Chapter 37. Nosocomial and Health Care-Associated Pneumonia

Whether in the emergency department or in the intensive care unit (ICU), pneumonia is a disease with which all treating clinicians must be familiar. In order to accurately recognize and appropriately manage pneumonia, one must first understand its varying classifications, affiliated pathogens, available diagnostic methods, treatment options, and methods of prevention.

Pneumonia classically has been dichotomized as either community-acquired pneumonia (CAP) or nosocomial pneumonia (NP), depending on where the patient becomes infected. NP must occur at least 48 hours after hospital admission and not be incubating at the time of admission.1 NP developing in an individual after requiring at least 48 hours of mechanical ventilation is termed ventilator-associated pneumonia (VAP). NP can also be categorized according to the timing of its onset following hospital admission. Early onset NP occurs within the first 4 days of hospitalization and can be due to antibiotic-susceptible bacteria found in the community (e.g., Pneumococcus, H. influenzae, Moraxella).2 Late-onset NP occurs after at least 5 hospital days, is generally caused by multidrug-resistant (MDR) organisms (e.g., P. aeruginosa, Acinetobacter spp.), and has a higher associated mortality.3 These definitions are clinically relevant because each is associated with a distinct set of typical infecting organisms, management strategies, and outcomes.

However, as the lines between typical outpatient and inpatient settings blur (e.g., due to increased utilization of outpatient dialysis, surgical, and rehabilitation centers), these standard definitions inadequately account for the evolving landscape of causative bacteria resulting in pneumonia. The type of pneumonia infecting individuals with exposure to health care settings is now termed health care–associated pneumonia (HCAP). Although this type of pneumonia typically presents in individuals technically living in the community, the bacterial flora and clinical outcomes affiliated with HCAP are generally similar to those of NP. Therefore, assessing a patient's risk factors for HCAP is essential for its prompt identification and subsequent antibiotic selection (Table 37-1).

Pneumonia exacts a great toll on the physical and financial health of both individuals and society. Compared with the estimated 10% mortality with CAP, one review of over 4,500 hospitalized patients demonstrated mortalities of 19.8% with HCAP and 18.8% with NP.4 NP, and in particular VAP, has an expectedly high mortality due to the invasive organisms infecting a very ill patient population. Some hypothesize that the similarly elevated mortality rate with HCAP may be due to the late recognition, and therefore inadequate antibiotic regimen for HCAP by clinicians not familiar with its risk factors.4

Occurring in 9–27% of all intubated patients,5 VAP accounts for approximately 90% of cases of NP6 and is the most lethal form of all hospital-acquired infections.7 VAP mortality typically ranges from 20% to 70%8,9 but commonly is >70% when caused by invasive MDR organisms.8 Although an accurate assessment of its lethality is difficult due to the numerous variables among studies on the topics of VAP (e.g., heterogeneity of patient populations, duration of mechanical ventilation prior to VAP onset, diagnostic methods used, infecting pathogens, time until appropriate antibiotic administration), its estimated in-hospital attributable mortality is 30%.8 Of note, these multiple confounders plague many of the attempted comparisons made among studies on the topic.

On average, each episode of NP increases the patient's hospital length of stay (LOS) by at least a week,10 and each VAP significantly increases the time on mechanical ventilation, ICU LOS, and hospital LOS. Additionally, VAP increases the cost of hospitalization by at least $40,000.11

To best develop strategies to diagnose, treat, and prevent NP, it is essential to first understand its underlying pathophysiology. HCAP and non-VAP NP generally originate from initial nasopharyngeal or oropharyngeal colonization of bacterial pathogens via either airborne passage or direct inoculation from an affected source.12 Potential sources include contact with medical staff or aspiration of gastrointestinal (GI) flora. Antibiotics that are often administered during a typical hospitalization eliminate indigenous GI bacteria, thus promoting the colonization of resistant strains. Additionally, acid-suppressive medications used for gastric ulcer prophylaxis increase gastric bacterial growth when gastric pH >4.6.13,14 The growing reservoir of increasingly resistant bacteria within the GI tract is particularly dangerous for individuals susceptible to aspiration, such as those with altered mental status, poor cough reflex, swallowing difficulties, or other circumstances diminishing proper airway protection. In fact, more than half of all critically ill patients routinely aspirate.15

In the presence of factors that adversely alter the body's antimicrobial milieu and normal host defenses, colonizing bacteria can subsequently gain access to the lower respiratory system, where they multiply and invade the affected area of lung. The pathogenesis of VAP shares some characteristics with those aforementioned pneumonias, but only VAP involves the presence of an endotracheal tube (ETT) in the upper airway, which directly bypasses normal airway protective mechanisms.

Intubation is the most important risk factor for the development of VAP.16 Although bacteria can access the lungs via inhalation into the respiratory tract from a colonized ventilator apparatus (e.g., humidifiers, filters, suction catheters), hematogenous spread, or direct extension from a parapneumonic process, tracheal colonization most commonly results from the leakage of bacteria-laden secretions pooling in the subglottic space around the cuff of the ETT.17 The insertion of nasal tubes (e.g., nasotracheal, nasogastric) can also displace bacteria colonizing the maxillary sinuses into the subglottic space, serving as a further risk factor for bacterial entry into the airway.18 Certain bacteria form a gelatinous substance called biofilm along the inner and outer sides of the ETT.19 Biofilm, which is found more frequently in individuals with VAP,20 may maintain a steady bacterial load that can serve as a source for recurrent pneumonia.21

Because of the relative ease of obtaining culture specimens from mechanically ventilated patients, more microbiologic data are available for VAP compared with other types of pneumonias. However, due to shared risk factors for multidrug resistance, the causative bacteria for both NP (including VAP) and HCAP are considered to be quite similar.1 In addition to the potential for infection from highly resistant organisms, individuals developing early onset NP (i.e., within 4 days of hospital admission) are also at risk for bacterial infection from community-acquired etiologies. Despite the possibility of infections from these more easily treated bacteria, patients with NP or HCAP should initially receive broad-spectrum antibiotics, making the treatment for community-acquired infections a moot point until culture data return.

Aerobic gram-negative bacilli (e.g., P. aeruginosa, E. coli, K. pneumoniae, Enterobacter spp., Acinetobacter spp.) are the most frequently implicated sources of NP, estimated to cause up to 60% of cases of VAP.5 Gram-positive cocci (e.g., S. aureus, Streptococcus spp.) account for most of the remaining cases of VAP.4 Moreover, up to half of S. aureus–induced cases of NP and HCAP are due to methicillin-resistant S. aureus (MRSA) strains.4,22 Various combinations of bacteria have been noted to coexist in up to 40% of cases of VAP.23 Viruses and fungi are exceedingly rare causes of VAP in immunocompetent patients.1 Although a considerable array and combination of bacteria may cause NP and HCAP, the typical microbial flora of the two types of pneumonia is quite similar and often resistant to multiple antibiotics. Therefore, acquisition of appropriate cultures prior to initiation of antibiotics is important to be able to later tailor the antibiotic regimen to cover only the infecting pathogens. Additionally, every community, hospital, and ICU has its own bacterial resistance patterns that continue to evolve along with antibiotic prescribing practices, so the initial antibiotic regimen must be suited to current local susceptibilities.24 Inadequately covering MDR bacteria in the treatment of NP or HCAP translates into increased mortality, particularly when those pathogens are Pseudomonas, Acinetobacter, or MRSA.25,26

Three objectives exist when evaluating for the possibility of NP: confirm its presence, grade its severity, and identify its microbiologic cause.9 The diagnosis of HCAP, NP, and VAP remains an ongoing contentious topic due to the lack of a diagnostic gold standard and a host of confounding variables among studies assessing different diagnostic techniques. The diagnosis of NP can be made clinically, microbiologically, or using some combination of the two methods.

A clinical diagnosis of NP may be made by the presence of a new or progressive infiltrate on chest radiography plus ≥ 2 clinical criteria that the infiltrate has an infectious origin (Table 37-2).1,27 A more complex 12-point scoring system called the clinical pulmonary infection score (CPIS) was developed in the early 1990s,28 but its inaccuracy (sensitivity 60–77%, specificity 42–75%) makes its use impractical.29,30 The currently recommended, broadly defined clinical diagnostic approach to NP gains sensitivity at the expense of specificity to avoid missing this potentially fatal disease. Unfortunately, this approach still inadequately identifies all cases of NP, as evidenced by one study demonstrating a sensitivity of 69% for the diagnosis of VAP when using two out of the three aforementioned clinical criteria compared with immediate postmortem lung tissue culture and histology.30 Complicating the matter further, however, is that diagnostic inconsistencies evident even with autopsy, histology, and lung aspirate cultures prevent the adoption of a diagnostic gold standard for VAP, and, therefore, NP and HCAP.31,32

Microbiologic diagnostic methods attempt to regain a suitable degree of specificity via the acquisition and culture of causative pathogens. Despite numerous studies attempting to answer this diagnostic question, the most accurate method of obtaining a bacterial specimen still is unclear. Many argue for culturing upper airway secretions from the trachea due the method's greater sensitivity, less invasive nature, and lower associated costs. One trial of 740 patients suspected of having late-onset VAP were randomized to either endotracheal aspiration without quantitative culture or bronchoalveolar lavage (BAL) with quantitative culture.33 Although the authors concluded that there was no significant benefit in clinical outcomes or antibiotic use with either diagnostic method, ≥40% of the screened patients were excluded for having a chronic disease, being immunocompromised, or having carbapenem-resistant bacteria. Due to the exclusion of a patient population that is so commonly encountered in the ICU setting, the applicability of the study's results is lessened. Others advocate culturing specimens from the lower airway due to enhanced specificity, which allows for more precise antibiotic de-escalation or cessation, and subsequently less antibiotic resistance in the future. In a multicenter study of over 400 patients suspected of having VAP, those who were randomized to fiberoptic-directed lower respiratory tract sampling with quantitative culture, as opposed to endotracheal aspiration without quantitative culture, experienced fewer deaths at 14 days (16.2% vs. 25.8%; P = .022) and decreased antibiotic use at 28 days (11.5 vs. 7.5 antibiotic-free days; P < .001).34 This reduction in 2-week mortality in patients receiving the more invasive diagnostic strategy is puzzling because adverse antibiotic reactions are rare and antibiotic stewardship would be expected to translate into longer-term, not short-term, benefits. One can postulate that the benefit of the more invasive, although more specific, test may force the clinician to more aggressively search for, identify, and then appropriately treat an alternative infectious source instead of continuing to treat an infection that would have been misdiagnosed as VAP based on the nonspecific criteria of a radiographic opacity, two clinical criteria, and a positive culture from an endotracheal aspirate.

Even those who believe that lower airway cultures more accurately reflect true infection disagree about what methods to use in obtaining airway secretions and how to most appropriately interpret culture data to distinguish colonization from infection. Techniques to acquire lower airway secretions can be categorized as either bronchoscopic or “blind.” Bronchoscopic modalities (which can also include the use of biopsy) utilize visually directed sampling, whereas blind ones entail the nondirected insertion of a catheter to some predetermined distance, usually until the catheter meets resistance. Lower airway secretions can be sampled with either a BAL, which involves flushing and subsequently aspirating a total of ≥120 mL of sterile saline through the tip of a bronchoscope or catheter that is wedged into a peripheral airway, or a protected specimen brush (PSB), which requires rubbing the airway wall with a brush that is otherwise contained within a protective sheath when it passes through the bronchoscope or catheter.

Quantitative cultures use a predefined logarithmic threshold of bacterial growth to attempt to differentiate colonization from infection. This diagnostic threshold for VAP varies according to the sampling method used: endotracheal aspirate 106 CFU/mL; BAL 104 CFU/mL; and PSB 103 CFU/mL.1,35 Because of inconsistencies among obtained sputum samples inherent to differing patient characteristics, type of bacteria, receipt of antibiotics prior to culture, techniques used to acquire sputum, volume of lavage fluid actually instilled and later aspirated, anatomic location of the sampled fluid, methods of bacterial analysis, and lack of a diagnostic gold standard for NP, the very development of arbitrary quantitative diagnostic thresholds is considered flawed math by some experts.36 Additionally, the accurate comparison of these thresholds among studies, which are plagued by so many confounding variables and technical differences, is exceedingly difficult.

Despite these diagnostic controversies, the diagnosis of NP should include the use of clinical criteria (radiographic evidence of a new or worsening infiltrate plus ≥2 abnormalities of: temperature, WBC count, or sputum) to initiate a diagnostic workup and antimicrobial treatment, followed by bacterial cultures from either the upper or lower airway, with or without quantitative cultures, to complete it.1

The cornerstones of treatment for NP are to obtain appropriate cultures, select initial antibiotics to treat the most likely pathogens based on the patient's risk factors and local susceptibility patterns, give empirical broad-spectrum antibiotics promptly, and de-escalate or stop empirical antibiotics based on the results of adequate culture data and clinical response. Although the same concept holds true for the treatment of HCAP, initially assessing a patient's risk factors for resistant organisms is of paramount importance to properly identify those individuals from the community with HCAP who require broad-spectrum antimicrobial therapy instead of treatment for CAP. Any delay in the administration of appropriate antibiotics results in increased mortality.37–39

Initial broad-spectrum antibiotics should be determined by an individual's risk for MDR pathogens (Table 37-1). In the absence of MDR risk factors and severe pneumonia, monotherapy can be instituted with a second- or third-generation cephalosporin (e.g., ceftriaxone), a β-lactam/β-lactamase inhibitor (e.g., ampicillin/sulbactam, piperacillin/tazobactam), ertapenem, or a fluoroquinolone (e.g., moxifloxacin, levofloxacin).1 Either a fluoroquinolone or the combination of clindamycin and aztreonam can be used in those who are allergic to penicillin. In the presence of MDR risk factors, combination therapy is important, not for synergy, but because it allows for a wider coverage of bacteria that are often resistant to one of the drug classes instituted. Therefore, therapy should include an antipseudomonal cephalosporin (e.g., cefepime, ceftazidime), carbapenem (e.g., imipenem, meropenem, doripenem), or β-lactam/β-lactamase inhibitor (e.g., piperacillin/tazobactam), plus either an antipseudomonal fluoroquinolone (e.g., ciprofloxacin, levofloxacin) or an aminoglycoside (e.g., gentamicin, tobramycin, amikacin), plus coverage for MRSA (e.g., linezolid, vancomycin). Confirmed infection with P. aeruginosa, however, warrants a 5-day combination of a β-lactam and an aminoglycoside, which can then be narrowed per culture data.1 For particularly drug-resistant gram-negative bacteria, polymyxin B and colistimethate (colistin) are treatment options. Additionally, the adjunctive use of inhaled antibiotics (e.g., aminoglycosides, colistin) shows promise, although limited data exist to routinely recommend it.40,41

Patients with VAP responding appropriately to the initial antibiotic regimen should exhibit clinical improvement within the first 6 days of the start of therapy, but some degree of recovery is usually seen well before then.1 If such resolution of disease severity is not seen within the first few days of antibiotics, then alternative pathogens and diagnoses should be aggressively sought. Successful attempts have been made at reducing the duration of therapy from what had historically required 10–14 days of antibiotics. For example, one trial of over 400 patients diagnosed with VAP by BAL and randomized to either an 8- or 15-day course of antibiotics had no significant difference in 28-day mortality but much less antibiotic use and MDR emergence, prompting recommendations for a shorter course of therapy.1,42 Another study supports the use of even a 7-day course of antibiotics, which also showed no difference in hospital mortality or LOS (Figure 37-1).43

Preventing NP hinges on properly addressing its underlying pathophysiology. Strategies used to prevent VAP include: utilize the least invasive method suitable for respiratory support, minimize the duration of mechanical ventilation, avoid unnecessary medications (e.g., stress ulcer prophylaxis, antibiotics), prevent the accumulation of secretions, inhibit bacterial colonization of airway equipment and pooled secretions, prevent the passage of secretions to the lower airway, enhance immune defenses, and enhance health care providers' compliance with these preventive measures.44

Because intubation is the biggest risk factor for the development of NP, a key component of prevention is the avoidance of intubation altogether. Early use of noninvasive positive pressure ventilation (NPPV), for example, bilevel positive airway pressure (BiPAP) or continuous positive airway pressure (CPAP), may obviate the need for intubation in certain circumstances.45 For those who require mechanical ventilation, aggressive ventilator weaning strategies should be instituted with the goal of prompt extubation,46 and for those unable to be extubated within a timely manner, early tracheostomy should be considered.47 Care should be exerted to minimize the chance of accidental extubation, as that significantly increases the development of VAP.48,49 Due to the risk of nasal and sinus carriage of potential pathogens, mechanically ventilated patients should be intubated via the orotracheal route when possible.50 A simple measure such as maintaining patients in the semirecumbent position (i.e., >30° above horizontal) is effective in reducing the chance of aspirating colonized gastric secretions.51,52 Moreover, acid-suppressive medications, particularly proton pump inhibitors, are associated with an approximately 30% increased odds of developing NP, so these should be used only when indicated.53

A trial of almost 6,000 ICU patients randomized to oral bacterial decontamination with topical antibiotics, digestive tract decontamination with topical antibiotics and 4 days of cefotaxime, or standard care demonstrated that both oral and gastric bacterial decolonization significantly reduced 28-day mortality (26.6% vs. 26.9% vs. 27.5%, respectively; P < .05).54 Some experts recommend oral decontamination with antiseptics in lieu of antibiotics in facilities with high antibiotic resistance, which is supported by studies that demonstrate a decreased incidence of VAP with the institution of oral chlorhexidine cleansing.55 To prevent leakage of colonized subglottic secretions around the ETT, endotracheal cuff pressures should remain between 20 and 30 cm H2O56,57 and continuous aspiration of those secretions with special ETTs has proven to be advantageous.58 By inhibiting bacterial colonization, silver-coated ETTs significantly decrease the incidence of VAP when compared with uncoated ETTs (4.8% vs. 7.5%; P = .03).59 When attempting to prevent NP, each of these interventions has modest effects when used alone; however, when implemented as a “ventilator bundle,” more dramatic effects have been demonstrated.60 The proactive application of preventive measures for NP can dramatically reduce mortality and health care expenditures savings for the individuals and society alike.