[Note: I’ve had a major revision in thinking about solutions to the O blood type problem. This blog post is no longer valid and will be replaced by a new one soon.]

Yesterday’s blog post showed that there is a shortage of O blood type kidneys for the patients who need them. There are two possible short-term solutions. One is to increase the number of altruistic O type donors in the exchange. The other is to encourage patients who are paired with an O type donor, but are not O type themselves, to enter the exchange.

Encouraging people with O blood type to become altruistic donors is not a solution

If you ever donate blood, you may be aware that blood centers are always asking for volunteers who are O- blood type. That’s because their blood is least likely to elicit an immune response from the recipient (assuming the recipient does not have high levels of HLA antibodies). Although people with O- blood constitute less than 10% of the U.S. population, I presume there are some who donate very frequently and that as a group these volunteers provide a significant proportion of the blood supply. Having lots of extra O- type blood available ensures there will never be a shortage of blood for patients with O type blood who need it.

This strategy cannot work for kidneys because a kidney donor can only donate once. This means people who are inclined to donate a kidney will do so and leave the pool. Currently, the annual number of altruistic kidney donors in the U.S. is about 200, so the number of O donors (both RhD+ and -) is under 100. To fill the current O type kidney shortfall would require about 4,000 one-time volunteers a year (this includes both the exchange and the UNOS waiting list). The number of altruistic donors is growing rapidly, but great effort would be needed to encourage a forty-fold increase in the number of healthy O type adults to donate. At some point this could lead to coercion. (For a taste of what a world dependent on “special people” could be like, read Kazuo Ishiguro’s novel, Never Let Me Go.)

Encouraging matching pairs to volunteer to enter the exchange is not a solution

There is a ready source of O type kidney donors, at least for exchanges. They are the donors in matched pairs. Currently, matched pairs never enter the exchange and may not even be aware that an exchange exists. Even if they are knowledgeable about exchanges, they would never consider participating in one because they have no need.

Further, it would be unethical to ask these matching pairs to enter into the exchange for the sole purpose of helping strangers get a match. We would be asking them to delay their transplant surgery in order to apply to the exchange, wait for a match to occur, and undergo additional compatibility testing. Then we would ask them to break an emotional bond between the pair. The recipient would have to take a kidney from a stranger and the donor would have to give a kidney to a stranger. The donor may back out of such an arrangement, leaving both recipients without a kidney. In return for their effort, the volunteering matched pair gets nothing in return (except the knowledge they’ve improved the life of a stranger). This would definitely not meet the ethical requirement of not harming patients either medically or psychologically.

Getting matched recipients to want to enter the exchange for their own benefit can work

Patients in unmatched pairs enter the exchange pool because they will never be worse off (they can exit the pool with the same donor they entered with at any time) and potentially better off (they can trade to get a match). They join the exchange to improve their lives.

There may be a way we can promise the same benefit to patients in a matched pair. That is, even though they match now, they may be willing to enter the exchange pool if there is a potential of finding someone who is a better match than their current partner. If no better match is found they can exit the pool at any time and continue with the transplant with the original donor.

This better match comes in the form of HLA compatibility. In the calculations of matching probabilities shown in yesterday’s post, we ignored HLA compatibility. Now we want to reintroduce it. As explained in an Oct 2008 blog post, there are 3 pairs of HLA genes that most strongly affect organ rejection. The benefits of finding a donor whose 6 HLAs are a subset of the recipient’s (called a 6-HLA match) are substantial. Getting a transplant from this donor can result in lower dosage of immunosuppressant medication. This will potentially leave the recipient’s immune system stronger, making them more resistant to infection and some cancers. A 6-HLA match kidney will also reduce the chance of organ rejection. Finally, if the recipient does require another transplant in the future, it will be easier to find a match since they will have developed fewer HLA antibodies.

It isn’t enough to promise patients a trade that produces a donor with a slightly lower level of HLA incompatibility; it has to be significantly better with proven medical benefits. Otherwise, we will be asking them to accept an exchange with a pair of strangers for the primary purpose of helping the strangers. Current research indicates that for patients who do not have any HLA sensitivity, a live transplant from a 4- or 5-HLA match doesn’t lead to significantly better results than one from a 1- or 2-HLA match. Thus, it may be necessary to restrict splitting of matched pair to cases only where a 6-HLA match is found.

Finding a 6-HLA match by chance is highly unlikely. So unless the donor is a sibling of the recipient, the donor-recipient pair are unlikely to be a 6-HLA match. Thus, if the recipient wants to find a 6-HLA match, their best bet to find one may be to enter the exchange pool.

Modification to existing matching algorithms

In the previous models for kidney exchange there was only one class of trader. All were looking for a minimum acceptable match. However, recipients (or their medical providers) hold private information regarding the advantages of seeking a 6-HLA match. They may decide that waiting for such a match will produce a better outcome (or greater utility to use economic parlance) than a shorter wait for a minimum acceptable match. It isn’t certain what proportion of kidney patients would enter the exchange for a chance to find a 6-HLA match. But my guess is that if the matching process was run once every two weeks or more frequently, then a large proportion would be willing to try it. I know that I would.

An article in Management Sci. Nov 2006 (subscription required) shows how using multiple queues, each with a different quality of organ and with longer waits for higher quality organs, can improve the allocation of cadaver organs. A similar approach can be used to encourage matched pairs to enter a kidney exchange.

In the proposal stated here, there will only be two queues available, one to get a minimum acceptable match and another to get a 6-HLA match. All participants would be told what the expected waiting time for each queue would be and could exit at any time. Patients with an unmatched donor may enter either queue. They could also switch queues at any time. Patients with a matched donor would only enter the second queue. There would be a constraint set so that traders in the second queue would never get split unless a 6-HLA match is found for the recipient. Toumas Sandholm of Carnegie Mellon University who has studied kidney exchanges and has been mentioned in this blog previously has proposed arrangements similar to this.

There are two ethical benefits from adding a separate queue to accommodate matched pairs in an exchange. First, all donors can be informed early that there is a chance that may not donate to recipient they are paired with and instead will enter into an exchange. Currently, donors are generally not informed of the existence of an exchange at the time they volunteer to be a donor. They may be surprised later to learn that the patient wants to enter an exchange. They may not want to donate to a stranger but feel an obligation to continue with the donation. Informing them early that there is a likelihood of entering an exchange allows them to decide if  they want to donate before they make a commitment and avoid anxiety and coercion later.

Second, the separate queues allows healthcare professionals and community outreach volunteers to inform all patients and potential donors of the benefits of an exchange. Under the current single queue exchange, matched pairs should not be informed of exchanges, much less encouraged to participate. The simple act of informing them about the exchange option would reveal a bias in favor of them. [Disclosure: I plan to become a community outreach volunteer after I complete my kidney donation and would like to be able to encourage matched pairs to enter the exchanges.]

Before a recommendation can be made to the kidney exchanges to add a second queue, evidence has to be provided that it will actually help patients in matched pairs find better kidneys. This would require simulations of two-queue exchanges to be conducted using generated data. I believe this is a promising avenue of research. With positive simulation results in hand, the National Kidney Registry, Alliance for Paired Donation, and UNOS could be persuaded to add a queue for matched pairs and help solve the shortage of O kidneys in exchanges.

[Update: There is a discussion of altruistic matched pairs in the Nephr. Times Feb 2010.]