INTRODUCTION: Grading systems for allograft rejection should be simple, reproducible, scientifically correct and predictive of outcome. Previous studies have shown the reproducibility of the Banff Schema for acute liver allograft rejection. This study prospectively examined the ability of the descriptive grading and RAI scoring of the Banff Schema for acute liver allograft rejection to predict liver injury tests, fibrosis or bile duct loss in subsequent biopsies and graft failure from acute and/or chronic rejection.
METHODS: Descriptive grading of acute liver allograft rejection (using the same criteria as the Banff Schema) has been prospectively applied to 3619 liver allograft biopsies from 1919 primary adult (>18 years old) hepatic allograft recipients at the University of Pittsburgh Medical Center. Starting on 8/21/1995 all biopsies (n=1028) from 499 adult primary hepatic allograft recipients were also assigned a Rejection Activity Index (RAI). The histopathologic grading and scoring of acute rejection were correlated each other and with liver injury test and with the development of liver fibrosis, bile duct loss and allograft failure from acute and/or chronic rejection.
RESULTS: 258/1028 (25%) of liver allograft biopsies assigned a complete Banff Grading and Scoring showed acute rejection (>= Indeterminate and RAI > 0). The distribution of descriptive grades was indeterminate (15%); mild (69%); moderate (14%) and severe (2%). The RAI scoring range for descriptive grades is as follows: indeterminate (RAI=1-4; mode=2); mild (RAI=1-6; mode=3); moderate (RAI=3-7; mode=5) and severe (RAI=6-7; mode=7). No biopsies had a total RAI > 7 and RAI component scores of "3" were rare: portal inflammation (0.74%); bile duct damage (0.4%) and venous endothelial inflammation (3.4%). Patients experiencing "severe" acute rejection and those with an RAI score of "3" for portal inflammation showed significantly greater serum ALT and AST levels compared to those with either no rejection or lesser grades of acute rejection. There was no correlation between the rejection grade or RAI score on biopsy and serum total bilirubin or gamma-GTP. However, biopsies with a RIA score of "3" for bile duct damage showed significantly higher serum alkaline phosphatase values.
Analysis of follow-up biopsies showed that patients having a portal inflammation score of >=2 were more likely to develop portal fibrosis in subsequent biopsies than those having none or mild portal inflammation, but the difference between the two groups was mild versus none, respectively. The RAI score for bile duct damage did not correlate with subsequent bile duct loss, but a venous endothelial inflammation score of >= 2 was more likely to result in central vein or perivenular fibrosis in follow-up biopsies than those with a score of 1(p<0.02).
Nine of the 497 (2%) primary allografts with at least one biopsy assigned complete Banff grading and RAI scoring failed primarily from acute and/or chronic rejection at a median of 434 days; range 11-5802 days. Proportional hazards regression analysis for risk of allograft failure from acute and/or chronic rejection according to the RAI components showed that only venous endothelial inflammation score of "3" predicted graft failure from acute and/or chronic rejection (p=0.067; hazard rate of 1.34). This corresponds to the "severe" lesion used for descriptive grading.
One hundred and twenty four 124 of the 1919 (6.5%) primary allograft recipients whose biopsies were assigned a descriptive grade of acute rejection ultimately failed from acute and/or chronic rejection. These biopsies were graded before RAI scoring was instituted. The median time until graft failure was 198 days (range 11-5802). The histopathologic grade was strongly correlated with graft failure from acute and/or chronic rejection (p<0.001).
CONCLUSIONS: Descriptive and RAI Banff grading of acute rejection correlates with liver injury tests and is predictive of graft failure from acute and/or chronic rejection. Given the significant association between the a RAI score of "3" for subendothelial inflammation, which corresponds to "diagnostic" lesion of "severe" acute rejection, the presence of centrilobular inflammation, necrosis and hepatocellular dropout (or lytic necrosis) deserves more weight in the total RAI scoring and is worthy of further study. The advantages of templated histology report forms and systems computerized systems developed for outcomes-based research will be discussed.