Arelatively small number of epidemiologic studies and clinical trials have examined the effects of food intake patterns on blood pressure. Vegetarians, for example, have been found to have lower blood pressures than persons with more traditional Western diets.1,2 The Dietary Approaches to Stop Hypertension (DASH) clinical trial demonstrated that, among adults with borderline high blood pressure, high intakes of fruits and vegetables and of low-fat dairy products lower blood pressure.3 Although higher intakes of fruits and vegetables alone reduced blood pressure in that study, the reduction was more striking when low-fat dairy products were added.
Sodium is probably the nutrient most often studied in relation to blood pressure. A meta-analysis of 56 trials of sodium and blood pressure concluded that sodium restriction may be of benefit for older hypertensive adults, but its benefit is less certain for normotensive individuals.4 The DASH-Sodium Trial was undertaken later to study the effect of the DASH diet on blood pressure among normotensive individuals who consumed varying levels of sodium. They found that the DASH diet lowered blood pressure at all levels of sodium consumption, although the combined effects of low sodium (target of 50 mmol per day) intake and the DASH diet were greater than either intervention alone.5 Epidemiologic studies have also implicated nutrients other than sodium in the etiology of adult hypertension, including potassium, calcium, magnesium, protein, fiber, and dietary fats.6–9 Clinical trials of individual nutrients have generally shown only modest effects.10,11
Although the effects of a diet rich in fruits, vegetables, and low-fat dairy products on blood pressure levels among children have not been studied, a comprehensive review of a number of studies of macronutrient and micronutrient effects among children suggests that higher sodium intakes are generally associated with increased blood pressure levels in children.12 This same review found fewer and more equivocal studies linking potassium, calcium, and magnesium to blood pressure. One dietary intervention among 8- to 11-year-old children with elevated low-density lipoprotein (LDL)-cholesterol levels demonstrated inverse effects of calcium, magnesium, and potassium on blood pressure.13
Although a diet higher in milk and milk products is richer in calcium, magnesium, and potassium and a diet higher in fruits and vegetables is higher in potassium and magnesium, it is unclear to what extent any beneficial effects of fruits, vegetables, and dairy products on blood pressure are attributable to these 4 nutrients. Fruits, vegetables, and dairy products contain many other substances. Therefore, the primary goal of this study was to estimate the independent and combined effects of fruits and vegetables and dairy products on childhood blood pressure from preschool to early adolescence.
The Framingham Children's Study is a longitudinal study of children (and their parents) who were 3.0 to 5.9 years of age at entry in 1986. We initially enrolled 106 families, 95 of whom provided data into early adolescence. The children and their parents were assessed at yearly clinic visits by means of interviews, questionnaires, and measurements of anthropometry, blood pressure, and blood lipids. Data collected through 12 years of age were used in these analyses, giving us longitudinal data from the preschool years into early adolescence.
Each child's diet was assessed repeatedly throughout the study using 3-day diet records. To obtain stable estimates of baseline dietary intake, we collected 4 sets of 3-day diet records during the first year of the study. In each subsequent year, we requested 1 or 2 sets of 3-day records. During the early years of the study, parents completed all diaries for the children as well as their own diaries. The study nutritionist instructed each family in the completion of the diaries, including how to use common household measures to estimate portion sizes. In later years, the child assisted in the collection of the dietary data.
At each annual clinic visit, we took 5 measurements of blood pressure using an automated blood pressure device (Critikon Dinamap; GE Medical Systems, Milwaukee, WI).14 Blood pressure measures were taken following a standard protocol. After the cuff was attached, the child sat quietly for 5 minutes before the first measure; a 1-minute rest interval between measures was programmed into the device. The automated blood pressure device uses oscillometric techniques rather than ascertaining Korotkoff (K) sounds for measuring blood pressure. The diastolic blood pressure estimated with the automated device is slightly lower than the K5 heart sound and will thus be somewhat lower on average than that estimated with a standard sphygmomanometer.15 The average of all 5 measures at each visit was used to estimate mean systolic and diastolic blood pressure. A second series of blood pressure measurements was taken in some years at the time of the blood-drawing clinics. All measures at a given age were averaged to provide more stable estimates.
For this analysis, we used data gathered on the following potential confounders: the child's age, sex, energy intake per day, percent of energy from fat and saturated fat, height, and physical activity, as well as each parent's age, height, weight, and education level. At each yearly clinic examination, weight (to the nearest 0.25 pound) was measured using a standard counterbalance scale, and height was measured (to the nearest 0.25 inch) using a measuring bar on the same scale. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared. Physical activity was assessed once or twice yearly using an electronic motion sensor (Caltrac accelerometer; Muscle Dynamics, Torrance, CA). This motion sensor has been shown to provide valid estimates of physical activity in children.16 The children wore the device for 3 to 5 consecutive days on 1 to 4 separate occasions each year. The device was programmed to provide unitless activity counts to give a measure of total movement. This information was combined with data on the amount of time the instrument was worn to estimate the average number of activity counts per hour at each age.
Daily dietary data from food records collected each year were entered into a nutrition calculation program, the Nutrition Data System (NDS) of the University of Minnesota.17 The data were entered using a standard protocol, and cleaned and verified for accuracy and consistency of coding by the study's nutritionist.
To classify each child according to intake of fruits and vegetables and dairy products, we calculated servings per day in each food group as defined by the Food Guide Pyramid of the U.S. Department of Agriculture (USDA).18 Specifically, 1 serving of fruit includes such items as whole fruit (an apple or banana), a grapefruit half, melon wedges, 0.75 cup of fruit juice, 0.5 cup of berries, 0.5 cup of chopped, cooked, or canned fruit, and 0.25 cup of dried fruit. A serving of vegetables includes 1 cup of raw leafy vegetables, 0.5 cup of other vegetables (raw or cooked), 0.75 cup of vegetable juice, and other definitions as needed for such things as dehydrated vegetables. We totaled the number of fruit and vegetable servings to create the fruits and vegetables intake group. For the dairy group, a serving is defined as 1 cup of milk or yogurt, 1.5–2.0 ounces of dry cheese, and 1 ounce of low-fat cheese. For other cheeses (eg, cottage cheese, cream cheese), a serving is defined by the USDA as the amount of the food that provides about the same amount of calcium as a cup of milk. Mixed dishes are separated into their component foods. For example, French fries would contribute both to servings of vegetables and grams of discretionary fats. Servings from a slice of fruit pie would be apportioned to servings from the unsweetened fruit, added sugar, discretionary fat (in the crust and filling), and refined grain (in the crust and the filling thickener).
We combined data from the food records collected in the Framingham Children's Study with data from the USDA's technical files from the Continuing Survey of Food Intake by Individuals (CSFII).19 We matched the CSFII foods with foods in NDS, linking their food codes. In cases for which information was insufficient to match the foods, we compared nutrient content data along with recipes and ingredients to classify the food type. This allowed us to link each food from NDS food intake data to the USDA's Food Guide Pyramid servings.
We classified children according to servings per day of fruits and vegetables (combined) as well as dairy products. We used mean servings in these 2 food groups during preschool (ages 3.0–5.9 years) and elementary school (ages 6.0–11.9 years) to classify children as having higher or lower intakes of dairy (<2 vs. ≥2 servings per day) and fruits and vegetables (<4 vs. ≥4 servings per day). Sensitivity analyses in one-fourth serving increments were carried out to assess the sensitivity of the findings to cutpoint changes. To evaluate the effect of a diet similar to the DASH diet on blood pressure, we classified each child into 1 of 4 categories according to the combined pattern of intake of fruits and vegetables and dairy products, as follows: 1) low fruit and vegetable plus low dairy, 2) low fruit and vegetable plus high dairy, 3) high fruit and vegetable plus low dairy, and 4) high fruit and vegetable plus high dairy.
Each child's blood pressure slope was calculated by separately regressing systolic and diastolic blood pressure on age. Children who dropped out of the study before the fifth examination cycle or who were otherwise missing data for 3 or more examinations were excluded from the calculation of the slopes. We examined the separate effects of fruits and vegetables and total dairy on the following outcomes: 1) the slope of systolic and diastolic blood pressure from 3 through 12 years of age and 2) the mean systolic and diastolic blood pressure in early adolescence. To estimate the mean early adolescent systolic and diastolic blood pressure, we took the mean of all the measures from ages 10.0 through 12.9 years. We used analysis of covariance methodology to estimate the adjusted mean blood pressures in early adolescence associated with fruit and vegetable and/or dairy intake.
We explored potential confounding in the multivariable models by parents’ education levels, ages, and BMIs, and the child's age, sex, total energy intake, percent of energy from fat and saturated fat, height, and physical activity level. We retained those factors that were independent predictors or actual confounders of blood pressure change (as measured by changing the adjusted mean systolic or diastolic blood pressure estimates by more than in the first decimal place).
To determine whether mineral content of fruits and vegetables or dairy products might explain any of the protective effect of these foods on childhood blood pressure, we explored the effect of including calcium, magnesium, sodium, and potassium in the multivariable models. If, for example, controlling for calcium intake in the model led to an attenuation of the dairy-blood pressure relations, we might conclude that the apparent protective effect of dairy intake was a consequence of the higher calcium content of dairy foods.
Finally, because dietary intake may have its effect on blood pressure through body fat, we explored the effect of including BMI in the final multivariable models. For those models that examined blood pressure slopes, we included BMI change (as a slope) from preschool to early adolescence. For those models examining blood pressure at the time of early adolescence, we included early adolescent BMI.
In Table 1, we show the characteristics of the preschool-aged children and their parents according to their pattern of intake of fruits and vegetables (<4 vs. ≥4 servings per day) and dairy products (<2 vs. ≥2 servings per day) at 3 to 6 years of age. There were few differences in the characteristics of children associated with their intake of fruits and vegetables or dairy products at baseline. Children consuming 2 or more servings of dairy per day were somewhat more active at baseline than those consuming fewer servings (11.7 vs. 10.7 counts per hour, respectively). There were no consistent differences in the baseline characteristics of the parents according to their children's intakes of fruits and vegetables and dairy products. Fathers of children with higher fruit and vegetable intakes did have somewhat lower systolic blood pressures than those whose children consumed less.
Table 2 provides data on the mean dietary intakes of the children according to their preschool intakes of fruits and vegetables and dairy products. Children consuming either 4 or more servings of fruits and vegetables or 2 or more servings of dairy per day had higher total energy intakes than those consuming less. Higher consumption of fruits and vegetables was associated with lower energy-adjusted intakes of total fat, saturated fat, calcium, sodium, and vitamin D, and with higher energy-adjusted intakes of carbohydrates and potassium. Conversely, children consuming more dairy products had higher energy-adjusted intakes of fat, saturated fat, protein, calcium, magnesium, potassium, and vitamin D.
In the bottom section of Table 2, we provide data on mean servings in each of the 5 major food groups. For those children in the high fruits and vegetables category, we see that the intake of fruit is nearly double that of vegetables. Children in this category consume a somewhat greater number of servings per day from the meat and grain groups. When the children are categorized according to dairy intake, fruit and vegetable intakes were identical. Those in the high dairy group did consume slightly more meat and approximately half-a-serving more grain.
Table 3 shows the influence of preschool fruit and vegetable and dairy intake on the yearly change in systolic and diastolic blood pressure from preschool to early adolescence. Only baseline blood pressure and mean activity levels were found to confound the diet-blood pressure relations; these factors were retained in the multivariable models. In univariate models, calcium, magnesium, and potassium were all inversely related to systolic blood pressure change, whereas sodium was directly related. None of these minerals was found to attenuate the diet-blood pressure relations. Because magnesium and sodium were independent predictors of systolic blood pressure, they were retained in the multivariable models. After adjusting for baseline blood pressures, mean activity level, and mean intakes of magnesium and sodium per day, children who had either higher intakes of fruits and vegetables or higher intakes of dairy products during the preschool years had smaller increases in systolic blood pressure than did children with lower intakes of these foods. Dairy intake had a slightly stronger protective effect on blood pressure than did fruit and vegetable intake. Finally, children who had both high fruit and vegetable intake and high dairy intake had the lowest systolic and diastolic blood pressure slopes (eg, 1.72 mm Hg per year increase in systolic blood pressure in the high fruit and vegetable/high dairy group vs. 3.03 mm Hg increase per year for those in the low fruit and vegetable/low dairy group).
The last 2 columns of Table 3 show the effects of adding the slope of BMI throughout childhood to the multivariable models. There is a small attenuation of the effects of the dietary patterns on blood pressure slope, suggesting that a small part of the protective effect of fruits, vegetables, and dairy products may be the result of differences in body fat.
To answer the question of whether children who consumed predominantly reduced-fat dairy products (ie, >50% of total dairy from reduced-fat products) showed an even greater reduction in blood pressure, we further classified children into 1 of the following 4 categories: 1) low dairy intake/predominantly high fat (n = 31), 2) low dairy intake/predominantly low fat (n = 25), 3) high dairy intake/predominantly high fat (n = 18), and 4) high dairy intake/predominantly low fat (n = 18). The mean systolic blood pressure slopes in the 4 groups were 3.1, 2.8, 2.1, and 2.2 mm Hg, respectively. Diastolic blood pressure results were similar. Because there was no consistent beneficial effect of low-fat products, we present total dairy servings for the remaining analyses.
Table 4 shows the individual and combined effects of the servings per day of fruits and vegetables and dairy products, during the preschool years, on mean blood pressures in early adolescence. For systolic blood pressure in particular, children who consumed more fruits and vegetables as well as children who consumed more dairy products had systolic blood pressures in early adolescence that were approximately 4 mm Hg lower, on average, than children whose intakes were lower. Once again, children whose diets were higher both in fruits and vegetables and in dairy products had the lowest systolic and diastolic blood pressure levels in early adolescence. In the last 2 data columns, we see again that there was a small amount of attenuation of the fruit and vegetable effects when adolescent BMI was added to the multivariable models.
In this study, fruit and vegetable intake at 3 to 6 years of age correlated strongly with intake at 6 to 12 years of age (Pearson correlation = 0.74). The corresponding correlation between early and later dairy intakes was 0.66. Nonetheless, dietary intake may change somewhat when children enter school. Therefore, we examined the effect of dietary intake in the elementary school years on early adolescent blood pressure (Table 5). The results were similar to those found for preschool dietary intake and blood pressure. Children who consumed 4 or more servings of fruits and vegetables per day or 2 or more dairy products from 6 to 12 years of age had lower systolic blood pressures and slightly lower diastolic blood pressures in early adolescence. As before, those children whose diets were higher in both fruits and vegetables and dairy products had the lowest blood pressure levels by the time of early adolescence. Conversely, children with lower intakes of fruits and vegetables and lower dairy intakes had the highest blood pressures at the end of follow up. Even after accounting for differences in BMI in early adolescence, those children consuming higher amounts of fruits and vegetables and higher amounts of dairy products during elementary school had a systolic blood pressure that was 7 mm Hg lower in early adolescence than that of children having lower intakes of both fruits and vegetables and dairy products.
This study provides new evidence that diets characterized by higher intakes of dairy (whether full-fat or reduced-fat products) and higher intakes of fruits and vegetables have beneficial effects on childhood blood pressure. These effects were stronger for systolic than for diastolic blood pressure. The combination of higher dairy consumption and higher fruit and vegetable consumption provided the greatest blood pressure benefit. Children with higher preschool intakes in both food groups had a mean systolic blood pressure, by the time of early adolescence, of 106 mm Hg, whereas those with lower intakes both of fruits and vegetables and of dairy had a mean systolic blood pressure of 113 mm Hg. Those with higher intakes of fruits and vegetables alone or dairy alone had intermediate levels of adolescent systolic blood pressure. The trends for diastolic blood pressure were weaker.
The effects of fruit and vegetable and dairy intakes were similar whether we looked at diet during the preschool years or during elementary school. Dietary patterns during these 2 age periods were highly correlated. These findings support the idea that the dietary habits established in the preschool years persist when children enter school and may have long-term effects on health outcomes.
The results of this study are consistent with the results of the DASH study.3 Although the DASH trial was carried out among adults with borderline high blood pressure, the current study suggests that the potential benefits of a dietary pattern characterized by higher intakes of fruits, vegetables, and dairy products on blood pressure may extend to normotensive children. Although only 1 family in this study had a lactovegetarian dietary pattern, our results may also be consistent with earlier studies showing a beneficial effect of a vegetarian diet.1,2
The mechanisms by which fruits, vegetables, and dairy products lower blood pressure are unclear. These foods are rich in minerals such as calcium, magnesium, and potassium, factors that have been associated with blood pressure reduction. In this study, however, controlling for total calcium, magnesium, potassium, and sodium in the diets did not explain the reduction in blood pressure associated with fruit, vegetable, and dairy intakes. Controlling for total fat and saturated fat did not diminish the effects.
Diets high in fruits, vegetables, and dairy products may be reflective of generally healthier dietary patterns. For example, children with higher fruit and vegetable and dairy intakes consumed slightly more of their total grains from whole grains (ie, 12.1% and 10.7% for the high and low fruit and vegetable groups, and 12.1% and 10.3% for high and low dairy intake groups). However, the mean intakes of legumes, nuts, seeds, and soy were the same or slightly lower in high (vs. low) fruit and vegetable group and in the high (vs. low) dairy group. Foods are complex packages of vitamins, minerals, and other compounds that may impact blood pressure through a variety of means. For example, peptides derived from milk proteins, especially fermented milk products, can function as angiotensin-converting enzymes, thereby lowering blood pressure.20
Because children who have healthier diets (perhaps reflected by higher intakes of fruits, vegetables, and dairy products) may be leaner, we were particularly interested in whether differences in body size might explain the lower blood pressures of children consuming more fruits and vegetables or dairy products. Controlling for BMI led to a slight attenuation of the effects, suggesting that a small part of the protective effect of diet may be explained by differences in body size.
This study's sample size limits our ability to examine sex-specific differences or to carry out other stratified analyses. Nevertheless, the precision of the estimates in this study was sufficient to detect important differences in blood pressure change and adolescent blood pressure associated with the intake of dairy products and fruits and vegetables. Although the effects were stronger for systolic blood pressure than for diastolic blood pressure, the consistency of the findings for both systolic and diastolic blood pressure is striking.
This study is strengthened by the detailed assessment of dietary intake. In the first year of the study alone, the parents provided diet record data for more than 9 days on average for each child. As has been demonstrated previously,21 compliance in this study was generally quite high. The study is further strengthened by the replicate yearly measures of blood pressure and the use of an automated blood pressure device. The covariates of interest were ascertained with similar precision (eg, repeated yearly measures of physical activity with an electronic motion sensor throughout childhood), allowing us careful control of potential confounding factors.
This study is unique is several ways. It is the first published study to examine the independent and combined effects of the intake of fruits and vegetables and dairy products on blood pressure change among children. It is also unique in its application of methodology for using dietary data derived from the Nutrition Data System to determine the child's pattern of intake of USDA Food Pyramid servings per day by linking with the CSFII technical files from the USDA.
These results suggest that a dietary pattern rich in fruits, vegetables, and dairy products may have beneficial effects on blood pressure change among children. Although it is unlikely that a large clinical trial of dietary intake and blood pressure change throughout childhood would be feasible, it will be important to replicate these findings using data from larger prospective studies of children.