Update of US FDA s Total Diet Study food list and diets

Transcription

1 Journal of Exposure Science and Environmental Epidemiology (2007) 17, r 2007 Nature Publishing Group All rights reserved /07/$ Update of US FDA s Total Diet Study food list and diets SARA KATHLEEN EGAN, PHILIP MICHAEL BOLGER AND CLARK DEWITT CARRINGTON Center for Food Safety and Applied Nutrition, US Food and Drug Administration, College Park, Maryland, USA The US Food and Drug Administration s (FDA) Total Diet Study (TDS) has been conducted continuously since the early 1960s to measures levels of various pesticide residues, contaminants, and nutrients in foods and to estimate the dietary exposures to these compounds. Both the TDS food list and the consumption amounts used for estimating exposures are based on results of nationwide food consumption surveys, and they are updated periodically to reflect changes in food consumption patterns. The most recent update was completed in 2003 using the same methodology employed in the previous update (1990). The updated food list includes approximately the same number of foods (285) as the previous list (290). Although most (75%) foods are the same in both versions, the new list reflects trends in consumption of foods containing less fat. The updated diets reflect an increase in total food consumption, with most notable increases in consumption of grains and beverages. A case study comparing cadmium exposures calculated from both the 1990 and 2003 versions of the TDS demonstrated the potential impact of changes in both the food list and consumption amounts on TDS exposure estimates. Journal of Exposure Science and Environmental Epidemiology (2007) 17, ; doi: /sj.jes ; published online 4 April 2007 Keywords: Total Diet Study (TDS), TDS food list, TDS diets, dietary exposure, monitoring. Introduction The US Food and Drug Administration s (FDA) Total Diet Study (TDS) has been an important component of the federal government s food safety and nutrition monitoring efforts since the early 1960s (Pennington and Gunderson, 1987; Pennington et al., 1996). The goal of the TDS is to monitor levels of selected substances (pesticide residues, elements, and chemical contaminants) in foods as consumed and to estimate the dietary exposure to these substances. This is accomplished through the routine collection and analyses of samples of foods that constitute the major components of the average American diet (TDS food list). The analytical results from the TDS are then combined with information on the amounts of foods consumed (TDS diets) to estimate dietary exposures for various age gender groups within the US population. Details of FDA s TDS are available on the website at 1. Abbreviations: CSFII, Continuing Survey of Food Intakes by Individuals; ND, nondetect; TDS, Total Diet Study; US FDA, United States Food and Drug Administration; USDA, United States Department of Agriculture 2. Address all correspondence to: SK Egan, HFS-301, Center for Food Safety and Applied Nutrition, US Food and Drug Administration, 5100 Paint Branch Parkway, College Park, MD , USA. Tel.: þ Fax: þ Katie.Egan@fda.hhs.gov Received 11 September 2006; accepted 4 December 2006; published online 4 April 2007 Both the TDS food list and diets are derived from the information collected in national food consumption surveys conducted by the US Department of Agriculture (USDA) and the Department of Health and Human Services (HHS). The food list and diets are updated periodically as new consumption data become available to reflect changes in food consumption patterns in the US (Pennington, 1983, 1992a, b). This paper reports on the methodology used for the most recent TDS update, which was completed in 2003, and compares the updated food list and diets to the previous versions, which were compiled in As cadmium levels in TDS foods have remained fairly stable over time, dietary exposure to cadmium was calculated from both versions of the TDS diets to demonstrate the potential impact of the revised food list and diets on exposure estimates. This is an important consideration when comparing exposure estimates over time, as they may be based on different versions of the TDS. Methods The methodology used in the recent update of the TDS food list and diets was similar to the methods used previously (Pennington, 1983, 1992a, b). This update involved a series of tasks in which detailed national food consumption survey data were aggregated so that it was possible to select a limited number of foods (less than 300) to represent the major components of the average American diet, at the same time ensuring that total consumption of all foods was accounted for when estimating dietary exposure to substances measured in the TDS.

2 Egan et al. Update of US FDA s TDS food list and diets The major tasks in the recent update included: Calculating the average consumption of each of 5891 foods as reported in USDA s , 1998 Continuing Survey of Food Intakes by Individuals ( CSFII) (USDA, 2000) for the total US population and 14 age/ gender groups; aggregating the survey foods into groups based on the similarity of their main ingredients; selecting the most representative food from each group of survey foods to be included in the TDS food list; and compiling the TDS diets by subtotaling the consumption amounts for each group of aggregated survey foods for each of the 15 TDS population groups. Theupdatedfoodlistanddietswerethencomparedtothe previous versions compiled in 1990, which were based on the USDA Nationwide Food Consumption Survey (NFCS) (USDA, 1990). Finally, dietary exposure to cadmium was calculated for the 15 TDS population groups using both the 1990 and 2003 versions of the TDS food list and diets to demonstrate the potential impact of the changes in food consumption patterns on exposure estimates. The CSFII was selected for this update because it was the only national dietary survey available at that time for which 2 days of dietary records were collected per individual. The survey data were analyzed using SAS s (SAS Institute Inc., 1999), applying the survey weights developed by USDA. Microsoft s Office Excel was used for all other tasks, including the aggregation of foods, compiling the consumption amounts and diets, and calculating dietary exposure. The methodology is described in more detail below. Calculating Average Consumption Amounts for Each Survey Food The CSFII involved collection of detailed dietary records on two nonconsecutive days for about 20,000 individuals of all ages. The dietary records included consumption information for 5891 different foods (including beverages). The 2-day average per-capita (all individuals) consumption amount was calculated for each food reported in the survey using the appropriate sampling weights developed by USDA for combining all years of the survey (USDA, 2000). This was carried out for each of the 15 TDS population groups (total US and 14 age/gender subgroups) (Table 1). In addition to consumption amount for each survey food, total food consumption and average body weights as reported by survey participants were calculated for each TDS population group. These results were compared to the consumption amounts and body weights that were the basis for the 1990 TDS food list and diets (Table 1). Aggregating the Survey Foods The first step in compiling the food list was to aggregate the 5891 foods reported in the CSFII into groups based on the similarity of their major ingredients. The foods were initially grouped using the aggregation method for the previous (1990) TDS food list as a guide, although changes Table 1. Average total food consumption and body weights for TDS population groups based on national food consumption surveys NFCS a CSFII b TDS population group Body weight (kg) Total food consumption (g/person/day) Body weight (kg) Total food consumption (g/person/day) 6 11 month infants MF 2 years MF 6 years MF 10 years F years M years F years M years F years M years F years M years F 70+ years M 70+ years Total US NA NA, not available. a USDA s Nationwide Food Consumption Survey; basis for the 1990 TDS update. b USDA s , 1998 Continuing Survey of Food Intakes by Individuals; basis for the 2003 TDS update. 574 Journal of Exposure Science and Environmental Epidemiology (2007) 17(6)

3 Update of US FDA s TDS food list and diets Egan et al. in consumption patterns were also taken into account. For the most part, the aggregation method was based on the similarity of primary ingredients (e.g., survey codes for white bread/rolls were grouped with the TDS food white bread and whole wheat bread/rolls were grouped with whole wheat bread ). The ingredient or food state (raw vs. cooked/ processed) was also considered. As an example, survey codes for cooked apple products (applesauce, baked apple, stewed apple) were grouped with the TDS food applesauce and raw apples and apple salad were grouped with the TDS food apple, raw. In some cases, particularly for foods that were consumed less frequently or in smaller quantities, several different types or varieties of foods were grouped together. For example, survey codes for several different legumes (chickpeas, white beans, black beans, cowpeas, lentils) were aggregated to form a single group of survey codes. Many foods reported in the survey were mixtures or had several components, with no one ingredient or component being predominant. This was the case for many entree items that contained a combination of meat or poultry with other ingredients such as a starch or grain product, vegetable, and sauce. Examples include lasagna, beef stew, chicken potpie, frozen entrees, and sandwiches. These foods were aggregated based on the similarity of the major components (e.g., meat dishes with rice, vegetable and sauce; meat dishes with noodles, vegetables, and sauce; and similar combinations with poultry or fish rather than meat). Selecting Representative Foods to be Included in the TDS Food List After the aggregation of survey foods was completed, the consumption amounts within each group of foods were reviewed and the food consumed in greatest quantity was selected to represent the group. This subset of survey foods constituted the updated TDS food list. The updated food list and the consumption amounts were also compared to the previous TDS food list to identify changes in consumption patterns. As an example, low fat cottage cheese replaced regular (whole milk) cottage cheese to reflect the change in consumption reported in the CSFII. In cases where more significant changes in consumption patterns were observed, a new TDS food was added (e.g., tomato salsa) or an existing food was deleted (e.g., evaporated milk). Consumption amounts for the US population as a whole were the basis for the initial aggregation and selection of the TDS foods. Consumption patterns for the other age gender groups were reviewed in a subsequent step. Although the TDS foods generally represent the major components of the US diet based on the average amounts consumed, an attempt was also made to include as broad a range of foods as possible. This was particularly true for fruits and vegetables, which encompass a wide variety of foods with varied consumption frequencies and amounts. As an example, the TDS food list includes asparagus and beets, which have low per-capita consumption when compared to other TDS vegetables such as green beans or corn. After the list of TDS foods was finalized, each food was assigned a number. To ensure consistency from one version of the TDS to another, foods that were retained from the 1990 food list also retained the same food number. When a TDS food was deleted, its number was not reassigned; when a new food was added, the next sequential number was assigned. Each TDS food was also assigned to one of 12 food categories (Table 2). Previous versions of the TDS assigned foods to one of 14 categories, using a unique alpha-numeric category code for each food (Pennington, 1992a, b). For the 2003 update, this approach was modified and the individual category codes were replaced by a general category designation based on either commodity (e.g., dairy, eggs, grains, vegetables) or food type (e.g., beverages, baby foods, and mixtures that are mainly soup and entree items). Two categories used previously (desserts and snacks) were excluded. The foods in these two categories represented more than one commodity group (e.g., desserts included ice cream, cakes, and gelatin), so they were reassigned to categories on the basis of the primary ingredient in order to be more consistent with the CSFII food coding structure Compiling the TDS Diets The TDS diets comprise a set of consumption amounts for each TDS food for the total US population and 14 age/ gender subgroups. These consumption amounts are combined with analytical results for TDS foods to calculate dietary exposure to TDS analytes. As discussed above, consumption values for each survey food were calculated from results of the CSFII for the total US population and each of 14 age/gender groups. Survey results for these 14 subgroups were aggregated, as they had been for the total US population in the previous steps. The consumption data for infants and toddlers were carefully reviewed to ensure that the TDS food list included a broad range of infant/toddler foods and formulas. Consumption data for the other age gender subgroups were also reviewed to ensure that no significant dietary component had been excluded from the TDS food list. Then the consumption amounts for each group of aggregated survey foods were summed to derive the TDS diet for each of the population groups. For example, the TDS diet amount for the TDS food applesauce is the sum of the per-capita consumption amounts for that group of aggregated survey foods for cooked apples (applesauce, baked apples, and stewed apples). This approach assumes that the analytical profile of the aggregated survey foods would be similar to that of the TDS food representing those foods, and thus the dietary exposure calculated from the Journal of Exposure Science and Environmental Epidemiology (2007) 17(6) 575

4 Egan et al. Update of US FDA s TDS food list and diets Table 2. TDS food categories food list 2003 food list TDS food category Category descriptions O O Dairy products All dairy products (e.g., butter, milk, cheese, ice cream). Previously included only milk and cheese O O Eggs Items that are primarily eggs (e.g., boiled egg, scrambled egg, omelet, egg salad) O O Baby foods All baby foods and infant formulas (excluding adult foods consumed by children) O O Meat, poultry, fish Items that are primarily meat, poultry, or fish (e.g., roasts, fried chicken, fish filets, luncheon meats) O O Legumes Legumes, nuts, and seeds O O Grains Items that are primarily grains. Now includes cookies and pastries, which were previously categorized as Desserts O O Fruits Fruits and fruit juices O O Vegetables Vegetables and vegetable juices O O Mixtures Primarily entree items containing mixtures of meat/poultry/fish, grains, and vegetables (no predominant ingredient) O Desserts Excluded from the 2003 update. Foods previously in this category are now grouped by their primary ingredient (e.g., ice cream is now categorized as a dairy product) O Snacks Excluded from the 2003 update. Foods previously in this category are now grouped by their primary ingredient (e.g., popcorn is now categorized as a grain) O O Sweets Sugars, sweeteners, syrups, candy, jelly, gelatin. Previously included condiments, which are now grouped with vegetables O O Fats/oils Vegetable fats and oils and salad dressings. Previously included some foods (e.g., butter, sour cream, cream) that are now categorized as dairy products O O Beverages Beverages except for fruit/vegetable juices TDS diets provides a reasonable estimate of exposure from the total diet. The updated diets for all 15 population groups were then compared to the 1990 diets on the basis of consumption by food category. As noted above, the 2003 update of the food list included several changes to the categorization of TDS foods. To ensure consistency when comparing the two versions of the food list and diets, the updated categorization was used for the 1990 version as well. Estimating Dietary Exposure using the TDS Diets Estimates of dietary exposure to cadmium were calculated from the 1990 and the 2003 versions of the TDS food list and diets for the total US and each of the 14 population subgroups. Cadmium was selected as a case study and example because it is a well-known environmental contaminant found at fairly consistent levels in a wide range of foods and could, therefore, demonstrate the potential impact of changes in food consumption patterns alone on dietary exposures estimated from the TDS diets. In calculating exposures, one set of cadmium concentration values (average of results for , assuming nondetect (ND) ¼ 0) were used for foods that were common to both versions of the TDS food list. For foods that were deleted from the food list in 2003, the average of cadmium results for were used; for foods that were added to the list in the 2003 update, the estimates were based on average cadmium concentrations for From both versions of the TDS diet and for each of the 15 population groups, exposure estimates were calculated for each TDS food and then summarized by category. The updated categorization of TDS foods was used for both sets of calculations. Results Comparison of Total Food Consumption and Body Weights Table 1 compares average total food consumption and body weights for the 1990 and 2003 versions of the TDS as calculated from the NFCS and CSFII, respectively, for each of the 15 TDS population groups. For the population as a whole, total food consumption reported in the CSFII was 18% greater than reported in the previous survey. The greatest relative increases in total food consumption were seen for infants 6 11 months (35%), boys years (34%), and men and years (26%). For most TDS population groups, self-reported body weights were the same or slightly higher when compared to the previous survey; average body weights of men 70 years and older decreased by nearly 10%. TDS Food List, Food Categories and Diets A description of each food category and notable changes between the 1990 and 2003 versions of the TDS are presented in Table 2. As noted above, two categories used previously (desserts and snacks) were excluded in the 2003 update and these foods were reassigned to categories on the basis of the primary ingredient in order to be more consistent with the CSFII food coding structure. 576 Journal of Exposure Science and Environmental Epidemiology (2007) 17(6)

5 Update of US FDA s TDS food list and diets Egan et al. Table 3 compares the 1990 and 2003 versions of the TDS food list and diets by food category, listing the total number of foods and consumption amounts for the total US population per category. The total number of TDS foods per list is similar (290 in the previous food list compared with 285 in the updated list). Note that the food list as revised and documented in 1990 included 265 foods; this number was increased to 290 in 1991 when additional baby foods were added to increase monitoring of foods consumed by children. Most (75%) of the TDS foods are common to both versions of the food list. As for changes in 2003, 72 foods from the previous list were deleted, and 68 new foods were added. Of the 68 new foods, about one-third are similar to a food that was deleted; in most cases, these are lower fat versions of foods included in the 1990 food list (e.g., cottage cheese with 2% rather than 4% milk fat, tuna packed in water rather than oil) or prepared foods rather than homemade (e.g., coleslaw from a delirather than prepared from a recipe). The remaining two-thirds of the foods added in 2003 are new foods reflecting changes in consumption patterns; examples include tomato salsa, chicken/turkey luncheon meats, catfish, and granola bars. A comprehensive table of all foods in both the 1990 and 2003 versions of the TDS food list is available on the TDS website at tds-toc.html. The major changes in consumption patterns by food category are also noted in Table 3. Total consumption of dairy products was similar for both versions, although consumption of whole milk decreased while low fat and skim milk consumption increased. In the fruit/fruit juice category, juices accounted for the majority of increased consumption in Increased consumption of pasta and rice was most significant for the grain category. For mixtures, substantial increases were seen for pasta items (spaghetti, lasagna), fast-food items (hamburgers, cheeseburgers, chicken sandwiches, breakfast muffins), Mexican foods (burritos and tortillas), and pizza. Total consumption of vegetables was similar in both versions of the TDS diets. Although there was a notable increase in consumption of tomato sauce and salsa in 2003, this was offset by slight decreases in consumption of a number of vegetables, especially potatoes. The detailed TDS diets for each population group, with consumption information for each TDS food, are available on FDA s TDS website. Table 4 summarizes and compares consumption amounts (total and by food category) for the 1990 and 2003 versions Table 3. Comparison of the 1990 and 2003 TDS food lists and diets for the total US Food category No. of foods per category Consumption (g/day) a No. of foods per category Consumption (g/day) a Change in consumption amounts (2003 vs. 1990) (%) Changes in consumption patterns Baby foods/infant formulas Decreased consumption of most foods/formulas on total US per-capita basis Beverages b Increased consumption of carbonated beverages and fruit drinks (accounts for 80% of the total increase in this category) Dairy Decreased consumption of milk; slight increase in consumption of cheese, yogurt, light ice cream; increased consumption of low-fat products Eggs No significant changes Fats/oils No significant changes Fruits/juices Increased consumption of juices Grains Increased consumption of most types of grain products, particularly rice and pastas Legumes/nuts Slight increase in consumption of legumes Mixtures Increased consumption of take-out and fast foods (sandwiches, pizza, Chinese and Mexican foods) and soups Meat, poultry, fish Increased consumption of poultry, fish; decreased consumption of meats Sweets Increased consumption of candies Vegetables Increased consumption of tomatoes (sauce, salsa); slight decrease for potatoes, other vegetables Total a Consumption amounts are for total US population. b Excluding drinking (tap or bottled) water. Journal of Exposure Science and Environmental Epidemiology (2007) 17(6) 577

6 Egan et al. Update of US FDA s TDS food list and diets Table 4. Comparison of 1990 and 2003 TDS diets by category. MF 6 11 months M/F 2 years M/F 6 years M/F 10 years Change from Change from Change from Change from 1990 Food category g/day g/day g/day % g/day g/day g/day % g/day g/day g/day % g/day g/day g/day % Baby foods Beverages Dairy Eggs Fats/oils Fruits/juices Grains Legumes/nuts Mixtures Meat, poultry, fish Sweets Vegetables Total F years M years F years M years Food category Change from Change from Change from Change from 1990 g/day g/day g/day % g/day g/day g/day % g/day g/day g/day % g/day g/day g/day % Baby foods Beverages Dairy Eggs Fats/oils Fruits/juices Grains Legumes/nuts Mixtures Meat, poultry, fish Sweets Vegetables Total F years M years F years M years Change from Change from Change from Change from 1990 Food category g/day g/day g/day % g/day g/day g/day % g/day g/day g/day % g/day g/day g/day % Baby foods Beverages Dairy Eggs Fats/oils Fruits/juices Grains Legumes/nuts Mixtures Meat, poultry, fish Sweets Vegetables Total Journal of Exposure Science and Environmental Epidemiology (2007) 17(6)

7 Update of US FDA s TDS food list and diets Egan et al. Table 4. Continued F 70 years M 70 years Change from Change from 1990 Food category g/day g/day g/day % g/day g/day g/day % Baby foods Beverages Dairy Eggs Fats/oils Fruits/juices Grains Legumes/nuts Mixtures Meat, poultry, fish Sweets Vegetables Total of the TDS diets for the 14 age/gender groups. The greatest increases in total food consumption in the 2003 diets were seen for infants 6 11 months (35%), boys years (34%) and men and years (26% for each). For infants, the majority of increased consumption is attributed to baby foods/infant formula, fruits/fruit juices, grains, and mixtures. For most population groups, there were substantial increases in consumption of beverages (mainly carbonated soft drinks) as well as fruits/fruit juices, grains, and mixtures. Beverage consumption alone increased by over 400 g/day for boys years and men years of age. Estimated Dietary Exposure to Cadmium Estimates of dietary exposure to cadmium based on the two versions of the TDS diets are presented in Table 5. As noted previously, the same cadmium concentration values were used for foods common to both versions of the food list and diets when calculating the exposures, which was the case for approximately 75% of the foods. As for the remaining 25%, about one-third of the foods were quite similar, differing mainly in fat content or cooking method. Thus, differences in the exposure estimates presented here primarily reflect the impact of changes in the TDS diets rather than change in cadmium levels in foods. Total exposure to cadmium based on the updated TDS diets increased for all population groups. The increases in exposure ranged from 0.9 mg/person/day for children 6 years of age to 3.73 mg/person/day for men years of age, which represent relative increases in daily exposure per person of 12% and 35%, respectively. For infants 6 11 months of age, exposure increased by 1.36 mg/person/day (88%). When expressed on a body weight basis, cadmium exposures increased by 16% or less for seven of 15 population groups. Infants aged 6 11 months had the highest relative increase (46%) followed by men years of age (33%). The trend in increased consumption of grains and mixtures as discussed above was reflected by the increase in cadmium exposures from those foods categories. For the total US population, those two categories combined accounted for nearly 60% of the total increase in cadmium exposure. In addition to the impact of changes in dietary patterns, the cadmium exposure example demonstrated the potential impact on exposure estimates of a change in the TDS food list. The food category legumes/nuts/seed accounts for a very small proportion (less than 1%) of total food consumption for the total US population and the increase in consumption between 1990 and 2003 was only 2 g/person/ day (Table 3). Nevertheless, this food category accounted for 19% of the total increase in cadmium exposure (i.e., 0.42 of 2.25 mg/day) (Table 5). The main source of the increased cadmium exposure from legumes/nuts/seeds was the TDS food sunflower seeds, which was added to the food list in Sunflower seeds, which are known to contain higher levels of cadmium than most foods (WHO, 2004), had the highest concentration (0.5 mg/kg) of all TDS foods included in the exposure estimate for Although the TDS diet consumption amount for sunflower seeds is less than 1 g/person/day for the total US population (refer to the TDS website), its addition to the 2003 food list and diets and the relatively high cadmium concentration had a significant impact on the exposure estimate. Discussion The TDS has been an on-going component of FDA s contaminant and nutrition monitoring programs for many Journal of Exposure Science and Environmental Epidemiology (2007) 17(6) 579

8 580 Journal of Exposure Science and Environmental Epidemiology (2007) 17(6) Table 5. Dietary exposure to cadmium based on TDS diets (1990 vs. 2003). Total US M/F 6 11 months M/F 2 years M/F 6 years Change Change Change Change Food category (mg/person/day) (mg/person/day) (mg/person/day) (mg/person/day) Baby foods Beverages Dairy Eggs Fats/oils Fruits/juices Grains Legumes/nuts/seeds Mixtures Meat, poultry, fish Sweets Vegetables Total exposure (mg/person/day) Relative change (per person/day) 26% 88% 36% 12% Body weight (kg) NA Total exposure (mg/kg body weight/day) Total exposure (% PTDI 1 ) 21% 22% 32% 10% 36% 45% 9% 33% 38% 4% Relative change (per kg body weight/day) 46% 26% 12% M/F 10 years F years M years F years Change Change Change Change Food category (mg/person/day) (mg/person/day) (mg/person/day) (mg/person/day) Baby foods Beverages Dairy Eggs Fats/oils Fruits/juices Grains Legumes/nuts/seeds Mixtures Meat, poultry, fish Sweets Vegetables Total exposure (mg/person/day) Relative change (per person/day) 10% 18% 28% 26% Body weight (kg) Total exposure (mg/kg body weight/day) Total exposure (% PTDI 1 ) 23% 26% 3% 15% 16% 1% 16% 20% 4% 13% 16% 3% Relative change (per kg body weight/day) 13% 10% 26% 24% Egan et al. Update of US FDA s TDS food list and diets

9 Table 5. Continued Journal of Exposure Science and Environmental Epidemiology (2007) 17(6) 581 M years F years M years F years Change Change Change Change Food category (mg/person/day) (mg/person/day) (mg/person/day) (mg/person/day) Baby foods Beverages Dairy Eggs Fats/oils Fruits/juices Grains Legumes/nuts/seeds Mixtures Meat, poultry, fish Sweets Vegetables Total exposure (mg/person/day) Relative change (per person/day) 35% 23% 20% 12% Body weight (kg) Total exposure (mg/kg body weight/day) Total exposure (% PTDI 1 ) 14% 18% 4% 13% 16% 3% 14% 17% 2% 13% 14% 1% Relative change (per kg body weight/day) 33% 23% 16% 10% M years F 70+ years M 70+ years Change Change Change Food category (mg/person/day) (mg/person/day) (mg/person/day) Baby foods Beverages Dairy Eggs Fats/oils Fruits/juices Grains Legumes/nuts/seeds Mixtures Meat, poultry, fish Sweets Vegetables Total exposure (mg/person/day) Relative change (per person/day) 17% 14% 10% Body weight (kg) Total exposure (mg/kg body weight/day) Total exposure (% PTDI 1 ) 14% 15% 2% 13% 14% 2% 14% 17% 3% Relative change (per kg body weight/day) 14% 13% 22% 1 PTDI ¼ 1 mg/kg body weight/day. Based on PTWI of mg/kg body weight/week. Update of US FDA s TDS food list and diets Egan et al.

10 Egan et al. Update of US FDA s TDS food list and diets years. The analytical results from the TDS provide a wealth of information about trends in levels of pesticides residues, contaminants, and nutrients in the food supply as well as the dietary exposures to these substances from a typical American diet. Periodic updates of TDS food list and diets are essential for ensuring that the program reflects current trends in the average American diet in terms of both content and quantity. At the same time, changes resulting from updates of the food list and diets add another variable in the estimation of dietary exposure. For that reason, it is important to consider the impact of such updates when comparing TDS dietary exposure estimates that have been calculated over time. Differences in body weights from one version of the TDS to another also have the potential to impact exposure estimates expressed on a body-weight basis, although the changes between 1990 and 2003 were minor for most TDS population groups (Table 1). The majority (75%) of TDS foods were included in both the 1990 and 2003 food lists, thus many foods have been staples in the average American diet since the late 1980s. The updated TDS diets demonstrate a clear trend in increased consumption of grains products and mixtures, which often contain grain as a major ingredient. Consumption of beverages F particularly soft drinks and fruit drinks F also increased substantially. These trends in consumption by children, adolescents, and adults were also noted in a series of publications by Enns et al. (1997) and Enns et al. (2002 and 2003). The cadmium dietary exposure example presented in this paper highlighted the potential impact of such dietary changes on exposure estimates calculated from different versions of the TDS diets. Dietary exposure to cadmium increased between 1990 and 2003 in large part because of the increased consumption of grains. Many changes in the food list were a reflection of trends in reducing intake of dietary fat by consuming foods containing less fat or preparing foods with less cooking fat. Such trends in consumption patterns will have an impact on exposure estimates, particularly for fat-soluble compounds, to the extent that these foods are sources of the TDS analyte for which an exposure might be estimated. Estimates of exposure to other substances that are found in a limited number F and different types F of foods (e.g., arsenic) would likely be affected less by the updated food list and diets. In recent years, FDA has made TDS information, including analytical results and the TDS diets, accessible through the TDS website at ~comm/tds-toc.html. This provides data users a means of downloading and analyzing individual TDS results rather than having to rely on published summary tables. It also enables data users to calculate dietary exposures based on the analytical results and the TDS diets. The files related to the 2003 TDS update provide detailed information about the aggregation of the survey foods, which helps to demonstrate howthetdsfoodlistanddietsrelatetotheresultsofthe CSFII. Finally, it is important to remember that the focus of the TDS is the typical year-round diet in the US. It is not designed to monitor levels of pesticide residues, contaminants, or nutrients in foods that are strictly regional or seasonal in nature, or foods that are consumed only by subsets of the population (e.g., ethnic groups). Likewise, the TDS does not provide estimates of exposures for individuals or distributions of exposure in a population. The TDS, however, provides an overall estimate of dietary exposures to pesticide residues, contaminants, and nutrients, and it does so over protracted periods of time. Acknowledgements We would like to acknowledge several FDA colleagues in the Center for Food Safety and Applied Nutrition for their contributions to the TDS update: Eric Hanson for his analysis of the CSFII data, and Donna Robie for her assistance in compiling the TDS diets. Especially we thank Kevin Cline of FDA s Kansas City District Laboratory for his assistance in updating the TDS recipes and shopping lists. References Enns C.W., Goldman J.D., and Cook A. Trends in food and nutrient intakes by adults: NFCS , CSFII , and CSFII Fam Econ Nutr Rev 1997: 10(4): Enns C.W., Mickle S.J., and Goldman J.D. Trends in food and nutrient intakes by children in the United States. Fam Econ Nutr Rev 2002: 14(2): Enns C.W., Mickle S.J., and Goldman J.D. Trends in food and nutrient intakes by adolescents in the United States. Fam Econ Nutr Rev 2003: 15(2): Pennington J.A.T. Revision of the Total Diet Study food list and diets. Journal of the American Dietetic Association 1983: 82(2): Pennington J.A.T., and Gunderson E.L. History of the Food and Drug Administration s Total Diet Study F 1961 to J AOAC Int 1987: 70(5): Pennington J.A.T. The 1990 revision of the FDA Total Diet Study. J Nutr Educ 1992a: 24(4): Pennington J.A.T. Revision of the Food and Drug Administration s Total Diet Study Food List and Diets, Appendices 1992b NTIS no. PB Pennington J.A.T., Capar S.J., Parfitt C.H., and Edwards C.W. History of the Food and Drug Administration s Total Diet Study (part II) F J AOAC Int 1996: 79(1): SAS Institute Inc. SASs, Version 8 SAS Institute Inc., Cary, NC, US Department of Agriculture (USDA). Human Nutrition Information Service. Nationwide Food Consumption Survey , Individual Intake Accession no. PB , National Technical Information Service, Washington, DC. US Department of Agriculture (USDA). Agricultural Research Service. Continuing Survey of Food Intakes by Individuals , 1998 [CD-ROM] NTIS no. PB , National Technical Information Service, Washington, DC. World Health Organization (WHO). Cadmium. In: Safety Evaluation of Certain Food Additives and Contaminant, Prepared by the Sixty-first Meeting of the Joint FAO/WHO Expert Committee on Food Additives (JECFA). WHO food additives series, Geneva, Journal of Exposure Science and Environmental Epidemiology (2007) 17(6)