[Our experience with revision total knee arthroplasty]

In the period from 1990 and June 2003, 885 total knee replacements were performed at the orthopedic ward of the Ceske Budejovice Hospital. Of these, 19 (2.14 %) patients underwent revision surgery; in addition, 25 patients who had had primary surgery in other hospitals were operated on. Of these 44 patients, 25 were followed up and evaluated. The aim of the study was to evaluate the pre-operative treatment including examination for bacterial infection by cultivation, the selection of an optimal procedure (one- or two-stage operation, surgical approach and implantation technique) and postoperative therapy.

The 25 evaluated patients were followed up for an average of 34.5 months (range, 6-109 months) after reimplantation. Loosening occurred in most of the commonly used types of primary implants. The group comprised five men and 19 women; the average age at the time of reoperation was 70 years (range, 51-78 years). Ten patients had repeat surgery on the left and 14 on the right knee at an average of 43.5 months (range, 4-120 months) after primary surgery. Fourteen patients were treated by one-stage and 10 patients by two-stage surgery. The Genesis system (Smith Nephew) was used in 13 patients, Sigma PFC revision implant (Johnson and Johnson) in eight, Walter-Motorlet implant (reuse of the primary implant at the time when a revision system was not available) in one and external fixator in five patients.

Revision arthroplasty was indicated on the basis of clinical symptoms and X-ray, scintigraphic and biochemical (CRP, WBC, FW) examinations. The use of recently adopted methods (procalcitonin, orosomucoid, alpha-1-antitrypsin, beta-2- macroglobulin, ceruloplasmin, PCR and PET) was not evaluated because of short-term applications. Patients in whom infection or colonization of the implant was suspected were treated by two-stage reimplantation, using a canalized spacer with a stem and a patellar pelota made of antibiotic-loaded cement. The average time between implant removal and reimplantation was 108 days (range, 60-244 days). Each removed implant was placed in a culture medium for 5 to 7 days. This resulted in a high occurrence of positive cultivation results even in the patients who, on the basis of previous examination, were first considered to have had aseptic loosening and had undergone one-stage surgery. All patients with positive tests received long-term antibiotic therapy, usually a combination of ciprofloxacin and rifampicin, according to the cultivation results.

Out of 14 one-stage reimplantations (indicated for by the negative results of all laboratory examinations), implant colonization was recorded in five cases, with a coagulase-negative staphylococcus being the most frequent infectious agent. No recurrent infection was found after the long-term antibiotic course. One patient with the implant infected with Staphylococcus aureus underwent primary arthrodesis. Out of 10 two-stage reimplantations (in patients with positive laboratory tests), recurrent infection was found in two cases and was caused by a pathogen different from the original one. The patients were treated by arthrodesis. Good outcomes, defined as a functional total knee replacement free from infection at least 6 months after reimplantation, were achieved in 79 % of the patients. Better functional results were obtained by onestage surgery. Patients with concomitant rheumatoid arthritis had aseptic loosening more frequently, and patients with impaired immunological status, due to diabetes mellitus, cytostatic drug or corticosteroid administration, more often showed septic loosening.

The 2.14 % loosening of total knee arthroplasty in our patients (19 out of 885) can be considered a good result. A functional joint was achieved in all patients (100 %) with aseptic loosening and in 69 % of those with infected or colonized implants. The results of routine biochemical tests and bacteriological cultivation did not allow us to distinguish aseptic from septic loosening with certainty. Therefore, we adopted new screening markers (PCR and PET) and a new method of microbiological examination of the removed implant and collected tissue. However, we could evaluate the role of these specialized tests only on the basis of literature data, since we had only short-term experience with them ourselves. Our results with the treatment of early loosening of total knee arthroplasty suggest that patients benefit more from the two-stage procedure.

We strongly emphasize the employment of all possible means to prevent loosening, i. e., to use an appropriate surgical technique for primary implantation, to observe aseptic principles and to administer antibiotic therapy in conditions suspected of bacteremia. The shorter the interval between the onset of complaints and the reimplantation, the better results. Early loosening should be treated by two-stage surgery. Our method of bacteriological examination gives good results. Because of complexity of the problem, patients with a loose knee prosthesis should be referred to orthopedic departments with experienced and skilled surgical teams and high-quality examination facilities.With the observation of appropriate procedures, there is a great chance of achieving good results. Arthrodesis is still regarded as a justified "salvage" operation, particularly in cases with pre-operative findings of Staphylococcus aureus. Procedures for repeat surgery following the failure of a reimplanted joint have so far yielded doubtful results and still await further development.