Journal of Chromatography A, 1261 (2012) 99 106
Contents lists available at SciVerse ScienceDirect
Journal of Chromatography A
jou rn al h om epage: www.elsevier.com/locat e/chroma
Optimization of solid phase microextraction for non-lethal in vivo determination
of selected pharmaceuticals in fish muscle using liquid chromatography mass
spectrometry
Oluranti P. Togundea, Ken D. Oakesb, Mark R. Servosb, Janusz Pawliszyna,"
a
Department of Chemistry, University of Waterloo, Waterloo, Ontario N2L 3G1, Canada
b
Department of Biology, University of Waterloo, Waterloo, Ontario N2L 3G1, Canada
a r t i c l e i n f o a b s t r a c t
Article history:
A new thin film microextraction (TFME) configuration based on C18 thin film was optimized to improve
Available online 26 July 2012
solid phase microextraction (SPME) sensitivity and extraction kinetics for in vivo determinations of trace
pharmaceuticals in fish tissue. Optimization of SPME involved increasing sampler surface area and vol-
Keywords:
ume of extraction phase to improve performance within in vitro applications such as agarose gel matrix
SPME
and spiked fish tissue, as well as in vivo determinations of pharmaceuticals in fish exposed to municipal
Thin film
wastewater. Rainbow trout (Oncorhynchus mykiss) were exposed for 4 d to effluents from a municipal
Pharmaceuticals
wastewater plant that was treated under three different configurations of a pilot plant. Fathead min-
In vivo sampling
now (Pimephales promelas) were caged for 14 d upstream and downstream of municipal wastewater
Fish
effluent discharges along Grand River at near Kitchener, ON. TFME and regular C18 fibers were inserted
into the dorsal-epaxial muscle of rainbow trout and fathead minnow, respectively, for 30 min sampling
intervals. Sample extracts obtained from fibers were desorbed in methanol:water (3:2) for 90 min under
1500 rpm agitation and analyzed using liquid chromatography coupled with tandem mass spectrometry
(LC/MS MS) to determine pharmaceutical bioconcentration. Most target pharmaceuticals were detected
in the wastewater with the exception of norfluoxetine and paroxetine. C18 TFME phase successfully
quantified fluoxetine, venlafaxine, sertraline, paroxetine, and carbamazapine in muscle of living fish
at concentrations ranging from 1.7 to 259 ng/g. Reproducibility of the method in spiked fish muscle
was 9 18% RSD with limits of detection and quantification ranging from 0.08 to 0.21 ng/g and 0.09 to
0.64 ng/g (respectively) for the analytes examined. Ibuprofen and gemfibrozil were not detected in fish
tissues, likely due to rapid excretion leading to low rates of bioconcentration. This study demonstrates
improved in vivo SPME sensitivity (recovery) and extraction rates during pre-equilibrium sampling of
target pharmaceuticals in fish tissue due to the improved C18 extraction phase geometry.
� 2012 Elsevier B.V. All rights reserved.
1. Introduction been proposed and utilized to determine the concentrations of
environmental pharmaceuticals in fish [10 15]. Traditional meth-
Recently, there has been sustained global scientific interest sur- ods include solid phase extraction (SPE) [8], liquid extraction (LE)
rounding the detection and potential impact of pharmaceuticals [10], and matrix solid phase dispersion (MSPD) [11]. However,
released into the environment on exposed organisms [1 4]. The none of these approaches can be applied in vivo to determine
detection of pharmaceutical residues in surface waters influenced concentrations of organic contaminants in fish. Traditional ana-
by municipal wastewater effluents (MWWEs) and agricultural lytical methods require lethal sampling of a large number of fish
runoff has raised concerns for regulatory agencies around the to obtain sufficient tissue and variance estimates to accurately
globe [5 7]. As a consequence of the continuous release of determine bioconcentration rates. From the perspective of animal
human pharmaceuticals in MWWEs, fish inhabiting these receiv- care, such sampling procedures are undesirable, and in some juris-
ing environments can bioconcentrate considerable amounts of dictions, unacceptable. Recently, SPME has proven itself a simple
these bioactive compounds in their tissues, despite their relatively alternative analytical method that is particularly relevant for the
water soluble properties [8,9]. Various analytical methods have non-lethal in vivo sampling of pharmaceutical residues in fish mus-
cle [16 20]. However, the small extraction surface area and volume
of PDMS fibers limited sensitivity and presented challenges for
" detecting fairly polar pharmaceuticals (i.e. carbamazepine) or those
Corresponding author. Tel.: [phone]; fax: [phone].
E-mail address: [e-mail] (J. Pawliszyn). with low bioconcentration factors under pre-equilibrium sampling
0021-9673/$  see front matter � 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.chroma.2012.07.053
100 O.P. Togunde et al. / J. Chromatogr. A 1261 (2012) 99 106
conditions. To overcome these limitations, a new SPME configura- from Toronto Research Chemical (ON, Canada). Isotopically
tion based on C18 thin film geometry (a modification of SPME) was labeled atorvastatin-d5, atrazine-d5, BPA-d16, carbamazepine-
proposed, with potential to improve the sensitivity of the method. d10, diazepam-d5, diclofenac-d4, fluoxetine-d5, gemfibrozil-d6,
13
Furthermore, the proposed SPME configuration has the potential to ibuprofen-d3, and C-naproxen-d3 were purchased from CND
improve extraction kinetics under pre-equilibrium sampling con- Isotope Inc (Point-Claire, QC, Canada) while sertraline-d3 and
ditions relative to traditional PDMS or C18 fiber configurations. paroxetine-d4 were purchased from Toronto Research Chemical.
The objective of this study is to investigate and optimize the new Chemical stock solutions were prepared in methanol and stored
ć%
octadecyl (C18) thin film geometry to improve SPME sensitivity and at -20 C while working solutions were diluted aliquots of these
in vivo extraction kinetics when quantifying pharmaceuticals in stocks. Dilution water was obtained from a Barnstead Nanopure
fish. diamond UV water purification system deionized to 18 . Ace-
tonitrile (HPLC grade), methanol (HPLC grade), and glacial acetic
acid were purchased from Fisher Scientific (Ottawa, ON, Canada).
2. Theoretical considerations
Phosphate-buffered saline (PBS) solution, pH 7.4 was prepared by
The volume and type of the extraction phase is a critical param- dissolving 8.0 g of sodium chloride, 0.2 g of potassium chloride,
0.2 g of potassium phosphate, and 1.44 g of sodium phosphate in
eter modifying the distribution constant between the extraction
1 L of purified water and by adjusting the pH to 7.4 using 1 M
phase and sample matrix, as well as the amount of the analyte
sodium hydroxide (NaOH). The C18 thin film extraction phase
extracted. SPME extraction kinetics and sensitivity (recovery) can
(45 m, 1.5 cm coating length) and traditional C18 fibers (45 m,
be improved by changing the configuration (geometry) of the
1.5 cm coating length) were obtained from Supelco (Bellefonte,
extraction phase. Theoretically, the initial rate of SPME extraction
PA, USA). A 10 mg/L mixed solution containing all the pharma-
is directly proportional to the surface area of the extraction phase
ceuticals was obtained by diluting the individual stocks with
[21] as described in Eq. (1):
D methanol.
dn A
s
= Cs (1)
dt 1

where dn/dt is the rate of extraction, Ds is the diffusion coefficient of 3.2. Fish exposure
the analyte in the sample matrix, A is the surface area of the extrac-
tion phase, and 1
is the thickness of the boundary layer surrounding Immature rainbow trout (Oncorhynchus mykiss) and was pur-
the extraction phase and Cs is the concentration of analyte. Based chased from Silvercreek Aquaculture (Erin, ON, Canada) and
on this equation, extraction kinetics can be improved (particularly fathead minnow (Pimephales promelas) was purchased from Sil-
during pre-equilibrium extraction) when the traditional cylindri- hanek Baitfish (Bobcaygeon, ON, Canada). All animal experimental
cal fiber geometry is changed to a thin film configuration which procedures were approved by the local Animal Care Committee
will provide a faster extraction rate for the analytes of interest. at the University of Waterloo (AUP # s 04-24, 08-08). For the
SPME sensitivity is a function of extraction phase type and volume laboratory exposure, rainbow trout were acclimatized to labo-
[22,23], as the amount of analyte extracted from a sample matrix at ratory conditions in de-chloraminated municipal water and fed
equilibrium is proportional to the volume of the extraction phase every other day with 2.0 Pt floating commercial trout ration
as shown in Eq. (2): (Martin s Feed Mill, Ontario) until 4 d prior to the onset of the
experiment. Subsequently, the fish were exposed to municipal
C0KfsVsVf
n = (2)
effluent diluted with de-chloraminated municipal water. In the
KfsVf + Vs
field caging experiment, fathead minnow (P. promelas), a small-
where n is the amount of analyte extracted at equilibrium, C0 is the
bodied fish were caged in the Grand River watershed (southern
original concentration of the analyte, Kfs is the partition coefficient
Ontario, Canada) adjacent the Doon Wastewater Treatment Plant
between the extraction phase and the sample matrix, Vs is the vol- (43ć%24 03.29 N; 80ć%25 12.04 W) at 2 upstream and 3 downstream
ume of sample and Vf is the volume of extraction phase. Eq. (2) can
sites for 14 d in October 2010. Following the exposure, pharma-
be further simplified to Eq. (3) when sampling large volumes, as in
ceuticals extracted from the fish tissues were desorbed in the
Eq. (3), such that KfsVf Vs:
desorption solution (methanol:water, 3:2) and analyzed by liquid
chromatography tandem mass spectrometry. The quantification
n = C0KfsVf (3)
of target pharmaceuticals extracted from free-moving rainbow
Based on Eq. (3), the amount of analyte extracted in fish can trout was performed using a well-established kinetic calibration
be related to the volume of the extraction phase and parti- approach [24,25]. Kinetic calibration is a pre-equilibrium calibra-
tion co-efficient between the extraction phase and fish sample tion approach utilizing known amounts of preloaded deuterated
matrix. Therefore, the use of C18 thin film geometry will theoret- analyte desorbing from the extraction phase, to calibrate the rate
ically improve pre-equilibrium SPME extraction kinetics relative of adsorption of the target analytes from the fish tissue to the
to traditional C18 coated fiber geometries for in vivo detection of extraction phase (based on the symmetric relationship between
pharmaceuticals in fish. extractions and desorption kinetics). The concentration of analytes
in fish tissue can be determined by Eq. (4):
3. Experimental
n
Cs = (4)
KVf (1 - Q/q0)
3.1. Chemicals and materials
All reagents and pharmaceuticals standards were of high- where Cs is the concentration of target analytes in the sample,
est purity grade available. Fluoxetine, diazepam, nordiazepam, n is the amount of the extracted analyte under pre-equilibrium
and diazepam-d5 were purchased as certified standards from sampling time, K is the distribution coefficient of the analyte
Cerilliant Corp (Round Rock, TX, USA). Gemfibrozil, ator- between the fiber coating and the sample matrix, Vf is the vol-
vastatin, ibuprofen, carbamazepine, diclofenac, naproxen, and ume of the fiber coating, Q is the standard remaining in the SPME
bisphenol-A (BPA) were obtained from Sigma Aldrich (Oakville, fiber coating after sampling, and q0 is the preloaded standard on
ON, Canada), while paroxetine and sertraline were purchased the fiber.
O.P. Togunde et al. / J. Chromatogr. A 1261 (2012) 99 106 101
ć%
water for 30 min. The metal blade is then dried in an oven at 150 C
for 30 min. The slurry of C18 particles (5 m discovery silica-based-
C18 particles, Supelco, PA) is made with a mixture of proprietary
biocompatible binder PAN (10, w/w) and N,N-dimethylformamide
(DMF) purchased from Caledon Laboratories (ON, Canada). Coatings
(mass of 0.28 ą 0.05 g) were immobilized on the metal fiber core
using a flask type sprayer connected to a nitrogen gas line (250 mL
Erlenmeyer flask with a sprayer head). After spraying, the coated
ć%
blade was dried at 180 C for 2 min, with the spraying and drying
processes repeated at least three more times until a desired coating
thickness was achieved. C18 fiber is cylindrical (metal fiber core has
200 m in diameter) while C18 thin film is rectangular with surface
area of 0.625 cm2. Fig. 1 illustrates the relative physical dimensions
of both C18 configurations.
3.4. Instrumental analysis LC/MS/MS
Analyses were performed on an AB-Sciex 3200 QTrap Mass
Spectrometer equipped with a Turbo Ion Spray source (Applied
Biosystems Sciex, Foster City, CA, USA). Liquid chromatography
(LC) was performed on an HP1100 HPLC system (Agilent Technolo-
gies) equipped with a degasser, binary pump, an autosampler, and
a column oven. Chromatographic separation was performed on a
Zorbax Eclipse XDB C18 (150 mm � 21 mm, 3.5 m) column which
was preceded by a C18 guard column at a flow rate of 0.8 mL/min
with mobile phase A (95% water, 5% MeOH, 0.1% acetic acid) and
B (95% MeOH, 5% water, 0.1% acetic acid). The injection volume
for analysis of all the samples was 20 L. The elution gradient was
programmed as follows: mobile phase B was increased from 10%
to 50% over 0.5 min and 50% to 100% over 7.5 min, held at 100%
for 2 min, and then reduced to 10% over 1 min. Analyst� version
1.4.2 software was used for the data analysis. Quality control sam-
ples (10 ng/mL of target compounds) were utilized at the beginning
and end of each run to ensure instrumental stability. All analytes
were analyzed using selective reaction monitoring (SRM) of the
transitions with electrospray ionization (ESI). Optimal MS/MS tran-
sitions for the target analytes were determined with either positive
or negative ionization mode. The precursor to product ion transi-
tion monitored for screening and quantitation are summarized in
Table S1.
3.5. In vitro evaluation of performance of thin film and fiber
geometries in spiked fish tissue
Fig. 1. Schematic diagram of the C18 SPME extraction phase geometries used in this
study. Both configurations utilize 45 m coating thicknesses over a 1.5 cm coating
Extraction efficiency of the two SPME configurations (C18 fiber
length. (A) C18 cylindrical fiber (core of 200 m diameter, length of 40 mm, extrac-
and C18 thin film) was evaluated in spiked fish tissue under pre-
tion phase surface area of 8.1 mm2 (shown in hypodermic needle); (B) C18 thin film
equilibrium sampling conditions. Evaluation of recovery by the
extraction phase (core of 40 mm � 2.1 mm � 0.07 mm and extraction phase surface
area of 65.4 mm2. extraction phases was carried out using homogenized (TeflonTM
homogenizer at 1100 rpm for 20 min) rainbow trout dorsal-epaxial
muscle (4.0 ą 0.2 g) spiked with 200 L of 2 mg/L solutions of target
3.3. Thin film extraction phase preparation compounds. Subsequently, the 100 ng/g spiked muscle sample was
vortexed for another 20 min to ensure complete analyte mixture
ć%
Traditional C18 SPME fibers were obtained from Supleco (Belle- within the tissue prior to 2 h incubation at 4 C. After incubation,
fonte, PA, USA) and consisted of a flexible stainless steel core with extraction was performed for 24 h in the spiked fish muscle sam-
a 1.5 cm coating immobilized on the bottom portion (Fig. 1a). In ple using C18 fiber and thin film extraction phases. After the 24 h
contrast, for the thin film extraction phase configuration, a slurry of extraction, both C18 extraction phase configurations were washed
biocompatible binder polyacrylonitrile (PAN) containing 5 m C18- with distilled water for 5 sec, and gently wiped with a Kimwipe�
coated porous silica particles were immobilized on the metal core tissue. The C18 fibers were immediately desorbed in 100 L of
(Fig. 1b) with coating thickness of 45 m and 1.5 cm coating length. desorption solvent (3:2 methanol:water) for 90 min at 1000 rpm
The flattened metal support for the thin film configuration was using a multi-tube vortex (model DVX-2500; VWR International,
obtained from Professional Analytical System Technology (PAS, Mississauga). For the thin film extraction phase, 300 L of the
Magdala, Germany), the surface upon which the C18 particles were same desorption solvent removed analytes from the extraction
immobilized as described in detail elsewhere [26]. Briefly, the blade phase. The extraction and desorption conditions are summarized
metal is etched with concentration HCl for 1 h prior to rinsing with in Table S2.
102 O.P. Togunde et al. / J. Chromatogr. A 1261 (2012) 99 106
3.6. Extraction kinetics of pharmaceuticals solvent (60:40 methanol:water) for 90 min; an aliquot of desorp-
tion solution was injected for LC/MS/MS instrumental analysis.
Extraction kinetics of both extraction phase configurations for After removal of the extraction phase from their muscle tissue,
drugs utilized in this study were determined in spiked 1% (w/v) fish were killed by spinal severance in accordance with protocols
agarose gel samples due to limited availability of fish muscle and approved by our local Animal Care Committee. The detailed work-
previous use of 1% (w/v) agarose gel in other studies determining flow for in vivo sampling using C18 thin film is summarized in
in vitro mass transfer phenomenon [22]. Agarose gel was used for Figs. S2 and S3.
in vitro study as both the gel and the fish muscle are semi-solids
of comparable porosity and tortuosity allowing free diffusion of
drugs (along their concentration gradient) with mass transfer gov- 3.8. Field caging of fathead minnow in municipal wastewater
erned by Fick s law in both sample matrixes [27,28]. The detailed
gel preparation procedure has been discussed elsewhere [17]. In Fathead minnow (P. promelas), a small-bodied fish
brief, agarose was weighed and dissolved in hot phosphate buffer (5.58 ą 0.03 cm; 1.43 ą 0.02 g, n = 300) were caged in the Grand
(pH 7.4) and allowed to solidify at room temperature for 3 h. Just River watershed (southern Ontario, Canada) adjacent the Doon
prior to solidification, a known volume of agarose gel was spiked Wastewater Treatment Plant (43ć%24 03.29 N; 80ć%25 12.04 W)
(100 ng/mL, n = 3) with a mixture of the selected drugs and vor- at 2 upstream and 3 downstream sites for 14 d in October 2010.
texed for 10 min in a 4 mL vial to homogenize the drug distribution The most upstream of the two reference sites is located 1.2 km
in the gel. After gel solidification, pre-equilibrium extraction time upstream of the Doon municipal effluent outfall, but 19.45 riverine
profiles were determined (20, 30, 40, 50, and 60 min) using C18 km downstream of the municipal wastewater discharge from
fibers and C18 thin film extraction phases, which simultaneously the City of Waterloo. The Waterloo WWTP serves a population
performed extractions from different vials for each time point. of more than 120,055 and the Kitchener MWWTP serves more
After each interval, both extraction phases were removed from than 190,000 individuals with daily discharge of 23,802 m3/d
their respective spiked gels, rinsed gently with distilled water, and 77,768 m3/d respectively. The second reference was 0.5 km
and patted dry with a Kimwipe�. Immediately thereafter, the upstream of the Doon outfall. The three downstream stations were
fibers and thin film extraction phases were desorbed in desorption 0.5, 1.7, and 5.6 km below the Doon wastewater effluent release,
solvent (60:40 methanol:water) for 90 min as previously described. respectively. At each site, two cages (RubbermaidTM containers)
Extracts from both C18 configurations at each extraction interval were deployed, each containing two commercial baitfish buckets
were quantified by LC/MS/MS. (FlowTroll�, Frabill Inc, Jackson, WI) holding 15 fish each (30
fish/cage). The Rubbermaid containers were perforated on all
3.7. Laboratory exposure of rainbow trout to wastewater effluent surfaces with 2 cm holes and contained a 60 � 60 cm concrete
patio stone beneath the bait buckets for weighting (Fig. S4). The
The juvenile rainbow trout used in this study (14.4 ą 0.34 cm; cages were anchored to the substrate with t-posts fastened to
25.4 ą 1.46 g, mean ą SE, n = 40) were acclimated to laboratory cable running through a homemade pipe frame. This caging design
conditions in 1400 L holding tanks continuously receiving de- exposes fish to minimal current, thereby reducing exposure stress
chlorinated clean water and fed every other day with floating from constant swimming, yet allowing water to pass freely through
commercial trout ration (Martin s Feed Mill, Elmira, ON). Accli- the enclosure. After 2 weeks of exposure, fish were anesthetized
mation within the 34 L exposure aquaria (containing previously and sampled as described for the laboratory-based rainbow trout
acclimated clean water) began 4 d prior to the onset of the exposures.
experiment. The experiment began when fish were exposed to
wastewater collected from three pilot treatment plants receiving
raw effluent generated from the City of Burlington (Burling- 4. Results and discussion
ton, ON). Wastewater samples were treated by Conventional
Activated Sludge (CAS), Conventional Activated Sludge with 4.1. Extraction efficiency of the method
Nitrification (CAS-N), or Conventional Activated Sludge with Bio-
logical Nitrifying Reactor (CAS-BNR). A detailed description of The influence of C18 fiber and thin film extraction phase geome-
the properties and process operation of each type of wastewa- try on method extraction efficiency was determined by the amount
ter treatment has been provided elsewhere [29]. The exposure of analyte extracted by each configuration under identical condi-
involved two controls and six different treatments: CAS-20%, CAS- tions. As the extraction phase/sample matrix partition coefficients
50%, CAS-N-20%, CAS-N-50%, CAS-BNR-20%, and CAS-BNR-50%. should be the same for both C18 configurations, the influence
Each 34 L aquaria housed 5 fish in each tank and the expo- of the extraction phase volume on extraction efficiency can be
sure duration was 4 d with exposure medium renewed every directly compared in spiked fish tissue. As shown in Table 1, the
48 h using the de-chlorinated acclimation water as diluents. amount of pharmaceuticals extracted by the thin film extraction
Fig. S1 shows the static laboratory exposure setup used in this phase was significantly higher than that of the fiber geometry,
study. Exposure water quality was checked daily and maintained attributable to its greater volume. According to SPME fundamen-
ć%
at conditions considered optimal for trout (12.5 ą 0.05 C; pH tal principles, as the volume of the extraction phase increases,
8.17 ą 0.06; ammonia 23.5 ą 1.5 g/L) under a 12 h:12 h light:dark the amount of extracted target analyte will increase proportion-
cycle. ally for equilibrium extraction. The C18 thin film geometry can
The thin film extraction phase was preloaded with deteurated extract more analyte than can be extracted by the cylindrical fiber
standard (50 ng/mL in spiked phosphate buffer solution) by direct configuration due to the increased surface area of the thin partic-
extraction under agitation (500 rpm) for 3 h on a shaker platform. ularly under pre-equilibrium sampling. The use of C18 thin film for
At the conclusion of the 4 d exposure period, in vivo sampling by in vitro sampling of selected pharmaceuticals in spiked fish dorsal-
thin film microextraction (TFME) was performed in dorsal-epaxial epaxial muscle demonstrates that both extraction efficiency and
muscle of anaesthetized (0.1% ethyl 3-amino benzoate methane- method sensitivity in terms of recovery can be enhanced compared
sulfonate) fish for 30 min (under pre-equilibrium conditions). After to traditional C18 fiber configurations. In both extraction phases,
30 min, the extraction phases were removed from fish muscle acceptable reproducibility (4 22% RSD) was achieved for the suite
and rinsed with distilled water prior to immersion in desorption of pharmaceuticals examined (Table 1).
O.P. Togunde et al. / J. Chromatogr. A 1261 (2012) 99 106 103
Table 1
Relative extraction efficiencies of C18 thin film and cylindrical fiber configurations using fish tissue spiked with 100 ng/g analytes of interest.
Analytes Fiber Thin film Ratio of mTFME/mSPME
Amount extracted (ng) RSD % (n = 4) Amount extracted (ng) RSD % (n = 4)
Carbamazepine 0.55 4 4.62 13 8
Fluoxetine 0.02 5 0.17 14 9
Diazepam 0.24 16 2.72 17 11
Norfluoxetine 0.03 22 0.19 11 6
Velanfaxin 0.03 11 0.33 16 11
Ibuprofen 0.31 5 2.38 9 8
Gemfibrozil 0.03 4 0.36 18 12
mTFME is the amount of analytes extracted by thin film extraction phase and mSPME is the amount of analytes extracted by solid phase fiber.
4.2. Pre-equilibrium extraction kinetics film extraction geometries would improve extraction kinetics of
pharmaceuticals, particularly under pre-equilibrium sampling con-
Based on Eq. (2), the extraction rate is expected to be propor- ditions due to increased surface area leading to faster extraction
tional to the surface area of the extraction phase. The extraction rates (Table S3). In addition, the amount of analyte extracted
kinetics of the thin film and fiber geometries were compared under pre-equilibrium conditions (30 min) by the thin film
in spiked gel with the aim to improve in vivo extraction rates extraction phase is higher than the fiber configuration due to
under pre-equilibrium conditions such as detecting pharmaceu- the larger surface area in contact with the sampled matrix
ticals in fish. The surface area of the rectangular thin film C18 (Table S4).
extraction phase (0.21 cm � 1.5 cm) is approximately 0.625 cm2
while the surface area of a C18 cylindrical fiber is approximately
4.3. In vivo determination of pharmaceuticals in fish
0.08 cm2; therefore, the surface area of the thin film extrac-
tion phase is approximately eight times greater than that of
In this study, the concentration of pharmaceuticals from fish
the fiber. Based upon the pre-equilibrium extraction of selected
muscle exposed for 4 d to wastewater effluent was successfully
drugs from the 100 ng/mL spiked gel (n = 3) for 30 min under
determined using thin film microextraction under pre-equilibrium
static conditions, it is evident that the extraction rate of car- conditions. The physico-chemical properties of the pharmaceuti-
bamazepine, ibuprofen, gemfibrozil and fluoxetine was higher
cals, as well as inter-fish variability in the uptake and inducible
using the C18 thin film configuration relative to the compa- depuration enzymes will affect the bioconcentration of the drugs
rable fiber of the same material (Fig. 2). The use of C18 thin
in fish tissue and contribute to the high fish-to-fish variability
Fig. 2. Extraction kinetics of fluoxetine (A) and carbamazepine (B) in spiked gel (1%, w/v) using C18 thin film and fiber extraction phase geometries. Extraction kinetics of
ibuprofen (C) and gemfibrozil (D) in spiked gel (1%, w/v) using C18 thin film and fiber extraction phase geometries.
104 O.P. Togunde et al. / J. Chromatogr. A 1261 (2012) 99 106
observed in this study (RSD 15 90%, n = 5). Consequently, the con-
centration of the pharmaceuticals in fish muscle is expressed as
minimum to maximum value, as summarized in Table 2. Fish
exposed to 100% effluent experienced high mortality within a few
hours of exposure due to high ammonia concentrations. Again, in
response to ammonia toxicity, fish exposed to 50% CAS-N (v/v) and
50% CAS (v/v) also did not survive the 4 d exposure. Bioconcentra-
tion factors (BCFs) were calculated in surviving fish using the ratio
of the analyte concentration in fish muscle to that of the exposure
water.
Following exposure to each treated wastewater effluent, the
anti-depressant drugs fluoxetine, sertraline and paroxetine were
detected in fish dorsal-epaxial muscle, demonstrating both the
efficacy of the proposed method and the recalcitrant nature of
these compounds. The concentration of fluoxetine detected in
fish muscle ranged from 101 ng/g in 20% CAS (v/v), to 111.9 ng/g
in 50% BNR (v/v) effluent; however, up to 259 ng/g were found
in fish exposed to 20% CAS-N (v/v), thereby suggesting all three
treatment paradigms allowed passage of significant quantities
of this drug for uptake by biota. Fish can readily bioconcen-
trate fluoxetine, as has been previously reported [30] and to
a lesser extent sertraline and paroxetine despite high rates
of prescription and parent compound excretion into wastew-
ater systems of the latter two drugs [31]. Notwithstanding
the relatively high concentrations of sertraline and paroxetine
entering Canadian wastewater treatment plants, final effluent
concentrations are lower than that of the less prescribed fluox-
etine, suggesting sertraline and paroxetine are rapidly removed
during wastewater treatment [31]. Fluoxetine, despite lower
prescription rates and relatively higher rates of metabolism in
humans, is frequently found in aquatic environments, and is
taken up by biota, presenting a potential environmental risk
[30,31].
All targeted pharmaceuticals were detected in the treated
municipal wastewater effluents used in this study, as summa-
rized in Table 2. The pharmaceuticals quantified in wastewater
were also detected in exposed fish muscle, with the exceptions of
gemfibrozil and ibuprofen. Sertraline was detected in fish dorsal-
epaxial muscle tissue up to 160.8 ng/g in 20% CAS (v/v) effluent,
along with paroxetine up to 18.8 ng/g. The detection of sertraline
in tissues is consistent with earlier reports in caged fish studies
where fathead minnow (P. promelas) accumulated 3.83 ą 1.8 g/kg,
although paroxetine was not bioconcentrated in that study [31].
This is somewhat surprising due to modeled rates of removal
during wastewater treatment of 85% and 28% for sertraline and
paroxetine, respectively (Table S5). Carbamazepine bioconcentra-
tion was low (<1.7 ng/g) which is a reflection of its relatively
high water solubility (log Kow 2.25) [32] and persistence (2.9%
removal during wastewater treatment). Similarly, the highly pre-
scribed antidepressant venlafaxine is predicted to be only sparingly
removed during wastewater treatment (8.9% removal) [32] and is
dominantly excreted as the parent compound from urine. How-
ever, as venlafaxine is fairly water soluble (log Kow of 3.28 [32],
excreted unmodified), its rapid excretion in fish also precludes
significant bioconcentration, although both venlafaxine and ser-
traline were detected in fish tissues in the 1 3 ng/g range in
other studies [31]. Gemfibrozil and ibuprofen were not detected
in muscle of fish exposed to treated effluent; as both pharma-
ceuticals are acidic drugs, they are easily be ionized in the pH
range (8 9) of the effluent, limiting their bioconcentration rela-
tive to that of their neutral form [30,31]. No target pharmaceuticals
were detected in the control fish, and for all compounds studied,
analyte concentrations in tissues after 4 d of exposure may have
been subjected to metabolism by inducible enzymes systems or
eliminated through conjugation with more hydrophilic moieties
[33 35].
NA
0.62
ą
0.01
NA
0.22
ą
0.02
NA
0.70
ą
0.01
NA
BNR 50% effluent
muscle (ng/g)
n = 3
ą
SD
NA
2.3 18.8
<0.06
0.30
ą
0.01
NA
<0.09
0.11
ą
0.01
NA
<0.84
0.23
ą
0.01
NA
<0.40
BNR 20% effluent
muscle (ng/g)
n = 3
ą
SD
NA
2.5 5.73
<0.06
0.29
ą
0.01
NA
<0.09
0.13
ą
0.02
NA
<0.84
0.23
ą
0.01
NA
<0.40
CAS 20% effluent
muscle (ng/g)
n = 3
ą
SD
NA
3.4 16.7
<0.06
0.23
ą
0.01
NA
<0.09
0.28
ą
0.03
NA
<0.40
1.30
ą
0.05
NA
<0.84
m
0.6 1.7
0.75
ą
0.03
1.0 2
0.4 0.5
0.81
ą
0.01
1
0.5 4.9
0.86
ą
0.01
5.0 49
0.2 2.4
0.23
ą
0.01
1.0 12
Conc. in fish
Water conc. ng/mL,
BCF
Conc. in fish
Water conc. ng/mL,
BCF
Conc. in fish
Water conc. ng/mL,
BCF
Conc. in fish
Water conc. ng/mL,
BCF
muscle (ng/g)
n = 3
ą
SD
Analytes
CAS-N 20% effluent
Fluoxetine
36.6 259
0.50
ą
0.05
73 518
28.8 101.1
0.45
ą
0.01
58 202
29.1 65.2
0.83
ą
0.01
37 79
1.7 119.9
0.97
ą
0.01
2 133
Velanfaxin
<0.09
Sertraline
7.5 160.8
0.36
ą
0.01
19 402
1.0 46.1
0.41
ą
0.02
20 334
13.5 39.5
0.37
ą
0.01
34 99
11.2 75.8
0.12
ą
0.01
112 758
Paroxetine
5.3 13.4
<0.06
CBZ
Ibuprofen
<0.84
Gemfibrozil
<0.40
Table 2
The measured bioconcentrations factors of target pharmaceuticals in dorsal-epaxial muscle (ng/g, n = 5) of rainbow trout (Oncorhynchus mykiss) using thin film SPME. The concentration is expressed as min max. Note that fish
exposed to 50% CAS-N (v/v) and 50% CAS (v/v) did not survive the 4 d exposure due to high concentrations of ammonia found in the effluent in these treatments.
CBZ, carbamazepine; BCF
, bioconcentration factor in muscle
Velanfaxin was detected in all exposed fish, but below the method limit of quantitation.
O.P. Togunde et al. / J. Chromatogr. A 1261 (2012) 99 106 105
4.4. Bioconcentration of pharmaceuticals in field-exposed fathead
minnow
In the case of field exposure of fathead minnow in the Grand
River, the antidepressant sertraline was also detected in fish mus-
cle at all sites except the intermediate downstream station (sites
3 and 4) with mean bioconcentration factors at the two upstream
sites of 186 and 418, respectively, while the last downstream sites
where sertraline was detected had a bioconcentration factor of
54 (Table 3). Similarly, fluoxetine was taken up by fish from the
wastewater and bioconcentrated in their muscle at sites 1, 2, and
5 with bioconcentration factors of 800, 425, and 753, respectively.
Conversely, carbamazepine and atorvastatin were not detected in
the muscle of caged fish at any of the sites, suggesting these drugs
have lower bioconcentration potential in fathead minnow. Carba-
mazepine in particular is very frequently detected in wastewater
effluents (Table 3), but with limited bioconcentration in fish tis-
sues due to its relatively high solubility (log Kow = 2.26) in water.
The antidepressant paroxetine was detected in the fish muscle in
the upstream sites 1 and 2 (reflecting the discharges of the Waterloo
municipal wastewater plant) with BCFs of 83 and 10, respectively,
but also was bioconcentrated at downstream sites 3 and 5 with
BCFs of 28 and 133, respectively. Although target pharmaceuti-
cals were detected in the muscle of field-exposed fathead minnow,
concentrations were generally lower than those observed during
the static laboratory exposure in which rainbow trout were used.
This may be due to inherent species differences, or fluctuations in
the bioavailability of the drugs in the river, which is very likely to
change over time with precipitation events, variation in hydraulic
and solids retention times in the wastewater treatment plants, dif-
fering rates of photolysis in receiving environments, and a myriad of
other environmental factors that are controlled within a laboratory
context.
5. Conclusion
For the first time, non-lethal in vivo thin film SPME was
used to quantitate target pharmaceuticals in fish muscle under
pre-equilibrium conditions. The C18 thin film geometry (as a
modification of SPME) improves in vivo extraction kinetics and
sensitivity, allowing for quantification of trace pharmaceuticals in
living fish dorsal-epaxial muscle following short-term wastewater
effluent exposure. The present study demonstrates the uptake of
pharmaceuticals in fish from wastewater in both laboratory and
field based exposures without the requirement of sacrificing the
fish. In vivo SPME based on thin film geometry is a promising non-
lethal technique which is suitable for determination of emerging
contaminants in fish muscle.
Acknowledgements
This work has been financially supported by the Natural Sci-
ence and Engineering Research Council of Canada and the Canadian
Water Network. We acknowledge the efforts and assistance of
Leslie Bragg, Lisa Bowron, and Tyrell Worrall from the Department
of Biology, University of Waterloo with analytical instrumentation
and setting up the exposure aquaria. The effluents were obtained
from Dr. Wayne Parker, Department of Civil Engineering, University
of Waterloo.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/
j.chroma.2012.07.053.
NA
Site 5
(ng/g)
n = 3 RSD (%)
NA
0.43
ą
1.0
0.008 (10)
54
Site 4
(ng/g)
n = 3 RSD (%)
Site 3
(ng/g)
n = 3 RSD (%)
Site 2
(ng/g)
n = 3 RSD (%)
Muscle conc.
Water conc. ng/mL, BCF
Muscle conc.
Water conc. ng/mL, BCF
Muscle conc.
Water conc. ng/mL, BCF
Muscle conc.
Water conc. ng/mL, BCF
Muscle conc.
Water conc. ng/mL, BCF
(ng/g)
n = 3 RSD (%)
Analytes
Site 1
CBZ
<0.40
0.022 (9)
NA
<0.40
0.024 (10)
NA
<0.40
0.117 (3)
NA
<0.40
0.109 (5)
NA
<0.40
0.045 (5)
NA
Fluoxetine
4.0
ą
2.6
0.005 (9)
800
1.7
ą
0.95
0.004 (11)
425
<0.64
0.005 (10)
NA
<0.64
0.004 (14)
NA
3.01
ą
1.70
0.004 (5)
753
Atorvastatin
<0.25
0.014 (10)
NA
<0.25
0.007 (11)
NA
<0.25
0.047 (16)
NA
<0.25
0.043 (14)
NA
<0.25
0.018 (7)
<0.25
Venlafaxin
17
ą
13
0.028 (6)
607
7.27
ą
4.0
0.024 (5)
303
<0.09
0.264 (3)
NA
<0.09
0.245 (5)
NA
2.30
ą
1.20
0.106 (2)
22
Norfluoxetine
<0.64
0.003 (12)
NA
<0.64
0.002 (14)
NA
<0.64
0.002 (11)
NA
<0.64
0.0021 (3)
NA
<0.64
ND
Sertraline
1.3
ą
1.0
0.007 (9)
186
2.09
ą
1.0
0.005 (10)
418
<0.24
0.013 (11)
NA
<0.24
0.01 (9)
Paroxetine
0.3
ą
0.2
0.004 (5)
82.5
0.37
ą
0.25
0.037 (7)
10
0.14
ą
0.20
0.005 (3)
28
<0.25
0.004 (6)
NA
0.40
ą
0.30
0.003 (8)
133
Table 3
Summary of the concentration (ng/mL,
ą
standard deviation) and bioconcentration factor of pharmaceuticals detected in fish muscle of fathead minnow caged in the field near a municipal effluent outfall (n = 6). Wastewater
samples were extracted by solid phase extraction (SPE). Sites 1 and 2 are 1.2 and 0.5 km upstream of the Doon outfall (respectively), while sites 3, 4, and 5 are 0.5, 1.7, and 5.6 km downstream, respectively.
NA, not applicable.
106 O.P. Togunde et al. / J. Chromatogr. A 1261 (2012) 99 106
References [20] G. Ouyang, K.D. Oakes, L. Bragg, S. Wang, H. Liu, S. Cui, M.R. Servos, D.G. Dixon,
J. Pawliszyn, Environ. Sci. Technol. 45 (2011) 7792.
[21] J. Koziel, M. Jia, J. Pawliszyn, Anal. Chem. 72 (21) (2000) 5178.
[1] C.J. Koester, A. Moulik, Anal. Chem. 77 (2005) 3737. [22] O.P. Togunde, K. Oakes, M.R. Servos, J. Pawliszyn, Anal. Chem. Acta 742 (2012)
[2] O.A.H. Jones, N. Voulvoulis, J.N. Lester, Environ. Technol. 22 (2001) 1383. 2.
[3] C.G. Daughton, T.A. Ternes, Environ. Health Perspect. 107 (1999) 907. [23] J. Pawliszyn, Application of Solid-Phase Microextraction, Royal Society of
[4] B. Halling-Sorensen, S.N. Nielsen, P.F. Lanzky, F. Ingerslev, H.C.H. Lutzhoft, S.E. Chemistry, Cambridge, UK, 1999.
Jorgensen, Chemosphere 36 (1998) 357. [24] Y. Chen, J. O Reilly, Y.X. Wang, J. Pawliszyn, Analyst 129 (129) (2004) 702.
[5] D.W. Kolpin, E.T. Furlong, M.T. Meyer, E.M. Thurman, S.D. Zaugg, L.B. Barber, [25] X. Zhang, A. Es-haghi, J. Cai, J. Pawliszyn, J. Chromatogr. A 1216 (2009) 7664.
H.T. Buxton, Environ. Sci. Technol. 36 (2002) 1202. [26] F.S. Mirnaghi, Y. Chen, L.M. Sidisky, J. Pawliszyn, Anal. Chem. 83 (2011) 6018.
[6] D.W. Kolpin, M. Skopec, M.T. Meyer, E.T. Furlong, S.D. Zaugg, Sci. Total Environ. [27] J. Pawliszyn, Solid-Phase Microextraction Theory and Practice, Wiley-VCH,
328 (2004) 119. New York, 1997.
[7] G.R. Boyd, H. Reemtsma, D.A. Grimm, S. Mitra, Sci. Total Environ. 311 (2003) [28] S.N. Zhou, X. Zhang, G. Ouyang, A. Es-Haghi, J. Pwaliszyn, Anal. Chem. 79 (2007)
135. 1221.
[8] A.J. Ramirez, M.A. Mottaleb, B.W. Brooks, C.K. Chambliss, Anal. Chem. 79 (2007) [29] V. Pileggi, N. Feisthauer, X. Chen, W. Parker, J. Parrott, G. Van Der Kraak, S.
3155. Tabe, S. Kleywegt, J. Schroeder, P. Yang, P. Seto, Water Environment Feder-
[9] M.M. Schultz, E.T. Furlong, Anal. Chem. 80 (2008) 1756. ation Technical Conference and Exhibition (WEFTEC), USA, October 15 19,
[10] S. Chu, C.D. Metcalfe, J. Chromatogr. A 1163 (2007) 112. 2011.
[11] C.C. Walker, H.M. Lott, S.A. Barker, J. Chromatogr. 642 (1993) 225. [30] K.D. Oakes, A. Coors, B.I. Escher, K. Fenner, J. Garric, M. Gust, T. Knacker, A.
[12] B.W. Brooks, C.K. Chambliss, J.K. Stanley, A.J. Ramirez, K.E. Banks, R.D. Johnson, Kuster, C. Kussatz, C.D. Metcalfe, S. Monteiro, T.W. Moon, J.A. Mennigen, J. Par-
R.J. Lewis, Environ. Toxicol. Chem. 24 (2005) 464. rott, A.R. P�ry, M. Ramil, I. Roennefahrt, J.V. Tarazona, P. S�nchez-Arg�ello, T.A.
[13] G. Stubbings, J. Tarbin, A. Cooper, M. Sharman, T. Bigwood, P. Robb, Anal. Chim. Ternes, V.L. Trudeau, T. Boucard, G.J. Van Der Kraak, M.R. Servos, Integr. Environ.
Acta 547 (2005) 262. Assess. Manag. 6 (2010) 524.
[14] Y. Wen, Y. Wang, Y.Q. Feng, Talanta 70 (2006) 153. [31] C.D. Matcalfe, S. Chu, C. Judt, H. Li, K.D. Oakes, M.R. Servos, D.M. Andrew, Envi-
[15] J. Schwaiger, H. Ferling, U. Mallow, H. Wintermayr, R.D. Negele, Aquat. Toxicol. ron. Toxicol. Chem. 29 (2010) 79.
68 (2004) 141. [32] Environmental Science Center of Syracuse Research Corporation (SRC), EPI
[16] X. Zhang, K.D. Oakes, S. Cui, L. Bragg, M.R. Servos, J. Pawliszyn, Environ. Sci. SuiteTM Estimation Software, U.S. Environmental Protection Agency (EPA),
Technol. 44 (2010) 3417. 2001, http://www.epa.gov/oppt/exposure/docs/episuitedl.htm (accessed
[17] S.N. Zhou, K.D. Oakes, M.R. Servos, J. Pawliszyn, Environ. Sci. Technol. 42 (2008) 20.02.2011).
6073. [33] Y. Nakamura, H. Yamamoto, J. Sekizawa, T. Kondo, N. Hirai, N. Tatarazako,
[18] S. Wang, K.D. Oakes, L.M. Bragg, J. Pawliszyn, G. Dixon, M.R. Servos, Chemso- Chemosphere 70 (2008) 865.
phere 85 (2011) 1472. [34] W. Fu, A. Franco, S. Trapp, Environ. Toxicol. Chem. 28 (2009) 1372.
[19] G. Ouyang, D. Vuckovi, J. Pawliszyn, Chem. Rev. 111 (2011) 2784. [35] H.O. Esser, P. Moser, Ecotoxicol. Environ. Saf. 6 (182) (1982) 131.