Autologous chondrocyte implantation has been used since March 1995 in a prospective cohort evaluation. One hundred sixty-nine patients, 13 to 58 years, have been treated as of December 1999. One hundred seven patients have greater than 12 months followup, and 56 have greater than 24 months followup. Overall 87% of patients improved. Patient assessment instruments include the modified Cincinnati knee rating scale, Short Form-36, Knee Society score, Western Ontario McMaster Universities Osteoarthritis Index score, and patient satisfaction survey. Treatment cases included the following categories: Simple (N = 12) isolated femoral condyles; Complex (N = 86) nonarthritic knees with multiple defects on the femur, or isolated lesions to the patella or tibia; and Salvage (N = 71) knees with early arthritic changes. The areas treated were large: Simple, 4.3 cm2 (one defect); Complex, 6.75 cm2 (4.5 cm2 per defect x 1.5 defects per case), and Salvage, 11.66 cm2 (5.3 cm2 per defect x 2.2 defects per case). Patients who had complex and salvage treatments (N = 107) with greater than 1 year followup frequently had adjuvant treatments including valgus tibial (N = 24) or tibial tubercle (N = 15) osteotomies or ligament reconstruction (N = 5). At the 2-year followup, statistically significant functional improvements occurred in the patients in the Simple and Complex categories (Cincinnati score; Simple, baseline 3.57, 24 months = 5.38; Complex, baseline 3.40, 24 months = 6.06). Patients in the Salvage category had statistically significant improvement in Short Form-36 quality of life scores (Physical summary, 24 months) and an increase in the Cincinnati rating scale when the patellofemoral joint was not involved. Patient satisfaction at 24 months for Simple, Complex, and Salvage categories was 60%, 70%, and 90%, respectively. There were 22 failures in 169 patients treated (13%), which was defined as no clinical improvement or graft failure.