To arrive at a recommendation, our committee examined the following questions:
1Why is living donation necessary, and is it different for patients with hepatocellular carcinoma (HCC)?
2What is double equipoise, and how is it affected by the diagnosis of HCC?
3Is a paired exchange appropriate if one or both recipients have HCC?
DDLT, deceased donor liver transplantation; HCC, hepatocellular carcinoma; LDLT, living donor liver transplantation; UCSF, University of California San Francisco.
MATERIALS AND METHODS
Literature searches were conducted, and additional references were provided by committee members. Conference calls and e-mail exchanges were used to debate the relevant issues because there is a paucity of relevant published data.
Question 1. Why Is Living Donation Necessary, and Is It Different for Patients With HCC?
It is well established that even in countries with adequate access to deceased donor liver transplantation (DDLT), living liver donation is felt to be appropriate because of organ shortages, increasingly long waiting lists, and the expectation that many patients listed for liver transplantation will die while they are waiting for a suitable organ. The Milan criteria (a single tumor <5 cm in diameter or 2-3 lesions with individual diameters ≤3 cm) describe the extent of tumors in a population of patients with HCC and cirrhosis who have low recurrence rates and excellent survival rates after liver transplantation.1 Intent-to-treat studies have demonstrated that the risk of removal due to HCC progression beyond the Milan criteria (eg, the dropout rate) is greater than the risk of death while patients are awaiting transplantation.2 For these reasons, in the United States, the Organ Procurement and Transplantation Network has developed an organ allocation policy for DDLT that allows for increased priority for liver transplant candidates with stage II HCC. This policy does not apply to living donor liver transplantation (LDLT) because other candidates without HCC are not being disadvantaged.
Ensuring the safety of the living organ donor and optimizing his or her care were the foci of an international forum held in 2005 in Vancouver, British Columbia. The Vancouver forum established a set of practice principles for LDLT that include the following points3:
Living liver donation should be performed only if the risk to the donor is justified by the expectation of an acceptable outcome for the recipient.
The indications for LDLT should be the same as those established for DDLT, except for institutionally approved protocol studies that consider LDLT preferential to DDLT.
In comparison with waiting for an acceptable deceased donor organ to become available for transplantation, LDLT should offer an overall advantage to the recipient. The decision to proceed with LDLT should be made after a careful analysis of the recipient's risk-to-benefit ratio: the severity of the patient's liver failure, the patient's quality of life, and the expected waiting time for a deceased donor should be considered.
Donor safety is of paramount importance in LDLT, yet living donor complications and deaths occur even with the most experienced hands.4, 5 In 2006, Trotter et al.6 reported 13 living liver donor deaths believed to be definitely related to the donor surgery, and they estimated that the mortality rate for living liver donation is 0.15% to 0.20%. Subsequently, Ringe and Strong7 identified 33 living liver donor deaths, 21 of which were felt to be related to the operative procedure. They concluded that the incidence of operative mortality for living donor partial hepatectomy (approximately 14,000 procedures worldwide) ranges from 0.15% to 0.30% and may be as high as 0.5% when the right hemiliver is used for adult-to-adult LDLT. In the past year, 2 additional deaths of living liver donors (both right lobe donors) have occurred in the United States, and they have been documented by the media.8, 9 Because there is no central reporting agency for living donor complications, the incidence of catastrophic complications (death or the need for transplantation) is imprecise, and the true risk is likely underestimated because near-miss events that do not result in catastrophe are not captured anywhere.
It is well established that the morbidity associated with LDLT is directly related to the amount of liver tissue removed from the otherwise healthy donor.3, 10, 11 Overall, approximately 10% of left lobe donors3 and one-third of all right lobe donors4 experience LDLT-related morbidity (ranging from minimal complications to serious ones). Deaths resulting from either technical failures or early postoperative problems underscore the reality that living donor hepatectomy is associated with the small but real possibility of donor death; the risk may approach 0.5% for right lobe donors and approximately 0.1% for left lobe donors.3
If we accept that LDLT is indicated for patients with non-HCC liver disease, HCC should also be an indication for LDLT. The debate then revolves around the relative merits of LDLT outside the accepted criteria. The Milan criteria have been criticized for being arbitrary and perhaps too restrictive and for thus excluding many patients from a lifesaving procedure. A number of reports have indicated that the criteria for LDLT can be safely expanded without any increases in patient death or disease recurrence rates.12–15 These reports have demonstrated that incremental changes can be made in the acceptable parameters for the size and number of tumors without a risk of recurrence that is significantly different from the risk with the Milan criteria. In these circumstances, LDLT is a reasonable treatment option for patients with a modest extension of the Milan criteria who do not receive any priority (supplemental Model for End-Stage Liver Disease points) for DDLT.
The question is whether acceptable indications for LDLT in patients with HCC should be identical in those countries with minimal or no access to DDLT. Geographic disparities not only exist in countries with various levels of access to deceased donor organs but also have been shown to play important roles in countries in which deceased donor organs are relatively plentiful.16 Several studies17, 18 have demonstrated long-term survival (>50% 3 years after transplantation) for patients with HCC exceeding the Milan criteria and the University of California San Francisco (UCSF) criteria (1 nodule ≤ 6.5 cm in diameter or 2-3 nodules with individual diameters ≤ 4.5 cm and a total tumor diameter ≤ 8 cm). LDLT is an optional treatment offering a chance for a cure to patients with otherwise uncontrolled disease. However, the risk to the donor and the degree of benefit to the recipient needed to justify LDLT for advanced HCC are still undetermined. Currently, there is no consensus about an acceptable HCC recurrence risk or an acceptable donor risk. More importantly, there is no sense of what risk to the living donor might be considered acceptable for a given risk of recurrence. In fact, this is confounded by evidence suggesting that although the recurrence rates for HCC may be higher with LDLT versus DDLT, the overall survival rates with LDLT may be superior.19
Furthermore, it can be argued that the acceptable risks for donors and recipients may vary with the availability of deceased donor organs within a given region or country. HCC within the Milan criteria is an excellent indication for LDLT in areas in which no priority is given to patients with HCC or in areas in which DDLT is readily available but the waiting time exceeds 6 months. Some patients with HCC beyond the Milan criteria (but within the UCSF criteria) may also benefit from LDLT, even in areas in which DDLT is available because LDLT can decrease the waiting time. However, until extended criteria are adequately defined and universally accepted, indications for LDLT should be individualized.
Question 2. What Is Double Equipoise, and How Is It Affected by the Diagnosis of HCC?
As long as there is a disparity between the supply of deceased donor livers and the demand, the allocation of organs and the system for their distribution must be based on the ethical principles of utility, justice, and equity. Utility represents the beneficial outcome of the transplant, so an allocation policy based solely on utility would prioritize patients likely to have the best survival. Justice implies that all candidates have equal access to transplantation. Equity, which is often interchanged with justice, ensures that the organ allocation policy is applied equally and that similar recipients have equal priority. The ethical principles of justice and equity used for deceased donation do not apply to living donation because the recipient is not competing with other listed candidates for organs. In fact, LDLT recipients may benefit all candidates awaiting DDLT by eliminating themselves from the waiting list.
Miller20 eloquently described the ethical dimensions of equipoise for LDLT with a graphic depiction of triangles; this depiction incorporates a careful assessment of the recipient need, the donor risk (safety), and the possibility of a good outcome for the recipient. The tension between these fundamental principles results in triangles of differing shapes and proportions that vary with the circumstances of each donor and each recipient. The concept of double equipoise describes the balance between the recipient's survival benefit with or without LDLT and the probability of mortality for the donor.21 When conventional solid organ transplantation is being performed with a deceased donor organ, the determination of the risk-benefit analysis is restricted to the recipient. The risk-benefit analysis for living donation, however, is not confined to the donor and the recipient but also includes balancing the donor risk and the recipient benefit. This dual balancing test defines double equipoise.22 This balance should be explicitly defined and accepted by all parties, including the recipient, the donor, the surgical team, and society.
The concept of double equipoise suggests that there clearly exists an area of excessive donor risk and unacceptably low recipient benefit. As physicians, we feel that it is not ethically defensible for a living donor to undergo an operation with a mortality risk greater than the expected rate of 0.5% (as reported by centers around the world). Nor do we feel that it is acceptable for a donor to undertake any risk if the recipient benefit will predictably be very low (eg, if the recipient has extrahepatic disease or a multifocal tumor). These situations fall within the zone of ethical unacceptability (Fig. 1). However, we should also take into account the donor's view, which, although it is affected by emotional attachment to some extent, is just as important as our own. Thus, although we may judge the risk-to-benefit ratio to be inappropriate in some situations, it is our duty as doctors to advise the donor and the recipient of the inherent risks and potential benefits so that they can make a balanced decision. It is not for us to impose our own values; our task is merely to advise them and guide them to what all feel is the correct decision.
Conversely, there are situations of donor-recipient balance that appear to be ethically acceptable. An adult donor who is providing a liver segment to a pediatric recipient is one example that falls within the zone of ethical acceptability. In this situation, the donor risk is well defined and small, and the recipient benefit is almost always a highly successful and durable transplant.23–25 Although these 2 extremes appear easily definable and defensible, the zone of ethical uncertainty is the most complex area. What are the minimal benefits to a recipient that warrant the use of a living donor, and to what extent must the recipient benefit (eg, 50% survival at 2 years) to justify the use of a living donor for extended indications and marginal recipient benefits?
To explain the conundrum, it is useful to consider 2 specific cases. The first case is Mrs. Smith, who is a 45-year-old mother of 3 children and has HCC beyond the UCSF criteria, even after downsizing. According to double equipoise, we have a healthy donor with an average risk (<0.5%) and a recipient who is less likely to medically benefit because her tumor is outside the UCSF and Milan criteria. The extent of Mrs. Smith's medical benefit will be relatively low, and because of the donor risk, her situation falls within the ethically uncertain category. If we consider additional recipient benefits beyond medical benefits, we find that Mrs. Smith's 3 young children will benefit significantly from her survival, and so will her husband. The social benefits for the recipient and her family arising from LDLT, which will prolong Mrs. Smith's life, push her far enough to the right on the double equipoise graph that liver transplantation becomes ethically acceptable in her case (Fig. 2). Indeed, in comparison with DDLT, LDLT could be suggested to more likely benefit Mrs. Smith because even though she is likely to experience disease recurrence at some point, current evidence suggests that her overall survival time is likely to be longer.
The second case is Mr. Jones, who is a 70-year-old man with HCC beyond the UCSF criteria. The potential donor in this case is a loving 20-year-old granddaughter. Again referencing the double equipoise graph, we have a healthy donor with an average risk (<0.5%) and a recipient whose likelihood of medical benefit is relatively low; Mr. Jones falls within the ethically uncertain category. First of all, his age negatively affects the recipient benefit calculation. Being 70 years old, Mr. Jones has a higher risk of experiencing comorbidities if he undergoes LDLT. Additionally, the relationship between the living donor (a granddaughter) and the recipient is less close than the relationship between Mrs. Smith and her husband. Lastly, we lack longitudinal data about long-term consequences for living donors and cannot guarantee that the granddaughter will not experience health issues later in her life that can be attributed to her liver donation. In this case, the recipient benefit is reduced, but the donor risk is increased. Consequently, Mr. Jones ultimately falls within the ethically unacceptable category, even though he may ultimately live longer with LDLT (Fig. 3).
Question 3. Is a Paired Exchange Appropriate If One or Both Recipients Have HCC?
As stated previously, the concept of double equipoise describes the balance between the recipient's survival benefit with or without LDLT and the probability of mortality for the donor. This is once again altered in the more unusual setting of a paired exchange. Consider the following: Mr. Smith is keen to donate to his wife Mrs. Smith for the aforementioned reasons, but his blood group is B, and his wife's blood group is A. Mr. Jones' blood group is B, but his granddaughter's blood group is A. Is it acceptable for a paired donor exchange to take place in which Mr. Smith donates to Mr. Jones and Mr. Jones' granddaughter donates to Mrs. Smith? In this situation, there are 2 potential benefits: the obvious ABO compatibility issue and the improved size matching. One reason for not going ahead with the original transplant for Mr. Jones is that his granddaughter would potentially be providing a small-for-size graft, but Mrs. Smith is smaller, so this risk would be reduced. In this scenario, both Mrs. Smith and Mr. Jones have a higher chance of benefiting from LDLT, and the paired exchange moves into the ethically acceptable category (Fig. 4).
SUMMARY AND RECOMMENDATIONS
1HCC within the Milan criteria is an appropriate indication for LDLT.
2Unlike DDLT, LDLT is governed by utility concerns that we describe as double equipoise; it is not governed by equity and justice considerations. One could argue that by reducing the number of recipients on the deceased donor waiting list, living liver donors potentially help each person remaining on the waiting list.
3Because LDLT is governed by utility concerns and not by justice and equity concerns, it can be supported for patients beyond the UCSF criteria: when a double equipoise decision is being made for the donor and the recipient, it is acceptable to include not only mortality and morbidity data but also quality of life, psychological, and social considerations related to the 2 parties.