For every two or three genuine positive blood cultures, there is usually one which represents skin contamination (in my laboratory at least). As laboratory microbiologists, we often need to make an assessment about the likelihood of this fact, in order to guide the clinicians.

The “offending” microbes are usually coagulase negative staphylococci, corynebacterium species, propionibacterium species, and bacillus species.

However the isolates listed above do not always represent contamination. How can we assess the likelihood of such an isolate being genuine?

The following factors should be considered:

• Time to Positivity: A blood culture signalling positive after 10 hours is more likely to be significant than one which signals positive after 30 hours. Bear in mind though that some organisms, such as propionibacterium species, are going to take a long time to become positive whether they are contaminants or not.
• Number of bottles positive: If both bottles in the set are positive for the same isolate, then this increases the chance of the isolate being genuine. However remember that sometimes there will only be one bottle in the set (paediatric blood culture), and some organisms will often only grow in one bottle (bacillus-aerobic bottle, propionibacterium-anaerobic bottle)
• Other BC sets taken and whether they are positive: If 6/6 bottles from 3 sets are positive with a coagulase negative staphylococcus then start looking hard for a geuine infection. On the other hand if only 1 bottle out of 6 is positive, then chances of this being a contaminant are high.
• Clinical details- Prosthetic material on board: Coagulase negative staphylococci, corynebacterium species, and propionibacterium species all love clinging to and infecting prosthetic material. Has the patient got a prosthetic joint, a pacemaker, orthopaedic metalware, a venous or arterial line, etc, etc. if they do then think twice before dismissing blood culture isolates as skin contaminants.
• Clinical details- Immunocompromise: I was “stung” once with a Bacillus species in a blood culture, which turned out to be genuine, as the patient was immunocompromised. The clue was in the fact that the isolate was present in two separate blood cultures…

All of these points above are just pieces of evidence in the puzzle. I advocate a systematic approach to assessing the potential significance/insignificance of blood culture isolates. Even if takes a couple of minutes per blood culture, it can be potentially very rewarding.

Even better is to look at ways of reducing the contamination rate of blood cultures. This new device looks very promising, essentially reducing the contamination rate by “discarding” the first part of the collection (the blood culture equivalent of an MSU!). I would have thought that such a device could potentially save both time and money on both the clinical and laboratory side. If those funding silos could only be broken down, I would love to try it!

Michael