BACKGROUND/AIMS: Liver stiffness (LS) as assessed by transient elastography (TE) can change longitudinally in patients with chronic hepatitis B (CHB). The aim of this study was to identify the factors that improve LS. METHODS: Between April 2007 and December 2012, 151 patients with CHB who underwent two TE procedures with an interval of about 2 years were enrolled. Ninety-six of the 151 patients were treated with nucleos(t)ide analogues [the antiviral therapy (+) group], while the remaining 55 patients were not [the antiviral therapy (-) group]. The two groups of patients were stratified according to whether they exhibited an improvement or a deterioration in LS during the study period (defined as an LS change of < or =0 or >0 kPa, respectively, over a 1-year period), and their data were compared. RESULTS: No differences were observed between the antiviral therapy (+) and (-) groups with respect to either their clinical characteristics or their initial LS. The observed LS improvement was significantly greater in the antiviral therapy (+) group than in the antiviral therapy (-) group (-3.0 vs. 0.98 kPa, P=0.011). In the antiviral therapy (+) group, the initial LS was higher in the LS improvement group (n=63) than in the LS deterioration group (n=33; 7.9 vs. 4.8 kPa, P<0.001). However, there were no differences in any other clinical characteristic. In the antiviral therapy (-) group, the initial LS was also higher in the LS improvement group (n=29) than in the LS deterioration group (n=26; 8.3 vs. 6.5 kPa, P=0.021), with no differences in any other clinical characteristic. CONCLUSIONS: A higher initial LS was the only factor associated with LS improvement in patients with CHB in this study.