**2.1 Geographical and temporal trends in access use**

Pisoni et al., (2002) in the Dialysis Outcomes and Practice Patterns Study (DOPPS) confirmed significant variations in access practice between Europe and the United States. The practice comparison found that AVFs were much more common in prevalent patients in Europe, while AVGs and catheters were more commonly used in the US. Arteriovenous fistulae were used in 80% of the European prevalent population compared with only 24% of US prevalent populations and the use was associated with younger male patients with fewer co-morbidities. Use in incident patients varied from 66% in Europe to only 15% in the US. Conversely, AVGs and catheters were more common in incident patients in the US compared to Europe (2% versus 24% and 60% versus 31% respectively). Dialysis catheters were the first modality of access at commencement of dialysis in the US (Pisoni et al., 2002).

Trends in vascular access have changed over time but have shown a progression towards AVF use. Data from DOPPS I (1996-2001), DOPPS II (2002-2004) and DOPPS III (2005-2007) were compared and found that trends towards increasing AVF use were observed in Australia, New Zealand and the United Kingdom. Australia and New Zealand have traditionally had higher rates of AVF use. Arteriovenous fistulae in these countries, along with Japan and most European countries (excluding the United Kingdom, Belgium and Sweden), are used in over 70% of prevalent haemodialysis patients. The use of AVFs had increased significantly in the US in the same time from 24% in DOPPS I to 47% by DOPPS III (Ethier et al., 2008). Most recent data available shows that AVF use in the US is now greater than 57% (www.fistulafirst.org). In all the countries studied, AVG use remained stable or declined. The US showed the greatest decline in prevalent patient use, falling from 58% in DOPPS I to 29% in DOPPS III (Ethier et al., 2008).

Despite efforts to improve outcomes for ESRD patients, dialysis catheters remain a predominant form of vascular access well into the 21st century. Dialysis catheters were observed in greater than 20% of prevalent patients in the UK, Belgium, Sweden, Canada, and the US. A 2- to 3-fold increase in catheter use was observed in Italy, Germany, France and Spain by DOPPS III (Ethier et al., 2008). Catheter use will never be completely eliminated as they have a significant role in those patients who require urgent dialysis and for whom no other access exists. There is often regional variation in access practice patterns. The reasons for this appear to be multifactorial and include variables such as patient preference, surgical wait times, surgical expertise, as well as physician and nursing factors (Polkinghorne, et al., 2004).

### **2.2 Epidemiological aspects of access type**

A number of studies have now shown an epidemiological association between access type and outcome. Use of venous catheters and AVGs over native AVFs has been shown to carry higher human costs. A number of studies have shown an association between catheter use

Vascular access remains a predominant cause of morbidity in haemodialysis patients. There is significant global variation in the use of the different types of haemodialysis access. There have now been a number of studies examining trends of access use (Ethier et al., 2008; Pisoni et al., 2002) and further, epidemiological associations between access type and outcomes (Dhingra et al., 2001; Ishani et al., 2005; Moist et al., 2008; Pastan et al., 2002; Polkinghorne et

Pisoni et al., (2002) in the Dialysis Outcomes and Practice Patterns Study (DOPPS) confirmed significant variations in access practice between Europe and the United States. The practice comparison found that AVFs were much more common in prevalent patients in Europe, while AVGs and catheters were more commonly used in the US. Arteriovenous fistulae were used in 80% of the European prevalent population compared with only 24% of US prevalent populations and the use was associated with younger male patients with fewer co-morbidities. Use in incident patients varied from 66% in Europe to only 15% in the US. Conversely, AVGs and catheters were more common in incident patients in the US compared to Europe (2% versus 24% and 60% versus 31% respectively). Dialysis catheters were the first modality of access at commencement of dialysis in the US (Pisoni et al., 2002). Trends in vascular access have changed over time but have shown a progression towards AVF use. Data from DOPPS I (1996-2001), DOPPS II (2002-2004) and DOPPS III (2005-2007) were compared and found that trends towards increasing AVF use were observed in Australia, New Zealand and the United Kingdom. Australia and New Zealand have traditionally had higher rates of AVF use. Arteriovenous fistulae in these countries, along with Japan and most European countries (excluding the United Kingdom, Belgium and Sweden), are used in over 70% of prevalent haemodialysis patients. The use of AVFs had increased significantly in the US in the same time from 24% in DOPPS I to 47% by DOPPS III (Ethier et al., 2008). Most recent data available shows that AVF use in the US is now greater than 57% (www.fistulafirst.org). In all the countries studied, AVG use remained stable or declined. The US showed the greatest decline in prevalent patient use, falling from 58% in

Despite efforts to improve outcomes for ESRD patients, dialysis catheters remain a predominant form of vascular access well into the 21st century. Dialysis catheters were observed in greater than 20% of prevalent patients in the UK, Belgium, Sweden, Canada, and the US. A 2- to 3-fold increase in catheter use was observed in Italy, Germany, France and Spain by DOPPS III (Ethier et al., 2008). Catheter use will never be completely eliminated as they have a significant role in those patients who require urgent dialysis and for whom no other access exists. There is often regional variation in access practice patterns. The reasons for this appear to be multifactorial and include variables such as patient preference, surgical wait times, surgical expertise, as well as physician and nursing factors

A number of studies have now shown an epidemiological association between access type and outcome. Use of venous catheters and AVGs over native AVFs has been shown to carry higher human costs. A number of studies have shown an association between catheter use

**2. Vascular access** 

al., 2004; Xue et al., 2003).

**2.1 Geographical and temporal trends in access use** 

DOPPS I to 29% in DOPPS III (Ethier et al., 2008).

**2.2 Epidemiological aspects of access type** 

(Polkinghorne, et al., 2004).

and increased mortality from both infective and non-infective causes (Dhingra et al., 2001; Pastan et al., 2002; Polkinghorne et al., 2004). One study analysed a random sample of patients from the U.S. Renal Data System Dialysis Morbidity and Mortality Study (USRDS) Wave 1. Both diabetics and non-diabetics with catheters demonstrated similar trends in survival. The best overall survival was observed with AVF over AVG. The poorest survival was seen in patients with catheters (Dhingra et al., 2001). This increased mortality observed with AVGs and catheters has been replicated in a number of studies (Pastan et al., 2002; Polkinghorne et al., 2004; Xue et al., 2003). One of the largest investigations of access-related mortality included over 60,000 patients from the United States (Xue et al., 2003). The use of catheters at iniation of dialysis was associated with the greatest mortality risk (catheter hazard ratio [HR] 1.70, 95% CI 1.59-1.81; AVGs HR 1.16, 95% CI 1.08-1.24; AVF reference). . In this study, greater than 50% of patients commenced dialysis with a catheter.

A *prospective* study of almost 1000 patients in France looked at the risk factors for development of bacteraemia in chronic haemodialysis patients (Hoen et al., 1998). This again confirmed that the greatest risk factor for bacteraemia was use of a dialysis catheter, with an incidence of 0.93 episodes of bacteraemia per 100 patient months. Multivariate analysis confirmed vascular access as a major risk factor for bacteraemia. Catheter use for haemodialysis carried a relative risk of bacteraemia of greater than 7 times that of an AVF (relative risk [RR] 7.6, 95% CI 3.7-15.6). Arteriovenous grafts carried only a marginally higher relative risk compared to AVFs.

Access-related bacteraemia has also been shown to be an important factor in the subsequent development of cardiovascular-related morbidity and mortality. Where cause-specific mortality was assessed, increases in both infectious deaths (Dhingra et al., 2001; Ishani et al., 2005; Pastan et al., 2002; Polkinghorne et al., 2004) and cardiac deaths were also observed (Dhingra et al., 2001; Ishani et al., 2005). Interestingly, in one study, non-diabetics using catheters at the inception of dialysis had a worse survival rate than those patients using permanent vascular access, with the difference being detectable after only 2 months of observation. The overall relative risk of infection-related death was approximately 2-fold higher in patients with central venous catheters over those with AVFs and was more marked in diabetics than non-diabetics. The risk of death from cardiac causes was approximately 1.5-fold higher in those with dialysis catheters (Dhingra et al., 2001). A prospective cohort study of incident dialysis patients in the U.S scrutinised the association between access modality and bacteraemia, and also the association between bacteraemia and cardiovascular events (Ishani et al., 2005). Cox regression analysis (n=2358) demonstrated that initial dialysis access was the main antecedent of septicaemia or bacteraemia. Long term dialysis catheters, temporary dialysis catheters and AVGs displayed HRs of 1.95, 1.76 and 1.05, respectively. The presence of bacteraemia or septicaemia was associated with heightened risks of subsequent cardiovascular morbidity and mortality. In those without defined coronary artery disease, a bacteraemic episode conferred a greater risk of death or acute cardiovascular event than those with preexisting cardiovascular disease.

A study undertaken in Australia examined incident haemodialysis patients between 1999 and 2002, and made further attempts to statistically adjust for the non-random nature of access selection. This study found that those patients starting dialysis with a dialysis catheter or AVG had a greater risk of dying in the first 6 months compared to those with AVF; catheters being the most life-limiting of all three. This trend continued with time. Dialysing via an AVF showed a mortality rate of 86 per 1000 person-years; AVGs had a

Hemodialysis Access Infections, Epidemiology, Pathogenesis and Prevention 91

overt infection, microbial colonisation of catheters may engender a chronic inflammatory state, which in turn increases the risk of erythropoietin-resistant anaemia, malnutrition

In those with AVF infection*, S. aureus* and *Staphylococcus epidermis* are most commonly responsible. Infection accounts for approximately one fifth of accesses being lost (Bhat et al., 1980). Microorganisms gain entry to the bloodstream during cannulation. In addition, the presence of pseudoaneuryms, peri-access haematomas, non-functioning clotted fistulae and manipulation of AVFs during non-dialysis interventions increase the risk of infection

Recent evidence suggests that cannulation technique may have an important effect on the rates of bacteraemia related to AVFs. Over the last 30 years, the buttonhole cannulation technique has become increasingly popular. This technique involves repetitive cannulation of a small number of puncture sites, with the aim of creating a tunnel track into which the needles can be easily inserted. There are a number of benefits associated with using buttonhole cannulation for haemodialysis, particularly in the home environment. These benefits include easier and quicker needle insertion, less painful cannulation with the elimination of anaesthetic, reduction in "bad sticks", and reduction in hematoma formation (Doss et al., 2008; Hartig and Smyth, 2009). The alternative cannulation method, referred to as the rope ladder technique, involves needle puncture along the length of the fistula and is more inclined to give rise to small dilatations over the length of the fistula. However, several studies have suggested that the buttonhole technique is associated with an increased risk of access-related infection compared with the rope ladder method. Birchenough et al. (2008) established a positive correlation between use of the buttonhole cannulation technique and an increased risk for infection in adult patients on haemodialysis. Nesrallah et al. (2010) observed a significantly increased risk of *S. aureus* bacteremia infection with potentially fatal metastatic complications in patients receiving home nocturnal haemodialysis with buttonhole cannulation. They recommended advising prospective patients of the infection risks, and, in the absence of more rigorous studies, giving consideration to topical Mupirocin prophylaxis. Other studies have similarly reported increased access-related infection rates in association with buttonhole

A subsequent retrospective observational cohort study in our unit involving 63 alternate nightly nocturnal haemodialysis and 172 conventional haemodialysis patients reported a statistically significant and clinically important increase in septic dialysis access events when nocturnal haemodialysis and buttonhole cannulation were used simultaneously (incidence rate ratio 3.0, 95% CI 1.04-8.66, p=0.04) (Van Eps et al., 2010). It is theorised that chronic bacterial colonisation of the buttonhole site may be the precursor to systemic infection, and that fibrosis surrounding the site may not provide as efficient a barrier as seen in those employing the rope ladder technique. This study highlights that increased infection control steps may be crucial in the setting of nocturnal haemodialysis and buttonhole technique.

The method of handling the dialysis catheter is crucial. Stringent aseptic techniques must be employed, with KDOQI 2006 recommending washing the access site with antibacterial scrub and water followed by cleansing of the skin with 2% chlorhexidine/alcohol or 70% alcohol.

and cardiovascular disease (Barraclough et al., 2009).

cannulation (Ludlow, 2010; Silva et al., 2010; Van Loon et al., 2010).

**4. Preventative measures against access-related infection** 

**3.2 Arterio-venous fistula bacteraemia** 

(Barraclough et al., 2009).

**4.1 Aseptic technique** 

mortality rate of 146 per 1000 person-years and catheters had the highest mortality rate of 261 per 1000 person-years. Catheter use conferred 1.5- to 3-fold increased risks in both infectious and all-cause mortality. A similar trend in both increased infectious and allcause mortality with AVGs was also observed but not significant on analysis (Polkinghorne et al., 2004).

Apart from catheter-related infectious mortality, proposed alternative mechanisms for the increased death rates in patients with catheters have included reduced dialysis doses delivered by central catheters and a higher prevalence of co-morbid conditions in patients who dialyse via catheters. However, the latter was not confirmed after controlling for vascular disease and congestive cardiac failure (Pastan et al., 2002). No patterns of catheter use associated with increasing age or existing co-morbidities were ascertained from the more recent DOPPS III analysis. The usage of dialysis catheters in younger (18-70 year old) non-diabetics increased 2-fold in the US and up to 3-fold in some European countries (France, Germany, Italy and Spain) (Ethier, et al., 2008).
