Empirically Derived Dietary Patterns and Hypertension Likelihood: A Meta-Analysis Open Access

Hypertension is one the most common chronic diseases, affecting about 25% of the adult population worldwide [1]. In China, the prevalence of hypertension has increased by almost one third from 1991 to 2002, which comprises approximately 200 million people [2,3]. In United States, hypertension is also a significant health problem, with an estimated one-third of the general population having hypertension [4]. It is well-known that hypertension is a multifactorial disease that results from complex interactions of some risk factors, including alcohol drinking, high salt, genetic factors, and dietary factors [5,6,7].

In past several decades, some epidemiological studies have particularly focused on dietary modification as an important risk factor in the development of hypertension, and examined the relationship between the intakes of individual foods or nutrients and the risk of hypertension [8,9,10,11,12]. However, because of the complexity of diet and the potentially synergistic effect of individual foods or nutrients on health, these analyses revealed limited impact of diets on hypertension [13]. Consequently, dietary pattern analysis has emerged as an alternative approach in examining the association between diet and diseases, considering the combined effects of foods and potentially facilitate nutritional recommendations [14].

Recently, several studies have indicated that there is a growing interest in medical research on the associations between overall dietary patterns and hypertension likelihood [15,16,17,18,19]. However, the findings from these studies remain inconsistent. Although some previous studies reported a positive association between healthy/prudent pattern and hypertension [15,16], other also reported an inverse association between this pattern and hypertension [18,19]. Therefore, we conducted this meta-analysis of studies published from January 1990 up to March 2016, to identify the potential relation between dietary patterns and the likelihood of hypertension.

An electronic literature search was performed in the MEDLINE and EBSCO databases to identify human studies written in the English and Chinese languages published from January 1990 up to March 2016, with the following keywords or phrases: “dietary pattern” OR “dietary patterns” OR “food pattern” OR “food patterns” OR “eating pattern” OR “eating patterns” OR “alcohol drinking” OR “alcohol consumption” AND “hypertension” OR “high blood pressure”. Furthermore, we manually reviewed the reference lists of relevant studies to search for additional studies.

Two reviewers (CJ Wang and Y Liu) independently read the abstracts of articles retrieved in the initial search to identify studies that examined the associations between dietary patterns and odds of hypertension. Differences between the two independent reviewers were resolved by consensus and referred to the third reviewer if necessary. When all agreed (CJ Wang, YX Shen and Y Liu), the full-text versions of articles were reviewed against inclusion criteria for this meta-analysis. To be eligible, a study must meet the following criteria: (1) It was an original report on the associations between dietary patterns and the likelihood of hypertension; (2)Dietary pattern was identified by factor analysis and/or principal component analysis; (3)Odds ratios, hazards ratio or relative risks and percentage of hypertension (or sufficient information to calculate them) had been presented; (4) If the data in original publication lacked sufficient detail, the corresponding author of the study was contacted for additional information by email.

To minimize error, the authors ensured that the selected dietary patterns were similar with regard to factor loadings of foods, which are consumed within those dietary patterns. For example, the healthy pattern is characterized to have high loadings of foods such as vegetables and fruit, whole grain, fish, soy and antioxidants(e.g. Vitamin C and E, flavonoids, and carotenoids) and low fat dairy. The western-style pattern is characterized to have high consumption of red and/or processed meat, refined grains, sweets, high-fat dairy products, butter, potatoes and high-fat gravy, and low intakes of fruits and vegetables. The heavy drinking pattern tend to have high loadings of beers, wines, and white spirits. The light-moderate drinking pattern is characterized to have moderate consumption of beers, wines, and white spirits. Finally, twenty-seven studies relevant to the role of dietary patterns and/or food and the odds of hypertension were included in this meta-analysis.

Data was extracted from all eligible studies, including the first author, the year of publication, country, study design, sample size, number of hypertension, dietary assessment method, identification of dietary patterns and the adjusted factors in the present analysis.

The different forms of alcohol consumption were converted into grams of ethanol per day. Alcohol consumption of >25g/d for men or >12.5 g/d for women was defined as a heavy drinking.alcohol consumption of <12.5g/d for men or <7.5g/d for women was defined as a low drinking, and alcohol consumption of >12.5 g/d and <25g/d for men or >7.5g/d and <12.5 g/d for women was defined as a light-moderate drinking [20].

The reviewers independently assessed the risk of bias using the Newcastle-Ottawa Quality Assessment scale for studies included in this meta-analysis [21]. A maximum of nine points was assigned to each study. Only these studies which the majority of the questions were deemed satisfactory (i.e. with a score of 6 or higher) were considered to be of high methodological quality.

All statistical analyses were performed using Review Manager, version 5.0 (Nordic Cochrane Center, Copenhagen, Denmark) and STATA software. version 12.0 (STATA Corp, College Station, TX, USA). P values less than 0.05 were considered significant, and statistical tests were two-sided. The original papers reported the results of dietary patterns in terms of tertiles, quartiles, and quintiles of dietary factor scores and the odds of hypertension. We used meta-analysis to evaluate the likelihood of hypertension in the highest versus the lowest categories of healthy, western-style and drinking patterns. Heterogeneity of study results was measured by Cochran's Q statistic and I² statistic. A P value of Q-test >0.10 or I²<50% indicated an absence of heterogeneity between studies, and a fixed-effects model was used to calculated the pooled ORs. If a P value of Q-test ≤0.10 or I²≥50% indicated a high degree of heterogeneity among studies, then a random-effects model (DerSimonnian and Laird method) was used [22]. OR and 95% CIs from individual studies were combined to produce an overall OR. Publication bias was assessed by inspection of the funnel plot and by formal testing for “funnel plot” asymmetry using Begg's test and Egger's test [23]. Subgroup analyses were carried out based on age(>50 and <50y), sample size(>5000 and <5000), race(white, and yellow and other) and study design (cross-sectional and cohort).

An electronic search in the databases of MEDLINE and EBSCO identified 894 papers, 867 of which were excluded based on the following reasons (in Figure 1). Twenty-seven articles [9,13,15,16,17,18,19,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43] met the inclusion criteria and were included in this meta-analysis, including16 cohort studies [13,25,27,30,31,32,34,35,36,37,38,39,40,41,42,43] and 11 cross-sectional studies [9,15,16,17,18,19,24,26,28,29,33]. Descriptive information of each included study was presented in Table 1.

The healthy pattern is characterized to have high loadings of foods such as vegetables, fruits, whole grains, olive oil, fish, soy, poultry and low fat dairy. There was obvious evidence of a decreased likelihood for hypertension in the highest compared with the lowest categories of healthy pattern in Figure 2 (OR=0.81; 95%CI: 0.67-0.97; P=0.02). A random-effects model is used to assess the including data. The heterogeneity was apparent in all the studies (P<0.00001; I²=94%).

The western-style pattern is characterized to have high intake of red and/or processed meat, refined grains, sweets, high-fat dairy products, butter, potatoes and high-fat gravy, and low intakes of fruits and vegetables. Figure 3 shows the forest plot for the likelihood of hypertension in the highest compared with the lowest category of western-style pattern. There was significant heterogeneity (I²=96%, P<0.00001) and hence the effect was assessed using the the random-effects model. The results demonstrated that western-style pattern was not associated with an increased risk of hypertension (OR=1.04; 95%CI: 0.83-1.31; P=0.74).

The drinking pattern is characterized by high consumption of alcohol containing beers, wines, and white spirits. Nine articles identified a heavy drinking pattern and six articles identified a light-moderate drinking pattern in this meta-analysis. An increased likelihood of hypertension was shown for the heavy drinking compared with non-drinking of heavy drinking pattern in Figure 4 (OR=1.62; 95%CI:1.16-2.26; P=0.004). Data from these studies were assessed using a random-effects model, and there was obvious evidence of heterogeneity (P<0.00001; I²=96%). Figure 5 showed no association between light-moderate drinking pattern and hypertension (OR=1.20, 95% CI:0.94-1.53, P=0.13).The data from six studies were assessed using the random-effects model, and there was significant heterogeneity (I²= 93%, P<0.00001).

To further detect the association between dietary patterns and hypertension, a subgroup analysis was performed. Age, sample size, race and study design had an effect on hypertension and dietary patterns. When the highest categories were compared with the lowest categories of healthy pattern, the inverse association between hypertension and dietary patterns was more obvious in these studies which were in cross-sectional studies and subjects were white. In contrast, when comparing the highest categories with the lowest categories in western-style patterns, the positive association between hypertension and dietary patterns was more obvious in which subjects were white. In addition, the positive association was obvious for those in the high alcohol consumption compared with non-alcohol consumption of heavy drinking pattern in studies which had large sample size and was cohort design, and subjects aged more than 50 years and white. Finally, when the results were analyzed by removing white subjects, the difference in the odds of hypertension was shown in light-moderate drinking pattern. As these variables have a strong effect on association between dietary patterns and odds of hypertension, their differences may partially explain the observed heterogeneity between studies (Table 2-Table 3).

Inspection of funnel plots revealed little evidence of asymmetry. Egger's tests for publication bias was not statistically significant(highest compared with lowest categories: healthy P=0.521; Western-style P=0.261; heavy drinking P=0.885; light-moderate drinking P=0.925.).

The quality of each included study in terms of population and sampling methods, description of exposure and outcomes, and statistical adjustment of data, is summarized in Table 4. All studies received a score of 6 or higher on the Newcastle-Ottawa Quality assessment scale and were considered to be of high methodological quality [9,13,15,16,18,19,24,25,26,27,28,31,32,34,35,37,38,39,40,41,42,43].

To the best of our knowledge, this is the latest meta-analysis reporting the associations between dietary patterns and the likelihood of hypertension. In the present meta-analysis, we have an update on the previous systematic review [6] and further assess the associations between drinking pattern and hypertension likelihood. The pooled results from this meta-analysis indicate that healthy pattern is significantly associated with a decreased odds of hypertension, whereas heavy drinking pattern is associated with an elevated odds of hypertension. Data from 27 studies involving 295799 participants were included in this meta-analysis. Our findings have further confirmed the significant associations between different dietary patterns and the odds of hypertension.

In our analyses, the healthy pattern was associated with a reduced odds of hypertension. This was in agreement with previously reported findings from China in which healthy/prudent patterns were reported to be associated with reduced odds of hypertension [15,16]. This favorable association might be explained by the beneficial effects of whole grains, vegetable and fruits on blood pressure. These foods are good sources of dietary fiber and antioxidants(e.g. Vitamins C and E, other carotenoid compounds). A previous meta-analysis of randomized, controlled clinical trails showed the beneficial effects of high fiber intake on blood pressure [44]. Similarly, Shu et al. [45] found that high dietary fiber consumption was associated with a reduced risk of obesity, which was considered as important risk for hypertension. Besides, several studies found that antioxidants (e.g. vitamin A, C, and E) could slow the rate of oxidation, protecting endothelial cells and vascular [46], thereby decreasing the risk of hypertension. Some previous studies have also indicated that the polyphenols, especially flavonoids found in vegetables and fruits may be contributing to the lowering effect on the blood pressure [47]. To our knowledge, flavonoids may exhibit a range of biological activities in vitro that may explain their potential cardioprotective properties, including anti-oxidant and anti-inflammatory effects and induction of apoptosis [48]. Furthermore, this pattern was characterized to have rich sources of Ca, Mg and K. Epidemiological studies and clinical trails indicated that these nutrients had vasodilation properties, and thus may decrease the level of blood pressure [49,50].

The western-style pattern was not associated with the odds of hypertension in this meta-analysis. Our results were inconsistent with some previous studies [17,27], which reported the positive associations between red and processed meat and the risk of hypertension. There are several plausible explanations that may explain this null association. Firstly, the consumption of foods with a rich source of calcium, potassium, and several vitamins along with “unhealthy” food may change the effect of western-style pattern to non-significant. Clinical and epidemiological studies have suggested that these beneficial nutrients like calcium, potassium and vitamins, play important roles in the prevention and the inhibition of blood pressure elevation [51,52]. Thus, these foods with high content of calcium, potassium, and several vitamins may counteract the harmful effects of other unhealthy items loaded on western-style pattern, which may explain the result [53]. Secondly, the BMI has been adjusted as a potential confounder in included studies. A number of previous studies has observed the significant association between obesity and the risk of hypertension [45,54]. Finally, the null association may be associated with some unknown or unmeasured potential confounders which were not considered in our analyses.

The heavy drinking pattern was associated with an increased odds of hypertension, whereas light-moderate drinking pattern was not associated with the odds of hypertension in our analyses. A previous systematic review and meta-analysis of alcohol consumption and risk of hypertension concluded that heavy alcohol consumption was significantly associated with an increased risk of hypertension [5]. In fact, high alcohol intake has been considered as an important risk factor for hypertension [55]. The mechanisms for increased odds of hypertension may be explained such as impairment of the baroreceptors, an imbalance of the central nervous system, stimulation of the renin-angio-tensin-aldosterone system and the endothelium to release vasoconstrictors, increased cortisol levels, increased vascular reactivity [56]. Moreover, Kim et al. [57] found that alcohol consumption was inversely associated with a reduction in eGFR in Korean men. Furthermore, to our knowledge, light-moderate alcohol intake may raise the concentration of HDL-C and decrease the levels of fasting insulin, which may be associated with the decreased risk of hypertension [58,59]. However, in our analyses, we did not observe the significant association between light-moderate drinking pattern and the odds of hypertension. This may be related to the amount of alcohol and physical activity level. Thus, alcohol consumption of light-moderate drinking pattern need to be defined in the different countries.

This meta-analysis holds its own strengths. Firstly, the cases of hypertension were confirmed based on clinician, avoiding misdiagnosis. Secondly, this is the latest meta-analysis reporting the associations between dietary patterns and the likelihood of hypertension. Besides, we further explored the associations between drinking pattern and the likelihood of hypertension. The pooled results were consistent with a previous systematic review and meta-analysis [5], which suggested that heavy alcohol consumption might increase the likelihood of hypertension. Thirdly, no signs of publication bias were evident in the funnel plot, and the statistical test for publication bias was non-significant. However, some limitations should be considered in this meta-analysis. Firstly, the principal limitation of this study was the use of potentially biased evidence. There was an inconsistent adjustment for potential confounders among the included studies. Consequently, the data included in our analyses might suffer from differing degrees of completeness and accuracy. Secondly, 16 of 27 studies used a cross-sectional design, which is more susceptible to recall and selection bias, especially dietary recall bias, than a cohort design. Thirdly, the geographical regions mainly covered in Asia, including China, Japan, and Korea in the present meta-analysis. Therefore, our findings have limited generalization to all the populations.

This meta-analysis indicated that the healthy pattern was associated with a decreased odds of hypertension, whereas heavy drinking pattern was associated with an elevated odds of hypertension. Our findings highlight the importance of whole-diet for the prevention of hypertension. Therefore, findings from the present meta-analysis add to existing literature by supporting the associations between different dietary patterns and the odds of hypertension.

The authors declared no conflict of interest.

The authors acknowledge department of Nursing, Linyi People's Hospital for their important contributions to the collection of data in this study.