• 
  

  Focused surveillance for health care–associated infections is the cornerstone of infection prevention activities in the ICU.

  
  
• 
  

  Commonly used invasive devices such as central venous and urinary catheters and endotracheal tubes are significant risk factors for health care–associated infection. Evidence-based ICU policies and procedures and staff education can reduce the risk of device-related infections.

  
  
• 
  

  Antibiotic resistance is an increasing problem, and its containment and prevention require a multifactorial approach, including adequate hand hygiene, surveillance for resistant pathogens, enforced infection control precautions, and prudent use of antibiotics.

  
  
• 
  

  Standard infection control precautions should be applied to all ICU patients. Precautions for contagious or epidemiologically significant pathogens are based on modes of transmission.

Health care–associated infections result in significant morbidity and mortality. Health care–associated infections have been reported to affect approximately 2 million hospitalized patients in the United States annually, at an estimated cost of $57.6 billion in 2000 and approximately 100,000 deaths.^1-3 ICU beds, while only accounting for 5% to 10% of all hospital beds, are responsible for 10% to 25% of health care costs generated.⁴ Patients admitted to the ICU have been shown to be at particular risk for health care–associated infections, with a prevalence as high as 30%.⁵ Given the increasing strain on health care resources in the United States and other countries, and the personal impact that these infections have on patients, the prevention of nosocomial infections in the ICU should be an important goal of any critical care clinician.

A likely explanation to account for the observation that ICU patients are more vulnerable to acquiring a health care–associated infection compared with other hospitalized patients is that critically ill patients frequently require invasive medical devices, such as urinary catheters, central venous and arterial catheters, and endotracheal tubes. Data on a sample of ICUs from the Centers for Disease Control and Prevention (CDC) show that adult ICU patients have central venous catheters in place and receive mechanical ventilation an average of 53% and 42% of their total time spent in the ICU, respectively.⁶ These devices result in infection by compromising the normal skin and mucosal barriers and serving as a nidus for the development of biofilms, which provide a protected environment for bacteria and fungi. In a survey of cases of ICU-acquired primary bacteremia, 47% were catheter related.⁷ While the increased severity of illness of ICU patients makes intuitive sense as a potential risk factor for health care–associated infection, few studies have shown a consistent relationship.⁸ This may be explained, however, by the fact that scoring systems were developed primarily to predict mortality and may not adequately capture markers for health care–associated infection, such as the need for prolonged parenteral nutrition.

Infection control in the ICU arose from hospital-wide infection control programs developed in response to the staphylococcal pandemic of the late 1950s and early 1960s. In 1976, the CDC initiated the Study on the Efficacy of Nosocomial Infection Control (SENIC) to better understand the impact of infection control programs on health care–associated infection rates in a random sample of 338 US hospitals.⁹ Programs that had the greatest impact in reducing health care–associated infection rates had the following components: organized surveillance and active intervention in patient care by infection control staff to reduce the risk of infection, a physician trained in infection control methods, a fixed ratio of infection control specialists to patient beds, and a system for reporting surgical infection rate to surgeons. In hospitals that implemented infection control programs meeting these criteria, the incidence of health care–associated infections decreased hospital wide by 32%, whereas in hospitals with ineffectual programs, infections increased by 18% over a 5-year period. These findings led to regulations requiring that hospitals demonstrate that their infection prevention programs meet the preceding criteria in order to maintain accreditation.

Surveillance for health care–associated infection is the cornerstone of effective infection control activity in the ICU. Surveillance activity serves several key functions, including the early detection of potential outbreaks, the identification of high endemic rates of infection as targets for intervention, and evaluation of the effectiveness of efforts to prevent infection. The process of surveillance itself involves the continuous and systematic collection, tabulation, analysis, and dissemination of information on the occurrence of nosocomial infections within the ICU. It was noted early in the development of infection control programs that feedback of nosocomial infection rates to clinicians, along with active intervention, is a necessary element to a successful program and that collection of surveillance information without this feedback is ineffectual at reducing infection rates.⁹ Health care–associated infection surveillance in the ICU involves the cooperation of both infection control and ICU personnel for both exchange of data and developing effective infection control measures.

Infection control surveillance, particularly when it involves chart review, can be labor intensive. Because of the reality that limited resources are available for infection control surveillance and intervention, a practice known as focused or targeted surveillance is commonly employed. This involves both hospital infection control and ICU personnel making determinations of the particular health care–associated infections to be monitored routinely. The factors involved in making the decision include the degree of morbidity or mortality that results from the infection, the frequency that the infection is known or perceived to occur in the ICU, the proportion of ICU patients at risk of becoming infected, the extent to which effective interventions can be implemented by the ICU team, the perception by both infection control and ICU personnel that a particular infection represents a significant problem for that unit, and finally, state public-reporting mandates.

Infection prevention surveillance data in the ICU typically are reported as the occurrence of a particular infection over a defined time period at risk (eg, cases of ventilator-associated pneumonia per total number of patient days spent on mechanical ventilation per month), also known as an incidence density. For infections that result from a point exposure (eg, the number of tracheostomy site infections per total number of tracheotomies performed in one quarter), a cumulative incidence can be determined. Cumulative incidence is reported less commonly in the ICU owing to the observation that infections in the ICU result primarily from invasive devices that are in place for days to weeks.

In order for the information collected for infection surveillance to be interpretable, a case definition for infection has to be developed. Criteria for the diagnosis of health care–associated infections have been developed by the CDC.¹⁰ These surveillance definitions are widely used for tracking infection incidence within ICUs. It is important to note that these definitions were developed to ensure standardized surveillance methods. Many states in the United States that mandate public reporting of health care–associated infections require these definitions be used.

Several methods exist for performing surveillance in the ICU, including the traditional methods of medical chart review and review of microbiology, radiology, and autopsy reports. More recently, the use of computerized expert systems and medical informatics in health care–associated infection surveillance has reduced the need for manual chart review, improved case ascertainment, and allowed for more resources to be used for intervention and prevention.^11,12

ICU DESIGN AND LAYOUT

While published architectural guidelines require that isolation rooms be included in the layout of critical care units, few data are available to address the impact of ICU design on prevention of nosocomial infection. Mulin and colleagues¹³ demonstrated a lower rate of bronchopulmonary colonization with Acinetobacter baumannii among mechanically ventilated patients in a surgical ICU after the unit was converted from one with a mixture of enclosed isolation rooms and open rooms to all enclosed rooms with hand washing facilities. Another study demonstrated a reduction in the incidence of ventilator-associated pneumonia and urinary tract infections in a pediatric ICU after it was converted from an open ward to separate isolation rooms, without a significant change in patient to staff ratios.¹⁴ Single, private rooms are the current trend in hospital planning and design, and single rooms may be associated with preventing health care–associated bloodstream infection.^15,16 Despite the lack of data, it is prudent to ensure that adequate access to hand hygiene exists for ICU personnel.¹⁵

NURSING STAFFING RATIOS

Much attention has been given to the issue of nurse staffing levels and the impact that this has on patient outcomes and complications, including infection. With increased workloads for registered nurses and the reliance on less trained health care personnel for the delivery of care, there is concern that lapses in infection prevention will occur, resulting in increased infections. In a pediatric cardiac ICU over a 1-year period, a decrease in nurse-to-patient staffing ratios correlated significantly with an increase in nosocomial infections.¹⁷ In a multicenter, retrospective cohort study among 2606 patients admitted to an ICU after abdominal aortic surgery, patients cared for in ICUs that reported nurse-to-patient ratio of 1:3 or greater on either day or night shifts were at greater risk of respiratory complications, including postoperative pneumonia.¹⁸ This relationship was independent of patient age, comorbidity, level of surgical urgency, ICU size, and hospital procedure volume. Another study demonstrated that lower nurse to patient ratio was associated with increase in the risk for late onset ventilator-associated pneumonia.¹⁹

These studies, along with several others, have limitations, including retrospective design, no determination of nursing experience or level of training, and no comment on the role that other types of health care worker staffing levels, such as respiratory therapists, have on health care–associated infection rates. Despite these limitations, a direct association between increased nursing workload and the occurrence of infections among ICU patients appears to exist. The optimal level of both nursing staffing and experience needed to minimize the risk of infection in ICUs remains to be determined but is unlikely to be uniform for every type of unit.

INFECTION PREVENTION POLICIES AND PROCEDURES

Given the complexity of delivering care to critically ill patients, policies and procedure are a necessary part of the organization of any ICU. These policies ensure that personnel perform certain procedures, such as central venous catheter insertion and care, in a consistent manner. Written ICU policies should incorporate evidence-based infection control practices. For policies to be effective, they should be clear, concise, and shared with the staff. Policies that are complex or unrealistic either will be ignored or will result in even wider variation in how care is delivered owing to individual interpretation. Most private and state hospital accreditation programs base their review of ICUs not only on whether ICUs have required policies but also on whether they actually follow them. Therefore, it is important that ICU physician and nursing staff review these policies on a routine basis in consultation with infection prevention practitioners or the hospital infection control committee and revise these policies when needed.

CENTRAL VENOUS CATHETERS/PULMONARY ARTERIAL CATHETERS

Catheter-associated bloodstream infection is one of the most common health care–associated infections seen in ICU patients. Approximately 80,000 of these infections have been estimated to occur annually in ICUs in the United States (excluding insertion-site infection and septic thrombophlebitis).²⁰ These infections are associated with increased ICU length of stay, health care costs, and use of broad-spectrum antibiotics.

Risk factors for catheter-associated bloodstream infections include the anatomic catheter insertion site, type of catheter used, and the patient population. The pathogenesis includes microbial colonization of the subcutaneous catheter tract by skin flora with subsequent colonization of the catheter and biofilm formation, as well as colonization of the catheter hub from microbes introduced during catheter use. While intravascular catheters can result in bloodstream infections by other means, such as the infusion of contaminated fluids, the mechanisms just mentioned are the primary means by which nontunneled central venous catheters in place for less than 2 to 4 weeks cause bloodstream infections.

Numerous infection control practices have been effective in preventing intravascular catheter–related infections (Table 4-1).³² Meticulous hand hygiene before and after handling intravascular catheters, along with maintaining an intact, nonsoiled dressing at the catheter insertion site, is essential to prevent device-related infections. Maximal sterile barrier precautions (ie, sterile gowns, gloves, surgical mask and hat, and a large surgical drape) during insertion reduce the incidence of infection.²¹ In one randomized trial, subclavian vein insertion was associated with a lower incidence of infectious complications and complete vessel thrombosis when compared with femoral insertion.²² However, another randomized controlled trial demonstrated no significant difference in the incidence of catheter-related bloodstream infection for either femoral or jugular hemodialysis catheter insertion sites for nontunneled hemodialysis catheters.²³

Use of a chlorhexidine-based antiseptic for skin preparation has been associated with reducing the incidence of catheter-related bloodstream infection.^24,25 Using an all-inclusive catheter insertion kit or cart is ideal.²⁶ Structured educational programs incorporating the use of maximal sterile precautions have reduced the incidence of catheter-associated bloodstream infections by 27% to 66%.^27,28 In a multicentered interventional study, optimizing combination of preventive measures (ie, maximal barrier precaution, avoidance of femoral vein as a insertion site, hand hygiene before catheter insertion and manipulation, and use of a chlorhexidine antiseptic for skin preparation and removing unnecessary catheter) and using a checklist led to dramatic decline in the incidence of catheter-related bloodstream infection.²⁹ It is also important to empower all health care workers to stop the procedure if sterile technique was not performed.

Several adjunctive approaches to prevent catheter-related bloodstream infection have been investigated. Chlorhexidine-containing sponge dressing is shown to be effective to reduce the incidence of catheter-related bloodstream infection in a randomized controlled trial.³⁰ Antibiotic-coated catheters are also effective.³¹ However, the unexplored issue of emerging resistance associated with the use of these catheters makes their role in an overall infection control strategy unclear. Antibiotic lock therapy may be considered in selective situations.³² Antibiotic lock therapy involves instilling a highly concentrated antimicrobial solution into an unused catheter lumen. A few potential concerns of antibiotic lock therapy include systemic side effects from leakage of lock solution or the emergence of drug-resistant organisms. Moreover, it might be impractical in the ICU settings because of the frequent need for continuous infusion of intravenous fluid or drugs.

ARTERIAL CATHETERS AND PRESSURE TRANSDUCERS

Compared with central venous catheters, the incidence of catheter-associated bloodstream infection attributable to arterial catheters has not been as well studied but is estimated to be roughly 1.5% per device or 2.9 cases per 1000 catheter-days.³³ Pressure transducer systems have been a common source of epidemic outbreaks of health care–associated infection. From 1977 to 1987, these devices were the most common source of epidemic bloodstream infection investigated by the CDC.³⁴ These outbreaks were prolonged (mean 11 months) and involved large numbers of patients (mean 24 patients). In each case, reusable transducers were either improperly disinfected or fitted with improperly sterilized domes. Preventive measures for arterial catheter-transducer systems are similar to those recommended for central venous catheters.³⁵ Recommendations for the proper care and use of both arterial catheters and pressure transducer systems are shown in Table 4-1.

URINARY CATHETERS

Urinary catheter use is the primary cause of urinary tract infections among critically ill patients. There are two clinical entities associated with urinary catheter use: catheter-associated asymptomatic bacteriuria (CA-ASB) and catheter-associated urinary tract infection (CA-UTI). These are differentiated by the presence of clinical symptoms (eg, new onset or worsening fever, rigors, altered mental status, flank pain).³⁶ A national survey of health care–associated infections in 112 medical ICUs found urinary tract infections to be responsible for 31% of all ICU-acquired infections, making urinary tract infections the most common health care–associated infection.³⁷ Most of these infections are asymptomatic, but between 0.4% and 3.6% of patients with a urinary tract infection develop a secondary bloodstream infection.^38,39 The most effective way to reduce CA-UTI and CA-ASB is restriction of urinary catheter use, removal of unnecessary urinary catheters, and hand hygiene before and after manipulating urinary catheters. In addition, potentially modifiable risk factors for infections with urinary catheters include avoiding the use of open urinary drainage systems and breaks in closed drainage systems, using condom urinary catheter for male patients and avoiding retrograde flow from collection bags into the bladder (Table 4-2).³⁶

RESPIRATORY THERAPY EQUIPMENT AND NASOGASTRIC TUBES

Respiratory failure requiring mechanical ventilation is one of the most common indications for ICU admission. Nasogastric tubes are used often for both gastric decompression and to permit feeding of ICU patients. Both mechanical ventilation and nasogastric intubation bypass the normal mucosal defenses of the upper and lower respiratory tract, which leaves patients at risk for health care–associated sinusitis and pneumonia.

Among ICU patients, the vast majority of health care–associated pneumonias are ventilator-associated pneumonias. Ventilator-associated pneumonia is most likely the result of aspiration of contaminated oropharyngeal and gastric secretions and contaminated condensate in the ventilator circuit. Risk factors include the supine position, sedation or impaired consciousness, and reduced gastric acidity.⁴⁰ Contaminated respiratory equipment has been implicated as the source of outbreaks of nosocomial pneumonia. Devices that generate aerosols, such as nebulizer reservoirs used for humidification⁴¹ and multidose medication nebulizers,⁴² have been associated with outbreaks caused by hydrophilic bacteria. Bronchoscopes have also been a source of health care–associated pneumonia in the ICU, usually as a result of an incomplete or compromised disinfection between procedures.⁴³

Guidelines for the prevention of ventilator-associated pneumonia focus on proper hand hygiene when handling respiratory equipment, nursing ventilated patients in a semirecumbent position (30° to 45°), and maintaining a closed ventilator circuit (Table 4-3).⁴⁴ The use of noninvasive ventilation in selected ICU patients results in a lower risk of pneumonia when compared with endotracheal mechanical ventilation. In a randomized, clinical trial, patients ventilated noninvasively had a significantly lower rate of both pneumonia (3% vs 25%) and sinusitis (0% vs 6%).⁴⁵ Subglottic suctioning of oropharyngeal secretions using a specially designed endotracheal tube has been shown to reduce the incidence of ventilator-associated pneumonia and is another method to reduce the risk of nosocomial pneumonia among ventilated patients.⁴⁶ The use of silver-coated endotracheal tubes has been shown to reduce the incidence of ventilator-associated pneumonia and mortality.^47,48

The presence of a foreign body in the nasopharynx, such as a nasogastric feeding tube, predisposes to upper airway infections, particularly sinusitis. Health care–associated sinusitis is often difficult to diagnose. First, the classic signs and symptoms of sinusitis (ie, sinus tenderness, pain, and fever) often are masked in the intubated and sedated patient. Also, sinus aspiration to determine if a sinus fluid collection is infected is performed only rarely. In a prospective study, paranasal sinus computed tomographic (CT) scans were obtained on all ICU patients with purulent nasal discharge and fever not attributable to another source, followed by sinus aspiration and culture of any fluid observed on CT scan.⁴⁹ Using this method, sinusitis was identified in 7.7% of patients. Risk factors for nosocomial sinusitis include nasotracheal intubation, feeding via nasogastric tube, and impaired mental status. These infections frequently are polymicrobial, with Pseudomonas aeruginosa and Staphylococcus aureus being the most common organisms identified.^49,50 Prevention involves avoiding nasotracheal intubation, placing feeding tubes through the mouth rather than through the nose, and minimizing sedative use.

INTRACRANIAL PRESSURE MONITORING DEVICES

Several infectious complications can result from the use of intracranial pressure monitoring devices, including scalp exit-site and tunnel infections, osteomyelitis of the calvarium, meningitis, and ventriculitis. The rate of infectious complications has been reported at 7.4% to 14.1% per procedure.^51,52 Intracerebral or intraventricular hemorrhage, cerebrospinal fluid leaks, open head injuries, monitors in place for greater than 5 days, breaks in the pressure transducer system, and use of intraventricular versus intraparenchymal monitors have been associated with increased risk of infection. Coagulase-negative staphylococci are the most common cause of intracranial pressure monitor–related infections, but gram-negative bacilli, such as Acinetobacter baumannii, Klebsiella pneumoniae, and Proteus mirabilis, have been reported in up to 50% of cases.⁵³ The role of prophylactic antibiotic therapy is unclear. Several cohort studies showed no impact on the incidence of infection among patients who received antibiotics during or after insertion of the devices,^52,54 but no randomized, controlled trials of sufficient sample size exist to address the question. Routine preventive measures including maximal barrier precaution during insertion, routine dressing changes, and use of antimicrobial impregnated intraventricular catheter may reduce catheter-related ventriculitis.⁵⁵

Interventions designed to reduce health care–associated infections typically focus on device-related infection, are evidence based, and often employ physician and nursing education or introduce a change in the process of care. An intervention consisting of a self-study module on risk factors for catheter-associated infections, a pre- and posttest assessment of knowledge, posters and handouts on the infection control practices related to central venous catheters, and didactic teaching were given to the nursing staff of a surgical ICU.²⁸ The authors reported a 66% reduction in the incidence of ICU-acquired bloodstream infections in the 18-month period after the intervention compared with the 18 months preintervention (p <0.001), without a significant change in the patient population. This study demonstrates that focused intervention in ICUs can reduce health care–associated infections, possibly through changes in ICU practice and staff behavior.