Table 14 Studies examining the relationship between physical activity and hypertension.

Rankinen et al 2007 [75]

To investigate the contributions of DNA sequence variation in candidate genes, PF and BMI, as well as their interactions to the incidence of hypertension.

• n = 629 cases; 605 controls

10 year follow up

PF showed the strongest association with HTN risk among all subjects as well as sex-specific models. Each 1- MET increment in PF was associated with 19% (12- 14%), 16% (9-22%), 32% (17- 45%) risk reduction in all subjects, men and women respectively.

PF is a significant predictor of the risk of hypertension.

USA

• Sex: Men and women

All subjects required to have 2 clinic visits at least 2 years apart.

Case control

• Age: Case: 43.3 (9.2) yr Control: 42.7 (8.9) yr

PF assessment: treadmill test (Blake protocol)

D & B score = 13

• Characteristics: Healthy with BP 134/86 mmHg or less at their first clinic visit. Cases: those who developed hypertension during the follow-up period. Controls were those who did not develop hypertension

Outcome measure: Incidence of hypertension during follow-up. Incident cases of hypertension were defined as physician diagnosed hypertension with medication or SBP ≥ 140 mmHg and/of DBP ≥ 90 mmHg t-tests and chi-square tests Logistic regression modelling

When divided into quartiles on the basis of sex specific MET cut-offs, the third and fourth quartiles had a 58% (41-71%) and 63% (47-75%) lower risk of hypertension compared to the 1^stquartile.

Pereira et al 1999 [76]

To examine PA and incident hypertension in men and women.

• n = 7,459

PA Assessment: Questionnaire for leisure, sport and work index, divided into quartiles

White Men

There is an inverse association between PA and incident hypertension in White middle aged men. White men in the highest quartiles of sport and leisure activity had statistically significant reductions in the odds of developing hypertension of 23 and 34% respectively, compared to men in the lower quartiles.

USA

• Sex: Men and women

Q1 = Lowest

Leisure Index Model 1

Prospective cohort

• Age: 45-65 yr

Q2

• Q1 = 1.00 (referent)

D & B score = 12

• Characteristics: No history of angina, MI, evidence of MI, angioplasty or other CV surgery or hypertension

Q3

• Q2 = 0.95 (0.70-1.28)

• Atherosclerosis Risk in Communities Study

Q4 = Highest

• Q3 = 0.83 (0.63-1.09)

Model 1 adjusted for: Age, education, baseline BP and study centre

• Q4 = 0.64 (0.46-0.89)

Model 2 adjusted for: Covariates in model 1 and smoking, alcohol consumption, parental history of hypertension, energy, sodium, potassium and caffeine intake, BMI, waist to hip ratio, menopausal status and hormone use

Trend p = 0.01

Outcome Measure: Incidence of hypertension as defined as a SBP 140 mmHg and/or a DBP 90 mmHg or use of antihypertensive medications.

Leisure Index Model 2

Unconditional logistic regression Orthogonal polynomial coefficients

• Q1 = 1.00 (referent)

• Q2 = 0.99 (0.72-1.35)

• Q3 = 0.86 (0.65-1.13)

• Q4 = 0.66 (0.47-0.94)

Trend p = 0.01

Women

Sport Index Model 1

• Q1 = 1.00 (referent)

• Q2 = 1.26 (0.78-2.05)

• Q3 = 1.06 (0.61-1.84)

• Q4 = 1.92 (1.12-3.29)

Trend p = 0.04

Men

Sport Index Model 1

• Q1 = 1.00 (referent)

• Q2 = 1.23 (0.91-1.66)

• Q3 = 0.92 (0.70-1.22)

• Q4 = 0.74 (0.54-1.02)

Trend p = 0.02

Sport Index Model 2

• Q1 = 1.00 (referent)

• Q2 = 1.26 (0.93-1.71)

• Q3 = 0.95 (0.71-1.26)

• Q4 = 0.77 (0.55-1.08)

Trend p = 0.05

Haapanen et al 1997 [77]

To assess the association between PA and hypertension.

• n = 732 men; 796 women

10 year follow up (1980 baseline)

Age adjusted incidence rates ofhypertension Total energy expenditure High as referent:

Increased EE during LTPA and increased intensity of these activities were associated with reduced risk for incident hypertension (age adjusted) in men but not women.

Finland

• Sex: Men and women

PA assessement: Questionnaire for EE (kcal/wk), divided into tertiles

Prospective cohort

• Age: 35-65 years

Male

Male:

D & B score = 11

• Characteristics: Free of hypertension at baseline. Excluded those unable to participate in regular PA due to poor health

   T1 = Low = 0-1100

• T1 = 1.00 (referent)

   T2 = Medium = 1101-1900

• T2 = 1.66

   T3 = High >1900

• T3 = 1.73

Trend p = 0.021

Female

Female:

   T1 = Low = 0-900

• T1 = 1.00 (referent)

   T2 = Medium = 901-1500

• T2 = 0.94

   T3 = High = >1500

• T3 = 1.16

Outcome measure: Incidence of hypertension through self reported diagnosis and death certificates

Trend p = 0.648

Cox proportional HR

Paffenbarger et al 1983 [78]

To examine the relationship of student and alumnus PA patterns and other characteristics with incident hypertension.

• n = 14,998

PA Assessment: Questionnaire for PA based on number of stairs ascended, blocks walked and hours per week of light and vigorous sports play, yard work etc.

There was no significantly reduced risk for hypertension in men who climbed 50 plus stairs per day (compared to < 50 stairs); who walked 5 plus blocks per day (compared to < 5 blocks); or who played light sports (compared to those who did not).

Contemporary vigorous exercise was inversely related to hypertension risk.

USA

• Sex: Men

Prospective cohort

• Age: 35-74 yr

Outcome measure: Diagnosis of hypertension by physicians using criteria of SBP > 160 mmHg and/or DBP > 95 mmHg

The 59% of men who did engage in vigorous sports were at 35% greater risk of hypertension than the 41% who did not.

RR = 1.35

Trend p = <0.001

D & B score = 12

• Characteristics: free of hypertension Harvard Alumni Study

Multivariate estimates

Alumni on the low side of the physical activity index (< 2000 kcal/wk) had a 30% increased risk of hypertension then those ≥ 2000 kcal/wk.

RR = 1.30

Trend p = 0.004

Paffenbarger et al 1997 [79]

To investigate the quantity and intensity of energy expenditure required to delay hypertension and prevent premature death.

• n = 6,390

PA Assessment: Questionnaire for weekly sports play, divided into tertiles

RR (95% CI)

Lack of vigorous sports play independently increased the risk of developing hypertension.

USA

• Sex: Men

T1 = None

• T1 = 1.00 (referent)

Prospective cohort

• Age: 45-84 yr

T2 = Light Only (< 4.5 METs)

• T2 = 1.04 (0.77-1.40)

D & B score = 12

• Characteristics: Free of hypertension, CHD, diabetes, COPD and potentially malignant cancer in 1977

T3 = Moderately vigorous (≥ 4.5 METs)

• T3 = 0.77 (0.62-0.96)

• Harvard Alumni Study

Outcome measure: Incident hypertension

Trend p = 0.004

Hu et al 2004 [81]

To discover whether regular PA can reduce the risk of hypertension in normal weight and overweight men and women.

• n = 8,302 men; 9,139 women

11 year follow up

Multivariate adjusted HR (95% CI), men

Regular PA can reduce the risk of hypertension. The protective effect of PA was observed in both sexes regardless of level of obesity.

Finland

• Sex: Men and women

PA assessement: Questionnaire for OPA, LTPA and commuting PA, divided into tertiles

• T1 = 1.00 (referent)

Prospective cohort

Age: 25-64 yr

T1 = Low

• T2 = 0.63

D & B score = 13

Characteristics: Healthy and free of hypertension at baseline

T2 = Medium

• T3 = 0.59

T3 = High

Trend p = < 0.001

Outcome Measure: Incidence of drug treated hypertension

Multivariate adjusted HR (95% CI), women

Cox proportional HR

• T1 = 1.00

• T2 = 0.82

• T3 = 0.71

Trend p = 0.005

Gu et al 2007 [82]

To determine the 8-year incidence of HTN and its risk factors among Chinese adults.

• n = 10,525

Baseline Examination in 1991 with 8 year follow up

RR (95% CI), men

Increasing PA has the potential to reduce incidence of hypertension.

China

• Sex: Men and women

PA assessment: Questionnaire administered by trained staff, divided into groups

• G1 = 1.00 (referent)

Prospective cohort

Age: ≥ 40 yr

G1 = Low

• G2 = 1.12 (0.86-1.46)

D & B score = 13

Characteristics: Healthy and free from hypertension at baseline.

G2 = Medium

• G3 = 1.27 (1.10-1.47)

G3 = High

RR (95% CI), women

Outcome measure: HTN as defined at SBP ≥ 140 mmHg and/or DBP ≥ 90 mmHg or current use of antihypertensive medication

• G1 = 1.00 (referent)

t-tests, chi squared tests, Cochran-Armitage modeling, Modified Poisson approach

• G2 = 1.14 (0.98-1.34)

• G3 = 1.22 (1.02-1.45)

Hayashi et al 1999 [83]

To investigate the association of the duration of the walk to work and LTPA with the risk for hypertension.

• n = 6,017

PA assessment: Questionnaire on health related behaviours and exercise Walk time to work

RR (95% CI) Frequency walk time to work (minutes

The duration of walk to work was associated with a decreased risk of hypertension even after adjustment.

Japan

• Sex: Men

T1 = 0-10 min

• T1 = 1.00 (referent)

Regular PA (at least once weekly) was inversely related to the risk of incident hypertension

Prospective cohort

• Age: 35-60 yr

T2 = 11-20 min

• T2 = 0.65 (0.47-0.90)

D & B score = 12

• Characteristics: Free from HTN at baseline. All employees at gas company in Osaka Japan. All had sedentary jobs.

T3 = ≥ 21 min

• T3 = 0.72 (0.59-0.88)

Outcome measure: Diagnosed with hypertension (as defined by a SBP ≥ 160 mmHg, a DBP ≥ 95 mmHg, or use of antihypertensive medication)

Trend p = < 0.001

Cox proportional HR

Nakanishi et al 2005 [84]

To examine the relationship of overall PA to the risk of developing hypertension in normotensive Japanese male office workers over a 7 year observation period.

• n = 2,548

7 year follow up

Multivariate adjusted RR (95% CI) by PA level only

The rate of rise in both SBP and DBP in each follow-up year decreased with higher EE and that the risk of developing hypertension decreased in a dose dependent manner with higher daily life activity level.

Japan

• Sex: Men

Q1 = 1.00 (referent)

Analysis stratified by the presence of or absence of a risk factor showed the negative association of daily life activity with the risk of developing hypertension for men at both low and high risk. This tendency was also observed among men in all 3 categories of normotension.

Prospective cohort

• Age: 35-59 yr

PA assessment: 1-day activity record and reported the type and frequency on a weekly basis of LTPA, divided into quartiles (kcal/kg/d)

Q2 = 0.84 (0.72-0.98)

D & B score = 12

• Characteristics Healthy at baseline. No hypertension or CHD. All office workers for a Japanese company

• Q1 = <33.3

• Q3 = 0.75 (0.63-0.88)

• Q2 = 33.3-36.9

• Q4 = 0.54 (0.45-0.64)

• Q3 = 37.0-40.3

Trend p = < 0.001

• Q4 = 40.4

Multivariate adjusted RR (95% CI) by PA level, low normal BP

3 categories of normotensive BP Low Normal: SBP < 120, DBP < 80 Normal: SBP 120-130, DBP 80- 85 High Normal: SBP 130-139 DBP 85-89

• Q1 = 1.00 (referent)

3 categories of normotensive BP Low Normal: SBP < 120, DBP < 80 Normal: SBP 120-130, DBP 80- 85 High Normal: SBP 130-139 DBP 85-89

• Q2 = 0.70 (0.47-1.05)

Cox proportional hazard model

• Q3 = 0.55 (0.37-0.83)

• Q4 = 0.43 (0.28-0.65)

Trend p = <0.001

Multivariate adjusted RR (95% CI) by PA level, normal

BP

• Q1 = 1.00 (referent)

• Q2 = 0.89 (0.68-1.16)

• Q3 = 0.69 (0.52-0.91)

• Q4 = 0.50 (0.37-0.68)

Trend p = <0.001

Multivariate adjusted RR (95% CI) by PA level, high normal BP

• Q1 = 1.00 (referent)

• Q2 = 0.86 (0.69-1.07)

• Q3 = 0.88 (0.69-1.11)

• Q4 = 0.60 (0.46-0.78)

Trend p = 0.001

Foy et al 2006 [85]

To examine whether insulin resistance is associated with the effect of vigorous or moderate PA on baseline BP.

• n = 1,599

Baseline examination in 1992-1993

Unadjusted OR (95% CI)

Participants who meet or exceed current caloric expenditure recommendations for VPA demonstrate significantly less hypertension than do sedentary or underactive individuals.

USA

• Sex: Men and women

PA assessment: VPA over the past year was determined via a 1-year recall of physical activity (kcal/d), divided into 3 groups

• T1 = 1.00 (referent)

Cross sectional

• Age: 40-69 yr

• T1 = O

• T2 = 0.69 (0.53-0.88)

D & B score = 12

• Characteristics: Community dwelling adults

• T2 = 1-149 kcal/day

• T3 = 0.57 (0.45-0.74)

• Insulin Resistance Atherosclerosis Study

• T3 = >150 kcal/day

• Trend p = < 0.001

Adjusted OR (95% CI)

• T1 = 1.00 (referent)

• T2 = 0.82 (0.62-1.09)

• T3 = 0.73 (0.55-0.98)

Trend p = 0.004

Folsom et al 1990 [270]

To examine the relationship between fat distribution and the 2-yr incidence of hypertension and stroke.

• n = 41,837

Baseline mailed survey in 1986: Pa assessment: Questionnaire for LTPA

• 978 cases

High PA reduced the risk of hypertension only before adjusting for other factors.

USA

• Sex: Women

T1 = Low

Age Adjusted RR (95% CI)

Prospective cohort

• Age: 55-69 years (yr)

T2 = Medium

• T1 = 1.00 (referent)

D & B score = 12

• Characteristics: All free of HTN at baseline

T3 = High

• T2 = 0.9 (0.7-1.1)

Mantel-Haenszel method

• T3 = 0.7 (0.6-0.9)

Multiple logistic regression

Levenstein et al 2001 [271]

To examine the effects of a variety of psychosocial factors on the development of HTN in men and women in the general population.

• n = 1,031 men, 1,326 women

Questionnaires in 1965 and 1974, cohort followed until 1994

LTPA predictor of hypertension OR (95% CI)

Risk of HTN was reduced with increases in LTPA in women.

USA

• Sex: Men and women

PA assessment: LTPA rated on a scale of 0-16 points and analysed as a continuous variable

• All Subjects: 0.94 (0.91-0.97)

Prospective cohort

• Characteristics: Free of hypertension at baseline

Outcome measure: Incidence of hypertension (defined as those who are taking antihypertensive medications)

• Women: 0.90 (0.87-0.94)

D & B score = 13

• Alameda cohort study

Logistic regression analysis

• Men: 0.98 (0.94-1.02)