What Is the Real Gain After Liver Transplantation?

LIVER TRANSPLANTATION 15:S1-S5, 9 AASLD/ILTS SYLLABUS What Is the Real Gain After Liver Transplantation? James Neuberger Organ Donation and Transplantation, NHS Blood and Transplant, Bristol, United Kingdom; and Liver Unit, Queen Elizabeth Hospital, Birmingham, United Kingdom Received April 3, 9; accepted August 6, 9. Key Points 1. For most liver allograft recipients, both the quality and length of life are greatly improved after transplantation. However, neither the quality of life nor the length of life in the survivors returns to that seen in age-matched and sex-matched normal subjects. 2. The gain in survival after transplantation can be estimated by a comparison of the actual outcome after transplantation and the predicted survival in the absence of transplantation. 3. The reduction in graft and patient survival, in comparison with a normal age-matched and sexmatched population, is determined by several factors: short-term survival is affected by the patient s condition pre-transplant and the quality of the graft, and for longer term survival, recurrent disease accounts for most of the differences seen between different indications. Some of the causes of premature death (such as infection, de novo malignancy, and cardiovascular and cerebrovascular disease) that are increased in the liver allograft recipient may be reduced by improved management with more aggressive surveillance and treatment. 4. The aims of selection and allocation vary in different health care systems: transparency, objectivity, equity of access, justice, mortality awaiting transplantation, utility, and transplant benefit are all important but often competing demands. Understanding the associated increase in survival will allow for a rational approach to this complex area. Liver Transpl 15: S1-S5, 9. 9 AASLD. Liver transplantation has become an accepted form of therapy for patients with end-stage liver disease: indications include an anticipated survival in the absence of transplantation of less than 1 year [as suggested by an adjusted Model for End-Stage Liver Disease (MELD) score above approximately 15] and a quality of life that is unacceptable to the patient. Over the last 3 decades, overall patient and graft survival has improved, despite a lowering of the quality of the grafts, an increase in the use of extended donor criteria grafts, and a reduction in the contraindications. Although there have been many advances in surgical and anesthetic techniques and an increase in the number and classes of immunosuppressive agents available to the clinician, most of the improvements in survival are associated with a reduction in early mortality (less than 6 months), 1 and rates of late graft loss have not improved greatly. Of course, there are many possible explanations for this failure to improve long-term outcomes, and they include changes in the case mix and in the quality of the grafts used. IMPROVEMENT IN SURVIVAL ASSOCIATED WITH TRANSPLANTATION For obvious reasons, there have been no formal randomized trials comparing outcomes of transplantation with those of conservative management. However, whether liver transplantation is associated with an increase in survival can be tested by a simple comparison of the observed outcome and the outcome predicted by prognostic models. Abbreviations: AIH, autoimmune hepatitis; ALD, alcoholic liver disease; FHF, fulminant hepatic failure; HCV, hepatitis C virus; MELD, Model for End-Stage Liver Disease; NASH, nonalcoholic steatohepatitis; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis; SIR, standardized incidence ratio (%). Address reprint requests to James Neuberger, Liver Unit, Queen Elizabeth Hospital, Birmingham, United Kingdom B15 2TH. Telephone: 44 121 627 2414; FAX: 44 121 627 2449; E-mail: j.m.neuberger@bham.ac.uk DOI 1.12/lt.21891 Published online in Wiley InterScience (www.interscience.wiley.com). Liver Transplantation, Vol 15, No 11, Suppl 2 (November), 9: pp S1-S5 S1

S2 NEUBERGER Figure 1. Actual and predicted survival of patients transplanted for primary biliary cirrhosis in the first cohort of patients (1982-199) and the second cohort of patients (1991-1999). The expected survival was calculated from a disease-specific model. Reprinted with permission from Hepatology. 2 Copyright 1, American Association for the Study of Liver Diseases. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.] How to Measure Improvement in Survival Assessing the survival benefit of transplantation can be done by a comparison of outcomes from the time of listing (intention to treat) or from the time of transplantation. The method of choice will depend on the question to be addressed. For instance, for a patient with small liver cell cancer and hepatitis C virus related cirrhosis, the comparison of outcomes of transplantation with those of resection should use survival from listing, whereas for an assessment of the survival gain from transplantation associated with, for example, cryptogenic cirrhosis or other indications for which there is no effective treatment, the comparison of outcomes should be from the date of transplantation. The use of prognostic models to assess survival in the absence of transplantation is attractive and is probably the best option but is subject to limitations, and caution must be applied when data generated from such prognostic models are used. These models are usually derived from historical and retrospectively collected series, so extrapolation to the transplant candidate population may lead to inappropriate conclusions; furthermore, the models often do not take into account advances in diagnosis and treatment or changes in referral. Additionally, in many models, those who were transplanted are censored and not included in the modeling set. Some models are developed to assess survival from a stage of disease well before transplantation, so variables associated with death or survival occurring, for example, 1 years before the endpoint may be different or have different weights at a time nearer the endpoint. However, despite these caveats, the comparisons do validate the very clear clinical impression that transplantation does increase survival. These comparisons between actual and predicted survival do provide an estimate of the survival benefit. For example, an early analysis of survival benefit in patients undergoing transplantation for primary biliary cirrhosis showed an improvement in overall survival between 2 cohorts (Fig. 1); however, the survival benefit, as demonstrated by the difference between observed survival after transplantation and anticipated survival (as judged from a disease-specific prognostic model), decreased over time. There are many potential explanations for this observation, but it demonstrates that the less ill patients are at the time of grafting, the less the benefit will be. 2 Similar conclusions can be drawn from a recent analysis of three 5-year cohorts of patients grafted in our unit: the outcomes are improving, yet the increase in survival as a result of transplantation is decreasing (Fig. 2). The MELD model was used for the prediction of survival in the absence of transplantation, but it should be noted that the MELD model was derived from and validated in a US population and does not apply that closely to UK patients. 3 Furthermore, the MELD model was developed for the estimation of short-term survival in those undergoing intrahepatic stenting; indeed, modifications have been suggested, and over time, the weights of the variables have been adjusted. Further improvements in accuracy may be made by the inclusion of serum sodium. It is also clear that the components of the model may be affected by a variety of factors, including gender, variations in the methods used for the measurement of analytes, and compounding factors such as hemolysis. Nonetheless, the model is effective and robust, and the data are clear. Those with a worse prognosis without transplantation have a greater transplant benefit (Fig. 3); because of careful donor-recipient matching by the clinician, outcomes are little different in comparison with the outcomes of those with a better prognosis (as shown by a lower MELD), although resource utilization may be greater. Survival benefit is affected by many factors, including donor factors (as shown by the donor risk index, for example) and the status of the patient at the time of transplantation. Furthermore, the transplant benefit is affected by the indication. 4,5 It must be recognized too that the benefit will vary according to the time period over which the benefit is calculated: selecting a benefit period of 5 or years will affect the selection of the transplant candidate. LOSS OF LIFE YEARS AFTER TRANSPLANTATION Liver allograft recipients, like other recipients of solid organ grafts, are at risk from a number of conditions,

WHAT IS THE REAL GAIN? S3 Figure 2. Actual survival after transplantation and predicted survival in the absence of transplantation (predicted with the unadjusted MELD model) for patients transplanted in Birmingham (A) between 1994 and 1998 and (B) between 4 and 8. Abbreviation: MELD, Model for End-Stage Liver Disease. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.] Figure 3. Actual and predicted survival (according to the unadjusted MELD score) of patients transplanted in Birmingham between 1994 and 8 with MELD scores (A) greater than 3 and (B) lower than 15. Abbreviation: MELD, Model for End-Stage Liver Disease. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley. com.] including disease recurrence and side effects of immunosuppression: the latter may be generic (such as an increase in infections and some de novo malignancies) or class-specific or drug-specific (such as renal impairment or diabetes). It is therefore not surprising that some of these conditions will affect patient survival. There is a great variation between the rates of recurrence and the rates of graft loss from recurrence (Fig. 4); graft loss due to disease recurrence is, in our experience, greatest for hepatitis C virus infection, primary sclerosing cholangitis, and autoimmune hepatitis. 6 Indeed, variations in the impact of recurrence is the major cause of the different rates of graft survival. Liver allograft recipients have an increased risk of and increased mortality from cardiovascular and cerebrovascular disease, infection, and de novo malignancy (Fig. 5). 6,7 Interventions, such as aggressive treatment of diabetes, hyperlipidemia, and hypertension, prevention of renal impairment, and screening for selected malignancies, may reduce the increased mortality rate. Against this background, it is not surprising that liver allograft recipients have reduced survival in comparison with an age-matched population. The extent of the life years can be readily estimated by a comparison of the outcomes of transplant recipients with those of age-matched and sex-matched normal subjects. Such an analysis must be tempered by the observation that the liver transplant recipients are, apart from the liver

S4 NEUBERGER Figure 4. Rate of graft loss from recurrent disease. Abbreviations: AIH, autoimmune hepatitis; ALD, alcoholic liver disease; FHF, fulminant hepatic failure; HCV, hepatitis C virus; NASH, nonalcoholic steatohepatitis; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis. Reprinted with permission from Transplant International. 6 Copyright 8, European Society for Organ Transplantation. 6 5 25.1 lifeyears 51.2 13.8 lifeyears Adult liver transplant recipients Equivalent UK population Life expectancy (years) 4 3 26.1 (18.8-36.2) 38.4 24.6 (19.5-31.1) 6.5 lifeyears 29.7 23.2 (19.5-27.6) 1.8 lifeyears. 21.8 (17. - 23.6) 4.2 lifeyears 12. 16.2 (9.1-15.8) 1 17-34 35-44 45-54 55-64 65-74 Age group Figure 5. Relative increase in de novo cancers after liver transplantation. Abbreviation: SIR, standardized incidence ratio (%). Reprinted with permission from Transplantation. 8 Copyright 5, Lippincott Williams & Wilkins. disease, healthy, and unlike the controls, those potential candidates with, for example, advanced cardiac or respiratory diseases will be excluded from transplantation. According to own analysis, UK adult liver transplant recipients who survive the first postoperative year will have an estimated 7.7-year reduction in life expectancy. 1 The extent of life years depends on several factors, including age, gender, and indications (Figs. 6-8). The younger recipients have the greatest number of life years ; this is possibly because the impact of immunosuppression (as discussed previously) is likely Figure 6. Predicted survival (in years) of adults who have received a liver transplant and survived 1 year and a matched UK population. Reprinted with permission from Gut. 1 Copyright 7, British Society of Gastroenterology. [Color figure can be viewed in the online issue, which is available at www. interscience.wiley.com.] to exist regardless of age. These data do provide an indication of life expectancy after transplantation and should be helpful to recipients and their families in understanding what to expect for the future and in planning for their future. Of course, caution must be taken in extrapolating these predictions, which are based on historical data, to an individual. It should be noted that these studies address survival and take no account of quality of life. Thus, a patient grafted for intractable pruritus or severe encephalopathy in the context of good graft function may have a dramatic improvement in the quality of life and a sig-

WHAT IS THE REAL GAIN? S5 Life expectancy (years) 6 5 4 3 1 17-34 years 35-44 years 45-54 years 55-64 years 65-74 years Similarly, for those with liver cell cancer, changes in selection and preoperative and postoperative management are likely to affect survival. When we look at gains in survival, it is not clear whether outcomes after transplantation have to be compared with outcomes after other treatment modalities such as resection, chemotherapy, and loco-regional therapies or after no treatment. In clinical practice, the different therapeutic options are usually complementary rather than competitive. There have been very few adequately powered prospective studies comparing transplantation with other therapies: the studies have been mainly retrospective comparisons of 2 treatment modalities or an evaluation of a settled protocol, so it has not been possible to do a full comparison. PBC PSC Crypto AIH HBV Alcohol HCV Cancer Figure 7. Life expectancy for males who have survived 1 year according to age and indication. Reprinted with permission from Gut. 1 Copyright 7, British Society of Gastroenterology. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.] Life-years 1-1 - -3-4 -5 PBC PSC Crypto AIH HBV Alcohol HCV Cancer Males nificant gain in quality-adjusted life years, although the gain in survival alone may be minimal. Issues surrounding quality of life are discussed elsewhere. Hepatitis B virus It is difficult to predict the survival benefit of those with hepatitis B virus infection, as the introduction of safe and effective agents to prevent infection of the graft means that historical data cannot be used to predict future life expectancy. Liver Cell Cancer Females 17-34 years 35-44 years 45-54 years 55-64 years 65-74 years Figure 8. Life years gained or according to age and indication in males who have survived 1 year post-transplant. Reprinted with permission from Gut. 1 Copyright 7, British Society of Gastroenterology. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley. com.] CONCLUSION An understanding of the gain in survival and quality of life after transplantation is important for providers of health care, for patients and their families, and for health care professionals. Furthermore, consideration of outcomes is a key component of any allocation system. A clear understanding of transplant benefit and those donor and pretransplant and posttransplant recipient factors that affect outcome will allow for a rational debate and evidence-based systems of selection and allocation. ACKNOWLEDGMENT The help from Bridget Gunson in generating some of the data is gratefully acknowledged. REFERENCES 1. Barber K, Blackwell J, Collett D, Neuberger J, for the UK Liver Transplant Advisory Group. Life expectancy for adult liver allograft recipients in the UK. Gut 7;56:279-282. 2. Liermann Garcia RF, Evangelista Garcia C, McMaster P, Neuberger J. Transplantation for primary biliary cirrhosis: retrospective analysis of 4 patients in a single center. Hepatology 1;33:22-27. 3. Freeman RB, Jamieson N, Schaubel DE, Porte RJ, Villiamil FG. Who should get a liver graft? J Hepatol 9; 5:664-673. 4. Schaubel DE, Guidinger MK, Biggins SW, Kalbfleisch JD, Pomfret EA, Sharma P, Merion RM. Survival benefit-based deceased-donor allocation. Am J Transplant 9;9:97-981. 5. Lucey M, Schaubel DE, Guidinger MK, Tome S, Merion R. Effects of alcoholic liver disease and hepatitis C infection on waiting list and posttransplant mortality and transplant survival benefit. Hepatology 9;5:4-46. 6. Rowe IA, Webb K, Gunson BK, Mehta N, Haque S, Neuberger J. The impact of disease recurrence on graft survival following liver transplantation: a single centre experience. Transpl Int 8;21:459-465. 7. Mells G, Neuberger J. Long term care of the liver allograft recipient. Semin Liver Dis 9;29:12-1. 8. Oo YH, Gunson BK, Lancashire RJ, Cheng KK, Neuberger J. Incidence of cancers following orthotopic liver transplantation in a single center: comparison with national cancer incidence rates for England and Wales. Transplantation 5;8:759-764.