Montgomery RA, Zachary AA, Ratner LE, Segev DL, Hiller JM, Houp J, Cooper M, Kavoussi L, Jarrett T, Burdick J, Maley WR, Melancon JK, Kozlowski T, Simpkins CE, Phillips M, Desai A, Collins V, Reeb B, Kraus E, Rabb H, Leffell MS, Warren DS. Clinical Results From Transplanting Incompatible Live Kidney Donor/Recipient
Pairs Using Kidney Paired Donation. JAMA. 2005;294(13):1655-1663. doi:10.1001/jama.294.13.1655
Author Affiliations: Departments of Surgery
(Drs Montgomery, Segev, Cooper, Burdick, Maley, Melancon, Kozlowski, Simpkins,
Phillips, and Warren, Mss Hiller, Collins, and Reeb, and Mr Desai), Medicine
(Drs Zachary, Kraus, Rabb, and Leffell and Ms Houp), and Urology (Drs Kavoussi
and Jarrett), Johns Hopkins University School of Medicine, Baltimore, Md;
and Department of Surgery, Columbia University, New York, NY (Dr Ratner).
Dr Cooper is now with the Department of Surgery, University of Maryland, Baltmore.
Dr Kozlowski is now with the Department of Surgery, University of North Carolina,
Chapel Hill. Dr Phillips is now with the Department of Surgery, University
of Virginia, Charlottesville. Mr Desai is now with the College of Medicine,
University of Illinois, Chicago.
Context First proposed 2 decades ago, live kidney paired donation (KPD) was
considered a promising new approach to addressing the shortage of organs for
transplantation. Ethical, administrative, and logistical barriers initially
proved formidable and prevented the implementation of KPD programs in the
Objective To determine the feasibility and effectiveness of KPD for the management
of patients with incompatible donors.
Design, Setting, and Patients Prospective series of paired donations matched and transplanted from
a pool of blood type or crossmatch incompatible donors and recipients with
end-stage renal disease (6 conventional and 4 unconventional KPD transplants)
at a US tertiary referral center (between June 2001 and November 2004) with
expertise in performing transplants in patients with high immunologic risk.
Intervention Kidney paired donation and live donor renal transplantation.
Main Outcome Measures Patient survival, graft survival, serum creatinine levels, rejection
Results A total of 22 patients received transplants through 10 paired donations
including 2 triple exchanges at Johns Hopkins Hospital. At a median follow-up
of 13 months (range, 1-42 months), the patient survival rate was 100% and
the graft survival rate was 95.5%. Twenty-one of the 22 patients have functioning
grafts with a median 6-month serum creatinine level of 1.2 mg/dL (range, 0.8-1.8
mg/dL) (106.1 μmol/L [range, 70.7-159.1 μmol/L]). There were no instances
of antibody-mediated rejection despite the inclusion of 5 patients who were
highly sensitized to HLA antigens due to previous exposure to foreign tissue.
Four patients developed acute cellular rejection (18%).
Conclusions This series of patients who received transplants from a single-center
KPD pool provides evidence that recipients with incompatible live donors,
even those with rare blood type combinations or high degrees of HLA antigen
sensitization, can receive transplants through KPD with graft survival rates
that appear to be equivalent to directed, compatible live donor transplants.
If these results can be generalized, broader availability of KPD to the estimated
6000 patients with incompatible donors could result in a large expansion of
the donor pool.
The number of patients waiting for a kidney transplant continues to
grow at an alarming pace1 and any significant
gains in closing the gap between organ supply and demand are likely to come
from the increased use of live donors. The 2 most significant barriers to
greater use of live donors are blood type incompatibility and HLA antigen
sensitization. Based on blood group frequencies in the United States, there
is a 36% probability that any 2 individuals will be blood type incompatible,
eliminating up to one third of the potential live donor pool.2,3 In
about 30% of the patients on the deceased donor waiting list, HLA antigen
sensitization is present due to exposure to foreign tissue in the form of
previous transplants, pregnancies, or blood transfusions. Approximately 7000
of these patients have a wide breadth of response to common HLA antigens as
measured by a panel reactive antibody (PRA) assay and are described as being
highly sensitized (PRA >80%).1 Patients who
are highly sensitized are likely to have a positive crossmatch with any given
donor, which would indicate that they harbor cytotoxic antibodies against
the donor that can result in an immediate, irreversible hyperacute or acute
antibody-mediated rejection (AMR).4- 6
Successful protocols for enabling incompatible transplants by removing
or neutralizing blood group or HLA-specific antibodies with plasmapheresis
and intravenous immunoglobulin (desensitization) are being performed at several
specialized centers but these procedures are expensive, labor intensive, and
have a variable response rate.7- 12 An
alternative strategy is kidney paired donation (KPD) transplantation. In KPD
transplants, incompatible donor/recipient pairs exchange kidneys so that each
recipient receives a compatible organ.
In this study, we present the results of our single-institution experience
with 22 patients involved in 10 live KPD transplants. The KPD transplant represents
a cost savings compared with desensitization, which in its own right is significantly
less costly than if an individual continues to undergo dialysis.13 While
logistically challenging, a broader implementation of KPD on a regional or
national scale could provide compatible organs for a substantial number of
the estimated 6000 patients on the waiting list who currently have incompatible
The KPD protocol was approved by the ethics committee and legal office
at Johns Hopkins University, Baltimore, Md. All operations for each KPD transplant
were performed simultaneously to reduce the possibility that 1 operation would
need to be aborted while the others were completed. All participants agreed
to the uncertainties inherent in a kidney donor exchange and to remain anonymous
to each other until after the operation. Johns Hopkins Hospital is a referral
center for patients with blood type and HLA antigen incompatibilities. All
patients were given the option of entering the KPD pool. However, some of
the patients with barriers that were more amenable to desensitization received
transplants successfully under a regimen of plasmapheresis and intravenous
immunoglobulin. Thus, patients who were more difficult to match because of
their blood types or broad HLA antigen reactivity were overrepresented in
our KPD pool, reducing the overall number of possible matches. Race/ethnicity
of the patients was recorded by the attending physician or nurse.
Twenty recipients received 0.1 mg/kg of tacrolimus daily (Prograf, Fujisawa
Healthcare Inc, Deerfield, Ill), 2 g of mycophenolate mofetil administered
in twice daily divided doses (Cellcept, Hoffmann-La Roche Inc, Nutley, NJ),
and 500 mg of methylprednisolone intraoperatively and then 125 mg every 6
hours for 6 doses, followed by 20 mg of prednisone daily beginning on the
day of transplantation. In addition, 8 patients who were considered to be
at higher risk for rejection (received previous transplants and/or patients
with high PRAs) also received induction therapy with 2 mg/kg of daclizumab
prior to reperfusion and then 1 mg/kg every other week for 5 total doses (Zenapax,
Hoffman-La Roche Inc). The target serum levels for tacrolimus were 8 to 10
ng/dL. The prednisone was rapidly tapered so that by 6 months most patients
were taking 5 mg/d. Unconventional KPD patients 2 and 9 were given sirolimus
daily (Rapamune, Wyeth Pharmaceuticals, Madison, NJ) due to complications
associated with tacrolimus during previous transplants.
One patient (unconventional KPD patient 8) received plasmapheresis using
a COBE Spectra (Gambro BCT, Lakewood, Colo) as previously described.9 This was a fourth transplant for this patient and
2 of his previous grafts were lost in the first week after transplantation
due to severe AMR. A splenectomy was performed 11 days prior to transplantation
and a single dose of anti-CD20 monoclonal antibody (375 mg/m2 of
rituximab [Rituxan], Genentech Inc, San Francisco, Calif) was administered
1 day prior to transplantation.
Recipients underwent percutaneous renal transplant biopsy for clinical
suspicion of acute rejection based on a decline in renal function. Standard
Banff criteria17,18 for acute
cellular and acute AMR were used for diagnosis. Patients with acute cellular
rejection were treated with 100 mg/d of dexamethasone for 3 days and then
a steroid taper or 1.5 mg/kg per day of antithymocyte globulin for 7 days
(Genzyme, Cambridge, Mass). There were no cases of AMR.
Isoagglutinin titers were determined by doubling dilutions of serum
using standard serological techniques.10 Crossmatch
techniques, including anti–human globulin-enhanced lymphocytotoxicity
crossmatch with T cells, one wash for lymphocytotoxicity with B cells, and
flow cytometry with T and B cells were performed as previously described.19 When present, anti-HLA antigen class I and class
II donor-specific antibody were identified by enzyme-linked immunosorbent
assay using soluble HLA antigens as targets (GTI Quik-ID and GTI Quik-ID Class
II; GTI Diagnostics, Waukesha, Wis). All titers for donor-specific antibody
represent IgG antibodies.
Probability calculations of highly sensitized patients receiving a kidney
from the deceased donor pool were performed using a previously published algorithm.20 The probability of finding an acceptable donor was
calculated as the frequency of donors with an acceptable blood type (column
2, Table 1), multiplied by the frequency
of donors with an acceptable maternal haplotype, multiplied by the frequency
of donors with an acceptable paternal haplotype. To calculate the frequency
of an acceptable maternal or paternal haplotype (column 5, Table 1), the summed allele frequencies of each unacceptable antigen
were subtracted from 1 (column 3, Table 1).
However, during this process, frequencies of haplotypes bearing 2 or more
unacceptable antigens (column 4, Table 1)
would be subtracted twice because an individual may carry unacceptable antigens
encoded by more than 1 locus. Therefore, these frequencies were added back
to the equation (column 5, Table 1).
Between June 2001 and November 2004, 6 conventional and 4 unconventional
KPD transplants were performed (Figure 1).
The term conventional is applied to KPD transplants
in which a blood type A and B donor/recipient is matched to a pair with the
opposite incompatibility. In an unconventional KPD,
recipients with blood type O can participate and derive mutual benefit, overcoming
incompatibilities of blood type and positive crossmatch or positive crossmatch
Eight of the KPD transplants, including all 6 conventional and 2 unconventional,
involved 2 donor/recipient pairs, while 2 unconventional KPD transplants included
3 donor/recipient pairs each. The KPD transplants were performed to avoid
blood group incompatibility (14 patients), eliminate a positive crossmatch
(4 patients), improve HLA antigen matching (1 patient who participated in
KPD primarily for altruistic reasons), avoid HLA antigens shared with a previous
transplant recipient (2 patients), or reduce the amount of donor-specific
anti-HLA antigen antibody (estimated by the strength of crossmatch reactivity)
to a level that could be easily removed by plasmapheresis (1 patient). In
the cases of 4 patients with a positive crossmatch, extremely high titers
of donor-specific antibody (dilution >1:1024) rendered desensitization using
plasmapheresis unfeasible based on our experience (R.A.M. et al, unpublished
data, 2005). Of the 4 patients, 3 were highly sensitized (PRA >80%).
Twenty-one of 22 recipients ultimately received a blood type compatible,
negative-flow cytometric crossmatch organ transplant. One patient (unconventional
KPD patient 8, PRA = 100%), whose cytotoxic crossmatch with the
intended donor had a titer of greater than 1024, had a much lower level of
anti-HLA antigen antibody (titer = 4) with his exchange donor but
required some pretransplant desensitization treatments. Patient characteristics
for all KPD exchanges appear in Table 2.
The median number of HLA antigen mismatches was the same between the intended
donor and the paired donor. However, 5 of the 22 patients had PRAs higher
than 80% and were matched on the basis of avoiding unacceptable antigens.
Outcomes and renal function are summarized in Table 3. One graft (conventional KPD patient 1) was lost on the
night of surgery due to renal vein thrombosis after the kidney had sustained
an injury to the hilar portion of the vein during the laparoscopic donor nephrectomy.
Four patients developed acute cellular rejection (18%) and all have responded
to a steroid pulse or antithymocyte globulin treatment. At a median follow-up
of 13 months (range, 1-42 months), the patient survival rate was 100%. Twenty-one
of the 22 patients currently have functioning grafts (95.5%) with a median
6-month serum creatinine level of 1.2 mg/dL (range, 0.8-1.8 mg/dL) (106.1 μmol/L
[range, 70.7-159.1 μmol/L]). Conventional KPD patient 9 (PRA = 73%)
was the only recipient of a conventional KPD transplant that had a PRA of
higher than 10%. However, the median peak PRA of the unconventional KPD transplant
cohort was 54.5% (range, 0%-100%) and 5 of the 10 patients had a PRA of higher
than 80%. Despite constituting a higher immunologic risk group, recipients
of an unconventional KPD transplant continue to display excellent graft function
with a median 6-month serum creatinine level of 1.4 mg/dL (range, 0.8-1.8
mg/dL) (123.8 μmol/L [range, 70.7-159.1 μmol/L]). There were no episodes
of hyperacute or acute AMR.
The complexity and potential benefits of an unconventional KPD transplant
are demonstrated by the triple exchange illustrated in Figure 2. All 3 patients in the exchange had high levels of donor-specific
antibody (titer >1024) on a cytotoxic crossmatch with their intended donor.
Furthermore, based on blood type and HLA antigen antibody reactivity, we calculated
the probability of finding a suitable donor in the deceased donor pool was
0.029 for unconventional KPD patient 8, 0.033 for unconventional KPD patient
9, and 0.008 for unconventional KPD patient 10 (Table 1).
Unconventional KPD patient 8 was a 42-year-old white man with blood
type A and end-stage renal disease as a result of diabetes mellitus and hypertensive
nephrosclerosis. He had received 3 previous kidney transplants. His first
allograft lasted 3 years. The second and third grafts were lost during the
first week after transplant due to severe AMR. He had spent a total of 14
years undergoing dialysis. His stepsister, who is blood type O, agreed to
serve as the donor for a fourth transplant but a positive cytotoxic crossmatch
with a titer greater than 1024 was identified. He had a PRA of 100% with strong
reactivity against common HLA antigens A1, A2, and A11. His intended donor
was mismatched at HLA antigens A1 and A11.Given the strength of the crossmatch,
the patient was deemed unsuitable for desensitization and was offered the
opportunity to participate in the KPD transplant program. Although our database
of potential exchange participants currently includes 86 donors and 71 recipients,
we could not identify any combination of donors and recipients that would
provide this patient with a negative crossmatch. However, he had a low titer
(dilution = 1:4) positive cytotoxic crossmatch against the intended
donor of unconventional KPD patient 10 (HLA antigen A3), a reactivity strength
amenable to desensitization. Five pretransplant and 9 posttransplant plasmapheresis
treatments were performed. He received both splenectomy and anti-CD20 prior
to transplantation.21 The patient eliminated
his donor-specific antibody and at 9 months after transplantation had a creatinine
level of 1.1 mg/dL (97.2 μmol/L).
Unconventional KPD patient 9 was a 31-year-old white woman with blood
type O and end-stage renal disease secondary to type 1 diabetes mellitus.
She continued to undergo dialysis for 2 years. She had a PRA of 98% and had
a high titer positive crossmatch (dilution >1:1024) with her father, the intended
donor. Among others, the patient had antibodies to HLA antigen A2 and Bw4.
She presented with 9 potential donors, all of whom had HLA antigens A2 or
Bw4. However, the intended donor of unconventional KPD patient 8 carried neither
HLA antigens A2 nor Bw4. The final flow crossmatch was negative and unconventional
KPD patient 9 underwent an unremarkable transplant. She had 1A cellular rejection
and was successfully treated with pulse steroids. At 9 months after transplantation,
she did not have additional episodes of rejection and had a stable serum creatinine
level of 1.3 mg/dL (114.9 μmol/L).
Unconventional KPD patient 10 was a 34-year-old white man with blood
type O and end-stage renal disease secondary to IgA nephropathy. He had a
previous deceased donor renal transplant in 1985 that was lost to chronic
rejection. He was highly sensitized due to his previous transplant and multiple
blood transfusions. His PRA was 82% and he had continued to undergo dialysis
for 4 years. He had high titer antibodies against HLA antigens DRw51, w53,
DQ1, and DQ3 and a positive cytotoxic crossmatch (dilution >1:1024) with his
wife, the intended donor, was identified. However, he was identical to the
father of unconventional KPD patient 9 at the HLA class II loci and had a
negative flow crossmatch with this donor. Following an unremarkable transplant,
the patient’s serum creatinine level increased on day 4 and a biopsy
revealed a 2B cellular rejection for which he received a course of anti–human
thymocyte globulin. He did not have any additional episodes of rejection and
at 9 months after transplantation had a stable serum creatinine level of 1.7
mg/dL (150.3 μmol/L).
The results of 6 conventional (blood groups A and B incompatible donor/recipient
pairs) and 4 unconventional (2 double and 2 triple exchanges) KPD transplants
performed at a single center were presented herein. The patient and graft
survival in this cohort was 100% and 95.5%, respectively, with a median follow-up
of 13 months. This compares favorably with the 2001 United Network for Organ
Sharing live donor 1-year adjusted patient and graft survival of 98.3% and
94.3%.1 For unconventional KPD transplants
in which the average PRA was higher than 50%, our graft survival was 100%
compared with the US Organ Procurement and Transplantation Network and the
Scientific Registry of Transplant Recipients 2004 rate of 92% for patients
with similar levels of sensitization. There was an 18% acute cellular rejection
rate and no episodes of AMR. This compares favorably with the 30% acute cellular
rejection rate reported by Fuller et al22 for
living unrelated transplant recipients. Seven (32%) of the 22 recipients had
prior transplants. All patients with positive crossmatches with their intended
donors had high titer donor-specific antibodies making them ineligible for
desensitization by our current acceptance criteria. Through the KPD transplant
program, 4 of 5 highly sensitized recipients were successfully paired with
donors for whom they had no reactivity on a flow crossmatch and required no
preconditioning. The remaining patient was matched with a donor who did not
have HLA molecules for which he showed strong reactivity but persisted with
a low–titer-positive crossmatch amenable to desensitization.
The concept of the KPD transplant was first described by Rapaport in
1986.23 Ross et al24 provided
the ethical construct for KPD transplant in 1997. Further refinements of the
ethics of various types of exchanges were introduced in subsequent publications
by this group.25- 27 The
ethical concerns abated and barriers to implementation were administrative
and logistical. Single centers did not have enough incompatible pairs to provide
a large enough pool to generate a significant number of matches and it became
clear that regional or national systems of listing and matching would be necessary.
The challenges inherent in organizing complex cooperative programs between
transplant centers (eg, should donors travel or kidneys be shipped) have dominated
the landscape and as a result only 53 patients have received a transplant
through KPD in the United States to date.1
Two types of paired donation have been performed by our group and others.26 In the conventional live KPD transplant, individuals
with blood types A and B are matched with a pair who have the opposite incompatibility.
Unfortunately, this is the rarest blood type combination and only affects
about 3% to 5% of the live donor/recipient pairs.2
The unconventional live KPD transplant, however, allows blood type O
donors and recipients to benefit from paired donation. This significantly
increases the impact of the KPD transplant on the live donor pool. In this
type of exchange, a blood type O recipient who has a positive crossmatch or
incompatible blood type with his/her intended donor is matched with a blood
type O donor of another crossmatch incompatible pair. More than 2 donor/recipient
pairs can participate in this exchange.28 This
type of KPD transplant solves the problem of the excess of blood type O recipients
with incompatiable blood type donors. It also allows patients who are broadly
sensitized to common antigens to have the opportunity to receive a kidney
with a negative crossmatch.
Immunologic risk varies depending on the donor/recipient profile (repeat
HLA antigen mismatches from previous transplants or strength of crossmatch
reactivity) between a sensitized patient and the intended donor.21 Likewise,
factors such as donor blood group (A1 vs A2 or B) and recipient blood group
antibody titers define immunologic risk for a patient undergoing desensitization
for an incompatible blood type transplant. When a patient presents with an
incompatible donor, the risk of AMR and graft loss can be estimated. In some
cases, this risk is so high that the patient would benefit from receiving
a kidney from another donor with more favorable HLA antigens or blood type.
By eliminating the requirement for a negative crossmatch or blood type compatibility
for all participants in a KPD transplant, patients could be matched with a
paired donor against whom they have lower immunologic risk and undergo pretransplant
desensitization. We are not aware of any other instances other than the one
presented herein in which the KPD transplant has been performed to facilitate
desensitization. We think this could have a dramatic impact on the field of
desensitization, yielding better results and lower cost of therapy.
This study demonstrates that KPD transplants can be performed with outcomes
similar to compatible living donor kidney transplants. The cost savings and
decrease in waiting time that could be realized by a wider application of
this concept are substantial.14 Because the
likelihood of finding a suitable match is dependent on the size of the pool,
a national list could enable many more transplants.13,29 We
estimate that about half of the incompatible pairs could receive transplants
using a national KPD transplant scheme with blood type compatible, negative
crossmatch kidneys, including as many as 14% of the highly sensitized patients.13 Patients unable to be matched by KPD could undergo
desensitization with their intended donor. Those who were not deemed acceptable
for desensitization due to high titer or immunologic risk could participate
in a less restricted KPD search in which a more favorable, but not completely
compatible, donor could be identified. This single-center experience demonstrates
that KPD is feasible, successful, and if applied to larger donor pools, capable
of expanding access to renal transplantation.
Corresponding Author: Robert A. Montgomery,
MD, DPhil, Department of Surgery, Johns Hopkins University School of Medicine,
720 Rutland Ave, Ross 765, Baltimore, MD 21205 (email@example.com).
Author Contributions: Dr Montgomery had full
access to all of the data in the study and takes responsibility for the integrity
of the data and the accuracy of the data analysis.
Study concept and design: Montgomery, Ratner.
Acquisition of data: Montgomery, Zachary, Hiller,
Houp, Cooper, Kavoussi, Jarrett, Burdick, Maley, Melancon, Kozlowski, Leffell,
Analysis and interpretation of data: Montgomery,
Zachary, Segev, Houp, Simpkins, Phillips, Desai, Collins, Reeb, Kraus, Rabb,
Drafting of the manuscript: Montgomery, Segev,
Phillips, Desai, Collins, Warren.
Critical revision of the manuscript for important
intellectual content: Montgomery, Zachary, Ratner, Segev, Hiller, Houp,
Cooper, Kavoussi, Jarrett, Burdick, Maley, Melancon, Kozlowski, Simpkins,
Reeb, Kraus, Rabb, Leffell.
Statistical analysis: Montgomery, Zachary,
Segev, Simpkins, Leffell.
Obtained funding: Montgomery, Ratner.
Administrative, technical, or material support:
Montgomery, Ratner, Segev, Hiller, Houp, Cooper, Kavoussi, Jarrett, Burdick,
Maley, Melancon, Kozlowski, Desai, Collins, Reeb, Kraus, Rabb, Warren.
Study supervision: Montgomery, Ratner, Warren.
Financial Disclosures: None reported.