Methods for Diagnosing Intravascular Device-Related Bloodstream Infections—And the Winner Is?
Abstract & Commentary
By J. Peter Donnelly, PhD, Clinical Microbiologist, University Hospital Nijmegen, The Netherlands, Section Editor, Microbiology, is Associate Editor for Infectious Disease Alert.
Dr. Donnelly is a consultant for Ortho Biotech, and does research for Janssen, Merck, Novartis, Numico, Pharmacia, and Pfizer.
Synopsis: Paired quantitative blood cultures are the most accurate means of diagnosing blood stream infection.
Source: Safdar N, et al. Meta-Analysis: Methods for Diagnosing Intravascular Device-Related Bloodstream Infection. Ann Intern Med. 2005;142:451-466. Erratum in: Ann Intern Med. 2005;142:803.
There exists no consensus about the best means for diagnosing bloodstream infections related to intravascular devices (IVD), leading Safdar and colleagues to conduct a formal meta-analysis using the data supplied by 51 of the 185 studies published between 1966 and July 31, 2004, that met the selection criteria. These criteria required studies to have compared a particular diagnostic method to a reference standard, and to have reported sufficient data to allow calculation of sensitivity and specificity. Eight methods were identified (see Table 2) and were divided into 3 requiring removal of the device, and the rest, which did not. Pooled sensitivity and specificity were calculated together with Q*, which is the upper leftmost point on the summary receiver-operating characteristic curve, with a maximum value of 1 and mean D, which is a log odds ratio. The paired quantitative blood culture was the most accurate test (Q* = 0.94 [95% CI, 0.88 to 1.0]), followed by quantitative blood culture drawn through the device (Q* = 0.89 [CI, 0.79 to 0.99]) and the acridine orange leukocyte cytospin test (Q* = 0.89 [CI, 0.79 to 0.91]). The most accurate catheter segment culture test was quantitative culture (Q* = 0.87 [CI, 0.81 to 0.93]), followed by semi-quantitative culture (Q* = 0.84 [CI, 0.80 to 0.88]). The pooled sensitivity and specificity exceeded 70%, but showed significant variation but, nonetheless, the negative predictive values all exceeded 99%. Importantly, the positive predictive value of all tests increased greatly when the pretest clinical probability or prevalence was high (see Figure 1 panel A), whereas the negative predictive value was only marginally affected (see Figure 1 panel B). Safdar et al emphasized the limitations of their analysis, including the fact that some studies only included patients already suspected of having IVD-related blood stream infections, whereas others looked only at the devices after removal. Hence, diagnosis should only be employed when there is suspicion of infection.
Commentary
Safdar et al are to be congratulated for taking the bull by the horns and trying to distil the essence out of the mists of confusion surrounding the diagnosis of IVD-related bloodstream infection. Aside from the fact that the term bloodstream infection is a misnomer, since the blood generally only transports the micro-organisms from one site to another, the issue of whether a device can diagnose the source of infection is important since a mistake can be costly. On the one hand, failure to remove a device which is infected results in avoidable morbidity, can lead to overuse of antibiotics, and puts the patient at risk of developing metastatic infection. On the other hand, premature removal of a device may result in difficult venous access at a critical time, forcing the caretakers to either draw blood more frequently from a peripheral vein, to use alternative devices that are less than optimal, or even to place a new device at a less than ideal site. It is also true to say that the issues of infection related to IVDs tends to arise when patients are critically ill with fever and sepsis syndrome and their caretakers are seeking the cause and usually also reassurance that the device is innocent.
Historically, microbiologists have been wary of the idea of drawing blood via the device for culture because of the risk of contamination. This is not unreasonable given that coagulase-negative staphylococci (eg, Staphylococcus epidermidis) are skin commensals, and the most frequent causes of bacteraemia related to intravascular devices. Furthermore, microbiologists have known for years that paired quantitative cultures of blood drawn through the device and from a peripheral vein simultaneously (ie, within 10 minutes of each other) would give optimum results. However, this was always considered too expensive compared with culturing catheter segments, usually the tip, and not much more accurate. The analysis of Safdar et al gently knock this argument off its pedestal since catheter segment cultures show the poorest specificity when prevalence is low. The clinicians will now feel fully vindicated, as they have long believed that catheter segment cultures are examined after the fact, and effectively vote with their feet in drawing blood through the device for culture. Unfortunately, the method used is qualitative and only marginally more specific than catheter segment cultures.
Scientifically and technically, the solution would seem obvious. Draw enough blood through the IVD and from a peripheral vein at the same time and submit it for quantitative culture. If only life were that simple. There are but 2 methods currently available for achieving this—pour plates and lysis centrifugation. Pour plates are cheap and simple but need a water bath since the agar has to be molten at ~45°C, mixed in a ratio of about 1 to 20 with fresh blood to a volume of about 20 mL, poured into a Petri dish, and allowed to set, all under aseptic conditions. The alternative technique requires drawing blood into a vial containing a cocktail of chemicals to lyse the blood cells, and then removing small amounts to spread on the surface of an agar plate. Both methods require skill and time and are sensitive to contamination. Nevertheless, it is to be hoped that this article will help give an impetus to microbiology laboratories that are reluctant to at least undertake prospective studies of paired quantitative blood cultures to see whether they really are cost effective. Failing that, Safdar et al have at least provided evidence that this approach probably offers timely and useful information to guide the clinicians in properly managing IVD related blood stream infections.
There exists no consensus about the best means for diagnosing bloodstream infections related to intravascular devices (IVD), leading Safdar and colleagues to conduct a formal meta-analysis using the data supplied by 51 of the 185 studies published between 1966 and July 31, 2004, that met the selection criteria.You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content