Sb

CEa

b

a

ARRAA

KCRBGLsP

1

cseotaebctlat

0d

Steroids 76 (2011) 616–625

Contents lists available at ScienceDirect

Steroids

journa l homepage: www.e lsev ier .com/ locate /s tero ids

tudies on the presence of natural and synthetic corticosteroids inovine urine

arolina Ferranti a,∗ , Fernanda delli Quadria , Luca Palleschia , Camilla Marchiafavaa , Marzia Pezzolatob ,lena Bozzettab, Maria Caramelli b, Rosa Draisci a

Istituto Superiore di Sanità, Viale Regina Elena 299, 00161 Roma, ItalyIstituto Zooprofilattico sperimentale del Piemonte, Liguria e Valle d’Aosta, Via Bologna 148, 10154 Torino, Italy

r t i c l e i n f o

rticle history:eceived 14 October 2010eceived in revised form 25 February 2011ccepted 26 February 2011vailable online 4 March 2011

eywords:orticosteroidsesiduesovine urinerowth-promotersiquid chromatography–tandem masspectrometry

a b s t r a c t

Natural and synthetic corticosteroids are widely used in veterinary medicine for their anti-inflammatoryproperties, but are also illegally used in animal breeding as growth-promoting agents: this latter appli-cation in livestock production has been banned within the European Union due to health concerns forthe consumer.

In this work urine samples collected from bovines experimentally treated with dexamethasone (0.4 mgof dexamethasone 21-disodium phosphate per capita/day for 20 consecutive days) and bovines bredunder strictly controlled conditions were investigated for the presence of natural and synthetic corti-costeroids, using a simple multi-residue liquid chromatography–tandem mass spectrometry method,developed and validated in accordance with the criteria of the Commission Decision 2002/657/EC.

The aim of this work is to investigate the effect of a low dosage and long term dexamethasone treatmenton the levels of endogenous corticosteroids in cattle and to evaluate the possible presence of pred-

rednisolonenisolone residues in bovines bred under strictly controlled conditions. Our findings confirm the high andrapid rate of dexamethasone urinary excretion. Dexamethasone treatment elicited an early reductionof hydrocortisone and cortisone, suggesting the disappearance of these two hormones as an indirectindicator of corticosteroid treatment in cattle. Prednisolone residues were found (concentration inter-val 0.4–1.4 ng mL−1) in urine samples collected from control bovines especially at the slaughterhouse,together with high levels of hydrocortisone and cortisone. Further studies are necessary to find out the

cretio

reason of unexplained ex

. Introduction

Natural corticosteroids are hormones secreted by the adrenalortex that are involved in a wide range of physiological processes,uch as stress response, inflammation, immune function, hydro-lectrolyte balance, reproduction and behavior [1,2]. The discoveryf their anti-inflammatory properties has led to the chemical syn-hesis of more active artificial corticosteroids, e.g. dexamethasonend prednisolone, that are used as therapeutic drugs. Dexam-thasone is a fluorinated hydrocortisone derivative, characterizedy an increased glucocorticoid potency associated with a nearlyomplete loss of mineralcorticoid activity; prednisolone is a struc-

ural isomer of cortisone, which exerts its pharmacological activityonger than cortisone but less than that of the longer-acting dex-methasone [3,4]. In veterinary medicine, the legal utilization ofhese compounds is strictly regulated, with withdrawal periods

∗ Corresponding author. Tel.: +39 06 49903150; fax: +39 06 49903079.E-mail address: carolina.ferranti@iss.it (C. Ferranti).

039-128X/$ – see front matter © 2011 Elsevier Inc. All rights reserved.oi:10.1016/j.steroids.2011.02.044

n of this hormone in urine samples of untreated bovines.© 2011 Elsevier Inc. All rights reserved.

between treatment and slaughtering and maximum residue lim-its (MRLs) established in edible biological matrices and milk forsome compounds [5]: MRLs have only been established for dexam-ethasone and betamethasone (2 �g kg−1 for liver, 0.75 �g kg−1 formuscle, 0.75 �g kg−1 for kidney and 0.3 �g kg−1 for milk), methyl-prednisolone (10 �g kg−1 for muscle, fat, liver and kidney) andprednisolone (4 �g kg−1 for muscle and fat, 6 �g kg−1 for milk and10 �g kg−1 for liver and kidney). On the other hand corticosteroids,especially at low concentration, are known to increase weightgain, to reduce the feed conversion ratio, and to have a synergeticeffect with other molecules like �-agonists or anabolic steroids[6,7]. Thus, these compounds have been used as growth promot-ers in cattle and their strong pharmacological activity makes theresidues of these molecules potentially dangerous for meat con-sumers. As a consequence, the administration of such drugs for

growth-promoting purposes is forbidden in Europe by the Coun-cil Directive 2003/74/EC [8]. National Surveillance Plans for steroidabuse adopted the detection of the parent compound in targetbiological matrices such as liver, kidney, muscle at slaughter-house, hair or urine at farms. In spite of the progress in analytical

eroids

tiawa

atatu[pdefrtphpbidta

kftpaoo

ccb

fCnttccseitt

deiriTupscsnaD

2.5. Sample extraction

C. Ferranti et al. / St

echniques, which allows to obtain very low detection limits forndividual residues, this strategy may be unsuccessful whenever

new compound is submitted to extensive biotransformation orhenever different steroids are administered at very low dosage

lone or in combination.Therefore, there is a requirement for sensitive multi-residue

nalytical methods for the quantification and confirmation ofhese compounds in biological samples. Over the years manynalytical methods have been developed to determine corticos-eroids in biological samples. Corticosteroids have been analysedsing either gas chromatography/mass spectrometry (GC–MS)9,10] or high performance liquid chromatography (HPLC) cou-led to UV or fluorescence detection [11]. Mass spectrometryetection is required for a method to be deemed confirmatory,arlier publications focused on GC–MS, GC–MS/MS techniquesor detection, these had advantages in that they offered goodesolution and sensitivity. However the drawback was that aime-consuming derivatization step was required for all these com-ounds, due to the poor thermal stability and volatility of steroidormones. More recently LC coupled to electrospray/atmosphericressure chemical ionization (ESI/APCI) mass spectrometry haseen used to determine corticosteroids quantitatively in bodily flu-

ds [12–24]. LC–MS/MS methods have become very widespreadue to their high selectivity, specificity and sensitivity and dueo the advantage of avoiding the derivatization step prior tonalysis.

Despite the extensive use of synthetic corticosteroids, little isnown about dexamethasone biotransformation in cattle and onlyew data are available about the kinetics of dexamethasone excre-ion in bovine urine; furthermore the major part of publishedapers deals with therapeutic dosages, administered either to onenimal or to a small number of individuals by the intravenous [25]r the oral route [26], to veal calves [27] or in combination withther drugs [28], respectively.

Only few studies have been devoted to investigate theorrelation between endogenous and administered syntheticorticosteroids (dexamethasone) in order to identify indirectiomarkers of illicit corticosteroid treatments [29–31].

In 2008, prednisolone, a synthetic corticosteroid, has beenound in some bovine urine samples within the Italian Residueontrol Plan. Moreover a recent paper [12] reported that pred-isolone residues were found in “blank” bovine urine samples athe levels of 13.0 ng mL−1, 5.9 ng mL−1 and 4.2 ng mL−1, respec-ively: no data are available about the history and breedingonditions of the animals. These data prompted us to furtherlarify if the presence of prednisolone residues in live andlaughtered bovine urines could be related to situations differ-nt from illicit treatment and to identify indirect biomarkers ofllicit corticosteroid treatment in order to provide useful evidencehat could possibly discriminate legal from illegal administra-ions.

The aim of this work is to investigate the effect of a lowosage and long term dexamethasone treatment on the lev-ls of endogenous corticosteroids (hydrocortisone and cortisone)n cattle and to evaluate the possible presence of prednisoloneesidues in bovines bred under strictly controlled conditions andts possible correlation with the levels of endogenous hormones.he presence of natural and synthetic corticosteroids was eval-ated using a multi-residue liquid chromatography–atmosphericressure chemical ionization-(heated nebulizer)–tandem masspectrometry (LC–APCI-(HN)–MS/MS) method for the unequivocalonfirmation of eight corticosteroids (dexamethasone, betametha-one, prednisolone, methylprednisolone, prednisone, methylpred-

isone, cortisone and hydrocortisone); the method was developednd validated in accordance with the criteria of the Commissionecision 2002/657/EC [32].

76 (2011) 616–625 617

2. Experimental

2.1. Animals and experimental protocol

The experimental plan was designed according to the guide-lines of the Italian law for care and use of experimental animals[33] and the study was approved by the Ministry of Health andthe local Committee for animal welfare. Twenty Friesian male vealcalves, were divided in two groups and farmed for 6 months undercontrolled experimentally conditions. The animals were fed a dietavailable on the market, usually employed in zootechnical practicewith ad libitum access to water. Feed ingredients were dairy-products, oils and fats, oilseed products and by-products, cerealproducts and by-products and minerals (21.5% proteins, 20% fats,0.3% fiber and 7.5% ash). Appropriate measures were taken to avoidany kind of cross contamination between the different animals.During the sixth month of breeding 0.4 mg/day of dexamethasone21-disodium phosphate (Desashock) was administered orally to 10animals for 20 consecutive days, the remaining 10 animals wereused as controls.

2.2. Samples collection

Urine samples of the two groups were collected one day beforethe first administration, then at the 10th, 20th, 21st, 22nd, 25th and30th day. The last samples were collected after slaughtering.

Urine were collected (taking care to prevent fecal contami-nation) after milk administration and waiting for spontaneousmicturition and were immediately stored in the dark at −20 ◦C untilanalysis.

2.3. Chemicals and reagents

All solvents were HPLC or analytical grade and purchasedfrom Riedel-de Haën (Seelze, Germany). Water was purified byMilli-Q System (Millipore, Bedford, MA, USA). Sodium acetateanhydrous and �-glucuronidase–arylsulphatase (Helix pomatia)were from Merck (Darmstadt, Germany), this latter was usedas supplied. Sodium hydroxide was from Applichem (Darmstadt,Germany) and acetic acid from Sigma–Aldrich (St. Louis, MO).OASIS HLB (Hydrophile Lipophile Balance) SPE (solid phase extrac-tion) cartridges 3 mL, 60 mg were supplied by Waters (Milford,MA). Dexamethasone, betamethasone, prednisolone, methylpred-nisolone and hydrocortisone were provided by Riedel-de Haën,methylprednisone, prednisone and cortisone were purchasedfrom Sigma–Aldrich (St. Louis, MO). Triamcinolone acetonide-d6was provided by RIVM, the Community Reference Laboratory inBilthoven, Netherlands.

The acetate buffer solution (ABS) was prepared by dissolving12.3 g of sodium acetate anhydrous in 800 mL of purified water:the pH was adjusted to 4.8 and then the solution was made up to afinal volume of 1000 mL.

2.4. Standard solutions

A 1 mg mL−1 stock solution of each standard was prepared inmethanol. From these solutions a 10 �g mL−1 dilution was preparedin methanol. A mixture of all standards is then prepared at dilutionsof 10 ng mL−1 and 100 ng mL−1. These working standards solutionwere stored at 4 ◦C and prepared on a daily basis.

Samples of bovine urine (5 mL) were spiked with triamcinoloneacetonide-d6 used as the internal standard (I.S.) and were addedwith 1 mL of ABS 1 M, pH = 4.8. The mixture was added with 30 �l of

6 roids

�e6tTa30oa1

2

S2upA(ptg10e

wnspatswSau

2

r1cphPvstsap

2

o[cpsma

18 C. Ferranti et al. / Ste

-glucuronidase–arylsulphatase enzyme solution (H. pomatia) (thenzyme activity was about 30 U/m for �-glucuronidase and about0 U/mL for arylsulphatase) and incubated overnight at 37 ◦C, cen-rifuged and then purified by using an OASIS HLB SPE cartridge [34].he cartridge was previously conditioned with 3.0 mL of methanolnd 3.0 mL of water and after the sample loading was washed with.0 mL of water and two times with 3.0 mL of sodium hydroxide.02 M/methanol (60:40, v/v). The analytes were eluted with 5.0 mLf methanol and the solvent removed using an evaporation blockt 40 ◦C under nitrogen atmosphere. The residue was dissolved in00 �l of methanol for LC–MS/MS analysis (10 �l injected).

.6. LC–MS/MS analysis

Analyses were carried out by using a LC system Perkin Elmereries 200 Micro Pump (Perkin Elmer, USA) with a PE Series00 autosampler. The chromatographic separations were obtainednder gradient conditions at room temperature using a reversedhase HPLC column (250 mm × 4.60 mm I.D., 4 (m POLAR-RP 80) Synergi C18 (Phenomenex, USA) with a C18 guard column

4 mm × 2 mm I.D.) (Security Guard, Phenomenex, USA). The mobilehase was composed of 1% acetic acid (mobile phase A) and ace-onitrile (mobile phase B) and the flow rate was 0.8 mL min−1. Theradient profile began at 70% A for 6 min, changing to 50% A in4 min and held for 1 min. Then it was changed from 50% A to% A in five min, held for 2 min and returning to 70% A in 1 min,quilibrating for other 6 min.

The API 3000 triple quadrupole mass spectrometer equippedith an Atmospheric Pressure Chemical Ionization (APCI)-heatedebulizer (HN) source, was set in negative ionization mode with aource temperature of 400 ◦C and a needle current of −3 �A; ultraure nitrogen was used as curtain and collision gas, while ultra pureir was used as nebulizer and auxiliary gas. The full identification ofhe analytes was achieved according to the criteria of the Commis-ion Decision 2002/657/EC [32] and the peak areas of the analytesere computed using the software Analyst version 1.4 from AB

ciex. For the statistical evaluation of the data the software pack-ge Statistical Package for Social Science (SPSS) version 13.0 wassed.

.7. Calibration and quantitation

Urine samples, previously tested and shown to contain noesidues of the compounds of interest, were spiked, daily, with0.0 ng mL−1 of internal standard followed by mixtures of corti-osteroids to obtain concentration intervals of 0.4–4.0 ng mL−1 forrednisolone and prednisone, 0.3–50.0 ng mL−1 for cortisone andydrocortisone and 0.4–10.0 ng mL−1 for the other corticosteroids.eak area ratios of the analyte to internal standard were plottedersus hormone concentrations and a least-squares linear regres-ion model was adopted to calculate matrix calibration curves, usedo estimate the analytes amounts in spiked and naturally incurredamples. A good linearity of calibration curves was confirmed forll the anabolic compounds at all the concentrations checked, asroved by the correlation coefficients, all greater than 0.9980.

.8. Method validation

The validation was performed according to the criteria and rec-mmendations of the European Commission Decision 2002/657/EC32] implementing the Council Directive 96/23/EC [35] and

oncerning the performance of analytical methods and the inter-retation of results. The fully validated matrix is a pooled urineample, taken from bovine different for age, gender and species. Theethod was validated on a multi-residue scale with simultaneous

nalysis of dexamethasone, betamethasone, prednisolone, methyl-

76 (2011) 616–625

prednisolone, prednisone, methylprednisone, hydrocortisone andcortisone. The calibration of the analytes was carried out usingtriamcinolone acetonide-d6 as the internal standard.

For each analyte, the performance of the method was assessedthrough its qualitative parameters, such as analyte specificity,molecular identification in term of retention time (RT), of signal-to-noise ratio and of transition ion ratios, and also through itsquantitative parameters, such as linearity, accuracy in term of true-ness and of precision expressed as the intra- and inter-day/seriesrepeatabilities, and analytical limits (decision limit CC˛ and detec-tion capability CCˇ).

2.8.1. Analyte specificityThe technique of LC–MS/MS itself offers a high degree of selec-

tivity and specificity. The selectivity/specificity of the method wasassessed directly onto the chromatograms obtained from blankand fortified urine matrices and also from naturally incurred urinematrices. It consists of both detecting any extra-peaks in the reten-tion time window of the analyte for the two multiple reactionmonitoring (MRM) transitions of interest onto the blank matrixchromatograms and also checking the matching of the relativeretention time observed for the spiked analytes compared to themethanolic standard analytes with a tolerance of ±2.5% accordingto 2002/657/EC [32].

2.8.2. Molecular identificationThe criteria for molecular identification are those of liquid chro-

matography coupled to tandem mass spectrometry. In terms ofrelative retention time, the analyte peaks in samples were found tobe within 2.5% tolerance when compared with standards. Moreovertwo transition from the analyte molecular peak were monitoredwith a signal-to-noise ratio greater than 3. All ion ratios of sam-ples were within the recommended tolerances as required by the2002/657/EC [32] when compared with standards.

2.8.3. LinearityThe correlation between concentration and the detector

response for each analyte was determined by using a lin-ear regression model. One matrix-match calibration curvewas built each day with five calibration levels in the con-centration intervals of 0.4–4.0 ng mL−1 for prednisolone andprednisone, 0.3–50.0 ng mL−1 for cortisone and hydrocortisone and0.4–10.0 ng mL−1 for the other corticosteroids. Three sets of repli-cates on three different days for a total of five concentrations pointsfor each curve were used to determine linearity.

2.8.4. Accuracy in terms of truenessTrueness was assessed through recovery of the analytes added

to blank matrix in three analytical series each with five levels ofconcentration in six replicates. The mean of calculated concentra-tions at each level of concentration divided by theoretical spikedconcentration gave the percentage of recovery that are in accor-dance with the maximal accepted values that are reported in theCommission Decision 2002/657/EC [32].

The estimated concentrations of analytes were calculated fromthe most intense transition using triamcinolone acetonide-d6 asinternal standard.

2.8.5. Accuracy in terms of precisionThe precision in terms of repeatability and intra-laboratory

reproducibility was evaluated by calculating the relative standard

deviation (R.S.D. %) of the results obtained for each analyte at fivelevels of concentration. Blank urine samples (5 mL) were forti-fied with analytes and I.S., resulting in three analytical series eachwith five corticosteroid concentrations (i.e. 0.4, 0.8, 1.0, 2.0 and4.0 ng mL−1 for prednisolone and prednisone, 0.3, 1.0, 5.0, 20.0 and

eroids

5acdwm

2

bta

sw

fa

oe

3

3

tcCtuifmPnnqfap[rs

qo

dqnat[

weiT

3

a2

C. Ferranti et al. / St

0.0 ng mL−1 for cortisone and hydrocortisone and 0.4, 1.0, 2.0, 5.0,nd 10.0 ng mL−1for the other corticosteroids) and six samples peroncentration. The series were analysed on each of three differentays to evaluate the analytical method’s repeatability (within-day),ithin-laboratory reproducibility (different operators and environ-ental conditions) and recovery (internal standard-corrected).

.8.6. Decision limit CC˛ and detection capability CCˇThe decision limit CC˛ is defined as the limit above which it can

e concluded with an error probability of ˛, that a sample containshe analyte. For prohibited substances an ˛ value equal to 1% ispplied.

The detection capability CCˇ is the smallest content of the sub-tance that may be detected, identified and quantified in a sample,ith a statistical certainty of 1 − ˇ, where ˇ is 5%.

The CC˛, was calculated as three times the signal-to-noise ratioor 20 representative blank bovine urines. The CCˇ is calculatednalysing 20 blank materials fortified with the analytes at the CC˛.

The value of the CC˛ plus 1.64 times the standard deviationf the within-laboratory reproducibility of the measured contentquals the CCˇ value.

. Results and discussion

.1. Optimization of mass spectrometry

The LC–MS/MS method was developed to provide confirma-ory data for the analysis of bovine urine for eight differentorticosteroids. For a method to be deemed confirmatory underommission Decision 2002/657/CE it must yield 4 identifica-ion points. In this method 4 identification points were obtainedsing MRM mode with one precursor ion and two product

ons (corresponding to strong and weak ion). We used a dif-erent ionization technique, both in negative and positive ion

ode, to detect corticosteroids and we chose the Atmosphericressure Chemical Ionization Heated Nebulizer source set inegative ionization mode for the higher intensity of the sig-al to noise ratio for all the analytes. APCI mass spectra (firstuadrupole Q1-scan) showed pseudomolecular ions [M+acetate]−

or triamcinolone acetonide-d6 and [M−H−CH2O]− for dex-methasone, betamethasone, prednisolone, methylprednisolone,rednisone, methylprednisone, hydrocortisone and cortisone. TheM−H−CH2O]− is generated from the loss of a formaldehyde group,esulting from cleavage of the C20–C21 bond on the corticosteroidtructure [36].

Precursor ions selected in the first quadrupole Q1 were subse-uently dissociated in the second quadrupole Q2. The product ionsbtained were registered by scanning the third quadrupole Q3.

As previously reported [36] some of the most abun-ant product ions obtained by scanning the thirduadrupole Q3 were [Precursor−H2O−CH4]− for pred-isolone, [Precursor−H2O−CH4−HF]− for dexamethasonend betamethasone, [Precursor−H2O−CH2O]− for hydrocor-isone, [fragment including ring A or C]− for cortisone andPrecursor−acetate−CH2O]− for triamcinolone acetodine-d6.

The collision energy (CE) and the declustering potential (DP)ere adjusted in MRM mode (Multiple Reaction Monitoring) for

ach transition monitored in order to reach the highest sensitiv-ty for all the analytes. The optimized parameters are reported inable 1.

.2. Performance characteristics of the method

The performance characteristics of the method were evalu-ted in accordance with the criteria of the Commission Decision002/657/CE [32].

76 (2011) 616–625 619

Blank and standard spiked samples were analysed and no sig-nals were observed on the MRM chromatograms of the blank atthe retention time in which eluted the compounds of interest.However we found a chromatographic peak that eluted just beforethe prednisolone chromatographic peak and had the same prod-uct ions of prednisolone but with different ion ratios: this peak didnot interfere with the quantification of prednisolone residues (thechromatographic separation was acceptable in term of resolution,the relative retention time was 11.9 min for the extra peak and12.4 min for prednisolone).

Confirmation of the analytes was carried out using four identi-fication points. One precursor ion refers to one identification pointand two transitions together represent three points. The ion ratiosin the standard solution and in the samples during the validationwere applied and the ion ratios of each spiked sample fell withinthe maximum permitted tolerances for positive identifications.

Good linearity was obtained for matrix calibration curves withcorrelation coefficients higher than 0.9980 for all compounds.

Since no CRMs are available, the trueness of measurements wasassessed through recovery of additions of known amounts of theanalytes to a blank matrix, the mean recovery were in accordancewith the maximal accepted values reported in the CommissionDecision. Inter-day recovery data ranged from 99.3% to 104.7% fordexamethasone, from 98.5% to 103.1% for betamethasone, from97.6% to 104.9% for hydrocortisone, from 97.2% to 105.2% for corti-sone, from 99.3% to 101.3% for prednisolone, from 97.2% to 102.0%for prednisone, from 95.0 to 102.9% for methylprednisolone andfrom 98.8% to 102.1% for methylprednisone. The coefficient of vari-ations were also satisfactory with values for inter-day repeatabilityvarying from 2.4% to 11.6% for dexamethasone, from 6.5% to 10.4%for betamethasone, from 4.0% to 13.8% for hydrocortisone, from4.1% to 11.3% for cortisone, from 3.9% to 7.6% for prednisolone, from7.8% to 10.3% for prednisone, from 9.0% to 11.0% for methylpred-nisolone and from 7.9% to 12.0% for methylprednisone.

The two critical concentrations CC˛ and CCˇ were deter-mined for all the analytes following the Commission Decision657/2002/CE [32]. The CC˛ and CCˇ values ranged from 0.2 ng mL−1

to 0.4 ng mL−1 and from 0.3 ng mL−1 to 0.5 ng mL−1, respectively(see Table 2).

3.3. Residues of natural and synthetic corticosteroids

Dexamethasone residues were detected (>CC˛) in all treatedbovine urine samples collected at 10th and 20th day of the treat-ment and the day after the last administration (21st day). At 22ndand 25th day of sampling dexamethasone residues were found inthe 90% and 50% of the analysed samples, respectively. Cortisoneand hydrocortisone were not detected in the same samples untiltwo days after the withdrawal of the treatment (with the excep-tion of two urine samples found positive for hydrocortisone at theconcentration of 0.4 ng mL−1). Our findings (see Table 3.) confirmthe high and rapid rate of dexamethasone urinary excretion [37]and show that low-dose and long-term dexamethasone treatmentinterfere with endogenous hydrocortisone and cortisone synthe-sis, as previously reported in the horse and in cattle (29,31,38,39];a daily dexamethasone administration could attain circulatingsteady state concentrations of this hormone able to influence theproduction of endogenous hydrocortisone in the adrenal cortex bya negative-feedback phenomenon on the hypothalamic-pituitary-adrenal axis. Dexamethasone residues were not detected in urinesamples collected from control animals and from treated animals

before the first administration. Prednisolone residues were notdetected in urine samples collected from treated animal at 10th,20th, 21st, 22nd, 25th and 30th day (the last day of breeding).

The data reported in Tables 3 and 4 were determined byaveraging the urine concentration values of each hormone (dexam-

620 C. Ferranti et al. / Steroids 76 (2011) 616–625

Table 1Transitions and method performances for the analysed steroids in MRM mode.

Compound Transition Collisionenergy (eV)

DeclusteringPotential (V)

Retentiontime (min)

CC˛(ng mL−1)

CCˇ(ng mL−1)

Dexamethasone 361 > 307a −27 −60 16.5 0.3 0.4361 > 292 −31 −60

Betamethasone 361 > 307a −27 −60 16.2 0.3 0.4361 > 292 −31 −60

Prednisolone 329 > 295a −30 −60 12.4 0.3 0.4329 > 280 −30 −60

Methylprednisolone 343 > 294a −35 −55 15.4 0.3 0.4343 > 309 −35 −60

Prednisone 327 > 149a −45 −50 14.0 0.3 0.4327 > 285 −25 −50

Methylprednisone 341 > 149a −35 −60 17.2 0.4 0.5341 > 123 −40 −55

Hydrocortisone 331 > 282a −35 −75 12.5 0.2 0.3331 > 297 −25 −75

Cortisone 329 > 137a −30 −60 14.6 0.2 0.3329 > 301 −20 −50

Triam. Acet-d6 499 > 419a −30 −40 19.1 n.d. n.d.499 > 317 −30 −40

a Most abundant product ion.

Table 2Inter-day validation statistics in bovine urines.

Parameter (mean value)a Validation samples level (ng mL−1)

Dexamethasone Betamethasone Hydrocortisone Cortisone

0.4 1.0 2.0 5.0 10.0 0.4 1.0 2.0 5.0 10.0 0.3 1.0 5.0 25.0 50.0 0.3 1.0 5.0 25.0 50.0

Mean recovery (%) 104.7 101.3 100.2 99.3 101.9 103.1 99.4 102.8 98.5 99.8 104.9 103.8 97.6 100.4 99.9 105.2 104.8 97.2 98.3 100.9Precision (CV %) 11.6 6.4 3.1 3.0 2.4 10.4 9.7 9.9 8.9 6.5 13.8 11.4 9.2 5.8 4.0 11.3 10.0 8.7 4.9 4.1

Parameter (mean value)a Validation samples level (ng mL−1)

Prednisolone Prednisone Methylprednisolone Methylprednisone

0.4 0.8 1.0 2.0 4.0 0.4 0.8 1.0 2.0 4.0 0.4 0.8 1.0 2.0 4.0 0.4 0.8 1.0 2.0 4.0

Mean recovery (%) 101.3 99.3 100.8 100.2 100.6 97.2 98.9 102.0 98.1 99.3 95.0 102.9 97.2 100.3 100.5 102.0 101.5 102.1 98.8 99.7Precision (CV%) 7.6 7.3 6.5 4.5 3.9 10.3 10.0 9.8 9.5 7.8 11.0 10.3 10.2 9.9 9.0 12.0 11.3 9.6 9.8 7.9

a n = 6 replicates.

Table 3Dexamethasone, hydrocortisone, cortisone and prednisolone residues in treated bovines.

Treated bovines

Sampling 0th 10th 20th 21st 22nd 25th 30th 30thslaughterhouse

Dexamethasone concentration in urine samples (ng mL−1)Average n/a 2.0 2.1 1.2 1.0 0.4 n/a n/aConcentration interval n/a 0.5–9.1 0.4–5.9 0.4–3.1 n.d.–2.5 n.d.–2.5 n/a n/a>CC˛ 0/7 10/10 10/10 10/10 9/10 5/10 0/10 0/10Hydrocortisone concentration in urine samples (ng mL−1)Average 4.2 n/a n/a n/a n/a 1.7 2.1 16.7Concentration interval 1.3–14.9 n.d.–0.4 n/a n.d.–0.4 n/a n.d.–3.1 n.d.–4.1 6.2–29.7>CC˛ 7/7 1/10 0/10 1/10 0/10 2/10 9/10 10/10Cortisone concentration in urine samples (ng mL−1)Average 5.4 n/a n/a n/a n/a 1.6 3.0 20.7Concentration interval 2.1–10.8 n/a n/a n/a n/a n.d.–3.7 n.d.–5.5 6.0–44.7>CC˛ 7/7 0/10 0/10 0/10 0/10 3/10 9/10 10/10Prednisolone concentration in urine samples (ng mL−1)Average n/a n/a n/a n/a n/a n/a n/a 0.6

n

Concentration interval n.d.–0.5 n/a n/a n/a>CC˛ 1/7 0/10 0/10 0/10

.d., not detected. n/a, not applicable.

n/a n/a n/a n.d.–1.50/10 0/10 0/10 9/10

C. Ferranti et al. / Steroids 76 (2011) 616–625 621

Table 4Dexamethasone, Hydrocortisone, cortisone and prednisolone residues in control bovines.

Control bovines

Sampling 0th 10th 20th 21st 22nd 25th 30th 30thslaughterhouse

Dexamethasone concentration in urine samples (ng mL−1)Average n/a n/a n/a n/a n/a n/a n/a n/aConcentration interval n/a n/a n/a n/a n/a n/a n/a n/a>CC˛ 0/8 0/9 0/10 010 0/10 0/10 0/8 0/9Hydrocortisone concentration in urine samples (ng mL−1)Average 3.3 2.5 1.1 2.3 1.7 4.3 3.9 21.5Concentration interval 0.7–8.4 0.5–12.8 n.d.–3.5 0.6–7.6 0.5–3.3 1.0–19.9 1.2–12.0 6.4–46.3>CC˛ 8/8 9/9 9/10 10/10 10/10 10/10 8/8 9/9Cortisone concentration in urine samples (ng mL−1)Average 4.5 4.1 1.6 3.6 3.1 5.1 7.1 26.6Concentration interval 1.0–13.2 1.0–20.8 0.5–4.7 1.1–12.8 1.0–5.8 0.5–21.6 1.4–21.7 7.6–43.3>CC˛ 8/8 9/9 10/10 10/10 10/10 10/10 8/8 9/9Prednisolone concentration in urine samples (ng mL−1)Average n/a n/a n/a n/a n/a 0.5 n/a 0.7

n/a0/1

n

ef

3a

lacico1tsbwphi0

c1tp

3a

sa0av2

1scao2t

Concentration interval n/a n.d.–0.7 n/a>CC˛ 0/8 1/9 0/10

.d., not detected. n/a, not applicable.

thasone, hydrocortisone, cortisone and prednisolone), obtainedrom treated and control animals at each interval time.

.3.1. Prednisolone residues in urine samples from live controlnd treated animals

Urine samples collected from live control animals and fromive treated animals before the first administration showed (7% ofnalysed samples, 5/72) prednisolone residues at an average con-entration of 0.6 ng mL−1 (concentration interval 0.5–0.7 ng mL−1):n these prednisolone positive samples, hydrocortisone andortisone residues were detected at an average concentrationf 12.9 ng mL−1 (concentration interval 5.0–19.9 ng mL−1) and6.2 ng mL−1 (concentration interval 6.1–21.7 ng mL−1), respec-ively. Some other samples (46% of analysed samples, 33/72)howed the presence of a chromatographic peak characterizedy a very low signal intensity, at the same retention time butithout the same ion ratio corresponding to the standard ofrednisolone; in these samples the average concentrations ofydrocortisone and cortisone were 2.9 ng mL−1 (concentration

nterval 0.3–8.4 ng mL−1) and 4.6 ng mL−1 (concentration interval.3–13.2 ng mL−1), respectively.

The concentration of hydrocortisone (average 1.3 ng mL−1,oncentration interval 0.3–4.6 ng mL−1) and cortisone (average.9 ng mL−1, concentration interval 0.5–5.2 ng mL−1) decreased inhe samples without prednisolone residues (47% of analysed sam-les, 34/72).

.3.2. Prednisolone residues in urine samples from slaughterednimals

Urine samples collected from the bladder of control animalshowed (56% of analysed samples, 5/9) prednisolone residues atn average concentration of 0.7 ng mL−1 (concentration interval.4–1.4 ng mL−1) and average concentrations of hydrocortisonend cortisone respectively of 29.9 ng mL−1 (concentration inter-al 16.6–46.3 ng mL−1) and 34.5 ng mL−1 (concentration interval0.6–43.3 ng mL−1).

Urine samples collected from the bladder of treated animals,0 days after the last treatment, evidenced (90% of analysedamples, 9/10) the presence of prednisolone at an average con-

entration of 0.6 ng mL−1 (concentration interval 0.4–1.5 ng mL−1),nd average concentrations of hydrocortisone and cortisonef 17.8 ng mL−1 (concentration interval 10.8–29.7 ng mL−1) and2.4 ng mL−1 (concentration interval 11.7–44.7 ng mL−1), respec-ively.

n/a n.d.–0.6 n.d.–0.4 n.d.–1.40 0/10 2/10 1/8 5/9

The remaining urine samples collected from the bladder oftreated and control bovine showed the chromatographic peakas previously reported above, with the same retention time ofprednisolone but with ion ratios not in accordance with that ofprednisolone spiked samples.

Data reported in Sections 3.3.1 and 3.3.2, related to averageconcentrations and concentration intervals for hydrocortisone andcortisone, refer only to prednisolone positive or negative samples.

None of the analysed samples showed the presence of pred-nisone.

We have also analysed [40] the feed used during the animalbreeding to verify the absence of steroid hormones like dexametha-sone, prednisolone and prednisone; we have found hydrocortisoneand cortisone, according to what has been reported in the scien-tific literature [41], at an average concentration of 0.7 ng mL−1 and0.3 ng mL−1, respectively.

Our findings suggest a possible correlation between the pres-ence in urine samples of prednisolone residues and higherconcentrations of hydrocortisone and cortisone (see Fig. 1a and b).The concentrations of these two hormones increase probably dueto natural stress particularly at the slaughterhouse, both in treatedand control animals.

The Pearson correlation between prednisolone, hydrocortisoneand cortisone was calculated and subsequently the P-value. ThePearson correlation between these three hormones ranges between0.7 and 0.96 (P-value of 0.01); in general a Pearson correlation from0.7 to 1.0 is considered as a strong association, from 0.3 to 0.7 as aweak association and from 0 to 0.3 as none or little association, nev-ertheless the criteria adopted are somewhat arbitrary and shouldnot be used to strictly.

In order to clarify the presence of prednisolone residues, urinesamples from live bovines (a 4 years old female and a 18 monthsyear old male) bred in the wild were also analysed. They showedthe presence of a chromatographic peak at the same retentiontime but without the same ion ratio corresponding to the stan-dard of prednisolone, preventing the unequivocal identificationof the compound in accordance with the criteria of the Commis-sion Decision 2002/657/EC [32]. These other results require furtherinvestigation in order to find out the identity and the origin of

this chromatographic peak and to confirm the correlation betweensynthetic and endogenous corticosteroids.

The presence of prednisolone residues could be explainedhypothesizing a conversion of endogenous corticosteroids, likehydrocortisone, to prednisolone by enzymes or microorganisms.

622 C. Ferranti et al. / Steroids 76 (2011) 616–625

Fig. 1. (a) Extract ion chromatogram of urine sample from a treated bovine. (b) Extract ion chromatogram of urine from a control bovine.

C. Ferranti et al. / Steroids 76 (2011) 616–625 623

Fig. 1. (Continued).

6 roids

cftbb

romwtc

pitf

spmt

rw

4

nHfttrlitwds

cebptrtg

A

otb

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

24 C. Ferranti et al. / Ste

The enzymatic reactions in the metabolism of the corti-osteroids in mammals, particularly of hydrocortisone, are theollowing: the reduction of the carbonyl group at the C20 position,he oxidation–reduction at the C11 position, the hydroxylation ofoth the C2 and C6 positions and finally the reduction of the doubleond in C4.

The enzymatic conversion of hydrocortisone to prednisoloneequires the oxidation at the C1 position of hydrocortisone, thepposite reaction has been reported in the literature [42]: Ver-eulen et al. investigated the metabolism of prednisolone in vitroith a rat liver preparation, demonstrating the bioconversion of

his hormone to naturally occurring corticosteroids, e.g. hydro-ortisone.

Several studies have been devoted to develop biotechnologicalrocesses to produce steroids which are widely used as anti-

nflammatory drugs including prednisolone and hydrocortisone;he bioconversion of hydrocortisone to prednisolone by isolatedungi and bacteria were extensively investigated [43,44].

Arioli and co-workers published [45] the results of an in vitrotudy demonstrating that urine contaminated with feces could beositive for prednisolone and prednisone residues as a result oficrobiological dehydrogenase activity on hydrocortisone and cor-

isone.This study does not explain the presence of prednisolone

esidues in the bovine bladder which is well known to be a sampleithout fecal contamination.

. Conclusions

Here we present an LC–MS/MS method for the simulta-eous determination of eight corticosteroids in bovine urine.ydrocortisone, cortisone, dexamethasone and prednisolone were

ound in urine samples, collected from bovines experimen-ally treated with dexamethasone and bovines bred accordingo strictly established control conditions. The results show aeduction of hydrocortisone and cortisone levels in urine col-ected from treated animals, that could be considered as anndirect indicator of oral, low-dose, long term dexamethasonereatment in cattle. Furthermore low prednisolone togetherith high hydrocortisone and cortisone residue levels wereetected in urine samples from control bovines particularly at thelaughterhouse.

Further studies are necessary to investigate the possibleorrelation between endogenous (hydrocortisone/cortisone) andxogenous corticosteroids (prednisolone) and to study the possi-le presence of microorganisms in bovine urines especially underarticular physiological conditions. Studies are ongoing in ordero find out the reason of unexplained excretion of prednisoloneesidues in urine samples of untreated bovines. We also suggesthe employing of a metabolomics profiling approach that wouldive a concrete explanations to the results obtained.

cknowledgements

The study was carried out within the project, “Developmentf an integrate diagnostic protocol for the control related to illicitreatment with growth promoters in ruminants”, that was fundedy Italian Ministry of Health (RF-IZP-2006-364645).

eferences

[1] Osamu N. Review: steroid analysis for medical diagnosis. J Chromatogr A2001;935:267–78.

[2] Schuerholz T, Keil O, Wagner T, Klinzing S, Sumpelmann R, Obule V, et al. Hydro-cortisone does not affect major platelet receptors in inflammation in vitro.Steroids 2007;72:609–13.

[

76 (2011) 616–625

[3] Ferguson DC, Hoenig M. Glucocorticoids, mineralcorticoids and steroid synthe-sis inhibitors. In: Adams HR, editor. Veterinary Pharmacology and Therapeutics.7th ed. Ames, IA: IOWA State University Press; 1995. p. 622–37.

[4] EMEA. Prednisolone (as free alcohol) summary report – Committee for Veteri-nary Medicinal Products. EMEA/MRL/629/99-FINAL; July 1999.

[5] Council Regulation (EEC) No 2377/90 of 26 June 1990 laying down a Commu-nity procedure for the establishment of maximum residue limits of veterinarymedicinal products in foodstuff of animal origin. O J L224; 18.8.1990.

[6] Courtheyn D, Le Bizec B, Brambilla G, De Brabander HF, Cobbaert E, Van De WieleM, et al. Recent developments in the use and abuse of growth promoters. AnalChim Acta 2002;473:71–82.

[7] Istasse L, De Haan V, Van Eneaeme C, Buts B, Baldwin P, Gielen M. Effect of dex-amethasone injections on performances in a pair of monozygotic cattle twins.J Anim Physiol Anim Nutr 1989;62:150–8.

[8] Directive 2003/74/EC of the European Parliament and of the Council of 22September 2003, amending Council Directive 96/22/EC concerning the pro-hibition on the use in stockfarming of certain substances having a hormonal orthyreostatic action and of beta-agonist. O J L262; 14.10.2003.

[9] Choi MH, Hamh RJ, Jung BH, Chung BC. Measurement of corticoids in thepatients with clinical features indicative of mineralcorticoid excess. Clin ChimActa 2002;320:95–9.

10] Delahaut Ph, Jacquemin P, Colemonts T, Dubois M, De Graeve J, DeluykerH. Quantitative determination of several synthetic corticosteroids by gaschromatography–mass spectrometry after purification by immunoaffinitychromatography. J Chromatogr B 1997;696:203–15.

11] Aburuz S, Millership J, Heaney L, McElnay J. Simple liquid chromatographymethod for the rapid simultaneous determination of prednisolone and cor-tisol in plasma and urine using hydrophilic lipophilic balanced solid phaseextraction cartridges. J Chromatogr B 2003;798:193–201.

12] Andersen HJ, Hansen GL, Pedersen M. Optimization of solid phase extractionclean up and validation of quantitative determination of corticosteroids inurine by liquid chromatography–tandem mass spectrometry. Anal Chim Acta2008;617:216–24.

13] Hauser B, Deschner T, Boesch C. Development of a liquidchromatography–tandem mass spectrometry method for the determi-nation of 23 endogenous steroids in small quantities of primate urine. JChromatogr B 2008;862:100–12.

14] Cho HJ, Kim JD, Lee WY, Chung BC, Choi MH. Quantitative metabolic profil-ing of 21 endogenous corticosteroids in urine by liquid chromatography–triplequadrupole-mass spectrometry. Anal Chim Acta 2009;632:101–8.

15] McDonald M, Granelli K, Sjoberg P. Rapid multi-residue method for the quan-titative determination and confirmation of glucocorticosteroids in bovine milkusing liquid chromatography–electrospray ionization-tandem mass spectrom-etry. Anal Chim Acta 2007;588:20–5.

16] Antignac JP, Bizec BL, Monteau F, Andrè F. Study of natural and artificial cor-ticosteroid phase II metabolites in bovine urine using HPLC–MS/MS. Steroids2002;67:873–82.

17] Keeffe MJO, Martin S, Regan L. Validation of a multiresidue liquidchromatography–tandem mass spectrometry method for the quantitation andconfirmation of corticosteroid residues in urine, according to the proposedSANCO 1085 criteria for banned substances. Anal Chim Acta 2003;483:341–50.

18] San Giorgio E, Curatolo M, Assassini W, Bozzoni E. Application of neutral lossmode in liquid chromatography–mass spectrometry for the determination ofcorticosteroids in bovine urine. Anal Chim Acta 2003;483:259–67.

19] Malone EM, Elliot CT, Kennedy DG, Regan L. Development of a rapid method forthe analysis of synthetic growth promoters in bovine muscle using liquid chro-matography tandem mass spectrometry. Anal Chim Acta 2009;637:112–20.

20] Malone EM, Dowling G, Elliot CT, Kennedy DG, Regan L. Development of arapid multi-class method for the confirmation analysis of anti-inflammatorydrugs in bovine milk using liquid chromatography tandem mass spectrometry.J Chromatogr A 2009;1216:8123–40.

21] Tölgyesi A, Sharma VK, Kovacsics L, Fekete J. Quantification of corticosteroidin bovine urine using selective solid phase extraction and reversed-phase liquid chromatography/tandem mass spectrometry. J Chromatogr B2010;878:1471–9.

22] Shao B, Cui X, Yang Y, Zhang J, Wu Y. Validation of solid-phase extraction andultra-performance liquid chromatography tandem mass spectrometry methodfor the detection of 16 glucocorticoids in pig tissues. J AOAC 2009;29:609–11.

23] Stolker AAM, Brinkman UATh. Analytical strategies for residues analysis of vet-erinary drugs and growth promoting agents in food producing animals – areview. J Chromatogr A 2005;1067:15–53.

24] Caretti F, Gentili A, Ambrosi A, Mainero Rocca L, Delfini M, Di Cocco ME, et al.Residue analysis of glucocorticoids in bovine milk by liquid chromatographytandem mass spectrometry. Anal Bioanal Chem 2010;397:2477–90.

25] Gaignage P, Lognay G, Bosson D, Vertogen D, Dreze P, Marlier M, et al.Dexamethasone bovine pharmacokinetics. Eur J Drug Metab Pharmacokinet1991;16:219–21.

26] Courtheyn D, Vercammen J, De Brabander H, Vandenreyt I, Batjoens P, Vanoost-huyze K, et al. Determination of dexamethasone in urine and faeces oftreated cattle with negative chemical ionization–mass spectrometry. Analyst

1994;119:2557–64.

27] Van Den Hauwe O, Schneider M, Sahin A, Van Peteghem CH, Naegeli H. Immuno-chemical screening and liquid chromatographic–tandem mass spectrometricconfirmation of drug residues in edible tissues of calves injected with thera-peutic dose of the synthetic glucocorticoids dexamethasone and flumethasone.J Agric Food Chem 2003;51:326–30.

eroids

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

C. Ferranti et al. / St

28] Calvarese S, Rubini P, Urbani G, Ferri N, Ramazza V, Zucchi M. Experimentaladministration of 19-nortestosterone and dexamethasone in cattle: elimina-tion of the two drugs in different biological matrices. Analyst 1994;119:2611–5.

29] Capolongo F, Tapparo M, Merlanti R, Ravarotto L, Tealdo E, Gallina G, et al.Illicit treatments in cattle and urinary 6beta-hydroxycortisol/cortisol ratio.Anal Chim Acta 2007;586:228–32.

30] Möstl E, Messman S, Bagu E, Robia C, Palme R. Measurement of glucocorti-coid metabolite concentrations in faeces of domestic livestock. J Vet Med A1999;46:621–32.

31] König T, Schuberth HJ, Leibold W, Zerbe H. Dexamethasone depresses theexpression of 1-selectin but not the in vivo migration of bovine neutrophilsinto the uterus. Theriogenology 2006;65(7):1227–41.

32] Commission of the European Communities. Commission Decision 2002/657/ECof 12 August 2002, implementing Council Directive 96/23/EC concerning theperformance of analytical methods and the interpretation of results. Off J EurCommun: Legis 2002;L221:8–36.

33] D. Lgs. 193/06 implementing the Directive 2004/28/EC of the European Parlia-ment and of the Council of 31st March 2004, on the Community code relatingto veterinary medicinal products, G.U. n. 121 of 26 May 2006, S.O.

34] O’Keeffe MJ, Martin S, Regan L. Validation of a multiresidue liquidchromatography–tandem mass spectrometric method for the quantitation andconfirmation of corticosteroid residues in urine, according to the proposedSANCO 1085 criteria for banned substances. Anal Chim Acta 2003;483:341–50.

35] Directive 96/23/EC of 29 April 1996, on measures to monitor certain substances

and residues thereof in live animals and animal products and repealing Direc-tives 85/358/EEC and 86/469/EEC and Decisions 89/187/EEC and 91/664/EEC.OJ L 125; 23.05.1996, p. 10.

36] Savu SR, Silvestro L, Haag A, Sörgel F. A confirmatory HPLC–MS/MSmethod for ten synthetic corticosteroids in bovine urines. J Mass Spectrom1996;31:1351–63.

[

76 (2011) 616–625 625

37] Vicenti M, Girolami F, Capra P, Pazzi M, Carletti M, Gardini M, et al. Study ofdexamethasone urinary excretion profile in cattle by LC–MS/MS: comparisonbetween therapeutic and growth-promoting administration. J Agric Food Chem2009;57:1299–306.

38] Soma LR, Uboh CE, Luo Y, Guan F, Moate PJ, Boston RC. Pharmacokinetics ofdexamethasone with pharmacokinetic/pharmacodynamic model of the effectof dexamethasone on endogenous hydrocortisone and cortisone in the horse.J Vet Pharmacol Ther 2005;20:71–80.

39] Soma LR, Uboh CE, Luo Y, Guan F, Moate PJ, Boston RC. Pharmacoki-netics of methylprednisolone acetate intra-articular administration and itseffect on endogenous hydrocortisone and cortisone secretion in horse. AJVR2006;67:654–62.

40] Fiori M, Pierdominici E, Longo F, Brambilla G. Identification ofmain corticosteroids as illegal feed additives replacers by liquidchromatography–atmospheric pressure chemical ionisation mass spec-trometry. J Chromatogr A 1998;807:219–27.

41] Koldovsky O, Thornburg W. Review: hormones in milk. J Pediatric GastroenterolNutr 1987;6:172–96.

42] Vermeulen A, Caspi E. The metabolism of Prednisolone by Homogenates of ratliver. J Biol Chem 1958;233:54.

43] Mahato SB, Subhadra G. Advances in microbial steroid biotransformation.Steroids 1997;62:332–45.

44] Manosroi J, Manosroi A, Saraphanchotiwittaya A, Abe M. Comparison of pred-nisolone production from cortexolone by free and immobilized isolated and

standard collection strains using mixed culture techniques. J Chem TechnolBiotechnol 1999;74:364–70.

45] Arioli F, Fidani M, Casati A, Fracchiolla ML, Pompa G. Investigation on pos-sible transformation of cortisol, cortisone and cortisol glucuronide in bovinefaecal matter using liquid chromatography–mass spectrometry. Steroids2010;75:350–4.