The hospital at which liver transplantation (LT) is performed has a substantial impact on post-LT outcomes. Center-specific outcome data are closely monitored not only by the centers themselves but also by patients and government regulatory agencies. However, the true magnitude of this center effect, apart from the effects of the region and donor service area (DSA) as well as recipient and donor determinants of graft survival, has not been examined. We analyzed data submitted to the Organ Procurement and Transplantation Network for all adult (age ≥ 18 years) primary LT recipients (2005-2008). Using a mixed effects, proportional hazards regression analysis, we modeled graft failure within 1 year after LT on the basis of center (de-identified), region, DSA, and donor and recipient characteristics. At 115 unique centers, 14,654 recipients underwent transplantation. Rates of graft loss within a year varied from 5.9% for the lowest quartile of centers to 20.2% for the highest quartile. Gauged by a comparison of the 75th and 25th percentiles of the data, the magnitude of the center effect on graft survival (1.49-fold change) was similar to that of the recipient Model for End-Stage Liver Disease (MELD) score (1.47) and the donor risk index (DRI; 1.45). The center effect was similar across the DRI and MELD score quartiles and was not associated with a center's annual LT volume. After stratification by region and DSA, the magnitude of the center effect, though decreased, remained significant and substantial (1.30-fold interquartile difference). In conclusion, the LT center is a significant predictor of graft failure that is independent of region and DSA as well as donor and recipient characteristics.
© 2013 American Association for the Study of Liver Diseases.