Exposure Assessment of Young Adults to Pesticides That Have Effects on the Thyroid—A Contribution to “One Health”

Abstract

The objective of this research was to evaluate the cumulative exposure of the population aged 10–24 years to pesticides that have a chronic effect on the thyroid. A consumption study covering fresh fruits and vegetables was collected from 377 respondents. In parallel, 2369 fruit and vegetable samples were chemically analyzed for pesticide residues. As a result, cumulative exposure was calculated for four different scenarios (as is, maximum residue level, 70% of MRL and below detection limit) using two recall methods. The results show that, depending on the scenario, cyprodinil from green lettuce and strawberries, fluopyram blueberries and strawberries, and fluxapyroxad, detected in grapes, contribute most to exposure. More stringent scenarios, with limits at 70% of the MRLs (0.7 MRL) and below the detection limit (“zero residue” approach), show that the estimated total margin of exposure increases by up to 50% in the “0.7 MRL” model, while levels almost triple in the “zero residue” model. The optimization of pesticide use has a beneficial effect on human health and the environment, contributing to the “One Health” approach.

1. Introduction

Current agricultural practices rely heavily on various chemical plant protection products (PPP) with their two main side effects on public health and the environment. Active substances from standalone chemicals or mixtures may have severe impacts on human health such as carcinogenic, reprotoxic, immunosuppressive and endocrine-disrupting effects [1,2]. To mitigate the potential adverse health effects of pesticide residues, regulatory agencies set critical limits, also known as maximum residue levels (MRLs) applicable to various agricultural commodities [3]. In the European Union (EU), Regulation 1107/2009/EC defines cut-off criteria for potential effects of active substances (AS) on health (mutagens, carcinogens, reproductive and endocrine disruptors) and the environment (persistent organic pollutants and persistent, bioaccumulative and toxic substances) [4]. In parallel, there are several initiatives, such as the European Green Deal, which supports sustainable food systems as part of its farm-to-fork strategy [5]. This strategy supports the reduction in dependence on pesticides and excessive fertilization and promotes organic farming [6]. There is limited organic production in Serbia.

The World Health Organization has introduced a “One Health” approach as an “integrated, unifying approach that aims to sustainably balance and optimize the health of people, animals and ecosystems”, including humans, animals, plants and the entire environment and its intertwining [7]. This concept has the potential to promote practices that optimize plant health protection [8], paving the road to sustainable food production in line with UN Sustainable development goals and ensuring food security [9,10].

Some retailers in the EU have the strictest demands in terms of MRL (between 33 and 70% of the prescribed EU limits) for the sale of their commodities in Central Europe (Germany, the Netherlands, Austria, Belgium), as well as in the UK and Switzerland [11] under the condition of certified good agricultural practices. In recent years, another concept related to PPP use has emerged. It is known as the “zero residue” concept, where residual pesticide levels are undetectable with analytical tools [12]. In contrast to the legal MRLs, the detection limit for most PPPs is usually ≤0.01 mg/kg [13], or in the case of higher analytical sensitivity, the levels can be even lower or more stringent [14]. To achieve this goal, food standards have been developed to promote the limited use of pesticides, such as the Pesticide Free Certification standard [13] and Zero/Controlled Pesticide Residue [12], which combines good agricultural practice auditing and laboratory testing of agricultural products. This approach is considered to add value to the primary sector, as the absence/reduced presence of PPP residues in food products should improve the environmental footprint and minimize health impacts [14].

The awareness of the health effects of pesticides in the EU has led the European Food Safety Authority (EFSA) to develop methods for calculating and analyzing the cumulative effects of multiple chemicals derived from pesticide residues [15]. Based on recent scientific evidence that pesticides can affect the thyroid gland through hypothyroidism (‘health effect #1’) and through hypertrophy, hyperplasia, and neoplasia of parafollicular cells (C cells) (‘health effect #2’), the EFSA has identified specific assessment groups of pesticides and proposed a methodology to calculate their health effects [16]. To validate its methodology, the EFSA conducted a cumulative dietary risk characterization of these pesticides associated with the thyroid [17].

Young adulthood is considered a special developmental period when the health and well-being of the young population is a high priority [18]. In accordance with the recommendation of the World Health Organization (WHO), adolescents are considered to be between 10 and 19 years old, while young adults are between 20 and 24 years old [19]. This target group falls into Generation ‘Z’ (Gen Z, born 1995–2010), which is represented by one third of the world’s population [20] and is the leading consumer group in food consumption worldwide [21]. Because hypothyroidism can be triggered by repeated exposure and should be part of a long-term cumulative assessment [16], the authors of this study focused on the young population. Latest EFSA study targeted three age groups—adults, toddlers and other children—while the young population was not analyzed per se [17]. At the same time, no cumulative exposure assessment related to thyroid has been conducted for this age group, which was identified as a research gap by the authors of this study.

The main objective of this study was to evaluate the cumulative exposure of the young population in Serbia to pesticides that have chronic effects on the thyroid, especially hypothyroidism, through the consumption of fresh fruits and vegetables and foods prepared from them. To achieve this objective, different scenarios were used to assess exposure. The working hypothesis is that new agricultural concept of “zero residue” may have a significant impact on reducing cumulative exposure and a potential to contribute to “One Health”.

2. Materials and Methods

2.1. Field Survey

For the purpose of this study, a consumption survey was conducted in the second half of 2022 and the first half of 2023. It was conducted in cities and villages and included both urban and rural population on the entire territory of the Republic of Serbia. The questionnaire for the consumption survey was developed in accordance with the recommendation of the European Food Safety Authority (EFSA) to collect data on the quantity and frequency of consumption of fruits and vegetables [22]. It consisted of two main sections: (i) basic demographic characteristics for this type of study—sex of respondents, their age, height (cm) and body weight (kg) [23]; (ii) recall method for fruits and vegetables consumed. In this study, the authors combined two recall methods—food frequency questionnaire (FFQ) and 24 h food recall. Combining two recall methods is preferable because it allows modeling of different consumption scenarios [24].

The survey was conducted face to face and targeted two groups of respondents: those over 18 years of age who could self-answer and parents of adolescents between 10 and 18 years of age. In both cases, verbal consent was obtained from respondents before the interview began, and they were fully informed about the purpose of the study and our professional commitment to uphold key ethical standards—anonymity, confidentiality and privacy [24]. Respondents were randomly selected when purchasing fresh produce at retail stores (mainly in urban areas) or by distributing the questionnaire through family and friend networks (in rural areas).

Body mass index (BMI) was calculated using data for respondents’ weight (kg) and height (m) (BMI = kg/m²) with the following categorization: (i) below 18.5—underweight; (ii) 18.5—24.9—healthy weight; (iii) 25.0 and above—overweight [25]. The reporting period for the FFQ recall methods was divided into several categories [26]: ‘on an annual basis’ (for food commodities that are available and/or consumed throughout the year) or ‘on a seasonal basis’, where the respondent could indicate the following frequencies: ‘daily’, ‘weekly’ (2–4 times per week), ‘monthly’ (up to 3 times per month) and ‘rarely’ (a few times per year or never). The FFQ part of the questionnaire included 51 different fruits and vegetables belonging to another 88 food categories: (i) bakery products, (ii) soups and stews, (iii) prepared meals, (iv) salads, (v) fruits. The respondents were asked to recall their dietary habits “in the last 12 months”.

As for dishes, as the most complex part of the questionnaire, consisting of a variety of ingredients, they were grouped according to culinary techniques and content. The size of a meal (dishes, pastries, soups and potages) and its content were also calculated based on the recommended diet of students, which provides guidance for over 350 dishes [27]. For fresh fruit consumption, one serving was set at 150 g [28].

The 24 h method allowed respondents to recall all types of meals, snacks, beverages and sweets they consumed ‘yesterday’ from morning until bedtime, including breakfast, lunch, dinner and in-between meals (various snacks). Respondents were given the opportunity to indicate the type of food (with additional description if needed) and the amount consumed as accurately as possible (in grams, pieces, cups, plates/portions, etc.). Attached to the questionnaire was a set of photographs with predefined portions, which were provided to respondents as a visual aid for both the FFQ and the 24 h recall.

2.2. Laboratory Food Analyses

Regarding the number of samples, from 2021 to 2023, more than 2369 fruit and vegetable samples were analyzed by two laboratories accredited according to ISO 17025 [29], including 1415 fruit samples, 834 vegetable samples and 120 miscellaneous product samples (herbs, spices, sprouts, etc.). The method used was according to EN 15662:2018 [30]. The extraction and clean-up of the samples were carried out using the QuEChERS procedure (quick, easy, cheap, effective, rugged and safe). The tested samples were simultaneously analyzed using GC-MS/MS gas chromatography (7890A Agilent Technologies GC System) coupled with a triple-quadrupole tandem mass spectrometer (7000A Agilent Technologies GC/MS Triple Quad), and LC-MS/MS utilizing an Agilent 1200 system (Agilent Technologies, Santa Clara, CA, USA) with a triple quadrupole mass spectrometer Agilent 6410B QQQ (Agilent Technologies, Santa Clara, CA, USA) equipped with an electrospray ionization (ESI) source, and the ESI source operated in positive ionization mode. The samples were a combination of official surveillance of the goods or part of the external control of the company. Samples with values below the detection limit (LOD) were counted as LOD/2 [26].

In total, 29 active substances were detected in 35 different fruits and vegetables, including 19 fruits—peaches, plums, blueberries, pears, blackberries, raspberries, strawberries, apples, cherries, lemons, grapes, bananas, oranges, tangerines, grapefruits, avocados, pomegranates, melons and apricots; and 16 vegetables—broccoli, carrots, green salad, bell pepper, Brussels sprouts, potato, ginger, tomato, celeriac, cherry tomato, cucumber, beetroot, parsley, chili pepper, onion and spinach.

2.3. Processing Factors

Some processing and culinary methods used to prepare meals affect pesticide levels and consequently consumer exposure. To account for these effects of the two recall methods, it was important to determine a processing factor (PF), which is a residue ratio between processed foods and raw commodities. PFs were collected from an EU database [31]. If PF was not available, values were extrapolated considering the active substance (or the closest member of the same group). Because the intake of some types of food was reported as a number (pieces, cups, plates, etc.), the proportion of food consumed was taken from the studies of Ruiz-Torralba et al. [32] and Zincke et al. [31].

2.4. Calculations of Specific Exposure Assessment Indicators

Exposure is usually calculated as a combination of food consumption data, data on the concentration of a particular chemical in a food and body weight of the population [33]. Exposure to “j” active substances was calculated by combining data on the consumption of “n” fruits and vegetables, the concentration of the active substance in the selected “n” fruits and vegetables, and the body weight of the respondents, as in Equation (1).

$${\mathrm{E}\mathrm{D}\mathrm{I}}_{\mathrm{j}}=\frac{\sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{F}\mathrm{V}}_{\mathrm{i}}\ast {\mathrm{C}}_{\mathrm{i}}\ast {\mathrm{P}\mathrm{F}}_{\mathrm{i}}}{\mathrm{b}\mathrm{w}}$$

(1)

EDI_j is the estimated daily intake of a given “j” active substance [μg/kg bw/day]; FV_i is the daily amount of “i” fruit/vegetable consumed [kg]; C_i is the active substance concentration in “i” fruit/vegetable [mg/kg]. If necessary, PF_i as processing factor has been considered for a specific food/active substance combination (made of “i” fruit/vegetables).

MOE_j is the margin of exposure to “j” active substance calculated by dividing the toxicological reference point (in our case NOAEL) of the “j” active substance and calculated EDI (Equation (2)). The NOAEL_j values (no observed adverse effect levels expressed in mg/kg bw per day) for “j” active substance were taken from [16]. Final risk characterization of the cumulative exposure was performed by calculating the total margin of exposure (MOET).

$${\mathrm{M}\mathrm{O}\mathrm{E}}_{\mathrm{j}}=\frac{\mathrm{N}\mathrm{O}\mathrm{A}\mathrm{E}\mathrm{L}\mathrm{j}}{{\mathrm{E}\mathrm{D}\mathrm{I}}_{\mathrm{j}}}$$

(2)

The total margin of exposure (MOET) was calculated using the direct method as the sum of all reciprocal individual MOEs contributing to the overall cumulative exposure, as in Equation (3) [17]:

$$\frac{1}{\mathrm{M}\mathrm{O}\mathrm{E}\mathrm{T}}=\frac{1}{{\mathrm{M}\mathrm{O}\mathrm{E}}_1}+\frac{1}{{\mathrm{M}\mathrm{O}\mathrm{E}}_2}+\cdots +\frac{1}{{\mathrm{M}\mathrm{O}\mathrm{E}}_{\mathrm{n}}}$$

(3)

Comparison with acceptable daily intakes (ADIs) was performed for obtained EDIs for each individual active substance as additional risk characterization.

In exposure assessments, uncertainties that could affect the results are as follows: (i) quality of water used for processing fruits/vegetables was not included; (ii) some of the respondents could have reported decreased or increased consumption amounts and frequencies due to different psychological reasons; (iii) contribution of metabolites and degradation product effect were not considered. Therefore, these constraints could result in overestimation/underestimation of some values [17].

2.5. Employed Scenarios

For the purposes of this study and interpretation of the results, four scenarios were used based on the fruit and vegetable samples analyzed and the active substances found in the commodities, namely the following:

-

Scenario 1—“as is”, using all fruit and vegetable samples, including those with levels above the MRL, as well as detection of unexpected pesticides (i.e., traces of pesticides not commonly used on that type of fruit and/or vegetable and pesticides not approved for general use or whose approved use has expired with 1 January 2020 as the start date).

-

Scenario 2—“1.0 MRL”, using only samples that were below the MRL and excluding samples with unexpected pesticide detections. This scenario was created assuming that primary producers have effective good agricultural practice (GAP) associated with an effective monitoring program and no samples are above the MRL.

-

Scenario 3—“0.7 MRL”, using samples that were below 70% of MRL, as required by some of the largest retailers in the EU [11].

-

Scenario 4—“zero”, defining the value of LOD as an ideal best-case scenario and a promising concept [14].

The fifth scenario—“Max”—was calculated as a worst-case scenario where the value of the samples is exactly at the MRL to determine the limit of MOET.

2.6. Statistical Analysis

The chi-square test for association was performed to determine if there were potential relationships between the consumption patterns (frequencies) and the demographic characteristics of the sample (sex and age of the population). The level of statistical significance was set at 0.05.

EDI_j for each of the active substance was calculated by deterministic approach using single values, or point estimates, as inputs, resulting in the output as a point value for exposure. Deterministic approaches produce a point estimate of exposure that falls somewhere within the full distribution of possible exposures. Depending on the input values, this approach can provide assessors with meaningful estimates of central tendency or high-end exposures within a defined population [34,35].

3. Results and Discussion

3.1. Characteristics of the Field Survey and Consumption Patterns

Characteristics of the sample are presented in

Table 1. Demographic profile of young adulthood population (N = 377).

Demographic Characteristics		Total
Sex	Male	119 (31.6%)
	Female	258 (68.4%)
Age	Adolescents (10–19 years of age)	129 (34.2%)
	Young adults (20–24 years of age)	248 (65.8%)
Average body weight (kg)	Male	60.4 ± 9.9
	Female	77.2 ± 16.4
	Total	65.7 ± 14.5
Body mass index	Underweight (BMI < 18.5)	32 (8.5%)
	Healthy weight (18.5 < BMI ≤ 24.9)	295 (78.2%)
	Overweight (BMI > 25.0)	50 (13.3%)
FFQ recall model	Average daily intake of fruits (g/day)	187.5 ± 184.0
	Average daily intake of vegetables (g/day)	236.9 ± 178.2
	Average daily intake of fruits and vegetables (g/day)	424.4 ± 424.1
24 h recall model	Average daily intake of fruits (g/day)	123.2 ± 135.9
	Average daily intake of vegetables (g/day)	150.2 ± 145.4
	Average daily intake of fruits and vegetables (g/day)	273.4 ± 202.9

. As the research focused on younger respondents, the sample was predetermined in terms of age (under 25) and parents of adolescents under 18 responded on behalf of their children. Therefore, the demographic characteristics were not stratified due to these predefined criteria. Since the number of adolescents and young adults in Serbia is estimated at about 1.2 million [36], the estimated sample size should be at 384 [37,38], with a confidence level of 95% and confidence interval of 5%. Our survey gathered 377 responses. Nevertheless, the sample can be considered suitable to assess fruit and vegetable dietary habits [26]. In general, young female respondents prevailed compared to males. Young adults (mainly students) participated in over two thirds of the sample. Considering BMI, 13.3% were overweight, with no case of BMI > 30, i.e., obesity [25]. Consumption of vegetables is slightly higher than the consumption of fruits. Based on the two recall methods, the FFQ recall model resulted in higher values compared to the 24 h recall.

Table 2. Analysis of the frequency of consumption in-between sex and age groups, based on FFQ results.

Food Commodity	In-between Sexes	In between Age Groups
Apple	χ2 = 1.978, p > 0.05	χ2 = 2.876, p > 0.05
Apricot	χ2 = 5.258, p > 0.05	χ2 = 6.012, p > 0.05
Banana	χ2 = 3.317, p > 0.05	χ2 = 2.632, p > 0.05
Blackberry	χ2 = 10.880, p < 0.05 *	χ2 = 4.217, p > 0.05
Blueberry	χ2 = 2.198, p > 0.05	χ2 = 1.127, p > 0.05
Cherry	χ2 = 4.819, p > 0.05	χ2 = 2.041, p > 0.05
Grapes	χ2 = 1.894, p > 0.05	χ2 = 6.040, p > 0.05
Melon	χ2 = 6.577, p > 0.05	χ2 = 6.645, p > 0.05
Orange	χ2 = 3.451, p > 0.05	χ2 = 10.543, p < 0.05 *
Peach	χ2 = 3.370, p > 0.05	χ2 = 1.757, p > 0.05
Pear	χ2 = 1.074, p > 0.05	χ2 = 2.433, p > 0.05
Plums	χ2 = 1.341, p > 0.05	χ2 = 6.676, p > 0.05
Raspberry	χ2 = 7.447, p > 0.05	χ2 = 3.455, p > 0.05
Strawberry	χ2 = 1.962, p > 0.05	χ2 = 10.949, p < 0.05 *
Tangerine	χ2 = 3.774, p > 0.05	χ2 = 15.121, p < 0.05 *
Carrots	χ2 = 1.027, p > 0.05	χ2 = 1.835, p > 0.05
Cherry tomato	χ2 = 5.451, p > 0.05	χ2 = 4.369, p > 0.05
Cucumber	χ2 = 3.486, p > 0.05	χ2 = 2.052, p > 0.05
Green salad	χ2 = 3.432, p > 0.05	χ2 = 0.619, p > 0.05
Onion	χ2 = 12.986, p < 0.05 *	χ2 = 8.748, p < 0.05 *
Tomato	χ2 = 7.667, p > 0.05	χ2 = 1.341, p > 0.05

Results show Pearson chi-square (χ²) value and significance level value (p); items with sign (*) are significantly different at level of 5%. FFQ—food frequency questionnaire. Frequency options: ‘daily’, ‘weekly’ (2–4 times per week), ‘monthly’ (up to 3 times per month) and ‘rarely’ (a few times per year or never).

depicts if there is statistically significant association between sex and frequency of consumption and age groups and frequency of consumption, as reported in the FFQ part of the questionnaire. There are sex related differences in the consumption frequency of blackberries and onion (p < 0.05), as female respondents reported more frequent consumption of the two commodities. The consumption of oranges, strawberries, tangerines and onions varies significantly between different age groups (p < 0.05). Younger population more often consumes these fruits while older population more frequently consumes onion.

3.2. Laboratory Analysis

Table 3. Overview of active substances causing ‘health effect #1’ found in fruits and vegetables.

Active Substance	Number of Types of Fruit	Number of Types of Vegetables	Number of Samples	Samples with Values above LOD	Samples with Values above MRL
Expected active substances
Boscalid	11	9	1277	130	1
Buprofezin	3	0	164	5	4
Cyantraniliprole	0	1	60	1	0
Cyproconazole	4	1	312	8	0
Cyprodinil	9	5	1012	92	0
Dithianon	1	0	45	4	0
Fenbuconazole	2	0	123	8	0
Fluopicolide	1	4	238	16	0
Fluopyram	6	6	792	36	0
Fluxapyroxad	7	2	575	20	1
Folpet	2	0	113	3	0
Myclobutanil	4	0	308	13	0
Pendimethalin	0	2	49	5	0
Pyrimethanil	17	2	1351	151	1
Spinosad	5	2	558	36	4
Spirodiclofen	1	0	61	4	0
Spiromesifen	1	3	323	9	0
Thiabendazole	7	0	392	46	0
Thiophanate-methyl	4	1	288	15	0
2,4-D	1	0	98	1	0
Unexpected active substances
Biternatol *	0	1	33	--	--
Carbaryl *	0	1	33	--	--
Chlorpropham *	1	3	257	--	--
Lufenuron *	1	0	60	--	--
Pyridalyl *	0	1	99	--	--
Terbuthylazine *	2	0	38	--	--
Tetraconazole *	0	1	105	--	--
Thiacloprid *	1	0	62	--	--
Thiamethoxam *	2	5	395	--	--

Legend: (*)—not expected to be found (not commonly used on that type of fruit and/or vegetable; pesticides not approved for general use or whose approved use has expired). ‘Health effect #1’—affecting the thyroid gland through hypothyroidism.

depicts information on the active substances analyzed, the number of commodities (fruit and/or vegetables) analyzed for each active substance, the number of sampled commodities, the number of samples with values above limit of detection (LOD) and the number of samples with values above maximum residue level (MRL). A total of 29 active substances with an effect on ‘health effect #1’ (hypothyroidism) as outlined by Crivellente et al. [16] and Craig et al. [17] were identified in 35 different types of fruits and vegetables. Out of these 29, only 20 were included in the subsequent modelling scenarios 2–4 (“1.0 MRL”, “0.7 MRL”, “zero”) as carbaryl, chlorpropham, bitertanol, lifenuron and thiamethoxam as well as pyridinyl, terbuthylazine, thiacloprid and tetraconazole were excluded from further modelling because they were either not approved for use or were present in products not typically intended for their use. In general, fruits were commodities with more active substances detected, opposed to vegetables. Peach was the fruit with traces of 13 active substances, while 10 active substances were detected in apples and strawberries. Onion and pepper were vegetables with seven active substances detected, while green salad had traces of six active substances. MRL was exceeded in 10 samples (buprofezine and spinosad in 4 samples, with pyrimethanil, boscalid and fluxapyroxad in 1 sample). Detailed description on occurrence of specific active substances in different types of fruits and vegetables is provided in Supplementary Table S1.

As for ‘health effect #2’, five active substances were considered in further calculations: buprofezine, imidacloprid, fenbuconazole, folpet, 2,4-D [16,17].

3.3. Exposure Assessment

3.3.1. Health Effect #01

Based on the results of consumption survey and laboratory analysis of agricultural products,

Table 4. Estimated daily intake of active substances for two recall periods, scenario “as is”, and ‘health effect #1’.

Active Substance	NOAEL (mg/kg bw/day)	ADI (mg/kg bw/day)	EDI Values (μg/kg bw/day) FFQ Recall		EDI Values (μg/kg bw/day) 24 h Recall
			Mean	95th Percentile	Mean	95th Percentile
Boscalid	22	0.044	0.047	0.232	0.023	0.088
Buprofezine	0.9	0.010	0.002	0.010	0.001	0.002
Cyantraniliprole	7	0.010	0.001	0.002	0.000	0.003
Cyproconazole	24.7	0.020	0.002	0.008	0.001	0.008
Cyprodinil	3.14	0.030	0.067	0.248	0.056	0.414
Dithianon	7.9	0.010	0.000	0.002	0.000	0.000
Fenbuconazole	3	0.006	0.005	0.022	0.003	0.007
Fluopicolide	32	0.080	0.003	0.010	0.003	0.022
Fluopyram	1.2	0.012	0.007	0.029	0.003	0.017
Fluxapyroxad	2.7	0.020	0.005	0.017	0.004	0.013
Folpet	68.4	0.100	0.002	0.006	0.002	0.009
Myclobutanil	15	0.025	0.003	0.010	0.002	0.011
Pendimethalin	43	0.125	0.001	0.003	0.000	0.002
Pyrimethanil	17	0.170	0.030	0.132	0.013	0.041
Spinosad	2.7	0.024	0.007	0.031	0.005	0.000
Spirodiclofen	19.88	0.015	0.002	0.007	0.002	0.013
Spiromesifen	6.5	0.030	0.003	0.012	0.001	0.009
Thiabendazole	10	0.100	0.010	0.047	0.005	0.033
Thiophanate-methyl	8	0.020	0.005	0.024	0.002	0.000
2,4-D	75	0.050	0.003	0.010	0.001	0.000
Biternatol *	100	0.003	0.000	0.000	0.000	0.000
Carbaryl *	60.2	0.007	0.000	0.000	0.000	0.000
Chlorpropham *	5	0.007	0.004	0.013	0.003	0.014
Lufenuron *	0.4	0.015	0.000	0.000	0.000	0.000
Pyridalyl *	47.4	0.030	0.000	0.000	0.000	0.000
Terbuthylazine *	6.97	0.004	0.000	0.002	0.000	0.000
Tetraconazole *	1	0.004	0.000	0.000	0.000	0.000
Thiacloprid *	1.2	0.010	0.000	0.002	0.000	0.000
Thiamethoxam *	198.6	0.026	0.001	0.005	0.001	0.005

Legend: (*) unexpected active substances; ADI—acceptable daily intake; NOAEL—no observed adverse effect levels. ‘Health effect #1’—affecting the thyroid gland through hypothyroidism.

shows the calculated EDI for the young population in Serbia for the “as is” scenario. The highest EDI was observed for cyprodinil (0.067 μg/kg bw/day—FFQ recall; 0.056 μg/kg bw/day—24 h recall) and boscalid (0.047 μg/kg bw/day—FFQ recall; 0.023 μg/kg bw/day—24 h recall). This fact is not surprising since anilinopyrimidine (AP) fungicide cyprodinil, currently classified as potential methionine biosynthesis inhibitors, and boscalid, fungicide belonging to the class of carboxamides which inhibits enzyme succinate dehydrogenase in fungal cells [39], are widely used in Serbia for the control of several diseases, especially grey mold (Botrytis cinerea) on grapevine, raspberries and strawberries; apple scab (Venturia inaequalis) and brown rot (Monilia spp.) of stone fruits. Estimated daily intake of active substances for two recall periods for differebt scenarios and ‘health effect #1’ are displayed in Supplementary Table S2a, S2b and S2c.

All calculate EDI values (expressed in μg/kg bw/day) are lower than ADI values (expressed in mg/kg bw/day). Acceptable Daily Intake values is provided in Supplementary Table S3.

The estimates of the MOETs and their mean values, as well as their 95th and 99th percentile of the exposure assessment are shown in Figure 1, for two recall scenarios—FFQ recall and 24 h recall. The 99th percentile in 10 European populations covering toddlers, other children and adults ranged between 155 and 401. This study included traces of active substances with effects on hypothyroidism captured from the official pesticide monitoring programs of the EU member states during the period 2014–2016 [17]. MOET results from our study for both recall methods are higher, indicating lower risk for ‘health effect #1’. However, it must be noted that hypothyroidism is now considered a common diseases worldwide [40] related to the thyroid gland and low thyroid hormone level, affecting quality of life in many ways [41]. In young adults, this disease can lead to lower IQ, affect memory, attention and behavior, cause hearing loss and cardiovascular problems [42].

The comparison between the scenarios for the mean values of the FFQ recall shows that the values for “as is” and “1.0 MRL” MOET are close to each other. However, the “0.7 MRL” scenario increases the MOET by over 40%, while the “zero” scenario doubles the MOET value. The results for 24 h recall show an even more pronounced increase in the MOET value (increase in MOET by 50% for “0.7 MRL” and almost a tripling of the value for “zero”).

Calculated MOET Values for scenario “Max” are as follows: FFQ recall (mean—117.9; 95th percentile—27.2; 99th percentile—13.6); 24 h recall (mean—194.8; 95th percentile—40.2; 99th percentile—9.1). It is considered that the MOET value below 100 at the 99.9th percentile of the exposure distribution is a threshold regulatory consideration [17]. For children and toddlers, the threshold was between 100 and 199, and for adults, it was between 264 and 314. In our case, for the ‘Max’ scenario, values were below 100.

In Figure 2, a Pareto distribution illustrates the primary sources contributing to exposure. In all scenarios with active substances (“as is”, “1.0 MRL” and “0.7 MRL”) and using both recall methods, cyprodinil is the predominant contributor, accounting for 41.3–63.5% of the total depending on the recall methods and scenario. Cyprodinil is mainly associated with green salads and strawberries. Fluopyram follows with a share of 8.8%–19.6%, mainly from blueberries and strawberries. In the “zero” scenario, fluopyram and fluxapyroxad are the main active substances of concern.

3.3.2. Health Effect #02

As for ‘health effect #2’, all values were above the threshold considering five active substances. MOET values are displayed in

Table 5. Total margin of exposure (MOET) for two recall methods for ‘health effect #2’.

	FFQ Recall Method			24 h Recall Method
Scenario	Mean	95th Percentile	99th Percentile	Mean	95th Percentile	99th Percentile
“As is”	476,457.5	108,193.9	55,578.1	894,564.3	339,755.4	32,238.1
“1.0 MRL”	477,869.9	109,992.3	55,757.1	908,235.0	330,349.0	32,371.4
“0.7 MRL”	568,088.4	165,251.6	84,368.2	1,248,736.9	205,654.2	95,512.3
“Zero”	831,586.6	236,249.1	121,582.4	1,281,031.9	206,130.9	102,705.4

Legend: MRL—maximum residue level; LOD—limit of detection; “as is”—all samples; “1.0 MRL”—samples below MRL; “0.7 MRL”—only samples below 0.7 MRL; “zero”—below LOD.

.

As it can be seen, the values confirm limited health effects of pesticides on hypertrophy, hyperplasia and neoplasia of parafollicular cells (C cells). This concurs with higher values of MOET for the 99th percentile that covered toddlers, other children and adults in 10 EU countries ranging between 1740 and 5010 [17]. This EFSA study covered 14 active substances with proven ‘health effect #2’. C-cell hyperplasia is a precursor for C-cell cancer of the thyroid gland which can remain unchanged for years as an indolent tumor, but can also become an aggressive variant with a high mortality rate [43].

As mentioned above, MOET values of 100 or lower at the 99.9th percentile are considered as thresholds [17]. Opposed to ‘heath effect #01’ results for the worst-case scenario when values at 99th percentile were below 100, for ‘health effect #2’, the ‘Max’ scenario at 99th percentile was 828.5 for FFQ recall method and 611.5 for 24 h recall.

4. Conclusions

Results for ‘health effect #1’ show that even a decrease in MRL for 30% (0.7 MRL) increases the MOET for 40% to 50% depending on the recall method while MOET values are doubled or almost tripled when the “zero” scenario is employed for the two recall methods. This confirms the working hypothesis that the “zero residue” concept has a significant potential in decreasing cumulative exposure effects and has a great sustainability potential. A certain limitation of the study is that organic food products were not sampled. As for ‘health effect #2’, the level of MOET for all four scenarios are considered as very low risk.

Future research should target other pesticides with chronic effects on other human organs and include other age groups. Other dimension of research could correlate overall sustainability and health effects of different agricultural practices.