Dietary Patterns and Blood Pressure in Adults: A Systematic Review and Meta-Analysis of Randomized Controlled Trials^1,2

Introduction

Hypertension or high blood pressure (BP)⁵ is a major risk factor for developing cardiovascular disease (CVD), stroke, and kidney disease, and it affects ∼1 billion people globally (1). In 2001, high BP was said to cause ∼54% of strokes, 47% of ischemic heart disease, 75% of hypertensive disease, and 25% of other CVDs, leading to ∼7.6 million deaths worldwide (2). It has been estimated that lowering systolic BP (SBP) and diastolic BP (DBP) in the US population by 5.0 and 3.0 mm Hg, respectively, would result in a 15% reduction in the incidence of coronary artery disease and a 27% reduction in stroke incidence (3). Conversely, an increase of 20 mm Hg in SBP and 10 mm Hg in DBP in adults aged 40–70 y has been estimated to double the risk of developing CVD (1).

There are various approaches to dealing with high BP in the community. The report from the Eighth Joint National Committee on recommendations for the management of high BP (4) endorses the lifestyle changes recommended by the Lifestyle Work Group (5). These include weight loss for those who are overweight or obese, engaging in regular physical activity, following a healthy eating plan such as the Dietary Approaches to Stop Hypertension (DASH) diet, and reducing dietary sodium intake.

The question of the relative effects of whole lifestyles, dietary patterns, and single nutrients drives a number of areas of research. Numerous studies, including randomized controlled trials, have investigated the effect of single nutrients on BP. For example, a recent meta-analysis demonstrated that a reduction of salt (sodium chloride) intake to a mean of 4.4 g/d (1716 mg sodium/d) led to a reduction in BP by 5/3 mm Hg in hypertensive subjects and 2/1 mm Hg in normotensive subjects (6). Similarly, a large meta-analysis involving 29 randomized clinical trials showed that increased potassium intake of ≥20 mmol/d (≥780 mg/d) led to a reduction by 4.9 mm Hg and 2.7 mm Hg in SBP and DBP, respectively, without the use of antihypertensive medication (7). Although investigating the effect of single nutrients on BP may be informative, these nutrients are delivered in food, and food is usually consumed as whole diets. In addition, multiple interactions may occur between components of whole diets, so there is value in assessing the effects of the entire dietary pattern (8). Furthermore, dietary patterns may vary between different cultural environments because of the differences in types of foods consumed (9).

In addition to studying single nutrients, various randomized controlled trials and observational studies have investigated the effect of individual foods on BP with conflicting results. These include fruit and vegetables (10–12), dairy and eggs (10, 13–15), nuts (16), pulses (17), meat (11, 13, 18), chocolate (19), tea (20), coffee (21), red wine (22), and fats and oils (23). Alcohol intake is associated with an increased incidence of hypertension (24), although the magnitude of the effect depends on the timing of BP measurement after intake, with peak levels registered after ∼10 h (25). Inconsistencies in single food-based analyses provide support to the argument that it is the context in which the food is eaten that may be important in better understanding diet-BP associations.

The effects of entire dietary patterns on BP have best been demonstrated by following the DASH diet (3) and the Nordic diet (26). Both of these dietary patterns emphasize consumption of different combinations of healthy foods to lower BP. Otherwise, there are little data on which types of dietary patterns are effective in lowering BP in hypertensive and normotensive adults. The aim of this review was to assess the effect of dietary patterns on BP in adults.

Methods

Results

Discussion

This meta-analysis of 17 randomized controlled trials evaluated the current evidence on the effect of dietary patterns on BP in adults. A significant reduction of 4.26 mm Hg in SBP and 2.38 mm Hg in DBP was observed in the present analysis. For the studies that had no weight loss, increased exercise, or sodium restrictions, SBP and DBP reduced by 4.25 and 2.27 mm Hg, respectively. A previous meta-analysis found a reduction of 6.74 and 3.59 mm Hg in SBP and DBP, respectively, from the DASH diet only (51). Heterogeneity was substantial for both SBP and DBP. Many differences existed between the populations studied in terms of age, sex, study methods, duration of intervention, number of participants, and difference in combination of foods included within the various dietary patterns. The existing evidence suggests that various dietary patterns are beneficial for BP and include the DASH diet, MD, and Nordic diet.

The DASH diet largely comprises of fruit and vegetables, low-fat dairy, whole grains, nuts, legumes, and seeds, with low intakes of meat and saturated fat (31). The DASH diet has been adopted in various cultural contexts such as in Brazil (40), Australia (43, 45), and Iran (48). Our results are consistent with an earlier analysis of BP and nutrient intakes in the United States, which showed that reduced consumption of dairy products and fruit and vegetable juices was a major predictor of hypertension (52). In observational studies, a high DASH score in the SU.VI.MAX (Supplementation en Vitamines et Mineraux Antioxydants) cohort was associated with lower BP at the first clinical examination and a lower increase in BP after 5 y (10). Although the DASH diet shows positive effects from the randomized controlled trials, the studies are short term in nature, and this may limit their generalization as a long-term intervention.

The Nordic diet is consumed in Nordic countries and consists of foods of Nordic origin such as whole grains, rapeseed oil, berries, fruits, vegetables, fish, nuts, and low-fat dairy products. This diet significantly reduced 24-h ambulatory DBP in comparison to a diet comprising the mean nutrient intake in Nordic countries (26). However, the effects of the diet may not be attributed to reductions in sodium or increases in potassium because neither of these electrolytes differed between intervention and control groups. One of the characteristics of the Nordic diet is that it is rich in berries. Animal studies have shown that Nordic wild blueberries lead to a reduction in BP (53), and likewise, randomized controlled trials have demonstrated that consumption of berries lowers BP (54, 55). The polyphenols, especially flavonoids, found in berries may be contributing to the BP-lowering effect of Nordic diets that typically include these fruits (56). Further research, however, is warranted on the effect of the Nordic diet on BP because only 3 studies were identified from our review.

Another dietary pattern, the MD, is generally high in plant foods such as whole-grain cereals, fresh fruits, vegetables, beans, nuts, and seeds (57, 58). It can have moderate amounts of dairy foods, fish, and poultry and low amounts of red meat, but there may be some variation in food composition between regions. In general, olive oil is the major source of fat, and wine is consumed moderately with meals. In the Prevencion con Dieta Mediterranea (PREDIMED) study, the MD in Spain was supplemented with extra virgin olive oil and with nuts, and significantly lower DBP was observed compared with a low-fat diet (59). Likewise, analysis of ambulatory BP in 235 participants of the PREDIMED study after 1 y showed significant reductions in ambulatory SBP by 4.0 and 4.3 mm Hg in the MD supplemented with extra virgin olive oil and nuts, respectively, and 1.9 mm Hg in DBP for both diets (47). These effects were also seen in studies of the MD in Italy, where significantly lower BP was observed (38, 39). However, no effect was found in one study from France (46) and the United States (37). In addition to possible differences in food and recipe composition, the study duration for these trials was <6 mo. Because diet effects tend to occur over longer periods of time, a long-term follow-up may be necessary to detect BP-lowering effects.

In comparison to randomized controlled trials, observational studies do not have control over dietary patterns, so not surprisingly, results from observational studies show more conflicting results for the MD. For example, an increased risk of developing hypertension (HR: 1.34; 95% CI: 1.04, 1.73; P = 0.002) was identified in adults with a higher adherence to the MD in Spain as identified through an updated modified MD score (60), but in the United States, an analysis of the Framingham Heart cohort found no association between the MD and BP after 7 y of follow-up in 2370 adults (61). In this case, and in comparison to the DASH diet, the analysis of the MD pattern did not distinguish between full-fat and low-fat dairy. Because low-fat dairy was used in the DASH diet, it has been assumed as contributing to the BP-lowering effect. Likewise, many of the randomized controlled trials chose to use low-fat dairy in their interventions and did not compare with full-fat dairy. Observational studies such as the Coronary Artery Risk Development in Young Adults study (13) did not distinguish between low- and full-fat dairy. Similarly, the analysis of nutrient intakes and BP in the US population did not discern any difference between low- and full-fat dairy (52). Further research is therefore warranted to investigate the effect of low-fat dairy compared with full-fat dairy on BP.

The research on the Tibetan diet was conducted in Germany and emphasized cereals from barley, wheat, rye, corn, rice, oat and buckwheat, and meat such as beef, mutton, roast hare, chicken, and venison (49). Compared with the usual Western diet in a 12-mo clinical trial, there was no significant change in BP between the 2 diets. Only one study on the Tibetan diet was identified, and this means that additional research is required before conclusions on its efficacy can be drawn.

In some studies, the amount of sodium was kept constant in both the intervention and control diets. In the original DASH study, sodium content was maintained at 3000 mg/d in all groups (31), whereas the Brazilian DASH diet was based on an intake of 2400 mg/d (40). In the original DASH diet study, a significant reduction in BP was found even when sodium intake was kept constant (31). When sodium reduction was added to the DASH diet (32), reduced BP was still seen at all sodium amounts, with a greater reduction in the high sodium amount. This suggests that regardless of the sodium content, the food pattern reflective of the DASH diet lowers BP. Sodium intake was also reduced in one of the Nordic diet studies (26) to <2730 mg/d and 2340 mg/d in men and women, respectively, making it difficult to separate out effects. These findings are consistent with recent studies that did not find a significant association between the amount of sodium intake and BP in normotensive populations (62, 63).

Various other nutrients characteristic of the DASH diet such as fiber (64), potassium (7), calcium (65), and magnesium (66) have been associated with reduced BP, but the findings are inconsistent. It is well known that nutrients have multiple interactions, for example, interactions between sodium, potassium, and calcium through either the sodium-to-potassium ratio or the BP-lowering effect of potassium and calcium due to their ability to increase sodium excretion by the kidneys (67). This understanding highlights the importance of considering the synergistic effects of whole foods and diets in examining effects.

Finally, the interface between the adoption of healthy diets and other behavioral lifestyle changes cannot be ignored. Our results indicate that cointerventions such as sodium restriction, exercise, and weight loss also led to significant BP reductions. However, a range of combinations of interventions may be effective for the management of BP, and further research on relative efficacy would be informative.

Various limitations in this review warrant discussion. It should be noted that the nature of intervention differed between studies, whereby some trials provided foods to the participants, whereas others did not. This can influence the amount of adherence to the intervention diet. Also, although the dietary patterns identified may have had some common basic characteristics, they were not homogeneous, and each diet was set in a unique cultural context. The use of BP as an endpoint for sodium intake has limitations. Recent studies using mortality or occurrence of CVD events as the composite outcomes for sodium intake have shown that lower (<7 g salt/d; 2730 mg sodium/d) or higher (>14 g salt/d; 5460 mg sodium/d) intakes are associated with increased risk of death and cardiovascular events (68, 69). In addition, the 2013 Institute of Medicine report (70) pointed out that higher sodium intakes are associated with the risk of CVD, and there is insufficient evidence on the benefit or risk of intake <2300 mg/d. The report did not assess the association of sodium intake with intermediate endpoints such as BP. In this review, we considered both studies that restricted sodium intake and those that did not in our subgroup analysis. The focus of our review, however, was the entire dietary pattern rather than sodium intake alone. Finally, there were limitations associated with individual studies used in the meta-analysis. Biases were identified, especially as-treated data analysis in 8 studies, and 2 studies that did not find a significant effect of the MD were not included. These features may have affected the overall results.

Conclusion

The consumption of dietary patterns characterized by high consumption of fruit, vegetables, whole grains, legumes, seeds, nuts, fish, and dairy and low consumption of meat, sweets, and alcohol resulted in significant reductions in BP. These reductions were found across a range of sodium intakes. Following a healthy lifestyle by incorporating exercise in addition to a healthy diet may also have an additional effect in lowering BP. Much of the population-level efforts to reduce BP are through salt reduction in the food supply. This meta-analysis supports the notion that dietary diversification may also play a role. The association between the consumption of the MD and BP warrants further investigation due to inconsistency in findings between studies. This research should consider an analysis of the effect of dietary patterns on BP in different cultures other than those identified in this review.