The Mathys^_¯ finger joint replacement system offers a novel fixation method into the proximal and distal medullary canals and a semi-constrained articulation. This comprises a separable, form-fit joint with a distal/proximal play of 0.7 mm, a lateral excursion of 10°, and rotation of up to 6°. This has theoretical advantages of increased radio-ulnar stability, preventing ulnar drift in patients with rheumatoid disease, and enhanced osseointegration offering implant longevity. In our unit, however, high failure rates were noted, prompting a review of cases to quantify our suspicions.

Case notes of all patients who underwent finger joint replacement using the Mathys^_¯ implant between 1999 and 2005 were retrieved. Twenty-two devices were implanted by a single surgeon during this period. Four were in finger proximal interphalangeal joints, 17 in finger metacarpophalangeal joints and one in a thumb carpometacarpal joint (CMCJ).

Indications for joint replacement included sixteen for rheumatoid arthritis (RA) and five for post-traumatic joint problems. The only thumb CMCJ was replaced for primary osteoarthritis. Patients were reviewed regularly and implant performance assessed critically along with survival of the implant to revision, infection or death of the patient. Mean follow up was 30 months.

Sixteen implants failed including the only thumb CMCJ. Modes of failure were rotation (11), poor range of motion (two), infection (two) and dislocation (the CMCJ). Similar ball and socket designs for uncemented thumb CMCJ replacement, such as the Ledoux prosthesis, have previously shown poor survival and have been withdrawn from the market. Of the finger implants in RA patients (n=16), 12 implants failed of which ten were due to rotation. Time to failure ranged from two to 48 months (mean 15 months).

The authors do not recommend the use of this implant, especially in cases of rheumatoid arthritis.

Total knee replacements and high tibial osteotomies are commonly performed orthopaedic operations with low complication rates. Both of these procedures involve surgery in close proximity to the popliteal artery with the use of power tools and sharp instruments. The behaviour of the popliteal artery during knee flexion, in particular the change in distance between itself and the posterior tibial cortex, is poorly understood. Many previous studies have been on stiff embalmed knees or with the patient lying supine, so as to subject the popliteal artery to an anterior pull from gravity.

We used duplex ultrasonography on 100 healthy knees to determine the distance of the popliteal artery from the posterior tibial surface at 0 and 90 degrees of flexion. One observer was used throughout. At 1–1.5cm below the joint line, we found the artery was closer to the posterior tibial surface in 24% of knees when the knee was flexion. This was also the case for 15% of knees at 1.5–2cm below the joint line. These two levels were chosen as they represent the usual positions for the tibial cuts performed in total knee replacement and tibial osteotomy. We provide an anatomical account to help explain our findings using cadaveric dissections, arteriography and static MRI studies. The first of our explanations for this posterior movement of the artery is the increase in the antero-posterior thickness of the popliteus muscle during knee flexion. We also observed a posterior pull on the popliteal artery from the sural vessels.

6% of the knees had a high branching anterior tibial branch. We highlight this anatomical variant as an example of an extremely vulnerable vessel. We review the existing literature regarding the popliteal artery dynamics, and conclude that 90 degrees of knee flexion is the safer position for tibial procedures, but repeat the warning that the surgeon must still take great care.