Treatment of hospital-acquired and ventilator-associated pneumonia in adults

INTRODUCTION — Hospital-acquired (or nosocomial) pneumonia (HAP) is an important cause of morbidity and mortality despite improved prevention, antimicrobial therapy, and supportive care [1].

The treatment of non-ventilator-associated HAP (nvHAP) and ventilator-associated pneumonia (VAP) will be reviewed here. The diagnosis, epidemiology, pathogenesis, microbiology, risk factors, and prevention of nvHAP and VAP are discussed separately. (See "Clinical presentation and diagnostic evaluation of ventilator-associated pneumonia" and "Epidemiology, pathogenesis, microbiology, and diagnosis of hospital-acquired and ventilator-associated pneumonia in adults" and "Risk factors and prevention of hospital-acquired and ventilator-associated pneumonia in adults".)

Treatment of respiratory infections in pregnant individuals is discussed separately. (See "Approach to the pregnant patient with a respiratory infection", section on 'Treatment'.)

DEFINITIONS

Pneumonia types — Pneumonia is frequently categorized based on site of acquisition (table 1).

●Hospital-acquired (or nosocomial) pneumonia (HAP) is pneumonia that occurs 48 hours or more after admission to the hospital and did not appear to be incubating at the time of admission.

●Ventilator-associated pneumonia (VAP) is a type of HAP that develops in intubated patients on mechanical ventilation for more than 48 hours. VAP also includes HAP that occurs within 48 hours of extubation.

●Non-ventilator-associated HAP (nvHAP) refers to HAP that develops in hospitalized patients who are not on mechanical ventilation nor underwent extubation within 48 hours before pneumonia developed. nvHAP can be divided into patients that ultimately require mechanical ventilation (vHAP) due to the pneumonia versus those that do not. vHAP is associated with particularly poor clinical outcomes.

The category of health care-associated pneumonia (HCAP) is no longer recognized as a separate category of pneumonia and was purposefully not included in the 2016 and 2019 American Thoracic Society/Infectious Diseases Society of America (ATS/IDSA) HAP [1] and CAP [2] guidelines or the combined 2017 European and Latin American HAP guidelines [3]. Historically, HCAP referred to pneumonia acquired in health care facilities such as nursing homes, hemodialysis centers, outpatient clinics, or within three months following a hospitalization [4]. This category was used to identify patients at risk for infection with multidrug-resistant (MDR) pathogens. However, this categorization may have been overly sensitive and may have led to increased, inappropriate broad antibiotic use. Although patients with recent contact with health care facilities are at increased risk for infection with MDR pathogens, this risk is small for most patients and the overall incidence of MDR pathogens in this population is low [5-11].

We manage patients who would have been previously classified as having HCAP in a similar way to those with community-acquired pneumonia (CAP), deciding whether to include therapy targeting MDR pathogens on a case-by-case basis depending upon each patient's specific risk factors and severity of illness [2]. Specific risk factors for resistance that should be assessed include known colonization with MDR pathogens, recent receipt of antimicrobials, comorbidities, functional status, and severity of illness [12,13]. (See "Treatment of community-acquired pneumonia in adults in the outpatient setting" and "Treatment of community-acquired pneumonia in adults who require hospitalization".)

Antimicrobial resistance — The United States Centers for Disease Control and Prevention (CDC) and the European Centre for Disease Prevention and Control (ECDC) have developed standard terminology for antimicrobial-resistant gram-negative bacilli, which are potential causes of nvHAP and VAP [14]:

●MDR refers to acquired nonsusceptibility to at least one agent in three different antimicrobial classes.

●Extensively drug resistant (XDR) refers to nonsusceptibility to at least one agent in all but two antimicrobial classes.

●Pandrug resistant (PDR) refers to nonsusceptibility to all antimicrobial agents that can be used for treatment.

Awareness of local resistance patterns is critical for decisions regarding empiric therapy for HAP and VAP [15]. All hospitals should regularly create and disseminate a local antibiogram, ideally one that is specific to the different units in the hospital (although small numbers of cases per unit may preclude this) [1].

Risk factors for multidrug resistance are discussed separately. (See 'Identifying risk factors for MDR pathogens' below and "Epidemiology, pathogenesis, microbiology, and diagnosis of hospital-acquired and ventilator-associated pneumonia in adults", section on 'MDR risk factors'.)

EMPIRIC THERAPY

Timing of antibiotics — Once non-ventilator-associated HAP (nvHAP) or ventilator-associated pneumonia (VAP) is suspected clinically, microbiological specimens should be obtained as soon as possible in all patients. In patients with signs of septic shock or rapidly progressive organ dysfunction, antimicrobial therapy should be started as soon as possible [1]. If the diagnosis of nvHAP and VAP is uncertain and the patient is not in sepsis or septic shock, then it appears to be safe and potentially beneficial to gather more data and await culture results before treating [16-19]. (See "Evaluation and management of suspected sepsis and septic shock in adults", section on 'Empiric antibiotic therapy (first hour)'.)

We pair early and aggressive treatment for patients with signs of sepsis or septic shock with early and aggressive de-escalation once the causative pathogen and susceptibilities are known or an alternative diagnosis is established (see 'Tailoring therapy' below). Long delays in treatment and/or failing to give a regimen with activity against the causative pathogens are both associated with higher mortality rates in patients with septic shock due to nvHAP or VAP [20-23]. However, broader regimens and longer treatment courses increase the risks of adverse drug effects, Clostridioides difficile infections, and antimicrobial resistance [24,25]. The difficulty in diagnosis may lead to overtreatment and its attendant risks of superinfection and antibiotic toxicity, but establishing the diagnosis of nvHAP and VAP can be difficult, especially for hospitalized patients in whom clinical, radiologic, and microbiologic findings can be due to numerous etiologies besides pneumonia. (See "Epidemiology, pathogenesis, microbiology, and diagnosis of hospital-acquired and ventilator-associated pneumonia in adults" and "Clinical presentation and diagnostic evaluation of ventilator-associated pneumonia".)

Identifying risk factors for MDR pathogens

Ventilator-associated pneumonia — Risk factors for multidrug-resistant (MDR) pathogens in patients with VAP include the following (table 2):

●Risk factors for MDR pathogens (including Pseudomonas aeruginosa, other gram-negative bacilli, and methicillin-resistant S. aureus [MRSA]):

•Intravenous (IV) antibiotic use within the previous 90 days

•Septic shock at the time of VAP

•Acute respiratory distress syndrome (ARDS) preceding VAP

•≥5 days of hospitalization prior to the occurrence of VAP

•Acute renal replacement therapy prior to VAP onset

●Risk factors specifically for MDR Pseudomonas aeruginosa and other gram-negative bacilli:

•Treatment in an intensive care unit (ICU) in which >10 percent of gram-negative bacilli associated with VAP are resistant to piperacillin-tazobactam and/or cefepime

•Treatment in an ICU in which local antimicrobial susceptibility rates among gram-negative bacilli are not known

•Colonization with and/or prior isolation of MDR Pseudomonas or other gram-negative bacilli on culture from any body site (but especially from the respiratory tract)

Early detection of resistance genes by a multiplex polymerase chain reaction (PCR) panel can also provide early information regarding the presence of MDR pathogens [26].

While presence of structural lung disease (eg, bronchiectasis, cystic fibrosis) in and of itself does not select for MDR pathogens, it can be a proxy for MDR risk since these patients are prone to frequent exposure to antimicrobials because of their increased risk for recurrent pneumonias. Repeated exposure to antimicrobials in turn increases risk for MDR pathogens. Generally, we treat patients with presence of structural lung disease with broad-spectrum antibiotics only if they have additional risk factors for VAP (such as known colonization with MDR pathogens or recent exposure to antibiotics), similar to the general population. It is important to be cautious about avoiding overuse of broad-spectrum antibiotics in this patient population to minimize further selection for resistance. Sometimes this may mean withholding antibiotics in patients who are not very sick or very frail until there is more diagnostic clarity of whether a pneumonia is truly present or not.

●Risk factors specifically for MRSA:

•Treatment in a unit in which >10 to 20 percent of S. aureus isolates associated with VAP are methicillin resistant

•Treatment in a unit in which the prevalence of MRSA is not known

•Colonization with and/or prior isolation of MRSA on culture from any body site (but especially the respiratory tract)

Risk factors for MDR VAP have been addressed in several studies. In a meta-analysis conducted by the 2016 American Thoracic Society/Infectious Diseases Society of America (ATS/IDSA) guidelines that included 15 studies, factors associated with an increased risk of MDR VAP were use of IV antibiotics in the past 90 days (odds ratio [OR] 12.3, 95% CI 6.48-23.35), ≥5 days of hospitalization prior to the occurrence of VAP, septic shock at the time of VAP (OR 2.01, 95% CI 1.12-3.61), ARDS before VAP (OR 3.1, 95% CI 1.88-5.1), and renal replacement therapy prior to VAP (OR 2.5, 95% CI 1.14-5.49) [1]. Coma present at the time of ICU admission was associated with lower risk of MDR VAP (OR 0.21, 95% CI 0.08-0.52).

Note that the prevalence of MRSA has been decreasing worldwide [27-29], which serves as an important reminder to consider the local prevalence when determining whether or not to include coverage for MRSA in an empiric regimen.

Non-ventilator-associated hospital-acquired pneumonia — Risk factors for MDR pathogens and/or mortality in patients with nvHAP include the following (table 3):

●Risk factors for increased mortality:

•Need for ventilatory support for hospital-acquired (or nosocomial) pneumonia (HAP)

•Septic shock

●Risk factor for MDR Pseudomonas, other gram-negative bacilli, and MRSA:

•IV antibiotics within the past 90 days

●Risk factors for MDR Pseudomonas and other gram-negative bacilli:

•Colonization with and/or prior isolation of MDR Pseudomonas or other gram-negative bacilli on culture at any body site (but especially the respiratory tract)

While presence of structural lung disease (eg, bronchiectasis, cystic fibrosis) in and of itself does not select for MDR pathogens, it can be a proxy for MDR risk since these patients are prone to frequent exposure to antimicrobials because of their increased risk for recurrent pneumonias. Repeated exposure to antimicrobials in turn increases risk for MDR pathogens. Generally, we treat patients with presence of structural lung disease with broad-spectrum antibiotics only if they have additional risk factors for VAP (such as known colonization with MDR pathogens or recent exposure to antibiotics), similar to the general population. It is important to be cautious about avoiding overuse of broad-spectrum antibiotics in this patient population to minimize further selection for resistance. Sometimes this may mean withholding antibiotics in patients who are not very sick or very frail until there is more diagnostic clarity of whether a pneumonia is truly present or not.

●Risk factors for MRSA:

•Treatment in a unit in which >20 percent of S. aureus isolates associated with HAP are methicillin resistant

•Treatment in a unit in which the prevalence of MRSA is not known

•Colonization with and/or prior isolation of MRSA on culture at any body site (but especially the respiratory tract)

Choosing an empiric regimen — Empiric regimens for VAP (algorithm 1) and nvHAP (algorithm 2) are outlined in the algorithms.

Empiric regimens for VAP and nvHAP are determined by multiple factors [1]:

●Local known distribution of pathogens causing VAP and nvHAP and their antimicrobial susceptibility patterns (table 2 and table 3) [15].

●Patient's individual risk factors for multidrug resistance (table 2 and table 3), including [1,30]:

•Recent antibiotic therapy (if patients have recently received antibiotics, empiric therapy should generally be with a drug from a different class since earlier treatment may have selected for pathogens resistant to the initial class).

•Prior microbiology data (colonization with or prior isolation of MDR pathogens on culture from any body site).

●Potential adverse effects of antimicrobial agents.

●Potential drug interactions.

●Patient's severity of illness – the necessity of immediate, active antibiotics is greatest for patients with septic shock. The margin for error is greater for patients with less severe illness, allowing for the possibility of starting with a narrower empiric regimen and monitoring closely.

●Results of a good-quality Gram stain – a good-quality Gram stain can be useful for guiding the choice of initial therapy in patients without septic shock in whom antibiotics can be delayed until there is better diagnostic certainty of nvHAP or VAP. As an example, in a multicenter randomized trial of 206 patients with VAP, clinical cure (response) was noninferior in the group whose empiric antibiotic regimen was guided by Gram stain results compared with the group whose treatment was guided by guidelines (77 versus 72 percent; risk difference 0.05, 95% CI -0.07 to 0.17) [31]. Furthermore, the Gram stain group had reduced use of empiric anti-pseudomonal (70 versus 100 percent) and anti-MRSA agents (61 versus 100 percent) without significant difference in 28-day ICU-free days, ventilator-free days, or mortality between the two groups.

●Results of PCR multiplex – Identification of a probable pathogen on sputum PCR multiplex can be useful for guiding the choice of initial therapy. However, it is not yet widely available in all centers.

●Drug cost and availability.

●Clinician's familiarity with different antibiotics.

In keeping with the 2016 HAP and VAP guidelines, we try to assure that ≥95 percent of patients with clear evidence of VAP and signs of severe illness receive empiric therapy with activity against the most likely pathogens, because patients with severe illness experience worse outcomes if initial antimicrobial therapy is ineffective against the causative pathogen [1]. In a meta-analysis of three small observational studies of patients with VAP, delays in appropriate therapy of >24 to 48 hours were associated with significantly increased odds of mortality (OR 3.03, 95% CI 1.12-8.19) [20].

Early and aggressive treatment should be paired with active efforts to narrow, tailor, or stop antibiotics as soon as it is safe to do so. Antimicrobial stewardship reduces rates of nosocomial infections (ie, C. difficile, MRSA, vancomycin-resistant enterococcal infections, and MDR gram-negative infections) and antimicrobial expenditures without increasing mortality or extending length of hospital stay [32-36]. (See "Antimicrobial stewardship in hospital settings".)

All hospitals should regularly create and disseminate local antibiograms, ideally ones that are specific to their different units. Ideally, local resistance rates should be determined for each hospital unit and derived from pulmonary culture results from patients with VAP [37]. The resistance rate calculation should account for both the frequency of pathogens causing VAP and their resistance rates, resulting in a blended estimate of the probability that any given antibiotic will be active [1,38]. Because most hospitals do not have sufficient numbers of VAP isolates or data management support to generate such estimates, unit-wide or hospital-wide MRSA and P. aeruginosa resistance rates are acceptable, albeit conservative, proxies.

The recommendations below are generally in keeping with the 2016 ATS/IDSA guidelines on the management of HAP and VAP [1] (see 'Society guideline links' below). Modifications to these recommendations have been made based on the availability of newer agents that were not widely available at the time of the guideline's publication; further modifications may be needed based on the local prevalence of pathogens and local and hospital antimicrobial resistance patterns. The 2017 combined European and Latin American guidelines differ somewhat in their approach to initial antibiotic selection, opting to reserve empiric treatment for Pseudomonas species for those who are critically ill or have specific risk factors in an effort to reduce antimicrobial resistance [3].

Spectrum of activity of empiric therapy — All empiric therapy regimens for VAP and nvHAP should include agents with activity against Staphylococcus aureus, Pseudomonas aeruginosa, and other gram-negative bacilli. Regimens are stratified based on the patient's presence of MDR pathogen risk factors. Occasionally, consideration for other pathogens is warranted, as is described below:

●Legionella spp – Patients who have compromised immune systems, diabetes mellitus, renal disease, structural lung disease, or have been recently treated with glucocorticoids are at increased susceptibility for Legionella spp pneumonia. However, since nosocomial Legionella pneumonia is quite rare, we generally only add empiric anti-Legionella therapy (eg, with azithromycin or a fluoroquinolone) for patients in whom the clinical syndrome is consistent with Legionella pneumonia (eg, severe illness, bilateral patchy infiltrates), when there is a known Legionella spp outbreak, or if the patient is not improving on initial empiric therapy (see 'No clinical improvement after 48 to 72 hours' below). Nosocomial cases of nvHAP and VAP due to Legionella spp attributable to contamination of the hospital water supply have been reported. (See "Treatment and prevention of Legionella infection", section on 'Nosocomial pneumonia'.)

●Anaerobes – Anaerobes are rarely implicated in VAP and anti-anaerobic therapy is generally not necessary [39]. We consider empiric treatment against anaerobes (beta-lactam-beta-lactamase inhibitor, a carbapenem, metronidazole, moxifloxacin, or clindamycin) in patients with frank and/or gross aspiration (inhalation of vomitus), poor dentition, recent abdominal surgery, presence of possible lung abscess, or if not improving on initial empiric therapy (see 'No clinical improvement after 48 to 72 hours' below). Anti-anaerobic empiric therapy may also be considered if a good-quality (eg, from BAL) Gram stain demonstrates polymicrobial organisms or polymicrobial oral flora as those findings increase the likelihood of anaerobic bacterial presence. (See "Anaerobic bacterial infections" and "Aspiration pneumonia in adults", section on 'Choice of regimen'.)

●Stenotrophomonas spp – We do not generally administer empiric therapy for Stenotrophomonas species unless it has been detected on the patient's previous respiratory cultures, there is an outbreak of Stenotrophomonas spp present at the institution, or the patient is not improving on initial empiric therapy. (See 'No clinical improvement after 48 to 72 hours' below and "Stenotrophomonas maltophilia", section on 'Pulmonary infection'.)

●Viruses and fungi – It is important to keep in mind that not all HAP/VAP is caused by bacteria. Viruses (eg, coronavirus disease-2019 [COVID-19], influenza) and fungi can cause similar clinical presentations, especially in patients with severe immunocompromise [40]. We routinely test for circulating respiratory viruses in patients with suspected HAP, particularly during respiratory viral surges and seasons. (See "Seasonal influenza in nonpregnant adults: Treatment", section on 'Hospitalized patients' and "COVID-19: Management in hospitalized adults" and "Nonresolving pneumonia".)

●Role of prior culture data – Prior respiratory isolates should also guide antibiotic selection. A patient with a history of a pathogen in their sputum (eg, Acinetobacter spp) should receive empiric therapy with activity against the previous isolate if they have severe pneumonia.

Specific regimens are presented in the following section.

Regimens — The dosing provided below is intended for patient with normal renal function; dosing will need to be adjusted in those with renal dysfunction. When available, choice of antibiotics should be guided based on local antibiogram resistance rates.

No risk factors present — For patients with nvHAP/VAP who have no known risk factors for MDR pathogens (table 2 and table 3) and no increased mortality risk for those with nvHAP, we suggest a regimen that has activity against Pseudomonas, other gram-negative bacilli, and methicillin-susceptible S. aureus (MSSA). Preferred intravenous empiric antibiotic regimens include one of the following:

●Piperacillin-tazobactam 4.5 g IV every 6 hours

●Cefepime 2 g IV every 8 hours

We generally prefer piperacillin-tazobactam or cefepime because they are more likely to have activity against gram-negative bacilli than the fluoroquinolones, although preference should be guided by local antibiograms, when available. However, levofloxacin 750 mg IV daily may be preferred if there is a high suspicion for Legionella spp infection and local resistance rates of S. aureus, P. aeruginosa, and other gram-negative bacilli to fluoroquinolones are low (see 'Spectrum of activity of empiric therapy' above). The ATS/IDSA guidelines also include imipenem and meropenem as options, but we generally reserve these agents for patients with a high likelihood of infection with an extended-spectrum beta-lactamase (ESBL)-producing gram-negative bacillus or for patients in units where local antibiograms favor these agents over other broad-spectrum beta-lactams. (See 'Gram-negative pathogens' below.)

We do not use ceftazidime as a preferred agent because it has less activity against MSSA than the other beta-lactams suggested above for HAP and VAP. When ceftazidime is used for empiric therapy, an additional agent with activity against S. aureus (eg, linezolid or vancomycin) should also be used.

In patients with renal insufficiency, cefepime has been associated with neurotoxicity including seizures. Patients at risk should be closely monitored or alternative beta-lactams should be used. (See "Beta-lactam antibiotics: Mechanisms of action and resistance and adverse effects", section on 'Neurologic reactions'.)

In general, HAP patients who are not in the ICU tend to be less severely ill than VAP patients, and, therefore, the negative consequences of initial inappropriate antibiotic therapy may be less pronounced with non-ICU nvHAP than with VAP. In addition, MDR pathogens tend to be less common in patients who develop HAP outside of the ICU, particularly early in the hospitalization course (see "Epidemiology, pathogenesis, microbiology, and diagnosis of hospital-acquired and ventilator-associated pneumonia in adults", section on 'Microbiology'). For this reason, the 2016 HAP and VAP guidelines suggest that a smaller subset of patients with HAP as compared with VAP require empiric treatment for MRSA and MDR gram-negative organisms [1]. A single agent active against both MSSA and P. aeruginosa is often sufficient.

Risk factors for MDR gram-negative bacilli

No history of carbapenem-resistant pathogens — For patients with VAP or nvHAP who have risk factors for MDR gram-negative bacilli and no history of carbapenem-resistant pathogens, we suggest one of the following (algorithm 1):

●Meropenem 1 g IV every eight hours

●Imipenem 500 mg IV every six hours

Patients at risk of seizures from imipenem (eg, renal insufficiency, underlying central nervous system [CNS] disease) should be closely monitored or alternative beta-lactams should be used (see "Beta-lactam antibiotics: Mechanisms of action and resistance and adverse effects", section on 'Neurologic reactions'). Patients who cannot tolerate carbapenems as a class can be given an alternative agent as discussed in the section below. (See 'History of carbapenem-resistant pathogens' below and 'Allergy to penicillins or cephalosporins' below.)

Risk factors for MDR gram-negative bacilli increase the risk for the presence of extended-spectrum beta-lactamase (ESBL)-producing and ampC-producing pathogens. Patients infected with these MDR pathogens have demonstrated better clinical outcomes when treated with carbapenems rather than other broad-spectrum antibiotics, such as piperacillin-tazobactam or cefepime [41]. As an example, in a randomized trial of 391 patients with ESBL-producing E.coli or K. pneumoniae bloodstream infections, 3-day mortality was significantly lower in those who received meropenem compared with piperacillin-tazobactam (4 versus 12 percent) [42].

History of carbapenem-resistant pathogens — For patients with prior respiratory culture data demonstrating gram-negative bacilli that are resistant to carbapenems, we suggest empiric monotherapy with one of the following agents:

●Ceftazidime-avibactam 2.5 g IV every eight hours

●Ceftolozane-tazobactam 3 g IV every eight hours

●Imipenem-cilastatin-relebactam 1.5 g IV every six hours

●Meropenem-vaborbactam 4 g every eight hours

Selection among these antibiotics is guided by availability and patient's prior culture and susceptibility data, when available. Ceftazidime-avibactam and ceftolozane-tazobactam have the most efficacy data and real-world experience. (See "Combination beta-lactamase inhibitors, carbapenems, and monobactams", section on 'Beta-lactamase inhibitor combinations'.)  

If the beta-lactam beta-lactamase agents listed above are not available and there is a high suspicion for carbapenem-resistant gram-negative bacilli, then the addition of one of the following agents along with a carbapenem is a reasonable alternative:

●Tobramycin 5 to 7 mg/kg IV daily (figure 1)

●Colistin 300mg colistin base activity (CBA) loading dose followed by 150 mg CBA every 12 hours

●Aztreonam 2 g IV every eight hours

●Levofloxacin 750 mg IV or orally daily

●Ciprofloxacin 750 mg IV every 12 hours or 400mg orally every eight hours

A more detailed discussion of the efficacy and appropriateness of each agent is discussed below:

●Ceftazidime-avibactam – Ceftazidime-avibactam is a cephalosporin-beta-lactamase inhibitor combination approved by the US Food and Drug Administration (FDA) for treatment of nosocomial pneumonia [43]. We generally reserve use of ceftazidime-avibactam for patients with HAP or VAP who are known to be colonized by carbapenem-resistant Enterobacterales (CRE) because it has been shown to improve survival in these patients compared with other antibiotics. As an example, in a prospective, multicenter study of 137 patients with CRE infections, those who received ceftazidime-avibactam had a lower adjusted all-cause 30-day mortality compared with those who received colistin (9 versus 32 percent, absolute difference 23 percent, 95% CI 9-35) [44]. Other smaller observational studies comparing ceftazidime-avibactam with other antibiotics have shown similar trends in mortality [45,46]. (See "Combination beta-lactamase inhibitors, carbapenems, and monobactams", section on 'Ceftazidime-avibactam'.)

●Ceftolozane-tazobactam – Ceftolozane-tazobactam is a cephalosporin-beta-lactamase inhibitor combination approved by the FDA for treatment of nosocomial pneumonia [47,48]. We generally reserve use of ceftolozane-tazobactam for patients with HAP or VAP who are known to be colonized by carbapenem-resistant Pseudomonas aeruginosa because this agent has greater in vitro activity and higher barrier to resistance against P. aeruginosa compared with ceftazidime-avibactam [49,50]. Ceftolozane-tazobactam has been shown to be more effective and safer than some of the older alternatives (eg, aminoglycosides, polymyxins) in the treatment of P. aeruginosa pneumonia. As an example, in a retrospective, multicenter cohort study of 200 patients with drug-resistant P. aeruginosa, patients who received ceftolozane-tazobactam were more likely to achieve clinical cure (81 versus 61 percent; OR 2.63, 95% CI 1.31-5.30) and have lower rates of acute kidney injury (AKI) (6 versus 34 percent; OR 0.08, 95% CI 0.03-0.22) compared with those who received aminoglycosides or colistin [51]. (See "Combination beta-lactamase inhibitors, carbapenems, and monobactams", section on 'Ceftolozane-tazobactam'.)

●Imipenem-cilastatin-relebactam – Imipenem-cilastatin-relebactam is a carbapenem-beta-lactamase inhibitor combination that restores imipenem's activity against some CRE and Pseudomonas isolates. It was approved for the treatment of hospital-acquired bacterial pneumonia (including VAP) following a randomized controlled trial that showed similar clinical response, mortality, and serious adverse event rates between imipenem-cilastatin-relebactam and piperacillin-tazobactam [52]. We generally reserve the use of this agent to patients with history of CRE. In a randomized, double-blinded trial of 47 patients with imipenem-resistant pathogens, those who received imipenem-cilastatin-relebactam had higher rates of clinical response (71.4 versus 40 percent; absolute difference 26.3 percent, 90% CI 1.3-51.5), a trend towards lower mortality rate (9.5 versus 30 percent; absolute difference -17.3 percent, 90% CI -46.4 to 6.7), and nephrotoxicity (10.3 versus 56.3 percent; absolute difference -45.9 percent, 90% CI -69.1 to 18.4) compared with patients who received imipenem plus colistin [53]. (See "Combination beta-lactamase inhibitors, carbapenems, and monobactams", section on 'Imipenem-cilastatin-relebactam'.)

●Meropenem-vaborbactam – Meropenem-vaborbactam is a carbapenem-beta-lactamase inhibitor that has been approved for the treatment of nosocomial pneumonia in Europe but not in the United States. In the United States, it is only approved for treatment of complicated urinary tract infections (UTIs) caused by Enterobacterales [54]. Meropenem-vaborbactam is not effective against carbapenem-resistant P. aeruginosa, Acinetobacter baumannii, or Stenotrophomonas maltophilia. Meropenem-vaborbactam is active against some CRE but clinical data on its performance in the treatment of HAP and VAP in particular are sparse [55]. A small observational study of 131 patients with CRE infections (37 percent with pneumonia) compared meropenem-vaborbactam to ceftazidime-avibactam and found similar 30-day and 90-day mortality as well as adverse event rates [56,57]. (See "Combination beta-lactamase inhibitors, carbapenems, and monobactams", section on 'Meropenem-vaborbactam'.)

●Aminoglycosides – Because the aminoglycosides have poor lung penetration, increased risk of nephrotoxicity and ototoxicity, and poorer clinical response rates compared with other antibiotic classes, aminoglycosides are not recommended as monotherapy for gram-negative infections. However, rates of susceptibility among gram-negative bacilli are high, so we sometimes use them as part of the initial empiric regimen in patients with septic shock or rapidly progressive disease when there are little or no alternative options. Notably, since P. aeruginosa is a concern in patients with nvHAP/VAP, we prefer the use of tobramycin to other aminoglycosides as part of combination empiric therapy for nvHAP/VAP as it has greater intrinsic antipseudomonal activity.

If aminoglycosides are used as part of a combination regimen and subsequent culture results indicate that the isolate is susceptible to a beta-lactam, the aminoglycoside should be discontinued. We will also discontinue the aminoglycoside after two or three days in patients who have improved clinically and in whom cultures are negative. In general, every effort should be made to minimize exposure to aminoglycosides since even very brief courses of aminoglycosides increase the risk of nephrotoxicity [58]. (See "Aminoglycosides", section on 'Toxicity'.)

●Polymyxins – Addition of an alternative agent, such as intravenous colistin or polymyxin B, may be appropriate if highly resistant Pseudomonas spp, Acinetobacter spp, Enterobacterales (including Klebsiella pneumoniae) is suspected or established and newer beta-lactam beta-lactamase agents are not available [59]. Polymyxins are used rarely given their significant nephrotoxicity and should be avoided if alternative agents with adequate activity against gram-negative bacilli are available [51,60]. When they are required, an infectious disease physician and/or pharmacist with expertise using these agents should be consulted. Dosing recommendations are provided separately. (See "Polymyxins: An overview", section on 'Nephrotoxicity'.)

●Aztreonam – Although the use of two beta-lactams is generally avoided, in the absence of other options, it is acceptable to use aztreonam as a second agent for gram-negative bacteria with another beta-lactam because it has different targets within the bacterial cell wall [1]. (See "Combination beta-lactamase inhibitors, carbapenems, and monobactams", section on 'Monobactams (aztreonam)'.)

●Anti-pseudomonal fluoroquinolones – Although the ATS/IDSA guidelines recommend either an antipseudomonal fluoroquinolone or an aminoglycoside for the second agent for gram-negative bacilli when needed to increase the probability of active coverage and they also state that aminoglycosides should be avoided if alternative agents with adequate activity against gram-negative bacilli are available [1], we generally prefer an aminoglycoside over a fluoroquinolone for patients with severe disease if there is no concern for Legionella, as aminoglycosides are more likely to have in vitro activity against gram-negative bacilli in those with risk factors for resistance. When possible, clinicians should consult their local antibiogram to help determine whether similar resistance patterns exist at their institutions. (See "Fluoroquinolones".)

For patients with history of highly resistant or pan-resistant gram-negative bacilli, consultation with a specialist with expertise in antimicrobial management of antibiotic-resistant infections is recommended.

Risk factors for MRSA — Our approach to the treatment of HAP or VAP caused by MRSA is largely similar to the 2016 ATS/IDSA guidelines on HAP and VAP and the 2011 IDSA guidelines for the treatment of MRSA infections, which recommend either linezolid or vancomycin for infections suspected or proven to be due to MRSA [1,61].

Preferred agents — For patients with VAP or nvHAP who have risk factors for MRSA, we suggest adding one of the following agents to the patient's empiric therapy regimen:

●Linezolid 600 mg IV every 12 hours, which may be administered orally when the patient is able to take oral medications. (See 'Methicillin-resistant S. aureus' below.)

●Vancomycin dosing as summarized in the following table (table 4). (See 'Methicillin-resistant S. aureus' below and "Vancomycin: Parenteral dosing, monitoring, and adverse effects in adults".)

Adding an agent with MRSA activity to the treatment regimen allows for more flexibility in the choice of gram-negative agents since the gram-negative agent does not need to include activity against S. aureus. Potential regimens include any of the antibiotics mentioned above (eg, piperacillin-tazobactam, cefepime, imipenem, meropenem) as well as ceftazidime, ciprofloxacin, levofloxacin, or aztreonam. Among these agents, we generally prefer piperacillin-tazobactam, cefepime, or ceftazidime because they are more likely to have activity against gram-negative bacilli than the fluoroquinolones or aztreonam. The ATS/IDSA guidelines also include imipenem and meropenem as options, but we generally reserve these agents for situations in which they are required, such as in patients with a high likelihood of infection with an ESBL-producing gram-negative bacillus.

Because clinical outcomes appear to be similar for linezolid and vancomycin [62-64], we select between these agents based on other factors such as renal function, monitoring convenience, potential drug interactions, blood cell counts, and quality of IV access (see 'Methicillin-resistant S. aureus' below). As examples, when all other factors are equal, we prefer linezolid to vancomycin in patients with limited IV access or difficulty achieving therapeutic serum concentrations of vancomycin and we prefer vancomycin to linezolid in patients receiving selective serotonin-reuptake inhibitors and for patients with cytopenias.

Vancomycin and anti-pseudomonal beta-lactams, particularly piperacillin-tazobactam, have been associated with acute kidney injury [65,66], although the risk is uncertain [67]. Patients receiving vancomycin with piperacillin-tazobactam should be closely monitored for renal dysfunction or alternative antibiotic combinations should be used. (See "Vancomycin: Parenteral dosing, monitoring, and adverse effects in adults", section on 'Acute kidney injury'.)

Alternative options — In the rare circumstance that vancomycin or linezolid cannot be used, other alternatives include:

●Ceftaroline 600 mg IV every 12 hours

●Tedizolid 200 mg IV once daily (may have lower clinical cure rates compared with linezolid [68]; may be administered orally when the patient is able to take oral medications)

●Telavancin 10 mg/kg IV every 24 hours (there are several boxed warnings that must be considered before choosing it)

●Ceftobiprole 500 mg IV every eight hours (use only for nvHAP [69]; not available in the United States)

●Clindamycin (use only if local antibiogram shows most MRSA isolates are susceptible to clindamycin)

All efforts should be made to use the preferred agents against MRSA mentioned above (see 'Preferred agents' above). Alternative agents have less data supporting their use in nvHAP/VAP and may have less efficacy against MRSA and/or other significant adverse effects. (See 'Methicillin-resistant S. aureus' below.)

Risk factors for MRSA and MDR gram-negative bacilli — For patients who have risk factors for both MRSA and MDR gram-negative bacilli (including P. aeruginosa), we suggest combining an agent against MRSA with an agent(s) against MDR gram-negative bacilli, as discussed above (algorithm 1 and algorithm 2). (See 'Risk factors for MRSA' above and 'Risk factors for MDR gram-negative bacilli' above.)

Increased risk of mortality in nvHAP — For patients with nvHAP who have an increased risk of mortality from nvHAP (eg, need for intubation, septic shock), we suggest choosing the same empiric therapy regimen as for patients with VAP who have risk factors for both MRSA and MDR gram-negative bacilli (algorithm 1 and algorithm 2). (See 'Risk factors for MRSA and MDR gram-negative bacilli' above.)

TAILORING THERAPY — All patients with non-ventilator-associated HAP (nvHAP) or ventilator-associated pneumonia (VAP) should be evaluated for clinical response and microbiologic results after initial empiric antimicrobial therapy. Reassessing a patient's status 48 to 72 hours after the initiation of therapy with consideration of discontinuing antibiotics or narrowing the regimen (de-escalating therapy) based upon appropriate culture results may reduce the selective pressure for antimicrobial resistance and appears to be safe [70]. (See 'Antimicrobial resistance' above.)

Early de-escalation based on nasal MRSA swab results — We send nasal methicillin-resistant S. aureus (MRSA) swabs on all patients with suspected nvHAP/VAP. If the nasal MRSA swab result (polymerase chain reaction [PCR] and/or culture) are negative, we suggest stopping anti-MRSA empiric therapy to reduce unnecessary antibiotic use. Observational studies [71-73] and at least one randomized trial [74] suggest that testing nasal swabs or bronchoalveolar lavage fluid for MRSA via culture or PCR can help decrease utilization of anti-MRSA therapy if negative results are used as a trigger to stop anti-MRSA treatment. In a randomized trial of 45 patients with suspected MRSA VAP who were randomized to have empiric anti-MRSA therapy discontinued based on rapid nasal MRSA swab results versus usual care, patients in the experimental group had lower in-hospital mortality, shorter duration of anti-MRSA treatment, and no differences in duration of mechanical ventilation, length of intensive care unit (ICU) or hospital stay, or adverse events [74]. Early discontinuation of MRSA coverage in patients with negative cultures and/or PCRs may be associated with better outcomes, including less renal failure and possibly lower mortality rates [74,75].

No pathogen identified and clinically improving — For patients who are clinically improving who do not have an identified pathogen, empiric treatment for MRSA, Pseudomonas aeruginosa, or multidrug-resistant (MDR) gram-negative bacilli can be discontinued if these organisms have not grown in culture from a high-quality sputum or bronchoalveolar lavage specimen within 48 to 72 hours.

Pathogen identified and clinically improving — For patients in whom a pathogen has been identified, the empiric regimen should be tailored to the pathogen's susceptibility pattern [1,76,77]. Tailoring antibiotic therapy when a pathogen has been identified and susceptibilities are known is generally very safe; there are no clear associations between tailoring antibiotics and increased mortality [70,78-80], recurrent pneumonia [70,79], or longer ICU admission [70,80].

The approach to tailored therapy for MRSA and methicillin-susceptible Staphylococcus aureus (MSSA) is discussed below. (See 'Methicillin-resistant S. aureus' below and 'Methicillin-susceptible S. aureus' below.)

The approach to treatment of pneumonia caused by other pathogens is discussed in greater detail separately.

Methicillin-resistant S. aureus — For patients with confirmed MRSA hospital-acquired (or nosocomial) pneumonia (HAP)/VAP, we suggest treatment with vancomycin or linezolid. The decision to use one or the other antibiotic depends on various factors, including the patient's other comorbidities, drug-drug interactions, risk for side effects, and cost. If neither vancomycin nor linezolid can be used, ceftaroline is an alternative, although it is not approved by the US Food and Drug Administration (FDA) for treatment of HAP/VAP and its efficacy in treating MRSA pneumonia is limited to case series [81-83]. Other anti-MRSA agents such as tedizolid and ceftobiprole are not as effective in treating MRSA pneumonia and telavancin and tigecycline have significant side effects which limit their use as first-line agents. Daptomycin is inactivated by lung surfactant and cannot be used to treat pneumonia.

●Linezolid and vancomycin – Several trials have compared linezolid and vancomycin for the treatment of HAP and VAP; clinical outcomes appear to be similar when comparing these two agents [62-64]. As an example, in a meta-analysis of nine randomized trials that compared linezolid to vancomycin for HAP, no differences in mortality, clinical response, or MRSA eradication were detected [64]. Linezolid was associated with a higher rate of gastrointestinal adverse effects, but there were no differences in rates of acute kidney injury (AKI), thrombocytopenia, or drug discontinuation due to adverse effects.

Several trials included in this meta-analysis used lower doses for vancomycin than those recommended by the American Thoracic Society (ATS) and Infectious Diseases Society of America (IDSA) [1,61,84,85], which may have resulted in lower efficacy rates for this agent. In one trial that did use optimal vancomycin dosing, clinical success rates were lower for vancomycin when compared with linezolid in the per-protocol analysis (47 versus 58 percent; 11 percent absolute difference, 95% CI 0.5-22) [86]. However, higher rates of mechanical ventilation, renal dysfunction, and bacteremia in the vancomycin per-protocol group may account for this difference. No differences in all-cause 60-day mortality or overall adverse events were detected in this trial, although nephrotoxicity did occur more frequently with vancomycin than linezolid (18 versus 8 percent, p value not reported). Vancomycin failure might also be related to vancomycin minimum inhibitory concentrations (MICs). Pharmacokinetic and pharmacodynamic analyses suggest that lung vancomycin area under time-concentration curve to minimum inhibitory concentration ratio (AUC:MIC) >400 may be difficult to achieve (particularly for isolates with MIC >1) [87-89].

Some clinicians favor using an alternative to vancomycin for treatment of pneumonia caused by MRSA strains with vancomycin MICs ≥2 mcg/mLalthough there are no data demonstrating better outcomes with linezolid or other alternatives compared to vancomycin in patients infected with MRSA with high vancomycin MICs. Cohort studies have reported worse outcomes in patients with HAP due to MRSA with higher MICs. In one prospective cohort of 95 patients, patients with MRSA isolates with MICs ≥2 mcg/mL had higher mortality rates than those with lower MICs (24 percent versus 10 percent) [90]. Another study of 158 intensive care patients reported a stepwise increase in mortality as the vancomycin MIC increased from 0.75 to 3 mcg/mL and was present even for strains within the susceptible range [91]. These findings are controversial, however, because studies of MSSA bacteremia have also correlated higher vancomycin MICs with higher mortality rates even though all patients were treated with beta-lactams [92,93]. These studies suggest that worse outcomes in patients with higher MICs may be due to clinical factors other than vancomycin failure alone.

●Other agents – There has been interest in using other agents for the treatment of MRSA pneumonia, but none of the following agents are recommended as first-line options for MRSA pneumonia:

•Ceftaroline – Ceftaroline is a broad-spectrum cephalosporin with activity against MRSA and gram-negative pathogens (but not P. aeruginosa). It has been approved by the FDA for community-acquired pneumonia (CAP) but not for CAP caused by MRSA, nor for empiric treatment of HAP or VAP. Clinical success with ceftaroline use for the treatment of MRSA pneumonia has been reported in cases series but comparative studies are lacking [81-83]. (See "Cephalosporins", section on 'Fifth generation'.)

•Tedizolid – Tedizolid is an antibiotic from the same family as linezolid that inhibits bacterial protein synthesis. It is active against gram-positive bacteria including MRSA. In a randomized trial comparing tedizolid with linezolid, 726 ventilated patients with HAP or VAP suspected to be due to a gram-positive pathogen were randomized to tedizolid 200 mg IV daily for 7 days versus linezolid 600 mg IV twice a day for 10 days [68]. Adjunctive coverage for gram-negatives was permitted per clinicians' discretion. Tedizolid was found to be noninferior to linezolid with regard to all-cause mortality (28.1 versus 26.4 percent; 1.8 percent difference, 95% CI -8.2 to 4.7) but was associated with a lower rate of clinical cure (56.3 versus 63.9 percent; 7.6 percent difference, 97.5% CI -15.7 to 0.5). It is unclear whether the difference in clinical cure rates was due to the difference in drugs or difference in duration of treatment. Drug-related adverse events occurred in 8.1 percent of patients who received tedizolid versus 11.9 percent of those who received linezolid. Thrombocytopenia was more common in patients randomized to linezolid (38 versus 28 percent). (See "Linezolid and tedizolid (oxazolidinones): An overview".)

•Telavancin – Telavancin is an antibiotic with activity against MRSA [94]. In 2013, telavancin was approved by the FDA for the treatment of HAP and VAP caused by S. aureus but not for other bacterial causes of HAP or VAP; it is recommended only when alternative agents cannot be used and should ideally be reserved for patients with normal renal function [95]. In patients with normal renal function, the dose of telavancin is 10 mg/kg IV every 24 hours.

The FDA has included the following boxed warnings for telavancin [96]:

-Patients with pre-existing moderate or severe renal impairment (creatinine clearance [CrCl] ≤50 mL/minute) who were treated with telavancin for HAP or VAP had increased mortality compared with vancomycin. Use of telavancin in patients with pre-existing moderate or severe renal impairment (CrCl ≤50 mL/minute) should therefore be considered only when the potential benefit to the patient outweighs the potential risk.

-New-onset or worsening renal impairment has occurred in patients receiving telavancin. Renal function should be monitored in all patients receiving telavancin.

-Adverse developmental outcomes were observed in three animal species at clinically relevant doses. These findings raise concerns about potential adverse developmental outcomes in humans. Women of childbearing potential should have a serum pregnancy test prior to administration of telavancin, and its use should be avoided during pregnancy unless the potential benefit to the patient outweighs the potential risk to the fetus.

The data on the performance of telavancin for HAP and VAP come from two randomized trials (the ATTAIN trials) comparing telavancin versus vancomycin for hospitalized patients with HAP or VAP (29 percent had VAP) caused by gram-positive pathogens, particularly S. aureus. Pooled results from these two studies revealed no differences in overall cure rates or mortality in the vancomycin versus telavancin groups [97]. However, in a subgroup analysis of patients with renal impairment (CrCl <50 mL/minute), cure rates were lower in patients who received telavancin when compared with vancomycin (47 versus 55 percent; 8 percent difference, 95% CI -17.5 to 1.9) [98].

•Ceftobiprole – Ceftobiprole is a broad-spectrum cephalosporin with activity against a broad range of gram-positive bacteria including MRSA and penicillin- and ceftriaxone-resistant pneumococci, as well as gram-negative bacteria. It has not been approved by the FDA, but it has been approved in Europe and Canada for treatment of HAP and CAP but not VAP. In a trial of patients with HAP, 781 patients with HAP (including 210 with VAP) were randomly assigned to receive either ceftobiprole or linezolid plus ceftazidime [69]. In the intention-to-treat population, overall cure rates for ceftobiprole versus linezolid plus ceftazidime were similar (50 versus 53 percent). Cure rates in HAP (excluding VAP) patients were also similar (60 versus 59 percent). However, cure rates in VAP patients were substantially lower in those who received ceftobiprole (23 versus 37 percent). Microbiologic eradication rates in HAP (excluding VAP) patients were 63 versus 68 percent (microbiologically evaluable [ME], 95% CI -16.7 to 7.6) and in VAP patients were 30 versus 50 percent (ME, 95% CI -38.8 to -0.4) [99]. (See "Cephalosporins", section on 'Fifth generation'.)

•Tigecycline – Tigecycline is a broad-spectrum antibiotic with activity against MRSA. It has been approved by the FDA for skin and skin structure infections and intra-abdominal infections caused by MRSA. It has also been approved for CAP but not for CAP caused by MRSA or for HAP or VAP. In 2010, the FDA issued a safety announcement regarding an increased mortality risk associated with the use of tigecycline compared with other drugs observed in a pooled analysis of 13 trials [100]. The increased risk was seen most clearly in patients treated for HAP, particularly VAP. In 2013, the FDA added a boxed warning in reaction to an additional analysis showing an increased risk of death associated with tigecycline use [101]. The boxed warning states that tigecycline should be reserved for use in situations when alternative agents are not suitable. In an analysis of 10 trials conducted for FDA-approved uses (CAP, complicated skin and skin structure infections, complicated intra-abdominal infections), tigecycline was associated with increased mortality compared with other antibacterial agents (2.5 versus 1.8 percent, adjusted risk difference 0.6 percent, 95% CI 0.0-1.2). Most deaths resulted from worsening infections, complications of infection, or underlying comorbidities. Randomized trials of patients with HAP have reported similar results [102,103]. (See "Tetracyclines".)

Methicillin-susceptible S. aureus — If sputum or bronchoalveolar lavage (BAL) culture reveals MSSA, empiric therapy for MRSA should be replaced with nafcillin (2 g IV every four hours), oxacillin (2 g IV every four hours), or cefazolin (2 g IV every eight hours) [1]. Of these, cefazolin is generally preferred since it is associated with fewer adverse effects (renal dysfunction, hepatotoxicity) compared to oxacillin or nafcillin. If no other organisms have been isolated then monotherapy with one of these agents is adequate. If culture data suggests a polymicrobial infection, then one can choose either monotherapy with an agent that covers both MSSA and gram-negatives (eg, cefepime, ceftriaxone) or the combination of an anti-gram-positive agent and an anti-gram-negative agent tailored to the recovered pathogens' antibiotic susceptibility profiles. In patients with concurrent MSSA bacteremia, the inclusion of a directed-MSSA agent (eg, cefazolin, oxacillin) is preferred over a broader-spectrum beta-lactam (see "Clinical approach to Staphylococcus aureus bacteremia in adults", section on 'Methicillin-sensitive S. aureus'). Ceftazidime has limited anti-gram-positive activity and should not be used in isolation for patients with MSSA pneumonia.

Streptococci and other gram-positive pathogens — For HAP/VAP caused by streptococci and other gram-positive pathogens (eg, S. lugdunensis), we tailor therapy to the pathogen isolated from culture. Although streptococci are susceptible to many different antibiotic classes, beta-lactams are preferred, especially penicillin and ceftriaxone. (See "Pneumococcal pneumonia in patients requiring hospitalization", section on 'Treatment' and "Infections due to the Streptococcus anginosus (Streptococcus milleri) group", section on 'Management' and "Group B streptococcal infections in nonpregnant adults", section on 'Treatment' and "Treatment of enterococcal infections", section on 'Antibiotic agents'.)

Gram-negative pathogens — For HAP/VAP caused by non-extended-spectrum beta-lactamase (ESBL) and non-carbapenem-resistant Enterobacterales (CRE) isolates, we favor choosing an antibiotic with the narrowest spectrum of activity that will be effective against the gram-negative pathogen based on its susceptibility profile. Common antibiotic selections include ampicillin-sulbactam, cefepime, ceftazidime, piperacillin-tazobactam, ciprofloxacin, or levofloxacin. Carbapenems are usually not indicated for pneumonia caused by highly susceptible gram-negative pathogens unless there is concern for allergies. (See 'Anaerobes' below and 'Allergy to penicillins or cephalosporins' below.)

The approach to treatment of pneumonia caused by Pseudomonas aeruginosa, ESBL and carbapenemase-producing gram-negative bacilli, Acinetobacter spp, Stenotrophomonas maltophilia, Haemophilus influenzae, and Legionella spp is discussed in greater detail separately.

●(See "Extended-spectrum beta-lactamases", section on 'Treatment options'.)

●(See "Carbapenem-resistant E. coli, K. pneumoniae, and other Enterobacterales (CRE)", section on 'Approach to treatment'.)

●(See "Acinetobacter infection: Treatment and prevention".)

●(See "Stenotrophomonas maltophilia".)

●(See "Treatment and prevention of Legionella infection", section on 'Treatment of Legionnaires' disease (Legionella pneumonia)'.)

●(See "Epidemiology, clinical manifestations, diagnosis, and treatment of Haemophilus influenzae", section on 'Directed treatment'.)

Anaerobes — nvHAP/VAP caused by anaerobes is uncommon. When it does occur, it generally is caused by aspiration of oral anaerobic flora. The presence of gut anaerobic pathogens in respiratory specimens, such as Bacteroides fragilis, should prompt suspicion and evaluation for intestinal pathology. (See "Aspiration pneumonia in adults" and "Anaerobic bacterial infections", section on 'Pleuropulmonary infections'.)

No clinical improvement after 48 to 72 hours — Patients who have not improved within 72 hours of starting empiric antibiotics or appropriate antibiotics for the causative pathogen should be evaluated for complications (eg, empyema, lung abscess), other sites of infection, and alternate diagnoses (eg, thromboembolic disease, pulmonary edema, malignancy, hypersensitivity reaction, etc).

If the diagnosis of bacterial pneumonia appears certain, there is no evidence of a pyogenic complication that requires drainage (eg, empyema, lung abscess), no evidence of untreated infections elsewhere in the body, and the patient has risk factors for drug-resistant pathogens (eg, prolonged hospitalization, recent exposure to multiple antibiotics), additional diagnostic pulmonary cultures should be obtained and the empiric regimen can be expanded to cover additional resistant organisms.

In patients with progressive HAP despite broad-spectrum antibiotic use, other potential hospital-acquired infections such as respiratory viral infections or Legionella should be considered. During outbreaks of highly drug-resistant organisms, such as Acinetobacter species or Stenotrophomonas species, including an antibiotic that targets these organisms in the empiric treatment regimen may be warranted. In immunocompromised patients, the differential diagnosis should be broad and include fungal, viral, parasitic, and less common bacterial pathogens. (See "Epidemiology of pulmonary infections in immunocompromised patients".)

DURATION — We suggest treating most patients with non-ventilator-associated HAP (nvHAP) or ventilator-associated pneumonia (VAP) for seven days, in agreement with the 2016 American Thoracic Society/Infectious Diseases Society of America (ATS/IDSA) guidelines and the combined 2017 European and Latin American guidelines on HAP and VAP [1,3]. Seven days appears to be as effective as longer durations in most circumstances and may reduce the emergence of resistant organisms [1,5].

For selected patients with metastatic infection, gram-positive bacteremia, slow response to therapy, immunocompromise, and pyogenic complications such as empyema or lung abscess, the duration of therapy should be individualized and courses longer than seven days may be warranted. (See "Clinical approach to Staphylococcus aureus bacteremia in adults", section on 'Completing antibiotic therapy' and "Pseudomonas aeruginosa pneumonia", section on 'Directed antimicrobial therapy'.)

Monitoring serial procalcitonin levels can help guide the decision to discontinue antibiotics. While the optimal approach to using procalcitonin in patients with hospital-acquired (or nosocomial) pneumonia (HAP) or VAP has not been determined, a low or declining procalcitonin level (eg, <0.25 ng/mL or ≥80 percent decrease from peak) in a patient who has clinically improved on antibiotics provides additional reassurance that antibiotics can be safely stopped [104-111]. (See "Procalcitonin use in lower respiratory tract infections", section on 'Ventilator-associated pneumonia'.)

A seven-day course of antimicrobial therapy is supported by two meta-analyses of up to six randomized trials evaluating over 1000 patients with HAP or VAP, in which short courses (7 to 8 days) of therapy were as effective as longer courses (10 to 15 days) [1,112,113]. In subgroup analysis of one of the meta-analyses [112], a higher rate of recurrent pneumonia was observed in patients with pneumonia caused by nonfermenting gram-negative rods, such as P. aeruginosa, who received shorter courses of therapy. This finding was primarily driven by a single large trial that used a microbiologic definition of pneumonia [114]; hence, this finding may have indicated microbiologic persistence rather than clinical failure. No differences in ventilator-free days, organ failure-free days, length of stay, or mortality were detected in patients with nonfermenting gram-negative bacillus infections randomized to short versus long courses [1,112,115]. Another multicenter, randomized trial of 186 patients with VAP due to P. aeruginosa demonstrated similar findings [116]. (See "Pseudomonas aeruginosa pneumonia", section on 'Duration of therapy'.)

Although a retrospective cohort study suggested that patients with suspected VAP who have stable and minimal ventilator settings (positive-end expiratory pressure <5 cm H₂O and FiO₂ ≤40 percent) have similar outcomes whether they are treated with ≤3 days or >3 days of antibiotics [117], these observations must be confirmed in a randomized trial before this strategy can be broadly applied.

CONVERSION TO ORAL ANTIBIOTICS — Generally, patients can be switched to oral therapy when they are hemodynamically stable, clinically improving, and able to tolerate oral medications [118]. If a pathogen has been identified, the choice of antibiotic for oral therapy should be based on the organism's susceptibility pattern. If a pathogen has not been identified, the oral antibiotic selected should be based on an appropriate de-escalation approach and generally can exclude coverage for MRSA and P. aeruginosa. All oral antibiotics selected for the treatment of pneumonia should have good lung penetration.

OTHER MANAGEMENT CONSIDERATIONS

Prolonged infusions of beta-lactams — We favor prolonged infusions of beta-lactams for empiric and targeted therapy of gram-negative bacilli with at least one of the following:

●In critically ill patients

●Elevated but susceptible minimal inhibitory concentrations (MICs) to the chosen agent (table 5)

Suitable agents for prolonged infusions include piperacillin-tazobactam, meropenem, imipenem, and cefepime. The decision to use this dosing strategy should also take into account logistical issues such as staffing or intravenous (IV) access availability.

Beta-lactam antibiotics demonstrate a time-dependent effect on bacterial eradication. Prolonged infusions achieve pharmacodynamic efficacy targets defined for beta-lactam antibiotics more effectively than traditional, short, intermittent infusions (eg, administered over 30 minutes). A prolonged infusion may therefore improve microbiologic and clinical cure, especially for pathogens with high MICs. Prolonged infusion for intravenous beta-lactams may include either a continuous infusion (over the entire dosing interval) or an extended infusion (over three to four hours).

Pharmacologic and clinical data indicate that patients who have an elevated risk of drug-resistant pathogens or who are critically ill in the setting of a severe infection are most likely to benefit from prolonged infusions. In a patient-level meta-analysis of 22 randomized trials including 1876 patients that compared prolonged versus rapid infusions of antipseudomonal beta-lactams in patients with sepsis, mortality was 30 percent lower in patients receiving prolonged infusions (risk ratio [RR] 0.70, 95% CI 0.56-0.87) [119]. Further discussion of prolonged infusions of beta-lactams is provided separately. (See "Prolonged infusions of beta-lactam antibiotics".)

The traditional intermittent dosing of each agent for hospital-acquired (or nosocomial) pneumonia (HAP) is described in this topic, but we favor prolonged infusions of antipseudomonal beta-lactams to optimize pharmacodynamics, especially in critically ill patients with infections caused by gram-negative bacilli and for patients with infections caused by gram-negative bacilli that have elevated but susceptible MICs to the chosen agent (table 5).

Aerosolized antibiotics — Knowledge on the most appropriate use of aerosolized antibiotics is evolving. We do not use aerosolized antibiotics as part of empiric therapy for the treatment of non-ventilator-associated HAP (nvHAP)/ventilator-associated pneumonia (VAP). Aerosolized colistin, polymyxin, or aminoglycosides can be used as adjunctive therapy (in combination with IV antibiotics) in patients with VAP or HAP caused by multidrug-resistant (MDR) gram-negative bacilli, such as A. baumannii or P. aeruginosa [1,120-126]. (See "Acinetobacter infection: Treatment and prevention", section on 'Pneumonia' and "Pseudomonas aeruginosa pneumonia", section on 'Inhaled antibiotics for selected patients'.)

Aerosolization may increase antibiotic concentrations at the site of infection and may be particularly useful for treatment of organisms that have high MICs to systemic antimicrobial agents [127]. However, the evidence supporting use of aerosolized antibiotics is not strong [128-130] and administration can be associated with adverse effects, particularly in hypoxemic patients [128,129]. As an example, in a meta-analysis of 13 randomized trials [131], adding aerosolized amikacin to IV antibiotics for gram-negative pneumonia versus treating with IV antibiotics alone was associated with higher rates of clinical cure in unblinded studies (9 studies, 700 patients; risk ratio 1.46, 95% CI 1.31-1.63) but not in blinded studies (4 studies, 750 patients; risk ratio 1.01, 95% CI 0.89-1.16). Similarly, adding aerosolized amikacin to IV treatment versus IV treatment alone was associated with no difference in duration of mechanical ventilation, intensive care unit (ICU) length-of-stay, or mortality. However, adjunctive aerosolized amikacin was associated with more bronchospasm, sometimes with hypoxemia (6.4 versus 2.4 percent of patients; risk ratio 2.55, 95% CI 1.40-4.66). There was no difference in the rate of renal dysfunction between treatment strategies. Other studies have reported occasional episodes of cardiac arrest when aerosolizing antibiotics due to obstruction of the ventilator circuit expiratory filter [120,122].

The addition of aerosolized colistin to intravenous colistin is rarely used with the advent of newer beta-lactam agents [132,133]. In a meta-analysis of three randomized trials including over 300 patients, addition of aerosolized colistin was associated with better microbiologic eradication rates but higher rates of bronchospasm [133]. Clinical response was similar between those who received IV monotherapy and those with aerosolized and intravenous colistin (66 versus 62 percent; OR 1.28, 95% CI 0.56-2.93).

Allergy to penicillins or cephalosporins — For patients who are allergic to penicillins or cephalosporins, the type and severity of reaction should be assessed. The great majority of patients who are allergic to penicillin can receive third-generation (and higher) cephalosporins or carbapenems using a simple graded challenge (1/10 dose followed by a one-hour period of observation; if no symptoms, give the full dose followed by another hour of observation). If a skin test for penicillin allergy is positive or if there is significant concern to warrant avoidance of a cephalosporin or carbapenem (IgE-mediated reaction), aztreonam (2 g IV every eight hours) is a reasonable alternative. It is important to note that aztreonam does not provide activity against gram-positive bacteria like S. aureus. When this agent is used for empiric therapy, an additional agent with activity against S. aureus (eg, another beta-lactam, linezolid, vancomycin) should also be used.

Approaches to the choice of antibiotics for patients with penicillin (algorithm 3) and cephalosporin (algorithm 4 and algorithm 5) allergies can be found in the algorithms. List of cephalosporins with different side chains is listed in the table (table 6). (See "Choice of antibiotics in penicillin-allergic hospitalized patients" and "Allergy evaluation for immediate penicillin allergy: Skin test-based diagnostic strategies and cross-reactivity with other beta-lactam antibiotics" and "Immediate cephalosporin hypersensitivity: Allergy evaluation, skin testing, and cross-reactivity with other beta-lactam antibiotics".)

PROGNOSIS — Despite high absolute mortality rates in hospital-acquired (or nosocomial) pneumonia (HAP) patients, the mortality attributable to the infection is difficult to gauge. Many studies report that HAP is associated with high crude mortality rates. However, many of these critically ill patients die from their underlying disease and not from pneumonia. While crude all-cause mortality associated with ventilator-associated pneumonia (VAP) has ranged from 20 to 50 percent in different studies [1], a meta-analysis of randomized trials of VAP prevention estimated the attributable mortality at 13 percent [134]. Another study estimated that eliminating VAP would lead to a relative decrease in 60-day intensive care unit mortality of 3.6 percent [135]. Similarly, a large study of non-ventilator-associated HAP (nvHAP) across 284 United States hospitals estimated that eliminating nvHAP would decrease overall hospital mortality rates by 7.3 percent [136].

Variables associated with increased mortality include [1,137-146]:

●Serious illness at the time of diagnosis (eg, high Acute Physiology and Chronic Health Evaluation [APACHE] score, shock, coma, respiratory failure, acute respiratory distress syndrome [ARDS])

●Bacteremia

●Severe underlying comorbid disease

●Infection caused by an organism associated with multidrug resistance (eg, P. aeruginosa, Acinetobacter spp, and Enterobacteriaceae, including K. pneumoniae)

●Multilobar, cavitating, or rapidly progressive infiltrates on lung imaging

●Delay in the institution of effective antimicrobial therapy

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Hospital-acquired pneumonia and ventilator-associated pneumonia in adults".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or email these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topic (see "Patient education: Hospital-acquired pneumonia (The Basics)")

SUMMARY AND RECOMMENDATIONS

●Definitions – Pneumonia is frequently categorized based on site of acquisition (table 1). (See 'Definitions' above.)

●Identifying risk factors for MDR pathogens and mortality in VAP and nvHAP – Risk factors for multidrug-resistant (MDR) pathogens and/or mortality in patients with ventilator-associated pneumonia (VAP) (table 2) and non-ventilator-associated HAP (nvHAP) (table 3) are listed in the tables. (See 'Identifying risk factors for MDR pathogens' above.)

●Choosing an empiric regimen – Empiric regimens for VAP (algorithm 1) and nvHAP (algorithm 2) are outlined in the algorithms. The choice of the antibiotic treatment regimen for hospital-acquired (or nosocomial) pneumonia (HAP) should be informed by the patient's recent antibiotic therapy (if any), the resident flora and resistance rates in the hospital or intensive care unit (ICU), the presence of underlying diseases, severity of illness, available culture data (including the patient's past microbiology data) and Gram stain, and additional risk factors for MDR pathogens. Further considerations include potential toxicities, potential drug interactions, cost, availability, and clinician familiarity with proposed medications. (See 'Choosing an empiric regimen' above.)

•No risk factors present – For patients with nvHAP/VAP who have no known risk factors for MDR pathogens (table 2 and table 3) and no increased mortality risk for those with nvHAP, we suggest a regimen that has activity against Pseudomonas, other gram-negative bacilli, and methicillin-susceptible S. aureus (MSSA), such as piperacillin-tazobactam or cefepime (Grade 2C). Levofloxacin is a reasonable alternative if there is a high suspicion for Legionella spp pneumonia and local resistance rates of S. aureus, P. aeruginosa, and other gram-negative bacilli to fluoroquinolones are low. (See 'No risk factors present' above.)

•Risk factors for MDR gram-negative bacilli

-No history of carbapenem-resistant pathogens – For patients with VAP or nvHAP who have risk factors for MDR gram-negative bacilli and no history of carbapenem-resistant pathogens, we suggest a carbapenem (eg, meropenem, imipenem) (Grade 2C). (See 'No history of carbapenem-resistant pathogens' above.)

-History of carbapenem-resistant pathogens present – For patients with high concern for or with prior respiratory culture data demonstrating carbapenem-resistant pathogens, we suggest empiric monotherapy with one of the newer beta-lactam beta-lactamase inhibitors rather than combination therapy with two different agents (Grade 2C). Appropriate beta-lactam beta-lactamase inhibitors include ceftazidime-avibactam, ceftolozane-tazobactam, imipenem-cilastatin-relebactam, and meropenem-vaborbactam. Selection among these antibiotics is guided by availability and the patient's prior culture and susceptibility data, when available.

If the beta-lactam beta-lactamase agents mentioned above are not available and there is a high suspicion for carbapenem-resistant gram-negative bacilli, then the addition of tobramycin, an anti-pseudomonal fluoroquinolone, aztreonam, or colistin along with a carbapenem is a reasonable alternative based on previous culture susceptibility patterns. (See 'History of carbapenem-resistant pathogens' above.)

•Risk factors for MRSA – For patients with VAP or nvHAP who have risk factors for methicillin-resistant S. aureus (MRSA), we recommend adding linezolid or vancomycin to the patient's empiric therapy regimen (Grade 1B). (See 'Risk factors for MRSA' above.)

•Risk factors for MRSA and MDR gram-negative bacilli – For patients who have risk factors for both MRSA and MDR gram-negative bacilli (including P. aeruginosa), we combine an agent against MRSA with an agent(s) against MDR gram-negative bacilli, as discussed above (algorithm 1 and algorithm 2). (See 'Risk factors for MRSA' above and 'Risk factors for MDR gram-negative bacilli' above.)

•Increased risk of mortality in nvHAP – For patients with nvHAP who have an increased risk of mortality, we choose the same empiric therapy regimen as for patients with VAP who have risk factors for both MRSA and MDR gram-negative bacilli (algorithm 1 and algorithm 2). (See 'Risk factors for MRSA and MDR gram-negative bacilli' above.)

●Tailoring therapy

•Early de-escalation based on nasal MRSA swab – We send nasal MRSA swabs on all patients with suspected MRSA nvHAP/VAP. If the nasal MRSA swab result (polymerase chain reaction [PCR] and/or culture) is negative, we suggest stopping anti-MRSA empiric therapy to reduce unnecessary antibiotic use (Grade 2B). (See 'Early de-escalation based on nasal MRSA swab results' above.)

•No pathogen identified and clinically improving – For patients who are clinically improving who do not have an identified pathogen, empiric treatment for MRSA, Pseudomonas aeruginosa, or MDR gram-negative bacilli can be discontinued if these organisms have not grown in culture from a high-quality sputum specimen within 48 to 72 hours. (See 'No pathogen identified and clinically improving' above.)

•Pathogen identified and clinically improving – For patients in whom a pathogen has been identified, the empiric regimen should be tailored to the pathogen's susceptibility pattern. (See 'Pathogen identified and clinically improving' above.)

•No clinical improvement after 48 to 72 hours – Patients who have not improved within 72 hours of starting empiric antibiotics or appropriate antibiotics for the causative pathogen should be evaluated for complications (eg, empyema, lung abscess), other sites of infection, and alternate diagnoses (eg, thromboembolic disease, pulmonary edema, malignancy, hypersensitivity reaction, etc). If further diagnostic evaluation is unrevealing, additional diagnostic pulmonary cultures should be obtained and the empiric regimen can be expanded to cover additional resistant organisms (eg, Legionella pneumoniae, Stenotrophomonas maltophilia, Acinetobacter spp). (See 'No clinical improvement after 48 to 72 hours' above.)

●Duration and conversion to oral antibiotics

•We suggest treating most patients with HAP or VAP for seven days (Grade 2B). For selected patients with metastatic infection, gram-positive bacteremia, slow response to therapy, immunocompromise, and pyogenic complications such as empyema or lung abscess, the duration of therapy should be individualized and courses longer than seven days may be warranted. (See 'Duration' above.)

•Generally, patients can be switched to oral therapy when they are hemodynamically stable, clinically improving, and able to tolerate oral medications. (See 'Conversion to oral antibiotics' above.)

ACKNOWLEDGMENT — UpToDate gratefully acknowledges John G Bartlett, MD (deceased), who contributed on earlier versions of this topic and was a founding Editor-in-Chief for UpToDate in Infectious Diseases.