Hepatocellular carcinoma and model for end-stage liver disease exceptions: The more we understand, the more challenging the allocation gets


  • Patrick G. Northup,

    Corresponding author
    1. Division of Gastroenterology and Hepatology, University of Virginia, Charlottesville, VA
    • Division of Gastroenterology and Hepatology, University of Virginia, P.O. Box 800708, Charlottesville, VA 22908
    Search for more papers by this author
    • Telephone: 434-243-2718; FAX: 434-244-7529

  • Carl L. Berg

    1. Division of Gastroenterology and Hepatology, University of Virginia, Charlottesville, VA
    Search for more papers by this author

  • See Article on Page 434


HCC, hepatocellular carcinoma; MELD, Model for End-Stage Liver Disease; OPTN, Organ Procurement and Transplantation Network.

Over the past 2 decades, hepatocellular carcinoma (HCC) has progressed from being a near contraindication to liver transplantation to occupying its current position as one of the most successful principle indications for liver transplantation. Approximately 4% of all patients on the liver transplant waiting list at any one time have an exception to the Model for End-Stage Liver Disease (MELD) score related to HCC.1 The advent of the MELD allocation system in February 2002 gave priority to patients with HCC, and this resulted in many more candidates with HCC successfully undergoing liver transplantation. Before the MELD era, 4.6% of all transplants were performed for candidates with HCC, whereas in 2002-2007, that number increased to 26%.2 The advent of liver transplantation for HCC has transformed a disease with a nearly universally fatal course into a disease with outcomes comparable or superior to those of many solid organ malignancies routinely treated with other medical or surgical therapies. Despite this success, questions about liver transplantation's role in the treatment of HCC persist. Much of the controversy arises because of the donor organ shortage in the United States and most other countries performing large volumes of liver transplants. Because the number of donor organs is limited, the increase in the number of candidates with HCC who successfully undergo transplantation has resulted in a corresponding decrease in the number of recipients with all other liver diseases. This raises a question within the transplant system: are patients with HCC provided unfair priority in the transplant system in comparison with all other patients awaiting transplantation?

The article by Goldberg et al.3 in this issue of Liver Transplantation adds strength to the argument that the sickest first policy may not be well served by our current allocation methods for HCC under the MELD system. The authors analyzed the US Organ Procurement and Transplantation Network (OPTN) database for adult liver transplant candidates who were listed for liver transplantation in the modern era of MELD exception points for T2 HCC tumors within the Milan criteria (2005-2009). Two cohorts were analyzed: (1) all patients listed with an HCC exception and (2) all other candidates without HCC or another specific MELD exception. The cohorts were compared for rates of wait-list dropout or death within 90 days of listing and for other demographics and mortality risk factors. There were some expected differences between the HCC group and the non-HCC group with respect to disease etiology, regional organ utilization, and demographics. Most significantly, viral hepatitis was the predominant etiology in the HCC group. However, the analysis comparing 90-day wait-list mortality (including removal due to clinical deterioration or tumor progression) yielded unexpected results. Only 4.2% of the HCC candidates receiving 22 MELD points were removed from the list because of death or disease progression, whereas for the non-HCC patients with a physiological MELD score of 22, the rate was 11.0%. In other words, in comparison with non-HCC patients with equivalent true MELD scores, patients with HCC exception points were 2.62 times less likely to die waiting. This difference in mortality became even greater when the authors compared the same cohorts as the MELD upgrades were applied to the HCC cohort over the accrued waiting time. In fact, there appeared to be no significant increase in mortality in the HCC cohort as the HCC patients were awarded the extra 3 MELD points every 3 months (as dictated by the current allocation policy). This mortality protection over time was not seen in the non-HCC cohort.

Goldberg et al.3 concluded that the OPTN policy giving 22 MELD points to HCC patients greatly overestimates the originally estimated 90-day mortality or wait-list dropout rate of 15%. Furthermore, the incremental increases awarded to HCC candidates do increase the chances of transplantation, but in reality, there is no perceptible increase in 90-day mortality or dropout in this cohort as they traverse the 3-month intervals on the waiting list. This is in direct contrast to non-HCC candidates, who suffer increasing death rates as their physiological MELD scores increase over time on the waiting list. The authors could not fully explain the lack of increasing mortality in the HCC cohort over time, but they speculated that both locoregional therapies and the self-selection of a favorable tumor biology play roles in this phenomenon.

Recognizing the valuable contributions that Goldberg et al.3 have made with respect to wait-list dropout and death for T2 HCC candidates within the Milan criteria in the United States, the transplant community must consider what steps to take in light of these new data in order to maintain an equitable allocation policy for all liver transplant candidates. This discussion can be held in the context of 2 possible paradigms: (1) the current liver allocation paradigm, which is a response to the final rule4 and is supported by recommendations from the Institute of Medicine5 and in which medical urgency is used as the primary determinant for the candidate listing status, or (2) a broader view in which both pretransplant mortality and posttransplant mortality are considered and maximized. We will defer the discussion of the latter scenario for another forum and focus on the consequences of changing the allotment of points to transplant candidates with HCC as a result of Goldberg et al.'s findings.

Goldberg et al.'s findings3 quite clearly demonstrate that the risk of death and wait-list dropout are overestimated both with the initial allocation of 22 MELD points and with the incremental upgrades to 25 and 28 points. With these objective data in hand, we can contemplate a few strong arguments against dropping the number of HCC exceptions awarded to T2 HCC wait-listed candidates at the time of listing as well as abolishing or reducing the magnitude of points awarded during sequential upgrades to transplant candidates with HCC. A strong precedent for such a change already exists in the MELD-based allocation era, with the OPTN implementing a reduction in HCC points for T2 HCC candidates within the Milan criteria in 2003 and again in 2005. One must think broadly, however, about the intended and unintended consequences of further changes in the assignment of MELD points for HCC candidates in the current allocation system. Although the most apparent impact of a reduction in HCC MELD points would be decreased priority for HCC candidates and increased availability of organs for candidates with biochemical MELD scores in the range of 22 to 28, one must recognize that the current allocation system is more complex and involves more than just the allocation of points to HCC and non-HCC liver transplant candidates. Furthermore, because the number of candidates with HCC MELD exceptions is high (more than 6000 candidates during the 4 years of the current analysis), the impact on the rest of the liver transplant waiting list would be considerable.4

OPTN policy assigns 22 MELD points to liver transplant candidates with hepatopulmonary syndrome, cholangiocarcinoma, cystic fibrosis, familial amyloidotic polyneuropathy, or portopulmonary syndrome. Thus, any change in the availability of organs that might be associated with HCC patients receiving lower MELD scores would influence the wait-list mortality for transplant candidates with MELD scores in the 22+ range. Although one could imagine that there would be increased organ availability for all of these wait-listed candidates, one needs to be certain that the MELD points assigned for these recognized exceptional diagnoses do not then subsequently overestimate the actual wait-list mortality and result in the development of the same systematic problem for these exception cases that Goldberg et al.3 have identified for HCC candidates. The consequences become even more complex when one takes into account non-HCC, nonstandard MELD exceptions, as outlined by Argo et al.6 and Massie et al.7

Thus, as we digest the messages in the data reported in the current issue of Liver Transplantation by Goldberg et al.3 and in the data previously reported by Washburn et al.,8 the transplant community must consider the impact of HCC MELD point allocation on the liver allocation system as a whole and not in isolation. By managing the allocation system in its totality rather than as isolated pieces, one can attempt to maximize the likelihood that the outcomes desired by the final rule will be achieved; that is, patients will be ranked on the waiting list according to the desire to minimize wait-list mortality. Grappling with the entire liver allocation system whenever another new analysis is reported is challenging, but the desire for fairness and for the equitable allocation of organs to all liver transplant candidates is paramount in our allocation paradigm. We must continue to strive to maintain an organ allocation system that is nimble and responsive to new information and that allows changes in allocation to occur in a timely manner when new data are developed. In an ideal world, the allocation system would be built upon complex models that take into account the impact of policy changes on all types of wait-list candidates and provide verification that the desired outcomes are in fact accomplished. Moreover, one could aspire to achieving this feedback in a more timely manner than Washburn et al. and Goldberg et al., who needed 5 years to identify the outcomes of HCC candidates on the wait list.