Preliminary Estimation of Perchlorate Dietary Exposure Based on FDA 2004/2005 Exploratory Data
FDA's Center for Food Safety and Applied Nutrition (CFSAN) estimated human exposure to perchlorate through consumption of 27 foods and beverages for which CFSAN has collected data on perchlorate levels. Data sources for the 27 foods and beverages are from exploratory surveys conducted by FDA in Fiscal Years 2004 (FY04) and 2005 (FY05). These exploratory surveys were conducted in regions where water sources are known to be contaminated with perchlorate and thus the perchlorate levels are likely biased high. Although the exploratory and limited nature of the surveys does not allow FDA to provide a representative exposure assessment, FDA conducted this preliminary exposure assessment because of significant public interest in the issue of perchlorate exposure from food. FDA plans to publish a comprehensive and nationally representative exposure assessment based on data obtained from FY05 and FY06 Total Diet Study(1) later this year.
Perchlorate Exposure Through Consumption of 27 Foods and Beverages Among Children Aged 2 Years and Over Through Adults
Perchlorate exposure from consumption of 27 foods and beverages was estimated via a Monte Carlo simulation using @Risk software (Environ, Inc., Arlington, VA)(2). The Monte Carlo simulation sums perchlorate exposure values based on food intake and perchlorate concentration levels randomly drawn from a range of possible values for each food from a particular iteration to yield a total exposure for an individual. Iterative individual exposure values result from the multiplication of a food or beverage consumption value (sampled from an Environ Dietary Exposure Assessment (EDEA)-determined cumulative food consumption distribution) with a perchlorate residue value (sampled from a discrete uniform distribution of perchlorate residue values, i.e., every perchlorate concentration value is used in the simulation, with equal probability of occurrence in any given iteration) for the food, taking into consideration the likelihood that a person eats that particular food (percent eaters). Food and beverage consumption values, incorporated in the EDEA software, are from the United States Department of Agriculture (USDA) Continuing Survey of Food Intake by Individuals (CSFII, 1994-1996 and 1998 Supplemental Children's Survey). A total of 5,000 iterations, enough to provide a stable exposure result (i.e., the change in statistics monitored every 50 iterations is less than one percent after 2300 iterations) at the mean and upper percentiles, were performed.
Food consumption based on United States Department of Agriculture (USDA) food commodity codes (ingredient use) (CSFII 1994-1996 and 1998 Supplemental Children's Survey) was used for those foods primarily consumed as an ingredient in a mixed food dish: lettuce, tomatoes, spinach, carrots, apples, cantaloupes, grapes, cucumbers, onions, potatoes, rice, oatmeal, wheat flour, and corn meal. Fruit juices were also calculated based on ingredient use. Food consumption based on USDA food codes (consumption as a whole food) was used for oranges, strawberries, watermelon, broccoli, cabbage, greens, green beans, sweet potatoes, milk, catfish, salmon, and shrimp. The foods were determined to be consumed primarily as an ingredient or as whole so as to maximize the reported intake of the food, i.e., whichever way gave the higher intake of perchlorate was used. For example, because spinach is consumed in higher amounts as an ingredient than as whole, it was categorized as an ingredient that resulted in higher perchlorate intake. Therefore, this preliminary assessment provides highly conservative perchlorate exposure estimates for the specific food commodities included in this analysis.
Table 1 shows the number of each of the 27 foods and beverages collected and analyzed for perchlorate under FY04 and FY05 exploratory surveys, and the average perchlorate levels in part per billion (ppb). Although a total of 125 milk samples were collected and analyzed for the survey, perchlorate levels in three farm raw milk samples were not included for estimating exposure from milk since farm raw milk is not expected to be consumed in any significant amount by the general U.S. population. In addition, apple and orange juices were combined into the fruit juices category.
Summary of exposure estimates
For this preliminary exposure estimate, three population groups were chosen. The first is all persons aged 2 and above and represents the general U.S. population. The second two groups were chosen to represent sensitive subgroups to the adverse health effects of perchlorate. They are: children aged 2-5 years and women of child bearing age (females aged 15-45 years). A summary of the estimated perchlorate dietary exposure values per body weight (at the mean and 90th percentile) from the 27 foods and beverages for the three population groups is displayed in Table 2. Perchlorate residue data for the 27 foods and beverages were combined with food consumption data from the USDA Continuing Survey of Food Intakes by Individuals (CSFII, 1994-1996 and 1998 Supplemental Children's Survey) to estimate perchlorate exposure. Consumption data for all 27 foods and beverages were extracted from CSFII data using the EDEA program. It is assumed that the food consumption survey data reflect long-term consumption of these foods for the whole population, albeit long-term consumption is more plausible for some foods (e.g., milk) than others (e.g., cabbage). Table 3 contains the mean population (all persons aged 2 and above) exposures for each of the individual foods and beverages listed in descending order of the mean population exposures and the cumulative percentage contribution for each.
Results in Table 2 show the estimated total mean population (all persons aged 2 and above) exposure from the 27 foods and beverages to be 0.053 µg/kg bw/day, which is below the Reference Dose (RfD) of 0.7 µg/kg bw/day recommended by the National Academy of Sciences (NAS) (see "Health Implications of Perchlorate Ingestion") and adopted by the Environmental Protection Agency (EPA) (see "EPA Sets Reference Dose for Perchlorate"). This estimate is similar to the geometric mean perchlorate dose of 0.066 µg/kg bw/day for the U.S. population (males and females aged 20 years and older) estimated by Blount et al. (Perchlorate Exposure of the US Population, 2001-2002. J Expo Sci Environ Epidemiol (2006), 1-8). The RfD is an estimate (with uncertainty spanning perhaps an order of magnitude) of a daily oral exposure to the human population (including sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects over a lifetime. The estimated total mean population exposures for children, 2-5 years, and females, 15-45 years, from the 27 foods and beverages are 0.17 µg/kg bw/day and 0.037 µg/kg bw/day, respectively, which are also below the RfD. Furthermore, the estimated total 90th percentile exposures for the three population groups (0.12 µg/kg bw/day for all ages 2+ years; 0.34 µg/kg bw/day for children, 2-5 years; 0.074 µg/kg bw/day for females, 15-45 years) were below the RfD.
The mean and the 90th percentile estimates represent exposure through consumption of 27 foods and beverages which comprise 32 percent or about one-third of the total diet for all ages 2+ years. These estimates pertain only to this limited survey and not to total exposure. However, because this preliminary survey is comprised of foods grown in areas known to have perchlorate present and/or foods with high water content (i.e., those more likely to contain perchlorate residues), the results from this assessment do not represent 1/3 of the total exposure to perchlorate. Many of the foods in the remainder of the diet will contain little or no perchlorate, particularly because they contain much lower levels of water which perchlorate has a particular affinity for. Thus, "scaling up" the mean and the 90th percentile estimates by a factor of 3 to estimate exposure for the total diet would be inappropriate because it would result in significant overestimation of exposure. Similarly, although the 27 foods and beverages comprise 42 percent of the total diet for children, 2-5 years, it would also be inappropriate to "scale up" the FDA estimates for children, 2-5 years by a factor of 2.4 to estimate exposure for the total diet.
For the mean population (all ages 2+ years) perchlorate exposure, Table 3 shows that among the 27 foods and beverages, milk contributed the highest exposure at 0.025 µg/kg bw/day or 47 percent of the estimated total exposure (0.053 µg/kg bw/day). Tomatoes contributed the next highest perchlorate exposure at 0.005 µg/kg bw/day or nine percent of the estimated total exposure. Fruit juices and spinach each contributed 0.004 µg/kg bw/day, or eight percent of the estimated total exposure, while the rest of the foods contributed progressively less to the total mean exposure.
This assessment provides one way of calculating an estimate of dietary exposure to a chemical and is a standard method used by FDA to estimate exposures to chemicals in foods. This method relies on food consumption information that is directly reported by survey respondents in USDA's CSFII. There are other methods to estimate exposure that rely on other CSFII-derived databases or use different kinds of residue data; use of these other methods may result in somewhat different estimates of exposure. For example, exposure estimates derived from perchlorate measurements obtained from FDA's ongoing Total Diet Study are expected to be released in the fall of 2007. This survey is more complete, samples a broader variety of foods, and is expected to provide more accurate estimates of exposure than the preliminary information provided here.
Uncertainties of Exposure Estimates
Exposure estimates through consumption of 27 foods and beverages
Because perchlorate levels in the 27 foods and beverages are exploratory data and consumption of the 27 foods and beverages represents only about 32 and 42 percent of the total diet for the U.S. population, ages 2 years and older and for children, 2-5 years, respectively, sources of uncertainty for this preliminary exposure estimate exist.
For example, because FDA does not have any information on the distribution of perchlorate in foods that constitute the remaining 68 percent of the total diet for U.S. population, ages 2 years and older and 58 percent of the total diet for children, 2-5 years, the true average intake of perchlorate could be higher than the average intake based on the data for the 27 foods and beverages. Further, the sampling set within any particular food type is not statistically representative of that consumed by any individual or available in the marketplace so its relationship to the true distribution of concentrations in the food as consumed in the population is not known. Moreover, many of the produce samples were collected primarily from regions where water sources are known to contain perchlorate (Environmental Working Group: Suspect Salads: Toxic rocket fuel found in samples of winter lettuce. 2003.) and thus, the perchlorate levels are likely to be biased high. Foods with high water content, such as produce, would be expected to contain higher levels of perchlorate, a highly water soluble chemical, and to contribute more to the overall dietary exposure than foods with lower water content.
In addition, a limited number of samples for certain foods (e.g., potatoes, sweet potatoes, seafood), a limited number of samples from specific regions of the country (e.g., all five orange juice samples were collected in a single city), and limited types of foods (i.e., representing about 32 and 42 percent of the total diet for the U.S. population, ages 2 years and older and for children, 2-5 years, respectively) were analyzed. Sampling of additional food types to increase representation of the total U.S. diet, collection of more samples within a food type, and collection of food types from wider regions of the country would better characterize perchlorate distribution in the U.S. food supply. Additional sampling such as the data expected from FDA's forthcoming Total Diet Study will provide a more precise assessment of the scope of perchlorate exposure and the public health implications for food with more reasonable certainty. Nevertheless, this exposure assessment suggests that the overall dietary exposure to perchlorate is likely to be below the RfD recommended by the National Academy of Sciences and adopted by the Environmental Protection Agency.
| Food Groups | Number of Data Points | Average Residue (ppb) |
|---|---|---|
| Lettuce | 137 | 10.3 |
| Milk | 125 | 5.81 |
| Tomatoes | 73 | 13.7 |
| Carrots | 59 | 15.8 |
| Spinach | 36 | 115 |
| Cantaloupes | 48 | 28.6 |
| Apples | 9 | 0.15* |
| Grapes | 12 | 8.58 |
| Oranges | 10 | 3.47 |
| Strawberries | 19 | 2.14 |
| Watermelon | 19 | 1.96 |
| Fruit Juices (Apple & Orange) | 14 | 2.31 |
| Broccoli | 14 | 8.49 |
| Cabbage | 13 | 8.80 |
| Greens | 14 | 92.4 |
| Cucumber | 20 | 6.64 |
| Green Beans | 19 | 6.12 |
| Onions | 12 | 0.53 |
| Potatoes | 6 | 0.15* |
| Sweet Potatoes | 6 | 1.24 |
| Corn Meal | 22 | 1.16 |
| Oatmeal | 22 | 3.96 |
| Rice (Brown & White) | 19 | 0.50* |
| Whole Wheat Flour | 19 | 4.27 |
| Catfish | 7 | 1.02 |
| Salmon | 11 | 1.06 |
| Shrimp | 5 | 19.83 |
| Total | 775 |
*Averages are 1/2 LOD (Limit of Detection) due to perchlorate results for these foods were all Non-Detects
| Population | Monte Carlo estimate using @Risk software with 5,000 iterations* (µg/kg-bw/d) | |
|---|---|---|
| Mean | 90th Percentile | |
| All ages 2+ Years | 0.053 | 0.12 |
| Children, 2-5 Years | 0.17 | 0.34 |
| Females, 15-45 Years | 0.037 | 0.074 |
*Rice (Brown & white) was not included in calculating the mean and 90th percentile total intake values because all rice samples were non-detects and their inclusion did not make a significant difference in the total intake.
a Perchlorate exposure estimates are through consumption of 27 foods and beverages representing 32 and 42 percent of the total diet for all ages 2+ years and children, 2-5 years, respectively. Because this preliminary survey is comprised of foods grown in areas known to have perchlorate present and/or foods with high water content (i.e., those more likely to contain perchlorate residues), the results from this assessment do not represent the total exposure to perchlorate. Many of the foods in the remainder of the diet will contain little or no perchlorate, particularly because they contain much lower levels of water which perchlorate has a particular affinity for. Thus, "scaling up" the mean and the 90th percentile estimates by a factor of 3 and 2.4 to estimate exposure for the total diet for all ages 2+ years and children, 2-5 years, respectively, would be inappropriate because it would result in significant overestimation of exposure.
| Food Groups | Mean Intake (µg/kg bw/day) |
Cumulative Percentage |
|---|---|---|
| Milk | 0.025 | 47 |
| Tomatoes | 0.005 | 56 |
| Fruit Juices (Apple & Orange) | 0.004 | 64 |
| Spinach | 0.004 | 72 |
| Carrots | 0.003 | 78 |
| Lettuce | 0.003 | 84 |
| Cantaloupes | 0.002 | 88 |
| Greens | 0.001 | 90 |
| Broccoli | 0.001 | 92 |
| Green Beans | 0.001 | 94 |
| Grapes | 0.001 | 96 |
| Cucumbers | 0.001 | 98 |
| Oranges, Whole Wheat Flour, Oatmeal, Cabbage, Watermelon, Strawberries, Apples, Onions, Potatoes, Sweet Potatoes, Rice (Brown & White), Corn Meal, Catfish, Salmon, Shrimp |
<> | 100 |
| Total | 0.053 | 100 |
(1) FDA is preparing an exposure assessment based on perchlorate data obtained from FDA's FY05 and FY06 Total Diet Study (TDS). TDS is FDA's ongoing market basket survey in which more than 280 core foods (TDS foods) in the U.S. food supply are collected and analyzed to determine levels of various contaminants and nutrients in those foods.
(2) For further information on the model used by FDA, contact Dr. Michael J. Dinovi at 301-436-1320.