Community duplicate diet methodology: A new tool for estimating dietary exposures to pesticides

Abstract

An observational field study was conducted to assess the feasibility of a community duplicate diet collection method; a dietary monitoring tool that is population-based. The purpose was to establish an alternative procedure to duplicate diet sampling that would be more efficient for a large, defined population, e.g., in the National Children's Study (NCS). Questionnaire data and food samples were collected in a residence so as not to lose the important component of storage, preparation, and handling in a contaminated microenvironment. The participants included nine Hispanic women of child bearing age living in Apopka, FL, USA. Foods highly consumed by Hispanic women were identified based on national food frequency questionnaires and prioritized by permethrin residue concentrations as measured for the Pesticide Data Program. Participants filled out questionnaires to determine if highly consumed foods were commonly eaten by them and to assess the collection protocol for the food samples. Measureable levels of permethrin were found in 54% of the samples. Questionnaire responses indicated that the collection of the community duplicate diet was feasible for a defined population.

---

Environmental impact

Alternative residential-based sampling procedures, but less direct methods than collecting each individual duplicate diet, are needed to alleviate burden to the participant and also decrease analytical costs for large scale studies. In addition, alternative procedures are needed to provide greater opportunities to establish sources of dietary exposure within a community that could be missed in survey responses. Collection of food samples in the field to determine intake of contaminants is justified for any large study, but a reasonable technique is needed. The Community Duplicate Diet Method allows for collection of foods within a home environment and utilizes national food frequency questionnaires. This approach provides a new method for collection of dietary exposure samples that could be utilized in large-scale studies.

1.0 Introduction

Traditionally, dietary exposure assessments have focused on collecting samples from individuals to mirror the diet consumed during a specific monitoring period.^1–6 Such approaches for multimedia, residential-based exposure measurements have been justified under the rationale that they: 1) provide data comparable to other pathways (e.g., inhalation and dermal) being monitored, 2) include the contribution of contaminants added during food storage and preparation in the residence and/or from non-market sources of foods, and 3) represent a reasonable measure of daily dietary intake of the adult individual during the monitoring period. However, in larger studies, involving many participants, traditional duplicate diet collections and analyses may be more burdensome and/or costly than available resources could support. Market-basket surveys offer cost savings, but the established indirect methods of assessing dietary intakes lack the accuracy of actual dietary exposure measurements and do not capture the important contributions of storage, preparation and consumption in the residence, or handling by a child.^7–10 Therefore, alternative residential-based sampling procedures, but less direct methods than collecting each participant's entire duplicate diet, may alleviate the burden to the participant and also decrease numbers of samples required to measure dietary exposure.¹¹ This, in turn, may lower both collection and analytical costs. More efficient procedures may lead to better measurements of dietary exposures of a population where there is reason to believe that eating habits follow normal practices, such as those of a target geographical area or well-defined cultural subgroups. In addition, alternative procedures may provide greater opportunities to establish sources of dietary exposure within a community that could be missed in survey responses. Collection of food samples in the field to determine intake of contaminants is justified for any large study, but a reasonable technique is needed.

An observational field study was conducted to assess the feasibility of a community duplicate diet collection method; a dietary monitoring procedure that is population-based. The purpose was to establish an alternative procedure to duplicate diet sampling that would be more efficient for a large population, e.g., in the National Children's Study (NCS), or in a defined geographical region (e.g., a city or state), or for a defined population (e.g., women of child bearing age). The most significant goal of the community duplicate diet approach was to enable the collection of food samples in a residence so as not to lose the important component of storage, preparation, and handling in a contaminated microenvironment. Analysis of food samples would allow determination of dietary exposure of a community. Finally, the community duplicate diet method would lower the burden and cost relative to currently established protocols. The objective of the field study was to determine if the foods identified through the use of national food frequency questionnaires, then collected within a home environment, were reasonable for a well defined community. Should it be found that this approach renders the required information, a new method for the collection of dietary exposure samples could be utilized in larger scale studies.

2.0 Materials and methods

This was an observational research study, as defined in 40 CFR Part 26.402.¹² The study protocol and procedures to obtain the informed consent of the participants were reviewed and approved by an independent institutional review board (IRB). The materials complied with all applicable requirements of the Common Rule¹² regarding additional protections for women of child bearing age.

2.1 Study population definition

The study population was defined as Hispanic women of child bearing age (18–49 years old) who formed, in some sense, a “community” from which duplicate diet food samples might be collected and composited. Extensive thought into the characteristics of this community focused on a desire to have clearly defined demographics (e.g., ethnicity, locality, food preparation techniques) and a desire that foods consumed among members of this community would be more homogeneous as compared to the general population. One such community was identified in Apopka, FL, USA, in which residents partnered with private, local, and state organizations in support of the Kids in Motion program.¹³ The study population was concerned about pesticide exposure from agricultural sources; agriculture operations are widespread in the Apopka area. In addition to intrusion of pesticides from agricultural uses into the homes from drift, track-in, and on the clothes of the farmers, the warm and humid climate of central Florida increases the probability that pesticides were used in the home; however, pesticide usage was not a criterion for inclusion in the study.

2.2 Food frequency data mining

Once the community of Apopka, FL, USA, was identified, food consumption and pesticide residue databases were used to guide participant recruitment in order to provide a study community with relatively similar dietary behaviors and to increase the chances for measurable pyrethroid levels in collected samples. To approach the community with some knowledge of dietary habits, information was gathered from existing national databases and used to identify the top 10–20 most commonly prepared and eaten foods for Hispanic women of child bearing age. The highly consumed foods were used during recruitment to identify individuals that fit into the definition of “community” used for this study, i.e., similar eating habits, therefore, analogous dietary exposures. Food consumption information was obtained from published food frequency questionnaire data contained in national surveys and was summarized as percentages of each food relative to all foods consumed by Hispanic women of child bearing age. Food intake data from the What We Eat In America (WWEIA) database and the National Health and Nutrition Examination Survey (NHANES)^14,15 from years 2001–2006 were used to determine the most frequently consumed foods for Hispanic women of child bearing age (18–49 yrs. old). Characterization of these NHANES subjects was performed with respect to cultural ethnic subgroups (e.g., Mexican), age subgroups, education, marital status, poverty income ratio, etc. The WWEIA and NHANES data allowed for calculation of the percent of foods consumed, among the most frequently eaten foods. Foods encompassed all meals eaten within a day (i.e., breakfast, lunch, dinner and snacks).

In addition to providing a study community which was relatively cohesive with respect to dietary behaviors, published databases were also consulted to determine foods which had been previously found to contain measureable levels of pyrethroid pesticides. Estimated instrumental limits of detection for gas chromatography/mass spectrometry (GC-MS) techniques were used as a guide to identify foods which might contain measureable levels so that the analytical results for each of the food samples would be more meaningful. A high proportion of non-detects among the measurements provides limited, if any, information on the utility of the analytical methodologies. In a large scale study, this may be less important, as the number of samples would provide a statistical probability of measuring pesticides in the food samples, but, in this pilot study of very modest size (nine individuals), the probability of collecting food samples with non-detectible levels of pesticides was deemed too great to be left to chance. Therefore, to adequately assess the collection protocols, the food sample choices were enhanced with residue information to enable differences to be shown. To this end, the food frequency and quantity information, for the most frequently consumed foods, were augmented with pesticide residue data from the Stochastic Human Exposure and Dose Simulation Model (SHEDS),¹⁶ the Pesticide Data Program (PDP)¹⁷ and the Food and Drug Administration's Total Diet Study (TDS)¹⁸ databases to identify the collected foods most likely to contain incurred pesticide residues.

The residue information in the databases is for raw commodities (e.g., tomato, lettuce, and broccoli). This may or may not be directly applicable to the foods collected as some or all may have been prepared, i.e., cooked, mixed, or peeled, etc. However, the databases provided the most comprehensive information regarding residues in foods. Final selection of the specific numbers and types of foods to be acquired from the participants in the field study was based on a combination of frequency, quantity and residue level for each food. Table 1 shows the foods that were selected for screening the field study participants. Because pesticide residue concentrations were available only for individual raw commodities (e.g., tomatoes), only individual foods are presented. While these most frequently consumed individual foods served as the basis for participant selection, the actual foods collected from the participants were to be prepared, containing these individual components, as presented “at the table.” Anecdotally, creating a ranking based on prepared foods would be extremely difficult given the number of possible combinations and preparations of individual foods. Two different reflections of food consumption frequency are presented in Table 1: 1) the percent of each food consumed among all foods consumed by the WWEIA study sample and 2) the product of the frequency from WWEIA and the mean pesticide concentration from the residue database(s). For residue information, the focus was on pyrethroid pesticides due to their current use and abundance of information obtained as compared to organophosphates. Table 1 is ordered by decreasing “Frequency × Residue” for the top 19 individual foods. Corn tortilla was the most frequently consumed food, accounting for more than 7% of the total. However, when pesticide residue was considered in the ranking, lettuce rose to the top, followed by tomatoes and apples. Interestingly, the top 19 foods contained no meats or dairy products, regardless of how the ranking was constructed. Because of the limited number of food types with pyrethroid pesticide residue data, there were eight foods for which a “substitute” had to be assigned. In this sense, “tomatoes” were used as a substitute for “salsa” because it had no known pesticide residue data.

Table 1 Foods Consumed Frequently and With Appreciable Pesticide Levels

Food Name/Type	Percent of all Foods Consumed	Pesticide	Mean Residue Concentration (μg/g)	Frequency × Residue
a pesticide residue data of frozen sweet corn. b pesticide residue data of tomatoes. c pesticide residue data of sweet bell peppers. d pesticide residue data of wheat flour. e pesticide residue data of broccoli. f pesticide residue data of oranges.
Lettuce	6.2	Cypermethrin	0.066	27.4
Tomatoes	5.4	Cyfluthrin	0.040	14.2
Apples	2.9	Cypermethrin	0.051	9.7
Corn tortilla	7.1	Cyfluthrin	0.019^a	9.1
Salsa	2.3	Cyfluthrin	0.040^b	6.2
Hot chili peppers	1.1	Cypermethrin	0.077^c	5.8
Onions	2.4	Deltamethrin	0.032	5.2
Carrots	1.8	Cypermethrin	0.025	3.0
Oranges	1.7	Cyfluthrin	0.026	3.0
Green tomato-chile sauce	1.1	Cyfluthrin	0.040^b	2.8
Bananas	3.2	Cyfluthrin	0.013	2.7
Flour tortilla	3.3	Cyfluthrin	0.012^d	2.6
Cantaloupe	0.4	Deltamethrin	0.088	2.4
Broccoli	0.5	Azinphos methyl	0.080	2.4
Cucumbers	0.8	Cyfluthrin	0.030	2.3
Salty snacks, corn tortilla chips	1.7	Cyfluthrin	0.019^a	2.2
Strawberries	0.7	Deltamethrin	0.041	1.9
Green cabbage	0.3	Azinphos methyl	0.080^e	1.8
Avocado	1.0	Cyfluthrin	0.026^f	1.8

---

2.3 Questionnaire development

To gauge the willingness of the community members to participate in the field study, their ability to collect the appropriate samples, and the overall understanding of the protocols used in the field study, three separate questionnaires were developed to: 1) aid in the screening and recruiting of participants, 2) determine the participants' dietary behavior history and capture information on the specific foods and their quantities prepared during their participation in the study, and 3) assess the participants' burden in completing the study. First, a Screening Questionnaire sought to determine if a person ate foods typical of those identified from the food frequency data mining. Only participants who were currently eating the listed foods regularly were selected and recruited (i.e., so that participants were not asked to prepare foods that were atypical of their normal dietary intake). The field study was to occur during a limited timeframe, so qualified participants would need to consume at least four of the listed food items within a week. The determination of at least four foods was computed statistically to obtain a reasonable distribution of food and to generate an acceptable number of food samples for analysis. Basic demographic information was also requested, such as age, geographic location and race.

Secondly, a Dietary History Questionnaire was given to determine the frequency (e.g., once a week), source (e.g., home garden, grocery store, restaurant, farmer's market), and type of food (e.g., canned, fresh, frozen, etc.) prepared and consumed over the previous 12 months. A Food Diary Information Form was used to collect information regarding the food saved for the field study, such as the name of the food, date prepared, quantity prepared, how it was prepared (e.g., raw, baked, fried, boiled, etc.), and serving size. The Dietary History Questionnaire and the Food Diary Information Form were administered concurrently to determine generalized eating behaviors. This dietary information obtained from the study participants was tabulated and compared directly to the information extracted from the NHANES/WWEIA database to discern the applicability of the a priori determination of foods on the screening list.

Lastly, a short Follow-up Questionnaire was developed to determine the burden on the participant. This instrument queried the participant on the time it took to collect the food samples, as well as the time required to fill out the food diary. It also attempted to subjectively discern the participant's comfort level with the study and their willingness to complete the required tasks. The responses to these questions were ordinally-scaled. This questionnaire also asked whether any pesticides were applied during the time period of the study and what the occupations were of the people living in the home. This information was used to explain possible sources of pesticides that might be found in the foods analyzed that were higher than incurred pyrethroid residues as determined from the extant databases.

2.4 Implementation of the field study

A recruitment letter, which described the study in detail, was mailed or handed to potential participants. The question of food consumption frequency was highly important for selecting participants after they had passed the screening questions pertaining to age, gender, and availability to perform the study. If a person responded that she ate a food at least once a week, she was considered for the study. Women were chosen that indicated consumption of at least four of the targeted foods once per week. Recruitment efforts continued at the community center in Apopka until nine participants had been screened and successfully enrolled into the study. If recruiting of participants had been difficult because of these criteria, it would have indicated that the consumption database information did not pertain to this community. The women were provided with an opportunity to meet with a field team member to ask questions and then were asked to sign a consent form. The field team was accompanied by an aide who was fluent in Spanish and could explain the study and obtain consent.

2.5 Food collection

Appointment times were coordinated with the participants, so that multiple homes were visited on the same day. Each participant was reminded of the foods they previously indicated were part of their normal diet from the Screening Questionnaire and was asked to complete sample collection activities within a one week period. During this period, participants were visited initially to obtain signed consent and to receive the food collection materials. Using foods determined from the food frequency assessment, participants were asked to collect a minimum of four of these foods per household over the four meals eaten (i.e., breakfast, lunch, dinner, and snacks) over the course of the study. Each participant was permitted to choose which days to collect foods and was encouraged to spread the collection over as many days as reasonable. It was anticipated that food, as consumed, would contain both cooked and uncooked food items – typical of how they were eaten. Each participant collected a portion of each food, individually, with notation on the food diary identifying the food, how it was prepared, and the approximate amount consumed. The quantity of each food sample provided (as determined by the collection bag) was determined by the field technician by weighing. A follow-up call was placed to the participants during the study to ensure they were collecting the foods correctly and to answer any questions they may have.

All samples were collected in high density polyethylene (phthalate-free) zip-seal bags and were stored in the participant's freezer until collected by the field staff. On the last day of the study, the field team visited the home a final time to collect all food samples and administer a follow-up questionnaire to assess their level of burden for completing the study. After collection of the food samples by the participant and transfer to the field staff, each food sample was labeled with a unique sample identification code without personal identifiers. This information was logged into a spreadsheet. Other information, such as date of collection and food descriptions, was also entered into the spreadsheet which became the samples' chain of custody record. Each sample was weighed and recorded onto this custody record upon receipt of the samples at the laboratory.

As compensation for collecting the foods over the course of four meals, each participant who completed all aspects of the monitoring received a gift card worth $55 at a local grocery store of their choice. Prorated amounts ($5 for each food sample collected, $10 for the Food Diary Information Form and Dietary History Questionnaire, and $5 for the Follow-up Questionnaire) were provided to participants who voluntarily withdrew from the study or did not complete the questionnaires.

2.6 Food analysis

To be fully confident that the community duplicate diet method can provide adequate samples to estimate dietary exposure, the foods collected were analyzed for the pyrethroid and organophosphate pesticides and phthalates listed in Table 2. These analytes were chosen because of their possible current use, past usage, and interest in the exposure science field. Also, all of the analytes were obtained from the same extraction procedure and all of the pesticides could be analyzed using a single method.

Table 2 Pesticides for Food Analysis

Pyrethroids

Allethrin (cis/trans)

Bifenthrin

Cyfluthrin (4 isomers)

Cypermethrin (4 isomers)

Deltamethrin

Esfenvalerate

Permethrin (cis/trans)

Piperonyl butoxide

Resmethrin

Sumithrin

Prallethrin

Organophosphates

Chlorpyrifos

Chlorpyrifos methyl

Diazinon

Phosmet

Azinphos methyl

Fonofos

Phthalate esters

Butylbenzyl phthalate

Di-n-butyl phthalate

Diethyl phthalate

Di(2-ethylhexyl) phthalate

Diisononyl phthalate

Diisodecyl phthalate

---

Food samples were analyzed using methods briefly described here with a more detailed description included in the Electronic Supplementary Information. The entire food sample was homogenized with dry ice. A 10 g aliquot was weighed out and fortified with surrogate recovery standards (SRSs). The food was extracted, centrifuged, dried, and concentrated to exactly 10 mL. A 1 mL aliquot was removed for phthalate ester analysis with no further cleanup, fortified with internal standard and analyzed using GC-MS (Agilent Technologies 6890 GC/5973). The remaining 9 mL extract was concentrated and cleaned up using solid phase extraction (SPE) cartridges. The final extract was concentrated to 0.2 mL, fortified with the internal standard, and analyzed using GC-MS in the selected ion monitoring (SIM) mode. Quality Assurance/Quality Control (QA/QC) samples including solvent method blanks, matrix duplicates, and matrix fortified samples were processed concurrently with each batch of five field samples. The GC-MS results were transferred electronically into Microsoft Excel spreadsheets for further data reduction. The QA/QC results were within the acceptance criteria and are included in the supplemental information and are summarized in Table S1.

3.0 Results

3.1 Participants

Nine participants met the inclusion criteria and were successfully recruited out of 13 that were screened. Two women were disqualified because one was above the age limit of 49 and one lived in the same home as another woman that was also screened for the study. Two participants were not selected because their food consumption frequencies were somewhat lower than those of other qualifying participants. Each participant collected eight individual, prepared foods (except for one that collected seven and one that collected four) for a total of 67 samples. All of the participants completed all portions of the study.

3.2 Questionnaire responses

3.2.1 Screening. Responses to the screening questionnaire were helpful in determining the potential problems with the protocols and set up of the field study. Out of 18 highly consumed foods identified (salty snack, corn tortilla chips was not included in the screening), all of them were part of most of the participants' normal diets. For the nine participants included in the study, eight of the foods were consumed at least once a week. This allowed for a variety of options of foods for the participant to collect which may have increased the probability of success of the study.

3.2.2 Dietary history and food diary information questionnaire. Information collected on the Dietary History Questionnaire was used to select respondents who were most likely to consume the frequently eaten foods on the list. Table 3 summarizes the responses of the nine participants. Responses of 1, 2, or 3 corresponding to once per week, twice per week, and three times per week, respectively, were the highest priority as this indicated that the food was indeed highly consumed and the participant would likely be able to collect the item within a short period of time. Responses of 4, 5, and 6 corresponding to once per month, twice per month, and year-round, respectively, indicated that the food was regularly consumed, but timing for collection may have been problematic. As evidenced by the vast majority of 1s, 2s, and 3s, the list of foods determined from the extant databases was more than adequate for screening this population.

Table 3 Summary of Responses to Dietary History Questionnaire^a,b

Participant ID#	Lettuce	Tomatoes	Hot Chili Peppers	Onions	Carrots	Broccoli	Cucumbers	Cabbage, green	Apples
a 1 = once per week, 2 = twice per week, 3 = 3 or more per week, 4 = once per month, 5 = twice per month, 6 = year round/all the time. b * = passes for study selection.
1	3*	3*	1*	6	2*	1*	3*	4	1*
2	3*	3*	2*	6	3*	3*	3*	1*	3*
3	3*	6		6	6	3*	3*	1*	1*
4	1*	3*	6	6	3*	4	4	4	1*
5	3*	3*		3*	3*	1*	2*	3*	3*
6	1*	6	6	2*	5	5	4		2*
7	3*	3*	4	3*	3*	3*	3*	3*	3*
8	2*	2*		3*	2*	1*	2*	1*	2*
9	3*	3*		3*	3*	4			3*

---

Participant ID#	Oranges	Bananas	Cantaloupe	Strawberries	Avocado	Corn Tortilla	Flour Tortilla	Salsa	Green Tomato-chile sauce
1	2*	6	3*	3*	2*	6	1*	6	2*
2	3*	1*	2*	2*	1*	3*	1*	3*	3*
3	1*	2*	1*	1*	2*			3*	3*
4	1*	1*	1*	1*	4	6	6	6	6
5	2*	2*	4	4	4	1*		4
6	3*	2*	2*	2*	3*	6	5	6	6
7	1*	1*	1*	1*	3*	3*	4	1*	1*
8	2*	2*	2*	1*	1*
9	4	3*	4	3*	3*	3*		3*

---

The diary information was used to identify foods as collected, amounts consumed, and preparation techniques. For mixtures, proportions of each targeted food were determined visually to aid in the analysis. Each sample was received in the analytical laboratory with no adverse incidents. One collected sample contained none of the listed foods.

3.2.3 Follow-up questionnaire. A summary of the participant responses to the follow-up questionnaire is shown in Table 4. The responses indicated that the participants felt the conduct of this study was sufficiently easy to warrant the procedures as acceptable with the notable exception of the use of the food diary. One participant indicated that it took 80 min to complete the food sampling and questionnaire. It was unclear if the participant interpreted this question as length of time to collect one food or all of the foods. However, this length of time was not considered too long. The initial visits with the field team took 10–40 min for each participant. Only the participant with the 40 min visit indicated that this was too long. As this was the first participant, the familiarity with the procedures may have caused an increase in administration time. Everyone indicated that they would participate in a study like this in the future and all agreed that the incentives were fair. No one indicated using any pesticides during the time of the study. The final visits to the homes took 15–30 min and everyone felt that it was an appropriate amount of time.

Table 4 Summary of Responses to Follow-up Questionnaire

QUESTION TEXT	Response (Number of Participants)
a Scale: 1–5; 1 = easy, 5 = very difficult. b range of time designated from the respondents.
How difficult was it to prepare the foods for the study?	1^a (7)	2 (1)	3 (1)
How difficult was it to fill the jars with the food?	1 (7)	2 (2)
How difficult was it to store the food that you collected?	1 (8)	2 (1)
How difficult was it to fill out the food diary?	1 (2)	2 (4)	3 (1)	4 (2)
How much total time did it take you to collect all the food samples and fill out the food diary? (minutes)	5–80^b
Was this too long?	No (9)
About right for you?	No (1)	Yes (8)
How much time did it take for the first visit from the field team (dropping off collection materials and going over the consent form)? (minutes)	10–40^b
Did you have to make a special trip to the grocery store to purchase these foods for this study?	No (8)	Yes (1)
Would you participate in a study like this in the future?	Yes (9)
Do you think the gift card amount was fair?	Yes (9)
Did you use any pesticides during the past week?	No (9)
How much time did it take for the final visit to pick up the food samples, food diary and complete this questionnaire? (minutes)	15–30^b
Was this too long?	No (9)
About right for you?	Yes (9)

---

3.3 Analysis of food

Chemical analysis was performed on the 67 individual food samples. All of the food samples collected had at least a component of the foods on the initial list, except for one. The results, shown in Table 5, provide some simple descriptive statistics including the 50th, 75th, 95th percentiles, and maximum for the concentrations of the pyrethroid and organophosphate (OP) pesticides and phthalates. If an analyte (from Table 2) is not included within Table 5, then all results were below method detection limits (MDL). Percentiles are reported as missing (--) in Table 5 (i.e., most of the OPs, cyfluthrin, cypermethrin, deltamethrin, resmethrin, prallethrin, etc.) when the small number of detectible levels did not allow for a complete distribution to be determined. Only two samples showed measurable levels of diazinon. Of the pyrethroid pesticides, permethrin, bifenthrin, cypermethrin, and esfenvalerate, along with piperonyl butoxide (a synergist), were most often detected. Chlorpyrifos, even though banned, is still being detected in food samples as demonstrated by the 16 detections out of 67 samples. Compared to the pyrethroids, the OPs showed a noticeably smaller number of samples with detectible levels. Each phthalate result was background corrected by the levels detected in blank samples analyzed simultaneously. The phthalates are, by far, the most commonly detectible contaminant in these food samples. This may be an indication of the type of packaging for these foods as purchased.

Table 5 Concentration Distributions of Pesticides and Phthalates in Individual Food Samples, ng/g

Pyrethroids	Concentration in Food sample, ng/g
	Individual Samples, n = 67
	Number of Samples > MDL	p50^c	p75^d	p95^e	max
a sum of isomers. b Detection limit. c 50th percentile. d 75th percentile. e 95th percentile.
Bifenthrin	30	0.07	0.48	9.52	135
Cyfluthrin^a	4	—	—	1.59	11.0
Cypermethrin^a	11	—	—	6.96	14.8
Deltamethrin	1	—	—	—	14.1
Esfenvalerate	10	—	—	5.81	19.1
Permethrin^a	33	0.19	0.47	2.25	34.6
Piperonyl butoxide	42	0.26	0.54	1.91	11.6
Resmethrin	1	—	—	—	1.34
Prallethrin	1	—	—	—	1.42

---

OP	Number of Samples > MDL	p50	p75	p95	Max
Chlorpyrifos	16	0.09	0.19	1.04	3.32
Chlorpyrifos methyl	3	—	—	—	0.36
Diazinon	2	—	—	—	0.20
Phosmet	3	—	—	—	1.00
Azinphos methyl	0	—	—	—	<DL^b
Fonofos	5	—	—	0.06	0.10

---

Phthalates	Number of Samples >MDL	p50	p75	p95	Max
Butylbenzyl phthalate	62	5.13	7.96	21.9	1410
Di-n-butyl phthalate	65	10.2	17.1	34.5	206
Diethyl phthalate	61	4.61	8.06	14.4	68.4
Bis(2-ethylhexyl) phthalate	67	26.5	78.5	162	270
Diisononyl phthalate	27	—	100	318	791
Diisodecyl phthalate	3	—	—	—	70.1

---

3.4 Observations from the field study

Several challenges were encountered during the conduct of the field effort which, if resolved in future studies, would improve the data collection. Participant participation was hindered by cultural and practical reasons, such as, some women weren't sure how their husbands would feel about their participation, they didn't cook that much at home, they were not comfortable with strangers coming to their homes, or their husbands cooked more than they did. The field team had to provide assistance to most of the participants to get them to completely fill out the food diaries. Specifically, the technician routinely needed to add the dates for each food type. Many of the participants were very unfamiliar with filling out charts like these diaries. Alternate diary formats and specific columns for the sample IDs and the date for each food should be considered for future field efforts. Specific forms, questionnaires, or other means of collecting these data should be field tested with group representatives of the target populations to identify such issues prior to study implementation. Some of the participants failed to collect foods from every meal type, even though this was explained at the initial visit and the participant appeared to understand and agreed to the request. Interpretation was required to define what some foods meant to individuals because of cultural and language factors. For example, some participants interpreted “salsa” as any kind of sauce, not just the tomato-based type used with tortilla chips in American-Mexican restaurants. One participant asked if Alfredo sauce would qualify, and required clarification that it needed to be a tomato-based sauce. A notable difference in the interpretation of “lunch,” “dinner,” and “breakfast” was found. For this reason, the language employed in future field efforts should be modified to “morning meal,” “afternoon meal,” and “night meal.” To insure that each food was assigned to the correct meal event, the field team verified the interpretation with each participant and indicated it on the forms. Conversations with the participants during the study suggested that literacy needed to be considered in relating to this particular community. In addition, cultural diversity was present based on Hispanic origin; some participants were of Mexican descent, while others were Columbian and Peruvian. The types of foods that were typically prepared and eaten were influenced by this diversity. Plantains and bananas were considered the same food item. An actual in-person visit in the home rather than just a telephone call at the midpoint of the collection cycle would have been preferable in promoting participant compliance with the intended sample collection. For future field efforts, it may be beneficial to make an instructional video to show how to prepare foods for collection during the study. An example of a fully completed diary with several meals might also be beneficial to help the participants understand the requirements.

4.0 Discussion and conclusion

The community duplicate diet methodology has potential for providing a dietary intake estimate for a group of well defined individuals with similar exposures. The relative ease of participant recruitment and the successful collection of the individual foods indicated that the extant consumption database provided acceptable information for this targeted community. The preparation prior to the field effort, i.e., use of the extant databases, resulted in an acceptable list of individual foods commonly consumed by the community because all screened participants ate several of the items on the list. The extant consumption database was specific for Hispanics; however, the ethnic diversity within the sampling group suggests that a less specific database would provide enough information to be useful in determining highly consumed food items for other communities which would be necessary in a larger scale study. Therefore, the community duplicate diet method should be applicable to other defined communities, as well.

The sample collection successfully obtained in this study indicated that the protocols were acceptable. The food measurements differed from the database information indicating the need to collect food samples from the residences to truly measure dietary exposure. This is especially true for banned pesticides still being measured because they are not found in current residue databases. Several OPs were detected in multiple samples that could not be predetermined through extant residue databases. The importance of knowing all of the potential pesticide exposures is relevant for determining cumulative risks. This will be even more important for communities in which multiple sources of potential contaminations exist. The community duplicate diet methodology will aid in measuring potential exposures to a variety of contaminants.

A limited number of participants were included in the field study due to financial constraints associated with pilot testing of the protocols. Even with the limited size of the study, some areas for improvement were noted, especially for the diary format. The food diary used during the field study was disliked by the participants. If participants do not like the tools, they may not be used properly. This may cause issues when trying to calculate dietary intakes with inaccurate or missing information. Terminology needs to be clear as differences existed between the various Hispanic dialects; therefore, care needs to be taken when developing written material to be sensitive to slight differences in terminology.

With respect to the analysis of the food samples, the number of samples with measureable levels of pyrethroid pesticides was far greater than those with OPs. Data presented in Table 5 show that pyrethroid pesticide concentrations are considerably higher than OP concentrations in foods collected in this study. Although data presented here provides no definitive information on the source of the pesticides found, it may suggest a trend toward increased agricultural usage of pyrethroids with respect to OPs. Furthermore, bifenthrin, permethrin, and piperonyl butoxide (a synergist found in pyrethroid mixtures) were detected in a large number of the samples (30 to 42 out of 67), indicating a possible increase in use in these pyrethroids vs. a more commonly used one, such as deltamethrin. Unfortunately, the metabolites of one of these alternative pyrethroids, bifenthrin, are not routinely measured during biomonitoring;¹⁹ therefore, a significant source of pyrethroid exposure may be missed if biomarkers alone were used to determine exposure. Exposure assessments need to consider the current usage of pesticides to fully determine the appropriate analyses. Phthalates were also measured in the food samples at relatively high levels, even with background subtraction. Phthalate exposure is a growing concern within the exposure science field and will need to be evaluated more thoroughly. The variety of contaminants able to be measured using the community duplicate diet methodology is apparent.

It was the intent of this study to demonstrate that the community duplicate diet method of collecting food samples in representative homes was feasible. With the limitations of participant size, difficulties in use of the food diary, and some language barriers, this method provided samples that suggested the need for direct food measurements. With the aid of extant databases, the sampling scheme was enhanced and the field work benefitted from the knowledge of the technicians prior to any participant contact. The community duplicate diet method has the potential of providing exposure science with a new tool for determining dietary exposure to contaminants for a large population.

Disclaimer

The United States Environmental Protection Agency, through its Office of Research and Development, funded and managed the research described here under contract EP-D-04-068 to Battelle, subcontracted to RTI International. It has been subjected to Agency review and approved for publication. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.

References

1. K. W. Thomas, L. S. Sheldon, E. D. Pellizzari, R. W. Handy, J. M. Roberds and M. R. Berry, Testing duplicate diet sample collection methods for measuring personal dietary exposures to chemical contaminants, J. Exposure Anal. Environ. Epidemiol., 1997, 7, 17–36.
2. L. J. Melnyk, M. R. Berry and L. S. Sheldon, Dietary Exposure from Pesticide Application on Farms in the Agricultural Health Pilot Study, J. Exposure Anal. Environ. Epidemiol., 1997, 7(1), 61–80.
3. K. W. Thomas, E. D. Pellizzari and M. R. Berry, Population-based dietary intakes and tap water concentrations for selected elements in the EPA region V National Human Exposure Assessment Survey (NHEXAS), J. Exposure Anal. Environ. Epidemiol., 1999, 9, 402–13.
4. N. K. Wilson, J. C. Chuang, C. Lyu, R. Menton and M. K. Morgan, Aggregate exposures of nine preschool children to persistent organic pollutants at day care and at home, J. Exposure Anal. Environ. Epidemiol., 2003, 13, 187–202.
5. C. Lu, D. B. Barr, M. Pearson, S. Bartell and R. Bravo, A longitudinal approach to assessing urban and suburban children's exposure to pyrethroid pesticides, Environ. Health Perspect., 2006, 114, 1419–1423.
6. A. Bradman, D. Whitaker, L. Quiros, R. Castorina, B. C. Henn, M. Nishioka, J. Morgan, D. B. Barr, M. Harnly, J. A. Brisbin, L. S. Sheldon, T. E. McKone and B. Eskenazi, Pesticides and their metabolites in the homes and urine of farmworker children living in the Salinas Valley, CA, J. Exposure Sci. Environ. Epidemiol., 2006, 17, 331–349.
7. G. Akland, E. D. Pellizzari, Y. Hu, J. M. Roberds, C. A. Rohrer, J. O. Leckie and M. R. Berry, Factors Influencing Total Dietary Exposures of Young Children, J. Exposure Anal. Environ. Epidemiol., 2000, 10, 710–722.
8. C. A. Rohrer, T. E. Hieber, L. J. Melnyk and M. R. Berry, Pesticide Transfer from Household Surfaces to Foods, J. Exposure Anal. Environ. Epidemiol., 2003, 13, 454–464.
9. A. P. Vonderheide, C. E. Bernard, T. E. Hieber, P. E. Kauffman, J. N. Morgan and L. J. Melnyk, Surface-to-food pesticide transfer as a function of moisture and fat content, J. Exposure Sci. Environ. Epidemiol., 2008, 19, 97–106.
10. L. J. Melnyk, T. E. Hieber, T. Turbeville, A. P. Vonderheide and J. N. Morgan, Influences on Transfer of Selected Synthetic Pyrethroids from Treated Formica® to Foods, J. Exposure Sci. Environ. Epidemiol., 2009, 21, 186–196.
11. K. C. Jordan, L. L. Knuth, L. J. Moyer-Mileur, R. R. Larson, B. E. Sherwood, S. McNutt, J. J. Quackenboss, S. M. Viet and L. J. Melnyk, Feasibility of Community Food Item Collection for the National Children's Study, Sci. Total Environ., 2010, 409, 307–313.
12. CFR (Code of Federal Regulations) 2009. 40 CFR Chapter I Environmental Protection Agency Part 26 Protection of Human Subjects. US code of Federal Regulations. Available: http://www.access.gpo.gov/nara/cfr/waisidx_09/40cfr26_09.html, [last accessed 6/25/10].
13. Kids in Motion Program, http://www.apopkafamilylearningcenter.com [last accessed 12/11/08].
14. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics, Hyattsville, MD. National Health and Nutrition Examination Survey (NHANES). 2007. Available from: http://www.cdc.gov/nchs/nhanes.htm [last accessed 12/11/08].
15. U.S. Department of Agriculture, Agricultural Research Service, Beltsville Human Nutrition Research Center, Food Surveys Research Group (Beltsville, MD) and U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics(Hyattsville, MD). What We Eat in America, NHANES 2003–2004 Data: Dietary Interview - Total Nutrients Intakes – First Day (DR1TOT_C). (2006, September). Available from: http://www.cdc.gov/nchs/about/major/nhanes/nhanes2003-2004/dr1tot_c.xpt [last accessed 12/11/08].
16. EPA/ORD/NERL, Details of SHEDS-Multimedia Version 3: Model to Estimate Aggregate and Cumulative Exposures to Chemicals. Available at http://www.epa.gov/heasd/products/sheds_multimedia/files/SHEDS%20Multimedia%20QAPP-Rev1-Final.pdf, [last accessed July 23, 2009].
17. U.S. Department of Agriculture, Agricultural Marketing Service, Agricultural Marketing Service, Science & Technology, Monitoring Programs Office, Manassas, VA. Pesticide Data Program. 2006. Available from: http://www.ams.usda.gov/AMSv1.0/ams.fetchTemplateData.do?template=TemplateC&navID=PesticideDataProgram&rightNav1=PesticideDataProgram&topNav=&leftNav=ScienceandLaboratories&page=PesticideDataProgram&resultType= [last accessed 12/11/08].
18. U.S. Department of Health and Human Services, U.S. Food and Drug Administration, Center for Food Safety and Applied Nutrition, Silver Spring, MD. Total Diet Study. 2007. Available from: http://www.fda.gov/Food/FoodSafety/FoodContaminantsAdulteration/TotalDietStudy/ucm184232.htm#fca [last accessed 12/11/08].
19. P. A. Smith, M. J. Thompson and J. W. Edwards, Estimating occupational exposure to the pyrethroid termiticide bifenthrin by measuring metabolites in urine, J. Chromatogr., B: Anal. Technol. Biomed. Life Sci., 2002, 778(1–2), 113–20.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c1em10611b

---