Content uploaded by Stéphane François Bermon
Author content
All content in this area was uploaded by Stéphane François Bermon on Jan 19, 2019
Content may be subject to copyright.
Page1of51
CirculatingTestosteroneastheHormonalBasisofSex1
DifferencesinAthleticPerformance2
Shorttitle:HyperandrogenismandFemaleAthletics3
4
DavidJ.Handelsman1,AngelicaLindénHirschberg2,StephaneBermon35
6
1ANZACResearchInstitute,UniversityofSydneyandDepartmentofAndrology,7
ConcordHospital,Sydney,NewSouthWales,Australia,8
2DepartmentofWomen´sandChildren´sHealth,KarolinskaInstitutetand9
DepartmentofGynecologyandReproductiveMedicine,KarolinskaUniversityHospital,10
Stockholm,Sweden11
3UniversitéCôted'Azur,LAMHESSNice,FranceandInternationalAssociationofAthletics12
Federations,HealthandScienceDepartment,Monaco13
14
Keywords:testosterone,sexdifference,athleticperformance,muscle,hemoglobin,disordersofsex15
differentiation16
17
Words:13,172 References:161 Tables:5 Figures:818
19
20
Correspondence:21
ProfDJHandelsman22
ANZACResearchInstitute23
SydneyNSW213924
Australia25
E:djh@anzac.edu.au26
27
28
29
Page2of51
Disclaimer:30
DJHisamedicalandscientificconsultantforInternationalAssociationofAthleticsFederations31
(IAAF)andmemberoftheInternationalOlympicCommittee(IOC)workinggroupon32
hyperandrogenicfemaleandtransgenderathletesandtotheAustralianSportsAnti‐Doping33
AgencyandisalsoamemberofWorldAnti‐DopingAgency(WADA)’sHealth,Medicineand34
ResearchCommittee.HehasreceivedinstitutionalgrantsupportfromBesinsHealthcareand35
Lawleyforinvestigator‐initiatedclinicalstudiesintestosteronepharmacologyandhasprovided36
experttestimonyintestosteronelitigation.37
ALHisamedicalandscientificconsultantfortheSwedishOlympicCommitteeandamemberof38
theIAAFandIOCworkinggroupsonhyperandrogenicfemaleathletesandtransgenderathletes.39
ShehasreceivedgrantsupportfromIAAFforastudyontestosteroneandphysicalperformancein40
women.41
SBisamedicalandscientificconsultantfortheIAAFandamemberoftheIAAFandIOCworking42
groupsonhyperandrogenicfemaleathletesandtransgenderathletes43
Theauthorshavenootherinvolvementwithanyentityhavingafinancialinterestinthematerial44
discussedinthemanuscript.Opinionsexpressedinthispaperarethepersonalviewoftheauthors45
anddonotrepresentthoseoftheIAAF,IOC,WADAorSwedishOlympicCommittee.46
47
48
49
Acknowledgements:TheauthorsaregratefulforhelpfulinsightsandcommentsfromAlan50
VernecandOsquelBarroso(WADA),PeterHarcourt(AFL,FIBA)andRichardBudgett(IOC).51
Page3of51
Abstract52
Eliteathleticcompetitionshaveseparatemaleandfemaleeventsduetomen’sphysicaladvantagesin53
strength,speedandendurancesothataprotectedfemalecategorywithobjectiveentrycriteriaisrequired.54
Priortopuberty,thereisnosexdifferenceincirculatingtestosteroneconcentrationsorathletic55
performancebutfrompubertyonwardssexdifferenceinathleticperformanceemergesascirculating56
testosteroneconcentrationsriseinmenbecausetestesproduce30timesmoretestosteronethanbefore57
pubertywithcirculatingtestosteroneexceeding15‐foldthoseofwomenatanyage.Thereisawidesex58
differenceincirculatingtestosteroneconcentrationsandreproducibledose‐responserelationshipbetween59
circulatingtestosteroneandmusclemassandstrengthaswellascirculatinghemoglobininbothmenand60
women.Thesedichotomieslargelyaccountsforthesexdifferencesinmusclemassandstrengthand61
circulatinghemoglobinlevelsresultinginatleastan8‐12%ergogenicadvantageinmen.Suppressionof62
elevatedcirculatingtestosteroneofhyperandrogenicathletesresultsinnegativeeffectsonperformance,63
whicharereversedwhensuppressionceases.Basedonthenon‐overlapping,bimodaldistributionof64
circulatingtestosteroneconcentration(measuredbyliquidchromatography‐massspectrometry)and65
makingallowanceforwomenwithmildhyperandrogenismincludingthatofpolycysticovariansyndrome,66
whoareover‐representedineliteathletics,theappropriateeligibilitycriterionforfemaleathleticevents67
shouldbeacirculatingtestosteroneoflessthan5.0nmol/L.Thiswouldincludeallwomenotherthanthose68
withuntreatedhyperandrogenicdisordersofsexualdevelopment(DSD),testosterone‐treatedfemale‐to‐69
male(F2M)transgender,noncompliantmale‐to‐female(M2F)transgenderorandrogendoping.70
71
Page4of51
1.Background72
73
Virtuallyallelitesportsaresegregatedintomaleandfemalecompetitions.Themainjustificationistoallow74
womenachancetowin,aswomenhavemajordisadvantagesagainstmenwhoare,onaverage,taller,75
stronger,fasterandhavegreaterenduranceduetotheirlarger,strongermusclesandbonesaswellasahigher76
circulatinghemoglobin.Hence,elitefemalecompetitionformsaprotectedcategorywithentrythatmustbe77
restrictedbyanobjectiveeligibilitycriterionrelatedbynecessitytotherelevantsex‐specificphysical78
advantages.Thepracticalneedtoestablishaneligibilitycriterionforelitefemaleathleticcompetitionledthe79
InternationalAssociationofAthleticFederations(IAAF)toestablisharulein2011,endorsedbythe80
InternationalOlympicFederation(IOC)in2012,forhyperandrogenicwomen.ThefirstIAAFregulationstated81
thatforathletestobeeligibletocompleteinfemaleevents,theathletemustbelegallyrecognisedasafemale82
and,unlessshehascompleteandrogeninsensitivity,maintainserumtestosteronelessthan10nmol/L.That83
IAAFeligibilityrulewaschallengedbyanathletetotheCourtforArbitrationinSports(CAS)whichruledin201584
that,althoughaneligibilitycriterionwasjustified,thescientificgroundsfortheoriginalIAAFrulewas85
consideredinsufficient,notablyintheextentofthecompetitiveadvantageenjoyedbyhyperandrogenic86
athleteswhohadcirculatingtestosteronegreaterthan10nmo/L.TheCASsuspendedthehyperandrogenism87
eligibilityrulependingreceiptofsuchevidence.Inthatcontext,thepresentpaperreviewstheavailable88
evidenceonthehormonalbasisforsexdifferencesinathleticperformance.Itconcludesthattheevidence89
justifiedarevisedeligibilitycriterionofathresholdcirculatingtestosteroneconcentrationof5nmol/L90
(measuredbyamassspectrometrymethod).91
92
2.Sex,FairnessandSegregationinSport93
94
Ifsportisdefinedastheorganizedplayingofcompetitivegamesaccordingtorules(1),fixedrulesare95
fundamentalinrepresentingtheboundariesoffairsportingcompetition.Rulebreaking,whetherby96
breachingeligibilityorcompetitionrules,suchasuseofbanneddrugs,illegalequipmentormatchfixing,97
createsunfaircompetitiveadvantagesthatviolatesfairplay.Cheatingconstitutesafraudagainstnotjust98
competitorsbutalsospectators,sponsors,thesportandthepublic.Intheabsenceofgenuinefair99
competition,elitesportwouldloseitswidepopularappealandabilitytocaptivateandinspirewiththe100
authenticattractionofgenuinecontestbetweenhighlytrainedathletes.101
102
Nevertheless,fairnessisanelusive,subjectiveconceptwithmalleableboundariesthatmaychangeover103
timeassocialconceptsoffairnessevolve.Forexample,untilthelate19thcenturywhenorganizedsports104
trainersemerged,trainingitselfwasconsideredabreachoffairnesssincecompetitionwasenvisagedat105
thattimeasacontestbasedsolelyonnaturalendowments.Similarly,sportsoncedistinguishedbetween106
amateursandprofessionals.Theconceptoffairnesshasdeepandcomplexphilosophicalrootsmainly107
Page5of51
focusedonnotionsofdistributivejustice.Theseconsiderationsimpactonsportthroughtheuniversal108
applicationofanti‐discriminationandhumanrightslegislation.Lessattentionisgiventothephilosophical109
basisoffaircompetitioninelitesportwheretheobjectivesarenotegalitarianbutaimtodiscovera110
hierarchyofachievementderivedfromamixtureofunequalnaturaltalentandindividualtrainingeffort.111
Excellent,insightfuldiscussionofthelegalandmoralcomplexitiesofsexandfaircompetitioninelitesports112
fromalegalscholarandformerelitefemaleathleteisavailable(2).113
114
Thetermssexandgenderareoftenconfusedandusedasifinterchangeable.Sexisanobjective,specific115
biologicaltermwithdistinct,fixedfacets,notablygenetic,chromosomal,gonadal,hormonaland116
phenotypic(includinggenital)sex,eachofwhichhasacharacteristicdefinedbinaryform.Whileallfacets117
ofbiologicalsexarealmostalwaysalignedsothatassignmentofsexatbirthisstraightforward,rare118
instanceswhereanytwoormorefacetsofbiologicalsexconflictconstituteanintersexstate,nowreferred119
toasDisorders(orDifferences)ofSexDevelopment(DSD)(3).Bycontrast,genderisasubjective,malleable,120
self‐identifiedsocialconstructwhichdefinesaperson’sindividualgenderroleandorientation.Prompted121
bybiological,personalandsocietalfactors,volitionalexpressionofgendercantakeonvirtuallyanyform122
limitedonlybytheimaginationwithsomeindividualsassertingtheyhavenotjustasinglenatalgenderbut123
twogenders,none,adistinctthirdgenderorgenderthatvaries(fluidly)fromtimetotime.Hence,while124
genderisusuallyconsistentwithbiologicalsexasassignedatbirth,inafewitcandifferduringlife.For125
example,ifgenderwerethebasisforeligibilityforfemalesports,anathletecouldconceivablybeeligibleto126
competeatthesameOlympicsinbothfemaleandmaleevents.Thesefeaturesrendertheunassailable127
personalassertionofgenderidentityincapableofformingafair,consistentsexclassificationinelitesport.128
129
Thestrongestjustificationforsexclassificationinelitesportisthatafterpubertymenproduce20times130
moretestosteronethanwomen(4‐7)resultingincirculatingtestosteroneconcentrations15timeshigher131
thaninchildrenorwomenofanyage.Age‐gradecompetitivesportingrecordsshownosex‐related132
advantagespriortopubertyonwards,whereasfromtheageofmalepubertyonwardsthereisastrongand133
ongoingmaleadvantage(8).Thestrikingmalepost‐pubertalincreaseincirculatingtestosteroneprovidesa134
major,ongoing,cumulativeanddurablephysicaladvantageinsportingcontestsbycreatinglargerand135
strongerbones,greatermusclemassandstrength,andhighercirculatinghemoglobinaswellaspossible136
psychological(behavioural)differences.Inconcert,theserenderwomen,onaverage,unabletocompete137
effectivelyagainstmeninpower‐basedorendurance‐basedsports.138
139
Sexclassificationinsportthereforerequiresproofofeligibilityasonlywomenshouldcompeteinthe140
protected(female)category.Thisdeceptivelysimplerequirementforfairnessistakenforgrantedbypeer141
femalecompetitorswhoregardparticipationbymales,orathleteswithphysicalfeaturesclosely142
resemblingmales,asunfair.Thismakespolicingofeligibilityinescapableforsportstoavoidunfairmale143
Page6of51
participationinfemaleevents.Yet,suchpolicinginevitablyintrudesintohighlypersonalmatterssothatit144
mustbeachievedwithrespectfordignityandprivacydemandinguseoftheleastinvasive,scientifically145
reliablemeans.Unsurprisinglythisdilemmahasalwaysbeenhighlycontentioussinceitfirstentered146
internationalelitesportsintheearly20thcenturyandithasbecomeincreasinglyprominentand147
contentiousinrecentdecades;nevertheless,therequirementtomaintainfairplayinfemaleeventswillnot148
disappearaslongasseparatefemalecompetitionsexist.Overrecentdecadestherehasbeenprogressively149
betterunderstandingofthecomplexbiologyofgeneticsexdeterminationandtheimpactofpubertal150
sexualmaturationinestablishingphenotypicsexualdichotomyinphysicalcapabilities.Thesesex151
dichotomousphysicalfeaturesformthebasisof,butremainquitedistinctfrom,adultgenderrolesand152
identity.Overthelastcenturyasknowledgegrewtheattemptstoformalizeascientificbasisforthe153
unavoidablenecessityofpolicingeligibilityforthefemalecategoryhavebeencontinuallychallenged.Most154
recently,theincreasingassertionofgenderself‐identificationasasocialcriterionhasfurtherchallenged155
thehegemonyofbiologyfordetermining“sportssex”,Coleman’saptterm(2).Allowingsubjectivegender156
self‐identificationtobecomethesolecriterionofsportssexwouldallowforgamingandperceptionsof157
systematicunfairnesstogrow.Thecaseforwomen’ssportsbeingdefinedbysexratherthangender,158
includingtheconsequencesofaccedingtogender‐basedclassificationhavebeenoutlined(9)inarguingthe159
importanceofpropermedicalmanagementofathletesintendingtocompeteinfemaleevents.160
161
Separatemaleandfemaleeventsinsportisadominantformofclassificationthatissuperimposedon162
othergraduatedagegroupandweight(e.g.weightlifting,powerlifting,wrestling,boxing,rowing)163
classifications,whichreflectdifferencesinstrength,power,speedtoensurefairnessintermsof164
opportunitytowinand,additionally,safetyincontactsports.Ageandweightclassificationsrelyon165
objectivecriteria(birthdate,weigh‐inweight)foreligibilityasnecessarilyshouldsexclassification.166
Nevertheless,somepowersportsdependentonexplosivestrengthandpower(egthrowingevents,167
sprinting)donotsegregateweightclasses,whileothersportswhereheightisanadvantage(egbasketball,168
jockeys)donothaveheightclassifications.Thesesportsdisproportionatelyattractathleteswithgreater169
weightand/orpower‐to‐weightratiooradvantageousstature,respectively.Ifsexclassificationwere170
eliminatedsuchopenormixedcompetitionswouldbedominatedalmostexclusivelybymen.Ittherefore171
seemshighlyunlikelythatsexclassificationwouldeverbediscardeddespitecallsonphilosophicalor172
sociologicalgroundstoend“gender”classificationinsport(10).173
174
3.Sexdifferenceincirculatingtestosteronelevels175
3.1Testosteronebiosynthesis,secretionandregulationinmenandwomen176
Anandrogenisahormonecapableofdevelopingandmaintainingmasculinecharacteristicsinreproductive177
tissues(notablythegenitaltract,andothertissuesandorgansassociatedwithsecondarysexual178
Page7of51
characteristicsandfertility)andcontributingtotheanabolicstatusofnon‐reproductivebodytissues(11).179
Thetwodominantbioactiveandrogenscirculatinginmaturemammals,includinghumans ‐‐testosterone180
anditsmorepotentmetabolite,dihydrotestosterone(DHT)‐‐accountforthedevelopmentandmaintenance181
ofallandrogen‐dependentcharacteristics,andtheircirculatinglevelsinmenandnon‐pregnantwomenarise182
fromsteroidssynthesizeddenovointhetestes,ovaryoradrenals(12).183
Thesexuallyundifferentiatedgonadsintheembryodevelopintoeitherovariesortestesaccordingto184
whetheraychromosome(oratleastthesrygene)ispresent.Afterbirthanduntilpubertycommences,185
circulatingtestosteroneconcentrationsareessentiallythesameinboysandgirls,otherthanbrieflyinthe186
neonatalperiodofboyswhenhigherlevelsprevail.Theonsetofmalepuberty,abrain‐drivenprocess187
triggeredbyastillmysterioushypothalamicorhighercerebralmechanism(13),initiatesthehormonal188
cascadeofpuberty.Inmalesthisleadstoenhancedpituitaryluteinizinghormone(LH)secretionthat189
stimulatesthe500millionLeydigcellsinthetestestosecrete3‐10mg(mean7mg)oftestosteronedaily(4,190
6,7,14,15).Thiscreatesaveryhighlocalconcentrationoftestosteronewithinthetestisaswellasasteep191
downhillconcentrationgradientintothebloodstreamthatmaintainscirculatingtestosteronelevelsatadult192
malelevels,whicharetightlyregulatedbystrongnegativehypothalamicfeedbackofcirculating193
testosterone.However,intheabsenceoftestesthesemechanismsdonotoccurinfemales.Ingirls,serum194
testosteroneincreasesduringpuberty(16),peakingatage20‐25yearsbeforedeclininggraduallywithage(17,195
18)butitremainslessthan2nmol/Latallages,asdeterminedbyareliablemethod(seebelow).Inadult196
women,circulatingtestosteroneisderivedfromthreeroughlyequalsources–directsecretionfromtheadrenal197
glandortheovaryaswellasindirectlyfromextra‐glandularconversion(inliver,kidney,muscle,fat,skin)from198
testosteroneprecursorssecretedbytheadrenalandovary.However,incombinationthesedifferentsources199
produceabout0.25mgoftestosteronedailysothatthroughoutlifewomenmaintaincirculatingtestosterone200
levelsoflessthan2nmol/L.Circulatingtestosteroneconcentrationsinwomenaresubjecttolittledynamic201
physiologicalregulation.Asaresult,circulatingtestosteroneconcentrationsinhealthypre‐menopausal202
womenarestable(non‐fluctuating)andnotsubjecttostrongnegativefeedbackbyexogenoustestosterone203
likemen.Eventhesmallrise(50%)atthetimeofthemid‐cycleLHsurgetriggeringovulation(19),remains204
withinthephysiologicalrangeforpre‐menopausalfemales.Insummary,onlywhencirculatingtestosterone205
concentrationsinmaleadolescentsrisesabovethecirculatingpre‐pubertalconcentrationsdoesthe206
virilisationcharacteristicofmencommence,progressandremainthroughoutadultlifeatleastuntiloldage207
(18).208
209
3.2Maleandfemalereferencerangesforcirculatingtestosterone210
211
Areliablethresholdforcirculatingtestosteronemustbesetusingmeasurementbythereferencemethodof212
liquid(orgas)chromatography‐massspectrometry(LC‐MS)ratherthanusingoneofthevariousavailable213
commercialtestosteroneimmunoassays.Thenecessaryrelianceonsteroidmassspectrometryforclinical214
Page8of51
applicationsinendocrinology,reproductivemedicineandsportsmedicineiswidelyrecognized.Ithasbeen215
standardfordecadesinanti‐dopingscience(20)andthegrowingconsensusisthatitisrequiredforhigh216
qualityclinicalresearchandpracticerecognizedbycognateprofessionalsocieties(21,22)andeditorialsin217
leadingclinicalendocrinology(23)andreproductivemedicine(24)journals.Theinherentlylimitedspecificity218
oftestosteroneimmunoassaysarisesfromantibodycross‐reactivitywithstructurallyrelatedsteroids(suchas219
precursorsandmetabolites)otherthantheintendedtarget.Asaresult,allsteroidimmunoassaysincluding220
fortestosteronedisplaymethod‐specificbiaswhereby,forexample,thelowerlimitofatestosterone221
referencerangeinhealthyyoungmenvariesfrom7.3to12.6nmol/Laccordingtotheimmunoassayused,so222
thatnoconsensusdefinitionofalowerlimitcouldbeobtainedindependentofthecommercialimmunoassay223
methodused(25).Further,testosteroneimmunoassaysareoptimizedforcirculatinglevelsinmenbutdisplay224
increasinginaccuracyatthelower,byanorderofmagnitude,circulatingtestosteroneconcentrationsin225
womenorchildren.Incontrasttoimmunoassays,LC‐MSbasedmethodsarehighlyspecificanddonot226
dependonproprietaryantibodies.UsingLC‐MS‐basedmeasurements,method‐specificbiascanbeavoided227
andafixedconsensuslowerreferencelimitdefined(seetable1).Hence,fortheprecisionrequiredin228
sportsmedicine,whetherforeligibilitycriteriaoranti‐dopingapplications,testosteroneinserummustbe229
measuredbyLC‐MSmethods.230
231
Priortopuberty,levelsofcirculatingtestosteroneasdeterminedbyLC‐MSarethesameinboysandgirls232
(16)aswellasremaininglowerthan2nmol/Linwomenofallages.However,fromtheonsetofmale233
pubertythetestessecrete20timesmoretestosteroneresultingincirculatingtestosteronelevelsthatare234
15timesgreaterinhealthyyoungmenthanagesimilarwomen.UsingLC‐MSmeasurement,circulating235
testosteroneinadultshasastrikingly,non‐overlappingbimodaldistributionwithwideandcomplete236
separationbetweenmenandwomen.Table1summarisesdatafromappropriatereportedstudiesusing237
MS‐basedmethodstomeasureserumtestosteroneinhealthymenandwomen.Basedonanumber‐238
weightedpoolingwithconventional95%two‐sidedconfidencelimitsoftheeightavailablestudiesusingLC‐239
MSmeasurementsofserumtestosterone,thelimitsofthereferencerangeforhealthyyoungmen(18to240
40years)is7.7nmol/Lto29.4nmol/L.Similarly,summarisingthenineavailablestudiesforhealthy241
menstruatingwomenunder40years,the95%(twosided)referencerangeis0to1.7nmol/L.These242
referencelimitsneglectfactorssuchasoralcontraceptiveuse(26,27),menstrualphase(19),SHBG(28,243
29),overweight(30,31),fastingandsmoking(32),aswellasdiet(31)andphysicalactivity(33,34)in244
womenandmen,allofwhichhavesmalleffectsoncirculatingtestosteronebutwithoutmaterially245
influencingthedivergencebetweenthenon‐overlappingbimodaldistributionofmaleandfemale246
referencerangesofcirculatingtestosterone.247
248
Increatingathresholdforeligibilityforfemaleeventsitisalsonecessarytomakeallowancefor249
hyperandrogenicwomenincludingwomenwithpolycysticovarysyndrome(PCOS)andnon‐classical250
Page9of51
adrenalhyperplasia.PCOSisarelativelycommondisorderamongwomenofreproductiveageswitha251
prevalenceof6‐10%,dependingonthediagnosticcriteriaused(35),inwhichmildhyperandrogenismisa252
keyclinicalfeatureandhashigherthanexpectedprevalenceamongelitefemaleathletes(26,36‐38).Non‐253
classicaladrenalhyperplasiaisamilderandlater(adult)onsetvariantofclassicalcongenitaladrenal254
hyperplasia(39)withamuchhigherbutstillrarepopulationprevalence(1:1000vs1:16,000fortheclassical255
variant(40).Table2summarisesclinicalstudies(n=16,≥40women)reportingserumtestosterone256
concentrationsmeasuredbyLC‐MSinsamplesfromwomenwithPCOS.Thepooleddatarevealsthatthe257
upperlimitofserumtestosteroneinwomenwithPCOSis3.1nmol/L(95%confidenceinterval,onesided)258
or4.8nmol/L(usinga99.99%confidenceinterval,onesided)(table3).Henceaconservativethresholdfor259
circulatingtestosteroneof5nmol/LmeasuredbyLC‐MSwouldidentifyfewerthan1:10,000womenwith260
PCOSasfalsepositives,basedoncirculatingtestosteronemeasurementalone.Circulatingtestosterone261
higherthanthisthresholdislikelytobeduetotestosterone‐secretingadrenalorovariantumors,262
intersex/DSD,badlycontrolledornon‐compliantM2Ftransgenderathletesortestosteronedoping.263
264
3.3Thephysiologicaleffectsoftestosteronedependonthecirculatingtestosterone,notitssource265
(endogenousorexogenous)266
Testosterone,whetherofnaturalendogenousormanufacturedexogenoussource,hasanidenticalchemical267
structureandbiologicaleffects,asidefromminordifferencesinisotopiccompositionwhicharebiologically268
insignificant.Regardlessofitssource,atequivalentdosesandcirculatinglevels,exogenoustestosterone269
exertsthesamebiologicalandclinicaleffectsoneveryknownandrogen‐responsivetissueororgan,apart270
fromeffectsonspermatogenesis,whichasdiscussedbelowisonlyamatterofdegree.Consequently,271
exogenoustestosteroneisafullyeffectivesubstituteforendogenoustestosteroneintherapeuticuse,272
counteringtheeffectsoftestosteronedeficiencyduetohypogonadism(reproductivesystemdisorders).Any273
purporteddifferencesbetweenendogenousandexogenoustestosteroneare,likethedifferencesbetween274
menandwomen,duetocorrespondingdifferencesintheendogenousproductionrateorexogenousdose.275
Suchdifferencesineffectiveexposureleadtocorrespondingdifferencesincirculatingtestosteronelevels276
anditseffectsaccordingtothedose‐responsecurvesfortestosterone.277
Likeallhormonesanddrugs,overtheireffectiverangeofbiologicalactivitythedose‐responserelationship278
fortestosteroneisusuallyasigmoidalcurvewithlowerandupperplateausjoinedbyamonotonicallyrising279
middleregion,whichmaybelinearinthenaturalscalebutmoreoftenlog‐linear(linearonthelogorsimilar280
transformedscale).Inthemiddleportionofthetypicalsigmoidaldose‐responsecurveforthesameincrease281
intestosteronedose(orconcentration),theresponsewouldbeincreasedinsimpleproportional(ielinear)282
butmoreoftenonalogarithmicscale.Bycontrast,atthelowerandupperplateausofdoseorconcentrations,283
changesintestosteroneexposuremayevokeminimalornoresponseontheendpoint.Forexample,in284
womenofanyagecirculatingtestosteroneconcentrationsarealongthelowerplateauofthedose‐response285
Page10of51
curve,sothatincreasesincirculatingtestosteroneconcentrationswithinthatlowerplateaumayhave286
minimalornoeffect.InfemaleathleteswiththemildhyperandrogenismofPCOS,higherperformancehas287
beenshown(38)withtheirmusclemassandpowerperformancecorrelatingwithandrogenlevels(26).288
However,beyondtheseeffectswhereendogenoustestosteroneconcentrationsareinthehigh‐normaladult289
femalerange,itisonlywhentheincreasesincirculatingtestosteroneconcentrationssubstantiallyand290
consistentlyexceedthoseprevailinginchildhood(<2nmol/L)andamongwomenincludingthosewithPCOS291
(<5nmol/L)thattheeffectswouldreplicaterisingtestosteroneconcentrationsofboy’sinmid‐ tolate292
puberty(typically>8nmol/l)whichcausethemasculinizingeffectsofincreasedmuscle,boneand293
hemoglobincharacteristicsofmen.Asshownabove,thecirculatingtestosteroneofmostwomennever294
reachesconsistentlyabove5nmol/L,alevelwhichboysmustsustainforsometimetoexhibitthe295
masculinizingeffectsofmalepuberty.296
Secondarily,theeffectsoftestosteronearemodulatedinaformoffinetuningbythepatternsofexposure,297
suchaswhetherthecirculatingtestosteroneisdeliveredintheun‐physiologicalsteady‐stateformat(e.g.298
quasi‐steadystatedeliverybyimplantortransdermalproducts)orbythepeak‐and‐troughdeliveryof299
injectionsasopposedtothenaturalstateofendogenousfluctuationsinserumtestosteronearoundthe300
averageadultmalelevels.However,theselatterpatterneffectsaresubtleandthedominanteffectremains301
thatofdoseandaveragetestosteroneconcentrationsinblood,howevertheyarise.Furthermore,thereis302
evidencethattheandrogensensitivityofresponsivetissuesdifferandmaybeoptimalatdifferentcirculating303
testosteroneconcentrations(41).304
Malesexualfunctionismaintainedbyendogenoustestosteroneatadultmalecirculatingconcentrations.305
Theseeffectscanbereplicatedbyexogenoustestosteroneifandonlyifitachievescomparablecirculating306
testosteroneconcentrations.Forexample,inawell‐controlledprospectivestudyofoldermenwithprostate307
cancer(42),androgendeprivationachievingcastratelevelsofcirculatingtestosteronesustainedover12308
monthsmarkedlysuppressedsexualdesireandfunction,whereasthoseeffectsdidnotoccurinage‐matched309
menhavingnon‐hormonaltreatmentforprostatecancerorthosewithoutprostatecancer.Inhealthy310
youngermenwhoseendogenoustestosteroneisfullysuppressed,theirsexualfunctioncompletelyrecovers311
whencirculatingtestosteronewasrestoredtothephysiologicalmalerangebyadministrationofexogenous312
testosterone(43).Similareffectswerealsoobservedinhealthy,middle‐agedmeninwhommalesexual313
functionwasfullymaintained(comparedwithplacebo)during2yearsoftreatmentwithanexogenous314
androgen(DHT)despiteitcausingsustained,completesuppressionofendogenoustestosterone(44).This315
furthersupportsthekeyinterpretationthatthebiologicaleffectsofexogenousorendogenoustestosterone316
arethesameatcomparablecirculatinglevels.317
318
Clinically,exogenoustestosteronereplicatesfullyalleffectsofendogenoustestosteroneonevery319
reproductiveandnon‐reproductiveorganortissue,withthesoleexceptionofthetestis.Spermproduction320
Page11of51
inthetestisrequiresaveryhighconcentrationoftestosterone(typically100timesgreaterthaninthegeneral321
bloodstream),whichisproducedinnatureonlybytheactionofthepituitaryhormoneLH.LHstimulatesthe322
Leydigcellsintheinterstitialspaceofthetestisbetweenseminiferoustubulestoproducehighintra‐testicular323
concentrationsoftestosterone,whicharenecessaryandsufficienttoinitiateandmaintainspermproduction324
intheadjacentseminiferoustubules.Thishighconcentrationoftestosteronealsoprovidesadownhill325
gradienttosupplytherestofthebody,wherecirculatingtestosteroneactsonandrogen‐responsivetissues326
tomaintainmasculinepatternsofandrogenization.Whenexogenoustestosterone(oranyotherandrogen)327
isadministeredtomen,pituitaryLHissuppressedbynegativefeedbackandthespermproductionhaltsfor328
aslongasexogenoustestosteroneorandrogenexposurecontinues,afterwhichitrecovers(45).However,329
eventhereductioninspermatogenesisandtestissizewhenmenaretreatedwithexogenoustestosteroneis330
onlyamatterofdegree.Itiswellestablishedinrodents(46,47)thatspermatogenesisisinducedby331
exogenoustestosteroneifthetestosteroneconcentrationsinthetestisarehighenoughtoreplicatewhat332
occursnaturallyviaLHstimulation(48).However,directreplicationthathighdosetestosteronealsoinitiates333
andmaintainsspermatogenesisinhumansisnotfeasibleasthesetestosteronedosesare10‐100times334
higherthancouldbesafelygiventohumans.Nevertheless,confirmatoryevidenceinhumansisavailable335
fromrarecasesofmenwithanactivatingmutationoftheCG/LHreceptor(49,50).Thismutationcauses336
autonomoustesticulartestosteronesecretionleadingtoprecociouspubertyarisingfromthepremature337
adultmalecirculatingtestosteroneconcentrationswhichleadtocompletesuppressionofcirculating338
gonadotropin(LH,FSH)secretion.Inthisillustrativecasethetestiswasexposedtonon‐physiologicallyhigh339
testosteroneconcentrations(butwithoutanygonadotropinstimulation)whichinducedspermproduction340
andallowedfornaturalpaternity(49).Thisindicatesthatevenforspermatogenesis,exogenoustestosterone341
canreplicateallbiologicaleffectsofendogenoustestosteroneinaccordancewiththerelevantdose‐response342
characteristics.343
Themostrealisticviewisthatincreasingcirculatingtestosteronefromthechildhoodorfemalerangetothe344
adultmalerangewillhavethesamephysiologicaleffectswhetherthesourceoftheadditionaltestosterone345
isendogenousorexogenous.Thisisstronglysupportedbywell‐establishedknowledgeabouttherelationship346
ofcirculatingtestosteroneconcentrationswiththetimingandmanifestationsofmalepuberty.The347
characteristicclinicalfeaturesofmasculinisation(musclegrowth,increasedheight,increasedhemoglobin,348
bodyhairdistribution,voicechangeetc)appearonlyifandwhencirculatingtestosteroneconcentrationsrise349
intotherangeofmalesatmid‐pubertywhicharehigherthaninwomenatanyageevenaftertherisein350
circulatingtestosteroneinfemalepuberty.Ifandonlyifthepubertalriseincirculatingtestosteronefails,351
themalesaffectedareclinicallyconsideredhypogonadal.Suchafailureofmalepubertymayoccurfor352
geneticreasons(arisingfrommutationsthatinactivateanyofthecascadeofproteinswhoseactivityiscritical353
inthehypothalamustotriggermalepuberty)orasaresultofacquiredconditions,causedbypathological354
disordersofthehypothalamusorpituitaryorfunctionaldefectsarisingfromseveredeficitsofenergyor355
Page12of51
nutrition(egextremeovertraining,undernutrition),thelatterbeingcomparablewithhypothalamic356
amenorrheaoranorexianervosainfemaleathletes/balletdancers.Ifmalepubertyfails,testosterone357
replacementtherapyisfullyeffectiveinreplicatingtheallthedistinctivemasculinefeaturesapartfrom358
spermatogenesis.359
360
361
3.4ElevatedcirculatingtestosteroneconcentrationcausedbyDSDs362
RaregeneticintersexconditionsknownasDSDscanleadtomarkedlyincreasedcirculatingtestosteronein363
womenand,whencoupledwithambiguousgenitaliaatbirth,appearingasundervirilizedmale,orvirilized364
females.Thiscancauseathleteswhowereraisedandidentifyaswomentohavecirculatingtestosterone365
levelscomparablewithmenandmuchexceedingthatofnon‐DSD(andnon‐doped)women,includingthose366
withPCOS.Keycongenitaldisordersinthiscategoryare46XYDSDsnamely5αreductasedeficiency(51),367
17β‐hydroxysteroiddehydrogenasetype3deficiency(52),androgeninsensitivity(53,54)aswellas368
congenitaladrenalhyperplasia(55),whichisa46XXDSD.Thereisevidencethatthefirstthreeconditions,369
componentsof46XYDSDs,are140timesmoreprevalentamongelitefemaleathletesthanexpectedinthe370
generalpopulation(56).371
Genetic5αreductasedeficiencyisduetoaninactivatingmutationinthe5αreductasetypeIIenzyme(51).372
ThisleadstoadeficitofDHTduringfetallifewhenDHTisrequiredforconvertingthesex‐undifferentiated373
embryonicandfetaltissuetoformthesex‐differentiatedmasculineformexternalgenitalia.Althoughgenetic374
males(46XY)with5αreductasedeficiencywilldeveloptestes,theyusuallyremainundescendedandlabial375
fusiontoformascrotumandphallicgrowthdoesnotoccur.Henceatbirththeexternalgenitaliamayappear376
feminine,leadingtoafemaleassignednatalsex.Thus,individualswith5α reductasedeficiencymayhave377
malechromosomalsex(46XY),gonadalsex(testes),andhormonalsex(adultmaletestosterone378
concentrations),butsuchseverelyunder‐virilizedgenitaliathataffectedindividualsmayberaisedfrombirth379
asfemalesratherthanasunder‐virilizedmales.However,fromtheonsetofmalepuberty,testicularLeydig380
cellsstartproducinglargeamountsoftestosterone,andthesteepriseincirculatingtestosteronetoadult381
malelevels(withthepermissiveroleof5αreductaseactivity)leadstomasculinevirilisation,includingmale382
patternsofmuscleandbonegrowth,hemoglobinlevelsandothermasculinebodyhabitusfeatures(hair383
growthpattern,voicechange),aswellasphallicgrowth(56).Suchchangesofmalepubertypromptaround384
halfaffectedindividualswhohadfemalesexassignedatbirthanddevelopedasgirlspriortopubertytoadopt385
amalegenderidentityandrole(57).Spermareformedinthetestessothat,usinginvitrofertilization,these386
individualsmayfatherchildren(58).387
Seventeenβ‐hydroxysteroiddehydrogenasetype3deficiency(52)hasasimilarnaturalhistoryto5α388
reductasedeficiency.Thisdisorderisduetoinactivatingmutationsinasteroidogenicenzymeexpressedonly389
inthetestisandwhichisessentialfortestosteroneformationinthefetus.Intheabsenceofafunctional390
Page13of51
enzyme,thetestismakeslittletestosteronebutinsteadsecreteslargeamountsofandrostenedione,the391
steroidimmediatelypriortotheenzymaticblock.Inthecirculation,theexcessofandrostenedioneis392
convertedtotestosterone(mainlybytheenzymeAKR1C3(12)).Althoughthecirculatingtestosteroneisthen393
convertedtocirculatingDHT,insufficientDHTisformedlocallywithintheurogenitalsinustovirilisegenitalia394
atbirth.Thiscausesthesamesevereunder‐virilisationoftheexternalgenitaliaofgeneticallymale395
individuals,leadingtoambiguousgenitaliaatbirthdespitemalechromosomal,gonadalandhormonalsex.396
Whenpubertyarrives,thetestesstartproducingtheadultmaletestosteroneoutputthisleadstomarked397
virilisationandsubsequentassumptionofamalegenderidentitybysomeaffectedindividuals,conflicting398
withafemaleassignednatalsexandchildhoodupbringing.399
Androgeninsensitivity,whicharisesfrommutationintheandrogenreceptor(AR),posesdifferentbut400
complexchallengesforeligibilityforfemaleathleticevents.AstheARislocatedontheXchromosome,401
geneticmales(46XY)arehemizygous,sothataninactivatingmutationintheARcanbepartiallyorfully402
insensitivetoandrogenaction.Affectedindividualshavemaleinternalgenitalia(testesintheinguinalcanal403
orabdomenwithWolffianducts)andconsequentlyadultmalecirculatingtestosteroneconcentrationsafter404
puberty.Thesenon‐lethalmutationshaveawidespectrumoffunctionaleffects,rangingfromfullresistance405
toallandrogenactionincompleteandrogeninsensitivitysyndrome(CAIS)whereindividualshaveafull406
femalephenotypewithnormalfemaleexternalgenitalia,topartialandrogeninsensitivitysyndrome(PAIS)407
wheresomeandrogenactionisstillexertedleadingtovariousdegreesofambiguousgenitalia,ortomild408
androgeninsensitivitywhichproducesaverymild,under‐virilisedmalephenotype(normalmalegenitaland409
somaticdevelopmentbutwithlittlebodyhairandnomalepatternbalding)(53).Testosterone(and410
dihydrotestosterone)havenoconsistenteffectofinducingnormalnitrogenretention(anabolic)responses411
inpatientswithCAIS(59‐62)althoughsomereducedandrogenresponsivenessisretainedbypatientswith412
PAIS(60,63‐66).AthleteswithCAIScanfairlycompeteasfemalesbecausethecirculatingtestosterone,413
althoughatadultmalelevels,hasnophysiologicaleffectsothat,intermsofandrogenactionandtheensuing414
physicalsomaticadvantagesofmalesex,affectedindividualsareindistinguishablefromfemalesandgainno415
benefitsofthesexdifferencearisingfromunimpededtestosteroneaction.Amorecomplexissueariseswith416
athleteshavingPAISreflectingthedegreeofincompleteimpairmentofARfunction.Residualandrogen417
actioninsuchARmutationsishardertocharacterisequantitativelyasthereisnostandardized,objectivein418
vitrotesttoquantifyARfunctionality.Hence,althoughindividualswithPAISmayhaveadultmalecirculating419
testosteroneconcentrationsbutvariableandrogensensitivity,atpresentthisrequiresacase‐by‐case420
evaluation,primarilybasedonthedegreeofvirilisation.Thecurrentbestavailableclinicalapproachto421
determiningthefunctionalimpact(degreeoffunctionality/sensitivity)ofanARmutationisbasedonthe422
degreeofsomatic,primarilygenital,virilisationassessedaccordingtotheQuigleyclassificationofgradeof423
androgensensitivity(67).424
Congenitaladrenalhyperplasia(CAH)isarelativelycommondefectinadrenalsteroidogenesisinthe425
Page14of51
enzymaticpathwayleadingtosynthesisofcortisol,aldosteroneandsexsteroidprecursors.Thedisease426
variesinseverityfromlife‐threatening(adrenalfailure)tomild(hirsutismandmenstrualirregularity),oreven427
asymptomaticandundiagnosed.ThemostcommonmutationscausingCAHoccurinthe21hydroxylase428
enzyme,accountingfor95%ofcases(55).Thedefectleadstoabottleneck,creatingamajorbackingupof429
precursorsteroidswhichthenoverflowintoothersteroidpathways,leadingtodiagnostichighlevelsof17430
hydroxyprogesteroneand,infemalepatients,excessivecirculatingtestosteroneorotheradrenal‐source431
androgenprecursors(egandrostenedione,DHEA)whichmaybeconvertedtotestosteroneintissues.A432
commonclinicalproblemwithmanagementofCAHisthatglucocorticoid/mineralocorticoidtreatmentisnot433
alwaysfullyeffectivepartlyduetovariablecompliance,whichmayleavehighcirculatingtestosterone,434
includingwellintoorevenabovethenormalmalerange(68).Itisunlikelythatmildnon‐classicalcongenital435
adrenalhyperplasiaisamajorcontributortothemildhyperandrogenismprevalentamongelitefemale436
athletes.TheprevalenceofPCOS(6‐16%)isabout100timeshigherthanmildnon‐classicalcongenitaladrenal437
hyperplasia(0.1%,(40))whileadisproportionatelyhighnumberofelitefemaleathletes(especiallyinpower438
sports)havePCOS(36).Inonestudyofhyperandrogenicfemaleathletes,evenmildNCAHwasruledoutby439
normal17hydroxyprogesterone(26)andinanother(38)reportedserumandrostenedioneandcortisoldid440
notdifferfromcontrols,rulingoutsignificantcongenitaladrenalhyperplasia..441
442
4.Sexdifferenceinmuscle,hemoglobin,boneandathleticperformancerelatingtoadult443
circulatingtestosteroneconcentrations444
445
Followingpuberty,testosteroneproductionincreases(16)butremainsbelow2nmol/Linwomenwhereas446
inmentestosteroneproductionincreases20‐fold(from0.3mgadayto7mgaday)leadingtoa15‐fold447
highercirculatingtestosteroneconcentrations(15vs1nmol/L).Thegreatermagnitudeofsexdifferencein448
testosteroneproduction(20fold)comparedwithcirculatinglevels(15fold)isduetowomen’shigher449
circulatingSHBG,whichretardstestosteroneclearancecreatingaslowercirculatinghalf‐timeof450
testosterone.Thisorderofmagnitudedifferenceincirculatingtestosteroneconcentrationsisthekeyfactor451
tomen’ssuperiorathleticperformanceduetoandrogeneffectsprincipallyonmuscle,boneandhemoglobin.452
453
4.1Muscle454
4.1.1Biology:455
Ithasbeenknownsinceancienttimesthatcastrationinfluencesmusclefunction.Modernknowledgeofthe456
molecularandcellularbasisforandrogeneffectsonskeletalmuscleinvolveseffectsduetoandrogen457
(testosterone,DHT)bindingtotheandrogenreceptorwhichthenreleaseschaperoneproteins,dimerizes458
andtranslocatesintothenucleustobindtoandrogenresponseelementsinthepromoterDNAofandrogen459
sensitivegenes.Thisleadstoincreasesin(a)musclefibrenumbersandsize,(b)musclesatellitecellnumbers,460
Page15of51
(c)numbersofmyonuclei,and(d)sizeofmotorneurons(69).Additionallythereisexperimentalevidence461
thattestosteroneincreasesskeletalmusclemyostatinexpression(70),mitochondrialbiogenesis(71),462
myoglobinexpression(72)andinsulin‐likegrowthfactor(IGF‐I)content(73)whichmayaugmentenergetic463
andpowergenerationofskeletalmuscularactivity.464
Customizedgeneticmousemodelscanprovideuniquephysiologicalinsightintargetingspecificmoleculesor465
theirreceptorstoprovideexperimentalinsightintomammalianphysiologywhichisunobtainablebyhuman466
experimentation.Thetightevolutionaryconservationofthemammalianreproductivesystemexplainswhy467
geneticmousemodelshaveprovidedconsistent,highfidelityreplicationofthehumanreproductivesystem468
(74,75).Geneticmales(46XY)withandrogeninsensitivitydisplayingsimilarfeaturesoccurthrough469
spontaneouslyoccurringinactivatingARmutationsinallmammalianspeciesstudiedincludinghuman,where470
theyareknownaswomenwithCAIS.Theconverse,geneticfemales(46XX)resistanttoallandrogenaction,471
cannotoccurnaturallyinhumansorothermammals.Thisisbecausefullyandrogenresistantfemalesmust472
havebothXchromosomescarryinganinactivatedAR.InturnthisrequiresacquiringoneXchromosomefrom473
theirfather.However,thepotentialfathersaresterileashemizygousmalesbearingasinglecopyanX474
chromosomewithaninactiveARproducenosperm,asafunctionalARisbiologicallyindispensablefor475
makingsperminanymammal.However,androgenresistantfemalescanbebredbygeneticengineering476
usingtheCre‐Loxsystem(76).Animportantfindingfromsuchstudiesisthatandrogen‐resistantfemalemice477
haveessentiallythesamemusclemassandfunctioncomparedwithwild‐typeandrogensensitivefemales478
bearingnormalARwhereasandrogen‐resistantmalemicehavesmallerandweakermusclemassand479
functionthanwild‐typemalesbutarecomparableinsteadwiththemuscleofwild‐typefemales(77).This480
indicatesthatandrogenaction,representedbycirculatingtestosterone,isthekeydeterminantofthehigher481
musclemassandstrengthcharacteristicofmalescomparedwithfemales.Furthermore,endogenous482
circulatingtestosteronehasminimaleffectsonskeletalmusclemassandstrengthinfemalemice.Although483
theseexperimentscannotbereplicatedinhumans,theirkeyinsightisthatthehighercirculatingtestosterone484
inmalesisthedeterminantofthemale’sgreatermusclemassandfunctioncomparedwithfemales.485
Nevertheless,thereisalsoevidencethathyperandrogenicwomen,mostlywithPCOS,haveincreasedmuscle486
massandstrengththatcorrelateswithmildlyincreasedcirculatingtestosteroneinthehigh‐normalfemale487
range(26,38).488
489
4.1.2Observationaldata:490
Thereisaclearsexdifferenceinbothmusclemassandstrength(78‐80)evenadjustingforsexdifferencesin491
heightandweight(80,81).Onaverage,womenhave50‐60%ofmen'supperarmmusclecross‐sectional492
area(CSA)and65‐70%ofmen'sthighmuscleCSA;andwomenhave50‐60%ofmen'supperlimbstrength493
and60‐80%ofmen'slegstrength(82).Youngmenhaveonaverageaskeletalmusclemassofover12kg494
greaterthanage‐matchedwomenatanygivenbodyweight(80,81).Whilenumerousgenesand495
Page16of51
environmentalfactors(includinggenetics,physicalactivityanddiet)maycontributetomusclemass,the496
majorcauseofthesexdifferenceinmusclemassandstrengthisthesexdifferenceincirculating497
testosterone.498
Age‐gradecompetitivesportsrecordsshowminimalornofemaledisadvantagepriortopuberty,whereas499
fromtheageofmalepubertyonwardsthereisastrongandongoingmaleadvantage.Correspondingtothe500
endogenouscirculatingtestosteroneincreasinginmalesafterpubertyto15‐20nmol/L(sharplydiverging501
fromthecirculatinglevelsthatremain<2nmol/Linfemales),maleathleticperformancesgofrombeingequal502
onaveragetothoseofage‐matchedfemalesto10‐12%betterinrunningandswimmingevents,and20%503
betterinjumpingevents(8)(figure1).CorroborativefindingsareprovidedbyaNorwegianstudythat504
examinedperformanceofadolescentsincertainathleticeventsbutwithoutreferencetocontemporaneous505
circulatingtestosteroneconcentrations(83).Thestrikingpost‐pubertalincreaseinmalecirculating506
testosteroneprovidesamajor,ongoing,cumulativeanddurableadvantageinsportingcontestsbycreating507
atleastgreatermusclemassandstrengthsuchthatthesesexdifferencesrenderwomenunabletocompete508
effectivelyagainstmen,especially(butnotonly)inpowersports.509
Thesefindingsaresupportedbystudiesofnon‐athleticwomenshowingthatmusclemassisincreasedin510
proportiontocirculatingtestosteroneinwomenwithmildlyelevatedtestosteronelevelsduetoPCOS(84,511
85),aconditionwhichismoreprevalentamongelitefemaleathleteswhoexhibitthesefeatures(26,36,38),512
oftenundiagnosed(37),butwhichmayprovideanergogenicadvantage(38),consistentwiththegraded513
effectsofcirculatingtestosteroneonexplosiveperformanceinmenandwomen(86).514
Studiesofelitefemaleathletesfurthercorroboratethesefindings.Onestudydemonstratesdose‐response515
effectsofbetterperformanceinsome(400m,400mhurdles,800mrunning,hammerthrow,polevault)but516
notallathleticeventscorrelatedwithsignificantlyhigherendogenoustestosteroneinfemale,butnotmale,517
athletes.Evenwithinthelowcirculatingtestosteronelevelsprevailingwithinthenormalfemalerange,in518
theseeventstherewasasignificantadvantageof1.8%to4.5%amongthoseinthehighestcomparedwith519
thelowesttertileofendogenoustestosterone(27).Afurtherstudyofelitefemaleathletescorroboratesand520
extendstheseobservationsinthatendogenousandrogensareassociatedwithamoreanabolicbody521
compositionaswellasenhancedmuscularperformance(26).Inthisstudy106SwedishOlympicfemale522
athleteswerecomparedwith117age‐andweight(BMI)‐matchedsedentarycontrolwomenfortheirmuscle523
andbonemass(bydualenergyX‐rayabsorptiometry,DEXA),theirmuscularstrength(squatand524
countermovementjumps),andtestosteroneandDHT,aswellasandrogenprecursors(DHEA,525
androstenedione)andurinaryandrogenglucuronidemetabolites(androsterone,etiocholanolone,3and17526
3α‐diols)measuredbyliquidchromatography‐massspectrometry(26).Theathletesdisplayedhighermuscle527
(andbone)massthanthesedentarycontrolwomen,withstrengthtestscorrelatingstronglywithmuscle528
masswhetherintotalorjustinthelegs.Inturn,musclemassandstrengthwerecorrelatedwithandrogens529
andandrogenprecursors.Consideringthatsuchstudiesmaybeconfoundedbyfactorssuchasmenstrual530
Page17of51
phaseanddysfunction,andheterogeneoussportsdisciplines,whichweakenthepowerofthestudy,these531
findingscanberegardedasquiterobust.532
4.1.3Interventionaldata:533
Dose‐responsestudiesshowthat,inmenwhoseendogenoustestosteroneisfullysuppressed,add‐back534
administrationofincreasingdosesoftestosteronethatproducegradedincreasesincirculatingtestosterone,535
causesadose‐dependent(whetherexpressedaccordingtotestosteronedoseorcirculatinglevels)increase536
inmusclemass(measuredasleanbodymass)andstrength(41,87).Takentogether,thesestudiesprovethat537
testosteronedosesleadingtocirculatingconcentrationsfromwellbelowtowellabovethenormalmale538
rangehaveunequivocaldose‐dependenteffectsonmusclemassandstrength.Thesedatastronglyand539
consistentlysuggestthatthesexdifferenceinleanbodymass(muscle)islargely,ifnotexclusively,dueto540
thedifferencesincirculatingtestosteronebetweenmenandwomen.Thesefindingshavestrongimplications541
forpower‐dependentsportperformanceandlargelyexplainthepotentefficacyofandrogendopinginsport.542
Thekeyfindingsprovidingconclusiveevidencethattestosteronehasprominentdose‐responseeffectsin543
menarereportedinstudiesbyBhasinetalthatprovedamonotonicdose‐response,extendingfromsub‐to544
supra‐physiologicalrangeformenfortestosteroneeffectsonmusclemass,sizeandstrengthinhealthy545
youngmen,findingsthathavebeenreplicatedandconfirmedbyanindependentgroup(41).Bothsetsof546
studiesusedacommondesignoffullysuppressingallendogenoustestosterone(tocastratelevels)forthe547
fulldurationoftheexperimentbyadministeringaGnRHanalog.IntheBhasinstudies,participantswerethen548
randomizedtofivegroupswhoreceivedweeklyinjectionsof25mg,50mg,125mg,300mgor600mgof549
testosteroneenanthatefor20weeks.Ineffectthiswastwosub‐andtwosupra‐physiologicaltestosterone550
doses.Inthesestudies,thelowesttestosteronedoseproducedameanserumtestosteroneof253ng/dl(8.8551
nmol/L)inyoungermenand176ng/dl(6.1nmol/L)inoldermen.Thestudiesshowedaconsistentdose‐552
responseformusclemassandstrengththatwasclearlyrelatedtotestosteronedoseandconsequentialblood553
testosteroneconcentrations(upperpanel,figure2).554
ThestudyofFinkelsteinetalinvolvedthesamedesignandinvolved400healthymenaged20to50yearsof555
agewhohadcompletesuppressionofendogenoustestosteroneforthe16weeksofthestudywith556
testosteroneaddedbackusingdailydosesof0,1.25g,2.5g,5gor10gofatopical1%testosteronegel(41).557
Thisagaincreatedagradeddose‐responsecurveforserumtestosteroneandformusclemassandstrength.558
Theinclusionofazero(placebo)dosealloweddifferentiationbetweenthezeroandlowesttestosterone559
dose.Theplacebo(zero)doseproducedaserumtestosteroneof0.7nmol/L,thetypicalmeanforcastrated560
men,childhood,andwomenofanyage.Meanwhilethelowesttestosteronedose(1.25ggelperday)561
producedaserumtestosteroneof6.9nmol/L,whichisequivalenttothatofamaleinearlytomid‐puberty.562
Akeyfindingforthisreviewisthat,fromthisstudyofmen,theincreaseinserumtestosteronefrommean563
ofnormalfemaleconcentration(0.9nmol/L)tosupra‐physiologicalfemaleconcentrations(6.9nmol/L)564
producedsignificantincreasesof2.3%fortotalbodylean(muscle)mass,3.0%forthighmusclearea,and565
Page18of51
5.5%increaseinlegpressstrength(digitiseddatapoolingbothcohortsfromlowerpanel,figure2).566
Studiesoftheergogeniceffectsofsupra‐physiologicalconcentrationsofcirculatingtestosteronerequire567
studiesadministeringgradeddosesofexogenoustestosteroneformonths.Duetoethicalconcerns568
regardingrisksofunwantedvirilisationandhormone‐dependentcancers,however,fewstudieshave569
administeredsupra‐physiologicaltestosteronedosestohealthywomen.Onewelldesigned,randomized570
placebo‐controlledstudyofpostmenopausalwomeninvestigatedtheeffectsofdifferenttestosteronedoses571
onmusclemassandperformanceandphysicalfunction(88).Sixty‐twowomen(meanage53)allhada572
standardestrogen‐replacementdoseadministeredduringa12weekrun‐inperiod(toeliminateany573
hypotheticalconfoundingeffectsofestrogendeficiency),afterwhichtheywererandomizedtooneoffive574
groupsreceivingweeklyinjectionsoftestosteroneenanthate(doses:0,3mg,6.25mg,12.5mg,and25mg575
respectively)for24weeks.Theincreasingdosesoftestosteroneproducedanexpecteddose‐responsein576
serumtestosteroneconcentrations(byLC‐MS)withthehighesttestosteronedose(25mg/week)produced577
ameannadirconcentrationof7.3nmol/L.Thewomenwhosetestosteroneconcentrationswereincreased578
to7.3nmol/Lachievedsignificantincreasesinmusclemassandstrength(table4),rangingfrom4.4%for579
muscle(lean)masstobetween12%and26%formeasuresofmusclestrength(chestandlegpress,loaded580
stairclimb).Asmusclestrengthmeasurementiseffort‐dependent,theplacebo‐controlleddesignofthe581
Huangstudysupportthefurtherinterpretationthatthehighestdoseoftestosteronealsohadprominent582
mentalmotivationaleffectsintheeffort‐dependenttestsofmusclestrength.Thesefindingsprovidesalient583
directevidenceoftheergogeniceffectsofhyperandrogenisminfemaleathletesconfirmingthatatleastup584
toaveragecirculatingtestosteroneconcentrationsof7.3nmol/L,womendisplayasimilardose‐response585
relationshipasdomenforsupra‐physiologicaltestosteronewithsignificantgainsinmusclemassandpower.586
587
Theseeffectsoftestosteroneadministrationoncirculatingtestosteroneconcentrationsinfemalesmaybe588
comparedwiththeeffectsinmalesfromtheFinkelsteinandBhasinstudies.Inmen,thelowesttestosterone589
dose(1.25g/day)increasedmeanserumtestosteroneto6.9nmol/Lequivalenttoearlytomid‐malepuberty590
resultinginsignificantincreasesoftotalbodylean(muscle)mass(2.3%),thighmusclearea(3.0%),andleg591
pressstrength(5.5%)comparedwiththeplacebodosewhichresultedinaserumtestosteroneof0.7nmol/L.592
IntheHuangstudy(figure3),musclemassandstrengthinpostmenopausalwomendisplayedaflatresponse593
atthe3lowerdoses,whencirculatingtestosteroneconcentrationsremainbelow5nmol/L,anddisplayeda594
significantincreaseonlywhenthemeancirculatingtestosteroneconcentrationproducedbythehighest595
testosteronedosefirstincreasedcirculatingtestosteroneconcentrationsabove5nmol/L.Thispattern,flat596
atlowerdosesandrisingathighestdose,representsthelowerplateauandtheearliestrisingportion,597
respectively,ofthesigmoidaldose‐responsecurveoftestosteroneformuscle.598
DatacorroboratingtheHuangstudyresultscomesfromanotherwell‐controlledstudyinwhichpost‐599
menopausalwomenwhowereadministeredmethyltestosteronefollowingarun‐inperiodofestrogen600
Page19of51
replacementdisplayedasignificantincreaseinlean(muscle)massaswellasupperandlowerlimbpower601
duringa16‐weekdouble‐blind,parallelgroupstudy(89).602
Similarly,twoprospectivestudiesofthefirst12monthstreatmentoftransmen(F2Mtransgender)showsa603
consistentmajorincreaseinmusclemassandstrengthduetotestosteroneadministration.Inonestudy604
testosteronetreatmentof17transmenachievingadultmalecirculatingtestosteronelevels(mean31nmol/L)605
increasedmusclemassby19.2%(90)whereas,conversely,testosteronesuppression(usinganestrogen‐606
basedtreatmentregimen)in20transwomenreducedcirculatingtestosteronelevels