Sepsis related to intravenous catheters is the commonest cause of bloodstream infections in Australia. The risk of infection is highest with percutaneous central venous catheters, somewhat lower with tunnelled or subcutaneous catheters, and lowest with peripheral intravenous catheters. The best prevention is removal of intravenous lines when they are no longer necessary. Optimal insertion techniques and line maintenance are also important. Once infection occurs, the line should generally be removed. Antibiotic therapy is directed against suspected micro-organisms (usually staphylococci) and modified with the results of cultures. If septic shock occurs general supportive measures including intravenous fluids, inotropic drugs and observation in an intensive care unit will also be necessary.



Intravenous catheters are indispensable in modern medicine and are no longer restricted to hospital inpatients. There is a growing number of patients on 'home' intravenous therapy, predominantly for total parenteral nutrition or cancer chemotherapy. However, these devices are increasingly associated with sepsis and are now the commonest cause of all bloodstream infections. These infections cause significant morbidity and mortality.


Rates of intravascular catheter-associated bloodstream infections

In Australia, there are at least 3500 cases of intravenous catheter-associated bloodstream infections annually. These are associated with a case fatality rate of 24%, and the mortality rate directly attributable to intravenous catheter sepsis is 12%. This equates to 1.5 bloodstream infections per 1000 admissions.1

Percutaneous central venous catheters are associated with 23 bloodstream infections per 1000 catheters. In contrast, catheters in peripheral veins are associated with only 0.36 bloodstream infections per 1000 catheters.2 Peripheral vein catheters remain in situ an average of 1.5 days, while central venous catheters remain in situ about four times longer (an average of 5.5 days).The daily infection risk with central venous catheters is about 20 times that of peripheral catheters. Tunnelled or surgically implanted catheters (Hickmans, Portacath) and peripherally inserted central venous catheters appear to havea quarter the daily risk of percutaneous central venous catheters, but they still pose a much higher risk than peripheral catheters.

Why there is such a disproportionate infection rate of central venous cathetersis unclear. It may reflect the poorer health of patients requiring this type of therapy as well as the longer duration of intravenous access in this group. The infusion of total parenteral nutrition, the use of triple lumen (versus single lumen) catheters, and some catheter insertion sites (jugular and femoral sites in particular) are independent risk factors. However, all these factors only partially account for the marked differences in daily infection risk rates.

Of concern, is the observation that a large number of intravenous catheters(including central venous catheters) in hospitalised patients are not in active use for prolonged periods of time but remain in situ 'just in case'. Also, some catheters are being used for interventions that are not necessary (for example, total parenteral nutrition when nasogastric feeding is possible).



There are several modes of colonisation with pathogens.3 At the skin entry site, the outer surface of the catheter can become colonised with organisms originating from the skin. These bacteria can then migrate proximally along the catheter's surface to reach the bloodstream. Alternatively, the catheter's inner surface may become colonised by introduction of organisms through the catheter hub (e.g. from the hands of hospital staff). Rarely, micro-organisms may be introduced by contaminated infusate (especially total parenteral nutrition).The point at which colonisation changes to invasive infection is unclear, but it is thought to be related to the number of organisms present on the catheter, and is time dependent (infection is rare within the first 48 hours of catheter placement).

The role of biofilms (collections of bacteria adherent to the catheter surface and organised within an extensive glycocalyx) is important. Although micro-organisms in biofilms are visible on microscopy, they are often unculturable, and are protected from the effects of antibiotics. If biofilms are present, cure is usually only possible by removing the catheter.

Table 1

Major pathogens with approximate frequencies2

Major pathogens with approximate frequencies


Skin-associated micro-organisms are the predominant isolated pathogens (see Table 1). Coagulase-negative staphylococci (e.g. Staphylococcus epidermidis) are the commonest, possibly because they appear to have the best adherence to inert surfaces. Staphylococcus aureus infections are second in frequency, with the infection risk being highest in patients with neutrophil defects orvenous thrombophlebitis. Enteric mucosal micro-organisms such as enterococci,Enterobacteriaceae, pseudomonas, and candida species may colonise the catheter either by colonising the skin, by colonising the infusate (especially total parenteral nutrition) or by haematogenous seeding from mucosal breaches.



In the majority of bloodstream infections associated with central venous catheters, there will be little or no evidence of sepsis at the insertion site (in contrast to infections associated with peripheral vein catheters).

The diagnosis of catheter-associated bloodstream infection requires a positive culture of blood from a peripheral vein and clear evidence implicating the catheter as the source. The culture of 15 or more colonies of a pathogen from a catheter tip is diagnostic of catheter-associated bloodstream infections. Unfortunately, this method only has a positive predictive value of 16-31% because most catheter tip cultures are negative.4

Another approach to diagnosis (which conserves the catheter) is simultaneous culture of blood drawn peripherally and blood drawn from the catheter. As the density of organisms is greatest in the catheter specimen (if it is the source of sepsis), the catheter blood culture will usually become positive at least two hours earlier than the peripheral blood culture (using the Bactec system).This technique has been reported to have a sensitivity and specificity of greater than 90% and a positive predictive value of approximately 80%.



Catheter removal is usually essential in all cases of catheter-associated bloodstream infections, with the exception being some cases associated with Hickmans or Portacath catheters. Even with these, catheter removal is still essential if Staphylococcus aureus or candidal septicaemia occurs and strongly recommended if Gram negative bacilli (due to likelihood of treatment failure)are isolated from blood cultures.

If low virulence organisms such as coagulase-negative staphylococci are isolated, removal of the line itself may be sufficient to resolve the infection, but usually the patient is also treated with one week of intravenous antibiotics. If a Hickmans or Portacath is involved and is not removed, the patient is treated with two weeks of intravenous antibiotics. This may control the infection in80% of cases, however, if the bacteraemia or fever persist despite appropriate antimicrobial therapy, the central venous catheter must be removed.

If bloodstream infection is suspected and the catheter is replaced, the new central venous catheter should not be passed over a guide-wire at the same vene puncture site. If it is, the new catheter will almost certainly be contaminated with the same organism (see Table 2 for further prevention issues).

While awaiting blood culture results, empiric therapy to cover staphylococci and Gram negative bacilli (i.e. vancomycin, or flucloxacillin in combination with an aminoglycoside) is the best initial treatment. The regimen may be modified once the pathogen is identified.

If Staphylococcus aureus is isolated, treat with antibiotics (e.g. flucloxacillin if sensitive) for a minimum of 14 days after catheter removal (4-6 weeks therapy if persistent fevers or a suspected distant focus of infection). If candida is isolated, treatment is generally with a triazole (e.g. fluconazole) for at least 14 days after the last positive blood culture. The fungal isolate should be fully identified, as species other than Candida albicans are often resistant to triazoles.

If the patient remains febrile after removal of the device, three sets of blood cultures should be obtained. Endocarditis or septic thrombophlebitis should be suspected if blood cultures remain positive for more than 48 hours after the device has been removed.


Prevention of catheter-related infection

Prevention of catheter-related infection


Catheter-related sepsis is a common complication of modern medical therapy. Reduction of this complication may be achieved by minimising intravenous access. If there is no absolute need for intravenous access, remove the intravenous line. Use peripheral access rather than central venous catheters wherever possible. When central venous catheter access is necessary, use peripherally inserted venous catheters or tunnelled/implanted lines if possible. If bloodstream infections occur, removal of the intravenous line is essential, with only a few exceptions (Hickmans- or Portacath-associated bloodstream infections with low virulence organisms such as coagulase-negative staphylococci).


Further Reading

Raad I. Intravascular-catheter-related infections. Lancet 1998;351:893-8.


Self-test questions

The following statements are either true or false.

1. Removal of the catheter is necessary in patients with catheter-related bloodstream infections caused by Staphylococcus aureus.
2. Central venous catheters inserted at the subclavian site have a higher riskof infection than those inserted into the femoral vein.

Answers to self-help questions

1. True

2. False


Robert Horvath and Peter Collignon

Infectious Diseases and Microbiology Unit, The Canberra Hospital, Canberra