Hospital-acquired bacteraemia--surveillance and guidelines for practice.

In the Surveillance Report, bacteraemia was not defined specifically within the document. However, it was implied that bacteraemia was present if blood cultures were positive. No clinical information concerning the effect of bacteraemia on the patients or the degree of haemodynamic support required was described. Nor was the movement of patients between specialties indicated. This information would have been helpful to clinicians because it would demonstrate the increased severity of illness experienced by patients and the concomitant increase in services required to meet these needs, although of course prevention should be the key response to the data provided. It is evident from the two reports that infection of the bloodstream is described utilizing different terms, i.e. bacteraemia and catheter-related bloodstream infection. This is potentially confusing and efforts should be made to encourage the use of consistent terminology across specialties, e.g. infection control, critical care, oncology The potential for confusion also arises where evidence is pooled utilizing both CRI and CR-BSI as endpoints for the research reviewed. The recommendation associated with flushing the CVC with heparinized saline solution does not consider a patient's coagulopathy to be a contraindication, only the manufacturer's recommendations. This appears to be a limitation of the guidelines and may place the patient at risk. The recommendation also does not indicate if this is only to be performed when the catheter or lumen is not in use, or if heparin should be administered if it is in use. Finally, the categories used to denote the level of evidence are not defined in the document. It can be assumed that Category 1 relates to randomized controlled trials, which demonstrate homogeneity or/and narrow confidence intervals; whilst categories 2 and 3 relate to cohort studies and pooled data. It should be expected that a document of this nature should establish criteria or refer the reader to the primary source for categorization. Eggimann and Pittet (2000) have undertaken an excellent review of central venous catheter related infections in intensive care units. The evidence accrued and recommendations made mirrors the two reports outlined above. It is imperative we take action to reduce the incidence of CR-BSI and it is hoped that this Editorial has provided a basis for discussion and action; and will stimulate debate.