sign in sign up
<<
Home , Central venous catheter, Sepsis

J Clin Pathol 1999;52:165–172 165 Diagnosis of central venous related —a critical look inside J Clin Pathol: first published as 10.1136/jcp.52.3.165 on 1 March 1999. Downloaded from

B M Dobbins, P Kite, M H Wilcox

The use of central venous for the Definitions administration of pharmaceutical agents, in- Perhaps one of the greatest diYculties in cluding regimens, inotropic reviewing the subject of support in the intensive care setting, intra- is the large variation in what is venous nutrition, cardiac , and as a considered to be an “infected catheter.” This means of maintaining long term , problem appears to have arisen as a result of has increased dramatically in the past three the many methods which have been described decades. Complications associated with central to culture catheters. In addition to the venous catheterisation include those associated multitude of catheter culture techniques much with insertion—for example, , debate still exists as to precisely what is a haemorrhage, nerve injury, catheter tip mis- significant quantity of bacterial growth. Cen- placement, and cardiac —and tral venous catheters are inserted through and those associated primarily with longer term reside in close proximity to skin containing approximately 105–106 /cm2. Further- use, including and infection. Cath- more, approximately 108 skin scales/person are eter related represent by far the shed daily, about 10% of which contain bacte- greatest risk associated with the use of central ria. Hence, separating infected, colonised, and venous catheters, and the rate of catheter contaminated central venous catheters can be related sepsis is variably reported to range from 1–4 extremely problematic. Furthermore there is 4% to 14%. The magnitude of this variance no gold standard method whereby all tech- reflects true diVerences in the of niques for the diagnosis of central venous cath- catheter related sepsis in some patient eter infection can be compared, and thus the groups—for example, the rate of catheter vast majority of sensitivities and specificities related sepsis in patients managed in quoted should be critically assessed and cannot intensive care is approximately 15-fold higher be taken at face value. Despite these shortcom- than in those with respiratory .5 Infec- ings, there is a generally accepted definition of tions associated with central venous catheters catheter related sepsis (or catheter related therefore represent approximately 30 000 and bacteraemia) which requires the following

400 000 cases in the and the three criteria to be present: http://jcp.bmj.com/ USA, respectively, each year.6 However, diVer- (1) A significantly positive catheter culture ences in the reported incidence of catheter (although the definition of “significant” is related sepsis also result from a lack of contentious). standardisation in diagnostic approach. (2) A positive peripheral culture taken Catheter related sepsis is associated with sig- before catheter removal. nificant morbidity and mortality, and with case (3) The same isolated in both 7 (1) and (2).

fatality as high as 10–20%. Major complica- on October 1, 2021 by guest. Protected copyright. tions of sepsis were reported in 32% of cases in By insisting on the presence of an associated one series.8 There has been a two- to threefold peripheral bacteraemia this allows for more increase in the cases of primary nosocomial accurate comparisons of methodology and in the last decade,9 the diagnosis of catheter related sepsis. However, the significance of a colonised catheter in the large proportion of which have been attributed absence of a systemically proven infection can to catheter infections. The incidence of hospi- be strongly debated. While it is accepted that a tal acquired infection in a recent surveillance positive catheter culture in the absence of study was sevenfold higher in patients with an Department of 10 peripheral bacteraemia may occasionally repre- invasive device. Catheter related sepsis repre- sent either poor peripheral blood sampling or Microbiology, sents a significant burden to the health service, University of Leeds, even a transient fall in peripheral bacterial load Leeds LS2 9JT, UK and the excess hospital cost associated with at the time of sampling, it is the only method by P Kite these bloodstream infections has been esti- which one can truly compare all of the methods 11 M H Wilcox mated at $40 000. Another report calculated for the detection of catheter related sepsis. the cost of a single episode of catheter related Department of Unfortunately, contamination by sepsis in patients on an skin is common, and recent , University of 12 Leeds (ICU) as up to $28 000. Episodes of catheter studies have highlighted the low positive B M Dobbins related sepsis cause a major proportion of the predictive value of blood cultures positive for septicaemias due to coagulase negative staphy- coagulase negative staphylococci.13 14 For ex- Correspondence to: lococci, , and spp. Dr Wilcox. aureus ample, of 89 blood cultures positive for skin email: This review aims to discuss contentious issues flora, 91% involved coagulase negative staphy- [email protected] relating to the aetiology and diagnosis of cath- lococci, and the incidence of significant and Accepted for publication eter related sepsis, and to challenge some indeterminate coagulase negative staphylococ- 2 November 1998 beliefs using recently available data. cal bacteraemia and of contamination was 166 Dobbins, Kite, Wilcox

found to be 25%, 12%, and 73%, respectively.14 peripheral blood should not theoretically be While part (3) of the above definition of cath- governed by flow through the catheter. Quali- eter related sepsis states that the “same” tative blood cultures are readily available in all J Clin Pathol: first published as 10.1136/jcp.52.3.165 on 1 March 1999. Downloaded from microorganism is isolated from the central hospitals and have the advantage of being both venous catheter and from peripheral blood cul- relatively inexpensive and easy to process. tures, it is often assumed that bacteria sharing However, quantitative blood cultures are con- the same antibiogram or biotype are indeed the sidered to be both more accurate and less likely same; we have found that of 21 coagulase to produce false positive results owing to negative staphylococcal pairs isolated from contamination during sampling. Quantitative peripheral blood cultures and central venous analysis of blood samples taken from adults catheters, five (24%) were in fact distinguish- and children clearly indicates that the magni- able by DNA fingerprinting despite most shar- tude of bacteraemia is greater in infants. Most ing the same biotype.15 episodes of clinically significant bacteraemia in Clinical definitions of catheter infections and adults are characterised by low numbers, and catheter related sepsis are not infrequently for example, Henry et al found < 1 colony used. These definitions rely on the absence of forming unit (cfu)/ml in 27%, 55%, and 62% any other demonstrable cause for a patient’s of cases of S aureus, P aeruginosa, and E coli sepsis. However, patients requiring central bacteraemia, respectively.17 Of 43 cases of venous access are often unwell, immunosup- catheter related sepsis from our studies, pressed, or have often undergone surgery and peripheral bacteraemic counts were < 1 cfu/ml are therefore likely to have alternative potential in 30% of cases (Kite P, Dobbins B, Wilcox M, sources of infection. Of those suspected and unpublished data). It is recommended there- subsequently proven to have catheter related fore that at least 10 ml of peripheral blood are sepsis, 50% had a documented potential alter- sampled for each blood culture to reduce the native focus of sepsis (that is, other than the likelihood of false negative results. catheter itself); 63% of the remainder who did The pour plate technique of quantitative not have proven catheter related sepsis also had blood culture involves mixing blood and a documented potential alternative focus of molten agar and then pouring them into a Petri infection.16 Furthermore, in a study of 109 dish.18 The spread plate technique involves cases of clinically suspected catheter related spreading blood across the surface of an agar sepsis only 40 were confirmed after microbio- plate.18 Both methods have the disadvantage of logical culture. Using both pyrexia and leuco- only allowing < 1 ml of blood/plate, and cytosis as clinical indicators, no significant dif- numerous plates (up to 10/sample) need to be ference was observed between groups proven plated for accurate assessment. A lysis centrifu- to have or not to have catheter related sepsis.16 gation technique for the analysis of blood Evidence of inflammation at the catheter skin cultures has been developed in the form of iso- entry site does not necessarily represent lator microbial tubes (Unipath). Developed catheter infection but may merely be caused by from an original technique described by Dorn 19 local irritation from the central venous cath- et al, this technique is designed to maximise http://jcp.bmj.com/ eter. Nevertheless localised infection does the detection of low magnitude bacteraemia or occur at the skin entry site and in its most florid fungaemia and to remove inhibitory factors form is seen as tunnel tract sepsis which may such as that may be present in give rise to formation. Such infections blood. The technique has proven benefits of should perhaps be considered as wound infec- isolation of a wide range microorganisms, and tions in the presence of a foreign body. has been shown to be superior to broth culture methods by increasing recovery of microorgan- Diagnostic methods for catheter related isms from 67% to 80%.20 21 It has also been on October 1, 2021 by guest. Protected copyright. sepsis shown to inhibit phagocytosis and enhance PERIPHERAL BLOOD CULTURES speed of culture, and it is less cumbersome As already indicated the confirmation of a than other quantitative techniques. However, peripheral bacteraemia is paramount in the variable rates of contamination from 1.4% to diagnosis of catheter related sepsis. The 10.9% have been reported,21 and these are peripheral blood sample should be obtained comparable with other surface spread tech- while the central venous catheter is in situ and niques. The technique is, however, expensive not after removal. The peripheral blood and time consuming in comparison with quali- cultures should ideally be taken while the tative blood cultures. patient is pyrexial, and blood samples should be obtained from separate peripheral sites, CENTRAL VENOUS CATHETER CULTURE although for practical purposes often only one Central venous catheter culture methods in- peripheral set of blood cultures is collected. volve culturing catheter segments, usually the Endoluminal colonisation of the central venous distal end but proximal (primarily subcutan- catheter results in the infusate flowing over eous) sections in some cases, and accurate heavily colonised endoluminal biofilm contain- comparison of diVering techniques can be ing both sessile and planktonic forms of bacte- extremely diYcult. Indeed, the length of ria. The planktonic bacteria enter the infusate catheter segments which are sampled is poorly and thereafter the systemic circulation. The standardised, and in the original description of infusate should not therefore be stopped before the Maki roll technique sections between 5 and peripheral blood sampling. In cases of catheter 7 cm in length were sampled.22 Techniques can related sepsis secondary to extraluminal colo- also be divided according to whether the exter- nisation alone (see below) the sampling of nal surface of the central venous catheter, the Diagnosis of central venous catheter related sepsis 167

endoluminal surface, or both environments are had Maki roll counts of > 1000 cfu were actu- sampled. While a positive catheter culture in ally from patients diagnosed as having catheter the presence of peripheral bacteraemia is all related sepsis, so indicating the low specificity J Clin Pathol: first published as 10.1136/jcp.52.3.165 on 1 March 1999. Downloaded from that is required to make a diagnosis of catheter of this method.22 The Maki roll technique and related sepsis, a negative catheter culture made the above tip culture techniques are prone to without assessing the whole of the catheter contamination of the external surface of the does not exclude the diagnosis. For practical catheter tip segment upon removal through purposes, in vitro sampling of the whole of the colonised skin entry sites, which—combined catheter (extraluminal and endoluminal sur- with the low cut oV values of 15 cfu often faces, from hub to tip) is not a cost–eVective or used—produces diYculty in interpreting cath- time–eVective method of assessing central eter tip culture results. Terms such as catheter venous catheters, and thus a compromise contamination, catheter colonisation, and cath- method of analysis has often been used. How- eter infection are often applied to interpret ever, in the study setting and in the interest of various degrees of catheter microbial bioload good science, assessment of the whole of the not associated with a peripheral bacteraemia, catheter should be undertaken, to act as a gold and are invariably confused and misinterpreted standard with which other central venous cath- as true cases of catheter related sepsis. A study eter culture methods can be compared. Unfor- looking at both clinically suspected and rou- tunately few studies have looked at all aspects tinely removed non-suspected triple lumen of central venous catheter colonisation. catheters (100 catheters studied) showed no significant diVerence in the proportion of Maki Specific central venous catheter tip positive results or the degree to which the culture techniques external portions of the catheter tips were QUALITATIVE CULTURE colonised, despite 25 cases of proven catheter A qualitative culture technique described by related sepsis in the suspected group and none Druskin and Siegel in 1963 involved the in the routinely removed central venous immersion of the catheter tip in broth, with any catheters.26 Furthermore, this study showed a growth of the catheter being significant.23 The strong correlation between the number of obvious disadvantage of this method is the microorganisms yielded by the Maki technique relatively high level of false positive results; and the magnitude of microbial growth after these led to development of both quantitative sampling skin around the catheter entry site. and semi-quantitative methods of catheter tip Elsewhere, the positive predictive value of the assessment. roll plate method has been found to be in the range of 10–75%.27 28 QUANTITATIVE CULTURE The accuracy and clinical significance of the Seligman first proposed the use of quantitative above techniques are more helpful when the culture of the catheter tip.24 The techniques of microorganism isolated is not a typical compo- microorganism retrieval and catheter culture nent of the skin microflora—for example,

have been modified over the ensuing years in and Gram negative bacilli—in which http://jcp.bmj.com/ order to improve the sensitivity and specificity cases the results are unlikely to represent of this approach. Flushing the catheter tip, vig- central venous catheter contamination. The orous agitation, sonication, vortexing, and a converse of this is illustrated in table 1, which combination of sonication and vortexing have summarises the composite results of colonised all been described as methods of releasing the and infected central venous catheters, as microorganisms entrapped in the catheter defined by two endoluminal methods and the biofilm.25 The threshold levels for these various Maki roll technique.29 It can be seen that techniques vary from between 100 and 1000 coagulase negative staphylococci were appar- on October 1, 2021 by guest. Protected copyright. cfu/ml,25 and as a result sensitivities and ently significantly more often associated with specificities are accordingly aVected. Both the colonisation as opposed to infection, whereas internal and external surfaces of the catheter the reverse was true for Gram negative bacilli tips are cultured with such qualitative and with a similar trend for S aureus. These results quantitative methods, and thus the relative probably arise from the excess proportion of contributions of each to catheter related sepsis central venous catheters identified as being cannot be deduced from these results. colonised by the Maki roll technique (92%) In 1977 Maki et al described a method of catheter culture which was to become the Table 1 Microorganisms causing catheter related sepsis and catheter colonisation (reproduced from reference 29) standard method for central venous catheter 22 analysis over the next 20 years. The original Number (%) of episodes study was performed primarily (87%) on Colonisation Catheter related peripherally inserted cannulas, 85% of which Microorganism¶ (n=76) sepsis (n=22) were short catheters (mean length 5.7 cm), but the technique oVered a simple, inexpensive, Coagulase negative staphylococci 63 (83)*** 9 (41)*** and seemingly accurate method of catheter 5 (7)* 4 (18)* culture. In this technique the external surface Enterococci 5 (7) 3 (14) of the catheter tip is sampled and as few as 15 Gram negative bacilli 1 (1)** 4 (18)** Candida spp 1 (1) 2 (9) cfu are considered as a significant bacterial Other 1 (1) 0 yield, although Maki himself acknowledged a higher value of 1000 cfu as more accurately ¶Predominant microorganism in the case of mixed growth. 21 22 *0.1>p>0.05; **p<0.01; ***p<0.001, statistical comparison of predicting cases of catheter related sepsis. the percentage of colonisation v catheter related sepsis episodes Furthermore, only four of 13 catheters which for each microorganism. 168 Dobbins, Kite, Wilcox

compared with those detected by the endolu- progress of the brush is detected owing to minal methods (43%).29 When the Maki roll catheter narrowing), and not through the cath- method has been employed, it has been eter tip as first described.36 Quantitative J Clin Pathol: first published as 10.1136/jcp.52.3.165 on 1 March 1999. Downloaded from estimated that up to 75–85% of central venous peripheral blood cultures before and after catheters are removed unnecessarily on clinical brushing in 43 cases of proven catheter related suspicion of catheter related sepsis,30 31 and the sepsis revealed no significant increase in the risks of complications following central venous mean bacterial count at either three minutes or catheter reinsertion are not insignificant.30 one hour after brushing.37 Unlike other meth- Nevertheless, an advantage of the technique is ods of catheter evaluation, this approach does that in cases of true catheter related sepsis the not require the needless sacrifice of catheters to source of infection is removed immediately, evaluate suspected cases of catheter related even though the diagnosis is only made 24–48 sepsis. However, diagnostic quandaries exist hours after removal. when investigating multilumen central venous catheters. In a study of 100 triple lumen cath- CATHETER TIP FLUSH TECHNIQUES eters in which catheter related sepsis was iden- The catheter tip flush technique was originally tified in 25 cases, 40%, 40%, and 20% of the described by Cleri et al.32 One modification, catheters had one, two, and three lumens aimed at removing microorganisms from the significantly colonised.26 Hence a decision has outer surface of the central venous catheter tip, to made about whether to sample the catheter has been shown to be as sensitive as the meth- lumen which has been used most often ods described above but in addition is more (particularly for total ) or specific for the diagnosis of catheter related to assess multiple lumens. sepsis.29 The enhanced specificity results either from the reduced false positive culture rate CATHETER HUB CULTURE associated with contamination of the external The catheter hub has often been treated as a surface of the catheter, or from the assumption separate part of the central venous catheter in that endoluminal as opposed to extraluminal many studies of catheter related sepsis. Results catheter colonisation is more significant in the of catheter hub studies have therefore been development of catheter related sepsis. used as a means of assessing the route taken by Attempts to overcome some of the inaccura- microorganisms in order to reach the catheter cies of catheter tip culture have included the tip, rather than as a possible source of catheter sampling of numerous segments of both the sepsis of itself. Sitges-Serra and colleagues catheter tip and the tunnelled segments of the showed that hub culture was positive (> 1000 catheter proximal to the tip.32 33 However, such cfu) in 14 of 20 documented cases of catheter techniques are complicated, diYcult to inter- related sepsis (70%).38 In our own studies of 60 pret and unless the extraluminal and endolu- cases of catheter related sepsis, culture of hub minal surfaces are diVerentially cultured, swabs yielded > 100 cfu in 70% of cases (Kite remain prone to contamination, as discussed P, Dobbins B, Wilcox M, unpublished data).

earlier. The hub is, however, often cleaned with an http://jcp.bmj.com/ and as a result culture may be nega- Other catheter sampling techniques tive but the lumen just distal may be signifi- ENDOLUMINAL BRUSH cantly colonised. The major advantages of both Endoluminal catheter sampling was first de- endoluminal and hub culture techniques are scribed by Grabe and Jakobsen in 1983,34 who that they can both be performed on central used a steel stiletto inserted into the lumen of venous catheters in situ, thereby reducing the the catheter; the technique was later improved number of catheters that are needlessly sacri- by the use of an endoluminal brush by Marcus ficed to enable other culture techniques to be on October 1, 2021 by guest. Protected copyright. and Buday in 1989.35 It involves introducing a performed. An iodine chamber/hub device has sterile endoluminal brush (FAS Medical) been shown to reduce catheter related sepsis through a Luer lock attached to the catheter rates fourfold.39 40 However, aseptic handling is hub and advancing it through the catheter. still required when attaching the device to the Upon removal the brush is clipped oV and, catheter hub. together with any adherent luminal biofilm, is vortexed in 1 ml of phosphate buVered Indirect methods of central venous and then quantitatively cultured (colony catheter culture counts of > 100 cfu being regarded as signifi- QUANTITATIVE BLOOD CULTURES cant). In 230 central venous catheters used Paired quantitative culture techniques rely on predominantly for total parenteral nutrition the principle that blood aspirated over a colo- (median dwell 9.5 days) we found that of 22 nised catheter lumen will contain proportion- (10%) causing sepsis, 17 yielded significant ately more organisms than peripheral blood growth from both surfaces, four from the inner cultures as a result of haemodilutional eVects surface alone, and one from the outer surface and, more importantly, because of the reticu- alone.29 In this cohort, the sensitivity and loendothelial system present in the and specificity of the endoluminal brush technique its ability to remove bacteria from the circulat- for the diagnosis of catheter related sepsis were ing blood. Various methods of quantification 95% and 84%, respectively. Recent re- of the blood cultures (described above) have evaluation of the technique has shown main- been used in the assessment of central venous tained sensitivity but improved specificity when catheters. The relative proportions of catheter the brush is passed to within approximately 5 to peripheral blood bacteraemia range from cm of the catheter tip (that is, until resistance to 1:4 and 1:30.41 42 Such techniques are accurate Diagnosis of central venous catheter related sepsis 169

for the diagnosis of catheter related sepsis and blockage, and also the requirement for special- again do not rely on catheter removal. ist equipment. However, in our experience blood cannot be J Clin Pathol: first published as 10.1136/jcp.52.3.165 on 1 March 1999. Downloaded from aspirated in between 12% and 25% of cases, Other catheter related sepsis detection and published data indicate that this may be approaches 33 the case in up to 50% ; thus the technique is ELECTRON MICROSCOPY not as accurate as it appears if all central Scanning electron microscopy has been used in venous catheters are assessed. Furthermore, the research setting to evaluate catheter coloni- right atrial blood can be assumed to have sation. Raad et al examined the extraluminal higher bacterial counts than peripheral blood and endoluminal surfaces of central venous in cases where there is a distant source of sep- catheters by semiquantitative electron sis as the blood has yet to be filtered by the microscopy.51 They visualised microbial coloni- pulmonary vascular system.43 sation of all central venous catheters examined, including 39 associated with catheter infection ENTRY SITE CULTURE and 26 culture negative controls. The extent of Skin entry site cultures have been used to both endoluminal surface coverage by biofilm (in predict and diagnose catheter infections.44 catheters obtained premortem) was greater Interpretation of positive results can be diY- than for extraluminal surfaces. While the extent cult as skin microbial colonisation may be of biofilm coverage on extraluminal surfaces transitory and can vary markedly between essentially remained unchanged with increas- patients. Not surprisingly an association with ing duration of central venous catheter place- positive skin culture and catheter tip culture ment, the percentage of colonised endoluminal has often been reported,45 46 but this may surface increased with time. Central venous simply reflect the contamination of the catheter catheters (n = 10) recovered postmortem gen- tip upon removal. Clinical signs of suppura- erally had a greater percentage of their extralu- tion, tract erythema, and even abscess forma- minal surfaces covered by microbial biofilm tion represent catheter wound infections and than catheters obtained premortem; the likely do not necessarily correlate with catheter explanation for this is that biofilm may be lost related sepsis; these should be viewed as sepa- owing to shearing upon removal of the central rate types of infective catheter . venous catheter. Our own electron microscopic studies of catheters causing sepsis have shown copious Rapid diagnostic techniques endoluminal but not extraluminal biofilms DIRECT CATHETER STAINING TECHNIQUES comprising microbes, extracellular material, Direct staining of the catheter tip or impression and fibrin (fig 1). Culture results confirmed the smears of catheter segments upon removal relative contributions of endoluminal and have been described, using both Gram and extraluminal catheter colonisation to catheter acridine orange stains.47–49 The obvious advan- related sepsis. Furthermore, these findings

tage of this approach is early diagnosis and were consistent regardless of the species of http://jcp.bmj.com/ therefore earlier treatment. However, as cath- bacteria or fungi isolated. In some cases eter removal is required, unnecessary sacrifice positive extraluminal cultures were obtained, is inevitable, and furthermore sus- but we were unable to demonstrate microor- ceptibilities still take a minimum of 24 to 48 ganisms on the external surface of the catheter hours from the time of removal. The technique by electron microscopy, possibly reflecting the is just as prone to contamination as other tip relative insensitivity of the latter approach or culture methods and the procedures are removal of biofilm during catheter withdrawal. tedious and time consuming, mainly owing to on October 1, 2021 by guest. Protected copyright. the requirement for focusing on a field of cellu- Recent evidence of the aetiology of lar material on a curved and often non- catheter related sepsis transparent material. While the aetiology of catheter related sepsis, and in particular the potential routes of ACRIDINE ORANGE STAINING OF CATHETER infection, has been extensively reviewed,33 52 53 BLOOD insight into the main contentious issue, namely Kite and colleagues described a rapid acridine the relative importance of the extraluminal and orange/Gram stain technique for the examina- endoluminal routes of infection, has recently tion of cytospun catheter blood samples.49 50 become available with the exciting develop- The technique requires as little as 50 ml of ment of antimicrobially impregnated/coated blood and results can be obtained in about 30 catheters.54 At first glance, the results of such minutes, thus allowing early diagnosis and studies have been conflicting and often confus- treatment of serious infections such as Gram ing. A large study of and silver negative bacterial and fungal sepsis. The sensi- sulphadiazine coated catheters suggested a tivity and specificity of this techniques were reduction in both colonisation and catheter 85% and 94%, respectively, when a group of 95 related sepsis rates,55 but later reports have neonates and infants was studied.50 The failed to substantiate the initial findings.56–58 method is also applicable to adults, with a sen- These catheters are coated externally alone, sitivity and specificity of 96% and 94%, which may explain the conflicting results. respectively, in 60 episodes of catheter related Interestingly, the use of impregnated cuVson sepsis.49 50 However, as stated previously the catheters, on the premise that extraluminal method is limited by the frequent unavailability catheter colonisation is the main source of of through line catheter samples owing to line catheter infection, also gave inconsistent 170 Dobbins, Kite, Wilcox

Such teams are common in the USA but not in the United Kingdom. It is assumed that this approach reduces the likelihood of central J Clin Pathol: first published as 10.1136/jcp.52.3.165 on 1 March 1999. Downloaded from venous catheter contamination during inser- tion and subsequent manipulations. It is inter- esting to note that despite taking maximal pre- cautions during central venous catheter insertion (experienced anaesthetists using strict aseptic technique in an operating thea- tre), Elliot and colleagues could demonstrate the presence of bacteria on five of 30 catheter tips (16%) within 90 minutes of placement into patients undergoing cardiac surgery. When central venous catheters were inserted through a protective Swan sheath the bacterial isolation rate was reduced to 1/30 (3%).66 67 The fate of bacteria introduced onto central venous cath- eter tips at the time of insertion remains unknown, but it is disturbing that relatively high incidences of catheter seeding can occur in the hands of experienced operators, and in the face of (given as surgical prophylaxis).

Conclusions In summary, the diverse methods which have emerged for the diagnosis of catheter related sepsis have not surprisingly yielded inconsist- ent findings. Critical appraisal of diagnostic methods tends to show a greater agreement for techniques that assess the catheter lumen as opposed to those which sample extraluminal surfaces. Clinical microbiologists and clini- cians should together review the value of their current practice for the diagnosis of catheter related sepsis. Without such a critical approach wasteful and misleading rituals may persist. For example, it was found that Maki roll

cultures of central venous catheters removed http://jcp.bmj.com/ from surgical ICU patients had no clinical impact in 96% of episodes, and indeed a new line was inserted in the great majority of cases (86%).68 It is clear that in the majority of cases where catheters are currently suspected of causing sepsis this is not confirmed, and hence Figure 1 (A) Scanning electron micrograph of a split triple lumen central venous catheter

the catheters are needlessly removed. Further- on October 1, 2021 by guest. Protected copyright. showing infected endoluminal biofilm in two lumens of a confirmed case of yeast more, a recent United Kingdom survey high- catheter related sepsis with negative extraluminal surface. (B) High magnification scanning electron micrograph of the same central venous catheter showing yeast cells enmeshed in a lighted the frequent practice of empirical blood fibrin matrix. central venous catheter replacement in the ICU setting, despite the absence of supportive, results.59–61 Recent published data on the prospective, randomised data.69 Unnecessary protective eYcacy of rifampin and minocycline catheter removal and replacement is costly and coated catheters have shown even more 62 63 not without risk to the patient. Diagnostic encouraging results. However, although methods for catheter related sepsis which are these catheters are coated on both outer and not reliant on line sacrifice are available, and inner surfaces, the manufacturers have failed to can be exploited to overcome the problems coat the extracorporeal portion of the catheter created by this practice. from the hub to the catheter manifold (as in the silver/chlorhexidine catheter).64 This area of the 1 Michel L, McMichen JC, Bachy JL. Microbial colonisation catheter represents over half the luminal of indwelling central venous catheters: statistical evaluation surface area, and therefore there is still of potential contaminating factors. Am J Surg 1979;137:745–8. potential for endoluminal biofilm formation 2 Powell C, Kudsk KA, Kulich PA, et al. EVect of frequent with consequent showering of bacteria into the guidewire change on triple lumen catheter sepsis. J Parenter Enter Nutr 1988;12:464–5. circulation. 3 Prager RL, Silva J. Colonisation of central venous catheters. One of the most dramatic reductions in South Med J 1984;77:458–61. 4 Gil RT, Kruse JA, Thill-Baharozian MC, Carlson RW. Tri- catheter infections and more specifically cath- ple versus single-lumen central venous catheters. A eter related sepsis has been the introduction of prospective study in a critically ill population. Arch Intern Med 1989;149:1139–43. dedicated teams to insert and manage central 5 Jarvis WR, Edwards JR, Culver DH, et al. Nosocomial venous catheters (for example, hub care, skin infection rates in adult and pediatric intensive care units in the . Am J Med 1991;91(suppl 3B):185– entry site , and giving set changes).65 91S. Diagnosis of central venous catheter related sepsis 171

6 Raad II, Darouiche RO. Catheter related septicaemia: risk eter related sepsis [abstract]. 37th Interscience Conference reduction. Infect Med 1996;13:807–12; 815–16; 823. on Agents and Chemotherapy, Toronto, 7 Maki DG, Cobb L, Garman JK, et al. An attachable silver- 1997:J189. impregnated cuV for the prevention of infection with 38 Sitges-Serra A, Linares J, Garau J. Catheter sepsis: the clue J Clin Pathol: first published as 10.1136/jcp.52.3.165 on 1 March 1999. Downloaded from central venous catheters: a prospective randomised multi- is the hub. Surgery 1985;97:355–7. centre trial. Am J Med 1988;85:307–14. 39 Segura M, Alvarez-Lerma F, Tellado JM, et al. A clinical 8 Arnow PM, Quimosing EM, Beach M. Consequences of trial on the prevention of catheter-related sepsis using a intravascular catheter sepsis. Clin Infect Dis 1993;16:778– new hub model. Ann Surg 1996;223:363–9. 84. 40 Salzman MB, Isenberg HD, Shapiro JF, et al. A prospective 9 Banerjee SN, Emori TG, Culver DH, et al. Secular trends in study of the catheter hub as a portal of entry for nosocomial primary bloodstream infections in the United microorganisms causing catheter related sepsis in neonates. States, 1980–1989. Am J Med 1991;91:87–9S. J Infect Dis 1993;167:487–90. 10 Glynn A, Ward V, Wilson J, et al. Hospital-acquired infection. 41 Capdevila J, Planes AM, Palomar M, et al. ValueofdiVeren- Surveillance, policies and practice. A report of the control of tial quantitative blood cultures in the diagnosis of catheter hospital-acquired infection in nineteen hospitals in Eng- related sepsis. Eur J Clin Microbiol Infect Dis 1992;11:403– land and Wales. London: Public Health Laboratory 7. Service, 1997. 42 Weightman NC, Simpson EM, Speller DCE, et al. 11 Pittet D, Tarara D, Wenzel RP. Nosocomial bloodstream Bacteraemia related to indwelling central venous catheters: infections in critically ill patients. Excess length of stay, prevention, diagnosis, and treatment. Eur J Clin Microbiol extra cost, and attributable mortality. JAMA 1994;271: Infect Dis 1998;17:125–9. 1598–601. 43 Rauncher HS, Hyatt AC, Barzilai A, et al. Quantitative 12 Heiselman D. Nosocomial bloodstream infections in the blood cultures in the evaluation of septicemia in children critically ill. JAMA 1994;272:1819–20. with broviac catheters. J Pediatr 1984;104:29–33. 13 Herwaldt LA, Geiss M, Kao C, et al. The positive predictive 44 Bjornson HS, Colley R, Bower RH, et al. Associations value of isolating coagulase-negative staphylococci from between microorganisms grown at the catheter insertion blood cultures. Clin Infect Dis 1996;22:14–20. site and colonisation of the catheter in patients receiving 14 Souvenir D, Anderson DE, Palpant S, et al. Blood cultures total parenteral nutrition. Surgery 1982;92:720–7. positive for coagulase-negative staphylococci: antisepsis, 45 Fan ST, Teoh-Chan CH, Lau KF, et al. Predictive value of pseudobacteraemia and therapy of patients. J Clin Microbiol surveillance skin and hub cultures in central venous 1998;36:1923–6. catheter sepsis. J Hosp Infect 1998;12:191–8. 15 Kindon AJL, Dobbins BM, Fitzgerald P, et al. DNA finger- 46 Atela I, Coll P, Rello J, et al. Serial surveillance cultures of prints of coagulase-negative staphylococci isolated from skin and catheter hub specimens from critically ill patients catheters using a novel endoluminal brush [abstract]. with central venous catheters: molecular epidemiology of Eighth International Symposium on Staphylococci and infection and implications for clinical management and Staphylococcal infections, Aix Les Bains, France, 1996: research. J Clin Microbiol 1997;35:1784–90. 17. 47 Cooper GL, Hopkins CC. Rapid diagnosis of intravascular 16 Dobbins BM, Tighe MJ, Kite P, et al. Should in-situ endo- catheter associated infection by direct gram staining of luminal brushing of central venous catheters become part catheter segments. N Engl J Med 1985;312:1142–7. of the standard septic screen? [abstract]Clin Nutr 1997;16 48 Coutlee F, Lemieux C, Paradis JF. Value of direct catheter (suppl 2):73. staining in the diagnosis of intravascular-catheter-related 17 Henry NK, McLimens CA, Wright AJ, et al. Microbiologi- infection. J Clin Microbiol 1988;26:1088–90. cal and clinical evaluation of the isolator lysis– 49 Kite P, Tighe MJ, Thomas D, et al. Rapid diagnosis of centrifugation blood culture tube. J Clin Microbiol 1983;17: catheter-related sepsis using the acridine orange leucocyte 864–9. cytospin test and an endoluminal brush. J Parenter Enter 18 Tilton RC. The laboratory approach to the detection of Nutr 1996;3:215–18. bacteraemia. Annu Rev Microbiol 1982;36:467–93. 50 Rushforth JA, Hoy CM, Kite P, et al. Rapid diagnosis of 19 Dorn GL, Burson GG, Haynes JR. Blood culture technique catheter-related sepsis. Lancet 1993;342:402–3. based on centrifugation; clinical evaluation. J Clin Microbiol 51 Raad I, Costerton W, Sabharwal U, et al. Ultrastructural 1976;3:258–63. analysis of indwelling vascular catheters: a quantitative 20 Dorn GL, Burson GG, Haynes JR. Improved blood culture relationship between luminal colonization and duration of technique based on centrifugation; clinical evaluation. J placement. J Infect Dis 1993;168:400–7. Clin Microbiol 1979;9:391–6. 52 Pearson ML, Hospital Infection Control Practices Advisory 21 Yagupsky P, Nolte FS. Quantitative aspects of septicaemia. Committee. Guideline for prevention of intravascular- Clin Microbiol Rev 1990;3:269–79. device-related infections. Infect Control Hosp Epidemiol 22 Maki DG, Weis CE, Sarafin HW. A semi-quantitative 1996;17:438–73. culture method for identifying intravenous-catheter-related 53 Raad I. Intravascular-catheter-related infections. Lancet infection. N Engl J Med 1977;296:1305–9. 1998;351:893–8.

23 Druskin MS, Siegel PD. Bacterial contamination of indwell- 54 Pearson ML, Abrutyn E. Reducing the risk for catheter- http://jcp.bmj.com/ ing intravenous polyethylene catheters. JAMA 1963;185: related infection: a new strategy. Ann Intern Med 1997;127: 966–8. 304–6. 24 Seligman SJ. Quantitative intravenous catheter cultures 55 Maki DG, Stolz SM, Wheeler S, et al. Prevention of venous identify focus of bacteraemia [abstract]. Abstracts of the catheter-related bloodstream infection by use of an Annual Meeting of the American Society of Microbiology, antiseptic-impregnated catheter. A randomised, controlled Chicago, 1974:M18. trial. Ann Intern Med 1997;127:257–66. 25 Kristinsson KG, Burnett IA, Spencer RC. Evaluation of 56 Logghe C, Van Ossel C, D’Hoore W, et al. Evaluation of three methods for culturing long intravascular catheters. J chlorhexidine and silver-sulfadiazine impregnated central Hosp Infect 1989;14:183–91. venous catheters for the prevention bloodstream infection 26 Kite P, Wilcox MH, Dobbins BM, et al. Prevalence of endo- in leukaemic patients: a randomised, controlled trial. J

luminal and extraluminal microorganisms in triple-lumen Hosp Infect 1997;37:145–56. on October 1, 2021 by guest. Protected copyright. catheters removed routinely or for suspected sepsis 57 Ciresi DL, Albrecht RM, Volkers PA, et al. Failure of [abstract]. 37th Interscience Conference on Antimicrobial antiseptic bonding to prevent central venous catheter- Agents and Chemotherapy, Toronto, 1997:J57. related infection and sepsis. Am Surg 1996;62:641–6. 27 Widmer AF. IV-related infections. In: Wenzel RP, ed. 58 Pemberton LB, Ross V, Cuddy P, et al. No diVerence in Prevention and control of nosocomial infections. Baltimore: catheter sepsis between standard and antiseptic central Williams and Wilkins, 1993:556–79. venous catheters. Arch Surg 1996;131:986–9. 28 Sitges-Serra A, Linares J. Limitations of semiquantitative 59 Maki DG, Cobb L, Garman JK, et al. An attachable silver- method for catheter culture. J Clin Microbiol 1988;26: impregnated cuV for the prevention of infection with 1074–5. central venous catheters: a prospective, randomised, multi- 29 Kite P, Dobbins BM, Wilcox MH, et al. Evaluation of a center trial. Am J Med 1988;85:307–14. novel endoluminal brush for the in-situ diagnosis of 60 Hasaniya NW, Angelis M, Brown MR, et al. EYcacyofsub- catheter related sepsis. J Clin Pathol 1997;50:278–82. cutaneous silver-impregnated cuVs in preventing central 30 Ryan JA, Abel RM, Abbott WM, et al. Catheter complica- venous catheter infections. Chest 1996;109:1030–2. tions in total parenteral nutrition; a prospective study 61 Groeger JS, Lucas AB, Coit D, et al. A prospective, of 200 consecutive patients. N Engl J Med 1974;290:757– randomised, evaluation of the eVect of silver-impregnated 61. subcutaneous cuVs for preventing tunnelled chronic 31 Padberg FT, Ruggiero J, Blackburn GL, et al. Central venous access infections in patients. Ann Surg 1993; venous catheterization for parenteral nutrition. Ann Surg 218:206–10. 1981;193:264–70. 62 Darouiche R, Raad I, Heard S, et al. A prospective, 32 Cleri DJ, Corrado ML, Seligman SJ. Quantitative culture of randomized, multicenter clinical trial comparing central intravenous catheters and other intravascular inserts. J venous catheters coated with minocycline and rifampin vs Infect Dis 1980;141:781–6. chlorhexidine gluconate and silver sulfadiazine [abstract]. 33 Sherertz RJ, Heard SO, Raad II. Diagnosis of triple-lumen 37th Interscience Conference on Antimicrobial Agents and catheter infection: comparison of roll plate, sonication and Chemotherapy, Toronto, 1997. flushing methodologies. J Clin Microbiol 1997;35:641–6. 63 Raad I, Darouiche R, Dupuis J, et al. Central venous 34 Grabe N, Jakobsen G. Bacterial contamination of subclavian catheters coated with minocycline and rifampin for the catheters: an intraluminal culture method. J Hosp prevention of catheter-related colonisation and blood- Infect 1983;4:291–5. stream infections. A randomised, double-blind trial. Ann 35 Marcus S, Buday S. Culturing indwelling central venous Intern Med 1997;127:267–74. catheters in situ. Infect Surg 1989;8:157–61. 64 Wilcox MH, Kite P, Dobbins B. Antimicrobial intravascular 36 Tighe MJ, Thomas D, Fawley WF, et al. An endoluminal catheters—which surface to coat? J Hosp Infect 1998;38: brush to detect the infected central venous catheter in situ: 322–4. a pilot study. BMJ 1996;313:1528–9. 65 Faubian WC, Wesley JR, Khaldi N, et al. Total parenteral 37 Dobbins BM, Kite P, Wilcox MH, et al. Clinical safety of the nutrition catheter sepsis: impact of the team approach. J endoluminal brush technique for in-situ diagnosis of cath- Parenter Enter Nutr 1986;10:642–5. 172 Dobbins, Kite, Wilcox

66 Elliot TSJ, Moss HA, Tebbs SE, et al. Novel approach to 68 Widmer A, Nettleman M, Flint K, et al. The clinical impact investigate a source of microbial contamination of central of culturing central venous catheters. A prospective study. venous catheters. Eur J Clin Microbiol Infect Dis 1997;16: Arch Intern Med 1992;152:1299–302. 210–13. 69 Cyna AM, Hovenden JL, Lehmann A, et al. Routine J Clin Pathol: first published as 10.1136/jcp.52.3.165 on 1 March 1999. Downloaded from 67 Livesley MA, Tebbs SE, Moss HA, et al. Use of pulsed field replacement of central venous catheters: telephone survey gel electrophoresis to determine the source of microbial of intensive care units in mainland Britain. BMJ 1998;316: contamination of central venous catheters. EurJClin 1944–5. Microbiol Infect Dis 1998;17:108–12.

Journal of Clinical Pathology - http://www.jclinpath.com

Visitors to the world wide web can now access the Journal of Clinical Pathology either through the BMJ Publishing Group’s home page (http://www.bmjpg.com) or directly by using its individual URL (http://www.jclinpath.com). There they will find the following: + Current contents list for the journal + Contents lists of previous issues + Members of the editorial board + Information for subscribers + Instructions for authors + Details of reprint services.

A hotlink gives access to: + BMJ Publishing Group home page + British Medical Association web site + Online books catalogue + BMJ Publishing Group books. http://jcp.bmj.com/ The web site is at a preliminary stage and there are plans to develop it into a more sophisticated site. Suggestions from visitors about features they would like to see are welcomed. They can be left via the opening page of the BMJ Publishing Group site or, alternatively, via the journal page, through “about this site”. on October 1, 2021 by guest. Protected copyright.