Fat or Fiction

Abstract

Purpose: Hepatic steatosis (or fatty liver) is common in the United States and frequent in patients with chronic hepatitis C (CHC), where it may lead to disease progression and hinder response to HCV therapy. Therefore, it is important to recognize fatty liver. Liver biopsy confirms the degree of fat, inflammation and fibrosis, but is invasive. Hepatic ultrasound is well-tolerated, non-invasive and provides good information. It is commonplace to trust the results of an ultrasound suggesting fatty liver, but there are conflicting reports on its accuracy. Therefore, we retrospectively examined liver biopsies and compared the histologic results to the ultrasound interpretations. We selected patients with chronic liver disease, primarily CHC, in order to evaluate the accuracy of ultrasound in identifying fatty liver in a typical clinical setting. Methods: Liver biopsies were reviewed on 131 patients who had a random liver biopsy performed for evaluation of chronic liver disease (89% had CHC). The biopsies were graded for fat (grades 0–3), inflammation (grades 0–4) and fibrosis (stages 0–4). Ultrasound interpretations were grouped into 3 categories- ‘normal’, ‘fatty liver’ and ‘non-specific’, and then compared to histologic results. Results: A ‘normal’ ultrasound interpretation was correct in excluding significant fat (grades 2–3) 95.8% of the time, but 25% had some fat (grades 1–3) on biopsy, representing false negatives. Furthermore, 37.5% had significant fibrosis (stages 2–4) and 8.3% had significant inflammation (grades 2–4). A ‘fatty liver’ interpretation was only correct in identifying some fat on biopsy in 36.4% and significant fat in 11.4%, yielding many false positives. In addition, 47.7% had significant fibrosis and 18.2% had significant inflammation. A ‘non-specific’ interpretation was associated with 25.6% significant fat, 53.8% significant fibrosis and 23.1% significant inflammation. The sensitivity of ultrasound for detecting fat ranged from 32.0–88.2 and the specificity ranged from 40.4–65.4, depending on the degree of fat on biopsy and the sonographic interpretation considered to be significant. Conclusions: Ultrasound is inaccurate for diagnosing hepatic steatosis in patients with chronic liver disease; echogenic alterations are more likely to be the result of fibrosis or inflammation in this setting. Therefore, in patients with chronic liver disease, the ultrasound diagnosis of fatty liver is more ‘fiction’ than ‘fat’.