"Keeping Pace with a Changing Food Supply"

23rd National Nutrient Databank Conference

April 16, 1999 $ Washington, DC

rainbow.gif (1314 bytes)


COMPARISON OF TWO NUTRIENT DATABASES WITH DIRECT FOOD ANALYSIS FOR ESTIMATING THE PHYLLOQUINONE (VITAMIN K1) CONTENT OF METABOLIC DIETS. Nicola McKeown, Helen Rasmussen, Kathleen McGann, Brian Kaszynski, Richard Wood, Sarah Booth. Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University, 711 Washington St., Boston, MA 02111.

Food and nutrient databases continually evolve to fulfill a wide variety of purposes, ranging from designing diets for metabolic studies to establishing diet-disease associations in epidemiologic studies. The purpose of this study was to determine if current nutrient databases that include phylloquinone (vitamin K1) can be used to design diets for vitamin K metabolic studies. The phylloquinone content of six diets were calculated using (1) the Nutrient Data System for Research (NDS-R; version 4.0) and (2) the USDA provisional table for vitamin K. The phylloquinone content of these diets was verified by analysis of homogenates using high performance liquid chromatography (HPLC). The mean " sd phylloquinone content of the six diets was 28.0 " 32.2 mg ( range: 10.4 to 93.1 mg), as determined by food analysis. Compared to food analysis, the phylloquinone contents of the six diets was underestimated by a mean " sd of 2.7 " 5 .5 mg (range of differences: -12.5 to +3.6 mg) using the provisional table, and by 4.3 " 6.0 mg (range of differences: -13.2 to +1.7 mg) using the NDS. For diets with a target phylloquinone content of <10 mg, the nutrient databases underestimate by as much as 60%. The mean " sd difference in phylloquinone content between the provisional table and the NDS was 1.6 " 3.2 mg. When designing metabolic diets for vitamin K studies, direct food analysis and verification of sources of vitamin K food composition is important for confirmation of actual phylloquinone intake.

DEVELOPMENT OF A NUTRIENT DATABASE FOR FRESH VEGETABLE AND FRUIT JUICES. Newman VA, Rock CL, Zoumas C, Faerber S, UCSD Cancer Prevention and Control, La Jolla, CA 92093-0901.

To augment the limited nutrient composition data available on fresh vegetable and fruit juices, a database was created, which includes the nutrient composition for 33 fresh vegetable and 17 fresh fruit juices. The associated software calculates the nutrient composition of any combination of juices found in this database. Nutrient composition of each juice in the database was calculated from the metric weight of raw produce required to produce 8 fl oz of juice adjusted for nutrients likely in the discarded pulp. The nutrient content of the raw produce was obtained from Nutrient Data System (NDS 2.92, University of Minnesota, Minneapolis, MN), augmented with an associated program to obtain updated carotenoid content (USDA-NCI carotenoid food composition database). The amount of fresh produce required to make a cup (8 fl. oz.) of juice was derived by weighing using a gram scale calibrated to the nearest 0.10 gram using two different juicer brands (Omega and MJ1000). Produce was weighed and juiced on three separate occasions using a standard protocol, and the mean weights were used. This study resulted in the development of a database and associated software that calculates the nutrient content of 50 fresh juices, or any combination of these juices.

MISCELLANEOUS VEGETABLES:  UPDATED COMPOSITES IMPROVE NUTRIENT CONTRIBUTIONS TO THE FOOD SUPPLY.  Bente, Lisa MS and Gerrior, Shirley PhD, United States Department of Agriculture, The Center for Nutrition Policy and Promotion, Washington D. C. 20036

The U.S. Food Supply Series is a historical data series beginning in 1909. It measures the amount of food and nutrients available for consumption in the United States. This study uses USDA=s Nutrient Database for Standard Reference 12 to calculate accurate and complete food supply nutrient estimates from miscellaneous vegetables. Historically, nutrient estimates for the miscellaneous vegetables did not represent the vegetable mix in the food supply, were calculated using a spreadsheet and subject to entry errors, and had no direct link to PDS codes. To make appropriate adjustments to the food supply nutrient file, canned and fresh vegetable composites were reviewed for their vegetable mix, PDS links, and nutrient estimates for food energy, iron, vitamins A and C, and folate. The new composites use a more representative mix of vegetables, and link each vegetable component to its appropriate PDS code. Revised estimates are different for some nutrients due to the change in the vegetable mix. For revised miscellaneous canned vegetable composites (1940-94) vitamin A is higher and iron is higher or the same for all years except 1965 and 1969, vitamin C is higher for 1970-94, and both food energy and folate are lower for 1940-94. For miscellaneous fresh vegetable composites vitamin C is lower for 1970-94 and folate is higher for 1909-94 reflective of new methods used to determine folate in foods. Other nutrients show little change during the period 1909-94. This study shows that the food supply nutrient estimates are improved by these updated vegetable composites.

ACCOMMODATING UNIQUE RECIPES: THE NDS-R USER-RECIPE FEATURE.   Lisa Harnack, Priscilla Goldstein, Mary Stevens, Nancy Van Heel, Gordon Weil, Nutrition Coordinating Center, Division of Epidemiology, University of Minnesota

Nutrition Data System for Research (NDS-R), a new Windows-based nutrient analysis program for personal computers, includes a user-recipe feature which allows the user to enter unique recipes for use and re-use in dietary records. The food and nutrient database linked with NDS-R includes commonly eaten foods prepared from recipes. Ingredient variables for such recipes may be selected during data entry to provide further specificity (e.g. ground beef percent fat). The NDS-R user-recipe feature allows for entry of recipes not included in the database such as recipes specific to various regions of the US; recipes modified for dietary interventions; and recipes that are unique due to ingredient adjustments. The user-recipe feature guides data entry with prompts for complete food descriptions and food preparation methods. Automatic conversion of ingredient amounts from raw-to-cooked values; ability to enter the portion with or without refuse; algorithms for fat and sodium uptake according to preparation method; and options for entering amounts in food-specific units and food shapes in addition to common units of volume and weight expedite entry of ingredients. A food group assignment may be selected for the recipe to assist in food group analysis. User-recipes may be analyzed individually or as part of other dietary records and may be recalculated in subsequent database versions to obtain nutrient updates. In addition to output files in ASCII text format, recipe reporting options include nutrient totals, nutrient values per individual ingredient and comparisons with Recommended Dietary Allowances and Daily Values for Nutrition Labeling.

DEVELOPMENT OF A NATIONWIDE SAMPLING PLAN FOR THE NATIONAL FOOD AND NUTRIENT ANALYSIS PROGRAM. P Pehrsson, D Trainer, D Haytowitz, J Holden, J Evans, C Perry, and D Beckler (SPON: G. Beecher). Nutrient Data Lab, BHNRC, ARS, USDA, Riverdale, MD and National Agricultural Statistics Service, USDA, Fairfax, VA.

The National Food and Nutrient Analysis Program (NFNAP) is a five-year research program designed to significantly improve the quality of food composition data in the USDA National Nutrient Data Bank. NFNAP consists of five Aims: 1) evaluation of existing data; 2) identification of foods (Key Foods) and nutrients for analysis; 3) development of nationally based sampling plans; 4) analysis of samples; and 5) calculation of representative estimates for components. For Aim 3, the US was divided into four regions, with nearly equal populations; each region was divided into three strata of nearly equal population. Generalized Consolidated Metropolitan Statistical Areas (gCMSAs) were selected in each stratum proportional to population size and supplemented with contiguous counties when the gCMSA contained less than 10 grocery stores. Grocery store lists were obtained through Trade Dimensions (Wilton, CT) for selection of primary and alternate outlets for food pickups. Individual brands and varying package sizes were selected using current market volume share data (as pounds consumed). Additional samples for determination of serving to serving variation were collected where considerable variation was expected for certain nutrients. The NFNAP sampling plan represents substantial advancements in providing more accurate, representative, and statistically robust estimates for components of important foods in the US food supply. Details of the study design and food composition data for selected foods are presented. Work supported by an USDA-NIH interagency agreement.

DEVELOPMENT OF REPRESENTATIVE NUTRIENT DATA FOR MARGARINE AND SPREADS UNDER THE NATIONAL FOOD AND NUTRIENT ANALYSIS PROGRAM. P. Pehrsson, J. Evans, D. Trainer, D. Haytowitz, J. Holden, C. Perry, and D. Beckler (SPON: G. Beecher). Nutrient Data Lab, BHNRC, ARS, USDA, Riverdale, MD and National Agricultural Statistics Service, USDA, Fairfax, VA.

The National Food and Nutrient Analysis Program (NFNAP) is a five-year research program designed to significantly improve the quality of food composition data in the USDA National Nutrient Data Bank. This involves evaluation of existing data, identification of foods (Key Foods) for analysis, development of nationally based sampling plans, analysis of samples, and calculation of representative estimates for components. Previously, margarines and spreads were identified as the leading contributors of fat in the US diet. However, major changes in product formulations emphasized the need for new composition data. Within the past decade, new products include those with fat levels that range from near 0 to 80%, and tremendous variation in the types of oils used, level of hydrogenation, moisture content, and physical form (soft tub to hard stick). Recently, NDL sampled selected margarines and spreads according to the population-based sampling plan for NFNAP; samples were picked up in each of 12 Generalized Consolidated Statistical Areas, across four regions of the US and three strata within each region. In order to sample these products, individual brands were categorized by fat range (fat-free, 10-30%, 37-50%, 52-65%, 70%, and 80%). Multiple nationwide composites by fat range and brand were developed; then analyzed for proximates, individual fatty acids, vitamins and minerals. Preliminary data are presented for these foods. Work supported by an USDA-NIH interagency agreement.

ISOFLAVONE DATABASE FOR FOODS. G. R. Beecher. S.Bhagwat. J. M. Holden. D. Haytowitz and P. A. Murphy. Beltsville Human Nutrition Reseach Center, ARS, USDA, Beltsville, MD 20705 and Department of Food Science and Human Nutrition, Iowa State University, Ames, IA 50011

Isoflavones, well known phytonutrients of soybeans and soy-containing foods, have low- and/or anti-estrogenic as well as antioxidant properties. However, they have not been directly linked with risk of chronic diseases for lack of a database of food values. We have developed such a database. About 30 scientific papers were identified which contained data on the isoflavone content of soy-based and a few other foods. Also, isoflavone data were generated from an extensive sampling of soy-containing foods in the U.S. Data for daidzein, genistein, glycitein and their glucosides were critically evaluated (Mangels et al., J Am Diet Assoc. 93:284-296, 1993). Total aglycone values of each isoflavone were calculated by converting glucoside forms into aglycones based on molecular weight ratios and adding them to their respective free aglycone values. Acceptable data for each isoflavone/food combination were combined to generate mean values for aglycone forms and total isoflavone content. The number of values used to calculate the mean, standard errors of the means (SEM) and Confidence Codes (reliability indicator) also were determined. Data for about 140 foods, including soy ingredients, were tabulated and assigned USDA Nutrient Database codes. The source of data for each food is shown and all references are listed. The database is available at http://www.nal.usda.gov/fnic/foodcomp. Partially funded by U.S. Army.

rainbow.gif (1314 bytes)

Nutrient Databank Conference Home Page