A systematic review comparing the relative effectiveness of antimicrobial-coated catheters in intensive care units

A systematic review comparing the relative effectiveness of antimicrobial-coated catheters in intensive care units Prabha Ramritu, MN,a Kate Halton, MSc,a,b Peter Collignon, MD, FRCPA,c David Cook, PhD, FFICANZCA,c David Fraenkel, BM, BS, FRACP,d Diana Battistutta, PhD,b Michael Whitby, MD, FRCP,a and Nicholas Graves, PhDa,b Brisbane and Canberra, Australia

Background: Bloodstream infection related to a central venous catheter is a substantial clinical and economic problem. To develop policy for managing the risks of these infections, all available evidence for prevention strategies should be synthesized and understood. Methods: We evaluate evidence (1985-2006) for short-term antimicrobial-coated central venous catheters in lowering rates of catheter-related bloodstream infection (CRBSI) in the adult intensive care unit. Evidence was appraised for inclusion against predefined criteria. Data extraction was by 2 independent reviewers. Thirty-four studies were included in the review. Antiseptic, antibiotic, and heparin-coated catheters were compared with uncoated catheters and one another. Metaanalysis was used to generate summary relative risks for CRBSI and catheter colonization by antimicrobial coating. Results: Externally impregnated chlorhexidine/silver sulfadiazine catheters reduce risk of CRBSI relative to uncoated catheters (RR, 0.66; 95% CI: 0.47-0.93). Minocycline and rifampicin-coated catheters are significantly more effective relative to CHG/SSD catheters (RR, 0.12; 95% CI: 0.02-0.67). The new generation chlorhexidine/silver sulfadiazine catheters and silver, platinum, and carboncoated catheters showed nonsignificant reductions in risk of CRBSI compared with uncoated catheters. Conclusion: Two decades of evidence describe the effectiveness of antimicrobial catheters in preventing CRBSI and provide useful information about which catheters are most effective. Questions surrounding their routine use will require supplementation of this trial evidence with information from more diverse sources. (Am J Infect Control 2008;36:104-17.)

Bloodstream infection related to a central venous catheter (CVC) in the intensive care unit (ICU) setting is a substantial clinical and economic problem.1,2 Attributable mortality is estimated between 2% and 35%,1-3 and length of stay in the ICU is thought to increase by 9.5 to 11.9 days,1,4 causing substantial economic cost and excess morbidity. Rates of catheter-related bloodstream infection (CRBSI) in the ICU are high at 5.0 From The Centre for Healthcare Related Infection Surveillance & Prevention,a Princess Alexandra Hospital, Brisbane, QLD; Institute of Health & Biomedical Innovation,b Queensland University of Technology, Brisbane, QLD; The Canberra Hospital,c Canberra, ACT; and Intensive Care Unit,d Princess Alexandra Hospital, Brisbane, QLD, Australia. Address correspondence to Kate Halton, MSc, Institute of Health & Biomedical Innovation, Queensland University of Technology, Kelvin Grove, 4059, Australia. E-mail: [email protected]. Supported by The Centre for Healthcare Related Infection Surveillance and Prevention (CHRISP), Queensland Health, which provided funding to the Queensland University of Technology for the development/ publication of this research, and by the National Health & Medical Research Council of Australia, which awarded a project grant for this research. 0196-6553/$34.00 Copyright ª 2008 by the Association for Professionals in Infection Control and Epidemiology, Inc. doi:10.1016/j.ajic.2007.02.012

104

per 1000 catheter-days5,6 because of a unique combination of factors including greater likelihood and extended duration of catheterization, more frequent catheter manipulation, greater likelihood of emergency catheter insertion and subsequent compromise of optimal asepsis, greater severity of underlying illness, and more immunosuppression.7,8 A proposed strategy to reduce risk in high-incidence settings is use of CVCs impregnated or coated with antimicrobial substances such as the antiseptics chlorhexidine and silver sulfadiazine or the antibiotics cefazolin9 or minocycline and rifampicin.10 Clinical trials of antimicrobial catheters have typically failed to estimate accurately their effectiveness because of small numbers of CRBSI and tend to draw conclusions from the surrogate outcome of catheter colonization. Systematic reviews incorporating metaanalysis have improved empirical estimates by combining information from trials undertaken in a variety of clinical contexts11-15; however, only one11 has attempted to assess the effectiveness of antimicrobial catheters specifically in the ICU context, despite higher infection and subsequent mortality rates in this patient population.16 In addition, existing reviews focus mainly on the comparison of antimicrobial catheters relative to nonantimicrobial catheters, and there are few data describing the effectiveness of the

Ramritu et al

Table 1. Electronic databases and evidence-based Internet sites searched Electronic databases MEDLINE CINAHL Current contents Current Contents Connect Australian Medical Index Biological Abstracts EMBASE Science Citation National Library of Medicine PubMed Dissertation Abstracts Database of Review of Abstracts of Effectiveness The Cochrane library Evidence-based Internet sites Health Services Technology (NHS, United Kingdom and United States) http://www.ncbi.nlm.nih.gov/books/bv.fcgi?rid5hstat National Clearing House (Agency for Healthcare Research and Quality) http://www.guideline.gov/ Centers for Disease Control and Prevention guidelines and reports http://www.cdc.gov/doc.do/id/0900f3ec8021ee7a Bandolier and Clinical Evidence (Wales BMJ Publishing group) http://www.jr2.ox.ac.uk/bandolier/index.html

different antimicrobial coatings available relative to one another. Our aim is to describe a systematic review and metaanalysis of 2 decades of research that evaluates the effectiveness of short-term antimicrobial CVCs, relative to one another and nonantimicrobial CVCs, in lowering rates of CRBSI in an adult ICU setting.

METHODS Search strategy Searches were made of electronic databases (Table 1) for studies published between January 1985 and January 2006 using the broad MeSH search term,‘‘catheterization—central-venous,’’ with combinations of the keywords ‘‘infection*,’’ ‘‘prevention*,’’ ‘‘bloodstream,’’ ‘‘antiseptic*,’’ ‘‘antibiotic*,’’ ‘‘antimicrobial*,’’ ‘‘silver*,’’ ‘‘impregnat*,’’ and ‘‘coat*.’’ No restrictions or search features on individual databases were used. Strategies for locating additional and unpublished studies included discussion with intensivists on the research team (reviewers D.C. and D.F.); electronic searches of dissertation abstract databases and evidence-based Internet sites (Table 1); and hand checking of reference lists of all relevant guidelines, systematic reviews, and studies that met all inclusion criteria.

Inclusion criteria All studies were independently reviewed against predefined inclusion criteria (Table 2) by 2 of 5

March 2008

105

Table 2. Inclusion criteria Inclusion criteria 1. Full report of observational study or randomized controlled trial (RCT) 2. Investigate centrally placed central venous catheters 3. Investigate short-term (,21 day) catheters 4. Investigate nontunneled catheters 5. Compare any antimicrobial- or heparin-coated catheter to either an uncoated catheter or another antimicrobial- or heparin-coated catheter 6. Report incidence of CRBSI and/or catheter colonization as an outcome 7. State definition of catheter colonization and CRBSI used 8. Conducted in adult ($18 yr of age) 9. Have at least 50% of included catheters managed in an intensive care unit setting 10. Written in English 11. Published/completed between January 1985 and January 2006 Reasons for exclusion of articles are given in Appendix 2.

reviewers (reviewers P.R., K.H., D.C., F.B., and N.G.). Abstracts were reviewed first, and, if required, the full article was consulted. Inclusion criteria required management of at least 50% of study catheters in the ICU or reporting of data separately for these catheters. The outcomes of CRBSI and catheter colonization were considered. Studies were required to state the definitions used and use a definition for CRBSI that required isolation of the same organisms from peripheral blood and catheter segment cultures. Where studies used multiple definitions, all results were extracted by reviewers, and this variation was explored in sensitivity analysis, with baseline analyses using counts for colonization defined by semiquantitative culture of a catheter segment showing $15 colony forming units and for CRBSI defined by the isolation of same organism from the catheter tip and peripheral blood.

Data extraction Two of 5 reviewers independently extracted information (Appendix 1) from each included study using validated, structured, data extraction instruments piloted prior to use and chosen for their comprehensiveness, clarity, and use in prior CVC-related reviews.17 Details of authors, institutional affiliation, and date and journal of publication were concealed during data extraction. Quality was assessed using Scottish Intercollegiate Guidelines Network (SIGN) checklists.18 Reviewers met to compare data extracted, and discrepancies were discussed with reference to the article to reach consensus. Interrater reliability was not formally assessed, but disagreement was usually over the level of detail rather than the content of extraction. Corresponding authors were contacted via e-mail for missing/ambiguous information. Two attempts were made within a 4-week interval. If the address failed, either a new address or a coauthor was sought for one further attempt.

106 Vol. 36 No. 2

Table 3. Matrix of catheter comparisons made in studies included in the review

Std Std

Appelgren (1996)

Other antimicrobial

Thornton (1996)

Corral (2003); Moretti (2005)

CH

Sherertz (1996)

Cefazolin Mino/rif

CH

SH/SS internal 1 external Cefazolin

Mino/ rif

n n

Bach (1999); Stoiser (2002); Dunser (2005)

Spc

CH/SS internal 1 external

Spc

CH/SS external

n

Heparin

CH/SS external

Silver ion/alloy

n

Other uncoated

Silver ion/alloy

Other Other uncoated Heparin antimicrobial

Bach (1996); Civetta (1996); Pemberton van Vliet (1996); George (1997); Loo (1997); (2001); Maki (1997); Heard (1998); Hannan (1999); Theaker Marik (1999); Hanley (2000); (2002) Sheng (2000); Richards (2003); Dunser (2005)

n

Ranucci (2003)

n n

Carrasco (2004)

Dunser (2005)

n

B-Buisson (2004); Rupp (2005)

n

Kamal (1991); Kamal (1998) Fraenkel (unpub)

Darouiche (1999); Marik (1999)

n

Std, standard polyurethane uncoated catheter; Other uncoated, uncoated catheters of other material, eg, smooth polymer; Spc, silver, platinum, and carbon; CH, chlormhexidine; CH/SS, chlorhexidine and silver sulfadiazine; Mino/rif, minocycline/rifampicin; CH/SS (internal 1 external), chlorhexidine and silver sulfadiazine (internal 1 external coating).

Ramritu et al

Raad (1997); Marik (1999); Chatzinikolaou (2003); Leon (2004); Yucel (2004)

n

Ramritu et al

Statistical methods and sensitivity analysis The risk ratios (RR) with 95% confidence intervals (CI) (RR 6 95% CI) for CRBSI and catheter colonization were calculated for each study. Episodes of catheter insertion were the unit used for analysis. Where this information was not provided,19 we assumed one episode of catheterization per patient. Data were pooled across studies for the same type of antimicrobial catheter and summary RR of CRBSI and catheter colonization, with 95% CI, were calculated using random effects metaanalyses run using the DerSimonian and Laird method,20 assuming normally distributed errors. Forest plots of results were drawn and Mantel-Haenzel tests of statistical heterogeneity calculated.21 Publication bias was assessed using funnel plots and tests of association between effect size and standard error.22 Variations in catheter insertion and management practices specified a priori as having potential to alter the efficacy of antimicrobial catheters were explored in subanalyses. The influence of duration of catheterization, year of publication, number of catheters, and baseline incidence of colonization and CRBSI on effect size was assessed using metaregression and visual inspection of the forest plot. Further sensitivity analyses explored the impact of using results obtained under alternative outcome definitions or rates (as opposed to counts). Stata 8.2 (StataCorp. Stat Statistical Software: Release 8.0, 2003; StataCorp LP, College Station, TX) software was used throughout.

RESULTS Study characteristics Thirty-four studies met inclusion criteria, including one unpublished study23 (unpublished as at finalization of data collection but subsequently published in March 2006) and one24 that provided results separately for ICU patients via e-mail. Studies were organized into 2 groups for analysis. The first compared antimicrobial coatings relative to uncoated catheters. The second group was comparisons between types of antimicrobial-coated catheter. The number of comparisons (38) is greater than the number of studies (34) because of the inclusion of 2, 3-arm trials25,26 (Table 3). There was considerable heterogeneity in study design, setting, patients, cointerventions used for catheter insertion and management (Appendix 3), and definitions of CRBSI and catheter colonization used (Table 4). Two studies gave multiple definitions for CRBSI.27,28 Two studies reported only CRBSI,24,29 and 2 reported only catheter colonization,26,30 with 30 studies considering both outcomes. In one study,31 only

March 2008

107

Table 4. Definitions of CRBSI used in studies included in the review Definition

Author

Year

SO

Loo Marik Carrasco Appelgren Hannan vanVliet Fraenkel Civetta Sherertz Heard Bach Moretti Rupp Bach George Maki Brun-Buisson Kamal Ranucci Kamal Hanley Sheng Theaker Richards Raad Darouiche Corral Yucel Leon Chatzinikolaou Pemberton Stoiser*

1997 1999 2004 1996 1999 2001 unpub 1996 1996 1998 1999 2005 2005 1996 1997 1997 2004 1991 2003 1998 2000 2000 2002 2003 1997 1999 2003 2004 2004 2003 1996 2002

¤ ¤ ¤ ¤ ¤ ¤ ¤ h h h h h h h O O O

SO NS

SO CS

SO/ NS/ CS

SO/ NS/ CS/ AB

SO/ NS/ RS

¤ ¤ ¤ ¤

h h ¤ ¤ ¤ ¤ ¤ h h h h O ¤ ¤

SO, same organism isolated from catheter and peripheral blood; NS, no other source of infection; CS, clinical symptoms of sepsis/infection; AB, physician instituted antimicrobial therapy; RS, resolution of symptoms on catheter removal. ¤, only intravascular catheter segment cultured; h, intravascular and subcutaneous catheter segments cultured; O, intravascular and subcutaneous and other (insertion site, hub, infusate) catheter segment cultured. *Stoiser et al used a weighted scoring system that considered, but did not require, the following elements: NS, CS, RS, Blood cultures, catheter cultures, and appearance of insertion site. Further details are given in the following reference: Lugauer et al. Sicherung der klinischen diagnose einer catheter-assoziierten infection anhand eines bewertungsscores. Infection 1998;26(Suppl 1):S53-59. A direct English translation of this article is provided in the same journal: Lugauer et al. A new scoring system for the clinical diagnosis of catheter-related infections. Infection 1999;27(Suppl 1):S49-53.

colonization results were retrieved for analysis because CRBSI results were given for a mixture of CVC types. Data for both outcomes are presented in Tables 5 and 6. We observed only a moderate linear correlation between the 2 outcomes (Appendix 4). Reporting and interpretation of results are therefore focused on estimates obtained for the clinically important outcome of CRBSI.

108

Ramritu et al

Vol. 36 No. 2

Table 5. Relative risk of CRBSI and catheter colonization associated with antimicrobial catheters as compared with uncoated catheters Catheter colonization Proportion of colonizations

CH-SS (external) Bach 1996 Civetta* 1996 1996 Pembertony George 1997 Loo 1997 Maki 1997 Heard 1998 Hannan 1999 Marik 1999 Hanley 2000 Sheng 2000 van Vliet 2001 Theaker 2002 Richards 2003 Dunser 2005 Summary CH-SS (internal and external) Brun-Buisson 2004 Rupp 2005 Summary Silver, platinum, carbon Corral 2003 Moretti 2005 Summary Silver ion/alloy Bach 1999 Stoiser 2002 Dunser 2005 Summary Minocycline/rifampicin Raad 1997 Marik 1999 Chatzinikolaou 2003 Leon 2004 Yucel 2004 Summary Chlorhexidine Sherertz 1996 Cefazolin Kamal 1991 Kamal 1998 Summary Vancomycin Thornton 1996 Heparin Appelgren 1996

Catheter-related bloodstream infection

Antimicrobial

Nonantimicrobial

RR (95% CI) using antimicrobial catheter

21/116 9/156

36/117 39/263

0.59 (0.37-0.94) 0.39 (0.19-0.78)

10/44 12/77 28/208 60/151 47/174 7/36

25/35 25/81 47/195 81/157 71/177 11/39

0.32 0.49 0.56 0.77 0.67 0.69

(0.18-0.57) (0.27-0.88) (0.37-0.85) (0.60-0.99) (0.50-0.91) (0.30-1.58)

9/113 6/48 40/101 14/237 12/165

25/122 10/46 55/131 30/223 19/160

0.39 0.58 0.94 0.44 0.61 0.59

(0.19-0.80) (0.23-1.45) (0.69-1.29) (0.24-0.81) (0.31-1.22) (0.50-0.71)

Proportion of CRBSIs Antimicrobial

Nonantimicrobial

RR (95% CI) using antimicrobial catheter

0/116

3/116

0.14 (0.01-2.76)

0/13 4/44 3/77 2/208 5/151 10/174 1/36 8/175 1/113 2/48 12/101 2/237

2/10 10/35 3/81 9/195 6/157 11/177 2/39 15/191 2/122 3/46 12/131 6/223

0.16 (0.01-2.95) 0.32 (0.11-0.93) 1.02 (0.21-4.84) 0.21 (0.05-0.95) 0.87 (0.27-2.78) 0.92 (0.40-2.12) 0.54 (0.05-5.72) 0.58 (0.25-1.34) 0.54 (0.05-5.87) 0.64 (0.11-3.65) 1.30 (0.61-2.76) 0.31 (0.06-1.54) 0.66 (0.47-0.93)

7/188 32/345

23/175 59/362

0.28 (0.12-0.64) 0.57 (0.38-0.85) 0.44 (0.23-0.85)

3/188 6/345

5/175 8/362

0.56 (0.14-2.30) 0.79 (0.28-2.24) 0.70 (0.30-1.62)

29/103 33/252

41/103 42/262

0.71 (0.48-1.04) 0.82 (0.54-1.25) 0.76 (0.57-1.01)

1/103 3/252

4/103 4/262

0.25 (0.03-2.20) 0.78 (0.18-3.45) 0.54 (0.16-1.85)

9/34 10/50 27/160

7/33 14/47 19/160

1.25 0.67 1.42 1.09

(0.53-2.96) (0.33-1.36) (0.82-2.45) (0.68-1.74)

2/34 3/50

2/33 3/47

0.97 (0.15-6.49) 0.94 (0.20-4.43)

0.32 0.37 0.79 0.55 0.14 0.40

(0.17-0.60) (0.13-1.07) (0.41-1.50) (0.41-0.74) (0.06-0.32) (0.23-0.67)

0/130 0/38 0/66 6/187 0/118

11/130 4/38 13/66 46/187 6/118

36/136 11/39 16/64 81/180 38/105

0.95 (0.29-3.16) 7/136 2/39 1/64 11/180 0/105

0.39 (0.17-0.92)

47/126

43/128

1.11 (0.80-1.55)

7/126

3/128

1/60 4/127

4/33 5/298

0.14 (0.02-1.18) 1.88 (0.51-6.88) 0.59 (0.04-7.72)

0/60 1/127

0/33 0/298

0/13

5/19

56/91

68/85

0.77 (0.63-0.93)

1/13

6/19

0.24 (0.03-1.79)

0.07 (0.00-1.21) 0.21 (0.01-4.14) 0.32 (0.01-7.79) 0.53 (0.20-1.39)

2.37 (0.63-8.96)

7.00 (0.29-170.9) Not possible

0.13 (0.01-2.16)

*Only data on colonization extracted as results for CRBSI included pulmonary arterial catheters. y Data on CRBSI provided for ICU patients only (as opposed to ward/ICU mix) via e-mail communication with authors.

The asymmetry seen in the funnel plots for risk of CRBSI against study size was not systematically related to underlying study characteristics such as the

baseline CRBSI rate, patient characteristics, study setting, publication date, or type of antimicrobial catheter. This gives some evidence for publication bias with a

Ramritu et al

March 2008

109

Table 6. Relative risk of CRBSI and catheter colonization in studies comparing 2 antimicrobial catheters Catheter colonization Proportion of colonizations

Antiseptic vs antiseptic Ranucci (2003) Dunser (2005) Antibiotic vs antiseptic Marik (1999) Darouiche (1999) Summary Fraenkel (2006) Antiseptic vs heparin Carrasco (2004)

Comparison

Intervention

Control

RR (95% CI) using intervention catheter

SPC vs benzalkonium CH-SS vs silver ion

50/268 12/165

82/277 27/160

MR vs CH-SS MR vs CH-SS

4/38 28/356

7/36 87/382

MR vs SPC

25/319

CH-SS vs heparin

13/128

Catheter-related bloodstream infection Proportion of CRBSIs Intervention

Control

RR (95% CI) using intervention catheter

0.63 (0.46-0.86) 0.43 (0.23-0.82)

9/268

12/277

0.78 (0.33-1.81)

0/38 1/356

1/36 13/382

43/327

0.54 (0.17-1.69) 0.35 (0.23-0.52) 0.36 (0.25-0.53) 0.60 (0.37-0.95)

8/319

10/327

0.32 (0.01-7.52) 0.08 (0.01-0.63) 0.12 (0.02-0.67) 0.82 (0.33-2.05)

29/132

0.46 (0.25-0.85)

3/128

4/132

0.77 (0.18-3.39)

Fig 1. Funnel plot of RR CRBSI against trial size. The absence of studies in the bottom right corner of the plot indicates the possible presence of publication bias among studies of antimicrobial catheters.

tendency for smaller studies to show a greater reduction in risk (Fig 1).

Chlorhexidine and silver-sulfadiazine coatings Chlorhexidine and silver-sulfadiazine (external). Fifteen studies of catheters coated externally with chlorhexidine and silver-sulfadiazine (CH-SS) provided information on 1524 CH-SS and 1493 uncoated catheters.23-29,31-39 Pooling results demonstrates a significant reduction in risk of CRBSI (RR, 0.66; 95% CI: 0.47-0.93) associated with CH-SS (Table 5 and Fig 2). There was no significant statistical heterogeneity between studies (x2 5 11.23, P 5 .509), but the pooled effect estimate was sensitive to 2 studies, both conducted

Fig 2. Forest plot of RR of CRBSI in trials comparing chlorhexidine/silver sulfadiazineimpregnated central venous catheters with nonimpregnated catheters. Note: Studies ordered by year of publication. For each study, the size of the square is proportional to the number of catheters. The diamond indicates the summary relative risk and 95% confidence interval. among high-risk groups. Theaker et al38 recruited longstay critically ill patients; removing this study from the metaanalysis increased the estimate of effectiveness (RR CRBSI, 0.55; 95% CI: 0.38-0.82). George et al33 conducted their study among immunocompromised transplant recipients; excluding this study reduced the pooled estimate of effectiveness (RR CRBSI, 0.72; 95% CI: 0.50-1.04). Excluding studies using modified uncoated catheters28,38 or studies with less stringent definitions of CRBSI from the analysis did not change results. An analysis limited to randomized controlled trials gave an increased estimate of effectiveness (RR CRBSI,

110

Ramritu et al

Vol. 36 No. 2

0.53; 95% CI: 0.33-0.88). Restricting analysis to studies conducted solely within the ICU supported a nonsignificant reduction in risk (RR CRBSI, 0.77; 95% CI: 0.531.13). Metaregression showed no association between RR of CRBSI and duration of catheterization, year of publication, study size, or baseline incidence of CRBSI. Subanalyses produced increased estimates of effectiveness when analyses were restricted to trials in which CH-SS catheters were used with either gauze (RR CRBSI, 0.27; 95% CI: 0.11-0.69) or transparent dressings (RR CRBSI, 0.37; 95% CI: 0.17-0.85); insertion protocol specified mask, gown, gloves, and drape (RR CRBSI, 0.50; 95% CI: 0.30-0.84); or guide wire exchange for CVC replacement was not allowed (RR CRBSI, 0.55; 95% CI: 0.32-0.97). CH-SS (internal and external). Two randomized controlled trials (RCTs) of the ‘‘new’’ CH-SS-coated catheter, impregnated both externally and internally with the antiseptic, were found40,41 giving information on 533 ‘‘new’’ CH-SS and 537 uncoated catheters. Both were multi-ICU trials, thereby increasing generalizability of results, and both used the same definition of CRBSI, although they differed as to which catheter segments were cultured. The study with twice the insertion time40 of the other41 reported a greater relative reduction in risk of CRBSI associated with use of these catheters. Metaanalysis suggests a reduction in risk of CRBSI (RR, 0.70; 95% CI: 0.30-1.62) relative to uncoated catheters, similar to that seen with the ‘‘old’’ externally coated CH-SS catheters (Table 5).

Silver coatings Silver, platinum, and carbon. Two RCTs compared 365 silver, platinum, and carbon-coated (SPC) and 355 standard catheters.42,43 Corral et al42 included only ICU patients, with a greater average duration of catheterization (13 vs 6 days, respectively), and a stricter definition of CRBSI than used in Moretti et al.43 Pooling results suggested a potential reduction in risk of CRBSI (RR, 0.54; 95% CI: 0.16-1.85) (Table 5) associated with use of these catheters, but a more precise estimation of the effect for CRBSI requires further trials. Silver ion/alloy. Three RCTs compared 240 silver ion-coated with 244 uncoated catheters19,26,44; however, one26 only reported catheter colonization, leaving 80 silver ion and 84 standard catheters compared for risk of CRBSI. The pooled results gave no indication that this coating was effective in preventing CRBSI (RR, 0.95; 95% CI: 0.29-3.16) (Table 5). This may indicate that the coating is ineffective or that differences among studies masked an effect. Stoiser et al used a scoring system44 to define CRBSI with a reported greater sensitivity than the CDC definition,45,46 which

Fig 3. Forest plot of RR of CRBSI in trials comparing minocycline/rifampicin-coated central venous catheters with noncoated catheters. Note: Studies ordered by year of publication. For each study, the size of the square is proportional to the number of catheters. The diamond indicates the summary relative risk and 95% confidence interval. could be measuring a different outcome to other studies. They also conducted their study in a group of immunocompromised patients who face a higher risk of CRBSI than the typical ICU case mix.

Minocycline/rifampicin coating Five RCTs compared 539 catheters coated with the antibiotics minocycline and rifampicin (MR) to 524 uncoated catheters.25,47-50 The pooled estimate of effectiveness for MR-coated catheters indicates a reduction in risk of CRBSI associated with this coating (RR, 0.39; 95% CI: 0.17-0.92) (Table 5 and Fig 3). Homogeneity in the definition of CRBSI used in these trials increases the validity of the pooled estimate of effectiveness, giving strong support for the effectiveness of MR CVCs in preventing CRBSI in the ICU.

Other antimicrobial coatings Five further studies were retrieved that compared use of chlorhexidine-,51 vancomycin-,30 cefazolin-,9,52 and heparin-coated53 catheters to uncoated catheters (Table 5). No conclusions can be drawn about the use of these coatings because none of these studies were statistically powered to detect a reduction in CRBSI or catheter colonization.

Antimicrobial comparisons Two RCTs directly compared the use of 394 MRcoated to 418 externally coated CH-SS catheters.10,25 Pooling results showed a significant reduction in risk

Ramritu et al

Fig 4. Forest plot of RR of CRBSI in trials comparing minocycline/rifampicin-coated central venous catheters with chlorhexidine/silver sulfadiazine-coated catheters. Note: Studies ordered by year of publication. For each study, the size of the square is proportional to the number of catheters. The diamond indicates the summary relative risk and 95% confidence interval. of CRBSI associated with MR relative to the old generation CH-SS-coated catheters (RR CRBSI, 0.12; 95% CI: 0.02-0.67) (Table 6 and Fig 4). This evidence is consistent with the indirect comparison provided by trials of MR (RR CRBSI, 0.39; 95% CI: 0.17-0.92) and external CH-SS (RR CRBSI, 0.66; 95% CI: 0.47-0.93) coatings relative to uncoated catheters (Table 5) generating strong evidence that MR-coated CVCs are more effective than externally coated CH-SS CVCs at preventing CRBSI. One RCT directly comparing 319 MR-coated to 327 SPC-coated catheters23 reported a reduction in CRBSI associated with MR relative to SPC catheters (RR, 0.82; 95% CI: 0.33-2.05) (Table 6), with low numbers of observed infections preventing any increase in precision around this estimate. Indirect evidence generated from comparisons of both MR (RR CRBSI, 0.39; 95% CI: 0.17-0.92) and SPC catheters (RR CRBSI, 0.54; 95% CI: 0.16-1.85) (Table 5) relative to uncoated catheters is consistent with the direct evidence, providing support for the proposition that MR-coated CVCs are more effective in preventing CRBSI than SPC-coated CVCs. That MR-coated CVCs achieve a greater reduction in risk relative to CH-SS- than SPCcoated CVCs may provide tentative indications that SPC-coated CVCs could be superior to CH-SS-coated CVCs when used in this setting. One RCTcompared use of 165 externally coated CHSS to 160 silver ion-impregnated catheters.26 The trial did not study CRBSI but indicated that the CH-SS catheters were less likely to become colonized (RR, 0.43; 95% CI: 0.23-0.82) (Table 6). An indirect comparison

March 2008

111

of these coatings, made by comparing results from trials of CH-SS (RR CRBSI, 0.66; 95% CI: 0.47-0.93) and silver ion coatings (RR CRBSI, 0.95; 95% CI: 0.29-3.16) relative to uncoated catheters (Table 5) is consistent with the suggestion that the CH-SS coating reduces risk of infection, whereas silver ions appear to have little effect. One multicenter RCT compared 268 SPC-coated catheters to 277 catheters coated with benzalkonium.54 It reported a nonsignificant reduction in risk of CRBSI (RR, 0.78; 95% CI: 0.33-1.81) associated with SPCcoated catheters (Table 6). The results of another RCT55 involving 260 catheters suggest a reduction in risk of CRBSI (RR, 0.77; 95% CI: 0.18-3.39) with the use of CH-SS external coating relative to heparin coating (Table 6). The lack of precision in these estimates reflects the paucity of evidence available to compare antimicrobial coatings.

DISCUSSION We evaluated the evidence for the effectiveness of short-term antimicrobial CVCs in preventing CRBSI, relative to one another and nonantimicrobial CVCs. This review supersedes previous reviews11,15,56 by including evidence on novel coatings,40,41 reporting direct and indirect comparisons of antimicrobial coatings, and including RCT and observational data that describe only the adult ICU population. This review will aid clinical decision making. Our finding that externally coated CH-SS CVCs are effective in preventing CRBSI relative to uncoated catheters confirms the conclusions from 5 previous reviews.11-15 Two published studies of the new generation internally coated CH-SS CVCs are available that suggest these catheters are at least as effective as the old generation CH-SS CVCs when compared with uncoated catheters. Further research would be justified to explore the potential benefits of this ‘‘new’’ CH-SS coating, including whether there is evidence for increased effectiveness as a consequence of increased antimicrobial capacity past 7 days. Support for this idea is given by the higher estimates of effectiveness produced by the study with longer insertion times. This finding should be treated cautiously, but the theory is supported in laboratory studies57 and is credible. We find that SPC-coated catheters outperform uncoated, benzalkonium-coated, and CH-SS-coated catheters in protecting against CRBSI, but the small numbers of infections precluded statistical support. Further investigation would be useful, particularly because SPC-coated catheters avoid problems of toxicity and antimicrobial resistance.58 Conversely, silver-ion/ alloys appear to offer no protection. Earlier reviews13,58-60 hypothesized that Staphylococci are not susceptible to silver ions and that silver ions interact

112

Vol. 36 No. 2

with the biologic environment and lose their antimicrobial activity. We do not recommend further investigation of silver-ion coatings. MR-coated catheters were effective in preventing CRBSI relative to uncoated and externally coated CH-SS catheters, confirming existing data14,58-60; however, widespread adoption is contentious because antibiotic catheters may accelerate development of resistance among microorganisms.56 This has been demonstrated in animal models,61 but evidence outside the laboratory is scarce.10,62,63 Limitations of this review include the exclusion of non-English language research and the suggestion of publication bias. This bias is unlikely to undermine conclusions because empirical studies have shown that larger studies usually have sufficient influence over the pooled result to compensate for this asymmetry.64 The inevitable heterogeneities in design and conduct of the studies have been discussed,56 with a lively debate on whether flaws in individual studies undermine the results of evidence synthesis.65,66 The subanalyses we conducted when pooling data from trials of the externally coated CH-SS relative to uncoated catheters indicated that variation in markers of study quality, eg, CRBSI definition, use of guide wire exchange, did not influence estimates of relative effectiveness. This is consistent with Geffers et al,15 who, when looking at the same comparison, also found no association between study quality and the relative reduction in risk of CRBSI achieved. The apparent increase in effectiveness for CH-SScoated CVCs when used with particular dressings or insertion techniques supports the postulate that the use of concurrent infection control strategies may have a synergistic effect in reducing infection risk, but it should be interpreted carefully because it may reflect variation in study reporting. Previous reviews of CVC dressing types,67 guide wire exchange,68 and use of maximal sterile barriers69 all identified data deficiencies and could draw no conclusions about the incremental effectiveness of these strategies. Despite the evidence for antimicrobial CVCs in the ICU, they are not standard practice, and current guidelines7 suggest implementation only when incidence rates are high ($3.3 CRBSI per 1000 catheter-days). This is, in part, due to reservations within the clinical community about the quality of the evidence. The first issue is variation in outcome definitions.56 Although we observed this, our results were robust to this parameter, and we are not convinced that this diversity is a good reason to question the evidence for antimicrobial CVCs. Crnich and Maki65 suggest it actually strengthens evidence for these catheters because its impact would be to dilute evidence of effectiveness. Recommendations for consistent use of definitions in future trials do not preclude use of earlier research.

Ramritu et al

Second is the question of whether colonization is a valid surrogate for CRBSI. We reported a moderate linear correlation that was robust to catheter coating and definition of CRBSI, and the effectiveness of CH-SS and MR-coatings was equivalent for both outcomes. The microbiologic evidence on the role of CVC colonization in the development of CRBSI70 is credible, and we support others who argue that catheter colonization is a good surrogate for CRBSI71 and can be used as a relevant outcome in assessing effectiveness. Further exploration of this relationship should focus on the application of statistical tests of surrogacy.72 Third are questions about whether the effectiveness of antimicrobial catheters is consistent across organisms. The consequences of CRBSI vary by organism,73,74 with resistant organisms generating higher cost and worse patient outcomes.75,76 Coagulase-negative Staphylococcus is the most common CRBSI isolate; however, the predictive value for true BSI of a positive blood culture has been questioned and, with it, the clinical significance of reducing rates of infection with this organism.56 All but 3 studies in this review reported the organisms involved, although several gave limited information, and resistant organisms were rarely discussed. However, most studies do not observe enough events to enable exploration of these issues. We recommend further research into the activity of these catheters against a range of organisms, in particular resistant organisms given reports of 60% of S aureus isolates being methicillin-resistant in ICUs.5 The decision process of whether to adopt antimicrobial CVCs remains unclear. The authors of 3 recent reviews15,56,60 argue that the decision should be based on cost-effectiveness criterion and that bigger trials are required to address uncertainty in key parameters, some of which are discussed here. We agree with this aim but not the method. Our view is that future trials would fail to provide the information required to inform the decision, given the number of alternative coatings on the market, the number of parameters for which data are required, the range of possible patient outcomes and the complexity of the clinical process under scrutiny, and the relatively low rates of infection observed. We predict that future ‘‘catch all’’ or ‘‘mega’’ trials would be too costly, would fail to observe sufficient numbers of events, and would raise ethical concerns. Instead, we argue that the adoption decision should be informed using an alternate approach: evidencebased, decision-analytic models. These are regularly used to address questions about the efficient allocation of health care resources77,78 by regulatory agencies in the United States, United Kingdom, and Australia.79-82 Rather than relying on results from a single RCT, modelling techniques integrate the best evidence from often disparate sources in a transparent and coherent

Ramritu et al

decision-making framework. Three earlier studies using decision-analytic models83-85 concluded that use of antimicrobial catheters was cost-effective; however, these findings have been questioned56 because of uncertainty in evidence used to parameterise the models. Rather than dismissing these earlier models, we suggest that they offer clues about what further evidence is required. For example, further RCTs of externally coated CH-SS and MR-coated catheters relative to uncoated catheters, where substantial evidence already exists, would not be useful because the new trial would be unlikely to cause substantial reduction in the uncertainty surrounding their effect. However, studies of their real-world effectiveness86 using rigorous prospective cohort studies could add valuable additional information. RCTs of SPC-coated catheters and direct comparisons of different types of antimicrobial catheter would be useful because evidence here is lacking/ ambiguous. Rare events (eg, hypersensitivity reactions, anaphylactic toxicity, and antibiotic resistance), which would not typically be observed in a study setting, may be best monitored through registry/surveillance systems, and estimates of attributable mortality could be obtained from survival analyses of patient cohorts. Studies of all designs can and should provide better reporting of clinically relevant end points (eg, duration of hospitalization and antimicrobial therapy) and key confounding variables (eg, severity of illness87 and degree of immunosuppression56) that impact on the decision of whether to use antimicrobial-coated catheters.

CONCLUSION Two decades of evidence is available on the effectiveness of antimicrobial catheters in preventing CRBSI and shows that externally coated CH-SS catheters significantly reduce risk of CRBSI in the ICU context and suggests that SPC-coated and the new generation CHSS-internally coated catheters do the same. Furthermore, MR-coated catheters significantly outperform the antiseptic catheters. Despite this, widespread use of these catheters is still contentious because of, in part, reservations about the quality of this trial evidence. We suggest that this research provides useful information and that it is not repetition of these trials that is required but supplementation with information from more diverse sources to address fully the questions about routine use of antimicrobial catheters. References 1. Pittet D, Tarara D, Wenzel RP. Nosocomial bloodstream infection in critically ill patients: excess length of stay, extra costs and attributable mortalilty. JAMA 1994;271:1598-601. 2. Eggimann P, Sax H, Pittet D. Catheter-related infections. Microbes Infect 2004;6:1033-42.

March 2008

113

3. Byers KE, Adal KA, Anglim AM, Farr BM. Case fatality rate for catheter-related bloodstream infections: a meta-analysis (abstract 43). 5th Annual Meeting of the Society of Healthcare Epidemiology of America. San Diego: Mt. Royal, NJ: The Society of Healthcare Epidemiology of America; 1995;16(4):23. 4. Rosenthal VD, Guzman S, Migone O, Crnich CJ. The attributable cost, length of hospital stay, and mortality of central line-associated bloodstream infection in intensive care departments in Argentina: a prospective, matched analysis. Am J Infect Control 2003;31:475-80. 5. NNIS. National Nosocomial Infections Surveillance (NNIS) System Report, data summary from January 1992 through June 2004, issued October 2004. Am J Infect Control 2004;32:470-85. 6. McLaws ML, Taylor PC. The Hospital Infection Standardized Surveillance HISS programme: analysis of a two-year pilot. J Hosp Infect 2003;53:259-67. 7. Centers for Disease Control and Prevention. Guidelines for the prevention of intravascular catheter-related infections. MMWR 2002; 51(RR-10):1-29. 8. Pittet D, Harbarth S, Ruef C, Francioli P, Sudre P, Petignat C, et al. Prevalence and risk factors for nosocomial infections in four university hospitals in Switzerland. Infect Control Hosp Epidemiol 1999;20:37-42. 9. Kamal GD, Pfaller MA, Rempe LE, Jebson PJR. Reduced intravascular catheter infection by antibiotic bonding: a prospective, randomized, controlled trial. JAMA 1991;265:2364-8. 10. Darouiche RO, Raad II, Heard SO, Thornby JI, Wenker OC, Gabrielli A, et al. A comparison of two antimicrobial-impregnated central venous catheters. Catheter Study Group. N Engl J Med 1999;340:1-8. 11. Veenstra DL, Saint S, Saha S, Lumley T, Sullivan SD. Efficacy of antisepticimpregnated central venous catheters in preventing catheterrelated bloodstream infection: a meta-analysis. JAMA 1999;281:261-7. 12. Marin MG, Lee JC, Skurnick JH. Prevention of nosocomial bloodstream infections: effectiveness of antimicrobial-impregnated and heparin-bonded central venous catheters. Crit Care Med 2000;28:3332-8. 13. Mermel LA. Prevention of intravacular catheter-related infections. Ann Intern Med 2000;132:391-402. 14. Walder B, Pittet D, Trame`r MR. Prevention of bloodstream infections with central venous catheters treated with anti-infective agents depends on catheter type and insertion time: evidence from a metaanalysis. Infect Control Hosp Epidemiol 2002;23:748-56. 15. Geffers C, Zuschneid I, Eckmanns T, Ru¨den H, Gastmeier P. The relationship between methodological trial quality and the effects of impregnated central venous catheters. Intensive Care Med 2003;29:403-9. 16. Rubinson L, Diette GB. Best practices for insertion of central venous catheters in intensive-care units to prevent catheter-related bloodstream infections. J Lab Clin Med 2004;143:5-13. 17. The Epic Project. Developing national evidence based guidelines for preventing hospital-acquired infections associated with the use of central venous catheters. London: Department of Health; 2000. 18. Scottish Intercollegiate Guidelines Network. Sign 50: a guideline developer’s handbook. Edinburgh, UK; 2001. 19. Bach A, Eberhardt H, Frick A, Schmidt H, Bo¨ttiger BW, Martin E. Efficacy of silver-coating central venous catheters in reducing bacterial colonization. Crit Care Med 1999;27:515-21. 20. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177-88. 21. Petitti DB. Meta-analysis, decision analysis and cost-effectiveness analysis. 2nd ed. New York: Oxford University Press; 2000. 22. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629-34. 23. Fraenkel DJ, Rickard C, Thomas P, Faoagali J, George N, Ware R. A prospective, randomized trial of rifampicin-minocycline-coated and silver-platinum-carbon impregnated central venous catheters. Crit Care Med 2006;34:668-75. 24. Pemberton LB, Ross V, Cuddy P, Kremer H, Fessler T, McGurk E. No difference in catheter sepsis between standard and antiseptic central venous catheters: a prospective randomized trial. Arch Surg 1996;131:986-9.

114

Vol. 36 No. 2

25. Marik PE, Abraham G, Careau P, Varon J, Fromm RE Jr. The ex vivo antimicrobial activity and colonization rate of two antimicrobialbonded central venous catheters. Crit Care Med 1999;27:1128-31. 26. Dunser MW, Mayr AJ, Hinterberger G, Florl CL, Ulmer H, Schmid S, et al. Central venous catheter colonization in critically ill patients: a prospective, randomized, controlled study comparing standard with two antiseptic-impregnated catheters. Anesth Analg 2005;101:1778-84. 27. Hannan M, Juste RN, Umasanker S, Glendenning A, Nightingale C, Azadian B, et al. Antiseptic-bonded central venous catheters and bacterial colonisation. Anaesthesia 1999;54:868-72. 28. van Vliet J, Leusink J. A comparison between two types of central venous catheters in the prevention of catheter related infections: the importance of performing all the relevant cultures. Clin Intensive Care 2001;12:135-40. 29. Hanley EM, Veeder A, Smith T, Drusano G, Currie E, Venezia RA. Evaluation of an antiseptic triple-lumen catheter in an intensive care unit. Crit Care Med 2000;28:366-70. 30. Thornton J, Todd NJ, Webster NR. Central venous line sepsis in the intensive care unit. Anaesthesia 1996;51:1018-20. 31. Civetta JM, Hudson-Civetta J, Ball S. Decreasing catheter-related infection and hospital costs by continuous quality improvement. Crit Care Med 1996;24:1660-5. 32. Bach A, Schmidt H, Bo¨ttiger B, Schreiber B, Bo¨hrer H, Motsch J, et al. Retention of antibacterial activity and bacterial colonization of antiseptic-bonded central venous catheters. J Antimicrob Chemother 1996; 37:315-22. 33. George SJ, Vuddamalay P, Boscoe MJ. Antiseptic-impregnated central venous catheters reduce the incidence of bacterial colonization and associated infection in immunocompromised transplant patients. Eur J Anaesth 1997;14:428-31. 34. Loo S, van Heerden PV, Gollege CL, Roberts BL, Power BM. Infection in central lines: antiseptic-impregnated vs standard non-impregnated catheters. Anaesth Intensive Care 1997;25:637-9. 35. Maki DG, Stolz SM, Wheeler S, Mermel LA. Prevention of central venous catheter-related bloodstream infection by use of an antisepticimpregnated catheter: a randomized, controlled trial. Ann Intern Med 1997;127:257-66. 36. Heard SO, Wagle M, Vijayakumar E, McLean S, Brueggemann A, Napolitano LM, et al. Influence of triple-lumen central venous catheters coated with chlorhexidine and silver sulfadiazine on the incidence of catheter-related bacteremia. Arch Intern Med 1998;158:81-7. 37. Sheng WH, Ko WJ, Wang JT, Chang SC, Hsueh PR, Luh KT. Evaluation of antiseptic-impregnated central venous catheters for prevention of catheter-related infection in intensive care unit patients. Diagn Microbiol Infect Dis 2000;38:1-5. 38. Theaker C, Juste R, Lucas N, Tallboys C, Azadian B, Soni N. Comparison of bacterial colonization rates of antiseptic impregnated and pure polymer central venous catheters in the critically ill. J Hosp Infect 2002;52:310-2. 39. Richards B, Chaboyer W, Bladen T, Schluter PJ. Effect of central venous catheter type on infections: a prospective clinical trial. J Hosp Infect 2003;54:10-7. 40. Brun-Buisson C, Doyon Fo, Sollet J-P, Cochard J-Fo, Cohen Y, Nitenberg Gr. Prevention of intravascular catheter-related infection with newer chlorhexidine-silver sulfadiazine-coated catheters: a randomized controlled trial. Intensive Care Med 2004;30:837-43. 41. Rupp ME, Lisco SJ, Lipsett PA, Perl TM, Keating K, Civetta JM, et al. Effect of a second-generation venous catheter impregnated with chlorhexidine and silver sulfadiazine on central catheter-related infections: a randomised controlled trial. Ann Intern Med 2005;143:570-80. 42. Corral L, Nolla-Salas M, Iban˜ez-Nolla J, Leo´n MA, Dı´az RM, Martin MC, et al. A prospective, randomized study in critically ill patients using the OLIGON VANTEX catheter. J Hosp Infect 2003;55:212-9. 43. Moretti EW, Ofstead CL, Kristy RM, Wetzler HP. Impact of central venous catheter type and methods on catheter-related colonization and bacteraemia. J Hosp Infect 2005;61:139-45.

Ramritu et al 44. Stoiser B, Kofler J, Staudinger T, Georgopoulos A, Lugauer S, Guggenbichler JP, et al. Contamination of central venous catheters in immunocompromised patients: a comparison between two different types of central venous catheters. J Hosp Infect 2002;50:202-61. 45. Lugauer S, Regenfus A, Bo¨swald M, Martus P, Geis C, Bechert T, et al. Sicherung der klinischen Diagnose einer Katheter-assoziierten Infektion anhand eines Bewertungsscores. Infection 1998;26(Suppl 1):S53-8. 46. Lugauer S, Regenfus A, Bo¨swald M, Martus P, Geis C, Bechert T, et al. A new scoring system for the clinical diagnosis of catheter-related infections. Infection 1999;27:S49-53. 47. Raad I, Darouiche R, Dupuis J, Abi-Said D, Gabrielli A, Hachem R, et al. Central venous catheters coated with minocycline and rifampin for the prevention of catheter-related colonization and bloodstream infections: a randomized, double-blind trial. The Texas Medical Center Catheter Study Group. Ann Intern Med 1997;127:267-74. 48. Chatzinikolaou I, Finkel K, Hanna H, Boktour M, Foringer J, Ho T, et al. Antibiotic-coated hemodialysis catheters for the prevention of vascular catheter-related infections: a prospective, randomized study. Am J Med 2003;115:352-7. 49. Leon C, Ruiz-Santana S, Rello J, de la Torre MV, Valles J, Alvarez-Lerma F, et al. Benefits of minocycline and rifampin-impregnated central venous catheters. Intensive Care Med 2004;30:1891-9. 50. Yucel N, Lefering R, Maegele M, Max M, Rossaint R, Koch A, et al. Reduced colonization and infection with miconazole-rifampicin modified central venous catheters: a randomized controlled clinical trial. J Antimicrob Chemother 2004;54:1109-15. 51. Sherertz RJ, Heard SO, Raad II, Gentry L, Bowton D, Scuderi P, et al. Gamma radiation-sterilized, triple-lumen catheters coated with a low concentration of chlorhexidine were not efficacious at preventing catheter infections in intensive care unit patients. Antimicrob Agents Chemother 1996;40:1995-7. 52. Kamal GD, Divishek D, Kumar GC, Porter BR, Tatman DJ, Adams JR. Reduced intravascular catheter-related infection by routine use of antibiotic-bonded catheters in a surgical intensive care unit. Diagn Microbiol Infect Dis 1998;30:145-52. 53. Appelgren P, Ransjo U, Bindslev L, Espersen F, Larm O. Surface heparinization of central venous catheters reduces microbial colonization in vitro and in vivo: results from a prospective, randomized trial. Crit Care Med 1996;24:1482-9. ˜ G, Giomarelli PP, Pavesi M, Luzzani A, Cattabriga I, 54. Ranucci M, Isgr A et al. Impact of oligon central venous catheters on catheter colonization and catheter-related bloodstream infection. Crit Care Med 2003; 31:52-9. 55. Carrasco MN, Bueno A, de las Cuevas C, Jimenez S, Salinas I, Sartorius A, et al. Evaluation of a triple-lumen central venous heparincoated catheter versus a catheter coated with chlorhexidine and silver sulfadiazine in critically ill patients. Intensive Care Med 2004;30:633-8. 56. McConnell SA, Gubbins PO, Anaissie EJ. Do antimicrobial-impregnated central venous catheters prevent catheter-related bloodstream infection? Clin Infect Dis 2003;37:65-72. 57. Bassetti S, Hu J, D’Agostino RBJ, Sherertz RJ. Prolonged antimicrobial activity of a catheter containing chlorhexidine-silver sulfadiazine extends protection against catheter infections in vivo. Antimicrob Agents Chemother 2001;45:1535-8. 58. Donelli G, Francolini I. Efficacy of antiadhesive, antibiotic and antiseptic coatings in preventing catheter-related infections: review. J Chemother 2001;13:595-606. 59. Pai MP, Pendland SL, Danziger LH. Antimicrobial-coated/bonded andimpregnated intravascular catheters. Ann Pharmacother 2001;35: 1255-63. 60. Gastmeier P, Zuschneid I, Geffers C. Antimicrobially impregnated catheters: an overview of randomised controlled trials. J Vasc Access 2003;4:102-10. 61. Sampath LA, Tambe SM, Modak SM. In vitro and in vivo efficacy of catheters impregnated with antiseptics or antibiotics: evaluation of the risk of bacterial resistance to the antimicrobials in the catheters. Infect Control Hosp Epidemiol 2001;22:640-6.

Ramritu et al 62. Chatzinikolaou I, Hanna H, Graviss L, Chaiban G, Perego C, Arbuckle R, et al. Clinical experience with minocycline and rifampin-impregnated central venous catheters in bone marrow transplantation recipients: efficacy and low risk of developing staphylococcal resistance. Infect Control Hosp Epidemiol 2003;24:961-3. 63. Hanna HA, Raad II, Hackett B, Wallace SK, Price KJ, Coyle DE, et al. Antibiotic-impregnated catheters associated with significant decrease in nosocomial and multidrug-resistant bacteremias in critically ill patients. Chest 2003;124:1030-8. 64. Sutton AJ, Duval SJ, Tweedie RL, Abrams KR, Jones DR. Empirical assessment of effect of publication bias on meta-analyses. BMJ 2000; 320:1574-7. 65. Crnich CJ, Maki DG. Are antimicrobial-impregnated catheters effective? Don’t throw out the baby with the bathwater. Clin Infect Dis 2004;38:1287-92. 66. McConnell SA, Gubbins PO, Anaissie EJ. Are antimicrobial-impregnated catheters effective? Replace the water and grab your washcloth, because we have a baby to wash. Clin Infect Dis 2004;39:1829-33. 67. Gillies D, O’Riordan E, Carr D, O’Brien L, Frost J, Gunning R. Central venous catheter dressings: a systematic review. J Adv Nurs 2003;44: 623-32. 68. Cook D, Kernerman P, Cupido C, King D, et al. Central venous catheter replacement strategies: a systematic review of the literature. Crit Care Med 1997;25:1417-24. 69. Hu KK, Veenstra DL, Lipsky BA, Saint S. Use of maximal sterile barriers during central venous catheter insertion: clinical and economic outcomes. Clin Infect Dis 2004;39:1441-5. 70. Safdar N, Maki DG. The pathogenesis of catheter-related bloodstream infection with noncuffed short-term central venous catheters. Intensive Care Med 2004;30:62-7. 71. Rijnders BJ, Van Wijngaerden E, Peetermans WE. Catheter-tip colonization as a surrogate end point in clinical studies on catheter-related bloodstream infection: how strong is the evidence? Clin Infect Dis 2002;35:1053-8. 72. Buyse M, Molenberghs G, Burzykowski T, Renard D, Geys H. The validation of surrogate endpoints in meta-analyses of randomized experiments. Biostatistics 2000;1:49-67. 73. Rello J, Ochagavia A, Sabanes E, Roque M, Mariscal D, Reynaga E, et al. Evaluation of outcome of intravenous catheter-related infections in critically ill patients. Am J Respir Crit Care Med 2000;162:1027-30. 74. Lark RL, Chenoweth C, Saint S, Zemencuk JK, Lipsky BA, Plorde JJ. Four-year prospective evaluation of nosocomial bacteremia: epidemiology, microbiology and patient outcome. Diagn Microbiol Infect Dis 2000;38:131-40. 75. Cooper BS, Stone SP, Kibbler CC, Cookson BD, Roberts JA, Medley GF, et al. Systematic review of isolation policies in the hospital management of methicillin-resistant Staphylococcus aureus: a review of the literature with epidemiological and economic modelling. Health Technol Assess 2003;7:1-194. 76. Salgado CD, Farr BM. Outcomes associated with vancomycin-resistant enterococci: a meta-analysis. Infect Control Hosp Epidemiol 2003;24: 690-8. 77. Buxton MJ, Drummond M, van Hout BA, Prince RL, Sheldon TA, Szucs T, et al. Modelling in economic evaluation: an unavoidable fact of life. Health Econ 1997;6:217-27. 78. Kuntz KM, Weinstein MC. Modelling in economic evaluation. In: Drummond M, McGuire A, editors. Economic evaluation in health care. Oxford, UK: Oxford University Press; 2001. p. 141-71. 79. Halpern MT, Luce BR, Brown RE, Geneste B. Health and economic outcomes modeling practices: a suggested framework. Value Health 1998;1:131-47. 80. Luce BR. What will it take to make cost-effectiveness analysis acceptable in the United States? Med Care 2005;43(Suppl 7):44-8. 81. Claxton K, Sculpher M, Drummond M. A rational framework for decision making by the National Institute for Clinical Excellence. Lancet 2002;360:711-5.

March 2008

115

82. George B, Harris A, Mitchell A. Cost-effectiveness analysis and the consistency of decision making: evidence from pharmaceutical reimbursement in Australia (1991 to 1996). Pharmacoeconomics 2001; 19:1103-9. 83. Veenstra DL, Saint S, Sullivan SD. Cost-effectiveness of antisepticimpregnated central venous catheters for the prevention of catheterrelated bloodstream infection. JAMA 1999;282:554-60. 84. Shorr AF, Humphreys CW, Helman DL. New choices for central venous catheters. Chest 2003;124:275-84. 85. Marciante KD, Veenstra DL, Lipsky BA, Saint S. Which antimicrobial impregnated central venous catheter should we use? Modeling the costs and outcomes of antimicrobial catheter use. Am J Infect Control 2003;31:1-8. 86. Borschel DM, Chenoweth CE, Kaufman SR, Vander Hyde K, VanDerElzen KA, Raghunathan TE, et al. Are antiseptic-coated central venous catheters effective in a real-world setting? Am J Infect Control 2006; 34:388-93. 87. Soufir L, Timsit JF, Mahe C, Carlet J, Regnier B, Chevret S. Attributable morbidity and mortality of catheter-related septicemia in critically ill patients: a matched, risk-adjusted, cohort study. Infect Control Hosp Epidemiol 1999;20:396-401.

Appendix 1. Data extracted from articles included in the review 1. 2. 3. 4. 5. 6. 7. 8.

Study setting Sample size and population (patients and catheters) Type of catheters compared Other preventative strategies used in catheter insertion and management Duration of catheterization Definitions of catheter colonization and CRBSI Incidence of catheter colonization and CRBSI Organisms responsible for catheter colonization and CRBSI

Appendix 2. Reasons for exclusion of articles 1. Nonresearch papers (review articles, case studies, summaries of trial published elsewhere and others). 2. Animal or in vitro laboratory studies of efficacy. 3. Abstract only available with no detail on trial methodology. 4. Studies investigated hemodialysis, tunneled, totally implantable, peripherally inserted central venous catheters, or arterial catheters and either did not include centrally placed CVCs or did not present data separately for these catheters. 5. Studies investigated long-term (.21 days) catheters and either did not include short-term CVCs or did not present data separately for these catheters. 6. Studies either conducted solely in non-ICU settings or in diverse settings with less than 50% catheters managed in the ICU and complete data not presented separately for ICU. 7. Studies included children or a mixture of children and adults and complete data not presented separately for adults.

Kamal Appelgren Bach Civetta Pemberton Sherertz Thornton George

1991 1996 1996 1996 1996 1996 1996 1997

3 1 2/3 3 1 3 3 Multi

Loo Maki Raad Heard Kamal Bach Darouiche Hannan Marik

1997 1997 1997 1998 1998 1999 1999 1999 1999

Hanley Sheng VanVliet Stoiser

Dressing

Dressing change regimen

Most common insertion site

Guidewire exchange .1 Catheter permitted? allowed?

Average duration catheterization coated/uncoated

M,G,G,D Sterile M,G,G,D M,G,G G,D n/s Asceptic Asceptic

Tegaderm Tegaderm Gauze Gauze Transparent n/s Transparent Transparent

24 hr Weekly 72 hr 72 hr Weekly 48 hr 48 hr 24 hr

Subclavian Subclavian Jugular n/s Subclavian Subclavian n/s Jugular

No No No Yes No No No No

Yes No No Yes No No Yes Yes

4.1 8.0 (med) 6.6 7.0 10.0 7.9 n/s n/s

5.0 12.0 (med) 6.6 4.7 11.0 8.8 n/s n/s

3 3 3 3 3 2 3 3 3

n/s PI PI/CH PI n/s Alcohol PI PI PI

Asceptic M,G,G,D M,G,G,D M,G,G,D M,G,G,D M,G,G,D M,G,G,D Asceptic M,G,G,D

n/s Gauze Gauze Polyskin2 Transparent Gauze n/s Tegaderm Transparent

n/s 48 hr 72 hr 48 hr Weekly 48 hr 48 hr 48 hr 48 hr

n/s Subclavian Subclavian Subclavian Subclavian Jugular Subclavian Jugular Jugular

n/s Yes No Yes Yes Yes No No No

Yes Yes Yes Yes Yes No Yes Yes No

6.6 6.0 6.0 (med) 9.0 2.9 4.1 8.2 7.6 6.0

2000 2000 2001 2002

3 3 2 3

PI PI CH Alcohol

M,G,G,D M,G,G,D M,G,G,D M,G,G,D

n/s Gauze Transparent Gauze

n/s 48 hr n/s n/s

Jugular Jugular Subclavian Subclavian

Yes No No No

No Yes No No

6.6 6.0 6.0 (med) 8.5 4.3 4.5 8.4 7.5 6.0 and 6.0 9.9 9.1 9.1 12.8*

Theaker Chatzinikolaou

2002 2003

4 2

n/s PI

Asceptic MSB

Tegaderm n/s

48 hr n/s

Jugular Femoral

n/s No

Yes No

Corral Ranucci Richards BrunBuisson Carrasco Leon Yucel Dunser

2003 2003 2003 2004 2004 2004 2004 2005

3 2 3 1/2 3 3 3 3/4

PI PI/CH 2% PI n/s PI PI Alcohol CH

M,G,G,D M,G,G M,G,G,D MSB M,G,G,D M,G,G,D M,G,G,D M,G,G,D

Gauze Transparent Gauze n/s Gauze Gauze Gauze Gauze

72 hr 48 hr 48 hr n/s 48 hr 48 hr 24 hr 24 hr

Subclavian Subclavian Subclavian Subclavian Femoral n/s Jugular Subclavian

Yes No Yes Yes No No No No

Yes No Yes Yes Yes No No Yes

Moretti Rupp Fraenkel

2005 2005 Unpub

n/s PI CH

M,G,G,D n/s M,G,G,D Transparent M,G,G,D Transparent

n/s 72 hr Weekly

Jugular Jugular Subclavian

No Yes No

No No No

3 3 3

7.4 8.0 12.0 9.0 8.4 10.5 8.8 10.3 7.5 9.3 and 9.7 6.7* 7.0* 6.0

10.7 8.2 10.2 12.1* 7.2 8.0 14.0 9.1 7.8 12.0 9.3 10.4 6.7 10.7 5.9* 6.8* 6.0

ICU ICU ICU ICU MIX ICU ICU Trauma ICU ICU ICU MIX ICU ICU ICU MIX ICU ICU

Patients

Study design

General General Cardiac General General tpn General General Transplant

RCT RCT RCT OBS RCT RCT RCT RCT

General General General General General Cardiac General General General

OBS RCT RCT RCT OBS RCT RCT RCT RCT

General General General Immunocompromised ICU Long stay Cancer Hemodialysis ICU ICU General MIX General MIX General ICU General ICU General ICU General MIX General ICU General

OBS RCT RCT RCT

RCT RCT OBS RCT RCT RCT RCT RCT

MIX ICU ICU

RCT RCT RCT

ICU ICU ICU ICU

General General General

RCT RCT

PI, povidone iodine; CH, chlorhexidine; M,G,G,D, mask, gown, gloves, drape; tpn, total parenteral nutrition; MSB, maximal sterile barriers; ICU, intensive care unit; MIX, ICU and ward; RCT, randomized-controlled trial; OBS, cohort observational study. *Means calculated from data in the published paper where medians are presented.

Ramritu et al

PI CH 0.5% Alcohol PI/CH 4% n/s n/s PI CH 2%

Study setting

Vol. 36 No. 2

Author

Sterile No. Skin insertion Year lumens preparation procedure

116

Appendix 3. Heterogeneity in study design, setting and patients included, and cointerventions used for catheter insertion and management

Ramritu et al

March 2008

117

Appendix 4. Catheter colonization as a surrogate for CRBSI A Spearman correlation coefficient was calculated, and a regression line with robust standard errors was fitted to a scatter plot of the incidence of catheter colonization and CRBSI in each study arm to investigate the appropriateness of colonization as a surrogate for CRBSI.72 Adjustment was made for catheter type and definition of CRBSI used. There was a moderate linear correlation (Fig 5) between the incidence of catheter colonization and incidence of CRBSI across the 59 arms of the 29 studies reporting both outcomes (r 5 0.63; r2 5 0.4; incidence of CRBSI 5 20.01 (SE, 0.01) 1 0.24 (SE, 0.07) 3 incidence of colonization; P 5 .003). This is similar to earlier findings,72 and the result was robust to adjustment for type of coating and definition of CRBSI (Table 4).

Fig 5. Scatter plot of incidence of catheter colonization against incidence of CRBSI. Note: Linear regression line with 95% CI marked. Incidence of CRBSI 5 20.01 1 0.24 3 incidence of colonization, P 5 .003.