Abstract Aims. To investigate whether hyperechogenicity of liver can reliably be interpreted as liver steatosis and if any concomitant or isolated fibrosis can be disclosed. Patients and methods. A series of 165 patients with no signs or symptoms of liver disease referred because of slightly to moderately raised aminotransferases (alanine aminotransferase and/or aspartate aminotransferase 0.7–5.0 μkat/l) for more than 6 months were prospectively investigated with a comprehensive laboratory profile, ultrasound examination of liver and percutaneous liver biopsy. Fibrosis was assessed quantitatively and according to Metavir. Steatosis was graded as none, mild, moderate or severe. Results. Of 98 (59.4%) patients with raised echogenicity 85 (86.7%) had liver steatosis of at least moderate degree, 9 patients with same degree of steatosis had normal echogenicity and 13 patients with no or only mild steatosis had a hyperechogenic liver (sensitivity 0.90, specificity 0.82, positive predictive value 0.87, negative predictive value 0.87). About the same relations were found regardless of body mass index and degree of fibrosis. With increased echogenicity together with high attenuation (n=59) and reduced portal vessel wall distinction (n=79), positive predictive value increased to 0.93 and 0.94, respectively. Quantitatively assessed fibrosis (mean ± SD) was 3.2±4.6% of biopsy area with normal and 2.3±1. 8% with raised echogenicity (ns). Echogenicity was normal in 5 out of 9 patients with septal fibrosis and in 4 out of 6 patients with cirrhosis. Any structural, non-homogenous findings at ultrasound were not associated with architectural fibrotic changes and none had nodular contours of liver surface. Conclusions. Assessment of liver echogenicity is of value for detection or exclusion of moderate to pronounced fatty infiltration (correct classification 86.6%) but cannot be relied upon in diagnosing fibrosis, not even cirrhosis in asymptomatic patients with mild to moderately elevated liver transaminases.