Three hundred and four consecutive cementless Ortholoc I TKAs were followed for 2 to 5 years. One patient had femoral and tibial component loosening. There were no cases of patellar component loosening, but six knees had failure of the patellar polyethylene surface due to wear. A major design change of the metal-polyethylene locking mechanism was effected to address this problem. The patellar component is now countersunk and the central post has been reduced in size. 79.1 per cent of the knees were graded excellent, 13.4 per cent good, 5.8 per cent fair, and 1.7 per cent poor. The Ortholoc I group was compared to a similar group of 424 Ortholoc II knees at 1 year postoperatively for pain relief. The Ortholoc II knee differs from the Ortholoc I primarily in tibial component fixation, using four peripheral 6.5-mm screws. The Ortholoc II group had significantly fewer painful knees, which suggests that the screws in the tibial component were effective in improving fixation and controlling load distribution. Survival analysis showed no tendency for the results to deteriorate over the time span studied. Cementless fixation of all components, including the tibia and patella, was highly successful and can be achieved without stress relief of the distal femur.