![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
ENVIRON International Corporation, Arlington, VA 22203
2To whom correspondence should be addressed. E-mail: jheimbach{at}environcorp.com
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONREFERENCES |
|---|
KEY WORDS: • dietary supplements • nutrition monitoring • nutritional status • dietary assessment • nutrient bioavailability
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONREFERENCES |
|---|
Of particular relevance is that many but not all dietary supplements have the potential to contribute significantly to the individual’s intake of one or more nutrients. For some dietary supplements, indeed, this is the primary function; multiple vitamin-mineral supplements come readily to mind. Other dietary supplements consist almost entirely of nutrients, such as vitamin C or zinc, but may be used by consumers for purposes—such as prevention of colds or boosting of immune function—that are not essentially nutritional as the term is normally used.
There is considerable interest in tracking consumer use of dietary supplements, and the question immediately arises as to how this can best be accomplished. There are various purposes for tracking consumer use, but the focus of this article is on nutritional and public health interests rather than on marketing or other applications. Most specifically, we address issues regarding use of the information systems that constitute the National Nutrition Monitoring and Related Research Program (NNMRRP)3for this purpose.
The NNMRRP was formally established by the National Nutrition Monitoring and Related Research Act of 1990 (Public Law 101-445), but to a great extent this merely formalized a collection of activities that had developed over many years—as far back as 1909 for the systematic collection of food supply data—for many different purposes. The NNMRRP is defined in the law as encompassing "the set of activities necessary to provide timely information about the role and status of factors that bear on the contribution that nutrition makes to the health of the people of the United States" (7 U.S.C. §5302). It includes five components: 1) dietary, nutritional and health status measurements; 2) food consumption measurements; 3) food composition measurements and nutrient data banks; 4) dietary knowledge and attitude measurements; and 5) food supply and demand determinations.
Although there are several dozen different data systems that are regarded as being part of the NNMRRP, there is a much smaller number that are potentially of use in tracking consumer use of dietary supplements. The keystone surveys of the NNMRRP are the National Health and Nutrition Examination Surveys (NHANES) and the Continuing Surveys of Food Intakes by Individuals (CSFII), both based on data reported by individual respondents in nationally representative samples of households.
The NHANES is conducted by the National Center for Health Statistics,
part of the Centers for Disease Control and Prevention in the
Department of Health and Human Services (CDC 1997
). The first NHANES
survey (NHANES I) was conducted in 1971–1974 after a nutrition
component was added to the earlier Health Examination Survey. NHANES II
was conducted in 1976–1980 and NHANES III in 1988–1994.
NHANES IV entered the field in 1999.
In NHANES III, nearly 34,000 people were interviewed and > 31,000
were examined. The survey obtained food consumption measurements for
1 d (24-h recall) along with dietary, nutritional and health
status measurements. A nonrandom subsample of 
5% of respondents
provided food consumption information for a 2nd d to allow estimation
of usual intakes.
The CSFII is conducted by the Agricultural Research Service of the U.S.
Department of Agriculture (USDA 2000
). The first CSFII was conducted in
1985; it was originally conceived as a smaller survey to fill in
between the large Nationwide Food Consumption Surveys (NFCS), which
were conducted every 10 y. The first of the surveys that evolved
into the NFCS was conducted in the 1930s; modern NFCS surveys were
conducted in 1976–1977 and 1987–1988. The early CSFII in 1985 and
1986 focused on women 19–50 y of age and their children 1–5 y of age;
one sample of men 19–50 y of age was also included. CSFII conducted in
1989–1991 and 1994–1996 included all members of sample households.
Respondents to the 1994–1996 CSFII were asked to provide information on all foods and beverages consumed on 2 nonconsecutive d, both via 24-h recall. A total of 16,000 respondents provided day 1 data and > 15,000 respondents provided data for both days. A 1998 supplement to the CSFII obtained similar data for 5559 children from birth to 10 y of age.
Even before the National Nutrition Monitoring and Related Research Act was enacted in 1990, there had been efforts to improve the correspondence between NHANES and CSFII. It appears that these efforts will reach fruition in 2001 when a unified NHANES/CSFII will commence as the National Food and Nutrition Survey (NFNS). The intentions of the cognizant agencies are that the NFNS will be continuous. That is, rather than having discrete surveys occurring in specific years with blackouts in between, the survey will enter the field on January 1, 2001, and continue indefinitely. Plans call for 10,000 respondents in a nationally representative sample of households to be interviewed every year. Of these, 5000 will be surveyed using computer-assisted telephone interviews and asked to provide 24-h recalls of food and beverage consumption on 2 nonconsecutive d, and 5000 will be surveyed in mobile examination centers and will provide a single 24-h recall of food and beverage consumption along with a variety of dietary, nutritional and health status measurements. A 10% subsample of the latter will provide dietary recall of a 2nd d.
The NHANES and CSFII together provide the best data available in the United States regarding individual consumption of foods and beverages. Through linkage with food composition data from the National Nutrient Data Bank, they allow estimation of individual consumption of substances found in foods, including protein and amino acids, carbohydrates and fiber, lipids, vitamins, minerals, trace elements and other substances with known or suspected nutritional or pharmacological effects. Even when complete information is not available for a substance of interest, approximate intakes can be estimated by linking published food composition data covering primary sources of the substance. ENVIRON and others have done this for such substances as lycopene, beta-sitosterol, inulin and other food constituents.
However, both surveys place relatively little emphasis on consumer use of dietary supplements, despite the fact that these supplements may contribute high proportions of an individual’s intake of certain nutrients. In CSFII, respondents are asked only, "How often, if at all, do you take any vitamin supplement in pill or liquid form? Would you say: every day or almost every day, every so often or not at all?" This single frequency question is then assumed to apply to all forms of dietary supplements to be identified in subsequent questions. NHANES respondents are asked, "Have you taken any vitamins or minerals in the past month?" If the response is in the affirmative, they are requested to identify each supplement taken, and for each one are then asked two questions: "How often did you take [product] in the past month?" and "How much [product] did you take each time you took it?"
As is immediately apparent, neither survey obtains data much beyond the anecdotal level on use of any supplements other than vitamin/mineral supplements, although the NHANES reporting does include respondents’ free mentions of consumption of amino acids or other nonvitamin/mineral dietary supplements. This, of course, dooms any effort to estimate consumption of such substances from diet and supplement use combined. There are currently no plans to expand coverage of dietary supplements in the NFNS scheduled to begin in 2001.
Even limiting consideration to vitamin/mineral supplements, the CSFII approach to dietary supplements allows little more than to classify respondents into those who regularly take supplements, those who occasionally do so and those who do not take supplements. The NHANES method provides a far better estimate of usual intake of nutrients from supplements, but estimating total nutrient intake from foods and supplements combined is still not really possible. The difficulty is that with dietary supplements, unlike foods, small differences may have large impacts because these supplements may have quite high content of vitamins and minerals.
The food consumption surveys have frequently been faulted for generally
underestimating consumers’ food (and, therefore, nutrient) intake. It
is not difficult to show that participants in metabolic ward studies
consume considerably more food than survey respondents report (Mertz et al. 1991
, Briefel et al. 1997
). However, the magnitude of
underreporting of food consumption pales in comparison to the likely
effects of underreporting of use of dietary supplements.
It must be acknowledged at the outset that obtaining detailed information regarding use of dietary supplements in consumer surveys is a daunting task. For foods, with the significant exceptions of breakfast cereals, the surveys have generally disregarded brand names. The implicit assumption has been that, overall, differences between brands of most foods are not nutritionally very significant—that the nutrient content of bacon or of white bread, for example, is sufficiently similar from brand to brand that no great violence is done to estimates of nutrient intake by ignoring brand information.
This has probably never been completely true, and with increasing use of fortification and the emergence of functional foods, it is likely to be even less true in the future. Nevertheless, it is reasonable to conclude that the small increase in precision that might accrue from obtaining brand information in the surveys is not worth the additional expense of data collection and, even more, of collecting and maintaining brand-specific nutrient composition data.
For dietary supplements, this is clearly not the case, because the interbrand variability is often extremely large. Vitamin C supplements, for example, are available at least in 250-, 500- and 1000-mg tablets. Because the average intake of vitamin C from food sources is well under 100 mg/d, if total vitamin C intake from foods and supplements combined is to be estimated, it is imperative that the survey respondent report whether a vitamin C supplement was taken and exactly which one. If one dose of a 500-mg vitamin C supplement is not reported, or if a 500-mg dose is reported when actually a 1000-mg tablet was taken, this is the equivalent of failing to report five glasses of orange juice. This is a fairly serious omission, with far larger potential impact on nutrient intake estimation than any of the food intake underreporting that has been identified.
Additionally, the formulation of dietary supplements challenges the surveyors because different brands of supplements may contain different forms of the same nutrient (such as d-alpha-tocopherol or dl-alpha-tocopherol) that differ widely in potential bioavailability, may use different carriers for nutrients or may contain nutrient combinations that create the likelihood of interactions that may either promote or inhibit absorption.
Although most dietary supplements are properly labeled so that the information regarding the forms and amounts of nutrients present is available, this is not always the case. Further, it is a challenging task for the interviewer to record what may be a large amount of technical information. The National Center for Health Statistics has developed and is working to expand a database of the composition of branded dietary supplements, but this is a truly Sisyphean task. Aside from the rapid proliferation of new products, existing products may change formulations frequently—often in quite significant ways.
Nevertheless, the fact that the magnitude of the challenge is high does not mean that it is not urgent that it be met. Estimates of individuals’ nutrient intakes may be greatly in error if nutrients from dietary supplements are omitted from the calculations, as may be shown by a single example—intake of iron by females.
As shown in Table 1
, women’s mean daily intake of iron from diet alone is considerably
less than their intake from diet and supplements combined. Overall,
total mean daily intake of iron from diet plus supplements is 50%
higher than that from diet alone. The iron added by use of supplements
is particularly significant for pregnant and lactating women: the mean
daily intake of iron from the diet by this group is 16 mg, whereas
the intake from the diet plus supplements approaches 50 mg/d.
|
). However, the 95th percentile intake of iron from diet plus
supplements is more than twice as high, 37.8 mg/d.
|
The NHANES and CSFII surveys, by bringing together food consumption and supplement use data, can also be called upon to determine whether individuals using dietary supplements tend to be those most in need of them or those already relatively well nourished. Again, a single example will suffice to illustrate the type of analysis that can be used.
We determined the vitamin C intake from diet alone of females 9 y of age and older who participated in the NHANES III. We then classified these individuals into three intake groups: those who achieved < 70% of the dietary reference intake (DRI) for vitamin C, those whose intake was between 70 and 100% of the DRI and those who obtained > 100% of the DRI from their diets alone. For each group, we determined the proportion who reported using at least one multivitamin product that included vitamin C and the proportion who reported using at least one single-vitamin preparation of vitamin C. (Those who reported using both single and multivitamin supplements were included in the single-vitamin group.)
The results of this analysis are shown in Table 3
. Although the relationship is not strong, there is a significant
(
df=22 = 16.07, P < .01) trend toward
increasing use of vitamin C supplements by those who already were
better nourished in this regard from their diets alone: use of vitamin
C supplements was reported by 20% of those in the group whose diets
already provided > 100% of the DRI for vitamin C, compared with
only 16% of those who obtained < 70% of the DRI for vitamin C
from their diets. It must be noted, of course, that this single example
may not be generalizable to other populations or other nutrients.
|
Consequently, we suggest that we must critically examine the data needs for dietary supplement ingredients and classify such ingredients into one of three categories: those for which detailed linkage with food/nutrient intake is necessary, those for which a more general linkage is needed and those for which a linkage with food/nutrient intake is not needed. By a detailed linkage with food/nutrient intake, we mean that knowledge is necessary of the dietary context in which the supplement is used—for example, if the supplement is taken along with a meal and the composition of the meal. By a general linkage, we mean that it is necessary to have the supplement use and the food intake data from the same individuals, but the dietary context of the supplement use is less critical.
In the first category would belong those supplement ingredients (such as iron), which interact with the diet—in this case, due to the dietary effects on the bioavailability of the iron in the supplement. In the second category would belong those ingredients—perhaps amino acids would be an example—for which an adequate assessment of intake requires knowledge of both food intake and supplement use.
One possible taxonomy is as follows: vitamins and minerals; other substances found primarily in foods; substances found in foods at low levels and concentrated in supplements; substances not found in foods, but with known or suspected interactions with foods or nutrients; and substances not found in foods and with no known or suspected interactions with foods or nutrients.
The NNMRRP can be a powerful tool, but a tremendous amount of thought and work is required if it is to address the difficult measurement issues posed by dietary supplements and fulfill its objective of providing "timely information about the role and status of factors that bear on the contribution that nutrition makes to the health of the people of the United States."
|
FOOTNOTES |
|---|
3 Abbreviations: NNMRRP, National Nutrition
Monitoring and Related Research Program; NHANES, National Health and
Nutrition Examination Surveys; CSFII, Continuing Surveys of Food
Intakes by Individuals; NFCS, Nationwide Food Consumption Surveys;
NFNS, National Food and Nutrition Survey; DRI, dietary reference
intake.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONREFERENCES |
|---|
1. 7 U.S.C. §5302 (National Nutrition Monitoring and Related Research Act, definitions)
2. 21U.S.C. §321(ff) (Dietary Supplement Health and Education Act, definition of dietary supplement)
3.
Briefel R. R., Sempos C. T., McDowell M. A., Chien S., Alaimo K. Dietary methods research in the Third National Health and Nutrition Examination Survey: Underreporting of energy intake. Am. J. Clin. Nutr. 1997;65:1203S-1209S
4. Centers for Disease Control and Prevention (CDC), National Center for Health Statistics(1997)Third National Health and Nutrition Examination Survey (NHANES III) Public-Use Data Files [CD-ROM], data and documentation. National Technical Information Service, Accession No. PB97-502959.
5.
Mertz W., Tsui J. C., Judd J. T., Reiser S., Hallfrisch J., Morris E. R., Steele P. D., Lashley E. What are people really eating? The relation between energy intake derived from estimated diet records and intake determined to maintain body weight. Am. J. Clin. Nutr. 1991;54:291-295
6. US Department of Agriculture, Agricultural Research Service(2000)1994–96, 1998 Continuing Survey of Food Intakes by Individuals [CD-ROM], data and documentation. National Technical Information Service, Accession No. PB2000-500027.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
