Secondary Logo

Journal Logo

Organ Procurement and Transplantation Network/United Network for Organ Sharing Histocompatibility Committee Collaborative Study to Evaluate Prediction of Crossmatch Results in Highly Sensitized Patients

Nikaein, Afzal1,2,18; Cherikh, Wida3; Nelson, Karen1,4; Baker, Timothy3; Leffell, Sue1,5; Bow, Laurine6; Crowe, Debbie7; Connick, Ketra8; Head, Mary Ann9; Kamoun, Malek10; Kimball, Pam11; Klohe, Ellen12; Noreen, Harriet1,13; Rebellato, Lorita14; Sell, Tom15; Sullivan, Karen16; Land, Geoffrey1,17

doi: 10.1097/TP.0b013e3181943c76
Clinical and Translational Research

Background. The requirement for a prospective crossmatch limits some organ allocation to local areas. The delay necessitated by the crossmatch restricts the distance across which offers can be made without unduly increasing the ischemia time. A collaborative study involving 14 transplant centers was undertaken by the Organ Procurement and Transplantation Network/United Network for Organ Sharing (OPTN/UNOS) Histocompatibility Committee to evaluate the accuracy with which the detection of unacceptable human leukocyte antigen (HLA) antigens by most advanced solid phase immunoassays can predict crossmatch results. In addition, using actual patients’ unacceptable HLA antigens, the number of compatible donors that would have been available from the OPTN deceased kidney donors during 2002 to 2004 were investigated.

Methods. Panel reactive antibodies were performed by conventional or solid phase assays, and crossmatches were performed by cytotoxicity or flow cytometry. Analyses were stratified for T and B cell and by method of identifying unacceptable HLA antigens and crossmatch techniques.

Results. Combination of solid phase immunoassays and flow cytometry crossmatches resulted in a higher prediction rates of positive T cell (86.1%–93.5%) and B-cell crossmatches (91%–97.8%). Prediction of negative crossmatches based on different combination of panel reactive antibodies and crossmatch techniques varied from 14.3% to 57.1%. Furthermore, numerous potential compatible donors were identified for each patient, regardless of their ethnicity, in the OPTN database, when predicted incompatible ones were excluded.

Conclusions. The above results showed that with the advent of solid phase immunoassays, HLA antibodies can now be accurately detected resulting in prediction of crossmatch outcome. This should facilitate organ allocation and prevents shipment of organs to distant incompatible recipients.

1 OPTN/UNOS Histocompatibility Committee, Richmond, VA.

2 Texas Medical Specialty, Inc. Dallas, TX.

3 Research Dept/UNOS, Richmond, VA.

4 Puget Sound Blood Center, Seattle, WA.

5 John Hopkins Medical Center, Baltimore, MD.

6 Hartford Transplant Center, Hartford, CT.

7 DCI Laboratory, Nashville, TN.

8 University of Utah School of Medicine, Salt Lake City, UT.

9 Oregon Health Sciences Center, Portland, OR.

10 University of Pennsylvania, School of Medicine, Philadelphia, PA.

11 Medical College of Virginia, Richmond, VA.

12 Inland Northwest Blood Center, Spokane, WA.

13 University of Minnesota, Fairview-University Medical Center, Minneapolis, MN.

14 ECU School of Medicine, Greenville, NC.

15 MeritCare Transplant Services, Fargo, ND.

16 Tulane University School of Medicine, New Orleans, LA.

17 The Methodist Hospital/Weill-Cornell Medical College, Houston, TX.

This work was supported wholly or in part by HRSA contract 234-2005-370011C.

The content is the responsibility of the authors alone and does not necessarily reflect the views or policies of the Department of HHS, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.

18 Address correspondence to: Afzal Nikaein, Ph.D., Texas Medical Specialty, 7777 Forest Lane, Building C, Suite 768, Dallas, TX 75230.


Received 24 June 2008. Revision requested 9 July 2008.

Accepted 26 September 2008.

Human leukocyte antigens (HLA) antibodies have been a major obstacle for transplanting highly sensitized patients (1–6). It has been hypothesized that identification of antibodies present in highly reactive sera and the use of this information in donor selection may increase the frequency of crossmatch negative donors for these patients (1, 2, 7–9). With the advent of innovative solid phase immunoassays using purified HLA molecules as targets, we have been able to better characterize the HLA antibodies in sera of the highly sensitized transplant candidates and identify their respective unacceptable HLA antigens. This information could be used to predict the crossmatch outcome (1, 2). To test this hypothesis on a large scale, the Organ Procurement and Transplantation Network/United Network for Organ Sharing (OPTN/UNOS) Histocompatibility Committee designed a collaborative study to evaluate the accuracy with which the identification of unacceptable HLA antigens could predict mainly, positive crossmatch results. This would lead to exclusion of potential positive crossmatch patients and avoiding the offers or shipment of organs to incompatible patients. Prediction of negative crossmatches was also investigated to evaluate the accuracy of prediction when unacceptable HLA antigens were identified by any of the solid phase immunoassays. In addition, based on the listed unacceptable HLA antigens, we determined for the first time, the number of potential compatible deceased kidney donors that would have been available from the OPTN data registry during 2002 to 2004, for the highly sensitized patients in this study.

Back to Top | Article Outline


Fourteen centers participated in initially submitting HLA types, panel reactive antibodies (PRA) values, and unacceptable HLA antigens for their highly sensitized patients, that is, those with PRA more than or equal to 80%. PRA was performed by complement mediated cytotoxicity with the addition of anti-human globulin (AHG) or by solid phase immunoassays, which included enzyme linked immunoassay (ELISA) (10) or single or multiple antigen coated beads/particles using cytometric analysis (flow cytometry, or Luminex) (11). Characterized sera were crossmatched with 10 to 50 random actual or surrogate donors. T and B lymphocyte crossmatches were then performed by flow cytometry or by cytotoxicity (National Institutes of Health [NIH] one wash, extended incubation or addition of AHG) and results were submitted to UNOS.

The study included data for 6620 crossmatches from 225 patients. Using patient unacceptable antigens and donor HLA-typings submitted by the participating centers, predicted crossmatches (whether positive or negative) were determined and then compared with the actual crossmatch results specified by the centers. Accuracy of prediction of crossmatch results (i.e., percent of correct positive or correct negative predictions) was determined by comparing the expected crossmatch results based on unacceptable antigens, with the actual results of crossmatch against living, deceased or surrogate donors submitted by the centers.

Analysis was done separately for T- and B-cell crossmatches and stratified by method of identifying unacceptable HLA antigens and by crossmatch technique. Study design was to evaluate how various laboratories perform in HLA antibody identification using their own methologies. Therefore, no standard protocol was required. As a result, there were intra and interlaboratory variations in PRA assessment. Some laboratories used a different method for the current serum when compared with peak; some used more than one technique to identify antibody specificities. Data analysis was based on the most sensitive technique used in antibody identification and crossmatching. In addition, all single antigen bead results were grouped into one cohort and all multiple antigen bead results were grouped into another, regardless of the method used for cytometric analysis (flow cytometer or Luminex). Chi-square test was used to compare the accuracy of prediction across methods for identifying unacceptable HLA antigens or crossmatch techniques.

In addition, we investigated whether there were any donors during 2002 to 2004, which would have been potentially compatible with these patients using the OPTN deceased kidney donor database. The potential compatible donors were selected, based on patient unacceptable antigens and donor HLA, stratified by ethnicity, PRA value, and PRA technique.

Back to Top | Article Outline


Results showed that accuracy of prediction of positive crossmatch was technique dependant. Prediction of positive T-cell crossmatch by flow cytometry was 86% and 93.5% (P<0.001) when unacceptable HLA antigens were detected by ELISA or single or multiple antigen beads, respectively (Table 1). Prediction of positive B-cell crossmatch by flow cytometry, including both class I and class II HLA antibodies was 91% and 97.8% when the unacceptable HLA antigens were identified by ELISA or single or multiple antigen beads, respectively. The two-way comparison was statistically significant (P<0.001). The combination of solid phase immunoassays for antibody detection and flow cytometry crossmatch was superior to the conventional cytotoxicity assays and resulted in a higher rate of positive crossmatch prediction. The overall prediction of positive T-cell crossmatch by cytotoxicity was 64.1% when PRAs were performed by cytotoxicity and 77.3% when PRA was performed by solid phase immunoassays. In contrast, the overall prediction of positive T-cell crossmatch by flow cytometry was 92.2% when PRAs were detected by solid phase immunoassays (Table 2). Similarly, the overall prediction of positive B-cell crossmatch by flow cytometry was much higher (96.6%) than cytotoxicity crossmatch (69.5%) when unacceptable HLA antigens were detected by solid phase immunoassays and the difference was statistically significant. No data were submitted for PRA values for HLA class II specific antibodies by cytotoxicity and this needs to be addressed.





Prediction of negative T cell flow cytometric crossmatches was not as accurate and was only 56.5% when unacceptable antigens were detected by ELISA, 50.7% when unacceptable antigens were detected by single or multiple antigen beads (Table 1). Prediction of negative B-cell crossmatches by flow cytometry was also low, 56.5% when antibodies were identified by ELISA and 27.8% when single or multiple antigen beads were used to detect unacceptable HLA antigens. The main reason for such a low prediction of negative crossmatches is that the centers did not report antibodies to HLA-C, DP or rare HLA antigens, which would have required additional testing. In addition, the study was not standardized in respect to the methodology and the sera which were used.

Prediction of negative T-cell crossmatch when combining all cytotoxic methods in one cohort and solid phase immunoassays in another is shown in Table 2. No flow cytometry crossmatches were performed when unacceptable antigens were detected by cytotoxicity and this needs to be addressed. Prediction of negative T-cell crossmatch by cytotoxicity alone was 66.7%. In contrast, prediction of negative T-cell crossmatch by flow cytometry was 50.9% and by cytotoxicity was 60.7% when unacceptable antigens were detected by solid phase immunoassays. Prediction of negative B-cell crossmatch by flow cytometry was 32% and by cytotoxicity was 57.7% when unacceptable HLA antigens were identified by solid phase immunoassays. The comparison difference was statistically significant. None of the laboratories reported negative B-cell crossmatches by either of the methods when unacceptable antigens were detected by cytotoxicity.

Using the OPTN deceased donor database during 2002 to 2004, the median numbers of potential compatible donors that would have been available for the highly sensitized patients within a given time period was calculated, stratified by patient’s ethnicity, PRA values (80%–89%, 90%–99%, and 100%), and PRA technique. Twelve patients whom their ethnicity were reported as “others” were excluded. The median number of potential compatible donors was lower when unacceptable HLA antigens were determined by single antigen beads when compared with ELISA, multiple antigen beads, or cytotoxicity-AHG (Figs. 1–4). This reflected a higher number of identified unacceptable HLA antigens for patients tested with the most sensitive method with a corresponding elimination of more potential donors in the database. Importantly, the ethnicity of the recipients did not affect the number of potential donors.









Back to Top | Article Outline


Sensitization is a barrier to transplantation, increases ischemic time and reduces graft survival (12–14). When a deceased donor is available, preliminary crossmatch is performed to exclude positive crossmatch patients. However, these preliminary crossmatches often are performed by an insensitive cytotoxicity method. In addition, the current system allows only initial crossmatching for T lymphocytes detecting class I antibodies against deceased donors at the local level. Shipped organs are not screened by preliminary crossmatch and therefore, many, with the exception of zero mismatched donors, are not transplanted into the intended sensitized patients.

One step which can be taken to fully maximize the chances of highly sensitized patients receiving a graft is to thoroughly define the specific antibody profile of these individuals. The purpose of this study was to determine whether the detection of antibody specificities can be more accurately determined by solid phase immunoassays and whether identification of unacceptable antigens would lead to the accurate prediction of crossmatch results in highly sensitized patients.

In this collaborative study, PRAs were defined by conventional cytotoxicity-AHG or solid phase immunoassays (ELISA, or, bead targets using the Luminex or flow cytometry instruments). Crossmatches were performed by cytotoxicity (NIH, one wash, or extended incubation, AHG) or flow cytometry methods. To investigate how centers performed by using different methodologies and with their own established protocols, no standardization was required for this study. Therefore, some laboratories used current and some peak sera. Crossmatches were also performed with living, deceased, or HLA typed volunteers as surrogate donors. Overall results showed a high rate of prediction of a positive crossmatch (>92%) when antibody specificities were determined by any of the solid phase immunoassays. This reflects the sensitivity and reproducibility of solid phase immunoassays for the detection of HLA antibody specificities. On the other hand, the low rate of prediction of a positive crossmatch resulting when PRAs were performed by cytotoxicity suggests that donor-specific antibodies that were either not detected or the specificity could not be defined by cytotoxicity methods.

The overall lower predictability of negative crossmatches could be due to several factors, including the lack of standard protocol, the lack of detection of HLA-C, DP or rare HLA antigens and antibodies, and by the lack of definition of non-HLA-specific antibodies that may have contributed to unexpected positive crossmatch results. To obtain higher predictive negative results, more extensive testing is needed as the number of negative crossmatches included in this study was small compared to the number of positive ones. Several individual centers have already reported the prediction of more than 90% negative crossmatch results with carefully analyzing their unacceptable antigens and using a standard internal policies, including HLA-C antibody identification (1, 2). In our study, better correlation of negative crossmatch results was obtained when the crossmatch was performed by cytotoxicity rather than by flow cytometry. Given the higher sensitivity of flow cytometry, this observation suggests that low titer antibodies or non-HLA antibodies may have contributed to the reduced correlation. Interestingly, better prediction of negative crossmatches was obtained when antibody studies were performed by ELISA platforms than with beads technologies. This is not surprising, considering that ELISA, using multiantigen platforms provide several HLA phenotypes that may bear the antigens in question, compared with beads displaying single HLA antigen. Moreover, the multi-antigen platforms may detect a combination of different, weak antibodies that could not be detected in the single format. It must also be noted that the clinical relevance of low levels of HLA-specific antibody has not been established. Recent reports indicate that antibodies that are only detectable by assays using single HLA antigens as targets do not necessarily have adverse effects on graft outcomes (15–18). More experience with these highly sensitive assays and collaborative studies is needed to resolve what antibody levels are critical.

Nonetheless, it is clear from this and other studies (1, 2, 19, 20) that positive crossmatches can be accurately predicted when solid phase immunoassays are used. The overall prediction among the highly sensitized patients in this study was high (92%–98%) when unacceptable antigens were determined by solid phase immunoassays. The most accurate results were obtained when solid phase assay results were paired with a flow cytometric crossmatch. It should be noted that this high predictability was obtained among laboratories using a diverse mixture of assays and crossmatch protocols. It is likely that increased utilization of sensitive antibody assays and consistency in the use of current or historic sera may lead to predictability of positive crossmatches approaching 100%. The inability to accurately identify class II HLA antibodies by cytotoxicity led to the lack of information in regard to flow cytometry crossmatch results when antibodies were detected by cytotoxicity.

The aim of the histocompatibility community is to facilitate organ allocation and placement through improved crossmatch prediction and to eliminate unnecessary shipment of organs to potentially incompatible recipients. This was the basis for revision of the OPTN policy to require listing of unacceptable antigens and to use the frequency of those antigens in the donor population as the definition of PRA, that is, a calculated PRA. This policy, approved by the Board of Directors in December 2006, will go into effect by end of 2008 and will require the use of at least one solid phase immunoassay for antibody detection (21). This policy is a step toward what has been termed “virtual crossmatching (1, 2, 22, 23),” or the strategy of determining incompatibility between donors and recipients without preliminary or screening crossmatches. Because negative crossmatches cannot be predicted with the same accuracy as positive crossmatches, absolute virtual crossmatching is not currently achievable. However, elimination of donors with defined unacceptable antigens will clearly facilitate organ allocation, reduce cold ischemic time, and eliminate much unnecessary organ shipment. This policy may also be applied to patients who receive desensitization treatment and can receive organs with positive crossmatch (24, 25). For these patients, the centers may only list the unacceptable antigens which the patients cannot receive.

In general, it is a widely-held concept that highly sensitized patients may only receive zero antigen mismatched organs (20). Thus, the OPTN/UNOS Histocompatibility Committee felt that it would be of interest to investigate whether, by identifying unacceptable HLA antigens with the solid phase immunoassays, there would be any potentially crossmatch compatible donors for these patients in the OPTN database. Looking at the OPTN deceased donor database for 2002 to 2004, we found a considerable number of potentially crossmatch compatible donors for these patients (Figs. 1–4). The median range was lowest when unacceptable HLA antigens were identified by single antigen beads and was highest with multiple antigen beads using the Luminex technology. This confirmed our hypothesis that more unacceptable antibodies were identified by single antigen solid phase assays and, therefore, correlated with exclusion of more donors. Interestingly, the data support the notion that the highly sensitized patients who have antibodies characterized by these newer, more specific technologies may have an excellent chance of finding compatible donors and, thus, may be no longer restricted to zero mismatched organs. Ethnicity did not influence the number of potentially compatible donors for an individual, indicating that the virtual crossmatch approach to allocation will not disadvantage the minority patients.

We must emphasize that the above analysis for the number of potentially compatible donors did not include the number of patients who would compete for each specific organ. Therefore, the actual number of donors for each highly sensitized patient may be far less than the numbers shown. However, our intention was to investigate whether there were any potentially compatible donors for each of these patients, by excluding the incompatible ones via virtual crossmatches algorithm. To obtain better information on the actual number of highly sensitized patients receiving transplants when compared with the numbers of the shipped organs, careful monitoring must be performed, should the virtual crossmatch strategy be adopted. Furthermore, a different strategy may apply for patients receiving desensitization treatment as often these patients may be successfully transplanted against positive donor crossmatches (24, 25).

In conclusion, this multicenter study demonstrates that positive crossmatches can be accurately predicted by definition of HLA-specific antibodies and supports the listing of unacceptable antigens to exclude the incompatible donors in organ allocation. Such a strategy will facilitate organ placement by reducing unnecessary shipment to crossmatch incompatible recipients and at the same time will facilitate transplantation of highly sensitized patients. The recently approved calculated PRA policy will encourage a more uniform approach to crossmatch and antibody screening techniques that may lead to future capability to achieve virtual crossmatching.

Back to Top | Article Outline


1.Bray RA, Nolen JD, Larsen C, et al. Transplanting the highly sensitized patients: The Emory algorithm. Am J Transpl 2006; 6: 2307.
2.Vaidya S, Partlow D, Susskind B, et al. Prediction of crossmatch outcome of highly sensitized patients by single and/or multiple antigen beads luminex assay. Transplantation 2006; 82: 1524.
3.Halloran PF, Wadgymar A, Ritchie S, et al. The significance of the anti-class I antibody response. I. Clinical and pathologic features of anti-class I-mediated rejection. Transplantation 1990; 49: 85.
4.Monteiro F, Buelow R, Mineiro C, et al. Identification of patients at high risk of graft loss by pre- and post-transplant monitoring of anti-HLA class I IgG antibodies by enzyme-linked immunosorbent assay. Transplantation 1997; 63: 542.
5.Claas FHJ, Witvliet MD, Duquesnoy RJ, et al. The acceptable mismatch program as fast tool for highly sensitized patients awaiting a cadaveric kidney transplantation: Short waiting time and excellent graft outcome. Transplantation 2004; 78: 190.
6.Smits JMA, Van Houwelingen HC, De Meester J, et al. Analysis of the renal transplantation waiting list: Application of a parametric competing risk method. Transplantation 1998; 66: 1146.
7.Laux G, Mytilineos J, Opelz G. Critical evaluation of the amino acid triplet-epitope matching concept in cadaver kidney transplantation. Transplantation 2004; 77: 801.
8.Doxiadis II, Duquesnoy RJ, Claas FHJ. Extending options for highly sensitized patients to receive a suitable kidney graft. Current Option in Immunol 2005; 17: 536.
9.Duquesnoy RJ, Takemoto S, de Lange P, et al. HLA matchmaker; A molecularly based algorithm for histocompatibility determination-III. Effect of matching at the HLA-A, B amino acid triplet level on kidney transplant survival. Transplantation 2003; 75: 884.
10.Zachary AA, Delaney NL, Lucas DP, et al. Characterization of HLA class I specific antibodies by ELISA using solubilized antigen targets. I. Evaluation of the GTI QuickID assay analysis of antibody patterns. Hum Immunol 2001; 62: 228.
11.Kao KJ, Scornik JC, Small SJ. Enzyme-linked immunoassay for anti-HLA antibodies: An alternative to panel studies by lymphocytotoxicity. Transplantation 1993; 55: 192.
12.Laundy GL, Bradley BA. The predictive value of epitope analysis in highly sensitized patients awaiting renal. Transplantation 1995; 59: 1207.
13.Kerman R, Orosz C, Lorber M. Clinical relevance of anti-HLA antibodies pre and post transplant. Am J Med Science 1997; 313: 275.
14.Opelz, G. Non-HLA transplantation immunity revealed by lymphocytotoxic antibodies. Lancet 2005; 365: 1570.
15.Zachary AA, Montgomery RA, Jordan SC, et al. Antibodies in transplantation: Destroying old dogma, creating new paradigms. Tissue Antigens (in press) (Report from 14th IHIWS).
16.Jindra PT, Zhang X, Mulder A, et al. Anti-HLA antibodies can induce endothelial cell survival or proliferation depending on their concentration. Transplantation 2006; 82: S33.
17.Pollinger HS, Stegall MD, Gloor JM, et al. Kidney transplantation in patients with antibodies against donor HLA class II. Am J Transplantation 2007; 7: 857.
18.Bielman D, Honger G, Lutz D, et al. Pretransplant risk assessment in renal allograft recipients using virtual crossmatching. Am J Transplantation 2007; 626.
19.Tyan DB, Li Va, Czer L, et al. Intravenous immunoglobulin suppression of HLA antibody in highly sensitized transplant candidates and transplantation with a histoincompatible organ. Transplantation 1994; 57: 553.
20.Claas FJH, Van Leewen A, Van Rood J. Hyperimmunized patients do not need to wait for an HLA identical donor. Tissue Antigens 1999; 34: 23.
21.Policy Allocation of Deceased kidneys: Sensitized wait list candidates—calculated PRA (CPRA). Policies of the OPTN UNOS. Available at: Accessed July 6, 2007.
22.Appel JZ III, Hartwig MG, Cantu E III, et al. Role of flow cytometry to define unacceptable HLA antigens in lung transplant recipients with HLA-specific antibodies. Transplantation 2006; 81: 1049.
23.Zangwill SD, Ellis TM, Zlotocha J, et al. The virtual crossmatch—A screening tool for sensitized pediatric heart transplant recipients. Pediatr Transplant 2006; 10: 38.
24.Warren DS, Zachary AA, Sonnenday CJ, et al. Successful renal transplantation across simultaneous ABO incompatible and positive crossmatch barriers. Am J Transpl 2004; 4: 561.
25.Zachary A, Montgomery R, Ratner L, et al. Specific and durable elimination of antibody to donor HLA antigens in renal transplant patients. Transplantation 2003; 76: 1519.

Prediction of crossmatch; HLA antibody identification; Solid phase immunoassays

© 2009 Lippincott Williams & Wilkins, Inc.