Transplantation of the Liver

Assessment of Disease Severity and Waitlist Stratification

In 1964 C. Garner Child and Jeremiah Turcotte of the University of Michigan proposed a scoring system that stratified the severity of end-stage liver disease. Their scoring system, the first of its kind, relied on five clinical parameters: total bilirubin, serum albumin, nutritional status, extent of ascites, and degree of hepatic encephalopathy. The Child-Turcotte scoring system was developed to preoperatively standardize the severity of chronic liver disease using a reliable set of clinical criteria, which allowed for the prediction of postoperative outcome. The scoring system was modified by Pugh in 1972, to allow for consideration of the bleeding tendency in patients with portal hypertension. The Child-Turcotte-Pugh (CTP) score replaced the criterion of nutritional status with prothrombin time. Lacking any other standardized method, the CTP score, combined with waiting time on the transplant list, was used to stratify liver transplant candidates for organ allocation. However, the CTP score required the subjective assessment of ascites and severity of encephalopathy. Dissatisfaction with the subjective components of the CTP score helped prompt the development of a completely objective scoring system for organ allocation.

The Model for End-Stage Liver Disease (MELD) score was originally developed to predict 3-month survival in patients undergoing transjugular intrahepatic portosystemic shunt. Based upon the patient's total serum bilirubin, serum creatinine, and the international normalized ratio (INR), the MELD score did not rely on the subjective assessment of ascites and encephalopathy severity inherent to the CTP score. The MELD score also proved to be superior to the CTP score in predicting transplant waitlist survival. In February 2002 the United Network of Organ Sharing (UNOS) adopted the MELD allocation system, and in doing so gave priority to the sickest patients based upon objective parameters, deemphasizing the importance of waitlist time. In 2008, Kim et al. demonstrated that a composite measure of serum sodium concentration and MELD, MELD-Na, was superior to MELD alone to predict waitlist mortality. MELD-Na allocation, approved in 2014 and implemented nationally in 2016, is expected to alter the score of approximately one third of the candidates on a waitlist. It should be noted that patients under the age of 12 years are prioritized on the waitlist according to the Pediatric End-Stage Liver Disease (PELD) score which is based on four clinical parameters: total bilirubin, INR, albumin, and extent of growth failure.

Although the MELD score has facilitated objective prioritization of candidates on a waitlist, geographic disparities in access to deceased donor livers exist because organ allocation and distribution for candidates with chronic liver disease predominantly occurs at the local level, followed by the regional and then national levels. The inequity may be best evidenced by the significant regional differences in the median MELD at which candidates achieve transplantation. Transplant centers in regions characterized by high, compared with low, median MELD scores exhibit more aggressive deceased donor organ utilization practices and more vigorous living donor transplant programs as strategies to meet need. Moreover, subsets of candidates, typically those with financial means, relocate to regions of relatively high organ supply to achieve transplantation.

Recently, the U.S. Health Resources and Services Administration charged the United Network for Organ Sharing (UNOS) to reconsider MELD allocation with the specific aim of mitigating geographic disparities in access to deceased donor livers. As a first step, in June 2013, “Share 35” allocation was implemented. Whereas prior to Share 35 implementation local candidates retained priority for liver grafts with the exception of Status 1 listed individuals, the new policy stipulates that liver grafts are first allocated to regional candidates with MELD scores greater than or equal to 35, ahead of local candidates with MELD less than 35. Analysis of the impact of Share 35, comparing the year immediately preceding to the year immediately following, has shown an increase in the overall number of transplants performed, increased frequency of regional sharing, increased distance of organ travel, no change in cold ischemia time, reduction in waitlist mortality, and no change in posttransplant outcomes. Currently, the UNOS Liver and Intestine Transplant Committee is considering more significant policy revisions by newly defining liver distribution units through mathematical modeling and optimization to further reduce waitlist mortality and geographic disparity. Fig. 51-1 documents the major changes in MELD allocation policies since its inception in 2002.

Candidates With Hepatocellular Carcinoma

Early results for the treatment of hepatocellular cancer (HCC) by liver transplantation were unfavorable, consisting of poor survival rates and a high incidence of posttransplantation recurrence. In 1996, Mazzaferro et al. reported results that have led to the codification of the “Milan criteria”, thereby defining the subset of patients with unresectable HCC for whom liver transplantation is the appropriate treatment. Transplantation within the Milan criteria (single tumor less than or equal to 5 cm in diameter, or two to three tumors less than or equal to 3 cm in diameter) resulted in overall and recurrence-free survival rates of 85% and 92%, respectively, at 4 years. Therefore, HCC with tumor burden within the Milan criteria gained wide acceptance as a legitimate indication for liver transplantation.

As patients with HCC often have preserved hepatic function, their calculated MELD score predicted a low risk of death from their liver disease alone. To provide HCC candidates access to liver transplantation prior to HCC progression beyond the Milan criteria, MELD exception points were awarded to this subset of liver transplant candidates. Under the initial HCC-adjusted MELD scheme of organ allocation, patients with Stage 1 HCC (one lesion <2 cm) were given a MELD score of 24, corresponding to an expected 15% 3-month dropout rate due to tumor progression beyond the Milan criteria. Those with Stage 2 HCC (one lesion 2 to 5 cm, two to three nodules all ≤3 cm) were assigned a MELD score of 29, which reflected an expected 30% 3-month dropout rate. For every 3 months on the waiting list, HCC candidates who remained within the Milan criteria were awarded additional exception points equivalent to a 10% increase in mortality risk. However, soon after its implementation, there was evidence that HCC was overvalued in this original MELD allocation, prompting two subsequent adjustments in the attribution of MELD points for HCC patients in April 2003 and again in January 2004. Stage 1 HCC candidates no longer received any MELD exception points, and Stage 2 HCC candidates were awarded a MELD score of 22.

Despite these downward adjustments in HCC exception points, mounting evidence showed that HCC candidates, compared with the non-HCC candidates, enjoyed a substantial transplant advantage irrespective of geographic location. Indeed, half of the HCC candidates on the national waitlist achieved liver transplant at a MELD score of 22, within 3 months of listing. To address the disparity in access to deceased donor livers between HCC and non-HCC candidates, the following two strategies were implemented in October 2015: (a) “HCC Delay”—instead of being listed with a MELD score of 22, candidates with HCC exception applications are registered with their calculated MELD score for the first 6 months. After this 6-month waiting period, candidates who continue to meet HCC listing criteria are awarded 28 exception points. (b) “HCC Cap”—HCC exception scores are capped at a MELD of 34 in light of the current Share 35 allocation policy.

Simultaneous to considering mechanisms to devalue HCC within the MELD economy, there are discussions to consider expansion of tumor criteria acceptable for transplantation beyond the current limits of the Milan criteria. In 2001, Yao et al. at UCSF proposed expansion of the Milan criteria for HCC liver transplant candidates. Termed the UCSF criteria, these patients with HCC had a single tumor less than or equal to 6.5 cm in diameter, or two to three tumors none exceeding 4.5 cm in diameter and whose sum of tumor diameters did not exceed 8 cm. This cohort that underwent liver transplantation following downstaging with locoregional therapy to within Milan Criteria achieved 1- and 5-year survival rates of 90% and 75%, respectively, thus demonstrating equivalent rates of long-term survival when compared with candidates presenting within the Milan criteria.

Indeed, with the success of HCC downstaging protocols, and continued advances in locoregional therapy, the upper boundaries of HCC tumor burden suitable for liver transplantation have come into question. Aside from gross evidence of macrovascular invasion, solely relying on tumor number/size has proved to be insufficient in a “one size fits all” system of granting HCC exception points. Consideration of AFP levels in conjunction with tumor size/volume, better identify HCC candidates best suited for transplant by excluding those candidates at highest risk for posttransplant recurrence. Data also suggests that waiting time serves as an effective, if not archaic, marker for tumor biology. HCC recipients emanating from UNOS regions with relatively longer waitlist times, compared with those with shorter waitlist times, have demonstrated improved long term posttransplant survival and a reduced risk for posttransplant HCC recurrence. Finally, an emerging theme in the literature is the prognostic utility of a candidate's individual response to locoregional therapy. Although traditionally utilized to reduce tumor burden with the aim of preventing waitlist dropout, achieving transplant eligibility (downstaging of patients who are outside of Milan criteria to within Milan criteria), and/or reducing posttransplant HCC recurrence, there is growing evidence that the tumor response to locoregional therapy—much like that of time on the waitlist—may allow for enhanced identification of candidates ideally suited for liver transplantation.