Highly sensitized (HS) patients, who have high levels of preformed anti-HLA (human leukocyte antigen) antibodies, face significant challenges in accessing kidney transplantation because of the limited pool of HLA-compatible donors. The definition of a candidate considered HS varies across countries and national consensus statements. Currently, two main criteria are used to assess the severity of sensitization: Virtual Panel-Reactive Antibodies ¹ and calculated Reaction Frequency (cRF) ² . Both estimate the proportion of deceased donors incompatible with a patient owing to pre-formed HLA antibodies [1]. This immunologic barrier leads to prolonged waiting times and increased morbidity and mortality while on a waiting list.

To address this issue, several complementary strategies exist worldwide: 1) national priorities for candidates with high cRF and long dialysis duration, including Acceptable Mismatch Programs (AMPs); 2) expansion of kidney-exchange programs and cross-border sharing, both expanding donor options without increasing rejection risk or compromising long-term graft survival; 3) immunologic forgiveness policies (unacceptable antigen delisting for antibodies that declined in level or disappeared over time or that do not activate complement); and 4) desensitization strategies.

In France, national priorities in the allocation rules are based on two principles: difficulty in access to transplantation and the excellence of HLA matching. A kidney recipient is considered HS if their peak cRF is ≥85%; these patients automatically receive national HS priority, provided that their current cRF remains >70%. The system operates via three sequential, prioritized allocation pathways, with systematic age matching and ABO blood-group compatibility exemption in certain circumstances (see below). The pathways are 1) full-match HLA-A, -B, and -DR priority; 2) the AMP; and 3) a priority allowing no more than one mismatch for HLA-A, -B, and -DR between the donor and the recipient. In 2024, 9% of new registrations and 17% of active prevalent candidates were considered hyperimmunized [2].

In some cases, a long list of unacceptable HLA antigens (Ag) or high cRF is insufficient to explain immunologic difficulties in accessing a transplant. This is particularly true because the national score for allocating kidney grafts favors a good HLA match, especially for class-II Ag in recipients <45 years old. This situation may be due to rare haplotypes, including those with no linkage disequilibrium or partial or complete homozygosity. Therefore, another indicator is used: the Potential Matched Donor (PMD). The PMD estimates the number of compatible and well-HLA matched donors among all utilized kidney donations after brain determination of death (DBDD) donors within the same blood group over the past 5 years [3, 4].

A low PMD score indicates that the patient has very few suitable donors. Consequently, these patients receive additional points in the French National Kidney Allocation Score (NKAS) whenever a compatible and well-matched donor becomes available (see Table 1 for a description of kidney transplantation allocation pathways). This final point explains why a certain number of HS patients undergo transplantation via the NKAS system.

France introduced its AMP in 2005 (Table 2). An initial evaluation in 2010 demonstrated improved access to kidney transplantation for HS patients, with no difference in graft survival from other patients. Within 2 years, access to transplantation for highly immunized incident patients increased, but access decreased for non-immunized or slightly immunized patients during the same period [5]. The AMP was significantly enhanced in 2015 with the adoption of advanced solid-phase assays, such as Luminex, which have revolutionized HLA antibody (Ab) screening and identification, particularly for class-II HLA Ag. Single Antigen Flow Bead (SAFB) assays enable precise, semi-quantitative detection of HLA Ab and acceptable HLA Ag, in accordance with national guidelines issued by the French Society of Histocompatibility and Immunogenetics (SFHI) [6].

The « virtual cross match » is routinely implemented in all transplant centers in France to decrease the duration of cold ischemia and the risk of delayed refusal for immunologic reasons [6]. The conditions for performing a pretransplant virtual crossmatch in an emergency are standardized by the national scientific committee of the Agence de la Biomédecine (official referral procedure 7 June 2016, 2016G-04 deliberation). HLA-C, -DP, and -DQa typing is now available for each deceased donor, which improves the pretransplant virtual crossmatch accuracy [6].

The aim of this study was to 1) evaluate the transplant activity of HS patients and 2) compare post-transplant outcomes of HS kidney transplant recipients according to whether they benefited from the French AMP or other allocation modalities for HS transplantation.

This study demonstrates that improved death-censored graft survival in HS patients is achieved with the French AMP, offering post-transplant survival comparable to that for non-HS recipients transplanted under the standard allocation scheme. Despite an increased risk of rejection, HS recipients transplanted under the AMP achieved significantly better 5-year graft survival than HS patients transplanted outside the AMP, without additional desensitization strategies. At comparable cRF levels, HS AMP candidates awaiting a first transplant had better access to transplantation than those awaiting retransplantation, which reflects higher class-II HLA immunization and fewer permissible class-II HLA Ag in the latter group. The profile of patients who benefited most from access programs included candidates with common HLA phenotypes with cRF >97%, first-transplant recipients, patients with lower class-II HLA sensitization, particularly against HLA-DQβ Ag, and those with a greater number of permissible mismatches in class-II HLA, including DR and DQβ. In contrast, HS non-AMP patients more frequently had a history of prior kidney transplantation. This observation reflects listing for retransplantation commonly being associated with increased class-II HLA immunization and DSAs. In this context, selected patients might be under an immunologic forgiveness policy to facilitate access to a subsequent graft.

Over the past 10 years, the proportion of HS patients on the active waiting list decreased markedly, from 29% to 17%. This improvement is mainly the result of two combined factors.

The first factor is better HLA matching strategies, especially the emphasis on high-quality class-II matching, which is included in our French National Kidney Allocation Score and is strongly weighted for younger recipients. This strategy, implemented since the nationalization of the kidney allocation score [3], helps reduce the creation of hyperimmunized patients after a previous graft failure by introducing immunologic sparing for the youngest recipients. HLA compatibility, particularly class-II HLA, carries a high coefficient in the score, especially for the youngest patients; the weight then decreases with recipient age. By promoting HLA class-II matching among young adult recipients, the number of hyperimmunized re-registrants has been limited and post-transplant sensitization reduced.

The second factor in the reduced proportion of HS patients on the active waiting list is improved allelic resolution of anti-HLA antibody testing and donor HLA typing and, a better interpretation based on SFHI and European Federation for Immunogenetics recommendations [8]. Recent recommendations have helped refine HLA antibody characterization by encouraging high-resolution, allele-level analysis while maintaining a critical approach to the risk of false-positive SAFB results, particularly via the evaluation of autoreactivity and assay artefacts [9–14]. These guidelines also emphasize avoiding the blanket exclusion of all DQβ Ag carrying a given DQα epitope and instead recommend focusing on the corresponding frequent (haplotype-based) DR-associated specificities in the presence of clinically relevant anti-DQα reactivity. Additional measures, such as improved technical standardization across laboratories, reduction of inter-assay variability, and the use of epitope-based interpretation tools, have been implemented, thereby enhancing the accuracy of unacceptable Ag assignment [12].

Collectively, these advances have substantially reduced the number of listed unacceptable HLA Ag and, consequently, the proportion of candidates classified as HS. These developments could be effectively implemented with the national adoption of the pretransplant virtual crossmatch, which serves as the final and essential safety check before accepting a graft. This implementation was facilitated by the integration, within the CRISTAL donor-management platform, of allele-level donor HLA typing together with national guidelines defining interpretation criteria and technical warning thresholds [6].

The French experience can be viewed as a succession of major immunologic milestones: 1) nationwide implementation of SAFB-based unacceptable Ag in 2009; 2) introduction of a cRF directly derived from SAFB-defined unacceptable Ag in late 2009; 3) incorporation of anti-HLA-DQ antibodies into kidney allocation and cRF calculation in 2009; 4) introduction of the new French Kidney Allocation System in 2015 with improved HLA matching, particularly for younger recipients; 5) nationwide implementation of virtual crossmatch and intermediate/high-resolution HLA typing from 2016 onward; and 6) progressive refinement of antibody interpretation by the exclusion of denatured HLA antibodies and isolated allele-specific sensitizations. Together, these successive developments likely explain why France experienced an early increase in the proportion of HS candidates, followed by a sustained decline over the last few years. Future changes in national consensus regarding the integration of anti-HLA-Cw and -DP antibodies into cRF calculation and the allocation algorithm for HS patients may again modify the epidemiology of HLA sensitization, particularly among retransplant candidates.

Although the cRF remains the cornerstone indicator for defining immunologic sensitization and access to national priority, it does not fully capture the structural difficulty of identifying a compatible donor for a given candidate. In this respect, complementary indicators such as the PMD, which estimates the number of compatible and well-matched donors among actually utilized deceased donors in recent years, provide additional insight into real-world transplantability. Patients with similar cRF values may exhibit markedly different PMD profiles, reflecting differences in HLA phenotype frequency and compatibility constraints. This distinction is particularly relevant for candidates with rare HLA phenotypes or limited acceptable mismatches, who may remain disadvantaged despite comparable cRF levels.

HS patients with a cRF >85% but <98% have good access to kidney transplantation. Their transplant rate increases markedly after 18 months on the waiting list and eventually reaches that of non-immunized first transplant candidates. Among candidates with cRF 98%–100%, most are awaiting retransplantation. Of the quarter of candidates who are first-transplant patients, most are women sensitized (90%) through pregnancy and/or transfusion, with broad class-I HLA immunization. In this group, first-time registrants have higher transplant rates, reaching 24% by 3 years. Overall, patients with common HLA phenotypes, particularly first-transplant candidates, including women, benefit most, because they have fewer unacceptable class-II Ag and more acceptable mismatches, especially regarding HLA-DQβ Ag. Finally, the AMP program also addresses sex inequities, because women are more frequently sensitized than men.

After adjustment, risk of death-censured graft failure was 31% higher for HS non-AMP than HS AMP recipients. Other significant risk factors included young recipient age, high body mass index, multiple comorbidities, prolonged dialysis duration (>3 years), retransplantation, and older donors with hypertension. Young recipient age has consistently emerged as a risk factor in numerous studies, because immune responsiveness tends to be more vigorous in younger individuals and this group also has a higher likelihood of non-adherence. Cold ischemia time also plays an important role, as it does in all kidney transplants, by inducing damage-associated molecular patterns via ischemia–reperfusion injury, leading to the overexpression of HLA Ag on endothelial cells. Moreover, recent mechanistic reviews highlight that ischemia–reperfusion injury triggers endothelial activation and complex innate–adaptive immune interactions that contribute to both acute and antibody-mediated rejection in kidney transplantation [15].

The French allocation framework has historically embedded the principle that priority for HS candidates should be granted only when the proposed transplant is expected to carry an acceptable immunologic risk and outcomes comparable to those observed under standard allocation rules. Therefore, national priority has never been conceived as an unconditional advantage but rather as a mechanism to restore equity while preserving optimal graft utility. In this context, the AMP fits precisely within this regulatory philosophy because it is able to expand donor options: it allows for improved access for difficult-to-match candidates without increasing rejection risk or compromising long-term graft survival. Several notable differences distinguish the French AMP from the Eurotransplant AMP. Both rely on a cRF threshold of 85% based on real donor data, although the Eurotransplant AMP is accessible after 2 years of waiting time on dialysis and centralizes eligibility assessment within a single reference laboratory in Leiden, where the immunologic profile is determined not just from SAFB testing but also includes the HLA Matchmaker [1, 16, 17]. This algorithm identifies acceptable Ag predicted to carry a very low risk of de novo DSAs formation, based on shared or non-permissive eplets, whereas the French AMP prioritizes identifying donors without any historical or current DSA on the day of transplantation. Despite these methodological differences, outcomes remain excellent in both programs, as shown in the most recent Eurotransplant reports [11, 18, 19]. A final distinction is that the French AMP does not incorporate anti-HLA-Cw or -DP antibodies in its exclusion criteria because of the lack of national consensus on their interpretation, particularly given the high proportion of false-positive anti-HLA-Cw results in first kidney transplant candidates [10, 20]. These antibodies are instead managed at the time of organ offer via pretransplant virtual crossmatch assessment and, when necessary, confirmatory cellular testing.

Nevertheless, a subset of candidates remains poorly served by current allocation strategies, including those with extremely high cRF values combined with rare HLA phenotypes, absence of common linkage-disequilibrium haplotypes, or homozygosity at multiple HLA loci. For these patients, even acceptable mismatch–based approaches may fail to generate compatible offers. These limitations underscore the need to shift the focus upstream, by preventing the emergence of such extreme immunologic profiles via improved HLA matching, particularly at the class-II level, beginning at the time of the first transplant for the youngest recipients. In the future, the number of such patients could be reduced if a minimum level of HLA class-II compatibility, or the inclusion of less immunogenic class-II matching, is systematically implemented in standard allocation systems. Another elegant approach would be to preferentially allocate both kidneys from a single donor with homozygous HLA class-II typing to HS candidates with cRF 99%–100% who have been on dialysis for several years. From the French CRISTAL Registry, we found 193 DBDDs aged 18–65 years who matched this hypothesis from 2021 to 2023. An analysis of the fate of these “golden” grafts showed that a significant proportion are often transplanted locally to less sensitized recipients with few or no years on dialysis, even though, because of the homozygous HLA DR-DQβ haplotype, the graft may represent the only compatible organ for HS patients waiting for >5 years.

Recent recommendations have also proposed strategies for these long-waiting patients without an immediately compatible donor, including immunologic forgiveness policies (Ag delisting after antibody decline or disappearance over time or antibodies that do not activate complement) [21, 22] or carefully monitored desensitization protocols [23, 24], based on simulations from deceased donor databases. Indeed, HS non-AMP recipients are at increased immunologic risk and require more intensive immunosuppressive regimens, with associated infectious or oncologic complications, yet outcomes remain acceptable, providing a viable option when no alternative is available.

A key limitation of this study is its observational design, which precludes definitive causal inference and leaves the possibility of residual confounding despite adjustment for measured covariates. In addition, mechanistic interpretations should be approached with caution because the present analysis could not reliably assess rejection episodes, DSAs, de novo DSAs, or immunologic causes of graft loss comprehensively. The CRISTAL registry, managed by the Agence de la biomédecine, offers nationwide coverage of organ allocation and transplantation activity in France but is primarily designed to support allocation processes and long-term outcome monitoring rather than detailed clinical research. As a result, although the registry includes fields for acute and chronic rejection, antibody-mediated rejection, and cellular rejection, with or without DSAs, completion of these fields is not mandatory and the data quality is not subject to the same level of control as allocation-related data. Likewise, the specific cause of graft loss is incompletely recorded. Consequently, the reporting of immunologic complications is not exhaustive, which limits the robustness of any mechanistic analysis between preformed sensitization and graft outcomes. To further explore this issue, we performed an additional analysis by the presence of historical preformed DSAs that had been subsequently removed from the list of unacceptable antigens during the waiting period. This analysis showed a significant difference in death-censored graft survival between HS AMP recipients, HS non-AMP recipients without such historical DSA, and HS non-AMP recipients with historical DSAs, with 5-year death-censored graft survival rates of 85.6% [83.4–87.5], 81.7% [79.0–84.0], and 75.4% [61.2–85.0], respectively (p = 0.024). In addition, among graft losses for which rejection was reported as acute or chronic, the corresponding proportions were 30%, 38%, and 43%, although these data should be interpreted with caution given incomplete reporting. Taken together, these findings are consistent with a possible role of immunologic memory and preformed DSAs, but they do not allow firm mechanistic conclusions.

In conclusion, the French AMP represents a major advancement in kidney transplantation for HS patients, substantially improving their access to suitable donors while maintaining post-transplant outcomes comparable to those of non-sensitized recipients. More broadly, our findings illustrate that acceptable mismatch–based allocation represents an effective compromise between equity and utility. By selectively expanding donor compatibility while maintaining stringent immunologic safety criteria, the AMP restores access to transplantation for HS candidates without compromising graft outcomes. This balance is particularly critical in systems in which organ scarcity mandates that increased equity for disadvantaged patients must not come at the expense of predictable graft failure. By facilitating transplantation for women, who are disproportionately represented among HS candidates due to prior pregnancies, the program also contributes to reducing sex-related inequities in organ allocation [25]. More broadly, these results underscore that the most effective long-term strategy remains careful immunologic management at the time of transplantation, particularly through high-quality HLA class-II matching in young recipients, which minimizes alloimmune risk and maximizes graft longevity. Overall, the AMP exemplifies how tailored allocation policies, informed by precise immunologic profiling, can optimize both equity and clinical outcomes in transplantation.