Chemosphere 119 (2015) 8–15

Contents lists available at ScienceDirect

Chemosphere
journal homepage: www.elsevier.com/locate/chemosphere

Liquid chromatography–tandem mass spectrometry multiclass method

for the determination of antibiotics residues in water samples
from water supply systems in food-producing animal farms
Malgorzata Gbylik-Sikorska ⇑, Andrzej Posyniak, Tomasz Sniegocki, Jan Zmudzki
National Veterinary Research Institute (NVRI), al. Partyzantow 57, 24-100 Pulawy, Poland

h i g h l i g h t s g r a p h i c a l a b s t r a c t

 Determination of 45 antibiotics from

9 different groups in water samples.
 Sample preparation by extraction and CONTROL
clean-up in a single step.
 Study of the occurrence of antibiotics LC-MS/MS

residues in water supply systems on water

sample
Veterinary inspection
Polish farms.

OUT OF CONTROL
adsorption

constant exposure

water with antibiotics

residues

environment

a r t i c l e i n f o a b s t r a c t

Article history: A sensitive liquid–chromatography–electrospray tandem mass spectrometry multiclass method for
Received 7 February 2014 determination of 45 veterinary compounds belonging to 9 different antibiotic groups, including amino-
Received in revised form 25 April 2014 glicosides (4), b-lactams (13), diaminopyrimidines (1), fluoroquinolones (10), lincosamides (1), macro-
Accepted 28 April 2014
lides (5), pleuromutilins (1), sulfonamides (6) and tetracyclines (4), in water from breeding animal
watering supply system has been developed. Isolation of the analytes was carried out by solid phase
Handling Editor: Klaus Kümmerer extraction with heptafluorobutyric acid as an ion-pair agent on the Strata-X reversed phase cartridges.
All analytes were determined simultaneously in one single run on a C18 column with gradient elution
Keywords: and short analysis time (13 min). Method was validated, average relative recoveries were in the range
Antibiotics of 84.3–109.3% with satisfactory precision are repeatability for all compounds are in the range of 4.7–
Water 12.2%, within-laboratory reproducibility are in the range of 4.4–13.5% for. The limit of quantitation
Multiclass (LOQ) of the method was in the range of 0.02–10 lg L1, depending of analyte. The applicability of the
LC–MS/MS method was tested by determining antimicrobial compounds in real water samples collected from water
supply systems in breeding animal farms. The average antibiotics concentration in real water samples
were, respectively, in the range of 0.14–1670 lg L1.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction

⇑ Corresponding author. Tel.: +48 81 889 31 27; fax: +48 81 886 25 95. Despite recommendations, the usage of antibiotics in food-
E-mail address: malgorzata.gbylik@piwet.pulawy.pl (M. Gbylik-Sikorska). producing animals therapy is still increasing (Kemper, 2008).

http://dx.doi.org/10.1016/j.chemosphere.2014.04.105
0045-6535/Ó 2014 Elsevier Ltd. All rights reserved.
 M. Gbylik-Sikorska et al. / Chemosphere 119 (2015) 8–15
Obviously, excluding the use of antibiotics in bacterial diseases in
breed animal farms is impossible. Therefore, many classes of anti-
biotics, such as aminoglicosides, b-lactams, diaminopyrimidines,
fluoroquinolones, lincosamides, macrolides, pleuromutilins, sul-
fonamides and tetracyclines are widely used in current veterinary
practice to treat animal diseases or illegally to promote animal
growth. In poultry and swine therapy, the antimicrobial com-
pounds are commonly added to animal feed or drinking water
sources. Due to the fact that many antibiotics are added to water,
some water supply systems can be contaminated and can spread
to the farm environment. Not using regular cleaning of water sup-
ply systems, especially after finish antibiotic treatment, deliberate
illegal administration of low doses of antibiotics in water to ani-
mals stimulation, as well as remains of unconscious application
of antibiotics, can lead to undesirable consequences. The constant
exposure of breeding animals to these compounds may lead to res-
idues in animal tissues and antimicrobial resistance of animal and
environment pathogens. Moreover, exposure of animal for trace
amounts of antibiotics can also interfere with the pharmacokinet-
ics of other compounds and can influence on action of other anti-
bacterial agents, which were administered intentionally in the
treatment of infection in animals. This may result in prolongation
of the withdrawal period of antibiotic, which was administrated
to treat animals diseases. In water supply systems, the same phys-
ical phenomenon as in the natural aquatic environment, such as
antibiotics accumulation, may also occur. Physicochemical proper-
ties of some antibacterial agents (tetracyclines, fluoroquinolones
and sulfonamides) enable them to absorb onto sewage sludge, soil
and sediments (Kemper, 2008; Kummerer, 2009). Thus, can adsorb
onto water supply system pipes and after then can be systemati-
cally eluted after animal treatment finish. This situation may be
harmful to the animals and consumers of food of animal origin
health and can lead to increase of allergy and bacteria resistance
to antibiotics use in human medicine. Non-controlled antibacterial
residues in water supply systems can also spread to the natural
environment with the farm sludge.
The aim of this study was to develop a non-invasive form of
control of antibiotic residues, which are present in water supply
systems after finishing their administration in water for the treat-
ment and the illegal administration in subliminal doses of antibac-
terial compounds in water for health animals. Therefore, it was
necessary to develop an analytical method for the determination
of antimicrobial compounds residues in water samples collected
from water supply systems in food-producing animals. Until
now, there have been many analytical methods for determining
antimicrobial compounds in different biological matrices, animal
tissues (Chico et al., 2008; Hurtaud-Pessel D. et al., 2011; Granelli
and Branzell, 2007; Smith et al., 2009; Gbylik et al., 2013) and feed
(Chafer-Pericas et al., 2011). In addition, there are many
approaches contributing to the realisation of studies on the occur-
rence of antibiotics in environmental water samples, such as
wastewater, river water, lake water, fish farms and ground water
(Kim and Carlson, 2007; Chang et al., 2010; Borecka et al., 2013;
Gros et al., 2013). Some of the reported methods included the anal-
ysis of only one group of the most important antibiotics used in
veterinary and human therapy, such as sulfonamides (Raich-
Montiu et al., 2007) or fluoroquinolones (Xiao et al., 2008). The
separate determination of antibiotics from different groups in
one sample is more expensive and time consuming. Furthermore,
more frequent detection of the trace amounts, as well as high con-
centrations of antimicrobial compounds from different classes in
the same water sample are observed. The multiclass analytical
methods for determination of antibacterial compounds belonging
to groups such as macrolides, sulfonamides, fluoroquinolones and
tetracyclines (Carlson and Kim, 2007; Tong et al., 2009), b-lactams
(Campagnolo et al., 2002), or lincozamides and thrimethoprim,
9
additionally (Silvia et al., 2005; Kasprzyk-Horden et al., 2007) in
environmental water samples have been described in the litera-
ture. However, most of the methods do not include the analysis
of aminoglycosides in the same analytical protocol (Campagnolo
et al., 2002; Carlson and Kim, 2007; Tong et al., 2009). Most of
them used the solid phase extraction (SPE) followed by liquid chro-
matography coupled with the tandem mass spectrometry (LC–MS/
MS). Typically, these methods used ethylenediaminetetraacetic
acid (EDTA) as a chelating agents and OASIS HLB SPE cartridges
(Campagnolo et al., 2002; Silvia et al., 2005; Karthikeyan and
Meyer, 2006; Carlson and Kim, 2007; Tong et al., 2009). Unfortu-
nately, none of them relates to determination and analysis of anti-
biotics in water samples from food-producing animal water supply
systems, which aim to control systems contaminations and there
are no published methods for simultaneous determination of
aminoglicosides, b-lactams, diaminopyrimidines, fluoroquino-
lones, lincosamides, macrolides, pleuromutilins, sulfonamides
and tetracyclines.
This study presents a simple and efficient multi-class method
for determination of 45 analytes belonging to 9 different antibacte-
rial compound groups in water collected from food-producing ani-
mals water supply systems. The sample preparation step contains
the ion-par agent addition and the Strata-X SPE cartridges usage.
The method was validated according to selected requirements of
the European Decision 2002/657/EC (C.D. 2002/657/EC), (linearity,
precision, recovery and selectivity), for analytical limits of the
method, limit of detection (LOD) and limit of quantitation (LOQ)
were estimated. A developed method is dedicated to determination
of antibiotics residues (trace amounts, in the range of 0.02–
1000 lg L1) after finishing the antibiotics application on
breeding-animal farms, but it cannot be used to determine of the
antibiotic concentrations during treatment (therapeutic doses).
The applicability of the method was evaluated by determining
antimicrobial compounds in water supply systems from food-
producing animals collected from a few different farms in Poland.
2. Materials and methods
2.1. Reagents
All reagents used were of analytical grade, >95% purity. Sodium
acetate and sodium hydroxide were from POCH (Gliwice, Poland),
acetonitrile and methanol were obtained from J.T. Baker (Deventer,
the Netherlands). Heptafluorobutyric acid (HFBA) was from Fluka,
(Newport News, VA, the United States). Water was deionised
(>18 MX cm1) by the Millipore system. Amoxicillin (AMOX),
ampicillin (AMPI), penicillin G (PEN G), nafcillin (NAF), dicloxacillin
(DICLOX), oxacillin (OXA), cephapirin (CFPI), ceftiofur (CFT), cef-
operazone (CFPE), cephalexin (CFLE), cefquinome (CFQ), cefazolin
(CFZ), cefalonium (CFLO), danofloxacin (DAN), difloxacin (DIF),
enrofloxacin (ENR), ciprofloxacin (CIP), norfloxacin (NOR), marbo-
floxacin (MAR), flumequine (FLU), sarafloxacin (SAR), oxolinic acid
(OXO), nalidix acid (NAL), chlortetracycline (CTC), tetracycline (TC),
doxycycline (DC), oxytetracycline (OTC), streptomycin (STRP),
dihydrostrepromycin (DISTRP), spectinomycin (SPEC), neomycin
(NEO), sulfamerazine (SME), sulfamethazine (SMT), sulfamethoxa-
zole (SMA), sulfamonomethoxine (SMM), sulfadimethoxine
(SDMX), sulfathiazole (SFT), trimethoprim (TMP), tylosin (TYL),
erythromycin (ERY), spiramycin (SPI), tilmicosin (TIL), josamycin
(JOS), lincomycin (LIN), tiamuline (TIM), sulfafenazole (SFF) –
internal standard (IS), were from Sigma–Aldrich (St. Louis, MO,
the United States). Strata X (100 mg, 6 mL) cartridges were
obtained from Phenomenex (Torrance, CA, the United States), syr-
inge filters 0.22 lm PVDF were from Restek (Bellefonte, PA, the
United States).
10 M. Gbylik-Sikorska et al. / Chemosphere 119 (2015) 8–15

2.2. Standard solutions for 3 min. The column returned to the initial composition and re-
equilibrated within 6 min before the next injection. The MS
Individual stock standard solutions (1 mg mL1) for macrolides instrument was operated in the positive ESI mode. The following
(TYL, ERY, SPI, TIL, JOS), tetracyclines (TC, CTC, DC, OTC), sulfona- parameters were used in the tune mode: resolution Q1 and Q3 –
mides (SME, SMT, SMA, SMM, SDMX, SFT) diaminopyrimidine unit; temperature – 500 °C, nebuliser gas (N2) – 40; curtain gas
(TMP), pleuromutylines (TIM) and lincozamides (LIN) were pre- (N2) – 20; collision gas (N2) – 3; auxiliary gas – 50; ion spray volt-
pared in methanol and stored in amber volumetric flasks at age – 5500 V. The Analyst 1.5 software controlled the LC–MS/MS
18 °C. Whereas for fluoroquinolones (DAN, DIF, ENR, CIP, NOR, system and processed the data. The data acquisition was in multi-
FLU, SAR, OXO, NAL, MAR) standard solutions were prepared in ple reactions monitoring (MRM) mode. The ion transitions and
methanol with addition of sodium hydroxide and stored in amber mass parameters monitored for each analyte are listed in Table 1.
volumetric flasks at 18 °C. For aminoglycosides (STRP, DISTRP,
SPEC, NEO) and b-lactams (AMOX, AMPI, OXA, DIKLOX, PEN G,
2.6. Method validation
NAF, CFPI, CFT, CFLE, CFQ, CFZ) standard solutions were prepared
in water; cefalosporines (CFLO and CFPE) were prepared in aceto-
Water samples free from antibiotics were used in the validation
nitrile and water (1:1, v/v) and stored in amber volumetric flasks
process. Linearity was performed by the matrix-matched calibra-
at 18 °C. Mixtures of working standard solutions were prepared
tion curve which had been prepared by fortifying blank water sam-
in water and stored in plastic flasks at 4 °C. Individual stock inter-
ples at 9 concentration levels (0.02, 0.05, 1, 10, 50, 100, 250, 500
nal standard (IS) solution (1 mg mL1) for SFF was prepared in
and 1000 lg L1). The 50 lL of IS was added to each sample.
water and stored in amber volumetric flasks at 18 °C. Working
Quantitative results evaluation was performed by comparing the
internal standard solution (2 lg mL1) was prepared in deionised
analyte/internal standard peak area ratio from matrix-matched
water and stored in amber volumetric flask at 4 °C.
calibration curve to the analyte/internal standard peak area ratio
in analysed samples.
2.3. Water samples
The repeatability was calculated as the coefficients of variation
(CV, %) of results obtained after fortifying of six blank samples at
For validation experiments, water samples were collected from
three concentration levels: 10, 100 and 500 lg L1. The spiked
water supply systems on local food-producing animal farm. Before
samples were analysed on the two subsequent days with the same
the experiment, water samples were respectively checked to be
instrument and the same operators.
free of the antibiotics. For the purpose of the matrix effect, water
The within-laboratory reproducibility was calculated as the
samples from water supply systems and wastewater samples were
overall coefficients of variation (CV, %) of the results obtained after
collected. For the main study, water samples were collected from
fortifying another two sets of blank samples at the same concen-
water supply systems of 25 food-producing animal farms in
tration levels of analysed compounds as for the repeatability and
Poland. Water samples were collected into dark plastic bottles
analysing on 2 d with the same instrument and another operators.
and kept in a cooler with ice until transportation to the laboratory.
Selectivity was checked by analysis of 20 blank water samples
Once back to the laboratory, water samples were stored at 18 °C
collected from different sources which allows to verify the appear-
until the analysis.
ance of possible presence of interfering substances around the
retention times of the compounds of interest.
2.4. Sample preparation
The percentage relative recovery was evaluated in the same
experiment as repeatability by comparing the mean measured con-
Before extraction, 50 lL of IS was added to a 250 mL water sam-
centration with the fortified concentration of the samples.
ple in polypropylene bottle (500 mL), mixed and left at room
The method detection limit (LOD) and the quantification limit
temperature in a dark place for 15 min, after that 6 mL of 0.5 M
(LOQ) were estimated after analysis of 20 fortified blank samples
sodium acetate pH = 5.6 and 30 lL of HFBA were added and shaken
at the minimum detectable concentration level of each analyte
briefly for 5 min. Such prepared samples were transferred to pre-
and measured the signal to noise ratios at 3 and 10 respectively.
conditioned Strata-X SPE cartridges (sequentially conditioned with
The stability of antibiotics in water samples were evaluated by
5 mL of MeOH, 5 mL H2O and 5 mL of 0.05 M HFBA). The flow was
analysing samples fortified at the 10 lg L1 concentration level
no faster than 1 drop/5 s. After then cartridges were vacuum-dried
concentration for each of 45 compounds. The stability of individual
for 5 min at a pressure ranged from 12 mmHg to 18 mmHg. The
stock standard solution for all analytes were estimated in the
analytes were eluted with 3 mL mixture of acetonitrile and
following order: 1 d, 1 and 2 weeks, 1, 3, 6 and 10 months. The
0.05 M HFBA (9:1, v/v) twice. Eluates were collected in 10 mL glass
standard solutions were stored in a dark place at two different
tubes and evaporated to dryness under a stream of nitrogen at
temperatures: 4 °C and 18 °C. The prepared samples were
45 ± 5 °C. The residues were dissolved in 500 lL of 0.025% HFBA
stored at 18 °C and analysed after 1, 7 and 14 d, 1, 2, 3 and
and filtered through 0.22 lm PVDF syringe filters into LC vials.
6 months.
The matrix effect was calculated by analysing 6 water samples
2.5. Liquid chromatography–mass spectrometry
from different sources spiked at 10 lg L1 concentration level for
each analyte and it was calculated by means of the Eq. (1). The
The LC–MS/MS analysis was performed on the Agilent 1200
matrix effect was evaluated as a percentage of signal intensity of
HPLC system (Agilent Technologies, Germany) with an automatic
water sample extract fortified after extraction (IWM) in relation
degasser, a binary pump and an autosampler connected to the
to signal intensity of deionised water fortified (IWD) at the same
AB Sciex API 4000 triple quadrupole mass spectrometer (AB Sciex,
concentration level.
Canada). The chromatographic separation was performed on the
 
Luna C18 (2) 100A column (50  3.0 mm, particle size 3 lm, Phe- IW M
nomenex), which was maintained at 30 °C. The flow rate of mobile Matrix effect ð%Þ ¼  100 ð1Þ
IW D
phase was 400 lL min1, the injection volume – 30 lL. The optimal
composition of mobile phase A and B was acetonitrile (A) and The robustness of the method was estimated for the four different
0.025% HFBA (B). The mobile phase gradient program started at possible factors that could have an influence on the method results.
85% of B, 60% B at 1 min, 40% B at 3 min and then 5% at 4 min, held Among possible factors are: the volume of solvent for conditioning
 M. Gbylik-Sikorska et al. / Chemosphere 119 (2015) 8–15 11

Table 1
MS/MS ion transitions and parameters.

Analyte Ion transition 1 (m/z) Ion transition 2 (m/z) Dwell time (ms) DP (V) CE (Ev)
SPEC 351.1/333.2 351.1/207.2 20 32 67
STRP 582.0/263.0 582.0/246.0 20 52 166
DISTRP 584.3/263.2 584.3/246.2 20 42 150
NEO 615.3/161.0 615.3/163.2 20 109 42
AMOX 366.1/349.1 366.1/208.0 20 14 45
PEN G 335.1/160.0 335.1/176.1 20 17 60
AMPI 350.1/106.0 350.1/160.0 15 58 27
DICLOX 470.0/160.0 470.0/311.0 15 50 22
NAF 415.0/199.0 415.0/171.0 15 48 20
OXA 402.0/160.0 402.0/243.0 15 52 25
CFPI 424.0/152.0 424.0/124.0 20 35 50
CFT 524.0/241.0 524.0/125.0 20 25 50
CFQ 529.0/134.0 529.0/125.0 20 25 50
CFLO 459.0/337.1 459.0/152.0 20 16 46
CFZ 455.0/323.0 455.0/156.0 20 15 50
CFLE 348.0/158.0 348.0/106.0 20 10 50
CFPE 646.0/530.0 646.0/530.0 20 17 60
TMP 292.1/262.2 292.1/231.3 15 52 36
CIP 332.0/314.0 332.0/231.0 15 28 65
ENR 360.0/342.0 360.0/286.0 15 33 100
DIF 400.5/382.1 400.5/356.0 15 30 50
DAN 358.0/340.0 358.0/255.0 15 60 33
FLU 262.1/244.0 262.1/202.0 15 44 25
OXO 262.0/244.0 262.0/216.0 15 53 25
NAL 233.0/215.0 233.0/187.0 15 42 30
MAR 363.0/345.0 363.0/320.0 15 70 30
SAR 385.8/368.1 385.8/348.0 15 50 31
NOR 320.0/302.0 320.0/231.0 15 30 50
ERY 734.0/576.5 734.0/158.2 15 28 75
TYL 916.0/174.0 916.0/772.5 15 52 110
TIL 869.6/696.5 869.6/174.2 15 135 61
JOS 828.2/173.9 828.2/229.0 15 80 46
SPI 843.5/540.4 843.5/174.2 15 120 44
LIN 407.2/126.1 407.2/359.3 15 74 28
TIM 494.4/192.2 494.4/118.8 15 128 30
SMT 279.2/156.0 279.2/108.0 10 25 50
SME 265.0/156.0 265.0/108.0 10 27 50
SDMX 311.0/156.0 311.0/108.0 10 23 50
SMA 254.0/107.8 254.0/155.9 10 24 40
SMM 281.0/156.0 281.0/108.0 10 35 50
SFT 256.0/156.0 256.0/108.0 10 53 20
DC 445.0/428.0 445.0/154.0 15 23 50
OTC 461.0/426.0 461.0/444.0 15 28 40
TC 445.0/410.0 445.0/427.0 15 27 55
CTC 479.0/444.0 479.0/462.0 15 29 60

of SPE cartridges and volume of elution solvent, the temperature of
the evaporation of the extracts, the size of syringe filters. The study
was performed with eight water samples spiked at the level of
10 lg L1. The results were evaluated on the basis of the Youden
approach, described by Plackett and Burman, 1946 and Van der
Heyden et al., 2001.
3. Results and discussion
3.1. LC–MS/MS conditions
The mass spectrometer settings were optimised with a direct
infusion of working standard solutions. For all compounds, the
two most abundant product ions produced from each precursor
ion were chosen as the ion transition in order to comply with the
criteria needed for qualitative and quantitative methods. Accord-
ing to the EU criteria (C.D. 2002/657/EC), two MRM transitions
were monitored. The characteristic MS/MS parameters: collision
energy (CE), declustering potential (DP) and dwell time have been
optimised separately for each analyte (Table 1). The analyses were
performed in the positive ionisation mode.
In order to improve the chromatographic separation of the
selected antibiotics from the possible matrix interferences, the LC
conditions were optimised. In the majority of the published meth-
ods, for the determination of the antibacterial compounds in water
samples, the sufficient chromatography separation was obtained
with a different C18 column (Raich-Montiu et al., 2007; Carlson
and Kim, 2007; Kasprzyk-Horden et al., 2007; Xiao et al., 2008;
Tong et al., 2009). This can be explained by the fact that highly
polar compounds are not retained and elute in the void volume
of column when reversed-phase separation is used. Ion-pair chro-
matography can be the solution to this problem, however, asym-
metric peak shape, low signal intensity also ion suppression may
occur. For this reason, C18 columns with different parameters (col-
umn size and particle size) were chosen for this experiment. It was
found that the best result (high selectivity, good peak shape, higher
intensity of most compounds and shorter retention times), was
obtained with the Luna C18 (2) 100A column (50  3.0, 3 lm). At
the second step, different compositions of the mobile phase includ-
ing ion-pair agents, were tested. Mobile phase compositions of
methanol and acetonitrile coupled with formic acid or water
including one of two ion-pair agents such as heplafluorobutyric
acid and pentafluoropionic acid were tested. The use of the ion-pair
12 M. Gbylik-Sikorska et al. / Chemosphere 119 (2015) 8–15
chromatography was necessary because of the polar character of
aminoglycosides (Karthikeyan and Meyer, 2006). Unfortunately,
the addition of ion-pair agent caused a significant reduction of sig-
nal intensity of ERY, but lower concentration of ion-pair agent
caused the lost of aminoglycosides elution from column. The best
results such as better aminoglycosides elution from chromatogra-
phy column, satisfactory peak shapes and reproducible retention
time were obtained with 0.025% HFBA combined with acetonitrile.
Because of this compromise between adequate ion-pair agent vol-
ume and pH of mobile phase, therefore the LOQ for ERY, NEO,
AMOX and PEN G are in the range of 5 lg L1 to 10 lg L1.
Furthermore, the mobile phase gradient was also optimised in
order to achieve the best chromatographic result with the mini-
mum analysis time. The last eluting analyte was detected at
7.66 min, the total run time was 13 min. In Fig. 1, total ion current
(TIC) chromatogram of real sample spiked on 5 lg L1 of mixture
containing 45 analytes is displayed.
3.2. Optimisation sample preparation
Water samples from water supply systems from breeding-
animals farms are specific kind of matrix, significantly differing
from the environmental water samples. Besides the antibiotics,
they can contain vitamins, nutritional formulas, special taunts,
feed, plant preparations (garlic), dyes and some other veterinary
pharmaceuticals. The analytes belonging to different antibacterial
compound groups isolation from such complex matrices is
difficult.
SPE extraction and concentration is a very important step in
water sample preparation. According to the literature, before the
process of SPE (Karthikeyan and Meyer, 2006; Raich-Montiu
et al., 2007; Carlson and Kim, 2007; Smith et al., 2007; Tong
et al., 2009), generally water samples are filtered through glass-fib-
ber filters (Campagnolo et al., 2002; Karthikeyan and Meyer, 2006)
and then the pH value of those samples were adjusted
(Campagnolo et al., 2002; Raich-Montiu et al., 2007; Tong et al.,
2009). In this study, a few experiments with application of
Fig. 1. Representative total ion current (TIC) chromatogram of real water sample spiked on 5 lg L1 concentration level by mixture of all analytes.
different pH values of the solution, filtration step, ion-pair agent
and chelating agent addition prior SPE extraction were conducted.
At first, the necessity of the filtration step was tested and no sig-
nificant influence on the SPE extraction was detected. Similarly to
reports of other authors, it is very important to adjust the pH val-
ues to acidic and neutrals compounds. The pH value is a critical
part of the sample preparation and SPE extraction process because
the pKa values of analytes are in the range of about 2.6–9.7. The
additions of a different buffer prepared with non-organic acid like,
hydrochloric acid and organic acids (formic acid, acetic acid) and
sodium acetate were tested. All buffers were tested with pH value
in the range of 3.0–6.0. It was difficult to find the compromise
between very sensitive to the presence of acid b-lactams and mac-
rolides, and tetracyclines, fluoroquinolones and sulfonamides, for
which the isolation is better in acidic solution. The most satisfac-
tory recoveries were achieved with 0.5 M sodium acetate pH = 5.6.
In the previous studies, addition of EDTA to water sample to
improve tetracyclines isolation was used (Xiao et al., 2008;
Chang et al., 2010; Gros et al., 2013). During the experiment, there
was no significant impact on tetracyclines recoveries with EDTA
addition into the sample.
In the next step the employing of the ion-pair agents which are
commonly recommended in the sample extraction to enable the
aminoglicosides extraction, took place. The selection of ion-pair
agent (heptafluorobutyric acid and pentafluoropionic acid) was
performed during the laboratory experiment. Both ion-pair agents
delivered good results but heptafluorobutyric acid gave better ami-
noglycosides isolation (recoveries) from water samples and more
satisfactory peak shapes at final chromatograms.
Afterwards, the extraction and clean up process was evaluated.
Clean up process and adequate SPE cartridges choice were very
important in sample preparation step. The SPE cartridges, includ-
ing Strata-X (100 mg, 200 mg), Strata-XL (500 mg), Oasis HLB and
C18 were tested (Silvia et al., 2005; Raich-Montiu et al., 2007;
Carlson and Kim, 2007; Xiao et al., 2008; Tong et al., 2009), because
only this kind of cartridges enables all different compounds to iso-
lation in neutral conditions. For the elution of the analytes metha-
nol, acetonitrile and the combination of them with the ion-pair
 M. Gbylik-Sikorska et al. / Chemosphere 119 (2015) 8–15 13

agent were checked. The best choice was found with Strata-X car- Table 3
tridges and the mixture of acetonitrile with HFBA (9:1, v/v) as the Matrix effects: negative values characterise the ion suppressions, and positive ones –
the ion enhancement () – ion suppression, (+) – ion enhancement.
elution solvent, because only these cartridges usage with ion pair
addition, enable the aminoglycosides elution. Overall, the Strata- Analyte Matrix effect (%) Analyte Matrix effect (%)
X (100 mg and 200 mg) cartridges gave the same recoveries better SPEC 11.6 ± 1.4 OXO 8.6 ± 0.5
then Strata-XL (500 mg). Finally, Strata-X (100 mg) was selected to STRP 10.2 ± 0.7 NAL 5.3 ± 0.4
be used at this study because of shorter time of the SPE sample DISTRP 8.8 ± 0.6 MAR 7.8 ± 0.6
NEO 9.8 ± 0.4 SAR 6.3 ± 0.2
loading step then Strata-X 200 mg. AMOX 9.6 ± 1.1 NOR 5.2 ± 0.4
To enhance the purification of the samples, final extract syringe PEN G 6.3 ± 0.4 ERY 12.4 ± 1.3
filters (PVDF and NYLON) were tested. PVDF syringe filters gave the AMPI 6.8 ± 0.2 TYL 11.8 ± 0.8
best ‘‘purity’’ of the extract without any influence on the analytes DICLOX 5.2 ± 0.6 TIL 10.7 ± 0.6
NAF 7.1 ± 1.4 JOS 9.1 ± 1.6
recoveries.
OXA 6.4 ± 0.4 SPI 10.3 ± 0.4
CFPI 3.6 ± 0.9 LIN 17 ± 0.5
3.3. Method validation CFT 2.1 ± 1.1 TIM 8.2 ± 0.6
CFQ 4.4 ± 0.2 SMT 7.5 ± 0.5
CFLO 3.6 ± 0.3 SME 8.9 ± 0.4
The validation results of the developed method are reported in CFZ 4.9 ± 0.6 SDMX 7.9 ± 1.3
Table 2. During the validation study parameters such as precision CFLE 2.9 ± 0.4 SMA 8.4 ± 0.9
(repeatability and within – laboratory reproducibility), selectivity CFPE 5.4 ± 0.5 SMM 10 ± 1.6
TMP 4.9 ± 0.6 SFT 8.3 ± 1.0
and linearity were established according to the European Decision
CIP 8.3 ± 0.7 DC 11.9 ± 1.3
2002/657/EC (C.D. 2002/657/EC). The matrix-matched calibration ENR 9.9 ± 0.4 OTC 13.6 ± 1.2
curves of 9 points achieved good linearity, greater than r2 = 0.99 DIF 7.1 ± 0.9 TC 10.5 ± 0.9
for all analytes. As can be seen in Table 2, the overall coefficient DAN 8.9 ± 0.5 CTC 11.3 ± 0.5
of variation (CV) for all compounds is in the range of 4.7–12.2% FLU 9.2 ± 0.3

Table 2
Validation report of developed method.

Analyte Repeatability (CV, %) Within-lab reproducibility (CV, %) LOD (lg L1) LOQ (lg L1) Recovery (%)
SPEC 6.1 ± 0.7 7.6 ± 0.7 0.35 1 94.1 ± 3.6
STRP 5.3 ± 0.8 7.1 ± 0.8 0.36 1 96.4 ± 2.4
DISTRP 6.0 ± 1.1 7.6 ± 0.5 0.73 2 91.3 ± 4.2
NEO 7.3 ± 0.7 6.8 ± 0.8 3.79 10 97.1 ± 3.5
AMOX 5.6 ± 1.0 4.0 ± 0.5 3.54 10 97.7 ± 5.1
PEN G 9.4 ± 2.1 12.7 ± 0.3 3.72 10 97.1 ± 3.1
AMPI 7.9 ± 0.6 12.2 ± 0.6 0.02 0.05 105.7 ± 2.5
DICLOX 5.7 ± 1.1 8.7 ± 0.6 0.02 0.05 104.8 ± 3.1
NAF 4.3 ± 0.4 9.2 ± 0.7 0.02 0.05 100.3 ± 1.7
OXA 9.8 ± 1.7 9.3 ± 0.4 0.02 0.05 90.3 ± 2.4
CFPI 8.0 ± 0.6 13.6 ± 0.4 0.02 0.05 92.1 ± 2.8
CFT 7.4 ± 0.5 8.8 ± 0.5 0.02 0.05 102.7 ± 1.9
CFQ 6.2 ± 0.8 9.7 ± 0.6 0.01 0.02 103.2 ± 3.2
CFLO 5.5 ± 1.0 11.2 ± 0.4 0.01 0.02 96.4 ± 1.7
CFZ 8.2 ± 0.8 7.9 ± 0.7 0.01 0.02 99.4 ± 2.4
CFLE 7.8 ± 0.9 13.9 ± 1.5 0.02 0.05 92.7 ± 1.2
CFPE 7.5 ± 1.2 9.9 ± 0.6 0.01 0.02 91.1 ± 2.2
TMP 5.4 ± 0.7 12.5 ± 0.8 0.02 0.05 95.5 ± 2.6
CIP 6.5 ± 0.3 10.5 ± 0.2 0.01 0.02 87.7 ± 2.7
ENR 7.3 ± 0.3 7.1 ± 0.5 0.01 0.02 89.5 ± 3.2
DIF 6.3 ± 0.5 6.1 ± 0.8 0.01 0.02 95.2 ± 1.4
DAN 8.4 ± 0.2 8.3 ± 0.6 0.01 0.02 96.0 ± 1.9
FLU 11.5 ± 0.3 10.2 ± 0.3 0.01 0.02 101.7 ± 3.1
OXO 9.1 ± 0.7 9.1 ± 0.4 0.01 0.02 105.1 ± 1.6
NAL 9.2 ± 0.2 11.1 ± 0.1 0.01 0.02 109.3 ± 2.8
MAR 5.2 ± 0.7 10.9 ± 0.3 0.01 0.02 85.7 ± 1.1
SAR 10.0 ± 0.7 9.1 ± 0.8 0.01 0.02 84.3 ± 1.7
NOR 10.2 ± 0.5 11.6 ± 0.7 0.01 0.02 85.3 ± 2.7
ERY 11.4 ± 0.4 9.4 ± 0.3 2.03 5 96.4 ± 4.6
TYL 11.3 ± 0.6 10.5 ± 0.3 0.01 0.02 91.4 ± 1.6
TIL 6.2 ± 0.4 12.2 ± 0.3 0.02 0.05 103.3 ± 3.1
JOS 7.7 ± 0.4 11.3 ± 0.6 0.02 0.05 103.6 ± 2.7
SPI 8.1 ± 0.4 8.0 ± 0.6 0.02 0.05 95.7 ± 2.9
LIN 4.7 ± 0.3 8.5 ± 0.2 0.01 0.02 98.9 ± 1.6
TIM 5.3 ± 0.6 7.3 ± 0.4 0.01 0.02 96.8 ± 1.6
SMT 6.3 ± 0.9 10.1 ± 0.2 0.01 0.02 106.9 ± 2.1
SME 8.2 ± 0.7 11.4 ± 0.6 0.01 0.02 104.5 ± 1.8
SDMX 8.3 ± 0.3 8.1 ± 0.3 0.01 0.02 98.9 ± 2.3
SMA 10.0 ± 0.7 7.0 ± 0.2 0.01 0.02 102.1 ± 1.6
SMM 9.1 ± 0.8 9.1 ± 0.5 0.01 0.02 91.0 ± 2.6
SFT 5.8 ± 0.8 7.9 ± 0.5 0.02 0.02 105.1 ± 2.2
DC 4.4 ± 0.6 13.3 ± 0.4 0.02 0.05 95.8 ± 1.8
OTC 11.4 ± 0.7 12.9 ± 0.3 0.01 0.02 99.3 ± 1.6
TC 11.0 ± 0.6 8.0 ± 0.3 0.02 0.05 95.9 ± 2.9
CTC 5.3 ± 0.5 10.5 ± 0.4 0.02 0.05 98.8 ± 2.3
14 M. Gbylik-Sikorska et al. / Chemosphere 119 (2015) 8–15

for repeatability and 4.4–13.5% for within – laboratory reproduc- and macrolides is 2 months at least because of the considerable
ibility. All analytes provided the acceptable results in the agree- decrease of concentration levels. All standard stock solution stored
ment with the criteria of the European Decision 2002/657/EC at 4 °C was stable not longer than for 3 months with the exception
(C.D. 2002/657/EC). The percentage of the relative recoveries was for penicillines, cephalosporins and macrolides which are stable for
calculated in terms of the internal standard and is in acceptable 1 month. The stability results for water samples spiced with fluoro-
range of the EU requirement with recovery values ranged from quinolones, sulfonamides, pleuromutilins, diaminopyrimidines at
70% to 120% respectively. The average recoveries for three valida- the 10 lg L1 concentration level for all tested antibiotics remain
tion levels are listed in Table 2. stable at the similar level stored at 18 °C at least for 6 months.
The selectivity of the method improves no interfering peaks co- For aminoglicosides, tetracyclines, macrolides, cefalosporines and
eluted at the same retention time of interest for analytes in any penicillines, significant reductions of concentrations of fresh
sample extracts. extracts were observed after 2 weeks.
The LOD values are in the range of 0.01–3.79 lg L1 and the LOQ The water samples from food-producing animal watering sys-
values are in the range of 0.02–10 lg L1, depending of analyte. tems are known to be rich in mechanical impurities, nutrients, feed
It was found that the stability of proven individual stock stan- residues, vitamin and minerals, which can greatly affect the perfor-
dard solution for fluoroquinolones, sulfonamides, pleuromutilins, mance of the method, especially in the ESI-MS analysis with signal
diaminopyrimidines and tetracyclines stored at 18 °C remains suppression and signal enhancement. The matrix effect was calcu-
unchanged for at least 6 months. The aminoglicosides individual lated for all analytes in different water samples (from water supply
stock standard solutions stored under the same conditions are systems and waste waters). The matrix effect is reported as the
stable for 3 months. The stability of penicillines, cephalosporins average results in Table 3.
As it was found, the average matrix effect expressed as ion sup-
Table 4 pression is in the range of 2.1–12.4%. CFT is characterised by the
The presence of antibiotics in real water samples. slight average value of signal suppressions – 2.1%. A low average
Analyte Antibiotic class Number Min Max signal enhancement was observed in the case of TMP – 6.9%. The
of concentration concentration highest average signal enhancement was calculated for tetracy-
positive level (lg L1) level (lg L1) clines (10.5–13.6%). The results of the studies indicate that the
samples matrix effect was observed for all compounds but it is still not
ENR Fuoroquinolone 19 1.6 1670 affected for determination of them in water samples. The signal
DC Tetracycline 13 0.21 1650 suppression and signal enhancement were minimised by usage
SMA Sulfonamide 3 0.19 58.7
of matrix-matched calibration curves and the internal standard.
TMP Diaminopyrimidine 3 0.14 17.8
TIL Macrolide 1 1.73 1.73 There were no significant differences for the matrix effect between
TIM Pleuromutyline 1 66.8 66.8 wastewater and water from water supply systems. This may indi-
FLU Fluoroquinolone 1 3.48 3.48 cate that other substances in water have no influence on the anti-
LIN Lincosamide 2 104 304 biotics determination with approached method usage.
NEO Aminoglicosyde 1 32 32
The robustness of the method becomes clear during the optimi-
sation of the sample preparation. First, 5, 6, 8 mL of methanol,

Fig. 2. Chromatograms of real water samples with mostly detected analytes: a – water sample contaminated with ENR = 22.6 lg L1 and DC = 75.1 lg L1, b – water sample
contaminated with ENR = 64.4 lg L1 and TIM = 68.8 lg L1, c – water sample contaminated with SMT = 22.5 lg L1 and TMP = 44.5 lg L1, d – contaminated water sample
contaminated with LIN = 304 lg L1.
 M. Gbylik-Sikorska et al. / Chemosphere 119 (2015) 8–15
water and 0.05 M HFBA for SPE cartridges conditioning step were
tested. Then, different values of elution mixture (2  3 mL,
2  4 mL, 2  5 mL) and the temperature evaporation (35–50 °C)
were checked. Additionally, some minor changes like different size
of syringe filters (0.22 lm and 0.45 lm) were investigated, too.
There were no significant differences observed at the evaluation
of robustness. In addition, no degradation of the compounds was
noted at low and higher temperature of evaporation. The syringe
filters resize was not affected for analytes concentration level.
3.4. Application of real samples
In order to check suitability of the proposed method, 50 water
samples obtained from different food-producing animal farms in
Poland were tested. Water samples were only collected on farms
provided with documentation of non-use of antibiotics during
sample collecting. Matrix-matched calibration curves were used
for the quantification of the antibiotics compounds in these sam-
ples, whereas retention times, comparing analyte/internal stan-
dard peak area ratio from matrix-matched calibration curve to
analyte/internal standard peak area ratio in analysed samples,
were also used for identification. The results of this study indicate
that 8 out of 45 antibiotics were detected in the analysed samples
of food-producing animal farms (Table 4).
Specifically, ENR was one of the most frequently detected anti-
biotics (in 19 samples), DC was found in 13 out of analysed sam-
ples. Twenty - six out of the analysed water samples were
contained antibiotics residues. The sample chromatograms of real
water samples are shown in Fig. 2. The presence of antibiotics
was found in 52% out of analysed water samples. This may be
caused by a wrong dosage of pharmaceuticals, as well as antibiot-
ics accumulation in water supply systems because of no cleaning
systems. Both cases provided the occurrence of antibiotics residues
in farm environment and constant exposure of animals to antibac-
terial agents. This situation can cause negative effect on the farm
environment, pass into animals organisms and promote dissemi-
nation of not only antibiotics resistant bacteria, but also the resis-
tance genes in bacterial populations. This is confirm of the
necessity to research and control the degree of water supply sys-
tems contamination by antibiotics.
4. Conclusion
The invented method, allows for a fast quantitative and qualita-
tive analysis of 45 veterinary antibiotics in water from food-pro-
ducing animals water supply systems. The method is based on
the SPE extraction and clean up subsequently analysed by LC–
MS/MS. The proposed method is simple, fast and easy to perform
and allows to determine a wide variety of antibiotics belonging
to 9 different groups in a single run method. The application of
the proposed method allows for a new and non-invasive control
of antibiotic residues in water supply systems. The concentration
of antibiotics residues in real water samples were respectively in
the range of 0.14–1670 lg L1.
Acknowledgements
The authors wish to thank Malgorzata Kotulska for her work in
the validation process, Anna Gajda and Marta Piatkowska for their
participation in sample analysis.
15
References
Borecka, M., Białk-Bielinska, A., Siedlewicz, G., Kornowska, K., Kumirska, J.,
Stepnowski, P., Pazdro, K., 2013. A new approach for the estimation of
expanded uncertainty of results of an analytical method developed for
determining antibiotics in seawater using solid-phase extraction disks and
liquid chromatography coupled with tandem mass spectrometry technique. J.
Chromatogr. A. 1304, 138–146.
Campagnolo, E.R., Johnson, K.R., Karpati, A., Rubin, C.S., Kolpin, S.W., Meyer, M.T.,
Esteban, J.E., Currier, R.W., Smith, K., Thu, K.M., Mcgeehin, M., 2002.
Antimicrobial residues in animal waste and water resources proximal to
large-scale swine and poultry feeding operations. Sci. Tot. Environ. 299, 89–95.
Carlson, K., Kim, S.C., 2007. Quantification of human and veterinary antibiotics in
waater and sediment using SPE/LC/MS/MS. Anal. Bioanal. Chem. 387, 1301–
1315.
Chafer-Pericas, C., Maquiera, A., Puchades, R., Miralles, J., Moreno, A., 2011.
Multiresidue determination of antibiotics in feed and fish samples for food
safety evaluation. Comparison of immunoassay vs LC–MS–MS. Food Control 22,
993–999.
Chang, X., Meyer, M.T., Liu, X., Zhao, Q., Chen, H., Chen, J., Qiu, Z., Yang, L., Cao, J., Shu,
W., 2010. Determination of antibiotics in sewage form hospitals, nursery and
slaughter house, wastewater treatment plant and source water Chongqing
region of three Gorge Resorvoir in China. Environ. Pollut. 158, 1444–1450.
Chico, J., Rubies, A., Centrich, F., Campanyo, R., Prat, M.D., Granados, M., 2008. High-
throughput multiclass method for antibiotics residue analysis by liquid
chromatography–tandem mass spectrometry. J. Chromatogr. A 1213, 189–199.
Commission Decision 2002/657/EC, Off. J Europ Commun L221, 2002 8.
Gbylik, M., Posyniak, A., Kamila, M., Bladek, T., Zmudzki, J., 2013. Multi-residue
determination of antibiotics infish by liquid chromatography–tandem mass
spectrometry. Food Addit. Contam. A 30, 940–948.
Granelli, K., Branzell, C., 2007. Rapid multi-residue screening of antibiotics in
muscle and kidney by liquid chromatography–electrospray ionization–tandem
mass spectrometry. Anal. Chim. Acta 586, 289–295.
Gros, M., Rodriguez-Mozaz, S., Barcelo, D., 2013. Rapid analysis of multiclass
antibiotic and some of their metabolites in hospital, urban wastewater and river
water by ultra-high-performance liquid chromatography coupled to
quadrupole-linear ion trap tandem mass spectrometry. J. Chromatogr. A 1292,
173–188.
Hurtaud-Pessel, D., Jagadeshwar-Reddy, T., Verdon, E., 2011. Development of new
screening method for the detection of 520 antibiotic residues in muscle tissues
using liquid chromatography and high resolution mass spectrometry with a LC-
LTQ-Orbitrap instrument. Food Addit. Contam. A 28 (10), 1340–1351.
Karthikeyan, K.G., Meyer, M.T., 2006. Occurrence of antibiotics in wastewater
treatment facilities in Wisconsin, USA. Sci. Tot. Environ. 361, 196–207.
Kasprzyk-Horden, B., Dinsdale, R.M., Guwy, A., 2007. Multi-residue method for the
determination of Basic/neutral pharmaceuticals and illicit drugs in surface
water by solid-phase extraction and ultra performance liquid chromatography–
positive electrospray ionisation tandem mass spectrometry. J. Chromatogr. A
1161, 132–145.
Kemper, N., 2008. Veterinary antibiotics in aquatic and terrestrial environment.
Ecol. Ind. 8, 1–13.
Kim, S.C., Carlson, K., 2007. Quantification of human and veterinary antibiotics in
water and sediment using SPE/LC/MS/MS. Anal. Bioanal. Chem. 387, 1301–
1315.
Kummerer, K., 2009. Antibiotics in aquatic environment – A – Part I. Chemosphere
75, 417–434.
Plackett, R.L., Burman, J.P., 1946. The design of optimal multifactorial experiments.
Biometrika 33, 305–325.
Raich-Montiu, J., Folch, J., Compano, R., Grandos, M., Prat, M.D., 2007. Analysis of
trace levels of sulfonamides in surface water and soil samples by liquid
chromatography-fluorescence. J. Chromatogr. A 1172, 186–196.
Silvia, M., Cruz, D., Barcelo, D., 2005. LC–MS2 trace analysis of antimicrobials in
water, sediment and soil. Trends Anal. Chem. 24, 7.
Smith, A.J., Ballam, J.L., Ward, A., 2007. The development of a rapid screening
technique to measure antibiotic ancticity on effluents and surface water
samples. Mar. Pollut. Bull. 54, 1940–1946.
Smith, S., Gieseker, C.H., Reimsschuessel, R., Decker, C.H., Carson, M.C., 2009.
Simultaneous screening and confirmation of multiple classes of drug residues in
fish by liquid chromatography–ion trap mass spectrometry. J. Chromatogr. A
1216, 8224–8232.
Tong, L., Li, P., Wang, Y., Zhu, K., 2009. Analysis of veterinary antibiotc residues in
swine wastewater and environmental water samples using optimized SPE-LC/
MS/MS. Chemosphere 74, 1090–1097.
Van der Heyden, Y., Nijhuis, A., Smeyers-Verbeke, J., Vandeginste, B.G.M., Massart,
D.L., 2001. Guidance for robustness: ruggedness tests in method Validation. J.
Pharm. Biomed. Anal. 24, 723–753.
Xiao, Y., Chang, H., Jia, A., Hu, J., 2008. Trace analysis of quinolone and
fluorquinolone antibiotics from wastewaters by liquid chromatography–
electrospray tandem mass spectrometry. J. Chromatogr. A 1214, 100–108.