DGJBL

AOaecCrv

DDtDMntIPS

RFasDPCM7

©P

onor-Recipient Gender and Size Mismatch Affectsraft Success after Kidney Transplantation

ennifer McGee, MD, Jeanette H Magnus, MD, PhD, Tareq M Islam, MBBS, MPH,ernard M Jaffe, MD, FACS, Rubin Zhang, MD, FASN, Sander S Florman, MD, FACS,Lee Hamm, MD, FACP, Navyata Mruthinti, MPH, Karen Sullivan, PhD, Douglas P Slakey, MD, MPH, FACS

BACKGROUND: Female recipients of male kidneys have an inferior graft survival, and patients receiving largerkidneys relative to their body size may have a graft survival advantage. Thus, graft survival maybe affected by both gender and kidney size mismatches. The objective of this study was toanalyze the possible confounding effect of body mass mismatch (body mass as proxy for kidneysize) between female recipients of male donor kidneys.

STUDY DESIGN: A total of 668 kidney transplantations between 1996 and 2005 at our center were studiedretrospectively. Graft and patient survival were determined by Kaplan-Meier estimation. Mul-tivariate Cox proportional analyses were performed to determine the hazards of graft loss.

RESULTS: There were 146 female recipients of male kidneys. Compared with all other gender combinations,this group had the lowest unadjusted graft survival (86%, 79%, and 78% vs 92%, 88%, and 86% at1, 2, and 3 years, respectively; log-rank p � 0.01). Donor body mass index (BMI) correlated withdonor kidney size (p � 0.001). Male kidneys were a risk factor of graft loss for female recipients(hazard ratio [HR] 3.45, 95% CI 1.40 to 8.51, p � 0.01), but male donors with a larger BMI relativeto female recipients’ significantly reduced the risk (HR 0.19, 95% CI 0.05 to 0.67, p � 0.01).

CONCLUSIONS: The inferior graft survival for female recipients of male donor kidneys is mitigated by male donors

with a larger BMI. (J Am Coll Surg 2010;210:718–727. © 2010 by the American College of Surgeons)

Tmbao

vthtturtthmn

MSWr

lthough current registry data from the United Network ofrgan Sharing report similar graft survival rates for males

nd females,1 a 2005 systematic review on gender differ-nces in kidney transplantation identified 14 studies withontradicting results.2 More recently, an analysis from theollaborative Transplant Study3 demonstrated that female

ecipients of male donor kidneys had the worst graft sur-ival after the first year and up to 10 years post-transplant.4

isclosure Information: Nothing to disclose.r McGee is partially supported by award number K12HD043451 from

he Eunice Kennedy Shriver National Institute of Child Health & Humanevelopment.anuscript content is solely the responsibility of the authors and does not

ecessarily represent the official views of the Eunice Kennedy Shriver Na-ional Institute of Child Health & Human Development or the Nationalnstitutes of Health.resented at Southern Surgical Association 121st Annual Meeting, Hotprings, VA, December 2009.

eceived December 23, 2009; Accepted December 28, 2009.rom the Departments of Surgery (McGee, Islam, Jaffe, Mruthinti, Slakey)nd Medicine (Hamm, Sullivan), and the Tulane Abdominal Transplant In-titute (Zhang, Florman), Tulane University School of Medicine; and theepartment of Community Health Sciences, Tulane University School of

ublic Health and Tropical Medicine (Magnus), New Orleans, LA.orrespondence address: Jennifer McGee, MD, Tulane University School ofedicine, Department of Surgery, 1430 Tulane Ave, SL22, New Orleans, LA

p0112-2669.

7182010 by the American College of Surgeons

ublished by Elsevier Inc.

he authors hypothesized that an alloimmune responseediated by H-Y minor histocompatibility antigens could

e responsible. H-Y antigens have been associated withcute rejection in smaller gender mismatch investigationsf bone marrow5 and corneal and kidney transplants.6-8

Several investigations have suggested a graft survival ad-antage for recipients receiving larger kidneys relative toheir body size.9-11 Compared with males, females usuallyave smaller kidneys.12-14 It has been theorized that whenhe recipient’s metabolic demand exceeds the capacity ofhe smaller donor kidney, hyperfiltration from nephronnderdosing could occur.15,16 Larger donor kidney mass inelation to smaller recipient mass diminishes hyperfiltra-ion injury, and subsequently, immune-mediated rejec-ion.9 Earlier studies of donor-recipient gender mismatchave not explored the possible confounding of body sizeismatch between female recipients of male donor kid-

eys, which was the purpose of this study.

ETHODStudy participantsith Institutional Review Board approval, we conducted a

etrospective chart review of 863 consecutive kidney trans-

lantations performed between January 1996 and August

ISSN 1072-7515/10/$36.00doi:10.1016/j.jamcollsurg.2009.12.032

2casftDcisit

IPrHpatnajv

CkDwPtsf

DaBTrnrn

SDwgmaf

wtaddu

tmadwatufcmid

as

RDcT(fspbiysCtvp

719Vol. 210, No. 5, May 2010 McGee et al Donor-Recipient Gender and Size Mismatch

005 at Tulane University Medical Center. Cases were ex-luded if donor or recipient age, race, gender, height,nd/or weight values were not documented. Six hundredixty-eight subjects met the inclusion criteria. Missing in-ormation on graft status and/or death was obtained fromhe United Network for Organ Sharing and the Nationaleath Index, respectively. A database for data entry and

leaning was created using Microsoft Access 2003. A qual-ty check of the data entry was performed by randomlyelecting 10% of the final sample for double entry. Anndividual was trained and assigned exclusively to performhe quality check. The discrepancy rate was less than 5%.

mmunosuppression therapyatients received standard triple immunosuppression: ste-oids, tacrolimus or cyclosporine, and mycophenolic acid.igh risk patients, including those with a previous trans-

lant, 6-antigen human leukocyte (HLA) mismatches,nd/or a panel reactive antibody �20%, received induc-ion therapy with the interleukin (IL)-2 receptor antago-ist basiliximab. All patients received standard antifungal,ntibacterial, and cytomegalovirus prophylaxis. Acute re-ection (AR) was confirmed by kidney biopsy, and the se-erity was graded according to the Banff classification.17

orrelating donor body mass index and donoridney sizeonor kidney surface area was calculated (kidney length �idth) from measurements provided by Louisiana Organrocurement Agency stored records. Univariate correla-ions between donor body mass index (BMI) and kidneyize using Pearson’s correlation coefficient were calculatedor 205 male donors and 155 female donors.

onor-recipient body mass index matchnd mismatchMI (kg/m2) was calculated for donors and recipients.here is no consensus as to what constitutes a match of donor-

ecipient BMI. In this study, this concept was operationalizedoting a match if the donor’s BMI fell within � 2 units of theecipient’s BMI (Appendix 1). BMI values outside the desig-

Abbreviations and Acronyms

AR � acute rejectionBMI � body mass indexHLA � human leukocyte antigenHR � hazard ratioMDFR group � male donor to female recipient transplantsPRA � panel-reactive lymphocytotoxic antibody

ated parameters were categorized as mismatches. T

tatistical method and outcomes analysisonor and recipient demographic and clinical parametersere stratified by gender. Significant differences betweenroups were ascertained by least squared means and maxi-um likelihood ratio for continuous and categorical vari-

bles, respectively, and subsequently for the male donor-emale recipient group versus all others.

Kaplan-Meier estimates of graft and patient survivalere calculated for donors and recipients by gender and for

he male donor-female recipient group versus others. Sub-nalyses on survival were performed for living and deceasedonor recipients by donor gender, recipient gender, andonor-recipient gender combination. Log-rank p value wassed as a test of significance.A multivariate Cox proportional hazards model ascer-

ained independent associations of graft loss. Covariates forodel adjustment included deceased donor and recipient

ge, deceased donor and recipient black race, hypertension,iabetes mellitus, number of HLA mismatches, cold andarm ischemia times, peak panel-reactive lymphocytotoxic

ntibody (PRA), previous kidney transplantation, donorype, 30-day acute rejection, donor BMI greater than 2nits compared with recipient BMI, and male donor-emale recipient versus others. Estimates of risk were alsoalculated after censoring for death. A second regressionodel to assess the risk of graft loss was stratified by recip-

ent gender and adjusted for previous covariates plus maleonor BMI 2 units larger and male versus female donor.All statistical analyses were performed using SAS 9.1.3,

nd p values less than 0.05 were considered statisticallyignificant.

ESULTSonor-recipient demographics and baselinelinical characteristicshere were 146 male donor-to-female recipient transplants

MDFR group) and 213, 179, and 130 male-to-male,emale-to-male, and female-to-female combinations, re-pectively. In this study, the latter 3 combinations wereooled for most analyses to form a single group (others)ecause no statistical differences in demographic character-stics or outcomes were identified in the exploratory anal-ses. Donor and recipient age, race, and BMI did not differignificantly between the MDFR group and others (Table 1).ompared with recipients among the others, recipients in

he MDFR group had a higher prevalence of diabetes (31%s 22%, p � 0.03) but were less sensitized (70% vs 65% foreak PRA � 50 and 30% vs 35%, for PRA � 50, p � 0.01.

able 1, bottom panel).

PaRdcr

00

MKgjaf

wcsaadtdetts7l

glw((tga9

BmSacKd(sprjrgpB

TSD

C

D

R

Bl

720 McGee et al Donor-Recipient Gender and Size Mismatch J Am Coll Surg

earson’s correlation of donor body mass indexnd donor kidney sizeegardless of calculation method, either by combined gen-er or single gender, donor BMI and kidney size directlyorrelated (combined male and female donors [n � 360],

able 1. Demographics and Clinical Characteristics in aingle-Center Study of 668 Kidney Transplantations byonor-Recipient Gender Combination

haracteristic

Donor-recipient gender

p ValueMale-female(n � 146)

Others(n � 522)

onorMean age, y (SD) 28 (15) 35 (16) NSRace, %

White 62 71Black 30 21 NSOther 8 8

Male,% 100 41 �0.01Mean BMI, kg/m2 (SD) 25 (7) 26 (7) NSDonor type, %

Living 14 20 NSDeceased 86 80

ecipientMean age, y (SD) 43 (16) 43 (17) NSRace, %

White 38 40Black 59 56 NSOther 3 4

Male,% 0 75 �0.01Mean BMI, kg/m2 (SD) 27 (7) 27 (7) NSPast medical history, %

Diabetes mellitus 31 22 0.03Hypertension 86 88 NSPolycystic kidney disease 7 7 NSCoronary artery disease 10 13 NSPeripheral vascular disease 5 6 NS

Previous kidneytransplant, % 10.6 11.6 NS

Mean cold ischemia time, h(SD) 9.4 (10.4) 10.4 (10.9) NS

Mean warm ischemia time,min (SD) 13.8 (17.9) 14.3 (19.0) NS

Peak PRA, %�50 70 65 �0.01�50 30 35

HLA mismatches, %0–1 20 262–4 47 42 NS5–6 33 32

MI, body mass index; HLA, human leukocyte antigen; PRA, panel-reactiveymphocytotoxic antibodies.

� 0.40, p � 0.0001; male donors only (n � 205), r � h

.40, p � 0.0001; and female donors only (n � 155), r �

.40, p � 0.0001).

DFR graft survival and the risk of graft lossaplan Meier estimates of graft survival for the MDFRroup were significantly lower than for the others. Unad-usted graft survivals at 1, 2, and 3 years were 86%, 79%,nd 78% for the MDFR group, and 92%, 88%, and 86%or the others (log-rank p � 0.01, Fig. 1).

Subanalyses of graft survival according to donor typeere also performed (data presented as text only). Overall,

ompared with deceased donors, living donors provided aignificantly better rate of graft survival at 1, 2, and 3 yearsfter transplantation (92%, 92%, and 91% vs 90%, 85%,nd 83%, respectively, log-rank p � 0.05). When livingonors were stratified by the MDFR group versus others,he rate of graft survival was consistently poorer for maleonor-to-female recipient transplants, although the differ-nce was not statistically significant. For deceased donorransplants analyzed by donor-recipient gender combina-ions, compared with the others, the MDFR group hadignificantly lower graft survival rates (86%, 78%, and7% at 1, 2, 3 years vs 91%, 87%, and 85%, respectively,

og-rank p � 0.03).A multivariate adjusted Cox proportional analysis of

raft loss demonstrated that the hazard ratio (HR) of graftoss for additional HLA mismatch among all study patientsas 1.24 (95% CI 1.08 to 1.44, p � 0.01) and was 1.29

95% CI 1.10 to 1.52, p � 0.01) after censoring for deathTable 2). When compared with others, a strong associa-ion was demonstrated between the MDFR group andraft loss (HR 1.78, 95% CI 1.12 to 2.85, p � 0.02). Thessociation increased after censorship for death (HR 1.97,5% CI 1.18 to 3.30, p � 0.01).

MI matches and mismatches, gender and sizeatches and mismatches, and the risk of graft loss

eparate sets of analyses were performed to identify associ-tions of graft loss between different donor-recipient BMIombination groups (data presented as text only). From aaplan Meier graft survival estimate, there was no survivalifference among 3 categories of BMI combinationsmatch, larger donor, and larger recipient).Three-year grafturvivals were 84%, 86%, and 84%, respectively, log-rank

value � 0.86). To further assess the effect of donor-ecipient BMI matching we performed a multivariate ad-usted Cox regression hazards model within each donor-ecipient gender combination subgroup. The donor-recipientender combinations that composed the other group were notooled for this analysis, and the donor-recipient matchingMI group acted as reference. The MDFR group displayed

azard ratios of graft loss among larger donors and larger

r0igr

btd8

TT

V

DRBBHDNCWPPD3DM

B

721Vol. 210, No. 5, May 2010 McGee et al Donor-Recipient Gender and Size Mismatch

ecipients of 0.41 (95% CI 0.09 to 1.87) and 0.96 (95% CI.26 to 3.59), respectively. Though not statistically signif-cant, except for the male donor-to-male recipient sub-roup, having a larger donor was associated with a lowerisk of graft loss compared with having a larger recipient.

able 2. Multivariate Adjusted Cox Proportional Hazardransplantations

ariable

onor age (continuous) 0ecipient age (continuous) 1lack race vs non-black race (donor) 0lack race vs non-black race (recipient) 1ypertension (yes vs no) 1iabetes mellitus (yes vs no) 0umber of HLA mismatches (continuous) 1old ischemia time (continuous) 0arm ischemia time (continuous) 1

eak PRA (�50 vs �50) 1revious kidney transplant (yes vs no) 1onor type (deceased vs living donor) 1

0-day acute rejection (yes vs no) 5onor BMI �2 units compared with recipient vs others 0ale donor-female recipient vs others 1

Figure 1. 1-, 2-, and 3-year graft survival in a singdonor-female recipient gender combination (86%,and 86%, respectively).

MI, body mass index; HLA, human leukocyte antigen; PRA, panel-reactive lymp

An additional multivariate model of graft loss stratifiedy recipient gender demonstrated that for female recipientshe risk of graft loss was more than 3 times higher if theonor was male versus female (HR 3.45 95% CI 1.40 to.51, p � 0.01). An analogous risk did not exist for male

f Graft Loss in a Single-Center Study of 668 Kidney

All patients Death–censoredard ratio p Value Hazard ratio p Value

.97–1.00) NS 0.99 (0.97–1.01) NS

.98–1.01) NS 0.98 (0.97–1.00) NS

.59–1.69) NS 0.81 (0.45–1.47) NS

.73–2.02) NS 1.21 (0.69–2.13) NS

.51–2.09) NS 1.22 (0.56–2.67) NS

.32–1.06) NS 0.76 (0.40–1.44) NS

.08–1.44) �0.01 1.29 (1.10–1.52) �0.01

.97–1.02) NS 0.99 (0.96–1.02) NS

.99–1.02) NS 1.00 (0.98–1.01) NS

.66–1.92) NS 0.84 (0.45–1.55) NS

.64–2.52) NS 1.61 (0.79–3.28) NS

.52–2.31) NS 1.26 (0.55–2.89) NS

.16–15.12) �0.01 5.10 (1.67–15.5) �0.01

.58–1.52) NS 0.84 (0.49–1.44) NS

.12–2.85) 0.02 1.97 (1.18–3.30) 0.01

ter study of 668 kidney transplantations by maleand 78%, respectively) versus others (92%, 88%,

s o

Haz

.98 (0

.00 (0

.95 (0

.21 (0

.03 (0

.58 (0

.24 (1

.99 (0

.00 (0

.13 (0

.27 (0

.10 (0

.72 (2

.94 (0

.78 (1

le-cen79%,

hocytotoxic antibody.

rana(arwti0

l00pde0kmt1

GTdwfp5

ssp

DOddpbetgB

ttrntHeiattgcl

TR

V

DRBBHDNCWPPD3MM

B lymp

722 McGee et al Donor-Recipient Gender and Size Mismatch J Am Coll Surg

ecipients (Table 3). After adjusting for all other risk factorsnd compared with others, female recipients of male do-ors whose BMI was 2 units larger than their own BMI hadstatistically significant reduction in the risk of graft loss

HR 0.19, 95% CI 0.05 to 0.67, p � 0.01). Again, annalogous risk did not exist for male recipients. In a sepa-ate model not presented in table form, female recipientshose donor BMI regardless of gender was 2 units larger

han their own BMI had a statistically significant reductionn the risk of graft loss (HR 0.23, 95% CI 0.08 to 0.68, p �.01).Increasing HLA mismatches augmented the risk of graft

oss for both male (HR 1.31, 95% CI 0.99 to 1.75, p �.03) and female (HR 1.44, 95% CI 1.10 to 1.87, p �.01) recipients. For the covariates cold ischemia time andrevious kidney transplantation, the hazards of graft lossiffered by gender. Compared with males, female recipi-nts appeared less susceptible to cold ischemia time (HR.95 95% CI 0.90 to 0.99, p � 0.02). Receiving a previousidney transplant did not significantly affect graft loss forale recipients, but female recipients had more than 4

imes the risk of graft failure (HR 4.46, 95% CI 1.43 to3.8, p � 0.01).

ender and acute rejectionhe prevalence of 30-day acute rejection (AR) did notiffer by male donor-female recipients versus others. ARas, however, associated with a significant risk of graft loss

or the whole sample (HR 5.72, 95% CI 2.16 to 15.12,� 0.01), as well as the death censored population (HR

able 3. Multivariate Adjusted Cox Proportional Hazards of Decipients and 276 Female Recipients

ariable

onor age (continuous)ecipient age (continuous)lack race vs non-black race (donor)lack race vs non-black race (recipient)ypertension (yes vs no)iabetes mellitus (yes vs no)umber of HLA mismatch (continuous)old ischemia time (continuous)arm ischemia time (continuous)

eak PRA (�50 vs �50)revious kidney transplant (yes vs no)onor type (deceased vs living donor)

0-day acute rejection (yes vs no)ale donor BMI �2 units compared with recipient vs othersale donor vs female donor

MI, body mass index; HLA, human leukocyte antigen; PRA, panel-reactive

.10, 95% CI 1.67 to 15.5, p � 0.01) (Table 2). When t

tratified by recipient gender, 30-day AR was a statisticallyignificant risk for males (HR 13.5, 95% CI 2.44 to 74.5,� 0.01) but not for females (Table 3).

ISCUSSIONur data support previous works reporting that the male

onor-to-female recipient transplant combination is an in-ependent risk factor for inferior graft survival. More im-ortantly, our study showed a confounding effect whenody mass mismatch (kidney size mismatch) was consid-red for male donor-female recipient kidney transplanta-ions. We demonstrated a significantly decreased risk ofraft loss for female recipients of male kidneys when donorMI was 2 or more units larger than recipient BMI.A seminal article by Gratwohl and colleagues4 retrospec-

ively studied almost 160,000 deceased donor kidneyransplant recipients and demonstrated in a multivariateegression model that female recipients of male donor kid-eys had an increased risk of poor graft survival. Investiga-ors in the study speculated that the H-Y antigens or the-Y effect were responsible. H-Y antigens are ubiquitously

xpressed18,19 and despite a “minor” classification, have beenmplicated in clinically relevant immunologic events suchs graft versus host disease after bone marrow transplan-ation5 and AR in solid organ transplantation (cornealransplants6 and kidney transplants7,8). Non-HLA anti-ens such has H-Y have been poorly characterized in theurrent literature. Most recently, however, Tan and col-eagues8 measured H-Y antibody expression in kidney

-Censored Graft Loss in a Single Center Study of 392 Male

Male recipients Female recipientszard ratio p Value Hazard ratio p Value

(0.97–1.02) NS 0.99 (0.96–1.02) NS(0.96–1.01) NS 0.97 (0.94–1.00) NS(0.13–1.55) NS 0.63 (0.27–1.46) NS(0.66–4.63) NS 1.20 (0.47–3.06) NS(0.31–3.24) NS 2.71 (0.64–11.40) NS(0.15–1.85) NS 1.34 (0.57–3.15) NS(0.99–1.75) 0.03 1.44 (1.10–1.87) 0.01(0.99–1.07) NS 0.95 (0.90–0.99) 0.02(0.95–1.00) NS 1.01 (0.99–1.04) NS(0.48–5.82) NS 0.42 (0.17–1.04) NS(0.18–2.48) NS 4.46 (1.43–13.8) 0.01(0.14–2.64) NS 2.67 (0.55–12.9) NS(2.44–74.5) �0.01 1.49 (0.24–9.29) NS(0.72–4.81) NS 0.19 (0.05–0.67) 0.01(0.35–2.33) NS 3.45 (1.40–8.51) 0.01

hocytotoxic antibody.

eath

Ha

0.990.980.441.751.010.521.311.030.981.680.660.6013.51.860.91

ransplant recipients. The investigators found that com-

pms(

lgaeddwgcMsbefsmaitmaanaant

dcwiritplatTftndco

n10dbrboc

tpot1smptrlwFlc

mcbawgb2dThtsrgHptd

acsd

723Vol. 210, No. 5, May 2010 McGee et al Donor-Recipient Gender and Size Mismatch

ared with all other donor-recipient gender groups, theale donor-female recipient population demonstrated a

ignificantly higher percentage of H-Y antibody production54% vs 7%).8

It is not known whether the H-Y effect is a short- or aong-term phenomenon. Several investigations have sug-ested that the H-Y effect can be clinically apparent as earlys 2 weeks post-transplantation.7,20 So, to assess the H-Yffect in our population we analyzed the frequency of 30-ay AR episodes by donor gender and donor-recipient gen-er combination. In our study the prevalence of 30-day ARas statistically similar for recipients of male and femalerafts. Also, 30-day AR events by donor-recipient genderombination failed to differ significantly between the

DFR group and the others. Therefore, our data do notupport an early H-Y effect after kidney transplantation,ut we recognize that our null result does not definitivelyxclude a late effect, especially the potential long-term in-luence. Although time to rejection was not reported in thetudy by Tan and associates,8 26 demographically similarale donor-to-female recipient transplantations were ex-

mined. The investigators found that compared with recip-ents not experiencing AR, recipients experiencing rejec-ion developed de novo anti-HY antibodies significantlyore often: 79% vs 8% (p � 0.00048).8 Because the H-Y

ntigens measured in the study were intracellular cytosolicnd nuclear, authors hypothesized that de novo antibodies,ot preformed, are responsible for the strong, H-Y initiatedlloimmune response associated with AR.8 Further, char-cterization of the H-Y minor histocompatibility antigen iseeded to fully understand its impact in kidney transplan-ation outcomes.

In our adjusted analyses looking at the influence of 30-ay AR on graft survival, the entire population was at in-reased risk for graft loss when AR occurred. Interestingly,hen stratified by recipient gender there was a significantly

ncreased risk for males, but not females. Of the 392 maleecipients in the study almost half had female donors. Thiss relevant with respect to early gender studies in kidneyransplantation because female donors were frequently im-licated in inferior outcomes. Vereerstraeten and col-

eagues21 reviewed 741 deceased donor transplantationsnd found after multivariate analysis that female donationo male recipients increased the risk of graft failure by 60%.he authors observed a greater number of technical failures

or the female donor-male recipient transplant group, andhey proposed greater immunogenicity among female do-ors.21 Similarly, a registry study of more than100,000 firsteceased donor kidney transplantations concluded thatompared with female recipients of female donors, the risk

f graft loss for male recipients of female donors was sig- t

ificantly greater (relative risk [RR] 1.22, 95% CI 1.16 to.29, p � 0.0001 vs RR 1.15, 95% CI 1.07 to 1.23, p �.0001).22 These investigators speculated that nephron un-erdosing and donor immunologic factors were responsi-le for the poor graft outcomes of the female donor to maleecipient group.22 To completely discern the relationshipetween the female donor-male recipient group and graftutcomes, we would need to include examination of asso-iated donor-recipient size mismatch.

Another group investigating the role of gender in kidneyransplantation studied 30,000 living donor kidney trans-lants, and found in an adjusted analysis, as we found forur deceased donors, that male donor-to-female recipientransplants had significantly increased risk of graft loss (HR.19, 95% CI 1.09 to 1.30).23 Interestingly, investigatorspeculated that in the context of nephron underdosing theale-to-female transplant combination should have im-

roved graft survival because male kidneys are characteris-ically larger.23 Although our living donor analysis did noteach statistical significance, compared with the others, theiving donor MDFR group demonstrated a trend toward aorse survival rate at all 3 years reviewed post-transplant.ailure of significance was most likely due to a smalleriving donor sample size for the MDFR group (n � 21)ompared with all others (n � 102).

Previous studies examining donor-recipient size mis-atch suggested a graft survival advantage for kidney re-

ipients who receive larger organs in relation to their ownody size.9-11 An early study described the magnitude ofdvantage comparable to the benefits for grafts implantedithin 48 hours of procurement as well as HLA-matchedrafts.15 Investigators have also described the advantage aseginning at 5 years after transplantation, with an almost0% difference in graft survival between recipients of largeronors compared with donors of similar or smaller size.24

he pathophysiology has been explained by larger kidneysaving more glomeruli,14,25 and more glomeruli translateso less susceptibility to progressive renal failure.13 As a con-equence, in the setting of a larger donor compared withecipient, there is less metabolic demand on the donorraft, and in turn, less hyperfiltration-induced injury.15,16,26

yperfiltration from nephron underdosing has been hy-othesized to initiate pathologic, structural kidney damageo the transplanted kidney, ultimately resulting in graftysfunction and failure.15,16

After confirming the correlation between donor BMInd kidney size in our population, donor BMI was ac-epted as a proxy for kidney size in concordance with earliertudies.12,14,25,27 We then tested the relationship betweenonor-recipient gender, donor-recipient size, and kidney

ransplantation outcomes. Multivariate analyses found that

aroftdOddoibmi

reimdimmnslttseihwhsct

efattbspd

die

ertt

A

SAADC

SAS

AMSPraokc

R

724 McGee et al Donor-Recipient Gender and Size Mismatch J Am Coll Surg

larger donor BMI compared with that of the recipient,egardless of donor gender, was protective against the riskf graft loss. This relationship was particularly strong foremale recipients of male donor kidneys. To our knowledgehis is the first report investigating the combined effects ofonor-recipient gender and size on kidney graft outcomes.ur results suggest that perhaps the donor-recipient gen-

er and size relationship should be considered in deceasedonor kidney allocation schemas and that female recipientsf male kidneys may benefit from protocol-driven, tailoredmmunosuppression to improve their outcomes. However,efore implementation of changes in clinical practice,uch more knowledge is needed on the impact of the

nteraction between donor-recipient gender and size.As expected, regardless of gender, our data describing the

elationship between HLA mismatches and adverse graftvents are consistent with established conclusions affirm-ng the risk of poor graft outcomes as the number of HLA

ismatches increases.28 Somewhat unexpectedly, genderifferences were identified for the cold ischemia variable. It

s difficult to explain why female recipients in our cohortay be less susceptible to longer cold ischemia times. Theajority of our female patients received male donor kid-

eys. Although controversial, male kidney grafts have beenhown to be less susceptible to cold ischemia.22 We specu-ate, however, that the result in our study has limited sta-istical and/or clinical relevance because the upper limit ofhe 95% confidence interval approaches unity. Our dataupport existing reports that compared with male recipi-nts, female recipients may have increased immunoreactiv-ty because of sensitizing events such as pregnancy andigher prevalence of autoimmune diseases,29 and comparedith first transplants, repeat kidney transplant recipientsave worse graft survival.30,31 Taken together, it was noturprising that compared with male recipients, female re-ipients in our study were at increased risk of graft loss ifhey had been a previous kidney transplant recipient.

This study was limited by its retrospective nature. How-ver, the single-center design minimized center-specific ef-ects such as differences in immunosuppression protocolsnd peri- and postoperative care. The percentage of pa-ients in this population receiving induction therapy oracrolimus versus cyclosporine did not significantly differetween the MDFR group and others, but because of limitedample size we were unable to analyze further by immunosup-ression regimens. Although it is unlikely based on prevalenceata, differing regimens might alter outcomes.

In conclusion, this study corroborated that the maleonor-to-female recipient kidney transplant combination

s an independent risk factor for poor graft survival. How-

ver, when an implanted kidney is from a larger donor this 1

ffect appears to be mitigated. As consideration is given toedefining the current kidney allocation system, perhapshe influence of donor size and gender effect should beaken into consideration.

uthor Contributions

tudy concept and design: McGee, Magnus, Jaffecquisition of data: McGee, Mruthinti, Sullivannalysis and interpretation of data: McGee, Magnus, Islamrafting of manuscript: McGee, Magnusritical revision: McGee, Magnus, Zhang, Jaffe, Florman,

Hamm, Slakeytatistical analysis: Islamdministrative, technical, or material support: McGee, Mruthintitudy supervision: McGee

cknowledgment: The authors would like to thank Paul Muntner, PhD for his early support in the project’s design;

udesh K Srivastav, PhD, for statistical assistance; Drs Anilaramesh, Mary Killackey, and Eddie L Hoover for criticallyeviewing the manuscript; Tiffany Haydel, Debbie Jeffreys,nd the staff at the Louisiana Organ Procurement Agency forrganizing the kidney size data; Joel Wells for abstracting theidney size data; and the many students dedicated to dataollection.

EFERENCES

1. OPTN data as of February 13, 2009. Available at: www.optn.org. Accessed February 13, 2009.

2. Jindal RM, Ryan JJ, Sajjad I, et al. Kidney transplantation andgender disparity. Am J Nephrol 2005;25:474–483.

3. Collaborative Transplant Study. Available at: http://www.ctstransplant.org/. Accessed February 3, 2010.

4. Gratwohl A, Dohler B, Stern M, Opelz G. H-Y as a minorhistocompatibility antigen in kidney transplantation: a retro-spective cohort study. Lancet 2008;372:49–53.

5. Voogt PJ, Goulmy E, Fibbe WE, et al. Minor histocompatibilityantigen H-Y is expressed on human hematopoietic progenitorcells. J Clin Invest 1988;82:906–912.

6. Bohringer D, Spierings E, Enczmann J, et al. Matching of theminor histocompatibility antigen HLA-A1/H-Y may improveprognosis in corneal transplantation. Transplantation 2006;82:1037–1041.

7. Pfeffer PF, Thorsby E. HLA-restricted cytotoxicity against male-specific (H-Y) antigen after acute rejection of an HLA-identicalsibling kidney: clonal distribution of the cytotoxic cells. Trans-plantation 1982;33:52–56.

8. Tan JC, Wadia PP, Coram M, et al. H-Y antibody developmentassociates with acute rejection in female patients with male kid-ney transplants. Transplantation 2008;86:75–81.

9. Sanchez-Fructuoso AI, Prats D, Marques M, et al. Does renalmass exert an independent effect on the determinants ofantigen-dependent injury? Transplantation 2001;71:381–386.

0. Poggio ED, Hila S, Stephany B, et al. Donor kidney volume and

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

D

Dg

nnd

cfc

othpwP

acos

cmrw

odas

rtAt

tauldselm

725Vol. 210, No. 5, May 2010 McGee et al Donor-Recipient Gender and Size Mismatch

outcomes following live donor kidney transplantation. Am JTransplant 2006;6:616–624.

1. Amante AJ, Pinon-Barretto SC. The correlation of renal allo-graft weight to metabolic index ratios and glomerular filtrationrate among living-unrelated kidney transplant patients: a cross-sectional study. Transplant Proc 2008;40:2313–2318.

2. Giral M, Nguyen JM, Karam G, et al. Impact of graft mass onthe clinical outcome of kidney transplants. J Am Soc Nephrol2005;16:261–268.

3. Hoy WE, Douglas-Denton RN, Hughson MD, et al. A stereo-logical study of glomerular number and volume: preliminaryfindings in a multiracial study of kidneys at autopsy. Kidney IntSuppl 2003:S3137.

4. Neugarten J, Kasiske B, Silbiger SR, Nyengaard JR. Effects ofsex on renal structure. Nephron 2002;90:139–144.

5. Brenner BM, Milford EL. Nephron underdosing: a pro-grammed cause of chronic renal allograft failure. Am J KidneyDis 1993;21:66–72.

6. Terasaki PI, Koyama H, Cecka JM, Gjertson DW. The hyper-filtration hypothesis in human renal transplantation. Transplan-tation 1994;57:1450–1454.

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

8. Billingham RE, Silvers WK. Studies on tolerance of the Y chro-mosome antigen in mice. J Immunol 1960;85:14–26.

9. Goulmy E. Human minor histocompatibility antigens. CurrOpin Immunol 1996;8:75–81.

0. Spierings E, Vermeulen CJ, Vogt MH, et al. Identification ofHLA class II-restricted H-Y-specific T-helper epitope evokingCD4� T-helper cells in H-Y-mismatched transplantation. Lan-cet 2003;362:610–615.

1. Vereerstraeten P, Wissing M, De Pauw L, et al. Male recipients ofkidneys from female donors are at increased risk of graft lossfrom both rejection and technical failure. Clin Transplant 1999;13:181–186.

2. Zeier M, Dohler B, Opelz G, Ritz E. The effect of donor genderon graft survival. J Am Soc Nephrol 2002;13:2570–2576.

3. Kayler LK, Rasmussen CS, Dykstra DM, et al. Gender imbal-ance and outcomes in living donor renal transplantation in theUnited States. Am J Transplant 2003;3:452–458.

4. el-Agroudy AE, Hassan NA, Bakr MA, et al. Effect of donor/recipient body weight mismatch on patient and graft outcome inliving-donor kidney transplantation. Am J Nephrol 2003;23:294–299.

5. Nyengaard JR, Bendtsen TF. Glomerular number and size inrelation to age, kidney weight, and body surface in normal man.Anat Rec 1992;232:194–201.

6. Oh CK, Lee BM, Jeon KO, et al. Gender-related differences ofrenal mass supply and metabolic demand after living donorkidney transplantation. Clin Transplant 2006;20:163–170.

7. Kasiske BL, Umen AJ. The influence of age, sex, race, and bodyhabitus on kidney weight in humans. Arch Pathol Lab Med1986;110:55–60.

8. Opelz G. New immunosuppressants and HLA matching. Trans-plant Proc 2001;33:467–468.

9. Meier-Kriesche HU, Ojo AO, Leavey SF, et al. Gender differ-ences in the risk for chronic renal allograft failure. Transplanta-tion 2001;71:429–432.

0. Pour-Reza-Gholi F, Nafar M, Saeedinia A, et al. Kidney retrans-plantation in comparison with first kidney transplantation.

Transplant Proc 2005;37:2962–2964. o

1. Magee JC, Barr ML, Basadonna GP, et al. Repeat organ trans-plantation in the United States, 1996-2005. Am J Transplant2007;7:1424–1433.

iscussion

R DANIEL J FREY (Lafayette, LA): I want to thank Dr Slakey foretting me the paper well in advance to review.

As he indicated, a transplant’s success revolves around the combi-ation of donor characteristics, recipient characteristics, and immu-osuppressive manipulation of the recipient in response to thoseonor characteristics.Numerous authors have looked at the female-to-male donor re-

ipient match as well as the size match. To my knowledge, this is theirst paper that’s looked at the combination of the two. I want toommend Dr Slakey for taking this on.

The paper is a retrospective review, and it has all of the limitationsf the retrospective review. He captured approximately 75% of theotal patients in his data analysis. I did notice that your definition ofigh risk in your immunologic monitoring was greater than 20%anel reactive antibodies (PRA); the cut-off in your statistical analysisas 50%, and I wonder if there’s any reason that you changed yourRA definition for high risk.Other details of the immunosuppressive management—how they man-

ged the rejection episodes and what their target drug levels are—is notlearlydefined.Itwouldbeverydifficult todoaretrospectivestudy,especiallyne spanning 10 years, with the changes that occurred in immunosuppres-ive care over that time. I have 4 questions for the authors.

I assume that Prograf was used in the later aspect of your study, andyclosporine was used in the initial part of your study, as that’s whatany of us have done. Did you subanalyze the male donor-to-female

ecipient so that you could see if Prograf may have had a protective effectith a questionable, more potent, immunosuppressive protocol?Along those same lines, now that you identified this, do you plan

n changing your immunosuppressive protocol to consider a maleonor-to-female recipient a high-risk transplant, consideration forntiantibody preparations initially, or possible use of Rapamune orome other immunologic alteration?

Do you plan on using BMI to match the male donor to femaleecipient in your personal practice habits? I must admit that I do tryo match according to BMI, now that the articles have come out.lthough under good immunologic conditions, I will go ahead and

ake an HLA identical match into a larger recipient.How do you treat rejection when it occurs early? Do you believe

hat increased steroids and increased baseline immunosuppressionre sufficient, or especially in the male-to-female group, should youse an antibody preparation? Finally, because the talk at the national

evel is about functional life years of transplantation in an organistribution model, do you believe that this should be initiated inome way into the distribution plan at the national level, ie, prefer-ntially give male kidneys to male recipients, or in another way toook at it, should there be some decrease in the points awarded to a

ale kidney to a female recipient so that only a good match would

ver-ride the BMI match?

A

P

B*

725.e1Vol. 210, No. 5, May 2010 McGee et al Donor-Recipient Gender and Size Mismatch

ppendix 1. Operationalizing Body Mass Index Match and Mismatch in Kidney Transplantation

atient

Example determinations of donor-recipient body mass index match*

Donor BMI, kg/m2 Recipient BMI, kg/m2Donor-recipient BMI

unit differenceBMI match or

mismatch Final determination

1 26 27 �1 M Match2 26 30 �4 MM Mismatch-larger recipient3 35 25 �10 MM Mismatch-larger donor

MI, body mass index; M, match; MM, mismatch.Donor-recipient BMI within � 2 units.