T he complete removal of the cement mantle at revision arthroplasty can be extremely difficult. Some authors advise a 'cement-within-cement' revision technique in which a new layer of cement is applied to the old before insertion of the femoral component. We could find no long-term clinical data regarding the success of this procedure. In a simple biomechanical study, we examined the strength of the cement-to-cement interface in conditions likely to prevail in vivo. We found that the presence of a thin layer of blood and marrow debris at the interface weakened the cement-to-cement bond by 80% to 85%. These biomechanical findings and additional photomicrographic evidence do not support the practice of cement-within-cement revision arthroplasty.

Synovial fistulae after arthroscopy and titanium-induced synovitis of the knee are both uncommon. We report a case in which a persistent fistula led to frank infection after a diagnostic arthroscopy during which a florid titaniuminduced synovitis of the knee had been discovered. We are not aware of any previous reports of this association. Case report. A 66-year-old man was admitted with pain and locking of his left knee two years after insertion of a Brigham unicondylar prosthesis for medial compartment osteoarthritis. The pain was worse around the patella and exacerbated by movement and weight-bearing. Radiographs were satisfactory apart from the presence of a cement loose body, and the WBC and ESR were normal. The provisional diagnosis was either mechanical impingement due to the loose body or patellar maltracking.Arthroscopy showed a florid black-stained titaniuminduced synovitis with a large quantity of metal debris in the synovial fluid. The polyethylene of the tibial component was completely worn, allowing metal-to-metal articulation at the tibiofemoral joint. The central part of the metal tibial tray had also worn through ( Fig. 1). Synovial biopsy showed a giant-cell granulomatous reaction due to titanium oxide.A decision was made to revise the knee arthroplasty in a few weeks, but the medial parapatellar portal failed to heal and became a persistent and troublesome synovial fistula (Fig. 2). This discharged black-stained sterile fluid for over three months and was complicated by a frank staphylococcal septic arthritis. Treatment by joint lavage and longterm antibiotic therapy was required before the infection settled. The knee revision was eventually performed using antibiotic-loaded cement. A blood test one week later showed a markedly raised titanium level of 420 g/l (normal ≤8 g/l) indicating titanium metallosis. Discussion. The incidence of persistent synovial fistula after arthroscopy is very low. Bamford et al (1993) reported only one case in 8500 arthroscopies. In our patient a large amount of titanium debris had been generated by the wearing of the chrome-cobalt femoral component of the Brigham prosthesis through the titanium tibial tray. The chronic effusion caused by the titanium particles may have resulted in persistent leakage of fluid which prevented healing, or the particulate titanium may have acted as a wound irritant. Whatever the exact mechanism, we are not aware of any previous report of an association between metal-induced synovitis and fistula formation.Without arthroscopy, a high index of suspicion is required to make the diagnosis of metal-induced synovitis. Other disorders, such as aseptic loosening, are a much more common cause of pain after knee arthroplasty. Weissman et al (1991) found that the presence of a dense line outlining a portion of the capsule (the so-called metal-line sign) was 322 BRIEF REPORTS Fig. 1 Fig. 2