Journal of Chromatography B, 941 (2013) 6980

Contents lists available at ScienceDirect

Journal of Chromatography B
journal homepage: www.elsevier.com/locate/chromb

A generic screening methodology for horse doping control by

LCTOF-MS, GCHRMS and GCMS
Maroula K. Kioussi a,b, , Emmanouil M. Lyris a , Yiannis S. Angelis a , Maria Tsivou a ,
Michael A. Koupparis b , Costas G. Georgakopoulos c
a

Doping Control Laboratory of Athens, Olympic Athletic Centre of Athens Spyros Louis (OAKA), Kissias 37, 15123 Maroussi, Greece
Laboratory of Analytical Chemistry, Department of Chemistry, University of Athens, 15771 Panepistimiopolis-Zographou, Athens, Greece
c
Anti Doping Laboratory, P.O Box 27775, Doha, Qatar
b

a r t i c l e

i n f o

Article history:
Received 18 June 2013
Accepted 8 October 2013
Available online 16 October 2013
Keywords:
Screening
Equine
Urine
Doping control
Mass spectrometry
Spectra deconvolution

a b s t r a c t
In the present study a general screening protocol was developed to detect prohibited substances and
metabolites for doping control purposes in equine sports. It was based on the establishment of a unied sample preparation and on the combined implementation of liquid and gas chromatographic MS
analysis. The sample pretreatment began with two parallel procedures: enzymatic hydrolysis of sulfate
and glucuronide conjugates, and methanolysis of the 17-sulfate steroid conjugates. The extracts were
treated for LCTOF-MS, GCHRMS and GCMS assays. The majority of the prohibited substances were
identied through a high mass accuracy technique, such as LCTOF-MS, without prior derivatization.
The sample preparation procedure included the formation of methylated and trimethylsilylated derivatives common in toxicological GCMS libraries. The screening method was enhanced by post-run library
searching using automated mass spectral deconvolution and identication system (AMDIS) combined
with deconvolution reporting software (DRS). The current methodology is able to detect the presence of
more than 350 target analytes in horse urine and may easily incorporate a lot of new substances without
changes in chromatography. The full scan acquisition allows retrospective identication of prohibited
substances in stored urine samples after reprocessing of the acquired data. Validation was performed
for sixty representative compounds and included limit of detection, matrix interference specicity,
extraction recovery, precision, mass accuracy, matrix effect and carry over contamination. The suitability
of the method was demonstrated with previously declared positive horse urine samples.
2013 Elsevier B.V. All rights reserved.

1. Introduction
Accredited horse racing doping control laboratories have to
examine hundreds of banned compounds in biological specimens.
The prohibited list of Fdration Equestre Internationale (FEI) [1] is
not limited to 1156 banned and controlled medication substances,
but also extends to other compounds with similar chemical structure or similar biological effect. Moreover, according to Article 6
of the International Agreement on Breeding, Racing and Wagering
(IABRW) published by the International Federation of Horseracing
Authorities (IFHA), masking agents, as well as any substance that
has the potential to affect the performance of a horse, acting on
one or more of the mammalian body systems are forbidden [2].

Corresponding author at: Doping Control Laboratory of Athens, Olympic Athletic

Centre of Athens Spyros Louis (OAKA), Kissias 37, 15123 Maroussi, Greece.
Tel.: +30 210 6853074; fax: +30 210 6834567.
E-mail address: maroulakiousi@yahoo.gr (M.K. Kioussi).
1570-0232/$ see front matter 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jchromb.2013.10.008

Furthermore, the presence of metabolites in urine may provide

additional information and evidence of drug abuse, even though
the parent drug may not be detected. Therefore, doping control
laboratories have to detect analytes with different physicochemical properties, with different ranges of polarity, acidbase
properties, molecular size or solubility and pharmaceutical nature.
Different types of mass spectrometers coupled with liquid and gas
chromatography accomplish the analytical requirements in the
particular eld and they have become indispensable tools for doping control analysis [314]. The available libraries of standardized
electron impact mass spectra are also useful for drug testing [15,16].
The establishment of comprehensive screening approaches that
allow the simultaneous detection of as many prohibited compounds as possible in equine urine is essential. Nonetheless,
procedures providing the simultaneous detection of several drug
classes, which were traditionally handled separately, appear to be
relatively limited in doping control in equine sports [1721]. The
development of a suitable sample preparation for broad screening
is an uphill task allowing for the numerous substances with

70

M.K. Kioussi et al. / J. Chromatogr. B 941 (2013) 6980

different properties to be examined, the interferences from the

complex horse urine matrix and the low concentrations of the drugs
to be screened.
The existing screening methods for the identication of
multi-class analytes in horse urine samples comprise a single
deconjugation procedure and subsequently either a simultaneous analysis by LCMS techniques with polarity switching [19,20]
or different separate analyses for basic and acidic/neutral fractions [17,18]. Regarding LCTOF-MS systems applied in drug
screening, the identication of the analytes is based on the accurate mass and the isotopic pattern of the detected compounds
by conducting library searching in databases [13,14]. Although
LCMS methods provide wide drug coverage, there are still compounds not detected due to low ionization efciency. GCMS
methods have also been adopted for the identication of individual groups of prohibited substances, such as anabolic steroids
[6].
The aim of the hereby study was the development of a generic
methodology for the screening of a large number of prohibited
compounds with different therapeutic activities and acidic/basic
properties. The current approach has the advantage of the determination of over 350 drug targets, by using a unied sample
preparation procedure that combines different deconjugation and
derivatization procedures. The analysis of doping agents was
accomplished by LCTOF-MS, GCHRMS and GCMS techniques
that acted complementarily. The majority of the prohibited substances were identied without prior derivatization and through
a high mass accuracy technique such as LCTOF-MS in full scan
mode which enables retrospective analysis. Furthermore, detection of prohibited drugs not targeted by the above screening
method was achieved through library searching by using automated mass spectral deconvolution and identication system
(AMDIS) incorporated into the deconvolution reporting software
(DRS).

2. Materials and methods

2.1. Materials
Reference materials were purchased from various suppliers
which are presented in detail in Supplement 1. Stock solutions of
the reference substances were prepared in methanol and stored
at 20 C. Multicompound solutions from the above stock solutions were also prepared. -Glucuronidase arylsulfatase isolated
from Helix pomatia was purchased from SigmaAldrich, Steinheim, Germany (Type HP-2; 197,114 units mL1 -glucuronidase
and 876 units mL1 arylsulfatase). Isolute C18(EC) solid phase
extraction columns (sorbent mass 500 mg) were obtained from
Biotage AB (Uppsala, Sweden). Solvents used in the urine extraction procedures were of analytical grade and were acquired
from Labscan (Dublin, Ireland). Sodium hydrogen carbonate
(NaHCO3 , 99.7100.3%) and sodium carbonate anhydrous (Na2 CO3 ,
99.8%) were supplied by Panreac (Barcelona, Spain). Sodium sulfate anhydrous (Na2 SO4 , >99%) was purchased from Lach-Ner,
S.r.o. (Neratovice, Czech Republic). Potassium carbonate (K2 CO3 ,
99101%), methanolic hydrochloride (3 N), iodomethane (CH3 I,
>99.0%), ammonium iodide (NH4 I, 99.999%), ammonium formate (HCOONH4 , 97%) and formic acid (HCOOH, 98100%) were
obtained from SigmaAldrich (Steinheim, Germany). N-methyl-Ntrimethylsilyltriuoroacetamide (MSTFA, 98100%) was acquired
from Pierce Biotechnology (Rockford, IL, USA). 1-Propanethiol
(PrSH) was supplied from Merck (Darmstadt, Germany). HPLCgrade water was obtained by purifying water in ltration system
(Millipore, Billerica, MA, USA) and LCMS grade acetonitrile was
from Carlo Erba Reagenti SpA (Rodano, Italy).

2.2. Instrumentation
2.2.1. LCTOF-MS
The liquid chromatograph was an Agilent 1200 Series rapid resolution LC system (Agilent Technologies, Waldbronn, Germany)
comprising a vacuum degasser, a high-pressure binary pump,
an auto-sampler with a cooled sample tray at 12 C and a column compartment kept at 35 C. Chromatographic separation was
carried out on a reversed phase Zorbax Eclipse Plus C18 column (100 2.1 mm i.d., 1.8 m particle size; Agilent Technologies)
protected with an in-line lter (0.5 m) (Agilent Technologies,
Waldbronn, Germany). The mobile phase was composed of 5 mM
ammonium formate/0.1% formic acid in water (solvent A) and
5 mM ammonium formate/0.1% formic acid in a mixture of acetonitrile:water, 90:10 (v/v) (solvent B). A linear gradient was run
at ow rate of 0.3 mL min1 , with 10% solvent B at initial condition (t = 0 min), increasing to 80% in 9 min. Then, the proportion of
solvent B increased to 100% from t = 9 min to t = 10 min, where it
was held for 3 min. The gradient was then returned to 10% solvent
B within 0.5 min and stabilized for 3.5 min (post-run equilibrium
time) before the next injection. The injection volume was 5 L.
Liquid chromatograph was coupled with an orthogonal
acceleration quadrupole time-of-ight mass spectrometer (6520
Accurate-Mass QTOF/MS; Agilent Technologies, Santa Clara, CA,
USA) through an orthogonal electrospray ionization (ESI) interface. The instrument was operated in positive mode. Nitrogen was
used both as a nebulizer gas (pressure 40 psi) and drying gas (temperature 330 C, ow rate 10 L min1 ). Capillary voltage of the ion
source was set at 3500 V and the fragmentor voltage at 140 V in
order to reduce the fragmentation of the pseudo-molecular ions of
the target analytes. The applied skimmer and octapole 1 RF voltages were 65 V and 750 V, respectively. All the other MS parameters
were automatically optimized by the instrument autotuning procedure, performed on a monthly basis. The mass spectral data were
acquired within the range of m/z 1001100.
Calibration of TOFMS was performed prior to each sequence
run covering the mass range of m/z 118.08631521.9715 by using
a calibration solution provided by the manufacturer (ESI-L Low
Concentration Tuning Mix, G1969-85000, Agilent Technologies).
The full width at half maximum (FWHM) mass resolution ranged
from 4700 (at m/z 118.0863) to 12600 (at m/z 2721.8948). Reference mass correction was accomplished during the analysis,
to achieve better mass accuracies, by continuously introducing
two reference compounds, hexakis (1H,1H,3H-tetrauoropropoxy)
phosphazine (Agilent Technologies) at m/z 922.0098 and benzyldimethylphenylammonium chloride (SigmaAldrich) at m/z
212.1434, simultaneously with the samples into the ESI source from
a second orthogonal nebulizer.
2.2.2. GCMS
2.2.2.1. GCMS for methylated derivatives. The analysis of the
methyl-derivatized target substances was performed on an Agilent 6890N GC system (Agilent Technologies, Palo Alto, CA, USA)
coupled with an Agilent 5973 inert Mass Selective Detector. Chromatographic separation was implemented with a bonded and
cross-linked 5% diphenyl, 95% dimethyl siloxan capillary column
(12 m 0.200 mm i.d., 0.33 m lm thickness, HP ULTRA 2) with
helium used as carrier gas at constant ow 0.6 mL min1 . A 2.0 L
sample volume was injected in split mode (8:1). The front inlet
and the MSD transfer line heater temperatures were set at 250
and 310 C, respectively. The column oven temperature was programmed to increase from 110 to 300 C at 15 C min1 and
maintained at 300 C for 3 min. The run time was 15.67 min, inclusive of equilibration time (0.5 min). MS source and quadrupole
temperatures were maintained at 230 and 150 C, respectively. The
MS system was acquiring data in full scan mode from 50 to 500 amu,

M.K. Kioussi et al. / J. Chromatogr. B 941 (2013) 6980

at a rate of 3.25 scans s1 and electron ionization energy of 70 eV.

Peruorotributylamine (PFTBA) mass-spectrometric grade (Agilent
Technologies) was used as tuning standard.
2.2.2.2. GCMS for trimethylsilylated derivatives. The trimethylsilylate (TMS) derivatized compounds were assayed by GCMS in full
scan mode (40550 amu), for library searching purposes utilizing
AMDIS and DRS programs. An Agilent 6890N GC system (Agilent
Technologies, Palo Alto, CA, USA) combined with an Agilent 5973
quadrupole mass selective detector (MSD) and equipped with a
bonded and cross-linked 5% diphenyl, 95% dimethyl siloxan capillary column (19 m 0.200 mm i.d., 0.33 m lm thickness, HP
ULTRA 2) was used. GC was operating at constant ow of helium
with a ow rate of 1.0 mL min1 and pressure of 10.43 psi. Two
microlitres of sample were injected in split mode of 15:1 at 250 C.
Initial oven temperature was 100 C, then ramped at 20 C min1 to
290 C, held for 5.0 min. MSD transfer line heater, MSD source and
quadrupole temperatures were maintained at 300, 230 and 150 C,
respectively. The run time was 14.50 min. Peruorotributylamine
(PFTBA) mass-spectrometric grade (Agilent Technologies) was used
as tuning standard.
2.2.3. GCHRMS
The assay of the TMS-derivatized target substances was carried out with a three-sector (electrostaticmagneticelectrostatic)
reverse geometry double focusing mass spectrometer (Micromass
AutoSpec-Ultima) coupled to an Agilent 6890N network GC system
(Agilent Technologies, Palo Alto, CA, USA). The GC was equipped
with a bonded and cross-linked 100% dimethylpolysiloxan capillary column (12 m 0.200 mm i.d, 0.33 m lm thickness, HP
ULTRA 1). Helium was used as carried gas at a ow rate of
1.1 mL min1 . The injector port was heated at 250 C and the transfer line at 350 C. One microlitre of sample was injected in split
mode (40:1). Initial oven temperature was 150 C for 0.5 min, then
ramped at 12.5 C min1 to 310 C and held for 2.5 min. The MS
source temperature was maintained at 220 C. The HRMS system
was acquiring data in 10,000 resolution in selected ion recording
(SIR) mode with eleven function groups. Peruorokerosene (PFK)
mass-spectrometric grade (Apollo, Manchester, UK) was used as
calibration reference substance.
2.3. Deconvolution software
Spectral deconvolution of the MS data is a mathematical technique which separates overlapping mass spectra and contributes
to the identication of analytes that are buried under co-eluting
matrix compounds. The fully automated deconvolution process
includes treatment of noise, correction for base line drift and extraction of pure component spectra and related information (e.g. peak
shape, retention time), from complex chromatograms. This individual information can be library searched and matched against
the spectra of the library.
In the present study an Agilent G1716 MSD-DRS reporting
application was used, combining results from the Agilent GC-MSD
ChemStation (G1701DA), the AMDIS32 (version 2.1) and the NIST
2005 Mass Spectral Search Program (NIST05, version 2.0d). DRS
worked with three reference libraries: Association of Ofcial Racing
Chemists AORC R6 Mass Spectral Library, NIST/EPA/NIH Mass Spectral Library (NIST05) and an in-house library comprising the locked
retention time and the mass spectrum of each entry. NIST05 Mass
Spectral Library contains 190,825 electron ionization (EI) spectra of
163,198 different chemical compounds [22]. AORC R6 Mass Spectral Library and in-house library contain 6840 and 377 entries,
respectively. DRS run in data reprocessing mode after GCMS full
scan analyses of the methyl- and TMS-derivatized samples.

71

The automated identication through library searching and

AMDIS-DRS programs was performed in duplicate. Primarily, the
acquired data les of methylated and TMS derivatives were evaluated by DRS in a simple type of analysis based on matching the
deconvoluted spectra with those of the reference libraries, with
a minimum match factor of 70 and without participation of the
retention times in the matching. At rst, AMDIS deconvoluted the
ChemStation data le and the resultant components were compared against the AORC R6 Mass Spectral Library. This comparison
used full scan spectra and was retention time independent. The
results from the AORC library were also searched against a second
library, NIST05 Mass Spectral Library. Finally, the generated DRSreport combined results from the processes described above, in one
easy-to-read format. The second automated detection by means of
AMDIS-DRS programs pertained only TMS-derivatized substances
and it was based on matching both the deconvoluted spectra and
the retention times with those of the in-house library. The identication was accomplished by incorporating the retention index
RI of the detected components in the search algorithm. The use
RI Calibration Data analysis type was selected which included the
deconvoluted spectra and the retention times in the matching. The
AMDIS-based data analysis was performed with a minimum match
factor of 30. A lter was also set in AMDIS, requiring the analytes
retention times to fall within a time window of 5 s. When the difference between the compared retention times was greater than
5 s, a penalty of 70 was abstracted from the match factor. Thus,
the match factor was lower than 30 and the particular substance
was not included in the detected compounds.
2.4. Sample preparation
Urine samples were prepared according to Supplementary Figure 1. Procedure A comprised preparation of 5.0 mL equine urine
fortied with etiocholanolone glucuronide (5 g), mefruside (1 g)
and nalidixic acid (3.5 g) as the internal standards (IS). For hydrolysis urine was adjusted to pH 4.85.5 with 1 M acetate buffer
and incubated at (50 2) C for 3 h after adding 100 L of glucuronidase/arylsulfatase from H. pomatia. After cooling, pH was
adjusted to 9.010.0 with a mixture of NaHCO3 :Na2 CO3 (10:1,
w/w). Basic liquidliquid extraction (LLE) was carried out with
7.0 mL ethylacetate, using anhydrous Na2 SO4 for salting out and
shaking for 10 min. The mixture was centrifuged for 10 min and
the organic phase was separated into three fractions (2.2 mL each)
and evaporated to dryness under a nitrogen stream at (50 5) C.
The rst fraction was reconstituted with 300 L of a mixture of solvent A/solvent B (80/20, v/v) and was analyzed by LCTOF-MS assay
in positive ESI mode. The second fraction was methyl-derivatized
with 2030 mg K2 CO3 , 150 L acetone and 40 L CH3 I under heating at hot block at (110 5) C for 30 min and then it was submitted
to GCMS analysis. The acquired data les were also reprocessed
by AMDIS and DRS algorithms. The remaining third fraction was
mixed with the extract from procedure B.
Regarding procedure B, a second aliquot of 2.5 mL horse urine
was fortied with androsterone sulfate (0.5 g) and methyltestosterone (0.25 g) as the IS. The urine was then diluted with 2.0 mL
acetate buffer 1 M and the pH was adjusted to 4.85.5. Solid phase
extraction (SPE) was performed on C-18 cartridges pre-conditioned
with 5.0 mL methanol and 5.0 mL acetate buffer (1 M, pH 5.2).
After loading the urine, the cartridge was washed with 5.0 mL
acetate buffer (1 M, pH 5.2) followed by 5.0 mL n-hexane and the
compounds of interest were eluted with 5.0 mL methanol. The eluate was evaporated under nitrogen at (60 5) C. Methanolysis
was accomplished by dissolving the residue in 1.0 mL anhydrous
methanolic HCl (3 M) and incubating at (60 5) C for 30 min. The
basic extraction was implemented with 4.0 mL diethylether after
pH adjustment at 9.510 with carbonate buffer (30%, w/v) and

72

M.K. Kioussi et al. / J. Chromatogr. B 941 (2013) 6980

addition of 700 L NaCl (4 M) for salting out. The organic extract

was mixed with the third fraction from procedure A and evaporated
to dryness under a nitrogen stream at (50 5) C. The residue was
dried in desiccator over P2 O5 overnight. TMS-derivatization was
performed by adding 150 L MSTFA:NH4 I:PrSH (1000:2:3, v/w/w)
and incubating at (110 5) C for 60 min. Per-TMS derivatives were
analyzed by GCHRMS. Additionally, the TMS-derivatized samples
were subjected to GCMS analysis and the acquired full-scan data
les were evaluated by AMDIS and DRS programs.
In both procedures, addition of CH3 COOH (3 M in ethylacetate)
solution preceded organic layer evaporation in order to prevent loss
of volatile compounds.
2.5. Method validation
The current screening method was validated according to
Eurachem guidelines for validation of qualitative chromatographic
methods [23]. Prior to validation, a tentative experimental day was
accomplished to two gelding horse urine samples fortied with the
reference compounds listed in Table 1 , at seven concentration levels (5, 10, 25, 50, 100, 250, 500 ng mL1 ). Afterwards, a minimum
concentration, Cmin , at which each compound was detected at both
urine samples, was determined.
2.5.1. Limit of detection method sensitivity
In order to determine the limit of detection (LOD) ten different blank equine urine substrates were fortied with the reference
compounds at six concentration levels (0.25 Cmin , 0.5 Cmin ,
1 Cmin , 2.5 Cmin , 5 Cmin , 10 Cmin ) and prepared along with
the respective blanks. The LOD was dened as the lowest concentration level at which a compound could be identied in all ten urine
samples with a signal-to-noise (S/N) ratio above 3 in the selected
ion chromatogram.
2.5.2. Matrix interference specicity
Ten different drug-free urine specimens from three mares, three
male and four gelding horses were prepared in order to check for
possible naturally occurring interfering substances at the retention
times of the examined analytes and at the selected extracted ion
masses.
2.5.3. Extraction recovery
Extraction recovery was investigated at three concentration levels (5 LOD, 10 LOD, 20 LOD) on three days by analyzing in
duplicate six different urine samples which had been spiked with
the target analytes either before or after sample extraction. Spiking
to the organic phase after LLE corresponded to 100% recovery. To
evaluate extraction recovery, the peak area response of the analyte
spiked initially before extraction into the urine sample was compared to the peak area of the same analyte spiked after extraction
into the blank urine extract.
2.5.4. Repeatability
The repeatability (intra-day and inter-day precision) of the
method was assessed by analyzing in duplicate six aliquots of
an equine specimen at three concentrations (5 LOD, 10 LOD,
20 LOD) on three different days over a two-month period. The
intra-day and inter-day precision were expressed as percentage
relative standard deviations (RSD) of the relative retention times
(RRTs) and the peak areas of the studied analytes.
2.5.5. Mass accuracy
Mass accuracy regarding only LCTOF-MS analysis, was evaluated on three days in six different urine substrates fortied at
10 LOD concentration.

2.5.6. Matrix effect

Matrix effect associated with LCMS analysis, was assessed at
three concentration levels (5 LOD, 10 LOD, 20 LOD) on three
days by analyzing in duplicate six different urine samples which
were rst extracted and then fortied with the studied compounds.
Matrix effect was expressed as the ratio of the peak response in the
presence of matrix to the peak response in the absence of matrix in
the neat reconstitution solvent [24].
2.5.7. Carry over contamination
Carry over contamination was examined in drug free horse urine
samples injected subsequently to fortied samples at high concentration level of 20 LOD in order to evaluate the presence of any
peak at the expected retention time. Regarding the evaluation of
carry over contamination, the peak area of the analyte detected at
the blank sample was expressed as a percentage of the peak area
of the same analyte detected at the previous sample.
2.6. Screening of 302 drugs in equine urine
The developed and validated method was further examined
for the identication of a large number of banned compounds
belonging to different therapeutic classes. The studied substances
presented in Supplementary Table 1 were fortied to blank equine
urine at concentrations of 100 or 200 ng mL1 . The fortied and the
blank horse urine samples were prepared and analyzed according
to Supplementary Figure 1.
2.7. Method applicability
In order to verify that the validated screening method was
effective to real positive samples, ten positive post-race horse
urine samples and ve urine specimens from the prociency
test of Association of Ofcial Racing Chemists (AORC) 2012
were analyzed. The samples had been previously analyzed with
established and validated methods and they had been reported
to contain amitriptyline, bumetanide, clanobutin, clenbuterol,
dexamethasone, unixin, furosemide, ketoprofen, mepivacaine,
modanil, morphine, nandrolone and its metabolite 5-estrane3,17-diol, phenylbutazone and its metabolite oxyphenbutazone,
procaine, phenobarbital, nordiazepam and oxazepam. The fteen
re-examined samples were prepared and analyzed along with
blank urine and quality control samples.
In addition to the above, two horse urine samples tested positive
with pentobarbital were also prepared and analyzed. The particular substance was not included among the target drugs of the
screening method and therefore the applicability of AMDIS-DRS for
automated evaluation of data les of real samples was examined,
as well.
3. Results and discussion
3.1. Method development
An extensive method development was performed in order to
establish a suitable sample pretreatment taking into account the
laboratory instrumentation. Thirty six compounds included in the
AORC list of prociency tests were used for the method development, in concentrations similar to the ones depicted in the list [25].
The selected compounds represent multi-class drugs with various molecular weights, polarities, acidicbasic properties, chemical
structures and biological effects. In brief, the method development
was performed by testing different deconjugation processes, different solvent extractions and mixtures of the nal extracts. The
examined deconjugation processes were (i) enzymatic hydrolysis

M.K. Kioussi et al. / J. Chromatogr. B 941 (2013) 6980

73

Table 1
Results of method validation for retention time (RT), limit of detection (LOD), extraction recovery (ER), repeatability on peak areas (intra-day and inter-day precision), mass
error and matrix effect (ME) with the corresponding mean values and the percentage relative standard deviations (RSD) in parenthesis. The subscripts 5, 10 and 20 correspond
to concentrations of 5 LOD, 10 LOD and 20 LOD, respectively.
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
27
54
55
56

Substance
LCTOF-MS analysis
1-(3-Chlorophenyl)
piperazine
16-OH-stanozolol
3-methoxytyraminee
Amitriptylined
Amphetamined
Anastrozole
Atenolold
Benzoylecgonined
Butorphanold
Caffeined
Capsaicin
Celecoxib
Chlorpromazined
Clenbuterold
Clonidine
Dexamethasoned
Diphenhydramine
Ephedrined
Ergonovine
Fentanyl
Flunixind
Fluphenazine
Hydrochlorothiazided
Hydrocortisonee
Imipramine
Isoxsuprine
Ketoprofend
Lidocained
Meloxicamd
Mepivacained
Methocarbamold
Methylphenidated
Morphined
Nikethamided
Nordiazepamd
Nortriptyline
Nylidrin
Oxazepamd
Pethidined
Phentermine
Pyrilamine
Ractopamine
Salbutamold
Terbutalined
Theobrominee
Timolold
Trenbolone
Triamcinolone acetonided
Triuoperazine
Verapamil
GCMS analysis
Diunisal
Flufenamic acidd
Furosemided
Ketoprofen
Phenobarbitald
Phenylbutazoned
GCHRMS analysis
5-estrane-3,17-diold , e
d

RTa (min)

LOD (ng mL1 )

% ER5 (RSD)b

Therapeutic activity

Molecular formula

Adduct/derivative

Target ion

Psychoactive

C10 H13 N2 Cl

[M+H]+ isotope

199.0812

4.90

62.5

72 (13)

Anabolic steroid
Neurotransmitter
Tricyclic antidepressant
Stimulant
Aromatase inhibitor
Beta blocker
Stimulant
Opiod
Cardiac-respiratory
stimulant; diuretic
Topical analgesic
NSAID
Sedative
Bronchodilator
Antihypertensive
Corticosteroid
Antihistamine
Stimulant
Vasoconstrictor
Opiod analgesic
NSAID
Antipsychotic
Diuretic
Corticosteroid
Antidepressant
Vasodilator
NSAID
Local anesthetic
NSAID
Local anesthetic
Muscle relaxant
Stimulant
Opiod analgesic
Stimulant
Tranquilizer
Antidepressant
Sympathomimetic
Anxiolytic
Opiod analgesic
Anticonvulsant
Antihistamine
Beta agonist
Bronchodilator
Bronchodilator
Vasodilator
Beta blocker
Anabolic steroid
Corticosteroid
Antipsychotic
Anti-arrhythmic

C21 H32 N2 O2
C9 H13 NO2
C20 H23 N
C9 H13 N
C17 H19 N5
C14 H22 N2 O3
C16 H19 NO4
C21 H29 NO2
C8 H10 N4 O2

[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+

345.2537
168.1019
278.1903
136.1121
294.1713
267.1703
290.1387
328.2271
195.0877

7.41
1.43
7.48
3.55
7.37
2.05
4.08
5.58
3.39

12.5
4000
1.25
1.25
2.5
2.5
25
1.25
62.5

64 (8.8)
26 (19)
57 (6.9)
55 (6.1)
77 (0.9)
38 (38)
13 (53)
68 (6.7)
81 (13)

C18 H27 NO3

C17 H14 F3 N3 O2 S
C17 H19 ClN2 S
C12 H18 Cl2 N2 O
C9 H9 Cl2 N3
C22 H29 FO5
C17 H21 NO
C10 H15 NO
C19 H23 N3 O2
C22 H28 N2 O
C14 H11 F3 N2 O2
C22 H26 F3 N3 OS
C7 H8 ClN3 O4 S2
C21 H30 O5
C19 H24 N2
C18 H23 NO3
C16 H14 O3
C14 H22 N2 O
C14 H13 N3 O4 S2
C15 H22 N2 O
C11 H15 NO5
C14 H19 NO2
C17 H19 NO3
C10 H14 N2 O
C15 H11 ClN2 O
C19 H21 N
C19 H25 NO2
C15 H11 ClN2 O2
C15 H21 NO2
C10 H15 N
C17 H23 N3 O
C18 H24 NO3
C13 H21 NO3
C12 H19 NO3
C7 H8 N4 O2
C13 H24 N4 O3 S
C18 H22 O2
C24 H31 FO6
C21 H24 F3 N3 S
C27 H38 N2 O4

[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+NH4 ]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+
[M+H]+

306.2064
382.0832
319.1030
277.0869
230.0246
393.2072
256.1696
166.1226
326.1863
337.2274
297.0845
438.1821
314.9978
363.2166
281.2012
302.1751
255.1016
235.1805
352.0420
247.1805
242.1023
234.1489
286.1438
179.1179
271.0633
264.1747
300.1958
287.0582
248.1645
150.1277
286.1914
302.1751
240.1594
226.1438
181.0720
317.1642
271.1693
435.2177
408.1716
455.2904

9.46
10.37
7.81
4.82
3.26
7.19
6.48
3.00
3.65
6.39
8.27
8.27
3.51
6.43
7.27
5.40
8.20
4.31
7.51
4.28
5.06
4.87
1.54
4.43
8.22
7.32
5.85
7.33
5.30
4.01
6.22
4.25
1.86
1.87
1.75
4.56
7.63
7.57
8.72
7.46

1.25
12.5
5.0
2.5
1.25
12.5
1.25
25
12.5
1.25
5.0
2.5
12.5
1000
1.25
1.25
25
1.25
1.25
1.25
5.0
1.25
2.5
2.5
1.25
12.5
12.5
2.5
1.25
2.5
1.25
5
25
10
2000
1.25
12.5
5.0
5.0
1.25

61 (4.6)
51 (6.7)
48 (9.2)
73 (4.4)
79 (9.6)
74 (5.9)
69 (5.8)
57 (8.7)
73 (7.4)
65 (0.8)
74 (5.1)
53 (9.9)
68 (6.9)
75 (5.7)
62 (9.6)
77 (2.1)
50 (25)
72 (9.4)
63 (13)
76 (6.8)
78 (7.4)
63 (5.8)
88 (0.6)
68 (9.0)
64 (5.8)
62 (16)
81 (0.6)
71 (7.2)
72 (5.6)
74 (4.3)
69 (9.0)
86 (3.0)
37 (28)
41 (27)
46 (13)
77 (6.1)
71 (8.2)
70 (8.0)
46 (30)
54 (3.9)

Analgesic
NSAID
Diuretic
NSAID
Sedative
NSAID

C13 H8 F2 O3
C14 H10 F3 NO2
C12 H11 ClN2 O5 S
C16 H14 O3
C12 H12 N2 O3
C19 H20 N2 O2

bis-Me
mono-Me
tris-Me
mono-Me
bis-Me
mono-Me

247; 278
263; 295
81; 372
209; 268
232; 117
322; 266

6.89
6.21
11.66
7.74
6.02
8.82

5.0
5.0
125
25
5.0
62.5

68 (11)
64 (5.7)
39 (15)
48 (28)
64 (14)
22 (7.0)

Anabolic steroid

C18 H30 O2

bis-TMS

332.2536;
407.2802
508.1884
315.0375;
317.0375
432.2880
237.0808

57
58

Bumetanide
Ethacrynic acid artifactd

Diuretic
Diuretic

C17 H20 N2 O5 S
C13 H12 Cl2 O4

bis-TMS
mono-TMS

59
60

Testosteronee
Theophylline

Anabolic steroid
Bronchodilator

C19 H28 O2
C7 H8 N4 O2

bis-TMS
mono-TMS

7.69

45

103 (6.3)

10.52
8.46

25
50

37 (6.4)
21 (20)

8.76
3.55

20
50

118 (11)
40 (17)

74

M.K. Kioussi et al. / J. Chromatogr. B 941 (2013) 6980

Table 1 (Continued )
No.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
27
54
55
56
57
58
59
60
a

% ER10
(RSD)b

% ER20
(RSD)b

LCTOF-MS analysis
74 (10.6)
82 (9.6)
57 (15)
70 (8.6)
54 (8.8)
58 (3.0)
71 (7.7)
35 (41)
14 (59)
64 (5.7)
68 (2.6)
58 (6.4)
48 (17)
47 (9.4)
71 (4.0)
71 (3.1)
70 (8.0)
66 (11)
60 (15)
67 (2.7)
57 (4.4)
71 (4.3)
45 (21)
68 (9.3)

64 (9.9)
74 (2.9)
86 (6.2)
44 (38)
18 (57)
79 (4.4)
84 (1.3)
72 (3.4)
65 (15)
59 (8.5)
89 (4.3)
83 (0.7)
87 (6.5)
73 (9.0)
69 (3.7)
76 (7.6)
68 (0.6)
90 (2.7)
51 (4.7)
83 (7.0)

59 (12)
70 (4.7)
52 (31)
68 (4.6)
61 (18)
71 (3.9)
73 (3.9)
62 (9.5)
64 (9.4)
61 (2.1)
63 (6.7)
60 (19)
77 (2.9)
72 (7.2)
69 (7.9)
75 (6.4)
66 (9.1)
79 (3.1)
43 (31)
54 (31)

71 (11)
83 (4.3)
63 (30)
81 (0.1)
80 (13)
89 (4.2)
92 (1.8)
73 (9.1)
79 (11)
82 (2.1)
80 (6.5)
68 (12)
88 (2.7)
89 (5.7)
80 (8.7)
84 (4.0)
78 (11)
92 (3.2)
54 (29)
59 (29)

72 (2.9)
89 (2.9)
69 (1.1)
82 (3.3)
66 (5.3)
83 (3.9)
37 (26)
40 (12)
51 (3.9)
61 (5.4)
GCMS analysis
61 (11)
75 (19)
64 (14)
79 (12)
36 (34)
44 (16)
51 (37)
59 (34)
68 (15)
82 (21)
20 (12)
27 (14)
GCHRMS analysis
34 (18)
20 (45)

35 (15)
19 (9.7)

40 (28)

50 (29)

Precision, % RSD

Intraday5 c
(interday5 )b

Intraday10 c
(interday10 )b

Intraday20 c
(interday20 )b

8.9 (13)
14 (11)
4.6 (1.9)
8.0 (5.2)
5.6 (20)
6.2 (9.4)
8.7 (8.6)
5.2 (24)
6.2 (8.3)
5.5 (9.0)
9.7 (3.0)
12 (19)
10 (12)
5.3 (6.0)
4.4 (5.3)
8.4 (18)
5.3 (8.7)
8.6 (4.4)
3.5 (18)
9.1 (7.1)
5.6 (3.4)
14 (21)
9.6 (5.0)
8.4 (3.7)
8.8 (6.8)
6.0 (4.9)
5.6 (8.0)
6.3 (11)
8.3 (14)
5.6 (23)
8.6 (16)
4.7 (5.7)
8.0 (3.0)
13 (15)
5.6 (10)
7.3 (23)
5.5 (5.6)
6.8 (20)
5.3 (10)
9.6 (41)
6.8 (6.2)
2.6 (28)
7.5 (6.4)
7.1 (21)
5.7 (8.7)
4.5 (2.7)
13 (5.6)
8.9 (9.9)
23 (35)
6.1 (8.9)

2.6 (12)
6.3 (6.8)

5.3 (13)
3.5 (16)

4.6 (3.4)
8.0 (19)
3.9 (17)
1.6 (10)
2.5 (16)
3.8 (3.4)
2.3 (6.1)
6.8 (7.2)
7.5 (10)
6.9 (8.7)
2.4 (6.5)
2.9 (10)
2.8 (15)
5.6 (5.8)
4.1 (8.9)
3.0 (17)
5.7 (4.2)
2.3 (2.3)
13 (8.6)
3.7 (5.8)

2.6 (8.9)
6.7 (23)
1.9 (21)
2.8 (15)
4.2 (12)
2.2 (9.2)
1.5 (11)
2.9 (12)
3.1 (15)
2.6 (18)
1.6 (11)
1.9 (16)
1.1 (15)
2.6 (12)
4.0 (12)
1.2 (14)
2.3 (11)
1.9 (7.7)
2.7 (14)
3.9 (10)

5.1 (13)
3.0 (9.3)
2.8 (8.1)
3.9 (3.8)
5.2 (1.5)
1.8 (16)
2.5 (4.4)
2.6 (7.8)
5.7 (6.5)
6.8 (8.6)
4.1 (6.6)
3.9 (23)
2.7 (9.9)
2.8 (27)
2.8 (9.6)
3.2 (36)
4.5 (11)
2.4 (19)
2.5 (7.2)
4.2 (25)

1.4 (16)
1.1 (10)
1.7 (18)
2.7 (4.1)
2.2 (8.7)
0.9 (7.1)
1.3 (7.5)
2.5 (13)
3.9 (11)
7.2 (2.7)
1.4 (9.8)
2.3 (24)
1.0 (11)
2.2 (34)
2.5 (15)
8.4 (25)
1.2 (14)
1.5 (9.7)
2.1 (7.6)
5.7 (22)

2.1 (0.9)
8.1 (16)
3.1 (15)
22 (22)
5.2 (5.8)

0.9 (5.2)
4.2 (19)
1.5 (19)
5.3 (27)
2.6 (7.8)

12 (61)
10 (26)
14 (39)
8.9 (27)
10 (32)
9.0 (27)

17 (51)
8.5 (26)
8.8 (32)
6.9 (25)
12 (30)
6.5 (27)

30 (35)
6.6 (27)
19 (31)
5.0 (24)
16 (22)
6.4 (27)

9.9 (37)
43 (39)

5.0 (49)
27 (39)

12 (13)

9.3 (18)

6.0 (46)
19 (31)
54 (58)
11 (18)
14 (14)

Mass
error10 ,
ppm

% ME5
(RSD) b

% ME10
(RSD) b

% ME20
(RSD) b

0.22
1.1
5.0
2.7
1.6
0.94
0.25
-0.5
1.2
2.2
1.2
0.07
2.4
-0.38
1.5
3.0
-2.8
2.5
0.53
1.2
2.9
1.6
1.4
0.4
1.0
0.94
2.0
-2.4
0.63
0.0
0.50
-1.2
1.5
1.2
0.64
0.23
0.39
0.79
0.16
-0.52
1.1
0.17
2.2
1.2
2.5
0.56
-0.10
1.7
1.3
1.5

51 (4.9)
21 (30)
76 (9.1)
55 (4.9)
47 (30)
39 (26)
84 (4.3)
35 (11)
69 (1.6)
44 (24)
53 (9.8)
45 (18)
55 (6.5)
61 (4.5)
69 (7.7)
47 (13)
52 (20)
78 (6.4)
61 (17)
65 (6.7)
35 (6.2)
55 (14)
61 (4.6)
73 (16)
48 (9.7)
63 (9.4)
60 (8.8)
62 (3.1)
42 (12)
64 (3.2)
55 (5.1)
63 (5.4)
50 (17)
49 (17)
53 (15)
43 (7.4)
72 (8.9)
159 (13)
63 (0.2)
18 (12)
89 (12)
38 (17)
85 (5.7)
77 (6.5)
58 (7.0)
62 (4.7)
42 (7.8)
43 (13)
56 (31)
59 (13)

58 (7.6)
25 (32)

60 (9.0)
28 (19)

57 (8.5)
69 (36)
42 (27)
75 (11)
39 (19)
68 (9.7)
46 (25)
50 (3.0)
42 (13)
82 (9.2)
61 (14)
71 (7.8)
45 (12)
58 (24)
86 (6.0)
60 (17)
63 (5.8)
37 (16)
100 (38)
68 (7.2)

58 (3.0)
52 (35)
46 (27)
79 (5.9)
46 (19)
71 (8.0)
51 (17)
51 (9.3)
38 (21)
136 (39)
68 (11)
72 (1.7)
47 (13)
53 (19)
87 (8.4)
68 (18)
63 (6.1)
40 (15)
108 (44)
76 (14)

52 (10)
58 (14)
64 (12)
72 (12)
45 (7.6)
68 (19)
58 (13)
76 (12)
35 (34)
77 (50)
62 (12)
48 (12)
76 (10)
166 (12)
72 (14)
24 (16)
90 (12)
43 (18)
87 (10)
94 (14)

53 (3.6)
61 (12)
69 (15)
66 (6.2)
49 (15)
68 (12)
63 (11)
67 (8.9)
35 (26)
57 (13)
69 (16)
52 (4.5)
78 (7.3)
165 (14)
60 (10)
19 (7.7)
83 (2.0)
46 (20)
97 (7.0)
111 (12)

61 (14)
42 (5.3)
40 (16)
115 (43)
58 (9.3)

66 (12)
48 (5.5)
42 (12)
107 (57)
59 (6.9)

n = 6 different horse urine matrices. For etiocholanolone (LCTOF-MS) RTIS = 9.98 min, for mefruside (GCMS) RTIS = 11.87 min and for methyltestosterone (GCHRMS)
RTIS = 9.28 min.
b
n = 3 different days.
c
n = 12; six aliquots of a horse urine analyzed in duplicate.
d
Substances used for the method development.
e
Threshold substances with threshold concentration presented in LOD column.

M.K. Kioussi et al. / J. Chromatogr. B 941 (2013) 6980

with -glucuronidase/arylsulfatase from H. pomatia (ii) methanolysis and (iii) basic hydrolysis with NaOH. The extraction solvents tested were (i) diethylether:ethylacetate:dichloromethane
(4:3:3.5) (ii) ethylacetate:dichloromethane (7:3) (iii) ethylacetate
and (iv) diethylether. Moreover, the extraction pH tested were 910
and 55.2.
The screening method was selected due to the greater number
of compounds detected and the greater mean extraction recovery obtained. The best results came from the sample pretreatment
which comprised enzymatic hydrolysis and LLE with ethylacetate
at pH 910. The 17-sulfate steroid conjugates were cleaved by
methanolysis after solid phase extraction and they were extracted
with diethylether at pH 9.510. A more detailed description is found
in Supplementary Figure 1.
3.2. Method validation
The developed method was validated for sixty prohibited compounds, listed in Table 1, which covered a large variety of drugs.
The studied substances were chosen on the basis of AORC list
of prociency test 2011 [25] and FEI prohibited substances list
[1] taking also into consideration the classication guidelines
for foreign substances of the Association of Racing Commissioners International [26]. In addition to the parent compounds,
metabolites of caffeine (theophylline), cocaine (benzoylecgonine),
stanozolol (16-hydroxystanozolol), amitriptyline (nortriptyline)
and diazepam (oxazepam, nordiazepam) were identied and
utilized as target compounds as well [27,28]. Moreover, terbutaline and ethacrynic acid artifacts caused by sample preparation
were detected. Terbutaline incorporated a methylene unit in the
presence of formaldehyde under acidic conditions during enzymatic hydrolysis, forming its tetrahydroisoquinoline analog [29].
In respect to ethacrynic acid, its artifact was generated during derivatization process. Furthermore, the selected compounds
for validation comprised ve threshold substances approved by
IFHA [2]. The threshold substances presented in Table 1, concern
endogenous compounds (5-estrane-3,17-diol in male horses,
testosterone, hydrocortisone, 3-methoxytyramine) or compounds
of dietary origin (theobromine) and therefore, they occur naturally
in horse urine. These substances were examined at the established
threshold concentration in the validation procedure.
The doping control method described was based on the combination of liquid and gas chromatography that covers probably
all small molecules for anti-doping interest. The method used two
technologies, LCTOF-MS and GCHRMS, which enabled accurate
mass determination at high resolving power allowing the target
analytes to be distinguished from the complex matrix. Besides,
LCTOF-MS and GCMS techniques offered acquisition in full scan
mode and thus the carrying out of retrospective analysis of the samples by reprocessing their initially acquired data les. Regarding
LCTOF-MS analysis, the detection criteria included retention time
of the monitoring ions (RT) and monoisotopic exact mass by applying mass tolerance 20 ppm. All the studied compounds in method
validation were ionized to their protonated molecules [M+H]+ in
positive ESI mode, except hydrochlorothiazide which was detected
as ammonium adduct [M+NH4 ]+ . Moreover, retention time and two
diagnostic ions, the molecular and a fragment ion, were used for the
identication of the methyl-derivatized substances. Furthermore,
the detection of TMS-derivatized compounds through GCHRMS
analysis was based both on retention time and selected ion recording of one or two diagnostic ions at 10,000 resolution.
The present method incorporated quality assurance measures
covering the deconjugation processes and the TMS-derivatization
step. Etiocholanolone-glucuronide was added in urine at procedure
A, prior to enzymatic hydrolysis, in order to evaluate the hydrolysis efciency. Moreover, the addition of androsterone-sulfate in

75

urine at procedure B allowed for the evaluation of methanolysis efciency. Besides, by monitoring the mono-TMS derivatized
etiocholanolone and androsterone ions at GCHRMS analysis, the
TMS-derivatization efciency could be estimated.
3.2.1. Limit of detection method sensitivity
Each compound of interest was identied by means of the technique in which the lower detection limit was achieved. When the
same detection limit was obtained by two techniques, LCTOFMS was the technique of choice (e.g. salbutamol and clenbuterol).
Regarding furosemide and phenylbutazone, methyl-derivatization
was preferred instead of their unstable per-TMS derivatives, while
for ethacrynic acid, the artifact formed during TMS derivatization
beneted GCHRMS analysis against methyl-derivatization and
GCMS analysis.
The estimated LOD values are given in Table 1. All fty ve target substances could be detected at 62.5 ng mL1 or lower, except
furosemide that presented an LOD at 125 ng mL1 . For thirty ve
analytes (64% of the examined compounds) the LOD ranged from
1.25 ng mL1 to 5.0 ng mL1 , whereas fourteen analytes (25% of the
studied compounds) showed LODs in the range 1025 ng mL1 .
The minimum detecting concentrations required by IFHA in Article
6 of the IABRW were met with the exception of benzoylecgonine (20 ng mL1 ), bumetanide (10 ng mL1 ) and dexamethasone
(2 ng mL1 ) [2].
3.2.2. Matrix interference selectivity
Ten negative post-race horse urine specimens were analyzed
and demonstrated that interferences from the different matrices
at the expected retention times of the target ions were negligible,
except ufenamic acid TMS derivative during GCHRMS analysis.
Thus, the particular analyte was detected by GCMS assay after
methyl-derivatization.
3.2.3. Extraction recovery
The extraction recovery data shown in Table 1 are the mean
extraction recoveries assessed on three experimental days (n = 3)
in six horse urine substrates per day. The mean calculated recovery
values ranged from 13% (benzoylecgonine) to 92% (ractopamine),
with overall average values (n = 55 analytes) of 62%, 59% and 71% at
the concentrations of 5 LOD, 10 LOD and 20 LOD, respectively.
Regarding threshold substances, the obtained mean extraction
recoveries varied between 26% (3-methoxytyramine) and 118%
(testosterone). At least forty seven of the sixty compounds studied
(78%) demonstrated mean recoveries above 50% at all three concentration levels. The hydrophilic and zwitterion benzoylecgonine
(pKa1 = 2.3, pKa2 = 11.2) as well as the acidic compounds phenylbutazone (pKa = 4.5) and ethacrynic acid (pKa = 3.5) were extracted
with recoveries around or below 20%. The poor extraction recovery of benzoylecgonine could be attributed to the degradation of its
ester structure under basic conditions during sample preparation
[30].
The calculated recovery values suggested recoverys dependence upon analyte concentration. Most of the examined compounds showed an increase of 10% in recovery at the highest
concentration, 20 LOD, comparatively to the lower concentrations 5 LOD and 10 LOD. Besides, the polar compounds
triuoperazine, atenolol, salbutamol, terbutaline, ketoprofen and
benzoylecgonine demonstrated variation in extraction between
days because their extraction is highly affected by slight alterations
in the pH during extraction. The particular validation parameter
was examined on six different urine substrates and the results
(not shown here) proved recoverys independence of the urine
substrate, except for individual analytes such as uphenazine, triuoperazine, bumetanide, theophylline, which exhibited RSD > 20%
for at least two experimental days. Moreover, recoverys RSD for

76

M.K. Kioussi et al. / J. Chromatogr. B 941 (2013) 6980

Fig. 1. Accurate mass extracted ion chromatograms obtained from LCTOF-MS analysis of (i) blank horse urine sample; rst column namely BLANK, (ii) quality control
horse urine sample fortied with amitriptyline at 500 ng mL1 , clanobutin at 200 ng mL1 , clenbuterol at 4 ng mL1 , dexamethasone at 50 ng mL1 , unixin at 200 ng mL1 ,
ketoprofen at 200 ng mL1 , mepivacaine at 50 ng mL1 , modanil at 500 ng mL1 , morphine at 1 g mL1 , nordiazepam at 100 ng mL1 , oxazepam at 100 ng mL1 and procaine
at 500 ng mL1 ; second column namely QUAL. CTRL, (iii) authentic horse urine samples; third column namely SAMPLE.

substances analyzed by GCHRMS and GCMS was higher than the

RSD of those analyzed by LCTOF-MS because of the additional step
of derivatization required for GC analysis.
3.2.4. Repeatability
Regarding the GCMS analysis of the methylated compounds,
the intra-day and inter-day precision of the RRTs were less than
0.08%. With respect to GCHRMS analysis of the TMS-derivatized
substances, the intra-day precision of the RRTs ranged from
0.03 to 0.16% whereas the respective inter-day precision ranged

between 0.03 and 0.36%. The RRTs for the selected compounds
analyzed by LCTOF-MS were also stable within day, as their RSD
values did not exceed 1%. Moreover, inter-day precision lower
than 1% was observed for forty one of the analytes detected by
LCTOF-MS, exclusive of salbutamol, terbutaline, hydrochlorothiazide, amphetamine, atenolol, ephedrine, morphine, clonidine,
ergonovine, 3-methoxytyramine, theobromine at overall concentration levels. The variation of RRTs for the latter analytes can be
attributed to their early elution in the beginning of the gradient or
to their peak shape e.g. asymmetrical and wide peaks.

M.K. Kioussi et al. / J. Chromatogr. B 941 (2013) 6980

77

Fig. 2. Extracted ion chromatograms of methylated derivatives (Me) obtained from GCMS analysis of (i) blank horse urine sample; rst column namely BLANK, (ii)
quality control horse urine sample fortied with furosemide at 500 ng mL1 , phenobarbital at 100 ng mL1 , phenylbutazone at 500 ng mL1 , oxyphenbutazone at 2 g mL1 ,
ketoprofen at 200 ng mL1 ; second column namely QUAL. CTRL, and (iii) authentic horse urine samples; third column namely SAMPLE.

The results of the precision (of peak areas stated in terms of RSD)
are summarized in Table 1. The intra-day precision for the compounds analyzed by LCTOF-MS was in the range of 0.9% (timolol)
to 23% (triuoperazine), with mean RSD values of 7.8%, 4.6% and
2.7% at the concentrations of 5 LOD, 10 LOD, 20 LOD, respectively. The inter-day precision varied from 0.9% (timolol) to 41%
(phentermine), with mean RSD values of 12%, 11% and 14% at the
concentrations of 5 LOD, 10 LOD, 20 LOD, respectively. The
poorest within-days repeatability was obtained for phentermine
because of the double chromatographic peak observed in the rst
and the second experimental day. For at least 81% of the compounds

analyzed by LCTOF-MS the inter-day precision was lower than 20%

at the examined concentrations.
For GCMS analysis, the intra-day precision of the methylated
substances ranged between 5.0% (ketoprofen) and 30% (diunisal)
and the inter-day precision was around 2461%. The highest variation was observed in diunisal due to the formation of mono- and
bis-methylated diunisal during the derivatization step.
In respect of GCHRMS analysis, the intra-day precision of
the trimethylsilylated target compounds was less than 25% at
the examined concentration levels except for the ethacrynic acid
artifact due to its non-constant formation during the sample

78

M.K. Kioussi et al. / J. Chromatogr. B 941 (2013) 6980

Fig. 3. Accurate mass extracted ion chromatograms of trimethylsilylated derivatives (TMS) obtained from GCHRMS analysis of i) blank horse urine sample; rst column
namely BLANK, (ii) quality control horse urine sample fortied with bumetanide at 20 ng mL1 , nandrolone at 25 ng mL1 and 5-estrane-3,17-diol at 50 ng mL1 ; second
column namely QUAL. CTRL, and (iii) authentic horse urine samples; third column namely SAMPLE.

preparation procedure. The inter-day precision ranged from 13%

(theophylline) to 58% (ethacrynic acid artifact).
Overall, the precision results derived from LCTOF-MS analysis were better than those obtained from GCMS and GCHRMS
analyses because of the absence of the derivatization step.
3.2.5. Mass accuracy
Mass accuracy was evaluated at a medium concentration level
of 10 LOD, at which most of the ion abundances were in the
range 1 104 5 105 . Bristow et al. [31] observed a clear deterioration of mass accuracy at both very low (<1 103 ) and very
high (>1 106 ) ion abundances. The mass error values presented
in Table 1 are the mean mass errors calculated on three days in a
period of four months, using lock mass correction. The mean mass
measurement errors ranged from 2.8 ppm (diphenhydramine) to
5.0 ppm (3-methoxytyramine) with an overall average value (n = 60
substances) of 1.0 ppm. 3-Methoxytyramine is a small molecule
(MW = 167) eluting in the beginning of the gradient (RT = 1.43 min)
along with many matrix compounds, resulting often in broad and
less symmetrical peak shapes which shifted the centroid mass to a
higher m/z value.
3.2.6. Matrix effect
The matrix effect results reported in Table 1 are the mean values
from the three experimental days with the corresponding relative standard deviations. A matrix effect value of 100% indicates

that the response in the mobile phase and the urine extract are
the same and thus no matrix effect is observed. A matrix effect
value greater than 100% indicates ionization enhancement and a
value smaller than 100% indicates ionization suppression. Signal
suppression, which is more common than signal enhancement in
LC-ESIMS, was the effect observed for almost all the analytes.
The highest ion suppression, about 20%, occurred in phentermine
(MW = 149), whereas the highest ion enhancement, about 160%,
was caused for oxazepam at all three concentration levels. Despite
the high ion suppression observed, phentermine could still be
detected at 2.5 ng mL1 . It is known that smaller and more polar
analytes are more susceptible to undergo ion suppression [32,33].
Regarding the antipsychotic compounds uphenazine and triuoperazine, the kind of matrix effect was not constant since both signal
suppression and signal enhancement were observed at different
concentrations.
The relative standard deviations of the matrix effect values were also calculated in order to evaluate the variation
between six different horse urine matrices (data not shown). The
particular RSD values demonstrated the dependence of matrix
effect upon substrate and especially for triamcinolone acetonide,
amphetamine, chlorpromazine, unixin, morphine, celecoxib,
trenbolone, uphenazine, triuoperazine, nortriptyline, phentermine, ergonovine and 16-hydroxystanozolol the matrix effect
varied considerably between different urine samples. Furthermore, the particular validation parameter was examined in three

M.K. Kioussi et al. / J. Chromatogr. B 941 (2013) 6980

concentration levels, by keeping constant the matrix amount.

For some substances, such as nortriptyline, hydrochlorothiazide,
nordiazepam, the extent of signal suppression was inversely
proportional to their concentration; the higher the concentration
of the target analyte, the lower the level of suppression. However,
there were substances, such as verapamil, dexamethasone, fentanyl, in which the matrix effect was constant and independent
of the concentration of the analyte.

3.2.7. Carry over contamination

Carry over contamination was found for furosemide (2%) and
bumetanide (1%) which were identied by GCMS and GCHRMS
respectively. Both analytes detected at the blank sample originated
from the former fortied sample at concentration of 20 LOD. No
carry over was detected for the rest of the compounds.

3.3. Identication of 302 drugs in equine urine

The hereby proposed method aimed at the determination of
a wide range of prohibited compounds in equine urine. Supplementary Table 1 shows the 302 analytes related to equine doping
control that were covered by the current screening method. The
majority of the examined compounds could be detected through
the LCTOF-MS analysis in full scan acquisition mode with high
mass accuracy. Some NSAIDs and carbonic anhydrase inhibitors
were detected through methyl-derivatization and GCMS analysis, while anabolic steroids were identied by LCTOF-MS [3440]
and GCHRMS assays. The studied analytes comprised parent compounds, phase 1 metabolites, degradation products and artifacts
[41].

3.4. Method applicability

Seventeen previously reported positive horse urine samples
were re-examined by applying the current screening method.
Figs. 13 present the extracted-ion-chromatograms (EIC) of the target analytes in the blank urine, the quality control sample and the
positive samples after LCTOF-MS, GCMS and GCHRMS analysis, respectively. The EIC of the blank urine did not show any peak
at the expected retention time. All previously reported prohibited substances were detected by the present method with good
S/N ratios and with retention times which did not vary from that
in the reference sample by more than 1.6% (morphine, Fig. 1).
There were compounds that could be identied by more than one
analysis (chromatograms not shown), such as bumetanide which
was detected by all the three analyses (LCTOF-MS, GCMS and
GCHRMS). Furosemide and clenbuterol were also detected after
TMS-derivatization and GCHRMS analysis, while phenylbutazone
and oxyphenbutazone could be identied by LCTOF-MS besides
GCMS analysis of their methylated derivatives. Regarding the
positive samples in nandrolone and phenylbutazone, the major urinary metabolites 5-estrane-3,17-diol and oxyphenbutazone
were clearly detected in addition to the corresponding parent analytes. Nandrolone and its metabolite 5-estrane-3,17-diol were
found in urine from female horse and therefore, doping control
was not based on threshold concentration. Moreover, the detection of oxazepam and nordiazepam indicated administration of
diazepam.
Furthermore, the applicability of the library searching using
AMDIS-DRS software was demonstrated by the detection of pentobarbital in two urine samples after GCMS analysis of the
methyl-derivatized samples. Pentobarbital was not detected by
the targeted screening method, since it was not included in the
examined analytes. The application of library searching using

79

AMDIS-DRS software opens the window of the detectable compounds.

4. Conclusions
A unied, sensitive and suitable method was developed for the
qualitative identication of a wide range of prohibited substances
in equine urine for doping control purposes. The phase 2 metabolites were cleaved through enzymatic hydrolysis and methanolysis,
while the solid phase and liquid-liquid extractions were implemented for the extraction of the compounds of interest from the
complex matrix. The analysis of the prepared urine samples was
performed in a complementary way by LCTOF-MS, GCHRMS and
GCMS techniques. Thus the substances with difculty in positive
ESI ionization could be detected by GC, whereas the thermo labile,
the nonvolatile molecules and those with marginal GC properties
even after derivatization, could be identied by LC. The method was
validated for sixty selected compounds and in addition it facilitated
a broad screening of 302 additional banned substances. Moreover,
the validated method was supported by automated identication of analytes through library searching by using the common
AMDIS-DRS program, which enabled the detection of banned
drugs not targeted by the screening method. Both approaches,
targeted and non-targeted doping control, were successfully
applied to previously reported positive authentic horse urine
samples.
Overall, this is a generic and robust racing animal protocol that
provides MS library searching, retrospective evaluation of analyzed
samples and easy incorporation of new released drugs.
Acknowledgements
The authors wish to thank Agilent Technologies for providing
AMDIS-DRS programs. S. Loui, F. Hlapana and I. Koutsouli are gratefully acknowledged for their technical assistance as well as P. Kiousi
and R. Fragkaki for their support in LCTOF-MS analysis.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/
j.jchromb.2013.10.008.
References
[1] Federation Equestre Internationale (FEI), Equine Prohibited SubList,
2013,
Available
at:
http://www.feicleansport.org/
stances
2013%20Equine%20Prohibited%20List.pdf (accessed on June 2013).
[2] International Federation of Horseracing Authorities, International Agreement on Breeding, Racing and Wagering, 2012 March, Available at:
http://www.horseracingintfed.com/resources/2012 choose eng.pdf (accessed
on June 2013).
[3] M. Thevis, W. Schnzer, Mass Spectrom. Rev. 26 (2007) 79.
[4] O.J. Pozo, J. Marcos, J. Segura, R. Ventura, Bioanalysis 4 (2) (2012) 197.
[5] I. Ojanper, M. Kolmonen, A. Pelander, Anal. Bioanal. Chem. 403 (2012) 1203.
[6] M. Yamada, S. Aramaki, M. Kurosawa, I. Kijima-Suda, K. Saito, H. Nakazawa,
Anal. Sci. 24 (2008) 1199.
[7] M.E. Touber, M.C. van Engelen, C. Georgakopoulos, J.A. van Rhijn, M.W.F. Nielen,
Anal. Chim. Acta 586 (2007) 137.
[8] C.H.F. Wong, D.K.K. Leung, F.P.W. Tang, J.K.Y. Wong, N.H. Yu, T.S.M. Wan, J.
Chromatogr. A 1232 (2012) 257.
[9] Y. Liu, C.E. Uboh, L.R. Soma, X. Li, F. Guan, Y. You, J.-W. Chen, Anal. Chem. 83
(2011) 6834.
[10] P. Garcia, A.-C. Paris, J. Gil, M.-A. Popot, Y. Bonnaire, Biomed. Chromatogr. 25
(2011) 147.
[11] R.J.B. Peters, A.A.M. Stolker, J.G.J. Mol, A. Lommen, E. Lyris, Y. Angelis, A. Vonaparti, M. Stamou, C.G. Georgakopoulos, M.W.F. Nielen, Trends Anal. Chem. 29
(11) (2010) 1250.
[12] A. Kaufmann, P. Butcher, K. Maden, M. Widmer, Anal. Chim. Acta 586 (2007)
13.
[13] R.G. Howitt, M. Pelzing, G.D. Beresford, Routine drug screening by accurate mass
using liquid chromatography/time of ight mass spectrometry, in: E. Houghton,

80

[14]

[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]

[24]
[25]

[26]

M.K. Kioussi et al. / J. Chromatogr. B 941 (2013) 6980

I. Kijima-Suda, R. Wada, J.F. Wade (Eds.), Proceedings of the 16th International Conference of Racing Analysts and Veterinarians, R & W Publications,
Newmarket, 2006, p. 222.
R.G. Howitt, G.D. Beresford, M. Pelzing, I. Krebs, Routine drug screening by accurate mass using liquid chromatography/time of ight mass spectrometry Part
2, in: E. Houghton, F. Ipek Keskin, J.F. Wade, N. Yazicioglu (Eds.), Proceedings
of the 17th International Conference of Racing Analysts and Veterinarians, R &
W Publications, Newmarket, 2008, p. 191.
M.R. Meyer, F.T. Peters, H.H. Maurer, Clin. Chem. 56 (4) (2010) 575.
M.R. Meyer, P. Du, F. Schuster, H.H. Maurer, J. Mass Spectrom. 45 (12) (2010)
1426.
G.N.W. Leung, E.W. Chung, E.N.M. Ho, W.H. Kwok, D.K.K. Leung, F.P.W. Tang,
T.S.M. Wan, N.H. Yu, J. Chromatogr. B 825 (2005) 47.
E.N.M. Ho, D.K.K. Leung, T.S.M. Wan, N.H. Yu, J. Chromatogr. A 1120 (2006) 38.
C.H.F. Wong, F.P.W. Tang, T.S.M. Wan, Anal. Chim. Acta 697 (2011) 48.
Y. Moulard, L. Bailly-Chouriberry, S. Boyer, P. Garcia, M.-A. Popot, Y. Bonnaire,
Anal. Chim. Acta 700 (2011) 126.
E.N.M. Ho, W.H. Kwok, A.S.Y. Wong, T.S.M. Wan, Drug Test. Anal. 2012 (5) (2013)
509.
NIST, available at: http://www.sisweb.com/software/ms/nist.htm (accessed on
June 2013).
Eurachem Guide, The Fitness for Purpose of Analytical Methods: A Laboratory
Guide to Method Validation and Related Topics, Eurachem Teddington, 1998,
Available at: http://www.eurachem.org/images/stories/Guides/pdf/valid.pdf
(accessed on June 2013).
B.K. Matuszewski, M.L. Constanzer, C.M. Chavez, Anal. Chem. 75 (2003) 3019.
AORC Prociency Testing Drug List 2011. Available at: http://www.aorc-online.
org/documents/aorc-prociency-testing-drug-list-2011 (accessed on June
2013).
Association of Racing Commissioners International, Inc. (ARCI), Uniform Classication Guidelines for Foreign Substances and Recommended Penalties and
Model Rule, 2013, Available at: http://www.arci.com/druglisting.pdf (accessed
on June 2013).

[27] J.P. Scarth, P. Teale, T. Kuuranne, Drug Test. Anal. 3 (1) (2011) 19.
[28] A.R. McKinney, C.J. Suann, A.J. Dunstan, S.L. Mulley, D.D. Ridley, A.M. Stenhouse,
J. Chromatogr. B 811 (2004) 75.
[29] M. Thevis, G. Opfermann, W. Schanzer, J. Mass Spectrom. 38 (2003) 1197.
[30] M. Kolmonen, A. Leinonen, T. Kuuranne, A. Pelander, I. Ojanper, Drug Test.
Anal. 1 (2009) 250.
[31] T. Bristow, J. Constantine, M. Harrison, F. Cavoit, Rapid Commun. Mass Spectrom. 22 (2008) 1213.
[32] R. Bonglio, R.C. King, T.V. Olah, K. Merkle, Rapid Commun. Mass Spectrom. 13
(1999) 1175.
[33] A. Kloepfer, J. Quintana, T. Reemtsma, J. Chromatogr. A 1067 (2005) 153.
[34] H.-W. Hagedorn, R. Schulz, A. Friedrich, J. Chromatogr. Biomed. Appl. 577 (1992)
195.
[35] M. Yamada, S. Aramaki, M. Kurosawa, K. Saito, H. Nakazawa, J. Anal. Toxicol. 32
(2008) 387.
[36] M. Yamada, S. Aramaki, T. Hosoe, M. Kurosawa, I. Kijima-Suda, K. Saito, H.
Nakazawa, Anal. Sci. 24 (2008) 911.
[37] J. Fox, C. Dumasia, S. Stanley, E. Houghton, The in vivo biotransformation of
normethandrolone in the horse, in: R.B. Williams, E. Houghton, J.F. Wade (Eds.),
Proceedings of the 13th International Conference of Racing Analysts and Veterinarians, R & W Publications, Newmarket, 2000, p. 479.
[38] E. Houghton, S. Maynard, Some aspects of doping and medication control
in equine sports, in: D. Thieme, P. Hemmersbach (Eds.), Doping in Sports:
Biochemical Principles, Affects and Analysis, Handbook of Experimental Pharmacology, vol. 195, Springer, Heidelberg, 2010, p. 369.
[39] G.N.M. Leung, F.P.W. Tang, T.S.M. Wan, C.H.F. Wong, K.K.H. Lam, B.D. Stewart,
Biomed. Chromatogr. 24 (2010) 744.
[40] M. Machnik, M. Thevis, C. Von Kuk, S. Guddat, W. Schnzer, Gestrinone analysis in equine urine by LCMS/MS, in: W. Schnzer, H. Geyer, A. Gotzmann, U.
Mareck (Eds.), Recent Advances in Doping Analysis (12), Sport and Buch Strau,
Cologne, 2004, p. 149.
[41] K. Deventer, O.J. Pozo, P. Van Eenoo, F.T. Delbeke, J. Chromatogr. A 1216 (2009)
2466.