The Nonalcoholic Fatty Liver Disease Activity Score (NAS) and the Histopathologic Diagnosis in Nonalcoholic Fatty Liver Disease: Distinct Clinicopathologic Meanings.
Elizabeth M Brunt, David E Kleiner, Laura A Wilson, Patricia H Belt, Brent A Neuschwander-Tetri. Washington University School of Medicine, St. Louis, MO; National Cancer Institute, Washington, DC; Johns Hopkins University, Baltimore, MD; Saint Louis University School of Medicine, MO
Background: The diagnosis (dx) of nonalcoholic steatohepatitis (NASH) is characterized by clinical features and the presence and pattern of specific histological lesions in liver biopsy. Some studies have used threshold values of the nonalcoholic fatty liver disease activity score (NAS), derived from unweighted sum of scores for steatosis (0-3), lobular inflammation (0-2) and ballooning (0-2), specifically NAS ≥ 5, as a surrogate for the histologic dx of NASH. Aim: To evaluate the validity of the NAS to replace pattern-based dx of NASH.
Design: Biopsy and clinical data from adults in the NASH Clinical Research Network (CRN) Database and clinical trials were reviewed. Biopsies had been blindly evaluated, scored and dxd by the NASH CRN Pathology Committee. Excluded were biopsies with cirrhosis, or the uncommon “pediatric” pattern of zone 1 accentuation (n=42).
Results: Among 934 biopsies, 543 (58%) were dxd as definite steatohepatitis (SH), 183 (20%) as borderline SH, and 208 (22%) “not SH”. The NAS was ≥ 5 in 51% and ≤ 4 in 49%. Among the biopsies of definite SH, 75% had NAS ≥ 5, as did 28% of borderline SH and 7% of "not SH" biopsies. Of all biopsies with NAS ≥ 5, 86% had SH and 3% were "not SH". Forty-two percent of NAS ≤ 4 were "not SH"; in the remaining 58% of NAS ≤ 4, 50% had SH. Thus, NAS ≤ 4 did not always equal a benign dx. Univariate linear regression analysis demonstrated that NAS≥5 and SH were individually associated with an increase in serum levels of ALT and AST (p<0.0001). When both the NAS≥5 and SH were included in a multiple linear regression model, both retained a significant association with ALT and AST. In contrast, SH was associated with diabetes (p<0.0001), metabolic syndrome (p=0.03), and insulin resistance by HOMA-IR (p<0.0001) and QUICKI (p=0.02), while the NAS≥5 was not associated with any of these.
Conclusions: The results of this study do not validate usage of NAS ≥ 5 as a means to dx SH. The process of histologic dx in SH, as in all forms of liver disease, is derived from evaluation of patterns as well as individual lesions, and differs from that of scoring. Thus, the diagnosis of definite SH does not always correlate with pre-determined values of the semiquantitative NAS.
Category: Liver & Pancreas
Tuesday, March 1, 2011 1:00 PM
Platform Session: Section C, Tuesday Afternoon