Restricted motion in flexion is a frequent TKA complication (0.1–5.3%). The aetiology has to be searched because adhesive knee arthritis is a rare pathology. Neglecting an implant malposition, an infection or a RSDS can lead to early recurrence of stiffness. After 8 weeks, it is very dangerous to try a knee manipulation under anaesthesia. Thus, we have the choose between two difficult arthrolysis: the open and the arthroscopic. We have developed the Less Invasive Arthrolysis (LIA) as a less aggressive technique to treat knee flexion stiffness. This subcutaneous procedure (performed by one or two arthroscopic portals) was already described on a short number of patients or as a part of the arthroscopic arthrolysis. In our institution, we have performed 3738 TKA, 144 knee manipulations (3.8%) and 67LIA after TKA (21% from other surgeons) between 1997 and 2009. We have reviewed retrospectively these 67 cases (Group A) and reviewed clinically 41 of these patients in a study consultation (Group B). Subjective results: from 41 patients (B), 66% have more flexion, 32% feel less pain (VAS: 6), 49% feel better than before LIA. 58% would undergo a new LIA if they had to do it again. Objective ROM (A):preoperative ROM: 88°, in the early postoperative period, we noted a 31° flexion improvement. At the last evaluation (6–120m after LIA), the flexion improvement was 17°(−15/+80) and the final ROM was 105°. The flexion falls of 45% in the first 6 months and became stable at the 7th month (until120m). We have isolated two particular subgroups: the first including the carriers of femoral implants positioned in internal rotation (< 5°) (6% of A), in which the flexion was only improved by 6°; the second including those who underwent a stiffness recurrence (9% of A) after knee manipulation, for which we obtained a stabilization of their flexion at 105° 1 year after LIA. Relative patellar mobility(B): 66% kept a free and painless patella. Mean clinical scores (B): the long term OXF-12 score (best=12) is 33 (−18%), the HSS (best=104) is 74 (+12%). Considering the delay between TKA and LIA (67 patients, mean 28m (2–120)), the best results were obtained when we performed 6 to 24m after TKA (flexion +19° in the 7–12m, +17° in the 13–24m, versus 14° in the > 25m group). No infection occurred (0/67). We never did twice the LIA in the same knee. The published series on open arthrolysis performed 17m after TKA show an improvement of flexion by 25°, 8° for extension. An arthroscopic arthrolysis performed 12m after TKA can lead to 20° of improvement in flexion (17–42) and 3° in extension. The gold operative indication is a flexion reduced to less than 90°, 6 months after TKA, with anterior knee pain. This study presents a reliable less invasive technique studied on a bigger group with a longer follow-up and approachable by the majority of surgeons.