Article pubs.acs.org/JAFC
Investigation into the Occurrence in Food of Veterinary Medicines, Pharmaceuticals, and Chemicals Used in Personal Care Products Richard J. Fussell,*,† Monica Garcia Lopez,† David N. Mortimer,§ Stuart Wright,† Monika Sehnalova,† Chris J. Sinclair,† Alwyn Fernandes,† and Matthew Sharman† †
Food and Environment Research Agency, Sand Hutton, York YO41 1LZ, United Kingdom Food Standards Agency, Aviation House, 125 Kingsway, London WC2B 6NH, United Kingdom
§
ABSTRACT: Human exposure to emerging contaminants by indirect routes is of increasing interest. This study assessed the contamination of food by chemicals used in human pharmaceuticals (HPs), veterinary medicines (VMs), and personal care products (PCPs). A prioritization study was undertaken to identify the chemicals and food-producing scenarios most likely to result in contamination of food. Around 400 samples of mushrooms, vegetables, aquaculture products, and animal tissues were collected from sites in the United Kingdom, along with aquaculture products imported from Southeast Asia. A number of multianalyte methods were developed and validated for the analysis of the prioritized compounds in these samples. The analysis of all sample−method combinations required approximately 18000 determinations. Around 325 individual residues, including parabens, musk compounds, and antibiotics, were detected in 118 individual samples, but mostly at low nanograms per gram concentrations. Results suggest that the limited contamination of target chemicals occurred in the realistic food-producing scenarios investigated. KEYWORDS: emerging contaminants, human pharmaceuticals, veterinary medicines, personal care products, food, uptake, LC-MS/MS, GC-HRMS
■
sulfonamides,5,7 levamisole,4 trimethoprim,4 florfenicol,4 diazinon,4 tylosin,7 and virginiamycin7 in vegetables. Less information is available about “real” scenarios,3,6 where both the concentrations and bioavailability of chemicals incurred into manure after the treatment of animals could be different. The objective of this study was to assess the potential for the contamination of food by HPs, VMs, and chemicals used in PCPs in real food production scenarios. First, a study of the currently available literature was undertaken to identify and prioritize the scenarios and the HPs, VMs, and PCPs most likely to contaminate food. The prioritization considered many factors including usage (except for PCPs), persistence, potential for uptake and bioaccumulation, results from previous prioritization exercises, and reported environmental occurrence.8,9 Suitable multianalyte/multiclass methods were developed and validated for the determination of 36 different VM and HP compounds. Separate chemical-class specific methods were developed for the determination of aminoglycosides, coccidiostats, parabens, and musks. Compared to the other compound classes, assessments of the uptake of parabens and musks in foods tend to be less frequent, even though they are known to occur in the environment.10−13 Parabens are widely used as bactericides and antimicrobial agents in PCP formulations and are approved for use as food preservatives in the European Union (EU).14 Polycyclic and nitro musk
INTRODUCTION Human exposure to “emerging contaminants” by indirect routes is becoming an ever-increasingly important issue. It is now understood that some groups of compounds previously not considered as a risk may enter the environment and, subsequently, the food chain by various pathways during their production, usage, or disposal. After use, human pharmaceuticals (HPs) and personal care products (PCPs) are eventually discharged into the sewage system and following treatment may be released into surface waters through the effluent from sewage treatment works.1,2 Increasing use and incomplete removal through conventional water treatments may contribute to the ubiquity of these chemicals in the environment. These classes of compounds can then enter agricultural soils through irrigation with contaminated surface water or through the application of biosolids containing HPs and PCPs. Similarly, the principal route for veterinary medicines (VMs) to enter agricultural soils is after the application of manure from farm animals.3 Theoretically, there is the potential for growing crops to take up compounds that are present in the contaminated soil. Another possibility for consideration is the direct uptake of VMs by mushrooms cultivated using compost based on poultry litter. If agricultural crops grown for food or animal fodder can take up chemicals from the environment, then it follows that there is the potential for human exposure via consumption of these food crops or meat from animals exposed to contaminated feed and/or water. The uptake of chemicals into crops cultivated in spiked soils3−6 or soils amended with manure mixed with antibiotics7 has been reported. The resultant concentrations of veterinary medicines in the growing substrate were relatively high. Uptake has been reported for coccidiostats,3,6,7 fluoroquinolones,4,5,7 Published 2014 by the American Chemical Society
Special Issue: 50th North American Chemical Residue Workshop Received: Revised: Accepted: Published: 3651
November 22, 2013 January 30, 2014 February 4, 2014 February 4, 2014 dx.doi.org/10.1021/jf4052418 | J. Agric. Food Chem. 2014, 62, 3651−3659
Journal of Agricultural and Food Chemistry
Article
Table 1. Overview of Compounds Analyzed by LC-MS/MSa methods and compounds
cone voltage (V)
LC-MS/MS Multiresidue (ES+) (a) 2-aminoflubendazole (ES) (a) cimetidine (ES) (a) ciprofloxacin (IS) (a) codeine (ES) (a) danofloxacin (IS) (a) decoquinate (ES) (a) diclofenac (IS) (a) difloxacin (IS) (a) dipyridamole (ES) (a) doxycycline (IS) (a) enrofloxacin (IS) (a) erythromycin (IS) (a) flubendazole (IS) (a) gliclazide (ES) (a) irbestan (ES) (a) lasalocid (ES) (a) lincomycin (IS) (a) maduramycin (ES) (a) mebeverine (ES) (a) mefenamic acid (ES) (a) sulfasalazine (ES) (a) sulfadiazine (IS) (a) tilmicosin (ES) (a) yrimethoprim (ES) (b) cvefalexin (ES) (b) chlorhexidine (ES) (b) chlortetracycline (IS) (b) dicyclanil (ES) (b) florfenicol amine (ES) (b) flucloxacillin (ES) (b) quinine (ES) (b) robenidine hydrochloride (ES) (b) salinomycin sodium (ES) (b) tetracycline (IS) (b) toltrazuril sulfoxide (ES) (b) tramadol (ES)
Aminoglycosides and Related Compounds (ESI+) (c) apramycin (ES) (c) dihydrostreptomycin (ES) (c) gentamicin C1 (ES) (c) gentamicin C1a (ES) (c) gentamicin C2 + C2a (ES) (c) kanamycin A (ES) (c) neomycin B (ES) (c) paromomycin (ES) (c) spectinomycin (ES) (c) streptomycin (ES)
Coccidiostats (ESI+) (a) diclazuril (IS) (a) dinitrocarbanilide (IS) (a) lasalocid (ES) (a) maduramycin (ES) (a) monensin (ES) (a) narasin (ES) (a) salinomycin (ES)
transition 1 *
CE 1 (eV)
transition 2
CE 2 (eV)
→ → → → → → → → → → → → → → → → → → → → → → → → → → → → → → → → → → → →
95 117 288 215 255 372 151 382 429 428 342 576 123 127 195 577 359 895 149 209 381 108 697 261 174 336 462 109 130 114 307 155 531 154 233 121
38 16 14 22 36 14 32 20 40 14 20 16 25 18 20 30 16 52 6 26 16 20 36 22 10 16 14 24 20 36 22 18 44 26 26 28
28 40 40 52 12 70 98 14 40 20 40 40 40 40 40 52 30 50 40 22 36 42 40 60 26 40 28 48 20 22 48 58 86 40 42 30
256 253 332 300 358 418 214 400 505 445 360 734 314 324 429 613 407 939 430 242 399 251 870 291 348 505 479 191 248 454 325 334 774 445 442 264
→ → → → → → → → → → → → → → → → → → → → → → → → → → → → → → → → → → → →
123 159 314 199 340 204 179 356 385 154 316 158 282 110 207 377 126 877 121 224 223 156 174 230 158 184 444 150 230 160 160 138 431 410 373 246
22 14 18 26 20 30 28 18 38 26 16 26 15 18 22 38 20 44 16 16 28 14 42 22 4 24 20 14 12 10 26 24 48 16 18 8
50 50 50 50 50 50 50 50 90 125
540 584 478 450 464 485 615 616 333 582
→ → → → → → → → → →
378 263 157 160 160 163 161 163 189 263
15 20 20 20 15 20 30 35 20 25
584 → 221
40
582 → 221
30
30 30 30 30 30 30 30
405 301 613 940 693 787 774
→ → → → → → →
334 137 577 878 675 431 431
20 15 33 35 40 52 51
407 301 613 940 693 787 774
→ → → → → → →
20 45 44 45 55 47 42
3652
256 253 332 300 358 418 214 400 505 445 360 734 314 324 429 613 407 939 430 242 399 251 870 291 348 505 479 191 248 454 325 334 774 445 442 264
336 107 377 896 461 531 531
dx.doi.org/10.1021/jf4052418 | J. Agric. Food Chem. 2014, 62, 3651−3659
Journal of Agricultural and Food Chemistry
Article
Table 1. continued methods and compounds Parabens (a) (a) (a) (a) (a) (a) (a)
(ESI−) butyl paraben (ES) benzyl paraben (ES) ethyl paraben (ES) methyl paraben (IS) isobutyl paraben (ES) isopropyl paraben (ES) propyl paraben (ES)
cone voltage (V) 50 42 46 52 54 54 40
transition 1 *
CE 1 (eV)
→ → → → → → →
22 22 20 20 22 22 22
193 227 165 151 193 179 179
92 92 92 92 92 92 92
transition 2
CE 2 (eV)
→ → → → → → →
16 16 14 12 16 16 14
193 227 165 151 193 179 179
136 136 136 136 136 136 136
a IS, internally standardized; ES, externally standardized; *, transition 1 used for quantification. Compounds were overspiked at (a) 1 ng/g, (b) 5 ng/ g, or (c) 10 ng/g.
and methanol, and magnesium sulfate, were purchased from SigmaAldrich. Doubly distilled grade dichloromethane, n-hexane, and ethyl acetate were acquired from Rathburn Chemicals (Walkerburn, UK). Polytetrafluoroethylene (PTFE, 0.22 μm) filters were obtained from QMX Laboratories. Dispersive solid-phase extraction (SPE) sorbents (Bondesil C18, NH2, and PSA) were supplied by Agilent (Wokingham, UK). Ultrapure water (18.2 MΩ cm) was obtained using a Purelab ultrapure water system (ELGA Purelab, UK). Sampling. Approximately 400 individual food/crop samples were collected for analysis. Sample preparation (including compositing and separation of component plant parts) produced around 200 samples for analysis. Samples were homogenized under cryogenic conditions using dry ice to minimize any loss of analytes. After cryogenic homogenization, the subsamples requiring screening for musk compounds were freeze-dried. Fish and Shrimp. A total of 30 samples of fish were collected from fish farms downstream of large urban sewage treatment effluent release points or from catchments close to areas of intensive livestock production. In addition, 21 samples of fish and 32 samples of shrimp imported from countries in Southeast Asia were also analyzed because of frequent detection of residues and reports issued by the EU Rapid Alert System for Food and Feed (RASFF). Fish were filleted, and the muscle including the subcutaneous fat was retained for analysis; the head, tail, and gut contents were not analyzed. Mushrooms. A total of 33 mushroom samples were collected from retail sources (sample origins labeled as UK, Ireland, or China) and growers in England. Mushroom samples (Agaricus bisporus) cultivated using compost from the two main commercial suppliers in the UK were obtained. The mushroom samples collected in the early stages of the study were brushed and rinsed (tap water) to remove compost adhering to the surface. A number of mushroom samples collected during the later stages in the study were peeled and destalked prior to extraction and these parts analyzed separately. In addition to mushrooms, associated compost material was also collected. Bovine Offal. Samples of bovine liver and bovine kidney were collected from abattoirs located in areas where the uptake of VMs or HPs and PCPs was considered a possibility: either rural areas associated with intensive livestock production or locations near major conurbations in the UK, respectively. Slices of kidney (typically 100 g) from 105 individual animals and slices of liver (typically 100 g) from another 95 individual animals were collected. On receipt at the laboratory the tissues were combined (5 samples representing 5 different animals, collected on the same day from the same abattoir) to produce a total of 40 composited samples (21 kidney and 19 liver) for analysis. This approach was employed to increase the chance of detecting residues because a greater number of animals would ultimately be tested. Root and Foliage Crops. A total of 22 samples of sugar beet, oilseed rape, and wheat were analyzed. Sugar beet leaves and roots were analyzed separately. Similarly, the stalks and heads of wheat were separated and treated as individual samples. For oilseed rape the oilseeds were separated from the foliage. These crops were collected from fields that had received applications of animal manures (pig slurry or cattle manure) or biosolids just before the sampled crop was sown.
compounds, which have been used extensively as fragrance fixatives in PCPs, are also recognized as common contaminants in the environment, especially in water.10 Previous concern arising from the use and persistence of nitro musks has led to restrictions on their use, and in recent times these compounds have been replaced by polycyclic musks, in particular galaxolide and tonalide. These are reported to account for around 90% of polycyclic musk usage in Europe during the past decade,15 but there is only limited information on their occurrence in the environment and their potential uptake into food.
■
MATERIALS AND METHODS
Prioritization. The scope of the prioritization undertaken was dependent on the food type and contaminant group of interest. Usage information was used to identify the most used HPs and VMs in the United Kingdom (UK), whereas appropriate usage data for PCPs could not be identified and, therefore, compound identification was based on expert opinion. Prioritizations were performed for nine different scenarios: HPs and PCPs in leaf crops, root crops, fish, and offal and VMs in leaf crops, root crops, fish, offal, and mushrooms. For example, the prioritization of HPs in leaf crops considered usage, adsorption to sewage sludge, sludge application to land, irrigation of contaminated surface waters, and subsequent uptake into root and translocation to leaves.16,17 The prioritization of VMs in farmed fish considered the potential surface water concentrations resulting from the treatment of intensively reared and pasture animals18 and the subsequent theoretical fish steady-state plasma concentration.19 Once the individual prioritizations were complete, the list was then finalized, taking into account the availability of methods for the analysis of the top-ranked compounds. The list of compounds analyzed by liquid chromatography−tandem mass spectrometry (LC-MS/MS) is given in Table 1. Also, six musk compounds (cashmeran, celestolide, galaxolide, tonalide, musk-xylene, and musk-ketone) were included in the study and analyzed by gas chromatography (GC)−high resolution (HR) MS. Reagents. HP, VM, and PCP standards were obtained from Qmx Laboratories (Thaxted, UK), Toronto Research (Toronto, Canada), Witega (Berlin, Germany), or Sigma-Aldrich (Gillingham, UK). Thirteen labeled analogues were used in this study: difloxacin-d3, ciprofloxacin-d8, enrofloxacin-d5, flubendazole-d3, and dinitrocarbanilide-d8, purchased from Witega; 13C-erythromycin-d3, danofloxacin-d3, lincomycin-d3, sulfadiazine-d4, and tetracycline-d6, obtained from Toronto Research; diclofenac-d4 and 13C-PCB77 from Cambridge Isotope Laboratories (Tewksbury, MA, USA); and 13C methyl paraben from Fluka (Gillingham, UK). Oxalic acid, glacial acetic acid, anhydrous sodium sulfate, aluminum oxide, sodium chloride, potassium phosphate monobasic, heptafluorobutyric acid (HFBA), and BakerBond CBX (500 mg, 6 mL) cartridges were acquired from Fisher Scientific (Loughborough, UK). Ethylenediaminetetraacetic acid (EDTA) disodium salt and trichloroacetic acid were purchased from Alfa Aesar (Heysham, UK). UPLC grade methanol, acetonitrile, and formic acid were obtained from Biosolve (Greyhound Chromatography, UK). Chromasolv acetonitrile 3653
dx.doi.org/10.1021/jf4052418 | J. Agric. Food Chem. 2014, 62, 3651−3659
Journal of Agricultural and Food Chemistry
Article
Table 2. Overview of Commodities and Compounds Classes Tested multiresidue imported fish imported shrimp trout mushrooms offal plant crops
yes yes yes yes yes yes
aminoglycosides
parabens
coccidiostats
yes yes yes yes
musks yes yes yes
yes yes yes
Sample Preparation. Multianalyte/Multiclass Method. The extraction method was based on the acetonitrile water extraction method described elsewhere.20 Briefly, cryogenically milled analytical portions (5 g) of the samples were spiked with 12 available deuterated/13C-labeled analogues. Water (2−5 mL depending on the sample) was added and the extraction performed using 1% oxalic acid in acetonitrile (15 mL). Anhydrous sodium sulfate (5 g) was added to reduce the water content prior to a cleanup step using dispersive solid-phase extraction (dSPE) with C18 and primary− secondary amine (PSA) to reduce the matrix coextractives. Finally, an aliquot (3 mL) of the extract was evaporated and reconstituted with acetonitrile/water (1:1, 1 mL). The final solvent composition was selected as a compromise between sufficient analyte solubility and acceptable chromatographic peak shapes for early-eluting polar compounds. All extracts were filtered through 0.22 μm PTFE filters. In the case of kidney and liver, 1% acetic acid in acetonitrile (instead of oxalic acid) was used because this modifier resulted in lower amounts of coextractives. Parabens. Cryogenically comminuted samples (5 g) spiked with 13 C-labeled methyl paraben were extracted with acetonitrile (10 mL). Magnesium sulfate (2 g) was added to reduce the volume of water in the final extract. Resultant extracts were purified by means of dSPE using C18 and PSA. Finally, an aliquot (4 mL) of the extract was concentrated to give a final volume of 1 mL in a mixture acetonitrile/ water (1:1). All extracts were filtered through 0.22 μm PTFE filters. Coccidiostats. Cryogenically milled mushroom samples (5 g), spiked with dinitrocarbanilide-d8, were extracted with 1% acetic acid in acetonitrile (15 mL). Anhydrous sodium sulfate (5 g) was added to reduce the volume of water, prior to dSPE cleanup with C18 and NH2. Finally, an aliquot (3 mL) of the cleaned up extract was evaporated and reconstituted with acetonitrile/water (3:1). All extracts were filtered through 0.22 μm PTFE filters. Aminoglycosides. Aminoglycoside residues were extracted from mushroom samples using a phosphate (20 mM KH2PO4) and EDTA (0.4 mM Na2EDTA) buffer containing trichloroacetic acid (2%) to precipitate proteins. The extract was then neutralized and cleaned up using a weak cation exchange solid-phase extraction cartridge (BakerBond CBX, 500 mg), which was eluted with 10% acetic acid in methanol (3 mL). The methanolic eluate was evaporated and reconstituted in an aqueous solution of an ion-pair reagent (400 μL of 20 mM HFBA). Musks. An aliquot (10 g) of the freeze-dried sample was homogenized with a mixture of dichloromethane/hexane (40:60, 30 mL) and placed in an ultrasonic bath for 30 min at 40 °C. The extract was cooled, filtered through anhydrous sodium sulfate, and then concentrated to ∼0.5 mL. The concentrated extract was purified on activated alumina (6 g) previously washed with hexane (100 mL) followed by ethyl acetate/hexane (1:10), before elution of musks with ethyl acetate/hexane (1:2) and ethyl acetate. The extracts were concentrated to ∼250 μL with the addition of 13C-labeled PCB77 internal standard. For some matrices, in particular kidneys, the final extracts were allowed to stand refrigerated, overnight/ and were then centrifuged to remove any solid material (e.g., salts) that may have precipitated, prior to analysis. Analytical Approach. Due to the diversity of samples analyzed and because individual samples, even of the same sample type, can exhibit different suppression effects, a sample “overspiking” approach (single-point standard addition) was employed to assess the validity of each individual LC-MS/MS commodity−analyte result. In this
approach each sample was analyzed twice, once without addition of analyte and once following addition of a known amount of analyte (typically equivalent to 1−10 ng/g in the sample, see Table 1). This enabled the detection of the analyte at the spiked concentration to be verified for each individual sample. Analyte concentrations were quantified against matrix-matched calibration standards. The calibration curve typically comprised five concentration levels prepared using a single sample known to be blank and of the type relevant to the sample batch. Blanks were subject to the same procedures as the samples and were spiked with the corresponding standards and internal standards at the end of the extraction and immediately before the evaporation step. Internal calibration was used for a selected number of compounds; others were quantified by external standardization (Table 1). Analytical standard solutions of each of the six musks were serially diluted in n-nonane and used to produce calibration standards for the quantification. The linearity of measurement with target musk compounds normalized to 13C-PCB77 was confirmed over a range corresponding to 6500 recovery determinations were carried out to support the validity of the methods used. The assessment of the results from the multiresidue method for the samples of fish and shrimp overspiked at 1 ng/g (24 analytes) and 5 ng/g (12 analytes) demonstrated that the method was fit for the purpose. The majority of recoveries were in the range of 60−120%, and associated relative standard deviations (RSDs) were generally
