OBJECTIVE The aim of the study was to evaluate the long-term results in a group of patients treated by proximal tibial valgus osteotomy. METHODS A group of 92 high tibial osteotomies performed in the 1985/1994 period were evaluated. All were indicated for primary osteoarthritis. Closing wedge osteotomy was the technique used. Staple-fixed osteotomy was immobilised in a plaster cast for 6 weeks in 44 cases. Osteotomy fixed with a plate in 35 or with a cerclage in 13 cases was immobilised for two weeks. METHODS Recurrent intensive pain and walking problems were regarded as a termination of the favourable effect of osteotomy. This was evaluated by the Kaplan-Meier survival analysis. The tibio-femoral angle 5 to 10 degrees of valgus was considered as an optimal deformity correction. The results were compared in relation to complications, correction, duration of immobilisation and patient age. RESULTS The mean age of the patients was 59.8 ± 8.7 years (range, 42 to 78). During surgery, medial unicompartmental arthritis, grade II or higher, was recorded in 59 patients (64.1%) and multicompartmental knee arthritis was found in 66 patients (71.7%). The absence of noticeable problems was reported by 80.4% of the patients at 10 years and by 30.4 % at 15 years after osteotomy. Surgery decreased the range of motion from 100 ± 6.9 to 94.5 ± 17.7 degrees flexion (p = 0.04). Poor correction and post-operative complications were found in 21 patients (22.8 %) who also experienced significantly worse outcomes (p = 0.003). Good results after 10 and 15 years were reported by 47.6% and 14.3% of the patients, respectively. Of 71 patients (77.2%) who had good correction and surgery without complications, the osteotomy showed a good effect in 90.1% at 10 post-operative years and in 35.2% at 15 years. A longer immobilisation resulted in significantly worse results (p = 0.04) and a restricted range of motion (p = 0.02). The patients younger than 60 years achieved better results than the elderly patients (p = 0.38), but the difference was not significant. CONCLUSIONS Good results were recorded up to 10 years after osteotomy, although some patients had had worse arthritis than recommended for this procedure. Inadequate correction and complications deteriorated the effect of osteotomy. With the opening wedge technique it is easier to achieve good correction, and stable fixation allows for active physical therapy. Better long-lasting results are achieved with total knee arthroplasty (TKA). However, in young active adults TKA can fail prematurely. Osteotomy can postpone the necessity of this implantation. If TKA is performed after osteotomy, functional outcomes are similar to those after primary implantation. CONCLUSIONS Corrective osteotomy is an effective method in patients below 60 years who have an early-stage osteoarthritis of the knee with axial mal-alignment. If the tibio-femoral angle is over-corrected to more than 10 degrees valgus, patelo-femoral pain will ensue. Post-operative active physical therapy is necessary. After bone healing, implants must be removed. Corrective osteotomy has not been overcome by the development of joint replacement techniques, but both methods are complementary in the treatment of osteoarthritis.