Pesticide detection in air samples from contrasted houses and in their inhabitants' hair

Science of the Total Environment 544 (2016) 845–852

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Pesticide detection in air samples from contrasted houses and in their inhabitants' hair Caroline Raeppel a,b, Guillaume Salquèbre b, Maurice Millet a,⁎, Brice M.R. Appenzeller b a b

Groupe de Physico-Chimie de l'Atmosphère, Institut de Chimie et Procédés pour l'Energie, l'Environnement et la Santé (UMR 7515 CNRS — Université de Strasbourg), Strasbourg, France Laboratory of Analytical Human Biomonitoring, CRP-Santé, Luxembourg, Luxembourg

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• Analysis of pesticides in air of homes by passive sampling • Several pesticides were detected in human hair • Pesticides in the two matrices were not necessarily associated • Exposure profiles varied between houses and between inhabitants of the same house

a r t i c l e

i n f o

Article history: Received 27 September 2015 Received in revised form 4 December 2015 Accepted 4 December 2015 Available online xxxx Editor: Adrian Covaci Keywords: Pesticide Indoor environment Hair analysis Air analysis Exposure

a b s t r a c t In order to identify associations between indoor air contamination and human exposure to pesticides, hair samples from 14 persons (9 adults and 5 children below 12 years) were collected simultaneously with the air of their 5 contrasted houses. Three houses were situated in Alsace (France), one in Lorraine (France) and one in Luxembourg (Luxembourg). Houses were located in urban (n = 3), semi-urban (n = 1) and rural areas (n = 1). Twenty five (25) pesticides were detected at least once in indoor air samples and 20 pesticides were detected at least once in hair samples. The comparison between hair and air samples for the same sampling periods shows that pesticides detected in the two matrices were not necessarily associated. Exposure profiles varied from one home to another but also between inhabitants of the same home, suggesting that exposure can be different between inhabitants of the same home. This study demonstrated the usefulness and the complementarity of hair analysis, for the personalized biomonitoring of people exposure to pesticides, and air analysis, for the identification of airborne exposure and house contamination. © 2015 Elsevier B.V. All rights reserved.

1. Introduction

⁎ Corresponding author at: Université de Strasbourg/CNRS, Institut de Chimie et Procédés pour l'Energie, l'Environnement et la Santé ICPEES UMR 7515, Groupe de Physico-Chimie de l'Atmosphère, 1 rue Blessig, F-67084 Strasbourg Cedex, France. E-mail address: [email protected] (M. Millet).

http://dx.doi.org/10.1016/j.scitotenv.2015.12.020 0048-9697/© 2015 Elsevier B.V. All rights reserved.

The processes and levels of air contamination induced by agricultural activities (maize crops, vineyards, etc) are currently well documented (Van Dijk and Guicherit, 1999; Peck and Hornbuckle, 2005; Scheyer et al., 2007; Schummer et al., 2010). Pesticides are however also used in other sectors for different treatments such as parks and

846

C. Raeppel et al. / Science of the Total Environment 544 (2016) 845–852

public roads maintenance, pest control in buildings, private gardens maintenance or pets care. These uses lead to outdoor contamination as well as indoor contamination when pesticides are applied in enclosed rooms, or when transfer from outdoor to indoor trough clothes, shoes or air exchange occurs (Bouvier et al., 2006). Several studies were carried out to evaluate occupational exposure of applicators (Delhomme et al., 2010; Harris et al., 2010; Cattani et al., 2001). Contamination of buildings and subsequent non-occupational exposure were also investigated. Different classes of pesticides such as organophosphates or pyrethroids applied in houses were studied (Lu and Fenske, 1998; Tulve et al., 2007; Williams et al., 2008) and transfer from the area treated to other surfaces was already reported (Stout and Mason, 2003). Such contamination is likely to increase children exposure due to extended contact with the floor as well as hand-to-mouth behaviour (Gurunathan et al., 1998; Harrad et al., 2010; Van Den Eede et al., 2011). Transfer towards indoors through pets after lawn treatment was also identified (Morgan et al., 2008). Furthermore, several studies pointed out the exposure to pesticides used for wood treatment such as pentachlorophenol or lindane (γ-HCH) (Neuber et al., 1999; Wilson et al., 2007). Human exposure can be evaluated through environmental monitoring such as air analysis. For the collection of air samples, passive sampling constitutes an alternative to traditional methods. This technique is based on the migration of chemicals from the air to the sampling support through molecular diffusion (Gorecki and Namiesnik, 2002). This simple method is easy to set up and costefficient, needs no power supply and allows large scale sampling, which is essential for providing a specific description of the spatial and temporal variations of air contamination. There are several kinds of passive air samplers (PAS) with different design and sampling support. PAS have already been used in studies focused on the contamination of air by pesticides in order to evaluate levels of contamination, distribution or transport of contaminants on long distance (Shen et al., 2005; Harner et al., 2006; Bohlin et al., 2008; Schummer et al., 2012a). Human exposure can also be evaluated through biomonitoring, which consists in the analysis of chemicals and/or their metabolites in biological matrices. For this purpose, urine is classically used and to a lesser extends blood. More recently, a growing interest is currently observed for hair analysis of organic pollutants such as pesticides for assessing human exposure. Contrary to blood or urine which provides information about substances recently absorbed, hair gives the possibility to evaluate the exposure to substances accumulated during weeks or months preceding the sampling time. Consequently, hair analysis is particularly adapted to study chronic exposure (Salquèbre et al., 2012). Furthermore, the sampling is easy to organise, does not required medical personnel and does not pose an infectious risk. Contrary to environmental analysis that requires considering transfer coefficient to assess exposure, biomonitoring provides direct information on the internal dose of pollutants (the dose that actually entered the body). This is the case for blood for example. However, for hair, it must be considered that a potential contribution from external exposure occurred (air, dust) and must be added to the internal dose of pollutants. On the other hand, biomonitoring does not allow the identification of the origin of the exposure and results integrate the contribution of the different sources of exposure. Among the different studies already carried out in the field, organochlorines were the most investigated pesticides, and lindane and DDT the most frequently detected (Covaci et al., 2008; Tsatsakis et al., 2008, 2014; Appenzeller and Tsatsakis, 2012). Hair analysis enabled to highlight differences in the level of contamination between different groups of population as for example higher exposure of inhabitants of urban area compared to rural area (Zhang et al., 2007). In the agricultural sector, occupational exposure was already demonstrated by hair analysis (Cirimele et al., 1999), and correspondences between the nature of the pesticides detected in hair of farmers and their agricultural activities were observed in a study carried out in Luxembourg (Schummer et al., 2012b). Hair analysis also allowed the

identification of significant exposure of pregnant women to pesticides due to domestic use (Ostrea et al., 2006). The aim of this pilot study was to identify possible associations between air contamination and human exposure to pesticides of inhabitants from 5 contrasted houses. Passive sampling was chosen for air sampling and human exposure was evaluated through hair samples from 14 persons collected during the same period as air. A method of extraction and analysis was developed for each kind of matrix. 31 and 26 pesticides of different chemical families were investigated in air and hair samples respectively. These pesticides were chosen to be representative of local pesticides application in agriculture, of non-agricultural application and of indoor domestic used.

2. Methods 2.1. Chemicals and reagents Pesticides standards of Pestanal® quality were purchased from Riedel de Haën (Sigma Aldrich, St. Quentin Fallavier, France). Internal standards were purchased from Dr. Ehrenstorfer (Augsburg, Germany). The pesticides stock solutions (1 g L−1) and a mix standards solution (10 mg L−1) were prepared in acetonitrile. A solution of internal standards was also prepared in acetonitrile (p,pʹ-DDE d8, pentachlorophenol 13C6, MCPA d6, permethrin d6 at 1 mg L−1 and trifluralin d14, γ-HCH d6, endosulfan-b d4 at 0.1 mg L−1). The purity of the standard pesticides and stable isotope labelled analogues was always above 98%. Pentafluorobenzyl bromide (PFBBr), N-(t-butyldimethylsilyl)-Nmethyltrifluoroacetamide (MtBSTFA) purity ≥97% and potassium carbonate anhydrous (K2CO3) from Fluka were purchased from Sigma Aldrich (St. Quentin Fallavier, France). Acetonitrile and ethyl acetate were supplied by Biosolve (Dieuze, France) and by Prolabo (LPCRSchiltigheim, France). Ultrapure water was produced using a water purification chain (Milli-Q A10 Advantage) from Millipore (Brussels, Belgium).

2.2. Samples collection Hair samples from 14 persons (9 adults and 5 children below 12 years) were simultaneously collected with the air of their 5 contrasted houses. Three houses (A, B, C) were situated in Strasbourg and its suburban area (Alsace, France), one in a town (D) near Metz (Lorraine, France) and one in (E) Luxembourg (Luxembourg). Houses were located in urban (n = 3), semi-urban (n = 1) and rural areas (n = 1). One home situated in Strasbourg centre was an old building with apparent wood structure on which protection treatments were classically applied in the past (but with no indication available). The other houses were possibly submitted to pesticides contamination due to private gardens or pesticides treatments surrounding their location. Air samples were collected using passive air samplers (PAS), which consisted in a Tenax® resin tube protected by a specially designed shelter allowing airflow. The shelters were produced in the laboratory on the basis of a model proposed by Wania et al. (2003) for XAD-2 resin. PAS were placed in a room and outdoor of each of the 5 houses during 2 to 13 weeks over a period of 1 year (October 2010–October 2011). Details on sampling periods and number of samples are described in Table 1. Hair samples were collected by a member of the laboratory staff, as previously described (Schummer et al., 2012b; Salquèbre et al., 2012). Strands of hair were sampled from the back of the head, as close as possible to the scalp. Hair samples were collected as possible on a monthly basis and were stored in aluminium foil at room temperature before extraction and analysis (Becker et al., 2014). Volunteers were informed about the objectives of the study and gave their authorization for the hair sampling. The study was approved by the Committee of the protection of persons East IV (France).

C. Raeppel et al. / Science of the Total Environment 544 (2016) 845–852

847

Table 1 Details of air sampling location and duration. Home

Type

Place

Location

Sampling area

Garden, field

Sampling period

A

Apartment in a very old building (18th century) Apartment in a recent building (about 10 years) Recent house (4 years) House House

Strasbourg, France

Urban

Living room

/

18/10/10–22/08/11 (n = 22)

Near Strasbourg, France

Semi-urban

Living room

Gardens and fields nearby

18/10/10–05/09/11 (n = 22)

Near Strasbourg, France Near Metz, France Luxembourg, Luxembourg

Semi-urban Rural Urban

Living room Living room Living room

Gardens and fields nearby Gardens and fields nearby /

18/10/10–22/08/11 (n = 21, one lost) 02/11/10–26/09/11 (n = 8) 02/11/10–18/10/11 (n = 7)

B C D E

2.3. Air analysis Analyses of passive air samplers were carried out by using an automatic thermal desorption unit (ATD 350, Perkin-Elmer Corp., Norwalk, CT, USA), connected to an Autosystem XL GC coupled to a Turbomass gold detector (Perkin-Elmer Corp., Norwalk, CT, USA). ATD 350 was coupled to GC–MS via a valve and a transfer line maintained at 300 °C and 280 °C, respectively. Separation have been performed on a Varian Factor-Four V5-MS (equivalent to 5% Phenyl, 95% Polydimethylsiloxane, Varian Les Ulis, France) capillary column (60 m × 0.25 mm i.d., 250 μm film thickness) as follows: 50 °C (5 min) to 150 °C at 25 °C/min, to 250 °C at 3 °C/min and to 300 °C (15 min) at 15 °C/min. Helium was used as carrier gas at 1.5 mL min−1 (regulated constant flow). Temperatures of the MS source and transfer line were maintained at 280 °C and 320 °C, respectively. Spectra of pesticides were obtained by electron impact ionisation (EI) at 70 eV. Depending on the pesticide, two of three ions were selected from the spectrum of each pesticide to quantify the response in the selected ion monitoring mode (SIM). For the ATD, a two-stage desorption was adopted to reduce the component bandwidths and improve the efficiency of the chromatographic separation. At first, the sample tube is heated at 300 °C for 30 min under a stream of He (desorb flow: 45 mL min−1) to extract pesticides from the sampling tube by thermal desorption and sweep them onto a cold focusing trap. This trap, which is empty, was maintained at −30 °C by the Peltier effect. The cold trap refocused all the components eluting from the sample tube. When primary desorption stage is complete, the cold trap is electrically heated at a rate of 5 °C s− 1 (trap low) to 300 °C (maintained for 5 min) so as to elute all the retained components in a vapour band as narrow as possible. The outlet-split flow is adjusted for high-resolution capillary column to 10 mL min − 1 leading to a fraction of 13% of the sample entering the column. Details on the method can be found in Raeppel (2012) and in Supplementary material S1.

2.4. Hair analysis 2.4.1. Hair treatment and derivatization step Before pulverization, hair samples were first washed with water and with acetonitrile, for 2 min under agitation (Schummer et al., 2012b). Samples were then dried in an oven for 5 min at 60 °C. Afterwards, hair samples were pulverized with a ball mill (Retsch, Haan, Germany) and 50 mg of hair powder were put in a 4 mL glass tube. Ten μL of internal standard mix (0.1 mg L− 1) and 990 μL of acetonitrile were added for extraction. Sealed tubes were placed on an incubator shaker heated at 40 °C for 16 h. Then, samples were centrifuged for 10 min at 5000 rpm and 700 μL of supernatant were isolated for analysis. Because some pesticides need a derivatization step before GC analysis, samples were split in two aliquots of 350 μL. One aliquot of each sample was used for analysis without derivatization. Before injection, samples were evaporated to dryness under nitrogen flow and re-suspended into 50 μL of ethyl acetate. The second aliquot

was used for pesticides derivatization with PFBBr. Aliquots were transferred in a glass tube and 50 μL of PFBBr, 30 mg of K2CO3 and 650 μL of acetonitrile were added. Sealed tubes were placed on an incubator shaker heated at 60 °C during 4 h. Samples were centrifuged for 3 min at 5000 rpm prior to remove the supernatant. Then, samples were evaporated to dryness at ambient temperature under nitrogen flow and reconstituted with 50 μL of ethyl acetate. 2.4.2. Gas chromatography mass spectrometry analysis Analyses were performed on an Agilent gas chromatograph (Agilent Technologies 7890 A GC System) equipped with a CTC Analytics Pas System autosampler and coupled to a triple quadrupole mass spectrometer (5975 C Inert XL MSD) operating in NCI ionization mode. Separation was done on an Agilent HP-5MS capillary column (30 m × 0.25 mm × 0.25 μm). The helium carrier gas flow was set at 1.2 mL min− 1. The reagent gas was methane set at 40 mL min− 1. The temperature of the injector was set at 250 °C. The following oven programme was used for analyses: 70 °C for 5 min, then an increase of 10 °C min− 1 to 200 °C, then an increase of 2 °C min− 1 to 240 °C and finally an increase of 10 °C min−1 to 300 °C (hold time 3 min). Because some pesticides need a derivatization step before GC analysis, two distinct injections were realized: one without derivatization and one after derivatization with PFBBr. The injection volume was 1 μL. The analyses were performed in the selected ion monitoring mode. Mass spectrometric details concerning the pesticides analysed and the internal standards are given in Supplementary material S2. 2.4.3. Calibration and validation of hair analysis method For the validation, 50 mg of pulverized hair samples were spiked with a solution containing a mix of the studied pesticides. The calibration was performed using internal standards and the concentration for the internal standards was 200 pg mg−1 for p,pʹ-DDE d8, pentachlorophenol 13C6, MCPA d6, permethrin d6 and 20 pg mg− 1 for trifluralin d14, γ-HCH d6, endosulfan b d4. The pesticides mix solution was added at 10 different concentrations (0.1, 0.5, 1; 5, 10, 20, 40, 60, 80 and 100 pg mg− 1) and 3 replicates for each concentration were carried out. The limits of quantification (LOQ) were determined as the lowest concentration with a precision and accuracy inferior or equal at 20% calculated on 5 replicates. The limits of detection (LOD) were calculated as 1/3 of the LOQ. Precision and accuracy was determined at two different concentrations (10 and 100 pg mg−1) with 5 replicates. Recovery was evaluated by comparing the mean of the responses of 5 hair samples spiked at 100 pg mg− 1 before the extraction to the mean of the responses of 5 hair samples spiked after the extraction. 3. Results and discussion 3.1. Air and hair analysis method The validation parameters of the method of air and hair analysis are presented in tables S1, S2 and 2 and are presented for air analysis in a previous paper (Raeppel, 2012). For hair analysis, the coefficients of determination of the calibration curves ranged from 0.962 to 0.997 and were higher than 0.98 for most pesticides. LOD varied from 0.15 to

848

C. Raeppel et al. / Science of the Total Environment 544 (2016) 845–852

6 pg mg−1 and LOQ from 0.5 to 20 pg mg−1. The LOQ determined for the pesticides analysed after derivatization were higher than for other pesticides because the background noise observed were much more important for the latter compounds. LOD were comparable to values presented in previous studies, in which pesticides of different chemical classes were analysed in hair using solid phase microextraction and GC–tandem mass spectrometry (Schummer et al., 2012b; Salquèbre et al., 2012; Tzatzarakis et al., 2014). For the majority of the pesticides, precision and accuracy were below 20%. The higher values observed for intraday precision and accuracy could be explained by partial loss of some pesticides during the extraction due to micro-leakage from some glass tubes. As no certified hair sample is available, recovery was determined using hair spiked with standard solution of pesticides. Recovery was quite satisfactory for pesticides analysed directly but was in general lower for compounds analysed after derivatization certainly because the derivatisation procedure do not reach 100% of efficiency. 3.2. Pesticides analysis In air samples, all the pesticides analysed were detected at least one time during the campaign (Supplementary material S3). Outdoor measured concentrations varied between bLOQ (bifenthrin, cyprodinil) and 738 (acetochlor) ng sampler−1. Higher levels are lower to those of a previous study conducted by Schummer et al. (2012a) which used XAD-2 based passive samplers exposed between one and three months (42 100 ng sampler−1 for dinocap). Indoor, the range of measured concentration is higher as concentrations varied between bLOQ (bifenthrin, cyprodinil, pp.'DDE, dichlobenil, α-endosulfan and flazasulfuron) and 24 109 ng sampler−1 (acetochlor). Some pesticides have been more frequently detected. Outdoor, acetochlor, MCPA, propachlor, trifloxystrobin and trifluralin have been more frequently observed (frequence of detection, calculated by home higher than 40%). Bifenthrin, β-endosulfan, fluroxypyr, metazachlor and triclopyr have also presented detection frequencies higher than 40% but these frequencies have been calculated for only four samples (houses D and E). Indoor, pesticides more frequently detected were pentachlorophenol (82%), γ-HCH (78%) and α-endosulfan (61%). Some pesticides frequently observed outside were also detected (acetochlor, MCPA or trifloxystrobin). Bifenthrin, cyprodinil, triclopyr and trifluralin have also a frequency of detection higher than 40% (calculated for 4 samples for triclopyr and trifluralin). In some cases, pesticides have been detected outdoor but not indoor (diflufenican, house A or picloram, house B). This observation could be the result of an indoor concentration below detection limit or by a low transfer potential from outdoor to indoor. The opposite case was also observed (oxadiazon, house C and picloram, house E). This behaviour is more difficult to interpret as these pesticides are used outdoor. Previous studies performed in Strasbourg by Scheyer et al. (2007) and Schummer et al. (2010) have shown that outdoor air is contaminated by diflufenican (0.1–1.22 ng m− 3), γ-HCH (0.03–1.6 ng m− 3), tebuconazole (0.2–8.2 ng m− 3), endosulfan (0.02–0.5 ng m− 3) and pendimethalin (0.3–7.8 ng m − 3). However, comparing these concentrations with Tenax® passive tubes is not possible as the sampling rate for these pesticides have not yet been determined. Some values obtained for home E can be compared with those obtained by Schummer et al. (2012a,b) with XAD-2® based passive samplers exposed between 1 to 3 months. Concentrations for bifenthrin (3.3–9.3 ng sampler− 1 ), cyprodinil (9.4 ng sampler− 1) and trifloxystrobin (17.5 ng sampler− 1) are in the same order of magnitude than those obtained by Schummer et al. (2012a). Mean values of pesticides concentrations calculated for outdoor and indoor for each houses are presented in Table 3. The number of pesticides detected in houses D and E is lower than other houses and can be explained by the short duration of the campaign in comparison to the other houses A, B and C. For these houses, the number of detected pesticides indoor is higher than outdoor and the maximum of detection

was for the home C (27 pesticides). For outdoor air, A, B and C presents an equivalent number of detected pesticides. However, it is known that a variability exist between urban and rural areas for airborne pesticides (Scheyer et al., 2007; Marlière, 2009). Houses A and B are situated in an urban areas while C has a typical peri-urban typology. These localizations could explain that the number of detected pesticides is equivalent and that home C shows the highest number of detected pesticides indoor. The comparison of mean values between the 5 experimented houses shows and important variability of concentrations (Table 2). Calculated coefficients of variations for each pesticide were, in most cases, higher than 50% for indoor and outdoor except for trifloxystrobin and pentachlorophenol where variations coefficients were respectively of 13% (indoor)–16% (outdoor) and 16%. Concentrations of some pesticides were always higher outdoor than indoor (i.e. cyprodinil and trifloxystrobin). Trifluralin was the only pesticide which was detected in all experimented houses indoor and outdoor. As this molecule was the more frequent detected pesticides outdoor (81%) in France (Marlière, 2009) but also indoor as reported by Atmos'Fair Bourgogne (2006), its presence in the present study is not surprising. Three other pesticides (bifenthrin, cyprodinil and triclopyr) have been detected outdoor on the different sites, except home D. Nine other molecules (acetochlor, fluroxypyr, flazasulfuron, MCPA, mecoprop-p, pentachlorophenol, picloram, propachlor and trifloxystrobine) have been detected in the three houses (A, B and C) which are geographically close while 2,4-D was detected only on the house B. Few pesticides have been quantified only indoor (propiconazole, pendimethalin, p,pʹDDE and γ-HCH). In hair samples, 20 pesticides were detected at least one time. The results for pesticides detection in hair samples are presented in Table 4. The concentrations measured (for 6 pesticides) varied between 0.8 (alpha-cypermethrin) and 104 pg mg−1 (2,4 D). Diflufenican was the most frequently detected but always below the LOQ. In previous study focused on the exposure of farmers, diflufenican was already described as one of the most frequently detected pesticides (Schummer et al., 2012b). The higher levels of concentration reported in the latter study (b 0.2–51.89 pg mg−1) can probably be explained individuals' exposure during their occupational activities. α- and β-endosulfan and MCPA were also frequently detected. This high rate of detection suggests regular exposure of individuals, whatever the place of residence. α- and β-endosulfan were already detected in hair samples in other studies (Salquèbre et al., 2012; Schummer et al., 2012b). Unfortunately, results for MCPA were not quantified due to the presence of an analytical artefact in the hair samples analysis. Alpha-cypermetrin, bifenthrin, γ-HCH, oxadiazon, pentachlorophenol and trifluralin were detected in more than half of the individuals but always under the LOQ except for alpha-cypermethrin and pentachlorophenol. γ-HCH was often analysed in previous studies and quantified up to 905.9 pg mg−1 (Appenzeller and Tsatsakis, 2012). Even if γ-HCH has been banned for several years, positive detection in hair samples were reported in recent studies (Salquèbre et al., 2012; Schummer et al., 2012b). Similar observation were made for pentachlorophenol, for which concentration levels observed in the present study were in the same order of magnitude as what was previously reported by Salquèbre et al. (2012). Other pesticides were detected in less people, suggesting more specific sources of exposure for each individual. For the pesticides presenting concentrations above the LOQ, the mean concentrations were calculated for each subject (attributing the value of ½ LOQ to samples below the LOQ). Concentration ranged from 27 to 66 pg mg−1 for 2,4 D, from 3 to 14 pg mg−1 for flazasulfuron, from 7 to 16 pg mg−1 for mecopropP, from 5 to 14 pg mg−1 for pentachlorophenol and from 12 to 34 for propachlor. The variability in the mean concentrations calculated for each subjects varied by a factor of 2.3 to 4.7. Some pesticides were detected in hair samples of inhabitants of one or several houses only. For example, 2,4 D was detected in inhabitants' hair of one home only,

C. Raeppel et al. / Science of the Total Environment 544 (2016) 845–852

849

Table 2 Validation parameters of the hair samples analysis method. Pesticide

Direct analysis Acetochlor Aldrin α-cypermethrin Bifenthrin Dieldrin Diflufenican α-Endosulfan β-Endosulfan γ-HCH Metazachlor Oxadiazon Pendimethalin pp’-DDE Propachlor Propiconazole Tau-fluvalinate Trifloxystrobin Trifluralin

R2

Domain of linearity [pg mg−1]

Intraday LOQ

Interday LOD

Precision [RSD%]

[pg mg−1]

Concentration [pg mg−1]

Recovery [%]

Accuracy [RSD%]

Precision [RSD%]

Accuracy [RSD%]

10

100

10

100

10

100

10

100

5–100 20–100 0.5–100 5–100 1–100 0.5–100 0.5–100 0.5–100 10–100 10–100 10–100 5–100 5–100 10–100 20–100 0.5–100 0.5–100 0.5–100

0.987 0.990 0.993 0.997 0.993 0.992 0.993 0.992 0.983 0.991 0.992 0.990 0.990 0.991 0.962 0.991 0.990 0.994

5 20 0.5 5 1 0.5 0.5 0.5 10 10 10 5 5 10 20 0.5 0.5 0.5

1.5 6 0.15 1.5 0.3 0.15 0.15 0.15 3 3 3 1.5 1.5 3 6 0.15 0.15 0.15

19.3 – 6.2 5.5 12.4 7.8 6.5 5.4 19.6 8.1 5.2 5.5 1.7 17.5 – 5.6 9.2 2.0

16.4 14.6 14.8 6.7 8.4 9.5 8.7 8.4 27.0 19.5 6.3 17.1 6.4 14.0 11.6 16.6 8.4 4.1

13.3 – 3.7 4.6 16.0 32.2 12.4 14.1 51.6 16.9 16.5 19.6 18.2 12.6 – 9.7 31.5 19.9

6.9 19.7 18.2 2.2 14.7 0.4 16.6 3.3 36.9 7.3 4.1 8.4 4.1 3.1 10.4 19.2 1.4 0.6

6.0 – 8.4 14.4 9.2 12.7 7.5 9.2 16.3 12.5 11.7 16.4 10.6 13.4 – 12.4 13.8 7.8

7.3 4.4 7.6 7.6 8.0 12.2 7.8 9.2 6.4 8.8 8.4 12.0 9.5 13.4 9.5 8.2 10.6 7.7

10.3 – 8.5 4.7 15.1 17.5 16.4 15.6 19.5 9.1 7.2 8.9 2.5 11.7 – 10.4 14.2 16.6

2.7 0.6 2.4 2.9 4.6 3.8 4.2 4.0 3.9 2.7 2.0 5.3 3.7 2.5 5.2 2.2 4.9 3.5

96 95 128 103 94 101 94 94 96 101 95 95 95 93 105 146 100 90

Analysis after derivatization 2,4-D 20–100 Bromoxynil 20–100 Flazasulfuron 5–100 Fluroxypyr 10–100 MCPA 20–100 Mecoprop-P 5–100 Pentachlorophenol 5–100 Triclopyr 10–100

0.985 0.989 0.984 0.977 0.987 0.991 0.992 0.992

20 20 5 10 20 5 5 10

6 6 1.5 3 6 1.5 1.5 3

– – 5.1 24.7 – 4.4 42.2 12.5

4.5 16.0 13.6 20.5 12.0 10.3 12.5 12.1

– – 0.6 53.8 – 44.9 54.8 28.2

10.7 39.7 31.4 10.1 4.4 4.3 7.0 9.1

– – 5.0 16.3 – 18.9 16.3 12.3

8.8 4.6 7.4 1.0 14.9 12.6 8.2 10.3

– – 17.8 8.1 – 5.5 5.7 4.4

3.9 1.1 5.5 0.6 7.4 6.2 2.7 7.7

8 25 91 11 27 40 38 18

Table 3 Mean concentrations (ng.sampler−1) measured outdoor and indoor for the 5 homes. A

B

C

D

E

Pesticide Indoor 2,4-D Acetochlor Aldrin Α-cypermethrin Bifenthrin Bromoxynil Clopyralid Cyprodinil p,p'DDE Dichlobenil Dieldrin Diflufenican α-Endosulfan β-Endosulfan Flazasulfuron Fluroxypyr γ-HCH MCPA Mecoprop-P Metazachlor Oxadiazon Pendimethalin Pentachlorophenol Picloram Propachlor Propiconazole Tau-fluvalinate Tebuconazole Triclopyr Trifloxystrobin Trifluralin

193 17 23 7 2 2 27

Outdoor 17 29 4 5 51

Indoor 32 43 45 1 208 2 3 2 45 173

Outdoor 73 103

2 30 8 23

Indoor

Outdoor

18

314

151 5 10 20 8 5 46 753

52 6 170 39 9

Indoor

Outdoor

Indoor

34 12

Outdoor 17

6

7

7

6

23 4

9

10 262

1734 15

51

11 21 42 5 5 11 38

110 4

3 21

2 25 3

3 17 5

3 5

12 4 22 6 6 61

276 162 21 3 5 11 8

17

115 431 3 21 14

13 2

138 18 4 13 6

1 20 5

44 3 25 4

57 236 12 3 13 15 9 4 3 90 3 14 233 23 10 16 9

1243 13 81 82 22 3 10

247 27 22

104 25 10

32

114 126 2

23 764

1

5 31 8

60 18 5

6 1 2

14 3

8

13 17 11

850

C. Raeppel et al. / Science of the Total Environment 544 (2016) 845–852

Table 4 Pesticides detected in hair samples. Pesticide

Number of subjects with positive detection (n = 14)a

Frequency of detection [%] (n = 78)

Concentration minb/max [pg mg−1]

Median [pg mg−1]

Standard deviation [pg mg−1]

2,4 D Aldrin Alpha-cypermethrin Bifenthrin pp’-DDE Dieldrin Diflufenican α-Endosulfan β-Endosulfan Flazasulfuron γ-HCH MCPA Mecoprop-P Oxadiazon Pendimethalin Pentachlorophenol Propachlor Propiconazole Trifloxystrobin Trifluralin

3 5 10 7 5 5 14 13 13 3 8 13 4 9 2 9 5 3 4 9

8 9 31 47 26 10 100 97 90 17 53 66 11 58 9 29 15 4 20 58

b10/104 b10 b0.25/0,8 b2.5 b2.5 b0.5 b0.25 b0.25 b0.25 b2.5/14 b5 b10 b2.5/18 b5 b2.5 b2.5/30 b5/57 b10 b0.25 b0.25

32.3 – 0,8 – – – – – – 13.6 – – 13.2 – – 9.9 19.4 – – –

32.0 – – – – – – – – 1.8 – – 2.6 – – 6,0 17.1 – – –

nd: non determined. A.M. Tsatsakis, M.N. Tzatzarakis, D. Koutroulakis, M. Toutoudaki & S. Sifakis 2009. Dialkyl phosphates in meconium as a biomarker of prenatal exposure to organophosphate pesticides: a study on pregnant women of rural areas in Crete, Greece. Xenobiotica, 5, 364–373. D. Koutroulakis, S. Sifakis, M.N. Tzatzarakis, A.K. Alegakis, E. Theodoropoulou, M.P. Kavvalakis, D. Kappou, A.M. Tsatsakis 2014. Dialkyl phosphates in amniotic fluid as a biomarker of foetal exposure to organophosphates in Crete, Greece; association with foetal growth. Reproductive Toxicology 46, 98–105. a Number of subjects for which pesticide has been detected at least once among the different hair samples provided by the subject. b Value = b LOQ/2.

probably due to a lawn treatment with 2,4 D applied at this place. These inhabitants also tested positive for oxadiazon, which can also be explained by lawn treatment applied before the sampling period. The presence of oxadiazon in inhabitants' hair of two others houses was also observed. In one case, although the source of exposure was not

identified, the presence of this compound in all inhabitants of the same house suggests a common source of exposure. γ-HCH was mainly found in hair of inhabitants living in old dwellings, possibly due to its release from treated woods. Houses could therefore be a shared source of exposure. (See Table 4).

Fig. 1. Comparison between detection in air and hair samples.

C. Raeppel et al. / Science of the Total Environment 544 (2016) 845–852

3.3. Comparison between air and hair results Among the pesticides analysed in hair in the current study, 20 were detected in hair and/or in air samples. The comparison between air and hair results pointed out differences in the pesticides detected in each matrix. Hair analyses showed that almost all inhabitants were exposed to diflufenican, α- and β-endosulfan. Diflufenican was however detected in only one air sample and air contamination by α- and β-endosulfan was not measured in all houses. The most frequently detected pesticides in outdoor air samples were not necessarily the same as in hair samples. For air analyses, the highest frequency of detection in outdoor air samples was observed for acetochlor, MCPA, propachlor, trifloxystrobin and trifluralin. Although acetochlor was frequently detected and sometimes at high level of concentration in PAS, it was never detected in hair samples. This pesticide could be degraded in hair or its concentration was below the LOD. Conversely, MCPA and trifluralin were detected in most of the inhabitants' hair. A part of the exposure to the latter pesticides could therefore be linked to the contamination of air. The most frequently detected pesticides in indoor air samples (pentachlorophenol, γ-HCH and α-endosulfan) were identified in more than half of the individuals, suggesting that exposure identified by hair analysis could be partially attributable to air contamination for these molecules. Results of the comparison between air and hair samples for the same sampling periods are presented in Fig. 1. In this figure, a dark blue square shows that all air samples from a home were associated with at least one hair sample, covering the sampling period of air, for each inhabitants of the home. A light blue square means that only some air samples were associated with hair samples (for a same period, an air contamination was observed but hair samples were not positive). Dark green means that all positive hair samples for each inhabitant were associated with at least one air contamination during the exposure period covered by the hair sample. Light green means that only some hair samples were associated with air contamination. Finally, a grey square means that no association were observed between air and hair samples. The comparison between the two matrices for the same sampling periods shows that pesticides detected in hair and in air were not necessarily associated. For example, 2,4-D was detected in indoor air of two houses but not the ones where positive hair samples were identified. Diflufenican, which was detected in hair samples of all the inhabitants, was however only detected in one air sample. Airborne exposure is therefore probably not the main source of exposure and some other potential source of exposure as food consumption can be hypothesised. In some cases, the same compounds were however detected both in air samples and in inhabitants' hair, suggesting that air could be the exposure pathway (e.g. γ-HCH for inhabitants and houses A and B or α-endosulfan except for inhabitants and home E). The lack of association in pesticide detection between the two matrices suggests other sources of exposure than home contamination although insufficient analytical sensitivity cannot be excluded as well as biological degradation (metabolization of parent compounds). 4. Conclusion Exposure profiles varied from one home to another but also between inhabitants of the same home, suggesting that exposure can be different between inhabitants of the same home and highlighting the necessity of personalised biomonitoring. Although association between air and hair samples was not systematic, the use of passive sampling could nevertheless provide information on exposure pathway due to home contamination. Acknowledgements The regional research program R.E.A.LI.SE, the “Région Alsace”, the ERICHE program from CNRS and the ANSES are gratefully acknowledged

851

for their financial support. Caroline RAEPPEL wants to particularly thank the ADEME for their support of a Ph.D grant. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.scitotenv.2015.12.020. References Appenzeller, B.M.R., Tsatsakis, A.M., 2012. Hair analysis for biomonitoring of environmental and occupational exposure to organic pollutants: state of the art, critical review and future needs. Toxicol. Lett. 210, 119–140. Atmos'Fair Bourgogne, 2006. Campagne de mesures de pesticides en air extérieur et intérieur. Une problématique de santé publique. (34 pp. http://www.atmosfairbourgogne.org/medias/fichiers_telechargement/campagne_de_mesures_de_ pesticides_en_air_exterieur.pdf). Becker, K., Seiwert, M., Casteleyn, J.R., Joas, A., Biot, P., Aerts, D., Castaño, A., Esteban, M., Angerer, J., Koch, H., Schoeters, G., Den Hond, E., Sepai, O., Exley, K., Knudsen, L., Horvat, M., Bloemen, L., Kolossa-Gehring, M., 2014. DEMOCOPHES consortium. A systematic approach for designing a HBM pilot study for Europe. Int. J. Hyg. Environ. Health 214, 312–322. Bohlin, P., Jones, K.C., Tovalin, H., Strandberg, B., 2008. Observations on persitent organic pollutants in indoor and outdoor air using passive polyurethane foam samplers. Atmos. Environ. 42, 7234–7241. Bouvier, G., Blanchard, O., Momas, I., Seta, N., 2006. Pesticide exposure of nonoccupationally exposed subjects compared to some occupational exposure: a French pilot study. Sci. Total Environ. 366, 74–91. Cattani, M., Cena, K., Edwards, J., Pisaniello, D., 2001. Potential dermal and inhalation exposure to chlorpyrifos in Australian pesticide workers. Ann. Occup. Hyg. 45, 299–308. Cirimele, V., Kintz, P., Ludes, B., 1999. Evidence of pesticide exposure by hair analysis. Acta Clin. Belg. 54, 59–63. Covaci, A., Hura, C., Gheorghe, A., Neels, H., Dirtu, A.C., 2008. Organochlorine contaminants in hair of adolescents from Iassy, Romania. Chemosphere 72, 16–20. Delhomme, O., Raeppel, C., Briand, O., Millet, M., 2010. Analytical method for assessing potential dermal exposure to pesticides of a non-agricultural occupationally exposed population. Anal. Bioanal. Chem. 399, 1325–1334. Gorecki, T., Namiesnik, J., 2002. Passive sampling. Trends Anal. Chem. 21, 276–291. Gurunathan, S., Robson, M., Freeman, N., Buckley, B., Roy, A., Meyer, R., Bukowski, J., Lioy, P.J., 1998. Accumulation of chlorpyrifos on residential surfaces and toys accessible to children. Environ. Health Perspect. 106, 9–16. Harner, T., Pozo, K., Gouin, T., Macdonald, A.-M., Hung, H., Cainey, J., Peters, A., 2006. Global pilot study for persistent organic pollutants (POPs) using PUF disk passive air samplers. Environ. Pollut. 144, 445–452. Harrad, S., Goosey, E., Desborough, J., Abdallah, M.A.-E., Roosens, L., Covaci, A., 2010. Dust from U.K. primary school classrooms and day care centers: the significance of dust as a pathway of exposure of young U.K. children to brominated flame retardants and polychlorinated biphenyls. Environ. Sci. Technol. 44, 4198–4202. Harris, S.A., Villeneuve, P.J., Crawley, C.D., Mays, J.E., Yeary, R.A., Hurto, K.A., Meeker, J.D., 2010. National study of exposure to pesticides among professional applicators: an investigation based on urinary biomarkers. J. Agric. Food Chem. 58, 10253–10261. Lu, C., Fenske, R.A., 1998. Air and surface chlorpyrifos residues following residential broadcast and aerosol pesticide applications. Environ. Sci. Technol. 32, 1386–1390. Marlière, F., 2009. Observation des niveaux de concentration en pesticides dans l'air ambiant. LCSQA, INERIS, France. Morgan, M.K., Stout, D.M., Jones, P.A., Barr, D.B., 2008. An observational study of the potential for human exposures to pet-borne diazinon residues following lawn applications. Environ. Res. 07, 336–342. Neuber, K., Merkel, G., Randow, F.F.E., 1999. Indoor air pollution by lindane and DDT indicated by head hair samples of children. Toxicol. Lett. 107, 189–192. Ostrea, E.M., Villanueva-Uy, E., Bielawski, D.M., Posecion, N.C., Corrion, M.L., Jin, Y., Janisse, J.J., Ager, J.W., 2006. Maternal hair — an appropriate matrix for detecting maternal exposure to pesticides during pregnancy. Environ. Res. 101, 312–322. Peck, A.M., Hornbuckle, K.C., 2005. Gas-phase concentrations of current-use pesticides in Iowa. Environ. Sci. Technol. 39, 2952. Raeppel, C., 2012. Ph.D thesis, University of Strasbourg, 204 pp. Salquèbre, G., Schummer, C., Millet, M., Briand, O., Appenzeller, B.M.R., 2012. Multi-class pesticide analysis in humain hair by gas chromatography tandem (triple quadrupole) mass spectrometry with solid phase microextraction and liquid injection. Anal. Chim. Acta 710, 65–74. Scheyer, A., Morville, S., Mirabel, P., Millet, M., 2007. Variability of atmospheric pesticide concentrations between urban and rural areas during intensive pesticide application. Atmos. Environ. 41, 3604–3618. Schummer, C., Mothiron, E., Appenzeller, B.M.R., Rizet, A.-L., Wennig, R., Millet, M., 2010. Temporal variations of concentrations of currently used pesticides in the atmosphere of Strasbourg, France. Environ. Pollut. 158, 576–584. Schummer, C., Tuduri, L., Briand, O., Appenzeller, B.M.R., Millet, M., 2012a. Application of XAD-2 resin-based passive samplers and SPME–GC–MS/MS analysis for the monitoring of spatial and temporal variations of atmospheric pesticides in Luxembourg. Environ. Pollut. 170, 88–94. Schummer, C., Salquèbre, G., Briand, O., Millet, M., Appenzeller, B.M.R., 2012b. Determination of farm workers' exposure to pesticides by hair analysis. Toxicol. Lett. 210, 203–210.

852

C. Raeppel et al. / Science of the Total Environment 544 (2016) 845–852

Shen, L., Wania, F., Lei, Y.D., Teixeira, C., Muir, D.C.G., Bidleman, T.F., 2005. Atmospheric distribution and long-range transport behavior of organochlorine pesticides in North America. Environ. Sci. Technol. 39, 409–420. Stout II, D.M., Mason, M.A., 2003. The distribution of chlorpyrifos following a crack and crevice type application in the US EPA Indoor Air Quality Research House. Atmos. Environ. 37, 5539–5549. Tsatsakis, A.M., Tzatzarakis, M.N., Tutudaki, M., Babatsikou, F., Alegakis, A.K., Koutis, C., 2008. Assessment of levels of organochlorine pesticides and their metabolites in the hair of a Greek rural human population. Hum. Exp. Toxicol. 27, 933–940. Tulve, N.S., Egeghy, P.P., Fortmann, R.C., Whitaker, D.A., Nishioka, M.G., Naeher, L.P., Hilliard, A., 2007. Multimedia measurements and activity patterns in an observational pilot study of nine young children. J. Expo. Sci. Environ. Epidemiol. 18, 31–44. Tzatzarakis, M.N., Barbounis, E.G., Kavvalakis, M.P., Vakonaki, E., Renieri, E., Vardavas, A.I., Tsatsakis, A.M., 2014. Rapid method for the simultaneous determination of DDTs and PCBs in hair of children by headspace solid phase microextraction and gas chromatography–mass spectrometry (HSSPME/GC-MS). Drug Test. Anal. 6 (Suppl. 1), 85–92. Van Den Eede, N., Dirtu, A.C., Neels, H., Covaci, A., 2011. Analytical developments and preliminary assessment of human exposure to organophosphate flame retardants from indoor dust. Environ. Int. 37, 454–461.

Van Dijk, H.F.G., Guicherit, R., 1999. Atmospheric dispersion of current-use pesticides: a review of the evidence from monitoring studies. Water Air Soil Pollut. 115, 21–70. Wania, F., Shen, L., Lei, Y.D., Teixeira, C., Muir, D.C.G., 2003. Development and calibration of a resin-based passive sampling system for monitoring persistent organic pollutants in the atmosphere. Environ. Sci. Technol. 37, 1352–1359. Williams, M.K., Rundle, A., Holmes, D., Reyes, M., Hoepner, L.A., Barr, D.B., Camann, D.E., Perera, F.P., Whyatt, R.M., 2008. Changes in pest infestation levels, self-reported pesticide use, and permethrin exposure during pregnancy after the 2000–2001 U.S. Environmental Protection Agency Restriction of Organophosphates. Environ. Health Perspect. 116, 1681–1688. Wilson, N.K., Chuang, J.C., Morgan, M.K., Lordo, R.A., Sheldon, L.S., 2007. An observational study of the potential exposures of preschool children to pentachlorophenol, bisphenol-A, and nonylphenol at home and daycare. Environ. Res. 103, 9–20. Zhang, H., Chai, Z., Sun, H., 2007. Human hair as a potential biomonitor for assessing persistent organic pollutants. Environ. Int. 33, 685–693.