Diagnosing Catheter-Related Bloodstream Infection

To the Editor—I want to congratulate Bouza et al. [1] for their effort to solve part of the puzzle regarding the diagnosis of catheter-related bloodstream infection without catheter removal in critically ill patients [1]. I would like to make 2 comments to place the results in a broader perspective. The authors excluded arterial and Swan-Ganz catheters from the analysis, notwithstanding the fact that probably 150% of critically ill patients have more than just the central venous catheter in place. What happened with these other catheters? Were they removed in all cases, to exclude or identify the arterial line as a source of non–central venous catheter– related bloodstream infection? If these other catheters were removed per protocol, then the conclusion of the authors should be that the evaluated techniques for the diagnosis of catheter-related bloodstream infection are useful when all other intravascular catheters are removed. If they were not removed, then the authors cannot exclude the possibility that other catheters were the source of the so-called non–catheter-related bloodstream infection. Furthermore, even if these diagnostic techniques are truly accurate in differentiating catheter-related bloodstream infection from other bloodstream infections, the problem remains that, in the large majority of cases in this study, the indication for catheter removal was fever or some other symptom in a patient with negative blood culture results. A total of 159 of the 204 patients included in the study reported by Bouza et al. [1] did not have bloodstream infection. None of the evaluated diagnostic techniques will overcome this problem of clinical over-diagnosis of catheter-related infection, because the techniques can only be used to differentiate between sources of bloodstream infection. I think that the only correct conclusion that can be drawn from this study is that, in the small subset of patients with suspected catheter-related infection who turn out to have bacteremia (45 of the 204 patients included in the study [1]), techniques exist to differentiate catheter-related bloodstream infection from other bloodstream infections. Future studies should evaluate the safety of protocols that try to avoid catheter removal in patients in whom bloodstream infection is suspected but unconfirmed [2]. If this policy is proven to be safe, we will be able to avoid unnecessary catheter removal in patients with unexplained fever.