DIFFICULT ASSESSMENT: Bone and joint infections are difficult to treat. Therapeutic success depends greatly on the diffusion of antibiotics into bone tissue. Few studies have been devoted to this subject and the variable nature of those reported hinders interpretation. Bone biopsies are generally obtained during orthopedic procedures. Antibiotic administration routes vary although intravenous infusion predominates. Agar gel diffusion is generally used for antibiotic assays but methodology varies depending of the study. The most recent reports use high-performance liquid chromatography. DIFFUSION STUDIES: The different studies examining antibiotic diffusion in bone tissue describe three classes: good diffusion (greater than 30%), moderate diffusion (between 15% and 30%), and low diffusion (less than 15%). Antibiotics in the good diffusion class include fluoroquinolones, teicoplanin, macrolides, rifampicin and trimethoprime. Antibiotics with moderate bone diffusion are ureidopenicillins, second and third generation cephalosporins, aminoglycosides, clindamycin, fosfomycin and vancomycin. Those with low bone diffusion are aminopenicillins, penicillin M and first generation cephalosporins. No data is available on the bone diffusion of pristinamycin. DATA INTERPRETATION: The clinical impact of these classifications must be interpreted with precaution when considering bone and joint infections as they were established on the basis of pharmacokinetic studies and not clinical trials. They would however appear to be useful in guiding antibiotic prophylaxis for orthopedic surgery in protocols with administration conditions and concentration goals similar to the experimental conditions. PRACTICAL ATTITUDES: These laboratory results could be used in clinical practice by comparing the MIC50 of the germs regularly encountered in bone infections (staphylococci, streptococci including enterococci, Gram negative bacilli including P. aeruginosa and H. influenzae) with concentrations obtained in the different studies, i.e. by calculating the inhibitor coefficient (IQ) of each antibiotic for each susceptible germ. This gives a classification by efficacy (excellent IQ > 10, good 1 < IQ < 10, poor IQ < 1) useful for guiding antibiotic choice in the difficult situation of bone and joint infection.