Should the liver transplant criteria for hepatocellular carcinoma be different for deceased donation and living donation?


  • Potential conflict of interest: Nothing to report.

Should the liver transplant criteria for hepatocellular carcinoma (HCC) be different for deceased donation and living donation? Living donor liver transplantation (LDLT) for HCC offers the opportunity to provide a neoadjuvant treatment organized around a scheduled time for the transplant and to restore excellent liver function with a high-quality graft. However, some centers have been reluctant to offer LDLT for HCC because of concerns about higher rates of cancer recurrence and inferior patient survival after LDLT versus deceased donor liver transplantation (DDLT).1-6 Theoretical reasons for concerns about the potential for higher rates of HCC recurrence after LDLT include the following: (1) the stimulation of residual cancer cells by growth factors in the regenerating liver7; (2) the relatively brief waiting time for LDLT, which may allow transplantation for patients whose aggressive or rapidly progressive HCC tumors might progress with a longer waiting time and exclude them from transplantation8; and (3) more limited oncological clearance with the inferior vena cava–sparing technique used for LDLT. Another explanation for the reports of higher recurrence rates after LDLT is the presence of programmatic biases that lead these centers to unknowingly offer LDLT to patients with a higher risk of HCC recurrence.

To determine whether the liver transplant criteria for HCC should be the same or different for deceased donation and living donation, we examine 2 questions:

  • 1Are recurrence or cancer-free survival rates after DDLT and LDLT for HCC so inherently different that different transplant criteria are justified?
  • 2Should LDLT be offered to patients whose tumor stage prevents the use of allocation exception points [eg, Model for End-Stage Liver Disease (MELD) points] and, therefore, limits the availability of deceased donor livers?

These questions are analyzed in this article.


A2ALL, Adult-to-Adult Living Donor Liver Transplantation Cohort Study; CEBM, Centre for Evidence-Based Medicine; DDLT, deceased donor liver transplantation; HCC, hepatocellular carcinoma; LDLT, living donor liver transplantation; MELD, Model for End-Stage Liver Disease; UCSF, University of California San Francisco; UNOS, United Network for Organ Sharing.


We undertook literature searches in June 2010 to identify English publications solely focused on comparisons of outcomes of LDLT and DDLT for HCC. Embase was searched with the following sets of keywords: (1) living, liver, donation, transplantation, hepatocellular, and cancer and (2) living, liver, donation, and ethics. Two hundred one references were identified.

PubMed was searched with the following search terms: hepatocellular, cancer, and transplantation. Another 1590 references were obtained.

The Cochrane Collection was searched with the following terms: hepatobiliary, hepatocellular, transplantation, and donation. No relevant articles were identified.

The obtained information was supplemented with additional references provided by committee members.

The quality of the evidence was evaluated and recommendations were made with the classification system developed by the Oxford Centre for Evidence-Based Medicine (CEBM).


Question 1. Are Recurrence or Cancer-Free Survival Rates After DDLT and LDLT for HCC So Inherently Different That Different Transplant Criteria Are Justified?

Literature Review

Six published articles comparing LDLT and DDLT for HCC with CEBM level 1b evidence were identified,1-6 and they are listed in Table 1.

Table 1. Three-Year HCC Recurrence Rates and Survival Data
AuthorsPatients (n)Overall Recurrence Rate (%)Overall Recurrence- Free Survival Rate (%)Overall Survival Rate (%)Recurrence Rate for Patients Within the Milan Criteria (%)Overall Survival Rate for Patients Within the Milan Criteria (%)Recurrence Rate for Patients Beyond the Milan Criteria (%)Overall Survival Rate for Patients Beyond the Milan Criteria (%)
  • n/a, not available.

  • *

    P < 0.05.

Kulik and Abecassis1 (A2ALL data)LDLT: 6327n/an/an/an/an/an/a
Hwang et al.2LDLT: 237n/a8073.2*n/a91.4n/a62.6
DDLT: 75 8061.1* 89.9 66.4
Fisher et al.3 (A2ALL data)LDLT: 5829*586726.3n/a34n/a
DDLT: 340*62630 0 
Lo et al.4LDLT: 4329*n/a8015.6n/a45n/a
DDLT: 170* 940 0 
Di Sandro et al.5LDLT: 25495.577.3n/an/an/an/a
DDLT: 15410.590.582.8    
Vakili et al.6LDLT: 2828.6*n/a80n/a87.1n/a80
DDLT: 7412.1* 70    

Quality of the Evidence (CEBM Level 2b)

All the studies included relatively small numbers of patients, although 1 study examined a consortium of multiple LDLT centers in the United States. The outcomes were not uniform. All reported slightly different endpoints. Most documented relatively brief periods of follow-up. The patient selection criteria were often not described in detail.

Question 2. Should LDLT Be Offered to Patients Whose Tumor Stage Prevents the Use of Allocation Exception Points (eg, MELD Points) and, Therefore, Limits the Availability of Deceased Donor Livers?

Literature Review and Quality of Evidence

The available literature includes only qualitative analyses and CEBM level 2b case-control series.


Question 1. Are Recurrence or Cancer-Free Survival Rates After DDLT and LDLT for HCC So Inherently Different That Different Transplant Criteria Are Justified?

Six published reports focusing on comparisons of the outcomes of DDLT and LDLT for HCC were identified and analyzed so that we could determine whether LDLT and DDLT for HCC inherently have prognoses sufficiently different to merit the use of different selection criteria.

Data from the Adult-to-Adult Living Donor Liver Transplantation Cohort Study (A2ALL), which was funded by the National Institutes of Health, were first reported by Kulik and Abecassis in 20041 and were later reported in more detail by Fisher et al.3 Kulik and Abecassis examined a limited cohort of patients undergoing transplantation for HCC at the Northwestern University Medical Center, and they reported a higher recurrence rate (stage for stage) in recipients whose transplants were accelerated (fast-tracked) by LDLT, especially in the pre-MELD era when patients with HCC were disadvantaged by the allocation algorithm. In the 2007 A2ALL follow-up study by Fisher et al. (see Table 1), the 3-year HCC recurrence rate was 29% for LDLT patients and 0% for DDLT patients (P = 0.002). However, because of differences between the 2 groups with respect to non–cancer-related deaths, there were no significant differences in the recurrence-free rates (58% versus 62%) or in the overall survival rates (67% versus 63%). The rate of recurrence reported by Fisher et al. for DDLT recipients was exceptionally low; this may reflect the fact that the waiting time was longer for DDLT recipients (with a higher dropout rate for non-LDLT recipients) and thus may constitute a patient selection bias.

In Hwang et al.'s study of the Seoul experience,2 the 3-year overall survival rate was greater for LDLT patients versus DDLT patients: 73.2% after LDLT and 61.1% after DDLT (P = 0.043). The 3-year HCC-free recurrence rate was 80% in both groups.

In Lo et al.'s study of the Hong Kong experience,4 the 3-year recurrence rates were 29% for LDLT patients and 0% for DDLT patients (P = 0.029), and they were identical to the rates of the A2ALL cohort. Like the overall survival rates in the A2ALL study,3 the overall survival rates in this study were not significantly different (80% versus 94%, P = 0.187). The rate of recurrence in DDLT recipients was also exceptionally low and may reflect a patient selection bias.

In Di Sandro et al.'s study of the Milan experience,5 the overall 3-year recurrence rates were 4% and 10.5% for LDLT and DDLT patients, respectively. The survival rates were 77.3% and 82.8% for LDLT and DDLT patients, respectively. No significant differences were observed.

In Vakili et al.'s study of the Lahey Clinic experience,6 there was a significant difference in the HCC recurrence rates (28.6% for LDLT and 12.1% for DDLT, P < 0.05), but the overall survival rates (80% versus 70%) were similar.

The different endpoints used in these studies are shown in Table 1. None of the studies provided detailed reports and statistical analyses of the different types of pretransplant treatments that their patients received or the effects of these treatments on posttransplant outcomes. The recurrence rate was significantly higher after LDLT in 3 studies,3, 4, 6 and in each of these studies, the increased rate of recurrence was noted stage for stage. The recurrence rate was not specifically provided by Hwang et al.,2 but the recurrence-free survival rates were equal for the 2 groups. Therefore, despite the higher recurrence rates in 3 studies, the overall survival rates after LDLT for HCC versus DDLT for HCC were not inferior in all the studies. We could anticipate that this difference in recurrence rates might eventually translate into lower long-term survival for the LDLT groups; however, this pessimistic perspective is challenged by Hwang et al.,2 who reported a higher overall survival rate for LDLT recipients.

Is LDLT itself to be blamed for the higher incidence of recurrence in 3 of the studies? The evolving evidence suggests that tumor characteristics are more important determinants of recurrence and survival rates than the graft type. In multivariate analyses in the published articles, the graft type and the waiting time were not significant risk factors for HCC recurrence after transplantation; instead, salvage transplantation, center experience, and tumor characteristics were significant factors (Table 2). These analyses are reinforced by (1) an abstract presented at the 2009 American Transplant Congress by the Toronto group, which reported no differences in the HCC recurrence rates with DDLT and LDLT (unpublished data, 2009), and (2) an updated analysis of the A2ALL cohorts presented at the 2011 meeting of the American Association for the Study of Liver Diseases, which included the prospective A2ALL cohort data for the MELD era and concluded that “differences in tumor characteristics and management of HCC in patients who received LDLT (rather than graft type) accounted for the higher HCC recurrence rates observed in their LDLT group.” In a multivariate analysis of the latter prospective cohort of 49 patients undergoing LDLT for HCC and 73 patients undergoing DDLT for HCC (including 35 patients listed for HCC who fell off the transplant waiting list), the graft type did not predict survival (CEBM level 1b).

Table 2. Significant Risk Factors for HCC Recurrence According to Multivariate Analyses
AuthorsRisk Factors
Hwang et al.2Tumor > 5 cm
Poor differentiation grade
Gross vascular invasion
Fisher et al.3 (A2ALL data)Transplant era
Transplant center experience
Alpha-fetoprotein level
Lo et al.4Salvage transplantation
Tumor status beyond the UCSF criteria

Is there an inherent advantage or disadvantage in the ability to offer LDLT more quickly than DDLT? Again, the limited available evidence suggests that the answer is no. First, as we have already discussed, the waiting time was not identified as a significant risk factor for HCC recurrence in the aforementioned studies (Table 2). Second, a recent study by the San Francisco group9 showed that among 100 patients with T2 HCC (21%-23% were understaged according to the final explant pathological examination) the HCC recurrence–free survival rates were equal for DDLT patients (90) and LDLT patients (10), regardless of the waiting time (<3 or >6 months; CEBM level 2a).

The outcomes in the series comparing LDLT and DDLT may be influenced by selection biases based on regional organ availability. For instance, the Northwestern University, A2ALL, and Lahey Clinic experiences may reflect organ availability in the United States and, more specifically, longer waiting times for DDLT recipients and thus higher dropout rates for patients with more aggressive HCC. Because LDLT is offered on a faster track than DDLT, it is conceivable that many LDLT recipients do not wait long enough for the biological behavior of their HCC to be determined. In contrast, patients who wait for DDLT and have biologically aggressive HCC are likely to suffer from disease progression while they are waiting and, therefore, drop from the waiting list; thus, only patients with less aggressive HCC are left to receive DDLT. In addition, in the Western world, LDLT is often pursued for patients who do not meet the stringent criteria for MELD exception points (the Milan criteria) and have tumors with an inherently worse prognosis. Thus, the answer to this riddle may lie in the point of reference. If the reference point is the time of diagnosis, LDLT may offer the advantage of a shorter waiting time and, therefore, a lower dropout rate, but the recurrence rate after LDLT may be higher. In contrast, if the reference point is switched to the time of transplantation, an artificially lower recurrence rate may result for DDLT recipients because some DDLT candidates may preferentially be withdrawn from the waiting list before transplantation. This paradigm, of course, would be affected by local, regional, and national allocation policies and by the availability of deceased donor livers.

How do these issues affect the choice of DDLT or LDLT for HCC? That depends on the perspective. The recipient is focused on the events before transplantation. From the recipient's perspective, transplantation is the preferred option because it is the only choice offering a chance for a cancer cure. From this perspective, undergoing transplantation as soon as possible is preferable to avoid disqualification for transplantation due to tumor progression. After transplantation, the risk of death due to cancer recurrence for DDLT patients is likely similar to the risk for LDLT patients, regardless of the tumor biology (good or bad). From the donor's perspective, the recipient's potential longevity with a transplant is a key issue. The donor wants the recipient to live, but he also wants to ensure that the risk of donation is justified by the recipient's longevity. Thus, from this viewpoint, treating recipients with an ablate and wait strategy8 is appealing because it improves the probability of survival benefits after transplantation through the identification of recipients with a favorable tumor biology. From the center's perspective, there are competing interests. On the one hand, the center wants to cure someone with cancer and cover the fixed operating costs. On the other hand, the center wants to minimize the risk of poor survival rates after liver transplantation due to high rates of tumor recurrence.

In conclusion, clinicians should help patients who have the option of either DDLT or LDLT for HCC to understand the uncertainty about the cancer recurrence risk associated with these 2 clinical options (a level 1 recommendation). Although there are no level 1a data (eg, randomized trials) for recurrence rates or cancer-free survival rates to guide the choice between DDLT and LDLT for HCC, the higher recurrence rates reported in the Hong Kong, A2ALL, and Lahey Clinic series cannot be ignored. Because of (1) differences in the allocation policies for deceased donor livers and in the availability of deceased donor livers, (2) local, regional, and national differences in the potential waiting times for DDLT, and (3) national legal requirements, patients with HCC and their donors should be offered LDLT with the following understanding: estimates of the waiting time for DDLT and the potential for dropout due to HCC progression need to be weighed against the potential for a lower dropout rate and a higher recurrence rate after LDLT. However, because the existing literature is far from definitive and the outcomes are far from uniform, patients should be informed that a true risk-versus-benefit balance for DDLT and LDLT for HCC cannot be fully and accurately assessed at this time. Nonetheless, the weight of the evolving evidence suggests that the tumor biology is likely the key factor driving HCC recurrence after DDLT or LDLT; the graft type and the waiting time are less likely to be important independent risk factors. In the future, understanding the biological behavior of HCC in individual patients may lead to more reliable measures of the risk of HCC recurrence.

Question 2. Should LDLT Be Offered to Patients Whose Tumor Stage Prevents the Use of Allocation Exception Points (eg, MELD Points) and, Therefore, Limits the Availability of Deceased Donor Livers?

The following discussion is predicated on the following assumptions: (1) neither DDLT nor LDLT should be performed when the results are deemed to be unacceptably poor, and (2) allocating scarce deceased donor livers to high-risk recipients disadvantages other patients on the waiting list with much better prognoses.10 Deceased donor availability, specifically for recipients with HCC tumors within or beyond the Milan criteria, varies widely. Within the United States, for example, exception points may be allocated differentially between United Network for Organ Sharing (UNOS) regions for patients who exceed the UNOS criteria (modified Milan criteria) but fall within the University of California San Francisco (UCSF) criteria or who have been down-staged. Thus, the decision in the United States to not allocate additional MELD points to patients with tumors beyond the UNOS criteria is based on the ethical principle of trying to achieve equipoise between utility, equity, and justice.

Societal concerns about utility are less relevant when there is a liver transplant recipient with advanced liver cancer who may never get the chance for DDLT because he or she does not qualify for extra MELD points, who faces death without a transplant, and who has a well-informed, low-risk donor who wishes to voluntarily donate even if the chance for a cure is low. Thus, in Asia, many LDLT programs accept patients with HCC beyond the Milan criteria for LDLT because they view a living donor organ as a private gift and the reported 3-year survival rates for such patients are higher than 50%.11-14 In North America, it is also considered ethically acceptable to use LDLT and DDLT for the treatment of patients with HCC tumors beyond the Milan criteria, particularly when the tumors have been successfully down-staged before transplantation15 or the recipients do not have unfavorable biological markers such as an alpha-fetoprotein level greater than >1000 ng/mL.16 The prognosis for patients with HCC with highly unfavorable factors (eg, macroscopic vascular invasion) is very poor, so these patients are poor candidates for either DDLT or LDLT outside clinical trials.11, 17, 18

Are there reasons to be cautious about preferentially or only offering LDLT for HCC? Yes, there are. Although transplant professionals are acting as stewards of a scarce resource in the case of DDLT, they are acting as moral agents in the case of LDLT. Health care providers have an overarching responsibility to protect living donors from harm. If extended criteria are adopted for LDLT only for HCC worldwide, many lives will be saved, but many donors will also be put at risk when the benefit for the recipients may not warrant the risk. The donor mortality rate is approximately 0.25% to 0.5%; the 5-year survival rate after LDLT (or DDLT) for HCC is dependent on several factors, including the tumor-node-metastasis stage.18 Although the published 3-year survival rates for LDLT recipients with HCC are approximately 80%, the decision to offer only LDLT for advanced HCC means at best the potential sacrifice of 1 donor for every 160 to 320 recipients saved. However, this ratio will worsen significantly with more advanced HCC and, therefore, lower 3-year survival rates if a constant donor mortality rate is assumed.

Arguments in favor of preferentially offering LDLT when legitimate societal concerns about the scarcity of deceased donors may warrant not offering DDLT include (1) potential benefits to the recipients (beneficence), (2) respect for people's right to make choices, and (3) donor autonomy (confirmed by the use of a very robust informed-consent process).19 Arguments against different criteria for LDLT and DDLT include the following: (1) the requirement for the health care team to respect nonmaleficence by overriding a patient's request and refusing to participate in a transplant if the team truly believes that the recipient survival rate is too low to justify the potential risk to the donor, (2) paternalism (the argument that doctors are imposing their values on other people who should have the freedom to make decisions about their own lives), (3) the high potential for donor coercion when LDLT is offered as the only lifesaving treatment, (4) the potential need to use deceased donor livers to salvage LDLT patients who develop arterial thrombosis, and (5) the risk of donor dissatisfaction when a high-risk cancer patient dies of recurrent cancer after LDLT.20 The lack of high-quality, longitudinal, long-term data for both physical and psychosocial outcomes of living liver donation makes it difficult to evaluate the importance of these concerns with particular relevance to the risks to the donor.

Can the competing viewpoints about the potential benefits and risks of preferentially offering LDLT be reconciled? A possible solution is to accept that the ethical conditions for LDLT are at least partly determined by locoregional factors, which include but are not limited to (1) the availability of organs from deceased donors, (2) the skill level and expertise of the donor center, (3) the rigor and independence of the donor assessment process, and (4) the willingness of the center to prospectively monitor and report clinical outcomes. Although this practical solution reflects the reality that life is often unfair, accepting such variability is highly disturbing in terms of justice.

In conclusion, the scientific evidence is lacking, poor in quality, or conflicting, so we cannot accurately assess the risk-benefit balance for donors when LDLT is offered preferentially for HCC. Nonetheless, there are conceptual reasons for concern about coercion and potential harm to the donor when only LDLT is offered for HCC (and especially advanced HCC). Thus, centers preferentially offering LDLT for HCC should prospectively monitor and share data for both donor and recipient outcomes and balance their responsibility to protect the living donor with their respect for autonomy (a level 1 recommendation).


Should the liver transplant criteria for HCC be different for DDLT and LDLT? We have found that there are no high-quality data justifying or refuting the use of different criteria.

According to locoregional circumstances and the availability of livers from deceased donors, it may sometimes be reasonable to use different acceptance criteria for DDLT and LDLT. For instance, in the United States, a patient with a single unresectable HCC tumor ≥ 5.0 cm in diameter or a tumor or tumors within the UCSF criteria would not meet the criteria for MELD exception points in most UNOS regions and, therefore, would not have access to DDLT; in this setting, LDLT may offer the only chance of a cure and provide excellent outcomes.6, 21 In many regions, deceased donation rates are poor or zero. In these areas, along with the philosophy of the living donor liver as a private gift, the option of LDLT should be considered in the context of paternalism and respect for the autonomy of the living donor, but this is a difficult balance to achieve.