Criteria abstracted from The Users' Guide to Medical Literature, from the Health Information Research Unit and Clinical Epidemiology and Biostatistics, McMaster University

Highlighted lines and questions below provide links to the pertinent description of criteria in The EBM User's Guide, now available at the Canadian Centres for Health Evidence

Article Reviewed:

Please visit the new Evidence Based Journal Club Reviews

Prevention of nosocomial bloodstream infections: effectiveness of antimicrobial-impregnated and heparin-bonded central venous catheters

Marin MG, Lee JC, Skurnick JH.

Crit Care Med. 2000 Sep;28(9):3332-8. [abstract]

Reviewed by Kshitij P. Mistry MD, Children's Hospital Boston

Review posted December 4, 2002

---

I. Are the results of the study valid?

A. Primary questions:

1. Did the overview address a focused clinical question?

Yes. The authors clearly state the purpose of performing this meta-analysis is to "examine the effectiveness of antimicrobial-impregnated and heparin-bonded catheters relative to standard central venous catheters in lessening catheter-related bloodstream infections." In addition, the authors also hoped to assess the potential cost-effectiveness of utilizing these treated catheters compared to standard catheters. Several other outcomes were also recorded, including the association of bloodstream infection with duration of catheter use, identification of microorganisms isolated in catheter-related bloodstream infections and subgroup analysis of antimicrobial-impregnated versus heparin-bonded central venous catheters.

2. Were the criteria used to select articles for inclusion appropriate?

Yes. The authors included all randomized, controlled studies comparing the use of antimicrobial-impregnated or heparin-bonded catheters and the use of standard central venous catheters. However, any studies relating to children were excluded (defined as subjects less than 17 years of age). Additionally, articles were excluded from consideration if they (i) did not contain original data of relevance to the central question, (ii) were nonrandomized or uncontrolled or (iii) related to animal subjects.

B. Secondary questions:

3. Is it unlikely that important, relevant studies were missed?

Yes. The authors performed computer searches of MEDLINE (1996 to December 1999), the Nursing and Allied Health (CINAHL) database (1982 to October 1999), Current Contents - Clinical Medicine and Life Sciences (January 12, 1998 to December 13, 1999) and the HealthSTAR database (1991 to December 1999) in addition to examining the reference lists of these identified articles. The authors also corresponded with each of the authors of the identified articles as well as with acknowledged experts.

Eleven trials were identified for full analysis.

4. Was the validity of the included studies appraised?

From each eligible study, the authors abstracted the type of control, the study setting and the type of experimental catheter used; however, there is no mention of evaluating each study's randomized group allocation, blinding methodology, patient characteristics or severity of illness between control and treatment groups or assessment of group treatment aside from the experimental intervention.

5. Were assessments of studies reproducible?

Not clear. The authors do not report the number of reviewers assessing potential studies, evaluating methodological quality or extracting pertinent information.

6. Were the results similar from study to study?

The authors utilized tests of homogeneity to determine if discrepancies among individual study results were secondary to chance or due to differences in patients, exposures, outcome measures or research methods. The more significant the test (p < 0.05), the less likely the observed differences are attributable to chance alone.

Four of the 11 studies included in the meta-analysis concluded that the experimental catheters did not have a significant protective effect against bloodstream infections. The authors attributed this absence of a statistically significant difference between catheter types to the small number of catheterized patients and the low occurrence of catheter-related bloodstream infections.

Additionally, the chi-square analysis of homogeneity suggested that the protective effect was not uniform across all studies (chi-square, 22.35; df = 10; p < 0.025). The authors subsequently performed subgroup analysis of the nine silver sulfadiazine and chlorohexidine group (SS-C) studies and the two studies utilizing either heparin bonding or coating with minocycline and rifampin (non-SS-C). The chi-square analysis of homogeneity of the odds ratio within the SS-C group yielded a chi-square value of 12.03 with eight degrees of freedom (p > 0.1).

The authors also performed a DerSimonian-Laird test of homogeneity of the differences in rates of bloodstream infections between the experimental and standard central venous catheters. This test of homogeneity did not find significant inter-study differences (chi-square, 13.93; df = 10; p = 0.18).

II. What are the results?

1. What are the overall results of the review?

Results were reported as the "chance" of developing a catheter-related bloodstream infection when using an antimicrobial-impregnated or heparin-bonded central venous catheter relative to a standard central venous catheter. The summary odds ratio calculated by the Peto method (ORp) for all studies was 0.58 (a value of less than one indicates a reduction in infections with the experimental catheters). The 95% confidence interval (for all studies in addition to subgroup analyses of the SS-C studies and the non-SS-C group) reportedly do not include one and therefore, indicates a significant reduction in the rates of infection utilizing the experimental catheters. The authors also report that subgroup summary odds ratios did not differ significantly from the ORp of the total group; although these specific values are not included, this interpretation is represented graphically in the paper's Figure 1.

The summary estimate of the rate difference (RDs) calculated by a general variance-based method is 2.32%. This positive value implies that the experimental catheters were associated with a reduction of bloodstream infections. The 95% confidence interval (for all studies in addition to subgroup analyses of the SS-C studies and the non-SS-C group) reportedly do not include zero and therefore, indicate a significant reduction in the rate of infections. Again, the authors do not provide specific values of all subgroup statistical analyses, but represent them graphically in Figure 2.1

Although no calculations are provided, the authors report that the catheters not impregnated with silver sulfadiazine and chlorohexidine (non-SS-C) had a significantly lower ORp than the SS-C central venous catheters.

The average infection rate of the control catheters was estimated to be 5.10% (sum of the number of infections observed with control catheters divided by the total number of control catheters). Based on the overall RDs of 2.32%, the experimental catheter infection rate was calculated as 2.78% (5.1% minus 2.32%).

The organisms reportedly isolated from infected catheters and catheter-related bloodstream infections were "similar" in both the control and experimental catheter groups; however, the antibiotic resistance patterns were not published in the original studies.

In nine of the 11 studies, information regarding catheter duration was provided. Based on this data, the duration of catheterization reportedly did not differ significantly between the control and experimental groups (no statistical analyses provided).

2. How precise were the results?

The 95% confidence interval (CI) for the summary odds ratio [0.58 (0.40, 0.84)] and the summary estimate of the rate difference [2.32% (1.04%, 3.61%)] of all studies were relatively narrow. Additionally, the 95% CI for the ORp did not include one, indicating significant reduction in infection risk with experimental catheters; the 95% CI for the RDs did not include zero, indicating a significant decrease in infection rates with the use of experimental central venous catheters.

III. Will the results help me in caring for my patients?

1. Can the results be applied to my patient care?

Probably. Although none of the reviewed studies included pediatric patients, there is no reason why these results could not be extrapolated to individuals admitted to the pediatric intensive care unit.

2. Were all clinically important outcomes considered?

In addition to the evaluating the reduction in infection rates utilizing antimicrobial-impregnated or heparin-bonded catheters versus standard central venous catheters, the authors also attempted to assess differences between SS-C and non-SS-C catheters, microbial resistance and cost-effectiveness. However, the small number of studies precludes any definitive conclusions regarding experimental catheter type (SS-C, heparin bonded or coating with minocycline and rifampin) and infection reduction. Additionally, insufficient data exists to conclusively state that the use of experimental catheters does not result in increased bacterial resistance.

3. Are the benefits worth the harms and costs?

The authors conclude that utilizing experimental catheters is more cost-effective than the use of a standard central venous catheter. This is based on (i) cost information obtained from University Hospital in Newark, NJ, (ii) the assumption that microorganisms associated with experimental catheters exhibit no greater virulence that those associated with standard catheters and (iii) that the cost of treating a catheter-related infection exceeds $3,495. The authors suggest that this savings may be more pronounced with the use of less expensive heparin bonded catheters, but also admit that this meta-analysis is insufficient in size to support this conclusion. Additionally, studies are necessary to definitively ascertain microbial resistance patterns with the use of experimental catheters and to evaluate the infection rate difference between antimicrobial-impregnated and heparin bonded catheters.

Comments

Back to the EB Journal Club Index