Assessment of a new hub design and the semiquantitative catheter culture method using an in vivo experimental model of catheter sepsis.

An in vivo model of hub-related catheter sepsis in rabbits is reported. The model was used to investigate the protection offered by a new hub design against external contamination by Pseudomonas aeruginosa or Staphylococcus epidermidis and to reassess the diagnostic value of the semiquantitative culture method in bacteremia of endoluminal origin. Contamination of conventional Luer-Lock connectors was followed by clinical sepsis, positive catheter segment cultures, or both, whereas contamination of the new hub was followed by complete protection. Clinical and bacteriological discrepancies observed between contamination with P. aeruginosa and S. epidermidis suggest that the virulence of microorganisms may account for differences in the natural history of hub-originated catheter sepsis. The semiquantitative extraluminal method for catheter culture yielded less than 15 CFU in three animals with proven bacteremia and should not be used as the "gold standard" to define catheter-related bacteremia.