Improved sample treatment for the determination of 17 strong sorbed

quinolone antibiotics from compost by ultra high performance liquid

chromatography tandem mass spectrometry
art ic l e

i nf o

Article history:
Received 14 December 2014
Received in revised form
3 March 2015
Accepted 6 March 2015
Available online 12 March 2015
Keywords:
Antibiotics
Quinolones
Compost
Microwave-assisted extraction
Matrix effect
UHPLCMS/MS

a b s t r a c t
The use of compost from sewage sludge for agricultural application is nowadays increa
sing, since
composting is recognized as one of the most important recycling options for this material, being
a source
of nutrients for plants but also of contamination by persistent pollutants. In the present work
, a multiresidue analytical method for the determination of 17 quinolone antibiotic residues in comp
ost using
multivariate optimization strategies and ultra high performance liquid chromatography
tandem mass
spectrometry has been developed. It is based on the use of microwave-assisted extraction
at drastic
conditions with ACN:m-phosphoric acid (1% w/v) for 5 min at 120 C, in order to achieve a
quantitative
extraction of the compounds (476% of extraction recovery). Extracts were cleaned-up by sa
lt-assisted
liquid liquid extraction (SALLE) with NaCl at pH 1.5 (with HClO4 ) and then using a dispersiv
e sorbent
(PSA). After LC separation, the MS conditions, in positive electrospray ionization mode (
ESI), were
individually optimized for each analyte to obtain maximum sensitivity in the selected reac
tion monitoring mode (SRM). The analytes were separated in less than 7 min. Cincophen was used a
s surrogate
1
standard. The limits of detection ranged from 0.2 to 0.5 ng g , and the limits of the quantica

1.

Introduction

tion from
1
0.5 to 1.5 ng g , while intra- and inter-day variability (% RSD) was under 7% in all cases. A
recovery assay
was performed with spiked samples. Recoveries ranging from 95.3% to 106.2% were obtaine
d. Cleanup
procedure reduced signicantly matrix effects, which constitutes an important achievement, co
nsidering
the important drawbacks of matrix components in quality and validation parameters. This met
hod was
applied to several commercial compost samples. Only 6 of the studied antibiotics were not d
etected in
any of the samples. The antibiotics with the highest concentrations were ciprooxacin (8
1
36 ng g ),
1
1
ooxacin (719 ng g ), and enrooxacin (674 ng g ), which were also the only ones found
in all the
analyzed samples. The results showed that this method could also be potentially adapted for the
analysis
of other strong sorbed basic pharmaceuticals in solid environmental matrices.
& 2015 Elsevier B.V. All rights reserve
d.

In recent years, antibiotics have been identied as an em

erging
class of potential contaminants. As these compounds are
highly
Antibiotics are widely used in human and veterinary me
dicine
[1]. After administration, from 10% to 90% of the drug is excre
ted in
urine or feces into sewage [2]. Subsequently, these substan
ces are

continually introduced into the environment, mainly throu

gh wastewater efuent from municipal treatment plants, hospitals
or livestock activities [3 5]. These compounds can also reach
terrestrial

environmental compartments when sludge is spread on the s bioactive, concerns have arisen about the presence and p
oil as
ossible
harmful effects of these substances [1]. Some scienti c studies
fertilizer.
have
demonstrated adverse effects from longstanding, low-dose
n
Corresponding author. Tel.: 34 958 24 07 98; fax: 34 958 24 33 28.
expoE-mail address: azafra@ugr.es (A. Zafra-Gmez).
sures in both aquatic and terrestrial wildlife. There is also a se
ttp://dx.doi.org/10.1016/j.talanta.2015.03.011
rious
0039-9140/& 2015 Elsevier B.V. All rights reserved.
interest over their role in enhancing antibiotic resistance
among
pathogenic bacteria, rendering current antibiotics ineffective i
n the
treatment of numerous diseases [6]. Antibiotics can also affe
ct the
endocrine system of sh and be toxic to algae and invertebrat
es.
Quinolones are one of the major classes of antimicrobials
that
are employed worldwide in human and veterinarian med
icine.
According to a categorization of antibiotics by the World H
ealth
Organization in 2011, ( uoro)quinolones are included as h
uman
248

N. Dorival-Garca et al. / Talanta 138 (2015) 247257

critically important antibiotics. The Spanish Agency of Medi

cines
and Health Products reported in 2009 [7] a continual use
of quinolones along the last 20 years. Ciprooxacin and ooxacin a
re the
most used quinolones in hospitals [8] and livestock [9],
while
enro oxacin is the most used drug in veterinary medicine
[6,9].
Administered quinolones are largely excreted as unchanged
compounds in urine, and consequently discharged into sewage [
1,10].
These antibiotics are not completely removed at WWTPs and
their
continuous introduction into the environment makes them
persistent compounds [11]. Quinolones that are initially pres
ent in
water bodies may be transferred without degradation and s
tored
in different environmental compartments [12]. Moreover, hu
mans
may be exposed to residues of drugs in the environmen
t by a
number of routes including the consumption of crops that
have
accumulated these substances from fertilized soils [13].
The utilization of sewage sludge for agricultural applicat
ion is
increasing [14], and composting has been recognized as one
of the
best sewage sludge recycling options [15]. Previous studie
s have
shown that the degradation of some pharmaceuticals and pe
rsonal
care products may take place during biosolids composting [16
], but
still very few systematic works concerning the degradation of
antimicrobials during sewage sludge composting have been deve
loped

[17 19]. Instead of developing technologies to assist the

degradation
of antibiotics in soil and preventing them from contaminating wat
ers
and crops, it is more effective to eliminate the antibiotics in comp
ost
before it is applied to agricultural land as fertilizer [20] such
that
composting process may provide a practical and economical soluti
on
for reducing the risk of pollution in the environment.
Although several papers on the determination of antibiotics
in
sewage sludge have been reported in the scienti c literature [21
24],
only few have been published about their determination in comp
ost
or treated sludge [25 28], or furthermore the specic extractio
n of
quinolones with good recoveries, considering the strong interactio
ns
that these compounds usually form with matrices with a
large
charge of organic matter [25]. The reported methods include
different sample treatment procedures, such as microwaveassisted
extraction (MAE), ultrasonic extraction (USE), pressurized li
quid
extraction (PLE), matrix solid phase dispersion (MSPD) or stir
bar
sorptive extraction (SBSE). However, most of these works appli
es a
further cleanup step by the traditional solid-phase extraction
(SPE),
which is a tedious, costly and time-consuming procedure. In the
last
years, modern sample preparation techniques, such as Quick,
Easy,
Cheap, Effective, Rugged and Safe (QuEChERS), based on the
minimization of organic solvents and less time consumption have b
een

developed [29]. This technique is based on salt-assisted liquid the occurrence and fate of these antibiotics during the composti
ng
liquid
process and their potential re-entry to the environment.
extraction (SALLE) and it has demonstrated to be a good alter
native
to SPE as a cleanup step. After extraction, the organic solvent
2. Materials and methods
phase
containing the analytes of interest is cleaned up again by addi
2.1. Chemicals and reagents
ng an
SPE sorbent (dispersive SPE, D-SPE) such as primary secondary
Water (18.2 M cm) was puried using a Milli-Q system fr
amine
(PSA) or C18, to remove interferences. Recently, some works f
om
or the
determination of antimicrobial agents using this techni Millipore (Bedford, MA, USA). Analytical grade standards of qui
noque in
environmental matrices have been published [30 34], in lones: pipemidic acid (PIP), enoxacin (ENO), noroxacin (N
OR),
cluding
quinolones [31,34]. However, the extraction procedure led ciprooxacin (CIP), ooxacin (OFL), enrooxacin (ENR), dio
xacin
to low
recoveries, which is indicative of the need of drastic ex (DIF), marbooxacin (MAR), danooxacin (DAN), saraoxacin (SA
R),
traction
cinoxacin, (CIN), lomeoxacin (LOM), moxi oxacin (MOX), nalidi
conditions.
The aim of the present study was to develop a multi- xic
acid (NAL), oxolinic acid (OXO), umequine, (FLU), piromidic
residue
analytical method for the determination of 17 quinolones, wiacid
(PIR); the surrogate 2-phenyl-4-quinoline carboxylic acid (cin
dely
used in human and veterinarian medicine, in compost sampl cophen, CIC), and the internal standard, caffeine (CAF) were purcha
es at
trace levels. The method includes an improved sample trea sed
from Sigma-Aldrich (St. Louis, MO, USA). Individual standard so
tment
that consists in microwave-assisted extraction with high lu1
tions of compounds (200 g mL ) were prepared in a w
recoveries for strong sorbed quinolones, followed by a cleanu ater/
p step
methanol mixture (1:4, v/v) and stored at 20 C. These solut
based on SALLE and D -SPE to reduce matrix effects and ions
consewere prepared fresh monthly. Working standard mixtures w
quently to increase the analytical sensitivity. The determi ere
prepared by diluting the individual stock solution in methanol or i
nation
was performed by UHPLC MS/MS. The method was appli n
the initial mobile phase immediately before use. These soluti
ed to
compost samples obtained from sewage sludge. This ana ons
lytical
were stored at 4 C and prepared fresh weekly. All solutions w
method is useful for the development of more in-depth
ere
studies on
stored in dark glass bottles to prevent photodegradation. Techni
cal
acetonitrile (min. 99% pure) used for extractions was purchased fr
om
VWR (Radnor, Pennsylvania, USA). m-Phosphoric acid (33.5 36.5%
as
HPO3 ), perchloric acid (60% w/w), anhydrous magnesium sulfate a
nd
sodium chloride were provided by Panreac (Darmstadt, Germa
ny).
PSA sorbent (primary secondary amine, 40 60 mm) was purcha
sed
from Scharlab (Barcelona, Spain) and BAKERBONDs octadecyl
C18
sorbent (40 mm particle size) was provided by J.T. Baker (Deve
nter,
The Netherlands). LC MS grade water and methanol, acetonit
rile,
sodium hydroxide, ammonia ( Z25%) and formic acid ( Z98%)
used
for the preparation of standards, mobile phases and pH adjustmen
ts
were purchased from Fluka (St. Louis, MO, USA).
2.2.

Instrumentation and software

The extraction was performed with a Milestone (ETHOS S

EL)

microwave solvent extraction Labstation (Sheldon, CT, USA), ACQUITY UPLC BEH C18 column (1.7 m ; 2.1 mm 100
opermm)
ating at 2455 MHz with a maximum delivered power of 10 (Waters, UK). A Xevo TQS tandem quadrupole mass spectrome
00 W.
ter
The time, temperature and microwave power control were adju (Waters) equipped with an orthogonal Z-spray electrosp
sted
ray
and controlled throughout the process using a control terminal ionization (ESI) source was used for antibiotic detection.
that
For pH measurements, a EUTECH PCD 650 digital pH-meter
runs EasyCONTROL software. Temperature was monitored with with
the
a combined glass Ag/AgCl (KCl 3 M) electrode (EUTECH
aid of a thermopar ATC-400 temperature sensor. This apparatus Instruments
was
Ltd, Singapore) was used. Compost samples were freeze-dried
equipped with 10 closed vessels made of teon, for whic using
a SCANVAC CoolSafe freeze dryer (Lynge, Denmark). A vor
h, the
maximum operating temperature was 300 C.
texUHPLC MS/MS analysis was performed using a Waters
mixer (IKA, Staufen, Germany), a Digicen 21 centrifuge (Orto Alre
Acquity
sa,
UPLC H-Class (Waters, Manchester, UK), consisting Madrid, Spain), an Ultrasons-HD ultrasound bath (Selecta,
of an
Barcelona,
ACQUITY UPLC binary solvent manager and an ACQUITY UPL Spain), a Spectrafuge 24D centrifuge from Labnet Internati
C
onal,
sample manager. Separation of compounds was obtained Inc. (New Jersey, USA) and a sample concentrator (Stuart, Staff
with
ordshire, UK) were also used. Statgraphics Plus software version
5.1
N. Dorival-Garca et al. / Talanta 138 (2015) 247257

(Statpoint Technologies Inc., Virginia, USA) was used for sta

tistical
treatment of data.
2.3.

Basic procedure

After collection, samples were freeze-dried and siev

ed to
r1.41 mm mesh size and introduced in dark glass bottl
es and
stored in the dark at 4 C until analysis.
2.3.1.

Preparation of spiked samples

For spiked samples preparation during optimization of extra

ction
parameters, 1 mL of a methanolic standard solution of analyte
s was
added to 1 g of compost to obtain a nal concentration of 80 n
1
gg
for each antibiotic. This volume allows the analytes to c
ome in
contact with the whole sample. In order to attain sorption
equilibrium and to allow complete evaporation of methanol from sa
mples,
the mixtures were shaken on a vortex mixer for 10 min an
d were
then left to stand for 24 h at room temperature in the dark
before
analysis.
For method validation (recovery assays, precision, and true
ness),
blank samples were spiked at different concentrations by
adding
1 mL of the spiking standard solution containing the analyt
es and
the surrogate (CIC) to 1 g of compost samples. Then, the
spiked
samples were treated as it was indicated before. The blank sa
mples
were previously analyzed in order to ensure the absence of an
alytes

249

or that these were below the LODs of the method. However,

none of
the analyzed compost samples from sewage sludge were fou
nd to be
completely free of analytes and therefore a different type
of commercial compost was used as blank. A compost type that est
ablished
very similar interactions with the target analytes was
selected,
which resulted in comparable extraction recoveries an
d matrix
effects, as shown in Fig. S1 (supplementary material) f
or a con1
centration of 15 ng g for each antibiotic.
2.3.2. Extraction procedure
A sample of 1.0 g of compost was placed in a microwave
vessel,
15 mL of the solvent, ACN:m-phosphoric acid 1% (7:3, v
/v) were
added. Samples were extracted for 5 min (10 min for ho
lding) at
120 C and 1000 W of power. Ten vessels were processe
d simultaneously. In the Ethos unit, the pressure is controll
ed automatically, and overpressure complications are thus avoid
ed. Only
one extraction cycle was required. After microwave irradia
tion, an
air ow cooled the vessels inside the microwave unit ( o4
0 C). In
order to decrease the matrix co-extractives in the ext
ract that
could cause the matrix effect, a cleanup of the extract
based on
SALLE and D-SPE was carried out. The extract was
transferred to a
50 mL Falcon conical tube containing 3 g of NaCl. Then,
75 mL of

phase, 80% (A), which was linearly decreased to 30% (A)

perchloric acid 60% was added. The mixture was handwithin
shaken for
2 min and centrifuged for 3 min at 2594g. The organic phase 3.0 min and to 0% within 0.1 min and held for 1.9 min to cl
ean the
(ACN)
containing the analytes was transferred to a new Falco column using 100% organic mobile phase. Finally, back to 80%
(A) in
n tube
containing 300 mg PSA and 600 mg anhydrous MgSO4 . The 0.1 min and kept for 1.9 min to equilibrate the column. Total run
time
mixture was hand-shaken for 1 min and centrifuged for 1 was 7 min.
The mass spectrometer (MS) was operated with electro
min at
2594g. The supernatant was decanted into a glass vial an spray
ionization (ESI) in positive ion mode and in multiple re
d, after
evaporation to dryness at 50 C under a N2 stream, the resid action
monitoring (MRM) mode. The MS/MS parameters were optimi
ue was
dissolved in 500 mL of a solution of ammonium formate 2 zed
1
individually for each antibiotic by infusion of 0.5 g mL
stan
5 mM
(pH 3.0):methanol (1:1, v/v), centrifuged at 16,300g for 30 m dard
in and
solution in the initial mobile phase, using [M H] as precursor
injected into the LC system.
ion,
under combined mode. Instrument parameters were as fo
2.4. Ultra high performance liquid chromatography tandem
llows:
capillary voltage, 0.60 kV; source temperature, 150 C; deso
mass
lvation
1
spectrometry conditions
temperature, 500 C; cone gas ow, 150 L h ; desolvation ga
s ow,
Chromatographic separation was performed using a binar 500 L h 1 ; collision gas ow, 0.15 mL min 1 , and nebulizer ga

y gras ow,
dient mobile phase consisting of ammonium formate 25 mM sol
7.0 bar. Nitrogen ( Z99.995%) was used as a cone and desol
ution
vation
at pH 3.0 (solvent A) and methanol (solvent B). The ow r
gas, and argon (99.999%) was used as a collision gas. After th
ate was
e pre1
300 mL min , the column was maintained at 40 C and the inj
cursor ions were selected, product ions were obtained with a
ection
comvolume was 7 mL. Gradient conditions were as follows: initial
bination of collision energies and cone voltages. Table 1 show
mobile
s the
two most sensitive transitions (one used for quantication an
d the
other for conrmation) selected. To obtain the maximum sensit
ivity,
the most abundant transition was used for quanti cation.
Dwell
time for each transition was 25 ms, and interscan delay was
set at
3 ms. Data acquisition was performed under time-segmented
conditions, based on the chromatographic separation of compound
s, to
maximize the sensitivity. The optimum collision energies,
cone
voltages for each transition, segment periods and retention time
s are
also summarized in Table 1.

3.
3.1.

Results and discussion

Optimization of the microwave-assisted extraction

The optimization of the extraction method was based on a

previous work about the determination of quinolones in sewage sl
udge
[23], which was improved by applying a cleanup process to the
MAE
extracts based on QuEChERS technique, since these extracts c
ontain
high amounts of matrix components due to the drastic cond
itions

that are applied during the extraction step. Briey, samples of 1 pound to control variations during chromatographic analysis, sin
ce it
g of
compost previously spiked with the analytes were extracted produces a very sensitive and constant signal, besides that it w
as not
for
17 min at 87 C and 1000 W using ACN:aqueous buffer McIlvaine naturally found in compost samples that were used for the
optipH
3 (1:1, v/v) as extraction solvent. Then, cleanup was performed u mization process.
It was observed that using this basic procedure, ver
sing
a mixture of 2 g of MgSO4 anhydrous and 1 g NaCl. The org y low
recoveries were obtained for the antibiotics in compost sa
anic
phase containing the analytes was separated and dried with 600 mples
( o20%). Therefore, it was decided to replace the buffer McIlvain
mg
MgSO4 . The extract was evaporated to dryness at 50 C under a e by
an aqueous solution of m-phosphoric acid 0.3% (w/v) that has
N2
stream; dissolved in 500 mL of ammonium formate 25 mM (pH 3. been
already applied for the determination of quinolones in s
0):
methanol, 1:1, v/v, containing caffeine as internal sta everal
matrices [35]. The use of a mixture of ACN:m-phosphoric acid
ndard
1
(50 ng mL ); centrifuged at 16,300g for 30 min and injected 0.3%
(w/v) as extraction solvent produced adequate results and recov
into
the LC system. Caffeine was used as internal standard during eries
increased around 70%. ACN was kept as organic solvent and
the
optimization process, because it was the most appropriated coother
extraction parameters were optimized using chemometric te
mchniques to improve antibiotic recoveries.
250

N. Dorival-Garca et al. / Talanta 138 (2015) 247257

Table 1
Optimized SRM conditions and retention times for quinolones, internal standard and surrogate.
Compound

SRM1

CV/CE

SRM2

CV/CE

Ion ratio

Segment (min)

RT (SD) (min)

CAF (IS)

195.1-137.9

32/18

195.1-109.9

32/22

0.79

2.04.0

3.02 (0.05)

CIC (SG)

250.2-127.9

4/34

250.2-222.0

4/26

0.62

4.07.0

5.51 (0.03)

Amphoteric quinolones
PIP

304.2-217.1

10/22

304.2-286.2

10/18

0.41

2.04.5

3.17 (0.04)

MAR
ENO
OFL
NOR
CIP
DAN
LOM
ENR
SAR
DIF
MOX

363.2-71.9
321.2-303.2
362.2-318.2
320.2-302.2
332.2-314.2
358.2-340.2
352.2-265.1
360.2-316.2
386.4-368.2
400.2-382.2
402.2-384.2

10/20
6/18
4/18
38/18
14/18
2/24
6/22
8/18
50/20
2/22
30/22

363.2-320.1
321.2-232.0
362.2-261.1
320.2-276.2
332.2-231.1
358.2-81.9
352.2-308.2
360.2-245.1
386.4-342.2
400.2-356.2
402.2-358.2

10/14
6/34
4/26
38/16
14/34
2/36
6/16
8/26
50/18
2/18
30/18

0.30
0.31
0.97
0.18
0.91
0.43
0.62
0.55
0.23
0.85
0.60

2.04.5
2.04.5
2.04.5
2.55.5
2.55.5
2.55.5
2.55.5
2.55.5
2.55.5
2.55.5
2.55.5

3.33
3.61
3.62
3.83
3.89
3.94
4.02
4.05
4.28
4.35
4.74

CIN

263.2-245.1

4/14

263.2-189.0

4/26

0.58

2.55.5

4.53 (0.03)

OXO
NAL
FLU
PIR

262.2-244.0
233.1-215.0
262.2-244.0
289.2-271.1

4/18
2/14
6/18
8/18

262.2-215
233.1-187.0
262.2-202.0
289.2-243.1

4/26
2/24
6/32
8/28

0.22
0.36
0.80
0.72

4.07.0
4.07.0
4.07.0
4.07.0

5.04
5.57
5.66
5.85

(0.03)
(0.06)
(0.03)
(0.05)
(0.04)
(0.04)
(0.03)
(0.04)
(0.05)
(0.03)
(0.02)

Acidic quinolones

(0.03)
(0.05)
(0.02)
(0.01)

IS: internal standard; SG: surrogate.

a
b

SRM1, transition used for quanti cation; SRM2, transition used for con rmation.
CV, cone voltage (V); CE, collision energy (eV).
temperature
RT, Retention time; SD, Standard deviation.

and extraction time were varied at different l

evels
(Fig. 1A). The studied response was the extraction recovery. Tabl
3.1.1. Screening by two-level half-fraction design.
e S1
First, in order to select the most in uential variables, a two (supplementary material) shows the used screening design mat
rix.
level
5 1
Pareto charts were obtained and statistically signi cant effect
half fractional factorial experimental design (2 ) was perf
ormed,
s of
resulting in 19 experiments (i.e. 163 central points). the variables were screened using a Student's t-test for AN
Solvent
OVA.
composition, pH of extraction solvent, extraction solvent v Variables having a con dence greater than 95.0% were consid
olume,
ered
c

to have a signicant effect on the extraction efciency. Fig. 1B

and C
shows the statistically signicant effect of each variable, pre (40% to 100% ACN), ve levels for m-phosphoric acid
sented
concentration
as the sum of recoveries of quinolones that were grouped into
(from 0.3% to 1% w/v) and three levels for extraction time (from
two
5 to
classes: amphoteric and acidic quinolones.
The evaluated variables and some of their interactions res 20 min) were considered. A wide temperature range was appl
ied,
ulted
signicant. Temperature and solvent composition were the considering the chemical stability of this family of antibiotics [
36].
most
The extraction conditions for each of the 23 experimental runs in
in uential parameters, followed by pH of the solvent, w
the
hereas
Doehlert matrix are shown in Table S2 (supplementary material)
extraction time was more signi cant for acidic quin
.
olones.
The data were evaluated by ANOVA and the test gave det
Although the volume of the extraction solvent was also signi
er2
cant
mination coefcients (R ) between 0.863 and 0.989. Since
in both cases, its interactions were not important. Therefore, i
the
t was
decided to x this variable to simplify the optimization P values for the lack-of- t test were 40.05 in all cases, the
model
of signi cant factors. Volume was set at 15 mL, corresponding appears to be satisfactory with the 95% of condence level.
Quadratic terms for temperature and extraction time resul
to the
ted
central point value. The most important interactions are asso
statistically signicant for all compounds. For solvent composit
ciated
to temperature, solvent composition and extraction time, ion,
quadratic terms were signicant only for amphoteric quinolo
being
the most signi cant the positive interaction between temper nes.
Solvent composition for acidic quinolones and m-phosphoric a
ature
cid
and solvent composition. In order to make practical the extra
concentration were signicant only in their lineal terms. The obser
ction
ved
procedure, the number of extraction cycles was set at 1.
tendency for both groups of analytes was that the solvent compo
3.1.2. Optimization of signi cant parameters by a Doehlert d sition for acidic quinolones, temperature, and extraction time ha
d a
esign
positive effect, whereas solvent composition for amphoteric qui
The most in uential variables (solvent composition, pH
noof the
lones had a negative effect. Important interactions were also obser
extraction solvent, temperature and extraction time) wer
ved
e optimized using a Doehlert experimental design. The Doehlert between solvent composition and temperature (positive), solv
ent
matrix
consisted of 23 experiments, including three central points. composition and m-phosphoric acid concentration (negative).
Seven
levels for temperature (from 50 to 120 C) and solvent compo 3.1.3. Optimization of multiple responses
The optimized conditions obtained using this approach usua
sition
lly
conicted with the aim of achieving good yields of analysis re
coveries for all the studied compounds, especially when some di
fferences between both the groups of quinolones were found. Thu
s, a
multicriteria-decision making process is essential and a total de
sirability function, D, was used to optimize all responses simultaneou
sly
(the recoveries of the 17 analytes). This function is a measu
re of
overall quality and provides conveniently a way to compare seve
ral
responses and to select the optimum with the most desirable pro
perties. Desirability ranges from zero to one. Under the mentio
ned
optimization criteria, the experimental conditions corresponding
to
one maximum in the desirability function (D0.921) were solv
ent
N. Dorival-Garca et al. / Talanta 138 (2015) 247257

251

5 1

Fig. 1. (A) Levels for the factors examined in the 2

screening design. Standardized main effect Pareto charts for (B) amphoteric quinolones and (C) acidi
c quinolones. ()
Positive effects on the response, () negative effects on the response. Vertical line shows the limit of decision to consider the signi cance of the fa
ctors (based on the
standardized effectestimated effect/standard error, at 95% of con dence level).

composition, 70% ACN; m-phosphoric acid concentration, 1

% w/v;
temperature, 120 C; extraction solvent volume, 15 mL; ex
traction
time, 5 min and 1 cycle of extraction.
The response surfaces obtained for the global desirability fun
ction
are presented in Fig. 2. These plots were obtained for a given
pair of
factors, while maintaining the other xed at their optimal valu
es.
The desirability is zero when solvent composition is higher
than
95% ACN (Fig. 2A), and between 40% and 50% ACN. This conr
ms the
importance of establishing an adequate ratio organic/aqueous p
hases
for the extraction solvent, considering the recognized water sol
ubility
of quinolones at the pH conditions that are extracted. On the
other
hand, higher desirability is observed at higher concentrati
ons of
m-phosphoric acid, which is an indication that strong acidic
condi-

tions are required for the efcient extraction of the target

analytes.
Temperature is an important factor (Fig. 2B), since as te
mperature
increases; the desirability also becomes signicantly higher,
probably
because higher temperature can break efciently the r
ecognized
strong interactions between quinolones and the matrix.
3.2.

Optimization of the cleanup procedures by SALLE and

D-SPE

sorbents
Since a drastic extraction technique as MAE was
used, the
resulted extract usually contains a signicant proportion of
matrix

components, and it requires cleanup prior to chromatograp phase. The separation is from homogeneous solution inste
hic
ad of
analysis in order to reduce matrix effects and increase met
through emulsion or suspension and centrifugation can facilitat
hod
e the
sensitivity. It was decided to apply a cleanup procedure based
process. Thus, matrix components, mainly water-soluble, cou
on
ld be
QuEChERS, which consists in two different steps: the rst one
separated from analytes that migrate to organic layer. Fig. 3A s
by
hows
SALLE, followed by the use of a dispersive SPE sorbent.
the results of the experiments performed to determine the effe
ct of
3.2.1. Optimization of cleanup by SALLE
the type of added salt on the partitioning of the antibiotics o
SALLE is a procedure that involves water-miscible solvents and n the
is
upper layer. Three salts were evaluated: NaCl, MgSO4 and (NH 4 )
based on the salting-out effect. The term also connotes reduction
2 SO4 ,
of
as well as combinations of them, as it happens when methods
mutual miscibility of two liquids by the addition of electroly
based
tes.
on QuEChERS are applied (NaCl MgSO4 ).
Although ACN is miscible with water in any proportion at ro
NaCl was selected as the optimum salt for SALLE. MgSO4 h
om
ad a
temperature, the addition of salt signicantly reduced the mu negative effect, even though this is a common-used saltin
tual
g-out
miscibility, even resulting in phase separation of ACN from aqueo
us
252

N. Dorival-Garca et al. / Talanta 138 (2015) 247257

Fig. 2. Response surface plots corresponding to the desirability function when optimizing the following pair of factors from MAE, while maintaining constant th
e remaining
two at their optimum values: (A) solvent composition (% ACN) vs m-phosphoric acid concentration; (B) extraction time vs temperature. Results were
evaluated using a 95%
con dence interval.

reagent due to its high ionic strength per unit concentration i biotics could be explained by the presence of Mg2 , because t
n the
his
aqueous phase. The low recoveries observed for the studied
divalent cation can be chelated by quinolones, as it has
antibeen

2
previously described [37]. The negative effect of Mg ion 3.2.2. Use of perchloric acid during SALLE
According to a previous work [38] and considering that extracti
s was
especially important for amphoteric quinolones (Fig. 3A), on
of quinolones is highly dependent of pH, the pH during SALLE
which
was
form complexes more easily with divalent and trivalent catio
also adjusted using HClO4 60% (w/w). This acid was speci
ns, in
cally
comparison with acidic quinolones (PIR, CIN, OXO, FLU, and
selected due that the chaotropic effect exhibited by inorg
NAL).
anic
The determination of the proper amount of salt is also i
counter-anions as perchlorate was found to be the most effec
mportive
tant, since this parameter can be used to control the percent under acidic conditions. In this study, the results may also imply t
age of
he
water in the organic phase (and vice versa). Anastassiade important role of ion-paring for the extraction of quinolones in
s et al.
the
[29] stated that the more NaCl is added to the system, th salting-out extraction procedure. The addition of 75 and 150 mL
e more
of
complete the phase separation becomes. Therefore, less HClO4 60% to the MAE extract was evaluated, which produc
water
ed a
remains in the ACN phase. Water in organic phase enables reduction of pH during SALLE from 3.0 (without acid) to 1.5 (7
a cer5 mL
tain degree of adjustment in the polarity of the phases, so th HClO4 ) and 1.0 (150 mL HClO4 ). When this acid was used to ac
idify
at the
extraction efciency of the analytes that are dissolved i the extract, followed by liquid liquid partitioning formed by
addition
nto the
organic layer depends greatly on the applied amount of sal of NaCl, an increase in the extraction recoveries was observed for
all
t into
the system. In this context, it has also to be considered t analytes when pHr1.5 (Fig. 3B). A volume of 75 mL of 60% (
w/w)
hat salt
perchloric acid as compromise value was selected.
can also produce the extraction of polar co-extractives,

decreasing
the recovery of the analytes. Some studies performed on pesti 3.2.3. Optimization of cleanup by a dispersive SPE sorbent
After the analytes were partitioned in the organic phase
cides
by
have also found that the amount of NaCl used during the
SALLE, the organic phase was further cleaned up and drie
partid by
tioning had a great in uence on the peak shapes and areas
mixing with SPE sorbents and anhydrous MgSO4 . The sorbents we
of the
re
analytes. This effect is also related to the amount and nature chosen to retain the matrix components and to enable the analyt
of the
es
co-extracted matrix components [28]. Mass of NaCl from 1 of interest to stay in the ACN phase. Traditionally, a dispersiveto 4 g
SPE
was evaluated, being selected 3 g NaCl as optimum value.
cleanup has been carried out in studies that employ QuEChERS.
Two sorbents were evaluated: PSA and C18. PSA acts like a we
ak
anionic exchanger and is able to form strong interactions with oth
er
N. Dorival-Garca et al. / Talanta 138 (2015) 247257

253

Fig. 3. Optimization of the cleanup steps. (A) Type of salt for salt-assisted liquidliquid extraction; (B) volume of HClO4 60% (v/v). Error bars indicate
standard deviation.
(C) Desirability function for the optimization of cleanup step with D-SPE sorbents: amount of PSA vs amount of C18. Results were evaluated using a 95%
con dence interval.

compounds as hydrogen bonds and dipole dipole forces.

amount (0, 150 and 300 mg), and PSA amount (0, 150 and 30
0 mg).
Moreover,
it shows a high chelating effect due to the presence of the pri The design matrix is shown in Table S3 (supplementary mater
ial).
mary
Fig. 3C shows the plot of the desirability function vers
and secondary amines. The result is the strong retention
us the
of acid
components such as fatty acids and other polar compounds amounts of PSA and C18 sorbents. PSA resulted the most statist
ically
in the
matrix [29,39]. However, in order to improve the cleanup proc signicant parameter for almost all compounds, with a positive
effect
for both, its lineal and quadratic terms. These results are accordi
ng to
previous works, because the use of PSA in the determinat
ion of
quinolones in similar environmental matrices has been a
lready
lipophilic
compounds from ACN extracts. Then, the optimization documented [30,33]. PSA demonstrated its ability to sel
ectively
of the
cleanup with sorbents was focused on the study of the compo remove acidic interferences that are main components of co
mpost,
sition
of the sorbents employed. Mixtures of PSA and C18, as well such as humic and fulvic acids, as well as other organic acids,
polar
as the
pigments, carbohydrates, sugars and fatty acids with hy
use of each one separately were checked. A three-level-full
drogen
factorial
design with two parameters was performed. The amount of M bonding properties [28]. However, although C18 was not in ue
ess of
extracts, especially when high content of organic matter an
d fat is
present, a mixture of PSA and C18 is usually applied [39]
. C18 is
specically used for removal of co-extracted fat and other

ntial
gSO4
anhydrous that is applied during this cleanup phase as a desi in most of the cases, it had a negative effect in 4 of the 17 quino
lones,
ccant,
in order to eliminate traces of water which complicate and an important negative interaction was observed between
both
sample
sorbents, so that the use of C18 was discarded. According
evaporation and concentration steps, was set at 600 mg.
The experimental design required 12 runs, including three to the
results of the design, 300 mg of PSA was selected as the sorbent
central points. The two variables, evaluated at three levels, w used
in the cleanup step, which was applied as a mixture with 600
ere C18
mg of
254

N. Dorival-Garca et al. / Talanta 138 (2015) 247257

MgSO4 . The extraction recoveries ranged from 72% to 96%

for the
3.2.4.
studied antibiotics (Fig. 4A).

Matrix effects

One of the major drawbacks of the use of electrospray made in triplicate and analyzed twice. Table 2 shows the statis
tical
is the
and the analytical parameters obtained for each studied antibiotic
suppression or enhancement of the analyte signal by co.
extracted
The matrix effects were also evaluated through the calibrat
substances of the matrix. This matrix effect was quanti ed
ion
for all
curves prepared in the compost extracts and in the initial mo
1
analytes at a concentration of 15 ng g for each antibiotic.
bile
Matrix
phase. Student's t-test was applied in order to compare the
effects were evaluated by calculating the percentage of
calisignal
bration curves. First, the variances estimated as S2y/x were comp
suppression in the extracts. The peak areas from the anal
ared
ysis of
by means of a Snedecor's F-test. Student's t-test showed statis
spiked compost extracts were compared with the ones
tical
corredifferences among slope values for the calibration curves of
sponding to the spiked solvent (mobile phase) at the sam
the
e contarget analytes and consequently, the existence of matrix effects
centration level. Average values for signal suppression are
was
comconrmed, so that it was decided to quantify the analytes in sam
piled in Fig. 4B. In most cases, matrix suppression was ob
ples
served.
The suppression ranged from 9% to 47%, and signal enhanc using the matrix-matched calibration in all cases.
The analytical method was validated in terms of linearity, sel
ement
which ranged from 1% to 29% for all the acidic quinolones, ectivity, sensitivity and accuracy (trueness and precision), accordin
MOX,
MAR and ENR was also observed. Signal enhancement hasg to
the protocols described in the US Food and Drugs Administrati
been
previously observed in the determination of quinolones in sl on
(FDA) guideline for Bioanalytical Method Validation [40].
udge
2
Linearity. The determination coefcient (R ) and the lackand similar matrices [31].
Using a previous work as a reference [24], signal supp of- t
2
test (Plof ) were evaluated. The values obtained for R ranged fr
ression
ranged between 30% and 80% for the determination of quinolo om
99.3% to 99.9%. Plof values were 45% in all cases. Therefore,
nes in
sewage sludge. Matrix effects have been reduced signicantly igood
linearity was observed within the concentration ranges.
n this
Selectivity. This parameter was demonstrated by LC
work for compost samples, which could be attributed to the a
MS/MS
pplication of these new alternatives of cleanup steps to MAE e analysis of blanks. A blank sample and a spiked blank sample wi
th
xtracts.
This difference could also be due to the type of matrix, the analytes were extracted and their chromatograms were co
mbecause
unlike sludge that only shows signal suppression, compost sa pared. No interferences were observed at the retention time of t
he
mples
exhibit both, signal suppression and signal enhancement. How analytes. These ndings suggest that the spectrometric conditio
ns
ever,
in both matrices there is a clear difference in matrix effects c ensure the high selectivity of the methods. Fig. 5A shows the S
RM
aused
by amphoteric and acidic quinolones, signal suppression chromatograms obtained from a spiked sample.
Sensitivity. The limit of detection (LOD) and quanti cat
is signi cantly lower for acidic quinolones in sludge samples, and o ion
(LOQ) were calculated by taking into consideration the stand
n the
other hand, acidic quinolones always show signal enhancem ard
deviation of residuals (Sy/x ), the slope (b) of the calibration gra
ent in
phs
compost samples.
and an estimate s0 obtained by extrapolation of the stan
dard
3.3. Method validation
deviation of the blank. The LOD was 3s0 and the LOQ was
10s0 .
A seven-point matrix-matched calibration curve was
1
Found limits of quanti cation ranged from 0.5 to 1.5 ng g . The
obtained for
each studied compound in the lineal dynamic ranges that ar se
e indiresults are also summarized in Table 2.
Accuracy (precision and trueness). The precision of the meth
cated in Table 2. Calibration curves were constructed using a
od
nalyte/
surrogate peak area ratio versus concentration of analyte. Cinc in terms of intra- and inter-day variability was evaluated us
ing
ophen
1
(50 ng g ) was selected as surrogate. Each calibration le spiked compost samples at three concentration levels (5, 28
and
vel was
1
44 ng g ) for each compound. Precision (expressed as rela
tive

standard deviation, %RSD) was determined from triplicate s samples during the same day and in 5 different days. The va
piked
lues
obtained are summarized in Table 3.

Fig. 4. (A) Extraction recoveries and (B) matrix effects (% signal suppression) for the target antibiotics for the whole optimized extraction method. E
rror bars indicate
standard deviation.
N. Dorival-Garca et al. / Talanta 138 (2015) 247257

255

Table 2
Analytical and statistical parameters.
Parameters

MOX

MAR

42
1

b (g ng )
1
Sb (g ng )
Sy/x
2
R (%)
Plof (%)
1
LOD (ng g )
LOQ (ng g1 )
1
LDR (ng g )

Parameters

42
2

ENR

42
2

LOM

42
2

CIP

42
2

ENO

42
2

NOR

42
3

PIP

42
3

42
3

1.1 10
5
6.2 10
3
6.1 10
99.88
32.2
0.3
0.9
0.944

1.5 10
4
1.2 10
3
8.5 10
99.75
11.1
0.3
1.0
1.564

2.2 10
4
1.6 10
3
8.2 10
99.82
19.9
0.2
0.7
1.256

1.5 10
5
7.9 10
3
7.8 10
99.90
35.1
0.3
0.9
0.956

1.0 10
5
4.7 10
3
4.6 10
99.92
47.5
0.2
0.8
0.856

5.7 10
5
3.5 10
3
4.4 10
99.84
26.7
0.4
1.2
1.256

1.6 10
5
1.5 10
3
1.5 10
99.66
11.2
0.5
1.5
1.544

9.7 10
4
2.3 10
2
1.1 10
98.85
11.2
0.4
1.3
1.844

DIF

SAR

DAN

PIR

CIN

OXO

FLU

NAL

42
1

b (g ng )
1
Sb (g ng )
Sy/x
2
R (%)
Plof (%)
1
LOD (ng g )
1
LOQ (ng g )
1
LDR (ng g )

OFL

42
2

2.6 10
5
9.4 10
3
8.3 10
99.95
73.8
0.2
0.6
0.644

42
2

1.3 10
5
5.7 10
3
5.6 10
99.92
47.6
0.2
0.7
0.744

42
2

2.2 10
4
1.3 10
2
1.3 10
99.87
28.1
0.3
1.0
1.056

42
2

5.5 10
4
1.5 10
2
2.2 10
99.97
14.6
0.2
0.6
0.664

42
1

1.1 10
4
4.3 10
2
4.3 10
99.95
72.1
0.2
0.6
0.656
2

42
2

5.5 10
4
1.9 10
2
1.9 10
99.96
82.8
0.2
0.6
0.656

2.5 10
5
2.6 10
3
1.2 10
99.32
18.9
0.4
1.3
1.944

42
2

5.3 10
4
1.6 10
2
1.6 10
99.97
11.4
0.2
0.5
0.556

6.7 10
4
3.4 10
2
4.4 10
99.88
37.9
0.3
1.0
1.056

bslope; Sbslope standard deviation; Sy/x regression standard deviation; R determination coef cient; %; Plof P value of lack-of-t test; LO
Dlimit of detection;
LOQlimit of quanti cation; LDRlinear dynamic range.
a

npoints of calibration.

Fig. 5. UHPLCMS/MS chromatograms of (A) a spiked, and (B) a positive compost sample with the target analytes containing the corresponding
surrogate (CIC).
256

N. Dorival-Garca et al. / Talanta 138 (2015) 247257

Table 3
RSD values fell between 0.1% and 6.4%. Inter-day
precision was
Accuracy of the method. Precision and trueness of target compounds in samples
lower than 15%. Therefore, all compounds were within the .
acceptable limits, which are considered r15% of the actual Compound Spiked level Recovery (%)
IntraPrecision
a
value,
day
Inter-day (%)
except at the LOQ, which it should not deviate by more than
20%.
The data indicate that the method is reproducible.
compost samples. The results are shown in Table 4. Quinol
Additionally, due to the absence of Certi ed Reference one
Matederivatives that are used in both, human and veterinarian m
rials, recovery assays were carried out to validate the meth ediods in
cine were found, which means that compost raw material inclu
terms of trueness. Recoveries were determined analyzi des
ng the
urban and farms sewage. CIP, OFL and ENR were the only qu
spiked blank samples and the concentration of each com inopound
lones that were present in all the analyzed samples, which
was determined by interpolation in the standard calibration c also
urve
coincides with the highest concentrations that were found,
within the linear dynamic range and compared with the a folmount
lowed by MOX and NOR. The rest of quinolones were pres
of analytes previously added to the samples. As shown in Ta ent
ble 3,
between 50% and 87.5% of the samples, being the rstthe recoveries were close to 100% (95.3-106.3%) in all
generation
cases.
quinolones NAL and CIN, the second-generation quinolone N
Precision and trueness data indicate that the met OR,
hod to
and the third-generation quinolone MOX, the most freque
determine the target compounds in compost samples is acc ntly
urate,
found antibiotics. LOM, ENO, PIP, DIF, SAR and DAN were not fou
and that the presence of co-extracted matrix components, nd
which
in any of the analyzed samples. Fig. 5B shows a chromatogram
typically suppress the analyte signal in mass spectrometry, d of a
id not
contaminated compost sample.
affect the performance of the method.
According to the European Surveillance of Antimicrobial Co
n3.4. Application of the method
sumption (ESAC) [41], CIP, OFL, LEV, NOR and MOX are among
the
The method was applied to the analysis of eight comm most used quinolones in Europe, which could be an indication t
ercial
hat

the high levels (hundreds of ng g ) that were found for

these
quinolones could be related with their high consumption.
Moreover, ESAC's study also con rms the continued substantial
use of
quinolones in Europe (e.g. MOX). On the other hand, other
study
[42] also reveals the much greater proportion of quinolone
s that
are consumed in Spain, especially by the elderly, although
these
antibiotics are not considered as rst-line treatment f
or the
majority of infections in primary care, according to
Spanish
guidelines. Another explanation for the high concentration
s that
were found for some quinolones in the analyzed compost sa
mples
could be their recognized persistence, which have been a
lready
proved in sludge [43], although there is not yet enough evi
dence
for this behavior for quinolones in compost. It is noteworthy
that
ENR and MAR were the only quinolones of exclusively veterin
arian
use that were found in samples, which conrms that ENR
is the
most used veterinarian quinolone and the extensive use of M
AR in
the treatment of a wide range of diseases in dogs and cats,
which
are the most common pets.

(%) (n6) (n30)

MOX

5
28
44

MAR

5
28
44

OFL

5
28
44

ENR

5
28
44

LOM

5
28
44

CIP

5
28
44

ENO

5
28
44

NOR

5
28
44

PIP

5
28
44

DIF

5
28
44

SAR

5
28
44

DAN

5
28
44

PIR

5
28
44

CIN

5
28
44

OXO

4.

5
28
44

Conclusions
FLU

103.2
104.0
101.5
99.6
100.3
100.6
99.4
98.8
99.7
99.0
98.5
100.0
101.4
99.6
99.7
98.5
100.5
102.3
95.3
97.7
97.5
96.1
100.3
98.9
97.0
103.1
101.2
100.1
98.5
99.8
100.7
101.6
100.0
100.1
99.5
99.9
100.6
98.9
99.4
102.5
99.7
99.3
99.7
101.2
99.8
100.1
100.1
100.5
106.3
100.4
98.9

2.6
0.8
0.4
3.1
2.2
0.2
1.3
1.0
1.3
0.1
1.1
0.5
3.0
0.7
0.6
1.0
0.1
2.4
0.9
0.3
0.8
0.7
2.6
1.1
1.5
1.7
0.9
0.9
0.2
0.4
1.7
1.1
0.3
0.8
0.1
0.3
1.4
0.1
0.3
0.1
0.4
0.1
0.1
0.2
0.2
0.1
0.1
0.1
0.5
0.6
0.4

4.1
1.6
1.4
1.8
1.8
1.3
1.3
2.8
1.8
1.1
3.4
0.7
4.3
0.8
1.3
1.0
1.2
2.6
6.4
0.8
3.1
0.7
1.6
1.0
0.6
1.0
1.1
2.5
1.2
1.1
1.2
1.9
1.4
1.4
1.9
1.6
1.3
1.3
0.7
1.9
1.6
0.2
1.2
1.5
0.9
0.6
0.4
0.6
1.1
1.1
1.3

The present work presents an effective multi-residue proc

28
edure
44
5
to determine the content of strong sorbed quinolones in co NAL
28
mpost
44
samples. This approach requires the use of a drastic extractio
a
RSD (%) percentages.
n procedure which was carefully optimized using chemometric strat
egies.
The selected extraction technique was MAE at low pH value combining a quick and simple cleanup step based on SALLE and
s proDvided by an extraction solvent containing m-phosphoric ac SPE with UHPLC MS/MS analysis, the obtained method is b

id. By
oth
selective and sensitive. This procedure enables the determination
of
1
target analytes at ng g levels. Finally, the method was successf
ully
1
validated, obtaining very low LOD (between 0.2 and 0.5 ng g ), h
igh
extraction recoveries (72 96%) and precision, besides the
signicant
reduction of matrix effects, which is an important achieve
ment
considering the strong interactions that form quinolones
with
matrices with high content in organic matter.
This method becomes a good alternative to the very few
and
inefcient existing methods, since it offers besides simplicity
and
rapidity of operation, very good extraction recoveries for the targ
et

antibiotics. The method was applied to the determination o may be used to perform screening studies about the presence a
f the
nd
levels of the target antibiotics in different compost samples a
nd it
N. Dorival-Garca et al. / Talanta 138 (2015) 247257

257

Table 4
1
Concentrations of quinolones (ng g ) in commercial compost samples.
Found amount in samples (ng g1 )

S1

S2

S3

S4

S5

S6

S7

S8

MOX

79.5 (0.6)

43.4 (0.4)

11.7 (0.1)

9.2 (0.2)

ND

ND

13.7 (0.3)

54.5 (0.8)

MAR
OFL
ENR
CIP
NOR
PIR
CIN
OXO
FLU
NAL

ND
246.5 (1.6)
22.7 (1.4)
667 (2)
58 (1)
ND
9.60 (0.05)
9.7 (0.3)
ND
3.0 (0.2)

ND
91.7 (0.8)
674.4 (2.3)
572 (2)
ND
4.0 (0.1)
3.0 (0.2)
9.0 (0.6)
3.7 (0.4)
3.2 (0.1)

ND
215.6 (2.2)
36.8 (0.7)
329 (2)
ND
ND
2.70 (0.07)
15.6 (1.2)
ND
23 (1)

ND
200.0 (2.4)
14.2 (0.4)
546 (2)
79 (1)
ND
1.5 (0.1)
6.6 (0.4)
ND
2.1 (0.3)

13.9 (0.1)
719.2 (1.6)
70.5 (1.2)
836 (2)
131 (2)
1.8 (0.1)
3.6 (0.3)
18.4 (1.3)
ND
16.5 (1.1)

13.5 (0.2)
66.1 (0.7)
18.0 (1.1)
167 (2)
ND
ND
3.8 (0.1)
ND
2.9 (0.3)
1.3 (0.1)

ND
10.7 (1.1)
9.0 (0.3)
152 (2)
43 (2)
6.1 (0.3)
3.5 (0.2)
2.2 (0.3)
1.7 (0.2)
ND

ND
119.7 (2.1)
24.5 (0.2)
279 (2)
50 (1)
1.2 (0.2)
ND
ND
5.8 (0.3)
8.4 (0.3)

Mean of 6 determinations; ND: not detected (oLOD). Standard deviations are in parentheses.

nal fate of these substances in the environment, as well as

in the
study of the efciency of the composting process in the remo
val of
this important family of antibiotics, taking into consideration
that
compost is used on agricultural land, where they pose a s
erious
environmental threat.

Acknowledgements
This study was supported by the Spanish Ministry of S
cience
and Innovation (Project no. CTQ2011-24210).

Appendix A.

Supplementary material

Supplementary material associated with this article c

an be
found in the online version at doi:10.1016/j.talanta.2015.03.
011.

References
[1] L. Fink, I. Dror, B. Berkowitz, Chemosphere 86 (2012) 144149.
[2] K. Kmmerer, Chemosphere 75 (2009) 417434.
[3] R.L. Oulton, T. Kohn, D.M. Cwiertny, J. Environ. Monit. 12 (2010) 1956
1978.
[4] J.M. Brausch, G.M. Rand, Chemosphere 82 (2011) 15181532.
[5] M. Gros, M. Petrovi , A. Ginebreda, D. Barcel, Environ. Int. 36 (2010)
15 26.
[6] V. Homem, L. Santos, J. Environ. Manag. 92 (2011) 23042347.
[7] Spanish Agency of Medicines and Health Products. Use of Antibiotics in
Spain,
2009. Available from: http://www.aemps.gob.es/medicamentosUsoHuma
no/
observatorio/docs/antibioticos.pdf (accessed 01.10.14).
[8] X. Chang, M.T. Meyer, X. Liu, Q. Zhao, H. Chen, J.A. Chen, Z. Qiu, L. Yang
, J. Cao,
W. Shu, Environ. Pollut. 158 (2010) 14441450.
[9] D.I. Mass, N.M.C. Saady, Y. Gilbert, Animals 4 (2014) 146163.
[10] E.M. Golet, A.C. Alder, W. Giger, Environ. Sci. Technol. 36 (2002) 3645
3651.
[11] A. Speltini, M. Sturini, F. Maraschi, A. Profumo, A. Albini, Trends Anal. Che
m. 30
(2011) 13371350.
[12] M. Sturini, A. Speltini, F. Maraschi, A. Profumo, L. Pretali, E. Fasani, A.
Albini,

Chemosphere 86 (2012) 130137.

[13] A.B.A. Boxall, P. Johnson, E.J. Smith, C.J. Sinclair, E. Stutt, L.S. Levi, J.
Agric. Food
Chem. 54 (2006) 22882297.
[14] M. Lillenberg, S. Yurchenko, K. Kipper, K. Herodes, V. Pihl, K. Sepp, R.
Lhmus,
L. Nei, J. Chromatogr. A 1216 (2009) 59495954.
[15] K. Hara, T. Mino, Waste Manag. 28 (2008) 26452652.
[16] F. Byksnmez, S. Sekeroglu, J. Residuals Sci. Technol. 2 (2005) 31
40.

[17] Y.B. Ho, M.P. Zakaria, P.A. Latif, N. Saari, Bioresour. Technol. 131 (2013)
Anal. Bioanal. Chem. 406 (2014) 12591266.
[31] M. Bourdat-Deschamps, S. Leang, N. Bernet, J.J. Daudin, S. Nlieu, J.
476484.
[18] K.R. Kim, G. Owens, Y.S. Ok, W.K. Park, D.B. Lee, S.I. Kwon, Waste Manag. Chromatogr. A 1349 (2014) 1123.
32
[32] M.B.R. Cerqueira, J.R. Guilherme, S.S. Caldas, M.L. Martins, R. Zanella,
(2012) 110116.
E.G. Primel, Chemosphere 107 (2014) 7482.
[19] M. Lillenberg, S. Yurchenko, K. Kipper, K. Herodes, V. Pihl, R. Lhmus, M. Iva
sk,
[33] M.V. Salvia, C. Cren-Oliv, E. Vulliet, J. Chromatogr. A 1315 (2013) 53
A. Kuu, S. Kutti, S.V. Litvin, L. Nei, Int. J. Environ. Sci. Technol. 7 (2010) 307 60.
[34] W. Peysson, E. Vulliet, J. Chromatogr. A 1290 (2013) 4661.
312.
[20] J. Ramaswamy, S.O. Prasher, R.M. Patel, S.A. Hussain, S.F. Barrington, Biores [35] M.P. Hermo, D. Barrn, J. Barbosa, Anal. Chim. Acta 539 (2005) 7782.
our.
[36] M.S. Daz-Cruz, D. Barcel, Trends Anal. Chem. 24 (2005) 645657.
[37] A.J. Carrasquillo, G.L. Bruland, A.A. MacKay, D. Vasudevan, Environ. Sci.
Technol. 101 (2010) 22942299.
Tech[21] J.L. Tadeo, C. Snchez-Brunete, B. Albero, A.I. Garca-Valcrcel, R.A. Prez,
Cent.
nol. 42 (2008) 76347642.
[38] W.H. Tsai, H.Y. Chuang, H.H. Chen, J.J. Huang, H.C. Chen, S.H. Cheng, T.C. H
Eur. J. Chem. 10 (2012) 480520.
uang,
[22] O. Zuloaga, P. Navarro, E. Bizkarguenaga, A. Iparraguirre, A. Vallejo,
Anal. Chim. Acta 656 (2009) 5662.
M. Olivares, A. Prieto, Anal. Chim. Acta 736 (2012) 729.
[23] A. Pamreddy, M. Hidalgo, J. Havel, V. Salvad, J. Chromatogr. A 1298 (2013 [39] O.D. Prestes, C.A. Friggi, M.B. Adaime, R. Zanella, Quim. Nova 32 (2009)
)
16201634.
[40] Guidance for Industry, Bioanalytical Method Validation, U.S. Department
6875.
of
[24] N. Dorival-Garca, A. Zafra-Gmez, F.J. Camino-Snchez, A. Navaln,
Health and Human Services, Food and Drug Administration, Center for D
J.L. Vlchez, Talanta 106 (2013) 104118.
[25] Y.B. Ho, M.P. Zakaria, P.A. Latif, N. Saari, J. Chromatogr. A 1262 (2012) 160 rug
Evaluation and Research (CDER) Center for Veterinary Medicine (CVM), 2
168.
001.
[26] B. Albero, C. Snchez-Brunete, E. Miguel, R. Aznar, J.L. Tadeo, J.
Chromatogr. A
[41] N. Adriaenssens, S. Coenen, A. Versporten, A. Muller, G. Minalu, C. Faes,
V. Vankerckhoven, M. Aerts, N. Hens, G. Molenberghs, H. Goossens, J. Ant
1336 (2014) 5258.
i[27] B.P. Chari, R.U. Halden, Water Res. 46 (2012) 48144824.
[28] J. Martn, J.L. Santos, I. Aparicio, E. Alonso, J. Sep. Sci. 33 (2010) 1760
microb. Chemother. 66 (2011) 4656.
[42] S. Malo, M.J. Rabanaque, C. Feja, M.J. Lallana, I. Aguilar, L. Bjerrum, Phar
1766.
[29] M. Anastassiades, S.J. Lehotay, D. Stajnbaher, F.J. Schenck, J. AOAC Int. 86 macol.
Toxicol. 115 (2014) 231236.
(2003) 412431.
[43] N. Dorival-Garca, A. Zafra-Gmez, A. Navaln, J. Gonzlez, J.L. Vlchez,
[30] A. Berlioz-Barbier, A. Vauchez, L. Wiest, R. Baudot, E. Vulliet, C. CrenSci.
Oliv,
Total Environ. 442 (2013) 317328.