Résumé : Clin Microbiol Infect ABSTRACT: The increase in the number of methicillin-resistant Staphylococcus aureus (MRSA) infections in children has prompted paediatricians to broaden th empirical treatment of common community-onset (CO) infections in children in several countries. Most European countries have reported low rates of CO-MRSA infection, but limited data on paediatric CO-MRSA infections are available. A prospective study was conducted from January 2002 to December 2004 in Brussels. CO-MRSA was defined as MRSA first detected by culture within 48 h of admission or in outpatients. Clinical and epidemiological data were recorded. CO-MRSA strains were genotyped by pulsed-field gel electrophoresis and multilocus sequence typing. Staphylococcal chromosomal cassette mec, toxin (Panton-Valentin leukocidin (PVL), toxic shock syndrome toxin 1, and Eta/b), enterotoxin and antibiotic resistance genes were detected by PCR. The antibiotic resistance phenotype was determined by disk diffusion. S. aureus was isolated in 1681 children. Among these, 107 harboured MRSA. Fifty-one children were colonized or infected by CO-MRSA, 20% of whom had no healthcare exposure. Twelve infants <3 months old and five cystic fibrosis patients were colonized. None of the 22 infected patients (59% with acute otitis media and 36% with skin and soft tissue infections (SSTIs)) required hospitalization. Two-thirds of them failed to respond to empirical antibiotic therapy. The 37 characterized CO-MRSA strains were genetically diverse. Most of them had healthcare-associated genotypes. Only six strains were PVL-positive, all of which were ciprofloxacin-susceptible and more common in children with SSTIs (p 0.001). CO-MRSA remains uncommon in our paediatric population. So far, there is no need to modify the empirical treatment of common S. aureus infections. Monitoring of MRSA rates in S. aureus CO infections remains mandatory, and further investigation is warranted to establish the source of colonization in young infants.