We report a prospective trial comparing the effectiveness of a post-operative flexion regime versus a standard extension regime on the early outcome of total knee arthroplasty. A total of 160 knees were allocated to two separate post-operative mobilisation regimes: a static flexion regime or an active extension regime. Patients were implanted with either an IBII or 913 total knee replacement systems. All operations were performed by the same surgeon. Patients were assessed pre-operatively and at 6 weeks, and were well matched with respect to age, sex, diagnosis and pre-op movement. Patients subjected to the flexion regime had a better maximum flexion and range of movement at 6 weeks and were also discharged earlier. These results were statistically significant (P < 0.05). Histograms of results at 6 weeks show that the flexion group had a more predictable outcome with smaller standard deviations (S.D.) than the extension group. Differences were also observed in the percentage of 'sub-optimal' results. No difference was observed in post-operative wound problems. We believe a static flexion regime is as good as reported results with CPM, better than an extension regime and does not result in an increase in wound problems.