| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
Center for Food and Nutrition Policy, Virginia Tech, Alexandria, Virginia
Address reprint requests to: Maureen L. Storey, Ph.D., Center for Food and Nutrition Policy, Virginia Tech, 1101 King Street, Suite 611, Alexandria, VA 22314. E-mail: Forshee{at}vt.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONDATA AND METHODSRESULTSCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
Design: Data from the U.S. Department of Agriculture’s Continuing Survey of Food Intake by Individuals 1994–96, 98 were used in the analyses. Age groupings (2–3, 4–8, 9–13, and 14–18 year olds) were based on the National Academy of Sciences recommendations for AI of calcium—500, 800, 1,300, and 1,300 mg calcium per day, respectively.
Results: Consumption of milk products was strongly and positively associated with calcium intake. Consumption of carbonated soft drinks and other non-dairy beverages was also positively associated with calcium intake, but this association was very weak. Beverage choices of African-American children and adolescents are significantly different than white and Hispanic children and adolescents. For example, African-American adolescent girls consume fewer milk products and more fruit drinks/ades. Average daily carbonated soft drink consumption is approximately 1.6 and 1.0 twelve ounce cans among 14–18 year old boys and girls, respectively.
Conclusion: Carbonated soft drink consumption among adolescent girls is modest and does not appear to be linked to decreased calcium intake. The analyses in this paper show that creative effective, efficient, and targeted policies should be considered to help adolescent girls increase calcium intake. Making low-fat milk products, flavored milks, calcium-fortified beverages and foods more attractive and available will help encourage girls to consume more of this important mineral. When adequate calcium intake is not achieved through foods, health professionals should consider recommending calcium supplements.
Key words: calcium, adequate intake, children, adolescents, soft drinks, milk
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONDATA AND METHODSRESULTSCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
One risk factor associated with osteoporosis is low bone mineral density. Therefore, maximizing bone mineral density during adolescence and early adulthood is crucial to good bone health later in life [13]. But while peak bone mass (PBM) is fully achieved during early adulthood [16], fractional calcium absorption is highest during early adolescence [17,18]. Moreover, a study conducted by Weaver and colleagues showed that adolescent girls absorbed and retained more calcium than young adult women [19]. Recently, another study showed that Chinese adolescent girls who drank milk had higher bone mineral density than those who did not [20]. Optimizing calcium intake along with the proper balance with phosphorus, vitamin D, and other key nutrients involved in bone health is particularly critical for pre-adolescent and adolescent girls due to the high incidence of osteoporosis among older women.
Low calcium intake especially among girls and women, is not a new phenomenon, however. The Third Report on Nutrition Monitoring in the United States reports, "Little change occurred in mean intakes of calcium between NFCS 1977–78 and CSFII 1989–91. Mean intakes of calcium remained well below the RDA of 800 mg/d for adult females" [21].
Nationwide government surveys conducted since the early 1970s consistently showed that calcium intake decreased with age with a sharp decline in the 30s and 40s. In particular, adult women consumed the lowest amounts of calcium of any group, including toddlers and very young children (ages 3–5 years). Data from the 1971–74 National Health and Nutrition Examination Survey I (NHANES I) showed that calcium intake among adult men (ages 20–29 to age 70 and over) ranged from 1,115 milligrams (mg) to 693 mg Ca/d, respectively; but calcium intake among adult women was much lower, ranging from 685 to 537 mg Ca/d (20–29 y to 70 y and over, respectively). Twenty years later, the Continuing Survey of Food Intake by Individuals (CSFII 1994–96, 98) indicated that calcium intake among adult women remained the same or rose slightly. For example, adult white women ages 20–39 y and 40–59 y consumed an average of 711 and 652 mg Ca/d, respectively, and adult black women in these age ranges consumed 549 and 513 mg Ca/d, respectively.
Among adolescents, data from the 1977–78 Nationwide Food Consumption Survey (NFCS) showed that the average calcium intake by adolescent boys—12–19 years of age—was about 1,145 mg Ca/d; whereas the same survey showed that average calcium intake by adolescent girls ages 12–15 and 16–19 years was only 849 and 716 mg Ca/d, respectively. A decade later, the 1989–91 CSFII showed that calcium consumption among boys and girls had remained fairly steady [22]. These trends are shown in Fig. 1a–c.
|
The objective of this study is to examine various factors associated with total calcium intake and percent adequate intake (% AI) of calcium by children and adolescents, with respect to age, gender, race/ethnicity, and diet and beverage choices. Specific research questions include:
Q1: Are certain non-milk beverages or food groups negatively associated with calcium consumption?
Q2: How strong is the relationship between consumption of milk and milk products and calcium consumption compared to other food groups and beverages?
Q3: What relationship do demographic factors (age, gender, race/ethnicity, and family income) have with calcium consumption?
|
DATA AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONDATA AND METHODSRESULTSCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
Group means in the descriptive section of the paper were estimated using multiple regressions with the constant suppressed. This procedure was analogous to an ANOVA (analysis of variance) model. The regressions were estimated in STATA 7 using the "svyreg" procedure to adjust for the complex design of the sample and the "subpop" option to calculate the group means for the age group, gender, and race/ethnicity sub-groups. This procedure used a Taylor linearization approach to correct the estimated standard errors for survey design effects [30]. The statistical significance of differences between group means (at the 0.05 level) was calculated using the STATA "test" procedure, which calculates the probability that any two estimated coefficients are equal to one another.
We conducted multiple regression analysis to determine the independent relationships of diet, beverage consumption, and selected demographic variables with calcium consumption (in mg) for the age groups 2–3 years (y), 4–8 y, and 9–18 y. The Adequate Intakes (AIs) for calcium in these age ranges were 500 milligrams (mg), 800 mg, and 1300 mg, respectively [31]. The models corrected for the complex survey design of the CSFII by using the "svyreg" procedure in STATA 7. The procedure generated heteroskedastic consistent standard errors, also known as robust standard errors. Diagnostic tests for multicollinearity and outliers were performed and did not identify any problems that affected the results in a significant way.
We tested the sensitivity of the specification by calculating several plausible alternative specifications. Specifically, we examined both quadratic and categorical specifications for the diet and beverage variables to test for non-linear relationships. We also estimated a step-wise regression algorithm on the full model as a diagnostic test to determine if any included variables were creating serious inefficiency (large variance) in the estimate of other coefficients. Note that step-wise regression was used as a diagnostic test only. None of the alternatives affected the substantive findings presented in this paper.
Age was included as an explanatory variable. Other demographic explanatory variables included sets of binary variables for gender, race/ethnicity, and family income as the percent of the poverty line. We estimated models to test for an association between calcium consumption and region of the country or residence in a city, suburb, or rural area, but none of these associations was statistically significant and they were eliminated from the final model.
Explanatory variables that were used in the total calcium consumption model include food group servings (g)—grains, vegetables, fruits, milk and milk products, lean meat, total fats and oils, and total sugars and sweets. Beverages included (in g): coffee, tea, regular fruit drinks/ades, diet fruit drinks/ades, regular carbonated soft drinks, diet carbonated soft drinks, and alcohol (alcohol was omitted from the model for ages 2–3 y).
Diet in this paper was modeled as the grams (g) of the major food groups (grains, vegetables, fruits, milk, meats, total fats and oils, and total sugars and sweets) as identified by CSFII and the Pyramid Servings Database and grams of beverage consumption (coffee, tea, regular fruit drinks/ades, diet fruit drinks/ades, regular carbonated soft drinks [RCSD], diet carbonated soft drinks [DCSD], and alcoholic beverages). Fruit drinks/ades include fruit punches and reconstituted powdered fruit drinks.
The food groups were based on the Pyramid Servings Database [32] as coded in CSFII. The grains group includes yeast bread and rolls, quick breads, rice, pasta, breakfast cereals, grain-based snacks, and baked goods. Foods with relatively high amounts of fat and sugar were counted toward the Pyramid Tip categories of discretionary fat or total sugars or sweets. The vegetable group includes dark-green, deep yellow, starchy, and other vegetables. This category included potato products such as fried potatoes and potato chips. The fruit group includes two subgroups—"citrus, melons, berries" and "other fruits." Fruit juices are included in this category. The milk and milk products group includes milk, yogurt, and cheese. Dairy foods that are primarily fat are excluded from this group. The lean meat group includes beef, pork, lamb, veal, game, poultry, fish, shellfish, frankfurters, sausages, bacon, luncheon meats, organ meats, and meat alternates. Only the cooked lean meat equivalents are included in this category. Excess fat in a meat product is counted in the discretionary fat category of the Pyramid Tip.
We kept RCSD and fruit drinks/ades as separate variables instead of combining them as a single "sweetened beverage" category. The main advantage of entering the two beverages as separate variables in the model is that it allows for the possibility that they may have different relationships with milk consumption. We believe that this is particularly important since there are important differences in the patterns of consumption for the two variables [33]. For example, African-Americans consume more fruit drinks/ades and less RCSD than do whites. Including the two variables in a single category would only be appropriate if there was strong reason to believe that they have the same relationship with milk consumption (since the new coefficient is constrained to be the same for both beverages). Combining the two beverages would increase the efficiency of the statistical estimation by reducing the number of parameters that were estimated and reducing multicollinearity. However, this would introduce bias if the relationships between the two beverage types and milk consumption are, in fact, different. Given the relatively large sample sizes of CSFII and the variance within each beverage category, we believe the data support the estimation of separate coefficients without undue loss of efficiency.
Water consumption is not included in the analysis. CSFII includes two water variables. One water variable measures the grams of water consumed based on the 24-hour recall, but this includes water from all sources, not just plain water consumed as a beverage. The other water variable is a food frequency style question that asks the respondent how many ounces of water he or she consumed the previous day. This collection technique is not directly comparable to the 24-hour recall, which includes many probing questions to improve the validity of the measure.
Demographic variables included age (years), female (1 if respondent is female, 0 if male), family income (as a percent of the poverty line), non-Hispanic African-American (1 if respondent is African-American and not of Hispanic origin, 0 otherwise), Hispanic (1 if respondent is of Hispanic origin, 0 otherwise), and non-Hispanic, other race (1 if respondent is not Caucasian, not African-American, and not of Hispanic origin, 0 otherwise).
The numbers of children and adolescents, gender, and race/ethnicity are shown in Table 1. After applying sample weights, the average household income was $42,870 and 49 percent of the sample was female.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONDATA AND METHODSRESULTSCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
|
|
|
|
|
|
Ages 2–3 Years
Boys and girls 2–3 years of age of all race/ethnicities exceed the AI of calcium (500 mg) for this age group. On average, boys and girls consume 808 and 774 mg Ca, corresponding to 162 and 155% AI of calcium, respectively (Table 2). Even at this age, total calcium intake, and therefore, % AI of calcium among African-American children is significantly lower than that of white and Hispanic children. Nevertheless, % AI of calcium for the African-American children exceeds 100 percent.
Differences in consumption of milk products and fruit drinks/ades are also beginning to emerge at this early age with African-American children consuming less milk products and more fruit drinks/ades than white and Hispanic children. Young African-American boys consume significantly more regular fruit drinks/ades than white and Hispanic boys (160 g vs. 98 and 103 g, respectively) and young African-American girls consume significantly more regular fruit drinks/ades (138 g) than white girls (96 g), but not more than Hispanic girls at this age (108 g). Consumption of RCSD is low at this age (boys, 74 g vs. girls, 69 g) and very little DCSD are consumed (boys, 6 g vs. girls, 6 g).
Ages 4–8 Years
At this age, average total calcium intake among boys and girls is greater than it is at age 2–3 years and boys consume about 100 mg more calcium than girls do (boys, [4–8 y] 923 vs. [2–3 y] 808 mg; girls, [4–8 y] 822 vs. [2–3 y] 774 mg). Moreover, on average, 4–8 year old children still exceed the AI of calcium (800 mg Ca/d). African-American boys and girls, however, have significantly lower total calcium intakes and % AI calcium, and they consume fewer milk products than do white or Hispanic boys and girls. For example, African-American boys consume about 150 fewer grams of milk products than do white boys; and African-American girls consume about 100 fewer grams of milk products than white girls do.
Consumption of RCSD and regular fruit drinks/ades begins to increase at this age. Average consumption of RCSD among boys and girls 4–8 years of age is nearly identical (120 g vs. 116 g, respectively). African-American boys and girls continue to consume significantly more regular fruit drinks/ades than all other race/ethnicity groups. There is no statistical difference in the consumption of RCSD across any of the race/ethnicity groups for girls, but African-American boys consume significantly less RCSD than do Hispanic boys.
Adolescents—Ages 9–13 and 14–18 Years
The average total calcium intake increases among 9–13 year old, pre-adolescent boys and girls (1,022 and 842 mg Ca/d, respectively). On average, pre-adolescent boys and girls consume 79 and 65% AI calcium, respectively. The low % AI of calcium is most striking among pre-adolescent African-American boys and girls getting 65 and 54% AI of calcium, respectively.
The AI of calcium increases by 62.5 percent, from 800 mg to 1,300 mg Ca/d, at age 9 to reflect changes in the physiologic requirement for calcium. It should be recognized therefore that the sudden decline in the mean % AI of calcium results because pre-adolescent boys and girls do not increase their calcium consumption in proportion to the increase in the AI of calcium. That is, the increase in the denominator (1,300 mg Ca/d) becomes much larger relative to the increase in the numerator—total calcium intake.
Among pre-adolescents, white boys consume significantly more milk products, and have higher calcium intakes and % AI of calcium than do African-American and Hispanic boys at this age. White, African-American, and Hispanic boys consume 1091, 846, and 961 mg Ca/d, achieving 84, 65, and 74% AI of calcium, respectively. Differences in milk product consumption are striking between the white and African-American pre-adolescent boys (483 vs. 334 g/d, respectively).
An interesting trade-off between RCSD and fruit drinks/ades emerges among white and African-American pre-adolescent boys at this age, with white boys consuming significantly more RCSD, but significantly less regular fruit drinks/ades than do the African-American boys. DCSD consumption remains low, especially among the Hispanic pre-adolescent boys.
Like the boys, African-American pre-adolescent girls consume significantly fewer grams of milk products (262 g/d) and have lower total calcium intake (705 mg Ca/d) and % AI of calcium (54 percent) than do white and Hispanic pre-adolescent girls. Unlike African-American pre-adolescent boys, however, RCSD and regular fruit drink/ades consumption among African-American pre-adolescent girls is not significantly different than that of the white or Hispanic pre-adolescent girls.
Consumption of DCSD among white pre-adolescent girls begins to increase at this age and is significantly greater than that of African-American and Hispanic pre-adolescent girls. Nevertheless, average DCSD consumption among the white girls is very low at this age—46 g DCSD/d or about one-tenth of a 355 mL can.
Among 14–18 year old adolescent boys of all races/ethnicities, average total calcium intake and % AI of calcium increases, but surprisingly, consumption of milk products is slightly less (28 g) than 9–13 year old boys. However, the older boys consume more of other foods that may contain calcium, such as grains and juices, which may account for this difference.
The race/ethnicity patterns remain. White adolescent boys consume significantly more milk products and get 1,241 mg Ca/d or nearly 100% AI of calcium. African-American and Hispanic adolescent boys consume significantly less milk products and calcium, and get about three-fourths of the AI for this age group. White adolescent boys drink significantly more RCSD than do African-American boys (708 vs. 393 g/d, respectively). African-American adolescent boys drink two times more regular fruit drinks/ades than white and Hispanic adolescent boys do, but the difference across the race/ethnic groups was not statistically significant.
Average total calcium intake and % AI of calcium among adolescent girls drops. Average total calcium intake falls to 704 mg Ca/d or 57% AI of calcium. White adolescent girls consume 755 mg Ca/d. African-American adolescent girls get less than half of the % AI of calcium and consume only 152 g of milk products.
White adolescent girls consume significantly more RCSD than African-American girls do (409, 231 g/d, respectively), which represents about 1.1 and 0.6 355 mL cans of RCSD for the white and African-American girls, respectively. African-American adolescent girls drink significantly more regular fruit drinks/ades than either white or Hispanic girls do.
Soft Drink Consumption in Categories
Some of the beverage consumption variables have a skewed distribution because there are a large number of non-consumers who cluster at 0 and a few heavy consumers who are at the far tail of the distribution. In order to appropriately interpret the means, we have calculated the proportion of respondents who fell into consumption categories corresponding roughly to Non-consumers (0 g), One serving or less (1–355 g), One to Two Servings (356 g–710 g), and More than Two Servings (>710 g) for fruit drinks/ades and RCSD, the two most relevant beverage categories.
In the 2–3 y age group, 52.2 percent consumed no fruit drinks/ades and 56.8 percent consumed no RCSD; 38.5 percent consumed one serving or less of fruit drinks/ades and 38.9 percent consumed one serving or less of RCSD; 8.0 percent consumed one to two servings of fruit drinks/ades and 4.0 percent consumed one to two servings of RCSD; 1.2 percent consumed more than two servings of fruit drinks/ades and 0.3 percent consumed more than two servings of RCSD.
In the 4–8 y age group, 46.8 percent consumed no fruit drinks/ades and 48.9 percent consumed no RCSD; 41.0 percent consumed one serving or less of fruit drinks/ades and 42.6 percent consumed one serving or less of RCSD; 10.6 percent consumed one to two servings of fruit drinks/ades and 7.6 percent consumed one to two servings of RCSD; 1.6 percent consumed more than two servings of fruit drinks/ades and 0.7 percent consumed more than two servings of RCSD.
In the 9–18 y age group, 58.2 percent consumed no fruit drinks/ades and 28.9 percent consumed no RCSD; 29.5 percent consumed one serving or less of fruit drinks/ades and 30.0 percent consumed one serving or less of RCSD; 9.3 percent consumed one to two servings of fruit drinks/ades and 25.4 percent consumed one to two servings of RCSD; 3.0 percent consumed more than two servings of fruit drinks/ades and 15.6 percent consumed more than two servings of RCSD.
Milk Consumption and Age
The pattern of milk products consumption across age groups is a function of gender, which is illustrated in Figure 2a–b. It is clear from the graphs that the consumption of milk and milk products declined much more rapidly for girls than it did for boys with the two curves beginning to diverge sharply around the 9–13 y age group. Boys had flat or increasing consumption of milk and milk products through the 9–13 y age group. The mean consumption of milk and milk products for boys declined in the 14–18 y age group. The decline for whites and African-Americans was modest, but it was more significant for Hispanics. Girls experienced a much steeper decline in consumption of milk and milk products. The mean consumption of milk and milk products for girls declines between the 4–8 y age group and the 9–13 y age group and falls sharply for the 14–18 y age group. These differences in milk consumption are expected to explain much of the differences in percent AI of calcium across these groups.
Regression Analyses
The multivariate regression models for calcium consumption (mg) provided a remarkably good fit to the data on calcium consumption. The models for ages 2–3, 4–8, and 9–18 years explained 85, 82, and 83 percent of the variance in calcium consumption, respectively. Consumption of milk products had the strongest relationship with calcium consumption. Other beverages including RCSD, DCSD, and regular and diet fruit drinks/ades had either no relationship or weakly positive relationships with calcium consumption.
Food Groups and Calcium Intake
Servings of the major food groups—as represented by the CSFII variables—had a strong relationship with calcium consumption. Not surprisingly, servings of milk products had the strongest relationship with calcium consumption. In fact, a model using only the servings of milk products (g) explained 78, 73, and 71 percent of the variance in calcium intake for 2–3, 4–8, and 9–18 years of age, respectively. For example, each additional gram of milk products was associated with a 1.2 mg increase in calcium consumption for ages 2–3 and 4–8 and a 1.3 mg increase for ages 9–18. Our model therefore predicts that an additional 8 oz. serving of 2 percent milk (244 g) would increase calcium consumption by 293 mg for ages 2–3 and 4–8 and by 317 mg for ages 9–18. Thus, our model is very consistent with the contribution made by an 8 oz. serving of 2 percent milk, which provides 298 mg of calcium.
Most of the other food group variables show positive and statistically significant associations with calcium intake. Total fats and oils has a strong positive association with calcium consumption for ages 2–3 and ages 4–8. A gram of total fats and oils is associated with a 2.3 and 1.1 mg increase in calcium consumption among children ages 2–3 and 4–8, respectively (p < 0.01). Consumption of discretionary fat is not associated with calcium consumption among adolescents ages 9–18.
Grains have the next largest set of coefficients. For ages 2–3 and ages 9–18, each gram of grains is associated with a 0.5 mg (p < 0.001 for each) increase in calcium consumption. For ages 4–8, each gram of grains is associated with a 0.6 mg (p < 0.001) increase in calcium consumption. Consumption of vegetables and lean meats was also associated with calcium consumption. The associations were statistically significant in all three models and ranged from 0.2–0.3 mg of calcium per gram of vegetable or lean meat. Consumption of fruits had a statistically significant association with calcium consumption, but the association was small—each gram of fruits was associated with a 0.1 mg increase in calcium consumption. Consumption of sugars and sweets was positively associated with calcium consumption for ages 2–3, but not for the other age ranges. Among children ages 2–3, each gram of total sugars and sweets was associated with a 0.3 mg increase in calcium. The full results are displayed in the Fig. 4 and shown in Table 6.
|
|
|
Demographics and Calcium Intake
Certain demographic variables were also associated with calcium consumption, but many of the striking differences observed in the descriptive statistics disappeared after controlling for diet and the other factors in the model. This is not surprising since calcium consumption is a function of diet, not race or gender.
Despite the large differences in calcium consumption between boys and girls in the descriptive statistics, there was only a small difference between boys and girls in the regression models after controlling for age, race/ethnicity, diet, and beverages. Being female was associated with 26 mg (p < 0.001) less calcium consumption among ages 4–8 and 42 mg (p < 0.001) less calcium consumption among ages 9–18. There was no statistically significant difference in calcium consumption between boys and girls in the 2–3 year old age group. This strongly suggests that the large differences in calcium consumption and % AI of calcium shown in the descriptive statistics were functions of the differences in the consumption of milk and milk products and the other variables in our model.
The race/ethnicity variables told a similar story. After controlling for the other variables in the model, many of the race/ethnicity differences were smaller or not statistically significant. Among ages 2–3 years, African-Americans consumed 33 mg less calcium than whites (p < 0.001), Hispanics consumed 26 mg less calcium than whites (p < 0.05), and respondents in the "non-Hispanic, other race" category consumed 83 mg less calcium than whites. Among ages 4–8, there was no statistically significant difference between African-Americans, Hispanics, and whites, but respondents in the "non-Hispanic, other race" category consumed 73 mg less calcium (p < 0.001) than whites. Among ages 9–18, there was also no statistically significant difference between African-Americans, Hispanics, and whites, but respondents in the "non-Hispanic, other race" category consumed 168 mg less calcium (p < 0.001) than whites.
Among adolescents ages 9–18, consumption of calcium increased 6 mg for each year of age (p < 0.01). Age was not significant in the other age groups.
Family income had a slight positive association with calcium consumption. A 100 percentage point increase in family income relative to the poverty line increased the predicted calcium consumption by about 7–11 mg, depending on the age group. For ages 2–3, 4–8, and 9–18 years, the increase was 12, 7, and 11 mg calcium, respectively. These were modest associations, but they were statistically significant (p < 0.05) for ages 2–3 and 4–8.
|
DISCUSSION |
|---|
One strategy that has been suggested to increase calcium intake is to discourage or restrict consumption of fruit drinks/ades and RCSD. The argument is that these beverages displace milk. If consumption of fruit drinks/ades and RCSD was reduced, the argument goes, fluid milk consumption would increase and therefore calcium intake would be higher. In this argument, fruit drinks/ades and RCSD have an indirect effect on calcium intake through their association with milk consumption.
It is very unlikely that perfect substitution of milk for soft drinks occurs. Milk and soft drinks are not close substitutes. Guenther has estimated a bivariate correlation between consumption of soft drinks and consumption of milk of -0.22, substantially less than the -1.0 that would reflect perfect substitution [34]. In our data, we found that the substitution pattern of RCSD and milk is far less than 1:1, and RCSD consumption explains very little of the variance in milk consumption. The relationship between milk consumption and soft drink consumption is stronger among young children than it is among adolescents. Based on a very simple model regressing RCSD consumption on milk and milk product consumption, an ounce of RCSD reduces consumption of milk and milk products by 0.28 oz. in the 2–3 y age group, by 0.25 oz. in the 4–8 y age group, and by 0.07 oz. in the 9–18 y age group. All of these relationships are statistically significant, but they explain very little of the variance in consumption of milk and milk products. The models explain 1.7 percent of the variance in the 2–3 y age group, 2.7 percent in the 4–8 y age group, and 1.1 percent in the 9–18 y age group. The results for fruit drinks/ades are very similar. Based on these estimates, it does not appear that RCSDs are a strong predictor of milk and milk product consumption, though there is clearly a negative relationship. Eliminating RCSDs completely from the diets of adolescents would only increase predicted milk and milk product consumption by a couple of ounces for the average consumer.
Our results differ from some previous studies [23–27] but are consistent with at least one other [28]. We believe there are two primary reasons that our findings differ from other work on this topic. First, we have used a model that provides comprehensive controls for the rest of the diet. A person’s diet is not a series of isolated food choices—one food choice may affect another. For example, if we choose to have a peanut butter sandwich we may select a glass of milk for our beverage, but if we choose to eat a slice of cheese pizza we select a carbonated beverage. By controlling for the other foods in a person’s diet we have isolated the independent relationship between beverages and calcium consumption and removed potential sources of bias. Second, we have modeled the impact of beverage consumption on calcium consumption when the rest of the diet remains the same. Some other papers have modeled the impact of sweetened beverages substituting for milk on a 1:1 basis. If the rest of the diet remains the same, changes in sweetened beverages have basically no impact on calcium consumption. As we discussed earlier in this paper, the tradeoff between RCSD and milk is far less than 1:1, for adolescents it is less than 10:1. Other models have implicitly assumed that sweetened beverages are substituted for milk and milk products on a 1:1 caloric basis. This assumption is the result of including total energy as a control variable in the models, thus creating an isocaloric restriction in the regression model. If one makes such an assumption about a 1:1 substitution, calcium consumption will obviously be reduced. But this is true for practically any beverage or food, and it is not consistent with the actual substitution patterns that we have observed.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONDATA AND METHODSRESULTSCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
One strategy is offering a variety of milk choices, including flavored milks as suggested by others [35,36] and calcium-fortified juices and fruit drinks through school foodservice and vending machines. Schools offering breakfast programs should offer calcium-fortified breakfast cereals along with reduced-fat milk and calcium-fortified juices. After-school snacks—whether offered at school, home, or daycare—are another way to increase calcium consumption by offering milk products, flavored milks, and calcium-fortified juices and snacks like breakfast cereals and cereal bars.
Finally, for some children and adolescents who can not or will not consume milk products, calcium supplements may be an important way to guarantee adequate calcium consumption. The highest risk population for inadequate calcium intake is adolescent girls, especially African-American girls. Experimental studies have shown calcium supplementation in foods [16] and tablets [37] increases bone mass and bone density in young and adolescent girls. Pediatricians, family physicians, and other allied health professionals should ask parents or their patients about diet choices, especially with regard to consumption of milk products. If reported milk product consumption is low, then the health professional should consider recommending calcium-fortified juices/beverages and foods. If adequate intake cannot be achieved through foods, a calcium supplement should be considered.
Calcium fortification and supplements are not a perfect solution to the inadequate calcium consumption of adolescent boys and girls. Milk and milk products contain other important vitamins, minerals, and nutrients in addition to calcium and would be the preferred means to increase calcium consumption. However, given the low % AI of calcium consumption for adolescents, especially adolescent girls, fortification and supplements should be considered as one practical, targeted alternative to increase calcium consumption.
The health consequences of inadequate calcium intake raise important policy questions about the factors that influence calcium intake. Once these factors are better understood, policies can then be implemented to efficiently and effectively target those segments of the population that need to increase calcium intake.
|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONDATA AND METHODSRESULTSCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
Received September 5, 2002.
Revised May 22, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONDATA AND METHODSRESULTSCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  L. R. Vartanian, M. B. Schwartz, and K. D. Brownell Effects of Soft Drink Consumption on Nutrition and Health: A Systematic Review and Meta-Analysis Am J Public Health, April 1, 2007; 97(4): 667 - 675. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. S Malik, M. B Schulze, and F. B Hu Intake of sugar-sweetened beverages and weight gain: a systematic review. Am. J. Clinical Nutrition, August 1, 2006; 84(2): 274 - 288. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. A. Forshee, P. A. Anderson, and M. L. Storey Changes in Calcium Intake and Association with Beverage Consumption and Demographics: Comparing Data from CSFII 1994-1996, 1998 and NHANES 1999-2002. J. Am. Coll. Nutr., April 1, 2006; 25(2): 108 - 116. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. l. Bronner, A. S. Hawkins, M. L. Holt, M. B. Hossain, R. H. Rowel, K. L. Sydnor, and S. P. Divers Models for Nutrition Education to Increase Consumption of Calcium and Dairy Products among African Americans J. Nutr., April 1, 2006; 136(4): 1103 - 1106. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. A. Forshee and M. L. Storey Reply to Barr and Johnson J. Nutr., May 1, 2005; 135(5): 1337 - 1337. [Full Text] [PDF]
|
|
|
|
|
|
M.-P. St-Onge Dietary fats, teas, dairy, and nuts: potential functional foods for weight control? Am. J. Clinical Nutrition, January 1, 2005; 81(1): 7 - 15. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. A. Forshee and M. L. Storey Controversy and Statistical Issues in the Use of Nutrient Densities in Assessing Diet Quality J. Nutr., October 1, 2004; 134(10): 2733 - 2737. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
