Aims: Prospective randomized single blind investigation was made to evaluate inßuence of two surgical approaches to the knee (antero-medial approach and subvastus approach) for total knee arthroplasty. Materials: Between 1995– 2002 215 consecutive total knee arthroplasties in 204 patients (192 women, 12 men) were made. There are following prosthesis implanted: PFC Ð 74, GSB Ð 45, PFC ∑ Ð 73, Link Ð 5, AGC Ð 18. All patients were evaluated. Patientsñ mean age was 65,7 years old. TKR with antero-medial approach was performed in 98 cases (45 per cent), whereas TKR with subvastus approach was performed in 117 cases (55 per cent). Methods: Functional status according to HSS scale at the 14 days and 6 months after procedure was regard as end results for both protocols. Results: In group with antero-medial approach there were: 26,4% excellent results, 69,8% good results and 3,8% fair results respectively. In group with subvastus approach there were: 24,3% excellent results, 70,2% good results, 3,3% fair and 2,2% poor results respectively. There are no statistical differences between early functional results of these two surgical approaches for total knee replacement. Conclusions: We assumed that subvastus approach is as good as anteromedial one in knee joint exposure but less harmful for extensor apparatus and joint proprioception.