Objective: Our objective was to evaluate the degree of change in left ventricular mass index (LVMI) regression after aortic valve replacement (AVR) using three different valves. Methods: Group 1 (n=17) included patients with bioprosthesis (Medtronic Hancock 2), Group 2 (n=21) included patients with mono-leaflet mechanical valve (Medtronic Hall), and Group 3 (n=17) included patients with bi-leaflet mechanical valve (St Jude). The mean ages of Group 1, 2 and 3 patients were 70.8 +/- 9.1, 61.6 +/- 13.7 and 56.2 +/- 18.3 years, respectively. In this observational study, patients were followed-up after surgery and left ventricular wall thickness and valvular functions were evaluated with echocardiography. The findings were compared with preoperative values. Statistical analyses were performed using one-way variance analysis (ANOVA), Kruskal Wallis, and Chi-square tests. Results: Statistically significant difference was observed among the three groups with respect to age (p=0.015). LVMI regressed in all groups; Group 1 from 232.74 +/- 53.36 g/m(2) (preoperative) to 174.64 +/- 46.33 g/m(2) (postoperative) (p=0.0001), Group 2-from 198.49 +/- 40.53 g/m(2) to 167.04 +/- 33.9 g/m(2) (p=0.0001), and Group 3-228.77 +/- 47.87 g/m(2) to 185.44 +/- 37.76 g/m(2) (p=0.0001). No statistically significant difference was observed among the groups with respect to LVMI regression (p=0.054, p=0.363). Conclusion: Mid-term results of AVR with three different aortic valve prosthesis revealed that all groups showed a similar regression of left ventricular mass. However, we advocate that long-term results of an increased number of patients should be evaluated for assessment in depth. (Anadolu Kardiyol Derg 2010; 10: 452-7)