Second, arguably the major problem with blood cultures is the relatively high proportion of false-positive results. The authors discussed this issue quite nicely in their article. In our ICU, to better deal with this pervasive problem, we have opted to define a “blood culture set” not as two blood culture bottles from one venipuncture site (as proposed by Shafazand and Weinacker1
), but as two anatomically separate venipuncture sites obtained at one point in time, with one or two blood culture bottles being filled from each site. In other words, in our ICU, a blood culture set refers to a point in time when blood cultures were obtained, and each set includes a minimum of two separate venipuncture sites. For example, if a patient is to have three blood culture sets drawn, this means that he or she will be sampled at three different points in time over a 24-h period, each time having blood drawn from two separate venipuncture sites (a total of six venipunctures for all three sets). Because anaerobic bottles are not routinely needed, for most patients the number of blood culture bottles to be processed would remain the same, but by this proposed definition each “set” would have its own built-in quality control: ie, a true-positive blood culture result requires that the culture from both sites sampled at the same point in time be positive for the same organism(s).