APPROACHES TO THE DETECTION OF
STEROID ABUSE IN VETERINARY SPECIES
1 of 331
APPROACHES TO THE DETECTION OF
STEROID ABUSE IN VETERINARY SPECIES
James P. Scarth
(Ghent University, Belgium and HFL Sport Science, UK)
Thesis for submission in fulfilment of the requirements
for the degree of Doctor (Ph.D) in Veterinary Sciences
Promoter: Prof. Dr. H. De Brabander (Ghent University, Belgium)
Co-promoter: Dr. J. Kay (University of Strathclyde, UK)
Co-promoter: Dr. L. Vanhaecke (Ghent University, Belgium)
Rector: Prof. Dr. P. Van Cauwenberge (Ghent University, Belgium)
3 of 331
TABLE OF CONTENTS
PART 1: DETECTION OF ‘ENDOGENOUS’ STEROID ABUSE IN
Chapter 1: Presence, metabolism
‘endogenous’ steroid hormones in food producing animals
Chapter 2: Validation and application of an analytical biomarker
approach for the detection of nandrolone abuse in the porcine
Chapter 3: Validation of analytical biomarker approaches for the
detection of androgen, oestrogen and progestagen abuse in the
PART 2: DETECTION OF ‘DESIGNER’ STEROID ABUSE IN
Chapter 4: Steroid metabolism and detection in the equine
Chapter 5: Assessment of the applicability of in vitro
technologies to study drug metabolism in the equine
Chapter 6: Metabolism of the ‘designer’ steroid estra-4,9-diene-
3,17-dione in the equine and comparison to human and canine
CHAPTER 7: GENERAL DISCUSSION
and detection of
5 of 331
atmospheric pressure ionization.
British Horseracing Authority.
a steroid based on the estrane nucleus.
a steroid based on the androstane nucleus.
a steroid based on the pregnane nucleus.
a steroid based on the cholane nucleus.
a steroid based on the cholestane nucleus.
collision activated dissociation.
European Community Reference Laboratory.
enzyme linked immunosorbent assay
enhanced product ion scan.
endogenous reference compound.
Food and Drug Administration.
Federation Equestre Internationale.
full width at half maximum height.
gas chromatography combustion isotope ratio mass spectrometry.
gas chromatography-mass spectrometry.
gas chromatography-tandem mass spectrometry.
Greyhound Board of Great Britain.
higher-energy collision decomposition.
high performance liquid chromatography.
high resolution-liquid chromatography-mass spectrometry.
International Conference of Racing Analysts and Veterinarians.
International Federation of Horseracing Authorities.
International Laboratory Accreditation Cooperation.
6 of 331
MSTFA – N-Methyl-n
MTBSTFA – N
International Union of Pure and Applied Chemistry.
liquid chromatography-mass spectrometry.
liquid chromatography-tandem mass spectrometry.
lower limit of quantification.
limit of detection.
limit of quantification.
linear trap quadrupole.
minimum required performance limit.
mass to charge ratio
nicotinamide adenine dinucleotide.
nicotinamide adenine dinucleotide phosphate.
nuclear magnetic resonance.
national monitoring programme.
programmable temperature vaporiser.
quantitative real time reverse transcriptase polymerases
chain reaction technology
relative standard deviation (also known as coefficient of
Scientific Committee on Veterinary Measures relating to
selected reaction monitoring.
surface plasmon resonance.
thin layer chromatography-fluorescence.
time of flight.
atomic mass unit
upper limit of quantification.
World Anti-Doping Agency.
7 of 331
Steroid structure and pharmacology
If one were to ask a member of the public what images the word ‘steroid’ conjured into
their imagination, the use of anabolic-androgenic steroids (AASs) in athletics or
bodybuilding would no doubt rank near the top of the list. However, steroids have a
range of structures and pharmacological actions that reach far beyond the anabolic
effects of AASs. The term ‘steroid’ itself refers to any compound possessing the basic
perhydrocyclopentanophenanthrene nucleus (Figure 1) (Makin, 1995).
Figure 1 – A) the perhydrocyclopentanophenanthrene nucleus, on which all steroids are
based and B) cholesterol as an example. Each carbon is assigned a number and the
four hydrocarbon rings are numbered A-D, as shown.
The nomenclature of this class of compounds is complex and large arrays of different
systems are used. These include; the official International Union of Pure and Applied
Chemistry (IUPAC) recommended systematic nomenclature (IUPAC, 2010), a range of
‘trivial’ or ‘common’ names and those of some proprietary preparations. Additionally,
many organisations use their own nomenclature (for example the company Steraloids).
However, these often deviate from the IUPAC recommendations. The choice of how to
8 of 331
best name a steroid in a particular situation is, therefore, dependent on a number of
factors. If one were to always use only the IUPAC systematic name, then this could
make the text difficult to read for a non-expert. However, inappropriate over-use of trivial
names does not always give enough information in order to inform the reader. Therefore,
a combination of systematic and trivial names is often employed as a pragmatic
compromise (such as described by Makin et al. 1995) and this will be used in the current
text. As an example of the different ways of naming a steroid, some options for
testosterone are given below:
Trivial name: testosterone.
IUPAC systematic name: 17β-hydroxy-androst-4-en-3-one.
Proprietary example (containing testosterone esters): Sustanon.
When depicted in the orientation shown in Figure 1, substituents on the steroid backbone
may protrude below or above the plane of the paper and are drawn as such using either
a dashed or solid wedge respectively (indicating the stereochemistry α and β
respectively). Hydrogens in positions 8, 9, 10, 13 and 14 (when present) take β, α, β, β,
and α orientation respectively in all steroids discussed in this manuscript so their
stereochemistry will not be shown in any of the subsequent diagrams. A substituent in
position 5 may take either the α or β form, so hydrogens in this position will always be
labelled. A wavy line indicates that stereochemistry is unspecified.
In order to aid in the systematic naming of steroids, a number of different hydrocarbon
backbones are specified for use by IUPAC. These differ in the number and orientation of
carbons, which range from the 17-carbon (C17) gonane nucleus to the 27-carbon (C27)
cholestane nucleus (on which cholesterol is based). The range of steroid backbones
used in systematic nomenclature is shown in Figure 2. In this text, when describing the
trivial name for the oestrogens, the English version will be used (as opposed to the USA
use of estrogens). However, when systematically naming steroids that are based on the
estrane nucleus, the ‘o’ will not be used (in accordance with IUPAC guidelines).
No endogenous and very few exogenous steroids are based on the gonane nucleus. The
oestrogens and nandrolone (17β-hydroxy-estr-4-en-3-one) are based on the estrane
nucleus. The majority of androgens are based on the androstane nucleus and the
majority of progestagens and corticosteroids are based on the pregnane nucleus. Most
9 of 331
of the bile acids are based on the cholane nucleus and sterols such as cholesterol are
based on the cholestane nucleus.
Gonane (C17) Estrane (C18)
Androstane (C19) Pregnane (C21)
Cholane (C24) Cholestane (C27)
Figure 2 – the range of hydrocarbon backbones used in steroid nomenclature.
10 of 331
While many steroids are known to be endogenous (discussed further in chapter 1), a
wide range of exogenous steroid structures have been synthesized by chemists over the
years in order to optimise their biological properties. Pharmacologically, steroids possess
a range of activities far more diverse than their seemingly similar structures may
suggest. The following discussion considers the major effects of different steroid classes
in mammals. There are some subtle differences between various species, but these will
not be considered here since it is only a general overview.
Cholesterol (cholest-5-en-3β-ol – Figure 1) is derived from dietary intake, but is also
synthesized in the body. Cholesterol acts to regulate the fluidity of cell membranes and is
the precursor to the endogenous androgens, oestrogens, progestagens, corticosteroids,
vitamin D, the bile acids and, in certain species, to pheromones such as the 16-
androstenes (Hadley and Levine, 2006).
Bile acids such as cholic acid (3α,7α,12α-trihydroxy-5β-cholan-24-oic acid – Figure 3)
are secreted by the gall bladder into the intestine where they aid the absorption of lipids
into the body by reducing their surface tension (Hadley and Levine, 2006).
Figure 3 – cholic acid (a bile acid).
Progestagens such as progesterone (pregn-4-ene-3,20-dione – Figure 4) are produced
by both males and females in the adrenal glands and gonads. However, they are
secreted in much higher concentrations by females during certain stages of the ovulatory
cycle (including by the corpus luteum) and during pregnancy (Hadley and Levine, 2006).
Progestagens produce the majority of their effects through agonism of the progesterone
receptor. This leads to an increased metabolic rate, changes in breast morphology and
development/maintenance of the uterus/oviduct before and during pregnancy (Hadley
and Levine, 2006). Progestagens (and synthetic progestins) may also be used as
11 of 331
contraceptives in females, which act by suppressing endogenous gonadotrophin release
and by inhibiting sperm penetration due to a change in viscosity of the cervical mucous
(Westhoff et al. 2010).
Figure 4 – progesterone (a progestagen).
Corticosteroids are produced by the adrenal cortex and fall into one of two broad
classes, depending on their predominant mechanism of action. However, there is some
overlap in the effects of the two classes. Glucocorticoids such as cortisol (11β,17α,21-
trihydroxy-pregn-4-en-3,20-dione – Figure 5a) agonise the glucocorticoid receptor and
act to regulate inflammation and immunity as well as fat, protein and carbohydrate
metabolism (Hadley and Levine, 2006). Mineralocorticoids such as aldosterone (11β,21-
dihydroxy-3,20-dioxo-pregn-4-en-18-al – Figure 5b) agonise the mineralocorticoid
receptor and act to maintain sodium and potassium balance (Hadley and Levine, 2006).
Figure 5 – A) cortisol (a glucocorticoid), B) aldosterone (a mineralocorticoid).
12 of 331
Androgens are produced by both males and females in the adrenal glands and gonads.
However, they are secreted in much higher concentrations by male gonads.
Testosterone (17β-hydroxy-androst-4-en-3-one – Figure 6) is the most abundant
circulating androgen in males, but requires reduction in position 5 to produce the fully
active androgen, 5α-dihydrotestosterone (17β-hydroxy-5α-androstan-3-one – Figure 6).
The androgenic effects of these steroids are produced by agonising the androgen
receptor (AR) (Hadley and Levine, 2006). Androgens produce both androgenic
(masculinising) and anabolic (growth promoting) effects to varying degrees. This leads to
their more correct classification as anabolic-androgenic steroids (AASs). The androgenic
effects are the characteristic male secondary sexual features such as facial/body hair
growth and deepening of the voice, while the anabolic effects are predominantly muscle
and bone growth (Hadley and Levine, 2006). Whether or not all the anabolic effects of
AASs are mediated through the AR is currently unknown. Another possible mechanism
of action is antagonism of the glucocorticoid receptor (subject discussed further in
Figure 6 – testosterone and its conversion to the more
active androgen 5α-dihydrotestosterone.
Oestrogens such as oestradiol (estra-1,3,5(10)-triene-3,17β-diol – Figure 7) are
produced by both males and females in the adrenal glands, gonads and adipose tissue.
However, they are secreted in much higher concentrations by females during certain
stages of the ovulatory cycle (including by the corpus luteum) and during pregnancy
(Hadley and Levine, 2006). Oestrogens produce the majority of their effects through
agonism of the oestrogen receptor. This leads to breast growth and redistribution of fat
within the body, development/maintenance of the uterus before and during pregnancy,
changes to skin morphology and they are also important for bone growth (Hadley and
Levine, 2006). Oestrogens are also used as female contraceptives, which act primarily
by suppressing endogenous gonadotrophin release (Westhoff et al. 2010).
13 of 331
Figure 7 – oestradiol (an oestrogen).
A wide range of synthetic AASs, oestrogens, progestagens and corticosteroids have
been produced over the years with the aim of enhancing their pharmaceutical qualities.
Modifications that have been applied to the majority of the steroid classes include
alkylation (in order to produce more orally active versions), esterification (to prolong
duration of action), acetylation (to enhance absorption) and halogenation (to enhance
potency) (Kicman, 2008). Specifically relating to AASs, the addition of a double bond at
position 1, the attachment of a pyrazole group to the A-ring or the removal of the 19
methyl group have been employed in order to increase the anabolic to androgenic ratio
and/or to inhibit their conversion to oestrogens (Kicman, 2008). Similar modifications to
glucocorticoids have been engineered in order to try and maximise the glucocorticoid to
mineralocorticoid effect ratio. Figure 8 shows a range of different synthetic AASs.
14 of 331
Figure 8 – structures of the synthetic AASs methyltestosterone (17β-hydroxy,17α-
dien-3-one), mesterolone (17β-hydroxy,1α-methyl,5α-androstan-3-one), stanozolol (17β-
Steroid use in food production and competitive sports
The above discussion highlighted the potential anabolic effects of AASs in mammals. In
some species, however, oestrogens and progestagens may also produce anabolic
effects. In addition to effects on muscle and bone, steroids may also affect the pattern of
fat deposition within the body, leading to differential partitioning of muscle and fat;
although this depends on the steroid, species and sex of animal in question (Heitzman,
1975, Lone, 1997). Corticosteroids may produce some positive metabolic effects
following initial administration, but long-term use of high doses produces a general state
of catabolism within the body (Hadley and Levine, 2006).
Because of their potential anabolic effects, some steroids have been used to boost the
mass and quality of animal carcasses in food production for economic reasons
15 of 331
(Heitzman, 1975, Lone, 1997, Kay, 2010). Although there are a number of steroid
preparations authorised for this purpose in countries such as the USA, the use of growth
promoters (also including non-steroidal products such as the oestrogenic compound
zeranol, growth hormone and the β2-agonist class of drugs) is banned within the EU (EU
Council Directive 96/22/EC). The reasons for this ban were highlighted in two reports
from the European Commission in 1999 and 2002, which concluded that the presence of
hormones in meat products may potentially be harmful to human health through
endocrine disrupting or carcinogenic mechanisms (SCVPH, 1999, SCVPH, 2002).
However, two subsequent opinions published by the UK Veterinary Products Committee
failed to agree with the findings of the earlier European Commission’s studies (VPC,
1999, VPC, 2006). For example, the latter of these two UK reports estimated that, as a
worst case scenario, a postmenopausal woman eating a kilogram of meat (kidney)
containing the highest concentration of oestradiol detected (56 ng/kg) following
administration of the steroid would experience an increased oestrogen level of only
0.01% of average endogenous production. Indeed, it has also been speculated that the
ban may have more to do with regulating trade, leading to official disputes between the
EU and USA (Charlier and Rainelli, 2002). Nonetheless, the hormone ban remains and
non-EU countries are, therefore, required to provide sufficient animal segregation and
residue testing schemes to ensure that treated animals are not sold in the EU.
In addition to their use in food production, steroids may also be used in competitive
human and animal sports in order to improve performance. The range of steroids used
for this purpose is generally limited to the AASs. These may enhance performance
through a number of mechanisms including increased muscle mass, enhanced recovery
from training, raised red blood cell count and heightened aggression (Kicman, 2008).
Because of their potential to affect performance, the use of AASs in the majority of
horseracing, greyhound racing and human sports is prohibited (IFHA, 2008, GBGB,
2009, FEI, 2010, WADA, 2009a). Protection of the welfare of individual competitors is
another reason for prohibiting these substances; an aspect that takes increased
importance in animal sports where trainers decide on the animal’s behalf what
substances are administered.
16 of 331
The detection of steroid abuse in food production and animal sports
In order to enforce the ban on hormone use in food production, EU Council Directive
96/23/EC (and EU Commission Decision 2002/657/EC) lay down the requirements for
residue testing. Enforcement of the ban on steroid use in competitive sports is not
regulated in law in the same way as in food production, but guidelines regarding the
analytical methods that must be followed by individual laboratories when confirming
cases of steroid abuse have been produced by both the animal (AORC, 2003, ILAC-G7,
2009) and human authorities (WADA, 2009b). A comparison of the regulations used in
the food residue and sports doping control arenas can be found in Van Eenoo and
The type of matrix used for steroid residue analysis in food and sports drug surveillance
differs by a number of variables including the country, the individual authority concerned,
whether samples are taken from live animals, at slaughter or from a food import
programme and whether the analyses for a particular analyte are suited to a specific
tissue. Other than food import programmes, where analysis of meat and organs are
typically required, urine and blood are the most common matrices for testing in both the
food and sports residue arenas (Wynne, 2004, Stolker et al. 2005). However, faeces and
hair are also important matrices in some countries. When dealing with blood or hair,
detection of unchanged ‘parent’ drug is often considered suitable for determination of
drug abuse. However, when dealing with urine or faeces, a large proportion of the
excreted dose can take the form of metabolites. This is a particularly important
consideration in the case of steroids, which are typically heavily metabolised (Scarth et
al. 2009). It is, therefore, often necessary to conduct metabolism studies in order to
determine the appropriate target metabolites for the detection of steroid abuse. The
metabolism of the steroid can be broadly categorised into phases 1 and 2. Phase 1
typically involves the modification of existing functional groups within the steroid
molecule (namely oxidation, reduction, hydrolysis etc.), whereas phase 2 involves
conjugation with, typically, polar moieties such as glucuronic or sulphuric acid in order to
increase water solubility and, therefore, aid excretion (see example in Figure 9).
17 of 331
Figure 9 – examples of theoretical phase 1 and 2 metabolic pathways for the boldenone
‘pro-drug’ boldione (androsta-1,4-diene-3,17-dione - top). A possible pathway of phase 1
metabolism is reduction of the 17-keto group to form boldenone (17β-hydroxy-androsta-
1,4-dien-3-one - middle). This may be followed by phase 2 conjugation with sulphate to
form boldenone-17-sulphate (17β-hydroxy-androsta-1,4-dien-3-one-17-sulphate
Because of the common aims of food residue and sport drug surveillance laboratories,
the development and application of analytical techniques for detecting steroid abuse has
been broadly similar between the two fields over the years. Indeed, many individual
laboratories across the world are involved in residue analysis within both of these fields.
Review articles concerning the analytical methods used specifically for veterinary steroid
18 of 331
analysis can be found in Stolker et al. 2005 and De Brabander et al. 2009 (food residue
analysis) and McKinney, 2009 (equine sports). Figure 10 and the following discussion
serves as a brief overview of the scientific evolution within these two fields.
Figure 10 – summary of the evolution of steroid screening techniques used in residue
Initially, thin layer chromatography-fluorescence detection (TLC-FL) was widely used
(Moss and Rylance, 1967, De Brabander and Verbeke, 1975). Immunoassay techniques
such as enzyme-linked immunosorbent assays (ELISA) became popular during the
1980s and 1990s, but were largely replaced in the late 1990s and early 2000s by more
definitive mass spectrometric-based techniques such as gas- and liquid-chromatography
mass spectrometry (GC- and LC-MS respectively) (McKinney, 2009). The ability of many
modern GC- and LC-MS instruments to carry out MSn experiments makes them
particularly useful for identifying compounds due to their high selectivity. Also, the recent
emergence of higher resolution LC equipment allowing the use of sub-2 m particle
sizes and high flow rates (ultra-pressure liquid chromatography or UPLC) means that
metabolites with similar molecular masses and retention times can now be more easily
resolved and that analytical run times are shorter (Plumb et al. 2009).
Although there has been a general shift from GC-MS to LC-MS for drug residue analysis
during the past decade, GC-MS has remained an important tool for analysing saturated
steroid metabolites. This is because saturated steroids generally suffer from poor
ionisation properties under the atmospheric pressure ionisation conditions of LC-MS
(McKinney et al. 2009, Teale and Houghton, 2010). Although the majority of current
19 of 331
urinary screening procedures are based on detection of the ‘free’ steroid (liberated from
its phase conjugates), it may be possible in the future to design assays based on the
analysis of intact conjugates. At present, progress in this area has been limited by a lack
of availability of the relevant analytical reference standards for animal specific
metabolites. However, the potential of this approach has already been demonstrated in
human sports, where the availability of reference material has allowed the development
of LC-MS/MS assays for intact steroid conjugates (Kuuranne et al. 2002, Hintikka et al.
2008, Pozo et al. 2008).
Most recently, robust high-resolution-accurate-mass LC-MS (HR-LC-MS) systems
operating at an increased level of resolution, typically ranging between 7,500 and
100,000 full width at half maximum height (FWHM) depending on the type of mass
analyser employed, have become commercially available and have started gaining
popularity for sports drug surveillance screening and research (Virus et al. 2008, Scarth
et al. 2010). Because the data acquired are full scan analyses of intact [M+H]+ or [M-H]-
species at very high resolution, a very large number of analytes can be simultaneously
monitored. Another advantage of using HR-LC-MS includes the ability to retrospectively
analyse data once new drug information comes to light.
The powerful technique of gas chromatography combustion isotope ratio mass
spectrometry studies (GC-C-IRMS) has been applied to the confirmation of endogenous
steroid abuse (discussed further in chapter 1) in both food and sports drug residue
analysis. However, it is not currently suitable as a screening technique due to its low
throughput nature (Piper et al. 2010).
In addition to the classical analytical chemistry techniques that are targeted toward the
detection of ‘parent’ steroids or their metabolites, a number of indirect techniques have
recently gained attention (discussed further in chapter 1). These include immunoassay
and receptor based biosensor assays as well as a range of ‘omics’ biomarker
approaches such as metabolomics, proteomics and transcriptomics (Scarth et al. 2006).
Because these techniques are targeted toward pharmacological activity rather than
individual drug structure, they produce complementary screening data that can be used
to indicate whether steroid abuse may have occurred. However, these techniques have
yet to find widespread application in the confirmation of steroid abuse, which is typically
still achieved by the direct measurement of a steroid or its metabolite.
20 of 331
While the above discussion served to summarise the range of instrumental techniques
that may be used to detect steroid abuse, the ability to detect the abuse of each
individual steroid is determined by a number of factors. Firstly, the sample needs to be
taken from the animal at a time close enough to the point of steroid administration for the
concentrations to be above the limits of instrumental sensitivity. If the sample is taken too
long after steroid administration, then traces of the drug may be too low to be detected.
Also, each individual steroid can be classified into one of three broad categories, which
impacts on the ability to detect the abuse of each:
‘Exogenous’ steroids are known marketed ‘classical’ steroids, such as stanozolol.
These contain synthetic structures that are thought not to occur naturally. Detection of
this class of steroid is relatively straightforward since a purely qualitative demonstration
of the presence of these synthetic steroids is all that is required in order to determine
‘Endogenous’ steroids are also known marketed steroids, such as testosterone, but
contain structures that are known to exist naturally. Detection of the abuse of
‘endogenous’ steroids is more complicated because they are, by definition, ‘natural’ to
some extent and so a simple qualitative demonstration of their presence is insufficient to
indicate abuse (discussed further in chapter 1). Some endogenous steroids such as
testosterone, progesterone and oestradiol are known to be ubiquitous amongst
mammals. However, the classification of a steroid as ‘endogenous’ is a grey area and
there are some steroids that may be considered ‘semi’-endogenous. This term signifies
that the steroid in question has been suspected to be endogenous, but only in certain
situations i.e. in a specific species or at particular time. Analytical sensitivities for
detecting steroids have increased significantly over the years, which has resulted in
more and more compounds being suspected as ‘endogenous’ or ‘semi-endogenous’ at
‘Designer’ steroids are previously unmarketed steroids that contain synthetic structures
that are thought not to occur naturally. The use of a designer steroid first came to the
public attention in 2003 when a syringe containing the novel steroid tetrahydrogestrinone
(THG) was handed to doping officials. This resulted in the disqualification of several
athletes after they were subsequently found to have used the steroid (Catlin et al. 2004).
Designer steroids have chemical structures based on previously marketed products, but
21 of 331
with minor modifications which make them undetectable by the majority of current
targeted mass spectrometric procedures. In the case of THG for example, the drug’s
structure is based upon gestrinone, but with the 17-alkyl side chain fully saturated such
that the relative molecular mass of THG is 4 atomic mass units (u) higher than
gestrinone. Designer steroids are synthesized either to deliberately evade detection, or
as appears more common, to enable them to be marketed freely on the Internet to
customers in some countries because their structures do not fall within the scope of legal
regulations that prevent the sale of defined steroidal products.
Detection of the abuse of the latter two classes of steroids in food production and animal
sports is very challenging. The development of analytical approaches to tackle these
associated issues forms the basis of the current thesis.
Aims and objectives of the current work
Within the author’s laboratory, analytical methods are already available to detect the
abuse of the majority of endogenous AASs in horseracing. However, the same is not true
for the majority of endogenous steroids in other food producing animals. ‘Designer’
steroids could in theory be abused in both horseracing and food production, but at
present the majority of work on designer steroids has been commissioned by the sports
regulatory authorities. One reason that this class of compounds has received more
attention in sports doping control is because of the proven use of designer steroids such
as THG by a number of athletes. Another factor may relate to the fact that human and
animal sports typically involve single individuals looking to gain marginal advantages for
significant financial and/or sociological gain, whereas food production involves large
herds of animals with smaller financial return relative to the risk. These differences in
return relative to risk could, therefore, be considered to make the abuse of relatively
expensive ‘exotic’ treatments such as designer steroids more likely in competitive sports
compared to food production. In light of the aforementioned discussion, the following
were determined as the overall aims and objectives of the research reported herein:
Overall aims: To develop novel analytical approaches for the detection of ‘endogenous’
steroid abuse in food-production (part 1) and of ‘designer’ steroid abuse in animal sports
(part 2). The primary focus of this thesis will relate to AASs such as nandrolone,
boldenone, testosterone and their synthetic ‘designer’ analogues. However, chapter 3
22 of 331
will also consider the important natural steroids oestradiol and progesterone in relation to
the detection of their abuse in the bovine.
Part 1 objectives:
- To review the literature regarding endogenous steroids and their detection
in food production, focussing mainly on endogenous AASs (chapter 1).
- To develop approaches for the detection of nandrolone abuse in the
porcine (chapter 2).
- To develop approaches for the detection of androgen, oestrogen and
progestagen abuse in the bovine (chapter 3).
Part 2 objectives:
- To review the literature regarding steroid metabolism and detection in the
equine (especially in relation to designer steroids) and to compare the
trends with those observed in other species (chapter 4).
- To develop and assess the suitability of in vitro techniques for conducting
equine drug metabolism studies (chapter 5).
- To use the newly developed in vitro methods to study the metabolism of a
novel ‘designer’ steroid in the equine (chapter 6).
Association of Official racing Chemists (AORC) (2003). Guidelines for the Minimum
Criteria for Identification by Chromatography and Mass Spectrometry. Accessed from:
http://www.aorc-online.org/AORC MS Criteria.pdf on 24th July 2010
Catlin, D. H., Sekera, M. H., Ahrens, B. D., Starcevic, B., Chang, Y. and Hatton, C. K.
(2004). Tetrahydrogestrinone: Discovery, synthesis and detection in urine. Rapid
Communications in Mass Spectrometry, 18(12), 1245-1249.
Charlier, C. and Rainelli, M. (2002). Hormones, risk management, precaution and
protectionism: An analysis of the dispute on hormone-treated beef between the
European Union and the United States. European Journal of Law and Economics, 14(2),
23 of 331
De Brabander, H. F. and Verbeke, R. (1975). Detection of antithyroid residues in meat
and some organs of slaughtered animals. Journal of Chromatography A, 108(1), 141-
De Brabander, H. F., Noppe, H., Verheyden, K., Vanden Bussche, J., Wille, K.,
Okerman, L., Vanhaecke, L., Reybroeck, W., Ooghe, S. and Croubels, S. (2009).
Residue analysis: Future trends from a historical perspective. Journal of
Chromatography A, 1216(46), 7964-7976.
EU Council Directive 96/22/EC. Official Journal of the European Union. L125
23/05/1996. Council Directive 96/22/EC of 29 April 1996 concerning the prohibition on
the use in stockfarming of certain substances having a hormonal or thyrostatic action
and of beta-agonists and replacing Directives 81/602/EEC, 88/146/EEC and
88/299/EEC. Pp. 3-9, Brussels, Belgium, 1996.
EU Council Directive 96/23/EC. Official Journal of the European Union, L125,
23/05/1996. Council 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
Directives 85/358/EEC and 86/469/EEC and Decision 89/187/EEC and 96/664/EEC. Pp.
10-32, Brussels, Belgium, 1996.
EU Commission Decision 2002/57/EC. Official Journal of the European Communities,
L221. Commission Decision (2002/57/EC) of 12 August 2002. Pp. 8-36. Brussels,
FEI (2010). Federation Equestre Internationale (FEI) prohibited substances list.
Accessed from www.feicleansport.org/ProhibitedSubstancesList_Jan2010.pdf on 12th
Greyhound Board of Great Britain (GBGB) (2009). Rules of Racing. Rule 217.
Hadley, M. E. and Levine, J. (2006). Endocrinology. 6th edition. Prentice Hall. ISBN:
Heitzman, R. J. (1975). The effectiveness of anabolic agents in increasing rate of growth
in farm animals; report on experiments in cattle. Environmental Quality and Safety.
Suppl. Vol. 5, 89-98.
Hintikka, L., Kuuranne, T., Leinonen, A., Thevis, M., Schänzer, W., Halket, J., Cowan,
D., Grosse, J., Hemmersbach, P., Neilen, M. W. F. and Kostiainen, R. (2008). Liquid
chromatographic-mass spectrometric analysis of glucuronide- conjugated anabolic
steroid metabolites: Method validation and interlaboratory comparison. Journal of Mass
Spectrometry, 43(7), 965-973.
International Laboratory Accreditation Cooperation-G7 (ILAC-G7). Accreditation
Requirements and Operating Criteria
www.ilac.org/documents/ILAC_G7_06_2009.pdf (accessed 19th October 2009).
International Federation of Horseracing Authorities (IFHA). (2008). International
Agreement on Breeding, Racing and Wagering. Article 6.
for Horseracing Laboratories.
24 of 331
http://www.chem.qmul.ac.uk/iupac/steroid/ on 17th July 2010.
Kay, J. (ed). (2010). Analyses for Hormonal Substances in Food-Producing Animals.
Kicman, A. T. (2008). Pharmacology of anabolic steroids. British Journal of
Pharmacology, 154(3), 502-521.
Kuuranne, T., Aitio, O., Vahermo, M., Elovaara, E. and Kostiainen, R. (2002). Enzyme-
assisted synthesis and structure characterization of glucuronide conjugates of
methyltestosterone (17α-methylandrost-4-en-17β-ol-3-one) and nandrolone (estr-4-en-
17β-ol-3-one) metabolites. Bioconjugate Chemistry, 13(2), 194-199.
Lone, K. P. (1997). Natural sex steroids and their xenobiotic analogs in animal
production: Growth, carcass quality, pharmacokinetics, metabolism, mode of action,
residues, methods and epidemiology. Critical Reviews in Food Science and Nutrition,
Makin, H. L. J., Gower, D. B. and Kirk, D. N. (Eds). (1995). 1st edition. Steroid Analysis.
Blackie Academic and Professional. ISBN: 0-7514-0128-5.
McKinney, A. R. (2009). Modern techniques for the determination of anabolic-androgenic
steroid doping in the horse. Bioanalysis. 1(4), 785-803.
Moss, M. S. and Rylance, H. J. (1967). A thin-layer chromatography study on the
metabolism of prednisolone in the horse. Journal of Endocrinology, 37(2), 129-137.
Piper, T., Geyer, H., Gougoulidis, V., Flenker, U. and Schänzer, W. (2010).
Determination of 13C/12C ratios of urinary excreted boldenone and its main metabolite
5β-androst-1-en-17β-ol-3-one. Drug Testing and Analysis, 2(5), 217-224.
Poetker, D. M. and Reh, D. D. (2010). A comprehensive review of the adverse effects of
systemic corticosteroids. Otolaryngologic Clinics of North America, 43(4), 753-768.
Pozo, O. J., Van Eenoo, P., Van Thuyne, W., Deventer, K. and Delbeke, F. T. (2008).
Direct quantification of steroid glucuronides in human urine by liquid chromatography-
electrospray tandem mass spectrometry. Journal of Chromatography A, 1183(1-2), 108-
Plumb, R. S., Rainville, P. D., Potts III, W. B., Johnson, K. A., Gika, E. and Wilson, I. D.
(2009). Application of ultra performance liquid chromatography-mass spectrometry to
profiling rat and dog bile. Journal of Proteome Research, 8(5), 2495-2500.
Scarth, J., Roberts, J., Teale, P. and Pleasance, S. (2006). Keeping London 2012 clean:
The fight against doping in sport. Biologist, 53(6), 305-310.
Scarth, J., Akre, C., Van Ginkel, L., Le Bizec, B., De Brabander, H., Korth, W., Points, J.,
Teale, P. and Kay, J. The presence and metabolism of endogenous androgenic-anabolic
steroid hormones in meat producing animals. A review. (2009). Food Additives and
Contaminants: part A. Vol. 26(5), 640-671.
(2010). The nomenclature of steroids. Accessed from
25 of 331
Scarth, J., Spencer, H., Timbers, S., Hudson, S. and Hillyer, L. (2010). The use of in vitro
technologies coupled with high-resolution/accurate-mass LC-MS for studying drug
metabolism in equine drug surveillance. Drug Testing and Analysis. 2(1): 1-10.
Scientific Committee on Veterinary Measures Relating to Public Health (SCVPH). (1999).
Assessment of potential risks to human health from hormone residues in bovine meat
and meat products. Download from http://ec.europa.eu/food/fs/sc/scv/out21_en.pdf on
21st September 2010.
Scientific Committee on Veterinary Measures Relating to Public Health (SCVPH). (2002).
Review of previous SCVPH opinions of 30 April 1999 and 3 May 2000 on the potential
risks to human health from hormone residues in bovine meat and meat products.
Download from http://ec.europa.eu/food/fs/sc/scv/out50_en.pdf on 21st September 2010.
Stolker, A.A. M. and Brinkman, U. A. Th. (2005). Analytical strategies for analysis of
veterinary drugs ang growth-promoting agents in food-producing animals – a review.
Journal of Chromatography A. 1067 (1-2), 15-33.
Teale P and Houghton E. (2010). Metabolism of anabolic steroids and their relevance to
drug detection in horseracing. Bioanalysis. 2(6), 1085-1107.
Van Eenoo, P. and Delbeke, F. T. (2004). Criteria in chromatography and mass
spectrometry - A comparison between regulations in the field of residue and doping
analysis. Chromatographia, 59(SUPPL.), S39-S44.
Veterinary Products Committee (VPC). (1999). Executive summary and critical
evaluation of the scientific reasoning and methods of argument adopted in the opinion of
the Scientific Committee on Veterinary Measures Relating to Public Health which
assessed the potential risks to human health from hormone residues in bovine meat and
meat products. Download from http://www.vmd.gov.uk/vpc/Working/finalrep.pdf on 21st
Veterinary Products Committee (VPC). (2006). Risks Associated with the Use of
Hormonal Substances in Food-Producing Animals. Report of the Veterinary Products
Committee. Download from http://www.vmd.gov.uk/vpc/Working/Hormones_report.pdf on
21st September 2010.
Virus, E. D., Sobolevsky, T. G. and Rodchenkov, G. M. (2008). Introduction of
HPLC/orbitrap mass spectrometry as screening method for doping control. Journal of
Mass Spectrometry, 43(7), 949-957.
World Anti-Doping Agency (WADA). (2009a). The 2010 Prohibited List International
Standard (the World Anti-Doping Code). World Anti-Doping Agency, Clause S1.
Accessed from http://www.wada-ama.org/Documents/World_Anti-
Doping_Program/WADP-Prohibited-list/WADA_Prohibited_List_2010_EN.pdf on 24th
WADA. (2009b). WADA international standard for laboratories. Accessed from
list/WADA_Prohibited_List_2010_EN.pdf on 24th July 2010.
26 of 331
Westhoff, C. L., Torgal, A. H., Mayeda, E. R., Pike, M. C. and Stanczyk, F. Z. (2010).
Pharmacokinetics of a combined oral contraceptive in obese and normal-weight women.
Contraception, 81(6), 474-480
Wynne, P. M., Batty, D. C., Vine, J. H. and Simpson, N. J. K. (2004). Approaches to the
solid-phase extraction of equine urine. Chromatographia, 59(SUPPL.), S51-S60.
PART 1: DETECTION OF ‘ENDOGENOUS’ STEROID ABUSE IN
29 of 331
Chapter 1: Presence, metabolism and detection of
‘endogenous’ steroid hormones in food producing
Scarth, J., Akre, C., Van Ginkel, L., Le Bizec, B., De Brabander, H., Korth, W., Points,
J., Teale, P. and Kay, J. The presence and metabolism of endogenous androgenic-
anabolic steroid hormones in meat producing animals. A review. (2009). Food Additives
and Contaminants: part A. Vol. 26(5), 640-671.
Scarth, J. and Akre, C. Book chapter - Presence and metabolism of endogenous steroid
hormones in meat producing animals. (2010) In: Analyses for Hormonal Substances in
Food-Producing Animals. Pg. 48-96. Ed. J. Kay. ISBN:978-0-85404-198-5.
Scarth, J., Teale, P. and Kay, J. Presence and metabolism of natural steroids in cattle,
sheep, swine, horse, deer and goat: current knowledge and potential strategies for
detecting their abuse. (2008). Pg. 1211-1215. Proceedings of Euroresidue VI.
As discussed in the introductory chapter, EU Council Directive 96/22/EC of 1996 states
that “substances having a hormonal action” are prohibited for use in animals intended for
meat production. As well as purely novel steroids not existing in nature, the directive also
covers synthetically produced versions of steroids that are known to occur naturally in
certain species under particular circumstances. However, in some countries, including
the USA, Canada and Australia, some (combinations of) steroids and a related synthetic
compound Zeranol are officially registered for use as hormonal growth promoting
compounds. Due to their anabolic and/or partitioning effect they increase the profit per
unit head for the farmer. EU Council Directive 96/23/EC (and EU Commission Decision
30 of 331
2002/657/EC) lays down the requirements for residue testing in order to ensure
compliance with the EU prohibition.
The steroid hormones considered in this review chapter are the androgenic-anabolic
steroids (AASs) that potentially derive from precursors within the body such as
cholesterol and pregnenolone (see figure 1). These include testosterone,
androstenedione, nandrolone, boldenone and dehydroepiandrosterone (DHEA), as well
as their numerous catabolic products and any precursor compounds that might
potentially lead to conversion to these steroids within the body. The major focus of this
chapter will relate to phase 1 steroid metabolites because there is much less information
relating to phase 2 metabolism. However, details of the phase 2 metabolism of some
steroids will be given where they have been shown to usefully distinguish situations of
abuse (for example boldenone). The task of detecting the abuse of synthetically
produced hormones that are also known to be endogenous under certain conditions,
dubbed ‘pseudo-endogenous’ or ‘grey zone substances’ due to their dual
synthetic/endogenous nature (Van Thuyne Wim 2006), is problematic for many reasons.
The most significant challenge arises due to the fact that when they are shown to occur
naturally within a particular type of animal, a simple qualitative demonstration of their
presence does not necessarily prove abuse. Most, but not all, steroid preparations are
ester versions of these potentially endogenous steroids. However, a simple
demonstration of the presence of the steroid ester as proof of abuse is not always
possible (with the exception of hair and injection/implant sites in some cases) due to a
large proportion of the steroid ester being cleaved by the time it reaches the test matrix
i.e. plasma or urine. Some type of quantitative uni- or multi-variate threshold approach is
therefore usually required in order to confirm abuse. Furthermore, as analytical limits of
detection decrease, the list of compounds that are suspected to be endogenous at low
concentrations increases. These and some further analytical and physiological
considerations are taken up again later in this review.
Analytical methods of various kinds have in the past been employed to identify and
quantify endogenous steroids, their metabolites and precursors, but their effectiveness
and the harmonisation of their application in different countries and situations is
questionable. For example, Van Ginkel et al. 1993 highlight the wide range of different
analytical methods and thresholds that have been applied in different EU countries in the
past. Since the author was aware of no comprehensive published review on the
Review: endogenous steroids in food-producing animals
31 of 331
concentrations and metabolism of such steroids in food producing animals, the overall
aim of the work reported herein was therefore to carry out a survey of the existing
literature. This then guided further practical work in order to increase knowledge and to
develop more effective testing methods (Chapters 2 and 3).
Figure 1 – schematic of the biosynthetic pathways for endogenous steroids in mammalian species. Many of the reactions that involve
oxidation or reduction of hydroxyl and ketone groups respectively are reversible. Wavy arrows indicate a putative pathway only.
Review: endogenous steroids in food-producing animals
33 of 331
1.2 Literature survey methods
The overall aim was to collect as much published and unpublished data as possible in
order to provide for the most comprehensive evaluation of the field. This review was
originally published in the journal Food Additives and Contaminants in June 2009.
However, studies reported after this date (up to October 2010) have now also been
reviewed and are discussed in the text. Literature searches were conducted using the
Pubmed facility of the USA National Centre for Biotechnology Information (NCBI)
(www.ncbi.nlm.nih.gov/), Scopus (www.scopus.com), the Web of Science
(www.scientific.thomson.com/products/wos/) and various ‘grey’ literature sources. Many
researchers were also contacted so that as much data as possible from individual
animals and particularly for values below the reported LODs, decision limit (CCα) or
detection capability (CCβ) could be obtained. This had the advantage that some
appreciation of concentrations could be obtained where analyte concentrations were
currently too challenging for fully rigorous quantitative analysis.
In the following sections, the occurrence of precursors and metabolites of testosterone,
nandrolone and boldenone in bovine, porcine and ovine matrices is reviewed in narrative
fashion. Also reviewed are precursors and metabolites of nandrolone and boldenone in
equine, cervine and caprine matrices. Ideally, a statistical analysis of results using a
meta-analysis (defined here as “the statistical analysis of a large collection of analysis
results for the purpose of integrating the findings” – from Glass, 1976) would be a
desirable outcome. However, due to a lack of sufficient data this approach could not be
used. Where differences between results in this review are stated to be “statistically
significant,” this refers to comparisons of controlled populations within a single study and
not between results of different studies. Unless otherwise stated, the results reviewed
derive from controlled studies where the use of banned steroids can be ruled out.
As many relevant matrices as possible have been considered in this review. However,
due to the magnitude of the literature and the overall scope of this review being
predominantly targeted at control of abuse rather than the safety implications, there is an
inevitable bias in the output toward plasma, urine, bile, faeces and hair over tissues such
as muscle and fat. Also, longitudinal studies using solid tissues are not often possible
because this usually means slaughter of the animal (hence only one sample can be
taken). In addition to the individual studies reviewed herein, several reviews dealing with
34 of 331
the concentrations of steroid hormones in different food products are also available
(Velle 1976, Henricks et al. 1976, Hartmann et al. 1998, Arnold 2000, Daxenberger et al.
2001, Stephany et al. 2004, Fritsche et al. 1999 and Mouw et al. 2006).
1.3 Physiological and analytical considerations regarding comparisons of
steroid concentrations within and between different species
A basic understanding of the analytical and physiological context of natural steroids is
assumed in this review. Nevertheless, some information of specific relevance is given
below and the biosynthetic and catabolic pathways of some representative natural
steroids are summarised in Figures 1 and 2 and Table 4. Further background information
on general analytical aspects can be found in Makin et al. 1995 and Stolker et al. 2005
while further physiological information can be found in Mason et al. 2002 and Hadley and
Although the background given here is separated into analytical and physiological
factors, there are areas of overlap between the two. A critical theme that will become
apparent is that the lack of reporting of sufficient method details (at least in a standard
format) often means that rigorous quantitative comparisons between different studies are
not possible. It was also necessary to limit the number of parameters chosen for study.
The remaining analytical parameters subject to full analysis were chosen by
consideration of a combination of their impact on any results as well as the frequency
and reliability of their reporting.
1.3.1 Analytical factors
Although most published methods rely on direct identification and/or quantification of
analytes, indirect approaches utilising biosensors, biomarkers, gas chromatography-
combustion-isotope ratio-mass spectrometry (GC-C-IRMS) or the detection of intact
steroid esters have also been investigated (discussed further in later sections). For the
purposes of the main body of this review chapter however, studies were limited to direct
detection/quantification using such techniques as immunoassay (IA), high performance-
liquid chromatography with ultraviolet detection (HPLC-UV) and liquid or gas
chromatography coupled to mass spectrometry (LC- and GC-MS respectively).
Figure 2 – schematic representation of the phase 1 metabolism of nandrolone in the three species for which most information is
available (bovine, porcine and equine). Sites of possible metabolic epimerisation are highlighted (*). For space purposes, and due to
the number of potential isomers, it is only possible to depict the major metabolites of nandrolone. For the same reasons it is also not
possible to give schematic metabolic pathway representations for all the individual steroids covered in this review. However, it is
worth noting that the same functional groups that are subject to metabolism in nandrolone are also liable to metabolism in other
36 of 331
When comparing data between studies, it becomes apparent that while ‘true’ differences
between data points and populations do exist, that variation can also be caused by
biases in sampling designs or the type of analysis used. In many cases, comparison of
data is further complicated by the reporting of different types of information i.e. LOD or
CCα/β are often not reported. Some examples of analytical aspects that can lead to
variation within the data are given below:
● Qualitative, semi-quantitative and fully quantitative data – While the ultimate aim of this
review was to consider concentrations of natural steroids in a quantitative fashion, it was
also recognised that a number of useful studies only reported data in a qualitative or
semi-quantitative fashion. While the results of these analyses were not subject to any
statistical analysis, they were considered useful in answering certain qualitatively
focussed questions e.g. does boldenone occur naturally at any concentration in species
X? Where qualitative or semi-quantitative data are analysed, this will be highlighted and
any assumptions stated. Even with studies that are reported to be ‘quantitative’ it is
important to understand that all data has a degree of uncertainty attached. However, tor
the majority of studies reviewed herein, insufficient validation data were available to fully
assess the degree of uncertainty of the results.
● Method of calibration line construction – When dealing with endogenous substances,
quantification can sometimes be complicated by the difficulty of finding a true blank
matrix. In cases where a blank matrix of the same type as the study samples is not
available then one can either use standard addition; where known amounts of steroids
are added ‘on top’ of the existing concentrations present, or alternatively a surrogate
matrix can be used. If using a surrogate matrix devoid of endogenous steroid, then
appropriate measures need to be taken to ensure the chosen matrix behaves in a similar
way to the actual sample matrix in order to control for any variation in the analytical
procedure. Neither of the two aforementioned measures is perfect and each can lead to
different reported concentrations for the same data set due to differential matrix effects
or recovery of analyte. In many published reports, the actual calibration range applied
was not explicitly given. This made it difficult to evaluate whether individual results fell
within a linear range. Due to the need to limit the number of factors that were being taken
into account in this review, adjustment of analytical data for recovery and matrix effects
was not attempted. In any case, many of the aforementioned parameters were not
always reported by authors.
Review: endogenous steroids in food-producing animals
37 of 331
● Limit of Detection (LOD) or CCα/β – For some steroids in certain physiological
situations, a large number of reported concentrations are ‘not detectable’ (ND). This
causes problems for two main reasons, one statistical and the other regulatory. From a
statistical angle, the existence of large numbers of concentrations below the LOD are
problematic because these values have to be effectively treated as zero, making it more
difficult to define the population distribution and hence set threshold levels. If the LOD is
not reported at all then this further complicates meta-analyses. From a regulatory point of
view, because sensitivities generally increase as technologies improve, this sometimes
means that steroids once thought to be purely synthetic appear to exist naturally at very
low levels. However, it is difficult to assess whether new clusters of positive findings at
such low levels are due to an increased abuse of steroids or natural occurrence. The
issue in this context is therefore not the LOD per se, but the LOD in the physiological,
analytical and regulatory context. Where possible, authors of published works were
contacted and information on LODs or CCα/βs was requested. Also requested (where
relevant) were any results that were quantified, but which were below the LOD or CCα.
In these cases, an estimation of the reliability of the additional trace concentration data
was also requested.
● Sample collection and subsequent preparation technique – Prior to analysis, most
techniques require some degree of sample preparation, which typically involves
extraction of the analytes of interest from unwanted or interfering matrix components.
The treatment of the sample once taken from the animal can influence the analytical
results in several ways, all of which highlight the need to stabilise samples appropriately
and to take into account any artefactual processes occurring prior to analysis. For
example, it is known that a number of meat producing species, e.g. bovine, ovine and
equine, but not porcine, have a propensity to convert 17β-hydroxy or ketone functions
into 17α-hydroxy compounds. (Gaiani et al. 1984). Bovine plasma in particular is known
to be especially active at catalysing this reaction and the addition of methanol to the
matrix has been shown to inhibit the activity (Gaiani et al. 1984).
It has also been shown that the new-born of the ovine, caprine and bovine display very
high rates of 20α-hydroxysteroid dehydrogenase activity (acting on progestagens and
corticosteroids) and that this activity diminishes rapidly with age; possibly due to the
replacement of fetal with adult erythrocytes (Nancarrow et al. 1983). Due to a general
dearth of knowledge on the metabolism of steroids in caprine, and ovine species, the
38 of 331
significance of this finding for surveillance of steroid abuse is currently difficult to assess.
However, for all the species concerned, the data suggests that it is important to choose
the correct age of the animals from a reference population that is used for establishing
thresholds. This is so that the ages reflect those of the animals likely to be encountered
in routine surveillance programmes.
Bovine faeces are known to be capable of producing boldenone and other 1-dehydro
steroids as metabolites from some steroidal precursors ex-vivo (Pompa et al. 2006). It is
therefore recommended that sampling of bovine urine be devoid of faecal contamination
in order to avoid boldenone false positives (De Brabander et al. 2004).
It has also been shown that nandrolone related compounds can be formed from
testosterone derivatives in human urine. The authors of this work showed that this
reaction can be partially stabilised by adding EDTA to the samples (Grosse et al. 2005).
Many steroids can also be conjugated with polar moieties such as sulphuric and/or
glucuronic acid. Samples are often hydrolysed prior to analysis in order to produce the
‘free’ steroid. Hydrolysis can be performed before or after preliminary extraction or group
separation and even then can be performed by a variety of methods. Helix pomatia
digestive juice is the most often applied enzymatic form of deconjugation and this
method affords hydrolysis of glucuronic acid conjugates and aryl sulphates at optimum
pH. However, it is also known to contain hydroxylase and oxidoreductase enzyme
activity that can artefactally oxidise or reduce some steroids (Houghton et al. 1992).
Another preparation that is frequently used is the β-glucuronidase enzyme from E. coli,
which as its name suggests cleaves glucuronic acid conjugates but not sulphate
conjugates (Houghton et al. 1992). In a ‘two fraction’ extraction, glucuronic acid
conjugates may be cleaved by enzymes from extracts of Helix pomatia, while sulphate
conjugates can be cleaved using acidified ethyl acetate:methanol (termed solvolysis; as
reported in Teale and Houghton, 1991). An alternative is to cleave both types of
conjugates simultaneously using acidified methanol (termed methanolysis; as reported in
Tang and Crone, 1989), but this can lead to more complex mixture of components
retained within extracts (James Scarth, personal observation). The use of a number of
different hydrolysis (or no hydrolysis at all) steps in the literature, all with varying
capacities to deconjugate steroids, is another factor that potentially leads to variation in
the reported concentrations.
Review: endogenous steroids in food-producing animals
39 of 331
A mixture of purification/concentration approaches were identified in the literature
including solid phase extraction, liquid-liquid extraction, protein precipitation,
immunoaffinity column chromatography, supercritical fluid extraction, accelerated solvent
extraction (ASE) and some very elaborate, but often effective, multi-step HPLC
fractionation processes. Results using these methods are generally not compared in this
review, unless there was specific relevance to a result.
● Type of analytical method used – A major factor leading to variation between reported
values lies in the type of end-point detection method used. These included (in
approximate descending order of reported use), immunoassay (IA), gas-
chromatography-mass spectrometry (GC-MS), liquid-chromatography-mass
spectrometry (LC-MS), high-performance-liquid chromatography-ultra-violet detection
(HPLC-UV) and thin-layer-chromatography-fluorescence detection (TLC-FL).
Immunoassay and mass spectrometry techniques generally afford higher sensitivity over
HPLC-UV or and TLC-FL and are also generally more selective. Mass spectrometry is
considered to offer more selectivity than IA, predominantly due to variable extents of
cross-reactivity of steroids against the IA antibody, although the impact of any cross-
reactivity can be reduced by performing HPLC separation of sample extracts prior to
analysis. Although generally considered very selective, mass spectrometry is still subject
to matrix effects such as ion suppression or enhancement (LC-MS generally more so
than GC-MS). However, these can usually be overcome through the use of matrix
matched standards (where available). As a general rule, it has been observed that IA
tends to overestimate oestrogen levels at low concentrations while underestimating them
at high concentrations (Stephany et al. 2004).
● Statistical analyses used within the studies reviewed –. Depending on a number of
factors, including the steroid, species, matrix and analytical LOD, a Gaussian distribution
of steroid concentration population data may or may not be determined. In this respect, a
large number of parametric and non-parametric approaches were reported by authors,
reflecting the different findings under varying conditions. With such major differences in
statistical reporting, such as mean vs median or standard deviation vs inter-quartile
range, it is very difficult to make quantitative comparisons between data sets. It is also
important to highlight a major difference between a statistical method being able to
discriminate a control from a steroid treated population (i.e. a T-test result) and a
statistical method that allows a workable threshold to be calculated (i.e. allowing a
40 of 331
degree of certainty that at a particular threshold a false positive will not occur). There
can be a significant amount of overlap in individual steroid concentrations from control
and treated steroid populations that can be discerned using a T-test, but this does not
necessarily mean they are significant enough differences to allow a realistic threshold to
It is also important to add that the uncertainty of measurement was very rarely reported
in the studies reviewed and was not easily calculated from the data available, further
adding to the difficulty in making quantitative comparisons between data sets.
1.3.2 Physiological factors
Some of the physiological considerations regarding steroid concentrations were given in
the analytical section above. In addition to inter-individual differences, there are many
further factors that lead to variation in the observed concentrations. For example,
Challenger (2004) has reviewed the peak ovarian cycle plasma/serum oestradiol and
progesterone concentrations in mammalian species. It was found that oestradiol
concentrations spanned around four orders of magnitude while those for progesterone
spanned three orders of magnitude. Oestradiol concentrations were on average two
orders of magnitude lower than progesterone concentrations and there were significant
differences between different animal orders. Maximum oestradiol concentrations were
more variable in artiodactyls and primates than in carnivores. Absolute oestradiol
concentrations were not correlated with dietary niche, but the progesterone to oestradiol
ratio was lower in artiodactyls and primates compared with carnivores. Although this
study refers to oestrogens and progestagens rather than androgens (a comparable study
for androgens could not be found by the author), it highlights the significant differences in
steroid concentrations between species and identifies the need to obtain endogenous
population data for hormones in each species before detection strategies for regulatory
surveillance are devised. As many references in this review will demonstrate, intra- and
inter-species genetic variation may be responsible for a large proportion of the observed
variation between animals.
● 4- vs 5-ene pathways – As well as the absolute differences in oestradiol and
progesterone described above, species are also known to vary in their utilisation of the
4- and 5-ene pathways for the production of steroids; which can be traced back to
Review: endogenous steroids in food-producing animals
41 of 331
differences in the substrate requirements of the CYP17 enzyme (Mason et al. 2002).
This means that some species produce more steroid precursors with a 4-ene group (e.g.
androstenedione) whereas others produce more with a 5-ene group (e.g. DHEA
[dehydroepiandrosterone]). Of relevance to meat producing animals, 5-ene precursors
are relatively high in bovine, porcine, ovine and equine species, while the cervine is
lower in 5-ene and higher in 4-ene steroids (Wichman et al. 1984).
● Pregnancy and pseudopregnancy – It is well known that pregnancy can lead to
extremely high concentrations of certain relevant steroids. Pregnant animals are
therefore usually excluded from threshold value calculations. However, a phenomenon
termed pseudo-pregnancy (also known as phantom pregnancy or pseudocyesis) also
exists. In some species this condition leads to the physiological appearance of a state of
pregnancy (including raised steroid concentrations), but without an actual fetus being
conceived (Johnson and Everitt 2000). The effect is certainly frequent in rodent and
canine species, but some references to its occurrence in the porcine (Pusateri et al.
1996), caprine (Lopes Junior et al. 2002) and ovine-caprine hybrids (Maclaren et al.
1993) were also obtained. While the condition does seem to occur naturally at a high
incidence in some caprine species, the porcine reports were of artificially induced
pseudo pregnancy by administering oestradiol. No reports of pseudopregnancy in bovine
species could be found in the published literature.
● Oestrous synchronisation – The effects of oestrous synchronisation devices are not
covered in this survey, but the subject has received comprehensive review in Rathbone
● Route of excretion – Endogenous and artificially administered steroids are
predominantly excreted from the body via the urine and faeces. The excretion of steroids
is species and compound dependent, with some species preferentially excreting in
faeces and some in urine. Consideration of whether urine, bile or faeces are the most
suitable choices for a particular steroid/species combination depends on a number of
factors (taken up later in this review), but their relative excretion in the form of recovered
radioactivity in urine versus faeces is one consideration. Although an important factor for
consideration, a predominance of radioactivity in one or other matrices does not always
imply greater suitability for that matrix since a smaller proportion of radioactivity present
as one analyte may be more useful than a larger proportion of radioactivity present as
42 of 331
many metabolites. Differences in the total volume of excreted material can influence
resulting concentrations. On the whole, urine generally suffers less analytical matrix
effects and residual ex-vivo metabolism than faeces. Figure 3 exemplifies the range of
different excretion patterns that have been observed for some steroids.
Figure 3 – Percentage excretion of radioactivity in different waste products after
intravenous infusion of testosterone into different species (adapted from Martin 1966,
Calvert et al. 1975, Velle 1976 and Palme et al. 1996).
● Hydration status– The concentrations of steroids in some matrices, especially urine,
can be affected by the hydration status of the animal (Wolfgang Korth – personal
observation). One could predict that this might be a particularly important factor in
countries that have experienced frequent droughts in recent years, for example Australia.
The adjustment of urinary steroid concentrations for the hydration status of the animal
(often measured as the specific gravity or the creatinine concentration of the urine)
therefore has potential to reduce the variation in steroid values among the population.
Like many physiological variables, it is also possible that dehydration may be a stressor
that affects minor metabolic pathways such as the rate of biosynthesis/catabolism of
steroids. However, the author is not aware of any studies that have assessed this
Review: endogenous steroids in food-producing animals
43 of 331
● Other variables – Many other physiological variables can affect the concentrations of
steroids in different animals. Previously proposed regulatory thresholds for natural
steroids in meat producing species (i.e Scippo et al. 1993, Arts et al. 1991) have taken
into account at least the age and sex of the animal when constructing thresholds. In the
current review, some of the factors that were analysed include the steroid in question,
matrix, age, sex, herd demographics, gestation and castration status, geographical
factors, housing conditions, season and time of day, disease, stress, medication,
housing conditions, diet and breed.
1.4 Natural androgenic-anabolic steroid concentrations in the bovine
1.4.1 General trends in the data
Figures 4 and 5 summarise the different analyte/matrix and analytical technique/analyte
combinations found for the bovine studies reviewed (as of June 2009 when the original
literature review on which this chapter is based was published in Food Additives and
Contaminants. Studies published after this date are reviewed in a narrative fashion in the
text, but do not appear in figures 4 and 5). This analysis was not repeated for other
species as it was apparent that a similar analysis of other species would not provide
sufficient data for a meaningful comparison.
Figure 4 – Summary of the use of different analytical techniques used in the bovine
studies reviewed in this report.
44 of 331
As Figure 4 shows, testosterone has most often been analysed using IA, whilst
nandrolone and boldenone by GC or LC-MS. The predominant use of IA (most often
radioimmunoassay [RIA], followed by enzyme immunoassay [EIA]) is mainly due to its
ease of application, its cost effectiveness, the fact that many studies were carried out in
research laboratories that do not have mass spectrometry facilities or only use them for
confirmatory analysis and because of its high sensitivity in determining analytes present
at low concentrations. Nandrolone and boldenone on the other hand are most often
analysed by GC-MS or LC-MS. This can be partially explained by the fact that
proportionally more research on these analytes is reported by residue screening
laboratories; which are more likely to use mass spectrometry than research departments
focussing on physiology. However, it may also be because of the ambiguous status of
these analytes and their metabolites i.e. are they endogenous or not?
Figure 5 – Summary of the matrices used in the bovine studies reviewed in this report
(all methods of analysis included).
Review: endogenous steroids in food-producing animals
45 of 331
As Figure 5 shows, testosterone has most often been analysed in plasma or serum,
whereas nandrolone and boldenone have most often been analysed from urine/bile or
urine/faeces respectively. As in the case of the explanation for the use of different
analytical techniques for different analytes, these differences can be in part explained by
quantitative biases in the type of research being carried out: either 1) physiology
research looking at matrices indicating relevant circulating concentrations (i.e. plasma)
for testosterone and 2) research for residue control in matrices more relevant to the
detection of abuse (i.e. concentrated amounts in urine, bile or faeces for nandrolone and
Data on the endogenous presence of androgenic-anabolic steroids in the bovine are
summarised in Table 3 while details of the major phase 1 metabolic products following
exogenous administration are given in Table 4.
1.4.2 Testosterone and related androgens in the bovine
126.96.36.199 Endogenous occurrence
As a general rule for all steroids, circulating plasma and tissues from non-excretory
organs contain relatively high concentrations of unchanged ‘parent’ steroid while
excretory products such as urine, bile or faeces contain relatively higher concentrations
of metabolites. As well as a relative difference in the proportion of each
steroid/metabolite present, excretory products generally contained higher absolute
concentrations of total analyte/metabolite due to a concentrating effect.
Testosterone and related steroids such as epitestosterone, androstenedione and DHEA
are ubiquitous among male and female animals of all mammalian species, so differences
among various groups and times are purely quantitative. When surveying the ranges of
mean, minimum and maximum values among the published studies (over 1,000 papers
for all species concerned), an approximate overall rank order of absolute concentrations
can be constructed. It must be stressed that some positions within this rank may be
caused by biases in the amount of information reported for each steroid in different
matrices. An approximate rank order for testosterone concentrations in the bovine is hair
> urine ~ fat ~ faeces ~ kidney > plasma > liver ~ muscle. An approximate rank order for
the significant testosterone metabolite epitestosterone is urine > faeces > plasma >
46 of 331
muscle > hair (no data in fat, liver or kidney). In terms of absolute values, testosterone
and epitestosterone were present at similar concentrations in muscle and plasma,
testosterone was at least a factor of 10 higher in hair and epitestosterone was around a
factor of 10 higher in urine and faeces (no data for fat, liver or kidney). There was more
variation among epitestosterone values relative to those for testosterone. As mentioned
elsewhere in this review, the majority of plasma results that contributed to the
aforementioned results do not use sample hydrolysis. However, Scippo et al. 1993
showed that while the maximum testosterone concentration found in bull plasma were
5.8 and 0.97 ng ml-1 for unconjugated and conjugated, respectively; the reverse was
seen for epitestosterone with values of 0.97 and 1.8 ng ml-1 for unconjugated and
conjugated, respectively. This could lead to artificially low reported concentrations of
epitestosterone in plasma relative to testosterone.
Several other precursors including DHEA, androstenediol isomers and androstenedione
were also occasionally quantified and there may be value in monitoring perturbations of
endogenous steroid feedback loops after exogenous steroid administration.
Existing EU guidelines for positive decision limits (as proposed by Heitzman 1994) in the
bovine already rely on separation of sex, age and gestation status as summarised in
Table 1 below.
Table 1 – EEC decision limits for testosterone in plasma (as proposed by Heitzmann
EEC decision limit in plasma (ng ml-1)
Age/sex of animal
Female (non-pregnant) 0.5
Male (< 6-months) 10
Male (> 6-months) 30
From the current review, ranges of mean plasma/serum concentrations of testosterone
and epitestosterone were found to be approximately 10-fold higher in intact mature
males relative to females (no data for steers). One significant finding was of a study that
stated that plasma testosterone was exceptionally high for a very brief time during the
late luteal phase of the normal female oestrous cycle exceeding 1.8 ng ml-1 (Dobson et
Review: endogenous steroids in food-producing animals
47 of 331
al. 1977). All other ranges of testosterone reported by this author were in-line with those
of other studies, so if real, this phenomenon could have a serious, negative impact on
the validity of the existing EU decision limit for females.
From the current review, ranges of mean urinary concentrations of testosterone and
epitestosterone were found to be approximately three-fold higher in mature males
relative to females (no data for steers). Ranges of mean muscle concentrations of
testosterone were found to be approximately 10-fold higher in mature males relative to
females or steers, although epitestosterone was similar between steers and bulls (no
data for females). Ranges of mean liver and kidney concentrations of testosterone were
found to be approximately 10-fold higher in mature males relative to females (no data for
androgens in steers or epitestosterone in any sex). Ranges of mean hair concentrations
of testosterone were found to be approximately three-fold higher in mature males relative
to females and steers (no data for epitestosterone). There were insufficient data to
compare testosterone concentrations by sex in faeces, fat or bile.
Several studies have assessed the effect of age on the plasma/serum concentrations of
testosterone in males, although the different ages, matrices and conditions under which
the animals were studied and a lack of standardisation in reporting the uncertainty of
measurement makes meaningful comparisons difficult. The results from three of the
most informative studies are summarised below:
1) Bagu et al. 2006 showed that mean male serum testosterone concentrations at 4-
weeks of age were around 0.1 ng ml-1. Concentrations then rose to 1.0 ng ml-1 at 20-
weeks, then dropped back to 0.4 ng ml-1 at 28-weeks and rose again to 1.1 ng ml-1 at
32-weeks. The authors of this study also referenced other studies that have shown a
trough in testosterone concentrations between 20 and 32 weeks of age.
2) Looking at older animals, Moura et al. 2001 reported that mean male serum
testosterone concentrations were 1.8 ng ml-1 at 26-weeks of age, rose to 8 ng ml-1 at
43 weeks and then dropped to 6.5 ng ml-1 at 52-weeks. In the same study,
androstenedione was 0.45 ng ml-1 at 17-weeks and dropped to 0.25 ng ml-1 at 52-
48 of 331
3) The most informative single study on the effect of age was published by Arts et al.
1990. Median male plasma testosterone concentrations at 15-weeks of age were 0.8
ng ml-1 and then rose to 1.3 ng ml-1 at 28-weeks. Concentrations of epitestosterone,
however, dropped from 7.1 ng ml-1 at 15-weeks to 0.8 ng ml-1 at 28-weeks. In the
same study, median male urinary testosterone concentrations were 1.0 ng ml-1 at 15-
weeks and rose to 3.7 ng ml-1 at 28-weeks. Epitestosterone concentrations did not
change with age and values 15- and 28-weeks were 40 and 41 ng ml-1 respectively.
As a result of the aforementioned testosterone and epitestosterone concentration
changes with age, the epitestosterone:testosterone ratio fell significantly from 15- to
28-weeks of age. On this note, the testosterone:epitestosterone ratio has been found
to be a good indicator of testosterone abuse in humans and horses (due to selective
elevation of testosterone after testosterone doping), but Angeletti et al. 2006 showed
it to be of less use in the bovine, probably due to the relatively high 17α-hydroxylase
Relatively fewer studies have analysed the effect of age on female testosterone
concentrations. Nakada et al. 2000 reported that mean female plasma testosterone
concentration immediately after birth was 0.075 ng ml-1 but then fell, ranging between
means of 0.015 and 0.021 ng ml-1 between birth and puberty. Mean (and standard error
of the mean) age to puberty was 43.3 (1.3) weeks, with a range of 38-55 weeks. The
same study by Arts et al. 1990 that reported male testosterone data by age also reported
female data. Median female plasma testosterone concentration was less than the LOD at
both 15- and 28-weeks, while median plasma epitestosterone was less than the LOD at
15-weeks and then rose to 0.2 ng ml-1 at 28-weeks. Median female urinary testosterone
concentration at 15-weeks was less than the LOD and then rose to 1.1 ng ml-1 at 28-
weeks. Median female urinary epitestosterone at 15- and 28-weeks were 6 and 17 ng ml-
1 respectively. There were insufficient data to compare the effect of age on testosterone
or related metabolites/precursors in faeces, liver, kidney, bile, muscle, hair or fat.
Data on the concentration of testosterone in any matrix from pregnant females was not
available, but it would be expected to be elevated relative to non-pregnant females in line
with other steroids (see later sections). However, mean plasma concentrations of DHEA
and androst-5-ene-3β,17β-diol were found to be approximately three-fold higher in
pregnant females (Gabai et al. 2004).
Review: endogenous steroids in food-producing animals Download full-text
49 of 331
In a 2001 study, Plusquellec et al. showed that lactating cows of the Herens breed
(artificially selected for fighting ability) had significantly higher (P<0.05) median plasma
testosterone concentrations compared to Brune des Alpes animals, with values of 0.21
and 0.11 ng ml-1 respectively. This conforms to the observation that aggression
correlates with increasing testosterone levels. The biochemical observations were also
borne out by secondary sexual characteristics, which were more prominent in the Herens
A study by Moura et al. 2001 showed that bulls suffering spermatic arrest had only
slightly lower serum testosterone concentrations than healthy controls. However, serum
androstenedione in one diseased animal was > 0.8 ng ml-1 at 12-months, relative to a
mean of 0.25 ng ml-1 in healthy controls. No reports on the effect of other factors known
to increase the androgen output in other species were found i.e. stress or congenital
No studies were found that directly compared concentrations of testosterone or related
precursors/metabolites in similar breeds under different housing conditions or in different
countries, nor of diet, time of day or season on testosterone or related
In most species long-term treatment with gonadotrophin-releasing hormone (GnRH)
agonists such as deslorelin decrease luteinzing hormone (LH) output (and therefore
testosterone secretion) due to desensitization of the pituitary gland. However, Aspden et
al. 1997a reported that testosterone concentrations in mature bulls are increased
following deslorelin administration, although another effect of this drug is that LH
pulsatility is lost, leading to a flat LH secretion profile. On the other hand, Renaville et al.
1996 showed that administration of GnRH to immature bulls between 70- and-150 days
of age delayed puberty relative to controls with mean pubarche ages of 180- and 120-
days respectively. No reports of the effects of other non-steroidal medications on
androgen concentrations were found, but several types of medication in other species
are known to affect increase or decrease in concentrations e.g. cytochrome P450
enzyme inducing inhibiting drugs.