Real Benefits From Virtual Crossmatches
Mary Sue Leffell · Transplantation · 2010
It is often said that the most pressing issue in solid organ transplantation is the critical shortage of deceased donor organs relative to the growing number of transplant candidates. This need has led to increased procurement of organs from expanded criteria donors (ECD) and from donors after cardiac death (DCD). In the United States, ECD and DCD kidneys now comprise one third of the organs procured for renal transplantation (1). Use of alternative donors is not without a price, however, because the incidence of delayed graft function and risk of graft loss are both increased with ECD and DCD organs (2, 3). Because prolonged cold ischemia time (CIT) is associated with delayed graft function (DGF) and other injury that may adversely impact allograft survival, there is great incentive to improve the efficiency of organ procurement and allocation, limiting as much as possible CIT (3–5). The article by Taylor et al. (6) in this issue of Transplantation addresses this goal with a 10-year prospective study of eliminating a pretransplant crossmatch in carefully selected patients, relying instead on a “virtual crossmatch” based on thorough analysis of human leukocyte antigen (HLA)-specific antibodies and candidate sensitization history. Taylor et al. reported the outcomes of 257 kidney transplants for which a prospective donor:recipient crossmatch was omitted on the basis of a predicted, “virtual” negative crossmatch. The virtual crossmatch recipients were among a total of 606 transplants performed from 1998 to 2008. The authors based their crossmatch policy on their experience of accurately predicting negative crossmatches using a combination of lymphocytotoxicity panel screening, enzyme-linked immunosorbent assays, and Luminex-based multiple- and single-antigen assays. Patients considered for omission of the prospective crossmatch were either nonsensitized or had only low levels of well-defined HLA-specific antibodies. They followed a strict policy for monitoring all potential sensitizing events. The only candidates included had no recent alloimmunizing events and a negative antibody-screening history for at least 6 months or had clearly defined antibodies that were stable or diminishing during the previous 6 months. Unacceptable HLA mismatches were defined by cytotoxicity tests or solid-phase immunoassays in sera from the previous 12 months. For the Luminex single-antigen assays, the criteria for assignment of unacceptable antigens were correlated with flow cytometric crossmatch results. HLA-specific antibodies with an adjusted median fluorescence intensity greater than 1500 were considered as unacceptable because this cutoff was a twofold dilution below that detectable by flow cytometry in their center. Using these conservative criteria, retrospective cytotoxicity crossmatches were found to be negative as predicted in all the cases. The authors considered three measures of outcome: DGF, time to graft failure or patient death, and probability of biopsy-proven acute rejection within the first posttransplant year. The mean CIT was 14.3 hr among the patients without a prospective crossmatch compared with 16.7 hr for those with a crossmatch. Among recipients of kidneys donated after brain death, there was a beneficial effect of crossmatch omission with a DGF rate of 18% compared with 28% with a prospective crossmatch. No significant difference in the incidence of delayed graft function was noted among recipients of DCD kidneys, which the authors attributed to the degree of acute tubular necrosis resulting from warm ischemic injury that negated any effect of reduction in CIT. Importantly, omission of the crossmatch was not associated with any impact on acute rejection rate or long-term graft survival. The strongest predictors of outcome in adjusted Cox regression models were acute rejection, donor age, and the number of HLA-DR mismatches, whereas omission of the pretransplant crossmatch remained nonsignificant. The conclusions of this study seem well justified by the data: namely, that a prospective crossmatch can be safely omitted in selected patients and that this policy can help reduce CIT and may decrease the DGF rate. But there are some issues that should be considered by others wishing to implement a similar policy. Importantly, Taylor et al. followed a well thought out approach. They established stringent patient and sensitization criteria and based their interpretation of solid-phase assays on correlations with crossmatch results. The authors stressed the importance of documenting potential alloimmunizing events and regular antibody screening. Transplant centers differ in their treatment regimens and levels of acceptable immunologic risk, and there is inherent interlaboratory variability in antibody assays. Therefore, it is vital that each center follow a similar approach. It should also be noted that omission of prospective crossmatches may not result in a reduction in CIT and DGF in all cases, particularly when histocompatibility testing can be performed on blood samples obtained before organ recovery or when there are unavoidable delays in transplantation such as unavailability of operating rooms. In cases where the crossmatch is not the rate-limiting factor and in cases where antibody screening is not current or the sensitization history is not complete, a cytotoxicity crossmatch remains a relatively fast and inexpensive safeguard. The virtual crossmatch is not new and is a topic of much current interest. Nonetheless, the study by Taylor et al. is the largest prospective study to date and provides important long-term outcome data and an excellent approach for transplantation of renal candidates with no or minimal sensitization. Their article adds support to the growing body of data indicating that the virtual crossmatch can be applied safely when based on regular and thorough antibody characterization. Other recent reports have shown that virtual crossmatches are highly accurate even for more highly sensitized patients (7–10), can be applied to nonrenal transplant candidates (11), and can increase the donor pool, both in numbers and geographic procurement area (11, 12). What Taylor et al. contributed additionally is the demonstration that the combination of stringent patient criteria with accurate crossmatch prediction can result in reduced CIT and improved transplant outcomes. Although these benefits may not be realized in all cases, with the increasing use of ECD and DCD organs that seem more susceptible to prolonged CIT and DGF, these practices merit serious consideration by other centers.