126 Prevention and Control of Nosocomial Pneumonia

Preventing pneumonia in the critically ill is a daunting task, and even controlling the incidence is difficult. Despite this, many in the patient safety movement have suggested that nosocomial pneumonia should be a “never” event. While complete prevention of nosocomial pneumonia is unlikely, substantial progress has been made in reducing the incidence.

Pneumonia is the most common nosocomial infection in the intensive care unit (ICU).¹ The frequency of ventilator-associated pneumonia (VAP) varies from 8% to 28%.² A large 1-day point prevalence study of pneumonia demonstrated that nearly 10% of ICU patients were being treated for pneumonia.¹ However, rather than overall rates, the incidence per day of mechanical ventilation is a more legitimate description. The National Nosocomial Infection Surveillance program reports VAPs/1000 ventilator days. However, the risk of VAP also does not remain static throughout the duration of ICU stay. The greatest risk is early in the course of mechanical ventilation, dropping from a daily hazard rate of 3.3% at day 5 to a 1.3% rate at day 15.³ The incidence also varies significantly among different types of ICU patients. Postoperative patients, especially those undergoing cardiothoracic and trauma-related surgery, appear to have the highest rates. Coronary care unit patients appear to have the lowest rates; medical, respiratory, and other surgical patients demonstrate intermediate rates.

The influence of endotracheal intubation is so dominant that ICU-acquired pneumonia is almost synonymous with VAP. Endotracheal intubation increases the rate of nosocomial pneumonia between 3- and 21-fold.² Research on hospital-acquired pneumonia has been dominated by VAP, and very little is known about pneumonia in nonintubated ICU patients. Because the effect of nosocomial pneumonia on morbidity and mortality in nonintubated patients is minor compared with that of VAP, concentration on VAP is appropriate.

A distinction should be made between prevention of all nosocomial pneumonia and prevention of life-threatening nosocomial pneumonia. The latter is almost exclusively VAP. The crude mortality rate for VAP ranges from 24% to 76%, with an estimated attributable mortality of 20% to 30%.^2,⁴ Early-onset VAP (within 5-7 days of intubation) has a minimal effect on mortality if any. The greatest crude and attributable mortality rates are associated with late-onset multidrug resistant (MDR) microorganisms such as Pseudomonas aeruginosa, Acinetobacter spp., and methicillin-resistant Staphylococcus aureus (MRSA). Unfortunately, the most effective and well-documented strategies to prevent pneumonia work predominantly or exclusively in early-onset VAP and therefore have not resulted in a significant improvement in mortality. Conversely, one of the most consistent adverse effects of VAP (including early onset) is a prolonged duration of mechanical ventilation. Because duration of ICU stay is the principal determinant of cost of care, prevention measures may be cost-effective even if they do not result in improved mortality.

Pathogenesis

The key to effective prevention and control strategies is a clear understanding of the underlying pathogenesis of nosocomial pneumonia. The essence of nosocomial pneumonia pathogenesis involves three basic steps:

1 Colonization of the oropharynx with pathogenic microorganisms

2 Aspiration

3 Overwhelming of the lower respiratory tract’s host defense mechanisms

Effective prevention and control measures can be analyzed by their effect on one or more of these steps.

Despite the convenience of this simple analysis, to assume that the pathogenesis of all types of nosocomial pneumonia and VAP is the same would be naive and incorrect. An example is the role of gastric colonization preceding oropharyngeal colonization, the basis for attention to enteral feedings and stress ulcer prophylaxis in VAP prevention. Although possibly an important factor for pneumonia due to Enterobacteriaceae, gastric and enteric colonization has no role in the pathogenesis of S. aureus or P. aeruginosa pneumonia, the two most common causes of VAP. Conversely, daily chlorhexidine baths did not prevent VAP in a trauma population but did significantly decrease VAP from MRSA.⁵ Therefore, prevention strategies should be individualized to the pathogens and mechanisms prevalent in a specific ICU.

Colonization with Pathogenic Microorganisms

The antecedent event to most nosocomial pneumonias is colonization of the oropharynx with pathogenic bacteria. The oropharynx is not sterile normally, but the character of the normal flora is remarkably constant. A variety of factors alter the normal flora, allowing more pathogenic microorganisms to appear and increase in number.

Time of exposure to these selective forces is a critical issue. Early-onset pneumonia, even early-onset VAP, tends to be caused by less pathogenic microorganisms such as streptococci, Hemophilus influenzae, or methicillin-sensitive S. aureus. Most of these selective forces are introduced in the hospital environment itself, rather than specifically in the ICU. Therefore, patients who develop pneumonia during the first few days of ICU admission or mechanical ventilation are at risk for MDR pathogens if the ICU admission was preceded by a 3- to 5-day hospital stay. Many of the same factors also operate in skilled-care nursing home facilities, blurring the distinction between hospital- and community-acquired pneumonia, and have led to a new designation of healthcare-associated pneumonia (HCAP).

Previously, colonization of the oropharynx by gram-negative enteric bacilli, generally from the Enterobacteriaceae family, was the major concern. These microorganisms are part of the normal bowel flora. Oropharyngeal colonization occurred by one of two main routes. The first is reflux of bacteria into the stomach from the duodenum, with subsequent gastroesophageal reflux into the esophagus and oropharynx. Colonization and proliferation in the stomach are critical intermediate steps in this pathway. Therefore, many prevention strategies logically target the stomach. The other route is self-inoculation by the fecal-oral route, through contamination of equipment or the hands of healthcare providers or the patient.

S. aureus is now the most common microorganism causing ICU-acquired pneumonia, with P. aeruginosa the next most common. In addition, Acinetobacter species have become a common cause of VAP in many institutions. None of these three microorganisms has a typical colonization pattern like that of the Enterobacteriaceae. S. aureus is a normal colonizer of the skin and the nasopharynx. Antegrade colonization of the oropharynx from the nose, especially with the use of nasogastric tubes in many critically ill patients, can occur quite easily. Similarly, Acinetobacter is found on moist body surfaces and in the gingival crevices of patients with poor oral hygiene. P. aeruginosa is usually not part of normal bowel flora but is ubiquitous in the environment. One of the unique aspects of Pseudomonas VAP is the appearance of tracheal colonization before oropharyngeal colonization.⁶ Because colonization of the stomach is not an important intermediary step for these pathogens, prevention measures directed at the stomach are not likely to affect pneumonia caused by these microorganisms. Both MRSA and Acinetobacter colonization can be decreased with the use of chlorhexidine whole-body bathing.⁵

Avoidance of Antibiotics

The most important factor that leads to increased colonization of the oropharynx with pathogenic microorganisms is the use of systemic antibiotics, especially broad-spectrum antibiotics.⁷ Antibiotic therapy results in alteration of the oropharyngeal flora and gives pathogens a selection advantage. The broader the antibiotic spectrum, the greater the likelihood normal flora will be affected. At the same time, some pathogens are also eliminated. For this reason, antibiotics function more as amplifying agents rather than as true causes of colonization. The pathogenic microorganisms must still reside in the area normally, such as nasopharyngeal carriage of S. aureus, or be transferred from other sites including the environment to colonize. Thus pneumonia can still occur despite avoidance of antibiotics. However, the causative microorganisms are more likely to be less virulent pathogens or even normal flora, such as α-hemolytic streptococci, and less likely to lead to life-threatening pneumonia.

Diagnostic strategies for fever in the ICU that result in the use of fewer antibiotics have been associated with lower mortality.⁸ Shorter courses and fewer antibiotics for documented infections in critically ill patients have also been associated with a decreased risk of superinfection.⁹^–¹¹ Although avoiding antibiotics may have only a small effect on the risk of developing the first episode of pneumonia, limiting their usage has a major effect on secondary pneumonia and infection-related death in the ICU.

Use of Topical Antibacterial Agents

In contrast to systemic antibiotics, the use of topical antibiotics for the prevention of colonization may be beneficial. In general, strategies rely on controlling pathogenic microorganisms at specific sites, despite the effect on normal flora. Topical agents generally do not have the toxicity of systemic agents, and although the use of topical antibiotics can lead to MDR isolates, the risk may not be as great as with systemic antibiotics.

Selective Digestive Tract Decontamination

By far the most extensively studied and most aggressive form of topical antibiotic strategy to prevent colonization is selective digestive tract decontamination. Although the specific agents used in different studies vary, the major focus is on controlling oropharyngeal colonization by almost sterilizing the bowel. Therefore, the antibiotics used are directed primarily at gram-negative bacilli (usually polymyxin B and an aminoglycoside) and Candida (usually amphotericin B). Most regimens include two components—topical antibiotics in the oropharynx, and nonabsorbable antibiotics via a gastric tube. Some also include an initial short course of systemic antibiotics.

Despite more than 40 randomized controlled trials and several meta-analyses,¹²^–¹³ the benefit of selective digestive tract decontamination remains unclear. However, several patterns have emerged. Selective digestive tract decontamination fairly consistently decreases the incidence of VAP when systemic antibiotics are used for the first 48 to 72 hours.¹⁴ The rationale for the use of systemic antibiotics is to prevent incipient endogenous infections until sterilization of the bowel occurs. However, an equivalent benefit has been found with a short course of prophylactic antibiotics alone.¹⁵

The efficacy of selective digestive tract decontamination in preventing life-threatening late-onset VAP is less clear. Most studies do not demonstrate lower mortality in the treated group, despite lower rates of VAP. Treatment is directed primarily against the Enterobacteriaceae and yeast in the gastrointestinal tract, but because these microorganisms do not cause the majority of cases of VAP in the ICU, its benefit in preventing VAP due to these microorganisms may be diluted by the many cases of pneumonia caused by organisms that are not specifically addressed by the regimen.

The major criticism of selective digestive tract decontamination is the potential for promoting antibiotic resistance. This theoretical risk has not been clearly demonstrated, even in ICUs that have used the regimen for prolonged periods.¹⁴ However, recent data which look at the whole ICU and non-ICU ecosystem suggest this may be an issue.¹⁶ The major determining factor is probably not the selective decontamination, but rather the concomitant systemic antibiotics. If selective digestive tract decontamination truly decreases the incidence of VAP (and possibly other nosocomial infections), the resultant decrease in systemic antibiotic use may cancel out the risk of selecting for resistant isolates.

Because the major benefit of selective digestive tract decontamination appears to be in preventing VAP due to Enterobacteriaceae, this strategy is probably best reserved for patient populations at increased risk for VAP due to these microorganisms. Postsurgical, trauma, and solid organ transplant patients are in this category. In addition, this approach appears to be very effective as part of the management of epidemics of antibiotic-resistant clones.

Topical Oropharyngeal Agents

Controlling colonization of the oropharynx alone has also generated interest. In a randomized controlled trial of open heart surgery patients, use of a chlorhexidine oral rinse lowered the risk of VAP from 9.4% to 2.9%, with the major effect being on gram-negative bacteria.¹⁷^–¹⁸ This primary finding was accompanied by decreases in all nosocomial infections, fewer nonprophylactic antibiotic prescriptions, and a trend toward lower mortality. Subsequent studies have confirmed the benefit of chlorhexidine topical oral treatments on risk of VAP.¹⁹ One advantage of oral decontamination only is no disruption of the normal bowel flora by treating only the primary area of concern. Conversely, chlorhexidine may not be able to prevent infection with MDR pathogens such as Pseudomonas and Acinetobacter.²⁰ Oral decontamination with other agents such as antimicrobial peptides²¹ has not been demonstrated to be of benefit.

Aerosolized Antibiotics

The earliest studied form of topical colonization prevention was aerosolized antibiotics. In the early era of mechanical ventilation, daily aerosolized polymyxin B resulted in a dramatic decrease in the rate of gram-negative VAP.²² Not surprisingly, routine use was soon complicated by the emergence of antibiotic-resistant microorganisms. This issue, combined with a lack of mortality benefit, led to abandonment of this strategy. Recently, aerosolized ceftazidime was not shown to decrease VAP rates in trauma patients, but also did not increase MDR pathogen colonization.²³ A recent variation is to use aerosolized antibiotics for purulent tracheobronchitis, thought to be a precursor to VAP.²⁴

Avoidance of Increased Gastric pH

The normally acidic environment of the gastric lumen is extremely effective in preventing colonization with either swallowed oropharyngeal flora or refluxed enteric flora. Several prevention strategies focus on this aspect of prevention of VAP.

Stress Ulcer Prophylaxis

At one time, gastrointestinal bleeding from stress ulceration was a substantial problem in ventilated patients and a major cause of death. Prophylaxis against stress ulceration was thus considered critical for ventilated patients. However, the incidence of stress mucosal ulceration has decreased markedly as a result of better hemodynamic resuscitation, improved ventilatory strategies, and earlier use of enteral nutrition.

The debate regarding optimal gastrointestinal bleeding prophylaxis has therefore evolved over the last few decades. Initially, antacids were found to be inferior to histamine type 2 blockers (H₂ blockers). In addition to increasing gastric pH, antacids increase gastric volume, which is probably an independent risk factor for VAP. Subsequently sucralfate was hypothesized to be superior to H₂ blockers because it did not affect gastric pH and might have intrinsic antibacterial properties. No clear-cut benefit of sucralfate over H₂ blockers in reducing VAP has been found, while a slight but consistent increase in gastrointestinal bleeding has been documented.²⁵ Proton pump inhibitors are also used frequently despite more limited data.

The major issue is whether stress ulcer prophylaxis is needed at all in most mechanically ventilated patients.²⁵ The few placebo-controlled trials suggest both H₂ blockers and sucralfate may lead to an increased risk of VAP. Several multivariate analyses found proton pump inhibitors to be associated with increased pneumonia rates, including HAP/VAP,²⁶ HCAP, and even community-acquired pneumonia. Ironically, use of gastrointestinal prophylaxis is actually encouraged as part of a ventilator/VAP bundle in many institutions. A subgroup of patients at increased risk for gastrointestinal hemorrhage can be identified and patients without these high risk factors may not need prophylaxis.²⁷

Enteral Nutrition Strategies

Malnutrition is clearly associated with an increased risk of pneumonia and increased mortality in the critically ill.²⁸ In addition to classic effects on cell-mediated immunity, an effect specific to pneumonia is increased binding of gram-negative bacilli, including Pseudomonas, to epithelial cells.⁶

Enteral administration of nutrition is the preferred route for treating and preventing malnutrition in the critically ill, although parenteral nutrition in high risk patients is preferable to no nutrition.²⁸ Meta-analysis has suggested that patients can even be fed soon after gastrointestinal surgery.²⁹^–³⁰ However, continuous enteral nutrition infusions may increase both gastric pH and gastric volume and theoretically increase VAP risk. Several multivariate studies have suggested that this potential risk is real.³¹^–³² A randomized trial found that the risk of VAP was increased with early aggressive feedings compared with low-level enteral nutrition (approximately 20% of goal feeding rate).³³ The lower rate was chosen to avoid atrophy of the microvilli of the enteric mucosa, a potential source of nosocomial infection. The increased risk of VAP was attributed to an increased risk of aspiration, which is also seen in surgical series.³⁰ Despite this, meta-analyses of early versus delayed enteral nutrition suggest a mortality benefit and probable decreased risk of VAP with early feedings.³⁴ A balance between potential risks would be early initiation of enteral feeding but avoidance of aggressive infusions that might cause high gastric residuals and gastric distention.