In the Netherlands, less than 1% of clinical isolates of Staphylococcus aureus are methicillin-resistant (MRSA). A national search and destroy policy prevents MRSA from becoming endemic. Some MRSA outbreaks cannot be related to patients at risk for MRSA carriage. This study was designed to measure the prevalence of MRSA among patients without risk factors for MRSA carriage at the time of admission to the hospital. In four Dutch hospitals, patients admitted to non-surgical departments in the period 1999-2000 were screened for MRSA nasal carriage. Nasal swabs were streaked on 5% sheep blood agar (BA), submerged in a selective broth, and incubated for two to three days at 35 degrees C. Colonies suspected of being S. aureus were identified with an agglutination test. Susceptibility testing was performed by an automated system and additional oxacillin disk diffusion. Methicillin resistance was confirmed by a DNA hybridization test and mecA PCR. MRSA strains were genotyped by pulsed-field gel electrophoresis (PFGE). Twenty-four percent (2332/9859) of the patients were S. aureus nasal carriers. Only three (0.03%) patients were MRSA carriers. These patients were not repatriated, nor known to be MRSA carriers before screening. Genotyping revealed that the strains were not clonally related and were not related to MRSA outbreaks in the hospital where the patients were admitted. We conclude that at routine admission to a Dutch hospital (excluding high-risk foreign admissions) the MRSA prevalence is low (0.03%), due to the Dutch search and destroy policy and restrictive antibiotic prescribing.