LC-MS in doping control - Pharmaceutical Press chapter.pdfdoping control inﬂuence analytical strategies in athletes drug testing: • The substance-based doping deﬁnition priori-tises

Introduction

Definition of doping

Doping analysis comprises a diversity of sub-stance classes with different pharmaceutical andchemical properties. Therefore, the discussion ofthe suitability of liquid chromatography-massspectrometry (LC-MS) in doping analysis needsto distinguish various categories.

According to its formal definition, a dopingviolation in sports can be caused by variousevents, e.g.:

• the detection of a prohibited substance ormetabolites or markers of that substance (asdefined by the recent document [1] of theWorld Anti-Doping Agency [WADA]) in theathlete’s specimen

• the use of prohibited substances or methods• possession or trafficking prohibited substances• refusing without compelling justification to

submit a sample.

This definition is clearly legally motivated anddoes not support the discussion of technicalissues.

The number and classes of prohibited sub-stances is very complex in human sports, whereselected stimulants, narcotics, hormones, β2-agonists, anti-oestrogenic agents and diuretics arecovered. The situation becomes even more con-fusing if the term ‘doping’ is extended to animal(e.g. equestrian) sports, where any application ofpharmaceutical drugs is totally prohibited andeven substance classes like muscle relaxants ormild stimulants are included.

Moreover, doping analysis is closely related to

adjacent fields with similar analytical prospects,like veterinary residue control (predominantlydealing with identification of growth promotersin various matrices), forensic sciences (the major-ity of doping-relevant substances are scheduledas controlled substances in most countries),environmental analysis (e.g. steroids in wastewater) or clinical chemistry (e.g. due to theincreasing relevance of steroid hormone replace-ment therapy).

This chapter describes the key fields of appli-cation of LC-MS in routine doping control (i.e.screening analysis, confirmation and quantifica-tion of positive results) extra to particularresearch activities. The latter are focused on theintended technical improvements (e.g. extensionof detection time windows, reduction of turn-around times and costs) of conventional analyti-cal procedures and, in particular, on the detectionof prohibited substances that cannot be ade-quately identified so far (e.g. growth hormone).

The arrangement follows mainly historicaland technical considerations, and does notnecessarily represent the frequency or relevanceof the application of LC-MS.

LC-MS in doping control – historical andtechnical aspects

Some peculiar legal and technical principles indoping control influence analytical strategies inathletes drug testing:

• The substance-based doping definition priori-tises target analyses compared with generalunknown screening procedures. Sensitive andspecific selected ion monitoring (SIM) or

9LC-MS in doping control

Detlef Thieme

193

LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 193

selected reaction monitoring (SRM) experi-ments are much more frequent than scanningexperiments.

• Urine, which is the preferred specimen fordoping control due to the ease of samplecollection and relatively high concentrationsof xenobiotics, requires a careful considerationof substance metabolism, including conjuga-tion. Minor biochemical pathways leading tolong-term metabolites are often more import-ant than active parent compounds. Therelevance of quantitative analyses is reducedto a few ‘threshold substances’. This groupcomprises compounds that are accepted belowcertain threshold concentrations, because lowamounts may be due to a permitted adminis-tration (e.g. inhalation of salbutamol) or anendogenous origin of the substance (e.g.natural levels of testosterone). In general,qualitative substance identification is asufficient proof of a doping offence.

• The differentiation between substance prohi-bition ‘in competition’ and ‘out of competi-tion’ requires modified analytical procedureswith respect to numbers of included sub-stances and threshold concentrations.

• A major analytical challenge consists in theverification of the prohibited administrationof endogenous substances like testosterone,human growth hormone (hGH) or erythro-poietin (EPO). In such cases, minor quantita-tive (e.g. amount of steroids compared withendogenous precursors or biochemicalbyproducts) or qualitative (e.g. glycosylationof proteins) deviations need to be identified.

• MS plays an outstanding role in dopinganalysis and was originally considered as amandatory analytical technique for confirma-tion of substance identity. Exceptions werelater acknowledged in the field of peptidehormones.

Approaches to the application of an LC-MScoupling in the framework of doping controlwere already reported in 1981 [2], when acombination of LC-MS equipped with a movingbelt was used for MS confirmation of corticos-teroids. The relevance of LC-MS application indoping control was ruled by practical demands,resulting in an early implementation of the tech-

nique in anti-doping research and routine [3]analyses. The issue of peptide hormones wasalready tackled in the mid-1990s, because gaschromatography (GC)-MS analysis could notsolve the problem of identification of macromol-ecular compounds. The potential of LC-MS to dif-ferentiate intact growth hormone obtained fromdifferent manufacturers, quantify the insulin-likegrowth factor (IGF-1) and characterise humanchorionic gonadotrophin (hCG) after trypticdigestion had already been reported in 1994 [4,5]. However, these ‘proofs of principles’ demon-strate the general usefulness of LC-MS for theidentification of peptide hormone doping, butare not used routinely to date, mainly due tosensitivity limitations.

The subsequent developments were mainlycharacterised by practical improvements. Sub-stances (e.g. mesocarb [6]) and substance groups(e.g. diuretics [7]) causing severe analyticalproblems (stability) or inconvenience (time-consuming derivatisation reactions) werecovered by efficient LC-MS assays, while otherscreenings remained unchanged, due to theavailability of well-established and validatedGC-MS procedures.

In contrast, there is an obvious preference touse LC-MS in cases of upcoming new substances(like the ‘designer steroid’ tetrahydrogestrinone[THG] and the stimulant modafinil) or substancegroups (e.g. corticosteroids, included in the list ofprohibited substances in 2003).

There is no preferred default LC-MS instru-mentation in doping control analyses. Almostany technical variant of ionisation – electrosprayionisation (ESI), atmospheric pressure chemicalionisation (APCI) or photo-ionisation (APPI) –has been applied in combination with quadru-pole (Q), ion trap (IT) or time-of-flight (TOF) massanalysers, whether as single or tandem massspectrometers.

Certain reports of related technical develop-ments of LC-MS seem to be just coincidentallyassociated with doping, e.g.:

• The introduction of isotope ratio MS linked toLC enables the identification of the origin ofsubstances. In particular, a differentiation ofendogenous production from synthetic mat-erial becomes possible in principle. However,

194 Chapter 9 • LC-MS in doping control


applications presented so far [8] are ratherinsensitive (requiring 400 ng substance on-column) and therefore of no practical value inroutine cases.

• The introduction of Fourier-transform ioncyclotron resonance (FTICR) MS to identifycorticosteroids [9] is probably technicallymotivated. The high expense of this techniquedoes not permit routine applications. Never-theless, it is clear that high-resolution (HR)-MSis essential for identification of multiplycharged intact peptide hormones (hGH) [10].Affordable routine instruments could greatlyimprove the detection and characterisation ofproteins.

Additionally, progress in chromatographic separ-ation (e.g. column switching [11], use of graphi-tised carbon [12] or chiral [13] columns) needs tobe achieved, particularly in the field of peptidehormones. The improvement of ionisationefficiency appears to be a crucial aspect of steroidanalysis. Derivatisation (dansylation [14]) ofsteroids and attempts to improve APPI ionisation(e.g. using anisol as dopant gas) are both specifi-cally focused on a sensitivity enhancement [15].

Small molecules

The majority of doping-relevant substancesbelong to the category of small molecules. Themost common definition of substance groups isbased on pharmaceutical activity, distinguishingstimulants, narcotics, cannabinoids, anabolicagents, β-agonists, anti-oestrogens, maskingagents and glucocorticoids. However, this sched-ule is not suitable for analytical consideration asstructurally and analytically unrelated speciesmay be arbitrarily grouped together (e.g. clen-buterol and testosterone as anabolic agents)because of their equal intended activity.However, one substance may be scheduled indifferent groups. The β2-agonist salbutamol, forinstance, may be considered as a permitted anti-asthmatic, as a stimulant or as an anabolic agent,depending on the occasion of the doping control(in/out of competition) and on its urinary con-centration. The group of masking agents is rather

complex too. It summarises any substance thatmay interfere with doping analysis by:

• diluting the urine and accelerating excretion(diuretics)

• suppression of reabsorption of xenobiotics(uricosurics, e.g. probenicid)

• manipulation of endogenous steroid profiles(administration of epitestosterone, used as anendogenous reference of urinary steroid con-centrations, is able to conceal elevated levelsof testosterone).

Diuretics

The main motivation to prohibit the use ofdiuretics in sports is the intended reduction ofbody mass in weight-classified sports. The secondreason is a masking effect. Due to forced diuresis,the clearance of prohibited substances (e.g. ana-bolic steroids) may be accelerated and the urinaryconcentration may drop below the detectionlimit or threshold. Different types of diuretics(thiazide-like, loop and potassium-sparingdiuretics) may be distinguished on the basis oftheir pharmacological properties (Table 9.1).However, these classes are not differentiated withregard to their doping relevance. The first twogroups are supposed to be most efficient indoping because of their high potency. They arecharacterised by acidic groups (acid amides, typi-cally sulphonamides) and are therefore suitablefor negative ionisation. In contrast to these com-pounds, the group of potassium-sparing diureticsis characterised by steroid structures (aldosteroneantagonists, e.g. canrenone) or cycloamidinestructures (triamterene), and is more appropriateto protonation and subsequent detection in pos-itive-ionisation mode.

Other compounds with diuretic (side) effects,like osmodiuretics (mannitol) or xanthine deriv-atives (caffeine), are no longer mentioned in thelist of doping substances, because prohibited usecould not be certainly discriminated from theirpermitted applications as food ingredients.

The identification of diuretics in dopingcontrol was originally carried out by GC-MS andLC-UV diode array detection (LC-DAD) [7, 16].The reduced polarity of the methyl derivatives

Small molecules 195


permitted a GC separation and subsequentidentification of the majority of diuretics in SIMmode. This approach includes analytical limi-tations, in addition to the requirement for atime-consuming derivatisation step. Certain sub-stances do not form stable, reproducible anduniform methyl derivatives. Chromatographicartefacts (Figure 9.1) need to be factored into thescreening [17] and there remained at least onediuretic (benzthiazide) undetectable in GC-MS.Therefore, LC-DAD was additionally applied as acomplementary analytical technique. Theobvious benefits of LC-ESI-MS are revocation ofderivatisation, improvement of comprehensive-ness, reduction of turn-around times and increase

of sensitivity [3, 18]. As stated earlier, there arestrongly acidic as well as basic compoundsamong the class of diuretics, requiring eithernegative- or positive-ionisation modes, respec-tively. Basic diuretics may well be combined withother screening procedures (e.g. anabolicsteroids) to avoid a re-injection of samples indifferent ionisation modes. Alternatively, theoption of a scan-to-scan polarity switching wasutilised to detect both groups of diuretics simul-taneously [19, 20].

The relatively high urinary concentrationscombined with the sensitivity of the techniquereduces chromatographic separation to aminimum and diminishes turn-around times to a


Table 9.1 Typical classes and examples of prohibited diuretics

Class Typical modifications Examples Chemical structure

Thiazide R1 = H, alkyl, subst. phenyl benzthiazideR2 = Cl, CF3 hydrochlorothiazide

Thiazide analogues R1 = subst. amide mefrusideR2 = H, OH xipamide

Loop diuretics R1 = Cl, phenoxy furosemide(furosemide type) R2 = amine bumetanide

Potassium-sparing diuretics, canrenonealdosterone antagonists spironolactone

Cycloamidine derivatives triamtereneamiloride

N

NHS

R2

H2NO2S

R1

O O

H

Cl

H2NO2S R1

R2

H2NO2S

R1

R2

COOH

O

O

O

N

N

H2N

Cl

NH2

O

NH

NH

NH2


few minutes. A combination with automatedsolid-phase extraction (SPE) may be applied toestablish high-throughput LC-MS screening [21].The assay is sufficiently selective to differentiateconcentrations above a minimum required per-formance level (MRPL [17]) of 250 µg/L from ablank matrix (Figure 9.2). The potential influenceof ion suppression is not critical, because thereare no threshold values and the quantification ofdiuretics is not required in screening analyses.

βββ2-Agonists

The classification of these sympathomimeticagents in doping control has been frequentlymodified. At present, there is a differentiationbetween substances that are permitted byinhalation to treat asthma (requiring a thera-peutic use exemption), while others are pro-hibited due to their potential stimulating oranabolic effect. Clenbuterol, which is supposedto be the most potent anabolic β-agonist, consti-tuted a particular analytical challenge as typicalurinary concentrations are of the order of magni-tude of 1 µg/L. Due to its two chlorine atoms, itis well suited for HR-MS and its trimethylsilyl(TMS) derivative was preferably identified byeither GC-HR-MS or GC-MS-MS. Typical LC-MSscreening procedures are slightly less sensitivethan GC-MS procedures, but more comprehen-sive, which appears to be more important in the

field of horse testing [22] because of the generalprohibition of all β2-agonists, regardless of theirintended action, concentration or adminis-tration route. Typically, β2-agonists (Table 9.2) areidentified in positive ESI mode using the proton-ated molecule as precursor ion. Characteristicfragmentation reactions are losses of the terminalisobutene group, resulting in [M – 56]+ fragments,whether or not combined with losses of water[23].

Approaches to differentiate between inhala-tional and prohibited systemic (e.g. oral) appli-cation of salbutamol were based on quantitativeexaminations (values greater than 1 mg/L were,according to the WADA regulations, consideredas an adverse finding) or investigations of salbu-tamol enantiomers. A discrimination functionwas derived from the higher amounts of the S(+)relative to R(–) isomer after oral administration.This approach requires the combination of chiralLC separation with MS detection, whether on-line or off-line [13].

ββ-Blockers

β-Adrenergic blocking agents (β-blockers) areprohibited due to their reduction of heart rate,blood pressure and hand tremor. Doping controlsare consequently restricted to competition con-trols in particular sports where steadiness isimportant (archery, shooting, etc.).

β-Blockers are characterised by a very similarchemical structure. With a few exceptions (e.g.sotalol or carvedilol), they represent derivativesof oxypropanolamine terminated by t-butyl orisopropyl groups and aromatic substituents(Table 9.3).

GC-MS was the conventional screeningtechnique for β-blockers in doping control.Derivatisation with N-methyl-N-trimethylsilyltri-fluoroacetamide (MSTFA), if necessary combinedwith N-methyl-bis-trifluoroacetamide (MBTFA),was applied after hydrolysis of the conjugates andisolation from urine [24]. The application ofLC-MS represents a useful alternative to avoidthis time-consuming derivatisation step, and theformation of unstable derivatives and artefactsin some cases (e.g. acebutolol). All β-blockerscontain a secondary amino group accounting for

Small molecules 197

SulphoxidationDealkylation/

ring opening

Figure 9.1 Diuretics with thiazide structure are convertedby hydrolysis or sulphoxidation during sample transporta-tion, storage or preparation, which may cause analyticalproblems in GC-MS. Respective artefacts need to be fac-tored into LC-MS screening.


their protonation and detection in positivemode. According to their high structural simi-larity, there are group-specific fragmentationreactions for both classes of β-blockers (Table 9.3).Substances with a terminal t-butyl group arecharacterised by loss of isobutene (M – 56),usually in combination with loss of water, whilemost of isopropyl terminated compoundsundergo a loss of an isopropylamino group

(M – 77) in addition to the formation of anN-isopropyl-propanolamine fragment (mass-to-charge [m/z] ratio 116) [25].

The proposal of a combined screening ofdiuretics and β-blockers agents, utilising ascan-to-scan polarity-switching technique [19], isvery promising for clinical purposes, becauseboth substance groups are frequently combinedin the treatment of hypertension. However, this


0.0

7.0e4 (a)

(b)

Inte

nsity (

cp

s)

Inte

nsity (

cps)

1.0 2.0

Time (min)

0.0

2.7e4

Figure 9.2 Multiple-reaction monitoring (MRM) experiments permit the screening of low amounts of prohibited diureticsin negative-ionisation mode. The turnaround time and sample preparation may be reduced to a minimum due to the out-standing selectivity. A mixture of 20 diuretics extracted from a urine matrix (a) is compared with a blank urine containingmefruside as internal standard (IS) (b).


approach seems to be less rational in dopingcontrol, because the intention of an abuse ofboth substance classes and the scope of theirprohibition (concerning sports, in competitionversus out of competition) are different.

Steroids

This substance class is characterised by a uniformstructure consisting of a modified sterane skele-ton (Table 9.4). Due to this apolar structure, the

Small molecules 199

Table 9.2 Chemical structures of typical β-agonists


Aniline R1, R2 = Cl, Br, CN, OH clenbuterolR3 = C(CH3)3, subst. phenyl brombuterol

fenoterol

Phenol R1 = OH, subst. alkyl salbutamolR2 = alkyl, subst. phenyl orciprenaline

Benzazepinone zilpaterol

H2N

R1

R2

OH

NH

R3

OH

NH

R1

R2

HO

N

NH

O

HO

HN

Table 9.3 Chemical structures of typical β-blocking agents prohibited in particular sports

Class Modifications Examples Chemical structure

Isopropylamine R1 = substituted aromatic rings atenololacebutololpropranolol

t-Butylamine R1 = substituted aromatic rings bupranololcarteololtimolol

NH

O

OH

R1

NH

OH

R1 O


efficiency of ionisation and suitability of LC-MSmainly depend on molecular substitutions,resulting in tremendous variations of sensitivityamong different steroids. Saturated steroid mol-ecules (e.g. by saturation of rings or reduction ofketo groups) may hardly be ionised by protona-tion or formation of ammonium adducts. Thepresence of conjugated double bonds (3-keto-4-ene, steroids), oxidation of the sterane moiety orother polar ring substitutions (e.g. the pyrazolring in stanozolol) improves the ionisation ratesignificantly. Therefore, this apparently uniformsubstance class needs to be divided into thefollowing groups accounting for their LC-MSproperties.

Synthetic anabolic steroids

The group of anabolic steroids still includes adiversity of similar structures, which cannot besystematically separated. The subgroups listed inTable 9.4 (biologically active anabolic steroids,precursors and metabolites) overlap each otherfrom legal as well as biochemical perspectives.These compounds (e.g. androstenedione) may beconsidered as precursors and metabolites in thebiosynthesis of steroids, but they also originatefrom prohibited application of synthetic ana-logues (pro-hormones) of endogenous steroids oreven from synthetic hormones. The analyticalresult of a urine analysis does not necessarilyallow the differentiation of an approvedmedication, the abuse of pro-hormones (legally


Table 9.4 Chemical structures of selected endogenous and synthetic steroids

Class Typical modifications Example Chemical structure

Endogenous oxidation/reduction in testosteronesteroids (including positions 3/17, saturationtheir precursors of 4 double bond, 5-α/βand metabolites) isomers

Synthetic 17α alkylation, double oxandrolonesteroids bonds at position 1–2, 4–5

or 5–6, A-ring condensation,4-chlorination, 19-demethylation

Esters acetate, propionate, testosteronedecanoate esters

O

OH

CH3

O

O

OH

12

3

45

6

12

161719

BA

DC

O

O

O

(CH2)nCH3


tolerated in certain countries) or the applicationof scheduled prohibited steroids.

The attempts to develop LC-MS methods forthe identification of anabolic steroids of the so-called ‘free fraction’ (i.e. slightly polar steroidswhich are excreted unconjugated in urine)demonstrated significant diversity of ionisationprinciples [26]. All anabolic steroids are detectedin positive mode. However, the appearance of themost abundant precursor ion depends signifi-cantly on structural modifications of the steraneskeleton and is almost unpredictable. Protonatedions [M + H]+, ammonium adducts [M + NH4]+

and fragments resulting from loss of up to threemolecules of water [M – nH2O]+ were reported asbase peaks. The balance between these ions isdetermined by proton affinity and thereforeconjugated double bonds (3 keto-4-ene steroids,pyrazol ring condensation, aromatic rings) arethe most obvious structural indicators for animproved protonation. Physical conditions inthe ion source appear to represent another sensi-tive factor of ionisation efficiency. ComparingESI, APCI and APPI for various anabolic steroids,it turned out that the choice of a precursordepends significantly on the ionisation tech-nique. Oxandrolone (Table 9.4), for example,formed predominantly the [M + H]+ ion in ESI,the adduct [M + NH4]+ in APCI and the [M – H2O]+

fragment in APPI [27].Moreover, there is no general consensus about

the preferences of various ionisation techniquesfor steroid analysis (Figure 9.3). Controversialevaluations of APPI [27, 28] suggest that techni-cal differences between manufacturers’ conceptsare more significant than physical constraints.The typical acquisition mode of these methods isMRM, including two to three fragmentationreactions per analyte. LC-MS has contributed tothe structural elucidation in the case of the newupcoming anabolic steroid tetrahydrogestrinone(THG) (Figure 9.3) [29] and provided an efficientcomplementary screening method, directed tothe identification of polar steroids (e.g. tren-bolone, THG, stanozolol).

However, unpredictable formation of pre-cursor ions, relatively low ion abundances, andunspecific fragmentation reactions in combin-ation with the existence of numerous isomericsteroids and metabolites complicate the design of

comprehensive and sensitive LC-MS methods forthe identification of anabolic steroids. Appli-cation in routine steroid analysis is mainlyfocused on selected polar steroid molecules(Figure 9.4), e.g. stanozolol [26, 30], boldenone[31], trenbolone [32, 33] or THG [29]. These polarsteroids often encounter analytical difficulties inGC-MS, due to the formation of instable deriva-tives and artefacts after silylation. The appli-cation of an additional derivatisation (e.g.methoxime derivatives of THG) followed by anextra GC-MS procedure would be required toidentify these substance groups. Therefore, LC-MS is a beneficial alternative to the extra effortof additional derivatisations or modified GCtechniques.

Endogenous steroids

The problem of endogenous steroids in drugtesting human athletes is mainly reduced to thequantitative balance between testosterone and itsbiochemical byproduct epitestosterone (see‘Steroid conjugates’). In addition, there is anupcoming interest in the quantification ofendogenous steroids related to the therapeuticadministration of steroids (treatment of testo-sterone deficiency [34]) and to their increasingrelevance as lifestyle drugs. Therefore, the quan-titative evaluation of endogenous steroid profilesusing LC-MS-MS is of increasing importance inclinical chemistry.

The growing number of precursors of en-dogenous anabolic steroids (so-called pro-hormones) that are widely available as nutritionsupplements and abused in sports and body-building constitutes another analytical chal-lenge. Biochemical precursors of testosterone(e.g. androstenedione, dehydroepiandrosterone)were the first products on the market, but theyhave been recently replaced by synthetic steroids,representing slight structural modifications ofthe endogenous compounds (e.g. 1-testosterone,1,5α-androstenedione, where the location of thedouble bond is shifted from the 4 to the 1 posi-tion). Analytical problems to identify thesecompounds by LC-MS are comparable to thoseencountered for their endogenous counterparts[35]. The technique is probably not sufficient forthe unambiguous identification of all potentially

Small molecules 201



XIC of +MRM: 313.2/159.1 amu

Max. 8.7e4 cps

0.0

8.5e4

5.31

(a)

XIC of +MRM: 313.2/159.1 amu

Max. 2250.0 cps

1.0 2.0 3.0 4.0 5.0 6.0

Time (min)

1.0 2.0 3.0 4.0 5.0 6.0

Time (min)

0.0

1.8e4

Inte

nsity (

cps)

Inte

nsity (

cp

s)

5.52

(b)

Epitestosterone-d3

IS

THG

OH

O M=312.1

Figure 9.3 Comparison of LC-MS-MS detection of epitestosterone and THG in (a) APPI and (b) ESI. Both chromatogramswere run under identical LC conditions. APPI shows an outstanding sensitivity for detection of THG, but is limited to steroidswith chromophoric molecular structures, while ESI appears to be the most versatile ionisation technique.


relevant steroids, but represents a helpful supple-ment to GC-MS.

Steroid esters

Steroids are typically administered as fatty acidesters by intramuscular injection. These com-pounds are stored in adipose tissues, rapidlyhydrolysed in blood, not markedly excreted in

urine as such and, therefore, not included inroutine doping analysis.

The general benefit of the identification ofsteroid esters is an unequivocal proof of illegaladministration of a synthetic compound. Anytrace amount of the exogenous esters provides aclear indication for a doping offence, while thecorresponding free steroids need to be distin-guished from natural endogenous levels.

Small molecules 203

269→161,

269→105

0.0

3e4

Inte

nsity (

cp

s)

Inte

nsity (

cp

s)

6.76Epimetendiol

345→345,

345→97

5.0 6.0 7.0 8.0

Time (min)

0.0

2e5

6.01

5.65

5.15

3′-OH-Stanozolol

HN

N

OH

OH

HO

CH3

CH3

OH

H

Figure 9.4 Threshold concentrations of 2 µg/L of 3�-OH-stanozolol and epimetendiol in extracts from a urinary matrix.The polar stanozolol metabolite is exceptionally well ionised and permits a sensitive identification, while detection of themedium polar epimetendiol is hampered by a limited intensity and specificity.


LC-APCI-MS-MS was reported to facilitate asensitive identification of testosterone [36] or19-nortestosterone esters [37] in equine plasma.The formation of dominant [M + H]+ precursorions was reported for all steroid esters.

Steroid conjugates

The detection of the abuse of endogenoussteroids (testosterone or its precursors) is basedon a quantitative evaluation of the urinaryconcentrations of testosterone and epitestos-terone. The latter steroid is a byproduct of thebiosynthesis of steroids and supposed to besuppressed after the administration of steroids.According to WADA criteria, further investi-gations to exclude physiological deviations aremandatory, if the ratio of testosterone/epitestos-terone exceeds a value of 6. This definition ismainly empirical and derived from the conven-tional analytical GC-MS procedure, which wastraditionally based on the quantification of bis-TMS derivatives of both epimers. Hydrolysis ofthe conjugates (mainly glucuronides and sul-phates, Table 9.5) was carried out after hydrolysis

with glucuronidase from Helix pomatia, whichdoes not cleave the corresponding sulphates.Logically, the testosterone/epitestosterone ratiosare by consensus referred to the total amount offree and glucuronidated steroids. It was suggestedthat elevated relative amounts of epitestosteronesulphate may cause increased testosterone/epitestosterone ratios when measured by conven-tional GC-MS. A potential racial bias of phase 2biotransformation was discussed.

The quantification of intact conjugates, whichis made possible by LC-MS, seems to be a con-clusive option to examine the individual influenceof glucurono- and sulpho-conjugation on testos-terone/epitestosterone ratios directly and circum-vent the uncertainty of hydrolysis recovery [38,39]. Steroid glucuronides were reported to formdifferent adducts (protonated, ammoniated andsodiated ions) in positive ESI. Depending on thedeclustering potentials, the [M + H]+ was found tobe most suitable due to its high signal-to-noise(S/N) ratios. Reasonably specific fragments(aglycone and its singly or doubly dehydratedfragment ions) were chosen for sensitive MRMexperiments. The use of a deuterated IS for each


Table 9.5 Chemical structures of steroid conjugates

Conjugation Examples Chemical structure

Glucuronides androsteroneglucuronide

Sulphates testosteronesulphate

O

O

O

COOH

HO

HO

OH

O

OSO3H


conjugate was found to be mandatory due to thepotential-ion suppression in the urinary matrix.The assumed influence of sulphation on elevatedtestosterone/epitestosterone ratios was not sup-ported by LC-MS-MS examinations [39].

Quantitative uncertainties of conjugatehydrolysis are less crucial in the case of exogenoussteroids, because anti-doping legislation does notrequire a quantitative threshold. Approaches toidentify glucuronides of synthetic steroids by LC-MS were directed to structural investigations ofsteroid glucuronides and automation of the ana-lytical process by application of on-line micro-extraction [40]. Steroid glucuronides proved toexhibit similar ionisation principles to their freeanalogues. The presence of conjugated doublebonds increases the proton affinity, resulting in anelevated abundance of [M + H]+ pseudo-molecularions, while saturated molecules tend to formammonium adducts rather than protonated ions.The steroid conformation (5α versus 5β linkage ofthe A and B rings of the sterane skeleton) wasreported to influence the fragmentation of A-ring-saturated steroids; 5β isomers were found to

produce considerable higher amounts of thedehydrated aglycone fragments (relative to thedeconjugated steroid) than 5α isomers [41].

Corticosteroids

Glucocorticosteroids (Table 9.6) are prohibitedwhen administered systemically (orally or byinjection), whereas all other administrationroutes require medical notification. Due to theirinfluence on protein and carbohydrate metab-olism, they are known to be abused as growthpromoters in food-producing animals and mayreasonably be abused in sports. At the moment,there is no analytical solution for a reliable differ-entiation of the administration pathway and ana-lytically positive cases may be rejected by theavailability of a therapeutic use exemption. ESI innegative mode was consistently reported to bethe most efficient ionisation for a corticosteroidscreening method [42, 43]. Typical fragmentationreactions are loss of the CH2OH moiety, water orhydrofluoric acid. The LC-MS assays are carriedout either in single-MS mode using diagnostic

Small molecules 205

Table 9.6 Chemical structures of synthetic corticosteroids


Cortisone 1–2 double bond prednisone

Cortisol 1–2 double bond prednisoloneR1 = H, OH betamethasoneR2 = H, OH, CH3 triamcinoloneR1, R2 = C(CH3)2R3, R4 = H, F

O

OH

OHO

O

O

R1

HO

R2

R3

R5

R4


fragments or in MS-MS mode, where [M +CH3COO]– adducts serve as precursor ions due tothe absence of [M – H]– pseudo-molecular ions.

Alternatively, the identification of intactcorticosteroid conjugates in bovine urine wasevaluated [44]. Base peaks detected in positive ESImodeare sodiumadducts,whileapredominant [M– H]– ion is observed in negative mode. Both pre-cursors exhibit a low fragmentation rate. MRMexperiments in negative-ionisation mode using a[M – H]– → [M – H]– pseudo-transition (monitoredat elevatedcollisionenergy)were foundtoproducespecific signals due to the high ion stability.

Other prohibited substances (stimulants,narcotics, cannabinoids)

There are numerous other prohibited substancesin doping control eligible for application of LC-MS, like stimulants or narcotics. Respectiveanalytical procedures are available in clinical,veterinary or forensic toxicology and may cer-tainly be adopted in doping analysis. However,conventional screening procedures based on GCwith nitrogen phosphorous-selective detection(NPD) or GC-MS are still state of the art, becausethe concentration and stability of respectivecompounds are sufficiently high. Moreover, thelist of prohibited stimulants and narcotics hasbeen revised and condensed recently, simplifyingthe analytical challenges.

Analytical development in these areas isfocused on upcoming (e.g. modafinil) or criticalsubstances (e.g. mesocarb). The latter stimulant isthermally instable, forms irreproducible artefactsand is therefore difficult to identify by conven-tional GC-MS. The long history of attempts to usevarious ionisation techniques (particle beam,thermospray, ESI, APCI) to detect mesocarb andelucidate its biotransformation [45] demonstratesthe analytical difficulties when tackling this sub-stance. Mesocarb does not fit analytically into itsnative pharmacological group of stimulants andis typically integrated into other LC-MS screeningprocedures (e.g. diuretics).

Other isolated substances causing GC prob-lems due to their high polarity and/or lowthermal stability (e.g. the anti-oestrogenicclomiphene or the cocaine metabolite benzoyl-

ecgonine) are comparable candidates for aninsertion into LC-MS screening procedures.

The identification of the major urinarymetabolite of ∆9-tetrahydrocannabinol (THC)(the glucuronide of 11-nor-9-carboxy-THC) indoping control does not represent any particularexception compared with forensic urine analysis(see Chapter 8). A threshold value of 15 µg/Lspecifies a potential doping violation; however,there are no uniform sanctions and cannabinoidswere downgraded to the group of ‘specified sub-stances, which are particularly susceptible tounintentional anti-doping rule violations . . . andwhich are less likely to be successfully abused asdoping agents’ [1]. Any eligible LC-MS assay forcannabinoids [46] may be applied in sports drugtesting without modifications.

The identification of alkylated xanthine deriv-atives is no longer essential in human dopinganalysis after its removal from the list of pro-hibited substances in 2004, although recentpublications have dealt with its identificationand quantification by LC-MS in horses [47, 48].According to the required performance specifica-tion of the Association of Official Racing Chem-ists (AORC), concentrations as low as 100 µg/Lneed to be detected in equine urine.

Among numerous pharmaceutical substancesthat are controlled in equine doping control,there are quaternary ammonium anti-cholinergicagents (e.g. isopropamide, glycopyrrolate) thatappear to be well suited for LC-MS. According totheir ionic structure, quaternary ammoniumdrugs require special pre-treatment (ion pair for-mation) to enable conventional liquid–liquidextraction (LLE) or LC separation. Examining theidentification of eight quaternary ammoniumdrugs in horse urine, the application of capillaryelectrophoresis was found to be more appropriate(enhanced sensitivity and separation power)than LC [49].

Large molecules

Introduction

Several proteins and peptide hormones areincluded in the list of prohibited substances,



due to their anabolic effect, as in the case ofhGH or hCG, or due to an increase in the oxygentransportation capacity of blood, like EPO andhaemoglobin-based oxygen carriers (HBOCs).These substances are synthetic or recombinantanalogues of endogenous hormones. The possi-bility to discriminate between an abuse of pro-hibited substances and endogenous productiondepends on the structural modifications of therespective exogenous compound. Possible ana-lytical approaches are:

• quantitative evaluations of compounds with asignificant concentration difference betweenbasal levels and exogenous administrations(hCG, IGF-1)

• identification of variations of the primarystructure (insulin, HBOCs)

• discrimination of mass variants of proteins(20- and 22-kDa isomers of hGH)

• investigation of charge variants resulting frompost-translational modifications (e.g. glyco-sylation of EPO).

In fact, the analytical approach depends on thecurrent availability of pharmaceutical substanceswhich are potentially abused. EPO, for instance,was originally available as recombinant proteinwith a primary structure identical to the en-dogenous hormone. The introduction of syn-thetic variants with modified amino acidsequences required an immediate adaptation ofthe analytical approach.

The practical application of LC-MS techniquesto identify large molecules in routine dopingcontrol is still restricted by sensitivity limitationsand the requirement of extensive and selectivesample clean-up. A shotgun approach based ondigestion of proteins followed by HR separationof the digests is most frequently applied. Selectiveclean-up procedures (e.g. immunoaffinity enrich-ment) of the resulting complex peptide mixturesneed to be applied to reduce ion suppression andto be able to identify relevant proteins at lowamounts.

The characterisation of structural particulari-ties based on the identification of intact proteinsrequires relatively high amounts of samplematerial and HR-MS for identification of themultiply charged precursor ions. Adequate ana-lytical techniques (e.g. FTICR MS combined with

linear ITs) are used for research investigations,but are not available for routine applications.

The identification of macromolecular com-pounds by LC-MS in routine doping analysis is sofar restricted to HBOCs and synthetic insulin.

Human chorionic gonadotrophin

This gonadotrophic hormone stimulates theendogenous production of testosterone and istherefore prohibited in men. Quantification ofhCG by two different immunological assays isrecognised by WADA as sufficient analytical tech-nique. Different conventional cut-off values of10 or 25 IU/L were suggested (but not officiallyadopted) to discriminate normal values andpathological situations or misuse. An LC-MS pro-cedure using an IT MS in positive ESI mode wasfound to be suitable for a sensitive quantificationprocedure of hCG [50]. Immunoaffinity extrac-tion was used to enrich the peptide from theurinary matrix prior to the tryptic digestion ofthe glycoprotein. A residue of the β-subunit con-taining 17 amino acids was chosen as a signifi-cant marker for hCG; in particular, a distinctstructural specificity compared with the similarsubunit of the luteinising hormone was con-firmed. MRM experiments of the doubly chargedpeptide allowed a quantification down to thresh-old concentrations of 5 IU/L. Examination of theglycosylation of hCG revealed a considerablemicro-heterogeneity [51], which does not affectthe evaluation of doping cases because a reliabledifferentiation is possible based on a wide con-centration gap between endogenous and abnor-mal hCG levels in men.

Human growth hormone

Human GH stimulates the production of IGF-I,leading to a promotion of protein synthesis,increase of muscle mass, reduction of the amountof stored fat and an induction of the growth oflong bones. The secretion from pituitary glandsoccurs in three to five daily pulses, during whichthe basal serum concentrations (around 3 µg/L)are temporarily greatly exceeded and returnrapidly (half-life around 15 min) to normal.

Large molecules 207


Therefore, hGH serum concentrations are notsuitable for the detection of its abuse. Instead, acombined evaluation of concentrations of IGF-I,its binding protein (IGF-BP3) and bone markerswas proposed to reveal abuse of growth hormonein human athletes [52].

In horses, a cut-off serum concentration of700 µg/L of IGF-I was anticipated as the criterionfor growth hormone administration [53]. A sen-sitive LC-ESI-MS quantification procedure (limitof quantification = 30 µg/L IGF-I) was reported tobe eligible for routine screening. The use of Arg3-IGF-I, which is derived from IGF-I by replacementof Glu3, as IS, appears to be essential, due to itssimilar properties in the affinity chromatographyclean-up. Full-scan MS was applied to detect theintact pseudo-molecular ion ([M + 7H]7+, m/z1093.4) of IGF-I, which is used for quantificationin equine serum samples [53].

Another analytical approach is based on theevaluation of qualitative variations of growthhormone. There are several mass variants of thepredominant protein containing 191 aminoacids corresponding to a molecular mass of22 kDa. The most interesting one is a 20-kDamolecule, which is derived by deletion of residues32–46 and exhibits comparable physiologicalactivity. The concentration ratio of both variantsdoes not depend significantly on gender, age,body height and weight [54]. Administration ofany hGH isoform suppresses the endogenoussecretion of both variants. Accordingly, abuse ofcommercially available hGH may be detected byan elevated 22/20 kDa ratio, because it containsexclusively the 22-kDa variant. Studies to provethe appropriateness of this approach based onenzyme-linked immunoassays are documented[55]. Current alternative quantification pro-cedures of low-level proteins by LC-MS, based onsophisticated analytical equipment, e.g. FTICRMS, require extensive sample preparation and donot yet achieve the required sensitivity [10, 56].Identification of two tryptic peptides of hGH(22 kDa), which was obtained from digestion ofLC fractionated plasma, provided limits of detec-tion 10-fold higher than relevant clinical plasmaconcentrations [57]. Another approach wascapable of identifying clinically relevant levels(5 µg/L) from unfractionated plasma, applyingtagged peptides, isolated by affinity chromatog-

raphy [10]. The detection was carried out by HR-LC-MS (FTICR), which was coupled to a linear IT.

In another study, qualitative investigations onhGH were carried out using LC-MS-MS (FTICR).Based on the isolation and accurate mass measur-ing of the [M + 17H]17+, a deviation of themolecular mass of 4 Da (compared with non-post-translationally modified hGH) could bedetected and attributed to the formation of twodisulphide bonds [56].

The identification of structural modificationslike disulphide linkage or variations of glycosyl-ation, influencing proper folding of proteins,may be of diagnostic value in doping analysis.Present technical constraints (relatively high cut-off values and/or high amounts of sample mater-ial, requirement of high mass ranges and MSresolution, and high MS accumulation times thatare incompatible with typical LC peak widths andprohibit on-line coupling) seem to impede theapplication of LC-MS to the identification ofhGH abuse in the near future.

Erythropoietin

EPO is a 34-kDa glycoprotein hormone control-ling red blood cell production and is therefore apotent doping agent to enhance the oxygentransport capacity of blood. Recombinant humanEPO (rhEPO) has been available since 1989 andcannot be directly distinguished from en-dogenous EPO due to its identical amino acidsequence. Indirect methods (e.g. haematologicalparameters) served as an indicator for a potentialabuse. The variable glycosylation is determinedby post-translational modification, whichdepends on the availability of enzymes in therespective cells, resulting in a wide diversity ofcarbohydrates, typically terminated by one ortwo sialic acid residues.

The number and location of sialic acid residuesper molecule (Figure 9.5) determine the formationof quaternary isoforms and influence the plasmahalf-life of EPO. Removal of sialic acid groupsleads to a total inactivity, due to a rapid clearancefrom blood circulation, while new EPO variants(i.e. darbepoetin) provide a prolonged activity,triggered by the inclusion of additional oligo-saccharides, which are terminated by sialic acid.



This alteration is based on a modification of fiveamino acids of the polypeptide and respectivesubstances may be distinguished from naturalEPO by analysis of the primary structure. Anothervariant of EPO (SEPO) consists of a syntheticpeptide that is conjugated with a polymer,aiming for a reduction of the biological diversity.

The identification of EPO variants in dopingcontrol urine samples is to date based on a chemi-luminescence detection of its isoforms afterimmunological enrichment and electrophoreticseparation [58]. The logical MS approaches to anidentification in combination with LC or capil-lary electrophoresis are hampered by the limitedsensitivity.

Application of capillary electrophoresis- andLC-IT MS enabled the quantitative characterisa-tion of EPO reference material after cleavageand derivatisation. The relative amount of sialicacid per molecule of rhEPO was found to be17.6 mol/mol. This was supposed to provide apotential parameter to differentiate recombinantfrom endogenous EPO [59]. Alternatively, acharacteristic sulphation of N-linked oligo-saccharides in EPO molecules from different celllines was detected, based on MS examinations(negative-ionisation mode) after LC separation ofthe protein digest using graphitised carbon

columns [60]. An LC-ESI-MS investigation ofvarious commercially available EPO forms(epoetin α, epoetin β, darbepoetin) resulted inastonishingly good LC separation. The full-scanMS of the three variants shows characteristic par-ticularities in ESI (on-line micro LC coupling) aswell as in matrix-assisted laser desorption ionisa-tion (MALDI) mode (examination of the LC frac-tions) [61]. The injected quantity of protein wasabout 100 ng and, hence, far beyond the amountavailable in routine analyses. The limited MSsensitivity is due to the heterogeneity of glyco-sylation, because the ion abundance is spreadover a large number of individual molecules. Adeglycosylation of the EPO molecule leads to areduction of diversity and increase of MS sensitiv-ity. Analyses of deglycosylated EPO (rhEPO anddarbepoetin) by MALDI showed a good match ofits molecular mass with the assumed structure[62]. LC-ESI-MS experiments allowed a sensitiveidentification of [M + nH]n+ pseudo-molecularions (n = 5–8) of rhEPO after application of250 fmol on-column. The additional identifi-cation of peptide, obtained from endoproteaseGlu-C digestion, provides complementary struc-tural information and is supposed to permit asufficient identification of rhEPO and darbe-poetin in race horses and greyhounds [62].

Large molecules 209

N N

N G

G

G

S

G S

�-Fucose

�-Acetylneuraminic acid

(sialic acid)

�-Acetylglucosamine

�-Mannose

�-Galactose

S

1

Asn-24

Asn-38

165

S

SG

N

N

NM

M M

M

MMM

N

N

N N N

N

N

G G G

G

G

S

S

S

S

S S

S

Asn-83

Ser-126

N

N

N

N

G

G

G

G

MM

M

N N

F

F

F

F

Figure 9.5 Chemical structure of EPO.


Insulin

The presumptive abuse of insulin among athletesand body-builders is due to its stimulation ofendogenous protein synthesis. Several cases ofhypoglycaemia (i.e. significantly reduced levelsof blood sugar) were observed in fatal incidents inbody-building, indicating its high popularity inthis field. Insulin (Figure 9.6) belongs to thegroup of prohibited peptide hormones accordingto the recent WADA definition [1]. There arevarious medications available, based on struc-tural alterations of insulin. The main purpose ofthese structural variations consists in the regu-lation of its bioavailability. Insulin tends to self-association to biologically inactive hexamers,which can be suppressed by structural modifica-tions like switching of the positions of lysine andproline (B28 versus B29, Humalog) or replace-ment of proline with an aspartic acid residue(B29, Novolog). Alternatively, long-actinginsulin was synthesised by elevation of its iso-electric point (modification in A and B chains,Lantus).

The identification of intact insulin derivatives,carried out on a QTOF instrument equipped witha nano-ESI ion source, revealed the predominantformation of corresponding multiply chargedmolecules (z = 5–7). This enables a mass-specificdifferentiation of the synthetic derivatives fromhuman insulin [63] (Figure 9.7) after LC

separation. Alternatively, the identification of theB chains can be applied after reduction of thedisulphide bonds.

Haemoglobin-based oxygen carriers

Similar to the administration of EPO or syntheticperfluorocarbons, the intention of an abuse ofHBOCs in sports is an elevation of the oxygentransportation capacity of blood. Variouscross-linked bovine or human haemoglobinpreparations (e.g. Hemopure, Hemolink, Oxy-globin) are approved for the treatment of animalsor humans. The intact oxyglobin was reportedto be inadequate for an efficient confirmation ofhaemoglobin by LC-MS in blood samples,because of the formation of one single dominantfragment that was interfered with by impuritiesfrom haemolysed plasma. Sufficient variationsbetween the primary structures of human andbovine haemoglobin permit the differentiationof both proteins, based on the identification ofdiagnostic tryptic peptides. Specific peptidesresulting from α and β chains of haemoglobinwere found to be suitable for a differentiation ofboth species [64, 65]. Subunits of the α (residue69–90, 2367 Da) and the β chain (residue 40–58,2090 Da) were commonly used as markers forbovine haemoglobin. Human haemoglobin wasidentified by recording of a diagnostic peptide


COO−

COO−

11

21

1

Ser

IleCys

Ser

LeuTyr

GlnLeu

GluAsn

Gly

Ile

ValGlu

Gln Cys Cys Thr

AsnCys

Tyr

Phe

Val

Asn

Gln

HisLeu Cys

Gly

Ser

His

Glu

Val

Leu

Cys

Val

Leu

Tyr

Leu

Ala

Gly

Gly

Glu

Phe

Arg

ThrTyr

Phe

Thr

Lys

Pro1

11

21

B-chain

A-chain

NH3+

NH3+

Figure 9.6 Chemical structure of insulin.


Large molecules 211

5.5 5.6 5.7 5.8 5.9 6.0 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7.0

6.05

6.14

(a)

(b)

20.0%

18.0%

16.0%

14.0%

12.0%

Perc

enta

ge a

bundance

m/z

Retention time (min)

10.0%

8.0%

6.0%

4.0%

2.0%

0.0%

800 900 1000 1100 1200 13001050 1150

1167.01

1162.42

966.03

968.86

830.60

1250850 950

6.14

100%

Rela

tive

abu

nd

an

ce

Lantus

Novolog

Humalog

[M + 5H]5+

[M + 7H]7+

[M + 6H]6+

Figure 9.7 ESI full-scan spectrum of human insulin (a) generating the multiply charged molecules [M + 5H]5+, [M + 6H]6+

and [M + 7H]7+ at m/z 1162.4, 968.9 and 830.6, respectively. Extracted ion chromatogram (b) of a plasma samplefortified with Humalog (m/z 1162), Novolog (m/z 1166) and Lantus (m/z 1213) at 10 pmol/mL. Insulins were analysedas intact proteins. (Reproduced from Thevis et al. [63] with permission.)


from the β chain (residue 42–60, 2059 Da).Doubly or triply charged ions of these peptidesare suitable for LC-ESI-MS-MS screening. Theassay is able to identify, confirm and quantifyHBOCs in human and equine plasma samples.Respective assays were established using eitherQTOF or triple-quadrupole (QqQ) MS [64, 65].Sample preparation requires a laborious combi-nation of SPE, filtration of macromolecules andenzymatic digestion.

Summary

LC-MS has been introduced in different branchesof doping analysis.

Qualitative progress in the analysis of highlypolar substances (diuretics) could be achieved incombination with a significant reduction ofsample preparation. Other low-mass pharma-ceutical substances (stimulants, narcotics, β-blockers) may certainly be analysed by LC-MS,but well-established and validated GC assays arestill dominant.

Progress in steroid analysis is mainly focusedon substances with high proton affinity (polarsubstituents, conjugated double bonds, corti-costeroids, steroid conjugates). The identificationof relevant steroids by LC-MS has considerablyadvanced the classic GC-MS screening foranabolic steroids, but an extensive replacementof the conventional surveys is not likely.

Finally, there is increasing application ofLC-MS to the analysis of doping-relevant pro-teins. Synthetic (e.g. insulin) or non-human (e.g.haemoglobin) peptides may be positively dis-criminated from endogenous analogues, whereasrecent approaches to an unequivocal identifi-cation of abuse of recombinant EPO or growthhormone are promising, but so far insufficientlysensitive and reliable for routine application.

References

1. World Anti-Doping Agency. The Prohibited List2004. International Standard. Montreal: WADA,2004.

2. Houghton E, Dumasia MC, Wellby JK. The use ofcombined high performance liquid chromatog-raphy negative ion chemical ionization massspectrometry to confirm the administration ofsynthetic corticosteroids to horses. Biomed MassSpectrom 1981; 8: 558–564.

3. Garbis SD, Hanley L, Kalita S. Detection of thiazide-based diuretics in equine urine by liquid chroma-tography/mass spectrometry. J AOAC Int 1998; 81:948–957.

4. Bowers LD. Analytical advances in detection of per-formance-enhancing compounds. Clin Chem 1997;43: 1299–1304.

5. Bowers LD, Fregien K. HPLC/MS confirmation ofpeptide hormones in urine: an evaluation of detec-tion limits. In: Donike M, Geyer H, Gotzmann A,et al., eds, Recent Advances in Doping Analysis 6.Cologne: Sport und Buch Strauß, 1994: 175–184.

6. Thieme D, Große J, Lang R, et al. Detection ofmesocarb metabolite by LC-TS/MS. In: Donike M,Geyer H, Gotzmann A, et al., eds, Recent Advances inDoping Analysis 2. Cologne: Sport und Buch Strauß,1995: 275–284.

7. Ventura R, Fraisse D, Becchi M, et al. Approach tothe analysis of diuretics and masking agents byhigh-performance liquid chromatography-massspectrometry in doping control. J Chromatogr 1991;562: 723–736.

8. Krummen M, Hilkert AW, Juchelka D, et al. A newconcept for isotope ratio monitoring liquidchromatography/mass spectrometry. RapidCommun Mass Spectrom 2004; 18: 2260–2266.

9. Greig MJ, Bolanos B, Quenzer T, et al. Fourier trans-form ion cyclotron resonance mass spectrometryusing atmospheric pressure photoionization forhigh-resolution analyses of corticosteroids. RapidCommun Mass Spectrom 2003; 17: 2763–2768.

10. Wu SL, Choudhary G, Ramstrom M, et al. Evalu-ation of shotgun sequencing for proteomic analysisof human plasma using HPLC coupled with eitherion trap or Fourier transform mass spectrometry.J Proteome Res 2003; 2: 383–393.

11. Magnusson MO, Sandstrom R. Quantitative analy-sis of eight testosterone metabolites using columnswitching and liquid chromatography/tandemmass spectrometry. Rapid Commun Mass Spectrom2004; 18: 1089–1094.

12. Kawasaki N, Ohta M, Itoh S, et al. Usefulness ofsugar mapping by liquid chromatography/mass



spectrometry in comparability assessments of glyco-protein products. Biologicals 2002; 30: 113–123.

13. Berges R, Segura J, Ventura R, et al. Discriminationof prohibited oral use of salbutamol fromauthorized inhaled asthma treatment. Clin Chem2000; 46: 1365–1375.

14. Nelson RE, Grebe SK, DJ OK, et al. Liquid chroma-tography-tandem mass spectrometry assay forsimultaneous measurement of estradiol andestrone in human plasma. Clin Chem 2004; 50:373–384.

15. Kauppila TJ, Kostiainen R, Bruins AP. Anisole, a newdopant for atmospheric pressure photoionizationmass spectrometry of low proton affinity, lowionization energy compounds. Rapid Commun MassSpectrom 2004; 18: 808–815.

16. Ventura R, Segura J. Detection of diuretic agents indoping control. J Chromatogr B 1996; 687: 127–144.

17. World Anti-Doping Agency. Technical DocumentTD2004 MRPL. Montreal: WADA, 2004: 1–2.

18. Thieme D, Grosse J, Lang R, et al. Screening, confir-mation and quantification of diuretics in urine fordoping control analysis by high-performanceliquid chromatography-atmospheric pressureionisation tandem mass spectrometry. J ChromatogrB 2001; 757: 49–57.

19. Deventer K, Van Eenoo P, Delbeke FT. Simultaneousdetermination of beta-blocking agents and diuret-ics in doping analysis by liquid chromatography/mass spectrometry with scan-to-scan polarityswitching. Rapid Commun Mass Spectrom 2004; 19:90–98.

20. Deventer K, Delbeke FT, Roels K, et al. Screening for18 diuretics and probenecid in doping analysis byliquid chromatography-tandem mass spectrom-etry. Biomed Chromatogr 2002; 16: 529–535.

21. Goebel C, Trout G, Kazlauskas R. Rapid screeningmethod for diuretics in doping control usingautomated solid phase extraction and liquidchromatography-electrospray tandem mass spec-trometry. Anal Chim Acta 2003; 502: 65–74.

22. Lehner AF, Harkins JD, Karpiesiuk W, et al. Clenbu-terol in the horse: confirmation and quantitationof serum clenbuterol by LC-MS-MS after oral andintratracheal administration. J Anal Toxicol 2001;25: 280–227.

23. Thevis M, Opfermann G, Schanzer W. Liquidchromatography/electrospray ionization tandemmass spectrometric screening and confirmation

methods for beta2-agonists in human or equineurine. J Mass Spectrom 2003; 38: 1197–206.

24. Segura J, Ventura R, Jurado C. Derivatization pro-cedures for gas chromatographic-mass spectro-metric determination of xenobiotics in biologicalsamples, with special attention to drugs of abuseand doping agents. J Chromatogr B 1998; 713:61–90.

25. Thevis M, Opfermann G, Schanzer W. High speeddetermination of beta-receptor blocking agents inhuman urine by liquid chromatography/tandemmass spectrometry. Biomed Chromatogr 2001; 15:393–402.

26. Leinonen A, Kuuranne T, Kotiaho T, et al. Screeningof free 17-alkyl-substituted anabolic steroids inhuman urine by liquid chromatography-electro-spray ionization tandem mass spectrometry.Steroids 2004; 69: 101–109.

27. Leinonen A, Kuuranne T, Kostiainen R. Liquidchromatography/mass spectrometry in anabolicsteroid analysis – optimization and comparison ofthree ionization techniques: electrospray ioniza-tion, atmospheric pressure chemical ionizationand atmospheric pressure photoionization. J MassSpectrom 2002; 37: 693–698.

28. Guo T, Chan M, Soldin SJ. Steroid profiles usingliquid chromatography-tandem mass spectrometrywith atmospheric pressure photoionization source.Arch Pathol Lab Med 2004; 128: 469–475.

29. Catlin DH, Sekera MH, Ahrens BD, et al. Tetra-hydrogestrinone: discovery, synthesis, and detec-tion in urine. Rapid Commun Mass Spectrom 2004;18: 1245–1249.

30. Draisci R, Palleschi L, Marchiafava C, et al. Confir-matory analysis of residues of stanozolol and itsmajor metabolite in bovine urine by liquidchromatography-tandem mass spectrometry. JChromatogr A 2001; 926: 69–77.

31. Draisci R, Palleschi L, Ferretti E, et al. Confirmatoryanalysis of 17beta-boldenone, 17alpha-boldenoneand androsta-1,4-diene-3,17-dione in bovine urineby liquid chromatography-tandem mass spectrom-etry. J Chromatogr B 2003; 789: 219–226.

32. Buiarelli F, Cartoni GP, Coccioli F, et al. Determina-tion of trenbolone and its metabolite in bovinefluids by liquid chromatography-tandem massspectrometry. J Chromatogr B 2003; 784: 1–15.

33. Van Poucke C, Van Peteghem C. Developmentand validation of a multi-analyte method for the

References 213


detection of anabolic steroids in bovine urine withliquid chromatography-tandem mass spectrom-etry. J Chromatogr B 2002; 772: 211–217.

34. Wang C, Catlin DH, Demers LM, et al. Measure-ment of total serum testosterone in adult men:comparison of current laboratory methods versusliquid chromatography-tandem mass spectrom-etry. J Clin Endocrinol Metab 2004; 89: 534–543.

35. Reilly CA, Crouch DJ. Analysis of the nutritionalsupplement 1AD, its metabolites, and relatedendogenous hormones in biological matrices usingliquid chromatography-tandem mass spectrom-etry. J Anal Toxicol 2004; 28: 1–10.

36. Shackleton CH, Chuang H, Kim J, et al. Electrospraymass spectrometry of testosterone esters: potentialfor use in doping control. Steroids 1997; 62:523–529.

37. Kim JY, Choi MH, Kim SJ, et al. Measurement of19-nortestosterone and its esters in equine plasmaby high-performance liquid chromatography withtandem mass spectrometry. Rapid Commun MassSpectrom 2000; 14: 1835–1840.

38. Borts DJ, Bowers LD. Direct measurement ofurinary testosterone and epitestosterone conju-gates using high-performance liquid chromatog-raphy/tandem mass spectrometry. J Mass Spectrom2000; 35: 50–61.

39. Bowers LD, Sanaullah. Direct measurement ofsteroid sulfate and glucuronide conjugates withhigh-performance liquid chromatography-massspectrometry. J Chromatogr B 1996; 687: 61–68.

40. Kuuranne T, Kotiaho T, Pedersen-Bjergaard S, etal. Feasibility of a liquid-phase microextractionsample clean-up and liquid chromatographic/massspectrometric screening method for selectedanabolic steroid glucuronides in biologicalsamples. J Mass Spectrom 2003; 38: 16–26.

41. Kuuranne T, Aitio O, Vahermo M, et al. Enzyme-assisted synthesis and structure characterization ofglucuronide conjugates of methyltestosterone (17alpha-methylandrost-4-en-17 beta-ol-3-one) andnandrolone (estr-4-en-17 beta-ol-3-one) metabo-lites. Bioconjug Chem 2002; 13: 194–199.

42. Deventer K, Delbeke FT. Validation of a screeningmethod for corticosteroids in doping analysis byliquid chromatography/tandem mass spectrom-etry. Rapid Commun Mass Spectrom 2003; 17:2107–2114.

43. Fluri K, Rivier L, Dienes-Nagy A, et al. Method for

confirmation of synthetic corticosteroids in dopingurine samples by liquid chromatography-electro-spray ionisation mass spectrometry. J Chromatogr A2001; 926: 87–95.

44. Antignac JP, Le Bizec B, Monteau F, et al. Study ofnatural and artificial corticosteroid phase IImetabolites in bovine urine using HPLC-MS/MS.Steroids 2002; 67: 873–882.

45. Appolonova SA, Shpak AV, Semenov VA. Liquidchromatography-electrospray ionization ion trapmass spectrometry for analysis of mesocarb and itsmetabolites in human urine. J Chromatogr B 2004;800: 281–289.

46. Maralikova B, Weinmann W. Simultaneousdetermination of delta9-tetrahydrocannabinol, 11-hydroxy-delta9-tetrahydrocannabinol and 11-nor-9-carboxy-delta9-tetrahydrocannabinol in humanplasma by high-performance liquid chromatog-raphy/tandem mass spectrometry. J Mass Spectrom2004; 39: 526–531.

47. Thevis M, Opfermann G, Krug O, et al. Electrosprayionization mass spectrometric characterizationand quantitation of xanthine derivatives usingisotopically labelled analogues: an application forequine doping control analysis. Rapid CommunMass Spectrom 2004; 18: 1553–1160.

48. Todi F, Mendonca M, Ryan M, et al. The confirma-tion and control of metabolic caffeine in standard-bred horses after administration of theophylline.J Vet Pharmacol Ther 1999; 22: 333–342.

49. Tang FP, Leung GN, Wan TS. Analyses of quaternaryammonium drugs in horse urine by capillaryelectrophoresis-mass spectrometry. Electrophoresis2001; 22: 2201–2209.

50. Liu C, Bowers LD. Mass spectrometric characteriza-tion of the beta-subunit of human chorionicgonadotropin. J Mass Spectrom 1997; 32: 33–42.

51. Jacoby ES, Kicman AT, Laidler P, et al. Determinationof the glycoforms of human chorionic gonado-tropin beta-core fragment by matrix-assisted laserdesorption/ionization time-of-flight mass spec-trometry. Clin Chem 2000; 46: 1796–1803.

52. Sonksen PH. Insulin, growth hormone and sport.J Endocrinol 2001; 170: 13–25.

53. de Kock SS, Rodgers JP, Swanepoel BC. Growthhormone abuse in the horse: preliminary assess-ment of a mass spectrometric procedure for IGF-1identification and quantitation. Rapid CommunMass Spectrom 2001; 15: 1191–1197.



54. Tsushima T, Katoh Y, Miyachi Y, et al. Serum con-centrations of 20K human growth hormone innormal adults and patients with various endocrinedisorders. Study Group of 20K hGH. Endocr J 2000;47(Suppl): S17–S21.

55. Wu Z, Bidlingmaier M, Dall R, et al. Detection ofdoping with human growth hormone. Lancet 1999;353: 895.

56. Wu SL, Jardine I, Hancock WS, et al. A new andsensitive on-line liquid chromatography/massspectrometric approach for top-down proteinanalysis: the comprehensive analysis of humangrowth hormone in an E. coli lysate using a hybridlinear ion trap/Fourier transform ion cyclotronresonance mass spectrometer. Rapid Commun MassSpectrom 2004; 18: 2201–2207.

57. Wu SL, Amato H, Biringer R, et al. Targeted pro-teomics of low-level proteins in human plasma byLC/MSn: using human growth hormone as a modelsystem. J Proteome Res 2002; 1: 459–465.

58. Lasne F, Martin L, Crepin N, et al. Detection ofisoelectric profiles of erythropoietin in urine:differentiation of natural and administered recom-binant hormones. Anal Biochem 2002; 311:119–126.

59. Che FY, Shao XX, Wang KY, et al. Characterizationof derivatization of sialic acid with 2-aminoacridone and determination of sialic acidcontent in glycoproteins by capillary electro-phoresis and high performance liquid chromatog-raphy-ion trap mass spectrometry. Electrophoresis1999; 20: 2930–2937.

60. Kawasaki N, Haishima Y, Ohta M, et al. Structuralanalysis of sulfated N-linked oligosaccharides inerythropoietin. Glycobiology 2001; 11: 1043–1049.

61. Caldini A, Moneti G, Fanelli A, et al. Epoetin alpha,epoetin beta and darbepoetin alfa: two-dimensionalgel electrophoresis isoforms characterization andmass spectrometry analysis. Proteomics 2003; 3:937–941.

62. Stanley SM, Poljak A. Matrix-assisted laser-desorption time-of flight ionisation and high-performance liquid chromatography-electrosprayionisation mass spectral analyses of two glycosy-lated recombinant epoetins. J Chromatogr B 2003;785: 205–218.

63. Thevis M, Ogorzalek Loo RR, et al. Mass spectro-metric identification of synthetic insulin. In:Schänzer W, Geyer H, Gotzmann A, et al., eds,Recent Advances in Doping Analysis 11. Cologne:Sport und Buch Strauß, 2003: 227–237.

64. Guan F, Uboh C, Soma L, et al. Unique tryptic pep-tides specific for bovine and human hemoglobin inthe detection and confirmation of hemoglobin-based oxygen carriers. Anal Chem 2004; 76:5118–5126.

65. Thevis M, Ogorzalek Loo RR, Loo JA, et al. Dopingcontrol analysis of bovine hemoglobin-basedoxygen therapeutics in human plasma by LC-electrospray ionization-MS/MS. Anal Chem 2003;75: 3287–3293.

References 215



Documents

LC-MS in doping control - Pharmaceutical Press chapter.pdfdoping control inﬂuence analytical strategies in athletes drug testing: • The substance-based doping deﬁnition priori-tises