Elsevier

The Lancet

Volume 354, Issue 9189, 30 October 1999, Pages 1504-1507
The Lancet

Articles
Rapid diagnosis of central-venous-catheter-related bloodstream infection without catheter removal

https://doi.org/10.1016/S0140-6736(99)04070-2Get rights and content

Summary

Background

Current methods for the diagnosis of bloodstream infection related to central venous catheters (CVC) are slow and in many cases require catheter removal. Since most CVC that are removed on suspicion of causing infection prove not to be infected, removal of catheters unnecessarily exposes patients to the risks associated with reinsertion.

Methods

The gram stain and acridine-orange leucocyte cytospin test (AOLC) is rapid (30 min), inexpensive, and requires only 100 μL catheter blood (treated with edetic acid) and the use of light and ultraviolet microscopy. We assessed the gram stain and AOLC test in suspected cases of catheter-related bloodstream infection, in comparison with two methods requiring catheter removal (tip roll and tip flush), and a third technique, done in situ (endoluminal brush) in conjunction with quantitative peripheral-blood cultures.

Findings

128 cases of suspected catheter-related bloodstream infection were assessed in 124 adult surgical patients (median duration of CVC placement was 16 days). In 112 (88%) cases CVC blood was obtainable. Catheter-related bloodstream infection was diagnosed in 50 cases (culture of the same organism from the catheter, in significant numbers, and from peripheral-blood culture). The sensitivity of the gram stain and AOLC test was 96% and the specificity was 92%, with a positive predictive value of 91% and a negative predictive value of 97%. By comparison, the tip-roll, tip-flush, and endoluminal-brush methods had sensitivities of 90%, 95%, and 92%, and specificities of 55%, 76%, and 98%, respectively.

Interpretation

The gram stain and AOLC test is a simple, and rapid method for the diagnosis of catheter-related bloodstream infection. This diagnostic method compares favourably with other diagnostic methods, particularly those that require the removal of the catheter, and can permit early targeted antimicrobial therapy.

Introduction

Central venous catheterisation is essential for the intensive management of both medical and surgical patients. An estimated 5 million central venous catheters (CVC) are used in the USA alone each year, a figure likely to increase as care of patients becomes more specialised and intensive.1 Complications of catheterisation include those associated with catheter insertion (pneumothorax and arterial and nerve injuries) and those associated with long catheter use (thrombosis and infection). CVC infection encompasses many definitions that may or may not represent a clinical concern. Catheter-related bloodstream infection is the most serious infection associated with catheters and occurs in 3–7% of patients with CVC; it is simply defined as peripheral bacteraemia caused by the same microorganism as cultured from the catheter.2 Conventional methods to diagnose catheter-related bloodstream infection generally require that the catheter is removed so that the microorganisms colonising it can be cultured. However, in up to 85% of patients the culture is negative, despite clinical suspicion of catheter-related infection.3, 4 Furthermore, culture results for peripheral blood and CVC are not available for at least 24 h, and hence even in-situ methods, such as quantitative blood cultures and endoluminal brushing, require clinicians to make empirical judgments on the need for CVC removal. Therefore, there is a need for a rapid method to diagnose catheter-related bloodstream infection that does not require removal of the catheter. The gram stain and acridine-orange leucocyte cytospin (AOLC) test, done on blood samples withdrawn through the CVC, is effective in the rapid diagnosis of bloodstream infection in neonates,5 but has yet to be proven in adults. This technique has the potential for both the rapid detection of microorganisms (bacteria or fungi), and permits early gram-stain classification of pathogens, allowing targeted antibiotic therapy. This study aimed to assess the accuracy of the gram stain and AOLC test in the detection of catheter-related bloodstream infection in a predominantly adult surgical population.

Section snippets

Methods

Between January, 1996, and December, 1997 we reviewed surgical patients with CVC every day for clinical signs of catheter-related bloodstream infection (temperature >37°C, white-blood-cell count >11×109/L, or evidence of infection at the site of catheter skin entry). The presence of other sources of infection was noted, but this diagnosis was not used to exclude the catheter as a potential source of infection. The type of CVC, duration of catheterisation, site of catheter insertion, and the

Results

128 cases of suspected catheter-related bloodstream infection were assessed in 124 adult surgical patients. In 16 (13%) cases we were not able to obtain throughcatheter blood samples and have subsequently excluded these from the analysis. The implicated catheters were 20 single-lumen lines, seven Hickman catheters, nine double-lumen renal-dialysis catheters, and 76 triplelumen catheters. 80% of these catheters had been used to deliver total parenteral nutrition. The median duration of CVC

Discussion

Bloodstream infection is associated with a significant degree of morbidity; case mortality of catheter-related bloodstream infection is 10–20%.9 There has been an increase of two to three fold in primary nosocomial bloodstream infections in the past decade,10 most caused by CVC. Such a large number of cases has large medical and financial implications for health-care institutions, with cost estimates of $28 000–40 000 for each case of bloodstream infection.11, 12 Since most CVC removed after

References (28)

  • DG Maki et al.

    A semi-quantitative culture method for identifying intravenous-catheter-related infection

    N Engl J Med

    (1997)
  • DJ Cleri et al.

    Quantitative culture of intravenous catheters and other intravascular inserts

    J Infect Dis

    (1980)
  • SN Banerjee et al.

    Secular trends in nosocomial primary bloodstream infections in the United States: 1980–1989

    Am J Med

    (1991)
  • D Pittet et al.

    Nosocomial bloodstream infections in critically ill patients: excess length of stay, extra cost, and attributable mortality

    JAMA

    (1994)
  • Cited by (0)

    View full text