Literature Update - June 2008
Compiled by Dr Leon Worth, ID Physician
Implementing a Surveillance Strategy for Dialysis-associated Bacteraemia: Results from a UK Centre
A high risk for bacteraemia is observed in patients receiving haemodialysis. An important risk factor is the presence of a venous catheter
for vascular access, and rates of bacteraemia are significantly higher when compared with patients with arteriovenous fistulae. In 1999, the
CDC established the Dialysis Surveillance Network. In 2002, equivalent methods were piloted at a UK centre – the Hammersmith Hospital dialysis
unit, London.
George A et al (BMJ 2006; 332:1435) report on outcomes following 31 months of continuous surveillance. Cases were identified whenever a
dialysis patient was admitted to hospital or started intravenous antibiotics. The type of vascular access and whether it was removed
was recorded, together with blood culture results and the presence or absence of criteria for infection. To enable calculation of a
denominator (‘per 100 patient months’), a census was completed for all patients dialysed in the first week of each surveillance month.
Access-related bacteraemia was defined as a positive blood culture in the presence of clinical signs of infection, with the patient
started on intravenous antibiotics or admitted to hospital, and the suspected source of the positive blood culture being the vascular
access or uncertain. Crude and adjusted (for access type) rates were compared with data obtained from US dialysis units. Rates were calculated
for the first 4 months of surveillance (June to September 2002), and compared with the last 4 months of surveillance (September to December 2004).
At the studied dialysis unit (112 patients dialysed per month), the surveillance strategy required 2 hours per month of a consultant’s time,
with the scheme embedded into day-to-day activities of the unit. 74% of the patients in the unit had tunnelled catheters. During the
first 4 months of surveillance, the rate of access-related bacteraemia was significantly higher than in the last 4 months of surveillance
(6.2 vs. 2.0 per 100 patient months). Antibiotic usage also decreased significantly, with the rate of commencing intravenous antibiotics
and the rate of commencing vancomycin being significantly higher in the first 4 months of surveillance, compared to the final 4 months
(p < 0.001). Similarly, the rate of admission to hospital due to access-related infection was higher in the first 4 months compared to the
final 4 months of surveillance (4.0 vs. 1.4 per 100 patient months). When these 4 outcomes were compared with all US centres, significantly
higher rates were observed in the first 4 months of surveillance, whereas rates were not statistically higher during the last 4 months of
monitoring.
The comparison of data with US data is likely to be unsatisfactory, given the potential for differences in case mix, staffing and local
practices to impact upon infection rates and antibiotic usage. Despite this limitation, the in-house comparisons provided by this study
do suggest a reduction in infection and antibiotic usage. Multi-centre national involvement would allow benchmarking of data and provide
opportunity for effects of targeted intervention to be demonstrated.
Systematic Review of Anti-infective-treated Central Venous Catheters for Patients Receiving TPN or Chemotherapy
To review the evidence for use of anti-infective-treated ventral venous catheters (CVCs) in patients receiving chemotherapy or TPN, Niel-Weise
BS et al (J Hosp Infect 2008: 69:114-23) retrieved publications up to October 2007, excluding studies of ICU and peri-operative patients in
acute care. The impact of using treated CVCs was evaluated using catheter-related bloodstream infection (CRBSI) as the outcome.
A total of 233 potentially relevant studies were identified. Of these, 214 were not relevant after titles and abstracts were reviewed.
After full-text retrieval and screening of references, 9 randomised controlled trials were included in the review. Three of the studies
were of patients requiring TPN, 4 studies were of patients with haemato-oncological malignancy, 1 study was of patients with haematological
malignancy, and 1 study was of outpatients with malignancy.
A single trial compared CVCs treated with antibiotics versus standard CVCs – an incidence density difference (IDD) of -1.03 per 1000 CVC days
was demonstrated (95% CI -1.76 to -0.30, p = 0.005). This was a trial of outpatients with malignancy, with CVC in situ for a mean of 9 weeks.
The remaining 8 trials studied CVCs treated with antiseptics versus standard CVCs – an IDD of -0.72 was demonstrated
(95% CI -2.23 to 0.80, p = 0.35). The mean duration of catheterisation in the control groups for these studies ranged from 11 to 20 days.
Subgroup analysis of trials studying patients receiving TPN (n=3) did not show a significant difference in rates of CRBSI when antiseptic-treated
and standard CVCs were compared. Similarly for trials studying patients receiving chemotherapy (n=5), no significant difference in rates for
CRBSI was evident.
Overall, a single trail was identified which demonstrated a reduced occurrence of CRBSI in outpatients treated with chemotherapy using
antibiotic-treated CVCs. A significant benefit of antiseptic coated CVCs was not demonstrated, although the duration of catheterisation
in these studies suggests that enrolled patients may not have been representative of the population receiving CVCs for chemotherapy or
TPN as a whole. Limitations of this review include heterogeneity of trials (including chosen case definition for CRBSI) and potential study
bias (allocation concealment for treatment groups). Therefore, results of this meta-analysis should be interpreted with caution.