Subscriber access provided by READING UNIV
Article
Occurrence of organophosphorus flame retardants and plasticizers (PFRs) in Belgian foodstuff and estimation of the dietary exposure of the adult population Giulia Poma, Carlos Sales, Bram Bruyland, Christina Christia, Severine Goscinny, Joris Van Loco, and Adrian Covaci Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b06395 • Publication Date (Web): 27 Jan 2018 Downloaded from http://pubs.acs.org on January 29, 2018
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 24
Environmental Science & Technology
1
Occurrence of organophosphorus flame retardants
2
and plasticizers (PFRs) in Belgian foodstuff and
3
estimation of the dietary exposure of the adult
4
population
5
Giulia Poma1*, Carlos Sales2, Bram Bruyland1, Christina Christia1, Séverine Goscinny3, Joris
6
Van Loco3, Adrian Covaci1*
7
1
Toxicological Center, University of Antwerp, Universiteitsplein 1, 2610 Wilrijk, Belgium
8
2
Research Institute for Pesticides and Water, University Jaume I, E-12071 Castellón, Spain
9
3
Food, Medicines and Consumer Safety, The Scientific Institute of Public Health, Juliette
10
Wytsmanstraat 14, 1050 Brussels, Belgium
11 12
*
corresponding author:
[email protected] 13
ACS Paragon Plus Environment
1
Environmental Science & Technology
Page 2 of 24
14
KEYWORDS
15
Organophosphorus flame retardants and plasticizers; food analysis; human dietary exposure; GC-
16
MS/MS; processed food
17
ABSTRACT
18
The occurrence of 14 organophosphorus flame retardants and plasticizers (PFRs) was
19
investigated in 165 composite food samples purchased from the Belgian market and divided in
20
14 food categories, including fish, crustaceans, mussels, meat, milk, cheese, dessert, food for
21
infants, fats and oils, grains, eggs, potatoes and derived products, other food (stocks) and
22
vegetables. Seven PFRs [namely, tri-n-butyl phosphate (TnBP), tris (2-chloroethyl) phosphate
23
(TCEP), tris (1-chloro-2-propyl) phosphate (TCIPP), tris(1,3-dichloro-2-propyl) phosphate
24
(TDCIPP), triphenyl phosphate (TPHP), 2-ethylhexyl diphenyl phosphate (EHDPHP) and tris (2-
25
ethylhexyl) phosphate (TEHP)] were detected at concentrations above quantification limits. Fats
26
and oils were the most contaminated category, with a total PFR concentration of 84.4 ng/g ww,
27
followed by grains (36.9 ng/g ww) and cheese (20.1 ng/g ww). Our results support the
28
hypothesis that PFR contamination may occur during industrial processing and manipulation of
29
food products (e.g. packaging, canning, drying, etc.). Considering the daily average intake of
30
food for the modal adult Belgian (15-64 y), the dietary exposure to sum PFRs was estimated up
31
to 7,500 ± 1,550 ng/day (103 ± 21 ng/kg bw/day). For individual PFRs, TPHP contributed on
32
average with 3,400 ng/day (46.6 ng/kg bw/day), TCIPP with 1,350 ng/day (18.5 ng/kg bw/day),
33
and EHDPHP with 1,090 ng/day (15 ng/kg bw/day), lower than their corresponding health-based
34
reference doses. The mean dietary exposure mainly originated from grains (39 %), followed by
35
fats and oils (21 %) and dairy products (20 %). No significant differences between the intakes of
36
adult men and women were observed.
ACS Paragon Plus Environment
2
Page 3 of 24
Environmental Science & Technology
37
1. Introduction
38
The phasing out of brominated flame retardants (BFRs) due to toxicity concerns1 and the
39
enlistment on the POP list of polybrominated diphenyl ethers (PBDEs) in 2009 and
40
hexabromocyclododecane (HBCDD) in 2013 has led to a surge in the usage of alternative flame
41
retardants2. Organophosphorus flame retardants and plasticizers (PFRs) have been implemented
42
as the foremost substitute of BFRs and have known a substantial growth in production volumes
43
over the last 15 years3. They are widely used in standard and engineering plastics, polyurethane
44
foams (PUF), thermoset coatings, textiles and polymer mixtures4. PFRs can be found inside the
45
metal and plastic walls of refrigerators and freezers, and are used as thermal insulation panels in
46
the construction sector or as casing for electronic equipment2, floor coverings, foam seating and
47
bedding products, and as food-wrapping film5.
48
PFRs are used as additive substances and are not chemically bound to the polymers they are
49
applied to4. This allows them to leak into the environment, mainly through volatilization and
50
abrasion5. Because of the structural differences among PFRs, leading to a variety of
51
physicochemical properties within this class of compounds, they have been already found in all
52
environmental compartments (like air6, water7, sediments and soil8,9, dust10,11, biota4,12,13) and
53
humans11,14. However, data on the environmental persistence and toxicity of PFRs are scarce and
54
few reports on adverse effects have been published2. For instance, tris (1-chloro-2-propyl)
55
phosphate (TCIPP) and tris(1,3-dichloro-2-propyl) phosphate (TDCIPP) are known carcinogens
56
in rats15; TDCIPP has also shown endocrine disrupting properties, altering the concentrations of
57
thyroid hormone thyroxine (T4) and triiodothyronine (T3) in zebrafish3 and decreasing human
58
semen quality16 and mobility17. Tris (2-chloroethyl) phosphate (TCEP), tri-n-butyl phosphate
59
(TnBP) and triphenyl phosphate (TPHP) have neurotoxic aspects2, while 2-ethylhexyl-diphenyl
ACS Paragon Plus Environment
3
Environmental Science & Technology
Page 4 of 24
60
phosphate (EHDPHP) is considered highly toxic to fish and aquatic plants and has potential to
61
bioaccumulate13. Because of the frequent occurrence of PFRs in the environment, and their
62
possible accumulation in biota samples, humans can be exposed to PFRs via different exposure
63
routes2: mainly ingestion (and inhalation) of indoor dust18 and ingestion of contaminated food19–
64
21
65
the presence of PFRs in food is due to migration from food packaging, bioaccumulation, uptake
66
from the agricultural environment, contamination during industrial processing or some other
67
sources is still not clear19,20. This leads to uncertain and insufficient information to estimate the
68
risk of human exposure to PFRs through diet.
69
In the present study, the concentrations of 14 PFRs (namely, TCIPP, TDCIPP, TCEP, TnBP,
70
TPHP, EHDPHP, tri-n-propyl phosphate (TnPP), tris(2-ethylhexyl) phosphate (TEHP), tritolyl
71
phosphates (TOTP, TMTP, TPTP), tris(2-isopropylphenyl) phosphate (T2IPPP), tris(3,5-
72
dimethylphenyl) phosphate (T35DMPP), tris(p-t-butylphenyl) phosphate (TBPP)) were
73
investigated in 165 food samples belonging to 14 different food groups (namely, fish,
74
crustaceans, mussels, meat, milk, cheese, dessert, food for infants, fats and oils, grains, eggs,
75
potatoes and derived products, other food (stocks) and vegetables), all purchased from the
76
Belgian food market. The dietary exposure to PFRs of the Belgian adult population was then
77
estimated based on consumption data from the most recent Belgian National Food Consumption
78
Survey22.
. However, there are few studies on the occurrence of PFRs in food and information on whether
79 80
2. Experimental section
81
2.1 Sample collection
ACS Paragon Plus Environment
4
Page 5 of 24
Environmental Science & Technology
82
The food samples were available as previously collected and prepared in the framework of the
83
FLAREFOOD project on the monitoring of the presence of BFRs in foodstuff23,24. In brief,
84
individual food items were purchased from supermarkets, discount retailers and specialized meat
85
and fish stores in main Belgian cities (Brussels and Antwerp) and in one village (Dessel) in
86
April-December 2015 and May-November 2016. The individual samples were ground and
87
homogenized, and equal amounts (±75 g) of the food items belonging to the same typology were
88
pooled to create a unique composite sample. The composite samples were then either freeze-
89
dried, analyzed fresh (samples with high fat content, e.g. included in “fats and oils” and “other
90
food” categories), or directly aliquoted for analysis (such as grains, potatoes, and milk powder
91
formulas for infants). A total of 165 samples, belonging to 14 food categories (overview in
92
Supporting Information, Table S1), were prepared for analysis. Pending analysis, the samples
93
were stored at -20°C in Falcon tubes (50 mL) previously tested to be free from the targeted
94
PFRs.
95 96
2.2 Analytical procedures
97
A schematic overview of the extraction procedure is given in Figure S1. About 300 mg of freeze-
98
dried sample was weighed in a 15 mL polypropylene (PP) tube with 100 mg NaCl. For high-fatty
99
matrices (>20 % lip), a lower amount (variable from 50-150 mg) of sample was taken to reduce
100
possible matrix effects. The lipid content of each sample was gravimetrically determined as
101
previously reported21 and is shown in Table S1. The sample was added with 5 mL of acetonitrile
102
(ACN) and 50 µL of internal standard (IS) mixture (containing TPHP-d15, TDCIPP-d15, TCEP-
103
d12 and TAP), vortexed 1 min, and centrifuged at 3500 rpm for 5 min. The supernatant was then
104
transferred to a pre-cleaned glass tube and evaporated to 2 mL under gentle nitrogen stream. A
ACS Paragon Plus Environment
5
Environmental Science & Technology
Page 6 of 24
105
dispersive-SPE (d-SPE) was performed by adding to the sample 100 mg C18 and 50 mg of
106
primary-secondary amine (PSA) sorbent powder to remove fats, sugars, fatty acids and most
107
polar pigments from the extract. After 1 min vortex and 5 min centrifugation at 3500 rpm, the
108
supernatant was transferred to a clean glass tube and evaporated to dryness. The sample was then
109
reconstituted in 0.5 mL hexane (n-Hex) and further cleaned-up by elution onto Florisil® cartridge
110
pre-conditioned with 6 mL ethyl acetate (ETAC) and 6 mL n-Hex. Fractionation was achieved
111
with 12 mL n-Hex:DCM (4:1, v/v) (F1, discarded) and 10 mL ETAC (F2, containing the target
112
compounds). F2 was evaporated to near-dryness, reconstituted in 0.5 mL n-Hex and further
113
cleaned-up by elution onto aminopropyl silica (APS) cartridge pre-conditioned with 6 mL n-
114
Hex:DCM (1:1, v/v) and 6 mL n-Hex. A second fractionation was achieved with 10 mL n-Hex
115
(F3, discarded) and 12 mL n-Hex:DCM (1:1, v/v) (F4, containing the target compounds). F4 was
116
evaporated, resolubilized in 50 µL iso-octane:ETAC (4:1, v/v) and 50 µL of recovery standard
117
(RS, CB-207), transferred to injection vials for GC-MS/MS analysis. Full information on the
118
chemicals and materials used in the present study is reported in Supporting Information, Section
119
S1.
120
The chromatographic analysis was performed using an Agilent 7890B gas chromatograph,
121
equipped with an Agilent 7693A autosampler with multimode inlet (MMI), coupled to a triple
122
quadrupole mass spectrometer, 7000C (Agilent Technologies Inc., Palo Alto, CA, USA), with a
123
EI source working in electron impact mode. The GC separation was performed using a Zebron
124
Semivolatile column (5% phenyl-arylene-95% dimethyl-polysiloxane) with a length of 20 m ×
125
0.18 mm ID and a film thickness of 0.18 µm (Phenomenex, Torrance, CA, USA) working at a
126
flow of 1.2 mL/min of helium (99.999%; Air Liquide, Liège, Belgium). The oven program was
127
set as follows: 90 °C (1.25 min); 10 °C/min to 220 °C; then 5 °C/min to 300 °C; then 40 °C/min
ACS Paragon Plus Environment
6
Page 7 of 24
Environmental Science & Technology
128
to 325 °C. The injection of 1 µL of sample extracts was performed in splitless mode at a
129
temperature of 110 °C. The pulse pressure was set to 50 psi, with a split purge flow of 50
130
mL/min and purge time of 1 min. TAP (IS) corrected for TnPP, TnBP and TEHP; TCEP-d12
131
corrected for TCEP and TCIPP; TDCIPP-d15 corrected for TDCIPP and TDBPP; TPHP-d15
132
corrected TPHP, EHDPHP, TOTP, TMTP, TPTP, T2IPPP and T35DMPP. The full list of
133
targeted PFRs and MS/MS quantitation parameters for their analysis is reported in Table S2.
134 135
2.3 Quality control
136
The quality control check was performed using the menhaden fish oil used in the first worldwide
137
interlaboratory study (ILS) on PFRs25. The measured values were within the range of the
138
reference concentration (± 15%), while the mean IS recoveries ranged between 74 and 130%,
139
with a relative standard deviation (RSD) below 15%. As blank contamination is the biggest
140
challenge in the analysis of PFRs, due to their ubiquitous presence in dust25, three procedural
141
blanks were run in parallel with every batch of samples (n = 21) to control any potential
142
background contamination. To limit the levels of PFRs in the blank samples, the working space
143
and all laboratory equipment was repeatedly cleaned with acetone; SPE and evaporation was
144
performed in a pre-cleaned fume hood; glass tubes were washed, pre-cleaned with n-hexane and
145
left in oven (at 300°C) for 2.5 hours to remove contaminant traces. Even following these
146
precautionary steps along the whole sample procedure, mean PFR values in the blank were found
147
at 1.3 ± 2.7 ng/g (range 1.6 pg/g to 21.2 ng/g). Average blank levels per batch were then
148
subtracted from the sample results and a value equal to 3*SD of the blank measurement was used
149
as limits of quantification (LOQs) before reporting results. For compounds not present in the
ACS Paragon Plus Environment
7
Environmental Science & Technology
Page 8 of 24
150
blanks, LOQs were based on a signal/noise ratio of 10. Due to the different weighed amount of
151
the analyzed samples, LOQs differed per matrix and per compounds and are listed in Table S3.
152 153
2.4 Dietary exposure estimation
154
Consumption data from the most recent Belgian National Food Consumption Survey (data from
155
2014) were used for the exposure assessment22. The target population included all Belgian
156
inhabitants from 3 to 64 year-old (divided in age categories), for a total of 3,200 participants,
157
randomly selected from the National Register by a multi-stage stratified procedure. The diet of
158
the participants was assessed by performing two non-consecutive 24-hour recalls (describing the
159
food and drinks consumed the day before the interview) using the EPIC-Soft program26,
160
developed to conduct standardized interactive 24-hour recall interviews or to enter dietary
161
records in a standardized way. In the present study, it was chosen to estimate the daily intake of
162
PFRs for the adult population, including adolescent (age group: 15-64 years). The individual
163
intake of PFRs from the Belgian adult population (ng/day) was estimated for each compound by
164
multiplying the average concentration of that compound in a particular food group (ng/g ww)
165
with the average consumption rate of that particular group (g/day) by the Belgian adult
166
population (Table S4). The total dietary intake of PFRs was obtained by summing the respective
167
intakes from each food category. The daily dietary intake was also calculated per body weight
168
(ng/kg body weight/day), based on an average body weight of 73 kg22. Since no consumption
169
data for age groups younger than 3 years old were available, no calculations concerning intake of
170
baby food products (mostly composed by food items for infants younger than 3 years old, such
171
as milk powder and baby meals) were provided in the present study.
172
ACS Paragon Plus Environment
8
Page 9 of 24
Environmental Science & Technology
173
2.5 Statistics
174
For data treatment and elaboration, statistical software IBM SPSS Statistics 24 was used. When
175
calculating the mean and median concentrations of PFRs in food composite samples, the non-
176
detects were substituted with half the quantification limits ( TCIPP (24 %) > EHDPHP (20 %) > TDCIPP (14 %) > TEHP and TnBP (6 %) > TCEP
197
(3 %) (Figure S2). However, the distribution of PFRs highly differed among the analyzed food
198
groups when they were considered individually (Fig. 1). TCIPP was the most representative
199
compound in fats and oils (FAT-OC), accounting for 46 % of the total measured concentration.
200
EHDPHP was mainly found in stock/other food (OTC = 12.5 ng/g ww, 66 %), dessert (DC = 7.2
201
ng/g ww, 63 %), food for infants (BC = 4.1 ng/g ww, 34 %) and potatoes (POC = 2.2 ng/g ww,
202
31 %), while TPHP was mainly found in grains (GRAC = 26.1 ng/g ww, 71 %), cheese (CHC =
203
10.9 ng/g ww, 54 %), mussels (MC = 6.3 ng/g ww, 80 %), and vegetables (VEC = 2.0 ng/g ww,
204
92 %). The remaining food groups were characterized by a more homogeneous distribution of
205
PFRs. The migration of EHDPHP from food packaging materials can be a possible source of
206
EHDPHP in the analyzed foodstuff, since it was approved by the US Food and Drug
207
Administration (US-FDA) for such application27. The presence of TPHP, used as flame retardant
208
and plasticizer, can derive from its application in electrical industrial equipment2 and its usually
209
high concentrations in indoor dust (which can reasonably deposit in industrial facilities)28–30, but
210
the source of the presence of the other PFRs remains unclear.
211
Considering all food categories, the most frequently contaminated groups were fats and oils
212
(FAT-OC = 40 %), grains (GRAC = 17 %), cheese and stock/other food (CHC and OTC = 9 %),
213
baby food (BC = 6 %) and dessert (DC = 5 %) (Figure S3). The main food groups from animal
214
origin, such as fish, crustaceans, mussels, meat, and eggs, contributed only marginally to the
215
overall PFR distribution (1 %, 1 %, 4 %, 2 %, and 0.3%, respectively) and this might be due to
216
the rapid metabolism/excretion of PFRs17,31,32. A similar trend was observed in our previous
217
study, where we suggested that the higher contamination with PFRs of certain food groups might
ACS Paragon Plus Environment
10
Page 11 of 24
Environmental Science & Technology
218
derive from industrial processing19. To further explore this hypothesis, the whole sample data-set
219
was divided in “non-processed food” and “processed food”. In the first group, we included all
220
foodstuff that had been only minimally altered from its natural state, for example by being
221
frozen, freshly cut, or directly packed (namely, vegetables, eggs, some fish, meat and milk
222
samples). Food items included in the second group were instead more heavily manipulated and
223
industrially altered, such as being canned, smoked, dried, fried, minced, etc. (namely, baby food,
224
stock, desserts, cheese, fats and oils, grains and potatoes, and the other samples of fish, meat and
225
milk). The full list of samples belonging to both categories is reported in Table S1. The total
226
distribution of PFRs in the two groups showed a clear predominance of PFRs in processed food
227
(89 %) vs non-processed food (11 %), with a similar and homogeneous contribution of each PFR
228
to the total (Figure 2a). The same elaboration was then carried out considering separately meat
229
(MEC), fish (including FC, CC, MC) and milk (LC), having both processed (e.g. prepared meat
230
or fish, minced meat, canned meat or fish, marinated fish, fried fish, cream, milk based drinks,
231
etc.) and non-processed (fresh/frozen meat and fish, fresh milk) samples. The results of the
232
distribution of PFRs in processed- and non-processed-meat, fish, and milk (Figure 2b) are very
233
similar among the three food groups and concordant with the general pattern (Figure 2a). These
234
outcomes support the hypothesis that PFR contamination of foodstuff might likely originate from
235
industrial processing and manipulation/alteration of food products.
236 237
3.2 Estimation of human dietary exposure to PFRs
238
The estimation of dietary exposure of the Belgian adult population to PFRs was based on the
239
measured MB concentrations of targeted compounds in 165 food samples belonging to food
240
groups representative of the general population food habit22. The category “baby food” (BC), not
ACS Paragon Plus Environment
11
Environmental Science & Technology
Page 12 of 24
241
consumed by the general adult population, was not considered in the elaborations. The intake of
242
fruits, beverages (e.g. water, soft-drinks, tea/coffee, and alcoholic drinks), and candy/sweets was
243
not included in this study since these food categories were not purchased in the frame of this
244
project.
245
Mean daily dietary intakes of PFRs were estimated for the overall Belgian adult population
246
(Table 2) as well as for men and women separately (Table S5). According to the consumption
247
data from the National Food Consumption Survey22, men generally consume more meat, eggs,
248
fish, grains and fats than women, while the consumption of the other food categories is similar
249
between the two genders (Table S4).
250
For the average adult population (15-64 years), the total dietary intake of PFRs was estimated at
251
7,500 ± 1,550 ng/day (103 ± 21 ng/kg bw/day) (Table 2) and no significant differences between
252
the PFR intakes of men and women were observed. These results were higher than those
253
estimated by Xu et al. (2017) by using a duplicate diet approach (i.e. each participant provided
254
diet samples by collecting duplicate portions of their meals over a 24-hour sampling period),
255
where the median dietary intake of the sum PFRs was estimated at 87 ng/kg bw/day20. However,
256
it is necessary to point out that this kind of approach might not be completely representative of
257
habitual diet and dilution effects arising from the typology of composite samples might reduce
258
the levels of contamination in food20.
259
The mean daily dietary intake of PFRs estimated in our study was similar to the one estimated in
260
our previous study for the Swedish adult population (5,700 ng/day and 85 ng/kg bw/day)19,
261
although the contribution of each PFR to the total intake is slightly different in the two studies.
262
This can be due to differences in the composition and preparation of each food group and the
263
number/typology of samples included in each category. Here, TPHP accounted for 45 % of the
ACS Paragon Plus Environment
12
Page 13 of 24
Environmental Science & Technology
264
total dietary intake of PFRs, followed by TCIPP (18 %), EHDPHP (15 %), TDCIPP (9 %),
265
TEHP and TnBP (5 %), and TCEP (3 %). In our previous study, the major contributors to the
266
total intake were in the order: EHDPHP (57%) > TDCIPP (14%) > TPHP (11%) > TCIPP (10%)
267
> TCEP (7%). On the other hand, the relative contribution of each food group to the dietary
268
intake of PFRs is very similar in both studies (considering the food groups analyzed in both
269
studies), and in the order: grains (39 and 36 % in this study and in Poma et al., 201719,
270
respectively), fats and oils (21 and 14 %), dairy products and desserts (20 and 23 %), meat (6 and
271
7 %), vegetables (6 % in both studies), fish and potatoes (4 and 3 %). More precisely, spaghetti
272
(GRAC-07) and stick bread (GRAC-04) contributed here for more than 20 % each to the PFR
273
dietary intake attributed to the grain category, while fish-oil food supplement (OTC-05)
274
contributed alone with more than 80% to the PFR dietary intake attributed to fats and oils.
275
It is likely, that the daily intake of PFRs through fat and oil consumption by the general adult
276
population might have been overestimated, due to the inclusion of the fish-oil supplement in the
277
FAT-OC food group, for which a total consumption of 19 g/day was estimated22. By considering
278
a more realistic daily intake for commercial fish-oil food supplement (e.g. 1 g/day)22, the dietary
279
intake of total PFRs of OTC-05 was thus estimated at 225 ng/day (3 ng/kg bw/day) and of FAT-
280
OC at 1200 ng/day (16 ng/kg bw/day) (not shown). In addition, the relatively high dietary intake
281
of PFRs of the FAT-OC category could have also been influenced by substituting the