Exposure Levels of Pyrethroids, Chlorpyrifos and Glyphosate in EU—An Overview of Human Biomonitoring Studies Published since 2000

Currently used pesticides are rapidly metabolised and excreted, primarily in urine, and urinary concentrations of pesticides/metabolites are therefore useful biomarkers for the integrated exposure from all sources. Pyrethroid insecticides, the organophosphate insecticide chlorpyrifos, and the herbicide glyphosate, were among the prioritised substances in the HBM4EU project and comparable human biomonitoring (HBM)-data were obtained from the HBM4EU Aligned Studies. The aim of this review was to supplement these data by presenting additional HBM studies of the priority pesticides across the HBM4EU partner countries published since 2000. We identified relevant studies (44 for pyrethroids, 23 for chlorpyrifos, 24 for glyphosate) by literature search using PubMed and Web of Science. Most studies were from the Western and Southern part of the EU and data were lacking from more than half of the HBM4EU-partner countries. Many studies were regional with relatively small sample size and few studies address residential and occupational exposure. Variation in urine sampling, analytical methods, and reporting of the HBM-data hampered the comparability of the results across studies. Despite these shortcomings, a widespread exposure to these substances in the general EU population with marked geographical differences was indicated. The findings emphasise the need for harmonisation of methods and reporting in future studies as initiated during HBM4EU.


Introduction
The general population is exposed to pesticides from residues in food items [1,2], but drifting from agricultural areas, and indoor use of biocides are other important sources of exposure [3][4][5][6]. In addition, some individuals are occupationally exposed. Currently used pesticides are metabolised and excreted, primarily in the urine, within a few days [7][8][9][10][11]. Urinary concentrations of pesticides or their metabolites are therefore useful as biomarkers for the integrated pesticide exposure from all sources. Within the European Human Biomonitoring Initiative (HBM4EU) the following pesticides were prioritised: pyrethroids (whole group), chlorpyrifos and dimethoate (organophosphate insecticides), fipronil (phenyl pyrazole insecticide), and glyphosate (organophosphate herbicide) in combination with polyethoxylated tallow amine (POEA) used as additive in glyphosate herbicide formulations. The prioritisation strategy has been described previously [12]. The primary aim was to get better information on the human internal exposure of these substances in the EU population(s), including potential differences between countries and population groups and time-trends. Another goal was to identify main sources and pathways of exposure across the member states.
Since no suitable urinary biomarkers were available for dimethoate, fipronil or POEA there was no existing European human biomonitoring (HBM)-data on these substances. For the remaining pesticides (pyrethroids, chlorpyrifos and glyphosate) the analytical methods were evaluated and harmonised and used for analysing urine samples collected in HBM4EU Aligned Studies [13,14]. These results have recently been published [14][15][16][17][18] or are in preparation for publication and were not available when the literature searches were completed for this review. The aim of this study was to present additional HBM studies on urinary metabolite concentrations of the priority pesticides: pyrethroids, chlorpyrifos, and glyphosate, across the HBM4EU partner countries published since 2000 in order to supplement the HBM-data obtained from the HBM4EU Aligned Studies. The findings in this review, combined with the results obtained from the HBM4EU-aligned studies, will provide a complete picture of the current HBM-data for these pesticides in Europe and might be useful for the planning of future HBM-studies.

Materials and Methods
Literature searches were performed in PubMed and Web of Science, for each of the search terms 'pyrethroid*', 'chlorpyrifos OR chlorpyrifos-methyl', and 'glyphosate OR AMPA' combined with 'urine AND human' or 'human biomonitoring' restricted to publications published between 01.01.2000 and 30.06.2022. We used no language restriction if an English abstract was provided. After exclusion of duplicates, all abstracts (and if necessary, method sections) were screened and only studies that presented HBM data based on urine samples collected in HBM4EU participating countries (i.e., EU Member States, as well as Norway, Iceland, Israel, Switzerland, and UK) were included. Publications with a focus on method development and/or validation were excluded if they presented HBM-data from less than 20 individuals and/or the participants were directly exposed to the pesticides as a part of the study. Table 1 presents the urinary pesticide biomarkers included. Data on limit of detection/quantification (LOD/LOQ), the frequency of detection (% above LOD/LOQ) and urine concentrations for the biomarkers were extracted. To make the studies as comparable as possible, the concentrations are presented as 50th percentiles (P50 or medians) and 95th percentiles (P95) in micrograms per litre (µg/L). Volume based concentrations were chosen because these values were provided in most studies and dilution adjustments (creatinine, osmolarity, relative density) varied between studies. If these data were not available other measures of central and upper concentrations were used as indicated in Tables. We also extracted information on some population characteristics and urine sampling method. Publications were not assessed for their overall quality but some information on quality control of the analytical methods was extracted, as was information on the analytical platform and deconjugation procedure, i.e., enzymatic hydrolyses (β-glucuronidase and/or sulfatase) or acidic hydrolysis, to help assess comparability of the urinary concentration between the studies. Finally, potential information on exposure determinants identified within the individual studies was included.

Pyrethroids
We identified 44 studies presenting urinary pyrethroid metabolite concentrations in European populations from 13 different countries (Table 2) mainly from the Western and Southern part of the EU. The datasets represent a range of different population groups including pregnant women (5 studies) and children (12 studies). Most of the studies were regional and only three studies claimed to be nationally representative for the respective population group, i.e., for pregnant women [19] and adults [20] in France and for children in Germany [21]. The majority of studies were based on single spot urine samples (25 studies) or first morning voids (FMV) (12 studies).
The generic pyrethroid metabolite, 3-PBA, representing the combined exposure to most pyrethroids, was included in most studies, while inclusion of more specific metabolites was more variable and only three studies included CFCA which is a metabolite of bifenthrin and λ-cyhalothrin [22][23][24]. The specific metabolites are formed in parallel to 3-PBA by ester cleavage of the parent compounds catalysed by carboxylesterase enzymes [25]. Accordingly, most studies showed highest detection frequency and urine concentration of 3-PBA.
The methodologies used to determine the metabolite concentrations varied both regarding pre-treatment of the urine sample and the analytical platform used for detection. Pyrethroid metabolites are present mainly as phase II conjugates (mainly glucuronide) in urine [26] and omitting a deconjugation step might underestimate the exposure level. Therefore, information on a deconjugation step or not, as well as the method used (i.e., enzymatic or acidic hydrolyses) were reported for each study (Table 2). Acidic hydrolyses is less specific than enzymatic deconjugation leading to potential release from other conjugates than glucuronide and thus slightly higher concentrations, as previously demonstrated for 3-PBA [27]. Of the included studies, 18 reported enzymatic deconjugation using β-glucuronidase, 18 reported acidic hydrolysis, and 8 studies did not mention any deconjugation step. For these studies, it was not possible to determine if the description was missing or deconjugation was not performed. Two of these datasets, based on the French PELAGIE cohort [28,29], and one from the Spanish INMA-Granada cohort [30] reported very low 3-PBA concentrations compared to the other studies. Thus, we suspect that these concentrations might be underestimated. The potential impact on the concentrations of the more specific pyrethroid metabolites is difficult to assess since these concentrations were more like those in other datasets.
Regarding the analytical platform, LC-MS/MS was used for all the metabolites in 19 studies. Two datasets from children and mothers from the PELAGIE cohort, respectively, used LC only for 3-PBA and 4-F-3-PBA while other metabolites were analysed by GC-MS/MS [28,29]. GC-MS/MS was used for all metabolites in 21 of the studies. One study used an immunoassay [31] and one study did not describe the analytical method but referred to an ISO9001 accredited lab [22]. Reported LOD/LOQs for, e.g., 3-PBA varied from a LOD of 0.004 to a LOQ of 0.8 µg/L between the included studies leading to large variation in detection frequencies.
The wide variation in urine sampling and analytical methods hamper the comparability of the results across studies and the possibility to assess time trends and geographical differences in pyrethroid exposure. Two studies from Sweden reported increasing 3-PBA concentrations between 2000 and 2017 among young adults [23] and between 2009 and 2014 among women after delivery [32]. Further, the highest 3-PBA concentrations in general population groups were reported from studies with urine samples collected after 2015, i.e., among children from Cyprus (median 1.93 µg/L) [33], the Valencia region in Spain for children (1.63 µg/L) [34] and lactating mothers (1.7 µg/L) [35] and among adolescents (0.87 µg/L) and children (0.98 µg/L) from Flanders in Belgium [36,37].
When urine samples were obtained from both children and adults within the same country and time period, the 3-PBA concentrations were in general higher in children than adults, e.g., medians of 0.56 vs. 0.24 µg/L, respectively, in Denmark [38,39], 0.29 vs. 0.23 µg/L in Poland [40], and 0.40 vs. 0.24 µg/L in Slovenia [41].
A limited number of the studies investigated exposure sources for pyrethroids in the general population. Regarding dietary exposure determinants, consumption of vegetables, fruits, and food items based on cereals (e.g., pasta and whole grain bread) [20,28,[42][43][44][45] and in some studies also fish [19,20] was associated with higher urinary pyrethroid metabolite concentrations. High organic food consumption was associated with lower urinary concentrations [2,28].
For non-dietary exposure determinants, indoor use of biocides [19,37,[43][44][45][46] including pet care products [40] was related to higher urinary metabolite concentrations. Two studies from France reported higher pyrethroid metabolite concentrations among pregnant women [19] and children [28] living in the vicinity of cultivated crops indicating some driftexposure of residents in agricultural areas. Further, higher concentrations were reported from children and parents in rural areas in Poland compared to urban residence [40] and among children whose parents were occupationally exposed to pesticides [28,38]. Only few, and mostly small, studies included occupationally exposed groups such as pesticide applicators [47], farmers [31,48,49] and greenhouse workers [50] or focused on residents after indoor use of pyrethroids [51,52]. Overall, these studies found higher urinary concentrations of pyrethroid metabolites related to recent occupational or residential exposure although a few studies were unable to demonstrate a difference, likely because of high LODs and correspondingly low numbers of participants with detectable concentrations [48,52].

Chlorpyrifos
The specific metabolite, TCPy, of chlorpyrifos/chlorpyrifos-methyl was included in 23 studies representing 12 different countries (Table 3). Of these, the urine samples were collected among pregnant women in six studies, women after delivery in two studies, and children in seven studies. Only one study, among pregnant women in Norway (MoBa), was reported to be nationwide [70]. One study collected repeated spot urine samples in each trimester of pregnancy [71], one study collected a 24 h urine sample [68] while the remaining studies were based on single spot urine samples (14 studies) or FMVs (7 studies).
TCPy was analysed by LC-MS/MS in 16 studies and by GS-MS/MS in six studies while one study used Gas-Liquid Chromatography. The reported LODs/LOQs varied between the studies from a LOD of 0.02 µg/L to a LOQ of 0.8 µg/L. A deconjugation step based on enzymatic (16 studies) or acidic (4 studies) hydrolyses was described in all except for three studies. One of these, a study from Spain [72], reported a low detection frequency and maximum concentration compared to most other studies, and an underestimation of the concentration cannot be excluded. However, another Spanish study [71] reported an even lower detection frequency and maximum concentration despite inclusion of enzymatic hydrolyses. The two studies included different population groups, i.e., adolescent males and pregnant women, respectively, but sampling years (2017-2019 and 2016-2017) and LODs were comparable. Other studies from the Valencia Region in Spain from the same time period reported considerably higher detection frequencies and urine concentrations [34,35]. These results indicate regional differences in exposure levels but, as for the pyrethroids, the variation in urine sampling and analytical methods hamper direct comparison of the results across the studies. Overall, the TCPy concentrations varied considerably between the studies. The highest median concentrations were reported among children from Cyprus (6.72 µg/L) [33], adolescents and children from Belgium (4.45 and 3.87 µg/L) [36,37], and among adults from the Amirim community in Israel (4.32 µg/L) [73] based on urine samples collected between 2013 and 2018 (Table 3).
Only one study investigated time-trends in TCPy concentrations and reported an increasing trend from 2001 to 2017 among adolescents in Sweden with the highest median concentration in 2009 and the highest P95 concentration in 2017 [23]. Few of the studies investigated dietary exposure determinants and reported higher TCPy concentrations associated with high vegetable consumption [69,73,74] and negatively associated with organic food intake [73]. Besides, TCPy was associated with higher education level and lower BMI in some studies [35,39,41] which might reflect a diet with high vegetable and fruit content. Higher TCPy concentrations were also related to farm working [49]. None of the studies included urine samples collected after the ban of chlorpyrifos/chlorpyrifos-methyl in the EU in 2020.    DF: detection frequency (%>LOD/LOQ); FMV: first morning void; German External Quality Assessment Scheme (G-EQUAS); NR: not reported; P50: 50th percentile (median); P25, P75, P90, P95: the respective percentile; GM: geometric mean; GSD: geometric standard derivation; SD: standard derivation; GW: gestational week; TCPy: 3,5,6-trichloro-2-pyridinol; crea: creatinine.

Glyphosate
Glyphosate (Gly) and its main environmental degradation product aminomethylphosphonic acid (AMPA) are excreted unchanged in urine. Since, they do not undergo phase II conjugation a deconjugation step is not necessary for urine analyses [78]. Gly was analysed in 23 studies of which 15 studies also analysed AMPA. A single study analysed AMPA but not Gly (Table 4). Two studies collected single 24 h urine samples while single spot urine samples or FMVs were used in 12 and 8 studies, respectively. Three studies did not report the urine sampling method. LC-MS/MS and GC-MS/MS was equally used in 10 studies each for quantification of Gly and/or AMPA, while four studies used Enzyme-linked immunosorbent assay (ELISA) for quantification of Gly. LODs/LOQs varied between 0.05 and 1.0 µg/L. The studies were performed in 11 different EU-countries mainly from the western and southern part of the EU while two studies included samples from several countries.
In general, detection frequencies and reported urinary concentrations of Gly were considerably higher in studies using ELISA than those using LC-or GC-MS/MS and the results were not considered to be directly comparable. Most studies using LC-or GC-MS/MS to analyse samples from the general population had detection frequencies below 50% and therefor medians could not be obtained. Among these studies, the highest P95 concentrations for Gly were reported among children from Cyprus (1.01 µg/L) [33], lactating mothers from the Valencia Region in Spain (0.62 µg/L) [79] a nationally representative group of children from Germany (0.51µg/L) [80] and young children from Germany in a regional study (0.97 µg/L) [80]. Urine samples in these studies were collected between 2014 and 2017. In general, the majority of studies were based on urine samples collected after 2010 but one study from Germany included urine samples collected between 2001 and 2015 and reported a continuous increase in the fraction of samples with detectable concentrations with a peak in 2012-2013 [81].
A few studies found associations between urinary Gly and/or AMPA and higher intake of specific food items, e.g., beer and fruit juice [82], pulses and mushrooms [83], eggs and fruit [79], nuts and whole grain rice [84], and self-produced vegetables [85]. However, most of the included studies did not investigate or were unable to identify specific exposure determinants for the general population (Table 4). Two small studies from Ireland included occupational exposures among amenity horticultural workers and found higher urinary Gly-concentrations after work exposure with peak values 3 h after exposure [86,87].

Discussion
In this review, we identified HBM-studies that measured the internal exposure to pyrethroids, chlorpyrifos, and glyphosate in European population groups by analysing urinary concentrations of suitable biomarkers. We included studies that have been published from 2000 until June 2022. For all three substance groups the number of studies increased during the years and the majority were published during the last ten years. Variation in analytical methods displaying different sensitivities impacted the reported frequencies of detection and the urinary concentrations. Besides, the urine sampling methods varied, and the quantitative data was reported differently. Thus, direct comparison of the urinary concentrations across the studies was not always possible although many of the studies participated in external quality control programs such as the German External Quality Assessment Scheme (G-EQUAS). Further, less than half of the HBM4EU participating countries were covered and especially studies from the eastern part of Europe were scarce. Many studies were regional with relatively small sample size. Despite these shortcomings, the results indicate a widespread exposure to these substances in the general EU population with marked geographical differences. Studies form Cyprus and the Valencia region in Spain reported the highest urinary concentrations for all the included pesticides. Thus, identification of the main exposure sources in these areas can be used to reduce exposure. In general, children had higher urinary concentrations of the pesticide metabolites than adults as also seen in studies from, e.g., the US [25]. An obvious explanation is a relatively higher food intake per kg body weight in children leading to higher exposure levels from pesticide residues in food, but also other physiological and behavioural differences may predispose children to elevated exposure [6,99,100].
The organophosphate chlorpyrifos was for decades one of the most widely used insecticides in agriculture worldwide [101] leading to a widespread exposure of the general population from residues in food, as reflected in the high detection frequency of TCPy in most of the included studies. However, the authorization for chlorpyrifos (and chlorpyrifosmethyl) in the EU was withdrawn by February 2020 because of concern for genotoxicity and developmental neurotoxicity [102]. Before this ban, acceptable daily intake (ADI) for chlorpyrifos was reduced from 0.01 to 0.001 mg/kg body weight/day in 2014. Because of parallel reductions in EU maximum residue levels (MRLs) in food items, the exposure level in the general population would be expected to have decreased in this period. A corresponding drop in urinary TCPy concentrations could not be documented from the included studies, but no urine samples collected after the ban in 2020 were included. Thus, the concentrations reported in the current studies can be used for comparison in future studies.
The studies on pyrethroids showed higher urinary concentrations of 3-PBA in samples collected in the most recent years indicating an increasing population exposure to these insecticides. A rise in exposure would be expected, since pyrethroids have replaced organophosphate insecticides in biocidal products and to some degree also as plant protection products. Accordingly, increasing urinary metabolite concentrations were also found in the human biomonitoring programs The National Health and Nutrition Examination Survey (NHANES) in the US [103] and the Canadian Health Measures Survey (CHMS) [104]. Indoor use of pyrethroids were associated with higher urinary metabolite concentrations in several of the included studies as also demonstrated in studies form the US [6,105].
Glyphosate is the most used pesticide worldwide and also one of the most widely used herbicides in agriculture in the EU [106] but HBM-data are limited, both from Europe and elsewhere. The included studies indicate a widespread low Gly exposure with rather low detection frequency in urine probably reflecting the low urinary excretion fraction of 1% for Gly estimated in humans after oral exposure [8,89]. In general, the knowledge on toxicokinetics of Gly and AMPA in humans is limited and more information on, e.g., uptake after inhalation exposure is needed.
For both pyrethroids and glyphosate there is a paucity of HBM-data regarding exposure levels in potentially higher exposed (sub)populations in occupational and environmental settings, e.g., studies focusing on exposure after indoor and outdoor residential use, exposure from living in vicinity to pesticide treated areas, occupational and paraoccupational exposure levels including take-home exposure after work. Such studies are complicated by the fact that currently used pesticides are rapidly metabolised and excreted within few days and therefor urinary concentrations reflect only recent exposure to the specific pesticides. Therefore, such studies require careful planning to obtain valid information on exposure levels and peak exposures.
The results from this review illustrate the need for harmonisation of the analytical methods as well as the reporting of HBM-data to enable comparisons of exposure levels across studies to obtain information on population differences in exposure sources and time-trends. Such a process was initiated within the HBM4EU initiative, in which harmonised HBM-data was obtained for the prioritised pesticides either by analysing new urine samples or by quality-assurance of data already collected in the HBM4EU Aligned Studies [13,107]. In this way, HBM-data on glyphosate was achieved among children from five countries, Slovenia, Germany, France, Belgium and Cyprus [16], and adults from Germany, Switzerland, France, and Iceland [15]. Data on pyrethroids and chlorpyrifos were obtained among children from Slovenia, Cyprus, France, Belgium, the Netherlands, and Israel and among adults from Germany, France, Switzerland and Israel [14,17,18]. These HBM-data from the HBM4EU Aligned Studies is available for visualization in the EU-HBM dashboard: https://www.hbm4eu.eu/what-we-do/european-hbm-platform/ eu-hbm-dashboard/ (accessed on 18 10 2022) along with HBM-data from some of the studies included in this review. However, more HBM-data are needed to get an EU-wide picture of the exposure and to evaluate differences between countries and population groups, time trends, and age-related differences in exposure levels and sources. The data presented in this review, combined with the HBM-data from the HBM4EU-aligned studies, can be used as a baseline for future studies of exposure to these pesticides. Furthermore, there is a need to establish harmonised and sensitive biomarkers for other frequently used pesticides/pesticide groups in order to get an overall picture of pesticide exposure in Europe. Funding: This review was performed as part of the HBM4EU initiative which is co-financed by EUs Horizon 2020 research and innovation programme (grant agreement No 733032).

Informed Consent Statement: Not applicable.
Data Availability Statement: Not applicable.

Conflicts of Interest:
The authors declare no conflict of interest.