KIDRADJ

N

SPaO

RFGMMa(CTF

©P

ORIGINAL SCIENTIFIC ARTICLES

idney Transplantation under Minimalmmunosuppression after Pretransplant Lymphoidepletion with Thymoglobulin or Campath

on Shapiro, MD, FACS, Amit Basu, MD, FACS, Henkie Tan, MD, PhD, FACS, Edward Gray, BS,khtar Kahn, MD, Parmjeet Randhawa, MD, Noriko Murase, MD, Adriana Zeevi, PhD, Alin Girnita, MD,iana Metes, MD, Roberta Ness, PhD, Debra C Bass, MS, Anthony J Demetris, MD,

ohn J Fung, MD, PhD, FACS, Amadeo Marcos, MD, Thomas E Starzl, MD, PhD, FACS

BACKGROUND: Multiple drug immunosuppression has allowed the near elimination of rejection, but without com-mensurate improvements in longterm graft survival and at the cost of quality of life. We havesuggested that transplantation outcomes can be improved by modifying the timing and dosage ofimmunosuppression to facilitate natural mechanisms of alloengraftment and acquired tolerance.

STUDY DESIGN: Two therapeutic principles were applied for kidney transplantation: pretransplant recipientconditioning with antilymphoid antibody preparations (Thymoglobulin [Sangstat] or Cam-path [ILEX Pharmaceuticals]), and minimal posttransplant immunosuppression with tacroli-mus monotherapy including “spaced weaning” of maintenance doses when possible. The resultsin Thymoglobulin- (n � 101) and Campath-pretreated renal transplantation recipients (n �90) were compared with those in 152 conventionally immunosuppressed recipients in theimmediately preceding era.

RESULTS: Spaced weaning was attempted in more than 90% of the kidney transplant recipients after pretreat-ment with both lymphoid-depleting agents, and is currently in effect in two-thirds of the survivors.Although there was a much higher rate of acute rejection in theThymoglobulin-pretreated recipientsthan in either the Campath-pretreated or historic control recipients, patient and graft survival inboth lymphoid depletion groups is at least equivalent to that of historic control patients. In theThymoglobulin-conditioned patients for whom followups are now 24 to 40 months, chronic allo-graft nephropathy (CAN) progressed at the same rate as in historic control patients. Selected patientson weaning developed donor-specific nonreactivity.

CONCLUSIONS: After lymphoid depletion, kidney transplantation can be readily accomplished under minimalimmunosuppression with less dependence on late maintenance immunosuppression and abetter quality of life. Campath was the more effective agent for pretreatment. Guidelines forspaced weaning need additional refinement. (J Am Coll Surg 2005;200:505–515. © 2005 by

the American College of Surgeons)

Kaprdlpattpat

o competing interests declared.

upported by NIH grant DK 64207-01.art of this work was presented at the IS Ravdin Lecture in the Basic Sciencest the American College of Surgeons 90th Annual Clinical Congress, Newrleans, LA, October 2004.

eceived November 24, 2004; Accepted December 30, 2004.rom the Thomas E Starzl Transplantation Institute (Shapiro, Basu, Tan,ray, Kahn, Randhawa, Murase, Zeevi, Girnita, Metes, Demetris, Fung,arcos, Starzl) and the Departments of Surgery (Shapiro, Basu, Tan, Kahn,urase, Metes, Fung, Marcos), Pathology (Randhawa, Zeevi, Girnita, Demetris),

nd Epidemiology (Ness, Bass), the University of Pittsburgh Medical CenterUPMC) and the University of Pittsburgh, PA.orrespondence address: Thomas E Starzl, MD, PhD, Thomas E Starzlransplantation Institute, University of Pittsburgh, Montefiore Hospital, 7th

ploor, Suite 723, Pittsburgh, PA 15213.

5052005 by the American College of Surgeons

ublished by Elsevier Inc.

idney transplantation became a practical clinical servicefter it was shown that rejections developing under azathio-rine were highly reversible with prednisone, and that theeversals often were succeeded by the emergence of variableonor-specific nonreactivity (ie, tolerance).1 In 1966, anti-

ymphoid globulin (ALG) was added to azathioprine andrednisone as a steroid-sparing adjunct.2 A short course ofntilymphoid globulin was begun preoperatively and con-inued for several posttransplant days or weeks.2,3 The pre-ransplant portion of the course was subsequently deem-hasized and omitted because of uncertainty about its valuend because the time constraints of cadaveric transplanta-ion made the pretreatment impractical. Instead, antilym-

hoid globulin usually was started on the day of, or day

ISSN 1072-7515/05/$30.00doi:10.1016/j.jamcollsurg.2004.12.024

aptt

msdiTtsatcpaz[itiOtc

MIMipctpt(amsPpaIamp

PNfenmtls21t(

wdpnwpg

IHPrttoainp

LL2tmapku1g

MT

506 Shapiro et al Transplantation with Reduced Immunosuppression J Am Coll Surg

fter, transplantation (induction therapy), and in addition,rednisone or other agents were increasingly instituted athis time in multiple drug regimens designed to eliminatehe threat of acute rejection.4

After elucidation of the mechanisms of alloengraft-ent,5,6 it was apparent that this treatment policy could

ubvert the seminal mechanism of clonal exhaustion-eletion.7 Consequently, we suggested modifications of

mmunosuppression in accordance with two principles.he first was lymphoid depletion before rather than after

ransplantation to reduce the anticipated donor-specific re-ponse into a more easily deletable range. The second wasvoidance of so much posttransplant immunosuppressionhat the immune activation-dependent mechanism oflonal exhaustion-deletion would be interdicted. Pretrans-lant conditioning was done with a single infusion of rabbitntithymocyte globulin (rATG, Thymoglobulin [Gen-ymel]),8,9 or alternatively, of alemtuzumab (Campath 1HILEX Pharmaceuticals]).10-14 Minimalistic posttransplantmmunosuppression was begun with relatively low doses ofacrolimus monotherapy with the intention of dose wean-ng after the first few months of highest immunologic risk.

ur initial experience15,16 and that reported here suggesthat this approach to management can be carried out effi-iently and safely.

ETHODSnstitutional review process

odifications in the timing and dosage of conventionalmmunosuppression were undertaken in July 2001. Therimary purpose was to improve the quality-of-life out-omes and patient and graft survival across the full spec-rum of all kinds of adult kidney recipients in our clinicalractice. The modifications were submitted in this contexto the University of Pittsburgh Institutional Review BoardIRB), which judged the changes to be within the bound-ries of historically based standard treatment. The treat-ent protocols were reviewed by the Presbyterian Univer-

ity Hospital Committee on Innovative Practices and theharmacy and Therapeutic Practices Committee, with ap-roval by both. All patients provided informed consent. Inddition, separate informed consent was obtained withRB approval for studies of immune variables not routinelyssayed in our conventional practice. Safety and efficacyonitoring were assured by formal weekly reviews of all

atients. d

atient selectiono adult kidney recipients were denied access to the re-

orms in management because of high risk factors. Recipi-nts who received previous, simultaneous, or subsequentonkidney solid organ allografts or bone marrow were re-oved from this analysis, as were kidney-only transplanta-

ions performed in the period of time when Thymoglobu-in and Campath pretreatment were both used. The threetudy populations were compiled during the eras of March000 to July 2001 (historic controls, no pretreatment: n �52); July 2001 to October 2002 (Thymoglobulin pre-reatment: n � 101); and March 2003 to September 2003Campath pretreatment: n � 90) (Table 1).

Differences in donor characteristics reflected nation-ide efforts to expand the donor pool by using oldereceased donors and more living donors. The large com-lement of nonrelated live donors (mostly spousal) isoteworthy. The mean histocompatibility match wasorst and the mean ischemic time best in the Campathretreatment series. Differences in recipient demo-raphic factors were not statistically significant.

mmunosuppressionistoric controlsatients treated between March 2000 and June 2001eceived multidrug immunosuppression that includedacrolimus, a 5-day intravenous and oral prednisoneaper (2003 40 mg in 40 mg decrements) followed byral prednisone (20 mg/d) and often a third agent (usu-lly mycophenolate mofetil or sirolimus). The multiagentmmunosuppression (particularly the steroid compo-ent) was weaned slowly throughout the first 6 to 12osttransplant months.

ymphoid depletionymphoid depletion between July 2001 and October002 was done with an infusion of 5 mg/kg rabbit anti-hymocyte globulin (rATG, Thymoglobulin). The Thy-oglobulin was administered over several hours before

llograft reperfusion. The antibody infusion was accom-anied by 1 or 2 g methylprednisolone to prevent cyto-ine release consequences. The same steroid doses weresed when alemtuzumab (anti-CD52 mAb, Campath-H) infusion of 30 mg was substituted for Thymo-lobulin as the conditioning agent.

inimalistic immunosuppressionhe lymphoid depleted patients were started on twice

aily tacrolimus (Prograf) on postoperative day 1, with a

tsmwnanzs

SAptdcm(ti

CPr

Tua

LTfiwsXvew

MCtit2

CT

T

C

nAFRRAPRDDAL

CA

SpcA

507Vol. 200, No. 4, April 2005 Shapiro et al Transplantation with Reduced Immunosuppression

arget 12-hour trough level of 10 ng/mL. In a few cases,irolimus (or cyclosporine) was substituted for tacroli-us because of nephro- or neurotoxicity. Renal functionas monitored primarily with serum creatinine determi-ations. Suspected rejection was confirmed by biopsynd treated with one or more boluses of methylpred-isolone, muronmonab-CD3 (OKT3), or alemtu-umab. Oral steroids or other secondary agents such asirolimus were added only as necessary.

paced weaningt some time after 3 to 4 months in lymphoid depletedatients who had been stable on tacrolimus monotherapy,he twice daily doses were consolidated to a single dailyose. For example, someone on 2 mg twice a day would beonverted to a single 4-mg dose. One and a half or moreonths later, weaning to every-other-day dosing was begun

ie, 4 mg every other day). Subsequent weaning to threeimes weekly, twice weekly, and once weekly tacrolimus wasnstituted on an individualized basis.

ellular immunologic monitoringeripheral blood mononuclear cells were obtained from

able 1. Population Characteristics

haracteristicHistoriccontrols

152ccrual dates 3/00 to 7/01ollowup, (mo) 39 to 54ecipient age, (y) 50.6 � 14.9ecipient gender (M/F), % 61/39frican-American recipients, % 14rimary Tx/Re-Tx, % 78/22ecipient PRA � 20%, % 26onor age, (y) 35.5 � 18.0onor gender (M/F), % 57/43frican-American donors, % 10iving/cadaveric donors, % 20/80Sibling-sibling 12Parent-offspring 1Offspring-parent 3Other related 3Nonrelated (incl. spouse) 11

adaveric ischemia time, (h) 27.7 � 8.69BDR mismatch 3.21 � 1.61

ignificant variations of the lymphoid depletion/minimum immunosuppressiranging from 0.04 to less than 0.001. All significant differences in the p

haracteristics were statistically similar.BDR, ; PRA, panel reactive antibodies; Tx, transplant; Re-Tx

ecipients who had been followed up for 1 year or more. d

he cells were separated by a standard gradient centrif-gation method17 and used fresh for in vitro functionalssays.

eukocyte surface markershree- or four-color flow cytometry with appropriate

lurochrome conjugated mAb combinations, andsotype-matched nonspecific mAbs (negative controls)ere used to determine lymphocyte subsets. Data acqui-

ition and analysis was performed on a Coulter EPICSL flow cytometer (Beckman Coulter Corp). In con-entionally gated cells for lymphocytes, 50,000 CD45�

vents were typically collected per sample and analyzedith EXP032 software (Applied Cytometry System).17,18

ixed lymphocyte reactiononventional unidirectional mixed lymphocyte reac-

ion (MLR) cultures were set up for 6 days using 1 � 105

rradiated stimulator cells (2000R). The degree of [3H]hymidine incorporation was assessed during the final0 hours of incubation.

ell-mediated lymphocytoxicityhe cytolytic activity of recipient lymphocytes toward

Thymoglobulinpretreatment

Campathpretreatment

101 907/01 to 10/02 3/03 to 9/03

23 to 39 12 to 1851.1 � 14.5 50.8 � 16.6

65/35 63/3714 18

85/15 87/1318 19

39.4 � 15.4 41.7 � 16.040/60 47/53

9 846/54 39/61

14 66 47 55 4

14 1626.4 � 6.37 20.8 � 7.643.12 � 1.72 3.64 � 1.57

pulations as compared with the historic controls are highlighted in bold withtions are attributable to the cadaveric donor subgroup, because live donor

on poopula

onor and third party targets was assessed in cell-

508 Shapiro et al Transplantation with Reduced Immunosuppression J Am Coll Surg

Figure 1. Patient and graft survival in historic control (oo), Thymoglobulin-pretreated (X), and Campath-pretreated (black squares) kidney recipients.

mlpr

LCllcrfapnfali3

ALcwHmsw

PPrjw

ab

dds

Fc

Tn

H

T

C

p

* result

509Vol. 200, No. 4, April 2005 Shapiro et al Transplantation with Reduced Immunosuppression

ediated lymphocytotoxicity. In these assays, effectorymphocytes were incubated with Cr-labeled donorhytohemagglutinin-induced target cells at various E:Tatios, ranging from 10:1 to 30:1.

imiting dilution assayytolytic T lymphocyte precursor frequency was ana-

yzed as previously described by Kaminski and col-eagues19 using recipient cells as responders and donorells as stimulators, with the addition on days 4 and 6 ofecombinant interleukin-2 to fresh medium to give ainal concentration of 10 U/mL. On day 10, a cytotoxicssay was carried out for each culture using 51Cr- labeledhytohemagglutinin-induced blast cells from the origi-al stimulator as the targets. Supernatants harvestedrom each well were measured for 51Cr along withppropriate positive and negative controls; cytolytic Tymphocyte precursor frequency was considered as min-mum or background with frequencies of less than 1 in00,000.20,21

ntibody monitoringymphocytotoxic crossmatches were negative in allases. The pre- and posttransplant sera of 92 recipientsere screened by ELISA for the presence of IgG anti-LA class I and class II alloantibodies according to theanufacturer’s instructions (One Lambda Inc). The re-

ulting optical densities were analyzed by LATTM soft-are for Windows (One Lambda Inc).22,23

athologic studiesosttransplant biopsies were not taken on protocol butather because of clinical evidence (or suspicion) of re-ection. But a few biopsies in the historic control patients

able 2. Mean Creatinine (mg/dL) at Successive Posttrans1 wk 1 mo 3 m

istoric controls (n � 113)Mean 4.45 2.03 1.7SD 4.13 1.47 0.8

hymoglobulin (n � 83)Mean 3.90 2.24 1.8SD 4.32 1.14 1.0

ampath (n � 84)Mean 1.95 1.74 1.5SD 1.60 0.71 0.5Value �0.001 0.025 0.0

Removal of 5% (n � 4) of the outlying results for the Thymoglobulin group

ere obtained in pursuit of unrelated research projects, (

nd some of the lymphoid depleted patients had baselineiopsies before beginning spaced weaning.Tissues were handled according to hospital proce-

ures.15 Biopsy findings were categorized by the stan-ardized Banff system24 and CAN was scored with acale of 0 to 3 for each of four kinds of abnormality:

igure 2. Incidence and time to first acute rejection in historicontrol (oo), Thymoglobulin-pretreated (X), and Campath-pretreated

Times of Allografts that Still Function6 mo 1 y 2 y 3 y Current

1.68 1.56 1.48 1.53 1.640.59 0.57 0.56 0.79 1.02

1.60 1.58 1.89* — 2.08*0.62 0.55 0.95 — 1.23

1.47 1.59 — — 1.690.54 0.75 — — 0.830.051 0.957 �0.001 — —

s in 2-y and current creatinines of 1.74 � 0.07 and 1.86 � 0.86, respectively.

planto

59

48

4873

black squares) recipients.

510 Shapiro et al Transplantation with Reduced Immunosuppression J Am Coll Surg

Figure 3. Variable preweaning courses and weaning outcomes in Thymoglobulin-pretreated kidney recipients. Solid shade � dailytacrolimus (Tac) dosing. Spaced weaning represented by spikes that indicate frequency and dose. (A) Uncomplicated weaning fromdaily to once weekly doses of tacrolimus between 4 and 10 posttransplant mo in a cadaver kidney recipient who has been rejectionfree for more than 3 y. (B) Spaced weaning begun after 7 mo after a difficult rejection at 2½ mo. Tacrolimus doses have been threetimes per week for the past year in this recipient of a kidney from a live unrelated donor. (C) Reweaning in a patient who developeda mild rejection after 9 mo on one dose of tacrolimus per week. The eventual (and current) dose is three times per week. (D)Irreversible rejection of a cadaver kidney 4 wk after an attempt to reduce tacrolimus doses to every other day. The patient’s original

disease was lupus nephritis.

ga

DAowqfdtscpaKp�

RSSSttOcu0ldt

GCstcptT

RTdTtt

TCaTa

WASrcToisjbs

nswri

pd2m

ASpttbrj

81c

MIia

511Vol. 200, No. 4, April 2005 Shapiro et al Transplantation with Reduced Immunosuppression

lomerulopathy (cg), interstitial fibrosis (ci), tubulartrophy (ct), and chronic vasculopathy (cv).

ata management and statistical analysisn honest brokering system approved by the Universityf Pittsburgh IRB was used for data management. Dataere extracted, related, reviewed, augmented (where re-uired) for accuracy and completeness, and deidentifiedor statistical analysis. Differences in means and stan-ard deviations calculated from participant characteris-ics by treatment group were evaluated for the statisticalignificance using t-tests and ANOVAs for continuousomparisons and chi-square tests for categorical com-arisons. Kaplan-Meier survival curves were generatednd evaluated for significance using a log-rank test.aplan-Meier comparisons were adjusted for significantopulation differences to ascertain their effect. A p value0.05 two sided was considered significant.

ESULTSurvival and graft functionurvivalurvival of live donor grafts (Fig. 1) was better in bothhe Thymoglobulin- and Campath-pretreated patientshan in the live donor historic controls (p � 0.037).therwise, patient and graft survival were not signifi-

antly different in either of the lymphoid depletion pop-lations versus the historic controls overall (p � 0.12 to.59) or in the subgroups of the respective cadaveric or

iver donor recipients (p � 0.25 to 0.61). These findingsid not change when survival was adjusted for popula-ion differences.

raft functionampath-pretreated patients had the best early mean

erum creatinine, but there was no difference in thehree populations at 1 year (see Table 2). Mean serumreatinine at 2 years and currently in Thymoglobulin-retreated patients was higher than in historic con-rols, mostly because of four outliers (see footnote toable 2).

ate of acute rejectionhe three groups showed a markedly different inci-ence and time to acute rejection (Fig. 2). In thehymoglobulin-pretreated patients, the onset of rejec-

ion was earlier (p � 0.001) and the incidence was higher

han in either the Campath or historic control recipients. l

he incidence of rejection during the first 6 months afterampath pretreatment was 1%. Rejections that occurred

fter 6 months were frequently associated in both thehymoglobulin- and Campath-pretreatment groups with

ttempts to space wean (see below).

eaningfter Thymoglobulin pretreatmentpaced weaning was attempted in 91 (90.1%) of the 101ecipients after a mean of 5.9 � 1.4 months. Clinicalourses were highly variable before and after weaning.he weaning process was uncomplicated in the majorityf cases in which it was attempted (Fig. 3A and B). Butn 45% of the patients in whom spaced weaning wastarted, daily therapy was resumed because of acute re-ection. If the rejection promptly responded to 1 or 2oluses of prednisone, less aggressive spaced weaningubsequently was resumed (Fig. 3C).

Weaning either could not be successfully done or wasever attempted in about one-third of the recipients. Inome of these patients and in others in whom spacedeaned was never tried, grafts were lost to nonreversible

ejection. In the patient depicted in Figure 3D, the orig-nal renal disease was lupus nephritis.

With followups of 24 to 39 months, 68% of the 83atients with currently surviving grafts are on spacedoses of maintenance immunosuppression. Another5% are on daily monotherapy. Only 7% are receivingore than one immunosuppressant (Table 3).

fter Campath pretreatmentpaced weaning was attempted in 83 (91.2%) of the 90atients after a mean of 6.3 � 1.2 months. The decisiono go forward with weaning was more straightforwardhan it was in the Thymoglobulin-conditioned patientsecause of the nearly complete absence of preweaningejection. In addition, the incidence of postweaning re-ection (20%) has been lower (Fig. 2).

With followups of 12 to 18 months, 62 (74%) of the4 Campath-pretreated patients are on spaced weaning,4% are on daily monotherapy, and only 12% are re-eiving more than a single agent (Table 3).

orbidityn the historic controls, the incidence of symptomaticnfections by cytomegalovirus and by BK virus was 4.5%nd 3.2%, respectively. The incidence of posttransplant

ymphoproliferative disease was 2%, and 9.7% had new

olo2p

PAAsawsfctntp

CTtT(frrnao

ptscw

doa

gth1b

ptpsttT

CPocmrcstTsn

tTw(4

T

DMSOTTE

512 Shapiro et al Transplantation with Reduced Immunosuppression J Am Coll Surg

nset diabetes. These complications were not seen in theymphoid depleted patients, except for a 2% incidencef BK virus infection in the Thymoglobulin series and a.2% incidence of new onset diabetes in the Campathatients.

athologycute rejectioncute cellular rejection was diagnosed and graded for

everity according to the Banff 1997 schema of renalllograft pathology.24 Patients were biopsied onlyhen there was a rise in creatinine. Several of these

pecimens showed intimal arteritis, glomerulitis, andocal or diffuse C4d deposition. The incidence of sub-linical rejection and its effect on graft morphology inhe nonbiopsied patients cannot be addressed becauseo protocol biopsies were performed. It is notablehat only one rejection episode occurred during thereweaning period after a Campath infusion.

hronic allograft nephropathyhe numbers of recipients having a baseline biopsy within

he first 30 days and later biopsies in the historic control,hymoglobulin- and Campath-pretreated groups were 58

38%), 52 (51%), and only 8 (9%), respectively. Theserequencies largely reflected the relative incidence of earlyenal dysfunction from primary graft dysfunction or acuteejection. Although these and subsequent samples were ino sense protocol biopsies, enough histopathology wasvailable to tentatively assess the posttransplant progressionf CAN.

Only a minority of the baseline (control) biopsy sam-les were completely free of findings that contribute tohe 12-point maximum CAN score. The baseline CANcores averaged from 1.3 to 2.0 to 2.6 in the three suc-essively compiled series. The acceptance of kidneys

able 3. Current Immunosuppression for Recipients with Su

Immunosuppression

Historic controls(n � 113)

n %

aily monotherapy 46 41ultidrug therapy 67 59

paced dose weaning —nce a week —wice a week —hree times a week —very other day —

ith an increasing incidence and severity of preexisting C

onor disease reflected nationwide efforts to expand thergan pool, ie, by relaxing the criteria for organcceptance.

During the first year posttransplantation, CAN pro-ressed from baseline by 2.5 points in recipients pre-reated with Thymoglobulin, and by 3.0 points in theistoric control patients. The 1-year increase was only.4 points in the recipients pretreated with Campath,ut this was based on a very small number of biopsies.In years 2 and 3 posttransplantation, average CAN

rogressed an additional 1.7 and 0.34 points, respec-ively, for the Thymoglobulin patients and 1.5 and 1.2oints, respectively, in the historic controls. In the ab-ence of protocol biopsies, an important caveat to thehird year value is that very few samples were available athis time point: 13 in the reference group and 7 in thehymoglobulin group.

ellular immunologic monitoringretreatment with 5 mg/kg Thymoglobulin, as previ-usly reported,14,25 resulted in profound depletion of Tells (CD3�) by postoperative day 1. T cell counts re-ained below baseline for 3 months, with a gradual

eturn toward baseline values. The CD8 population re-overed more promptly than the CD4 population, re-ulting in an inverted CD4/CD8 ratio that persisted upo 6 months posttransplant. B cells were not depleted.he T cell depletion was considerably greater and more

ustained with Campath, and in addition, B cells wereotably depleted.Approximately 1 year after transplantation, more de-

ailed studies were obtained in eight of thehymoglobulin-pretreated patients. Seven of the eightere on spaced weaning, with doses of one per week

n � 2), two per week (n � 1), and three per week (n �). The eighth patient was on multiple drugs. The mean

ng GraftsThymoglobulin

(n � 83)Campath(n � 84)

n % n %

21 25 12 146 7 10 12

56 68 62 7412 —18 617 409 16

rvivi

D4/CD8 ratio, which had been 2.3 before lymphoid

dp

iwhstfmmwh

Msmwpwddtt

ASTnIappwsoafo

tafac

a2

oaotts

DTsbvcctgcd

qratmgl(eopttww

csw2actkhipc

513Vol. 200, No. 4, April 2005 Shapiro et al Transplantation with Reduced Immunosuppression

epletion, remained less than 1.0 in five of the eightatients between 1.38 and 1.5 in the other three.At 1 year, all seven of the patients on spaced wean-

ng had donor-specific hyporeactivity demonstratedith MLR; the patient on multiple drugs had globalyporesponsiveness. In six of the seven hyporespon-ive patients, kidney function was excellent at theime of testing. The kidney of the seventh patient wasailing at the time of testing (serum creatinine 3.7g/dL). Biopsy of this allograft revealed recurrentembranous glomerulonephritis similar to thathich had destroyed the native kidney, but with noistopathologic evidence of rejection.In additional studies of four of the seven patients withLR evidence of donor-specific hyporeactivity, the ab-

ence of donor cell killing was confirmed by cell-ediated lymphotoxicity assay, the essential absenceith limiting dilution assay of cytolytic T lymphocyterecursor, or both findings. But one of these patients,ho was on one dose per week at the time, subsequentlyeveloped a mild rejection that was treated with a singleose of prednisone and temporary reinstitution of dailyacrolimus. This patient is presently on three doses ofacrolimus per week.

ntibody monitoringerial samples were available from 92 of the 101hymoglobulin-pretreated recipients, all of whom hadegative conventional crossmatches with their donors.n 59 (64%) of these patients, nonspecific anti-HLAntibodies were never detectable before or after trans-lantation. In 14 of the other 33, anticlass I antibodiesredated transplantation (n � 12) or developed after-ard (n � 2). These disappeared in 9 who had adequate

ubsequent samples, and they were known to persist innly 2 of the others. Similarly, the isolated finding ofnticlass II antibodies before or after transplantation wasrequently temporary or it waxed and waned. There werenly three examples of a class switch or addition.

Combined anticlass I and class II antibodies predatedransplantation in five patients and developed de novofter 2, 4, and 12 months in the three others. Of interest,our of the five patients with both kinds of preformedntibodies received zero mismatched kidneys. In twoases, the antibodies disappeared after transplantation.

Except for a trend to poorer results in patients withnticlass II antibodies (ie, graft losses or death at 8, 21,

4, and 28 months), the preexistence or de novo devel-

pment of the anti-HLA antibodies did not appear to bessociated with clinical outcomes. C4D staining on bi-psy samples was not routinely carried out in these pa-ients. So in these patients, a meaningful correlation ofhe serum antibody and tissue C4D complement depo-ition was not possible.

ISCUSSIONhe efficacy of the multiple drug protocols of immuno-

uppression in wide use for kidney transplantation haseen judged primarily by how well these regimens pre-ent acute rejection. The extent to which this objectivean be accomplished was exemplified by our historicontrol group. Despite the low incidence of early rejec-ion, however, there was a steady erosion of patient andraft survival that usually was related in some way tohronic rejection, organ-specific drug toxicity, immuneepression per se, or combinations of these factors.With our revised strategy, improvement in the

uality of recipient life replaced avoidance of acuteejection as the highest priority. Because the immunectivation that can proceed to organ rejection also ishe mandatory first step of the seminal toleranceechanism of clonal exhaustion-deletion,5-7,26 the

reater than 50% rate of early and delayed acute cel-ular rejection in Thymoglobulin-pretreated patientsFig. 2) was not necessarily viewed with alarm. Nev-rtheless, the frequent need for urgent intensificationf immunosuppression mandated unusually closehysician surveillance. Despite vigilant supervision,here were examples of nonreversible acute rejectionhat proceeded to graft loss. These cases included, butere not limited to, patients who were on spacedeaning.The obvious question raised by these observations

oncerned the risk-to-benefit ratio of the tolerogenictrategy overall, with particular reference to the spacedeaning in the Thymoglobulin series. With followups ofto more than 3 years, patient survival, graft survival,

nd graft function of the Thymoglobulin-pretreated re-ipients are equivalent overall to the results in the his-oric controls. In the subgroup of recipients of live donoridneys, survival parameters are superior to those in theistoric controls. Although these results are encourag-

ng, the effect of the high rate of rejection on longtermrognosis in the Thymoglobulin-pretreated recipientsannot yet be definitively evaluated.

Protocol biopsies for histopathologic studies were not

sTtngdiTwdtcirtmt

atfmboTmserbac

weIuwtepcte

spq

ofwdawsttcvaar

dikpBmgotpme

ASA

ADCSS

AtPLAwoBF

514 Shapiro et al Transplantation with Reduced Immunosuppression J Am Coll Surg

ystematically obtained in either the historic control orhymoglobulin-pretreated patients. Instead, most of

he biopsies were done to confirm the clinically diag-osed rejections that were most common in the Thymo-lobulin cohort. In cases in which multiple biopsies wereone, the severity and progression of CAN were similar

n the Thymoglobulin and historic control recipients.he lesions under both kinds of immunosuppressionere particularly prevalent in kidneys from cadavericonors. This was not surprising because it is well knownhat preexisting donor disease and ischemic injury mayontribute to CAN27-29 and aggravate drug nephrotoxic-ty.30 It also is well established that any degree of acuteejection (even of the subclinical variety) may contributeo the arteriopathy, fibrosis, tubular atrophy, and glo-erular loss that were the criteria for our semiquantita-

ive histopathologic scale of CAN.30,31

Concerns about acute rejection were largely eliminatedfter Campath was substituted for Thymoglobulin as pre-reatment. With Campath, the incidence of rejection be-ore spaced weaning was 1%, and even afterward, the cu-ulative total has been only 20%. In addition, the trend of

etter patient and graft survival relative to historic controlsut to 1 to 2 years, are much the same as in thehymoglobulin-pretreated cohort. The superior perfor-ance of Campath may reflect, in part, the benefit of les-

ons previously accrued with the Thymoglobulin experi-nce. For example, because of rejections associated with tooapid weaning in the Thymoglobulin series, spaced dosingeyond every other day or three times a week is no longerttempted until at least 1 year unless there are specific indi-ations (eg, drug nephrotoxicity or neurotoxicity).

It should be emphasized that it is not yet knownhether the exceptionally complete elimination of acute

arly rejection with Campath will carry a delayed price.n addition to its long biologic effect (6 to 12 months),nbound Campath remains in the circulation for 1 to 2eeks after infusion (information from ILEX Inc), and

heoretically could erode the mechanism of clonalxhaustion-deletion in the same way as multidrugosttransplant therapy.7 If so, clinically silent CANould develop and not be detected, particularly becausehe rate of biopsy sampling has been very low with thefficient avoidance of acute rejection.

The most striking benefit of the tolerogenic immuno-uppression in both the Thymoglobulin- and Campath-retreatment series was the improvement in recipient

uality of life. This was reflected in a very low incidence J

f infection and of de novo malignancies, and freedomrom new onset insulin-dependent diabetes. The gainsere clearly associated with reduced exposure to chronicaily immunosuppression. In both the Thymoglobulinnd Campath series, only about 10% of the recipientsho still bear functioning grafts are on more than a

ingle immunosuppressive agent. The vast majority ofhe others are on spaced dose schedules of the mono-herapy, including 12 patients in the Thymoglobulinohort who are on one dose per week. Based on obser-ations in adults, the Thymoglobulin-based strategy wasdopted in April 2003 for all pediatric kidney recipientst our center, and more recently, Thymoglobulin waseplaced with Campath.

In conclusion, recipient pretreatment by lymphoidepletion combined with minimalistic posttransplant

mmunosuppression is an acceptable way to manageidney transplant recipients. Campath currently ap-ears to be the most effective means of pretreatment.ut the strategy, which is designed to permit naturalechanisms of tolerogenesis, is neither drug nor or-

an specific. It is most easily applied for live donorrgan transplantation, but can be readily used forransplantation of cadaveric organs. Additional im-rovements should be possible, including develop-ent of better guidelines for the optimal timing and

xtent of drug weaning.

uthor Contributionstudy conception and design: Starzlcquisition of data: Shapiro, Basu, Tan, Khan,Randhawa, Murase, Zeevi, Girnita, Metes, Demetris

nalysis and interpretation of data: Gray, Bass, Nessrafting of manuscript: Shapiroritical revision: Starzltatistical expertise: Bass, Nessupervision: Starzl, Fung, Marcos

cknowledgment: We gratefully acknowledge the contribu-ions of transplant surgeons who are no longer at University ofittsburgh Medical Center: Velma P Scantlebury, MD, MarkJordan, MD, Carlos Vivas, MD, Potdar Santosh, MD, andshok Jain, MD. The work would not have been possibleithout the indispensable help of Dr Jennifer Woodward andf the renal transplant coordinators: Cindy Anderson, Angelaarber, Linda Bonazza, Denise Boris, Cheryl Buzzard, Jareenlohr, Jami Gandy, Janice Glidewell, Debbie Good, Gerri

ames, Jackie Lever, Cordie McFeaters, Kim Meyer, Annie

Sf

R

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

515Vol. 200, No. 4, April 2005 Shapiro et al Transplantation with Reduced Immunosuppression

mith, Maureen Vekasy. We finally thank Ms Terry L Manganor the skillful preparation of the article.

EFERENCES

1. Starzl TE, Marchioro TL, Waddell WR. The reversal of rejectionin human renal homografts with subsequent development ofhomograft tolerance. Surg Gynecol Obstet 1963;117:385–395.

2. Starzl TE, Marchioro TL, Porter KA, et al. The use of heterolo-gous antilymphoid agents in canine renal and liver homotrans-plantation and in human renal homotransplantation. Surg Gy-necol Obstet 1967;124:301–318.

3. Starzl TE, Porter KA, Iwasaki Y, et al. The use of antilymphocyteglobulin in human renal homotransplantation. In: Wolsten-holme GEW, O’Connor M, eds. Antilymphocytic Serum.London: J and A Churchill Limited; 1967:4–34.

4. Starzl TE, Murase N, Demetris AJ, et al. Lessons of organ-induced tolerance learned from historical clinical experience.Transplantation 2004;77:926–929.

5. Starzl TE, Demetris AJ, Murase N, et al. Cell migration, chi-merism, and graft acceptance. Lancet 1992;339:1579–1582.

6. Starzl TE, Zinkernagel R. Antigen localization and migration inimmunity and tolerance. New Engl J Med 1998;339:1905–1913.

7. Starzl TE, Zinkernagel R. Transplantation tolerance from a his-torical perspective. NATURE Reviews: Immunology 2001;1:233–239.

8. Préville X, Flacher M, LeMauff B, et al. Mechanisms involved inantithymocyte globulin immunosuppressive activity in a non-human primate model. Transplantation 2001;71:460–468.

9. Mueller TF. Thymoglobulin: an immunologic overview. CurrOpin Organ Transplant 2003;8:305–312.

0. Hale G, Waldmann H, Dyer M. Specificity of monoclonal an-tibody Campath-1. Bone Marrow Transplant 1988;3:237–239.

1. Calne R, Friend P, Moffatt S, et al. Prope tolerance, periopera-tive campath 1H, and low-dose cyclosporin monotherapy inrenal allograft recipients. Lancet 1998;351:1701–1702. Erra-tum in: Lancet 1998;352:408.

2. Stuart FP, Leventhal JR, Kaufman DB, et al. Alemtuzumab fa-cilitates prednisone free immunosuppression in kidney trans-plant recipients with no early rejection. Am J Transplant 2002;2(suppl 3):397.

3. Knechtle SJ, Pirsch JD, H Fechner J Jr, et al. Campath-1Hinduction plus rapamycin monotherapy for renal transplanta-tion: results of a pilot study. Am J Transplant 2003;3:722–730.

4. Kirk AD, Hale DA, Mannon RB, et al. Results from a humanrenal allograft tolerance trial evaluating the humanized CD52-specific monoclonal antibody alemtuzumab (CAMPATH-1H).Transplantation 2003;76:120–129.

5. Starzl TE, Murase N, Abu-Elmagd K, et al. Tolerogenic immu-nosuppression for organ transplantation. Lancet 2003;361:1502–1510.

6. Shapiro R, Jordan M, Basu A, et al. Kidney transplantation

under a tolerogenic regimen of recipient pre-treatment andlow-dose postoperative immunosuppression, with subsequentweaning. Ann Surg 2003;238:520–527.

7. Boyum A. Separation of leukocytes from blood and bone mar-row. Introduction. Scand J Clin Lab Invest 1968;Suppl 97:7.

8. Metes D, Logar A, Rudert WA, et al. Four-color flow cytometricanalysis of peripheral blood donor cell chimerism. Hum Immu-nol 2003;64:787–795.

9. Kaminski E, Hows J, Goldman J, Batchelor R. Optimising alimiting dilution culture system for quantitating frequencies ofalloreactive cytotoxic T lymphocyte precursors. Cell Immunol1991;137:88–95.

0. Fussell ST, Donnellan M, Cooley MA, Farrell C. Cytotoxic Tlymphocyte precursor frequency does not correlate with eitherthe incidence or severity of graft-versus-host disease aftermatched unrelated donor bone marrow transplantation. Trans-plantation 1994;57:673–676.

1. Kaminski E, Hows J, Man S, et al. Prediction of graft versus hostdisease by frequency analysis of cytotoxic T cells after unrelateddonor bone marrow transplantation. Transplantation 1989;48:608–613.

2. Suciu-Foca N, Reed E, D’Agati VD, et al. Soluble HLA anti-gens, anti-HLA antibodies, and antiidiotypic antibodies in thecirculation of renal transplant recipients. Transplantation 1991;51:593–601.

3. Monteiro F, Buelow R, Mineiro C, et al. Identification of pa-tients at high risk of graft loss by pre- and posttransplant mon-itoring of anti-HLA class I IgG antibodies by enzyme-linkedimmunosorbent assay. Transplantation 1997;63:542–546.

4. Racusen LC, Solez K, Colvin RB, et al. The Banff 97 workingclassification of renal allograft pathology. Kidney Int 1999;55:713–723.

5. McCurry K, Iacano A, Zeevi A, et al. Early outcomes in humanlung transplantation utilizing thymoglobulin or campath 1Hfor recipient pretreatment followed by posttransplant tacrolimusnear-monotherapy. J Thorac Cardiovasc Surg, in press.

6. Starzl TE, Demetris AJ, Trucco M, et al. Cell migration andchimerism after whole-organ transplantation: The basis of graftacceptance. Hepatology 1993;17:1127–1152.

7. Isoniemi H, Taskinen E, Hayry R. Histological chronic allograftdamage index accurately predicts chronic renal allograft rejec-tion. Transplantation 1994;58:1195–1198.

8. Tullius SG, Tilney NL. Both akkiabtugeb-dependent and-independent factors influence chronic allograft rejection.Transplantation 1995;59:313.

9. Halloran RF, Melk A, Barth C. Rethinking chronic allograftnephropathy: the concept of accelerated senescence. J Am SocNephrol 1999;10:167–181.

0. Nankivell BJ, Borrows RJ, Fung CL-S, et al. Calcineurin inhib-itor nephrotoxicity: longitudinal assessment by protocol histol-ogy. Transplantation 2004;78:557–565.

1. Nankivell BJ, Borrows RJ, Fung CL, et al. The natural history ofchronic allograft nephropathy. N Engl J Med 2003;349:2326–

2333.