BACKGROUND: The objective was to compare the effect of total knee replacement (TKR) followed by a 3-month non-surgical treatment with the non-surgical treatment alone in reducing pain sensitization and other pain-related measures in patients with knee osteoarthritis. METHODS: One hundred patients were randomized to (1) TKR followed by a non-surgical treatment of neuromuscular exercise, education, diet, insoles and pain medication or (2) the non-surgical treatment alone. Outcomes assessed at baseline and after 3 months were as follows: (1) pain sensitization assessed as pressure-pain thresholds (PPTs) at the knee (localized sensitization) and the lower leg (spreading sensitization), (2) peak pain intensity during the previous 24 h, (3) pain intensity after 30 min of walking, (4) pain location and pattern, (5) spreading of pain on a region-divided body chart and (6) the usage of pain medication. RESULTS: There was a statistical significant mean difference (95% CI) in change in PPTs from baseline to 3 months between groups in the crude analysis of 71 kPa (21-121) and of 75 kPa (33-117) when adjusting for baseline PPT, age, gender and body mass index, favouring the group having TKR. There were no significant between-group differences in change in the pain-related measures from baseline to 3 months (p = 0.15-0.27). Both groups improved in most of the pain-related measures (p < 0.05). CONCLUSIONS: At 3 months, TKR followed by non-surgical treatment is more effective in reducing localized and spreading pain sensitization than non-surgical treatment alone. Both treatments are equally efficacious in reducing the pain-related measures of this study. WHAT DOES THIS STUDY ADD?: Knee replacement followed by non-surgical treatment is more effective in reducing pain sensitization, but not other pain-related measures, as compared to non-surgical treatment alone at 3 months.