Managing Hospital- and Health Care-Associated Pneumonia

Approach to Management

The terms hospital-acquired pneumonia (HAP) and health care–associated pneumonia (HCAP) are often used interchangeably, although HCAP encompasses not only patients hospitalized within the past 3 months but also nursing home or extended-care residence, receipt of chronic hemodialysis, or home wound care. The outcome of HAP can be improved with early, appropriate empiric therapy. Several studies have demonstrated that mortality rates for HAP are higher if the initial antimicrobial therapy was inadequate as demonstrated by subsequent invasive culture results. There is also evidence to suggest that if the initial empiric choice is incorrect, subsequent correction to a more appropriate choice does not improve the mortality rate in patients with ventilator-associated pneumonia (VAP). Inappropriate therapy and delayed initiation of appropriate therapy increases the mortality of VAP. It appears that in severely ill patients, the initial empiric therapy is the most important decision a physician can make.

An approach to the management of these patients that has gained wide acceptance includes an assessment of the severity of illness and risk factors that may predispose the patient to specific respiratory pathogens. Categorizing patients in this manner allows identification of likely pathogens and renders the choice of antimicrobials more rational. This approach has been used by the American Thoracic Society (ATS) and Infectious Diseases Society of America (IDSA) guidelines.

Enteric gram-negative bacilli and S aureus are the bacterial pathogens most commonly associated with HAP (Table 7-1). In patients with VAP, several reports have suggested that the infection may be polymicrobial. Bacterial pathogens predominate in this clinical setting but viral and fungal pneumonias also occur occasionally. The diagnosis of viral pneumonia is often missed because there are few unique clinical features to separate it from bacterial pneumonia; the traditional investigations (viral cultures, viral serology, rapid viral antigen tests) are rarely ordered, but multiplex polymerase chain reaction (mPCR) testing may improve detection of viral pathogens. Viral pathogens should be considered during epidemic nosocomial infections, especially if there is a concomitant community outbreak. This scenario was observed during the COVID-19 epidemic.

Another approach that is helpful is to consider the likely pathogens according to the time after hospitalization in which pneumonia developed (Table 7-2). Among patients with early-onset pneumonia (occurring within the first 5 days of hospitalization) relatively simple-to-treat organisms predominate. In contrast, late-onset pneumonia is usually characterized by infection with more resistant organisms.

A core group of bacterial pathogens has been identified. These organisms are common in all groups of patients and need to be covered in all circumstances. Among these core organisms are enteric gram-negative bacilli, such as Enterobacter spp, E coli, Klebsiella spp, Proteus spp and S marcescens, as well as H influenzae and gram-positive organisms, including methicillin-sensitive Staphylococcus aureus (MSSA), methicillin-resistant Staphylococcus aureus (MRSA) and S pneumoniae (Table 7-3). Multiresistant gram-negative organisms, such as P aeruginosa, do not fall into the category of core organisms.

The definition of severe HAP has been adapted from that used for severe community-acquired pneumonia (CAP) (Table 7-4). Any patient residing in the intensive care unit (ICU) or being admitted to the ICU falls into this category. Other factors that have been designated include the presence of:

• Respiratory failure
• Rapid radiographic progression
• Multilobar pneumonia
• Cavitation of a lung infiltrate
• Evidence of severe sepsis with hypotension
• End-organ damage.

Patients should be classified according to the severity of illness, the presence of specific risk factors for multidrug-resistant pathogens, and the time of onset of the illness in relation to the day of hospital admission. This allows patients to be divided into two major groups:

• Patients without risk factors for multidrug-resistant pathogens and with early-onset pneumonia (<5 days) (Table 7-5 and Table 7-6).
• Patients with specific risk factors with onset at any time during hospitalization or with late-onset pneumonia (>5 days) and patients with severe HAP, either of early onset with specific risk factors or of late onset (Table 7-7).

The usual pathogens associated with mild-to-moderate pneumonia occurring at any time during hospitalization in patients without risk factors are the core pathogens identified above. For patients falling into this category, a nonpseudomonal third-generation cephalosporin (ceftriaxone, cefotaxime), a fourth- or fifth generation cephalosporin (cefepime, cefiderocol), or a β-lactam/β-lactamase–inhibitor combination (ampicillin/sulbactam, ticarcillin/clavulanate, or piperacillin/tazobactam) can be used effectively (Table 7-5). In penicillin-allergic patients, a fluoroquinolone or clindamycin/aztreonam combination is a reasonable alternative. In the event of a suspected Pseudomonas infection, combination therapy is always recommended.

Enlarge 

Enlarge 

Enlarge 

Enlarge 

Enlarge 

Therapeutic Options

Despite earlier hypotheses, there is no convincing linkage between presenting symptoms, findings on physical examination, or laboratory test results and specific pathogens. Thus microbiologic techniques have been developed to permit recognition of the causative pathogen. It is preferable to identify the pathogen since such identification makes optimal antimicrobial selection possible. This is particularly relevant in this era of increasing antimicrobial resistance. Knowledge of the pathogen also limits the consequences of antibiotic abuse, including:

• Higher costs
• Increased resistance
• Adverse drug reactions.

Identifying pathogens of potential epidemiologic significance, such as Legionella or inducible β-lactamase–producing Enterobacteriaceae, would have a profound effect on antimicrobial selection and outcome. In comparison to the total costs of hospitalization, the costs associated with diagnostic testing are relatively trivial.

Unfortunately, the results of diagnostic testing have not proved to be sufficiently reliable to make therapeutic decisions. Several studies have indicated that there is reasonable concordance between clinical and microbiologic criteria for pneumonia. Some investigators have argued that antibiotic treatment is indicated in most cases with intermediate likelihood of pneumonia and in all cases with high likelihood, even if low colony counts from invasive testing are found. In general, initial antibiotic therapy should not be delayed or withheld because delay in therapy may increase risk of mortality. Any diagnostic testing must be done rapidly in order to facilitate therapy; testing rapidity increased dramatically in the last decade with increasing availability of mPCR platforms and rapid antigen tests.

The current (2016) ATS-IDSA guideline provides a “weak” (conditional) recommendation that treatment of patients with suspected HAP (non-VAP) commence on the basis of microbiologic studies of noninvasively collected samples rather empirically. If empiric treatment is chosen for HAP/VAP, the guideline recommends that it be informed by the local distribution and antimicrobial susceptibilities of HAP/VAP-associated pathogens. Knowledge of the usual microbial etiologies for HAP makes antimicrobial selection based on local antibiograms a strong approach. Factors that predict unusual pathogens, such as duration of hospitalization or previous use of antimicrobials, should be identified. A proposed management algorithm for HAP is shown in Figure 7-1.

Enlarge 

Monotherapy vs Multidrug Therapy

Single-agent therapy with β-lactams, first-generation cephalosporins, or macrolides is not feasible simply because these agents do not have a sufficiently broad spectrum of activity. In general, clinicians have relied upon combination antibiotic therapy to ensure broad coverage and to prevent the emergence of resistant organisms.

Aminoglycosides demonstrate good activity against the target pathogens but, unfortunately, as agents for HAP, exhibit significant flaws. Aminoglycoside levels in pulmonary secretions are 10-45% of simultaneously measured serum aminoglycoside levels. Since toxic side effects limit the dosing capabilities of the aminoglycosides, levels achievable in respiratory tissues may be inadequate to eradicate most Enterobacteriaceae. Aminoglycosides perform poorly in an acidic environment with an 8-fold reduction in antibacterial activity at pH 6.4 when compared with its activity at pH 7.4. The acidic environment associated with pneumonia, particularly in areas of necrosis or abscess, may inactivate aminoglycosides.

Several randomized controlled trials have demonstrated an association between the use of synergistic antibiotic combinations and improved therapeutic outcomes. However, many broad-spectrum antibiotics can cover the usual pathogens associated with HAP. Initial therapy should cover the enteric gram-negative organisms and
S aureus. Among the antibiotics that are active against these pathogens are:

• Third-generation cephalosporins:
  • Ceftazidime
  • Ceftazidime/avibactam
  • Ceftriaxone
  • Cefotaxime
  • Cefoperazone
• Fourth-generation cephalosporins:
  • Cefepime
  • Cefpirome
• Fifth-generation cephalosporins:
  • Ceftolozane/tazobactam
  • Cefiderocol
• Older fluoroquinolones:
  • Ciprofloxacin
• Newer fluoroquinolones:
  • Levofloxacin
  • Moxifloxacin
• Carbapenems:
  • Imipenem/cilastatin
  • Imipenem/cilastatin/relebactam
  • Meropenem
  • Ertapenem
  • Doripenem
• An ESBL/β-lactamase inhibitor:
  • Ticarcillin/clavulanate
  • Piperacillin/tazobactam
• Glycylcyclines:
  • Tigecycline.

Among these, only ceftazidime, cefepime, cefpirome, ceftolozane, cefiderocol, doripenem, imipenem/cilastatin (with or without relebactam), meropenem, levofloxacin and ciprofloxacin would be considered to be active against P aeruginosa. Clinical efficacy of single-drug empiric therapy for HAP has generally been equivalent to combination regimens. However, reports of efficacy in HAP should not be extrapolated to pneumonia caused by P aeruginosa simply because most of these studies have not included enough patients infected with this organism to definitively demonstrate equivalence. Other problems that have been identified with single-agent therapy include emergence of antibiotic-resistant bacteria while on therapy and, occasionally, development of serious bacterial superinfections.

Given these considerations in selecting antibiotics for initial empiric therapy, it is prudent to use monotherapy only in those without risk factors and with early onset. It must be noted that use of monotherapy in hospitalized patients with pneumonia is rarely feasible. Multidrug therapy is warranted for those with risk factors or late-onset pneumonia primarily because the multidrug regimen improves the probability that the initial regimen will cover the causative pathogen. Once the causative agent has been identified, the regimen can be streamlined.

Problems With Drug-Resistant Organisms

Methicillin-Resistant Staphylococcus Aureus (MRSA)

Staphylococci have been identified as important agents associated with community- and hospital-acquired infections for decades. In 1968, the first large outbreak of MRSA infections in the United States was reported at the Boston City Hospital. Once established in institutions, MRSA is difficult to eradicate. It currently accounts for up to 50% of all nosocomial S aureus isolates. Infections due to MRSA commonly occur in surgical and critically ill patients. The National Nosocomial Infection Surveillance System identified S aureus as the second most common HAP-causative pathogen after P aeruginosa. Community-acquired cases of MRSA in adult patients have often followed influenza and were associated with increased mortality. The incidence of MRSA has increased so dramatically that it must now be considered as a possibility in any patient with pneumonia in the ICU for >72 hours. The resolution of MRSA VAP, regardless of the appropriateness of initial antibiotic therapy, is associated with longer respiratory support.

The identified risk factors for MRSA infection include increased length of stay and previous antibiotic therapy. Compared with other etiologies of HAP, patients who are infected with S aureus are more likely to be >25 years of age, in a coma, not users of corticosteroids and to have antecedent trauma. Using a step-forward logistic-regression analysis, only coma is defined as significantly influencing the development of S aureus pneumonia. Other investigators have identified neurosurgical patients, especially those treated for cerebral edema and burn patients are at increased risk for S aureus pneumonia. MRSA-infected patients are more likely to:

• Be >25 years of age
• Have received corticosteroids before developing infection
• Have been ventilated >6 days
• Have received antibiotics
• Have had preceding COPD compared with patients infected with MSSA.

The presence of bacteremia, septic shock and mortality is higher in the MRSA group. Hematogenous spread of S aureus to the lungs has been reported in patients with right-sided endocarditis, in users of illicit IV drugs, and in those with skin and soft tissue infection, burns, pyomyositis and infected IV catheters. Transmission of MRSA is related to transient carriage on the hands of hospital personnel.

In MSSA infections, the semisynthetic penicillins are the most active agents. These include methicillin, oxacillin and nafcillin. Other useful agents include an extended spectrum b-lactamase (ESBL/β)-lactamase inhibitor (ticarcillin/clavulanate, ampicillin/sulbactam, piperacillin/tazobactam) and the cephalosporins. The first- and second-generation cephalosporins (cefazolin, cefuroxime) are more potent than the third-generation cephalosporins against MSSA. The treatment of choice for MRSA pneumonia is vancomycin or linezolid. There appears to be no great advantage of one vs the other. Alternatives to vancomycin are few, but include teicoplanin, trimethoprim/sulfamethoxazole (TMP/SMX) (in limited cases especially community-acquired MRSA), and the new streptogramins. RP59500 (quinupristin/dalfopristin) is composed of two streptogramins, A and B, which act synergistically to inhibit protein synthesis. The efficacy of these new agents is currently being investigated. Tigecycline, a new first-in-class glycylcycline antibiotic with very good in vitro antistaphylococcal activity, is another alternative agent that should be considered.

Multidrug-Resistant Gram-Negative Rods

With the introduction of the β-lactamase-stable cephalosporins (second-generation and later), stable, multiresistant, derepressed mutants of gram-negative bacteria have emerged during therapy. Failure of therapy or relapse after discontinuation of therapy has been reported with these agents. These organisms have been responsible for hospital outbreaks even in patients not receiving the cephalosporins.

These inducible β-lactamases are chromosomally mediated and not related to plasmids. In normal circumstances, these enzymes are repressed, which leads the casual observer to conclude that the organisms are sensitive to the tested antibiotics (usually cephalosporins or other stable β-lactams). However, with exposure to a β-lactam, increased levels of β-lactamases are produced, leading to resistance to multiple b-lactam antibiotics and clinical failures. The nonfastidious gram-negative bacilli that possess inducible β-lactamases include Enterobacter spp (most commonly), Serratia spp, C freundii, Proteus spp, Providencia spp, Morganella spp and P aeruginosa. In this case, simple exposure to a cephalosporin may lead to rapid emergence of resistance. When Enterobacter spp are isolated from clinical samples, it may be prudent to avoid newer cephalosporins regardless of in vitro susceptibility. Resistance has emerged in 14% to 56% of treated patients, while the combination of clinical failure and resistance has been identified in 7% to 30% of patients. Emergence of resistance is seen more often with treatment of respiratory tract and bone and soft tissue infections than with infections of the urinary tract.

The addition of another drug, such as an aminoglycoside or another β-lactam, has not prevented the emergence of resistance. Several studies have suggested that there is good correlation between the use of second- and third-generation cephalosporins and the occurrence of multiresistant gram-negative bacilli. Selective pressure leads to the emergence of resistance, but these organisms have a propensity for secondary spread within an institution. Early detection of these strains, judicious use of certain cephalosporins, and the use of barrier precautions in infected patients are necessary to prevent widespread hospital outbreaks. Use of a different class of antimicrobials should be considered.

Patients With Specific Risk Factors With Onset During Hospitalization

In patients at risk for anaerobic infection (recent abdominal surgery, witnessed aspiration), the addition of clindamycin to the usual third-generation cephalosporin is recommended (Table 7-6). A β-lactam/β-lactamase–inhibitor combination or a new fluoroquinolone with anaerobic activity (moxifloxacin) could be selected as a single agent. While aspiration is the common mechanism leading to the development of pneumonia, gross aspiration leading to anaerobic lung infection is uncommon. There are almost no data for the use of an antibiotic with anaerobic activity for pneumonia in the absence of lung necrosis or abscess formation.

In individuals at increased risk for S aureus infection (those with coma, head trauma, diabetes mellitus, renal failure), the possibility of methicillin resistance drives the consideration of vancomycin in addition to the core antibiotics. Among patients at risk for infection with Legionella spp (eg, those who have received high-dose corticosteroids), a macrolide or a fluoroquinolone with or without rifampin should be considered.

Enlarge 

Patients With Severe HAP (Early Onset With Specific Risk Factors or Late Onset)

Patients with severe HAP with risk factors or patients with late-onset pneumonia should be initially treated with broad-spectrum antibiotics that will cover multiresistant pathogens, such as P aeruginosa or Acinetobacter spp (Table 7-7). A regimen consisting of an aminoglycoside or fluoroquinolone (high-dose levofloxacin or ciprofloxacin) plus one of an antipseudomonal penicillin, β-lactam/β-lactamase inhibitor, cefpirome or cefoperazone, imipenem, doripenem, or meropenem should be selected. Sulbactam/durlobactam is a combination specifically approved for the treatment of HAP/VAP caused by bacteria of the Acinetobacter baumannii-calcoaceticus (ABC) complex.

In a randomized prospective trial of severe pneumonia, ciprofloxacin was compared with imipenem as single-agent therapy. Ciprofloxacin-treated patients had a higher bacteriologic eradication rate and a higher clinical response rate. When P aeruginosa was recovered from initial respiratory tract cultures, failure to achieve bacteriologic eradication and development of resistance during therapy were common in both groups. This study demonstrated that monotherapy for severe pneumonia is safe and effective but that other agents are required if P aeruginosa is suspected. The newer fluoroquinolones do not have the same activity as ciprofloxacin against
P aeruginosa.

These recommendations are for initial empiric management of patients with HAP. Once the pathogen is identified or a good clinical response is obtained, therapy can be simplified. It remains unclear whether routine follow-up microbiologic investigations are required. Certainly, in patients failing to respond to the initial empiric regimen, further investigations and a reexamination of the antibiotic selection are required.

Enlarge