A knowledge of bacteriological profiles in previously treated cases is helpful as a guide to management of infected joint replacements, especially in those cases where bacteriology results are not available. The object of this study was to assess the changing trends of the bacteriology of infected total knee replacements (TKR) over 2 decades.

The records of 79 patients undergoing revision for infected TKR between 1979 and 1999 were reviewed. There were 30 males and 49 females, average age 63 years, range 36 – 82 years. The types and sensitivities of bacteria identified, and the use of prophylactic systematic and cement antibiotics, was recorded. The Chi-squared test was used to test statistical significance.

70 organisms were identified in 62 patients: 29 Staphylococcus aureus; 27 Coagulase Negative Staphylococci (CNS); 8 Streptococci; 6 Coliforms. In the 1980s S. aureus accounted for 55% of infections, CNS 25%. In the 1990s S. aureus 38%, CNS 41%. Following the use of systematic antibiotics (3 x cefuroxime) or antibiotic cement (bacitracin/erythromycin/colistin – BEC) at primary TKR, fewer CNS infections were seen (p< 0. 05). There was only 1 case of methicillin resistant S. aureus. Coagulase negative staphylococci had a 36% resistance to flucloxacillin. With BEC cement there was a tendency to increased erythromycin resistance in CNS, but this did not achieve statistical significance.

At revision for infected TKR, different bacterial profiles were observed depending on prophylactic antibiotic usage. As CNS now causes > 40% of infections, patients undergoing revision TKR should have antibiotics effective against CNS until definitive results are available.

We reviewed 1567 elective knee replacements performed between 1980 and 1990, using either the Total Condylar prosthesis with an all-plastic tibial component, or the Kinematic prosthesis which has a metal tibial tray. The ten-year probability of survival was 92.1% for the Total Condylar design and 87.9% for the Kinematic. The difference was mainly due to 16 revisions required in the Kinematic series for fracture of the metal base-plate. This was the most common cause of aseptic failure in this group.

These fractures were strongly associated with a preoperative varus deformity (hazard ratio (HR) 8.8) and there was a slightly increased risk in males (HR 1.9) and in osteoarthritic knees (HR 1.8). In the nine fractures which occurred within four years of primary implantation (group 1), failure to correct adequately a preoperative varus deformity and the use of a bone graft to correct such a deformity were both strongly associated with fracture (HR 13.9 and 15.8, respectively). In eight fractures which occurred more than five years after primary replacement (group 2) we could detect no significant risk factors.

Early complications occurred in two patients after the 16 revision procedures for tray fracture. One had a deep infection and the other refracture of the tray.