Addition of Anti-thymocyte Globulin in Allogeneic Stem Cell Transplantation With Peripheral Stem Cells From Matched Unrelated Donors Improves Graft-Versus-Host Disease and Relapse Free Survival

Addition of Anti-thymocyte Globulin in Allogeneic Stem Cell Transplantation With Peripheral Stem Cells From Matched Unrelated Donors Improves Graft-Versus-Host Disease and

Relapse Free Survival

M.M. Ali,

B. Grønvold,

M. Remberger,

I.W. Abrahamsen,

A.E. Myhre,

G.E. Tjønnfjord,

Y. Fløisand,

T. Gedde-Dahl

Abstract

In 2014 we introduced anti-thymocyte globulin (ATG) to the graft-versus-host disease (GvHD) prophylaxis regimeninallogeneicstemcelltransplantation(Allo-HSCT)withperipheralstemcells(PBSC)frommatched unrelateddonors (MUD).Weanalysed theoutcomesof415patientswhowentthroughMUDallo-HSCTand receivedPBSCwithorwithoutATG.WereportdramaticreductionoftheincidenceofchronicGvHDandour studyillustratesthebenefitofATGinadditiontostandardGvHDprophylaxis.

Anti-thymocyte globulin (ATG) is commonly used to prevent graft-versus-host disease (GvHD) after allogeneic hematopoieticstemcelltransplantation(allo-HSCT). ToevaluatetheimpactofATGaspartoftheGvHDprophylaxisin ourinstitution,wereporttheoutcomeof415patientswithmatchedunrelateddonors(MUD)transplantedforhemato- logicalmalignancieswithorwithoutATGfrom2005to2019atOsloUniversityHospital,Norway.Thefollowinggroups werecompared:(1)154patientstransplantedwithperipheralbloodstemcells(PBSC)withoutATG2005-2014.(2)137 patientstransplantedwithbonemarrowstemcells(BMSC)2005-2019.(3)124patientstransplantedwithPBSCand ATG(PBSC+ATG)2014-2019. Threeyearssurvivalwassimilarinthegroups,61%followingallograftingwithPBSC, 54%withBMSC,and59%withPBSC+ATG.AcuteGvHDgradeIII-IVwas14%,14%,and7%;chronicGvHDwas81%, 32,and26%;andextensivecGvHD44%,15%,and6%inthecorrespondinggroups.BothacuteandchronicGvHDwere significantlyreducedinthePBSC+ATG-versusthePBSCgroup(p<0.05andp <0.001respectively).Transplant- relatedmortality(TRM)was33%,25%,and17%(p=0.18).Graftversushostdiseaseandrelapsefreesurvival(GRFS) at3yearswas43%,43%,and64%inthegroups.AddingATGtotheGvHDprophylaxisregimenofMUDallo-HSCTwith PBSCresultedinasubstantialreductionofbothacuteandchronicGvHDwithoutcompromisingthediseasecontrol, reflectedinasuperior3yearsGRFS.

ClinicalLymphoma,MyelomaandLeukemia,Vol.21,No.9,598–605© 2021TheAuthor(s).PublishedbyElsevierInc.

ThisisanopenaccessarticleundertheCCBYlicense(http://creativecommons.org/licenses/by/4.0/) KeyWords:MUD,ATG,GvHD-prophylaxis,Allo-HSCT,GvHD

Introduction

Allo-HSCThasthe potentialto curehematological malignan- cies.However,itisassociatedwithseveremorbidityandmortality.^1,2 Graft-versus-hostdisease(GvHD)andrelapsearethemajorobsta-

1DepartmentofHaematology,OsloUniversityHospital,Oslo,Norway

2DepartmentofMedicalSciences,UppsalaUniversityandKFUE,UppsalaUniversity Hospital,Uppsala,Sweden

3InstituteofClinicalMedicine,UniversityofOslo,Norway

4CentreforCancerCellReprogramming,InstituteofClinicalMedicine,Universityof Oslo,Oslo,Norway

5TheClatterbridgeCancerCenterNHSFoundationTrust,Liverpool,United Kingdom

Submitted:Feb23,2021;Revised:Apr29,2021;Accepted:May3,2021;Epub:11 May2021

Addressforcorrespondence:MaryanMAli,DepartmentofHaematology,OsloUniver- sityHospital

E-mailcontact:[email protected]

clesforcureandmaintaininggoodqualityof lifeaftertransplan- tation.^3-5Individualthoroughriskassessmentbeforeproceedingto transplantisrequired.AnHLA-matchedrelateddonor(MRD)is stillconsideredthebestdonorchoice,althoughoutcomesfollowing transplantswithmatchedunrelateddonor(MUD)areapproaching thoseofMRDs.^6-9

The mostfrequentlyavailablesourceofallogeneicstem cellsis currentlyPBSCfromaMUD.BothPBSCcomparedtoBMSCand MUDcomparedtoMRDareassociatedwithincreasedfrequency and severity of GvHD.^10-12 Early studies suggested that adding ATGtoGvHDprophylaxiswasassociated withdelayedimmune reconstitution and increased risk of viral infections, particularly cytomegalovirus(CMV)andEpstein-Barrvirus(EBV)reactivation inadditiontoanincreasedriskofrelapse.¹⁰However,additionof ATGtoGvHDprophylaxishasbeenshowninseveralrandomized 2152-2650/$-seefrontmatter© 2021TheAuthor(s).PublishedbyElsevierInc.

ThisisanopenaccessarticleundertheCCBYlicense

prospective trials tosignificantlyreducecGvHDand toimprove GvHDandrelapsefreesurvival(GRFS).^11-16Inthe onlydouble- blind trial in MUD/MAC, allo-HSCT ATG lowered moderate toseverecGvHD,butprogressionfreesurvival(PFS)andoverall survival (OS) also were lower, indicating that additional analy- ses are needed to understand the appropriate role for ATG in allo-HSCT.¹⁹

ThereisalsoevidencethatadditionofATGmayreducecGvHD intheMRDsetting.^14,17

ThestandardbackboneregimentopreventGvHDisthecombi- nationof a calcineurin inhibitorand methotrexate.¹⁸ Ex-vivo T- celldepletion,monoclonalantibodies,andinvivoT-celldepletion orpost-transplantcyclophosphamidearesupplementaryalternatives

19-21.

TheuseofATGhasincreasedtremendouslyinthedifferenttrans- plantmodalities^10,22-24 andisnowpartoftheGvHDprophylaxis inmost Europeantransplant centers, both for PBSC-as wellas BMSC/MUDtransplants.

ATGproductsdifferintheirmanufacturingprocess,including thecelllineusedforimmunizationaswellastheanimalsusedfor production.Anti-thymoglobulinisapolyclonalIgGimmunoglob- ulingeneratedinrabbitsbyimmunizingwitheitherhumanthymo- cytes (Thymoglobulin: Sanofi, Paris, France) or with the Jurkat T-lymphoblastoid cell line (ATG-Fresenius, rebranded now as Grafalon, Neovii Biotech, GMBH, Grӓfelfing, Germany). ATG bindsdirectlytoT-cellsurfaceepitopesandinducescomplement- dependent cell lysis, thereby suppressing the recipients T-cell function.Duetoitslonghalf-life,ATGalsointeractswiththetrans- planteddonor T-cellsand evolvingimmunecellsduring engraft- ment. Itssuppressiveaction ondonor T-cellsisthe rationalefor useasGvHDprophylaxis.ATGdelayspost-transplantlymphocyte recoveryandincreasestheriskforinfections,especiallyviralinfec- tions.^25,26

Inreducedintensityconditioning(RIC),thecurativepotentialof theallo-HSCTislargelydependentonthedonorT-cellgraft-versus- leukemia(GVL)effect.Soifferetalhaveshownanincreasedriskof relapseinRICallo-HSCTwithATG,althoughintheirstudythey usedbothhorseATGandhighdosesofrabbitATG(7mg/kg).²² Both Baron et al and Devillier et al showed thatlow doses (<

6mg/kg)ofrabbitATG(Thymoglobulin)areeffectiveasGvHD prophylaxiswithoutcompromisingdiseasecontrolandoutcomesin RIC.^23,24

Severalreviewsandmeta-analysishaveconfirmedthebenefitof ATGinpreventingGvHD.^27-29However,thetrialsareveryhetero- geneous regarding conditioning regimens, patient characteristics, donortype,graftsource,ATGbrand,dosage,andtimingofATG administration, making interpretation and comparison challeng- ing.³⁰

In 2014 we added ATG to the GvHD prophylaxis regimen for patients undergoing allo-HSCT with PBSC from MUD in ourinstitution.Wehaveperformed aretrospectiveanalysisofthe outcome in 415 patients who went through 10/10 MUD allo- HSCT with or without ATG for hematological malignancies at OsloUniversityHospital.Theaimofthisretrospectiveanalysiswas toevaluatethe resultsof addingATGtothe GvHDprophylaxis regimenintheMUDwithPBSC.

Patients and Methods

Weretrospectivelycollecteddatafrom415adultpatientstrans- plantedwith10/10MUDforhematologicalmalignancies(Table1) between2005and2019atOsloUniversityHospital,Norway.

The patientswere categorizedinthree groups: (1) ThePBSC group:154patientswhoreceivedPBSCwithnoATGbetween2005 and2014priortoimplementationofATGaspartofthe GvHD prophylaxis.(2)TheBMSCgroup:137patientsreceivingBMSC without ATG between2005 and 2019. (3) The PBSC + ATG group:124patientsreceivingPBSCwithATGprophylaxisbetween 2014and2019.ThepatientcharacteristicsaregiveninTable1.

ConditioningandStemCellSource

Myeloablativeconditioning(MAC,n=233)consistedofbusul- fan16mg/kgincombinationwithcyclophosphamide120mg/kg, orfludarabin150mg/m²,orTBI13Gyandcyclophosphamide120 mg/kg.Reducedintensityconditioning(RIC,n=182)consisted offludarabin150mg/m²incombinationwithtreosulfan42g/m², busulfan8mg/kgorcyclophosphamide600mg/m²,orTBI2Gy.

EightpatientsreceivedotherMACandRICprotocols.Priortothe introductionofATG,westrivedtouseBMSCinMACtransplants becauseofthedocumentedloweroccurrenceofcGvHD.Sincethe curativepotentialofRICallo-HSCTislargelydependentonthe donorT-cellgraft-versus-leukemia(GVL)effect³¹,weroutinelyuse PBSCintheRICsetting.

Donors

Forboth donorsand recipients, weperformed high-resolution HLA matching for the HLA-A, -B, -C, -DRB1, and -DQB1 gene locion two different time points for each recipient-donor pair.Atleastonetyping wasdoneinourhospital’s HLAlabora- tory accredited by the European Federation of Immunogenetics.

TheothertypingwasperformedeitherinourHLAlaboratoryor ina donor registry contractHLAlaboratory. Duetochanges in typingmethodsandtechnologyfrom2005to2018,combinations of various genotyping techniques (eg, sequence-specific oligonu- cleotide typing, sequence-specific primer typing, and sequence- basedtyping)wereusedtoachievehigh-resolutiontypingresults.

GVHDProphylaxis

GvHD prophylaxisconsisted of cyclosporine (CsA) incombi- nationwithmethotrexate(MTX)onday+1,+3,+6(and+11) (n = 339), mycophenolate mofetil day 0 to +28 (n = 9), or sirolimusday-1to+2(n=60).SevenpatientsreceivedCsAalone.

We administered a low dose ATG (Thymoglobulin) of 2 mg/kg/dayin2days(totaldose4mg/kg)forPBSC10/10MUD.

InfectionProphylaxis

All patients received antifungal and herpes virus prophylaxis accordingtotheEuropeanConferenceonInfectionsinLeukemia (ECIL)guidelinesinforceatthattime.³²PatientsatriskforEBV- orCMVreactivationwerescreenedforviralreplicationaccordingto institutionalguidelines.Preemptivetreatmentwasinitiatedaccord- ingtothedefinitionsgivenbelow.

Table1 PatientandTransplantCharacteristics

PBSC PBSC+ATG BMSC p-value

n 154 124 137

Age(median,Q1-Q3) 53(42-60) 61(51-65) 42(29-51) <0.001

Sex(M/F) 92/62 87/37 84/53 0.17

Diagnosis: <0.001

AML/ALL 74 44 107

CML 8 2 11

Lymphoma/CLL 46 21 5

MDS/MPN 25 56 14

Stage: 0.02

CR1/CP1 41 42 75

CR2-3/CP2-3 41 26 38

Later 9 6 5

Unknown 63 50 19

Donorage: 30(24-38) 25(22-33) 29(22-36) <0.001

Sexmismatch:

FtoM 25(16%) 19(15%) 22(16%) 0.98

CD34dose(median,Q1-Q3) 6.7(5.2-8.4) 6.9(5.7-8.7) – 0.50

TNC(median,Q1-Q3) – – 2.7(2.1-3.4)

Conditioning:

MAC/RIC 92/62 11/113 130/7 <0.001

TBI-based 3(2%) 5(4%) 17(12%) <0.001

GVHDprophylaxis: <0.001

CsA+MTX 106 100 133

CsA+MMF 4 4 1

CsA+Sirolimus 39 20 1

CsAalone 5 0 2

CMVsero-statuspreSCT:

Match/Mismatch 90/59 76/46 82/48 0.89

Mindicatesmale;Ffemale;AMLacutemyeloidleukemia;ALLacutelymphoblasticleukemia;CLLchroniclymphocyticleukemia;MDSmyelodysplasticsyndrome;MPNmyeloproliferativeneoplasia;CR completeremission;CPchronicphase;FtoMfemaletomale;SCstemcell;BMSCbonemarrowstemcell;PBSCperipheralbloodstemcell;TNCtotalnucleatedcells;MACmyeloablativeconditioning;

RICreducedintensityconditioning;TBItotalbodyirradiation;CsAcyclosporine;MTXmethotrexate;MMFmycophenolatemofetil.

Definitions

Engraftmentwasdefinedasabsoluteneutrophilcountofatleast 0.5 ×10⁹/Lfor 3 daysand plateletcount atleast 20 ×10⁹/L for5dayswithouttransfusions.CMVreactivationwasdefinedas quantitative valueof CMV-DNAabove200IU/mland preemp- tivetreatmentwithgancicloviralternativelyfoscavirarestandardof care.EBVreactivation wasdefined asquantitativevalueofEBV- DNAabove1000IU/ml.Accordingtoourinstitutionalguidelines, thethresholdforEBVdiseaseandpreemptiveEBVtreatmentwith Rituximab375mg/m^bweeklyupto4consecutiveweekswasEBV DNAlevels≥20000IU/ml.

TheprocedureswereperformedinaccordancewiththeHelsinki declaration,andthestudywasapprovedbyTheRegionalCommit- teeforMedicalandHealthResearchEthicsSouthEastNorwayand theDataProtectionOfficer,OsloUniversityHospital.

Statistics

TheprimaryobjectiveofthestudywastoanalyzeOS,GvHD, GRFS,RFS,TRM,infections,CMV,andEBVreactivationsinthe 3patientcohorts.

The analysis was performed on September 19, 2019. OS, RFS,andGRFSwerecalculatedusingthe Kaplan-Meiermethod and compared with the log-ranktest. For GRFS,relapse, death, severeacuteGvHD(gradesIII-IV),andextensivechronicGvHD, whichever came first, were considered the events. Transplant- relatedmortality(TRM),GvHD,andrelapseincidence(RI)were estimatedusinganonparametricestimatorofcumulativeincidence curvestakingcompetingeventsintoconsideration.Correctedmulti- variateriskfactoranalysesforTRM,RI,andGvHDwereperformed using the proportional subdistribution hazard regression model developed byFineand Gray.Multivariatemodeling for OS and RFSwascarriedoutusingCoxregressionmodels(toestimatehazard ratios(HRs)).Correctionsweremadeforthepretransplantfactors thatweredifferentbetweenthethreegroups(patientanddonorage, diagnose,diseasestage,RIC/MACandGvHDprophylaxis).Asthe studycoveredalongperiod(2005through2019),wealsoincluded

“yearofHSCT” tocorrectforchangestreatmentperformedduring thistimeframe.Allp-valuesweretwo-tailed.Categoricalparameters werecomparedusingtheChi-squaretestandcontinuousvariables werecomparedusingtheMann-WhitneytestortheKruskal-Wallis

test.AnalysiswasperformedusingStatistica13software(StatSoft, Tulsa,OK)andtheEZRstatisticalsoftware.

Results

There was no significant difference in disease stage at trans- plantation, gender mismatch, stem cell dose, or CMV serologi- calstatus pretransplantbetweenthethree groups.Patientsinthe PBSC+ATGgroupweresignificantlyolder,receivedgraftsfrom younger donors, had the highest proportion of myelodysplastic syndrome, and a significantlyhigher number receivedRIC. The patientcharacteristicsaregiveninTable1.

NeutrophilEngraftmentandGraftFailure

The median timeto neutrophil engraftment was 13 days for thepatientswhoreceivedPBSCwithoutATG,interquartilerange (IQR)9daysto18 days.Fivepatientsinthisgroupdiedbefore engraftment,and1experiencedprimarygraftfailure.

ForpatientswhoreceivedBMSC,themediantimetoneutrophil engraftmentwas21days,IQR18daysto25days.Threepatients in this group died before engraftment, and 2 suffered primary graftfailure.The mediantimetoneutrophilengraftment forthe PBSC + ATG patients was 15 days, IQR 14 days to 18 days.

One patient in this group died before engraftment, and 2 had primarygraftfailure.Timetoneutrophilengraftmentwassignifi- cantlylongerintheBMSCgroupcomparedtoboththePBSC(p

<0.001)andthePBSC+ATGgroup(p<0.001).ATGdelayed engraftmentinpatientsreceivingPBSC(p=0.01).

Infections

Invasivefungalinfections(IFI)wereobservedin12%ofpatients inthePBSCgroupcomparedto19%and14%intheBMSCand PBSC+ ATGgroups(p=0.26).CMVreactivationoccurredin 44%,32%,and42%inthethreegroups,respectively(p=0.08).

ThecumulativeincidenceofEBVreactivationwas71%following PBSC+ATG,whichwassignificantlyhighercomparedwiththe BMSCandPBSCat9%and11%,respectively(p<0.001).Inspite ofthehighEBVreactivationrateinthePBSC+ATGgroup,only 4 patients received rituximab for EBV reactivation preemptively whereas2patientsreceivedrituximabforpost-transplantlympho- proliferative disease (PTLD). In one case, PTLD wasdiagnosed postmortem. For a number of patients in the non-ATG group (approximately 50%)and only9% ofthe PBSC + ATGgroup, EBV-PCRwasnotmeasuredduetoinstitutionalstandard ofcare priortoimplementationofATG.

AcuteandChronicGvHD

ThecumulativeincidenceofaGvHDgradesII-IVinthePBSC group was 42%, 41% in the BMSC group, and 27% in the PBSC+ATGgroup(p=0.007)(Figure1A).

Thecumulativeincidenceofsevere acuteGvHDgrades III-IV inthePBSC,BMSC,andPBSC+ATGgroupwere14%,14%, and 7%,respectively (Figure1B).ThePBSC +ATGgrouphad significantlylesssevereaGvHDcomparedbothtotheBMSCgroup (p<0.05)andPBSCgroup(p=0.045).

ThecumulativeincidenceofcGvHDwassignificantlyhigherin thePBSCgroup(81%)comparedtotheBMSCgroup(32%,p<

Table2 CausesofDeath

CoD BMSC PBSC PBSC+ATG

Relapse 31(23%) 16(10%) 16(13%)

Infection 8(6%) 19(12%) 4(3%)

GVHD 4(3%) 18(12%) 6(5%)

MOF 14(10%) 11(7%) 6(5%)

Other 8(6%) 12(8%) 3(3%)

CoDcausesofdeath;MOFmultipleorganfailure

0.001)andthePBSC+ATGgroup(26%,p<0.001)(Figure1C).

TheincidenceofextensivecGvHDwasalsosignificantlyhigherin thePBSCgroup(44%)comparedtotheothertwogroups(15%

and6%,respectively,p=0.03).

TreatmentRelatedMortality(TRM)

TherewasnodifferencebetweenthePBSC+ATGgroupandthe BMSCgroupregarding3-yeartreatment-relatedmortality(17%vs 25%,p=0.18).Onthecontrary,the3-yearTRMwassignificantly lowerinthePBSC+ATGgroupthaninthe PBSCgroup,17%

and33%respectively(p=0.024)(Figure2A).TheTRM>3years inthePBSCgroupiscausedbyinfections(n=5),secondarymalig- nancies(n=3),suicide(n=1),and“other” (n=3).Amongthese 12lateTRM,10sufferedfromextensivecGvHD,whichpredispose tothebefore-mentionedcausesofdeath.

GvHDandRelapse-freeSurvival(GRFS)

Three years after transplantation, the rate of GRFSwas 43%

forboththePBSC groupandtheBMSC groupbut64%inthe PBSC+ATGgroup,p=0.005.TheadditionofATGtoPBSC increasedGRFSsignificantly(Figure2B).

Relapse

The cumulative incidence of relapse 3 years after transplanta- tionwas11%inthePBSC,24%intheBMSC,and19%inthe PBSC+ATGgroup(Figure2C).Therelapsedifferencebetween PBSC and BMSC was significant only inunivariant analysis (p

<0.01),andtherewasnorelapsedifferenceinPBSC+ATGvs BMSC(p=0.14)(Figure2C).

Survival

Threeyearsaftertransplantation,theprobabilityofsurvivalwas 61%inthePBSCgroup,54%intheBMSCgroup,and59%inthe PBSC+ATGgroup(Figure2D).Therewasnosignificantdiffer- enceeitherinRFSorOSinthePBSC +ATGgroupcompared tothetwoothergroups.CausesofdeatharepresentedinTable2. ThemaincausesofdeathbothintheBMSCandthePBSC+ATG groupsarerelapses.Thereis,however,nodifferenceinrelapserates inthePBSCandthePBSC+ATGgroups(Table2).

MultivariateAnalysis

Inthecorrectedmultivariateanalysis,wefoundthattheaddition of ATG to PBSC resulted in less chronic GvHD compared to PBSC withoutATG (Table 3).GRFSwassignificantly betterin thePBSC+ATGgroupcomparedtotheBMSC(Table3).There

Figure1 Cumulativeincidenceof(A)acuteGvHDII-IV,(B)acuteGvHDIII-IV,and(C)chronicGvHDafterMUDHSCTinpatients receivingPBSC,BMSC,orPBSC+ATG.

was a trendforless acute GvHDII-IV and betterGRFSinthe PBSC + ATG group compared to the PBSC group. Chronic GvHDtendedtobehigherintheBMSCgroupcomparedtothe PBSC+ATGgroup.

Discussion

GvHD and relapse are the major obstacles encountered after allo-HSCT.ThestandardregimenforpreventionofGvHDisthe combinationofacalcineurininhibitorandmethotrexate.Duetoan unacceptablehighincidenceofcGvHDinpatientstransplantedup to2014withPBSCfromMUDinourinstitution,weimplemented ATGaspartoftheGvHDprophylaxisregimenintheMUDPBSC settingfromFebruary2014.

Theincidenceof cGvHDwasdramaticallyreducedfrom 81%

to 26% in patients receiving MUD PBSC after the introduc- tion ofATGintheGvHD prophylaxisregimenwithoutincreas- ingtheincidenceofrelapseorleadingtomoregraftfailure.This is in harmony with published data.12,14,23,25,33,34 In fact, in our material there is a trend toward lower incidence of cGvHD in the PBSC + ATGgroupcomparedwith the BMSCgroup.The incidenceofacuteGvHDwasalsosignificantlyreducedbyadding ATG. Thisisalsoinaccordancewithpreviousstudies^29,35-37; asa

matteroffact,theincidenceofsevereaGvHDinthePBSC+ATG groupislowerthanintheBMSCgroup.Weraisethequestionasto theoutcomewhenATGisaddedtotheGvHDprophylaxisregimen in patients allografted with BMSC. The available studies show conflictingresults.Bacigalupoetalshowedsignificantreductionin both severe aGvHDand cGvHD amongMUDBMSC patients who receivedATG7.5-10mg/kg.¹¹ Onthe otherhand,Ravinet etalreportednosignificant impactofATGinMUDBMSC.In thisstudy,varioustotaldosesofATGwereusedatthediscretionof theattendingphysician(range<5to≥10mg/kg,only10patients receivedATGlessthan5mg/kg).³⁸Inbothstudies,Thymoglobulin wasused,asinourstudy.

Inthemultivariateanalysiscorrectedforpatientanddonorage, diagnosis,diseasestage,conditioningregimen,andGvHDprophy- laxis,the additionof ATGtoMUDallo-HSCTwith PBSC did notincreaserelapserate.Ourregistrydatadidnotincludesufficient datatodefinethediseaseriskindexinallpatients.Weacknowledge thatthediseaseriskindexwouldhaveallowedbettercomparability betweengroupsandwithpublisheddata.

Cumulativeincidence(CI)ofTRMforallthreegroupsappearsto notreachaplateau,mostevidentforthePBSCgroup(Figure2A).

This may be explained by the observation that the majority of

Figure2 (A)CumulativeincidenceofTransplant-relatedmortality(TRM),(B)probabilityofGRFS,and(C)cumulativeincidence ofrelapseand(D)probabilityofsurvival(OS),afterMUDHSCTinpatientsreceivingPBSC,BMSC,orPBSC+ATG.

patientsinthePBSCgroupdevelopedextensivecGvHD,making themmorepronetoinfectionsanddeath.

GRFSisanincreasinglyrecognizedoutcomemeasureafterallo- HSCT.³⁹GRFSisdefined aslackofsevereaGvHDgradeIII-IV, cGvHDthatnecessitatessystemictreatment,relapse,anddeath.^39-41 OurfindingsdemonstratethatATGincreasesGRFSbyreducing severeacuteandchronicGvHDwhichmostlikelyincreasesquality of life (QoL).Severe aGVHD and cGVHD have a detrimental impactonQoLandpatientsuffering.⁴²

EBVreactivationandEBV-relatedpost-transplantlymphoprolif- erativedisease(PTLD)havebeen showntobeassociatedwithT- celldepletionusingATGformulations^43,44andpretransplantEBV sero-status.As expected,the occurrence of EBVreactivation was significantlyhigherinthe PBSC+ ATGgroupcomparedtothe BMSC and PBSC groups. We now perform weekly monitoring with quantitative EBV-PCR inblood until 4 monthsafter allo- HSCTforallpatientswhoreceiveATG.Aftertheimplementation ofweeklymonitoringandpreemptivetreatmentwithrituximab,we

haveencounteredonlyonepatientwithverifiedPLTDamongthe PBSC+ATGgroup.

Supportivecare before,during, andafterHSCThasimproved inrecentyears⁴⁵ andcouldhavehadanimpactonthestudy,but inthemultivariateanalysisweincluded“yearofHSCT” tocorrect fortreatmentchangesperformedduringthistimeframe.Theretro- spectivenatureofourdataisalimitation.Ourstudypopulationis heterogeneousintermsofdiseasetype,diseasestage,andcondition- ingregimen.Comparingourstudywithdatafromavailableclinical trialsusingATGaspartoftheGvHDprophylaxisischallengingdue todifferencesamongstudiesintermsofATGdosing,timing,length ofadministration,typeofpreparation,andtransplantcharacteristics includingdiseasegroup,donor type,graftsource, andcondition- ingregimen.^23,27,28Nonetheless,ourdatasupportpreviousreports that a dramatic reduction in cGvHD may be accomplished by ATG.

Ourstudyshowsthattheadditionoflow-dose(4mg/kg)ATGin allo-HSCTdecreasestheincidenceofGvHDandnotablyincreases

Table3 Multivariate Analysis for Various Outcome Variables Corrected for Differences Between the Groups. Corrected for (Patient and Donor Age, Diagnosis, Disease Stage, RIC/MAC and GvHD Prophylaxis),andYearofHSCT

HR,95%CI,p-value Mortality

BMSC 0.96,0.37-2.48,0.93

PBSC 0.90,0.38-2.13,0.80

PBSC+ATG Ref

TRM

BMSC 0.96,0.37-2.48,0.92

PBSC 0.90,0.38-2.13,0.80

PBSC+ATG Ref

Relapse

BMSC 2.11,0.81-5.49,0.13

PBSC 1.13,0.44-2.93,0.80

PBSC+ATG Ref

AcuteGVHDII-IV

BMSC 1.65,0.81-3.35,0.17

PBSC 1.81,0.92-3.57,0.08

PBSC+ATG Ref

AcuteGVHDIII-IV

BMSC 1.81,0.47-6.95,0.39

PBSC 1.84,0.51-6.66,0.35

PBSC+ATG Ref

ChronicGVHD

BMSC 1.87,0.90-3.87,0.09

PBSC 5.49,2.82-10.7,<0.001

PBSC+ATG Ref

GRFS

BMSC 2.00,1.05-3.79,0.03

PBSC 1.67,0.92-3.05,0.09

PBSC+ATG Ref

HRhazardratio;CIconfidenceinterval.

GRFS,thusillustratingthebenefitofATGinadditiontostandard GvHDprophylaxisinallo-HSCT.

ClinicalPracticePoints

• Allogeneic stem cell transplantation (allo-HSCT) is associated with severe morbidity and mortality. Graft-versus-host disease (GvHD)andrelapsearemajorobstaclesforcureandmaintain- inggoodqualityoflifeaftertransplantation.

• Anti-thymocyte globulin (ATG) is commonly used to prevent GvHDafterallo-HSCT.In2014weaddedATGtotheGvHD prophylaxis regimen for patients undergoing allo-HSCT with peripheral stem cells (PBSC) from matched unrelated donor (MUD) at our institution. We administered a lowdose ATG (Thymoglobulin)of2mg/kg/dayin2days(totaldose4mg/kg) forPBSC10/10MUD.

• Wecomparetheoutcomeof415adultpatientswithhaematolog- icalmalignanciesundergoingallo-HSCTwithMUDandPBSC withorwithoutATG.

• Therewasnosignificantdifferenceindiseasestageattransplanta- tion,gendermismatch,stemcelldose,orCMVserologicalstatus pretransplantbetweenthepatientswhoreceivedandthosewho didnotreceive ATG. Patientswho receivedATGweresignifi- cantlyolder,receivedgrafts fromyoungerdonors, andhad the highest proportionof myelodysplasticsyndrome, and a signif- icantlyhigher number received reducedintensity conditioning (RIC)regimen.

• The incidence of chronic GvHD was dramatically reduced in patientsreceivingMUDPBSCwithlow-doseATGintheGvHD prophylaxisregimenwithoutincreasingtheincidenceofrelapse orleadingtomoregraftfailure.

• OurdatasupportpreviousreportsthattheadditionofATGto thestandardGvHDprophylaxisdecreasestheincidenceofboth acuteandchronicGvHDandincreasestheGvHD-free,relapse- freesurvivalsignificantly.

Declaration of Competing Interests

None.

References

1.AppelbaumFR.Haematopoieticcelltransplantationasimmunotherapy.Nature. 2001;411:385–389.

2.Sorror ML,Estey E. Allogeneichematopoietic cell transplantation foracute myeloidleukemiainolderadults.HematologyAmSocHematolEducProgram. 2014;2014:21–33.

3.Socié G,StoneJV,WingardJR,etal.Long-termsurvivalandlatedeathsafter allogeneicbonemarrowtransplantation.LateEffectsWorkingCommitteeofthe InternationalBoneMarrowTransplantRegistry.NEnglJMed.1999;341:14–21. 4.FilipovichAH,WeisdorfD,PavleticS,etal.NationalInstitutesofHealthconsen-

susdevelopmentprojectoncriteriaforclinicaltrialsinchronicgraft-versus-host disease:I.Diagnosisandstagingworkinggroupreport.BiolBloodMarrowTrans- plant.2005;11:945–956.

5.MartinPJ,CountsJrGW,AppelbaumFR,etal.Lifeexpectancyinpatients survivingmorethan5yearsafterhematopoieticcelltransplantation.JClinOncol. 2010;28:1011–1016.

6.Yakoub-AghaI,MesnilF,KuentzM,etal.Allogeneicmarrowstem-celltrans- plantationfromhumanleukocyteantigen-identicalsiblingsversushumanleuko- cyteantigen-allelic-matchedunrelateddonors(10/10)inpatientswithstandard- -riskhematologicmalignancy:aprospectivestudyfromtheFrenchSocietyofBone MarrowTransplantationandCellTherapy.JClinOncol.2006;24:5695–5702. 7.WoolfreyA,LeeSJ,GooleyTA,etal. HLA-allelematchedunrelateddonors

comparedtoHLA-matchedsiblingdonors:roleofcellsourceanddiseaserisk category.BiolBloodMarrowTransplant.2010;16:1382–1387.

8.SaberW,OpieS,RizzoJD,etal.Outcomesaftermatchedunrelateddonorversus identicalsiblinghematopoieticcelltransplantationinadultswithacutemyeloge- nousleukemia.Blood.2012;119:3908–3916.

9.ShouvalR,FeinJA,LabopinM,etal.Outcomesofallogeneichaematopoietic stemcelltransplantationfromHLA-matchedandalternativedonors:aEuropean SocietyforBlood andMarrow Transplantationregistryretrospectiveanalysis.

LancetHaematol.2019;6e573-e84.

10.RembergerM,SvahnBM,HentschkeP,etal.Effectoncytokinereleaseand graft-versus-hostdiseaseofdifferentanti-Tcellantibodiesduringconditioning forunrelatedhaematopoieticstemcelltransplantation.BoneMarrowTransplant. 1999;24:823–830.

11.BacigalupoA,LamparelliT,GualandiF,etal.Prophylacticantithymocyteglobu- linreducestheriskofchronicgraft-versus-hostdiseaseinalternative-donorbone marrowtransplants.BiolBloodMarrowTransplant.2002;8:656–661.

12.FinkeJ,BethgeWA,SchmoorC,etal.Standardgraft-versus-hostdiseaseprophy- laxiswithorwithoutanti-T-cellglobulininhaematopoieticcelltransplantation frommatchedunrelateddonors:arandomised,open-label,multicentrephase3 trial.LancetOncol.2009;10:855–864.

13.WalkerI,PanzarellaT,CoubanS,etal.Additionofanti-thymocyteglobulin tostandardgraft-versus-hostdiseaseprophylaxisversusstandardtreatmentalone inpatientswithhaematologicalmalignanciesundergoingtransplantationfrom unrelateddonors:finalanalysisofarandomised,open-label,multicentre,phase 3trial.LancetHaematol.2020;7e100-e11.

14.KrögerN,SolanoC,WolschkeC,etal.Antilymphocyteglobulinforprevention ofchronicgraft-versus-hostdisease.NEnglJMed.2016;374:43–53.

15.FinkeJ,SchmoorC,BethgeWA,et al. Long-termoutcomesafter standard graft-versus-hostdiseaseprophylaxiswithorwithoutanti-human-T-lymphocyte immunoglobulininhaemopoieticcelltransplantationfrommatchedunrelated donors:finalresultsofarandomisedcontrolledtrial.LancetHaematology.2017;4 e293-e301.

16.SoifferRJ,KimHT,McGuirkJ,etal.Prospective,randomized,double-blind, PhaseIIIclinicaltrialofanti-T-lymphocyteglobulintoassessimpactonchronic graft-versus-host disease-free survival in patients undergoing HLA-matched unrelated myeloablative hematopoietic cell transplantation. J Clin Oncol. 2017;35:4003–4011.

17.OthmanJ,GreenwoodM,MooreJ,LarsenS,WatsonAM,ArthurC.Unrelated donor transplantrecipients giventhymoglobuline have superiorGRFS when comparedtomatchedrelateddonorrecipientsundergoingtransplantationwithout ATG.BiolBloodMarrowTransplant.2020;26:1868–1875.

18.HandbookE.In:CarrerasE,DufourC,MohtyM,KrögerN,eds.TheEBMT Handbook: HematopoieticStem CellTransplantationandCellularTherapies.

Cham(CH):Springer,2019.

19.Champlin R.T-celldepletion to prevent graft-versus-hostdisease afterbone marrowtransplantation.HematolOncolClinNorthAm.1990;4:687–698. 20.SimpsonD.Newdevelopmentsintheprophylaxisandtreatmentofgraftversus

hostdisease.ExpertOpinPharmacother.2001;2:1109–1117.

21.BuscaA,AversaF.In-vivoorex-vivoTcelldepletionorbothtopreventgraft-ver- sus-hostdiseaseafterhematopoieticstemcelltransplantation.ExpertOpinBiol Ther.2017;17:1401–1415.

22.Soiffer RJ, Lerademacher J, Ho V, et al. Impact of immune modulation with anti-T-cell antibodies on the outcome of reduced-intensity allogeneic hematopoietic stemcell transplantation forhematologicmalignancies.Blood. 2011;117:6963–6970.

23.BaronF,LabopinM,BlaiseD,etal.ImpactofinvivoT-celldepletiononoutcome ofAMLpatientsinfirstCRgivenperipheralbloodstemcellsandreduced-inten- sityconditioningallo-SCTfromaHLA-identicalsiblingdonor:areportfromthe AcuteLeukemiaWorkingPartyoftheEuropeanGroupforBloodandMarrow Transplantation.BoneMarrowTransplant.2014;49:389–396.

24.DevillierR,LabopinM,ChevallierP,etal.Impactofantithymocyteglobulindoses inreducedintensityconditioningbeforeallogeneictransplantationfrommatched siblingdonorforpatientswithacutemyeloidleukemia:areportfromtheacute leukemiaworkingpartyofEuropeangroupofBoneMarrowTransplantation.Bone MarrowTransplant.2018;53:431–437.

25.BacigalupoA.Antilymphocyte/thymocyteglobulinforgraftversushostdisease prophylaxis:efficacyandsideeffects.BoneMarrowTransplant.2005;35:225–231. 26.MohtyM.Mechanismsofactionofantithymocyteglobulin:T-celldepletionand

beyond.Leukemia.2007;21:1387–1394.

27.Nishihori T, Al-Kadhimi Z, Hamadani M, et al. Antithymocyte globu- lin in allogeneichematopoietic cell transplantation:benefits and limitations.

Immunotherapy.2016;8:435–447.

28.YuanJ,PeiR,SuW,etal.Meta-analysisoftheactionsofantithymocyteglobulin inpatientsundergoingallogeneichematopoieticcelltransplantation.Oncotarget. 2017;8:10871–10882.

29.KumarA,ReljicT,HamadaniM,etal.Antithymocyteglobulinforgraft-ver- sus-hostdiseaseprophylaxis:anupdatedsystematicreviewandmeta-analysis.Bone MarrowTransplant.2019;54:1094–1106.

30.WalkerI,PanzarellaT,CoubanS,etal.Pretreatmentwithanti-thymocyteglobu- linversus noanti-thymocyteglobulininpatientswith haematologicalmalig- nanciesundergoing haemopoietic cell transplantation fromunrelated donors:

arandomised,controlled,open-label,phase3,multicentretrial.LancetOncol. 2016;17:164–173.

31.ByrneM,SavaniBN,MohtyM,etal.Peripheralbloodstemcellversusbone marrowtransplantation:AperspectivefromtheAcuteLeukemiaWorkingParty oftheEuropeanSocietyforBloodandMarrowTransplantation.ExpHematol. 2016;44:567–573.

32.MaertensJA,GirmeniaC,BrüggemannRJ,etal.Europeanguidelines forprimary antifungalprophylaxisinadulthaematologypatients:summaryoftheupdated recommendationsfromtheEuropeanConferenceonInfectionsinLeukaemia.J AntimicrobChemother.2018;73:3221–3230.

33.MohtyM,LabopinM,BalèreML,etal.Antithymocyteglobulinsandchronic graft-vs-hostdiseaseaftermyeloablativeallogeneicstemcelltransplantationfrom HLA-matchedunrelateddonors:areportfromtheSociéteFrançaisedeGreffede MoelleetdeThérapieCellulaire.Leukemia.2010;24:1867–1874.

34.Afram G,Simón JAP,Remberger M,et al. Reducedintensity conditioning increasesriskofseverecGVHD:identificationofriskfactorsforcGVHDina multicentersetting.MedOncol.2018;35:79.

35.AraiY,JoT,MatsuiH,etal.Efficacyofantithymocyteglobulinforallogeneic hematopoieticcelltransplantation:asystematicreviewandmeta-analysis.Leuk Lymphoma.2017;58:1840–1848.

36.BonifaziF,SolanoC,WolschkeC,etal.AcuteGVHDprophylaxisplusATLG aftermyeloablativeallogeneichaemopoieticperipheralbloodstem-celltransplan- tationfromHLA-identicalsiblingsinpatientswithacutemyeloidleukaemiain remission:finalresultsofqualityoflifeandlong-termoutcomeanalysisofaphase 3randomisedstudy.LancetHaematol.2019;6e89-e99.

37.BailénR,KwonM,Pascual-CascónMJ,etal.Post-transplantcyclophosphamide forGVHDprophylaxiscomparedtoATG-basedprophylaxisinunrelateddonor transplantation.AnnHematol.2021;100:541–553.

38.RavinetA,CabrespineA,Socié G,etal.Impactofthymoglobulinbystemcell source(peripheralbloodstemcellorbonemarrow)aftermyeloablativestemcell transplantationfromHLA10/10-matchedunrelateddonors:areportfromthe Société Françaisede Greffede Moelle etde ThérapieCellulaire.Transplantation. 2016;100:1732–1739.

39.HoltanSG,DeForTE,LazaryanA,etal.Compositeendpointofgraft-versus-host disease-free,relapse-freesurvivalafterallogeneichematopoieticcelltransplantation.

Blood.2015;125:1333–1338.

40.LiuYC,ChienSH,FanNW,etal.Prognosticfactorsonthegraft-versus-host disease-freeandrelapse-freesurvivalafteradultallogeneichematopoieticstemcell transplantation.StemCellsInt.2016;2016.

41.InamotoY,KimuraF,KandaJ,etal.Comparisonofgraft-versus-hostdisease-free, relapse-freesurvivalaccordingtoavarietyofgraftsources:antithymocyteglobulin andsinglecordbloodprovidefavorableoutcomesinsomesubgroups.Haematolog- ica.2016;101:1592–1602.

42.PalluaS,GiesingerJ,OberguggenbergerA,etal.ImpactofGvHDonqualityoflife inlong-termsurvivorsofhaematopoietictransplantation.BoneMarrowTransplant. 2010;45:1534–1539.

43.Landgren O, Gilbert ES, Rizzo JD, et al. Risk factors for lymphopro- liferative disorders after allogeneic hematopoietic cell transplantation. Blood. 2009;113:4992–5001.

44.Hoegh-PetersenM,GoodyearD,GeddesMN,etal.Highincidence ofposttrans- plantlymphoproliferativedisorderafterantithymocyteglobulin-basedcondition- ingandineffectivepredictionbyday28EBV-specificTlymphocytecounts.Bone MarrowTransplant.2011;46:1104–1112.

45.CooperJP,StorerBE,GranotN.Allogeneichematopoieticcelltransplantation withnon-myeloablativeconditioningforpatientswithhematologicmalignancies:

Improvedoutcomesovertwodecades.Haematologica.2020.