14th International HLA and Immunogenetics Workshop: report on the Prospective Chronic Rejection Project.

An international collaborative study of 45 transplant centers was undertaken at the 14th International HLA (human leukocyte antigen) and Immunogenetics Workshop to see if HLA antibodies detected posttransplant are predictive of chronic graft failure. With the newly developed assay, MICA (major histocompatibility complex class I-related chain A) antibodies were also measured and their effect analyzed. Total of 5219 sera from patients who were more than 6 months posttransplant with functioning graft were tested for HLA antibodies by enzyme-linked immunosorbent assay, flow cytometry, or Luminex. HLA antibodies were found in 27.2% of kidney patients, 23.6% in the liver, 52.7% in the heart, and 21.7% in the lung. The method of antibody testing did not have a marked influence on the frequency of antibodies detected. MICA antibodies were detected in 15% of kidney patients, 30% of heart patients, and 31% of liver patients. Among 948 kidney patients who had HLA antibodies, 7.3% had rejected their graft within 1 year of testing, compared with 1.7% in 2615 patients without HLA antibodies (P= 0.8 x 10(-17)). Death occurred in 1.4% of total kidney patients and did not correlate to the presence of antibodies. We conclude that patients with posttransplant HLA antibodies indeed have a higher rate of chronic graft failure and that posttransplant antibodies are predictive of chronic rejection.

14th International HLA and Immunogenetics Workshop Prospective Chronic Rejection Project: a three-year follow-up analysis.

The three-year follow-up of 4,144 patients of the 14th International Workshop Prospective Chronic Rejection study has reinforced the evidence that post-transplant HLA antibodies are predictive of long-term graft loss. Three years after a single testing for HLA antibodies, 10% of kidney recipients who were antibody-positive had lost their grafts, in contrast to only 5% of antibody-negative patients (p<0.0001). The adverse effect of post-transplant antibodies on graft survival was also observed in lung, heart, and liver transplants. Donor-specific antibodies and 'strong' non-DSA had stronger association with graft loss than 'moderate' non-DSA. Periodic antibody monitoring, combined with specificity and strength analysis, would help in the early identification of allograft recipients who are at high risk of graft failure.

Management of the sensitized cardiac recipient: the use of plasmapheresis and intravenous immunoglobulin.

Previously, we reported that the combination of plasmapheresis (PP) and intravenous immunoglobulin (IVIg) allow sensitized patients to undergo orthotopic heart transplantation (OHT), even across a positive crossmatch. In the current study, the effect of that combination, PP+IVIg, on survival of a larger group of such recipients is investigated. The latter group (I) consisted of 35 sensitized patients who received PP+IVIG together with standard immunosuppressive drugs. Rejection was seen in 11 patients, findings strongly suggestive of a vascular (humoral) being identified in five of those cases. Four deaths occurred, two of them in the immediate post-operative period, one after almost six months, and one after almost two yr post-OHT. Follow-up range 4.5 months to 7.8 yr post-OHT (average=1.1 yr). Patient survival was analyzed after generation of a Kaplan-Meier plot. Comparison with a control OHT group (II) given standard immunosuppressive drugs only (N=276) showed enhanced survival of group I (p=0.0414 by log-rank test). We conclude that the combination of PP and IVIG (i) is associated with declines in T- and B-percent-reactive antibody and in crossmatch positivity, and (ii) is very useful in the management of the sensitized cardiac patient undergoing OHT, often allowing a successful outcome to transplantation in the face of a positive crossmatch.

Risk of early renal allograft failure is increased for patients with antiphospholipid antibodies.

Renal allograft thrombosis can cause transplant failure. Because antiphospholipid antibodies (aPA) are associated with thrombosis, we investigated pretransplant sera from patients with early renal allograft failure to determine if aPA were present. Fifty-six final cross-match (FxM) sera from patients whose transplant failed within 16 days were compared to FxM sera from the next sequential transplant patients. The sera were tested for IgG, IgM, and IgA antibodies to cardiolipin, phosphatidylserine, and phosphatidylethanolamine. aPA were identified in 57% of FxM sera from patients with early non-function versus 35% of FxM sera from patients with functioning grafts (P = 0.02). Historical sera from 11 aPA-positive patients contained aPA up to 18 months prior to transplantation. Since aPA were present in historical sera, testing for aPA can identify certain patients at risk for early allograft failure. The involvement of aPA in early allograft loss is supported by studies demonstrating aPA recovery from an explanted failed transplant.

Antiphospholipid antibodies are a risk factor for early renal allograft failure.

BACKGROUND: Biopsy specimens of transplanted kidneys that fail to function reveal cellular infiltrates, infarcts, and thrombi. Because antibodies to phospholipids (aPA) and/or phospholipid-binding proteins have been associated with thrombosis, we asked whether aPA are a risk factor for early allograft failure. METHODS: Final crossmatch sera from 56 patients with primary nonfunctioning renal allografts were tested for aPA. Serum from the next consecutive patient to undergo transplantation served as transplantation controls. Both groups were compared with aPA values obtained from testing 252 control individuals. The ELISA was designed to detect IgG, IgM, and IgA antibodies to phosphatidylserine, cardiolipin, and phosphatidylethanolamine. RESULTS: Patients were evaluated based upon the aPA ELISA findings. aPA were present in 57% of the patients with early nonfunction renal allografts and 35% of the patients with functioning grafts (P=0.0234). aPA in previously hemodialyzed patients did not predict allograft failure or success (P=0.3766). In contrast, all nonhemodialysis patients who had aPA at the time of transplantation experienced early allograft failure (P=0.0022). CONCLUSIONS: These data show that aPA are an important risk factor for early renal allograft failure. Furthermore, aPA-positive patients who have no history of hemodialysis are at the greatest risk. Pretransplantation aPA screening of renal transplant candidates forewarns of early graft failure and indicates which patients may benefit from anticoagulant therapy.

The LifeLink Foundation and cadaver kidney transplantation in Tampa.

1. LifeLink Foundation, a not-for-profit organization, has been the driving force and absolutely essential entity for kidney and liver transplantation in Tampa providing all the components (patient, organs and clinicians) save for inpatient hospitalization. It also plays a big role in the heart transplant program. LifeLink has increased the kidney transplant rate from the first 1,000 done in 17 years to the second 1,000 in 7 years and is on a pace for the third 1,000 in 5 1/2 years. 2. Because of its innovative programs, cadaver donor procurement by the Tampa LifeLink OPO has been roughly double the national average for the past 10 years. Because of cadaver kidney availability the median wait time from activation on the wait list to transplantation over the past 5 years was 159 days. The recent transplant rate is 14.7-22.7% higher than the national average, dependent upon the parameter measured. Similar results are seen for Tampa patients awaiting heart and liver transplantation. 3. The overall outcome of 1,184 cadaver kidney transplants performed in the decade 1989-98 was similar to that reported from the UNOS database in this series of publications. a) One- and 2-year graft survival increased 2% per year over the decade with a recent one-year graft survival rate of 96%. The overall T1/2 was 10 years. b) Our disastrous 1994 results were quickly reversed by a more intense pretransplant medical evaluation, the introduction of mycophenolate mofetil, more aggressive and earlier treatment of rejection episodes, and mandatory T- and B-cell flow cytometry crossmatching for all transplants. The incidence of rejection episodes decreased from 40 to 20%, and the first year immunological graft loss decreased from 5%, to 1.9%, to 0.8%, to 1.4% and 0% over the succeeding 4 years. 4. Individual factors affecting allograft survival were strikingly similar to national data, although all did not react statistical significance probably due to the smaller numbers. a) Primary and second grafts had similar survival rates (p = 0.97) whereas the third or subsequent graft survival was 7-32% poorer (p = 0.02). b) Black recipients had survival rates 10-13% lower than Caucasians and other races (p = 0.003). c) Patients with a peak PRA > 50 had survival values 4-13% poorer than those with < 50 PRA (p = 0.14). d) Patients with 2-4 HLA mismatches had graft survival rates 4-10% poorer than those with 0-1 mismatch (p = 0.12), whereas those with 5-6 mismatches had rates 6-17% poorer (p = 0.04). e) Although 22% of our transplants were to patients > 60 years of age, there was no difference (p = 0.81 to 0.90) in graft survival for the age groups 0-40, 41-60 and > 61. However, the proportion of grafts lost due to patient death compared with all allografts lost, was very different at 21% in the youngest group, 43% in those 41-60 years of age, and 63% in recipients > 61 years. 5. The rate of delayed graft function with imported kidneys was higher (27 vs. 16%, p = 0.006) but essentially the same as local kidneys with the same ischemia times. However, 41% of local kidneys were transplanted within 12 hours of procurement. Totally, 78% of local kidneys were transplanted within 18 hours (11% DGF rate) versus 79% of imports being transplanted at > 18 hours (32% DGF rate). Ischemia time, not the kidney source is the key issue since: a) There was no difference in overall graft survival of imported versus local kidneys (p = 0.95) nor in comparing local versus import kidneys with (p = 0.66) or without (p = 0.69) DGF. b) There was, however, a 11-17% overall poorer graft survival over 3 years in kidneys with DGF (p < 0.001) seen with both local (9-18% poorer, p = 0.0002) and imported (12-19% poorer, p = 0.008) kidneys. c) Kidneys displaying DGF came from older donors (40 vs. 34 years, p = 0.023) and had longer ischemia times (21 vs. 15 hours, p < 0.0005). 6. Dual kidney transplants were started in late 1996 with older or marginal donors to provide a better chance of success fo

New alleles of the HLA-B15 family.

In a previous study of B locus alleles by sequence-specific oligonucleotide probe (SSOP) hybridization, we observed 18 novel patterns in a panel of 360 individuals. Four of these novel patterns were caused by alleles of the human leukocyte antigen (HLA)-B15 group, and three were available for this study. These alleles were found in Oriental, Latin American, African American, and Caucasian individuals. In addition, we analyzed a Caucasian subject who was found by serology to have an unusual B15 specificity. We sequenced these four samples by performing amplification from genomic DNA using polymerase chain reaction primers designed to obtain HLA class I products that included exon 2 and exon 3 as well as the intervening intron. The amplified segments were cloned and identified by colony hybridization with nonradioactive SSOP. Nucleotide sequences were obtained using an automated DNA sequencer. The allele B*1530 differs from B*1501 by a substitution of Asp for Asn in position 114 and Ser for Tyr in codon 116. The new allele B*1531 differs from B*1502 at amino acids 94, 95, and 152. The variant B*1524 was found to have N-77, I-80, A-81, L-82, R-83. A similar motif exists in B locus alleles that have the supertypic specificity Bw4 and in B*1513, B*1516, B*1517, and B*1523; it is likely to have been generated by gene conversion. Finally, the novel allele B*1527 is similar to B*1501 except for the presence of Phe instead of Tyr at position 99. Because this change exists also in B*1506, it is possible that B*1506 was derived from B*1501 through B*1527. It is of interest that a similar substitution (Cys for Tyr at position 99) distinguishes A*0201 from A*0207 and is known to determine an epitope recognized by T cells. Thus, B*1527 may also carry a change that is functionally relevant in cell-mediated immunity.

Flow cytometry crossmatching and primary cadaver kidney graft outcome: relevance of T and B cell targets, historic sera and autologous controls.

There is limited information regarding the role of flow cytometry crossmatching (FCXM) in primary cadaver kidney allografting and even less about B cell reactivity and graft survival (GS). Furthermore, there is little or no published data concerning reaction strength (cutoff value), the effect of historic sera reactions, and the usefulness of performing autologous crossmatches (XMs) on GS. These factors were examined retrospectively on 214 primary transplants performed from August 1991 to January 1994 with follow-up to July 1995. Three-color FCXMs were done on a 1024-channel BD-FACScan, and the shift in median channel fluorescence (MCF) over the negative control was calculated. All patients had a negative T cell (AHG) and warm B cell (2 was, extended incubation) cytotoxicity XM, and none was excluded in calculating GS. A quantitative effect was noted as stronger MCF shifts vs. T or B cells correlated with decreased GS (r = 0.98 and 0.92, respectively). Significant differences were seen with cutoff values of T = 50 and B = 110 which were 1.7-1.8 times the SD above the mean MCF of normal sera controls T neg patients (n = 198) and 1- and 3-yr actuarial GS of 86% and 79% compared to T pos patients (n = 16) of 75% and 49%, p = 0.008. B neg patients (n = 177) had 1- and 3-yr GS od 86% and 81% compared to B pos patients (n = 37) of 78% and 47%, p = 0.005. Most informative was the analysis of combined T and B cell FCXM results. Three years GS for T neg - B neg patients (n = 171) was 81% and for T pos - B neg patients (n = 6), it was 83%, p = 0.98. The 27 T neg - B pos group's GS was lower at 62% but did reach significance. Poorest GS was seen for T pos - B pos patients (n = 10) at 23%, p = 0.0001. Reaction patterns showed that T cells detected only HLA Class I antibodies, whereas B cells detected both Class I and II. Historic sera (> or = 1 month old) reactivity influenced GS. Patients with > or = 2 past sera positive but current serum negative reactions vs. T or T plus B cells (n = 7) had a poor 29% GS, while those historically positive only vs. B cells (n = 7) had 100% GS. On the other hand, patients positive only with the current serum (n = 16) had 2-yr GS of 100% (false positive test?), while patients whose current and historic sera reactions were positive (n = 21) had a 25-50% GS (true positive test?). About 1 in 15 patients (19%) displayed positive autologous FCXM reactions. Subtraction of autologous MCF shift values from those vs. the donor converted 17 patients to the T neg - B neg or T pos - B neg group whose 2-yr actual GS was not significantly different (p > 0.8) from those initially testing T neg B neg vs. their donors.

Hemoconcentration prior to serology testing in hemodiluted cadaver bone and tissue donors.

Concern about false negative serology tests for infectious diseases in hemodiluted cadaver donors resulted in issuance of new regulations and guidelines from the FDA and CDC. Bone or tissue from donors receiving > or = 4 units of blood, blood products, colloids or crystalloids within 48 h of sampling must be quarantined unless: (a) a pretransfusion serum is available; or (b) an adequate algorithm is employed to ensure hemodilution is insufficient to alter test results, i.e. cause false negatives. Left undefined in these regulations is, what is an adequate algorithm and what amount of hemodilution would cause false negatives. A pretransfusion sample was not available for about 20% of our donors and many had incomplete infusion histories. We used the unambiguous quantitation of serum albumin and total protein to define hemodilution and, if present, hemoconcentration of sera by ultrafiltration to normal protein levels prior to serology testing. Control experiments showed excellent correlation between serum dilution and protein concentration (r > 0.99) and a quantitative recovery of 96.9 +/- 1.4% upon hemoconcentration. Known positive sera (CMV-Ab; HTLV-1Ab; HIV-1,2Ab; HBsAb; HCV-Ab; HBsAg) were spiked into normal sera and diluted up to 1:2000, well beyond detectable levels. A qualitative recovery of 100% and a quantitative recovery of 97.6 +/- 7.5% of antibody or antigen reactivity was achieved upon hemoconcentration and retesting. In two studies, 14% (30/210) and 43% (23/54) of cadaver donors had serum proteins below normal limits and their sera was hemoconcentrated.(ABSTRACT TRUNCATED AT 250 WORDS)

Positive crossmatches against mandatorily shared kidneys.

From October 1987 through February 1990 approximately 8.5% (29/341) of all donor kidneys shipped under the UNOS Mandatory Sharing Policy were denied to 27 intended recipients due to a positive final crossmatch [XM(+)]. The intended recipients included 18.5% hispanics and 7.4% blacks compared to 2.4% and 1.6%, respectively, for XM(-) recipients (1-3). Further, more were highly sensitized with 81% having a current PRA greater than 10% and 56% with a peak PRA greater than 80% compared to 65% and 14%, respectively, for XM(-) recipients. More importantly, 19/27 (70%) of the recipient candidates may have had irrelevant positive XMs. The XM(+) patients were classified into five categories defined by: I) autoantibodies; II) transfusions in the 2 weeks prior to the availability of the donor; III) the XM technique; IV) highly sensitized regraft candidates with current and peak PRAs greater than 85% and V) antibody to unreported MHC antigens. Of these, 70% may have been denied a transplant due to IgM autoantibodies or the use of XM techniques lacking extensive evaluation. The authors propose that all XM(+) mandatorily-shared kidneys be examined for IgM autoantibody and that kidneys not be denied to potential recipients due to IgM autoantibody. In addition, to minimize exclusions based on positive B-cell XMs, it is proposed that mandatorily-shared kidneys be shared on the basis of the DR subtypes, insofar as is currently practical.

Hyperacute and acute kidney graft rejection due to antibodies against B cells.

Because of the perception of its uncertain clinical significance, the B cell crossmatch is not universally performed before renal transplantation. Even though sporadic cases of hyperacute rejection associated with B cell antibodies have been reported, doubts remain in light of other studies suggesting no effect on graft survival. This report describes 4 cases of graft rejection (3 hyperacute and 1 acute) that occurred in patients with anti-B-cell antibodies specific against donor HLA-DR or DQ antigens. Absence of anti-donor class I antibodies was confirmed in all cases by 2-color flow cytometry. Strong evidence for an antibody-mediated mechanism was found in one patient with anti-class I and anti-class II antibodies in serum transplanted with a class II mismatched kidney. In this case, only anti-class II antibodies were recovered in the eluate of the nephrectomy specimen. These four cases were compiled from three different institutions over a four-year period, which confirms the infrequent occurrence of these events. While anti-class II antibodies may not always be detrimental for graft survival, these results also confirm that they have the potential to cause hyperacute or acute graft loss. We conclude that the information provided by the B cell crossmatch should be available at the time that a decision to proceed with a renal transplant is made.

Relative performance of specific and nonspecific fluorescence polarization immunoassay for cyclosporine in transplant patients.

The analytical performance of specific and nonspecific fluorescence polarization immunoassays (FPIAs, Abbott Laboratories) for cyclosporine was compared. Both specific and nonspecific FPIAs demonstrated excellent between-run coefficients of variation (5.9% vs. 3.9%) at three levels of control, and a high degree of between-center reproducibility (r2 greater than 0.96). In addition, the correlation between cyclosporine levels measured by specific and nonspecific FPIAs was statistically significant, though imperfect, in both renal (r2 = 0.70) and cardiac transplant patients (r2 = 0.55). In kidney transplant patients, the nonspecific/specific ratio was significantly higher in patients with serum bilirubin concentration exceeding 3 mg/dl (5.9 +/- 2.6 vs. 2.8 +/- 1.1), due to impaired elimination of cyclosporine metabolites in the bile. The nonspecific/specific ratio was also significantly higher in heart transplant patients early (less than 1 month) posttransplant compared with patients in the late posttransplant period (3.4 +/- 0.8 vs. 2.9 +/- 0.8). The Abbott FPIA provides a highly precise method for measuring cyclosporine, with a turnaround time of 15-20 min. The specific monoclonal FPIA has the additional advantage of measuring primarily unchanged cyclosporine and thus has an imperfect correlation with the nonspecific polyclonal FPIA. Together with clinical data, the use of FPIAs may help to improve the efficiency of cyclosporine therapeutic drug monitoring.

Uses of orthoclone OKT3 for prophylaxis of rejection and induction in initial nonfunction in kidney transplantation.

We have shown that OKT3 can be safely administered to a varied group of transplant recipients, including a high percentage of diabetics and older patients. Used as a prophylactic agent, OKT3 was associated with a decreased incidence and delayed onset of acute rejection. Effectiveness may be reduced by shortening the course of treatment and/or lowering the dose of concomitant azathioprine. Use of OKT3 in patients with ATN (thereby avoiding use of CyA) may be a beneficial strategy. Stimulation of significant titers of antimurine antibody has not been a common event in this study.

OKT3 prophylaxis versus conventional drug therapy: single-center perspective, part of a multicenter trial.

As part of a collaborative study, Orthoclone OKT3 (Ortho Pharmaceutical Corporation, Raritan, NJ) monoclonal anti-T-cell antibody was used prophylactically for 14 days following cadaveric renal transplantation. Patients were randomized before treatment and compared with a control group treated with triple-drug immunosuppressant therapy consisting of cyclosporine, azathioprine, and prednisone. The OKT3 group (also treated with prednisone and azathioprine) had significantly fewer patients with acute rejection during the first year (27% v 60%), significantly delayed onset of acute rejection (median number of days to rejection, 56 v 10), and shortened duration of delayed initial graft function (median number of days on nonfunction, 5.5 v 9.0). OKT3 was safely administered intraoperatively. The study group included diabetics and patients over the age of 50 years. There was an increased incidence of benign fungal infection in patients treated with OKT3. The outcome data of both the prophylactic and control groups in this single-center trial was different from the multicenter experience in that patient and graft survival was 100% in both groups. The only difference in the immunosuppressive protocol was the use of 2 g intravenous methylprednisolone intraoperatively. The combination of intraoperative OKT3 with 2 g methylprednisolone may merit further study. It was concluded that OKT3 prophylaxis reduces the incidence of rejection episodes, delays the occurrence of rejection, and may reduce the duration of delayed initial graft function. It is not associated with an increase in serious infection and can be administered safely to a variety of cadaver transplant recipients.

Four-year experience with exclusive use of cytomegalovirus antibody (CMV-Ab)-negative donors for CMV-Ab-negative kidney recipients.

Because of unacceptable CMV related morbidity, mortality, and graft loss, we adopted a policy that (-) patients would receive renal allografts only from (-) donors. This policy has resulted in a significant decrease in (1) morbidity (10.6% v 1.7%, P less than .0001), (2) mortality (3.7%) v 0%, P = .002), and (3) graft loss (2.5% v 0%, P = .016). CMV-Ab-negative patients (1) constitute 26% of all patients tested (compared with 36% of all donors tested), (2) receive transplants at the same rate as (+) patients, (3) do not have a prolonged waiting time, and (4) received greater HLA-A, B, -DR mismatched kidneys (3.6 v 3.1, P less than .01).

Organ sharing for good HLA-A,B, and DR matching improves cadaver renal graft survival in SEOPF: retrospective and prospective studies considering delayed graft function, race, center effects, cyclosporine, and other factors.

1. HLA matching is associated significantly with factors including CsA use, ALS use, recipient race, prior graft loss, presensitization, preservation time and most strongly, with organ sharing. However, HLA match is not directly associated with delayed graft function. 2. By univariate and multivariate analyses, good HLA matching provides significant benefits in graft survival regardless of CsA use, organ source or other potentially confounding factors. 3. HLA-A,B, and DR matching have independent and essentially equivalent benefits on graft survival in CsA-treated patients, whereas HLA-A,B matching has a greater benefit in non-CsA-treated patients. 4. Organ sharing, per se, provides no direct detrimental effect on graft survival by univariate or multivariate analysis. 5. By multivariate and univariate analyses, shared/well-matched kidneys provide significantly better graft survival than local/poorly matched kidneys. 6. Delayed graft function is associated in a complex relationship with organ sharing, prior graft failure, presensitization, and CsA use. 7. The increased rate of delayed graft function associated with organ sharing is overcome by the benefit of good HLA matching. 8. Since April 1986, purposeful organ sharing at SEOPF centers for good HLA matching has been associated with improved graft survival, especially in patients at high risk due to presensitization or prior graft failure.

Lymphocytotoxicity crossmatch proficiency testing among regional transplant centers.

During an eight and a half-year period (1976-1984), 408 combinations of different cells and sera involving 12,652 lymphocytotoxicity crossmatch reactions performed by 21-37 histocompatibility laboratories were studied in 20 proficiency tests conducted by the South-Eastern Organ Procurement Foundation. Consensus by 75% or more laboratories were obtained on 336 (82.4%) of these test samples: of the 10,410 reactions examined with the use of these consensus cells and sera, only 649 (6.2%) were discordant. Considering positive and negative reactions separately, the discordant (false) positive and negative rates were similar (6.1%, 6.4%, respectively). Varying specific factors in given tests indicated that discordant results were significantly reduced by (1) dispensing sera in trays centrally rather than locally; (2) using a standard one-wash procedure rather than local or three-wash methods; and (3) using reagent (monospecific) rather than patient (polyspecific) sera. Use of standard versus local rabbit complement did not significantly influence the concordance of results.

Antibodies to crossreactive HLA antigens. Evaluation by cytotoxicity, flow cytometry, and inhibition of monoclonal antibody binding.

Broadly sensitized patients have antibodies that react with cells from most individuals except those having HLA antigens similar to their own. In this work we addressed the question of whether crossreactive antigens could also be considered compatible. Broadly reactive sera were tested in multiple experiments against lymphocytes selectively mismatched for only one HLA-A, B antigen. Antibodies, measured by cytotoxicity and flow cytometry, were detected in most cases, and their prevalence was the same, regardless of whether the mismatched HLA antigen was crossreactive with the patient's. There were several patients bearing one of the A2, B5, or B7 crossreactive group antigens with antibodies against another antigen of the group. Anti-HLA-A2 reactivity was further evaluated by testing the inhibitory effect of broadly reactive sera on the binding of an anti-HLA-A2 monoclonal antibody. Anti-A2 reactivity measured in this assay was detected in A28-positive patients as frequently as in A28-negative patients. These results suggest that antigens of several crossreactive antigen groups are significantly immunogenic and elicit antibodies as often as noncrossreactive antigens. Further studies are necessary to evaluate other less-frequent antigen combinations.