Liver Transplant for Hepatocellular Carcinoma in the United States

Evolving Trends Over the Last Three Decades

Marc Puigvehí; Dana Hashim; Philipp K. Haber; Amreen Dinani; Thomas D. Schiano; Amon Asgharpour; Tatyana Kushner; Gaurav Kakked; Parissa Tabrizian; Myron Schwartz; Ahmet Gurakar; Douglas Dieterich; Paolo Boffetta; Scott L. Friedman; Josep M. Llovet; Behnam Saberi


American Journal of Transplantation. 2020;20(1):220-230. 

In This Article


In this study, we retrospectively evaluated the evolving trends of HCC and the underlying diagnosis of liver disease in deceased donor LT recipients over the last three decades (1987 to September 2017) using the SRTR database. In addition to including patients with the diagnosis of HCC, we included recipients whose providers had sought HCC MELD exception points and were approved, even if the diagnosis of HCC had not been entered. This allowed us to identify cases with HCC as an indication for LT more broadly. Therefore, our reported HCC cases/per year are higher than those noted in previous publications.[17] Our study shows a trend toward a decrease in HCV-related LT and a parallel increase in NAFLD-related LT following the implementation of DAAs in clinical practice. Moreover, we show a significant improvement in HCV patients' survival in the DAA era, being now comparable to NAFLD patients, whereas survival for the latter group remains unchanged.

Our study shows an increase in HCC-related LT, particularly in the MELD era. The trends and peaks of HCC over time are reflective of changes in the way patients with HCC are prioritized for LT. Following the implementation of MELD exception points in 2002, HCC has grown as an indication for LT and accounts for 21% of the total number of deceased donor LTs over a 30-year period. As a result of a series of analyses indicating that the MELD exception scores advantaged HCC patients, the system was modified in 2003, 2004, and 2005 to reduce the priority accorded to these patients.[6,18,19] Despite these modifications, we show that the rate of LT for HCC continued to rise.

In 2013 "Share 35" policy was implemented with the goal of allowing an increased proportion of patients with a MELD > 35 to undergo LT and thus decreasing death on the waiting list. A study evaluating the effect of "Share 35" on patients who underwent LT for HCC demonstrated no change in the proportion of LT performed or overall waiting time. However, a higher rate of death/delisting was observed.[20] Similarly, in our study, we did not observe any significant change in the rate of LT for HCC in the Share 35 era, although our study did not specifically evaluate United Network for Organ Sharing regional differences. Only after the most recent modification in 2015 in which HCC patients received no priority for 6 months and the MELD exception score was capped at 34 points[6,7] has the rate of LT for HCC and proportion of HCC candidates undergoing LT declined (Figure 1). The most recent MELD policy change awards exception points equal to median MELD score of a DSA (donation service area) region minus 3 using a calculation based on a 250 nautical mile (NM) circle around each donor hospital that is recalculated every 180 days.[21] Although this policy only was recently implemented, it is speculated that it may decrease the rate of LT and increase the dropout rate for HCC candidates.

Our study confirms that HCV has been the leading etiology for HCC as an indication for LT over the last 25 years, accounting for almost half of the cases. Even during the current IFN-free DAA era in which HCV is routinely cured, HCV remains the predominant etiology of liver disease in HCC patients, although there is a downward trend. It is worth noting that the SRTR database cannot distinguish active from cured-HCV in the LT candidates or recipients. Model-based simulation studies have predicted that HCC will continue to increase over the next decade.[22] In addition, in some liver transplant centers patients with HCV and HCC are treated with DAA after LT to increase their chance of receiving a HCV positive organ and decreasing the waiting list time.[10] Although achieving SVR decreases the risk of all-cause and liver-related mortality, the risk of developing HCC persists, more so in those with cirrhosis, in which the annual incidence of HCC in post-SVR patients is 1.82/100 person per year in patients with cirrhosis compared to 0.34/100 person per year in those without cirrhosis.[23] Although early reports suggested that DAA treatment may lead to increased risk of cancer occurrence/recurrence, that concern has proven unfounded.[24,25] From a public health perspective, it will be several years before there will be a decline in the rates of HCC secondary to HCV.[26]

Based on the analysis herein presented, NAFLD-related HCC, which was first reported as a diagnosis in the SRTR database in 2003, is shown to be the most rising etiology of liver disease in HCC patients undergoing LT.[27] It is estimated that the incidence of HCC secondary to NAFLD will increase by 137% by 2030.[28] Unlike other groups analyzing the SRTR database, we did not include patients with cryptogenic cirrhosis or unknown diagnosis with DM or elevated BMI > 30 in the NAFLD group.[17,29] This is of critical importance, as our analysis shows that up to 50% of HCC-related LT patients have underlying ascites, which falsely raises the BMI. Of note, cryptogenic cirrhosis accounted for less than 3% of cases.

As no patients with NAFLD codification were transplanted until 2003, we performed our survival analysis in the MELD era (2002–2017). In our study, HCV was a main determinant of death in the pre-DAA era (2002–2013) but was not associated with decreased survival in the DAA era (2014–2017), confirming that HCV widespread cure has significantly improved the prognosis of HCV patients undergoing LT for HCC.[30] During the 2002–2013 period, HBV patients showed the best survival, and HCV patients had an impaired survival compared to NAFLD patients. Even though diabetes was far more prevalent in the NAFLD population, and having diabetes was a strong predictor of mortality, survival was significantly worse for HCV-infected patients. However, this strong effect of HCV infection on survival disappeared in 2014, in concert with the rise of DAAs, and no differences in survival have been noted between HCV and NAFLD patients since then. Even though the follow-up is still short and median survivals have not been reached, our study is, to the best of our knowledge, the first to show the changing trends on etiology and their impact on survival in HCC-related LT in the United States in the recent years. Of note, the causes of death were different among HCV and NAFLD patients, and as expected, more related to cardiovascular events in the latter. The higher rate of graft-related death among HCV patients may be explained by differences in recipient and donor characteristics or, plausibly, be related to post-LT HCV recurrence and related graft failure before the DAA era. Remarkably, graft-related deaths were comparable in patients with HBV and HCV.

There are several limitations to our study. Due to the nature of the database, the determination of the underlying etiology of chronic liver disease, by the primary and secondary diagnoses, is based on how the diagnosis codes were entered into the database. Therefore, the HCC cases could be under or overreported. Data regarding HCV-RNA are not available in the database. Therefore it is unclear whether HCV recipients were post-SVR or had active HCV. Furthermore, a detailed history regarding the amount of alcohol use is not available. Besides, there are missing data, specifically prior to 2002, and cases with unknown etiology within the database. Patients with HCC with unknown etiology were noted to have the highest mortality compared to other groups. However, in recent years, these cases accounted for only 5% of the total.

In summary, changes in the MELD exception policy have over time led to a decrease in the proportion of LT for HCC candidates after an initial significant increase with the adoption of the MELD score for organ allocation. Although HCV remains the most common etiology of HCC-related LT, the availability of DAA is decreasing its burden. Conversely, there is in the same timeframe a steady increase in patients undergoing LT with NAFLD-related HCC. Whereas in the pre DAA era HCV infection was one of the strongest determinants of death in the HCC-related LT population, NAFLD and HCV patients have similar survival in the DAA era, and HCV is no longer an independent predictor of an adverse outcome. The rate of death for cardiovascular disease is higher in NAFLD patients, whereas the rate of graft-related death is higher among HCV individuals. Further studies in the next years will be of high importance in order to confirm these changing trends.