.99 (0.78–1.26),
1.31 (1.02–1.67)
140–159 mmHg: 1.21 (0.97–1.52),
1.25 (1.02–1.55), 1.39 (1.11–1.74),
1.73 (1.37–2.19)
>160 mmHg: 1.82 (1.46–2.28),
1.76 (1.42–2.17), 1.84 (1.45–2.34),
2.24 (1.78–2.83)
Women
<120 mmHg: 0.52 (0.28–0.97),
1.00 (0.61–1.65), 1.08 (0.62–1.86),
1.43 (1.84–2.44)
120–139 mmHg: 1.00 (Reference),
1.12 (0.80–1.57), 1.18 (0.81–1.73),
1.79 (1.26–2.53)
140–159 mmHg: 1.47 (1.04–2.09),
1.54 (1.12–2.12), 1.66 (1.17–2.34),
1.93 (1.39–2.69)
>160 mmHg: 1.77 (1.26–2.54),
2.49 (1.84–3.37), 2.60 (1.87–3.60),
2.41 (1.76–3.30)

Data presented herein are results from multivariate analyses for all studies; unadjusted results are not shown in this table. +Only results for SBP are presented here; a similar pattern for
DBP was observed [42]. HPA, high physical activity; LPA, low physical activity; MPA, moderate physical activity.

1282 www.jhypertension.com Volume 30  Number 7  July 2012

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 Blood pressure, physical activity and mortality

regularly active þ regular hard activity), Engström et al. [30] 1.78, 2.83; women: RR 2.41, 95% CI 1.76, 3.30) of cardio-
found a significant lower risk of all-cause mortality [relative vascular mortality compared to very active participants with
risk (RR) 0.43, 95% confidence interval (CI) 0.22, 0.82] and lower blood pressure. Thus, all six studies have shown an
cardiovascular mortality (RR 0.33, 95% CI 0.11, 0.94) when inverse relationship between physical activity and cardio-
compared with hypertensive patients who did not engage vascular or all-cause mortality.
in vigorous physical activity. The same authors found that
level of physical activity did not make a difference in Risk of bias assessment and publication bias
mortality amongst normotensive participants [30]. Fang The results of this evaluation can be found in Table 1. Final
et al. [31] showed that patients with prehypertension scores ranged between 8 and 13 [mean  standard devi-
who were active for less than 30 min/day (hazard ation (SD) 11.4  1.9, median 12). Four studies received
ratio ¼ 0.79, 95% CI 0.65, 0.97), but not those with higher high scores (12/15). Overall the studies rated well in the
levels of physical activity (hazard ratio ¼ 0.93, 95% CI 0.74, reporting category, with an average of seven of eight
1.18), had a decreased risk of all-cause mortality. In hyper- questions receiving full points. The studies received poor
tensive patients, active individuals had a 14–20% lower scores for external validity (average 1 point out of 3).
risk of cardiovascular death and similarly 12–17% lower Information regarding the representativeness of the sample
all-cause mortality risk compared to their least active was generally unavailable. Additionally, whether or not the
counterparts. type of care provided was typical for the patients was not
Overall, the active group from the LIFE [32] sample had a addressed. Three of the four questions assessing internal
significant decrease in cardiovascular mortality compared validity were given full points for each article. When appro-
to the sedentary group. A nonsignificant decrease of 20% priate, most studies did indicate if analysis was adjusted for
was noted for those who participated in 30 min or less of the length of follow-up. There was no discernable differ-
activity/week. Compared to the sedentary groups, active ence in reported outcome between the high and low-
men (hazard ratio ¼ 0.45, 95% CI 0.33–0.61) and women scoring studies. Figure 2 is a funnel plot of sample size
(hazard ratio ¼ 0.55, 95% CI 0.38–0.79) had a reduced risk and log hazard ratio for all studies which provided indi-
of cardiovascular death. Men participating in 30 min/week vidual group sample sizes [30–32,37,42]. Generally, the
or less of physical activity also had lower risk of cardio- sample sizes for each group varied (lowest n ¼ 31, highest
vascular mortality (hazard ratio ¼ 0.65, 95% CI 0.47–0.90); n ¼ 7689). Overall there is no recognizable difference in
however, there was no difference for moderately active symmetry for both cardiovascular and all-cause mortality,
women. Similar results were observed for all-cause which suggests the absence of publication bias.
mortality.
Hu et al. [37] demonstrated that hypertensive patients DISCUSSION
who engaged in moderate (some activity >4 h/week) or
high (vigorous activity >3 h/week) levels of leisure time This systematic review examined the impact of physical
physical activity had a graded lower risk of cardiovascular activity on cardiovascular and all-cause mortality in patients
death than those who engaged in the lowest category of with high blood pressure. An extensive literature search
leisure time physical activity. Of note, similar results were yielded six studies which addressed this question in pro-
observed in separate analyses for both sexes. Men and spective cohorts. Overall, the studies indicated that physical
women who engaged in moderate activity had a 16 and activity was inversely related with mortality in hypertensive
22% lower risk of cardiovascular mortality, respectively. patients, meaning patients with hypertension who were
Likewise, the most active groups showed further reductions more active showed a lower cardiovascular (16–67%
in risk, totalling 27 and 26% decreased risk of cardiovascular decrease) and all-cause (17–57% decrease) mortality.
death for men and women, respectively. The results from The results indicated that inactive men and women with
the University of Pennsylvania College Alumni cohort indi- high SBP had more than double the risk of cardiovascular
cated that hypertensive patients who engaged in combined death.
light and vigorous sport participation had a 27% reduced
risk of all-cause mortality [39]. Additionally, Paffenbarger Mechanisms
et al. [39] found that the men who engaged in only vigorous Physical activity has been shown to have an inverse
sport participation displayed a 37% decrease in all-cause relationship with blood pressure, as well as other cardio-
death. No decrease in mortality was observed with vascular disease risk factors and mortality in the general
participation in only light activities [39]. population [7,8,48]. Previous studies examining physical
An extensive analysis by Vatten et al. [42] stratified risk activity have demonstrated up to nearly 40% decreased risk
across four categories of blood pressure and four levels of of mortality in women and 35% decreased risk in men
physical activity for both men and women, SBP and DBP across all age groups [7]. The results of this systematic
alike, ultimately showed that regular physical activity was review also showed that this statement is true for patients
beneficial for patients with moderate hypertension in terms with elevated blood pressure and/or hypertension. How-
of lowering cardiovascular risk. Generally, the data dis- ever, the mechanisms by which physical activity may exert
played a pattern of increased risk with increasing blood this effect remain unclear. Meta-analyses have indicated
pressure categories (SBP and DBP) and decreasing levels of that regular aerobic exercise [13,14,49] and resistance train-
physical activity. The participants in the highest blood ing [15,16] decrease blood pressure between 2 and
pressure group (SBP >160 mmHg) who were inactive dis- 6 mmHg. Similar modest decreases in blood pressure have
played greater than double the risk (men: RR 2.24, 95% CI been shown to decrease risk of cardiovascular events and

Journal of Hypertension www.jhypertension.com 1283

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Rossi et al.

Cardiovascular mortality All-cause mortality

8000 8000

7000 7000

6000 6000
Sample size (n)

Sample size (n)

5000 5000

4000 4000

3000 3000

2000 2000

1000 1000

0 0
−0.6 −0.4 −0.2 0 0.2 0.4 0.6 −0.6 −0.4 −0.2 0 0.2
Log HR Log HR
FIGURE 2 Funnel plots of sample size versus log Hazard ratio (HR) in 46 groups for cardiovascular mortality (diamond markers, black lines) and 12 groups for all-cause
mortality (round markers, grey lines). Solid vertical lines represent the mean log HR and dashed vertical lines indicate the median log HR.

cardiovascular mortality [50] by magnitudes comparable to participate in regular physical activity are protected against
those observed with physical activity in this review. Thus it the typical increases in arterial stiffness seen with aging [68].
is possible that the blood pressure-lowering effect of Thus the benefits of physical activity in mediating the
regular physical activity and exercise can account for relationship between blood pressure and mortality are
decreases in cardiovascular and all-cause mortality. Yet, likely a result of changes in cardiovascular risk factors
the effects of physical activity on mortality may be con- and overall arterial health. Nevertheless, evidence from
comitantly exerted through the reduction of other cardio- the eligible reports suggests that physical activity can be
vascular risk factors, for example, improved glucose employed for the primary prevention and management of
tolerance [51], lower BMI [52], reduced platelet activity hypertension and reducing risk of mortality.
[53], and reducing risk of comorbid diseases, for example,
type 2 diabetes mellitus [54]. A review by Arakawa [55] Sex differences
highlighted changes in total peripheral resistance and a Amongst the six eligible studies, four included women in
decrease in plasma volume and/or cardiac index as their sample populations and of those only three indicated
possible mechanisms, amongst several others, though there risk for men and women separately [32,37,42]. The findings
is not enough evidence available to draw strong con- from all studies indicated that physical activity is protective
clusions. Fagard [56] has also suggested a decrease in for both men and women with elevated SBP or DBP. These
vascular resistance, driven by the sympathetic nervous benefits are similar in magnitude for both sexes, where
system and renin–angiotensin systems, as the main mech- highly active men gain between 27 and 45% reduced risk of
anism by which aerobic exercise reduces blood pressure. cardiovascular mortality and women approximately 26–
Patients with different types of hypertension, for example 55% reduced risk. Correspondingly, inactive men with
essential hypertension versus preeclampsia, have an altered elevated blood pressure have more than double the risk
inflammatory profile [57,58]. The sympathetic nervous of cardiovascular mortality, whereas the risk for women is
system and renin-angiotensin system are impacted by almost two and a half times that of the active women with
anti-inflammatory [e.g. interleukin-6 (IL-6)] and pro- lower blood pressure. However, little to no consideration
inflammatory [e.g. tumor necrosis factor-a (TNF-a)] was given to potential differences between sexes in the
markers [59]. Moreover, IL-6 and TNF-a can affect endo- remaining cohorts. Sex-related differences are especially
thelial cells and alter vascular function [59], which is also important to consider given that the average age of partici-
implicated in the pathogenesis of hypertension [60]. pants ranged from 20 to 66 years and blood pressure has
Physical activity has been shown to improve endothelial been shown to differ between men and women across the
function [61] even in clinical populations, for example, lifespan [69,70]. Through adulthood, women typically have
those with obesity [62], coronary artery disease [63], or lower blood pressure levels than men [69,70]. However,
exaggerated inflammation [64]. Thus these pathways may during menopause and subsequently throughout the fol-
mediate the benefits of physical activity on blood pressure lowing decades, there is a shift in this trend, whereby the
and mortality. difference in incidence of hypertension between sexes
Other measures of arterial health, for example arterial narrows and is eventually higher in women [69,70].
stiffness, which are inversely related to mortality [65] are Additionally, age-adjusted comparison of the three phases
also improved with exercise [66]. Improvements have also of the NHANES survey has indicated that prevalence of
been observed in hypertensive patients after 4 weeks of hypertension tends to be higher in adult women compared
aerobic exercise training [67]. Additionally, women who to adult men [71]. Also, from 1988 to 2000 the change in

1284 www.jhypertension.com Volume 30  Number 7  July 2012

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 Blood pressure, physical activity and mortality

prevalence increased in women to a greater extent than in as well. Albeit, most of the data collected in the cohorts
adult men [71]. The mechanisms by which these shifts occur presented here predate the advent of these new technol-
are not yet understood; however, hormonal changes are ogies; however, moving forward, this should be taken
thought to play a significant role [69,70]. into consideration.

Measurement of physical activity Measurement of blood pressure and

The influence of the findings of these studies, or any other
for that matter, rests in the quality of the measurement
hypertension
Another point of inconsistency between the studies was the
tools used to acquire relevant information. In this case,
definition of hypertension. Again, this is likely a result of the
we are concerned with the quantification and classification
recruitment and/or follow-up periods predating standar-
of physical activity. Physical activity was consistently
dized guidelines, which have since significantly evolved.
measured by self-report, whether through questionnaires
However, this adds a level of confusion when comparing
or a structured interview.
the results. For example, the individuals classified as hyper-
Self-report is a poor measure of physical activity
tensive in the study by Engström et al. [30] (SBP
because the data collected are subject to ‘recall bias’ and
160 mmHg and DBP 95 mmHg) would instead be classi-
can be dependent on the participants’ health, mood, and
fied as having ‘moderate or severe’ hypertension according
depression amongst other psychological factors [72].
Despite being easy to administer, particularly when col- to the scheme used by Hu et al. [37]. As noted in the 2007
Guidelines for the Management of Arterial Hypertension of
lecting data in large cohorts, these methods do not suffi-
the European Society of Hypertension and European
ciently capture vital information such as frequency,
duration, intensity or volume of activity. Additionally, each Society of Cardiology [4], previous research has shown that
the relationship between SBP and DBP and cardiovascular
of the studies had defined levels of activity based on the
risk is linear only upwards of 110–115 mmHg and 70–
information available from their respective questionnaires,
75 mmHg, respectively, thus creating a somewhat arbitrary
as opposed to standardized tools, which may not have
cut-off point, designating anything above this point as
been available at the time of data collection. Thus, what
hypertension. Risk of death from ischemic heart disease
may be considered as a high level of activity according to
and stroke is linear upwards of these values and risk of
one study, for example at least 30 min of activity per week
death is increased two-fold for every increase in 20 mmHg
[24], does not equate with the definition of high activity
for SBP and 10 mmHg for DBP [5]. The grading of isolated
in another study, for example, vigorous activity for more
systolic hypertension adds yet another level of complexity.
3 h/week [37]. These discrepancies make the direct com-
parison of results across studies virtually impossible and Established guidelines vary according to the governing
society and are continuously being revised, which has
make it very difficult to identify the optimal frequency,
made and will continue to make room for confusion.
duration, intensity, and volume of activity necessary to
Though there are differences between European [4], Amer-
reduce the risk of death. Despite this shortcoming, the
ican [5], and Canadian [6] guidelines it is important for
results still consistently indicate that there is a decrease in
researchers to follow recognized guidelines when classify-
risk of mortality with increasing levels of activity, no matter
ing patients in order to establish some consistency across
how the latter are defined.
the literature.
Another important aspect to consider when assessing
The quality of blood pressure measurement, again likely
physical activity in large cohorts is the age or age range of
subject to the era of data collection, was relatively sufficient
participants. Age is a determinant of activity energy
across the studies, although not always described in the
expenditure and accordingly, older adults tend to spend
methods [31,32,37,39], and, ultimately, had to be obtained
most of their active time engaged in low-intensity activi-
through other sources [43,44,46]. Blood pressure can vary
ties compared to younger age groups [73]. Measurement
scales need to be sensitive enough to perceive these depending on the time of the day measured, the position of
the patient, and the type of equipment used [5]; as such, it is
patterns. It has been shown that self-report tools validated
important for these details to be specified. In the case of
to measure physical activity in younger adults are erro-
large cohort, multisite trials it is important to standardize
neous when used in older populations [74]. Thus, tools
measurement techniques and ensure all operators have
need to be customized specifically for the populations
been properly trained to take accurate measurements using
in question.
a standardized protocol. It may very well be that clear
To overcome the fundamental flaws inherent in the use
standardized protocols were followed in the measurement
of self-report instruments, researchers are now recom-
of blood pressure in the studies presented; however, the
mending and standardizing the use of objective measures
protocols were generally not well reported.
of physical activity, such as accelerometers [75]. These
instruments allow better characterization of the key
physical activity parameters; frequency, duration, intensity, Limitations
and volume [76], and if used consistently will allow better Several limitations warrant consideration. Firstly, no meta-
comparison between studies. Additionally, accelerometers analysis was performed here due to the heterogeneity of
or similar devices can be worn throughout the waking day the identified studies. Three of the six studies ranked low
and over the course of several days, hence providing an (<12/15) in the risk of bias assessment. External and
excellent opportunity to capture not only leisure time internal validity sections were generally rated with poor
physical activity, but commuting and occupational activity scores. Additionally, the reporting of general participant

Journal of Hypertension www.jhypertension.com 1285

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Rossi et al.

characteristics, such as age, height, weight, race/ethnicity, 4. Mancia G, De Backer G, Dominiczak A, Cifkova R, Fagard R, Germano
G, et al. 2007 Guidelines for the management of arterial hypertension.
smoking status, alcohol consumption, and dietary factors, Eur Heart J 2007; 28:1462–1536.
which are important with respect to the prevention and 5. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL
management of hypertension [5], needs to be improved. Jr, et al. Seventh Report of the Joint National Committee on Prevention,
Of the six studies selected only one, by Fossum et al. [32], Detection, Evaluation, and Treatment of High Blood Pressure. Hyper-
described the pharmacological agents prescribed to the tension 2003; 42:1206–1252.
6. Rabi DM, Daskalopoulou SS, Padwal RS, Khan NA, Grover SA, Hackam
hypertensive participants. It is valuable for researchers to DG, et al. The 2011 Canadian Hypertension Education Program
indicate medication usage so the readers are aware of the Recommendations for the Management of Hypertension: blood pres-
other treatments administered to these patients. This should sure measurement, diagnosis, assessment of risk, and therapy. Can J
be a standard component of reporting data for hypertensive Cardiol 27:415–33.e2.
patients. The major limitations regarding quantification of 7. Andersen LB, Schnohr P, Schroll M, Hein HO. All-cause mortality
associated with physical activity during leisure time, work, sports,
physical activity, blood pressure classification, and sex differ- and cycling to work. Arch Intern Med 2000; 160:1621–1628.
ences have been discussed above. The main driving factor 8. Paffenbarger RS, Hyde RT, Hsieh C-C, Wing AL. Physical activity, other
for this may be the period in which these data were collected, life-style patterns, cardiovascular disease and longevity. Acta Med
beginning as early as 1962 [39]. However, moving forward, Scand 1986; 220:85–91.
9. Salonen JT, Slater JS, Tuomilehto J, Rauramaa R. Leisure time and
this information should be considered necessary and meth- occupational physical activity: risk of death from ischemic heart dis-
odological concerns should be addressed in future studies. ease. Am J Epidemiol 1988; 127:87–94.
Lastly, all of the selected studies are observational in nature 10. Lee I-M, Skerrett PJ. Physical activity and all-cause mortality: what
and therefore any conclusions drawn herein do not infer is the dose-response relation? Med Sci Sports Exerc 2001; 33:S459–S471.
causation. However, given that all of the results of these 11. Blair SN, Kohl HW, Paffenbarger RS, Clark DG, Cooper KH, Gibbons
LW. Physical fitness and all-cause mortality. JAMA 1989; 262:2395–
studies favor physical activity as beneficial for minimizing the 2401.
risk of mortality related to high blood pressure, we consider 12. Blair SN, Kohl HW, Barlow CE, Gibbons LW. Physical fitness and
this strong support for the role of exercise. To properly judge all-cause mortality in hypertensive men. Ann Med 1991; 23:307–
the causative role of physical activity in minimizing risk of 312.
13. Whelton SP, Chin A, Xin X, He J. Effect of aerobic exercise on blood
mortality related to elevated blood pressure, randomized pressure. Ann Intern Med 2002; 136:493–503.
controlled trials would be required. 14. Cornelissen VA, Fagard RH. Effects of endurance training on blood
In conclusion, according to the authors’ knowledge, this pressure, blood pressure–regulating mechanisms, and cardiovascular
is the first systematic review evaluating the impact of risk factors. Hypertension 2005; 46:667–675.
15. Kelley GA, Kelley KS. Progressive resistance exercise and resting blood
physical activity on mortality in individuals with high blood
pressure: a meta-analysis of randomized controlled trials. Hypertension
pressure. Following an exhaustive literature search, six 2000; 35:838–843.
articles were reviewed. Overall, the results indicate that 16. Cornelissen VA, Fagard RH, Coeckelberghs E, Vanhees L. Impact of
there is an inverse relationship between physical activity resistance training on blood pressure and other cardiovascular risk
and blood pressure in hypertensive patients. More research factors. Hypertension 2011; 58:950–958.
17. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis
is warranted to determine the influence of activity fre- JPA, et al. The PRISMA statement for reporting systematic reviews and
quency, duration, intensity and volume on mortality in meta-analyses of studies that evaluate healthcare interventions: expla-
participants with high blood pressure. nation and elaboration. BMJ 2009; 339.
18. Downs SH, Black N. The feasibility of creating a checklist for the
ACKNOWLEDGEMENTS assessment of the methodological quality both of randomised and
nonrandomised studies of healthcare interventions. J Epidemiol Com-
The authors would like to acknowledge Angella Lambrou mun Health 1998; 52:377–384.
(Liason Librarian, McGill University) and Dubravka Kapa 19. Warburton D, Charlesworth S, Ivey A, Nettlefold L, Bredin S. A system-
atic review of the evidence for Canada’s Physical Activity Guidelines for
(Director, Vanier Library, Concordia University) for their Adults. Int J Behav Nutr Phys Activity 2010; 7:39.
assistance with the literature search. 20. Prince S, Adamo K, Hamel M, Hardt J, Gorber S, Tremblay M. A
The authors would like to acknowledge support from comparison of direct versus self-report measures for assessing physical
the Fonds de la Recherche en Santé du Québec (Chercheur- activity in adults: a systematic review. Int J Behav Nutr Phys Activity
2008; 5:56.
Boursier: SLB and Chercheur-Boursier-Clinicien: S.S.D.) as 21. Al-Khalili F, Janszky I, Andersson A, Svane B, Schenck-Gustafsson K.
well as the Canadian Institutes of Health Research (Vanier Physical activity and exercise performance predict long-term prognosis
Canada Graduate Scholarship: A.R.). in middle-aged women surviving acute coronary syndrome. J Intern
Med 2007; 261:178–187.
Conflicts of interest 22. Apullan FJ, Bourassa MG, Tardif J-C, Fortier A, Gayda M, Nigam A.
Usefulness of self-reported leisure-time physical activity to predict
There are no conflicts of interest. long-term survival in patients with coronary heart disease. Am J Cardiol
2008; 102:375–379.
REFERENCES 23. Batty GD, Shipley MJ, Marmot M, Davey Smith G. Physical activity and
1. Ezzati M, Lopez AD, Rodgers A, Vander Hoorn S, Murray CJL. Selected cause-specific mortality in men with type 2 diabetes/impaired glucose
major risk factors and global and regional burden of disease. Lancet tolerance: evidence from the Whitehall study. Diabet Med 2002;
2002; 360:1347–1360. 19:580–588.
2. Kearney PM, Whelton M, Reynolds K, Muntner P, Whelton PK, He J. 24. Boman K, Gerdts E, Wachtell K, Dahlöf B, Nieminen MS, Olofsson M,
Global burden of hypertension: analysis of worldwide data. Lancet et al. Exercise and cardiovascular outcomes in hypertensive patients in
2005; 365:217–223. relation to structure and function of left ventricular hypertrophy: the
3. Lewington S, Clarke R, Qizilbash N, Peto R, Collins R. Age-specific LIFE study. J Cardiovasc Risk 2009; 16:242–248.
relevance of usual blood pressure to vascular mortality: a meta-analysis 25. Carlsson S, Andersson T, Wolk A, Ahlbom A. Low physical activity and
of individual data for one million adults in 61 prospective studies. mortality in women: baseline lifestyle and health as alternative expla-
Lancet 2002; 360:1903–1913. nations. Scand J Public Health 2006; 34:480–487.

1286 www.jhypertension.com Volume 30  Number 7  July 2012

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 Blood pressure, physical activity and mortality

26. Charlton KE, Lambert EV, Kreft J. Physical activity, change in blood 46. Kuulasmaa K, Hense H-W, Tolonen H, Project ftM. Quality assessment
pressure and predictors of mortality in older South Africans - a 2-year of data on blood pressure in the WHO MONICA Project. 1998 [cited
follow-up study. S Afr Med J 1997; 87:1124–1130. 2011 June 6]; Available from: http://www.ktl.fi/publications/monica/
27. De Backer G, Kornitzer M, Dramaix M, Kittel F, Tiolly C, Graffar M, et al. bp/bpqa.htm.
The Belgian Heart Disease Prevention Project: 10-year mortality follow- 47. Dahlöf B, Devereux RB, Kjeldsen SE, Julius S, Beevers G, de Faire U,
up. Eur Heart J 1988; 9:238–242. et al. Cardiovascular morbidity and mortality in the Losartan Inter-
28. Dorn JP, Cerny FJ, Epstein LH, Naughton J, Vena JE, Winkelstein W Jr, vention For Endpoint reduction in hypertension study (LIFE): a
et al. Work and leisure time physical activity and mortality in men and randomised trial against atenolol. Lancet 2002; 359:995–1003.
women from a general population sample. Ann Epidemiol 1999; 48. Haapanen N, Miilunpalo S, Vuori I, Oja P, Pasanen M. Association of
9:366–373. leisure time physical activity with the risk of coronary heart disease,
29. Eaton CB, Medalie JH, Flocke SA, Zyzanski SJ, Yaari S, Goldbourt U. hypertension and diabetes in middle-aged men and women. Int J
Self-reported physical activity predicts long-term coronary heart Epidemiol 1997; 26:739–747.
disease and all-cause mortalities: twenty-one-year follow-up of 49. Fagard RH. Exercise characteristics and the blood pressure response to
The Israeli Ischemic Heart Disease Study. Arch Fam Med 1995; 4: dynamic physical training. Med Sci Sports Exerc 2001; 33:S484–S492.
323–329. 50. Staessen JA, Wang J-G, Thijs L. Cardiovascular protection and blood
30. Engström G, Hedblad B, Janzon L. Hypertensive men who exercise pressure reduction: a meta-analysis. Lancet 2001; 358:1305–1315.
regularly have lower rate of cardiovascular mortality. J Hypertens 1999; 51. Heath GW, Gavin JR, Hinderliter JM, Hagberg JM, Bloomfield SA,
17:737–742. Holloszy JO. Effects of exercise and lack of exercise on glucose
31. Fang J, Wylie-Rosett J, Alderman MH. Exercise and cardiovascular tolerance and insulin sensitivity. J Appl Physiol 1983; 55:512–517.
outcomes by hypertensive status: NHANES I epidemiological follow- 52. Hu G, Barengo NC, Tuomilehto J, Lakka TA, Nissinen A, Jousilahti P.
up study, 1971–1992. Am J Hypertens 2005; 18:751–758. Relationship of physical activity and body mass index to the risk of
32. Fossum E, Gleim GW, Kjeldsen SE, Kizer JR, Julius S, Devereux RB, hypertension: a prospective study in Finland. Hypertension 2004;
et al. The effect of baseline physical activity on cardiovascular out- 43:25–30.
comes and new-onset diabetes in patients treated for hypertension and 53. Rauramaa R, Salonen J, Seppanen K, Salonen R, Venalainen J, Ihanai-
left ventricular hypertrophy: the LIFE study. J Intern Med 2007; nen M, et al. Inhibition of platelet aggregability by moderate-intensity
262:439–448. physical exercise: a randomized clinical trial in overweight men.
33. Glynn RJ, Field TS, Hebert PR, Taylor JO, Hennekens CH, Rosner B. Circulation 1986; 74:939–944.
Evidence for a positive linear relation between blood pressure and 54. Hu G, Qiao Q, Silventoinen K, Eriksson JG, Jousilahti P, Lindström J,
mortality in elderly people. Lancet 1995; 345:825–829. et al. Occupational, commuting, and leisure-time physical activity in
34. Holtermann A, Mortensen OS, Burr H, Søgaard K, Gyntelberg F, relation to risk for type 2 diabetes in middle-aged Finnish men and
Suadicani P. Fitness, work, and leisure-time physical activity and women. Diabetologia 2003; 46:322–329.
ischaemic heart disease and all-cause mortality among men with 55. Arakawa K. Antihypertensive mechanism of exercise. J Hypertens 1993;
preexisting cardiovascular disease. Scand J Work Environ Health 11:H45.
2010; 36:366–372. 56. Fagard RH. Exercise is good for your blood pressure: effects of
35. Holtermann A, Mortensen OS, Burr H, Søgaard K, Gyntelberg F, endurance training and resistance training. Clin Exp Pharmacol Physiol
Suadicani P. Physical work demands, hypertension status, and 2006; 33:853–856.
risk of ischemic heart disease and all-cause mortality in the Copen- 57. Peeters ACTM, Netea MG, Janssen MCH, Kullberg BJ, Van der Meer
hagen Male Study. Scand J Work Environ Health 2010; 36:466– JWM, Thien T. Pro-inflammatory cytokines in patients with essential
472. hypertension. Eur J Clin Invest 2001; 31:31–36.
36. Holtermann A, Mortensen OS, Burr H, Søgaard K, Gyntelberg F, 58. LaMarca B, Ryan M, Gilbert J, Murphy S, Granger J. Inflammatory
Suadicani P. Physical demands at work, physical fitness, and 30-year cytokines in the pathophysiology of hypertension during preeclamp-
ischaemic heart disease and all-cause mortality in the Copenhagen sia. Curr Hypertens Rep 2007; 9:480–485.
Male Study. Scand J Work Environ Health 2010; 36:357–365. 59. Granger JP. An emerging role for inflammatory cytokines in
37. Hu G, Jousilahti P, Antikainen R, Tuomilehto J. Occupational, commut- hypertension. Am J Physiol Heart Circulat Physiol 2006; 290:H923–
ing, and leisure-time physical activity in relation to cardiovascular H924.
mortality among Finnish subjects with hypertension. Am J Hypertens 60. Taddei S, Virdis A, Ghiadoni L, Sudano I, Salvetti A. Endothelial
2007; 20:1242–1250. dysfunction in hypertension. J Cardiovasc Pharmacol 2001; 38:S11–
38. Martinson BC, O’Connor PJ, Pronk NP. Physical inactivity and short- S14.
term all-cause mortality in adults with chronic disease. Arch Intern Med 61. Moyna NM, Thompson PD. The effect of physical activity on endo-
2001; 161:1173–1180. thelial function in man. Acta Physiol Scand 2004; 180:113–123.
39. Paffenbarger RS, Jung DL, Leung RW, Hyde RT. Physical activity and 62. Sciacqua A, Candigliota M, Ceravolo R, Scozzafava A, Sinopoli F,
hypertension: an epidemiological view. Ann Med 1991; 23:319– Corsonello A, et al. Weight loss in combination with physical activity
327. improves endothelial dysfunction in human obesity. Diabetes Care
40. Paffenbarger RS, Lee I. Intensity of physical activity related to incidence 2003; 26:1673–1678.
of hypertension and all-cause mortality: an epidemiological view. 63. Hambrecht R, Adams V, Erbs S, Linke A, Kränkel N, Shu Y, et al. Regular
Blood Press Monit 1997; 2:115–123. physical activity improves endothelial function in patients with cor-
41. Savela S, Koistinen P, Tilvis R, Strandberg A, Pitkälä K, Salomaa V, et al. onary artery disease by increasing phosphorylation of endothelial nitric
Leisure-time physical activity, cardiovascular risk factors and mortality oxide synthase. Circulation 2003; 107:3152–3158.
during a 34-year follow-up in men. Eur J Epidemiol 2010; 25:619– 64. Ford ES. Does exercise reduce inflammation? Physical activity and
625. C-reactive protein among U.S. adults. Epidemiology 2002; 13:561–
42. Vatten LJ, Nilsen TI, Holmen J. Combined effect of blood pressure and 568.
physical activity on cardiovascular mortality. J Hypertens 2006; 24: 65. Vlachopoulos C, Aznaouridis K, Stefanadis C. Prediction of cardiovas-
1939–1946. cular events and all-cause mortality with arterial stiffness: a systematic
43. National Health and Nutrition Examination Survey. Public Use review and meta-analysis. J Am Coll Cardiol 2010; 55:1318–1327.
Data Tape Documentation: Medical Examination Ages 1-74 Tape 66. Kingwell BA, Berry KL, Cameron JD, Jennings GL, Dart AM. Arterial
Number 4233. U.S. Department of Health and Human Services; compliance increases after moderate-intensity cycling. Am J Physiol
1981. Heart Circulat Physiol 1997; 273:H2186–H2191.
44. Dahlöf B, Devereux R, Faire Ud, Fyhrquist F, Hedner T, Ibsen H, et al. 67. Collier SR, Kanaley JA, Carhart R Jr, Frechette V, Tobin MM, Hall AK,
The Losartan Intervention For Endpoint Reduction (LIFE) in Hyper- et al. Effect of 4 weeks of aerobic or resistance exercise training on
tension Study Rationale, Design, and Methods. Am J Hypertens 1997; arterial stiffness, blood flow and blood pressure in pre and stage-1
10:705–713. hypertensives. J Hum Hypertens 2008; 22:678–686.
45. Dahlöf B, Devereux RB, Julius S, Kjeldsen SE, Beevers G, de Faire U, 68. Tanaka H, DeSouza CA, Seals DR. Absence of age-related increase in
et al. Characteristics of 9194 patients with left ventricular hypertrophy: central arterial stiffness in physically active women. Arterioscler
The LIFE Study. Hypertension 1998; 32:989–997. Thromb Vac Biol 1998; 18:127–132.

Journal of Hypertension www.jhypertension.com 1287

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Rossi et al.

69. Oparil S, Miller AP. Gender and blood pressure. J Clin Hypertens 2005; 73. Westerterp KR. Physical activity as determinant of daily energy expen-
7:300–309. diture. Physiol Behav 2008; 93:1039–1043.
70. Reckelhoff JF. Gender differences in the regulation of blood pressure. 74. Washburn RA. Assessment of physical activity in older adults. Res Q
Hypertension 2001; 37:1199–1208. Exerc Sport 2000; 71:79.
71. Hajjar I, Kotchen TA. Trends in prevalence, awareness, treatment, and 75. Murphy SL. Review of physical activity measurement using acceler-
control of hypertension in the United States, 1988–2000. JAMA 2003; ometers in older adults: considerations for research design and con-
290:199–206. duct. Prev Med 2009; 48:108–114.
72. Rikli RE. Reliability, validity, and methodological issues in 76. Plasqui G, Westerterp KR. Physical activity assessment with acceler-
assessing physical activity in older adults. Res Q Exerc Sport 2000; ometers: an evaluation against doubly labeled water. Obesity 2007;
71:89. 15:2371–2379.

1288 www.jhypertension.com Volume 30  Number 7  July 2012

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.