OBJECTIVE To study the possible role of contaminated environmental surfaces as a reservoir of methicillin-resistant Staphylococcus aureus (MRSA) in hospitals. METHODS A prospective culture survey of inanimate objects in the rooms of patients with MRSA. METHODS A 200-bed university-affiliated teaching hospital. METHODS Thirty-eight consecutive patients colonized or infected with MRSA. Patients represented endemic MRSA cases. RESULTS Ninety-six (27%) of 350 surfaces sampled in the rooms of affected patients were contaminated with MRSA. When patients had MRSA in a wound or urine, 36% of surfaces were contaminated. In contrast, when MRSA was isolated from other body sites, only 6% of surfaces were contaminated (odds ratio, 8.8; 95% confidence interval, 3.7-25.5; P < .0001). Environmental contamination occurred in the rooms of 73% of infected patients and 69% of colonized patients. Frequently contaminated objects included the floor, bed linens, the patient's gown, overbed tables, and blood pressure cuffs. Sixty-five percent of nurses who had performed morning patient-care activities on patients with MRSA in a wound or urine contaminated their nursing uniforms or gowns with MRSA. Forty-two percent of personnel who had no direct contact with such patients, but had touched contaminated surfaces, contaminated their gloves with MRSA. CONCLUSIONS We concluded that inanimate surfaces near affected patients commonly become contaminated with MRSA and that the frequency of contamination is affected by the body site at which patients are colonized or infected. That personnel may contaminate their gloves (or possibly their hands) by touching such surfaces suggests that contaminated environmental surfaces may serve as a reservoir of MRSA in hospitals.