Intended for healthcare professionals


An endoluminal brush to detect the infected central venous catheter in situ: a pilot study

BMJ 1996; 313 doi: (Published 14 December 1996) Cite this as: BMJ 1996;313:1528
  1. Mark J Tighe, research fellowa,
  2. Peter Kite, consultant microbiologistb,
  3. Warren N Fawley, research assistantb,
  4. Daniel Thomas, research assistantb,
  5. Michael J McMahon, professor of surgerya
  1. a Nutrition Support Service and University Department of Surgery, General Infirmary, Leeds LS1 3EX
  2. b Department of Microbiology, General Infirmary, Leeds
  1. Correspondence to: Professor McMahon.
  • Accepted 20 August 1996

Catheter related sepsis is a potentially life threatening infection caused by an indwelling intravenous catheter, with the same organism cultured from peripheral blood and from the removed catheter. If such sepsis is suspected the catheter is usually removed, although subsequent bacterial culture shows most catheters to be sterile.1 2 We evaluated a new brush (Endoluminal, FAS Medical, London) for detecting infection in the lumen of a central venous catheter. The brush has nylon bristles wound tightly around the distal end of a stainless steel wire and slides along the lumen of the catheter to its distal (inner) end. It is then removed for culture. The technique is based on the principle that bacteria collect on the fibrin sleeve on the catheter's inner surface; fibrin thus becomes enmeshed in the brush's bristles.

Methods and results

We studied 115 catheters from 112 ward based surgical patients (16–91 (median 65) years) receiving intravenous nutrition whose central venous catheters (diameter 14–18 mm) were going to be removed either on termination of treatment (n=44) or if the catheter was thought to be the source of infection, causing either sepsis or unexplained fever (n=71). We obtained informed consent from each patient, and the study was approved by our local ethics committee.

The brush was passed along then withdrawn from the catheter with the catheter in situ. The catheter was then removed and the tip sent for culture. Blood was aspirated from a peripheral vein (a) before brushing to determine if catheter related sepsis was present and (b) one minute after brushing to determine if a bacteraemia had been induced by brushing a colonised catheter (growth of >/=100 colony-forming units externally3 or internally4). We assessed patients immediately after and one hour after removal of the brush, for evidence of systemic upset.

The external surface of the catheter was cultured by using the Maki roll technique3 and the internal surface by a modified vortex and sonication technique.4 The number of colony forming units at 24 hours were counted and recorded. The brush was rolled back and forth across a 5% horse blood agar plate at least four times.


The organism responsible for most of the colonised catheters was the coagulase negative staphylococcus (18 catheters). Other organisms were Escherichia coli (4), Staphylococcus aureus (3), Streptococcus faecalis (2), Candida albicans (3), and a mixed growth (9). Brush and tip organisms were phenotypically similar in all positive cases.

The brush was positive in 14/15 (93%) (table 1) cases of catheter related sepsis. In colonised catheters, including catheter related sepsis, the sensitivity of the brush was 0.77 when compared with the Maki roll and sonication. When a negative brush culture was compared with sterile (no intraluminal or extraluminal growth) or contaminated (growth of <100 colony-forming units3) catheters, the specificity was 1.00. The brush therefore delivered a positive predictive value of 1.0 when compared with colonisation of the catheter tip, and a negative predictive value of 0.89.

Table 1

Performance of brush roll cultures compared with external (Maki roll) and internal (sonication) tip cultures. Values are numbers of catheters

View this table:

Altogether, 36/71 (51%) catheters were found to be sterile in the group of patients thought to have infection caused by the catheter.

No patient recorded symptoms or signs of systemic upset during or after the procedure. A bacteraemia was induced in 3/50 (6%) patients after brushing, although repeat cultures at 24 hours were negative in all cases (and no antibiotics were used).


A reliable test not requiring removal of the catheter is needed to identify catheter related sepsis. In our study, the brush detected organisms in 77% of colonised catheters and 93% of cases of sepsis. It also detected organisms in 11.4% of patients in whom infection was not suspected. Although the brush missed 11% of colonised catheters, this compared favourably with the Maki roll (12%) and sonication (13%).

The brush may have the potential to reduce the removal of sterile catheters wrongly suspected of harbouring infection.


  • Funding FAS Medical provided brushes and equipment to perform microbiological testing.

  • Conflict of interest None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
View Abstract