Outcome Predictors in Transplantation

Model for End-Stage Liver Disease and Posttransplant Survival

Although the MELD system is accurate at predicting 3-month waiting list mortality from liver disease, the MELD score alone is not able to accurately predict posttransplant deaths. Although early studies showed progressively lower patient survival benefits with rising MELD scores, recent studies have validated no direct correlation between MELD score and posttransplant survival. The MELD prediction of 3- month posttransplant mortality has a C statistic of only 0.62 (95% confidence interval, 0.55 to 0.69). Models with a C statistic value greater than 0.70 are considered clinically acceptable. Successful development of preoperative models that accurately predict patient survival after liver transplantation has therefore become a crucial concern of the liver transplant community. Such models will facilitate the clinical decision-making process by (1) balancing outcomes with severity of disease to avoid “futile” transplantation and (2) providing guidelines based on objective criteria for “delisting” of transplant candidates when expected OLT outcomes fall below an established survival threshold, as proposed by an amendment to the final rule of the Department of Health and Human Services (DHHS).

Several authors have proposed predictive models using the UNOS registry data. The survival outcomes following liver transplantation (SOFT) score used 18 risk factors (recipient, donor, and operative) to successfully predict 3-month recipient survival following OLT ( Tables 98-1 and 98-2 , Fig. 98-1 ). The most significant risk factors were a history of previous transplantation and the need for life support before transplantation. Another model, balance of risk (BAR) score, identified six strongest predictors of posttransplant survival: recipient MELD score, cold ischemia time, recipient age, previous OLT, and life support before transplant. Although these models may provide clinicians an overall guide, analysis of registry data suffers lack of granularity for the level of expertise of a particular transplant center in the medical and surgical care of patients with the highest acuity of illness. Therefore it is imperative that the implementation of clinical decision pathways be based on each transplant center’s medical capacity, experience and depth as well as the ability to match donor-recipient risk factors under the current organ allocation system to optimize outcomes after OLT.

High Acuity of Illness and Posttransplant Survival

To date the balance between serving the sickest first, as stipulated in the DHHS final rule, strikes a delicate balance with avoiding futile transplantation. During the initial implementation of the MELD allocation system, a consensus was reached to cap the MELD score at a maximum of 40 points so that candidates with MELD scores higher than 40 do not receive additional priority. This policy allows transplant centers to attempt transplantation of desperately ill candidates without providing an additional advantage of organ allocation because it was thought that transplant candidates with MELD scores higher than 40 were at high risk for death at OLT.

Although patients with MELD scores of 40 or higher have the highest 3-month waiting list mortality rate (80% to 100%), these patients receive the greatest survival benefit from OLT, compared to patients with lower MELD scores. Allocation of organs to the sickest first has changed the characteristics of liver transplant recipients during the last decade and has brought new medical and economic challenges to bring these patients through OLT. Patients undergoing OLT today have more severe end-stage liver disease, are older, and have greater comorbidity, compared with the pre-MELD era. With implementation of broader sharing of deceased donor livers to patients with high acuity of illness, status 1, and MELD scores of 35 or higher, it is paramount to provide not only excellent medical care and achieve optimal outcomes in high-risk patients but also avoid futile transplantation in the face of organ crisis.

The UCLA group proposed that a cumulative futility risk score (UCLA Futility Risk Scoring System) that included disease-specific factors, including laboratory MELD score, cardiac risk, previous septic shock, and comorbidities, predicted posttransplant futility, rather than demographic, donor, or operative factors. Table 98-3 and Figure 98-2 show the cumulative UCLA Futility Risk Scoring System and long-term posttransplant patient survival outcomes. Cardiac risk and age-adjusted comorbidities were associated with the highest risk for futile outcome in the study. Cardiac risk factors that included elevated pretransplant serum troponin levels or history of cardiovascular disease and coronary artery disease were highly predictive of mortality and graft loss within 1 year after OLT, The scoring system included the Charlson Comorbidity Index (CCI), a validated instrument to assess whether a patient will live long enough to benefit from a specific screening measure or medical intervention. Age-adjusted CCI was highly predictive of futility, with no 5-year survivors among recipients with an age-adjusted CCI of 6 or higher. Another study also showed that both the CCI and individual comorbidities, including coronary artery disease, diabetes mellitus, chronic obstructive pulmonary disease, connective tissue disorders, and renal insufficiency, were associated with decreased posttransplant survival. Therefore cardiac and other aggravated comorbidities are probably the most important predictors of futile liver transplantation in the sickest patients.

Based on the four identified independent futility predictors (MELD score, septic shock, cardiac risk, and age-adjusted CCI), the model had a good discriminatory ability with a C statistic of 0.75 compared to other established risk scores such as SOFT score (C statistic of 0.70), BAR score (C statistic of 0.68), and Donor Risk Index (DRI) (C statistic of 0.53). The proposed UCLA Futility Risk Score can be applied for risk stratification of the sickest patients considered for OLT. Liver transplant candidates with MELD scores of 40 or higher who have septic shock, cardiac risk, and significant comorbidities should not undergo liver transplantation because the predicted risk for futile outcome is greater than 75%. Based on this model excluding futility, the 5-year patient survival rate of our high MELD group at 75% was comparable to reported rates for all recipients in the United States (73%) and Europe (71%).