Diet and Health: Implications for Reducing Chronic Disease Risk (1989)

Page 41

3—
Dietary Intake and Nutritional Status: Trends and Assessment

Throughout most of history, the quest for sufficient food was the chief occupation of the earth's people. The diet of Paleolithic hunter-gatherers, before the development of agriculture, is believed to have consisted of approximately 35% meat and 65% plant foods; no dairy products and practically no cereal grains were consumed. Meat from wild animals contains low levels of fats (4% in this early diet compared to 25 to 30% fat in today's domesticated animals), and the plant foods in this early diet consisted of a variety of vegetables and fruits (Eaton and Konner, 1985). The high-meat diet resulted in a high protein intake, but dietary fat was relatively low and contained more polyunsaturated fats than saturated fats. The intake of cholesterol, dietary fiber, calcium, and ascorbic acid is believed to have been high, but sodium intake was remarkably low. The accuracy of these estimates of the diet of hunter-gatherers cannot be established, however.

Two notable revolutions caused major changes in food supplies. The first occurred around 10,000 B.C., when people began to give up their nomadic ways in favor of living on specific plots of land, existing chiefly on plants they grew and animals they domesticated. For the first time, dairy products and cereal grains became a part of the diet. Agricultural innovations evolved slowly at first, but accelerated greatly with the onset of the second important revolution—the Industrial Revolution of the 1800's. Industrialization gave rise to two new socioeconomic classes: a new middle class of merchants and managers, who demanded a variety of socially desirable foods, and a new class of industrial workers, who could afford only the cheapest foods. Although the poverty, poor sanitary conditions, malnutrition, and disease that prevailed among workers in the industrial cities and towns was a blight on the Industrial Revolution, resources were soon mobilized to meet the food demands of the middle classes. Eventually the poor also benefited, as increased production and new  techniques made cheaper foods available to them (Tannahill, 1973).

When large numbers of people left farming to work for wages in factories or to become entrepreneurs, there was a marked change in the kinds and quantities of food that were readily accessible. In the years since the Industrial Revolution, the U.S. diet has again undergone very large changes. In 1800, 95% of all Americans consumed minimally processed foods produced chiefly on their own small farms, but by 1900, only 60% of the population remained on farms (Hampe and Wittenberg, 1964). In less than 175 years, nearly all Americans have become dependent on others to produce and distribute food to supermarkets where their ability to obtain items they desire is determined largely by their financial resources.

The construction of railroads across the country in the mid-1880s was responsible for changes in

Page 42

the character of the food supply. Foods were no longer strictly seasonal in nature, because they could be shipped from different climates. This trend accelerated with the advent of refrigerated railcars and trucks. Innovations in food processing were also important. In 1869, processed foods consisted chiefly of milled flour and cornmeal, refined sugar, cured meats, and processed dairy products. Today, in addition to these foods, the consumer finds canned, frozen, fermented, and dehydrated foods, as well as foods fabricated in the laboratory to resemble traditional foods. These include drinks resembling fruit juices, but containing no fruit juice, and analogs of meat or fish made from soybeans or wheat gluten. Innovations such as sugared breakfast cereals and a variety of snack items were unheard of before World War II. Hampe and Wittenberg (1964) estimate that 60% of the items on supermarket shelves in 1960 came into existence during the 15 years after the end of World War II. Home refrigerators and freezers also increased the homemaker's ability to select and store a variety of foods. Today's large supermarkets carry as many as 15,000 different items from which consumers must choose, complicating the task of nutrition educators.

The next section focuses on changes in the food supply during the twentieth century and describes national surveys to determine the U.S. population's intake of foods, nutrients, and, to a limited extent, pesticides and industrial chemicals. This is followed by a discussion of the limitations of the studies and a section on consumption trends.

National Surveys of Dietary Intake and Nutritional Status

Surveys Conducted by the U.S. Department of Agriculture (USDA)

The Food Supply: Historical Data

Changes in foods available to the public from 1909 to the present have been ascertained from USDA data based on the disappearance of foods into wholesale and retail markets. Annually, foods available to the civilian population are estimated by subtracting data on exports, year-end inventories, nonfood use, and military procurement from data on total production, imports, and beginning-of-the-year inventories. These quantities are larger than those actually consumed, because they fail to take into account losses that occur during processing, marketing, and home use. Since they do not represent actual consumption, they are referred to here as availability or use of foods or nutrients.

The USDA estimates per-capita use of foods or food groups by dividing total available food by the population of the 50 states and the District of Columbia. The nutritive value of the food supply is calculated from per-capita use by using nutritive values found in food composition tables. Although these data provide no information on how foods are distributed among individuals or population groups, or on changes in patterns of waste and other losses, they nevertheless reflect changes in overall patterns of foods available over time. Furthermore, these data are similar to data produced in many other countries, and they have been useful in epidemiologic research across countries, such as studies of dietary lipids and atherosclerotic diseases (Stamler, 1979).

The Nationwide Food Consumption Surveys (NFCS)

NFCS focuses on the food use of households and the dietary intakes and patterns of individuals. These surveys have been conducted approximately every 10 years since 1935 by USDA's Human Nutrition Information Service (HNIS), but the first four surveys (in 1935, 1942, 1948, and 1955) obtained information only on household food use over a 7-day period. These data reflect food use in an economic sense only and do not take into account food waste or how food is distributed among household members. Beginning in 1965, data have been collected on intakes by individuals. Surveys were conducted in 1965-1966 and in 1977-1978; separate surveys were conducted in 1977-1978 in Puerto Rico, Alaska, and Hawaii, and among low-income and elderly populations. The most recent NFCS, 1985 and 1986, were the Continuing Surveys of Food Intakes of Individuals (CSFII), designed to be conducted annually. The household screening procedures for CSFII were designed to provide three separate samples: (1) women 19 to 50 years of age and their children 1 to 5 years of age—the core group; (2) a similar age sample of low-income women and children; and (3) men 19 to 50 years of age. Data have been published on both the 1985 and 1986 surveys (USDA, 1985, 1986a,b, 1987a,b,c, 1988).

Surveys Conducted by the U.S. Department of Health and Human Services (DHHS)

In the Ten-State Nutrition Survey conducted during 1968-1970, DHHS studied low-income

Page 43

populations in 10 states (DHEW, 1972). In the biennial Food Label and Package Survey (Woteki, 1986), DHHS studies a statistically representative sample of packaged food products to obtain information on ingredients and on the extent of nutrient labeling. Two other DHHS studies—the Total Diet Study, conducted by the Food and Drug Administration (FDA), and the National Health and Nutrition Examination Survey (NHANES), conducted by the National Center for Health Statistics (NCHS)—are of greatest interest in the present report.

Total Diet Study

The only national system for studying average intakes of pesticides, toxic substances, radionuclides, and industrial chemicals is FDA's Total Diet Study. That study also provides estimates of dietary intakes of certain essential elements: iodine, iron, sodium, potassium, copper, magnesium, and zinc. The extent to which selected age-sex groups (males and females age 6 to 11 months, 2 years, 14 to 16 years, 25 to 30 years, and 60 to 65 years) are exposed to harmful substances and to essential minerals through diet can be determined from the results of this annual study.

Four times a year, foods representative of U.S. diets are purchased in grocery stores across the nation and are individually analyzed in FDA laboratories for the constituents mentioned above. The food items used in the Total Diet Study through April 1982 were based on data from the 1965 NFCS. Since 1982, the food items have been based on data from the 1977-1978 NFCS and the Second National Health and Nutrition Examination Survey (NHANES II), conducted during 1976-1980 and described below. Revisions to the list of food items have been described by Pennington (1983).

An example of findings from the Total Diet Study was the observation that iodine was present in the food supply in larger-than-recommended amounts, chiefly because of a higher-than-usual iodine content of milk and cereal grain products (Park et al., 1981). These findings are discussed in greater detail in the Minerals subsection of Trends in the Food Supply and Dietary Intakes, below.

Another important finding of that study in the early 1970s was that polychlorinated biphenyls (PCBs) were migrating into foods through paperboard packaging. Such packaging materials were immediately banned. Since that time, PCBs have been detected in this study only in minute amounts and then only sporadically (E. Gunderson, FDA, personal communication, 1987).

A 1987 survey conducted by the Food Marketing Institute indicated that 76% of the food shoppers questioned believed that pesticides in foods constitute a ''serious hazard" (Food Marketing Institute, 1987). No pesticide examined in the Total Diet Study as far back as 1961 has been found in the diet above tolerance levels. However, the FDA's laboratory methods did not permit analyses of all pesticides that might contaminate foods, and so few samples were taken that rare but high contamination levels could be missed entirely. According to a National Research Council (NRC) report, 71 to 80% of pesticides on U.S. markets have been insufficiently tested for carcinogenesis, 90% have never been tested for damage to the nervous system, and 50 to 61% have not been tested for teratogenicity (NRC, 1984).

A 1987 report from the NRC Board on Agriculture pointed out that government regulation of herbicides, fungicides, and insecticides needs to be greatly improved to protect consumers from cancer risks due to the presence of these contaminants in food. Consistent standards are not applied to old and new pesticides, with the result that continued use of some pesticides is permitted, despite the fact that newer alternative compounds posing smaller cancer risks are available (NRC, 1987).

National Health and Nutrition Examination Survey (NHANES)

NHANES is conducted by NCHS, in part to monitor the overall nutritional status of the U.S. population through health and medical histories, dietary interviews, physical examinations, and laboratory measurements. Information is obtained about many medical conditions, including nutrition-related disorders. Among these are obesity, growth retardation, anemia, diabetes, atherosclerotic cardiovascular diseases, hypertension, and deficiencies of vitamins or minerals. NHANES I was conducted between 1971 and 1974, NHANES II between 1976 and 1980, and the Hispanic HANES (HHANES) between 1982 and 1984. NHANES III, which began in 1988, includes a potential for following people throughout their lives, surveying them at regular intervals, and using a national death certificate system to establish cause and date of death. It may also be possible in NHANES III to reexamine respondents to earlier studies. During 1982-1984, reexamination of respondents who were 25 to 74 years old during NHANES I provided a unique research opportunity for epidemiologists (Madans et al., 1986).

Page 44

Nutrition Monitoring in the United States

National Nutrition Monitoring System (NNMS)

In 1977, Congress directed USDA and DHHS to integrate their surveys, and by 1981, the two departments had developed a Joint Implementation Plan for a National Nutrition Monitoring System (NNMS). The plan was to design a system for coordinating survey methods and reporting survey findings to Congress through reports from the Joint Nutrition Monitoring Evaluation Committee (JNMEC) established by the two departments in 1983. In the first report, issued in 1986 (DHHS/USDA, 1986), food intake data from the 1977-1978 NFCS, biochemical analyses from NHANES II (1976-1980), and USDA's historical food supply data were used to determine food components of public health importance. That report categorizes some food components as "warranting public health monitoring priority status." Those components are discussed below in the section on Trends in the Food Supply and Dietary Intakes. Additional details on certain aspects of NNMS were recently published (DHHS/USDA, 1987), and the second JNMEC report will be published in 1989. USDA and DHHS are also working together to coordinate their survey methods, to publish results promptly, to conduct the NFCS more frequently, and to add longitudinal aspects to data collection in NHANES.

Coordinated State Surveillance System (CSSS)

The Centers for Disease Control (CDC) contribute to nutrition monitoring through CSSS, in which the nutritional status of the high-risk pediatric population and pregnant women is monitored on the basis of information obtained from service delivery programs operated by selected state and metropolitan health jurisdictions. The CSSS provides information about the prevalence of overweight, underweight, retarded growth, and anemia among high-risk children. Among pregnant women, data are gathered on anemia, abnormal weight changes, fetal survival, birth weights, and infant feeding practices. In 1986, 34 states, the District of Columbia, and Puerto Rico participated in the pediatric survey, and 14 states, the District of Columbia, and Puerto Rico participated in the pregnancy survey.

Limitations of NFCS AND NHANES

NFCS and NHANES systematically provide valuable data on the dietary and nutritional status of Americans. Through them, desirable and undesirable trends in dietary patterns can be monitored, and data can be used to evaluate the need for group interventions. The limitations in methods used and differences in the design of the two surveys, however, influence the interpretations and conclusions that can be drawn from the data.

Populations Represented by NFCS and NHANES

The sampling units in NFCS and NHANES are households and individuals within those households; the samples are designed to represent the civilian, noninstitutionalized population. Excluded are the homeless and residents of hotels, rooming or boarding houses, dormitories, Indian reservations, military posts, prisons, hospitals, and residential treatment centers for drug addiction, alcoholism, and obesity. Clearly these surveys were not designed to represent the entire U.S. population. Omission of homeless and noninstitutionalized people underrepresents the population at greatest risk of nutritional deficiencies, but the magnitude of this bias has yet to be determined (NRC, 1986).

Data from NFCS generally are drawn from the 48 conterminous states; data from Alaska and Hawaii, when obtained, have come from separate surveys and are reported separately. NHANES II, on the other hand, included Alaska and Hawaii. NHANES oversamples low-income groups, but in the past included no respondents over age 75. A recent NFCS of low-income women and their children has been completed as part of CSFII.

How well do NFCS and NHANES represent the civilian, noninstitutionalized population? The 1977-1978 NFCS was based on a stratified area probability sample of households in the 48 conterminous states. Of the 20,812 households in the original sample, 14,930 (72%) completed the household questionnaire. Furthermore, of eligible individuals in participating households, 81% provided 3 days of dietary information. It is important to know  whether nonparticipating households (28% of the original sample) differ greatly from participating households. Similarly, there would be reason for concern if the additional 19% of eligible people in participating households who failed to contribute 3-day food intake data were (for example) less educated or had lower incomes (NRC, 1986). Weighting factors were applied to make the sample more representative; however, in the absence of studies to determine the bias of nonresponders, the extent to which the 1977-

Page 45

1978 NFCS is representative of the noninstitutionalized population cannot be known.

The 1985 CSFII was designed to include a stratified area probability sample in the 48 conterminous states. The first sample of women was obtained from 1,893 households containing at least one age-eligible woman (19 to 50 years old); 1,341 households (71%) provided usable data. For men 19 to 50 years old, one sample was drawn from all income levels, and an independent sample was drawn to represent households containing at least one man in that age group whose income was at or below 130% of the poverty guidelines. Of the 744 all-income sample households, 71% participated, compared with 67% of the 149 low-income sample households. USDA interviewed neighbors to determine whether people who did not participate in the 1985 CSFII differed from participants (R. Rizek, USDA, personal communication, 1987). Weighting factors were applied to adjust for nonresponders.

The response rate for NHANES is tabulated for people who participated in the initial interview held in the home, when medical histories and sociodemographic data were obtained, and who furnished dietary information and were physically examined in mobile examination centers. In NHANES I (1971-1974), 74% of the original sample completed both parts of the survey. Two reinterview surveys were conducted to determine reasons for nonparticipation (Forthofer, 1983). Of 27,801 people in the original NHANES II sample, 25,286 (91%) participated in the initial interview and 20,322 (73%) completed all aspects of the study. The investigation of potential nonresponse bias was more extensive than for NHANES I and indicated that the poststratification and nonresponse adjustments made by NCHS removed most factors that were potential sources of bias (Carroll et al., 1983; Forthofer, 1983). It was possible to do only a limited analysis of the characteristics of the 9% of the original sample who failed to participate in even the first interview. Even when studies of nonrespondents indicate minimal bias, individual users of these data should be aware that data are missing on 27 and 28% of the samples in NHANES II and the 1977-1978 NFCS, respectively, and 29 to 33% of the sample in the 1985 CSFII (NRC, 1986).

Comparability of NFCS and NHANES Data

Data from these two surveys are difficult to compare because of differences in the survey designs and methods of data collection. These differences lie chiefly, but not exclusively, in methods of estimating dietary intake, food composition data used to estimate nutrient intake, and standards used for determining dietary adequacy.

The 1-day (24-hour) dietary recall method has been used in NFCS and in NHANES. In the 1965-1966 NFCS, food intake data were recalled, not necessarily by the specific individual identified to respond to the questions, but by one designated respondent in each household, usually the person who shopped for and prepared the food. In the 1977-1978 NFCS, all eligible people in a household (except for their small children whose intakes were reported by adults) recalled their intake on the day before the interview and then recorded their intake on the day of the interview and on the following day, thus providing information on intake for 3 consecutive days. Standard measuring cups, spoons, and a ruler were used as aids in estimating quantities of foods consumed. Interviews were conducted on all days of the week; therefore, the data for some people included weekend days. Studies show that dietary intake on weekends may differ from  that on  weekdays (Acheson et al., 1980; Beaton et al., 1979; Richards and Roberge, 1982).

In the 1985 CSFII, data were collected by using the 1-day dietary recall only. Data on women and their children were obtained on 6 separate days throughout the year at intervals of approximately 2 months. The initial 1-day recall was obtained by in-person interview; later, recall data were obtained by telephone. At the time of this writing, data on only 4 nonconsecutive days of intake have been published. Men in the 1985 CSFII supplied recall data by in-person interview for only 1 day.

Respondents in the 1965-1966 and the 1985 NFCS were not notified in advance about the interviews, but they were notified by letter in the 1977-1978 NFCS. Notification ahead of time may have resulted in more accurate recalls or in modification of dietary intake, since the letter suggested that respondents should begin to keep records of foods purchased for the household. This could have affected some people's ability to recall their food intake.

An important change in method was introduced in the 1985 CSFII. Interviewers probed specifically to learn brand names of processed foods; whether fat on meat or skin on poultry was consumed; whether salt or butter was added to food during

Page 46

cooking or at the table; whether food items were eaten during cooking or cleaning up; and if snack items or beverages had been forgotten. Although these changes were made for appropriate reasons, one needs to keep these differences in mind when attempting to use NFCS data to study changes in food consumption over time. In addition, we do not know the extent to which differences in NFCS and NHANES estimates of nutrient intake are due to differences in the extent to which interviewers probed for specific information.

In both NHANES I and NHANES II, 1-day recalls were obtained through in-person interviews conducted in a mobile unit, where laboratory measures (e.g., blood pressure) were also obtained. Three-dimensional food models were used as aids in estimating quantities consumed. Dietary interviews were conducted in such a way that 1-day intakes were for weekdays only, and the method of obtaining intakes was the same for NHANES I and II. Respondents supplied information on the frequency with which foods had been consumed during the preceding 3-month period. This was an attempt to ascertain the usual pattern of food consumption. In both NHANES I and II, a letter announcing the survey was sent to each household 1 week before the first interview, but the letter informed respondents only of the general purpose of the survey and said nothing about the plan to obtain information on food intake.

Research is needed on the extent to which differences in  methods used by NFCS and NHANES affect the results reported. These differences include days of the week for which data are collected; the use of probing questions and methods to aid respondents in estimating portion sizes; assignment of food codes; and privacy during the interview. In NFCS, other household members could have been present. In NHANES, only the respondent was present with the interviewer (Woteki, 1985).

One of the most difficult tasks in nutrition research is documenting the actual or habitual food and nutrient intake of individuals or groups. A single 24-hour recall cannot be used to estimate the habitual intake of a person, although it can be used to estimate the average intake of a group. Problems involved in estimating food intake are discussed at length in Chapter 2 and are not repeated here.

A recent publication comments on errors in reporting dietary intake and on differences in the distributions of intakes reported in several recent large surveys (NRC, 1986). These differences were inconsistent across nutrients and suggested bias in either food intake estimates or food composition data. Systematic biases can affect estimates of nutrient intake; for example, the failure of the 1977-1978 NFCS to record supplement use resulted in underestimating the intake of vitamins and minerals.

In both NHANES and NFCS, food composition data (discussed below) are used to calculate the nutritive value of food consumed by respondents. The nutrients and dietary fiber reported by the major national surveys are shown in Table 3-1.

The major repository of nutrient composition data for individual foods in the United States is USDA's Nutrient Data Bank (NDB), sections of which have been published in a revision of USDA Agriculture Handbook No. 8, Composition of Foods—Raw, Processed, and Prepared. Sources of data for the NDB include studies in the scientific literature, unpublished reports from federal government and university laboratories, studies contracted by USDA, and data from industry for foods bearing nutrition labels. Generally, NDB data used in the 1977-1978 NFCS pertained to nutrients in food as purchased (not as actually consumed) and therefore did not account for losses or modifications due to preparation or processing. In addition to the NDB data, NHANES has also used data from industry on the composition of new food products and brand-name products of unique formulation. The data bases of both surveys are updated as new information is obtained. For example, for the 1985 and  1986 CSFII, corrections were made for changes in moisture and fat and for retention of nutrients during preparation (F.N. Hepburn, USDA, personal communication, 1987).

Little is known about the quantity of some nutrients and nonnutrients in foods because of inadequate analytical methods. The presence of nonnutrients, such as dietary fiber, in foods has sparked scientific interest, but data on these food constituents are not yet complete. Table 3-2 makes it clear that there are large gaps in food composition data. When possible, amounts of nutrients in foods noted in the tables come from actual chemical analyses, but if such data are unavailable, the amounts in the table are imputed. For nutrients that have been tracked for a long time, such as calcium and protein, the proportion of analytical data in the tables, as opposed to

Page 47

Page 48

imputed data, is high, but the proportion is low for nutrients recently added to USDA surveys, such as dietary fiber and a-tocopherol (Hepburn, 1987). Table 3-2 indicates, for example, that only 64% of the data on vitamin B₆ in all foods in the data set are analytical values, but a higher proportion (72%) of the data on foods that are best sources of vitamin B₆ represent analytical as opposed to imputed data.

Although there are ample data on nutrients in commodities, there is little information on highly processed or manufactured foods such as snack foods, baked products, convenience foods, restaurant meals, fast foods, and frozen dinners. In addition, knowledge is limited  regarding the amounts of vitamin B₆, pantothenic acid, folacin, vitamin E, zinc, copper, magnesium, manganese, chromium, and selenium in foods (Beecher and Vanderslice, 1984). Information on other substances of interest, such as carotenoids and dietary fiber, is rapidly accumulating. Lack of information on nonnutrients is of particular concern for those studying the relationship between diet and cancer.

Food composition tables necessarily show typical values, but the nutrient composition of a specific food portion depends on many factors; for example, the composition of fresh fruits and vegetables depends on the variety, extent of exposure to sun, maturity, and transport and storage conditions. Biases in food composition data may result when inappropriate analytical methods are used (e.g., such as certain methods used in gathering data on fiber) and when a food item is incorrectly identified (e.g., as skim milk instead of whole milk) (NRC, 1986).

In both NHANES and NFCS, the nutrient values are calculated by matching each reported food to a description in the survey's food composition data base, then selecting and assigning the appropriate food code. In NHANES II, interviewers coded the responses on site. In NFCS, coding was done at a central location by persons other than the interviewers. The extent to which this methodological difference in assigning the codes resulted in different estimates of nutrient intakes in the two surveys is unknown (Woteki, 1985).

Use and Misuse of Standards of Dietary Adequacy

Standards used in national surveys to judge dietary adequacy differ. In NFCS, the 1980 Recommended Dietary Allowances (RDAs) were used. NHANES has its own set of standards that are developed by an ad hoc advisory group and differ from the RDAs in several ways; in particular, standards for vitamin A and calcium are lower.

RDAs have been set for protein, certain vitamins (A, D, E, B₆, B₁₂, thiamin, riboflavin, niacin, and folacin), and certain minerals (calcium, phosphorus, magnesium, iron, zinc, and iodine). Because data were inadequate in 1980 to set RDAs for other nutrients, ranges of Estimated Safe and Adequate Daily Dietary Intakes (ESADDI) were given for three vitamins (biotin, pantothenic acid, and vitamin K) and for several minerals (copper, manganese, fluoride, chromium, selenium, molybdenum, sodium, potassium, and chloride).

The RDAs have been often used to interpret survey data on food intake obtained by a single

Page 49

24-hour recall or by 1- to 3-day food intake records. In discussing the results of such studies, investigators may use cutoff points to report the number or percentage of respondents whose intakes fall within specific percentiles or cutoff points of the RDA, e.g., two-thirds or 70% of the RDA for a nutrient, implying that certain segments of the respondent population have inadequate intakes of the nutrient, or that they are "at nutritional risk." Because these cutoff points are arbitrary, this practice leads to incorrect estimates of the frequency of adequate and inadequate intakes in the population and is based on misunderstandings of the appropriate uses of the RDAs and of a single 24-hour recall or 3-day record in food intake studies.

The RDAs are not requirements below which deficiency diseases are apt to develop. Rather, for many nutrients they are set at sufficiently high levels to cover the needs of practically all healthy people. Since individuals differ in their requirements for specific nutrients, however, it is impossible to know from a dietary survey which person requires at least the RDA and which one requires less or possibly even more. Therefore, all cutoff points are misleading, even if the dietary method used provides accurate data on the customary intake of each person. Many people who rank below the cutoff point actually have adequate intakes because they require less, whereas some above the cutoff point have too low an intake to meet their needs. The magnitude and direction of the errors involved in the misuse of the RDAs are not known (NRC, 1986).

As pointed out in Chapter 2, because of the enormous day-to-day variation in the amounts and kinds of foods eaten by one person, 1 or 3 consecutive days of intake are not representative of usual or customary intake over an extended period. Thus, although a single 24-hour recall or 1- to 3-day record, if carefully done, may be useful in assessing the average or median intake of a population group, their use in ranking individuals is inappropriate (Garn et al., 1978; Hegsted, 1972).

In summary, NFCS and NHANES dietary data obtained at different times must be compared cautiously and with full knowledge of the differences in methods used to gather and summarize data. Food intake data can be compared with somewhat less difficulty than nutrient intake data.

NFCS and NHANES samples are sufficiently large to detect major public health problems but not to uncover clinical illness scattered throughout the population. A deficiency or other nutritional disorder would have to affect approximately 1% of the population—about 2 million people—to be reasonably sure of being detected by these surveys (DHHS/USDA, 1986).

USDA and DHHS recognize the problems described and are attempting to devise solutions to improve  comparability of the two surveys. NHANES III, to be conducted between 1988 and 1994, will oversample (i.e., sample more people in a subgroup than warranted by their percentage in the general population) the elderly, blacks, Hispanics, and the very young. Data for the 19871988 NFCS are currently being gathered, and CSFII will cover all sex and age groups beginning in 1989. In planned nutrition monitoring activities, data on dietary intakes will be collected through 1996. Efforts will be made to expedite publication of data from these surveys (DHHS/ USDA, 1987).

Problems in Assessing Nutritional Status

Nutritional status has been defined as an individual's health condition as it is influenced by the intake and utilization of nutrients (Todhunter, 1970). In theory, optimal nutritional status should be attained by consuming sufficient, but not excessive, sources of energy, essential nutrients, and other food components (such as dietary fiber) not containing toxins or contaminants.

Traditionally, efforts to detect poor nutritional status have centered on nutritional deficiencies in populations, since defining or assessing optimal health is difficult. Nutritional deficiency follows a pattern starting with low intake or utilization of one or more nutrients, then progressing to biochemical abnormalities, abnormal growth, abnormal body mass, and, eventually, to full-blown deficiency. Poor nutritional status is not confined to undernutrition. It may also result from excessive intake or inadequate expenditure of food energy, or from excessive intakes of specific nutrients, resulting in acute toxicity or chronic diseases.

A major problem in interpreting national dietary surveys and their relationship to nutritional assessment for populations, especially with regard to chronic diseases, is the use of fixed cutoff points, such as a fixed percentage of the RDAs, as criteria for judging the adequacy of dietary intakes. The Food and Nutrition Board Subcommittee on Criteria for Dietary Evaluation (NRC, 1986) proposed that multiple criteria be used for assessing adequacy of dietary intake contingent on the intended outcome. Thus, adequate intake levels might range progressively from those required to maintain high

Page 50

tissue concentrations of a nutrient to lower levels needed to just maintain normal metabolic functions or to still lower levels required to prevent clinical deficiency (NRC, 1986). Clinical or laboratory indicators would be developed for each level of nutriture so that they could be used in population assessments.

Cutoff points frequently are used in nutritional assessment studies, but as noted earlier in relation to dietary intake, no single cutoff can separate adequately nourished people from those with nutritional deficiencies. As Figure 3-1 illustrates, regardless of the cutoff used, the nutritional status of some people will be erroneously classified as deficient, whereas some individuals with nutritional deficiencies will be classified as adequately nourished. The Food and Nutrition Board Subcommittee on Criteria for Dietary Evaluation proposed that problems stemming from use of a cutoff point might be overcome if the distribution of nutrient requirements was compared with the distribution of nutrient intakes in a sample population (NRC, 1986). This probability approach would allow investigators to make better estimates of the prevalence of inadequate intakes in a population, but would still not permit the identification of people with adequate nourishment. The success of this method will depend on development of more accurate estimates of the mean requirement for each nutrient and its variability in the population, as well as the improvement of methods of assessing dietary intakes. Analogous conceptual approaches may be used for biochemical assessment of nutritional status (Beaton, 1986).

Many investigators have used regression or correlation analyses to examine the relationship of dietary intake to biochemical or other indicators of nutritional status within populations, and some have reported no relationship or only a weak one. Beaton (1986) identified several possible causative factors for this. First is the failure to determine usual dietary intake, which cannot be accomplished by obtaining only one 24-hour recall from each respondent (as was done in NHANES I and II) because of the large day-to-day variability in intake by individuals (Block, 1982). Second is the known biologic variability in nutrient requirements and laboratory indicators at a given level of nutritional status. Third is the variable sensitivity of some nutritional status indicators across different levels of nutriture. For example, as iron stores increase with high iron intakes, hemoglobin is no longer a sensitive indicator of iron nutritional status. Thus, only if nutritional status among

FIGURE 3-1
Difficulties when cutoff points are used to assess nutritional status. Eleven grams
of hemoglobin per deciliter is an arbitrary cutoff for assessing iron status. The lower 
left-hand curve represents the distribution of hemoglobin levels among individuals known to
respond to increased iron intake. The upper right-hand curve represents the distribution among
those known to have adequate iron intakes. The cross-hatched area above 11 g represents
individuals who are anemic but classified as normal by this cutoff point; the stippled
area below 11 g represents those classified as anemic by this cutoff, but who are not
responsive to increased iron intakes. Since the two distributions overlap, no single cutoff
point can separate adequately from inadequately nourished individuals. From Beaton (1986).

subjects varies greatly would it be possible to demonstrate a strong relationship between intake and the laboratory indicator. If all subjects are adequately nourished with respect to some nutrient, the variability in amounts of the nutrient required and the normal physiological variability that subjects show in the laboratory indicator may obscure any association between intake and the laboratory indicator. Beaton (1986) notes that much of the controversy regarding relationships between dietary intake and nutritional status is due to flawed concepts, which can obscure relationships when they exist and produce spurious evidence of relationships when they do not. New conceptual frameworks are needed to overcome these problems.

In the past, evaluations of nutritional status have focused chiefly on criteria for prevention of nutrient deficiencies. Today, there is substantial interest in  the  association  of nutrition  with chronic diseases. Excessive intakes of nutrients can work through normal biologic or metabolic functions to produce some chronic diseases or risk factors for disease. The Food and Nutrition Board Subcommittee on Criteria for Dietary Evaluation

Page 51

FIGURE 3-2
The curve on the left, based on a cumulative distribution of requirements, indicates increasing risk of inadequate intake.
The curve on the right indicates increasing risk of excess. The safe range of intake is between these levels. From Beaton (1986).

(NRC, 1986) noted that the probability approach can be used to analyze excessive intakes of nutrients or food components just as it can be used to assess inadequate intakes (Figure 3-2). To date, little attention has been paid to the frequency distribution of intakes that are or may be detrimental; this must be remedied before this method can be used. Other relationships between diet and disease may not fit into this conventional conceptual view, because metabolic functions—which form the basis of the conceptual framework described above—may not be involved (Beaton, 1986).

Data gathered to assess nutritional status, as in NHANES, are of limited value in evaluating possible relationships between diet and many cancers because of the long latency periods between exposure and clinical manifestation. In most cancers—for example, breast cancer—information about intake during the distant past, rather than present intake, is needed. Problems in obtaining such information are discussed in Chapter 2. Furthermore, causative or protective dietary constituents—such as nonessential trace elements (cadmium), contaminants (aflatoxins, pesticides), and carotenoids with no vitamin A activity—may be among many substances that are included in surveys and current food composition tables. Methods by which foods are stored or prepared may play a role in the causation of cancer, but such data are not generally collected through NFCS or NHANES.

Objectives and priorities for research on diet and cancer have been identified by a National Research Council committee in its report Diet, Nutrition, and Cancer: Directions for Research (NRC, 1983). Among other priorities, that committee noted a pressing need for short-term tests to identify early biologic indicators of exposure to dietary constituents that affect carcinogenesis. The complexities of studying the relationships of diet to cancer are just beginning to be appreciated, but it is already obvious that studies must be designed specifically to test hypotheses regarding this relationship.

NHANES and Nutritional Assessment

NHANES is the only national survey providing information on the nutritional status of the population. Four types of NHANES data are of importance in studies of diet and health: dietary intake, to determine kind and amount of food consumed and its nutrient quality; anthropometry, to describe growth and body weight patterns; biochemical tests, to determine nutrient levels in blood and urine; and clinical examinations, to detect signs of nutritional problems. NHANES data have been used to evaluate the proportion of the population at risk for deficiencies of vitamin A, vitamin C, folate, iron, zinc, and protein.

Experience gained in NHANES I led to changes in the collection of biochemical data in NHANES II. For example, because vitamin A deficiency was judged not to be a problem among older age groups in NHANES I, biochemical values for this vitamin were obtained only for children 3 to 11 years of age in NHANES II. The current interest in vitamin A relative to cancer risk had not yet emerged. Blood levels of zinc and copper were obtained in NHANES II but not in NHANES I, and assessment of anemia was intensified in NHANES II by adding several biochemical tests and gathering additional information in the medical history and physical examination. No attempt was made in either NHANES to gather data on toxins and contaminants in the diet.

Problems in Using NHANES Data to Study Diet-Chronic Disease Relationships

Although NHANES was designed to examine nutrition and health status in the United States—not to study hypotheses regarding diet and chronic diseases—some investigators have used NHANES

Page 52

data to evaluate dietary intake and certain risk factors for cardiovascular diseases (e.g., blood pressure and serum cholesterol levels) (Harlan et al., 1983, 1984; McCarron et al., 1984; Sempos et al., 1986). Although NHANES data are extensive and derive from a broad range of measurements, there are some limitations, which affect attempts to use NHANES data in the study of chronic diseases. The following discussion of limitations is based on publications by Yetley and Johnson (1987) and by Murphy and Michael (1982).

NHANES provides only cross-sectional, periodic data, which are not suitable for studying causal relationships. Only longitudinal studies can supply data appropriate for determining causality. Furthermore, because many respondents are under treatment for medical conditions at the time of the survey, they may have altered their dietary habits and thereby certain biochemical parameters. In a longitudinal study, their prediagnostic dietary and biochemical data would be known, making possible more accurate tracking of the influence of dietary intake on the disease under study. Such people must be excluded from analyses of relationships between diet and disease when using NHANES data.

Certain measurements needed for such studies of diet and chronic diseases may not have been made in NHANES, or the methods used may have been inappropriate. Furthermore, biochemical measurements are usually evaluated in NHANES according to specified cutoff points, which are as inappropriate for studies of chronic diseases as they are for judging dietary adequacy. Erroneous estimates of nutritional risk can result if there are substantial differences in physiological status at a given biochemical level (e.g., blood level of nutrients).

Investigators who use NHANES data in studies of chronic diseases need to evaluate carefully whether the potential nonresponse bias (discussed above) and measurement bias will affect their application of the data. In addition, the complex sample designs and weighting factors must be taken into account when data are analyzed. The use of statistical programs that assume simple random sampling is rarely appropriate; NHANES investigators have developed computer programs that are appropriate for the data gathered.

Trends in the Food Supply and Dietary Intakes

Changes in dietary patterns since the turn of the century have been extensive and include changes in sources of calories; the composition of foods; consumption of. specific food groups, including nonalcoholic and alcoholic beverages; and eating patterns, such as snacking and eating away from home and the selection of diets that differ from those of the average American. A discussion of these changes, below, is based on historical food supply data, results from NFCS and NHANES, and other sources.

Food Energy

Total caloric intake is of interest in epidemiologic studies not only because of its association with body weight, but also because of implications that it may be involved in the relationship between nutrients and chronic diseases. Whether or not an individual gains, loses, or maintains body weight depends on the balance between caloric intake and physical activity, body size, body composition, and probably metabolic efficiency (Sims et al., 1973) (see Chapter 6). These factors largely explain the differences in energy intake among individuals.

Because individuals vary in their energy needs, there is no satisfactory dietary standard for assessing adequacy of energy intake. Fully recognizing this fact, investigators in the 1977-1978 NFCS chose the midpoints of the ranges of the Recommended Energy Intakes (REls) given by the Food and Nutrition Board in 1980 (NRC, 1980). Food energy intakes reported in that survey averaged 84% of the REI midpoints. Approximately 25% of the respondents had intakes of the REI midpoints or greater, and 52% had at least 80% of the REI midpoints.

A comparison of data from national surveys indicates that reported caloric intakes have decreased over time, whereas body weights have increased. A comparison of 1971-1974 NHANES I data with data from the 1960-1962 Health and Examination Survey shows average weight increases of 3 lb for females and 6 lb for males. Heights also increased over the same period (Abraham, 1979). The 1977-1978 NFCS data obtained by 24-hour recall showed a decline in caloric intake for men and women 19 to 50 years of age when compared with 1965 NFCS data (USDA, 1984). The reported energy intakes for males 9 to 64 years old were 10 to 17% lower in 1977-1978 than in 1965; for females 23 to 50 years old, energy intakes were 9% lower. The average intake reported for females (1,500 to 1,600 kcal/ day) is of concern because of the difficulty in incorporating all nutrients at recommended levels in a diet so low in calories (Mertz and Kelsay, 1984).

Page 53

As pointed out above, some differences in survey data over time may reflect differences in methods of obtaining data, yet the reported decline in caloric intake accompanied by weight gain requires some explanation. Decline in physical activity may have occurred but was not monitored in these surveys. In addition, studies indicate that there may have been some underreporting of food intake. For example, the USDA's Beltsville Human Nutrition Research Center reported that subjects in nutrition studies conducted over the past 10 years generally required more calories to maintain body weight on a controlled dietary regimen (as encountered in the studies) than they reported in dietary records of their usual self-selected diets collected for 7 days before they entered the studies (Mertz and Kelsay, 1984). Hallfrisch et al. (1982) reported that compared with caloric intake reported on 7-day records of self-selected diets, male subjects required an average of 500 additional calories and females 900 additional calories to maintain constant weight on an 18-week controlled diet. The observed differences may have been due to differences in physical activity, although subjects were cautioned to maintain usual activity. Subjects in this study were fed in a ''modified gorging" pattern, receiving 25% of calories at breakfast and 75% at the evening meal—a definite change in eating pattern; the influence of this pattern on caloric need, if any, is unknown. Another possible explanation of the observed difference in caloric intake is that portion sizes were underestimated on the records or that the requirement to record intake resulted in modification of usual eating patterns (see Chapter 2).

In the 1977-1978 NFCS, the average daily food energy intake for all survey participants was estimated to be 1,826 kcal based on a 3-day intake. The group with the highest average daily intake, 2,568 kcal, was 15- to 18-year-old males. Men 75 years and older averaged 1,866 kcal. The highest average intake for females was 1,849 kcal for 9- to 11-year-olds. This declined with age to a low of 1,417 kcal for women 75 years and older based on a 3-day intake. The intake of women 19 to 50 years of age averaged 1,588 kcal/day in the 1986 CSFII based on a 1-day intake, compared with 1,528 kcal/day based on a 4-day intake in the 1985 CSFII and 1,573 kcal/day in the 19771978 NFCS.

Sources of Food Energy

As noted above, historical food supply data represent amounts of foods that disappear into the food distribution system. These data have provided a way to assess trends in the availability of foods

Page 54

FIGURE 3-3
 Percentage of calories from protein, carbohydrates, and
fat from 1909-1913 to 1985. These data fail to present the
entire picture, since food supply data do not include alcohol
or grains used in production of alcoholic beverages.
From R.M. Marston, USDA/HNIS, unpublished data, 1986.

and nutrients in the food supply since early in this century.

The distribution of food energy between carbohydrates and fats in the food supply has changed since the first decades of this century. The availability of fats has steadily increased (Table 3-3), whereas carbohydrate levels fell from 1909 to 1969 but have increased by approximately 9%  since then. Of the percentages of calories provided by the macronutrients, protein has contributed approximately 11% from  1909 to 1985, while the percentage from fats increased from 32 to 43% and the percentage available from carbohydrates fell from 57 to 46% (Figure 3-3).

Food groups consisting of grain products and of meat, poultry, or fish supply the largest share of calories in the U.S. diet. Fats, sweets, and beverages combined contribute about as many calories as fruits plus vegetables or milk plus milk products, and substantially more than the combination of eggs, legumes, nuts, and seeds (Figure 3-4; see figure caption for definitions of groups).

Fats, Fatty Acids, and Cholesterol

Dietary fat is of concern in relation to coronary heart disease (CHD), cancer, and possibly obesity. The associations of saturated, monounsaturated, and polyunsaturated fatty acid intakes with CHD have received particular attention.

The per-capita availability of fatty acids in the food supply increased from  1909 to 1985. This increase was markedly greater for linoleic acid (which increased 19 g/day) and oleic acid (a 20-g/ day increase) than for saturated fatty acids (a 7-g/ day increase) (Figure 3-5). In 1985, linoleic acid in the food supply accounted for 7%  of total calories, oleic acid for 17%, and saturated fatty acids for 15% (Figure 3-6). Changes since 1909 resulted in a large increase in the percentage of total fat calories from linoleic acid and a decline in the percentage of fat calories from saturated fatty acids (Figure 3-7). Nevertheless, saturated fatty acids and oleic acid still make up the highest percentage of calories from fats and of total calories from fatty acids in the food.

Fat Intake

Does actual fat consumption reflect these food supply changes? Data from NFCS indicate that mean fat intakes by individuals were lower during 1977-1978 (40.3%  of calories) than in 1965 (42.1% calories). The 1985 CSFII for women and men 19 to 50 years of age indicated that fat intakes continued to decline. As shown in Table 3-4, however, data from  NHANES contradict these results. These data indicate that for women 19 to

Page 55

FIGURE 3-4
 Food sources of energy and energy nutrients in percentages per person per day,
as indicated in the 1977-1978 NFCS. Based on data from USDA (1984).

50 years of age, no difference in fat intake occurred between the 1971-1974 and 1976-1980 surveys (NHANES I and II). On the other hand, data from NFCS show a higher intake of fat in 1977-1978 than NHANES data indicated in 1976-1980 and a decline in intakes in the 1985 and 1986 CSFII. Thus, all surveys except the 1977-1978 NFCS seem to provide similar values. The discrepancies between NFCS and NHANES may be explained by systematic biases inherent in methods used in NHANES, NFCS, and CSFII (Jacobs et al., 1985). Differences in fat intake observed in the 1977-1978 NFCS and 1985 and 1986 CSFIIs may have been due to differences in the type and depth of questions asked in the two surveys. For example, the CSFII included specific questions concerning the type of fat (e.g., meat fat, poultry skin) consumed, but the 1977-1978 NFCS did not. Thus, the national surveys fail to tell us whether or not fat intake has really decreased over time.

At this writing, the most recent estimates of usual intake are those from the 1985 CSFII survey of women 19 to 50 years old and their children between the ages of 1 and 5. In this survey, estimates of usual intake were based on 4 nonconsecutive days of intake obtained by 24-hour recall rather than on only one 24-hour recall (USDA, 1987b). Women consumed a mean of 36.8% of their calories as fat—13.3% from saturated fat, 13.6% from monounsaturated fat, and 7.4% from polyunsaturated fat. Children consumed a mean of 34.7% of their calories as fat—13.7% from saturated fat, 12.7% from monounsaturated fat, and 5.9% from polyunsaturated fat. These are mean values; many respondents consumed higher or lower percentages.

Dietary Food Sources of Fat

Changes in levels of specific fatty acids in the food supply can be attributed to several changes in the availability of foods. The increase in linoleic acid supplies was due chiefly to the remarkable increase in the use of salad and cooking oils from 2 to 25 lb per capita from 1909 to 1985 (Table 3-5). The increased use of margarine and shortening containing vegetable fat also contributed to supplies of linoleic acid as well as to oleic acid.

Page 56

FIGURE 3-5
 Trends in per-capita intake of oleic acid, saturated fatty acids, and
linoleic acid in the U.S. food supply from 1909-1913 to 1985. From
R.M. Marston, USDA/HNIS, unpublished data, 1986.

The contribution of poultry to fat intake greatly increased after 1940. In the meat, poultry, and fish group, however, pork has contributed the most fat since 1909-1913; beef contributed the second highest amount. Fluid whole milk was the major source of fat in the dairy products group until 1980, when cheese became the chief source. Table 3-5 indicates that consumption of fluid whole milk declined by almost 50% since 1967-1969, whereas the consumption of low-fat milk almost doubled.

Data from the 1977-1978 NFCS and the 19761980 NHANES show that the meat, poultry, and fish group is the primary source of dietary fats. Sources of fats reported in the 1977-1978 NFCS are shown in Figure 3-4. Fats contributed by grain products come chiefly from added fats or oils in cakes, pies, pastries, and other baked products. Because the two surveys differed in the ways foods were assigned to food groups and in the data bases used for calculating nutrients, detailed direct comparisons cannot be made.

NHANES II data for the total adult population surveyed indicate that on the 1 day surveyed,

FIGURE 3-6
Trends in percentage of total calories from selected fatty acids in the U.S.
food supply. From R.M. Marston, USDA/HNIS, unpublished data, 1986.

13% of total fat and 16% of saturated fat came from a food group made up of hamburgers, cheeseburgers, meat loaf, hot dogs, ham, and luncheon meats (Block et al., 1985). Beef items furnished more than 15% of total fat. Approximately one-third of the saturated fat reported came from meats, and about one-fourth was provided by milk, milk products, and nondairy creamers. Mayonnaise and salad dressings contributed approximately 15% of the linoleic acid reported; margarine contributed 10%. French fries and other forms of fried potatoes supplied almost 8% of the linoleic acid consumed (Block et al., 1985).

FIGURE 3-7
Trends in contribution of selected fatty acids to percentage of calories from
fat in U.S. food supply. From  R.M. Marston, USDA/HNIS, unpublished data, 1986.

Page 57

Dietary Trans Fatty Acids

Unsaturated fatty acids in natural foods exist chiefly in the as form, in which the hydrogen atoms are on the same side of the double bond. When fats and oils are partially hydrogenated during commercial processing, varying amounts of trans isomers form. In this form, the hydrogen atoms are on opposite sides of the double bond. Although small amounts of trans isomers occur naturally in milk and butter, the large increase in usage of partially hydrogenated vegetable oils has resulted in increases in trans isomers in foods during past decades, raising questions concerning possible adverse effects of these isomers (see Chapter 7).

As yet there are no reliable data on the trans fatty acid intake by the U.S. population, but Hunter and Applewhite (1986) estimate, based on market share and product composition data, that per-capita availability (not actual consumption) is about 7.6 g/ person per day. An Ad Hoc Review Panel of the Federation of American Societies for Experimental Biology estimated 8.3 g/person per day based on USDA food supply data and published analytical values (Senti, 1985). A higher estimate of 12.1 g/ person per day (8% of total fat) was based on 1972 food supply data (Enig et al., 1978).

Page 58

Since estimates of fatty acids available in the food supply are higher than amounts actually consumed, studies of actual consumption are of interest. In one such study, the fatty acid content of the diet as reported by 7-day food intakes of eight adolescent females were quantitated by using gas chromatography to analyze duplicate meals. The average daily consumption of trans fatty acids was estimated to be 3.1 g (6.5% of total fatty acids) (van den Reek et al., 1986). In another study based on gas chromatographic analysis, Aitchison et al. (1977) found that in the self-selected diets of 11 women 25 to 35 years of age, trans isomers accounted for an average of 5% of total fatty acids. Since the numbers of subjects were small in these studies and fails to represent the U.S. population, further research is needed to ascertain the usual intake of these isomers.

Trans-octadecanoic acid (18:1t) is the predominant trans isomer in foods. In an analysis of 220 samples of 35 food types, Enig et al. (1983) found that most samples of mayonnaise, salad dressings, and salad and cooking oils contained no 18:1t, unless the label indicated partial hydrogenation of the oil. Margarines varied in 18:1t content as follows (by weight of fat): stick margarines, 16 to 31%; tub margarines, 7 to 18%; and diet margarines, 11 to 13%. In contrast, cakes, candies, and frostings varied from 3 to 33%; cream substitutes, from 0.4 to 12%; cookies and crackers, from 2 to 34%; breads and rolls, from 0.2 to 24%; and puddings, from 28 to 35%. Van den Reek et al. (1986) observed that approximately two-thirds of the 18:1t consumed by their eight adolescent subjects could be calculated using the analyses of Enig et al. (1983).

The association of dietary cholesterol with CHD is also of interest. Cholesterol in the food supply increased from 500 mg during 1909-1913 to its highest point, 570 mg, during 1947-1949, but declined by 16% to 480 mg/person per day during 1977-1979, where it has remained (Table 3-3). The decline in the availability of cholesterol was due chiefly to reduced use of eggs—from a peak of 49 lb/person per year in 1951 (Welsh and Marston, 1982) to 32 lb/person in 1985 (Table 3-5). The 1985 food supply data show that equal amounts of cholesterol (40%) came from eggs and from the meat, poultry, and fish group; 14% came from dairy products (excluding butter); and 5% from fats and oils (including butter) (Marston and Raper, 1987). Beef contributes more cholesterol to the food supply than do other meats (Marston and Raper, 1987).

The average cholesterol intake found in the 1977-1978 NFCS (USDA, 1984) was 385 mg/day, or 214 mg per 1,000 calories. In terms of caloric intake, cholesterol levels were lowest for respondents under age 19; higher for blacks than for whites; higher for those below the poverty level; and highest in the South and West. The relationship of dietary fats and cholesterol to blood cholesterol levels is discussed in Chapter 7.

Carbohydrates and Caloric Sweeteners

Carbohydrates are often categorized as complex carbohydrates (polysaccharides, consisting chiefly of starches), dietary fiber, and mono- and disaccharides (sugars). The major sugars relevant to this discussion are sucrose (table sugar) and fructose.

Food Supply Data

Carbohydrate availability has declined since 1909-1913, both in absolute amounts (Table 3-3) and as a percentage of total calories (Figure 3-3). A striking change occurred in the relative proportions of total carbohydrates available from starches and from sugars. In 1909-1913, the proportion was approximately two-thirds starch and one-third sugar. By 1980, sugars furnished a little more than one-half the carbohydrates in the food supply. The decline in starches was due to the marked decrease in use of grain products and potatoes; at the same time, the use of refined sugars, syrups, and other sweeteners dramatically increased (Table 3-5).

The availability of specific sugars also changed. Sucrose peaked at 102 lb/person per year in 1971-1972 (Glinsmann et al., 1986), but declined to 63 lb in 1985 (Table 3-5). This decline is attributable to the replacement of sucrose in soft drinks and other products by corn sweeteners, which increased in the food supply from 21 lb/person per year in 1972 to 58 lb/person in 1984 (Glinsmann et al., 1986). This increase was due chiefly to the greater availability of high-fructose corn syrup (HFCS)—from 1 lb/person per year in 1972 to 36 lb/person by 1984 (Glinsmann et al., 1986). By 1985, HFCS in the food supply had increased another 20% (Marston and Raper, 1987). Thus, over the century, sugar use has changed—from primarily sucrose to a mixture of sucrose, glucose (largely from corn syrup), and fructose (from HFCS).

Page 59

Because of concern about possible effects of increased fructose use on certain chronic diseases or on carbohydrate metabolism, the FDA established a Sugars Task Force to review and interpret recent scientific studies relative to the health effects of sugars and sweeteners added to foods. In its report (Glinsmann et al., 1986), the task force pointed out that the true increase in fructose availability over the past 10 years is more than the food supply data suggest, because approximately 60% of sucrose added to acidic beverages is converted to glucose and fructose. Such beverages consequently contain more fructose and less sucrose than food supply data indicate.

Carbohydrate Intake

The 1977-1978 NFCS found that carbohydrate intakes averaged 43% of calories. (These estimates do not include calories from alcohol.) For children 1 to 8 years of age, carbohydrates averaged 47% of calories, whereas for males and females 9 to 18 years old, they averaged 45 and 46%, respectively. Individuals below poverty levels had higher carbohydrate intakes than those above (DHHS/USDA, 1986).

Caloric Sweetener Intake

Using intake data from the 1977-1978 NFCS and a specially developed data base on the sugar content of foods, the FDA's Sugars Task Force estimated the intake of added, naturally occurring, and total sugars (Glinsmann et al., 1986). Summaries of the task force's estimates appear in Tables 3-6 and 3-7. Since lactose was not a part of the safety evaluation, figures are given for total sugars minus lactose, as well as for total sugars including lactose. The 14 age-sex groups chosen were those used in the Food and Nutrition Board's report Recommended Dietary Allowances (NRC, 1980).

The average daily intake of total sugars (minus lactose) within age-sex groups ranged from 31 to 116 g/day (mean, 80 g/day). As a percentage of calories, total sugars (minus lactose) averaged 18% of calories, about half of which was contributed by sugars added to foods. The 90th percentile level of daily intake of total sugars (minus lactose) was 139 g/day (range, 65 to 193 g), or a mean of 27% of calories.

The mean and the 90th percentile values for added sugars were 53 g/day and 104 g/day, respectively, or 11 and 20% of caloric intake, respectively. Table 3-7 indicates that the mean percentage of total calories from added sugars for all age-sex groups was 2% for fructose, 6% for sucrose, and 4% for HFCS. The average daily intake of added sugars was highest (13 to 14% of caloric intake) for children 4 to 10 years old, for males 11 to 18 years, and females 11 to 22 years (Glinsmann et al., 1986).

Dietary Sources of Carbohydrates

The 1977-1978 NFCS data indicate that grain products provide more carbohydrates than other food groups. Grain products include cereals and

Page 60

pasta as well as baked products, which contribute both starch and sugar. The sweets group contributed only half the amount of carbohydrates supplied by either fruits or vegetables, but sugars are added to foods in the grain products, fruit, and beverage groups (Figure 3-4). Beverages and fruits contribute similar proportions to carbohydrate intake.

Beverage Consumption

One of the most striking changes in food consumption patterns in the past two decades is the increased consumption of soft drinks, citrus juices, beer, and wine, accompanied by decreased consumption of coffee and milk. Data in Table 3-8 indicate that over the past 20 years, the availability of citrus juices and soft drinks increased by 138% each, wine by 123%, and beer by 23%. Coffee availability declined by 29% and milk by 18%. Data from the 1977-1978 NFCS indicate that on the 3 days surveyed, coffee was consumed by 51% of the respondents, tea by 39%, fruit ades and drinks by 14%, and decaffeinated coffee by 7% (Pao et al., 1982).

Alcohol Consumption

The data in Table 3-8 on alcohol availability overestimate actual consumption, because they are based on industry production data and fail to account for wasted beverages due to bottle breakage, spillage, alcohol evaporation when beverages are used in cooking, and other losses. On the other hand, data from NFCS are underestimates due to underreporting by some individuals and the inability to assess usual intake over long periods, which results in underreporting by heavy drinkers. The

Page 61

1977-1978 NFCS indicated that 8% of individuals reported consuming beer or ale at least once in the 3 days surveyed; 5% reported wine intake; and 3%, distilled liquor (Pao et al., 1982). On the other hand, NHANES II data indicated that on the day surveyed, a higher proportion of the population  consumed alcohol and in larger amounts than were reported by the 1977-1978 NFCS. One possible reason for the discrepancy may be that respondents felt more comfortable disclosing their alcohol consumption during NHANES interviews because family members were not present (Woteki, 1985). Furthermore, intake on 1 day does not reflect the usual intake of individuals.

In the Alcohol and Health Practices Survey—a component of the National Health Interview Survey by DHHS—investigators gathered data on consumption of alcoholic beverages during a 2-week period by adults 20 years of age and older (Schoenborn and Cohen, 1986). Fifty percent of males and 23%  of females reported they had consumed five or more drinks in 1 day at least once during the year. Thirty percent of the respondents were classified as "lighter" drinkers (0.01 to 0.21 oz of ethanol/day), 21% were "moderate" drinkers (0.22 to 0.99 oz of ethanol/day), while 10% were ''heavier" drinkers (1 or more oz of ethanol/day). "Moderate" drinkers, by this definition, consumed between 5.5 and 24 oz of beer, between 1.5 and 6.5 oz of wine, or 0.5 to 2 oz of distilled liquor/day. Men were four times more likely to be "heavier" drinkers than women. About one-third of respondents were abstainers.

Whites were less likely to be abstainers than blacks and other racial minorities. Younger people

Page 62

were more likely to be drinkers, and they drank more heavily than older people. The more highly educated and those with higher incomes were more apt to drink and apt to drink more than those with less education and income except that 43% of those in the lowest income group—<$7,000 year—had consumed five drinks or more on at least 1 day in the past year, compared with 36% of those whose income was $40,000 or more (Schoenborn and Cohen, 1986).

Heavy alcohol consumption can drastically alter the proportion of calories obtained from carbohydrates, fats, and protein. An individual who drinks a fifth of a gallon of whiskey daily derives 2,120 kcal from ethanol or 58% of a 3,600 kcal diet (Scheig, 1970). A study of the alcohol intake of respondents in the 1977-1978 NFCS indicated that on the 3 days surveyed, alcohol supplied an average of 19% of total calories consumed by drinkers (Windham et al., 1983). There is a need for better documentation of actual alcohol consumption through national surveys. The present estimates of the percentage of calories from macronutrients in the diet are in error because of inadequate estimates of alcohol intake.

Alcohol was among the food components given high-priority monitoring status by the Joint Nutrition Monitoring Evaluation Committee (DHHS/ USDA, 1986). Chapter 16 discusses the health aspects of alcohol consumption.

Drinking Water

Respondents in the 1977-1978 NFCS (USDA, 1984) reported drinking an average of 3.3 cups (8 fl oz/cup) of water per day on the 3 days surveyed. The median intake was 2.8 cups.

Dietary Fiber

Dietary fiber is composed of complex plant substances that resist digestion by secretions of the human intestinal tract. The chief difficulty in relating dietary fiber to the occurrence of chronic diseases in populations is the paucity of data on the amount and kinds of dietary fiber in foods. Until the recent past, data were available in food composition tables only for "crude fiber"—the residue resulting after foods are treated with acids and alkalies. Since this method of analysis destroys many components of dietary fiber, crude fiber is an inadequate indicator of dietary fiber in foods.

There are few data on dietary fiber in the food supply or in diets consumed by respondents in 1977-1978 NFCS or 1976-1980 NHANES. The first surveys to include an estimate of dietary fiber were the 1985 and 1986 CSFIIs, but USDA data are as yet limited regarding dietary fiber in foods. In these surveys, dietary fiber included both the insoluble fraction (neutral detergent fiber) and soluble fraction (such as gums in cereal grains and pectin in fruits and vegetables). The values for these fractions were based chiefly on the method of Englyst et al. (1982) and to a lesser extent on that of Prosky et al. (1985). In 1985, average intake of dietary fiber per day for women 19 to 50 years of age was 10.9 g, for children 1 to 5 years old, 9.8 g (both based on 4 days of intake), and for men 19 to 50 years old, 18 g (based on a 1-day intake). The 1986 CSFII (USDA, 1987c) indicated that women in the West and Midwest had higher intakes of dietary fiber than those in the South or Northeast.

Information on dietary fiber in foods is building rapidly as the USDA data base expands. The new USDA Agriculture Handbook No. 8 when completed will provide more up-to-date data on dietary fiber. Foods highest in dietary fiber include whole (unrefined) grains and breads made from them, legumes, vegetables, fruits, nuts, and seeds.

Protein

As mentioned above, Figure 3-3 indicates that the food supply has provided about 11% of calories as protein since 1909-1913, or about 100 g of protein per person per day (Table 3-3). A major change over the years in the food supply is the increased use of animal over plant protein sources. During 1909-1913, about 52% of protein came from animal sources compared with about 68% in 1982. This change was a result of increased use of meats, poultry, fish, dairy products, and eggs, accompanied by decreased use of flour, cereal products, and potatoes. Animal products provide almost three-fourths of the eight essential amino acids in the food supply and contributed 70% of the total protein in the 1977-1978 NFCS (USDA, 1984). Protein available in the food supply is much higher than the RDA for protein, which is 56 g/ day for males 15 years and older weighing 70 kg and 44 g/day for females in the same age group weighing 55 kg (NRC, 1980).

Protein Intakes

According to the 1977-1978 NFCS, protein intakes averaged 74.3 g/day for all respondents and exceeded the RDA for all 22 age-sex groups. Race, poverty status, region, urbanization, and season had little influence on dietary protein levels. Protein contributed an average of 17% of total calories

Page 63

FIGURE 3-8
Food sources of vitamins in percentage per individual
per day. Data from the 1977-1978 NFCS (USDA, 1984).

in both males and females in the 1977-1978 NFCS and in the 1985 and 1986 CSFII. Annual per-capita consumption of beef and veal changed from 32 kg in 1960 to 36 kg in 1982; consumption of pork remained stable; and poultry consumption increased from 16 kg in 1960 to 29 kg in 1982 (USDA, 1981, 1983). Other surveys, such as the CSFII (USDA, 1986b), have found slightly different results based on a 1-day recall. For example, 42% of males 19 to 50 ate beef on a 1-day recall in 1977 whereas 28.3% of this same population ate beef during this period in 1985. The intake of poultry remained the same while fish intake increased from 8.5% in 1977 to 11.4% in 1985. In 1985, as compared to 1977, meat intakes decreased as income increased. Low-income people eat more meat than high-income people.

In the 1977-1978 NFCS, meats, poultry, and fish contributed by far the largest amount of dietary protein (49%); dairy products and grain products each supplied 18%. The largest part of the food dollar was used to purchase meat, poultry, and fish; however, lower income households obtained more protein per dollar than higher income households.

Vitamins

Vitamins most often reported in food supply data and in food consumption surveys include vitamins A and C, thiamin, riboflavin, preformed niacin, vitamin B₆, and vitamin B₁₂ (Table 3-1). Data regarding amounts of vitamin E and folic acid in foods are at present incomplete and of doubtful validity.

Table 3-9 summarizes the 1980 RDAs for specific vitamins, major food sources, availability in the food supply, intakes by survey populations, and conclusions by the JNMEC regarding the adequacy of intakes and current knowledge of the vitamin status of the population. In Tables 3-9 and 3-10, average intakes are expressed as percentages of the RDAs; the intent is to indicate the relative amount. Intakes below the RDA should not to be construed as inadequate. Figure 3-8 presents food sources of vitamins reported in the 1977-1978 NFCS.

As summarized in Table 3-9, the JNMEC accorded priority status to only one vitamin—vitamin C—due to low serum levels among some segments of the population in NHANES II. The group also concluded that vitamin A, thiamin, riboflavin, and niacin warrant continued monitoring and that further investigation of the relationship between dietary intake and nutritional status is needed relative to vitamin B₆ and folacin.

Minerals

Minerals included in the 1977-1978 NFCS were calcium, phosphorus, magnesium and iron, whereas NHANES II included all these except magnesium. CSFII included those in the 1977-1978 NFCS as well as sodium, potassium, copper, and zinc. The Total Diet Study is the only national study at present that includes all those plus iodine, manganese, selenium, and chromium.

Table 3-10 summarizes information about minerals, including the 1980 RDAs, major food sources, availability in the food supply, intakes by survey populations, and nutritional status. Figure 3-9 shows food sources of minerals reported in the 1977-1978 NFCS.

Among national surveys, only the Total Diet Study included manganese, selenium, and iodine

Page 64

Page 65

Page 66

Page 67

Page 68

Page 69

Page 70

Page 71

Page 72

Page 73

Page 74

FIGURE 3-9
Food sources of minerals shown as percentage per person
per day. Data from the 1977-1978 NFCS (USDA, 1984).

(Pennington et al., 1986). The iodine content of foods purchased in grocery stores between 1982 and 1984 was sufficient to provide dietary intakes markedly higher than the iodine RDA for all age-sex groups. For example, infants 6 to 11 months of age consumed 400%, 2-year-olds, 657%, and boys 14 to 16 years old, 473% of the RDA for iodine.

Although no evidence of adverse effects of present iodine intakes have been observed, the Food and Nutrition Board's report Recommended Dietary Allowances states that ''any additional increases should be viewed with concern" (NRC, 1980). The report also contains a recommendation that iodophors used in the dairy industry, dough conditioners containing iodine, alginates, and certain coloring dyes be replaced, if possible, with substances not containing iodine (NRC, 1980).

Iodine is present in seafoods and in foods grown on high iodine soils. During the past 15 years or so, dairy products have become the chief iodine source in U.S. diets due to the use of iodine-containing substances to clean and disinfect equipment and to the addition of iodine to the feed of dairy cows. Some dough conditioners used in breadmaking and Red Dye No. 3 used in candies, breakfast cereals, and vitamin pills contain high levels of iodine. However, the iodine in Red Dye No. 3 is believed not to be absorbed. Many fast foods are also high in iodine (Taylor, 1981). Iodized table salt also contributes to iodine intake.

Selenium intakes were judged to be within the Estimated Safe and Adequate Daily Dietary Intake (ESADDI) in the Total Diet Study (Pennington et al., 1986). Manganese intakes were judged to be high in 6- to 11-month-old infants and low in teenage girls and women up to age 65.

Fluoride is important in the prevention of dental caries, but fluoride intakes have not been determined in any of the large surveys. Dietary fluoride is provided by seafood and tea leaves, but occurs in relatively small amounts in other foods. The fluoride content of drinking water, either naturally or artificially fluoridated, is probably the best indicator of dietary exposure. Not all Americans live in areas where the water supply is fluoridated, however. JNMEC concluded that fluoride should receive public health monitoring status, because not everyone in the U.S. population may consume amounts adequate to prevent dental caries (DHHS/USDA, 1986).

In summary, JNMEC listed sodium, calcium, iron, and fluoride among the nutrients that merit priority in considerations of diet and public health. It also noted that magnesium and zinc warrant further investigation (DHHS/USDA, 1986).

Fortification and Enrichment of Foods

The term fortification is usually used to designate the addition of nutrients not naturally present in a food (e.g., the fortification of milk with vitamin D), whereas the term enrichment generally means the addition of nutrients already present in a food. Often the terms are used interchangeably, however (Quick and Murphy, 1982).

The FDA sets standards specifying the nutrients and their amounts to be added to enrich flour, cereals, and bread products; standards are also set for the addition of vitamins A  and D to milk products and of vitamin A to margarine. State agencies, not the FDA, are responsible for requiring that specific foods be enriched or fortified. Enrichment of bread and white flour is mandatory in about two-thirds of the states, but in fact, nearly all white flour in the United States is enriched with certain B vitamins and iron.

In 1980, the FDA published guidelines to promote the rational fortification of food. Sugars, candies, carbonated beverages, and other snack foods were considered to be inappropriate for fortification. Fortification was considered to be

Page 75

appropriate (1) to correct a dietary inadequacy recognized by the scientific community to result in a deficiency disease; (2) to restore nutrient levels to those present in a food before conventional processing and storage; (3) to adjust the protein, vitamin, and mineral content of the food to meet specific caloric requirements (i.e., a food must furnish a specific number of calories per serving); and (4) to ensure that a substitute food is nutritionally similar to the traditional food it replaces (Quick and Murphy, 1982).

USDA prohibits the direct fortification of meat and poultry, taking the view that these are highly nutritious foods that do not need additional nutrients. Enriched flour or cereals may be used in meat mixtures, however.

A direct relationship between food fortification and improved nutritional status is difficult to establish because of problems in controlling all the factors that may contribute to such improvements. Some attempts have been made to assess the influence of enrichment and fortification on nutrient intake. A  survey covering 1966 to 1970 (Friend, 1972) indicated that nutrients added to the food supply increased the availability of thiamin by 40%, iron by 25%, niacin by 20%, riboflavin by 15%, and 10% each for vitamin A and ascorbic acid. Among foods supplemented at that time were flour and baked products, cereal products, beverages, milk, infant formulas, margarine, and formulated meal replacements.

In a recent study, Cook and Welsh (1987) used food consumption data from the 1977-1978 NFCS to study the effect of enriched and fortified grain products on nutrient intake. Enrichment and fortification of grain products were found to provide 32, 18, 20, and 19%  of the total intakes of thiamin, riboflavin, niacin, and iron, respectively (Cook and Welsh, 1987). Grain enrichment led to thiamin intakes averaging 110%  of the RDA, rather than the former level of 74%. Fortified breakfast cereals provided approximately 20% of the added thiamin, niacin, and iron, and about 25% of the added riboflavin for all respondents.

Fortification and enrichment clearly have made important contributions to the nutrient intake of the U.S. population. Nevertheless, many nutrition scientists have warned against indiscriminate fortification of foods (Mertz, 1984).

Changes in Eating Patterns

Marked changes in the availability and nutrient content of foods in the food supply have occurred in parallel with lifestyle changes affecting the kinds and amounts of foods consumed. The 1977-1978 NFCS (USDA, 1984) assessed eating patterns only on 3 consecutive days, and the 1986 NFCS (USDA, 1987c) was based on 1 day of reported intake. The resulting data are inadequate to assess the habitual eating patterns of individuals, but can be used to assess eating patterns of groups.

Eating Occasions

The 1977-1978 NFCS (USDA, 1984) indicated that fewer males and females 19 to 34 years old ate breakfast at least once in 3 days than did people in other age groups. Children under 12 and adults 65 years old and over were most likely to eat breakfast. In 1985, 53% of women 19 to 50 years of age and 85% of children 1 to 5 years old ate breakfast on all 4 days of the survey (USDA, 1987a).

Respondents ate alone on 29% of eating occasions, primarily breakfasts and snacks (USDA, 1984). Nine percent of respondents packed lunches to eat away from home. Twenty percent of all eating occasions away from home (food obtained and eaten away from home) were at restaurants, 19%  at work, 16%  at school, 16% at someone else's home, and 13% at fast-food places (USDA, 1984).

Eating Away from Home

In the United States, people have been eating away from home at an increasing rate for many years. Data based on a 1-day intake obtained in 1986 (USDA, 1987c) indicated that 57% of women ages 19 to 50 obtained and ate some food away from home, compared with 45% in 1977 to 1978. On the basis of a 4-day intake in 1985, 88% of women 19 to 50 years of age reported eating some food away from home on at least 1 of the 4 days (USDA, 1987b). Forty-five percent of children 1 to 5 years old consumed some food away from home in 1986, compared with 30% in 1977 (USDA, 1987c). A larger proportion of men also reported obtaining and eating food away from home in 1985 than in 1977 (69 vs 53%).

Compared to the foods prepared and eaten at home, the nutrient densities (amount of a given nutrient per 1,000 kcal) of foods eaten away from home during the 3 days surveyed were lower for nearly all nutrients (USDA, 1984). Ries et al. (1987) reported the nutritive value of foods consumed at restaurants, fast-food places, and snack bars in grocery or drug stores by respondents 15

Page 76

years of age and older who were interviewed in the 1977-1978 NFCS. Nutrient densities were lower for practically all nutrients in these foods than in food eaten at home except for fat, which was higher in the foods eaten "out." This survey did not determine the habitual behavior of individual respondents. The authors concluded that the respondents did not eat outside the home frequently enough to significantly diminish the adequacy of their diets, but cautioned that individuals who do eat frequently in such places (particularly teenagers and the elderly) are at risk of low intakes of calcium, vitamin A, vitamin B₆, and vitamin C (Ries et al., 1987).

Snacking

Eating defined by respondents as "a snack, a coffee break, or a beverage break" was practiced by 77% of all respondents at least once during the 3 days reported in the 1977-1978 NFCS (USDA, 1984). Such snacks provided an average of 18% of the day's calories and a higher proportion of carbohydrates (23%) than fats (15%) or protein (11%). Preschoolers and  teenagers obtained about 20%  of their total caloric intake from snacks. The increased frequency of eating away from home and increased snacking practices appear to have affected the nutritive quality of diets in a negative way.

Dieting

Respondents in the 1977-1978 NFCS were asked whether or not their food intake on the day of the interview differed from their usual intake. If so, they were asked whether or not they were on a diet to lose weight. Among females, 11% of those 15 to 18 years old and 16 to 19% of those 19 to 64 years old answered "yes" to this question. Twelve percent of women 65 to 74 years of age reported they were on weight-reduction diets.

Alternative Diets

The surveys discussed in this chapter indicate that the U.S. population consumes relatively large amounts of meat and sugar, more refined than whole-grain products, and larger amounts of commercially processed than fresh foods. In contrast, most of the world's population today subsists on vegetarian or near-vegetarian diets for reasons that are economic, philosophical, religious, cultural, or ecological. Indeed, humans appear to have subsisted for most of their history on near-vegetarian diets (ADA, 1980).

During the 1960s and 1970s, interest in diets other than the usual intakes heightened in the United States. Some young people for the first time became vegetarians. Others following alternative diets include users of whole (minimally processed) foods or organically grown foods, as well as those who simply try to avoid food additives. Many obtain a large percentage of their food from health food stores or from small food cooperatives, rather than from supermarkets. Wolff (1973) noted that such a group in Hawaii avoided such foods as refined sugar, bread and other products made with white flour, white rice, processed foods containing additives, soft drinks, processed meats and cereals, and coffee. Instead, they bought (in descending order of frequency) whole-grain products, whole-wheat bread, fresh vegetables and fruits, raw nuts, wheat germ, brown rice, honey, yogurt, dried fruit, brewer's yeast, and seeds (Wolff, 1973). Their chief reason for adhering to their alternative diet was that they believed it to be healthier. Today, increased demand for these kinds of foods has prompted many large supermarkets to offer such foods as brown rice, whole-grain flour, unsalted nuts, seeds, and dried fruit, often in bulk, as well as additive-free whole-grain breads.

People who call themselves vegetarians vary from  those who avoid only red meat, but eat poultry or fish, to lacto-ovovegetarians (who eat milk and eggs but no flesh foods), to total vegetarians (who eat no foods of animal origin). Vegetarian organizations generally do not classify as vegetarian those who avoid only red meat. Technically, vegans are those who not only avoid all animal foods, but use no other kinds of animal products such as wool, silk, or leather.

The first national survey to recognize the increasing interest in vegetarianism was the 19771978 NFCS in which respondents were asked: "Are you a vegetarian?" Of the 37,135 people surveyed, 464 (1.2%) answered "yes," but since some of these people reported consumption of flesh foods during the 3 days on which dietary information was obtained, it is not clear how vegetarianism was defined by the respondents. Data from this survey indicated that vegetarians obtained 15.5% of their calories from protein, 37.6% from fats, and 47.3% from carbohydrates, whereas nonvegetarians obtained 16.6, 40.6, and 42.6%  of calories from protein, fats, and carbohydrates, respectively (USDA, 1984). The nutritive values of the food intakes of vegetarians and nonvegetarians are

Page 77

shown in Tables 3-11 and 3-12 as percentages of the 1980 RDAs for various nutrients.

Vegetarians had higher caloric intakes than nonvegetarians, except for women 35 years of age and older, who consumed 72%  of the energy requirements for their subgroup, which is 2,000 kcal for women ages 35 to 50; 1,840 kcal for those ages 51 to 64; 1,800 kcal for those ages 65 to 74; and 1,600 kcal for those 75 and over (NRC, 1980). Although mean heights of vegetarians were similar to their nonvegetarian counterparts, mean weights tended to be lower in vegetarians.

A comparison of Tables 3-11 and 3-12 indicates that vegetarians had lower intakes of protein, preformed niacin, and vitamin B₁₂ than nonvegetarians, but that their average intakes of all three nutrients were above the RDAs. All other nutrients were, on average, at the same level or higher in vegetarian than in nonvegetarian diets. Average intakes of vitamin B₆ and magnesium were below the RDA for vegetarians and nonvegetarians, but the nutrient data base used to estimate these nutrients is less well established than that for other nutrients. Intakes of calcium, vitamin A, and vitamin C averaged 20 to 24% higher, and magnesium 14% higher, in vegetarian than in nonvegetarian diets.

Since the RDA for iron is highest (18 mg/day) for females in their reproductive years, it is instructive to compare average intakes of vegetarian women 19 to 34 years of age with the comparable group of nonvegetarians women. Although these two groups had comparable energy intakes (78% of the RDA) and comparable iron intakes (60% and 61% of the RDA), vegetarian women had higher mean intakes of calcium, magnesium, phosphorus, vitamin A, riboflavin, vitamin B₁₂, and vitamin C and slightly higher intakes of vitamin B₆ and thiamin. Nonvegetarians had higher mean intakes of protein and preformed niacin, but these nutrients still met or exceeded the RDA in vegetarian diets.

It is possible that the iron in vegetarian diets (all inorganic, if no flesh foods are consumed) is less well absorbed than iron in nonvegetarian diets, which include heme iron in meat. However, the absorption of inorganic iron is enhanced by the simultaneous consumption of vitamin C. Iron deficiency anemia appears to be no more prevalent among vegetarian women than among nonvegetarian women, but further study of iron bioavailability in vegetarian diets is needed.

Groups with the lowest caloric intakes were female vegetarians 35 years of age and older whose energy intake averaged 72% of the RDA. The diets of vegetarian women 65 years old and above were nutritionally inferior to those of nonvegetarian women of similar age (Tables 3-11 and 3-12). In this age group, vegetarians had lower intakes of protein, iron, magnesium, phosphorus, thiamin, preformed niacin, vitamin B₆, vitamin B₁₂, and vitamin C than did nonvegetarians. However, each of these nutrients met or exceeded 88% of the RDA in vegetarian diets except magnesium (68%) and vitamin B₆ (52%) for which the intake was also below the RDA for nonvegetarians (75 and 62%, respectively.)

Do Food Choices in the United States Reflect a Concern about Health?

Scholars of dietary behavior have long recognized that although biologic functioning is related to one's lifetime food intake, complex economic, social, political, and cultural factors govern food choices. Specialists who seek the cooperation of designated respondents in nutrition surveys or who endeavor to change food habits to provide good health must appreciate fully the fact that food choices have strong symbolic, emotional, and cultural meanings (Sanjur, 1982).

One of many factors affecting food choices is the individual's belief about the health or nutritional benefits or harm associated with specific choices. Investigators working in the area of diet and chronic diseases consequently have focused their attention on ways to modify belief systems to effect desired changes in food choices. In recent years, national campaigns have been launched to inform the public about the association between dietary salt/sodium and hypertension; between dietary fat/ saturated fat and cholesterol and heart disease; and between dietary fats and fiber-containing foods and certain cancers. The National High Blood Pressure Education Program began at the National Heart, Lung, and Blood Institute (NHLBI) in 1972, and major efforts to inform the public about the link between sodium/salt and hypertension began in 1981. The NHLBI's National Cholesterol Education Program was launched in 1984 to educate the public about the relationship of dietary fat/saturated fat and cholesterol to high blood cholesterol—a major risk factor for CHD. The National Cancer Institute began a program in fall 1984 to encourage the public to reduce fat intake and increase fiber intake in an effort to reduce the risks for breast and colorectal cancer.

Page 78

Page 79

Page 80

Public Opinion Surveys about Diet and Health

FDA and the National Institutes of Health (NIH) periodically have conducted public opinion surveys to evaluate the effectiveness of these public information programs. These telephone surveys reach a sample of approximately 4,000 consumers chosen to be representative of the U.S. population. These results indicate that by 1986, two-thirds of adults in the United States had heard about a relationship between salt/sodium and hypertension; 44% reported that they studied package ingredient lists to avoid or limit salt/sodium; and 61% reported purchasing low-sodium products at least once. Many food manufacturers responded to consumer demands not by lowering salt/sodium in their basic line of products, but by introducing new sodium-reduced products (Heimbach, 1985).

Public perceptions that risks for heart disease may be affected by diet also grew: from 58% of the respondents in 1982 to 76% in 1986. Furthermore, beliefs that dietary fats and cholesterol are related to heart disease risks became widespread among the respondents: from 29% for fats and 26% for cholesterol in 1982 to 43 and 40%, respectively, in 1986. In response to the question, "What things that people eat and drink might make them more likely to get cancer?" 19% named fats or fat-containing foods in 1986, compared with only 12% two years earlier. To the question, "What things that people eat or drink are likely to prevent cancer?" 22% in 1986 compared with 10% in 1984 answered vegetables/fruits; 32% in 1986 compared with 9% in 1984 answered fiber/bran/roughage/ whole grains. The latter response was attributed to the $75 million advertising campaign for high fiber breakfast cereals launched in fall 1984 (Heimbach, 1985).

Although self-reports of behavior are apt to be misstated, sales data show declining use of red meats, butter, whole milk, and eggs and increased sales of fresh produce and high-fiber cereals (but not whole-wheat bread). In addition, 61%  of respondents in the FDA and NIH public opinion surveys (Health and Diet Surveys) reported major dietary changes between 1984 and 1986 in an effort to prevent heart disease or cancer. Reported changes included reducing fat intake, primarily by eating less meat, and reducing intakes of salt, cholesterol, and sugar. At the same time, the respondents reported consuming more fresh vegetables, fruits, fish, poultry, grain products, and bran. The survey indicated that those most knowledgeable and most apt to avoid or limit consumption of substances believed to be linked to health problems were better educated, had higher incomes, and were between 30 and 45 years of age (Heimbach, 1985).

A recent study of NHANES II data (19761980) indicates that on the 1 day surveyed, a relatively small percentage of respondents consumed foods recommended as possibly protective against cancer by a committee of the National Research Council (NRC, 1982) and the American Cancer Society (ACS, 1984). Only 18% of the respondents reported they consumed cruciferous vegetables (e.g., cabbage, Brussels sprouts, cauliflower), 21% reported eating fruits and vegetables high in vitamin A, and 16% reported the consumption of breads and cereals high in dietary fiber (Patterson and Block, 1988). A larger percentage reported they consumed red meats (55%); 43% reported eating bacon and luncheon meats. Diets of females were closer to the recommended guidelines than those of males. Older people were closer to the guidelines than younger people, and blacks were closer than whites, because blacks consumed more vegetables, fruits, fish, and poultry than did whites. These data do not represent the usual dietary intake of individual respondents, since only 1 day of dietary intake was obtained. Surveys being conducted by USDA  and DHHS will provide additional data on food choices relative to dietary recommendations.

Summary

In the United States, food patterns have changed significantly since 1909, when USDA began to collect data on the food supply. These data represent foods, excluding alcohol, that disappear into civilian markets—not actual consumption—but they reflect changes in overall patterns of food use by the population over time. Between 1909 and 1985 the percentage of calories available in the food supply from fats increased from 32 to 43%, the percentage from carbohydrates declined from 57 to 46%, and the percentage from protein remained unchanged at 11%.

Saturated and monounsaturated fatty acids provide the highest percentage of calories from fat in the food supply, although the availability of polyunsaturated fatty acids greatly increased over the years due to wider use of oils and margarines. Compared to 1909-1913, the food supply in 1985 furnished larger amounts of beef, poultry, fish, dairy products, fats, oils, fruits, sugars, and sweet-

Page 81

eners, but furnished lower amounts of eggs, vegetables, potatoes, and grain products.

Public health agencies and several expert groups concerned with diet and health in the United States have urged the public to decrease total fat intake to approximately 30%  of total calories. Whether or not actual consumption has decreased over the years is uncertain, however, because data from NFCS and NHANES fail to agree on this point.

NHANES monitors the overall nutritional status of the U.S. population, whereas NFCS determines the food use of households and the dietary intakes and patterns of individuals. Although these surveys have provided valuable data regarding the dietary intake and nutritional status of the population, some limitations in methods used and differences in design of the two surveys must be taken into account when interpreting their results and drawing conclusions. Furthermore, cross-sectional data traditionally provided by these surveys are unsuitable for studying causal relationships between dietary factors and chronic diseases. Data appropriate for studying causality can be supplied only by longitudinal studies, which are included in plans for NHANES III.

JNMEC was established by USDA and DHHS in 1983 to coordinate the survey methods used by the two departments to report their findings. The first report, issued in 1986, provides food intake data from the 1977-1978 NFCS and information on nutritional status based on biochemical analyses from NHANES II (1976-1980). JNMEC concluded that food components deserving high-priority monitoring status because of high consumption by a considerable portion of the population include food energy, total fats, saturated fat, cholesterol, sodium, and alcohol. Nutrients deserving high-priority monitoring because some portions of the population appear to have low intakes are vitamin C, calcium, iron, and fluoride. The group concluded that protein, vitamin A, thiamin, riboflavin, niacin, total carbohydrates, vitamin B₁₂, and phosphorus should continue to be monitored. In addition, added caloric sweeteners, fiber, vitamin B₆, folacin, magnesium, and zinc were judged to require further investigation since data regarding intake and nutritional status are inadequate at present.

Information on food composition and criteria for assessing nutritional status are incomplete for most nutrients. JNMEC noted that the most complete information exists for food energy, vitamin C, iron, protein, and vitamin A. For all other nutrients listed above, there is a need for more accurate information about the occurrence of the nutrient in foods and methods of nutritional assessment.

Women in their reproductive years constitute a group at risk for nutritional inadequacies because their total caloric intakes tend to be low, while their needs for certain nutrients may be high because of menstrual losses and increased requirements during pregnancy and lactation. Consequently, USDA  surveys (NFCS and CSFII) in 1985 and 1986 focused on women 19 to 50 years of age and their children 1 to 5 years of age.

The USDA used the 1980 RDAs as a standard of comparison in reporting their data, fully recognizing that failure to reach this standard does not indicate inadequate intake, since the RDAs for many nutrients are deliberately set to exceed the requirements of most individuals. According to the 1985 survey (based on 4 nonconsecutive days of intake), women's mean caloric intake was only 1,528 kcal, but their mean intakes were above the RDA  for eight nutrients (protein, vitamin A, ascorbic acid, thiamin, riboflavin, preformed niacin, vitamin B₁₂ and phosphorus). Their mean intakes were below the RDA for vitamin E (87%), calcium  (74%), magnesium (67%), vitamin B₆ (57%), iron (56%), zinc (56%), and folacin (46%). On the basis of only one 24-hour dietary recall, male respondents 19 to 50 years of age reported a mean caloric intake of 2,838 kcal and intakes equal to 98% or more of the RDA for all nutrients listed above except magnesium (94%), zinc (94%), vitamin B₆ (85%), and folacin (76%). Clearly, a major reason that the diets of women are relatively lower in many nutrients than the diets of males is that total caloric intake by women is lower.

Although better-educated, higher-income people appear to be altering their diets in the direction advocated by public health experts, national surveys and other studies indicate that intakes of total fats and saturated fats generally are higher than recommended. White bread is by far the favorite kind of bread, and sweet baked products such as cookies and cakes are very popular. Carbonated soft drinks containing either caloric or noncaloric sweeteners are consumed in large amounts. Alcoholic beverages also contribute calories to the diet, but the extent of actual consumption is uncertain at present because of reporting methods. Fruits, vegetables, and other foods high in dietary fiber are consumed in relatively low amounts. Consumption of cruciferous vegetables such as cabbage, Brussels sprouts, broccoli, and cauliflower is relatively low

Page 82

as are intakes of carotenoid-containing foods such as carrots, sweet potatoes, and winter squash.

Extensive changes in eating patterns have occurred over the century, including marked increases in eating away from home and in snacking. Dieting to lose weight is practiced by many people, especially females. Deviations from recommended dietary guidelines have persisted, and in some instances increased, despite the overall growth in variety of commonly available food, improved transportation and storage of fresh foods, increased disposable income, and greater public and professional knowledge about dietary needs to maintain good health and nutrition.

References

Abraham, S., C.L. Johnson, and M.F. Najjar. 1979. Weight and Height of Adults 18-74 Years of Age: United States, 1971-74. Vital and Health Statistics, Ser. 11, No. 211. DHEW  Publ. No. PHS-79-1659. National Center for Health Statistics, Public Health Service, U.S. Department of Health, Education, and Welfare, Hyattsville, Md. 49 pp.

Acheson, K.J., I.T. Campbell, O.G. Edholm, D.S. Miller, and M.J. Stock. 1980. The measurement of food and energy intake in man—an evaluation of some techniques. Am. J. Clin. Nutr. 33:1147-1154.

ACS (American Cancer Society). 1984. Nutrition and Cancer: Cause and Prevention. American Cancer Society Special Report. American Cancer Society, New York. 10 pp.

ADA (American Dietetic Association). 1980. Position paper on the vegetarian approach to eating. J. Am. Diet. Assoc. 77:61-69.

Aitchison, J.M., W.L. Dunkley, N.L. Canolty, and L.M. Smith. 1977. Influence of diet on trans fatty acids in human milk. Am. J. Clin. Nutr. 30:2006-2015.

Beaton, G.H. 1986. Toward harmonization of dietary, biochemical, and clinical assessments: the meanings of nutritional status and requirements. Nutr. Rev. 44:349-358.

Beaton, G.H., J. Milner, P. Corey, V. McGuire, M. Cousins, E. Stewart, M. de Ramos, D. Hewitt, P.V. Grambsch, N. Kassim, and J.A. Little. 1979. Sources of variance in 24-hour dietary recall data: implications for nutrition study design and interpretation. Am. J. Clin. Nutr. 32:2546-2559.

Beecher, G.R., and J.T. Vanderslice. 1984. Determination of nutrients in foods: factors that must be considered. Pp. 29-55 in K.K. Stewart and J.R. Whitaker, eds. Modern Methods of Food Analysis. Avi Publishing Co., Westport, Conn.

Block, G. 1982. A review of validations of dietary assessment methods. Am. J. Epidemiol. 115:492-505.

Block, G., C.M. Dresser, A.M. Hartman, and M.D. Carroll. 1985. Nutrient sources in the American diet: quantitative data from the NHANES II survey. II. Macronutrients and fats. Am. J. Epidemiol. 122:27-40.

Bunch, K. 1987. Food Consumption, Prices, and Expenditures, 1985. Bulletin No. 749. Economic Research Service, U.S. Department of Agriculture, Washington, D.C. 128 pp.

Carroll, M.D., S. Abraham, and CM. Dresser. 1983. Dietary Intake Source Data: United States, 1976-80. Vital and Health Statistics, Ser. 11, No. 231. DHHS Publ. No. (PHS) 83-1681. National Center for Health Statistics, Public Health Service, U.S. Department of Health and Human Services, Hyattsville, Md. 483 pp.

Cook, D.A., and S.O. Welsh. 1987. The effect of enriched and fortified grain products on nutrient intake. Cereal Foods World 32:191-196.

DHEW (Department of Health, Education, and Welfare). 1972. Ten-State Nutrition Survey 1968-1970, Vols. I-VI. DHEW Publ. No. HSM-72-8130, 72-8131, 72-8132, 72-8133, 72-8134. Center for Disease Control, Health Services and Mental Health Administration, Atlanta. (various pagings)

DHHS/USDA (Department of Health and Human Services/ U.S. Department of Agriculture). 1986. Nutrition Monitoring in the United States-A Progress Report from the Joint Nutrition Monitoring Evaluation Committee. DHHS Publ. No. (PHS) 86-1255. National Center for Health Statistics, Public Health Service, U.S. Department of Health and Human Services, Hyattsville, Md. 356 pp.

DHHS/USDA (Department of Health and Human Services/ U.S. Department of Agriculture). 1987. Operational Plan for the National Nutrition Monitoring System: Report to Congress. Public Health Service, U.S. Department of Health and Human Services, Bethesda, Md. 47 pp.

Eaton, S.B., and M. Konner. 1985. Paleolithic nutrition: a consideration of its nature and current implications. N. Engl. J. Med. 312:283-289.

Englyst, H., H.S. Wiggins, and J.H. Cummings. 1982. Determination of the non-starch polysaccharides in plant foods by gas-liquid chromatography of constituent sugars as alditol acetates. Analyst 107:307-318.

Enig, M.G., R.J. Munn, and M. Kenney. 1978. Dietary fat and cancer trends—a critique. Fed. Proc. 37:2215-2220.

Enig, M.G., LA. Pallansch, J. Sampugna, and M. Keeney. 1983. Fatty acid composition of the fat in selected food items with emphasis on trans components. J. Am. Oil. Chem. Soc. 60:1788-1795.

Food Marketing Institute. 1987. Trends: Consumer Attitudes and the Supermarket. Food Marketing Institute, Washington, D.C. 50 pp.

Forthofer, R.N. 1983. Investigation of nonresponse bias in NHANES II. Am. J. Epidemiol. 117:507-515.

Fregly, M.J. 1985. Attempts to estimate sodium intake in humans. Pp. 33-112 in M.J. Horan, M. Blaustein, J.B. Dunbar, W. Kachadorian, N.M. Kaplan, and A.P. Simopoulos, eds. NIH Workshop on Nutrition and Hypertension. Proceedings from a Symposium, Bethesda, Md., March 12-14, 1984. Biomedical Information Corp., New York.

Friend, B. 1972. Enrichment and Fortification of Foods, 1966-70. National Food Situation (142):29-33.

Garn, S.M., F.A. Larkin, and P.E. Cole. 1978. The real problem with 1-day diet records. Am. J. Clin. Nutr. 31: 1114-1116.

Glinsmann, W.H., H. Irausquin, and Y.K. Park. 1986. Evaluation of health aspects of sugars contained in carbohydrate sweeteners: report of Sugars Task Force, 1986. J. Nutr. 116:S1-S216.

Hallfrisch, J., P. Steele, and L. Cohen. 1982. Comparison of seven-day diet record with measured food intake of twenty-four subjects. Nutr. Res. 2:263-273.

Hampe, E.C., Jr., and M. Wittenberg. 1964. The Lifeline of America; Development of the Food Industry. McGraw-Hill, New York. 390 pp.

Harlan, W.R., A.L. Hull, R.P. Schmouder, F.E. Thompson,

Page 83

F.A. Larkin, and J.R. Landis. 1983. Dietary Intake and Cardiovascular Risk Factors, Part II. Serum Urate, Serum Cholesterol, and Correlates. Vital and Health Statistics, Ser. 11, No. 227. DHHS Publ. No. (PHS) 83-1677. National Center for Health Statistics, Public Health Service, U.S. Department of Health and Human Services, Hyattsville, Md. 94 pp.

Harlan, W.R., A.L. Hull, R.L. Schmouder, Jr., J.R. Landis, F.E. Thompson, and F.A. Larkin. 1984. Blood pressure and nutrition in adults. The National Health and Nutrition Examination Survey. Am. J. Epidemiol. 120:17-28.

Hegsted, D.M. 1972. Problems in the use and interpretation of the recommended dietary allowances. Ecol. Food Nutr. 1: 255-265.

Heimbach, J.T. 1985. Cardiovascular disease and diet: the public view. Public Health Rep. 100:5-12.

Hepburn, F.N. 1987. Food Consumption/Food Composition Interrelationships. Human Nutrition Information Service, U.S. Department of Agriculture, HNIS Report No. Adm-382. Hyattsville, Md. 77 pp.

Hunter, J.E., and T.H. Applewhite. 1986. Isomeric fatty acids in the U.S. diet: levels and health perspectives. Am. J. Clin. Nutr. 44:707-717.

Jacobs, D.R., Jr., P.J. Elmer, D. Gorder, Y. Hall, and D. Moss. 1985. Comparison of nutrient calculation systems. Am. J. Epidemiol. 121:580-592.

LSRO (Life Sciences Research Office). 1984a. Assessment of the Folate Nutritional Status of the U.S. Population Based on Data Collected in the Second National Health and Nutrition Examination Survey, 1976-1980. Federation of American Societies for Experimental Biology. Bethesda, Md. 96 pp.

LSRO (Life Sciences Research Office). 1984b. Assessment of the Iron Nutritional Status of the U.S. Population Based on Data Collected in the Second National Health and Nutrition Examination Survey, 1976-1980. Federation of American Societies for Experimental Biology. Bethesda, Md. 120 pp.

LSRO (Life Sciences Research Office). 1984c. Assessment of the Zinc Nutritional Status of the U.S. Population Based on Data Collected in the Second National Health and Nutrition Examination Survey, 1976-1980. Federation of American Societies for Experimental Biology. Bethesda, Md. 82 pp.

Madans, J.H., J.C. Kleinman, C.S. Cox, H.E. Barbano, J.J. Feldman, B. Cohen, F.F. Finucane, and J. Cornoni-Huntley. 1986. 10 Years after NHANES I: report of initial follow-up, 1982-84. Public Health Rep. 101:465-473.

Marston, R., and N. Raper. 1987. Nutrient content of the U.S. food supply. National Food Review, Winter-Spring, NFR-36:18-23.

McCarron, D.A., C.D. Morris, H.J. Henry, and J.L Stanton. 1984. Blood pressure and nutrient intake in the United States. Science 224:1392-1398.

Mertz, W. 1984. Foods and nutrients. J. Am. Diet. Assoc. 84: 769-770.

Mertz, W., and J.L Kelsay. 1984. Rationale and design of the Beltsville one-year dietary intake study. Am. J. Clin. Nutr. 40 suppl. 6:1323-1326.

Murphy, R.S., and G.A. Michael. 1982. Methodologic considerations of the National Health and Nutrition Examination Survey. Am. J. Clin. Nutr. 35 suppl. 5:1255-1258.

NRC (National Research Council). 1980. Recommended Dietary Allowances, 9th ed. Report of the Committee on Dietary Allowances, Food and Nutrition Board, Division of Biological Sciences, Assembly of Life Sciences. National Academy Press, Washington, D.C. 185 pp.

NRC (National Research Council). 1982. Diet, Nutrition, and Cancer. Report of the Committee on Diet, Nutrition, and Cancer, Assembly of Life Sciences. National Academy Press, Washington, D.C. 478 pp.

NRC (National Research Council). 1983. Diet, Nutrition, and Cancer: Directions for Research. Report of the Committee on Diet, Nutrition, and Cancer, Commission on Life Sciences. National Academy Press, Washington, D.C. 74 pp.

NRC (National Research Council). 1984. Toxicity Testing: Strategies to Determine Needs and Priorities. Report of the Steering Committee on Identification of Toxic and Potentially Toxic Chemicals for Consideration by the National Toxicology Program, Board on Toxicology and Environmental Health Hazards, Commission on Life Sciences. National Academy Press, Washington, D.C. 382 pp.

NRC (National Research Council). 1986. Nutrient Adequacy: Assessment Using Food Consumption Surveys. Report of the Subcommittee on Criteria for Dietary Evaluation, Coordinating Committee on Evaluation of Food Consumption Surveys, Food and Nutrition Board, Commission on Life Sciences. National Academy Press, Washington, D.C. 146 pp.

NRC (National Research Council). 1987. Regulating Pesticides in Food: The Delaney Paradox. Report of the Committee on Scientific and Regulatory Issues Underlying Pesticide Use Patterns and Agricultural Innovation, Board on Agriculture. National Academy Press, Washington, D.C. 272 pp.

Pao, E.M., K.H. Fleming, P.M. Guenther, and S.J. Mickle. 1982. Foods Commonly Eaten by Individuals: Amount Per Day and Per Eating Occasion. Home Economics Research Report No. 44. Human Nutrition Information Service, U.S. Department of Agriculture, Hyattsville, Md. 431 pp.

Park, Y.K., B.F. Harland, J.E. Vanderveen, F.R. Shank, and L. Prosky. 1981. Estimation of dietary iodine intake of Americans in recent years. J. Am. Diet. Assoc. 79:17-24.

Patterson, B.H., and G. Block. 1988. Food choices and the cancer guidelines. Am. J. Public Health 78:282-286.

Patterson, K.Y., J.T. Holbrook, J.E. Bodner, J.L. Kelsay, J.C. Smith, Jr., and C. Veillon. 1984. Zinc, copper, and manganese intake and balance for adults consuming self-selected diets. Am. J. Clin. Nutr. 40 suppl. 6:1397-1403.

Pennington, J.A.T. 1983. Revision of the total diet study food list and diets. J. Am. Diet. Assoc. 82:166-173.

Pennington, J.A.T., B.E. Young, D.B. Wilson, R.D. Johnson, and J.E. Vanderveen. 1986. Mineral content of foods and total diets: the selected minerals in foods survey, 1982 to 1984. J. Am. Diet. Assoc. 86:876-891.

Prosky, L., N.G. Asp, I. Furda, J.W. DeVries, T.F. Schweizer, and B.F. Harland. 1985. Determination of total dietary fiber in foods and food products: collaborative study. J. Assoc. Off. Anal. Chem. 68:677-679.

Quick, J.A., and E.W. Murphy. 1982. The Fortification of Foods: A Review. Agriculture Handbook No. 598. Food Safety and Inspection Service. U.S. Department of Agriculture, Washington, D.C. 39 pp.

Richards, L., and A.G. Roberge. 1982. Comparison of caloric and nutrient intake of adults during week and weekend days. Nutr. Res. 2:661-668.

Ries, C.P., K. Kline, and S.O. Weaver. 1987. Impact of

Page 84

commercial eating on nutrient adequacy. J. Am. Diet. Assoc. 87:463-468.

Sanjur, D. 1982. Food and food intake patterns—central issues in their conceptualization and measurement. Mass. Agric. Exp. Stn. Bull. (675):73-103.

Scheig, R. 1970. Effects of ethanol on the liver. Am. J. Clin. Nutr. 23:467-473.

Schoenborn, C.A., and B.H. Cohen. 1986. Trends in Smoking, Alcohol Consumption, and Other Health Practices Among U.S. Adults, 1977 and 1983: Advance Data from Vital and Health Statistics of the National Center for Health Statistics, No. 118. DHHS Publ. No. (PHS) 861250. Public Health Service, U.S. Department of Health and Human Services, Hyattsville, Md. 16 pp.

Sempos, C., R. Cooper, M.G. Kovar, C. Johnson, T. Drizd, and E. Yetley. 1986. Dietary calcium and blood pressure in National Health and Nutrition Examination Surveys I and II. Hypertension 8:1067-1074.

Senti, F.R., ed. 1985. Health Aspects of Dietary trans Fatty Acids. Federation of American Societies for Experimental Biology, Bethesda, Md. 148 pp.

Sims, E.A.H., E. Danforth, Jr., E.S. Horton, G.A. Bray, J.A. Glennon, and L.B. Salans. 1973. Endocrine and metabolic effects of experimental obesity in man. Recent Prog. Horm. Res. 29:457-496.

Smith, J.C., Jr., E.R. Morris, and R. Ellis. 1983. Zinc: requirements, bioavailabilities and recommended dietary allowances. Pp. 147-169 in A.S. Prasad, A.O. Çavdar, G.J. Brewer, and P.J. Aggett, eds. Progress in Clinical and Biological Research, Vol. 129: Zinc Deficiency in Human Subjects. Alan R. Liss, New York.

Stamler, J. 1979. Population studies. Pp. 25-88 in RI. Levy, B.M. Rifkind, B.H. Dennis, and N. Ernst, eds. Nutrition, Lipids, and Coronary Heart Disease. Raven Press, New York.

Tannahill, R. 1973. Food in History. Stein and Day, Briarcliff Manor, N.Y. 448 pp.

Taylor, F. 1981. Iodine: going from hypo to hyper. FDA Consumer 15:14-18.

Todhunter, E.N. 1970. A Guide to Nutrition Terminology for Indexing and Retrieval. National Institutes of Health, Public Health Service, U.S. Department of Health, Education, and Welfare, Bethesda, Md. 270 pp.

USDA (U.S. Department of Agriculture). 1981. Food Consumption, Prices, and Expenditures, 1960-1980. Statistical Bulletin No. 672. Economic Research Service, U.S. Department of Agriculture, Washington, D.C. 106 pp.

USDA (U.S. Department of Agriculture). 1983. Food Consumption, Prices and Expenditures, 1962-1982. Statistical Bulletin No. 102. Economic Research Service, U.S. Department of Agriculture, Washington, D.C. 103 pp.

USDA (U.S. Department of Agriculture). 1984. Nationwide Food Consumption Survey. Nutrient Intakes: Individuals in 48 States, Year 1977-78. Report No. 1-2. Consumer Nutrition Division, Human Nutrition Information Service, Hyattsville, Md. 439 pp.

USDA (U.S. Department of Agriculture). 1985. Nationwide Food Consumption Survey. Continuing Survey of Food Intakes of Individuals. Women 19-50 Years and Their Children 1-5 Years, I Day, 1985. Report No. 85-1. Nutrition Monitoring Division, Human Nutrition Information Service, Hyattsville, Md. 102 pp.

USDA (U.S. Department of Agriculture). 1986a. Nationwide Food Consumption Survey. Continuing Survey of Food Intakes of Individuals. Low-Income Women 19-50 Years and Their Children 1-5 Years, 1 Day, 1985. Report No. 85-2. Nutrition Monitoring Division, Human Nutrition Information Service, Hyattsville, Md. 186 pp.

USDA (U.S. Department of Agriculture). 1986b. Nationwide Food Consumption Survey. Continuing Survey of Food Intakes of Individuals. Men 19-50 Years, 1 Day, 1985. Report No. 85-3. Nutrition Monitoring Division, Human Nutrition Information Service, Hyattsville, Md. 94 pp.

USDA (U.S. Department of Agriculture). 1987a. Nationwide Food Consumption Survey. Continuing Survey of Food Intakes of Individuals. Low-Income Women 19-50 Years and Their Children 1-5 Years, 1 Day, 1986. Report No. 86-2. Nutrition Monitoring Division, Human Nutrition 106 pp.

USDA (U.S. Department of Agriculture). 1987b. Nationwide Food Consumption Survey. Continuing Survey of Food Intakes of Individuals. Women 19-50 Years and Their Children 1-5 Years, 1 Day, 1986. Report No. 86-1. Nutrition Monitoring Division, Human Nutrition Information Service, Hyattsville, Md. 98 pp.

USDA (U.S. Department of Agriculture). 1987c. Nationwide Food Consumption Survey. Continuing Survey of Food Intakes of Individuals. Women 19-50 Years and Their Children 1-5 Years, 4 Days, 1985. Report No. 85-4. Nutrition Monitoring Division, Human Nutrition Information Service, Hyattsville, Md. 182 pp.

USDA (U.S. Department of Agriculture). 1988. Nationwide Food Consumption Survey. Continuing Survey of Food Intakes of Individuals. Low-Income Women 19-50 Years and Their Children 1-5 Years, 4 days, 1985. Report No. 85-5. Nutrition Monitoring Division, Human Nutrition Information Service, Hyattsville, Md. 220 pp.

USDA (U.S. Department of Agriculture). In press. USDA Agricultural Handbook No. 8. Comparison of Foods—Raw, Processed, and Prepared. Nutrient Data Research, Human Nutrition Information Service, United States Department of Agriculture. Hyattsville, Md. (various pagings)

van den Reek, M.M., M.C. Craig-Schmidt, J.D. Weete, and A.J. Clark. 1986. Fat in the diets of adolescent girls with emphasis on isomeric fatty acids. Am. J. Clin. Nutr. 43: 530-537.

Welsh, S.O., and R.M. Marston. 1982. Review of trends in food use in the United States, 1909 to 1980. J. Am. Diet. Assoc. 81:120-125.

Windham, C.T., B.W. Wyse, and R.G. Hansen. 1983. Alcohol consumption and nutrient density of diets in the Nationwide Food Consumption Survey. J. Am. Diet. Assoc. 82:364-373.

Wolff, R.J. 1973. Who eats for health? Am. J. Clin. Nutr. 26: 438-445.

Woteki, C.E. 1985. Improving estimates of food and nutrient intake: applications to individuals and groups. J. Am. Diet. Assoc. 85:295-296.

Woteki, C.E. 1986. Dietary survey data: sources and limits to interpretation. Nutr. Rev. suppl. 44:204-213.

Yetley, E., and C. Johnson. 1987: Nutritional applications of the Health and Nutrition Examination Surveys (HANES). Annu. Rev. Nutr. 7:441-463.