Sue K. Gao, Annette L. Fitzpatrick, Bruce Psaty, Rui Jiang, Wendy Post, Jeffrey Cutler, Matthew L. Maciejewski. Suboptimal Nutritional Intake for Hypertension Control in 4 Ethnic Groups. Arch Intern Med. 2009;169(7):702–707. doi:10.1001/archinternmed.2009.17
This study compared intake of specific nutrients based on the Dietary Approaches to Stop Hypertension (DASH) guidelines for hypertension management among multiethnic middle-aged and older adults.
We conducted quantitative analysis using baseline data of a prospective cohort study of 5972 adults aged 45 to 84 years recruited between July 2000 and August 2002 who participated in the Multi-Ethnic Study of Atherosclerosis (MESA). Diet information was collected using a 120-item food frequency questionnaire. Bivariate and multivariate methods were used to evaluate associations between DASH-accordant intake of each nutrient (fat, saturated fat, cholesterol, protein, fiber, calcium, magnesium, and potassium) with ethnicity and hypertension status.
Less than 30% of MESA participants met any DASH nutrient target. DASH accordance was lowest in saturated fat intake and highest in cholesterol intake (5.3% and 29.5% of the participants, respectively). Multivariate analyses showed significant ethnic differences in DASH accordance in all nutrients but saturated fat. Compared with white participants, Chinese American participants had greater DASH accordance in cholesterol (odds ratio [OR], 1.37; 95% confidence interval [CI], 1.13-1.67) and protein intake (2.32; 1.55-3.49) but less in total fat (0.47; 0.30-0.74), magnesium (0.58; 0.51-0.67), and potassium intake (0.40; 0.20-0.81); African Americans and Hispanics had greater DASH accordance in fiber intake (1.36; 1.13-1.62; and 2.23; 1.53-3.23, respectively) but less in calcium intake (0.44; 0.37-0.52; and 0.79; 0.68-0.91, respectively). Diagnosed and uncontrolled hypertension was associated with less DASH accordance in saturated fat (OR, 0.80; 95% CI, 0.70-0.91) and magnesium (0.80; 0.71-0.91). DASH accordance differed significantly with and without inclusion of dietary supplements in the analysis.
There is significant variation in DASH goal attainment among different ethnic groups. Assessments of nutrient intake that exclude dietary supplements may be underestimating DASH accordance.
Hypertension is a major cause of stroke, coronary heart disease, heart failure, and end-stage renal disease.1 The Dietary Approaches to Stop Hypertension (DASH) diet has been promoted as 1 of the most effective strategies for lowering blood pressure.2 It consists of an eating plan low in total fat, saturated fat, and cholesterol and rich in fruits, vegetables, and low-fat dairy products.3 The underlying nutrient targets are reduced fat, saturated fat, and cholesterol intake and increased protein, fiber, calcium, magnesium, and potassium intake.3 Literature suggests that these nutrients are associated with blood pressure control.3- 7
Although the DASH diet has been shown to be effective among white and African American populations,3,8 it has not been critically examined in other ethnic groups. This study examined variation in accordance with DASH nutrient intake in white, African American, Chinese American, and Hispanic individuals. We hypothesized that Hispanics and Chinese Americans would differ in their DASH accordance from whites because of ethnic-specific eating patterns that stem from different food choices and food preparation techniques. These patterns have been supported by previous studies on diet and ethnicity.9,10 We also examined whether accordance with DASH nutrient intake varied by hypertension status in this multiethnic population. We hypothesized that normotensive individuals would have best DASH accordance based on a recent National Health and Nutrition Examination Survey (NHANES) that showed a declining trend in DASH accordance over time among people with hypertension.11
Moreover, DASH guidelines have promoted optimal nutrition via whole food intake and have not accounted for dietary supplement intake, even though many people use dietary supplements to obtain key nutrients not otherwise provided in their diet.12 In this study, we measured DASH-accordant intake of nutrients obtained from both whole food and dietary supplements. To understand the importance of supplements in achieving DASH accordance, we also assessed the proportion of DASH nutrient intake attributable to dietary supplements and whole food.
For the purpose of this study, DASH accordance was defined as intake of a nutrient that falls within the range specified by DASH trials. The term compliance or adherence is not used because we had no information on participant knowledge or attitude toward the DASH diet and were unable to determine if these overt eating behaviors have been consciously modified.
The Multi-Ethnic Study of Atherosclerosis (MESA) is a population-based, cohort study designed to track the characteristics of subclinical cardiovascular disease in multiple ethnic groups in the US population.13 Between July 2000 and August 2002, 6814 men and women were recruited from 6 US sites (Baltimore County, Maryland; Chicago, Illinois; Forsyth County, North Carolina; Los Angeles County, California; Northern Manhattan, New York; and St Paul, Minnesota). MESA participants were 45 to 84 years old, free of any clinical cardiovascular disease, and self-identified as white, African American, Hispanic, or Chinese American. Informed consent was obtained from all participants, and study protocol was approved by the institutional review board at all 6 sites.
This analysis used the cross-sectional data collected from MESA's baseline visit. Among a total of 6814 MESA participants, 6236 participants had a valid baseline dietary assessment and valid information on hypertension status. Those with missing data on comorbidities, education level, or income were also excluded, resulting in a final sample size of 5972, including 777 Chinese Americans, 1318 Hispanics, 2391 whites, and 1486 African Americans. Major baseline characteristics of the people with missing data did not differ from the rest of the group.
To collect information on each participant's usual diet pattern during the previous year, a 120-item quantitative food frequency questionnaire (FFQ) was administered to participants during the baseline visit. The questionnaire was modified from the validated Multi-Ethnic Insulin Resistance Atherosclerosis Study FFQ to accommodate the ethnic diversity of the MESA participants. Fourteen typical Chinese and Hispanic ethnic foods were included to better assess the dietary behaviors of these study subpopulations. All forms were processed by the DietSys Nutrient Analysis program (HHHQ-DietSys Analysis Software, Version 4.02, National Cancer Institute, 1999), which assigned average daily intakes of nutrients based on the responses to FFQ survey items. The underlying nutrient and portion size databases used by DietSys were completely reviewed and updated for MESA.
We evaluated intake of each DASH nutrient separately, so there were 8 outcome variables: total fat, saturated fat, protein, cholesterol, fiber, potassium, magnesium, and calcium. For the 3 nutrients that should be controlled—total fat, saturated fat, and cholesterol—DASH accordance was defined as intake value lower than the cap used in DASH trials. For the 5 nutrients that should be promoted—protein, fiber, potassium, magnesium, and calcium—DASH accordance was defined as an intake value higher than the threshold used in DASH trials (Table 1).3 DASH-accordant nutrient intake was assessed based on the combined intake from both food and dietary supplements in all the main analyses.
The explanatory variables of interest were ethnicity and hypertension status. MESA specifically tracked individuals with self-reported ethnicity of white, African American, Hispanic American, or Chinese American. At each site visit, after a 5-minute rest, seated blood pressure was measured 3 times at 1-minute intervals using a Dianmap PRO 100 automated oscillometric device (GE Healthcare; Chalfont St Giles, England). In this analysis, we used the average of the second and third measurements of blood pressure. The first measurement was discarded because of the common “white-coat syndrome.”
Five categories of hypertension status were defined using up-to-date classifications of the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure.2 Prehypertension was defined as a systolic blood pressure between 120 and 139 mm Hg or a diastolic blood pressure between 80 and 89 mm Hg without use of antihypertensive medication. Diagnosed hypertension was defined by self-reported hypertension or self-reported use of any antihypertensive medications. The diagnosed hypertension group was then further divided into the controlled group and the uncontrolled group based on blood pressure measures (cutoff points: systolic blood pressure ≥140 mm Hg and diastolic blood pressure ≥90 mm Hg). Undiagnosed hypertension was defined as systolic blood pressure equal to or greater than 140 mm Hg or diastolic blood pressure equal to or greater than 90 mm Hg but with no self-report on the presence of hypertension or use of antihypertensive medication. Adults with normotension served as the reference group. This classification for hypertension status appropriately assigned all but 1 participant, who was taking an antihypertensive medication but did not self-report having hypertension.
We generated descriptive statistics of the main baseline characteristics of the MESA cohort, and differences by ethnicity were tested using the χ2 test for categorical variables and analysis of variance for continuous variables. The number and percentage of people who met each DASH target were calculated by ethnicity and hypertension status, and differences were tested using χ2 tests. To evaluate the relationship among ethnicity, hypertension status, and DASH accordance and address the complexity regarding ethnicity, a logistic regression of DASH-accordant intake in each nutrient was estimated on ethnicity and hypertension status, controlling for self-reported age, sex, self-reported conditions (diabetes mellitus, hyperlipidemia, hepatitis, and cancer), education level, and income. We built our models hierarchically to address the concerns about potential overadjustment of education and income.14,15
To assess the role of dietary supplements in DASH intake, we compared the proportion of DASH participants who met calcium, magnesium, and potassium targets based on intake from food only (PF) and the proportion who met these targets based on intake from food and dietary supplements (PF+S). The differences in proportions (PF+S − PF) represent the underestimate in DASH accordance that is obtained when only considering nutrient intake from whole foods.
Multiplicity were handled using Bonferroni corrections in bivariate estimates; for multivariate analyses, we used a more conservative significance threshold (P < .01). All multivariate models took account of clustering by geographic sites. All statistical analyses were performed with a commercially available software program (Stata, version 8; StataCorp, College Station, Texas).
Table 2 gives the baseline characteristics of the MESA participants. The mean age was 62.6 years, and slightly more than half (52.7%) were female (Table 2). The median education level was some college education, and the median household income was $35 000 to $49 999. Nearly 35% of the overall sample were normotensive, 18% were prehypertensive, 22% had diagnosed controlled hypertension, and almost 17% had diagnosed uncontrolled hypertension. Education level and income differed significantly by ethnicity (P < .001).
When both food and supplements are combined to assess dietary accordance, less than one-third of MESA participants had DASH-accordant intake in any of the 8 nutrients (Table 3). DASH accordance was worst for saturated fat (5.3% of total sample) and greatest for cholesterol intake (29.5% of total sample).
A higher proportion of whites were DASH accordant in total fat (18.4%) and calcium intake (37.3%) than the other ethnic groups (P < .01) A higher proportion of Hispanics had DASH-accordant intake in fiber (14.1%; P < .01) than whites, but a lower proportion had DASH-accordant saturated fat intake (3.6%; P < .01). A higher proportion of Chinese Americans had DASH-accordant intake in protein (45.2%) and cholesterol (34.1%) than whites, but a lower proportion had DASH-accordant potassium (3.1%) and magnesium intake (7.7%) (all P < .01). Differences of DASH-accordant nutrient intake by hypertension status were less consistent (results not presented; data available from S.K.G.).
In multivariate analyses with whites as the reference group, variation in DASH accordance due to ethnicity was found in 7 of 8 nutrients.4 Chinese Americans had more DASH-accordant intake in cholesterol (Table 4) (odds ratio [OR], 1.37; 95% confidence interval [CI], 1.13-1.67) and protein (2.32; 1.55-3.49) but less DASH-accordant intake in total fat (0.47; 0.30-0.74), magnesium (0.58; 0.51-0.67), and potassium (0.40; 0.20-0.81) targets. African Americans (OR, 1.36; 95% CI, 1.13-1.62) and Hispanics (2.23; 1.53-3.23) had better DASH accordance in fiber than whites but had less DASH accordance in calcium (African Americans: 0.44; 0.37-0.52; Hispanics: 0.79; 0.68-0.91). Ethnicities did not differ significantly in DASH accordance in saturated fat intake (P > .1).
Hypertension status was a significant predictor of DASH accordance in saturated fat and magnesium. The diagnosed and uncontrolled group was less likely to meet DASH targets in saturated fat (OR, 0.80; 95% CI, 0.70-0.91) and magnesium (0.80; 0.71-0.91) than the normotensive group. In saturated fat, undiagnosed hypertension was also associated with worse DASH accordance (OR, 0.77; 95% CI, 0.63-0.93), whereas diagnosed and controlled hypertension was associated with better accordance (1.31; 1.22-1.40).
When we compared the proportion of individuals with DASH accordance based on whole food and supplements (PF+S) or just whole food (PF), we found significant differences in all 3 nutrients (calcium, magnesium, and potassium). In the overall sample, the proportion of individuals with DASH accordance in calcium decreased from 29.4% to 13.5%, DASH accordance in magnesium decreased from 13.3% to 4.6%, and DASH accordance in potassium decreased from 7.1% to 6.8% once supplements were excluded. Differences in the proportion of MESA participants with DASH-accordant intake for calcium and magnesium (PF+S − PF) were greatest in whites (17.5% for calcium, P < .001; 8.5% for magnesium, P = .13).
We examined variation in accordance with DASH nutrient intake in white, African American, Chinese American, and Hispanic individuals aged 45 to 84 years from the MESA study and found significant differences in DASH-accordant intake in 7 of 8 nutrients by ethnicity. Ethnic differences in dietary behavior have been well documented in nutritional research.9,10,15,16 Similar to NHANES III data and the Healthy Eating Index report, we also found significantly lower intake of calcium among the African American and Hispanic participants.9,15 However, our finding that both African Americans and Hispanics were more likely to be DASH accordant in fiber intake than whites was contradictory to the mainstream literature in which minority ethnicity was associated with lower vegetable intake.9,15 Our dichotomous measure of DASH accordance was a different coding system than most diet quality indices that focus on the mean intake of nutrients or food groups. However, the FFQ used in this study was specially modified to accommodate ethnic foods and should be a better reflection of the real diets of ethnic minorities.17 Chinese American participants were less likely to meet total fat, potassium, and magnesium targets than whites but more likely to meet DASH goals in protein. Diet statistics for the Chinese Americans at the national level are rare and often do not account for the regional, origin, language, and cultural heterogeneity of the Chinese American population; these factors could potentially explain the findings of this study. Most of the observed DASH accordance differences by hypertension status were observed in the diagnosed and uncontrolled hypertension group. Compared with the diagnosed and uncontrolled hypertension group, people with diagnosed and controlled hypertension were more likely to meet DASH optimal intake in all nutrients, a result which could owe to either a previous diagnosis of hypertension that led to diet modification or diet modification that was initiated and resulted in greater reduction of blood pressure. If hypertension awareness led to behavior change in the diagnosed and controlled hypertension group, then the diagnosed and uncontrolled hypertension group would be made up of those who did not modify their original diet, a result that is consistent with our findings that the diagnosed and uncontrolled hypertension group behaved like the undiagnosed group in saturated fat intake. We could not assess the reverse causality that better diet led to better control status because of the cross-sectional nature of data.
When we compared DASH accordance based on whole food or whole food and dietary supplements in the overall sample, significant differences were found in the proportion of MESA participants who were DASH accordant in calcium and magnesium intake. A greater proportion of whites had less DASH-accordant intake when dietary supplements were excluded than other ethnic groups. This result was consistent with previous findings that whites were more likely to use multivitamins and other dietary supplements.12 These results suggest that assessments of nutrient intake that exclude dietary supplements may be underestimating DASH accordance.
This study has several limitations. First, the cross-sectional nature of data precluded causal inferences about ethnicity and hypertension status on DASH-accordant diet behaviors. Unobserved confounders not captured in MESA, such as social support, living condition, and physical environment, might bias the hypertension-diet relationship. Second, measurement errors in the FFQ and DietSys system could be affecting our results, and hypertension identified in part based on measurements at 1 clinic visit may not really correspond to a clinical diagnosis. Third, MESA participants are recruited from 6 geographic locations across the nation and oversampled ethnic minorities. MESA is an ongoing study from 1999 and onward. These time and space limitations need to be considered when generalizing MESA results to other populations.
Findings from MESA may provide insights into how to customize DASH messages for different ethnic groups. For the Chinese American group, the core message could be to promote the intake of micronutrients and fiber or their whole food equivalents of fruits and vegetables and dairy products. For Hispanics and African Americans, the core message could be to promote protein- and calcium-rich foods, such as seafood, beans and peas, and lean meat, and to consider the use of calcium-fortified foods.
This is one of the first studies to examine the nutrient intake patterns of multiple ethnic groups using the DASH targets and to examine the role of dietary supplements on DASH accordance. Since DASH dietary guidelines are widely recommended for hypertension control and for general health, future research should evaluate whether dietary patterns in broader multiethnic samples and samples with greater hypertension severity are in accordance with DASH recommendations.
Correspondence: Sue K. Gao, MPH, PhD, Amgen Inc, 1 Amgen Center Dr, MS28-3A, Thousand Oaks, CA 91320-1799 (email@example.com).
Accepted for Publication: November 14, 2008.
Author Contributions:Study concept and design: Gao, Psaty, and Maciejewski. Acquisition of data: Gao, Psaty, and Post. Analysis and interpretation of data: Gao, Fitzpatrick, Jiang, Post, Cutler, and Maciejewski. Drafting of the manuscript: Gao. Critical revision of the manuscript for important intellectual content: Fitzpatrick, Psaty, Jiang, Post, Cutler, and Maciejewski. Statistical analysis: Gao, Fitzpatrick, Jiang, and Maciejewski. Obtained funding: Fitzpatrick. Study supervision: Fitzpatrick and Maciejewski.
Financial Disclosure: None reported.
Funding/Support: This research was supported by contracts N01-HC-95159 through N01-HC-95165 and N01-HC-95169 from the National Heart, Lung, and Blood Institute and grant 1 R21 AT002152-01 from the National Center for Complementary and Alternative Medicine.
Additional Information: A full list of participating MESA investigators and institutions can be found at http://www.mesa-nhlbi.org.
Additional Contributions: Laura Svetkey, PhD, reviewed the manuscript. We thank the other investigators, staff, and participants of MESA for their valuable contributions.