Living donor liver transplantation in high-risk vs. low-risk patients: optimization using statistical models.

Living donors represent a recognized alternative for facilitating the access to transplantation in a period of organ shortage. However, which candidates should be preferentially considered for living-donor liver transplantation (LDLT) is debated. The aim of this study was to create statistical models to determine which strategies of selection for LDLT provide the most efficient contribution. The study included 331 patients listed for deceased-donor transplantation (DDLT) and 128 transplanted with living donors. Statistical models predicting the events following listing were created and combined in a multistate model allowing the testing of different strategies of selection for LDLT and to compare their results. Taking 3-yr survival after listing as the principal end-point, selecting the 20% patients at highest risk of death on the waiting list gave better results than selecting the 20% patients at lowest risk of death after LDLT (70% vs. 64%, respectively). These strategies resulted in waiting list mortality rates of 17% and 8%, respectively. One-year survival after LDLT was lower in high-risk patients (85%) than in low-risk patients (91%). However, the 1-yr survival benefit derived from LDLT was 75% in high-risk patients while it was nil in low-risk patients. In conclusion, LDLT is more effective for overcoming the consequences of organ shortage when performed in patients at high risk of death on the waiting list. On an individual basis, the sickest patients are those who derive the most important benefit from LDLT. This study provides incentives for considering LDLT in high-risk patients. Copyright 2006 AASLD