Difficult to treat: are there organism-dependent differences and overall risk factors in success rates for two-stage knee revision?

Failure after two-stage procedure for periprosthetic joint infection (PJI) is a rare, but devastating complication. Some authors assume a correlation of underlying organisms and recurrence rate. Methicillin-resistant Staphylococci (MRS) and other organisms (quinolone-resistant Gram-negative bacteria, rifampicin-resistant Staphylococcus, Enterococcus, and Candida) are meant to be "difficult to treat" (DTT) with an inferior outcome for two-stage revision. In addition to the type of bacteria, some more risk factors seem to be present. The aim of this study was (1) to detect a difference of reinfection rates between reinfection-causing groups of bacteria ["difficult to treat" (DTT), "easy to treat" (ETT) and methicillin-resistant staphylococci (MRS)] after a two-stage procedure, and (2) find overall risk factors for reinfection in a standardized long (spacer insertion for at least 6 weeks) two-stage procedure for periprosthetic knee infection. One hundred and thirty-seven two-stage revisions for periprosthetic knee infection were performed at one tertiary referral center. Finally, 96 patients could be included for analyses. Possible risk factors (comorbidities, prior surgery, etc.) and the types of organisms were documented. Quinolone-resistant Gram-negative bacteria, rifampicin-resistant Staphylococcus, Enterococcus, and Candida were classified as "difficult to treat" (DTT). Methicillin-resistant Staphylococci were summarized as "MRS", all other organism are summarized as "easy to treat" (ETT). Statistical analyses included univariate analysis (t test, Fisher's exact test, Chi square test) and logistic regression analysis. There were no statistical significant differences in recurrent infection rates between organism groups (DTT vs. ETT, p = 0.674; DTT vs. MRS, p = 0.705; ETT vs. MRS, p = 0.537). Risk factors seem to be "need of revision after first stage" (p = 0.019, OR 5.62) or completed second stage (p = 0.000, OR 29.07), numbers of surgeries (p = 0.028) and alcohol abuse (p = 0.019, OR 5.62). Revision needed during or after a two-stage exchange, numbers of surgeries and alcohol abuse are risk factors for recurrence, a different recurrence rates between organism-groups cannot be shown. The absence of significant differences in recurrence rates points to the importance of the individuality of each periprosthetic infection case: a reduction of necessary surgeries (with a thorough debridement, appropriate antibiotic addition to spacers) and the control of comorbidities (alcohol abuse) appear to be essential components of a two-stage exchange.