Nosocomial infection

Published on 27/02/2015 by admin

Last modified 22/04/2025

Print this page

This article have been viewed 1753 times

Chapter 62 Nosocomial infection

Nosocomial or hospital-acquired infections are a major problem in hospitals, affecting up to 9% of inpatients at any one time. Intensive care units (ICUs) represent 2–10% of hospital beds, but are responsible for 25% of all nosocomial blood stream and pulmonary infections. In the European Prevalence of Infection in Intensive Care (EPIC) snapshot of prevalence the infection rate in ICU was 20.62%.¹ Nosocomial infection is, at least in theory, a preventable cause of morbidity and mortality (Table 62.1).

Table 62.1 Principles of diagnosis of nosocomial infections

• Diagnosis of infection usually requires the combination of both clinical findings and the results of diagnostic tests

• Clinical diagnosis of infection from direct observation at surgery, endoscopy or other diagnostic procedure is an acceptable criterion for an infection

• It must be hospital-acquired. There must be no evidence that the infection was present or incubating at the time of hospital admission. Infection acquired in hospital, but only evident after hospital discharge, also fulfils the criteria

• Usually no specific time during or after hospitalisation is given to determine whether an infection is nosocomial or community-acquired. Each infection is examined for evidence that links it to hospitalisation (this is a matter of controversy)

EPIDEMIOLOGY

The prevalence of nosocomial infection is reported as being between 3 and 12% in most institutions but varies considerably between different sites within each institution.² The vulnerability of the patient population, the nature of interventions and cross-infection are but three of many factors. This is seen clearly if one compares the range between ophthalmology and critical care – 0–23%.³

The site of infection varies with location so that, whereas the urinary tract and the chest are common throughout the hospital, within the ICU surgical wound infection, pneumonia and blood stream infection are far more common (8–12%).

The impact of nosocomial infection is impressive. Ventilator-associated pneumonia (VAP) is common, has significant morbidity with increased length of stay, associated costs and a twofold increase in mortality.⁴ It has been suggested that blood stream infections, surgical wound infections and nosocomial pneumonia result in 14, 12 and 13 attributable extra hospital days respectively.⁵ Catheter-related blood stream infection (CR-BSI) was also associated with major morbidity although, curiously, not necessarily mortality.⁶ The mortality rates directly due to these infections are hard to separate from the mortality attributed to the presenting severity of illness, which in its own right may have predisposed to infection. What is clear is that nosocomial infection is associated with increased mortality, and huge financial and resource costs.²

THE MECHANISMS INVOLVED IN NOSOCOMIAL INFECTION

A range of factors come together to enable nosocomial infection to occur. Some may be risk factors in their own right whereas other may simply represent an identifier of a sicker and therefore more vulnerable population (Table 62.2).

Table 62.2 Risk factors for nosocomial infection

Patient

Severity of illness

Underlying diseases

Nutritional state

Immunosuppression

Open wounds

Invasive devices

Multiple procedures

Prolonged stay

Ventilation

Multiple or prolonged antibiotics

Blood transfusion

Environment

Changes in procedures or protocols

Multiple changes in staff; new staff

Poor aseptic practice – poor hand-washing

Patient-to-patient: busy, crowded unit, staff shortages

The organism

Resistance

Resilience in terms of survival

Formation of slime or ability to adhere

Pathogenicity

Prevalence

HOST

The vulnerability of a potential host will be determined by several factors:

• the acute illness or problem that required ICU admission

• the chronic health status and underlying disease. Conditions such as diabetes predispose to infective problems

• the status of their immunocompetence, which may involve medications, such as steroids

• the integrity of natural defences, such as the skin and mucous membranes. This may be disrupted by burns, surgery or by breakdown of tissues such as pressure areas. The presence of invasive devices (e.g. endotracheal tube, intravenous devices) provides a conduit past the natural defences

• antibiotics given to the patient which may encourage emergence of particular strains or resistant organisms

• blood transfusion has been suggested as a risk factor ⁷

ENVIRONMENT

Local environmental pressures play their part. The combination of antibiotics, in particular multiple antibiotics, and cross-infection predisposes a vulnerable population to pseudomembranous colitis from Clostridium difficile toxin.⁸ Epidemiological patterns, such as the prevalence of Enterococcus faecalis as a common pathogen in the surgical population, may be linked to widespread cephalosporin usage. Much of the multiresistance problem probably originates from antibiotic pressures.⁹ Cross-infection is the biggest single problem in intensive care and transmission is by various means, but still the most common is by hands.¹⁰

ORGANISM

The host usually lives in synergistic or symbiotic tranquillity with a huge range of organisms (Table 62.3). Antibiotics suppress many normal organisms and allow the emergence and overgrowth of a usually insignificant organism or resistant organism of the same type. For example, an intrinsic organism such as Candida will flourish in the presence of broad-spectrum antibiotics and this overgrowth may result in symptomatic or even invasive candidiasis. Cephalosporin use may encourage the intrinsically resistant but quiescent enterococci to emerge as a dominant and problematic organism.

Table 62.3 Common commensals that may cause infection in a vulnerable host

Site	Common commensal organisms
Skin	Staphylococcus epidermidis, streptococci, Corynebacterium (diphtheroids), Candida
Throat	Streptococcus viridans, diphtheroids
Mouth	Streptococcus viridans, Moraxella catarrhalis, Actinomyces, spirochaetes
Respiratory tract	Streptococcus viridans, Moraxella, diphtheroids, micrococci
Vagina	Lactobacilli, diphtheroids, streptococci, yeast
Intestines	Bacteroides, anaerobic streptococci, Clostridium perfringens, Escherichia coli, Klebsiella, Proteus, enterococci

Extrinsic organisms may be introduced from the environment, from other patients, from staff or from surfaces. These may be organisms which are thriving in that environment because of local pressures (e.g. antibiotics), or from poor hygiene. Examples include Acinetobacter and, of course, meticillin-resistant Staphylococcus aureus (MRSA). On admission, patients will be carrying a range of organisms that have the potential to cause problems, but during their stay they are likely to acquire a new ecology from their surroundings. In a hospital that ecology may be quite hostile, with multiresistance being common.

The individual characteristics of the organism are important. These include their resilience in the local environment, the ease of transmission and the individual pathogenicity. This clearly interacts with the vulnerability of the host, as some usually innocuous organisms, such as Candida or Serratia marcescens, will only cause problems in vulnerable hosts whereas others, such as some strains of Staphylococcus aureus, Acinetobacter or Clostridium difficile, may be intrinsically more virulent.

THE ORGANISMS

A vast range of organisms can cause nosocomial infection (Table 62.4). It must be emphasised that each hospital and each ICU will have its own local ecology and knowing this ecology is important. Regional, national and international surveys give indications of general trends but this does not supplant local knowledge.

Table 62.4 Organisms responsible for the majority of nosocomial infections

Methicillin-resistant Staphylococcus aureus (MRSA)

Coagulase-negative Staphylococcus (CNS)

Enterococcus spp. (E. faecalis, E. faecium)

Pseudomonas aeruginosa

Acinetobacter baumanii

Stenotrophomonas maltophilia

Enterobacter spp.

Klebsiella spp.

Escherichia coli

Serratia marcescens

Proteus spp.

Candida spp. (C. albicans, C. glabrata, C. krusei)

Other organisms may be a problem in the severely immunocompromised, such as those with acquired immunodeficiency syndrome (AIDS: see Chapter 60)

Nosocomial infection is dynamic in that it is influenced by many environmental factors, the type of patient, type of surgery or illness, the antibiotic usage profile and many other variables. This dynamism is illustrated by the Gram-positive infections of the 1950s and 1960s, giving way to the Gram-negative infections of the 1970s and, while the multiresistant Gram-positive organisms are a major anxiety currently, they are already being superseded by superresistant organisms, such as Stenotrophomonas and Acinetobacter. The ease of transmission or development of resistance is a key factor in the current explosion in multiresistance.

The combination of sick patients and widespread use of potent antibiotics selects out problematic organisms and, as this epitomises intensive care practice, it is in ICU where multiresistance is common.

The organisms causing nosocomial infection may be endogenous or exogenous. Illness and antibiotics may both encourage the emergence and overgrowth of endogenous organisms that were normally suppressed and hence ecological change takes place in the skin, nasopharynx and gut. Alternatively the same factors influence colonisation with exogenous organisms from the environment. Cross-infection plays a significant role, as does the size of the local reservoir of exogenous organisms. The colonising organisms are well placed to invade or be introduced by invasive procedures, by devices or simply through areas of injury. Infection follows.

MULTIRESISTANT ORGANISMS

Many of the organisms that cause nosocomial infection are characterised by multiresistance. There are several mechanisms involved in resistance and in its spread. Enzymes, such as the β-lactamases, render a large array of antibiotics useless. Class 1 β-lactamase is effective against some β-lactam-containing antibiotics but extended-spectrum β-lactamases (ESBL), which incorporate enzymes, such as TEM-24, will produce cross-resistance to multiple classes of antibiotics, including fluoroquinolones and aminoglycosides. Resistance may be produced by a combination of mechanisms, such as in Pseudomonas aeruginosa, where the resistance is due to a combination of derepression of protein efflux systems, cephalosporinases and derepression of AmpC enzyme.

Resistance is acquired in a variety of ways. Mutation of any gene occurs at a rate of one cell in 10⁷ and, if this cell is then presented with antibiotics that it can survive, it will become a dominant cell, reproducing at a rate of 10⁹ overnight. An example of this might be the development of AmpC β-lactamase in Enterobacter, when patients are treated with third-generation cephalosporins. A similar example is the loss of porin OprD in P. aeruginosa in the presence of imipenem.

The big problem, particularly with ESBL, is that they can spread by plasmid transmission which is very rapid. The enzyme production is encoded chromosomally within an organism, and can then be transferred between bacteria by plasmids. Transposons transfer genes between plasmids. Examples of this include:

1 SHV-1 plasmid β-lactamase from Klebsiella

2 plasmid AmpC β-lactamases from Citrobacter

3 aminoglycoside-modifying enzymes from aminoglycoside-producing streptomycetes

The phenomenon of induction is also seen. This is the process whereby the presence of an antibiotic appears to ‘induce’ or speed up the production of the relevant enzyme so that the organism rapidly becomes resistant.

Staphylococcal resistance to meticillin occurs due to an altered penicillin-binding protein, which has low affinity for all β-lactam agents. It is linked to a MecA gene. This gene does not develop readily and spread of meticillin resistance is by vector transmission, not de novo production of resistance ⁷ (Table 62.5).

Table 62.5 The influence of extended-spectrum β-lactamases (ESBL) on resistance in Klebsiella

Antibiotics	ESBL-negative (% resistant)	ESBL-positive (% resistant)
Gentamicin	8	76
Amikacin	3	52
Ciprofloxacin	3	31
All the above	0	5

(Reproduced from Livermore DM, Yuan M. Antibiotic resistance and production of extended-spectrum beta-lactamases amongst Klebsiella spp. from intensive care units in Europe. J Antimicrob Chemother 1996; 38: 409–24.)

SOME COMMON ORGANISMS

See Table 62.4 for organisms responsible for the majority of nosocomial infections.

COLONISATION AND INFECTION

Many organisms both colonise and infect, but differentiating between colonisation and infection is difficult. In essence the presence of the organism may not indicate infection and, in complex patients with multiple colonisations and infection, determining which the culprit is may be difficult or even impossible. As a consequence, treatment is often based on probability or the weight of evidence, rather than certainty.

MRSA colonises very readily and causes infection in a significant proportion of cases but is often easily identified and treated. VRE is still relatively uncommon. It does colonise, although the incidence of overt infection appears low, at present, but is difficult to treat when it occurs. Acinetobacter colonises and infects very readily, may take days to identify and is difficult to treat.²⁰

SITES OF INFECTION

The main sites of nosocomial infection are the chest, wounds and intravenous lines. Urinary tract infection, often quoted as common, seems to be seen rarely in the critically ill.

NOSOCOMIAL PNEUMONIA

Nosocomial pneumonia is a common problem in the critically ill, particularly in ventilated patients, with an incidence of 15–30%.²¹^–²³

Aetiology

There are several possible mechanisms, including aspiration from the nasopharynx, local spread or haematogenous spread of infection. Some 45% of healthy adults aspirate in their sleep. In the sick the nasopharynx colonises rapidly with a wide range of organisms, usually Gram-negatives, and aspiration is encouraged by the unconscious state, the presence of a nasogastric tube or by endotracheal intubation. Pneumonia will develop in up to 25% of colonised patients, compared with a 3% incidence in non-colonised patients. It may also be related to the colonisation of the upper gastrointestinal tract, which may in turn colonise the nasopharynx.

Factors that are likely to encourage colonisation of the pharyngeal areas with organisms include antibiotics and nasogastric tubes. Also implicated are alterations in the host defences in the pharynx, changes in the local pH, the amount of surface mucin and impairment of other local immune defence mechanisms.

Patient factors predisposing to nosocomial pneumonia include acute severity of illness; chronic illness, especially chronic lung disease; diabetes; immunosuppression; advanced age; recent surgery to thorax or abdomen; intubation; and bronchoscopy. Environmental factors include broad-spectrum and prolonged use of antibiotics; potential pathogens in the vicinity; bacterial properties such as ability to adhere to surfaces; cross-infection; 24-hour ventilator tubing changes; and foreign bodies such as nasogastric tubes. Intubated and ventilated patients have a higher rate of nosocomial infection than patients receiving non-invasive ventilation. However, this may be due to different patient populations with different underlying problems and background morbidity.²⁴ The role of neutralisation by H₂ antagonists and proton pump inhibitors is still debated.²⁵^–²⁷

Diagnosis

The general criteria for diagnosis are general signs of infection; clinical signs of a chest infection; purulent sputum; radiological evidence, such as new pulmonary infiltrates; and positive cultures, from sputum or blood.^21,^23,^28,²⁹ These produce a non-specific diagnosis. Expectorated sputum is difficult to assess, but should contain < 25 polymorphs and > 10 squamous epithelial cells per low-power field. In ICU it is more usual to provide samples from the endotracheal tube and then to use quantitative techniques. This can achieved with protected brush specimens (PBS) bronchoalveolar lavage (BAL) and protected BAL. These are difficult and time-consuming and it appears that non-bronchoscopic techniques such as blind tracheal aspirates through the endotracheal tube are both practical and reasonably effective from a clinical – if not a research – viewpoint^23,³⁰ (Table 62.6).

Table 62.6 Principles for the diagnosis of nosocomial pneumonia

Crackles on auscultation or dullness to percussion on physical examinaion of the chest and any of the following:

• New onset of purulent sputum

• Organism isolated from blood cultures

• Isolation of pathogen from specimen obtained by transtracheal aspirate, bronchial brushing or biopsy

• Chest radiography examination shows no or progressive infiltration, consolidation, cavitation or pleural effusion

• Isolation of virus or detection of virus antigen in respiratory secretions

• Diagnostic single antibody titre (IgM) or fourfold increase in paired serum samples (IgG) for pathogen

• Histopathological evidence of pneumonia

IgM, immunoglobulin M.

The Organisms

Aerobic Gram-negatives predominate, including Pseudomonas aeruginosa, Enterobacter spp., Klebsiella, Escherichia coli, Serratia marcescens and Proteus; between them, these account for about 70% of infections. Staphylococci, in particular MRSA, account for a small but significant number of nosocomial pneumonias. Early-onset VAP, between 48 hours and 5 days, may be due to community pathogens, such as methicillin-sensitive Staphylococcus aureus, Streptococcus pneumoniae and Haemophilus infiuenzae, as well as Gram-negative enteric bacilli (GNEB). This has been described as early endogenous infection. In contrast, later onset, after 5–7 days, involves more resistant species such as MRSA, Pseudomonas aeruginosa, Acinetobacter baumannii and Stenotrophomonas maltophilia.^31,³² This has been called secondary endogenous and exogenous infection.

Prevention

Prevention is by methods that reduce aspiration, cross-infection and contamination of respiratory devices.²⁹ Several recommended methods are:

• hand-washing

• limiting antibiotic use

• changing ventilator circuits at longer intervals – 1 week

• orotracheal, rather than nasotracheal intubation

• removing nasogastric tube

• semirecumbent position – this reduces aspiration

• avoiding muscle relaxants

• avoiding reintubation

• using non-invasive ventilation, where possible

Other more controversial methods include:

• antibiotic prophylaxis after intubation

• continuous suctioning of subglottic secretions

• early enteral nutrition: benefits of the colonisation of the gastrointestinal tract might be offset by problems with nasogastric tubes and nasopharyngeal colonisation

• the use of heat moisture exchangers

• closed suctioning: with some limited evidence it reduces VAP

Treatment

The use of inappropriate or inadequate antibiotics leads to an increased mortality, so current trends are towards initial broad-spectrum antibiotics followed by reassessment and de-escalation.^33,³⁴ The principle is that an empirical broad-spectrum approach will cover the relevant organisms but once the specific organism is identified the antibiotics can be rationalised to focus on that organism; this is clearly harder to apply in practice than in theory. As the spectrum of resistance grows it is highly likely that it will be increasingly difficult to apply this broad-spectrum empiric treatment, especially with late-onset VAP. Surveillance of current colonisation, awareness of an individual ICU’s ecology and targeted treatment will replace empiric regimens. At present, combination therapy is recommended, but there is little evidence to support it over monotherapy.^23,^33,³⁵^–³⁷

Another area of interest is duration of treatment with antibiotics. Many regimens are traditional rather then evidence-based and may result in unnecessarily prolonged treatment. There is a trend to reducing courses. For VAP an 8-day course may be adequate.³⁰ There has not yet been a move towards using a clinical or bacterial response to treatment as an indicator for the duration of treatment, although it may be on the horizon.

The attributable mortality with nosocomial pneumonia is difficult to determine because patients requiring ventilation often have a high intrinsic mortality. Rates of 30% have been quoted: pneumonia accounts for 60% of deaths from nosocomial infections. In addition, nosocomial pneumonia increases length of stay by 12–23 days.^4,^38,³⁹

WOUND INFECTION

Wound infections are common and represent 20% of all nosocomial infection. The organisms involved may be those introduced at the time or from later contamination.

RISK FACTORS

• The procedure itself and the level of contamination prior to or during surgery. In clean elective surgery the incidence should be less than 10%, but where contamination may occur (e.g. bowel surgery) it may be as high as 15%. In contaminated procedures, the rate rises to around 20%, and where infection already exists it may be 40% or higher.

• The surgeon’s technical ability ⁴⁰

• Host factors: see Table 62.2

• The use of antibiotic prophylaxis

• Duration of stay in ICU

THE ORGANISMS

The organisms are frequently determined by the type of surgery and procedure. Skin commensals are common, such as staphylococci or streptococci, and are usually early infections. These are characterised by fever, pain and redness around the site. With streptococcal infection this may be quite confluent. Gram-negative organisms may also cause infection. Drainage is important.

PREVENTION

Clean theatres, good operating technique and thorough asepsis, both in the theatre and postoperatively, are important.

ANTIBIOTIC PROPHYLAXIS

This is a difficult and contentious area. The unnecessary use of broad-spectrum antibiotics should be avoided, but there are some areas of surgery where prophylaxis has a proven place. If there is no risk of infection with a clean procedure there is no place for antibiotics. If contamination is either seen or likely to occur, the use of antibiotics is to provide cover for the spillage. There may be other circumstances where the patient is particularly vulnerable to bacteraemia, such as with valve disease or where the consequences of infection would be disastrous. Effective prophylaxis requires an antibiotic:

• which covers the most likely organisms

• to be given prior to contamination

• at peak dosage at the time of contamination

A single dose should suffice, although a second dose is recommended if the procedure extends beyond 3 hours. Prolonged administration:

• increases the chance of antibiotic resistance

• unnecessarily exposes the patient to adverse effects from the drugs

• encourages colonisation with resistant organisms

• ensures that if a late infection occurs it will be with an organism resistant to the prophylaxis

In many circumstances the evidence supporting prophylaxis is minimal, and frequently the antibiotics used are inappropriate to the perceived risk. Local factors, including the local ecology, will determine specific requirements. With grafts, mesh or prostheses the morbidity from infection is so great as to justify using prophylaxis even if the evidence is marginal (Table 62.7).

Table 62.7 Surgical prophylaxis

Type of surgery
Abdominal wall	Insertion of mesh. Staphylococcal or streptococcal cover. In the groin, Gram-negative cover may be needed
Cardiac	Protection of valves and grafts. Staphylococci may be resistant
Vascular	Protection of grafts. Staphylococci are a major consideration. If the groin is involved, it may require Gram-negative cover
Orthopaedic	Prostheses. Staphylococci are an increasing problem
Biliary upper GI	Usually Gram-negative and anaerobic cover, with an awareness of resistant enterococci
Colorectal	Protection against faecal flora. Cephalosporins and metronidazole have been popular but predispose to developing enterococci such as Enterococcus faecalis or E. faecium, which may be multiresistant
Gynaecological	Conventionally broad-spectrum involving cephalosporins or, currently, augmentin. Metronidazole is also favoured
Urological	Protection from instrumentation of the urinary tract, Gram-negative cover. Awareness of any existing infection

GI, gastrointestinal.

LINE SEPSIS

The use of intravascular devices in hospital practice is ubiquitous. There is both a significant morbidity, resulting in prolongation of hospital stay, and an attributable mortality associated with their use due to infection.⁴¹^–⁴³

PERIPHERAL LINES

The use of peripheral lines is commonly associated with phlebitis, which may be chemical but is frequently a precursor of infection. Usually peripheral lines should only be in situ for less than 72 hours.

ARTERIAL LINES

Curiously, arterial lines seem to become colonised and infected less frequently than other peripheral devices. Nevertheless they do become colonised and should be regularly changed. Complications include infected radial aneurysms or thrombosis.

PULMONARY ARTERY CATHETERS

The introducers are a potential source of infection, as is the part of the catheter in the introducer. They are frequently left in situ when they should have been removed earlier.

CENTRAL VENOUS CATHETERS

Colonisation of catheters is common and it is likely that this is a precursor of infection. Colonisation rates are in the order of 5–40%, determined in part by the risk factors involved (see below). Infection rates are approximately 10% of the colonised catheters.⁴⁴

DEFINITIONS

1 Colonised catheter: growth of > 15 colony-forming units (semiquantitative) or 10³ (quantitative) from a proximal or distal catheter segment in the absence of accompanying clinical symptoms and signs.

2 CR-BSI: isolation of the same organism from the catheter segment (see above) as from a peripheral blood culture in a patient with signs of infection and in the absence of another source. In the absence of laboratory corroboration of infection then effervescence of infection after removal of the catheter may be taken as circumstantial evidence.

Extrinsic mechanisms associated with developing catheter sepsis include infection from the skin and insertion site; contamination from the hub and then internally spread; and contamination of drugs or fluids administered through the catheter. Bacteraemia seeding to catheter is an intrinsic mechanism.

The actual mechanics by which the organisms colonise the catheter is important. Some organisms, such as CNS, produce a polysaccharide film of ‘slime’ that develops on the catheter while the host’s proteins such as fibronectin may also provide a matrix in which the organism can adhere, and which may provide a protective barrier against both white cells and antibiotics. Polyvinyl chloride or polythene is more prone to this film developing than some other materials such as silicone (Table 62.8).

Table 62.8 Risk factors for catheter infection

Host risk factors

Site: subclavian is a lower risk than internal jugular and femoral

Catheter material: antibacterial catheters may reduce infection, antiseptic catheters reduce colonisation

Number of lumens: multilumen catheters increase the infection risk³⁰

Number of administrations through the lines

Dressing type: frequency of changes

Skin preparation

Experience of technique of personnel

Occurrence of bacteraemia

Tunnelling: often used for long-term access but the data are contentious³¹

THE ORGANISMS

The organisms involved in catheter sepsis are many and varied. There is a 25% incidence of CNS, which are increasingly recognised as pathogens. Staphylococcus aureus and MRSA are prevalent. There may be seeding, resulting in vertebral osteomyelitis and a variably reported incidence of endocarditis. Enterococci are increasingly seen. Occasionally fungi are found and may again be associated with seeding to the eyes or heart.

TREATMENT OF CATHETER INFECTION

The most important aspect of treatment is a high index of suspicion that leads to removal of the device if infection is present either locally or systemically.⁴² Although fever and bacteraemia are likely to resolve rapidly after removal of the line, appropriate antibiotics are indicated. The recommended duration of treatment varies: 5–7 days for CNS, 10–14 days for S. aureus, Gram-negative organisms and fungi, and 4–6 weeks if there is evidence of endocarditis, infected thrombus or osteomyelitis, or clinical line sepsis is still present after 3 days.⁴² Lines removed should be cultured. If an infection is present it is best to avoid replacing the central venous catheter if possible for a few days. In situations where it is uncertain if the line is implicated in infection, some advocate replacing a new line over a wire. If the removed catheter is subsequently shown to be infected, then it must be removed.

The situation is slightly different with Broviac or Hickman catheters. Occasionally administration of antibiotics through the catheter will eradicate the infection and save the line. Unfortunately, it is not always successful: the risk of failure should be balanced against the value of the line. Some medical catastrophes have been the consequence of persisting with infected Hickman lines.