NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Clinical Guideline Centre (UK). Hypertension: The Clinical Management of Primary Hypertension in Adults: Update of Clinical Guidelines 18 and 34 [Internet]. London: Royal College of Physicians (UK); 2011 Aug. (NICE Clinical Guidelines, No. 127.)

Cover of Hypertension

Hypertension: The Clinical Management of Primary Hypertension in Adults: Update of Clinical Guidelines 18 and 34 [Internet].

Show details

11Lifestyle interventions

11.1. Overview

A vast epidemiological literature describes an apparent relationship between raised blood pressure and lifestyle choices and habits. For example, observational studies have shown that people with raised blood pressure tend also to have low dietary calcium627. Does inadequate intake of dietary calcium promote raised blood pressure or is the relationship a spurious one, arising from inadequate adjustment for other hard-to-measure influences (a common problem in observational studies). There is similar controversy about the role of diet, exercise, alcohol, caffeine, potassium and magnesium supplements, sodium (table) salt and relaxation therapies. Cause and effect can only be established by repeated and methodologically sound randomized controlled trials, supported by evidence of a plausible biological mechanism, particularly when the potential benefit is small.

Randomized controlled trials, enrolling patients who had raised average blood pressure defined as systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥85 mmHg, analysing either blood pressure or major cardiovascular endpoints on an intention-to-treat basis, of eight weeks or more follow-up, are included in this review. However, none of the studies identified were designed to quantify significant changes in rates of death or cardiovascular events due to lifestyle interventions: instead they relied on the surrogate endpoint of reduced blood pressure with its epidemiological link to reduced rates of disease. Thus the evidence is less direct than for drug interventions which show reductions in morbidity directly. The requirement that trials have a follow-up of at least eight weeks is arbitrary but it reflects the belief that shorter time frames cannot usefully inform us about enduring changes in blood pressure.

We searched electronic databases (Medline, Embase, CENTRAL) from 1998 to July 2003 for reports of relevant randomised controlled trials; articles published before 1998 were identified from hypertension guidelines, systematic reviews and meta-analyses31,118,187,192,214,293,366,388,37,117,153,204,205,238,239,248,251,268,279,299,300,319–323,444,489,632–634, 152,241,350,407. Though there were a number of trials informing most of the areas of interest, the trials were commonly small and the intervention of short duration (several months) relative to the progression of raised blood pressure and cardiovascular disease. The quality of reporting of studies was commonly poor (Table 48) and this may reflect poor methodological conduct, further weakening the strength of evidence and consequent recommendations for clinical care.

Table 48. Summary characteristics of trials of lifestyle interventions.

Table 48

Summary characteristics of trials of lifestyle interventions.

In overview, 98 trials including 7,993 participants were combined to provide principal findings on lifestyle interventions (see Figure 4) although these were augmented with a number of other trials and reviews. Statistically significant reductions in blood pressure were found, in the short term for improved diet and exercise, relaxation therapies, and sodium and alcohol reduction. For example, our best estimate is that a multiple intervention addressing diet and exercise can reduce systolic and diastolic blood pressure in a cohort of patients, on average, by about 5 mmHg. However this estimate is based on a limited number of patients and is uncertain. The 95% confidence interval shows that (19 times out of 20) the true average reduction may be anywhere between about 2 and 9 mmHg. Individual patients may achieve a greater or lesser reduction than the average and for a combined diet and exercise intervention the best guess is that about one quarter of patients will achieve a reduction in systolic blood pressure of at least 10 mmHg.

Figure 4. Overview of lifestyle interventions: effect on systolic and diastolic blood pressure in randomised trials of patients with raised blood pressure (≥140/85mmHg).

Figure 4

Overview of lifestyle interventions: effect on systolic and diastolic blood pressure in randomised trials of patients with raised blood pressure (≥140/85mmHg). All estimates are DerSimonian-Laird Weighted Mean Differences, see individual meta-analyses (more...)

Most areas featured considerable heterogeneity (i.e. study findings were inconsistent, some positive and some negative) over and above the variation expected by the normal play of chance. This heterogeneity tends to limit the strength of recommendation that can be made about any course of action.

11.1.1. Managing changes in lifestyle

Our systolic (and to a lesser extent our diastolic) blood pressure tends to increase as we grow older. It is unhelpful to think of a single threshold above which we suddenly have problematically high blood pressure, although such thresholds can be useful to spur us into action. A review of our lifestyle helps us to identify changes we can make which may reduce our blood pressure and thus delay, reduce or remove the need for long term drug therapy as well as leading to a healthier life. The cumulative trial evidence suggests that individuals who develop improved habits of regular exercise, sensible diet and relaxation can reduce their blood pressure. Forming these habits will take determination and support. Health care professionals can provide advice, encouragement and materials but ultimately may have limited scope to influence poor dietary habits and inadequate exercise which result in part from the busy and stressful pace of life and in part from personal choice. Much of the research evidence for lifestyle change uses regular time spent together in groups for support and encouragement. Patient and healthcare organisations may be able to help provide patients with, or point them to local groups which encourage lifestyle change, particularly those promoting healthy eating and regular exercise.

11.1.2. Diet

Fourteen randomised controlled trials, including 1,474 participants, met the review inclusion criteria. 18,45,84,138,144,235,262,295,310,406,508,520,545,577,617, 380,495,499,502. Studies most commonly compared low calorie diets, aimed at overweight patients, with either the patients’ usual diet or with a prescribed ‘usual care’ diet. In addition, one study compared fish oil capsules with olive oil capsules (as a control); one study compared diets supplemented with fibre from oats and wheat; one study compared soy milk with skimmed cows’ milk; these studies are discussed separately498, 158, 510.

The mean age of study participants was 48 years and 62% were male. Only four studies reported ethnicity and in these about 45% of the participants were white. The median duration of both treatment and follow-up was 26 weeks, ranging from eight weeks to one year.

Randomisation could be confirmed as adequate in only three studies (21%) and concealment of allocation as adequate in only one (7%). Blinding was confirmed as adequate in six studies (43%). Treatment and control groups were confirmed as comparable at baseline, with regard to age, sex and initial blood pressure in 12 studies (86%).

Studies varied in their methods and in definitions of diets prescribed. Some focussed primarily on low saturated fat, others primarily on weight reduction but in practice there was considerable overlap of content. Patients were sometimes given advice on other aspects of lifestyle, such as exercise. Dieticians, nurses or counsellors generally delivered interventions although in two studies doctors were primarily involved. Two of the studies provided meals for the participants406,520. Contact between participants and the treatment providers varied considerably from several times weekly through to occasionally. Crucially, we could identify no clear system for sub-grouping diet studies: there were too many confounding influences.

There was generally little change in the weight of people in the control groups, whereas average study losses in dietary intervention groups were between two and nine kilograms.

Average changes in blood pressure, when comparing treatment and control groups, are shown in Figure 5. Overall, with dietary intervention there was a significant reduction in both systolic (6.0 mmHg, 95% CI: 3.4 to 8.6) and diastolic (4.8 mmHg, 95%CI: 2.7 to 6.9) blood pressure. There was no evidence of reporting bias, but significant heterogeneity existed between studies. Forty percent (95%CI: 33% to 47%) of patients put on diets were likely to show at least a 10 mmHg reduction in systolic blood pressure. There was no overall difference in withdrawal when comparing diet and control arms of studies (treatment vs. control, risk difference 3.6%, 95%CI: −0.1% to 7.2%), although studies varied.

N:\NCGC Guidelines\Hypertension update_2009\6-Guideline drafts and Final Documents\Images\evidencef8.png

Figure 5

Effect of diet on systolic and diastolic blood pressure in randomised trials of patients with raised blood pressure. †DerSimonian-Laird Weighted Mean Difference Systolic BP: DL= −6.0 (95% CI: −8.6 to −3.4); Q:p = <0.001; (more...)

Omission of a study which enrolled abnormally hypertensive patients (mean baseline BP: 170/110 mmHg)508 resulted in a more modest estimate of reduced blood pressure due to diet: systolic 5.0 mmHg (95% CI: 3.1 to 7.0) and diastolic 3.7 mmHg (95%CI: 2.4 to 5.1).

While soy milk appeared to lower blood pressure when compared to skimmed cows’ milk510 and fish oil appeared to lower blood pressure when compared to olive oil135, these findings were from single small short-term studies and require substantiation by other independent studies. In one small study, supplementing the diet with oats did not appear to lower blood pressure when compared to wheat158.

The Cochrane Collaboration415 carried out a review which had different inclusion criteria (it included simple interventions reported up to June 1998, had no restriction on length of follow up and also used weight loss as an end point) leaving only four studies common to both reviews. Nevertheless, its conclusions were similar. The recent Canadian guideline reviewed studies between 1966 and 1996355. Although without a formal meta-analysis, it likewise concluded that overweight hypertensive patients should be advised to reduce their weight.

11.1.3. Exercise

Seventeen randomised controlled trials of parallel design84,85,162,184,235,246,249,261,341,18,45,231,391,513,559,575,583,585 including 1,357 participants, met the review inclusion criteria. Studies most commonly enrolled overweight patients and compared no intervention with a weekly schedule of three to five sessions of aerobic exercise. One study249 offered advice to participants whereas all others provided facilities. Three further studies could not be included because of missing data274,327,604.

The mean age of study participants was 53 years and 58% were male. Only five studies reported ethnicity and in these about 80% of the participants were white. The median duration of both intervention and follow-up was 17 weeks, ranging from eight weeks to one year.

Randomisation could be confirmed as adequate in only one study (6%), and concealment of allocation as adequate in none (0%). Blinding was confirmed as adequate in one study (6%). Treatment and control groups were confirmed as comparable at baseline, with regard to age, sex and initial blood pressure in 13 studies (76%).

Overall, patients receiving exercise-promoting interventions achieved a modest reduction in both systolic (3.1 mmHg, 95%CI: 0.7 to 5.5) and diastolic (1.8 mmHg, 95% CI: 0.2 to 3.5) blood pressure compared to those in control groups (see Figure 6). There was no evidence of reporting bias. Significant heterogeneity existed between studies, although there was no obvious underlying cause for this. There were not enough studies to explore the relative merits of weight training compared to aerobics or differences between low and medium intensity aerobics. Thirty-one percent (95% CI: 23% to 38%) of patients receiving exercise interventions were likely to show at least 10 mmHg reduction in systolic blood pressure. People in the exercise arms were more likely to withdraw from the studies than those in the control arms (treatment vs. control, risk difference: 5.9%, 95%CI: 0.1% to 11.1%), although studies varied.

Figure 6. Effect of exercise on systolic and diastolic blood pressure in randomised trials of patients with raised blood pressure.

Figure 6

Effect of exercise on systolic and diastolic blood pressure in randomised trials of patients with raised blood pressure. †DerSimonian-Laird Weighted Mean Difference Systolic BP: DL= −3.1 (95%CI: −5.5 to −0.7); Q:p = 0.007; (more...)

A recent systematic review of studies of the effect of exercise on blood pressure187 included seven studies between 1966 and 1995, all with at least 26 weeks follow-up, and including normotensive and hypertensive participants. The review found exercise had a small and statistically non-significant effect on blood pressure (−0.7/0.3 mmHg in 4 studies with hypertensive participants), but noted the poor quality of studies.

The recent Canadian guideline reviewed studies between 1966 and 1997132. Although without a formal meta-analysis, it reported short term reductions in blood pressure of 5 to 10 mmHg and recommended 50–60 minutes of moderate intensity exercise three or four times per week.

11.1.4. Relaxation therapies

Twenty-three randomised controlled trials of parallel design, including 1,481 participants, met the review inclusion criteria. RCTs of relaxation interventions32,33,31,34,69,95,115,120,142,221,265,276,277,289,304,367,397,477–479,525,533,610,661. Twelve further trials could not be included because of missing data128,232,245,345,398,586, 36,80,92,288,418.

The mean age of study participants was 49 years and 62% were male. Only six studies reported ethnicity and in these about 84% of the participants were white. The median duration of intervention was 8 weeks, ranging from four weeks to six months; the median duration of follow-up 17 weeks, ranging from eight weeks to four years, reflecting that studies often assessed the longer term impact of interventions well after formal therapy had ceased.

Randomisation could be confirmed as adequate in only seven studies (30%), and concealment of allocation as adequate in only one (4%). Blinding was confirmed as adequate in seven studies (30%). Treatment and control groups were confirmed as comparable at baseline, with regard to age, sex and initial blood pressure in 16 studies (70%).

The common component in studies was a strategy to promote relaxation although this could be oriented through education, physical techniques (such as breathing or progressive muscle relaxation), talk therapies, stress management or some combination. Additionally some studies used biofeedback, where the participant received auditory or visual information about their heart rate, peripheral temperature or some other physical marker. There was variation in content, with individual studies incorporating (for example) forms of cognitive training, breathing management, meditation, yoga, behavioural contracts, assertiveness training and anger control techniques. Similarly, delivery varied, being provided by a range of health professionals, most commonly to groups but in a few studies to individuals. Most treatment sessions were about an hour in length (varying from 30 to 90 minutes) and were usually conducted once a week.

Control groups received care varying from no intervention to sham group therapy excluding components that investigators believed to be the effective aspects of therapy. Some studies included both types of control groups.

Overall relaxation interventions were associated with statistically significant reductions in systolic (3.7 mmHg, 95%CI: 1.3 to 6.0) and diastolic (3.5 mmHg, 95%CI: 1.9 to 5.1) blood pressure (see Figure 7). There was no evidence of reporting bias. However, significant heterogeneity existed between studies. Analysis of the additional value of biofeedback as a component of the intervention was inconclusive when comparing studies that did or didn’t include it, or when comparing alternative interventions within trials. Thirty-three percent (95%CI: 25% to 40%) of patients receiving relaxation therapies were likely to show at least a 10 mmHg reduction in systolic blood pressure in the short term. Based on 12 of the studies, there was no significant difference in withdrawal when comparing treatment or control arms of studies (treatment vs. control, risk difference: 3.4%, 95%CI: 0.0% to 6.8%), although studies varied.

Figure 7. Impact of relaxation interventions on blood pressure: findings from randomised controlled trials.

Figure 7

Impact of relaxation interventions on blood pressure: findings from randomised controlled trials. †DerSimonian-Laird Weighted Mean Difference Systolic BP: DL= −3.7 (95%CI: −6.0 to −1.3); Q:p = <0.001; Size: p = (more...)

A recent systematic review of studies of the effect of stress reduction on blood pressure187 included seven studies between 1966 and 1995, all with at least 26 weeks follow-up, and including hypertensive participants. Although the inclusion criteria differed from ours, the review found a small and statistically non-significant effect on blood pressure (−1.0/−1.1 mmHg) consistent with longer follow-up studies reported here. The review similarly found considerable heterogeneity between studies.

The recent Canadian guideline reviewed studies between 1966 and 1997550. It concluded that multifaceted interventions to reduce stress were more likely to be effective than single component therapies and favoured the use of cognitive behavioural therapy, based on the findings of three meta-analyses192,293,366. For hypertensive patients in whom stress appears to be an important issue, they recommended that stress management including individualized cognitive behavioural therapy may be appropriate.

11.1.5. Multiple lifestyle interventions

Six randomised controlled trials, including 413 participants, met the review inclusion criteria. RCTs of multifaceted interventions45,47,84,294,337,337,408,599. Three of the studies essentially provided a therapeutic intervention combining group exercise and diet strategies similar to the lifestyle interventions found in the previous sections45,47,84,337, 599; one study also included relaxation and restriction of intake of common salt337; one study combined a weight loss diet, relaxation and salt restriction294; and one study combined a weight loss diet, exercise and salt restriction408. A further trial, which delivered a health education package to a British population with angina, did not meet our inclusion criteria for blood pressure and so was excluded from the meta-analysis and is considered separately146. Three further trials could not be included because of missing data274,309,334.

The mean age of participants was 52 years, 66% were male and the median follow-up of studies was six months. Five studies reported ethnicity and in these about 75% of the participants were white.

Randomisation was confirmed as adequate in only two studies (33%). Concealment of allocation was inadequate or unclear in all six studies. Blinding was confirmed as adequate in four studies (67%). Treatment and control groups were confirmed as comparable at baseline, with regard to age, sex and initial blood pressure in five studies (83%).

Overall, multifaceted interventions caused a modest reduction in both systolic (5.5, 95%CI: 2.3 to 8.8) and diastolic (4.5 mmHg, 95% CI: 2.0 to 6.9) blood pressure (see Figure 8). However heterogeneity existed between studies: the study of Jacob (1985) did not demonstrate a reduction in blood pressure. Twenty-six percent (95%CI: 2% to 49%) of patients receiving combined interventions were likely to show at least a 10 mmHg reduction in systolic blood pressure. Data from five studies found no statistically significant difference in withdrawal from treatment and control groups (treatment versus control, risk difference: 4.9%, 95%CI: −2.6% to 12.4%).

Figure 8. Impact of combined lifestyle interventions on blood pressure: findings from randomised controlled trials.

Figure 8

Impact of combined lifestyle interventions on blood pressure: findings from randomised controlled trials. †DerSimonian-Laird Weighted Mean Difference Systolic BP: DL= −5.5 (95% CI: −8.8 to −2.3); Q:p = 0.07; Size: p = (more...)

It was not possible to assess from the available data whether the effects of diet and exercise were additive or whether the combination was no better than either diet or exercise on its own.

The large British health promotion study, of 688 participants, lasted longer (two years) and was of older people (mean age 63 years) than the therapeutic studies. It did not show any reduction in blood pressure in response to health advice, but nevertheless reported fewer deaths among those receiving advice (29 in control group and 13 in treatment group), providing a relative reduction in mortality of 55%, an absolute reduction in mortality of 4.6% (95%CI: 1.0% to 8.4%) or a Number Needed to Treat of 22 to prevent a death during two years of follow-up. Patients in this trial, suffering from angina, were at higher risk than most other patients enrolled in lifestyle trials, leading to greater levels of morbidity and mortality. However, the benefit of health promotion in this trial does not appear mediated by reduced blood pressure or any other obvious prognostic marker (smoking, cholesterol or body mass index), and thus needs confirmation from further research.

A recent systematic review of studies of multiple interventions for preventing coronary heart disease; included nine studies of normotensive and hypertensive participants, published between 1966 and 1995, and with at least 26 weeks follow-up186. The review found an overall reduction of 4.2/2.7mmHg, but no significant reductions in morbidity and mortality in studies not including drug interventions.

11.1.6. Alcohol

The epidemiological link between alcohol consumption, blood pressure, cardiovascular disease and all-cause mortality has been studied extensively181,263,497,596. While moderate consumption may do no harm, the literature consistently finds that the move from moderate to excessive drinking (men: more than 21 units/week; women: more than 14 units/week) is associated both with raised blood pressure and a poorer prognosis. (Approximately: one half-pint of beer, glass of wine or a single measure of spirits equals one unit of alcohol or one standard drink and contains 8g or 10ml of alcohol287).

Three randomised controlled trials, including 397 participants, met the review inclusion criteria and examined the effect of changes in alcohol consumption on blood pressure148,382,502. Interventions varied in their content but commonly featured a number of visits to a health care practitioner for advice on reducing intake of alcohol. At baseline, patients typically reported drinking 300 to 600 ml of alcohol, or 30–60 standard drinks, per week. Although alcoholism was not formally defined, very heavy drinkers were commonly excluded. A further cluster randomized trial with 93 participants was identified and included in a secondary analysis348.

The mean age of study participants was 53 years; in the two studies that provided the details all participants were male and three quarters were white. The PATHS study148, with 6 months treatment duration, two year follow-up and 59% of patients, differed in scale from the two other shorter and smaller trials.

Randomisation could be confirmed as adequate only in the PATHS study, and concealment of allocation as adequate in none. Blinding was confirmed as adequate in two studies. Treatment and control groups were confirmed as comparable at baseline, with regard to age, sex and initial blood pressure in all three studies, with the possible exception of PATHS which did not report the proportions of men and women in the treatment and control groups. No studies were designed to assess the impact of alcohol reduction on cardiovascular endpoints.

Overall, interventions to reduce alcohol consumption caused small but statistically significant reductions in both systolic (3.4 mmHg, 95%CI: 0.9 to 6.0) and diastolic (3.4 mmHg, 95%CI: 1.5 to 5.4) blood pressure. Thirty percent (95%CI: 21% to 39%) of patients receiving a structured intervention to reduce alcohol consumption were likely to achieve a reduction of at least 10 mmHg in systolic blood pressure. No harmful effects of intervention were reported in these trials; withdrawal rates were reported in only one small trial. Inclusion of the single cluster randomized study did not alter qualitatively the summary reduction in systolic (3.7 mmHg, 95% CI: 1.3 to 6.1) or diastolic (3.2 mmHg, 95%CI: 1.4 to 5.0) blood pressure, (see Figure 9).

Figure 9. Impact of alcohol reduction on blood pressure: findings from randomised controlled trials.

Figure 9

Impact of alcohol reduction on blood pressure: findings from randomised controlled trials. †DerSimonian-Laird Weighted Mean Difference Systolic BP: DL= −3.7 (95% CI: −6.1 to −1.3); Q:p = 0.52; Size: p = 0.74

The recent Canadian guideline reviewed studies between 1966 and 1996113. Although without a formal meta-analysis, it recommended that alcohol consumption be limited in patients with hypertension to two or fewer standard drinks per day, with consumption not exceeding 14 standard drinks per week for men and nine standard drinks per week for women.

For recommendations on preventing the development of hazardous and harmful drinking, see NICE Public Health guidance 24 (http://guidance.nice.org.uk/PH24).

11.1.7. Coffee

Although coffee is a complex beverage containing many chemicals, only the effect of caffeine has been studied extensively516. According to personal taste and type of coffee, the amount of caffeine varies, but typically coffee contains 60 to 120 mg per 150ml cup. This can be compared with tea (20 to 40 mg per 150ml cup) and cola drinks (30 to 50 mg per 330ml can)444, 130.

Caffeine consumption has long being associated with raised blood pressure and can demonstrate a dose-related increase of 5–15 mmHg systolic and 5–10 mmHg diastolic for several hours following consumption. The most likely mode of action of caffeine is as an adenosine receptor antagonist, which results in vasoconstriction and raises blood pressure. The half life of caffeine in the body is typically about five hours297.

We identified no randomised controlled trials examining the impact of coffee or caffeine intake on patients with hypertension, which provided at least eight weeks follow-up. A published systematic review included normotensive as well as hypertensive participants, and shorter durations of follow-up299. Eleven trials with a total of 522 participants and a median duration of eight weeks (range 2 to 11 weeks) were included. Control groups drank a median of five caffeinated cups of coffee a day, with treatment groups receiving no, or decaffeinated, coffee. The reported overall effect of coffee was an increase in systolic (2.4 mmHg, 95%CI: 1.0 to 3.7) and diastolic (1.2 mmHg, 95%CI: 0.4 to 2.1) blood pressure.

Identifying the influence of coffee upon blood pressure, or identifying groups at particular risk, is problematic in the presence of confounding factors such as age, lifestyle, and cardiovascular disease. The small sample sizes and durations of existing trials do not provide an adequate evidence base to infer the long term effects of routine caffeine consumption.

11.1.8. Reducing sodium (salt) intake

Practical steps to reduce sodium intake include choosing low-salt foods (e.g. choosing fresh fruits and vegetables and avoiding processed foods) and reducing or substituting its use in cooking and seasoning. Much dietary salt comes from processed foods whose content should be labelled helping to monitor intake.

Five randomised controlled trials (four of parallel design125,212,311,544, one of crossover design10,11), examining the effect of sodium reduction on blood pressure, met the review inclusion criteria and included 420 patients. The findings of one Italian trial in young adults are considered separately141. A further trial could not be included because of missing data395.

The mean age of study participants was 52 years and 81% were male. The ethnicity of participants was not reported in any of the studies. The median duration of both intervention and follow-up was 12 weeks.

One trial (17%) was double-blinded; blinding could not be confirmed in any of the other studies. Randomisation and concealment of allocation could not be confirmed to be adequate in any of the studies. Treatment and control groups were confirmed as comparable at baseline, with regard to age, sex and initial blood pressure in 2 studies of parallel design (40%); the crossover study did not report on carryover effects.

The studies advised participants to change their diet so as to restrict their sodium intake to below 70–100 mmol/day (4.2 – 6.0g of salt). The Scientific Advisory Committee on Nutrition target for all adults is 6 grams/day532 and NICE public health guidance on the prevention of cardiovascular diseases recommends people aim for a maximum intake of 6 grams per day per adult by 2015 and 3 grams by 2025.

Average changes in blood pressure, when comparing treatment and control groups, are shown in Figure 10. Sodium reduction was associated with a statistically significant reductions in systolic (3.4 mmHg, 95%CI: 2.3 to 4.5) and diastolic (2.2 mmHg, 95%CI: 1.5 to 3.0) blood pressure. Twenty-three percent (95%CI: 17% to 30%) of patients who reduced their salt intake were likely to show at least a 10 mmHg reduction in systolic blood pressure. Based on two studies, there was no difference in withdrawal when comparing treatment and control arms of studies (treatment versus control, risk difference: −0.6%, 95%CI: −6.5% to 5.4%).

Figure 10. Impact of sodium reduction on blood pressure: findings from randomised controlled trials.

Figure 10

Impact of sodium reduction on blood pressure: findings from randomised controlled trials. †DerSimonian-Laird Weighted Mean Difference Systolic BP: DL= −3.4 (95% CI: −4.5 to −2.3); Q:p = 0.60; Size: p = 0.50

One Italian trial enrolled young, borderline hypertensive participants, aged 16–31 years. This trial found a dramatic reduction in systolic (18.4 mmHg, 95%CI: 10.1 to 26.7) blood pressure. The trial was poorly described and it is unclear whether the reduction in systolic blood pressure is due solely to the intervention. The authors note that the benefit was found mostly in participants less than 20 years of age. The inclusion of the trial in the meta-analysis increased the average benefit of salt reduction on systolic blood pressure (7.1 mmHg, 95%CI: 2.9 to 11.3), but introduced considerable statistical heterogeneity (Q: p=0.007).

Two recent systematic reviews have evaluated advice to reduce salt intake in normotensive and hypertensive adults, in trials with at least 6 months follow-up187,279. The inclusion criteria used in these reviews differ from ours, notably they included studies where the dose of antihypertensive drugs was allowed to vary. Regardless, both reviews found statistically significant reductions in blood pressure in studies with hypertensive participants, of 2.5/1.2 (up to one year follow-up) and 1.1/0.6 (one to six years follow-up)279 and 2.9/2.1 mmHg187, suggesting that reductions in blood pressure tend to diminish over time.

The recent Canadian guideline220, citing a previous systematic review, concluded that sodium restriction in adults over 44 years of age resulted in a reduction in blood pressure of 6.3/2.2 mmHg per 100 mmol/day reduction in sodium. Recommendations were made for clinicians to determine salt intake by interview; aim for a target range of 90–130 mmol per day (3–7 grams per day); provide advice on choosing low-salt foods (e.g. choosing fresh fruits and vegetables and avoiding pre-prepared foods) and reduce usage in cooking and seasoning.

11.1.9. Calcium supplements

Eleven randomised controlled trials (three of parallel design242,378,442, eight of crossover design227,318,396,571,581,584,627,660), examining the effect of calcium supplementation on blood pressure, met the review inclusion criteria and included 414 patients. Another trial, carried out in patients who were undergoing dialysis, was excluded after consideration of their unusual calcium metabolism but its details are tabulated487. A further trial could not be included because of missing data414.

The mean age of study participants was 45 years and 68% were male. Only four studies reported ethnicity and in these 46% of the participants were white. The median duration of both intervention and follow-up was eight weeks.

Randomisation could be confirmed as adequate in only two studies (18%) and concealment of allocation as adequate in only one (9%); nine studies (82%) studies were double-blinded treatment and control groups were confirmed as comparable at baseline, with regard to age, sex and initial blood pressure in one study (33%) of parallel design; three studies (37%) of crossover design confirmed no carryover effect.

The intervention was provided as a simple oral supplement taken several times a day.

Average changes in blood pressure, when comparing treatment and control groups, are shown in Figure 11. Calcium supplementation was associated with a small reduction in systolic blood pressure 2.3 mmHg, 95%CI: 0.3 to 4.4) which was statistically significant but not robust to minor changes in the reported blood pressure of the participants, and no difference in diastolic blood pressure (−0.8 mmHg, 95%CI: −2.1 to 0.6). No harmful effects of intervention were reported in these trials; withdrawal rates were on average around 10% in both treatment and control groups. The trials were unable to identify sub-groups of patients that might benefit from calcium.

Figure 11. Impact of calcium supplementation on blood pressure: findings from randomised controlled trials.

Figure 11

Impact of calcium supplementation on blood pressure: findings from randomised controlled trials. †DerSimonian-Laird Weighted Mean Deference Systolic BP: DL= −2.3 (95% CI: −4.4 to −0.3); Q:p = 0.03; Size: p = 0.82

11.1.10. Magnesium supplements

Eleven randomised controlled trials (nine of parallel design215,270,365, 91,443,475,621,646,659] 2 of crossover design [317,645), examining the effect of magnesium supplementation on blood pressure, met the review inclusion criteria and included 504 patients.

The mean age of study participants was 55 years and 44% were male. Only two studies reported ethnicity and in these 11% of the participants were white. The median duration of both intervention and follow-up was 12 weeks.

Ten studies (91%) studies were single or double blinded. Randomisation and concealment of allocation were confirmed to be adequate in one study (9%) and no studies respectively. Treatment and control groups were confirmed as comparable at baseline, with regard to age, sex and initial blood pressure in six studies (67%) of parallel design; neither of the studies of crossover design reported on carryover effects.

The intervention was provided as a simple oral supplement taken several times a day.

Average changes in blood pressure, when comparing treatment and control groups, are shown in Figure 12. Magnesium supplementation was associated with little change in systolic (−1.0 mmHg, 95%CI: −4.1 to 2.1) but a statistically significant reduction in diastolic (−2.1 mmHg, 95%CI: −3.5 to −0.7) blood pressure. No harmful effects of intervention were reported in these trials; withdrawal rates were reported in only eight studies, where these were on average around 7% in both treatment and control groups. The trials were unable to identify sub-groups of patients that might benefit from magnesium.

Figure 12. Impact of magnesium supplementation on blood pressure: findings from randomised controlled trials.

Figure 12

Impact of magnesium supplementation on blood pressure: findings from randomised controlled trials. †DerSimonian-Laird Weighted Mean Difference Systolic BP: DL= −1.0 (95%CI: −4.1 to 2.1); Q:p = 0.001; Size: p = 0.16

11.1.11. Potassium supplementation

Five randomised controlled trials (four of parallel design107,543,543, 578, one of crossover design470), examining the effect of potassium supplementation on blood pressure, met the review inclusion criteria and included 410 patients. The findings of one African trial are considered separately455. A further trial could not be included because of missing data149.

The mean age of study participants was 51 years and 76% were male. Only one study reported ethnicity and in this 86% of the participants were white. The median duration of both intervention and follow-up was 12 weeks.

Two studies were triple blinded, two were assessment blinded and one was unclear. Randomisation and concealment of allocation were confirmed to be adequate in one (20%) and two (40%) studies respectively. Treatment and control groups were confirmed as comparable at baseline, with regard to age, sex and initial blood pressure in two studies (50%) of parallel design; the crossover study did not report on carryover effects.

The intervention was provided as a simple oral supplement taken several times a day in all but one trial, where dietary advice was provided to increase intake of foods rich in potassium125.

Average changes in blood pressure, when comparing treatment and control groups, are shown in Figure 13. Potassium supplementation was not associated with any significant change in systolic (−3.5 mmHg, 95%CI: −7.9 to 0.9) or diastolic (−0.7 mmHg, 95%CI: −4.9 to 3.6) blood pressure. The findings of the studies were heterogeneous and there are no obvious reasons for this that can be deduced from the limited available evidence. No harmful effects of intervention were reported in these trials; average withdrawal rates of 6–8% were similar in both treatment and control groups.

Figure 13. Impact of potassium supplementation on blood pressure: findings from randomised controlled trials.

Figure 13

Impact of potassium supplementation on blood pressure: findings from randomised controlled trials. †DerSimonian-Laird Weighted Mean Difference Systolic BP: DL= −3.5 (95% CI: −7.9 to 0.9): Q:p = 0.02; Size: p = 0.80

One trial, which enrolled treatment naïve and hypertensive Kenyan participants (DBP 90–109 mmHg and SBP>160 mmHg) reported an average reduction of 39/17 mmHg. Although the effect of various salts upon certain ethnic groups is known to vary, a reduction of this magnitude exceeds our understanding and requires confirmation from further independent research.

A meta-analysis by Whelton and colleagues found that oral potassium supplementation was associated with a significant reduction in both systolic blood pressure and diastolic blood pressure633, based on 12 trials in normotensive people and 21 in hypertensive people, with a duration ranging from four days to three years (median five weeks). The review found that the blood pressure lowering effect was greater in hypertensive than normotensive people, although the statistical significance of findings in the hypertensive subgroup is not reported. The review also found that the effect was more pronounced in people eating a diet high in sodium chloride (common salt) and therefore recommended potassium supplementation for both prevention and treatment of hypertension, especially in people unable to reduce their intake of sodium.

In contrast, our restriction to trials of at least 8 weeks duration, enrolling only hypertensive patients, resulted in inclusion of only 5 trials with a median duration of 12 weeks and found that the blood pressure lowering effect of oral potassium supplementation was not statistically significant. The group concluded that there is not sufficient relevant evidence to recommend oral potassium supplementation for hypertension.

11.1.12. Combined salt supplements

Two randomised controlled trials studied combinations of the potassium, magnesium, sodium and calcium salts considered individually in previous sections.

One study used paired supplements comparing two of calcium, potassium and magnesium with placebo519. None of the combined supplements reduced blood pressure when compared with placebo (see Figure 14). This was consistent with the findings for the individual supplements.

Figure 14. Impact of combined supplements on blood pressure: findings from randomised controlled trials.

Figure 14

Impact of combined supplements on blood pressure: findings from randomised controlled trials. +F/U Duration of follow up in weeks, months or years; and N: Number randomised Calcium (Ca), Magnesium (Mg), Potassium (K) and Sodium (Na) salts

A second study compared a mineral (reduced sodium) salt containing sodium, potassium and magnesium with common sodium table salt. The mineral salt was used in prepared food as well as for seasoning229. The reduction of blood pressure by about 5/4 mmHg consistent with that found with strategies to reduce sodium salt intake.

The recent Canadian guideline reviewed studies between 1966 and 1996108. Although without a formal meta-analysis, it recommended against supplementing calcium, magnesium or potassium intake amongst hypertensive participants above the recommended normal daily levels.

11.1.13. Drug therapy versus lifestyle change

Five small randomised controlled trials enrolling 233 patients directly compared the effects of lifestyle interventions and drugs for the treatment of mild to moderate hypertension. Goldstein et al 232, Murugesan et al 418, Kostis et al 337, MacMahon et al 380, 381, Koopman et al 333. An additional quasi-randomised trial, which allocated participants to treatments on the basis of their birth date rather than at random, was also considered (Berglund et al72).

All trials were small (between 38 and 66 participants), of short duration (between eight and 52 weeks) and were not designed to assess cardiovascular endpoints. Randomisation and concealment of allocation were either inadequate or not clearly reported in all trials. The outcome assessor was blinded to the treatment status of the participants in three trials333,337,380; blinding was not reported in two trials232,418, and there was no blinding in one trial72. One trial was poorly reported and did not state the total number of participants418. In two trials the confidence intervals on the effects of treatment could not be estimated, as either the numbers in each treatment group418 or the standard error of the treatment effects were not reported232.

The populations studied in the trials differed in: (i) age – participants in one trial333 were older, which probably accounted for their higher baseline blood pressure compared to participants in the other trials; (ii) treatment status at the point of recruitment – participants were currently untreated or treatment naïve in four trials72,232,333,380, currently treated in one trial337, or treatment status was not reported418.

The trials compared different drugs with different lifestyle interventions. Typically either a diuretic or a beta-blocker was the class of drug used, although one trial allowed a choice of drugs. Four trials used a low calorie diet: one used diet alone; one combined a low calorie intake with a low sodium and high potassium diet; one used a multiple intervention combining weight loss, a low calorie and low sodium diet, exercise, and relaxation and one combined weight reduction with restricted sodium and alcohol intake. Two trials had relaxation interventions: one considered two separate relaxation interventions (biofeedback and muscular relaxation/breathing exercises); the other used yoga.

Five trials reported comparable blood pressure at baseline in both treatment groups and for one trial this was unclear. Within each study, findings for systolic and diastolic blood pressure were similar.

Trials comparing diet with drugs provided conflicting evidence (see Figure 15). In the trial of older participants333 who had not received treatment before and had a high baseline blood pressure, drug treatment appears more effective than diet in lowering blood pressure, whereas in a trial of younger participants381 who were currently untreated and had a lower initial blood pressure, diet appears significantly more effective than drug treatment in lowering blood pressure. The one trial337 comparing multiple lifestyle interventions with drugs found both treatments had similar effects on lowering blood pressure. Two trials found drugs to be more effective than relaxation although the confidence intervals on the treatment effects could not be determined418.

Figure 15. Comparison of lifestyle and drug interventions: findings from randomised controlled trials.

Figure 15

Comparison of lifestyle and drug interventions: findings from randomised controlled trials. +F/U: Duration of follow up in months, and N: Number randomised Components: alcohol restriction; biofeedback; diet; exercise; potassium enhancement;

Participants receiving dietary interventions improved their total cholesterol profiles in all four trials compared to participants receiving drugs. Cholesterol levels were not reported in either relaxation trial. Although it was a post hoc exercise, we combined cholesterol reductions found in the dietary trials by imputing missing standard deviations. Using a random effects model, the average reduction in cholesterol was 0.52 mmol/l (95% CI −0.34 to −0.7).

Withdrawals were reported in five trials: rates of withdrawal were similar for lifestyle and drug treatments.

The current evidence cannot determine whether a lifestyle intervention is generally better than drug treatment for reducing blood pressure. Although cholesterol levels were not a prespecified outcome, it was observed that, in all four trials with diet interventions, diets were better than antihypertensive drugs at reducing cholesterol. As reduced cholesterol levels are likely to lower the risk of cardiovascular morbidity or mortality irrespective of any change in blood pressure643, a healthier diet may reduce, delay or remove the need for long-term drug therapy in some patients. Thus it seems important that patients are encouraged to try lifestyle changes before proceeding to or increasing drug therapy.

11.1.14. Smoking cessation

A review of the health consequences of smoking and benefit of smoking cessation is not included in this guideline, since there is no direct link to raised blood pressure. However smoking reduces life expectancy and is associated with poor cardiovascular and pulmonary outcomes179,180,357,410,488,648. The NHS website www.smokefree.nhs.uk has facts and information about giving up smoking.

Refer to NICE’s public health guidance on smoking cessation services in primary care, pharmacies, local authorities and workplaces, particularly for manual working groups, pregnant women and hard to reach communities for more information (www.guidance.nice.org.uk/PH10).

11.1.15. Recommendations

For NICE guidance on the prevention of obesity and cardiovascular disease see ‘Obesity’ (NICE clinical guideline 43, 2006) and ‘Prevention of cardiovascular disease at population level’ (NICE public health guidance 25, 2010).

30.

Lifestyle advice should be offered initially and then periodically to people undergoing assessment or treatment for hypertension. [2004]

31.

Ascertain people’s diet and exercise patterns because a healthy diet and regular exercise can reduce blood pressure. Offer appropriate guidance and written or audiovisual materials to promote lifestyle changes. [2004]

32.

Relaxation therapies can reduce blood pressure and people may wish to pursue these as part of their treatment. However, routine provision by primary care teams is not currently recommended. [2004]

33.

Ascertain people’s alcohol consumption and encourage a reduced intake if they drink excessively, because this can reduce blood pressure and has broader health benefits. [2004]

34.

Discourage excessive consumption of coffee and other caffeine-rich products.

35.

Encourage people to keep their dietary sodium intake low, either by reducing or substituting sodium salt, as this can reduce blood pressure.[2004]

36.

Do not offer calcium, magnesium or potassium supplements as a method for reducing blood pressure. [2004]

37.

The best current evidence does not show that combinations of potassium, magnesium and calcium supplements reduce blood pressure. [2004]

38.

Offer advice and help to smokers to stop smoking. [2004]

39.

A common aspect of studies for motivating lifestyle change is the use of group working. Inform people about local initiatives by, for example, healthcare teams or patient organisations that provide support and promote healthy lifestyle change. [2004]

Copyright © 2011, National Clinical Guideline Centre.

Apart from any fair dealing for the purposes of research or private study, criticism or review, as permitted under the Copyright, Designs and Patents Act, 1988, no part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior written permission of the publisher or, in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK. Enquiries concerning reproduction outside the terms stated here should be sent to the publisher at the UK address printed on this page.

The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore for general use.

The rights of National Clinical Guideline Centre to be identified as Author of this work have been asserted by them in accordance with the Copyright, Designs and Patents Act, 1988.

Bookshelf ID: NBK83262

Views

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...