國立臺灣大學公共衛生學院環境衛生研究所 碩士論文
Institute of Environmental Health College of Public Health National Taiwan University
Master Thesis
以極致液相層析/串聯式質譜儀檢測魚體組織中鄰苯二甲 酸二乙酯與個人保健品
Determination of Diethyl Phthalate and Personal Care Products in Fish Tissues with Ultra-performance Liquid
Chromatography/tandem Mass Spectrometry 洪婉華
Wan-Hua Hung
指導教授：陳家揚 博士 Advisor: Chia-Yang Chen, Ph.D.
中華民國 107 年 7 月
July 2018
謝辭
感謝陳家揚老師這四年來的耐心教導與鼓勵，並在研究遇到困難時提供建議，
指引我前進的方向；此外，老師對生活和學術認真嚴謹的態度，也是值得學習的典 範，四年下來，不論是平常做事還是研究方面，我均成長許多。
感謝實驗室的所有成員在我研究過程中提供許多協助，包括銘聰學長、妍秀學 姐、信宏學長、彥均、冠萍、則穎、祐辰、耕文、儒佑、采築、瑞敏、蕭鵬，也謝 謝我的同窗和好友們，包括于茹、宜蓁、瀅安、佩婷、秋霖，在我需要幫助時鼎力 相助以及在我實驗撞牆期時傾聽我的煩惱並給予鼓勵，是我能堅持下去完成研究 不可或缺的一環；此外也謝謝蔡詩偉老師和陳鑫昌老師願意撥冗擔任口試委員，並 提供許多專業及寶貴的意見；最後謝謝我的家人，在這四年間給予許多協助、鼓勵 和支持，讓我能無後顧之憂地進行研究。
中文摘要
鄰苯二甲酸酯(phthalate esters)廣泛使用於工業和消費性產品，個人保健品 (personal care products)則包含民眾基於日常健康照護、驅蟲、美容等目的而大量使 用的各式各樣的化合物。許多研究顯示鄰苯二甲酸酯和一些個人保健品具有發育 毒性且為內分泌干擾物質。這兩類化合物普遍存在於環境中；污水處理廠是它們進 入環境的主要途徑之一。一些研究顯示這些化合物可能會累積在魚體內，然而少有 研究同時分析不同魚體組織器官裡的鄰苯二甲酸酯或個人保健品成分，且能在魚 體組織器官同時檢測這兩大類化合物的分析方法目前仍相當有限。因此，本研究開 發檢測魚肉和魚肝內的鄰苯二甲酸二乙酯和十一種個人保健品成分之分析方法。
樣本前處理採用基質固相分散法(matrix solid-phase dispersion, MSPD)，以 C8 作為 分散劑，並依序用 5 毫升甲醇和丙酮沖提樣本管柱和其下所接的矽膠樣本淨化管 柱；樣本經濃縮後以極致液相層析串聯式質譜儀，在多重反應監測模式下獲取質荷 比資訊，並搭配同位素稀釋技術進行定量分析。鄰苯二甲酸二乙酯和鹼性個人保健 品以正離子電灑游離法游離，並使用 Ascentis Express F5 管柱搭配移動相(A) 5 mM 醋酸氨水溶液(pH = 6.40)、(B)甲醇，進行梯度層析；酸性個人保健品則以負離子電 灑游離法游離，並使用 Waters CORTECS UPLC C18 管柱搭配移動相(A) 0.04%乙 酸(pH = 3.45)、(B)甲醇，進行梯度層析。
儀器方法最佳化方面，正離子電灑游離法比較了兩種有機移動相和四個游離 源溫度，負離子電灑游離法則比較了兩種不同的層析條件。前處理最佳化方面，本 研究比較了：(1)兩種沖提溶劑組合；(2)兩種沖提溶劑體積；(3)兩種 MSPD 分散劑；
(4)兩種淨化吸附劑；(5)不同量的淨化吸附劑；(6)不同體積的最終樣本萃取液。魚 肉和魚肝的基質效應因子大部分分別落在 70.3-95.6% 與 24.3-61.9%；魚肉的萃取 效率有一半落在 62.1-76.6%，魚肝則大部分落在 31.6-71.2%。待測物在魚肉和魚
肝的方法偵測極限分別為 0.57-15.0 ng/g（濕重）及 4.37-104 ng/g（濕重）。本分 析方法待測物的定量偏差大多低於 30%，相對標準偏差則均低於 20%。
關鍵字：鄰苯二甲酸二乙酯、個人保健品、魚肉、魚肝、基質固相分散法、極致液 相層析串聯式質譜儀
Abstract
Phthalate esters (PAEs) are widely used in industrial and consumer products;
personal care products (PCPs) contain diverse chemicals used at a large scale for daily
lives or personal hygiene, which include analgesics, insect repellents, UV filters, and so
on. Previous studies indicate that PAEs and some PCP ingredients have developmental
toxicity and could disrupt endocrine systems. The two groups of compounds are
ubiquitous in the environment, and wastewater treatment plants are one of the major
emission sources. Some studies show that PAEs and PCPs may accumulate in fish tissues;
however, limited studies determined PAEs or PCPs in different fish tissues
simultaneously. Furthermore, few methods are available to analyze PAEs and PCPs
together in fish tissues. Thus, this study developed and validated a method to
simultaneously determine diethyl phthalate (DEP) and 11 PCPs ingredients in fish muscle
and liver. Samples were extracted with matrix solid-phase dispersion (MSPD) using C8
adsorbent; 5-mL methanol and acetone were sequentially passed through the tandem
system of a MSPD cartridge piggyback on a silica gel cartridge for cleanup. After
concentration, the eluents were analyzed using ultra-performance liquid
chromatography/tandem mass spectrometry (UPLC-MS/MS) with multiple reaction
monitoring (MRM) and were quantified with isotope dilution techniques. DEP and the
basic PCPs were separated on an Ascentis Express F5 column (30 × 2.1 mm, 2.0 μm)
with the mobile phases consisting of (A) 5 mM ammonium acetate(aq) (pH = 6.40) and (B)
methanol, and were ionized at positive electrospray ionization mode (ESI+). The acidic
analytes were separated on a Waters CORTECS UPLC C18 column (30 × 2.1 mm, 1.6 μm) with the mobile phases consisting of (A) 0.04% acetic acid(aq) (pH = 3.45) and (B)
methanol, and were ionized at negative electrospray ionization mode (ESI-).
The optimization of the instrumental analysis included the tests of two organic
mobile phases and four source temperatures for ESI+, and two chromatographic
conditions for ESI-. The sample preparation method was optimized by testing two elution
solvent combinations, two elution volumes of solvents at each portion, two adsorbents
for MSPD, two adsorbents for cleanup, the amount of cleanup adsorbents, and volumes
of the final residues.
The matrix effect factors of most analytes in fish muscle and liver ranged from 70.3-
95.6% and 24.3-61.9%, respectively. The extraction efficiencies of half the analytes in
muscle were 62.1-76.6%, and those of most analytes in liver were 31.6-71.2%. The limits
of detection (LODs) of analytes were 0.57-15.0 ng/g weight wet (w.w) for muscle and
4.37-104 ng/g w.w. for liver, respectively. Most of the quantitative biases were below
30%, and all the relative standard deviations were below 20%.
Keywords: diethyl phthalate, personal care products, fish muscle, fish liver, matrix solid-
phase dispersion, UPLC-MS/MS
Contents
謝辭 ... I 中文摘要 ... III Abstract ... V List of figures ... X List of tables ... XI
Chapter 1. Introduction ... 1
1.1. Phthalate esters ... 1
1.2. Personal care products ... 3
1.3. Analytical methods for PAEs and PCPs in fish tissues ... 5
1.4. Objectives ... 7
Chapter 2. Methods ... 9
2.1. Reagents and materials ... 9
2.2. Sample collection ... 10
2.3. Sample preparation ... 11
2.4. Water content of samples ... 13
2.5. Instrumental analysis ... 14
2.5.1. Liquid chromatography ... 14
2.5.2. Tandem mass spectrometry... 15
2.6. Method validation ... 16
2.6.1. Extraction efficiency and matrix effect... 17
2.6.2. Accuracy and precision ... 17
2.6.3. Identification, quantification and data analysis ... 18
2.6.4. Quality assurance and quality control ... 19
Chapter 3. Results and discussion ... 21
3.1. Optimization of chromatography ... 21
3.2. Optimization of mass spectrometric parameters ... 22
3.3. Optimization of sample preparation ... 23
3.4. Method validation ... 30
Chapter 4. Conclusions ... 37
Reference ... 39
Figures ... 47
Tables ... 59
List of figures
Figure 1. Chromatograms of ESI+ of a chemical standard solution ... 47 Figure 2. Chromatograms of ESI- of a chemical standard solution ... 48 Figure 3. Chromatograms of the first and second product ions of acetaminophen in liver
samples ... 48 Figure 4. Signal intensities of DEP and basic PCPs with different organic mobile phases
... 49 Figure 5. Chromatograms of benzophenone with different organic mobile phases ... 49 Figure 6. Chromatograms of acetaminophen with different initial organic mobile phase
proportions ... 50 Figure 7. Signal intensities of acidic PCPs under different chromatographic conditions
... 50 Figure 8. Chromatograms of ESI- with different chromatographic conditions ... 51 Figure 9. Signal intensities of DEP and basic PCPs with different ionization source
temperatures... 51 Figure 10. Elution efficiencies (%) of analytes with different combinations of elution
solvents on C18 non-endcapped adsorbent ... 52 Figure 11. Elution efficiencies (%) of analytes with different elution volumes at each
portion ... 52 Figure 12. Elution efficiencies (%) of analytes with different combinations of elution
solvents on alumina cartridge ... 53 Figure 13. Elution efficiency (%) of analytes from different adsorbents ... 53 Figure 14. Elution efficiency (%) of analytes from different cleanup adsorbents ... 54 Figure 15. Background levels of DEP and benzophenone using the silica gel with and
without the pre-wash step ... 54 Figure 16. Peak areas of analytes using the silica gel with and without pre-wash step . 55 Figure 17. Appearance of eluents of liver samples using different amount of silica gel
for cleanup ... 55 Figure 18. Peak areas of analytes in the 500-μL final residues of liver samples (original
samples) and their two-fold dilution samples ... 56 Figure 19. Matrix effect factors (%) of analytes in matrices ... 57 Figure 20. Extraction efficiencies (%) of analytes in matrices... 57
List of tables
Table 1. Chemical structures and molecular weights of analytes... 59
Table 2. The LC conditions of separating analytes ... 61
Table 3. Tandem mass parameters ... 62
Table 4. Different chromatographic conditions for ESI- ... 63
Table 5. Matrix effect factors (%) of analytes in matrices ... 64
Table 6. Extraction efficiencies (%) of analytes in matrices ... 64
Table 7. IDLs, IQLs, linear ranges and r2 of calibration curves ... 65
Table 8. The limits of detection (LODs) and limits of quantification (LOQs) ... 66
Table 9. Accuracy and precision in fish muscle and liver ... 67
Chapter 1. Introduction
Phthalates esters (PAEs) are widely used in consumer products manufacturing, and
personal care products (PCPs) are extensively used in human daily lives. Some of both
groups of the compounds have developmental toxicity and endocrine disrupting
properties [1, 2]. PAEs and PCPs are ubiquitous in the environment, and water is one of
the major media for their distribution in the environment [3, 4]. Some of these chemicals
are relatively lipophilic and have the potential to bioaccumulate in fish [2, 3]. In addition,
the continuous exposures of fish to PAEs and PCPs may increase the potential for
accumulation of these compounds in fish tissues because effective exposure duration is
increased [5, 6]. Therefore, it is important to investigate the levels of the above
compounds in fish. However, the existing methods are limited for simultaneously
analyzing PAEs and PCPs in fish tissues. It warrants the method development for
determining these compounds in fish tissues.
1.1. Phthalate esters
PAEs, a group of di-esters of 1,2-benzenedicarboxylic acid, are used in several
consumer products [7]. Lower molecular weight PAEs like dimethyl phthalate (DMP),
diethyl phthalate (DEP) and Di-n-butyl phthalate (DnBP) are mainly applied to adhesives,
waxes, insecticides and cosmetics; higher molecular weight PAEs, such as di-(2-
ethylhexyl) phthalate (DEHP) and di-isononyl phthalate (DiNP) are primarly used as
plasticizers in products of polyvinyl chloride plastics (PVC) [8]. Because PAEs are not
chemically bound to the products where they are applied, they constantly leach into the
environment during manufacture and from final products. Moreover, PAEs are used
extensively; to date, worldwide annual consumption of PAEs is 6−8 million tons. They
have been detected in air, water, sediment, soil and biota worldwide [9]. Wastewater
treatment plants (WWTP) is one of the important sources of discharging PAEs to the
environment because most conventional WWTPs are not designed for removing these
micropollutants [4, 10]. The concentrations of PAEs are reported from <LOQ (limits of
quantification) to thoudands of μg/L in surface water [9, 11], and from <LOQ to tens of
mg/kg (dry weight, d.w.) in sediment, respectively [9, 12, 13]. In fish, the concentrations
of PAEs are reported from <LOQ to hundreds of mg/kg (d.w.) in muscle [14], and ranged
from 0.39 to 6.90 mg/kg (d.w.) in livers [15].
Regarding the impact of PAEs on aquatic organisms, PAEs are known to have
developmental toxicity and endocrine disrupting effects [1, 16]. Furthermore, some
studies showed that PAEs may bioacumulate in fish muscle and liver [9, 15, 17]. Valton
et al. found that the bioaccumulation factors (BAF) of di-iso-butyl phthalate (DiBP),
DnBP, butylbenzyl phthalate (BBP) and di-n-octyl phthalate (DNOP) for fish musle and
the BAFs of DMP, DEP, DiBP, DnBP, BBP, DEHP and DNOP for livers were higher
than 2,000 [15, 18].
1.2. Personal care products
Personal care products (PCPs) contain various compounds used in large quantities for
personal hygiene or daily life including analgesics, stimulating drinks, insect repellants,
UV filters, preservatives, to name a few [3]. Many PCPs enter domestic sewage after
usage; however, their removal rates in wastewater vary widely depending on their
chemical properties, the operating conditions and the treatment technologies of
wastewater treatment [19-23]. Although some PCPs can be removed almost completely
and the environmental half-lives of most PCPs are short, the extensitve use and continous
emission make most PCPs behave as “pseudo-persistent” in the aquatic environment [24].
PCPs usually exist in surface water at ng/L to μg/L levels [2, 25]. Aquatic creatures
may have life-cycle and even multigeneration exposure to PCPs, and PCPs may
potentially bioaccumulate in fish and affect these organisms even at low concentrations
in the surface water. In addition, PCPs are designed to be biogially active, which may
result in unintended consequences on aquatic organisms [3].
Analgesics give relief from pains without causing anesthesia. They are increasingly
used because of population aging, population growth, and their easy accessibility.
Concentrations of analgesics are found to range from <LOD to tens of ng/g (d.w.) in fish
muscle [26, 27].
Caffeine is the most consumed stimulant in the world [28]. Alvarez-Munoz et al.
reported caffeine in fish muscle at tens of ng/g (d.w.) [29]. Concentrations of caffeine are
found to range from <LOD to 4.5 ng/g (d.w.) in whole fish [30].
DEET (N,N-Diethyl-meta-toluamide) is the most common active ingredient of
insect repellents [31]; however, it may pose neurotoxicity to humans [32]. There are few
research on analyzing DEET in fish tissues. Tanuoe et al. analyzed DEET in Japanese
wild fish muscle, liver, brain, and kidney, and there were just few positive samples [33].
UV filters are commonly used in sunscreen and cosmetics [34]. Because of the
lipophilic characteristics of these compounds, many UV filters have potential for
bioaccumulation and biomagnification through food chain [35]. In addition, some UV
filters would be estrogenic [36]. Concentrations of UV filters in fish filet and livers were
found to range from <LOD to 182 ng/g (d.w.) and range from <20 to 1,037 ng/g (d.w.),
respectively [37, 38].
Parabens, the alkyl esters of p-hydroxybenzoic acid, are antimicrobial and are used
as preservatives in cosmetics, food and pharmaceuticals [3]. The most common parabens
are methyl, ethyl, propyl, butyl and benzyl parabens. Some studies reported parabens as
endocrine disrupters [39, 40]. The concentrations of parabens found in fish muscle are
from <LOD to 18.5 ng/g (d.w.) [29, 41].
1.3. Analytical methods for PAEs and PCPs in fish tissues
PAEs or PCPs in fish tissues are usually extracted using ultrasonic extraction [15, 38,
42-44], pressurized liquid extraction (PLE) [14, 24, 29, 45], and Soxhlet extraction [46-
49]. Liquid chromatography coupled with tandem mass spectrometry (LC–MS/MS) is the
major technique of instrumental analysis because of its good detection sensitivity and
selectivity [26, 29, 30, 33, 47, 50-53].
Most of the above sample preparation methods take long time or use a large amount
of organic solvents or require expensive apparatus. Matrix solid-phase dispersion
(MSPD), which was first introduced by Barker et al. in 1989 and was designed to disrupt
and extract semi-solid and solid matrices [54], requires less time, organic solvents, and
does not need special equipment. The procedure of MSPD basically consists of three main
steps as follows: (1) blend the sample with the dispersant material; (2) transfer the
homogenized mixture into a solid-phase extraction (SPE) cartridge; (3) elute analytes
with an appropriate solvent or sequence of solvents. The main feature of MSPD is the
dispersion of matrices over a huge solid sorbent surface, which increases the contact
surface area between matrices and adsorbents as well as the elution solvents. Additionally,
in MSPD, extraction and cleanup may be performed in one step, which could reduce the
used amount of organic solvents and simplifies the procedure. Furthermore, the extraction
process in MSPD takes place under ambient conditions and does not require any
expensive equipment. Because of the feasibility, flexibility, versatility and low costs of
MSPD, it is wildly applied to viscous, semi-solid or solid matrices, and various groups of
analytes [55-57]. For example, Ocaña-Rios et al. used MSPD and gas chromatography-
mass spectrometry to analyze two polycyclic aromatic musks and five UV filters in fish
muscle, and the limit of detections (LODs) were 0.004-0.012 μg/g d.w. Freitas et al. used
MSPD and gas chromatography with electron-capture detection (GC-ECD) to determine
five pesticide residues in tropical fruits, and the LODs were 5.0-25 μg/kg [58].
1.4. Objectives
Most studies of PAEs or PCPs focus on the levels of these contaminants in water or
sediment. Although some studies reported the concentrations of these compounds in fish,
most of them only focused on muscle [24, 38, 48, 59-62]. It is vital to analyze other tissues
such as liver, which is important for metabolism and detoxification, to learn more about
the potential health effects of these chemicals on fish.
There are some existing methods for analyzing PAEs or PCPs in fish tissues;
however, few studies investigated PAEs and PCPs simultaneously. Furthermore, some
sample preparation methods are solvent consuming, tedious, laborious, and in high-cost.
Therefore, an analytical assay is desired that is capable of analyzing these compounds in
fish tissues efficiently and simultaneously.
The present study aimed to develop and validate an analytical method for
simultaneous analysis of diethyl phthalate (DEP) and 11 PCPs (Table 1, page 59) in fish
muscle and liver. The 11 PCPs are acetaminophen, caffeine, DEET, benzophenone,
oxybenzone, methyl paraben, ethyl paraben, propyl paraben, butyl paraben, ketoprofen,
and ibuprofen. The 12 analytes were chosen according to three criteria: (1) high usage or
production amount, (2) high concentrations and being frequently detected in the aquatic
environment, and (3) lacking field-derived information about bioaccumulation. For
instance, the selected PCPs in this study has been shown ubiquitous in the aquatic
environment in Taiwan [63-66].
This study tested different elution solvent combinations, elution volumes of solvents
at each portion, adsorbents for MSPD, adsorbents for cleanup, amounts of cleanup
adsorbents, and volumes of the final residues for optimization of sample treatment. The
instrumental analysis was conducted on ultra-performance liquid
chromatography/tandem mass spectrometry (UPLC-MS/MS) at electrospray ionization
(ESI). MS parameters, mobile phases, chromatography columns, and LC conditions were
optimized. The method was validated using Tilapia (Oreochromis niloticus) because it
can be acquired easily from markets in Taiwan and it is one of the most consumed fish in
the world. Matrix effect, extraction efficiency, accuracy and precision were evaluated.
Chapter 2. Methods
2.1. Reagents and materials
Diethyl phthalate (99.9%, 5,000 μg/mL in methanol) was purchased from
AccuStandard (New Haven, CT, USA). Acetaminophen (99.6%, powder) was bought
from United States Pharmacopeia (Rockville, MD, USA). DEET (N,N-diethyl-3- methylbenzamide) (purity≧98%, 250 mg), caffeine, oxybenzone, methyl paraben, ethyl
paraben, propyl paraben, ketoprofen, ketoprofen-2D3 and ibuprofen (purity ≧ 98%,
powder) were obtained from Sigma-Aldrich (St. Louis, MO, USA). Benzophenone and
butyl paraben were bought from Alfa Aesar (Heysham, Lancashire, U.K.; purity≧99%,
powder). Diethyl phthalate-2D4 (DEP-2D4), benzophenone-2D10 (purity ≧ 98%, 100 μg/mL in nonane), caffeine-13C3 (purity≧98%, 100 μg/mL in methanol), DEET-2D6
(purity≧98%, 100 μg/mL in dichloromethane-2D6), oxybenzone-13C6, ibuprofen-13C3
(purity≧98%, 100 μg/mL in acetonitrile), methyl paraben-13C6 and butyl paraben-13C6
(purity ≧ 98%, 1000 μg/mL in methanol) were purchased from Cambridge Isotope
Laboratories (Andover, MA, USA). Acetaminophen-2D4 (100 μg/mL in methanol) was
obtained from LGC Standards (Teddington, Middlesex, England, U.K.).
HPLC-grade acetone and methanol, and LC/MS-grade methanol were provided by
Merck (Darmstadt, Hesse, Germany). HPLC-grade n-heptane and LC/MS-grade
acetonitrile were bought from J.T Baker (Philipsburg, NJ, USA). Ammonium acetate (5 M in H2O) and acetic acid (≧99.8%) were purchased from Sigma-Aldrich. Milli-Q water
was from a Milli-Q integral water purification system (Merck Millipore, Darmstadt,
Germany).
Solid and liquid state standards were dissolved in methanol as stock solutions, and
the concentrations were as below: acetaminophen, caffeine, DEET, benzophenone,
oxybenzone and ketoprofen-2D3 at 1 mg/mL; methyl paraben, ethyl paraben, propyl
paraben, butyl paraben, ketoprofen and ibuprofen at 4 mg/mL. Regarding the stable
isotope-labeled internal standards originally in nonane, because nonane is not miscible
with methanol, we used acetone to make their respective two-fold dilutions as stock
solutions, then used methanol or acetone to make the subsequent dilutions, which
depended on the concentrations we needed. All stock solutions and the commercialized
standard solutions of the rest chemicals were stored at 4℃.
2.2. Sample collection
The tilapia (Oreochromis niloticus) samples for method development and validation
were bought from local markets in New Taipei City. The samples were purchased in the
morning, stored at 4°C, and extracted on the next day.
2.3. Sample preparation
Fish muscle and liver samples were homogenized first using a blender (Waring
Commercial, Stamford, CT, USA) individually. One gram (wet weight, w.w.) of
homogenized samples were weighted into an IKA DT-20 homogenization tube
(Wilmington, NC, USA) and stable isotope-labeled internal standards (ISTDs) were
added. Regarding muscle samples, 25-μL DEP-2D4 at 20 μg/mL in acetone and 25 μL of
rest ISTDs in methanol were spiked; the spiked concentrations in methanol were as follows: 20 μg/mL of butyl paraben-13C6; 5 μg/mL of acetaminophen-2D4, caffeine-13C3,
DEET-2D6, benzophenone-2D10, oxybenzone-13C6, methyl paraben-13C6, ketoprofen-2D3
and ibuprofen-13C3. Liver samples were spiked with 50-μL mixture of benzophenone-
2D10 and DEP-2D4 at 20 μg/mL in acetone, and 40 μL of rest ISTDs in methanol; the
spiked concentrations in methanol were as follows: 25 μg/mL of butyl paraben-13C6,
ketoprofen-2D3 and ibuprofen-13C3; 10 μg/mL of acetaminophen-2D4, caffeine-13C3,
DEET-2D6, oxybenzone-13C6 and methyl paraben-13C6.
After spiking, the samples were homogenized with 4 g of C8 adsorbent (particle size 40–63 μm, carbon content ≧ 11.6%, SiliCycle, Quebec City, Quebec, Canada) using an
IKA ULTRA-TURRAX Tube Drive. Afterwards, the mixture was transferred into a 12-
mL SPE cartridge with a polyethylene frit at the bottom, then it was compressed and
covered with another polyethylene frit. A 6-mL silica gel cartridge (bed weight 2 g,
PerkinElmer, Waltham, MA, USA) was for cleanup to adsorb amino acids, fatty acids
and organic acids, which was washed twice by 4-mL acetone before used. The sample
cartridge was piggybacked on a washed silica gel cartridge and analytes were eluted twice
with 5-mL methanol and 5-mL acetone, respectively; the first 5-mL methanol was from
the solvent for rinsing the used homogenization tube. The elution rate was at 1-2 drops/sec.
The eluents were concentrated to approximately 5 mL at 45℃ by a Savant SPD 1010
SpeedVac (Thermo Fisher, Waltham, MA, USA.), and then were centrifuged by a
KUBOTA 2010 centrifuge (Kubota, Bunkyo-ku, Tokyo, Japan) at 3,000 rpm (1409 × g)
for 5 minutes. The supernatants were transferred to 15-mL deactivated (silanized) glass
centrifuge tubes, concentrated to 1 mL, and were filtered through a methanol-washed
Millex Samplicity Filter (hydrophilic PTFE, pore size 0.20 µm, diameter 33 mm) with a
Samplicity Filtration System (Merck Millipore, Darmstadt, Germany). Regarding muscle
samples, the filtrates were further concentrated to 250 µL and were refrigerated at 4°C
overnight. Afterwards, the filtrates were centrifuged at 3,000 rpm (1107 × g) for 5 minutes
and the supernatants were transferred to 300-μL inserts. For liver samples, the filtrates
were concentrated to 500 µL and were refrigerated at 4°C overnight. Thereafter, the
filtrates were centrifuged by a Centrifuge 5415 R (Eppendorf, Stevenage, Hertfordshire,
UK) at 13,200 rpm (16100 × g) at 4°C for 10 minutes. The subnatants were filtered
through a methanol-washed hydrophilic PTFE filter (0.20 μm) with a Samplicity
Filtration System again. Four microliters of the final residues were injected onto the
UPLC-MS/MS for analysis.
2.4. Water content of samples
Fish muscle and liver were weighted, then the tissues were dried in a Circulator Oven
DO45 (Deng Yng, Taishan Dist., New Taipei City, Taiwan) at 105°C until their weights
remained constant. After drying, the samples were transferred to a desiccator for cooling.
Afterwards, the samples were weighted again. The water contents of samples were
determined by using the formula below:
Water content (%) = × 100 W1 - W2 W1
Where,
W1 = Weight (gram) of sample before drying.
W2 = Weight (gram) of sample after drying.
The water contents of fish muscle and liver were 78.5% and 67.3%, respectively.
2.5. Instrumental analysis
2.5.1. Liquid chromatography
The chromatographic separation was performed on a Waters ACQUITY UPLC
system (Waters, Milford, MA, USA). DEP and the basic PCPs (acetaminophen, caffeine,
DEET, benzophenone and oxybenzone) were separated on an Ascentis Express F5
column (30 × 2.1 mm, 2.0 μm). The mobile phases were composed of (A) 5 mM
ammonium acetate(aq) (pH = 6.40) and (B) methanol, and the flow rate was 0.65 mL/min.
The column oven temperature and sample chamber temperature were set at 40℃ and
20℃, respectively. The injection volume was four microliters. The chromatographic
gradient began from 5% B for 0.5 minutes, then increased to 95% B in 4 minutes, held
for 1 minute, and back to the initial compositions in 0.3 minutes. The column was re-
equilibrated for 2 minutes and the total chromatographic time was 7.8 minutes (Table 2,
page 61).
A CORTECS UPLC C18 column (30 mm × 2.1 mm, 1.6 μm, Waters) was used for
the separation of the acidic PCPs (methyl paraben, ethyl paraben, propyl paraben, butyl
paraben, ketoprofen and ibuprofen). The mobile phases consisted of (A) 0.04% acetic
acid(aq) (pH = 3.45) and (B) methanol, and the flow rate was 0.5 mL/min. The column
oven temperature and sample chamber temperature were 30℃ and 20℃, respectively.
The gradient started with 15% B for 0.5 minutes, and was increased to 100% B in 2.5
minutes, with holding for 0.5 minutes before returning to 15% in 0.5 minutes, and the
column was re-equilibrated for 1.7 minutes. The total chromatographic time took 5.7
minutes (Table 2, page 61). The chromatograms of the analytes were shown in Figure 1
and Figure 2 (page 47 and 48).
2.5.2. Tandem mass spectrometry
After the chromatographic separation, analytes were detected by a Waters Quattro
Premier XE triple-quadrupole mass spectrometer (Waters, Milford, MA, USA) at
multiple-reaction monitoring (MRM) mode; the two most abundant ion transitions of
each analyte were for quantification and confirmation, respectively. For ketoprofen and
ibuprofen, only one ion transition was monitored because they only form one intensive
and stable product ion. The MRM parameters were optimized by using a syringe pump
to inject 1.0 µg/mL standard solutions of individual analyte directly to the mass
spectrometer. The MRM transitions and parameters of each analyte are shown in Table 3
(page 62).
DEP and the basic PCPs were ionized by positive electrospray ionization (ESI+)
with the capillary voltage 2 kV, extractor voltage 4 V, source temperature 120℃,
desolvation temperature 500℃, cone gas flow 150 L/hr, desolvation gas flow 900 L/hr,
and collision cell pressure 3.37 × 10-3 mbar. The desolvation and collision gas were
nitrogen and argon, respectively.
Acidic PCPs were ionized by negative electrospray ionization (ESI-) with the
capillary voltage 3.0 kV, extractor voltage 3 V, source temperature 120℃, desolvation
temperature 450℃, cone gas flow 100 L/hr, desolvation gas flow 900 L/hr, and collision
cell pressure 3.37 × 10-3 mbar. The desolvation and collision gas were nitrogen and argon,
respectively.
2.6. Method validation
2.6.1. Extraction efficiency and matrix effect
Extraction efficiency was defined as the peak area ratios of pre-spiked samples to
those of post-spike samples at the same levels of analytes. Matrix effect factors were
calculated as the peak area ratios of post-spiked samples to those of the same
concentrations of chemical standards in methanol. The spiked level of each analyte in
muscle samples was 200 ng/g w.w. For liver samples, the spiked level of DEP and
acetaminophen was 1,000 ng/g w.w., and those of other analytes were 400 ng/g w.w. The
samples were done in four duplicates (n = 4). The spiked level of acetaminophen was
higher than most of other analytes in liver samples because there was a peak (retention
time = 0.62 min) near acetaminophen (retention time = 0.74) and interfered with the
quantification (Figure 3, page 48). Regarding DEP, studies found that PAEs tend to
accumulate more in fish livers than in muscle [15, 42], and the endogenous high
concentrations of DEP in liver samples may influence the quantification. To prevent the
quantification of the two analytes from interference, the spiked level in liver samples was
elevated.
2.6.2. Accuracy and precision
Accuracy and precision were evaluated using pre-spiked fish muscle at three
spiked levels with four duplicates (n = 4) at each level, and pre-spiked fish liver at 1,000
ng/g w.w. with four duplicates. The spiked levels of muscle samples were 62.5, 200, and
500 ng/g w.w.
2.6.3. Identification, quantification and data analysis
The instrumental detection limits (IDL) and instrumental quantification limits (IQL)
were determined by analyzing low concentrations of chemical standards in methanol. IDL
and IQL were defined as the signal-to-noise ratio (S/N ratio) of the confirmatory ion at 3
and the S/N ratio of the quantitative ion at 10, respectively. If calculated IDLs were higher
than IQLs based on the above definition, the IQLs were reported as the same values of
IDLs. The limits of detection (LODs) and limits of quantification (LOQs) were defined
as the S/N ratio of confirmatory ion at 3 and the S/N ratio of quantitative ion at 10 in
spiked matrix samples, respectively. If LODs were higher than LOQs based on the above
definition, the LOQs were reported as the same values of LODs.
The calibration curves were established using MassLynx 4.1 (Waters) by
normalizing the peak areas of native analyte standards to those of their individual isotope-
labeled internal standards. The curves were made by linear regression with the weighting factor of 1/χ. There were at least 6 points in each curve which the concentrations of
analytes ranged from 1 to 4,000 ng/mL. The r2 of all analytes were higher than 0.99.
Further data analysis were done using Microsoft Excel 2013.
2.6.4. Quality assurance and quality control
All the glassware that would contacted the samples was deactivated (silanized) to
prevent the analytes from adsorbing on the glass surface. All glassware, homogenization
tubes, and cartridges were rinsed with methanol and acetone before use. C8 adsorbent and
cartridge frits were sonicated with methanol and acetone sequentially, and were dried in
a chemical fume hood before use. After use, the glassware was washed with detergent
and tap water, and then was rinsed with methanol and acetone. The anatomical tools and
blender were washed with tap water, and were rinsed with Milli-Q water, acetone, n-
heptane, acetone and methanol sequentially. Other labware was washed by tap water, and
was rinsed with Milli-Q water, methanol and acetone sequentially. All cleaned labware
was dried by air flow in a chemical fume hood and was covered with aluminum foil to
avoid contamination.
A Waters Isolator column (50 × 2.1 mm, 3.5 μm) with an extension tube was
installed onto the UPLC system to eliminate the influences of the background DEP from
the UPLC system and mobile phases. Caffeine, DEET and oxybenzone were detected in
reagent blanks at approximately 1 ng/g, and DEP, benzophenone and methyl paraben
were found at tens of ng/g.
Chapter 3. Results and discussion
3.1. Optimization of chromatography
The chromatographic method for ESI+ in the present research was modified from
previous methods developed by our team [66, 67]. Different organic mobile phases
(acetonitrile and methanol) were tested for better signal intensities. Methanol as the organic mobile phase provided 1.9-40 times higher signal intensities of DEP and the basic
PCPs than those of acetonitrile (ACN) (Figure 4, page 49). Furthermore, when 0.25 μg/mL of benzophenone was injected at 4 μL, the signal intensity reached 1 × 105 with
methanol but no apparent peak using ACN (Figure 5, page 49). Methanol is a protic
solvent and would be able to facilitate the protonation of basic analytes. As a result,
methanol was chosen as the organic mobile phase for ESI+.
Regarding the LC gradient for ESI+, initial organic mobile phases at 5% and 10%
methanol were tested for better retention of acetaminophen, which was first eluted analyte
from the column. 5% of methanol retained acetaminophen (retention time, RT = 0.69
minute) better comparing with 10% of methanol (RT = 0.47 minute) (Figure 6, page 50).
Thus, 5% of methanol was adopted.
For ESI-, the chromatographic method was also modified from a previous method
developed by our team [66]. The combination of 0.04% acetic acid(aq) (pH 3.45) as the aqueous mobile phase, a CORTECS UPLC C18 column (30 × 2.1 mm, 1.6 μm) and
Gradient 1 (the details of the gradient was shown in Table 4, page 63) offered 1.6-3.5
times higher of signal intensities of the acidic PCPs compared with the combination of
10 mM N-methylmorpholine(aq) (pH 9.60), a Kinetex EVO C18 column (50 × 2.1 mm, 1.7 μm, Phenomenex, Torrance, CA, USA) and Gradient 2 (the details of the gradient was
shown in Table 4, page 63; Figure 7, page 50). The faster slope of Gradient 1 (34%/min)
than that of Gradient 2 (22.9%/min), plus the shorter column length of the CORTECS
UPLC C18 column than that of the Kinetex EVO C18 column, made better ionization and
sharper peaks (Figure 8, page 51). Consequently, the former chromatographic condition
was chosen for ESI-.
3.2. Optimization of mass spectrometric parameters
Because DEP was added in this study and some compounds ionized at ESI+ were
removed from the previous method developed by our team [66], the source temperature
for ESI+ was reevaluated. Four source temperatures (120, 130, 140 and 150℃) were
tested for signal intensities at ESI+. The signal intensities of most analytes at 120℃ were
1.0 to 1.3 times higher than those at 130℃, 140℃ and 150℃ (Figure 9, page 51).
Moreover, a lower source temperature caused less heating pressure on the ionization
source. Thus, 120℃ was chosen as the source temperature for ESI+.
3.3. Optimization of sample preparation
The MSPD procedure employed in this study was modified from a previous
validated method developed by our team for analyzing PCPs, NPAHs and OPAHs in
sediment, fish muscle, and liver [66]. Because DEP was added in this study and NPAHs,
OPAHs and some PCPs were removed, some parameters were reevaluated for better
performance: the type and volume of elution solvents, the type of MSPD adsorbents, the
type and amount of cleanup adsorbents, and the volume of final residues for instrumental
analysis.
Regarding elution solvents of MSPD on C18 non-endcapped adsorbent (particle size 40-63 µm, carbon content 23%, SiliCycle, Quebec City, Quebec, Canada), two types of
solvent sequences (methanol and acetone versus methanol and dichloromethane) were
passed through the adsorbent containing spiked standards for comparing the elution efficiency. The combination of methanol and acetone gave eight analytes 1.0-2.0 times
higher elution efficiencies than those of methanol and dichloromethane (DCM) except for acetaminophen, ketoprofen and ibuprofen (1.2-2.4 times lower) (Figure 10, page 52).
The elution efficiencies of DEP and benzophenone were partly influenced by
backgrounds from reagent blanks. Studies have shown that background contamination is
a common problem when analyzing PAEs and UV filters [45, 68]. When background
levels were deducted, the elution efficiencies of methanol/acetone and methanol/DCM
for DEP were 33.5% and 18.7%, respectively; the elution efficiencies of
methanol/acetone and methanol/DCM for benzophenone were 30.7% and 18.7%,
respectively. Overall, the combination of methanol and acetone was chosen.
Elution volume is crucial to the elution efficiency. Because the elution efficiencies
of some analytes were still not ideal using methanol and acetone, such as acetaminophen
(5.05%) and oxybenzone (15.4%), elution volumes of 5 mL and 7.5 mL at each portion
were tested. The elution efficiencies of both volumes were similar on most analytes except for the parabens (4.7-15% better) (Figure 11, page 52). The elution efficiencies
on DEP and benzophenone were partly influenced by backgrounds; when the background
levels were deducted, the elution efficiencies on DEP and benzophenone were about 26%
and 33%, respectively. Increase of the elution volume did not improve the elution on most
analytes, and thus 5 mL aliquots of elution were decided.
A tandem column system was used for cleanup; therefore, the type of elution solvent
combinations (methanol/acetone and methanol/DCM) was also tested on the cleanup
cartridge. The test was done firstly on 6-mL acidic alumina cartridges (bed weight 2 g,
Sigma-Aldrich, St. Louis, MO, USA), which was used as cleanup adsorbent in the previous method [66]. Elution efficiencies of methanol/acetone were 1.1-1.9 times higher
on DEP, DEET, benzophenone and methyl paraben than those of methanol/DCM, but were 1.1-2.2 times lower on acetaminophen, caffeine, ethyl paraben, propyl paraben, and
butyl paraben than those of methanol/DCM. Both combinations could not elute
oxybenzone, ketoprofen and ibuprofen from the alumina adsorbent (Figure 12, page 53).
Ketoprofen and ibuprofen contain a carboxyl group, which may tend to bind with alumina.
Again, the elution efficiencies of DEP and benzophenone were partly influenced by
backgrounds from reagent blanks. When background levels were deducted, the elution
efficiencies on DEP using methanol/acetone and methanol/DCM were 53.5% and 31.2%,
respectively; those of benzophenone were 40.9% and 17.1%, respectively. In brief, for
the most analytes, methanol/DCM did not provide better elution efficiencies than
methanol/acetone. Consequently, methanol/acetone were chosen as the elution solvents
to avoid using chlorinated solvent.
In addition to elution solvent strength and volumes, the retention of adsorbents is
also crucial on elution efficiency. Because elution efficiencies were not good on some
analytes, the combination of C8 for MSPD and silica gel for cleanup was further
investigated. The elution efficiencies on C8 for most analytes were 1.3 to 5.8 times higher
than those on C18 (Figure 13, page 53); the elution efficiencies on silica gel for most
analytes were 18 to 90% higher than those for alumina (Figure 14, page 54). Besides, it
deserves to be mentioned that the elution efficiencies of oxybenzone, ketoprofen and
ibuprofen were improved a lot using silica gel, which were not able to be eluted from
alumina. Once again, the elution efficiencies of DEP and benzophenone were influenced
by backgrounds. When the background levels were deducted, the elution efficiencies for
DEP became 26.2%, 54.0%, 53.5%, and 45.6% on C18, C8, alumina and silica gel,
respectively. Regarding benzophenone, the elution efficiencies became 33.5%, 59.1%,
40.9% and 48.8% on C18, C8, alumina and silica, respectively. Accordingly, for better
elution, C8 and silica gel were selected as MSPD and cleanup adsorbents, respectively.
Regarding the backgrounds of DEP and benzophenone observed in the reagent blank
in above test done on silica gel, silica gel cartridges might be one of the sources. The
background levels of both DEP and benzophenone decreased a lot if silica gel cartridges
were pre-washed with acetone (Figure 15, page 54); the concentrations of DEP reduced,
and the backgrounds of benzophenone was eliminated.
To test if the activity of silica gel was influenced by the pre-wash step with acetone,
we eluted an MSPD cartridge of non-spiked fish muscle sample and collected the
methanol portion and acetone portion of the eluent, respectively. Thereafter, chemical
standards of the analytes were spiked to the methanol portion of the eluent, and then
passed the spiked methanol eluent and acetone portion of eluent sequentially through
silica gel cartridges with and without the pre-wash step, respectively. The peak areas of
analytes in the final residues of the two groups were similar (Figure 16, page 55), which
demonstrated that the activity of silica gel was not affected by the pre-wash step.
The amount of silica gel was increased to 2 g to remove pigment completely from
liver samples. As indicated in Figure 17 (page 55), the extract cleaned up by 2-g silica gel
was much cleaner than that by 1 g.
The previous method concentrated the eluents to 100 µL by Savant SPD 1010
SpeedVac and then injected four microliters of the samples onto the UPLC-MS/MS for
analysis. However, it was difficult to quantify the small volume precisely, which might
influence the precision of the method. We modified the protocol and evaporated the eluent to nearly dry and then reconstituted it with 100 μL of methanol for better control on the
final volume. Nevertheless, some viscous and light yellow material presented when the eluents of muscle samples were concentrated to around 20-50 μL. In addition, the
residues were not able to be completely reconstituted with 100 μL of methanol, which
looked like lipids. To solve the above problem, muscle samples were not concentrated to
nearly dry and the volume of final residues was increased to 250 μL.
Regarding liver samples, the eluents became difficult to be concentrated to lower
volume when concentrated to nearly 100-200 μL, and it would take much time to
concentrate liver samples to nearly dry, which might cause more analytes to evaporate.
In addition, the sample cleanup of liver was worse than that of muscle. Thus, the volume
of final residues of liver samples was increased to more than 250 μL to avoid time-
consuming concentration and serious matrix effect, and the peak areas of analytes in 500- μL final residues of post-spiked liver samples were compared with those of the two-fold
dilution samples of the 500-μL final residues. All the area ratios of the two-fold dilution
samples to the original samples were > 0.5 but < 1, indicating that two-fold dilution
improved matrix effect limitedly (Figure 18, page 56). For better sensitivity, 500 μL was
chosen as the volume of final residues of liver samples.
In terms of the 500-μL final residues of liver samples, there were some suspended
solids and lipids after storing at 4°C even though the eluents had been filtered by 0.20- μm PTFE syringe filters when the volume was one mL before concentrated to 500 μL. To
remove these materials, the 500-μL residues were centrifuged at 13,200 rpm (16100 × g)
at 4°C for 10 minutes after refrigerated at 4°C overnight. Some solids precipitated after
the centrifugation, but lipids and other solids still suspended on the surface of the final
residues. Thus, only the subnatants (not included the precipitates) were taken and were
filtered again through PTFE filtered (0.20 μm) before analysis.
Regarding the chromatograms of matrix blank samples of liver (Figure 3, page 48),
there was a peak (RT = 0.62 minute) near acetaminophen (RT = 0.74 minute) and
interfered with the quantification. The signal intensity of this unknown signal in the
chromatogram of the second product ion of acetaminophen was about 21 times lower than
that in the chromatogram of the first product ion. To reduce its impact on quantifying
acetaminophen in liver, we used the second abundant product ion as the quantitative ion
and the most abundant product ion as the confirmatory ion, respectively.
3.4. Method validation
The matrix effect factors of analytes in fish muscle and liver ranged from 13.4 to
95.6% and ranged from 4.52 to 61.9%, respectively (Figure 19, page 57; Table 5, page
64). All analytes had lower matrix effect factors in liver than in muscle, indicating that
the matrix effects of all analytes were more serious in liver than those of muscle. The
matrix effect factors of most analytes in liver ranged from 24.3 to 61.9%, and ranged from
70.3 to 95.6% in muscle; therefore, most analytes did not suffer significant matrix effect
in muscle. The matrix effect factors of acetaminophen and caffeine in both muscle (13.4%
and 44.2%) and liver (4.52% and 19.8%), plus that of methyl paraben in liver (18.9%)
were much lower than most of other analytes, which could be attributed to their earlier
elution with other polar compounds from the column.
The extraction efficiencies of analytes in fish muscle and liver were 10.9-76.6% and
1.86-71.2%, respectively (Figure 20, page 57; Table 6, page 64). The extraction
efficiencies of DEP, benzophenone, oxybenzone, propyl paraben, butyl paraben and
ibuprofen were < 40% in both matrices. Most analytes had lower extraction efficiencies
in liver than those of muscle, especially on DEP, ethyl paraben and propyl paraben. The
lower extraction efficiencies of most analytes in liver might be partly because the
concentration time of liver samples was longer than that of muscle samples (about two
extra hours), which might cause more analytes to evaporate. In general, extraction
efficiencies of half of the analytes in muscle ranged from 62.1 to 76.6%, and those of
most analyes in liver ranged from 31.6 to 71.2%.
The IDLs of DEP, analgesics, caffeine, DEET, UV filters and parabens were 4.04 pg, 2.86-8.52 pg (except for ketoprofen at 47.7 pg), 12.5 pg, 0.43 pg, 0.64-17.3 pg, 1.29-
7.63 pg, respectively. The IQLs of DEP, analgesics, caffeine, DEET, UV filters and parabens were 10.6 pg, 9.53-9.57 pg (except for ketoprofen at 159 pg), 22.9 pg, 1.03 pg,
1.95-37.1 pg, 2.14-7.63 pg, respectively. (Table 7, page 65)
The LODs of analytes were 0.57 to 15.0 ng/g w.w. (2.65 to 69.9 ng/g d.w.) for fish
muscle and 4.37 to 104 ng/g w.w. (13.4 to 319 ng/g d.w.) for fish liver, respectively (Table
8, page 66); the LOQs of analytes ranged from 1.04 to 34.9 ng/g w.w. (4.86 to 163 ng/g
d.w.) for fish muscle and ranged from 10.6 to 281 ng/g w.w. (32.4 to 861 ng/g d.w.) for
fish liver, respectively. Both the LODs and LOQs of analytes in liver were higher than
those of muscle, which might be explained by higher ion suppression on all analytes and
lower extraction efficiencies on most analytes in liver. Some analytes had lower LODs or
LOQs than those of most other analytes, which might be attributed to their higher IDLs
or IQLs, lower extraction efficiencies, and lower matrix effect factors. For example,
acetaminophen had highest LOD and LOQ in muscle, which might be explained by its
lowest matrix effect factor (13.4%) and slightly higher IDL and IQL than most of other
analytes; the lower extraction efficiencies of ethyl paraben (3.29%) and butyl paraben (2.72%) in liver than most of other analytes (14.4%-71.2%) might result in their higher
LODs and LOQs in liver than many of other analytes.
Our LODs or LOQs of some analytes were higher than those of some previous
reports, but some were similar to or lower than those of some previous reports. Kwon et
al. determined pharmaceuticals and PCPs in fish livers using LLE with LC-MS, and the
LOQ of oxybenzone was 2.0 times lower than that in our study [69]. Our LOQ of DEP
was similar to that of Cheng et al. (5 ng/g w.w.) who analyzed PAEs in fish muscle using
Soxhlet extraction with GC-MS [48], and slightly higher than that of Guo et al. (2 ng/g
w.w.) who investigated PAEs in seafood using liquid-liquid extraction (LLE) with GC-
MS [70]. The LODs of DEET and methyl paraben in liver, and LODs of ethyl parabens,
propyl paraben and butyl paraben in muscle in our study were 1.4-2.8 times higher than
those of Tanoue et al., and LODs of ethyl parabens, propyl paraben and butyl paraben in liver in our study were 24-74 times higher than those of theirs; however, our LODs of
DEET and methyl paraben in fish muscle were 5.6 times and 2.7 times lower than those
of theirs, respectively [33]. Carmona et al, reported a method analyzing multiple organic
compounds in fish muscle using solid-liquid extraction (SLE) and ultra-high performance
liquid chromatography coupled to tandem mass spectrometry (UHPLC–MS/MS), and the LOQs of parabens and ibuprofen were 1.4-7.5 times higher than those in our study [52].
Ramirez et al. determined pharmaceuticals and PCPs in fish fillet and livers using
ultrasonic extraction with HPLC-MS/MS and GC-MS/MS; their LOD of acetaminophen
in muscle was 3.4 times lower than that in our study while their LOD of acetaminophen
in liver and LODs of caffeine in both matrices were 1.6-3.2 times higher than ours.
Furthermore, their LODs of ibuprofen in muscle and liver were 38 times and 25 times
higher than ours, respectively [5].
Different LODs and LOQs might result from different extraction techniques,
instrumental methods, injection volumes, and definitions of LODs and LOQs. For
instance, we calculated LODs based on the confirmatory ions rather than the most
intensive product ions. Furthermore, this study dealt with analtytes at a wide range of
physical and chemical properties, so we had to comprise parameters of sample treatment
and instrumental analysis.
The relative standard deviations (%RSD) of all analytes were below 20%, and the
%RSDs of most analytes were below 8%, which showed that the method provided good
precision for all analytes. Most of the quantitative biases (%bias) were below 30% in fish
muscle at three spiked levels (62.5, 200, 500 ng/g) and in liver spiked at 1,000 ng/g (Table
9, page 67), indicating that the method offered good accuracy for most analytes. The
%bias of propyl paraben was higher than those of most analytes would result from the
use of methyl paraben-13C6 as its isotope-labeled internal standard rather than its own
isotope internal standard.
The accuracy of some analytes was influenced by the backgrounds from labware or
endogenous amount in matrices. The %bias of DEP at the lowest spiked level and %bias
of methyl paraben at all spiked levels in muscle were higher than those of most analytes
might be due to both backgrounds from labware and endogenous amount in matrices. The
concentrations of DEP and methyl paraben in the reagent blank were 35.5 and 23.6 ng/g,
respectively; the concentrations of DEP and methyl paraben in the matrix blank were 64.3
and 34.6 ng/g, respectively. The %bias of benzophenone at the lowest and medium spiked
level in muscle were higher than those of most analytes, which might be explained by
backgrounds from labware because the concentrations of benzophenone in the reagent
blank and the matrix blank were similar (78.7 ng/g and 71.3 ng/g). When the background
levels were deducted, the %biases of these analytes were below 30%.
Chapter 4. Conclusions
This study developed and validated a method for simultaneously determining DEP
and 11 PCPs in fish muscle and livers. Although the cleanup was not so effective on liver
samples, the method offered reproducible analytical results on all analytes with %RSD
below 20% and accurate analytical results on most of the analytes with quantitative biases
below 30% by using isotope dilution techniques. In addition, the LODs of most analytes
ranged from to sub-ng/g to tens of ng/g w.w., and some LODs were lower than those in
other previous reports, indicating that better sensitivity was acquired for some analytes.
Furthermore, the MSPD method consumed only small volumes of organic solvents and
did not need expensive devices for extraction. The developed method is able to be applied
to the determination of these compounds in wild fish samples to acquire more information
about the levels of these chemicals in fish tissues and possible health effects on fish.
Since the cleanup effect on liver samples was not ideal with silica gel, the protocol
for preparing liver samples was more complicated than that for muscle samples. Further
studies are desired to improve the cleanup, for example, elimination of lipids by freezing
them in, or tests of other adsorbents such as Enhanced Matrix Removal-Lipid (EMR-
Lipid).
Reference
1. Mankidy, R., Wiseman, S., Ma, H., and Giesy, J.P., Biological impact of phthalates. Toxicol Lett, 2013. 217(1): p. 50-58.
2. Brausch, J.M. and Rand, G.M., A review of personal care products in the aquatic environment: environmental concentrations and toxicity. Chemosphere, 2011.
82(11): p. 1518-1532.
3. Daughton, C.G. and Ternes, T.A., Pharmaceuticals and personal care products in the environment: Agents of subtle change? Environ Health Persp, 1999. 107: p.
906-938.
4. Jackson, J. and Sutton, R., Sources of endocrine-disrupting chemicals in urban wastewater, Oakland, CA. Sci Total Environ, 2008. 405(1-3): p. 153-160.
5. Ramirez, A.J., Brain, R.A., Usenko, S., Mottaleb, M.A., O'Donnell, J.G., Stahl, L.L., Wathen, J.B., Snyder, B.D., Pitt, J.L., Perez-Hurtado, P., Dobbins, L.L., Brooks, B.W., and Chambliss, C.K., Occurrence of pharmaceuticals and personal care products in fish: results of national pilot study in the United States. Environ Toxicol Chem, 2009. 28(12): p. 2587-2597.
6. Du, B., Haddad, S.P., Luek, A., Scott, W.C., Saari, G.N., Kristofco, L.A., Connors, K.A., Rash, C., Rasmussen, J.B., Chambliss, C.K., and Brooks, B.W., Bioaccumulation and trophic dilution of human pharmaceuticals across trophic positions of an effluent-dependent wadeable stream. Philos Trans R Soc Lond B Biol Sci, 2014. 369(1656): p. 10.
7. Feng, Y.L., Liao, X., Grenier, G., Nguyen, N., and Chan, P., Determination of 18 phthalate metabolites in human urine using a liquid chromatography-tandem mass spectrometer equipped with a core–shell column for rapid separation. Anal Methods-UK, 2015. 7(19): p. 8048-8059.
8. Schettler, T., Human exposure to phthalates via consumer products. Int J Androl, 2006. 29(1): p. 134-139.
9. Net, S., Sempere, R., Delmont, A., Paluselli, A., and Ouddane, B., Occurrence, fate, behavior and ecotoxicological state of phthalates in different environmental matrices. Environ Sci Technol, 2015. 49(7): p. 4019-4035.
10. Kong, M., Song, Y., Zhang, Y., Liu, R., Wei, J., and Zheng, L., Fate of phthalate esters in municipal wastewater treatment plant and their environmental impact.
Water Sci Technol, 2016. 73(6): p. 1395-400.
Adeniyi, A., Phthalate ester plasticizers in freshwater systems of Venda, South Africa and potential health effects. Water Sa, 2010. 36(1): p. 117-125.
12. Yuan, S.Y., Liu, C., Liao, C.S., and Chang, B.V., Occurrence and microbial degradation of phthalate esters in Taiwan river sediments. Chemosphere, 2002.
49(10): p. 1295-1299.
13. Zeng, F., Cui, K., Xie, Z., Liu, M., Li, Y., Lin, Y., Zeng, Z., and Li, F., Occurrence of phthalate esters in water and sediment of urban lakes in a subtropical city, Guangzhou, South China. Environ Int, 2008. 34(3): p. 372-380.
14. Huang, P.C., Tien, C.J., Sun, Y.M., Hsieh, C.Y., and Lee, C.C., Occurrence of phthalates in sediment and biota: relationship to aquatic factors and the biota- sediment accumulation factor. Chemosphere, 2008. 73(4): p. 539-544.
15. Valton, A.S., Serre-Dargnat, C., Blanchard, M., Alliot, F., Chevreuil, M., and Teil, M.J., Determination of phthalates and their by-products in tissues of roach (Rutilus rutilus) from the Orge river (France). Environ Sci Pollut Res Int, 2014.
21(22): p. 12723-12730.
16. Chen, X., Xu, S., Tan, T., Lee, S.T., Cheng, S.H., Lee, F.W., Xu, S.J., and Ho, K.C., Toxicity and estrogenic endocrine disrupting activity of phthalates and their mixtures. Int J Environ Res Public Health, 2014. 11(3): p. 3156-3168.
17. Staples, C.A., Peterson, D.R., Parkerton, T.F., and Adams, W.J., The environmental fate of phthalate esters: A literature review. Chemosphere, 1997.
35(4): p. 667-749.
18. Arnot, J.A. and Gobas, F.A.P.C., A review of bioconcentration factor (BCF) and bioaccumulation factor (BAF) assessments for organic chemicals in aquatic organisms. Environ Rev, 2006. 14(4): p. 257-297.
19. Miege, C., Choubert, J.M., Ribeiro, L., Eusebe, M., and Coquery, M., Fate of pharmaceuticals and personal care products in wastewater treatment plants-- conception of a database and first results. Environ Pollut, 2009. 157(5): p. 1721- 1726.
20. Oulton, R.L., Kohn, T., and Cwiertny, D.M., Pharmaceuticals and personal care products in effluent matrices: A survey of transformation and removal during wastewater treatment and implications for wastewater management. J Environ Monit, 2010. 12(11): p. 1956-1978.
21. Verlicchi, P., Al Aukidy, M., and Zambello, E., Occurrence of pharmaceutical compounds in urban wastewater: removal, mass load and environmental risk after
22. Blair, B.D., Crago, J.P., Hedman, C.J., Treguer, R.J., Magruder, C., Royer, L.S., and Klaper, R.D., Evaluation of a model for the removal of pharmaceuticals, personal care products, and hormones from wastewater. Sci Total Environ, 2013.
444: p. 515-521.
23. Blair, B.D., Crago, J.P., Hedman, C.J., and Klaper, R.D., Pharmaceuticals and personal care products found in the Great Lakes above concentrations of environmental concern. Chemosphere, 2013. 93(9): p. 2116-2123.
24. Subedi, B., Du, B., Chambliss, C.K., Koschorreck, J., Rudel, H., Quack, M., Brooks, B.W., and Usenko, S., Occurrence of pharmaceuticals and personal care products in German fish tissue: A national study. Environ Sci Technol, 2012.
46(16): p. 9047-9054.
25. Yang, Y., Ok, Y.S., Kim, K.H., Kwon, E.E., and Tsang, Y.F., Occurrences and removal of pharmaceuticals and personal care products (PPCPs) in drinking water and water/sewage treatment plants: A review. Sci Total Environ, 2017. 596-597:
p. 303-320.
26. Xie, Z., Lu, G., Liu, J., Yan, Z., Ma, B., Zhang, Z., and Chen, W., Occurrence, bioaccumulation, and trophic magnification of pharmaceutically active compounds in Taihu Lake, China. Chemosphere, 2015. 138: p. 140-147.
27. Moreno-Gonzalez, R., Rodriguez-Mozaz, S., Huerta, B., Barcelo, D., and Leon, V.M., Do pharmaceuticals bioaccumulate in marine molluscs and fish from a coastal lagoon? Environ Res, 2016. 146: p. 282-298.
28. Beydoun, M.A., Gamaldo, A.A., Beydoun, H.A., Tanaka, T., Tucker, K.L., Talegawkar, S.A., Ferrucci, L., and Zonderman, A.B., Caffeine and alcohol intakes and overall nutrient adequacy are associated with longitudinal cognitive performance among U.S. adults. J Nutr, 2014. 144(6): p. 890-901.
29. Alvarez-Munoz, D., Rodriguez-Mozaz, S., Maulvault, A.L., Tediosi, A., Fernandez-Tejedor, M., Van den Heuvel, F., Kotterman, M., Marques, A., and Barcelo, D., Occurrence of pharmaceuticals and endocrine disrupting compounds in macroalgaes, bivalves, and fish from coastal areas in Europe. Environ Res, 2015.
143(Pt B): p. 56-64.
30. Wang, J. and Gardinali, P.R., Analysis of selected pharmaceuticals in fish and the fresh water bodies directly affected by reclaimed water using liquid chromatography-tandem mass spectrometry. Anal Bioanal Chem, 2012. 404(9): p.
2711-2720.
