Highly Sensitized Kidney
Transplant Candidates

Two questions, not one

In a highly sensitized candidate, the calculated panel-reactive antibody (cPRA) tells you how hard it will be to find any compatible donor; the donor-specific antibody (DSA) and the crossmatch tell you the risk of proceeding with one particular donor. These are different problems, and a safe access strategy has to optimize both at once. Almost every avoidable error in this population comes from collapsing the two — treating a very high cPRA as if it were a rejection risk, or treating a low cPRA as if it guaranteed donor-specific safety.

This guide is clinician decision-support and referral guidance. It is not a consent script, a desensitization protocol, or an organ-acceptance directive. Its aim is that a general nephrologist, internist, dialysis physician, or transplant coordinator who does not work daily in a histocompatibility laboratory can read an immunologic record, understand what each number does and does not mean, know which access strategy sits where in the hierarchy, and know which decisions must stay inside a transplant center with an accredited human leukocyte antigen (HLA) laboratory.

The two axes combine into a simple triage that orients every later decision:

Take-away. Before you look at any single result, decide which question it answers: is this an access number (cPRA) or a donor-specific risk number (DSA, crossmatch)? They are optimized by different strategies and must not be traded against each other.

What makes a candidate sensitized — and how to prevent it

Sensitization is immunologic memory of foreign HLA. It is acquired through three classical exposures — pregnancy (fetal paternal HLA), transfusion of cellular blood products, and a previous transplant — and each repeat exposure deepens it. The failing or failed allograft is a potent and under-recognized sensitizing context: allograft nephrectomy and, especially, withdrawal of maintenance immunosuppression after graft failure can unmask and amplify anti-HLA antibody, converting a re-transplant candidate into a highly sensitized one. This matters because the mechanism is not simple antibody accumulation — long-lived plasma cells and memory B cells persist, so an antibody that has fallen below the detection threshold can rebound rapidly on re-exposure. Infection and vaccination are not HLA alloexposures and do not sensitize, though an intercurrent inflammatory event can transiently perturb antibody assays and should be noted when interpreting a sample.

Because endowment of donors is fixed and memory is durable, the most reliable lever a referring physician holds is prevention of avoidable sensitization. In a transplant-eligible patient with chronic kidney disease (CKD), that means treating anemia with iron and erythropoiesis-stimulating agents rather than reflexive transfusion, applying patient blood management, and reserving cellular products for genuine indications. A unit of blood given casually to a young dialysis patient can narrow their donor pool for life.

Take-away. The cheapest way to help a highly sensitized candidate is to have prevented the sensitization — avoid the nonessential transfusion, and never stop a failed graft's immunosuppression without a transplant-center conversation.

The vocabulary that prevents errors

Most miscommunication between a referring clinician and an HLA laboratory is vocabulary drift — the same word carrying a stronger meaning at the bedside than it does at the bench. The table below fixes each term to what it does mean and, just as importantly, what it does not.

How antibody injures a graft — the chain, and the number that does not measure it

HLA molecules are inherited cell-surface identity labels; a donor kidney carries a particular set, and a sensitized recipient may hold antibody that reads some of those labels as foreign. When a clinically important DSA binds HLA on graft endothelium it can fix complement, recruit fragment-crystallizable (Fc)-receptor-bearing immune cells, activate endothelium, and drive microvascular inflammation. The result ranges from no measurable injury, through subclinical injury, to hyperacute or active antibody-mediated rejection (AMR) and, later, chronic transplant glomerulopathy. Crucially, whether injury occurs depends on far more than a bead's MFI: antibody amount and avidity, complement- and Fc-effector function, HLA density and locus, the memory response, the depth of immunosuppression, and the graft's own susceptibility all contribute.

Take-away. Fix the vocabulary before the decision: cPRA is a population estimate, MFI is an assay signal, DSA is a risk factor — and none of the three is a diagnosis of rejection, which is a clinicopathologic call.

How single-antigen bead testing works — and why the number can mislead

The single-antigen bead (SAB) assay is the workhorse of modern sensitization testing, and understanding its mechanics is what lets a clinician read its output honestly. Each bead is a microscopic test target coated with one HLA antigen; the assay reports how much antibody bound to each target under its own conditions. In order: patient serum is exposed to a panel of beads, each carrying a single HLA antigen; bound IgG is detected with a fluorescent secondary reagent; the instrument reports an MFI for every bead; and the laboratory then interprets the pattern against controls, background, the patient's history, and the donor's typing. The MFI is the third step — an intermediate signal, not the conclusion.

A number of well-described artifacts can make the raw MFI mislead in either direction, which is why the pattern — not the isolated value — is what counts:

• Prozone / complement interference — a neat (undiluted) serum can read falsely low; treatment with ethylenediaminetetraacetic acid (EDTA), heat, or dilution unmasks the true signal.
• Bead saturation — a neat-serum MFI can plateau, so two sera with very different true antibody burdens can report similar values; titration separates them.
• Shared epitopes (eplets) — antibody to one shared structural feature lights up several beads, producing distributed or unexpected patterns.
• Denatured antigen reactivity — some beads present denatured HLA whose in vivo relevance is uncertain.
• Lot, platform, and inter-laboratory variation — the same serum can read differently across kits and centers.
• Locus and allele gaps — HLA-DP, HLA-C, and the alpha chains DQA1/DPA1 are incompletely covered, and allele-level typing gaps hide real specificities.
• Cumulative versus immunodominant DSA — a summed load reads differently from a single dominant specificity.
• Recent IVIG or antibody-directed therapy, autoantibody, or nonspecific background — all distort the picture.
• Historical peak versus current serum — the two answer different clinical questions and must not be conflated.

Take-away. Interpret the SAB pattern — with history, dilution, and crossmatch — not a single fluorescence value, and never treat a borderline MFI as a decision on its own.

Read the immunologic record as a timeline, not a snapshot

Sensitization has a natural history: a sensitizing exposure produces a peak antibody level, which then declines — sometimes below the program's detection threshold — before a donor is offered, and which may rebound after transplantation because the memory compartment was never removed. Reading a single current serum without that trajectory is like reading one creatinine without a baseline. Three clinically distinct states fall out of the timeline, and they carry very different risk:

1. No historical or current DSA to the donor — the lowest-risk state.
2. Historical DSA, now below the program threshold or undetectable — reduced but not zero risk, because memory persists and can rebound.
3. Persistent current DSA, especially if strong and/or crossmatch-positive — the highest-risk state, and the one delisting cannot fix.

Take-away. Classify every candidate–donor pair into one of the three states — no DSA, historical-cleared, or persistent — because the safe strategy differs for each, and the classification requires the whole serum history, not the latest sample alone.

What the 2026 delisting cohort actually found

The most useful recent data on antigen delisting come from a retrospective, multicenter French cohort (Usureau and colleagues, Kidney International 2026) covering automated delisting across eight Île-de-France transplant centers from 2019 to March 2025. It is the anchor for a conservative, history-aware delisting concept — and it earns that role precisely because it separates historical from persistent antibody rather than lumping them.

Of 2,418 candidates in the access cohort, 901 were transplanted and 804 were analyzed after transplant. The access signal was concentrated in the hardest-to-match candidates: a reproducible benefit emerged only above a baseline cPRA of roughly 96.6%, and among candidates with baseline cPRA ≥98%, achieving a cPRA fold-change ≥3 was associated with a 2-year transplant probability of 26.7% versus 16.4%. A deliberately conservative simulation — a 3-year antibody look-back combined with an MFI 2000 rule — reached the target cPRA fold-change in 42.4% of candidates with baseline cPRA above 99%.

The safety analysis is where the history-dependence shows. Recipients fell into three groups: no historical DSA above 2000 MFI (n=661), historical DSA that had cleared below threshold by transplant (n=102), and persistent DSA above 2000 at transplant (n=41). Early post-transplant DSA above 2000 was detected in 24.5% of the historical-cleared group versus 2.7% of those with no such history — and 91.3% of that early DSA was rebound of a historical specificity, exactly what the memory-persistence physiology predicts. The composite of graft loss or death was 11.5%, 9.8%, and 24.4% across the three groups respectively; persistent current DSA remained independently associated with the composite.

A counterweight belongs beside it. Cucchiari and colleagues (Kidney International 2025), a multicenter prospective cohort of active delisting in candidates with cPRA ≥99.9%, found that delisting can indeed increase transplantation but at the cost of a meaningful rejection burden when the strategy is aggressive. Read together, the two cohorts point the same way: the access benefit is real and largest at the extreme of sensitization, but it is bought with immunologic risk that scales with how permissively the antibody threshold is set.

The arithmetic of access — teaching scale, not predicting an offer

Two calculations make the access problem concrete. If cPRA is expressed as a proportion, the fraction of random donors expected to be compatible on the modeled antigen list is 1 − cPRA, and the chance of at least one compatible donor among n independent donors is 1 − cPRAⁿ. The study's own access metric, the cPRA fold-change, is (1 − final cPRA) ÷ (1 − baseline cPRA) — the fold increase in the modeled fraction of compatible donors. A worked example: delisting that moves cPRA from 99% to 97% changes the compatible fraction from 0.01 to 0.03, a 3-fold access gain.

Take-away. The delisting evidence supports a conservative, time-aware strategy for the most sensitized candidates and warns against permissive MFI-only rules; the access arithmetic tells you how steep the mountain is, not whether a given climb is safe.

Delisting: what it changes, and what it does not

Delisting is an administrative access strategy: it removes selected HLA specificities from a candidate's unacceptable-antigen list so more donors are offered. It is the single easiest concept to overstate, because it changes a computer allocation record without changing the patient's biology at all.

Two mechanically different approaches exist. Time-dependent delisting ignores or reclassifies specificities not detected within a defined look-back period, while retaining the full historical record for risk review. MFI-dependent delisting raises the operational signal threshold above which an antigen is called unacceptable. The physiology of memory favors the first: a rule that requires current antibody clearance before a specificity is set aside is more defensible than one that simply tolerates persistent high-strength DSA — but both require local validation, and neither is a licence to cross a persistent strong DSA.

Take-away. Time-dependent delisting that demands current antibody clearance is the more biologically defensible design; whichever is used, delisting expands access and leaves the crossmatch and the antibody exactly where they were.

The hierarchy of access strategies

Because lower immunologic risk should generally be exhausted before higher risk is accepted, the strategies to expand access for a highly sensitized candidate form a hierarchy. The ordering is a default, not a rule — a candidate with negligible match probability should not languish indefinitely at a lower rung — but the presumption runs from the safest, cheapest interventions upward:

The evidence for placing KPE ahead of desensitization is a guideline position with modest underlying certainty: the European Society for Organ Transplantation (ESOT) 2022 HLA-antibody guideline recommends kidney paired exchange as the preferred option over desensitization when feasible — a recommendation the guideline itself bases on low-certainty evidence (it grades its statements with the GRADE system; confirm the exact strength wording against the source before quoting a grade). The rationale is mechanistic and pragmatic rather than trial-proven: a compatible kidney obtained through exchange avoids the antibody problem entirely, whereas desensitization manages it at cost and risk. The counterpoint, equally important, is that a highly sensitized candidate with a negligible match probability in the exchange pool should not be left there indefinitely — the hierarchy is a starting order, and futility at a lower rung is itself a reason to move up.

Why selected incompatible transplantation can still be the right call rests on older but landmark data: Orandi and colleagues (New England Journal of Medicine 2016), a 22-center observational study of HLA-incompatible live-donor recipients, found a survival advantage over matched candidates who remained on the waiting list or waited for a compatible deceased donor. That is an observational comparison, geography- and era-dependent, and not a head-to-head test of specific modern protocols — but it is why "wait for perfect" is not automatically safer than a well-supported incompatible transplant for a patient facing many more years of dialysis.

Take-away. Work up the staircase, not down it — but let demonstrated futility at a lower rung, not dogma, decide when to climb. KPE before desensitization is an ESOT recommendation on low-certainty evidence, and incompatible transplantation can still beat indefinite waiting for the right patient.

Desensitization in plain clinical terms

Desensitization is the set of therapies that reduce or control antibody-mediated incompatibility so a transplant can proceed. It is best understood by purpose — what each agent is trying to do to the antibody or its effects — and by its principal limitation. None of these agents confers durable tolerance; the antibody-producing system remains, so rebound is the rule and maintenance is intrinsic.

Take-away. Match the agent to its purpose — removal (PLEX, immunoadsorption), production (rituximab, proteasome/CD38 agents), or downstream injury (complement blockade, imlifidase's IgG cleavage) — and remember that every route leaves the memory intact, so none is a one-time cure.

Imlifidase: rapid IgG cleavage, not immune erasure

Imlifidase is an immunoglobulin G–degrading enzyme of Streptococcus pyogenes that cleaves IgG below the hinge, so the antibody can no longer link its antigen-binding arms to the complement- and Fc-mediated effector functions that injure a graft. Given hours before transplantation it can convert a positive crossmatch to negative quickly enough to permit the operation. What it does not do is remove the antibody-producing machinery: the recipient's plasma cells and memory persist, and DSA can rebound within days as newly synthesized IgG appears and the drug is cleared.

The clinical implications follow directly from the mechanism. Because imlifidase also cleaves therapeutic IgG antibodies, induction and any monoclonal agents must be sequenced around the drug per the licensed expert protocol; early AMR surveillance and a rescue plan must be in place; and an immediate crossmatch conversion must never be read as long-term tolerance.

Take-away. Imlifidase buys a negative crossmatch for a narrow, expert-managed group; it does not erase the immune response, DSA rebounds, and its access and approval are jurisdiction-specific — treat immediate crossmatch conversion as a window, not a cure.

The donor-offer immunologic review

When a possible DSA appears on an offer, the task is structured review, not a reflex. Work four domains — candidate context, the antibody record, the donor and crossmatch, and program capability — and integrate them; no single field, and no automatic score, produces an accept or decline.

A shared risk language for the conference table

Replace "positive versus negative" with four graded domains, so the conference reasons about a candidate rather than a label:

Take-away. Structure beats reflex: review candidate context, antibody record, donor/crossmatch, and program capability; grade the four risk domains; and let the multidisciplinary team — never a number or a checklist score — accept or decline.

After transplantation: monitoring without chasing MFI alone

Post-transplant surveillance in these recipients follows graft function, proteinuria, immunosuppressant exposure and adherence, and DSA — but DSA read in context, never as a solo trigger. DSA evolution is informative chiefly when paired with graft phenotype, and biopsy remains the best available way to assess suspected antibody-mediated tissue injury. MFI changes can reflect analytic variability, so each laboratory should define what counts as a meaningful change rather than reacting to noise. For stable recipients, the ESOT consensus on subclinical DSA monitoring offers a pragmatic scheme — DSA at roughly 3–6 months and annually thereafter — while acknowledging that local protocols and individual risk should modify it.

When de novo DSA is detected, first optimize adherence and maintenance immunosuppression; additional treatment should generally be guided by biopsy rather than by the antibody alone, and neither complement-binding status nor MFI by itself should decide whether to biopsy a patient with subclinical de novo DSA. This is the same discipline as the pre-transplant phase: the serology raises a question, the histology answers it.

Take-away. Monitor function, proteinuria, adherence, and DSA together; biopsy — not MFI — diagnoses AMR; and give every historical-cleared recipient an explicit early-rebound surveillance plan.

What the referring nephrologist should do today

Most of the leverage in this population is upstream of the transplant center, in the hands of the general nephrologist and the dialysis physician:

Take-away. Refer early, protect the donor pool by avoiding needless transfusion, send the full sensitization history, and leave MFI interpretation to the histocompatibility laboratory.

Cases for applied learning

Each case gives what matters, what can mislead, the next multidisciplinary question, and what not to do.

How the evidence is graded here

Claims in this guide are tiered, and the verbs track the tier. A Guideline / consensus position carries the source's own grade where one exists. Comparative outcome evidence means randomized or adjusted multicenter comparison. An Observational cohort shows its sample and endpoint and is reported with "was associated with," not "caused." Mechanistic / early evidence — laboratory work, small series, investigational therapy — is labeled as such and never presented as proof of patient benefit. Access benefit, perioperative AMR risk, and long-term graft survival are kept as three separate outcomes, because a strategy can improve one while worsening another.

Take-away. Keep access benefit, AMR risk, and long-term survival as distinct outcomes; state observational findings as associations; and treat the delisting evidence as conservative support, not a safety guarantee.