Organ allocation matching

Why does the patient who has waited longest for a kidney sometimes not receive the one that becomes available?

A queue seems like the fairest thing in the world. Turn up, take a ticket, be served in order. Nobody has to decide who deserves more, and nobody can be accused of favouritism.

Yet no national kidney allocation system works that way, and the reason is not that the designers were careless about fairness. So let us take the queue seriously and ask what it assumes. A queue is fair only if the thing being handed out is the same for everyone in it. Is a kidney?

Start with the obvious barrier. Blood group antigens must be compatible, or the graft is destroyed almost immediately. So already the queue has failed: the person at the front may simply be unable to receive this organ, and a rule that hands it to them is not a rule at all.

Push further, because the real difficulty is a step past blood groups. Every person carries HLA molecules — the surface proteins by which the immune system distinguishes self. A recipient who has previously been pregnant, transfused, or transplanted may have been exposed to foreign HLA and made antibodies against them. Those antibodies persist. Offer such a person a kidney bearing an antigen they are primed against and the graft can be attacked within minutes.

Now notice what follows, because this is where the queue does not merely fail but inverts. The fraction of donors a candidate has antibodies against is summarised as a calculated panel reactive antibody value, cPRA. Someone with cPRA of zero is compatible with essentially the whole donor pool. Someone at 99% is compatible with one donor in a hundred. Hold the offer rate fixed and the arithmetic is immediate: the second person needs roughly a hundred times as many organs to pass through the system before one of them can be theirs. Under pure waiting-time ordering, they would move up the list forever and never reach the front, because reaching the front is not what limits them. This is exactly why allocation systems grant priority points for high sensitisation — not to override fairness, but because a queue quietly starves the people for whom compatible organs are rarest.

There is a second pressure, and it comes from the clock rather than the immune system. A recovered kidney is perishable. It sits in cold storage, and graft function deteriorates as that cold ischaemia time lengthens. So the allocation cannot be a leisurely search for the ideal recipient: it must be run, offered, accepted, and transplanted within hours. If the top-ranked candidate cannot be reached, has an infection that day, or turns out on final crossmatch testing to have antibodies against this specific donor, the organ moves down the list — not as a slight, but because waiting for them costs the organ itself.

A third consideration is more contested. Kidneys differ in how long they are likely to function, candidates in how long they are likely to live, and some systems deliberately steer the longest-lasting organs toward the longest-surviving candidates so that a kidney with decades left is not buried two years later. That is an efficiency argument trading openly against age-blind fairness, and allocation policies have been rewritten repeatedly as that disagreement moves.

What would refute the case for matching over queueing. Two observations, either of which would be decisive. If graft survival turned out to be independent of the degree of HLA mismatch and of cold ischaemia time, the efficiency argument for matching would have nothing left, and a queue would be both fairer and no worse. And if highly sensitised candidates under priority schemes received transplants at no higher rate than before those schemes existed, the access argument would collapse too. The first is genuinely partly true, and it is the honest complication: modern immunosuppression has weakened the effect of HLA mismatch considerably compared with the era in which matching rules were written, and how much weight matching now deserves is actively argued.

Two further honesty notes. HLA matching has documented equity consequences, because HLA frequencies differ across ancestral populations and a donor pool drawn disproportionately from one group produces better matches for that group — allocation rules have been revised to blunt this, and the adequacy of those revisions is disputed. And nothing here decides an individual case: which patient receives which organ is set by allocation policy and by the transplant clinicians who can assess both donor and candidate, and no reasoning about mechanism substitutes for that.

The useful correction is about what kind of good an organ is. A queue is the right instrument for a fungible resource: any unit serves any claimant equally, so only order is in question. A kidney is not fungible. Its value is a property of the pairing, not of the organ, and can be zero for one candidate and decades of life for the next.

That is also the fixed difference from emergency triage, its nearest neighbour in this collection. Triage reorders a queue for a treatment that would work on anyone in it — the question is purely who goes first when capacity is short. Allocation asks something else: given that this specific item works only for some of the queue, how do we combine compatibility, urgency, expected benefit, and time waited into one ranking? Waiting time does not disappear from that ranking. It stops being the whole of it.

• How paired kidney exchange chains let two incompatible donor-recipient pairs each obtain a compatible organ, and how long such chains can run.
• Why deceased-donor and living-donor allocation face almost none of the same constraints, and what that reveals about which pressures are really binding.

Every term the collection defines is gathered in the glossary.