Incidence

Nosocomial pneumonia is the second most common nosocomial infection and the leading cause of death from nosocomial infections for critically ill patients. The incidence of nosocomial pneumonia is age dependent, with 5 of every 1000 cases affecting hospitalized patients younger than 35 years of age and up to 15 of 1000 in hospitalized patients older than 65.⁵ In earlier reports, nosocomial pneumonia increased hospital stay by 7 to 9 days per patient, accounting for up to 25% of all ICU infections and for more than 50% of the antibiotics prescribed.⁶ A Spanish report by Sopena et al.⁷ in 2005 studied the epidemiology of nosocomial pneumonia in 186 non-ICU patients from 12 hospitals and found that nosocomial pneumonia was observed mostly in elderly patients with underlying diseases and was primarily caused by S. pneumoniae, Legionella pneumophila, Aspergillus spp., and P. aeruginosa. The mortality rate was 26%, with an attributable mortality of 13%. Cook et al.⁸ estimated that the risk of VAP is 1% per day on mechanical ventilation⁸; they also demonstrated that the risk changes over time, being 3% the first 5 days on mechanical ventilation, 2% from the 5th to 10th day, and 1% for the remaining days. Considering that most invasively ventilated patients are intubated for less than a week, nearly half of VAP cases occur during the first days of mechanical ventilation.

Mortality

The crude mortality from nosocomial pneumonia may be as high as 30% to 70%, although several cofactors influence mortality during critical illness and make it extremely difficult to determine attributable mortality.⁹ Mortality caused by VAP has been defined as the percentage of deaths that would not have occurred in the absence of the infection. Several case-matching studies have estimated that one third to one half of all VAP-related deaths are the direct result of the infection, with a higher mortality rate in cases caused by P. aeruginosa or Acinetobacter spp. and associated with bacteremia.¹⁰ The development of VAP is accompanied by a 1.8- to 4-fold increase in the risk of death. A multicenter French study⁹ evaluated the attributable mortality and risk factors for death for late-onset pneumonia. They evaluated 764 patients admitted to the ICU for more than 96 hours and found a 47% mortality in patients with VAP versus 22% in the total population. Moreover, mortality was inversely related to adequacy of the initial empirical therapy. Similarly, Luna et al.¹¹ assessed the appropriateness and delay of antibiotic therapy in 76 mechanically ventilated patients with bacteriologically confirmed VAP and found an overall mortality of 52.6%. Based on current evidence, nosocomial pneumonia is associated with a high mortality rate, but precise estimates of mortality attributable to this condition are not possible, owing to heterogeneity between patient populations, microbial patterns, antibiotic treatment, and diagnostic methods.¹²

Role of Tracheal Tube in the Pathogenesis of ventilator-associated pneumonia

Pulmonary aspiration of colonized oropharyngeal secretions across the tracheal tube cuff is the main pathogenic mechanism for development of VAP. The endotracheal tube (ETT), commonly used in the ICU for long-term mechanically ventilated patients, includes a high-volume, low-pressure (HVLP) cuff. HVLP cuffs were originally designed to control pressure exerted against the tracheal wall and prevent tracheal injury.^13–15 However, the potential diameter of the HVLP cuff is two to three times larger than the tracheal diameter, so when the tracheal cuff is inflated within the trachea, folds invariably form along the cuff surface, causing consistent micro- and macroaspiration of oropharyngeal secretions.¹⁶

Pathogens may also grow on the internal surface of the ETT and ultimately translocate into the lungs. The ETT is commonly made of polyvinyl chloride (PVC), and bacteria easily adhere to its internal surface to form a complex structure called biofilm^17,18 (Figure 67-1). Biofilm is composed of sessile bacteria embedded within a self-produced exopolysaccharide matrix.¹⁹ Biofilm on the internal surface of an ETT can be identified early following tracheal intubation.^20,21 Sessile bacteria undergo phenotypic differentiation from their planktonic counterparts, and most of such differentiation constitutes a survival advantage. Indeed, bacteria within the biofilm are difficult to eradicate, and antibacterial efficacy of the host’s immune response and antibiotics are largely reduced. During mechanical ventilation, biofilm particles may dislodge into the airways as a result of inspiratory airflow¹⁷ and invasive medical interventions such as tracheal aspiration²² and bronchoscopy. Several studies have assessed the role of bacterial biofilm on pathogenesis of VAP and confirmed that the ETT biofilm is difficult to eradicate and constitutes a persistent source of colonization. Adair et al.²³ studied 40 tracheal tubes obtained from critically ill patients with and without VAP and compared the genotype of bacteria retrieved from the lower airways and the ETT. In 70% of the samples obtained from patients with VAP, the authors found the same genotype in bacteria from ETT biofilm and the patients’ airways. Moreover, they confirmed that antibiotic susceptibility was lower in pathogens isolated from within the biofilm.

Figure 67-1 Laboratory studies to assess biofilm formation on internal surface of tracheal tube, following oropharyngeal challenge in pigs of Pseudomonas aeruginosa (strain PAO1) and 72 hours of mechanical ventilation. A, Internal surface of tracheal tube at extubation, largely covered by respiratory secretions. B, Cross-section of tracheal tube stained with LIVE/DEAD BacLight bacterial viability kit and imaged with confocal scanning laser microscopy. Bacterial biofilm adheres to internal ETT surface. White arrows indicate bacteria embedded into biofilm matrix. ETT, endotracheal tube. C, Scanning electron microscopy of tracheal tube lumen. Note presence of amorphous deposits on most of surface. D, Higher magnification of tracheal tube lumen through scanning electron microscopy. P. aeruginosa sessile cells are clearly visible within biofilm extracellular polymeric substance.

Sources of Colonization

Tracheally intubated patients in the ICU can be colonized either exogenously or endogenously. Patients are colonized exogenously by contaminated respiratory equipment, the ICU environment, and the hands of the ICU staff. Several reports have described ICU outbreaks due to colonized bronchoscopes,^24,25 water supply,^26,27 respiratory equipment,^28,29 humidifiers,³⁰ ventilator temperature sensors,^31,32 respiratory nebulizers,^33,34 and contaminated environment.³⁵ Several factors play a significant role in reducing risks for cross-transmission of pathogens. Indeed, in every ICU an adequate ratio of single rooms to open beds should be provided; healthcare personnel should be adequately trained on infection control and preventive strategies; strict sterilization protocols and hand washing with alcohol-based solutions should be implemented, and finally, lower patient-to-nurse ratios are advantageous.

Endogenous colonization is believed to be the primary pathogenic mechanism for VAP development. In the critically ill patient, the oral flora shifts early to a predominance of aerobic gram-negative pathogens,^36,37 Pseudomonas aeruginosa, and methicillin-resistant Staphylococcus aureus (MRSA). Therefore, pulmonary aspiration of oropharyngeal contents increases the risk for airway colonization and infection. Following aspiration and colonization of the airways, the occurrence of VAP primarily depends on the size of the inoculum, functional status, and the competency of host defenses. There is still controversy regarding the exact sequence of colonization and sources of infection in the pathogenesis of VAP. Early studies by Feldman et al.³⁸ found that in patients undergoing mechanical ventilation, the oropharynx is the first site to be colonized by pathogens (36 hours), followed by the stomach (36-60 hours), the lower respiratory tract (60-84 hours), and thereafter the tracheal tube (60-96 hours).

Impairment of Respiratory Defense During Critical Illness and Tracheal Intubation

In healthy subjects, the physiologic adduction of the true and false vocal folds provides full closure of the airways, and efficiently prevents aspiration of pathogen-laden oropharyngeal contents. The airways are additionally protected by the epiglottis, which moves over the top of the larynx to divert any fluid or solids from passing into the airways. Following intubation, the tracheal tube completely bypasses these anatomic barriers and creates a direct conduit for bacteria to be aspirated and reach lower airways. Cough is one of the most efficient mechanisms to prevent further translocation of pathogens that may have gained access into airways. In the healthy human, cough begins with an inspiratory effort followed by a forced expiratory effort against a closed glottis and ultimately, opening of the glottis to generate rapid expiratory airflow. Tracheal intubation prevents closure of the glottis, hence it fully hinders cough; moreover, intubated patients are often sedated and unable to generate high expiratory flows. Mucociliary clearance is the primary innate airway defense mechanism to clear pathogens. In young, healthy nonsmokers, the mucociliary velocity ranges between 10 and 15 mm/min. Studies in animals have consistently shown that inflation of the tracheal tube cuff within the trachea lowers mucociliary velocity by 37% within an hour, and 52% after 2 hours.⁶⁴ Clinical studies⁶⁵ in critically ill, tracheally intubated patients have confirmed those results and found that mucociliary velocity is very low (0.8-1.4 mm/min); lower mucociliary clearance has been associated with higher risks for pulmonary complications.

Although many ICU patients develop tracheal bacterial colonization during the course of mechanical ventilation, only a small proportion subsequently develop VAP. As previously mentioned, the daily hazard rate for developing VAP is higher during the first days of mechanical ventilation.⁸ Investigators have found that a temporary immunoparalysis can be found early in the course of the critical illness and admission to the ICU.⁶⁶ In particular, researchers have focused on assessing human leukocyte antigen DR (HLA-DR) expression on peripheral monocytes as a marker of immune function.^67–69 According to the rationale that impaired immune function may predispose to the development of VAP, low levels of HLA-DR expression have been found in patients who subsequently developed nosocomial pneumonia.⁷⁰

Etiologic Agents

Nosocomial pneumonia may be caused by a variety of pathogens and, in many patients, more than one pathogen may be isolated. Microorganisms responsible for nosocomial pneumonia differ according to the ICU population, the duration of hospital and ICU stays, and the specific diagnostic method(s) used. VAP is commonly caused by aerobic gram-negative bacilli such as P. aeruginosa, Escherichia coli, Klebsiella pneumoniae, or Acinetobacter species, while S. aureus is the predominant isolated gram-positive pathogen.^1,71–74 Data from 7087 infected patients (63.5% with respiratory tract infection) from the Extended Prevalence of Infection in Intensive Care (EPIC II) study⁷⁵ have confirmed that Pseudomonas spp. and S. aureus are the most common isolated pathogens in intensive care units.

Few studies have assessed whether the pathogens that cause pneumonia in ventilated patients differ from those in patients who are not mechanically ventilated. Weber et al.⁷⁶ evaluated 158,519 patients admitted to a single center over a 4-year period and identified 327 episodes of VAP and 261 episodes of nosocomial pneumonia. The infecting flora in ventilated patients mostly included MRSA (17.75%) and gram-negative bacilli such as P. aeruginosa (59.0%), Stenotrophomonas maltophilia (17.50%), and Acinetobacter species (6.75%). Similarly, in 20.37% of the patients not requiring mechanical ventilation, MRSA was identified, while a lower incidence of nosocomial pneumonia due to P. aeruginosa, Acinetobacter spp., and S. maltophilia was found. Nevertheless, the overall frequency of infection with these gram-negative pathogens was sufficiently high to warrant the use of empirical therapy likely to be active against them.

The high rate of polymicrobial infection in VAP has been shown repeatedly. Combes and colleagues⁷⁷ studied 124 ICU patients, of whom 65 (52%) had monomicrobial VAP and 59 had (48%) polymicrobial VAP. In most patients, two different bacteria were isolated (42 patients, 34%); however, up to four different bacteria coexisted in 7 patients (6%). Interestingly, no differences were detected in mortality rate at 30 days between patients with polymicrobial or monomicrobial infection. A study by Teixeira et al.⁷⁸ investigated risk factors for inadequate empirical antimicrobial therapy in 151 ICU patients and found that 69 (45.7%) patients with a clinical diagnosis of VAP received inadequate empirical antimicrobial treatment. Multiple logistic regression analysis revealed that inadequate antimicrobial treatment was associated with polymicrobial VAP (OR, 3.67; 95% confidence interval [CI], 1.21-11.12; P = 0.02), and importantly, inadequate antimicrobial treatment was associated with higher mortality for patients with VAP.

Underlying diseases may predispose patients to infection with specific organisms, Patients with chronic obstructive pulmonary disease (COPD), for example, are at increased risk for Haemophilus influenzae, Moraxella catarrhalis, P. aeruginosa, or Streptococcus pneumoniae infections.^79,80 Patients with acute respiratory distress syndrome (ARDS) are at higher risk for developing VAP caused by S. aureus, P. aeruginosa, and Acinetobacter baumannii, and often in these patients, VAP is caused by multiple pathogens.^81,82 Finally, trauma and neurologic patients are at increased risk for S. aureus, Haemophilus, and S. pneumoniae infections.^83–85

It is important to identify MDR pathogens in order to guide appropriate antibiotic treatment. Causative pathogens of VAP that are potentially multiresistant are P. aeruginosa, MRSA, Acinetobacter spp., S. maltophilia, Burkholderia cepacia, and extended-spectrum β-lactamase (ESBL+) Klebsiella pneumonia. Conversely, S. pneumoniae, H. influenzae, methicillin-sensitive S. aureus (MSSA), and antibiotic-sensitive Enterobacteriaceae are not considered MDR pathogens. Patients at risk of being colonized by MDR pathogens are extremely heterogeneous, often present several comorbidities, and many receive antibiotics prior to and during the course of their hospitalization. Therefore, it is extremely challenging to accurately define risk factors for carrying MDR pathogens. Langer et al.³ tried to better classify patients who develop VAP, in order to provide data for guiding empirical antibiotic treatment. They compared early-onset and late-onset pneumonia and found that early-onset pneumonia is rarely caused by MDR pathogens, is less severe, and is associated with better outcome. However, recent reports are challenging such conclusions and demonstrate no association between MDR pathogens and time of onset of pneumonia.^86,87 Those data suggest the need for additional studies to accurately identify risk factors for harboring MDR pathogens, rather than risk stratification based on nonspecific factors such as severity of pneumonia and time of onset. The incidence of MDR pathogens is also closely linked to local factors and varies widely from one institution to another. Consequently, each ICU has to continuously collect accurate epidemiologic data. Rello et al.⁸⁸ analyzed variations of VAP etiology among three Spanish ICUs and compared them with data collected in Paris. The authors concluded that VAP pathogens varied widely among the four clinical centers, with marked differences in the microorganisms isolated from VAP episodes in Spanish centers as compared with the French site. Therefore, clinicians must be aware of the common microorganisms associated with both early-onset and late-onset VAP in their own hospitals to avoid the administration of inadequate initial antimicrobial therapy.

Legionella pneumophila as a cause of nosocomial pneumonia should be considered, particularly in immunocompromised patients.⁸⁹ Often the source of legionellosis outbreaks within the hospital is a water system that has become colonized by the microorganism.⁹⁰

The role of anaerobes in the pathogenesis of VAP requires further assessment, since the primary mechanism for VAP development is through aspiration of oropharyngeal contents, and the oropharynx is highly colonized by anaerobes. Robert et al.⁹¹ studied 26 mechanically ventilated patients and found that 15 patients became colonized with 28 different anaerobic strains. Similarly Dore et al.⁹² found anaerobic bacteria in 30 (23%) of 130 patients diagnosed with VAP, but always in association with aerobic pathogens. Importantly, empirical antibiotic therapy active against anaerobic bacteria appears to improve short-term outcomes in patients with VAP.⁹³ Nevertheless, several authors^94,95 were unable to reproduce those data, and the role of anaerobes in VAP is still considered controversial. In particular, Marik et al.⁹⁵ studied microbiology of 185 episodes of suspected VAP through blind protected specimen brush sampling and mini-BAL and were unable to identify anaerobes as the causative pathogens of VAP.

Rarely, the causative organism of VAP is a fungus. Candida spp. and Aspergillus fumigatus are the most common isolated fungi, predominately in immunocompromised patients. In mechanically ventilated patients, the clinical significance of respiratory tract colonization by Candida is controversial. In a retrospective analysis of 639 patients from a Canadian study or VAP, 114 patients had Candida colonization of the respiratory tract.⁹⁶ Interestingly, patients with Candida colonization had a significant increase in hospital mortality (34% versus 21% in patients without Candida colonization, P = 0.003). However, it is still unclear whether Candida colonization is associated with or responsible for worse outcomes. Moreover, a recent report showed that in ICU patients, isolation of Candida species in respiratory samples demonstrates only colonization rather than Candida pneumonia.⁹⁷

It is commonly reported that VAP is infrequently due to viruses; however, it should also be acknowledged that patients with clinical suspicion of VAP are rarely screened for viruses. Daubin et al.⁹⁸ studied 139 patients mechanically ventilated for more than 48 hours, of which 39 (28%) developed VAP. Although P. aeruginosa and MRSA still accounted for most of the VAP cases, herpes simplex virus type 1 was found in 12 cases of VAP and cytomegalovirus (CMV) in 1 case. Several studies have reported a high incidence of active CMV infection in mechanically ventilated patients.^99–101 Recently, Chiche et al.¹⁰² studied 242 immunocompetent ICU patients and found active CMV infection in 39 (16%). At 28 days, only 15% of the patients with active CMV infection were weaned and alive, in comparison to 52% of patients free of CMV infection (P < 0.001).

General Prophylactic Measures

Maintaining high levels of education among ICU personnel relating to VAP pathophysiology and preventive strategies can be effective in reducing incidence of this problem.^107–110 Respiratory care practitioners and nurses should be the primary recipients of education programs, and frequent performance feedback and compliance assessment should be undertaken.¹¹¹ Interestingly, Needleman et al.¹¹² studied administrative data from 799 hospitals in 11 states (covering 5,075,969 discharges of medical patients and 1,104,659 discharges of surgical patients) and found that a higher proportion of hours of care per day provided by registered nurses, compared to licensed practical nurses and nurses’ aides, were associated with lower incidence of pneumonia.

Adherence to simple infection-control measures such as alcohol-based hand disinfection effectively reduces cross-transmission of pathogens and incidence of VAP.¹¹³ The World Health Organization has endorsed hand hygiene as the single most important element of strategies to prevent healthcare-associated infections.¹¹⁴ Overall, most studies conducted in ICUs have shown consistent results and temporal association between implementation of alcohol-based hand hygiene and reduction of nosocomial infections.^115–117

Kollef et al.¹¹⁸ demonstrated that patient transport outside the ICU was associated with increased risks for VAP. Clinicians and nursing staff should carefully carry out transport of intubated patients. In particular, the internal pressure of the ETT cuff should be always kept within the recommended range, particularly when the patient is expected to be maintained supine during diagnostic or therapeutic procedures; ventilator circuits should be carefully manipulated in order to prevent aspiration of colonized fluids from within the circuit.

Daily interruption or lightening of sedation to avoid constant impairment of respiratory defenses, as well as avoidance of paralytic agents, is highly recommended. It is well acknowledged that prolonged tracheal intubation is associated with VAP.⁸ A report by Kress et al.¹¹⁹ was recently confirmed by Schweickert et al.,¹²⁰ who studied 128 mechanically ventilated patients randomized to continuous infusions of sedation with or without daily interruption. The authors demonstrated reduction of duration of mechanical ventilation and length of stay in the ICU when patients were allowed to wake up daily. Moreover, a trial by Schweickert et al.¹²¹ has shown that early physical and occupational therapy during critical illness is associated with more ventilator-free days. In a study by Strøm et al.,¹²² 140 critically ill adult patients expected to be intubated for more than 24 hours were randomized to receive no sedation or sedation with daily interruption until awake. Both groups were treated with bolus doses of morphine. In that study, patients receiving no sedation had significantly more days without ventilation and shorter stay in the ICU. Results from these clinical trials are challenging standard sedation protocols for intubated patients and hold promise for reducing length of stay on mechanical ventilation and, ultimately, risks for VAP.

There is evidence of shorter length of mechanical ventilation, reduced rate of failed extubation, and decreased incidence of VAP when protocol-driven weaning from the ventilator is implemented.^123,124 Marelich et al.¹²³ randomized 385 patients to receive either a protocol-driven weaning procedure or standard care and found that duration of mechanical ventilation was decreased from a median of 124 hours for the control group to 68 hours in the protocol-driven weaning group (P = 0.0001). Moreover, a trend toward less VAP was found in the treatment group (P = 0.061).

Noninvasive Ventilation

Tracheal intubation and mechanical ventilation account for the main risk for nosocomial pneumonia and therefore should be avoided whenever possible. Noninvasive ventilation (NIV) is an attractive alternative for patients with acute exacerbations of COPD or acute hypoxemic respiratory failure and for some immunocompromised patients with pulmonary infiltrates and respiratory failure.^125–128 NIV can also be safely used to facilitate early extubation and avoid continued invasive weaning. A meta-analysis¹²⁹ evaluated 12 trials enrolling 530 participants, mostly with chronic obstructive pulmonary disease, and confirmed that noninvasive weaning is significantly associated with reduced mortality, VAP, and length of stay in the ICU and hospital. Another report¹³⁰ emphasized the role of NIV in preventing reintubation in recently extubated patients at risk for relapse and respiratory failure. Kohlenberg et al.¹³¹ pooled data of 400 ICUs in Germany and found mean pneumonia incidence of 1.58 and 5.44 cases per 1000 ventilator days for NIV and invasive mechanical ventilation, respectively. Therefore, when indicated, NIV should be attempted to avoid tracheal intubation and reduce overall duration of tracheal intubation.

Tracheal Tube Cuff

Several strategies have been applied to improve the design of tracheal tubes and reduce the likelihood of aspiration of pathogen-laden secretions across the cuff. Novel ETT cuffs made of new materials such as polyurethane,¹⁶ silicone,¹³² and latex^133,134 have been developed and tested in laboratory and clinical trials. Particularly, the polyurethane cuff has a thickness of 5 to 10 µm in comparison to 50 µm of PVC cuffs; hence, upon inflation, smaller folds form, and aspiration of secretions above the cuff can be prevented or reduced. Lorente et al.¹³⁵ compared a standard ETT to an ETT incorporating an ultrathin polyurethane cuff and intermittent aspiration of subglottic secretions and found a reduction of incidence of VAP from 22.1% to 7.9% between the standard and new tubes, respectively (P = 0.001). A single-center study by Miller et al.¹³⁶ tested the use of a polyurethane-cuff ETT versus a standard PVC cuff ETT, with a before and after design, and found that VAP rates decreased from 5.3 per 1000 ventilator days before the use of the polyurethane-cuffed ETT to 2.8 per 1000 ventilator days during the intervention year (P = 0.0138). The polyurethane-cuffed ETT has also shown benefits in reducing early postoperative pneumonia in cardiac surgical patients. Poelaert et al.¹³⁷ studied 134 cardiothoracic surgery patients and demonstrated that the incidence of early postoperative pneumonia was significantly lower in the polyurethane group than in the polyvinyl chloride group (23% versus 42%, P < 0.03). Silicone^132,138 and latex^133,134,139 cuffs are low-volume, low-pressure cuffs and are promising alternatives to PVC cuffs. Upon inflation, folds are never formed, yet compliance of those materials is extremely high; thus they allow reliable control of the pressure exerted against the trachea. In a clinical trial on patients undergoing anesthesia or admitted to the ICU, Young et al.¹³² demonstrated high effectiveness of a silicone cuff in reducing pulmonary aspiration.

The shape of the cuff plays an important role in prevention of aspiration.^133,140 In comparison to standard cuffs with cylindrical shapes, cuffs designed with a smooth, tapering shape allow elimination of folds for a full circumference of the trachea/cuff contact zone, irrespective of the cuff material.

It is important to maintain the internal pressure of ETT cuff pressure between 25 and 30 cm H₂O, particularly when no positive end-expiratory pressure (PEEP) is applied; this serves to prevent either macroleakage of contaminated secretions into the lower airways or tracheal injury. A recent study¹⁴¹ demonstrated that frequently the ETT cuff was deflated or hyperinflated using standard management.

Ventilatory settings may play a role in pathogenesis of VAP. In particular, PEEP may decrease the incidence of VAP by counteracting hydrostatic pressure exerted by oropharyngeal secretions above the ETT cuff, hence reducing pulmonary aspiration.¹⁴² A recent report¹⁴³ assessed effects of 5 to 8 cm H₂O in normoxemic ventilated patients and showed reduction of the rate of VAP (PEEP group 9.4%, control patients 25.4%, relative risk, 0.37; 95% CI = 0.15-0.84; P = 0.017).

Tracheal Tubes Coated with Antimicrobial Agents

Coating the ETT with antimicrobial agents such as silver is a promising strategy to prevent biofilm formation within its internal surface and VAP. Olson et al.¹⁴⁴ examined a silver-coated tracheal tube in comparison to standard tube in intubated dogs. The dogs were challenged with P. aeruginosa into the oropharynx. Using the new tube, the investigators were able to postpone colonization of the ETT inner surface (3.2 ± 0.8 versus 1.8 ± 0.4 days; P = 0.02) and reduce bacterial burden in the lung parenchyma (4.8 ± 0.8 versus 5.4 ± 9 log colony-forming unit [CFU]/g lung tissue; P = 0.01). Similarly, Berra et al.²⁰ randomized 16 sheep to be intubated with a standard ETT or a silver sulfadiazine/chlorhexidine–coated ETT. After 24 hours of mechanical ventilation, all eight ETTs and ventilatory circuits in the control group were heavily colonized, and biofilm was found within the ETT. Pathogenic bacteria colonized the trachea and the lungs in five of eight sheep (up to 10⁹ CFU/g). In the study group, seven of eight ETTs and ventilator circuits showed no growth and no biofilm; moreover, there was no bacterial growth in the lungs and bronchi, except for one bronchus in one sheep. Interestingly, the efficacy of silver-based coatings seems to decrease over time. Indeed, animal studies consistently reported no colonization and biofilm formation after 24 hours of mechanical ventilation. However, heavy ETT colonization was reported when studies were prolonged after 72 hours. To date, only one laboratory study¹⁴⁵ has reported the absence of ETT colonization and biofilm formation following up to 168 hours of mechanical ventilation. In that study, the authors used ETTs internally coated with silver-sulfadiazine that were regularly cleaned with a novel concentric inflatable silicone device, the Mucus Shaver,¹⁴⁶ devised to keep the ETT lumen free of mucus. The North American Silver-Coated Endotracheal Tube (NASCENT) randomized trial⁷² studied 1509 patients expected to require mechanical ventilation for more than 24 hours and randomized to be intubated with either a silver-coated or a conventional tube. The silver-coated ETT was associated with a lower incidence of microbiologically confirmed VAP (37/766 (4.8%) versus 56/743 (7.5%); P = 0.03), for a relative risk reduction of 35.9%. More importantly, the silver-coated tube had its greatest impact during the first 10 days of tracheal intubation. A retrospective cohort analysis by Afessa et al.,¹⁴⁷ based on the NASCENT study, showed that the silver-coated ETT was associated with reduced mortality in patients with VAP (silver ETT versus control, 5/37 [14%] versus 20/56 [36%], P = 0.03), but mortality was higher in those without VAP (silver versus control, 228/729 [31%] versus 178/687 [26%], P = 0.03). In conclusion, there is promising evidence that ETTs coated with antimicrobial agents could reduce incidence of VAP. Nevertheless, clinicians should carefully consider benefits and limitations of these new ETTs and properly direct the use of silver-coated tubes to patients expected to be ventilated for longer periods of time and with higher associated risks for nosocomial pneumonia.

Shorr et al.¹⁴⁸ analyzed cost-effectiveness of the silver-coated ETT as a preventive tool for VAP. Based on the NASCENT trial, the authors assumed a reduction in the relative risk of VAP from 35.9% to 24%. Assuming marginal VAP costs of $16,620 and costs of $90.00 for coated and $2.00 for uncoated ETTs, the authors found that savings per case of VAP prevented were $12,840.

Aspiration of Subglottic Secretions

Aspiration of colonized subglottic secretions through dedicated ETTs reduces hydrostatic pressure exerted above the cuff and potentially prevents macroleakage across the cuff. A meta-analysis¹⁴⁹ comprising data from five studies and 896 patients has shown that subglottic secretion drainage reduced the incidence of VAP by nearly half (risk ratio [RR] = 0.51; 95% CI, 0.37-0.71), primarily by reducing early-onset pneumonia. Likewise, in the trial by Bouza et al.¹⁵⁰ of 690 patients undergoing major cardiac surgery and on mechanical ventilation for more than 48 hours, the use of ETT tubes with aspiration of subglottic secretions was able to reduce incidence of VAP, median length of ICU stay, and antibiotic use and led to overall cost savings. In the multicenter trial by Lacherade et al.,¹⁵¹ 333 patients were randomized to be intubated with either an ETT that allowed drainage of subglottic secretions or a standard ETT. Microbiologically confirmed VAP occurred in 14.8% of the patients in the treatment group, compared to 25.6% of the patients intubated with standard tube (P = 0.02). Importantly, this was the first trial reporting efficacy of subglottic secretions aspiration in reducing both early- and late-onset VAP in comparison to earlier studies (late onset VAP in 18.6% of patients in treatment group versus 33.0% of the patients in control group, P = 0.01). In more recent studies, aspiration of subglottic secretions was performed intermittently.^135,151 The benefits of intermittent subglottic suction were similar to studies in which continuous aspiration of subglottic secretions was used. Hence, based on current evidence, intermittent aspiration (every 4-6 hours) is advisable to avoid potential risks for tracheal injury using continuous aspiration.¹⁵²

Tracheostomy

Tracheostomized patients present the same risks for aspiration of pathogen-laden secretions pooled above the cuff^153,154 as do orotracheally intubated patients. An observational study by Ibrahim et al.¹⁵⁵ on 880 mechanically ventilated patients demonstrated an association between tracheostomy and higher incidence of VAP (adjusted OR, 6.71; 95% CI, 3.91-11.50; P < 0.001). Conversely, a case-control study by Nseir et al.¹⁵⁶ on 354 patients mechanically ventilated for more than 7 days showed a lower rate of nosocomial pneumonia associated with tracheostomy (4.8 versus 9.2 episodes per 1000 ventilator days in patients with or without tracheostomy, respectively). Meta-analyses have assessed outcomes of early versus late tracheostomy.^157–159 Unfortunately, all included studies were highly heterogeneous owing to differences in studied populations and no clear classification of “early” versus “late” tracheostomy. Nevertheless, the meta-analyses failed to demonstrate benefits of early tracheostomy on reduction of VAP incidence. To date, the latest multicenter randomized trial¹⁶⁰ enrolled 419 mechanically ventilated patients to undergo either early or late tracheostomy. Patients in the early tracheostomy group underwent tracheostomy after a mean of 7 days, whereas patients in the late tracheostomy group underwent tracheostomy after a mean of 14 days. Although the authors found shorter length of mechanical ventilation and ICU stay with early tracheostomy, they were able to demonstrate only a trend toward a lower incidence of pneumonia and no difference in survival. Clinicians should consider that early tracheostomy may offer several benefits for mechanically ventilated patients: improved patient comfort, ability to communicate, capability for oral feeding, less need for sedation and analgesia, and reduced airway resistance in comparison to standard ETTs, which could be extremely important during the weaning period to shorten the duration of tracheal intubation.

Ventilator Circuit Management

Decreased frequency of ventilator circuit changes, replacement of heated humidifiers by heat and moisture exchangers, decreased frequency of heat and moisture exchanger changes, and closed suctioning systems have been tested as measures for preventing VAP. Results from clinical trials in adults^161–167 and meta-analyses^168,169 yield consistent evidence that routine change of the ventilator circuit does not decrease risks for VAP and costs. Therefore, circuits should not be changed unless the circuit is soiled or damaged. Importantly, inadvertent flushing of the contaminated condensate into the lower airways or nebulizers should be always avoided by careful emptying of ventilator circuits and water traps.^170–172

Two meta-analyses assessed the effects of heated humidifiers (HH) and heat and moisture exchangers (HME) on prevention of nosocomial pneumonia. Kola et al.¹⁷³ pooled data from nine clinical trials on 1378 patients and found that the use of HME decreased the rate of VAP (relative risk = 0.7; 95% CI = 0.50-0.94). Conversely, a meta-analysis by Siempos et al.¹⁷⁴ including 13 studies on 2580 patients found no difference between HME and HH patients in the prevention of VAP and secondary outcomes such as ICU mortality, length ICU stay, duration of mechanical ventilation, or episodes of airway occlusion. Hence, to date there are no consistent data showing reduction in the incidence of VAP and better outcome using either HME or HH. Based on the ongoing controversy, neither humidification strategy can be recommended as a pneumonia prevention tool. However, it is rational to deliver inspiratory gases at body temperature or slightly below and at the highest relative humidity in order to prevent loss of heat and moisture from the airways and, more importantly, change in rheologic properties of secretions and impairment of mucociliary clearance.¹⁷⁵ Therefore, the use of HH is indicated particularly in patients with hypothermia, prolonged mechanical ventilation, thick secretions, and chronic respiratory disorders. Finally, studies that have evaluated the effect of less frequent changes of heat and moisture exchangers on the development of VAP have found no increased risks.^176–180 However, it is important to emphasize that when HMEs are used for prolonged periods of time, the technical performance of the devices should be periodically checked.

Closed tracheal suctioning systems have been introduced in clinical settings to avoid adverse events associated with ventilator disconnection during open tracheal suctioning and exogenous contamination of suction catheters entering the ETT. Three meta-analyses^181–183 have compared the closed tracheal suction system to the open tracheal suction system in mechanically ventilated patients and found no benefits on VAP prevention. One meta-analysis¹⁸² evaluated nine randomized trials comprising 1292 patients and found no difference in the incidence of VAP between patients suctioned with closed or open systems (OR = 0.96, 95% CI 0.72-1.28). Moreover, data pooled from four of these nine studies showed a higher incidence of respiratory tract colonization in the group managed with a closed system (OR = 2.88, 95% CI 1.50-5.52).

The use of a saline solution instilled into the tracheal tube before tracheal suctioning remains controversial. Caruso et al.¹⁸⁴ published a report on 262 patients randomized to receive either isotonic saline instillation before tracheal suctioning or no treatment. The authors found a lower incidence of microbiologically proven VAP (saline instillation versus no treatment: 23.5% versus 10.8%; P = 0.008), and no significant differences were found in secondary outcomes such as the incidence of ETT obstruction, pulmonary and lobar atelectasis, mortality, and duration of mechanical ventilation and ICU stay. Theoretically, in sedated patients in the semirecumbent position and with the internal surface of the ETT highly colonized, saline instillation may increase risks for translocation of pathogens into the airways, so current limited evidence suggests that routine saline instillation should not be recommended.

Body Position

Early studies clearly demonstrated that intubated patients are at higher risk for gastropulmonary aspiration when placed in the supine position (0 degrees) as compared with a semirecumbent position (45 degrees).^185,186 One randomized trial¹⁸⁷ demonstrated a reduction in the incidence of VAP in patients positioned in the semirecumbent position compared with patients treated completely supine. Moreover, the trial confirmed increased risk for VAP in patients enterally fed. A later randomized trial¹⁸⁸ studied the feasibility of maintaining the head of the bed oriented 45 degrees during mechanical ventilation. This study found that patients were positioned on average only 28 degrees above horizontal, and no difference on VAP incidence was found. Thus, as strongly suggested by the American¹ and European¹⁸⁹ guidelines, intubated patients should be preferentially kept in the semirecumbent position (30-45 degrees) rather than supine (0 degrees) to prevent aspiration, especially when receiving enteral feeding.

Laboratory reports^190,191 challenge the role of the semirecumbent position in patients with oropharyngeal colonization due to tracheal intubation or during the course of mechanical ventilation. Theoretically, in such patients a tracheal orientation above horizontal, as in the semirecumbent position, might facilitate aspiration across the tracheal tube cuff. Laboratory studies in animals^190,191 consistently found that tracheal orientation and body position to avoid aspiration across the ETT cuff enhance mucus drainage and decrease risks for VAP; however, such results need to be confirmed in humans.

Rotating Bed

Normal healthy people change body position, even during sleep, every few minutes. Conversely, when critically ill patients are tracheally intubated and on mechanical ventilation, they are maintained in the supine, semirecumbent position for days with few or no changes in body position. Several ICU beds allow rotation of patients in the longitudinal axis from one lateral position to the other and seem to reduce extravascular lung water, improve ventilation-perfusion ratio, and enhance mobilization of airway secretions.¹⁹² Several studies have evaluated the effects of rotational therapy on VAP; however, most of these studies present limitations. For example, most studies used a clinical diagnosis of pneumonia, lack of standardization of VAP preventive measures, and included heterogeneity on the duration and type of rotation therapy. Three meta-analyses^193–195 showed significant reduction in the incidence of VAP in patients undergoing rotation therapy, but they consistently failed to show beneficial effects on secondary outcomes such as duration of mechanical ventilation, length of stay, and mortality. An article by Staudinger et al.¹⁹⁶ studied the effects of continuous lateral rotation therapy on the incidence of microbiologically confirmed VAP in 3 medical ICUs and found an incidence of 11% in the rotation group and 23% in the control group (P = 0.048), respectively. The authors also found that the duration of ventilation (8 ± 5 versus 14 ± 23 days, P = 0.02) and length of stay (25 ± 22 days versus 39 ± 45 days, P = 0.01) were significantly shorter in the rotation group. In conclusion, in patients at higher risk for prolonged immobilization and respiratory infection, continuous lateral rotation therapy should be considered as a feasible method exerting additive effects to other preventive measures for VAP.

Stress Ulcer Prophylaxis and Enteral Feeding

There is clear evidence that in intubated and mechanically ventilated patients, the stomach is often colonized by pathogens. Early studies showed that in tracheally intubated patients, gastric pH higher than 4 was consistently associated with gastric colonization.^49,50 Alkalinization of gastric contents due to drugs for stress ulcer prophylaxis and continuous enteral nutrition were the main risk factors for gastric colonization.

In the ICU, stress ulcer prophylaxis is usually achieved with either sucralfate, histamine type 2 blockers (H₂ blockers), or proton pump inhibitors (PPI). Sucralfate is the only treatment that potentially prevents stress GI ulceration without raising gastric pH. Several randomized studies have compared the effects of drugs that alkalinize gastric contents to sucralfate.^{52–57,197,198} Two studies compared either H₂ blockers⁵¹ or sucralfate¹⁹⁹ to placebo. Nine studies assessed gastric pH and colonization and consistently found lower pH and gastric colonization with regular use of sucralfate. Conversely, studies conducted by Bonten et al.⁵⁴ and Thomason et al.⁵⁶ found gastric luminal alkalinization and colonization, irrespective of the use of sucralfate. Finally, Eddleston et al.¹⁹⁹ found no differences in gastric colonization when sucralfate was compared to placebo.

Randomized trials reported inconsistent results regarding stress ulcer prophylaxis and the incidence of VAP. In particular, early studies found a higher incidence of pneumonia in patients with alkalinized gastric contents,^51–53 while more recent studies have not found such an association.^54,55,124 Cook et al.⁵⁵ studied 1200 patients randomized to receive either H₂ blockers or sucralfate for stress ulcer prophylaxis. The authors found a higher risk for GI bleeding using sucralfate and no significant difference in VAP incidence, 19.1% and 16.2% in patients treated with H₂ blockers or sucralfate, respectively. A recent meta-analysis²⁰⁰ on the efficacy and safety of PPIs in comparison with H₂ blockers pooled data by seven randomized controlled trials on 936 patients and found no difference between PPI and H₂-blocker therapy on the risk for pneumonia and ICU mortality. In conclusion, GI bleeding is a serious complication in critically ill patients at high risk for stress ulcers (i.e., patients with coagulopathy, need for prolonged mechanical ventilation, and history of GI ulceration or bleeding). The actual risk for VAP is unknown when accurate methods of enteral feeding (i.e., avoiding large gastric residual volumes) or other preventive measures are used in combination with stress ulcer prophylaxis. Therefore, clinicians must weigh the potential benefit of sucralfate (with potentially less VAP and more GI bleeding) versus H₂ blockers/PPI (with potentially more VAP and less GI bleeding) and probably limit stress ulcer prophylaxis to high-risk patients.

Enteral nutrition has been considered a risk factor for the development of nosocomial pneumonia, mainly because of increased risks for alkalinization of gastric contents, gastro-esophageal reflux, and gastropulmonary aspiration. However, its alternative, parenteral nutrition, is associated with higher risks for catheter-related infections, complications of line insertions, higher costs, and loss of intestinal villous architecture, which may facilitate enteral microbial translocation. A large meta-analysis²⁰¹ of 15 studies comprising 753 patients admitted to the ICU for trauma, head injury, burns, and abdominal surgery found a significantly lower incidence of infections and a reduced length of hospital stay associated with early enteral feeding. Conversely, studies in medical ICU patients have proven higher risk for VAP with early enteral feeding.^202,203 Nonetheless, in a study by Artinian et al.²⁰² the increased risk of VAP associated with early enteral feeding did not translate into an increased risk of death. Therefore, in medical ICU patients, the benefits of early nutrition should be balanced with associated increased risks for VAP.

A large number of studies have evaluated the risks for ICU-acquired pneumonia in patients randomized to either gastric or postpyloric feeding. Theoretically, many ICU patients present impaired gastric emptying; hence placement of the feeding tube beyond the pylorus has the potential to achieve nutrition goals without increased risks for gastropulmonary aspiration. A meta-analysis by Heyland et al.²⁰⁴ found that small-bowel feedings were associated with a lower incidence of pneumonia (RR, 0.77; 95% CI, 0.60-1.0); conversely, a meta-analysis by Ho et al.²⁰⁵ found no significant benefit on the risk of diarrhea, length of ICU stay, mortality, or risk of pneumonia. Therefore, ICU physicians should preferentially indicate postpyloric feeding in critically ill patients who have impaired gastric emptying.

Modulation of Oropharyngeal and Gastrointestinal Colonization

One of the most important factors in the pathogenesis of nosocomial pneumonia is the early shift of oral flora following tracheal intubation into a predominance of aerobic gram-negative pathogens. Therefore, extensive efforts have been devoted to modulating oropharyngeal flora of ICU patients and reducing the risks for aspiration of pathogens. Several antiseptics for oropharyngeal decontamination have been evaluated: chlorhexidine gluconate, iseganan, or povidone iodine; chlorhexidine has been the focus of most research. Chlorhexidine is a cationic chlorophenyl bis-biguanide antiseptic that has long been used as an inhibitor of dental plaque formation and gingivitis. Meta-analyses^206–208 of studies assessing the benefits of chlorhexidine on reduction of VAP have shown good results, particularly in cardiothoracic ICU patients. Results in noncardiac ICU populations are more uncertain. Most of the aforementioned studies used chlorhexidine concentrations of 0.12% and 0.2%. However, recent studies in general ICU patients have demonstrated significant reductions in VAP rates when chlorhexidine concentration was increased to 2%.^209,210 Therefore, oral decontamination with chlorhexidine should be routinely used, particularly in cardiothoracic patients. The usefulness of chlorhexidine as VAP preventive strategy in other ICU populations still requires more evidence before being put into general practice, but the use of higher chlorhexidine concentrations has showed promising results.

Since the original studies published by Stoutenbeek and coworkers,^211,212 selective decontamination of the digestive tract (SDD) has been used as a preventive strategy for nosocomial pneumonia for almost 3 decades. SDD comprises a combination of nonabsorbable antibiotics against gram-negative pathogens (i.e., tobramycin, polymyxin E) plus either amphotericin B or nystatin administered into the GI tract in order to prevent oropharyngeal and gastric colonization with aerobic gram-negative bacilli and Candida spp., while preserving the anaerobic flora. Some regimens include a short course of systemic antibiotics (most commonly cefotaxime) in addition to nonabsorbable GI antibiotics. Randomized clinical trials^213–215 and meta-analyses^216–218 confirm results of earlier studies and suggest that SDD confers protection against pneumonia. Interestingly, SDD is the only preventive strategy for VAP that has shown reduction of mortality rates. A clinical trial by de Smet et al.²¹⁴ evaluated the effectiveness of SDD and selective oropharyngeal decontamination in a crossover study using cluster randomization in 13 ICUs and applied SDD, oropharyngeal decontamination, or standard care in random order for 6 months. SDD consisted of 4 days of intravenous cefotaxime and topical application of tobramycin, colistin, and amphotericin B in the oropharynx and stomach. Oropharyngeal decontamination consisted of oropharyngeal application only of the same antibiotics. The authors enrolled a total of 5939 patients. Post hoc analysis in a random-effects logistic-regression model found that the odds ratio for death at day 28 in the oropharyngeal decontamination and SDD groups, as compared with the standard-care group, were 0.86 (95% CI, 0.74-0.99, P = 0.045) and 0.83 (95% CI, 0.72-0.97, P = 0.02), respectively.

It is important to acknowledge that prophylactic use of antibiotics to modulate GI flora may potentially increase risks for antibiotic resistance, and results from randomized clinical trials still remain controversial; moreover, standard SDD is aimed at preventing overgrowth of aerobic gram-negative bacteria, yet it could increase colonization by gram-positive bacteria such as MRSA and Enterococcus spp. A large Dutch randomized controlled trial²¹³ demonstrated that carriage of multiresistant gram-negative bacteria was actually reduced in patients who had SDD, compared with controls (RR, 0.61; 95% CI, 0.46-0.81). Overall, emergence of resistance of gram-negative bacteria was consistently a rare event with SDD. Recently, Bonten’s group reported data²¹⁹ on bacterial ecology in 13 ICUs that participated in their previous study of SDD.²¹⁴ Rectal and respiratory samples were analyzed once monthly in all ICU patients and showed that during SDD, average proportions of patients with intestinal colonization with GNB resistant to either ceftazidime, tobramycin, or ciprofloxacin increased from 5%, 7%, and 7%, respectively to 15%, 13%, and 13% post intervention (P <0.05). During SDD/SOD, resistance levels in the respiratory tract were not more than 6% for all three antibiotics but increased gradually (for ceftazidime, P <0.05 for trend) during intervention and to levels of 10% or more for all three antibiotics post intervention (P <0.05).

Conclusions on the effect of SDD on VAP, based on the results of several meta-analyses and clinical trials, may be summarized as follows:

Early attempts at VAP prophylaxis using parenteral antibiotics were unsuccessful.²²² Only one study²²³ showed that a short course of cefuroxime upon emergent intubation and during 48 hours following intubation in patients with structural coma or severe burns was an effective prophylactic strategy to decrease the VAP rate. However, routine use of parenteral antibiotics is not recommended until more data become available.

Several clinical trials have attempted to modify GI and oropharyngeal growth of pathogens through the use of probiotics. Probiotics are microorganisms that can be administered either as individual strains or in various combinations. These microorganisms are often administered with nondigestible food ingredients that facilitate bacterial growth and/or activity (prebiotics); products containing both probiotics and prebiotics are called synbiotics. A meta-analysis²²⁴ on the effects of probiotics, pooling data from five randomized controlled trials in 689 patients, showed a lower incidence of VAP. The use of probiotics is a promising strategy for VAP; however, additional evidence is required before recommending its use in all mechanically ventilated patients, particularly owing to the heterogeneity of previous studies.

Evaluation of Diagnostic Strategies

Four randomized controlled trials^{250,261–263} have assessed the impact of diagnostic strategies on antibiotic use and outcome in patients with clinical suspicion of nosocomial pneumonia. In three small studies,^250,261,262 invasive diagnostic techniques resulted in a greater number of antibiotic changes than noninvasive techniques; however, no differences in mortality and morbidity were found when either invasive (PSB and/or BAL) or noninvasive (quantitative endotracheal aspirate cultures) techniques were used. By contrast, a larger trial²⁶³ showed a reduction in mortality, better Sequential Organ Failure Assessment (SOFA) score at follow-up, reduced use of antibiotics, and increased number of antibiotic-free days using invasive diagnostic techniques. This study was limited, however, by the use of qualitative cultures of tracheal aspirates, thereby limiting comparison with other clinical trials. A meta-analysis by Shorr et al.²⁶⁴ pooled data from these randomized studies on 628 patients and found that overall, an invasive approach did not alter mortality (OR, 0.89; 95% CI, 0.56-1.41). Invasive testing, though, affected antibiotic utilization (OR for change in antibiotic management after invasive sampling, 2.85; 95% CI, 1.45-5.59). Importantly, it should be realized that diagnostic cultures of pulmonary secretions after initiation of new antibiotic therapy in patients with suspicion of HAP can lead to a high number of false-negative results, irrespective of the sampling technique. In this clinical setting, a lower threshold should be used to define a positive quantitative result.^265,266 Nevertheless, it is strongly recommended that diagnostic sampling of the respiratory tract be obtained before starting any new antibiotic or changing previous antimicrobial therapy.

A clinical trial²⁶⁷ compared quantitative culture of BAL fluid and culture of endotracheal aspirate in critically ill patients with suspected VAP. This study was part of a larger 2-by-2 factorial design also comparing empirical antimicrobial monotherapy (a carbapenem) and combination therapy (a carbapenem plus a fluoroquinolone). A total of 740 patients in 28 ICUs in Canada and the United States were enrolled, and the authors found no difference in the 28-day mortality rate between the BAL group and the endotracheal aspiration group (18.9% and 18.4%, respectively; P = 0.94). The BAL group and the endotracheal aspiration group also had similar rates of targeted therapy (74.2% and 74.6%, respectively; P = 0.90), days alive without antibiotics (10.4 ± 7.5 and 10.6 ± 7.9; P = 0.86), and maximum organ-dysfunction scores (mean [±SD], 8.3 ± 3.6 and 8.6 ± 4.0; P = 0.26). The two groups did not differ significantly in the length of stay in the ICU or hospital. Unfortunately, at least 40% of the screened patients were excluded because they were at risk for colonization with Pseudomonas spp. or MRSA or were immunosuppressed. Therefore, translation of these findings into clinical practice is a major concern, because many ICU patients evaluated for suspected VAP fall into these categories.

Practical Implementation of a Diagnostic Strategy in Suspected Ventilator-Associated Pneumonia

In practice, the development of local clinical guidelines can combine both clinical and bacteriologic strategies (Table 67-2). The diagnostic protocol begins with clinical suspicion of nosocomial respiratory infection (Figure 67-4). In mechanically ventilated patients, the presence of an infiltrate on chest radiograph differentiates between the possible presence of pneumonia and tracheobronchitis. The next step is to sample the lower respiratory tract (see Table 67-2) to identify the causative microorganism. Sampling should be performed before initiation or change of antibiotic treatment, even though it should not delay the administration of antibiotic therapy, particularly for septic patients. Respiratory tract specimens can be obtained through expectoration, bronchial aspirate, BAL, or PSB. The latter two techniques can be performed with bronchoscopy or blindly. Several other samples should also be collected, as noted in Table 67-2. With clinical suspicion of pneumonia, CPIS²³⁵ should be calculated to improve objective assessment of the clinical parameters (see Table 67-1).

TABLE 67-2 Diagnostic Protocol to Combine Clinical and Bacteriologic Strategies for the Diagnosis of Ventilator-Associated Pneumonia

* Samples should be sent to the microbiology department, or if not available, maintained in refrigerator at 4°C (only respiratory samples) for a maximum of 1 hour for Gram staining, intracellular organism counting (only in BAL and mini-BAL), and quantitative cultures. The collection of lower respiratory secretion samples should not delay the initiation of empirical treatment in patients with severe sepsis.

** These techniques may be performed by bronchoscopy or blind procedures. Quantitative cultures are performed with the respiratory secretions obtained by BAS, BAL, or PBS. The cutoff count to diagnose pneumonia is the following: BAS 10⁶ CFU/mL; BAL 10⁴ CFU/mL, and PSB 10³ CFU/mL.

Figure 67-4 Clinical suspicion of nosocomial respiratory infection.

Likely Etiologic Microorganisms

The microorganisms most frequently isolated from the bronchial secretions of patients with VAP are S. aureus and P. aeruginosa, comprising around 50% of the isolates. These are followed, in order of frequency, by Enterobacteriaceae (E. coli, Klebsiella spp., Enterobacter spp., Citrobacter spp., Serratia spp., and Proteus spp.) representing 15%, nonfermentative gram-negative bacilli other than P. aeruginosa (Acinetobacter spp., Stenotrophomonas spp., and Burkholderia spp.) in 10%, and H. influenzae and S. pneumoniae (among others) in the remaining cases.¹⁰³

The microorganisms causing VAP generally come from the oropharyngeal flora of the patient. Underlying chronic diseases,²⁶⁸ specific risk factors, acute inflammatory processes, and factors specific to each hospital or ICU can facilitate abnormal bacterial colonization of the oropharynx and may predispose patients to infection with specific organisms.^1,269 Therefore, the selection of initial antimicrobial therapy must be tailored to the local prevalence of pathogens and antimicrobial patterns of resistance of each institution.^88,270 Healthy subjects rarely have significant colonization with gram-negative bacilli in the oropharynx, even after prolonged exposure to the hospital or ICU environment. Conversely, elderly individuals and patients with comorbidities and/or previous exposure to antibiotics may be at increased risk for abnormal oropharyngeal colonization.²⁷¹ The dynamics of change of oropharyngeal flora during hospital stay can be described as follows (Figure 67-5):

Figure 67-5 Evolution of potentially pathogenic microorganisms present in oropharyngeal flora, related to comorbidity, antibiotic treatment, and colonization pressure.

*Transitorily present in healthy carriers.

**Producers of ESBL or with type ampC chromosomal β-lactamases.

***Pseudomonas aeruginosa, Stenotrophomonas spp., Acinetobacter spp., Burkholderia spp.

ESBL, extended-spectrum β-lactamse; GNB, gram-negative bacilli; MDR, multidrug-resistant; MRSA, methicillin-resistant Staphylococcus aureus; MSSA, methicillin-sensitive S. aureus.

Changes in oropharyngeal flora tend to occur progressively such that the presence of microorganisms during one stage often overlaps with the next stage.

Choice of Empirical Antimicrobials Likely to be Active Against Causative Microorganisms

The latest guidelines of the American Thoracic Society and Infectious Diseases Society of America (ATS/IDSA) for the management of adult patients with nosocomial pneumonia¹ recommend that the selection of empirical antibiotic therapy for each patient should be based on the timing of onset and presence of risk factors for MDR pathogens. Risk factors for MDR pathogens defined by the ATS/IDSA guidelines are summarized in Box 67-2. An algorithm for the initial management of patients with nosocomial respiratory infection and selection of appropriate antimicrobials is shown in Figure 67-6. The antibiotics recommended by the current ATS/IDSA guidelines are shown in Tables 67-3 and 67-4. Adequate dosing of antibiotics for empirical therapy is summarized in Table 67-5. Broad-spectrum empirical antibiotic therapy should be rapidly deescalated as soon as microbiological data become available in order to limit the emergence of resistance in the hospital. In brief, initial empiric therapy should be based on patient’s risk of colonization by MDR organisms and managed as follows:

Box 67-2

Risk Factors for Multidrug-Resistant Pathogens Causing Nosocomial Pneumonia

• Antimicrobial therapy in preceding 90 days

• Current hospitalization of 5 days or more

• High frequency of antibiotic resistance in the community or in the specific hospital unit

• Presence of risk factors for healthcare-associated pneumonia:

Hospitalization for 2 days or more in the preceding 90 days

Residence in a nursing home or extended care facility

Home infusion therapy (including antibiotics)

Chronic dialysis within 30 days

Home wound care

Family member with multidrug-resistant pathogen

• Immunosuppressive disease and/or therapy

Adapted from American Thoracic Society. Guidelines for the management of adults with hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit Care Med 2005;171:388-416.

Figure 67-6 Algorithm for treatment of patients with suspicion of nosocomial respiratory infection. Systemic inflammatory response syndrome (SIRS) comprises at least two of the following: temperature >38°C or <36°C; heart rate > 90 beats/min; respiratory rate >20 breaths/min or PaCO₂ <32 mm Hg; and leukocytes > 12,000/mm³, <4000/mm³, or the preference of >10% immature neutrophils.

TABLE 67-3 Initial Empirical Antibiotic Treatment in Nosocomial and Ventilator-Associated Pneumonia of Early Onset in Patients Without Risk Factors for Infection by Multidrug-Resistant Pathogens

Adapted from American Thoracic Society. Guidelines for the management of adults with hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit Care Med 2005;171:388-416.

TABLE 67-4 Initial Empirical Antibiotic Treatment for Nosocomial and Ventilator-Associated Pneumonia of Late Onset or in Patients with Risk Factors for Infection by Multidrug-Resistant Pathogens and Any Degree of Severity

ESBL, extended-spectrum β-lactamase; GNB, gram-negative bacilli.

* The choice of β-lactam is made as follows: patients who have not received any antipseudomonal β-lactam within the last 30 days should be administered piperacillin-tazobactam or an antipseudomonal cephalosporin. Patients who have received these drugs should be given empirical therapy with a carbapenem. Patients with infection by ESBL-producing microorganisms should be treated with carbapenem regardless of the results of the antibiogram.

** For combined empirical therapy for multidrug-resistant GNB, an antipseudomonal fluoroquinolone should be used in cases of renal failure or concomitant use of vancomycin. In other settings, combined empirical therapy is initiated with amikacin and maintained for a 5-day period.

*** Empirical therapy aimed against MRSA is initiated in patients with proven colonization (ψ), previous infection by this microorganism, or implementation of MV for more than 6 days. The antibiotic of choice is either vancomycin (except in patients allergic to this medication, with creatinine values ≥ 1.6 mg/dL, or in patients presenting signs of empirical treatment failure after 48 hours of antibiotic therapy) or linezolid. (Ψ) For epidemiologic surveillance, nasal and perineal cultures should be performed on admission and at 1-week intervals thereafter while remaining in the ICU.

^† If an ESBL+ strain such as K. pneumoniae or Acinetobacter spp. is suspected, a carbapenem is the first choice.

^‡ If L. pneumophila is suspected, the combination antibiotic regimen should include a macrolide (e.g., azithromycin), or a fluoroquinolone (e.g., ciprofloxacin, levofloxacin) should be used rather than an aminoglycoside.

Antimicrobial Therapy in Special Situations

The addition of antibiotics with activity against MRSA depends on the local prevalence of MRSA, the presence of risk factors for MRSA, and the severity of infection. In geographic areas with documented presence of community-acquired MRSA, severe pneumonia with radiologic images of cavitation and presence of gram-positive cocci in respiratory secretions, empirical treatment with linezolid or vancomycin may be appropriate. Recently an outbreak of MRSA and linezolid-resistant S. aureus (LRSA) was reported in an intensive care department of a 1000-bed tertiary care university teaching hospital in Madrid, Spain, and was associated with nosocomial transmission and extensive usage of linezolid.²⁷² In that report, 12 patients with LRSA were identified, and a mortality of 50% was reported. CFR-mediated linezolid resistance was demonstrated in all isolates. Tigecycline may be a useful alternative in this setting, although clinical experience is scanty.

Infections by L. pneumophila serogroup 1 can be diagnosed by a Legionella urinary antigen test. This test should be routinely obtained if the hospital water supply is known to be colonized with L. pneumophila serogroup 1. A fluoroquinolone or a macrolide would be appropriate treatment for L. pneumophila infection.

Modifications of Therapy and Duration of Treatment

A suggested flowchart for follow-up of patients with nosocomial pneumonia is shown in Figure 67-7. After 72 hours, treatment should be adjusted based on microbiological results. The initial β-lactam should be continued if the microorganism is susceptible to the empirical β-lactam originally prescribed. If not, another β-lactam, possibly a carbapenem, may be introduced. The empirical antibiotic against MRSA should be discontinued if the presence of this pathogen is not confirmed by cultures. Discontinuation of the fluoroquinolone and especially the aminoglycoside should be considered after 3 to 5 days of treatment. The bactericidal activity of aminoglycosides and fluoroquinolones leads to a rapid reduction in the bacterial load during the first days of treatment. After this time, monotherapy may be sufficient. This approach would decrease emergence of resistant mutants and minimize nephrotoxicity caused by aminoglycosides.

Figure 67-7 Suggested flowchart for follow-up of patients with nosocomial pneumonia and VAP.

*Criteria of treatment failure taken from Ioanas M, Ferrer M, Cavalcanti M et al. Causes and predictors of non-response to treatment of the ICU-acquired pneumonia. Crit Care Med 2004;32:938-45.

^†In cases in which the etiologic agent is Pseudomonas aeruginosa or Acinetobacter spp., treatment should be maintained for 14 days.

^‡Patients with criteria of treatment failure and in whom MRSA is isolated should be administered linezolid. If a GNB is isolated, consultation is recommended.

The majority of infections can be effectively treated with regimens lasting up to 8 days. Four situations may justify prolonged treatment: (1) infection by microorganisms that may multiply in the cellular cytoplasm, such as Legionella spp.; (2) the presence of biofilms or prosthetic devices; (3) the development of tissue necrosis, the formation of abscesses, or infection within a closed cavity, such as empyema; and (4) persistence of the original infection (such as perforation or endocarditis). If the clinical course from the pneumonia is favorable—as defined by defervescence, improvement in PaO₂/FIO₂, and reduction in C-reactive protein (CRP) levels within the first 3 to 5 days of antimicrobial therapy—treatment may be withdrawn after the completion of 7 days. If the causative microorganism is a nonfermenting gram-negative bacillus, the treatment can be extended beyond 14 days. A large prospective, multicenter, randomized trial study comparing the efficacy of 8-day and 15-day antibiotic regimens for treating VAP suggested that an 8-day regimen reduces antibiotic use and decreases the emergence of multiresistant bacteria in the lung, without modification of the prognosis.²⁷³ However, this study observed that in cases of pneumonia produced by nonfermenting gram-negative bacilli, eradication of these microorganisms from bronchial secretions was lower with the shorter regimen.²⁷³ On the other hand, the 14-day treatment regimen was associated with a greater trend to colonization by multiresistant flora and a greater frequency of reinfection.²⁷⁴

In patients with clinical suspicion of ICU-acquired pneumonia who have a CPIS lower than 6 on the third day of treatment, the treatment may be withdrawn. In this setting, the patient probably does not have pneumonia, or the pneumonia is sufficiently mild such that prolonged antibiotic treatment is not required.²³⁵

Focus of Guidelines on Prevention and Management of VAP to Improve Outcome

Preventive approaches to VAP have focused on reducing cross-transmission, pulmonary aspiration across the ETT cuff, and reducing bacterial load in the oropharynx. Several strategies with proven efficacy in reducing morbidity and mortality related to mechanical ventilation have been grouped by the Institute for Healthcare Improvement as the “Ventilator Bundle.” Although the bundle was not designed to specifically prevent VAP, interventions such as semirecumbent position¹⁸⁷ and sedation vacation^119,124 have proven to significantly reduce VAP rates. The bundle was later modified to specifically address VAP by adding two strategies: daily oral use of chlorhexidine and subglottic secretion drainage. A 2-year multifaceted program to prevent VAP with eight targeted measures based on well-recognized published guidelines showed that increasing compliance with these measures was followed by an important reduction in the incidence of VAP.¹¹¹

In conclusion, in an effort to improve survival of patients with nosocomial pneumonia and considering the impact of antibiotic therapy on outcome, clinicians should consider the following points:

Key Points