Are simple noninvasive scoring systems for fibrosis reliable in patients with NAFLD and normal ALT levels?

Abstract

Nonalcoholic fatty liver disease (NAFLD) is common and many affected individuals have normal-range alanine aminotransferase (ALT) levels. There is a need for a robust screening tool to triage individuals with advanced fibrosis for specialist care. The aim of this study was to assess the performance of noninvasive fibrosis tests in patients with biopsy-proven NAFLD and normal levels of ALT. Patients presenting at a fatty liver clinic between 1999 and 2009 were included in the study. Liver biopsies were assessed using the Kleiner score. The aspartate aminotransferase (AST)/ALT ratio, BARD, FIB-4 and NAFLD fibrosis scores were calculated. A total of 305 patients were included [70 with normal ALT levels (women: ALT≤30 IU/l, men: ALT≤45 IU/l) and 235 with elevated levels]. In total, 24% of patients with normal ALT levels and 17% of those with elevated ALT levels had advanced fibrosis (Kleiner stage 3-4). The FIB-4 performed best in identifying advanced fibrosis in patients with normal ALT (area under receiver operating characteristic curve=0.86, 82% sensitivity, 77% specificity and 92% negative predictive value). The sensitivity of the AST/ALT ratio and BARD and NAFLD fibrosis scores for advanced fibrosis was good in patients with normal ALT levels (94, 94 and 82%, respectively), but the specificity was low (44, 26 and 51%, respectively). The FIB-4 yielded best results in patients with elevated ALT levels. Using the FIB-4, 61% of patients with normal ALT levels and 63% of those with elevated ALT levels could avoid liver biopsy to exclude advanced fibrosis. In contrast, AST/ALT ratio and BARD and NAFLD scores would have led to a high proportion of patients with mild disease having to undergo a biopsy. The FIB-4 yielded good results in patients with normal or elevated ALT levels, reliably excluding advanced fibrosis and reducing the need for liver biopsy.