Reduction in the colonization of central venous cannulae by mupirocin

In an in-vitro simulation of an intravascular cannula enclosed in a fibrin sheath, 0.03 mg1(-1) of mupirocin prevented significant colonization [greater than 15 colony forming units (cfu)] by two clinical isolates of Staphylococcus epidermidis and one each of S. saprophyticus, S. hominis and S. haemolyticus. This suggests that in vivo, where protein binding of mupirocin is 95-97%, 1 mg 1(-1) of mupirocin at the cannula surface would be required to prevent colonization. These results support the findings of our previously published prospective controlled trial, in which mupirocin applied to the insertion sites of 172 internal jugular cannulae reduced the rate of colonization of cannula tips to 5%, compared with 25% for the 186 controls (P less than 0.001). Of the 46 colonized cannula tips from 110 control patients, the same species was isolated from the skin of the insertion site in 67% and from the lumen flush in only 15%. Analysed by patient, mupirocin reduced the proportion of patients with colonized tips from 17% to 3% after 24 h of infection, and from 35 to 10% after 48 h (P = 0.002). The use of agar containing charcoal, as a mupirocin neutralizer, and the incubation of tip-culture plates flooded with the Oxford staphylococcus, gave no evidence of carry over of mupirocin onto cannulae removed from mupirocin-treated patients.