Survival rate after septic total knee arthroplasty- a comparison of two treatment algorithms

AIMS AND OBJECTIVES: To date there exists no golden standard of treatment of late periprosthetic joint infection. Different forms of treatment comprise single-stage or two-stage or multiple revisions. Respectively, mid- to long-term results are rare. MATERIALS AND METHODS: We retrospectively analysed 66 patients with septic total knee arthroplasty treated by two different therapy algorithms: Until 2009 patients were treated routinely by implant removal and implantation of an articulating spacer in combination with at least 6 weeks of antibiotic administration. After an antibiotic-free period of two weeks persistent infection was ruled out by punctation. In case of negative antimicrobial cultures after another 14 days patients were treated by reimplantation (n=36, group 1). After 2009 persistent infection was ruled out by open revision and change of the articulating spacer instead of a punctuation. In case of negative antimicrobial cultures after another 14 days patients were treated by reimplantation (n=30, group 2). After a follow-up of 61 months (group 1), respectively 36 months (group 2), patients and their family doctors were contacted by phone concerning the survival rate of the new implant. RESULTS: In group 1 persistent infection was ruled out in every patient by punctuation. During reimplantation tissue samples of 8 patients (22%) showed positive microbiological results again, within 4 patients with a change of the underlying bacteria. After a follow-up of 61 months 31 patients showed no signs of reinfection of their implant and were not operatively revised (86%). In group 2 all patients were treated with open revision and change of the articulating spacer for at least once. 8 patients showed a change of the underlying bacteria during these procedures. During reimplantation tissue samples of 8 patients (27%) showed positive microbiological results again. After a follow-up of 36 months 27 patients showed no signs of reinfection of their implant and were not operatively revised (90%). The difference between both groups was not significant (p>0.05). Concerning the overall complication rate no significant differences between both groups could be found (p>0.05). CONCLUSION: Concerning reinfection rate we could not detect significant differences between both groups. Patients of group 2 were treated operatively significantly more often and showed a significantly more frequent change of the underlying bacteria. If these results are confirmed in big cohort studies the elaborate treatment algorithm of open revision and change of the articulating spacer might become dispensable.