One hundred twenty patients (22 men, 98 women; 144 knees) with uncemented Freeman-Samuelson total knee arthroplasty were followed prospectively. Eighty-one patients had rheumatoid arthritis and 39 patients had osteoarthrosis. The mean follow-up period was 6.8 years. Three different types of tibial components were used: a high-density polyethylene component without stem, a metal-backed tibial component without stem, and a metal-backed tibial component with stem. Progressive varus tilting turned out to be an early sign of failure and occurred in 22% of the tibial components. Revision of the tibial component was done in 17 knees. Survival analysis with revision as endpoint revealed a survival rate of 79% at a follow-up period of 10 years. Cementless fixation of this design using macrointerlocking pegs and no other stabilization resulted in poor fixation and a high revision rate and cannot be recommended.