There is currently no consensus on the use of suppressive antibiotic therapy (SAT) in prosthetic joint infections (PJI) (1). We describe herein the experience of a French Reference Centre for Complex Osteo-Articular Infections on use of oral cyclines (doxycline and minocycline) for SAT.

A retrospective analysis was performed on consecutive patients with PJI who received oral cyclines (doxycycline or minocycline) for SAT between January 2006 and June 2014. All patients had surgical management, followed by systemic antibiotic treatment and SAT instauration thereafter. Remission was defined as an asymptomatic patient with a functioning prosthesis.

Seventy-nine patients with a mean age of 63.8 ± 16.8 years were included. Sixteen patients (20%) had neoplasia, 9 (11%) diabetes mellitus, 10 (13%) rheumatoid arthritis, and 6 patients (8%) were receiving corticosteroids or chemotherapy.

There were 37 knee (47%), 36 hip (46%), 4 elbow (5%), and 2 shoulder (3%) infections, with a mean delay from implantation of 7.37 ± 6.94 months (range 1–27). Surgical management consisted in debridement and implant retention for 60 patients (76%), or in implant exchange for 19 patients (24%).

Main pathogens were coagulase-negative staphylococci (37%) and Staphylococcus aureus (41%); 23 patients had polymicrobial infection (29%).

The most frequent initial antibiotic regimens debuted before SAT were rifampicin combinations (70%). Mean duration of curative antibiotic therapy was 103 ± 75 days.

Indications of SAT were (i) patients unsuitable for or refusing further surgery (n=23), suboptimal (ii) surgery (n=26) or (iii) curative antibiotic therapy (n=11), (iv) complex orthopaedic surgery (n=11), and (v) immunosuppressive status (n=8). Seventy-three patients received doxycycline and 6 patients received minocycline as SAT (n=48). Mean SAT duration was 625± 536 days (range 30–2900), with a mean follow-up of 765 ± 572 days.

Adverse events were reported in 13 patients (16%), leading to SAT discontinuation in 5 (6%).

During follow-up, 59 patients were considered in remission (75%), and 20 failed including 13 relapses (16%) and 7 reinfections (9%). Among failure patients, 10 pathogens resistant to doxy/minocycline were identified, including 5 with acquisition of cycline resistance.

In our study, SAT with cyclines is associated to a 75% remission rate, with an acceptable tolerability.

Further studies are warranted to determine ideal regimens and optimal duration of SAT.

We would like to thank Dron Hospital and Lille University Hospital medical teams.

The authors declare that there are no conflicts of interest.

In France, 5% of men and 7% of women aged more than 60 years have a joint prosthesis (JP). The incidence of H-PJI following BSI remains unknown (1–2). The aim of this study was to determine prospectively the clinical characteristics of patients with JP and the incidence of H-PJI following a BSI.

A prospective observational multicentric study was performed in two French General Hospitals, from December 2012 to April 2015. Each patient with JP, in whom a BSI was diagnosed, was evaluated prospectively by an ID specialist. Data regarding clinical and microbiological characteristics were collected. A follow-up by phone call was performed monthly during 6 months to determine the incidence of H-PJI following BSI.

During the study period, 97 patients of mean age ± SD of 82.1 ± 10.4 years were identified, with a predominance of women (n=61). Nineteen patients (20%) had neoplasia, and 32 diabetes mellitus (33%). Most patients had one (n=61; 63%) or two JP (n=29; 30%); with a predominance of hip arthroplasty (n=77; 79%). Predominant pathogens were E. coli (n=41; 42%), S. aureus (n=23; 23%) and S. pneumoniae (n=8; 8%).

At the onset of BSI, the JP was concomitantly infected in 10 (10.3%) patients (including 8 S. aureus, 1 E. coli and 1 P. mirabilis), thus 87 were studied for the incidence of H-PJI following BSI of another source. Among these 87 patients, no H-PJI was detected, with a complete 6-month follow-up available for 29 patients (34%), incomplete follow-up for 26 patients (30%), loss of follow-up for 3 patients (3%), and death occurring in 29 patients (34%). The comparison between the patients with no H-PJI detected (« No Event Group ») and the deceased patients (« Death Group ») showed that patients of the « No Event Group » had a lower rate of neoplasia (14% vs 34%; P=0.025).

Our preliminary results show that patients with JP in whom a BSI occurred were old, and had a high mortality rate. In our study, the incidence of secondary H-PJI appears to be low, since no event was detected during the follow-up. The incidence of H-PJI may have been underestimated due to the high mortality rate.

We would like to thank Dron Hospital and Bethune Hospital medical teams.

The authors declare that there are no conflicts of interest.