We evaluated the short- to mid-term results of an unconstrained total ankle prosthesis (S. T. A. R.) with uncemented fixation. Fifty consecutive ankle replacements were performed in 48 patients between 1996 and 1999. The initial diagnosis was posttraumatic osteoarthrosis in 31 cases (62 %), primary osteoarthrosis in 8 cases (16 %), and systemic joint affection in 11 cases (22 %), e. g. rheumatoid arthritis (6 cases), hemochromatosis (2 cases), psoriasis arthritis (1 case), lupus erythematodes (1 case), and sclerodermia (1 case). There were two perioperative complications: one superficial wound dehiscence that healed uneventfully, and one injury to the dorsal foot artery that necessitated primary reconstruction. Seven revisions, all in cases of posttraumatic arthrosis, were necessary: local revision of the fibula for painful lateral impingement (3 cases), posteromedial soft tissue revision for painful restriction of dorsiflexion (2 cases), percutaneous lengthening of the Achilles tendon (1 case), and osteotomy and callus distraction for angular correction after stress fracture of the distal tibia (1 case). At the last follow-up control, 21 patients (62 %) were very satisfied, 10 patients (29 %) were satisfied, and 3 patients (9 %) were satisfied with reservations. The obtained range of motion was 30 ° (range, 15 to 55 °), with a maximal plantarflexion of 25 ° (range, 15 to 45 °) and dorsiflexion of 5 ° (-3 to 20 °). When the AOFAS-Hindfoot-Score was applied, the 34 patients scored 84.1 points (range, 44 to 100 points). After settling of the implants within 6 weeks, no migration was noted in any case, and all implants were considered to be stable. The favorable results were considered to be a result of the mechanical properties of the S. T. A. R. total ankle prosthesis that allows for unconstrained motion of the polyethylene inlay on the tibial component, e. g. anteroposterior translation, mediolateral translation and axial rotation. The success of implantation may depend on exact technique, correct hindfoot alignement, sufficient capsuloligamentous stability of the ankle, and a solid bone stock. Although our first results are very encouraging, a longer follow-up is mandatory to answer the question whether ankle replacement is a viable alternative to ankle arthrodesis.