Purpose of the study: Factors affecting the course of Staphylococcus aureus (SA) infected total hip arthroplasty (THA) and total knee arthroplasty (TKA) are poorly understood.

Material and methods: The microbiology laboratory database on patients hospitalised from June 2001 to June 2006 for medical and surgical treatment of an SA infected THA or TKA were reviewed. Data collected were: clinical, radiological, and biological (CRP) status known at least two years after the end of the antibiotic treatment. Success was defined as normal findings in all three domains at last follow-up and failure for any other situation.

Results: The series included 87 patients (43 male and 54 female, mean age 66.5 years, 62 THA, 35 TKA). Material was removed in 51 patients (52.6%). The infection was related to met-R SA in 14 patients (14.4%) and polymicrobial infection in 24 patients (24.7%). Mean duration of intravenous treatment was 7.3±9.4 days and total duration of antibiotic treatment was 115.472.2 d. At mean follow-up of 54.1±19.4 d, the overall success was 62/97 (63.9%). The only parameters significantly associated with failure were delay to management after first signs revealing longer infection (67.9 days vs 144.8 days) and an antibiotic therapy not adapted bacteriological results to immediately after surgery. (28.6% vs 3/28; 4.8%). In this series, met-R SA was not a risk factor of failure.

Conclusion: This work suggests that delayed surgical management and quality of antibiotic treatment for the first revision are factors affecting the prognosis of SA infected THA and TKA. These results are in favour of a multidisciplinary approach to care for these patients.

Purpose of the study: Because of its specific properties, linezolid (LZD) is an alternative to glycopeptides for the treatment of BJI causes by resistant Gram-positive cocci. The efficacy and tolerance of long-term (> 28d) administration remain to be determined with precision.

Material and methods: This was a retrospective analysis of the files of patients who received long-term LZD for BJI with a post-treatment follow-up of at least twelve months.

Results: Between January 2001 and December 2003, 68 patients (mean age 54.6 years) received LZD (600 mg/12h) intravenously then orally for a mean duration of 14.6 weeks on average (range 6–36 weeks). The orthopedic implants were infected in 31 patients (45.6%, 27 prostheses and 4 external fixators). A fistulization was diagnosed in 29 patients (42.6%). Medical and surgical treatment was undertaken in 51 patients (75%). Germs isolated from deep samples were staphylococci (n=57, 51 meti-R), and strepotococci (n=11). A polymicrobial infection was recorded in six patients (8.8%). Clinical success was achieved in 52 patients (76.5%) at the end of treatment, a result which was sustained for 48 of them to mean follow-up of 16.8 months (range 12–48). Sixty-three episodes of side effects were noted: anemia (n=29), peripheral neuropathy (n=9), nausea (n=14), headache (n=2), neutropenia (n=2), and thrombopenia (n=1), leading to discontinuation of LZD in 23 patients (33.8%) on average 11.9 weeks (range 6–36) after onset of treatment. Fifteen patients (22.1%) required transfusion.

Discussion: Among the patients in this series with BJI, most caused by a multiresistant germ, the rate of success was satisfactory (48/68, 70.6%) with LZD. Treatment was associated with a high frequency of serious adverse effects leading to discontinuation in one-third of patients.

Conclusion: Long-term treatment of BJI with LZD appears to be effective, including in patients with infected orthopedic material. Conversely, tolerance problems can limit is use to infections caused by mul-tiresistant Gram-positive cocci unresponsive to other compounds.