Introduction: To establish the diagnostic usefulness of liver biopsy (LB) and its influence on the therapeutic approach in patients with persistent abnormal liver tests of unknown cause.
Methods: The 1135 LB performed between January 1999 and January 2007 were retrospectively evaluated. Patients with a strongly suspected diagnosis were excluded. One hundred and twelve patients with chronic elevation of aminotransferases (103 patients), gamma-glutamyltransferase (GGT) (7 patients) and ferritin (2 patients) were included in the analysis.
Results: The most common diagnoses were normal liver (35 patients, group 1), minimal changes (15 patients, group 2), bland steatosis (23 patients, group 3), non-alcoholic steatohepatitis (NASH) (17 patients, group 4), and chronic hepatitis (eight patients, group 5). Three patients had tuberculosis of the liver and the remaining 11 had other diagnoses. Ferritinemia was the only parameter significantly lower in group 1 than in groups 2 (p = 0.038), 3 (p = 0.023), and 4 (p < 0.001). Transaminase levels lacked discriminatory value except in chronic hepatitis (p = 0.008). Alkaline phosphatase levels (p = 0.003) were lower in group 4 than in group 1. Triglyceride levels were higher in group 3 (p = 0.009) and group 4 (p = 0.008) than in group 1. Ultrasonography detected steatosis in 28 of the 40 patients with fatty liver (specificity = 0.94; sensitivity = 0.70).
Conclusions: LB modified the therapeutic approach only in the three patients with hepatic tuberculosis. LB confirmed ultrasonographic findings of steatosis and differentiated bland steatosis from NASH, but did not influence the therapeutic approach. Most patients with normal findings on ultrasonography had normal or near-normal biopsies. The indication for LB should be individualized.