In France, the incidence of native joint infections is about 10 per 100 000 person-years, most commonly caused by S.aureus followed by b-haemolytic streptococci. French and international antibiotic guidelines, based on expert advice and retrospective studies, recommend intravenous antibiotics for two weeks, then oral for 4 weeks without evident link between intravenous, prolonged oral treatment and cure. Long term exposure to antibiotics increases bacterial resistance, a major problem of public health. Several studies show that serious infectious can be treated safely by a shorter treatment and with oral antibiotics. There is no randomized controlled trial to establish the duration of antibiotics in native joint infections. Moreover, no consensus prevails on the administration route and duration of antimicrobial therapy. Although most clinicians acknowledge the interest of oral antibiotics and shorter treatment duration, randomized controlled trials are necessary to evaluate this practice. The SHASAR project aims to evaluate whether a shorter antibiotic treatment (3 week treatment) is safe and not inferior to the conventional 6 week treatment in native joint infections. If successful, this would represent a major advance in terms of patients' quality of life; decreased rate of health-care-related infections and complications, bacterial resistance and cost.

This trial is a nationwide, non inferiority, multicenter French randomized, open-label, controlled trial comparing two treatment durations, 3 versus 6 weeks, in septic arthritis of native joints. Patients who fulfill inclusion criteria will be randomized between day0 and day5. Day 0 is the time when effective intravenous antibiotic treatment is started. Effective treatment is defined by active antibiotics on the identified bacteria according to the susceptibility. The randomisation (1:1 ratio) will be stratified on early planned drainage. Follow-up will include 6 visits and will consist in clinical, biological, radiological, health quality of life (EQ-5D-3L score) and adverse events record. After collection, validation of data and population description, analysis will be conducted on the per-protocol population (patients receiving the planned duration of antibiotic +/-3days). This study will not be carried out blind for feasibility reasons. According to statistical considerations (by accounting for 5% of subjects lost to follow-up), the required sample size will be of 350 patients overall.

Inclusion Criteria:

• Septic arthritis of native joint defined by: 1) clinical signs (hot and/or swollen and/or tender and painful joint (measured by a visual analogue scale)) and 2) microbiologically confirmed pyogenic arthritis (microorganism cultured in the synovial fluid joint or in the blood with complete bacterial susceptibility)
• Patients aged of 18 years or older
• Informed, written consent obtained from the patient
• Patient having the rights to French social insurance

Exclusion Criteria:

• Prosthetic joints
• Septic arthritis in the past 12 months
• Osteomyelitis
• Diabetic foot
• Septic choc
• Arthritis due to bacteria resistant to available oral antibiotics
• Arthritis due to the following microorganisms: Mycobacterium, fungi, Brucella, Borrelia, Neisseria gonorrhoeae, Neisseria meningitidis, Nocardia, Mycoplasma spp, Pseudomonas aeruginosa.
• Glomerular filtration rate < 30ml/min/1,73m2
• Neutrophils < 500/mm3
• Difficulties regarding compliance with oral antibiotics
• Contraindication to oral antibiotics necessary to treat joint infection
• Pregnancy or lactating woman
• Curator or guardianship
• Participation in other interventional research during the study