Dear Editor,

Central venous catheter infections are an important cause of morbidity and mortality in patients on haemodialysis. The KDOQI describes an infection frequency of between 3.8 and 6.6 episodes/1,000 days of catheter for non-tunnelled catheters and between 1.6 to 5.5/1,000 days of catheter for tunnelled catheters.1 The frequency of bacteraemia varies in various studies between 1.6 and 7.7/1,000 days of catheter for non-tunnelled catheters and between 0.2 and 0.5/1,000 days of catheter for tunnelled catheters.2-6

Our goal was to review the catheter infections in our Haemodialysis Unit. Patients and methods: we examined the catheter infections over a period of six years (1 January 2001 to 31 December 2006). We have implanted 168 catheters, 90 of which were nontunnelled and 78 tunnelled, in 70 patients. Mean age at the time of catheter implantation was 70 ± 14 years old and the average time on haemodialysis 10 months (1 day-17 years). Approximately 25.7% of the patients suffered from kidney disease of unknown aetiology and 22.8% had diabetic nephropathy. Cephazolin was administered before the implantation of tunnelled catheters in all the cases.

Results: The tunnelled catheters were left in place for a median time of 4 (0-45) months and the non-tunnelled catheters were left for 1 (1-6) months. Diabetic patients did not experience more infections (32.4 vs. 34.2%; p = NS) or bacteraemia (2.6 vs. 3.8%; p = NS) than the rest of the patients. The incidence of infections was 2.33/1,000 days of catheter for nontunnelled catheters and 3.10/1,000 days of catheter for tunnelled catheters (p = NS). The most frequent type of infection was that of the exit site both in tunnelled catheters (44/57 [77.2%]) and non-tunnelled catheters (7/9 (77.8%]); p = NS. The incidence of bacteraemia of the non-tunnelled catheters was not greater than that of the tunnelled catheters (0.78/1.000 versus 0.22/1,000 days of catheter; p = 0.08). Twenty-one tunnelled catheters (26.9%) were implanted over a guidewire in non-tunnelled catheters but these did not suffer more infections (8/29 versus 13/49, p = NS).

Cephazolin (55.3%) was the empiricallyusedantibiotic in the majority of the infections.

Staphylococcus was the predominant type, identified in blood cultures in 100% of the cases and in 79.2% of the exit site swabs. The prevalence of Methicillinresistant species was 60%.

After the microbiological results, the initial antibiotic was changed in 22.7% of the cases (N = 15). In almost half of them (46.7%) the antibiotic was switched to Vancomycin.

All patients with sepsis (N = 9) were hospitalized. One died of septic shock and the rest recovered fully.

Conclusions: The non-tunnelled catheters were used for one month without any negative impact on the number of infections. In our opinion, they are safe as a temporary access site for patients that are waiting for the construction or maturing of a definitive access.

Due to the high prevalence of Methicillinresistant species in our centre led to the subsequent inclusion of Vancomycin in our protocol for the management of catheter infections.