BackgroundStaphylococcus aureus (SA) colonisation is associated with development of bloodstream infection (BSI), with the majority of colonising and infecting strains identical by pulsed-field gel electrophoresis (PFGE). We examined SA colonisation in patients with SABSI to delineate better the relationship between the two.MethodsPatients with SABSI were swabbed in the nose, throat, groin, axilla and rectum. Isolates were typed using PFGE. Logistic regression was performed to determine factors associated with positive swabs.Results79 patients with SABSI had swabs taken. 46 (58%) had ≥ 1 screening swab positive for S. aureus; of these 37 (80%) were in the nose, 11 (24%) in the throat, 12 (26%) in the groin, 11 (24%) in the axilla and 8 (17%) in the rectum. On multivariate analysis, days from blood culture to screening swabs (OR 0.5, 95% CI 0.32-0.78, P = 0.003) and methicillin resistance (OR 9.5, 95% CI 1.07-84.73, P = 0.04) were associated with having positive swabs. Of 46 participants who had a blood sample and 1 other sample subtyped, 33 (72%, 95% CI 57-84%) had all identical subtypes, 1 (2%) had subtypes varying by 1–3 bands and 12 (26%) had subtypes ≥ 3 bands different. 30/36 (83%) blood-nose pairs were identical.ConclusionOverall, 58% of patients with SABSI had positive screening swabs. Of these, only 80% had a positive nose swab ie less than half (37/79, 47%) of all SABSI patients were nasally colonised. This may explain why nasal mupirocin alone has not been effective in preventing SA infection. Measures to eradicate non-nasal carriage should also be included.