The object of this study was to investigate the efficacy of a methicillin-resistant Staphylococcus aureus (MRSA) multisite carriage decolonization in 32 hospitalized carriers--25 from surgical and seven from medical wards. Twenty-four of the patients had wounds (e.g. chronic ulcers, surgical sites) and 17 were spinal cord injury patients. Decolonization was performed by intranasal application of mupirocin, combined with an octenidine dihydrochloride bodywash over a period of five days. Samples from the nose, forehead, neck, axilla and groin were taken 24-48 h before beginning decolonization (sample point I, N=32) and 24-48 h afterwards (sample point II, N=32). Further samples, were taken seven to nine days after the procedure (sample point III, N=25). Contact sheep blood agar plates (24 cm2) were used to quantify MRSA colonies on forehead and neck. MRSA from other sample sites was determined semi-quantitatively. All patients were proven to be MRSA positive at one or more extranasal site(s); 18.8% did not have nasal carriage. The overall decolonization rate for all sites was 53.1% (sample point II) and 64% (sample point III), respectively. The reduction was significant for every site, showing a rate of 88.5% for nose (II, III) and of 56.3% (II) and 68% (III) for all extranasal sites together. Of 32 patients, a median of 6.5 cfu MRSA/24 cm2 was obtained for the forehead before decolonization and 0.5 cfu MRSA/24 cm2 for the neck. A significant reduction (0 cfu MRSA/24 cm2) from both sites was shown after treatment. Before decolonization procedures, median MRSA levels for the nose, groin and axilla were 55, 6 and 0 cfu/swab. After treatment, MRSA from each of these sites was significantly reduced. We conclude that nasal mupirocin combined with octenidine dihydrochloride whole-body wash is effective in eradicating MRSA from patients with variable site colonization.