Journal of Hospital Infection (2002) 50: 77–81
`
`Letters to the Editor
`
`doi:10.1053/jhin.2001.1134, available online at
`http://www.idealibrary.com on
`
`Needleless connectors—the way forward
`in the prevention of catheter-related infections?
`
`Sir,
`in
`Intravascular catheters have become integral
`patient management. Indeed, in the UK, 200 000
`central venous catheters (CVC) are used annually.1
`However, indwelling intravascular devices are also
`associated with many complications
`including
`infection. Micro-organisms, in particular coagulase-
`negative staphylococci (CNS), originating from the
`skin of healthcare workers or patients may gain access
`to intravascular catheters via contamination of hubs
`and stopcock entry ports. The micro-organisms
`migrate to the catheter distal tip resulting in coloni-
`zation and subsequent infection. In a recent study,
`22% of stopcock entry ports and 31% of arterial line
`hubs were contaminated with micro-organisms
`within 72 h of placement.2 In an attempt to reduce
`hub colonization, daily disinfection with 70% ethanol
`and the use of a povidone-iodine gauze hub cover
`have been employed.3,4 Needleless connectors have
`recently been introduced into clinical practice to
`reduce the risk of needlestick injuries. These devices
`may also facilitate aseptic technique and reduce the
`time spent on manipulation of intravenous connec-
`tions. The potential infection risk of needleless con-
`nectors is, at present, unclear with studies yielding
`conflicting results.5,6
`We have investigated the external and internal
`microbial contamination rates of two types of needle-
`less connector—Clave1 and Posiflow1 (Becton
`Dickinson, Utah, USA)—in a prospective rando-
`mized study (Figure 1). Thirty-eight patients from
`the cardiac and haematology departments at the
`University Hospital Birmingham NHS Trust who
`required a CVC as part of their clinical management
`were recruited into the study and assigned to have
`either the Clave or Posiflow attached to the hubs/
`stopcocks of their CVC. The needleless connectors
`were disinfected with chlorhexidine gluconate 0.5%
`w/v in industrial methylated spirit BP 70% w/w spray
`(Adams Healthcare, Leeds, UK) prior to and fol-
`lowing each manipulation. All administrations and
`
`Figure 1 Triple lumen central venous catheter complete with 3-way
`1
`, standard
`taps and needleless connectors. From left to right; Posiflow
`1
`luer caps and Clave
`.
`
`aspirations made through the connectors 72 h post
`CVC insertion were recorded. Each device remained
`in situ attached to a CVC for 72 h after which they
`were removed and investigated for microbial con-
`tamination. Two hundred and fifty-five needleless
`connectors (128 Clave, 127 Posiflow) were studied.
`The external surface of the silicone compression seals
`were pressed onto 7% blood agar (Oxoid, Basing-
`stoke, UK) 10 times after which the plates were
`incubated at 37(cid:14)C in air for 48 h. The silicone com-
`pression seals were then cleaned with a 70% v/v iso-
`propyl alcohol swab (Sterets1, Seton Prebble Ltd,
`Merseyside, UK). The internal microbial contami-
`nation rate of the needleless connectors was assessed
`by flushing with 100 mL of brain–heart infusion
`broth (Oxoid, Basingstoke, UK) which was inocu-
`lated on to 7% blood agar (Oxoid) and incubated in
`air at 37(cid:14)C for 48 h. The external and internal
`microbial contamination rates of the Clave and
`Posiflow were compared by Fisher’s exact test.
`
`1095-6701/02/010077(cid:135)05 $35.00/0
`
`& 2002 The Hospital Infection Society
`
`BAXTER EXHIBIT 1017
`Page 1 of 3
`
`

`

`78
`
`Letters to the Editor
`
`There was no significant difference between the
`rate of external microbial contamination of the
`Posiflow (28.3%) and Clave (39%) (P (cid:136) 0.085). Skin
`and environmental micro-organisms were recovered
`from the external silicone compression seal of the
`connectors including CNS (91.9%), environmental
`Gram-negative rods (8.1%), Staphylococcus aureus
`(4.7%) and Gram-positive rods
`(predominantly
`Corynebacterium spp.) (3.5%). There was also no
`significant difference between the internal micro-
`bial contamination rates of the devices, Posiflow
`(P (cid:136) 0.571). Micro-
`(5.5%) versus Clave (4%)
`organisms recovered from the internal cultures of
`the connectors included CNS (91.7%), Staph. aureus
`(8.3%) and Gram-positive Rods (predominantly
`Corynebacterium spp.) (8.3%). All of the Clave devi-
`ces with internal microbial contamination also had
`external contamination, however identical micro-
`organisms were not always recovered. In compar-
`ison, 57.1% of Posiflow connectors were both
`internally and externally contaminated. The mean
`number of manipulations per connector for each
`type of device during the 72 h period was 1.6
`(range (cid:136) 0 to 17). There was no correlation between
`the number of manipulations and the presence of
`internal microbial contamination (P (cid:136) 0.5, Spearman
`Rank Test) and no correlation between the number
`of manipulations and the extent of internal micro-
`bial contamination (P (cid:136) 0.5, Spearman Rank Test).
`Seventy-five percent of the devices with micro-
`organisms present internally were used for con-
`tinuous infusions or for central venous pressure
`(CVP) monitoring, which would have resulted in
`the connectors being continually in use during the
`study period. Drugs and fluids including propofol,
`adrenaline, dopamine, 5% dextrose and blood were
`administered through the contaminated devices
`and may,
`therefore, have supported microbial
`growth.7
`The number of needleless connectors through
`which each drug was administered was similar for
`both types of device. Although the average internal
`contamination rate of the devices was 4.8%, this was
`significantly lower than the contamination rates of
`open ports reported in a previous study2 which
`suggests that the devices, if cleaned appropriately,
`may reduce contamination and subsequent infection
`rates. Furthermore, the Posiflow device has the
`added feature of maintaining positive pressure in the
`catheter. Therefore, when the syringe is removed,
`backflow of fluid up the catheter is prevented,
`maintaining catheter patency.
`
`The microbial contamination rate associated with
`needleless connectors may be further reduced if the
`devices are used in conjunction with pre-filled syr-
`inges.8 The reduced levels of microbial conta-
`mination associated with pre-filled syringes and
`needleless connectors compared with conventional
`systems is encouraging, and the use of these devices
`may be the way forward in preventing sepsis
`associated with intravascular catheters.
`The authors would like to thank: David Quinn for
`recruiting patients into the study, all the staff from
`cardiac theatres and everyone in the intensive care,
`cardiac surgery and haematology/bone marrow
`transplant units.
`This work was supported by an educational grant
`from Becton Dickinson, UK.
`
`A. L. Casey*,
`M. K. Spare*,
`T. Worthington*,
`M. H. Faroquiy,
`E. Trotterz,
`P. A. Lambertxand
`T. S. J. Elliott*
`
`*Department of Clinical
`Microbiology; yDepartment of
`Anaesthetics and Intensive
`Care; zHaematology
`Department, University
`Hospital Birmingham
`NHS Trust, Queen Elizabeth
`Hospital, Edgbaston,
`Birmingham, B15 2TH
`and xDepartment of
`Pharmaceutical and Biological
`Sciences, Aston University,
`Aston Triangle, Birmingham,
`B4 7ET, UK
`
`References
`
`1. Elliott TSJ. Line-associated bacteraemias. Commun Dis
`Rep 1993; 3: 7.
`2. Tebbs SE, Ghose A, Elliott TSJ. Microbial con-
`tamination of intravenous and arterial catheters. Inten-
`sive Care Med 1996; 22(3): 272–273.
`3. Salzman MB, Isenberg HD, Rubin LG. Use of disin-
`fectants to reduce microbial contamination of hubs of
`vascular catheters. J Clin Microbiol 1993; 31: 475–479.
`4. Sitges-Serra A, Linares J, Perez JL, Jaurrieta E,
`Lorente L. A randomised trial on the effect of tubing
`changes on hub contamination and catheter sepsis during
`parenteral nutrition. J Parenter Enter 1985; 9: 322–325.
`5. Kellermen S, Shay DK, Howard J et al. Bloodstream
`infections in home infusion patients: The influence of
`race and needleless intravascular access systems. J
`Pediatr 1996; 129: 711–717.
`6. Brown JD, Moss HA, Elliott TS. The potential for
`catheter microbial contamination from a needleless
`connector. J Hosp Infect 1997; 36(3): 181–189.
`7. Graystone S, Wells MF, Farrell DJ. Do intensive care
`drug infusions support microbial growth? Anaesth
`Intens Care 1997; 25(6): 640–642.
`
`BAXTER EXHIBIT 1017
`Page 2 of 3
`
`

`

`Letters to the Editor
`
`79
`
`8. Worthington T, Tebbs S, Moss H, Bevan V, Kilburn,
`Elliott TSJ. Are contaminated flush solutions an
`overlooked source for catheter related sepsis? J Hosp
`Infect 2001; 49: 81–83.
`
`doi:10.1053/jhin.2001.1055, available online at
`http://www.idealibrary.com on
`
`An outbreak of vancomycin-resistant enterococci
`associated with major ward refurbishment
`
`Sir,
`
`During 1998, an outbreak of infection caused by
`vancomycin-resistant enterococci (VRE) occurred
`on a renal unit, following major ward refurbishment.
`In January 1998, patients were relocated temporarily
`to allow the refurbishment, which included re-siting
`of the sluice and toilets, installation of new hand-
`wash basins and showers, and the provision of a new
`bathroom. The refurbished ward was re-opened over
`the last weekend in May 1998, but members of the
`infection control team (ICT) were unable to inspect
`the ward because this took place over a Bank Holiday
`weekend, and access had been restricted on the pre-
`ceding Friday due to ongoing work. When patients
`were moved back, however, some work was still in
`progress.
`Whilst on the temporary ward, just before the
`move back to the refurbished unit, VRE were cul-
`tured from peritoneal dialysis effluent in one patient
`with peritonitis and from the urine of another. In
`early June, VRE were isolated from two more
`patients on the refurbished ward and an outbreak
`control group was formed.
`Several major problems were identified on the
`refurbished ward. Handwashing facilities were poor:
`several new handwash basins were not plumbed in
`and others had no taps. At functioning handwash
`basins, no liquid soap was available because new
`dispensers, which could not accept the soap car-
`tridges used throughout the remainder of the hos-
`pital, had been fitted. At some basins, paper towel
`dispensers had not been fitted and paper towels were
`held on the rim of basins where they could be easily
`contaminated by splashing during hand washing.
`The ward cleaners’ room was undergoing conversion
`to an office (not included in the original plans),
`necessitating removal of cleaning equipment to the
`sluice, where it was stored next to the bedpan washer.
`Floor polisher buffer pads were being cleaned in the
`sluice with a toilet brush.
`
`Infected patients were isolated, and non-func-
`tional handwash basins were reinstated. Soap com-
`patible with the new dispensers was ordered urgently
`and alcohol hand rub was provided in the interim. An
`intensive ward clean was initiated using neutral
`detergent and water, followed by a 1000 ppm aqu-
`eous hypochlorite solution. The cleaners’ room was
`reinstated. Use of vancomycin as first-line therapy
`for Clostridium difficile infection and for empiric
`therapy for suspected vascular access device-asso-
`ciated infection was prohibited.
`Environmental screening for VRE on the refur-
`bished ward, the maintenance haemodialysis unit
`and temporary ward, and screening for faecal car-
`riage of VRE in renal inpatients was instituted. Iso-
`lates were typed using pulsed-field gel electrophoresis
`(PFGE) of Sma1-digested DNA.1
`All outbreak isolates were identified as E. faecium
`and displayed the VanA phenotype. The isolate from
`the presumed index case with CAPD peritonitis
`yielded a PFGE banding pattern designated ‘A’. A
`blood culture from a patient with haematological
`malignancy and peritoneal fluid from another epi-
`sode of CAPD peritonitis also yielded isolates
`which were ‘closely related’ and ‘possibly related’ to
`outbreak strain A, respectively using the criteria of
`Tenover et al.2 Strain A was not recovered from
`any environmental swabs. In the second chrono-
`logical case, VRE was cultured from multiple ulcer
`swabs, multiple superficial sites and faeces as well as
`urine. The PFGE banding pattern of this strain
`differed from that of the index case and was desig-
`nated ‘B’. This was isolated subsequently from the
`blood and pleural fluid, respectively, of two addi-
`tional patients.
`Of the 51 environmental sites sampled initially on
`the refurbished ward, 13 (25%) were positive for
`VRE (strain B only). Only one site was positive on
`the haemodialysis unit and none on the temporary
`ward. Screening for faecal carriage of VRE among
`acute renal inpatients identified two further carriers
`who had not previously been detected by routinely
`obtained clinical samples.
`Before the outbreak described in this report, no
`faecal carriers of VRE had been identified among
`19/70 CAPD patients and 91/120 maintenance
`haemodialysis patients. However, not all of these
`patients in this incomplete screening exercise had
`been in recent contact with the ward in which the
`outbreak occurred. Nevertheless, the apparent lack
`of faecal colonization and the rarity of VRE from
`samples obtained for routine diagnostic purposes,
`
`BAXTER EXHIBIT 1017
`Page 3 of 3
`
`