Bone Mineral Density in Transgender Adolescents Treated With Puberty Suppression and Subsequent Gender-Affirming Hormones.

AND RELEVANCEInthiscohortstudy,afterlong-termuseofGAH, zscoresin individualstreatedwithpubertysuppressioncaughtupwithpretreatmentlevels,exceptfor thelumbarspineinparticipantsassignedmaleatbirth,whichmighthavebeenduetolow estradiolconcentrations.ThesefindingssuggestthattreatmentwithGnRHagonistsfollowed bylong-termGAHissafewithregardtobonehealthintransgenderpersonsreceiving testosterone,butbonehealthintransgenderpersonsreceivingestrogenrequiresextra attentionandfurtherstudy.Estrogentreatmentshouldbeoptimizedandlifestylecounseling providedtomaximizebonedevelopmentinindividualsassignedmaleatbirth. JAMA Pediatr.2023;177(12):1332-1341.doi: 10.1001/jamapediatrics.2023.4588 PublishedonlineOctober30,2023. Supplemental content Author Affiliations:Author affiliationsarelistedattheendofthis article. Corresponding Author:MariaAnna TheodoraCatharinavanderLoos, MD,DepartmentofEndocrinology andCenterofExpertiseonGender Dysphoria,AmsterdamUniversity MedicalCenter,POBox7057, Amsterdam1007MB,the Netherlands(m.vanderloos@ amsterdamumc.nl). Research JAMA Pediatrics | Original Investigation 1332 (Reprinted) jamapediatrics.com © 2023 American Medical Association. All rights reserved. Downloaded from jamanetwork.com by University of Amsterdam user on 09/29/2025 B onemassacquisitionoccursduringchildhoodandado- lescenceandreachesapeakattheendofthesecondde- cade of life. Inadequate acquisition of peak bone mass (PBM)mayresultinagreaterriskoffracturesandosteoporosis inadulthood. 1Peakbonemassandmaintenanceofbonemin- eral density (BMD) are influenced by a number of factors, of whichsexsteroidsareakeyfactor. 2Forexample,treatmentwith agonadotropin-releasinghormone(GnRH)agonisttosuppress estrogenproductioninadultswithendometriosiswasfoundto reduceBMD. 3Inchildren,GnRHagonistshavemainlybeenused totreatcentralprecociouspuberty.Previousresearch4 inthisfield revealed a decrease in bone mineral accrual during treatment. However,aftertreatmentcessation,accrualresumedandBMD wassimilartopeersbylateadolescence. 4 A relatively new field wherein GnRH agonists are used is thetreatmentoftransgenderadolescents.Transgenderpeople perceivegreatdistressduetoamismatchbetweentheirgen- der identity and sex assigned at birth, known as gender dys- phoria. For some, the development of secondary sex charac- teristicsduringpubertyimposesasubstantialstrain.Around the year 2000, treatment with GnRH agonists was intro- duced as a means of puberty suspension to lighten the bur- den of undesired physical changes while extending the time forexplorationoftreatmentwisheswithsubsequentgender- affirming hormones (GAH). A decrease in BMDzscores (ie, a decreasecomparedwiththemeanBMDofage-,sex-,andeth- nicity-matchedpersons)wasseenduringGnRHagonisttreat- mentintransgenderadolescents. 5-10 However,zscoresatage 22 years did not return to pretreatment scores after addition ofGAH, 11whichcouldindicateapermanentlossofbonemin- eralaccrualpotentialduetopostponementofpubertyormight be a reflection of delayed achievement of PBM. Evidence on long-termbonehealthintransgenderadolescentstreatedwith aGnRHagonistandsubsequentGAHislimited.Duetothepos- siblyincreasedriskofosteoporosisandassociatedincreased fracture risk in the case of lower PBM, it is important to gain more long-term insight into the bone health of transgender adolescentswhoweretreatedwithaGnRHagonist.Inthisco- hortstudy,wethereforeinvestigatedBMDdevelopmentand BMD zscoresintransgenderadultsafterlong-termGAHwith prior use of a GnRH agonist in adolescence. Additionally, we investigated whether estradiol, testosterone, or vitamin D concentrations;BMI;anddurationofmonoGnRHagonisttreat- mentwereassociatedwith zscoresatfollow-up.

This study was submitted for review of ethical and legal aspects to the local medical ethics committee of Amsterdam University Medical Center (UMC), Vrije Universiteit Amster- dam, Amsterdam, the Netherlands. The committee con- firmed that the Medical Research Involving Human Subjects Actdidnotapplytoourstudyowingtoalackofinterventions. 12 Hence,anofficialapprovalofourstudywasnotrequired.All participantsprovidedwritteninformedconsent.Thisstudyfol- lowedtheStrengtheningtheReportingofObservationalStud- iesinEpidemiology( STROBE)reportingguideline. Study Design and Population Thisstudywasconductedatthesamecenterasthestudyby Klink et al11 and contained some overlapping participants. However, the currentstudy population was larger and had a markedly extended follow-up time. In this prospective fol- low-up cohort study, transgender persons were invited to participateiftheyhadstartedmedicaltransitionwithaGnRH agonistbeforetheageof18yearsandhadsubsequentlyused GAH for at least 9 years. Individuals were selected from the Amsterdam Cohort of Gender Dysphoria, 13 a retrospectively builtdatabasecontainingallpeoplevisitingthegenderiden- tity clinic of Amsterdam UMC from 1972 until December 31, 2018. People with disorders of sex development or preco- cious puberty were not included. Eligible participants were approachedbymail,email,and/ortelephone.Recruitmenttook placebetweenMarch31,2020,andAugust31,2021. During the study visit at the outpatient clinic of the Am- sterdam UMC, clinical data (including ethnicity, age, Tanner stage, body mass index [BMI; calculated as weight in kilo- gramsdividedbyheightinmeterssquared]atthestartoftreat- ment,andwhethergonadectomyhadbeenperformed),blood samples,andadual-energyx-rayabsorptiometry(DXA)scan wereobtained.Ethnicitydatawererelevanttothestudyout- comebecauseBMD zscoresareethnicitydependent. Toassesschangeovertime,retrospectiveDXAscanswere collected from the medical records. At our center, the medi- cal protocol for transgender adolescents includes prescribed DXAaspartofstandardcareatthestartofGnRHagonistand GAHtreatmentandatages22and25years. ThemainoutcomeswerearealBMDandBMD zscoresfor 3 regions of interest at 4 time points. Additionally, we ana- lyzedBMDTscoresatfollow-up(ie,comparedwiththemean BMDofyoungsex-andethnicity-matchedpersons).Tscores werenotcalculatedatthestartofGnRHagonisttreatmentbe- causeparticipantshadnotreachedPBMyet. Medical Treatment Protocol Ifdiagnosedwithgenderdysphoria(basedoncriteriafromthe Diagnostic and Statistical Manual of Mental Disorders(Fourth Key Points Question Howdobonemineraldensity(BMD)andBMD zscores developintransgenderadultsafteraminimumof9yearsof gender-affirminghormonetreatmentwithprioruseofa gonadotropin-releasinghormone(GnRH)agonistinadolescence? Findings Inthiscohortstudyof75individualsdiagnosedwith genderdysphoria,amongthoseassignedmaleatbirth,BMD z scoreswerenotdifferentatlong-termfollow-upcomparedwith thestartofGnRHagonisttreatmentatthetotalhipandfemoral neck,incontrasttothelumbarspine.Inindividualsassigned femaleatbirth,BMD zscorescaughtupwithpretreatmentlevels atthelumbarspine,totalhip,andfemoralneck. Meaning ThesefindingssuggestthattreatmentwithaGnRH agonistfollowedbylong-termgender-affirminghormonesissafe regardingbonehealthintransgenderpersonsreceiving testosterone,butbonehealthintransgenderpersonsreceiving estrogenrequiresextraattentionandfurtherstudy. BoneMineralDensityinTransgenderAdolescents Original Investigation Research jamapediatrics.com (Reprinted) JAMA Pediatrics December2023 Volume177,Number12 1333 © 2023 American Medical Association. All rights reserved. Downloaded from jamanetwork.com by University of Amsterdam user on 09/29/2025 Edition, Text Revision)14 and the Diagnostic and Statistical Manual of Mental Disorders(FifthEdition),15adolescentscould start puberty suppression consisting of subcutaneous or in- tramusculartriptorelin(aGnRHagonist)3.75mg,every4weeks or 11.25 mg every 10 to 12 weeks. To start treatment, Tanner genital stage 2 to 3 was required for people assigned male at birth, and Tanner breast stage 2 was required for people as- signedfemaleatbirth. Providedthatgenderdysphoriapersisted,pubertyconsis- tent with the gender identity was induced with GAH in incre- mental dosages when adolescents turned age 16 years. Over time, the protocol was adapted, enabling those aged 15 years to start GAH under certain conditions. Individuals assigned maleatbirthwereusuallygiven5μg/kgbodyweightoral17β- estradiol, which was gradually increased to a daily mainte- nancedoseof2to4mg.Forthoseassignedfemaleatbirth,an intramuscular testosterone ester mixture starting at a dose of 25mg/m 2 bodysurfaceareaevery2weekswasused,building uptoamaintenancedoseof250mgevery3to4weeks. Afteratleast1yearofGAHandaminimumageof18years, persons could become eligible for gonadectomy. Treatment withGnRHagonistswasdiscontinuedifgonadectomywasper- formed or, in those assigned female at birth, when the full maintenancedoseoftestosteronewasreached.Treatmentwith GAHisusuallycontinuedthroughoutlife. Bone Densitometry Bone density was assessed by a densitometric system (Ho- logicDelphi;Hologic,Inc),whichwasupdatedin2004.InFeb- ruary 2011, the appliance was replaced by a different system (Hologic Discovery; Hologic, Inc). The software on this de- vicewasupdatedin2012and2015.Anewdevice(HologicHo- rizon; Hologic, Inc) was installed in December 2020. Cross- calibration with an imaging phantom allowed for direct comparisonofBMDafterupdatesandbetweenalldevices. TheDXAscanswereperformedbetween2001and2018. Anatomicalregionsofinterestwerethelumbarspine,totalhip, and femoral neck of the nondominant hip. All T scores andz scores were calculated at the time of study analyses. Al- though BMD z scores were calculated for both the sex as- signedatbirthandtheexperiencedgender,thesexassigned atbirthwasusedasreferenceintheanalysesassessing zscores overtimetoallowcomparisonswithbaselinevalues.Tscores were calculated using the sex assigned at birth as well as the experiencedgender.BonemineraldensityvaluesfromtheNa- tionalHealthandNutritionExaminationSurveywereusedas reference. 16 In children and adolescents, osteoporosis is de- fined as az score of −2 or lower in combination with a clini- cally meaningful fracture history. In adults, osteoporosis is diagnosediftheTscoreis−2.5orlower. 17 Biochemical Assays UntilJanuary2010,estradiolwasmeasuredusingaradioim- munoassay(RIA;DiaSorin;lowerlimitofquantitation[LLOQ]: 4.90 pg/mL [to convert to picomoles per liter, multiply by 3.671]; interassay coefficient of variation [CV]: 10%). Be- tweenJanuary2010andJuly2014acompetitiveimmunoas- say(Delfia;PerkinElmerWallacOy;LLOQ:5.45pg/mL;inter- assay CV: 10%-13%) was used for estradiol measurements. A conversion formula was applied to compare concentrations (Delfia concentration = 1.267 × [Diasorin concentration] – 28.870).AfterJuly2014,thisassaywasreplacedbyliquidchro- matographytandemmassspectrometry(LC-MS/MS;Amster- dam UMC; LLOQ: 5.45 pg/mL; interassay CV: <7%). Another formula(LC-MS/MSconcentration = 1.60 × [Delfiaconcentra- tion]–29.00)wasappliedforconvertingmeasurements. For testosterone, an RIA (Coat-A-Count; Siemens Medi- calSolutions;LLOQ:28.82ng/dL[toconverttonanomolesper liter, multiply by0.0347]; interassay CV: 7%-20%) was used untilJanuary2013.Thereafter,acompetitiveimmunoassay(Ar- chitect;Abbott;LLOQ:2.88ng/dL;interassayCV:6%-10%)was used. The formula applied to convert measurements de- pended on the testosterone concentration (testosterone 8 nmol/L: Architect concentra- tion = 1.34 ×[RIAconcentration]–1.65). Between 2012 and 2015, 25-hydroxyvitamin D was mea- suredusinganLC-MS/MSmethod(LLOQ:1.6ng/mL[toconvert tonanomolesperliter,multiplyby2.496];interassayCV:8%). 18 Themethodwasadjustedin2015withoutnecessitatingcon- versionofyieldedconcentrations. 19Luteinizinghormone(LH) was measured using an immunometric assay (Architect; Ab- bott;LLOQ:2.0mIU/mL[toconverttointernationalunitsper liter,multiplyby1.0];interassayCV:<6%). Statistical Analysis ContinuousvariableswerereportedasmeanswithSDsfornor- mallydistributeddataandasmedianswithIQRsfornonnor- mallydistributeddata.Dichotomousvariableswerereported aspercentages. Analyses were done separately for those assigned male at birthandthoseassignedfemaleatbirth.Linearmixedmodels were used to analyze areal BMD (in grams per centimeter squared) and BMDz scores. Repeated measures were nested within participants. A random intercept was included in the model. We did not add a random slope or adjustment covari- ates.Inthefirstmodel,timewasmadeintoacategoricalsingle- yearvariablebetweenthestartofGnRHagonisttreatmentand thestartofGAHtreatmentandbetweenthestartofGAHtreat- mentandfollow-up.Scansmade6yearsormoreafterthestart ofGAHweretakentogetherin2-yearintervalstoensureappro- priategroupsize.Formostparticipants,DXAscansweremade onthesamedaysasthestartofGnRHagonisttreatmentandthe start of GAH treatment. Otherwise, a DXA scan made within a 3-month range before or after start of treatments was used for thesetimepoints.IfpeoplehadusedaGnRHagonistformore than2yearsbeforeGAHinitiation,therangewasstretchedto6 monthsbeforeand3monthsafterthestartofGAH. AdifferentlinearmixedmodelwasusedtoevaluateBMD z score development at 4 distinct time points. In this model, timewasdividedinto4categories:startofGnRHagonist,start ofGAH,short-termfollow-uparoundage22years,andlong- term follow-up. The criteria described in the previous para- graphwerefollowedregardingwhichDXAscanswereusedat thestartoftreatments.FortheDXAscanatshort-termfollow- up,scansmadebetweenages21and24yearswereused.The Research Original Investigation BoneMineralDensityinTransgenderAdolescents 1334 JAMA Pediatrics December2023 Volume177,Number12 (Reprinted) jamapediatrics.com © 2023 American Medical Association. All rights reserved. Downloaded from jamanetwork.com by University of Amsterdam user on 09/29/2025 long-termfollow-upDXAscanwasthescanprospectivelymade atthestudyvisitafteratleast9yearsofGAH.Alinearregres- sion model was used to estimate the association of sex ste- roids,LH,and25-hydroxyvitaminDconcentrations;BMI;and durationofmonoGnRHagonisttreatment(inyears)withlong- termfollow-upBMD zscores.Allindependentvariableswere analyzed separately. People who had not undergone gonad- ectomyandwerestillusingaGnRHagonistwereleftoutofthe analysisregardingLH. Statasoftware,version15.1(StataCorpLLC),wasusedfor dataanalyses.Statisticalsignificancewasdefinedasa95%CI excluding0.

Overall Intotal,143individualswereeligibletoparticipate.Ofthose,30 individuals (21%) did not provide informed consent, 27 (19%) could not be reached, 6 (4%) discontinued GAH, 2 (1%) lived abroad,2(1%)hadpsychologicalreasonsfornonparticipation, and1(1%)hadinsufficientdata.Theremaining75participants (52%;25assignedmaleatbirthand50assignedfemaleatbirth) wereincluded.Theinclusionprocessandreasonsfornotpartici- pating are shown in the flowchart (eFigure inSupplement 1). Characteristicsoftheparticipantsareprovidedin Table1.The median (IQR) age at long-term follow-up was 28.2 (27.0-30.8) yearsinparticipantsassignedmaleatbirthand28.2(26.6-30.6) yearsinparticipantsassignedfemaleatbirth.Themedian(IQR) durationofGAHtreatmentwas11.6(10.1-14.7)yearsamongthose assignedfemaleatbirthand11.9(10.2-13.8)yearsamongthose assignedfemaleatbirth.DataonBMI,hormoneconcentrations, BMDTscores,andBMD zscoresatlong-termfollow-upareshown inTable2.Characteristicsofpersonsnotincludedareprovided ineTable1in Supplement1.Peoplenotincludedweresimilarto thoseincluded,exceptforageatthestartofGAHinindividuals assignedfemaleatbirth. BMD Development in Participants Assigned Male at Birth Areal BMD and BMD z score development over time esti- matedbythemixed-modelanalysisareshownin Figure1.Re- sults from the mixed-model analyses in whichz scores were assessed at 4 specific time points (ie, start of GnRH agonist, startofGAH,short-termfollow-up,andlong-termfollow-up) areshownin Figure2 . Inparticipantsassignedmaleatbirth,BMDremainedstable during GnRH agonist treatment but increased during GAH (Figure1).The zscoreswerealreadylowerthan0atthestart of a GnRH agonist and further decreased during GnRH ago- nisttreatment(Figure2).DuringGAHtreatment,thelumbar spine zscoreremainedstable(mean[SD],−1.34[1.16]),result- inginadecreased zscoreatlong-termfollow-up(−0.87;95% CI,−1.15to−0.59)comparedwithstartofGnRHagonisttreat- ment. Thez scores at the total hip (mean [SD], −0.66 [0.75]) andfemoralneck(mean[SD],−0.54[0.84])increasedduring GAHtreatment,andthe zscoreatfollow-upwasnotdifferent fromthepretreatment zscore(totalhip:−0.12[95%CI,−0.31 to 0.07]; femoral neck: 0.01 [95% CI, −0.20 to 0.22]). Mean changes in BMDz scores between all 4 time points are listed ineTable2in Supplement1. Hormone Concentrations, BMI, and Duration of Mono GnRH Agonist Treatment in Participants Assigned Male at Birth Theassociationsofestradiol,testosterone,LHandvitaminD concentrations;BMI;anddurationofmonoGnRHagonisttreat- mentwithlong-termBMD zscoresareshownin Figure3.The BMD zscoresatlong-termfollow-upweregreaterwithincreas- ing estradiol concentrations, although the score differences werenotsignificant(lumbarspine:0.27per27.24pg/mL[95% CI, −0.01 to 0.55 per 27.24 pg/mL]; total hip: 0.18 per 27.24 pg/mL[95%CI,−0.02to0.37per27.24pg/mL];femoralneck: 0.16 per 27.24 pg/mL [95% CI, −0.06 to 0.38 per 27.24 pg/ mL]).TheLHandvitaminDconcentrationsandtheduration of mono GnRH agonist treatment were not associated with long-termfollow-up zscores.Apositiveassociationwasfound between BMI andz scores at follow-up at the total hip (0.05 perkg/m 2;95%CI,0.01-0.10perkg/m 2)andatthefemoralneck (0.08 per kg/m2; 95% CI, 0.04-0.13 per kg/m2) but not at the lumbar spine. Because testosterone was low in all partici- pants assigned male at birth at follow-up (range, 14.41-31.70 ng/dL),thisanalysiswasomitted. BMD Development in Participants Assigned Female at Birth In participants assigned female at birth, BMD increased over timeatall3regions(Figure1).TheBMD zscoresdecreasedaf- ter the start of GnRH agonist treatment (Figure 2).However, during GAH treatment, thez scores increased, resulting in a similar zscorecomparedwiththestartofGnRHagonisttreat- ment. After long-term GAH, the mean (SD)z score was 0.20 (1.05;changefromstartofGnRHagonist:0.09;95%CI,−0.09 to 0.27) at the lumbar spine, 0.07 (0.91; change from start of GnRHagonist:0.10;95%CI,−0.06to0.26)atthetotalhip,and −0.19 (0.94; change from start of GnRH agonist: −0.20; 95% CI,−0.26to0.06)atthefemoralneck.MeanchangesinBMD z scores between all 4 time points are listed in eTable 2 in Supplement1. Hormone Concentrations, BMI, and Duration of Mono GnRH Agonist Treatment in Participants Assigned Female at Birth At follow-up, there was no association of estradiol, testoste- rone,orvitaminDconcentrationsordurationofmonoGnRH agonist treatment with BMDz scores at any of the regions of interestinparticipantsassignedfemaleatbirth(Figure3).The LHwasnegativelyassociatedwiththelumbarspine zscoreat follow-up(−0.03per1.0mIU/mL;95%CI,−0.06to−0.01per 1.0mIU/mL).Similartoparticipantsassignedmaleatbirth,BMI inthoseassignedfemaleatbirthwaspositivelyassociatedwith zscoresatthetotalhip(0.10perkg/m 2;95%CI,0.02-0.17per kg/m2)andfemoralneck(0.11kg/m 2;95%CI,0.03-0.18perkg/ m2)butnotatthelumbarspine.

This cohort study found that BMDz scores declined during GnRHagonisttreatmentbutcaughtupwithpretreatmentlev- BoneMineralDensityinTransgenderAdolescents Original Investigation Research jamapediatrics.com (Reprinted) JAMA Pediatrics December2023 Volume177,Number12 1335 © 2023 American Medical Association. All rights reserved. Downloaded from jamanetwork.com by University of Amsterdam user on 09/29/2025 els at long-term follow-up at all regions of interest except for the lumbar spine in participants assigned male at birth. At the total hip and femoral neck, BMI was positively associated with z scores at long-term follow-up in both participants assigned male at birth and assigned female at birth. In those assigned female at birth, a negative association between LH and zscores at long-term follow-up was found. Our study provided evidence that bone mineral accrual is temporarily suspended by the use of puberty suppression but, due to an increase during GAH treatment, BMD catches up with pretreatment levels at long-term follow-up, except for the lumbar spine in individuals assigned male at birth. Because the median age at long-term follow-up was 28.2 years, and PBM is generally reached around this age, it was Table 1. Characteristics of Study Participants Characteristic Assigned male at birth (n = 25) Assigned female at birth (n = 50) At start of GnRH agonist treatment Age, median (IQR), y 14.5 (13.4-15.7) 14.9 (13.0-16.4) BMI, median (IQR) 19.3 (18.0-21.9) 20.0 (17.9-23.5) European ethnicity, No. (%) 19 (76) 43 (86) Tanner stage, No. (%) B2-B3 NA 9 (18) B4-B5 NA 40 (80) G2-G3 5 (20) NA G4-G5 20 (80) NA Testicular volume, median (IQR), mL 20 (13-20) NA Experienced menarche, No. (%) NA 38 (76) LH, median (IQR), mIU/mL 2.0 (1.7-3.5) 2.9 (1.5-5.5) Estradiol, median (IQR), pg/mL 15.53 (6.27-22.61) 38.95 (10.90-90.98) Testosterone, median (IQR), ng/dL 461.10 (317.00-576.37) 37.46 (37.46-43.23) Lumbar spine z score, mean (SD) Male reference −0.37 (1.07) 0.83 (1.16) Female reference −1.07 (1.07) 0.11 (0.91) Lumbar spine z score <−2.0, No./total No. (%) Male reference 2/22 (9) 0/47 Female reference 4/22 (18) 1/47 (2) Total hip z score, mean (SD) Male reference −0.53 (1.00) −0.34 (0.90) Female reference −0.23 (1.04) −0.05 (0.86) Total hip z score <−2.0, No./total No. (%) Male reference 1/21 (5) 1/47 (2) Female reference 1/21 (5) 0/47 Femoral neck z score, mean (SD) Male reference −0.58 (0.99) −0.43 (0.88) Female reference −0.28 (1.01) −0.11 (0.87) Femoral neck z score <−2.0, No./total No. (%) Male reference 2/21 (10) 0/47 Female reference 1/21 (5) 0/47 Age at start of GAH treatment, median (IQR), y 16.0 (16.0-16.8) 16.1 (16.0-17.6) Age at short-term follow-up, median (IQR), y 22.1 (21.7-22.6) 22.1 (21.9-22.5) Age at long-term follow-up, median (IQR), y 28.2 (27.0-30.8) 28.2 (26.6-30.6) Duration of mono GnRH agonist treatment, median (IQR), y 1.5 (0.7-2.6) 1.5 (0.7-3.1) Duration of GAH treatment, median (IQR), y At short-term follow-up 5.3 (4.3-6.3) 5.7 (4.7-6.1) At long-term follow-up 11.6 (10.1-14.7) 11.9 (10.2-13.8) Gonadectomy, No. (%) 24 (96) 50 (100) Age at gonadectomy, median (IQR), y 19.1 (18.8-21.0) 18.8 (18.3-19.5) Abbreviations:BMI,bodymassindex (calculatedasweightinkilograms dividedbyheightinmeterssquared); GAH,gender-affirminghormones; GnRH,gonadotropin-releasing hormone;LH,luteinizinghormone; NA,notapplicable. SIconversionfactors:Toconvert estradiolfrompicogramspermilliliter topicomolesperliter,multiplyby 3.671;toconvertLHfrom milli-internationalunitspermilliliter tointernationalunitsperliter, multiplyby1.0;toconvert testosteronefromnanogramsper decilitertonanomolesperliter, multiplyby0.0347. Research Original Investigation BoneMineralDensityinTransgenderAdolescents 1336 JAMA Pediatrics December2023 Volume177,Number12 (Reprinted) jamapediatrics.com © 2023 American Medical Association. All rights reserved. Downloaded from jamanetwork.com by University of Amsterdam user on 09/29/2025 not possible to assess whether PBM was decreased or achievement was delayed. Research has consistently found that bone mineral ac- crualdecreasesduringGnRHagonisttreatment.5-11Accrualgen- erally increases during GAH use, but after short-term treat- ment, z scores do not catch up to levels similar to pre-GnRH agonist initiation at every region of interest. 8,9,11 Klink et al11 foundthatthelumbarspine zscoreatage22yearsinbothin- dividualsassignedmaleatbirthandthoseassignedfemaleat birth was still significantly lower compared with thez score atthestartofGnRHagonisttreatment.Thisdecreasewasnot seenwhencomparingthe zscoreatthestartofGnRHagonist Table 2. Body Mass Index, Biochemical Values, Bone Mineral Density T andz Scores, and Percentage With Osteoporosis at 3 Regions of Interest After Long- Term Use of Gender-Affirming Hormones Variable Assigned male at birth (n = 25) Assigned female at birth (n = 50) BMI, median (IQR) 24.7 (21.1-28.7) 24.3 (22.4-26.3) LH, median (IQR), mIU/mL 16.0 (12.5-20.5) 3.7 (0.1-18.0) Estradiol, median (IQR), pg/mL 58.57 (34.32-89.08) 20.43 (16.89-37.32) Testosterone, median (IQR), ng/dL 20.17 (14.41-25.94) 533.14 (403.46-778.10) 25-Hydroxyvitamin D, median (IQR), ng/mL 25.64 (17.23-32.85) 22.04 (15.22-26.84) Lumbar spine T score, mean (SD) Male reference −1.34 (1.16) −0.24 (1.06) Female reference −0.94 (1.16) 0.16 (1.06) z Score, mean (SD) Male reference −1.34 (1.16) −0.23 (1.05) Female reference −0.90 (1.15) 0.20 (1.05) T score <−2.5, No./total No. (%) Male reference 4/25 (16) 0 Female reference 4/25 (16) 0 z Score <−2.0, No./total No. (%) Male reference 9/25 (36) 2 Female reference 4/25 (16) 0 Total hip T score, mean (SD) Male reference −0.70 (0.75) −0.56 (0.74) Female reference −0.12 (0.92) 0.06 (0.91) z Score, mean (SD) Male reference −0.66 (0.75) −0.51 (0.74) Female reference −0.10 (0.92) 0.07 (0.91) T score <−2.5, No./total No. (%) Male reference 0/23 0/50 Female reference 0/23 0/50 z Score <−2.0, %, No./total No. (%) Male reference 0/23 2/50 (4) Female reference 0/23 1/50 (2) Femoral neck T score, mean (SD) Male reference −0.67 (0.85) −0.80 (0.77) Female reference −0.09 (1.04) −0.25 (0.94) z Score, mean (SD) Male reference −0.54 (0.84) −0.68 (0.77) Female reference −0.03 (1.04) −0.19 (0.94) T score <−2.5, No./total No. (%) Male reference 0/23 0/50 Female reference 0/23 0/50 z Score <−2.0, No./total No. (%) Male reference 1/23 (4) 4/50 (8) Female reference 0/23 2/50 (4) Abbreviations:BMI,bodymassindex (calculatedasweightinkilograms dividedbyheightinmeterssquared); LH,luteinizinghormone. SIconversionfactors:Toconvert estradiolfrompicogramspermilliliter topicomolesperliter,multiplyby 3.671;toconvertLHfrom milli-internationalunitspermilliliter tointernationalunitsperliter, multiplyby1.0;toconvert testosteronefromnanogramsper decilitertonanomolesperliter, multiplyby0.0347;toconvert 25-hydroxyvitaminDfrom nanogramspermilliliterto nanomolesperliter,multiplyby 2.496. BoneMineralDensityinTransgenderAdolescents Original Investigation Research jamapediatrics.com (Reprinted) JAMA Pediatrics December2023 Volume177,Number12 1337 © 2023 American Medical Association. All rights reserved. Downloaded from jamanetwork.com by University of Amsterdam user on 09/29/2025 treatmentandage22yearsatthefemoralneckinindividuals assignedeithermaleatbirthorfemaleatbirth. 11 In individuals assigned male at birth,z scores were al- readylowerthan0atthestartofGnRHagonisttreatment.The zscoresmighthaveatendencytodeclineovertimedueto,for example,lifestylefactorsbeingdifferentfrompeers.Thishy- pothesis is in line with findings of low pretreatmentz scores inadulttransgenderwomen. 20 However,informationonthe naturalcourseofBMDdevelopmentintransgenderpeopleis notavailable. ArelativelylowestradioldosageinthefirstperiodofGAH treatmentmightplayaroleinthelackofimprovementinthe lumbarspine zscoreinindividualsassignedmaleatbirth.This possibility is supported by the pattern we found forz scores Figure 1. Areal Bone Mineral Density (BMD) and BMDz Score Development Over Time at 3 Regions of Interest 1.1 1.2 1.0 0.9 0.8 0.7 Areal BMD, g/cm2 Duration of treatment, y Assigned male at birthA Start of GnRHa Start of GAH 1 2 1 2 3 4 5 6 15-1613-1411-129-107-8 Lumbar spine –2.5 0 0.5 1.0 –1.0 –0.5 –2.0 –1.5 BMD z score 1.1 1.2 1.0 0.9 0.8 0.7 Areal BMD, g/cm2 Duration of treatment, y Assigned female at birthB 1 2 3 Start of GnRHa Start of GAH 1 2 3 4 5 6 15-1613-1411-129-107-8 Lumbar spine –2.5 0 0.5 1.0 –1.0 –0.5 –2.0 –1.5 BMD z score 1.1 1.0 0.9 0.8 0.7 Areal BMD, g/cm2 Duration of treatment, yStart of GnRHa Start of GAH 1 2 1 2 3 4 5 6 15-1613-1411-129-107-8 Total hip –1.5 –2.0 1.0 0 0.5 –1.0 –0.5 BMD z score 1.1 1.0 0.7 Areal BMD, g/cm2 Duration of treatment, y 1 2 3 Start of GnRHa Start of GAH 1 2 3 4 5 6 15-1613-1411-129-107-8 Total hip –1.5 –2.0 1.0 0 0.5 –1.0 –0.5 BMD z score 1.0 0.9 0.8 0.7 Areal BMD, g/cm2 Duration of treatment, yStart of GnRHa Start of GAH 1 2 1 2 3 4 5 6 15-1613-1411-129-107-8 Femoral neck –1.5 –2.0 1.0 0 0.5 –1.0 –0.5 BMD z score 1.0 0.9 0.8 0.7 Areal BMD, g/cm2 Duration of treatment, y 1 2 3 Start of GnRHa Start of GAH 1 2 3 4 5 6 15-1613-1411-129-107-8 Femoral neck –1.5 –2.0 1.0 0 0.5 –1.0 –0.5 BMD z score BMD z score Areal BMD 0.9 0.8 Estimatedbymixed-modelanalyses.Amean(SD)of7(2)scanswereavailableperparticipant.Whiskersindicate95%CIs.GAHindicatesgender-affirm ing hormonesandGnRHa,gonadotropin-releasinghormoneagonist. Research Original Investigation BoneMineralDensityinTransgenderAdolescents 1338 JAMA Pediatrics December2023 Volume177,Number12 (Reprinted) jamapediatrics.com © 2023 American Medical Association. All rights reserved. Downloaded from jamanetwork.com by University of Amsterdam user on 09/29/2025 at long-term follow-up to be greater with higher estradiol concentrations and is consistent with the fact that the spine is particularly sensitive to sex steroids. At follow-up, LH was in the higher range in the majority of participants assigned male at birth, which may also indicate insufficient estrogen supplementation. Interestingly, in contrast to testosterone concentrations, which varied widely due to variation in time since last testos- teroneinjection,LHwasnegativelyassociatedwithlong-term follow-up lumbar spine z score in participants assigned female at birth. This finding might imply that a higher dose of testosterone leading to more suppression of LH might be ben- eficial for BMD sustainment in those assigned female at birth, as has also been reported in adults. 20 Body mass index was positively associated with higherz scores at the total hip and femoral neck but not the lumbar spine. This association is likelyduetothefactthatthespinemainlyconsistsoftrabecu- lar bone, whereas the hip contains a greater part of cortical bone. Cortical bone is known to be affected by mechanical loading more than trabecular bone. 21 At follow-up, when participants were in their late 20s (around 28 years), the majority hadz scores within the nor- malrangewhenusingreferencedataoftheaffirmedgender, as is recommended by the International Society for Clinical Densitometry. 17However,becausethelumbarspine zscorein participantsassignedmaleatbirthdidnotcatchup,extraat- tention should be paid to maximize bone mineral accrual by optimizing mechanical loading of the bone (eg, by motivat- ingtoplaysports)andfurthermorebyadvisingagainstsmok- ing,ensuringadequatecalciumintakeorsupplementationand sunlightexposureorvitaminDsupplementationinthisgroup, and optimizing estrogen treatment. Given the results of our study, monitoring of BMD using DXA scans is not necessary inindividualsassignedfemaleatbirthwhostartedtreatment inlatepuberty,butfollow-upwithDXAremainsimportantin thoseassignedmaleatbirth. Figure 2. Bone Mineral Density (BMD)z Score Development and Mean Change at 4 Time Points and 3 Regions of Interest 0.5 –0.5 0 –1.0 –1.5 –2.0 BMD z score Time point Assigned male at birthA Start of GnRHa Start of GAH Short-term FU Long-term FU Lumbar spine 0.5 0 –0.5 –1.0 –1.5 –2.0 BMD z score Time point Assigned female at birthB Start of GnRHa Start of GAH Short-term FU Long-term FU Lumbar spine 0.5 0 –0.5 –1.0 –1.5 –2.0 BMD z score Time point Start of GnRHa Start of GAH Short-term FU Long-term FU Total hip 0.5 0 –0.5 –1.0 –1.5 –2.0 BMD z score Time point Start of GnRHa Start of GAH Short-term FU Long-term FU Total hip 0.5 0 –0.5 –1.0 –1.5 –2.0 BMD z score Time point Start of GnRHa Start of GAH Short-term FU Long-term FU Femoral neck 0.5 0 –0.5 –1.0 –1.5 –2.0 BMD z score Time point Start of GnRHa Start of GAH Short-term FU Long-term FU Femoral neck Estimatedbymixed-modelanalyses. Themedian(IQR)durationof gender-affirminghormone(GAH) treatmentatshort-termfollow-up (FU)was5.3(4.3-6.3)yearsin individualsassignedmaleatbirthand 5.7(4.7-6.1)yearsinindividuals assignedfemaleatbirth.Themedian (IQR)durationofGAHtreatmentat long-termFUwas11.6(10.1-14.7)years inindividualsassignedmaleatbirth and11.9(10.2-13.8)yearsin individualsassignedfemaleatbirth. Atthestartofgonadotropin- releasinghormoneagonist(GnRHa) treatment,3individualsassigned maleatbirthand2individuals assignedfemaleatbirthhadmissing dual-energyx-rayabsorptiometry (DXA)scans.AtthestartofGAH treatment,DXAscansweremissingin 5individualsassignedmaleatbirth and8individualsassignedfemaleat birth.Atshort-termFU,DXAscans werenotavailablefor11individuals assignedmaleatbirthand20 individualsassignedfemaleatbirth. NoDXAscansweremissingat long-termFU.Circlesindicateareal BMDandwhiskers,95%CIs.Mean changesinBMD zscoresbetweenall 4timepointsarelistedineTable2. BoneMineralDensityinTransgenderAdolescents Original Investigation Research jamapediatrics.com (Reprinted) JAMA Pediatrics December2023 Volume177,Number12 1339 © 2023 American Medical Association. All rights reserved. Downloaded from jamanetwork.com by University of Amsterdam user on 09/29/2025 Toourknowledge,thisstudyisthefirsttoreportonlong- term bone health in transgender adolescents treated with a GnRH agonist and subsequent GAH treatment. With a me- dianfollow-upofalmost12yearsofGAHtreatment,thisstudy yields new and important data. While treatment with pu- bertysuppressionforadolescentsdiagnosedwithgenderdys- phoriahasbecomecontroversial,especiallybecauseofques- tionsaboutthelong-termsafety,wepresentedevidencethat GnRHagonisttreatmentisgenerallysaferegardingbonehealth. In future studies, it would be informative to assess the asso- ciation between estradiol and BMD in a larger study popula- tion.Additionally,thenaturalcourseofBMDdevelopmentin transgender people should be studied because BMDz scores before the start of GnRH agonist treatment were already de- creased in participants assigned male at birth. This research would help clarify to what extent the finding that follow-up measurementsdidnotcatchupwithpretreatmentlevelsatthe lumbarspinecanbeattributedtothehormonaltreatmentand towhatextenttootherfactors.

There are some limitations to this study. Overall, 19% of eli- gibleparticipantscouldnotbereached.Another21%didnot provideinformedconsentforvariousreasons.TheDXAdata werenotavailableforallparticipantsateverytimepoint.By using mixed-model analyses, we tried to have the least pos- siblebiasduetomissingdata. 22 Additionally,wewereunable to stratify participants for puberty stage at the start of GnRH agonisttreatmentbecausesofewwereinearlypubertyatthe startoftreatment.Individualsstartingpubertysuppressionin late puberty have inherently gained more bone mass before the start of treatment compared with individuals starting in early puberty. It can be speculated that this might be benefi- cial for long-term bone health, although a previous study 8 found that those starting puberty suppression in early pu- bertyactuallyhadsimilarorhigherBMD zscoresafter3years ofGAHcomparedwiththosestartinginlatepuberty.Further- more, the association of sex steroids and vitamin D concen- trations with BMDz scores could not be assessed over time because these variables were only sufficiently available at short- and long-term follow-up. Ideally, to account for sea- sonalfluctuation,multiplemeasurementsofvitaminDwould havebeenusedperparticipant.Likewise,thetestosteronecon- centration depends on the time since last administration in individuals assigned female at birth who are receiving intra- musculartestosterone. Figure 3. Regression Coefficients for the Associations of Biochemical Values, Body Mass Index (BMI), and Duration of Mono Gonadotropin-Releasing Hormone Agonist (GnRHa) Treatment With Bone Mineral Density (BMD)z Scores at Follow-up at 3 Regions of Interest –0.5 0.50 BMD z score Source Estradiol (per 100 pmol/L) Lumbar spine Total hip Femoral neck Lumbar spine Total hip Femoral neck Lumbar spine Total hip Femoral neck Lumbar spine Total hip Femoral neck Lumbar spine Total hip Femoral neck LH (per 1 U/L) 25–Hydroxyvitamin D (per 10 nmol/L) BMI (per kg/m 2) Mono GnRHa duration (per y) BMD z score (95% CI) 0.00 (–0.03 to 0.04) –0.01 (–0.04 to 0.03) 0.01 (–0.20 to 0.22) –0.01 (–0.15 to 0.13) –0.01 (–0.17 to 0.15) 0.05 (–0.03 to 0.13) 0.05 (0.01 to 0.10) 0.08 (0.04 to 0.13) 0.02 (–0.49 to 0.54) 0.07 (–0.28 to 0.43) –0.02 (–0.43 to 0.38) 0.27 (–0.01 to 0.55) 0.18 (–0.02 to 0.37) 0.16 (–0.06 to 0.38) 0.01 (–0.06 to 0.04) Assigned male at birthA –0.5 0.50 BMD z score Source Estradiol (per 100 pmol/L) Lumbar spine Total hip Femoral neck Lumbar spine Total hip Femoral neck Lumbar spine Total hip Femoral neck Lumbar spine Total hip Femoral neck Lumbar spine Total hip Femoral neck Lumbar spine Total hip Femoral neck Testosterone (per 10 nmol/L) LH (per 1 U/L) 25–Hydroxyvitamin D (per 10 nmol/L) BMI (per kg/m 2) Mono GnRHa duration (per y) BMD z score (95% CI) –0.10 (–0.26 to 0.06) –0.11 (–0.27 to 0.06) –0.03 (–0.06 to –0.01) –0.02 (–0.04 to 0.00) –0.01 (–0.04 to 0.01) 0.07 (–0.07 to 0.21) 0.04 (–0.09 to 0.17) 0.06 (–0.06 to 0.19) 0.07 (–0.02 to 0.17) 0.10 (0.02 to 0.17) 0.11 (0.03 to 0.18) 0.03 (–0.20 to 0.25) 0.00 (–0.20 to 0.18) 0.00 (–0.24 to 0.15) 0.30 (–0.21 to 0.81) –0.03 (–0.48 to 0.41) –0.04 (–0.51 to 0.42) –0.01 (–0.19 to 0.18) Assigned female at birth B Estimatedbyregressionanalyses.Theassociationbetweentestosteroneand BMD zscoresinindividualsassignedmaleatbirthwasnotassessedbecause testosteronewaslow(range,0.5-1.1nmol/L)inallindividualsassignedmaleat birthatfollow-up.Squares,triangles,andcirclesrepresentregression coefficients(β);whiskersrepresent95%CIs.SIconversionfactors:Toconvert estradiolfrompicomolesperlitertopicogramspermilliliter,divideby3.671;to convertLHfrominternationalunitsperlitertomilli-Internationalunitsper milliliter,divideby1.0;toconverttestosteronefromnanomolesperliterto nanogramsperdeciliter,divideby0.0347;toconvert25-hydroxyvitaminD fromnanomolesperlitertonanogramspermilliliter,divideby2.496.BMIis calculatedasweightinkilogramsdividedbyheightinmeterssquared.LH indicatesluteinizinghormone. Research Original Investigation BoneMineralDensityinTransgenderAdolescents 1340 JAMA Pediatrics December2023 Volume177,Number12 (Reprinted) jamapediatrics.com © 2023 American Medical Association. All rights reserved. Downloaded from jamanetwork.com by University of Amsterdam user on 09/29/2025

This cohort study found that, after long-term use of GAH,z scoresinindividualstreatedwithpubertysuppressioncaught upwithpretreatmentlevels,exceptforthelumbarspineinpar- ticipantsassignedmaleatbirth,possiblyduetolowestradiol concentrations.ThesefindingssuggestthatuseofaGnRHago- nistfollowedbylong-termGAHissaferegardingbonehealth intransgenderpersonsreceivingtestosterone,butbonehealth in transgender persons receiving estrogen requires extra at- tention and further study. Estrogen treatment should be op- timized and lifestyle counseling provided to maximize bone developmentinindividualsassignedmaleatbirth. ARTICLE INFORMATION Accepted for Publication:August30,2023. Published Online:October30,2023. doi:10.1001/jamapediatrics.2023.4588 Author Affiliations:DepartmentofEndocrinology andMetabolism,AmsterdamUniversityMedical Center(UMC)atVrijeUniversiteitAmsterdam, Amsterdam,theNetherlands(vanderLoos, denHeijer,Wiepjes);CenterofExpertiseonGender Dysphoria,AmsterdamUMCatVrijeUniversiteit Amsterdam,Amsterdam,theNetherlands (vanderLoos,Hannema,denHeijer,Wiepjes); DepartmentofInternalMedicine,Hospital StJansdal,Harderwijk/Lelystad,theNetherlands (Vlot);DivisionofPediatricEndocrinology, DepartmentofPediatrics,GhentUniversity Hospital,Ghent,Belgium(Klink);Departmentof Pediatrics,AmsterdamUMCatVrijeUniversiteit Amsterdam,Amsterdam,theNetherlands (Hannema). Author Contributions:DrsvanderLoosand Wiepjeshadfullaccesstoallofthedatainthestudy andtakeresponsibilityfortheintegrityofthedata andtheaccuracyofthedataanalysis. Concept and design:vanderLoos,Klink,denHeijer, Wiepjes. Acquisition, analysis, or interpretation of data: vanderLoos,Vlot,Hannema,Wiepjes. Drafting of the manuscript:vanderLoos. Critical review of the manuscript for important intellectual content:Vlot,Klink,Hannema, denHeijer,Wiepjes. Statistical analysis:vanderLoos,Wiepjes. Administrative, technical, or material support: Wiepjes. Supervision:Vlot,Klink,Hannema,denHeijer, Wiepjes. Conflict of Interest Disclosures:DrHannema reportedreceivingpersonalfeesfromPfizer outsidethesubmittedwork.Nootherdisclosures werereported. Funding/Support:Nonereported. Data Sharing Statement:SeeSupplement2.

1. BonjourJP,ChevalleyT.Pubertaltiming,bone acquisition,andriskoffracturethroughoutlife. Endocr Rev.2014;35(5):820-847.doi:10.1210/ er.2014-1007 2. GordonCM,ZemelBS,WrenTA,etal.The determinantsofpeakbonemass. J Pediatr.2017; 180:261-269.doi:10.1016/j.jpeds.2016.09.056 3. SurreyES,HornsteinMD.ProlongedGnRH agonistandadd-backtherapyforsymptomatic endometriosis:long-termfollow-up. Obstet Gynecol. 2002;99(5Pt1):709-719.doi: 10.1097/ 00006250-200205000-00008 4. AntoniazziF,ZamboniG,BertoldoF,LauriolaS, TatòL.BonedevelopmentduringGHandGnRH analogtreatment. Eur J Endocrinol.2004;151(suppl 1):S47-S54.doi:10.1530/eje.0.151s047 5. CarmichaelP,ButlerG,MasicU,etal.Short-term outcomesofpubertalsuppressioninaselected cohortof12to15yearoldyoungpeoplewith persistentgenderdysphoriaintheUK. PLoS One. 2021;16(2):e0243894.doi:10.1371/journal. pone.0243894 6. JosephT,TingJ,ButlerG.TheeffectofGnRH analoguetreatmentonbonemineraldensityin youngadolescentswithgenderdysphoria:findings from a large national cohort.J Pediatr Endocrinol Metab.2019;32(10):1077-1081.doi: 10.1515/j pem-2019-0046 7. NavabiB,TangK,KhatchadourianK,LawsonML. Pubertalsuppression,bonemass,andbody compositioninyouthwithgenderdysphoria. Pediatrics.2021;148(4):e2020039339. doi:10.1542/peds.2020-039339 8. SchagenSEE,WoutersFM,Cohen-KettenisPT, GoorenLJ,HannemaSE.Bonedevelopmentin transgenderadolescentstreatedwithGnRH analoguesandsubsequentgender-affirming hormones. J Clin Endocrinol Metab.2020;105(12): e4252-e4263.doi:10.1210/clinem/dgaa604 9. StoffersIE,deVriesMC,HannemaSE.Physical changes,laboratoryparameters,andbonemineral densityduringtestosteronetreatmentinadolescents withgenderdysphoria. J Sex Med.2019;16(9): 1459-1468.doi:10.1016/j.jsxm.2019.06.014 10. VlotMC,KlinkDT,denHeijerM,Blankenstein MA,RotteveelJ,HeijboerAC.Effectofpubertal suppressionandcross-sexhormonetherapyon boneturnovermarkersandbonemineralapparent density(BMAD)intransgenderadolescents. Bone . 2017;95:11-19.doi:10.1016/j.bone.2016.11.008 11. KlinkD,CarisM,HeijboerA,vanTrotsenburgM, RotteveelJ.Bonemassinyoungadulthood followinggonadotropin-releasinghormoneanalog treatmentandcross-sexhormonetreatmentin adolescentswithgenderdysphoria. J Clin Endocrinol Metab.2015;100(2):E270-E275. doi:10.1210/jc.2014-2439 12. GovernmentoftheNetherlands.Wet Medisch-WetenschappelijkOnderzoekMet Mensen.PublishedonlineJanuary1,2020. AccessedSeptember25,2023. https://wetten. overheid.nl/BWBR0009408/2020-01-01 13. WiepjesCM,NotaNM,deBlokCJM,etal.The AmsterdamCohortofGenderDysphoriaStudy (1972-2015):trendsinprevalence,treatment,and regrets. J Sex Med.2018;15(4):582-590.doi:10. 1016/j.jsxm.2018.01.016 14. AmericanPsychiatricAssociation. Diagnostic and Statistical Manual of Mental Disorders(Fourth Edition,TextRevision).AmericanPsychiatric Publishing;2000. 15. AmericanPsychiatricAssociation. Diagnostic and Statistical Manual of Mental Disorders(Fifth Edition).AmericanPsychiatricAssociation;2013. 16. KellyTL,WilsonKE,HeymsfieldSB.Dual energyX-rayabsorptiometrybodycomposition referencevaluesfromNHANES. PLoS One.2009;4 (9):e7038.doi:10.1371/journal.pone.0007038 17. ShuhartCR,YeapSS,AndersonPA,etal. Executivesummaryofthe2019ISCDPosition DevelopmentConferenceonmonitoring treatment,DXAcross-calibrationandleast significantchange,spinalcordinjury, peri-prostheticandorthopedicbonehealth, transgendermedicine,andpediatrics. J Clin Densitom. 2019;22(4):453-471.doi:10.1016/j.jocd.2019.07.001 18. HeijboerAC,BlankensteinMA,KemaIP,Buijs MM.Accuracyof6routine25-hydroxyvitaminD assays:influenceofvitaminDbindingprotein concentration. Clin Chem.2012;58(3):543-548. doi:10.1373/clinchem.2011.176545 19. DirksNF,VesperHW,vanHerwaardenAE,etal. Variouscalibrationproceduresresultinoptimal standardizationofroutinelyused25(OH)D ID-LC-MS/MSmethods. Clin Chim Acta.2016;462: 49-54.doi:10.1016/j.cca.2016.08.016 20. WiepjesCM,deJonghRT,deBlokCJ,etal. Bonesafetyduringthefirsttenyearsof gender-affirminghormonaltreatmentin transwomenandtransmen. J Bone Miner Res.2019; 34(3):447-454.doi:10.1002/jbmr.3612 21. OzakiE,MatsukawaM,ManoI,etal.Growthof corticalbonethicknessandtrabecularbonedensity inJapanesechildren. Bone.2020;141:115669. doi:10.1016/j.bone.2020.115669 22. TwiskJ,deBoerM,deVenteW,HeymansM. Multipleimputationofmissingvalueswasnot necessarybeforeperformingalongitudinal mixed-modelanalysis. J Clin Epidemiol.2013;66(9): 1022-1028.doi:10.1016/j.jclinepi.2013.03.017 BoneMineralDensityinTransgenderAdolescents Original Investigation Research jamapediatrics.com (Reprinted) JAMA Pediatrics December2023 Volume177,Number12 1341 © 2023 American Medical Association. All rights reserved. Downloaded from jamanetwork.com by University of Amsterdam user on 09/29/2025