Healthcare-Associated Pneumonia
Dennis C.J.J. Bergmans
Marc J.M. Bonten
The substantial clinical and financial impact of healthcareassociated pneumonia makes this an important topic for healthcare epidemiologists. According to surveillance data from the National Nosocomial Infections Surveillance system of the Centers for Disease Control and Prevention (CDC), pneumonia is the second most common healthcare-associated infection overall (1) and the most common infection in intensive care units (ICUs) (2). Additionally, pneumonia is associated with significant mortality and considerable costs of care (3). The widespread use of tracheal intubation and mechanical ventilation (MV) to support the critically ill has defined an expanding group of patients who are at particularly high risk for developing healthcare-associated pneumonia. In this group of patients, the infection is usually called ventilator-associated pneumonia (VAP). Unfortunately, both the diagnosis and the prevention of VAP have proven to be difficult (4).
HISTORICAL ASPECTS
During the last four decades, much has been learned about the epidemiology of healthcare-associated pneumonia. In the 1960s Pierce, Sanford, and others investigated the relationship of epidemic necrotizing gram-negative pneumonia to contaminated reservoir nebulizers in respiratory therapy devices and described effective disinfection measures (5, 6, 7). During the 1970s and early 1980s, additional work described the continuing association of healthcareassociated pneumonia with respiratory therapy equipment (8), risk factors for postoperative pneumonia (9), and the relationship of healthcare-associated pneumonia with oropharyngeal (10) and gastric (11) gram-negative bacillary colonization. During the 1980s and 1990s, several preventive strategies were designed and tested with varying success, such as the use of sucralfate for stress ulcer prophylaxis (12), selective decontamination of the digestive tract (SDD) (13), and continuous subglottic aspiration (14). Moreover, controversies developed over the relevance of gastric colonization in the pathogenesis of VAP (4,15,16), the usefulness of SDD (17,18), and the necessity of bronchoscopic techniques for diagnosing VAP (19,20). In the most recent years, more general approaches for patient management, not only directed to VAP, have been applied in ICUs, such as strict control of blood glucose levels, noninvasive ventilation, sedation strategies to reduce duration of ventilation, and bundles of care. This chapter summarizes current knowledge of diagnosis, epidemiology, and prevention of healthcare-associated pneumonia.
DIAGNOSIS
Pneumonia refers to inflammation of the distal lung caused by infection with microorganisms and is characterized histologically by the accumulation of neutrophils in the distal bronchioles, alveoli, and interstitium. Three types of healthcare-associated pneumonia can be distinguished: hospital-acquired pneumonia, early-onset VAP, and late-onset VAP. Pneumonia is defined as VAP when diagnosed in an intubated, mechanically ventilated patient after more than 48 hours of ventilation; early-onset VAP occurs within the first 4 days of MV, and late-onset VAP occurs thereafter (21). The relevance of a rapid and accurate diagnosis of healthcare-associated pneumonia is obvious. The goals are to prescribe the optimal antibiotic therapy to only those patients with infection of the lungs. Using a technique with a low sensitivity, patients will remain untreated and may suffer significant morbidity and increased risk for mortality. In contrast, a technique with high sensitivity but low specificity will lead to unnecessary use of antibiotics, resulting in unnecessary exposure of patients to toxicity, a potential delay in diagnosing the real etiology of infection, increased hospital costs, and, most importantly, unnecessary selection and induction of resistant pathogenic microorganisms.
Clinical and Radiographic Findings
Traditionally, clinical and radiographic criteria have been used to identify cases of healthcare-associated pneumonia among patients who are not mechanically ventilated. Patients with healthcare-associated pneumonia are likely to have fever, purulent sputum, signs of pulmonary consolidation, and new or progressive radiographic infiltrates. Although they may complain of dyspnea, cough, and pleuritic chest pain, many patients with healthcare-associated pneumonia are unable to give a helpful history because of neurologic impairment or severity of illness. The CDC definitions of healthcare-associated pneumonia have been widely used for infection control surveillance and rely predominantly
on clinical and radiographic criteria, although the results of other diagnostic tests may also be used (22).
on clinical and radiographic criteria, although the results of other diagnostic tests may also be used (22).
Diagnosing VAP is even more problematic than diagnosing healthcare-associated pneumonia in nonventilated patients (23). For scientific purposes, VAP is usually diagnosed using a modified version of the CDC’s definitions (24). These criteria are a new or progressive radiographic infiltrate that has persisted for at least 48 hours plus at least three of the following: a temperature above 38.5°C or below 35.0°C, a leukocyte count of >10,000/mm3 or <5,000/mm3, purulent sputum, or isolation of pathogenic bacteria from an endotracheal aspirate. An alternative for the modified CDC criteria is the Clinical Pulmonary Infection Score (CPIS) as defined by Pugin et al. (25) (Table 22-1). This scoring system, basically, uses the same criteria as in the modified CDC criteria. The range of the score is from 0 to 12, with VAP defined by a score of 7 or more.
TABLE 22-1 CPIS Used for the Diagnosis of Ventilator-Associated Pneumonia | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Although each of these criteria may have a reasonable sensitivity for VAP, specificity is poor. Radiographic infiltrates are sensitive indicators of VAP; however, more specific radiographic findings such as single air bronchograms, fissure abutment, or rapid progress to cavitation are infrequently present (26). Unfortunately, even the combination of clinical and radiographic criteria (as in the modified CDC criteria and the CPIS) has been unable to reliably diagnose cases of healthcare-associated pneumonia diagnosed by autopsy (26,27), histopathology (28), or other stringent criteria (29, 30, 31, 32). In an autopsy study of ventilated patients, no radiographic sign predicted pneumonia more than 68% of the time, and no radiographic signs predicted pneumonia in the subgroup of patients with the adult respiratory distress syndrome (ARDS); when clinical and sputum culture results were added to the model, the diagnostic efficiency rose only to 72% (26). Two other reports have documented that fever and pulmonary infiltrates in mechanically ventilated patients were caused by processes other than pneumonia in 49% to 69% of cases (29,30). A variety of other conditions (such as drug reactions, atelectasis, chemical aspiration, congestive heart failure), alone or in combination, can mimic the clinical and radiographic presentation of healthcare-associated pneumonia (33); these conditions are not uncommon in patients with significant underlying medical illness. Furthermore, cultures of sputum and tracheal aspirate do not reliably identify pathogens causing healthcare-associated pneumonia (33, 34, 35, 36) although surveillance cultures of sputum seem indicative for the pathogen eventually causing VAP and are superior in terms of choice of empiric antibiotic therapy compared to guidelines (37).
Interpretation of the usefulness of diagnostic techniques is seriously hampered, because populations studied varied widely, and the use of antibiotics, which strongly influences the yield of bacteriologic procedures, was not always taken into account. The largest problem, however, is the gold standard used to verify the value of diagnostic procedures. Histologic and bacteriologic examination of lung tissue remains the optimal standard to establish the diagnosis of pneumonia. However, these techniques require an open-lung biopsy or autopsy. Although histology and bacteriology of lung tissue, usually at autopsy, have been used as the gold standard in several studies, one should realize that patients who died form a subgroup of all patients with pneumonia. In addition, in several studies, autopsy findings are compared to diagnostic procedures performed several days earlier. In such circumstances, the physiologic host response to infection, usually in combination with antibiotic therapy, may have influenced the ultimate findings. These considerations also account for the evaluation of newer diagnostic techniques.
Quantitative Bronchoscopic Techniques
Diagnostic techniques have been improved by utilizing quantitative cultures of lower airway samples to more accurately diagnose healthcare-associated pneumonia and identify causative pathogens (28,35,38, 39, 40, 41). The rationale of bronchoscopy is to avoid contamination of culture samples with material from the upper respiratory tract. Three bronchoscopic techniques have been used to diagnose VAP: protected specimen brush (PSB), bronchoalveolar lavage (BAL), and protected bronchoalveolar lavage (PBAL).
The PSB technique uses a double-lumen bronchoscopic catheter with a telescoping cannula and distal plug (38). A bacterial burden of <104 colony-forming units (CFU)/g in lung tissue has been associated with histologic pneumonia (42). Since the sample size with PSB is ±0.001 mL, a quantitative culture of 103 CFU/mL reflects about 106 CFU/mL at the site of infection and is generally used as a cutoff point for VAP. Quantitative cultures of PSB specimens have been shown to accurately identify pneumonia and its causative pathogens in mechanically ventilated patients undergoing open lung biopsy (28) and in mechanically ventilated baboons with diffuse lung injury (35). Studies in normal hosts (43), patients with chronic bronchitis (44), and mechanically ventilated patients with suspected pneumonia (28,29,36) indicate that quantitative PSB cultures are considerably more specific than clinical and radiographic criteria for diagnosing healthcare-associated pneumonia. Based on the pooled results of published studies (28,29,35,45), one review (33) reported PSB to have a sensitivity of 83% and a specificity of 91%. False-negative results have been attributed to sampling error and to antibiotic use (36,41,45). In several studies, PSB results were compared to histologic and microbiologic evaluation of lung tissue, with sensitivities ranging from 36% to 100% and specificities ranging from 50% to 95% (28,46, 47, 48, 49).
Some investigators have also used quantitative cultures of BAL specimens to sample a larger portion of the lung, including alveoli (39,40,50). BAL entails sampling of an area of 106 alveoli. The tip of the bronchoscope is wedged, under visual control, into a third- or fourth-generation midsize bronchus, according to chest radiograph appearance. Lavage is carried out using at least 120 mL of sterile isotonic saline in several aliquot portions. The dilution of alveolar secretions in the lavage fluid is 10- to 100-fold, so a colony count of 104 CFU/mL in lavage fluid represents 105 to 106 bacteria per milliliter of alveolar secretion (51). In mechanically ventilated baboons, this method appears to be sensitive and provides the best correlation with culture of lung tissue (50). Like PSB, quantitative culture of BAL specimens appears to be more accurate than clinical and radiographic criteria for diagnosing healthcareassociated pneumonia (39,40,52); however, most studies evaluating quantitative BAL have not used rigorous criteria for determining whether pneumonia was actually present. BAL samples can become contaminated with bacterial flora from the upper respiratory tract, and reduced specificity may result (39). As with PSB, antibiotic use may diminish the sensitivity of BAL (40,50). In studies using histology and microbiologic examination of lung tissue as the gold standard, sensitivities of BAL ranged from 47% to 91%, and specificities ranged from 45% to 100% (46, 47, 48, 49).
To reduce upper airway contamination while still obtaining a representative sample of affected alveoli, Meduri et al. (41) developed a technique to obtain PBAL specimens using a balloon-tipped telescoping catheter. In a study of 46 patients, including 25 with suspected VAP, quantitative PBAL culture proved to have a sensitivity of 92% and a specificity of 97% for diagnosing bacterial pneumonia (41). Castella and associates (53) applied a similar PBAL technique and found that PBAL displayed improved sensitivity (85%) compared with PSB (62%) and improved specificity (83%) compared with BAL (44%). PBAL appears to be somewhat more demanding and timeconsuming than unprotected BAL, and has been studied less frequently than PSB and BAL.
After several years of experience with these bronchoscopic techniques, a consensus conference proposed a new definition for diagnosing VAP, which became known as the Memphis Ventilator-Associated Pneumonia Consensus Conference criteria (Table 22-2) (42). Definite VAP is only present with radiographic evidence of abscess and a positive needle aspirate, or if there is histologic proof of pneumonia at biopsy or autopsy. Probable VAP requires either positive quantitative or semiquantitative cultures from PSB or BAL, or blood or pleural fluid cultures of a microorganism found within 48 hours of isolation in the sputum, or abscess formation or consolidation with polymorphonuclear-cell infiltration at histologic examination.
Several technical and safety considerations pertaining to these bronchoscopic techniques warrant comment. The technique of the bronchoscopist may influence the results. The passage of the bronchoscope through the upper airway results in bacterial contamination of the suction channel, and the injection of topical anesthetic agents through the bronchoscope carries contaminants into the distal airways (38). Therefore, it is prudent to avoid suctioning or the injection of topical anesthetics prior to obtaining bronchoscopic samples. Other operator-dependent variables may also affect quantitative culture results. Elevated temperature, hypoxemia, increased radiographic infiltrates, bleeding, arrhythmias, and pneumothorax may be related to these bronchoscopic procedures, but serious or lasting complications appear to be rare (19,20). The use of a high FiO2, careful monitoring of exhaled volumes and oxygen saturation, and prebronchoscopy assessment of risk should help prevent serious complications.
The most important question remains whether these bronchoscopic techniques influence (preferably improve) patient care in the ICU. Several studies have suggested that quantitative bronchoscopic sampling of the distal airways by PSB, BAL, or PBAL substantially improves the accuracy of diagnosing healthcare-associated pneumonia when compared to clinical and radiographic criteria. In this way, these techniques may help to distinguish between patients who do and patients who do not need antibiotic therapy. In one study, VAP was simultaneously diagnosed according to the modified CDC criteria, CPIS, and quantitative bronchoscopic sampling (54). Incidences of VAP were 22% using the modified CDC criteria, 20% using CPIS, 9% using the Memphis criteria for probable VAP, and 0.4% using the Memphis criteria for definite VAP. In another study, only 50% of all patients fulfilling the modified CDC criteria for VAP met the definition of probable VAP with the Memphis criteria (55).
Six studies determined the effects of withholding or withdrawing empiric antibiotic therapy if a clinical suspicion was not confirmed by a diagnostic test. In three Spanish studies and a large Canadian study, patients with a clinical suspicion of VAP were randomized to an invasive or noninvasive strategy (56, 57, 58, 59). Because empirical therapy was not discontinued in any patient, regardless whether the suspicion was microbiologically confirmed or not, these studies cannot answer the question whether addition of these techniques to the diagnostic workup changes patient outcome, and have, therefore, not been included here.
TABLE 22-2 Recommended Definitions for VAP from the Memphis Ventilator-Associated Pneumonia Consensus Conference | |||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
In a landmark study, Fagon and collaborators randomized 413 mechanically ventilated patients with a clinical suspicion of VAP to a diagnostic strategy based on semiquantitative cultures of endotracheal aspirates (n = 209) or an invasive strategy (n = 204) (60). The invasive strategy consisted of bronchoscopy with direct microscopic examination of specimens (see above) and quantitative cultures. Quantitative cultures (PSB cutoff point >103 CFU/mL, BAL cutoff point >104 CFU/mL) were used to adjust or withdraw empirical treatment. In 117 of 204 patients, direct examinations yielded no bacteria. However, 20 of these patients had signs of severe sepsis and received antibiotics, although the suspicion of VAP was not confirmed. Of the remaining 97 patients, seven had significant quantitative cultures and antibiotics were started when culture results were available. So in all, the clinical suspicion of VAP was not confirmed in 90 patients. When comparing antibiotic use between patients randomized to either of the two invasive strategies, patients undergoing bronchoscopy had more antibiotic free days. Twenty-nine patients did not receive antibiotics up to day 28 (compared to four in the clinical management group), and patient survival was better among those randomized to the invasive strategy as well.
A different approach was used by Singh et al. (61). They determined CPIS scores of patients with a clinical suspicion of VAP and randomized those with a CPIS score ≤ six to different therapeutic strategies; antibiotics for 10 to 21 days (n = 42) or ciprofloxacin for 3 days (n = 39), which could be compared with withdrawing empirical therapy. Patients randomized to a short course had less antibiotic use, lower antibiotic costs, fewer superinfections with
antibiotic resistant pathogens, and no increase in length of stay or mortality. So, it seems safe to withdraw antibiotics after 3 days in patients with CPIS scores ≤6. However, the question is whether patients with such a low CPIS score should receive antimicrobial treatment in the first place.
antibiotic resistant pathogens, and no increase in length of stay or mortality. So, it seems safe to withdraw antibiotics after 3 days in patients with CPIS scores ≤6. However, the question is whether patients with such a low CPIS score should receive antimicrobial treatment in the first place.
Investigational Methods
A number of adjunctive or alternative methods for diagnosing healthcare-associated pneumonia have been proposed. Several reports have evaluated nonbronchoscopic techniques for sampling distal airway secretions, including brush (36), BAL (25,49,62,63), and endotracheal aspirates (45). Although such techniques might reduce costs, they have not been adequately validated or standardized, and sampling error can occur. In one study, however, good diagnostic agreement was demonstrated between quantitative cultures from PSB and mini-BAL done by respiratory therapists in patients with suspected VAP (64). And in another study, using postmortem analysis as the gold standard, blind bronchial sampling had a higher sensitivity for VAP than PSB (49).
The identification of intracellular microorganisms (ICOs) by Giemsa stain of BAL specimens has been reported to be highly predictive of healthcare-associated pneumonia (25,41,65). This technique has been compared to histologic and microbiologic examination of lung tissue at autopsy in two studies, with contradictory results. Specificities were high in both studies, 100% and 89% (46,47). However, sensitivity was 91% in one study when >5% of leukocytes containing intracellular bacteria was considered positive for VAP (46). In the other study, sensitivity was only 37%, when any leukocyte with intracellular bacteria was considered as a threshold for diagnosis of VAP (47). Since the presence of ICOs is indicative for the cellular response of the lung to invading microorganisms, it must have discriminatory value differentiating between infection and colonization. Determination of ICOs as an additional, early, indicator of VAP is being used increasingly, the cutoff value of 2% of cells with ICOs is used most frequently as diagnostic for VAP. An additional advantage of the determination of ICOs is that it is not influenced by recent antibiotic use (60,66,67).
Measurement of cytokines or inflammatory mediators may be adjunctive tools for diagnosing VAP in the future. Although blood levels of inflammatory mediators poorly correlated with quantitative results from bronchoscopic samples (68,69), elevated concentrations of endotoxin (>5 endotoxin units/mL) in BAL fluid had a sensitivity of 100% and specificity of 75% for diagnosing VAP (70). Soluble Triggering Receptor Expressed on Myeloid cells-1 (sTREM-1) has proven to be a biomarker for sepsis; however, for the diagnosis of VAP, the determination of sTREM-1 in BAL fluid showed high sensitivity and specificity in the first study (71). Unfortunately, this could not be repeated in several later studies (72). Procalcitonin is secreted as part of an inflammatory response but only when triggered by infection as shown in severe sepsis. The value of procalcitonin in pulmonary infection remains unclear and studies aiming to establish a diagnostic cutoff value of procalcitonin in case of VAP show contradictory results; sequential measurements might be more useful (73).
DESCRIPTIVE EPIDEMIOLOGY
Incidence
Incidence rates of VAP among ICU patients depend on the type of ICU, the severity of illness of patients studied, and the criteria for diagnosis. The overall incidence of pneumonia decreases when the definition of pneumonia becomes more strict. Therefore, whether investigators used bronchoscopic techniques in their diagnosis of VAP or just clinical and radiographic parameters is important with regard to incidence rates of VAP. In a number of studies aiming to ascertain incidences of VAP, or to evaluate modalities to diagnose VAP, or studies in which risk factors for VAP were assessed, the cumulative incidences of VAP range from 8.6% to 64.7%. Moreover, in studies on the effect of preventive measures on the occurrence of VAP, incidences of up to 78% have been reported in the control groups (77). In the Extended Prevalence of Infection in Intensive Care study (EPIC II), 13,796 patients in 1,265 ICUs were studied on a single day, 51% of the patients were considered infected, and 64% of them had pneumonia (78).
The cumulative risk for developing VAP during ICU stay increases until day 5. In one study, the calculated rates for VAP were 3% per day in the first week, 2% per day in the second week, and 1% per day thereafter (79). Two other studies suggest that there is a relatively constant 1% to 3% risk per day for developing VAP while MV continues for medical and surgical ICU patients (equivalent to 10 to 30 cases per 10,000 patient-ventilator-days) (80,81).
RISK FACTORS
The strongest risk factor for healthcare-associated pneumonia appears to be tracheal intubation and MV, which results in a 3- to 21-fold increase in the risk of developing healthcare-associated pneumonia (8,81,85, 86, 87, 88). Because other pathogenetic factors may be different, it is useful to consider nonventilated and ventilated patients (Table 22-3) separately when discussing risk factors.
When nonventilated or broad hospital populations are considered, factors found by multivariate analysis to significantly increase the risk of healthcare-associated pneumonia include chronic lung disease (85,119), severity of illness (119), upper abdominal or thoracic surgery (85,119,120), duration of surgery (119), age (85), poor nutritional state (120,121), immunosuppressive therapy (120,122), depressed level of consciousness (85,123), large volume of aspiration (85), impaired airway reflexes or difficulty handling secretions (123), nasoenteric intubation
(123), neuromuscular disease (121), and male gender (119). Additional risk factors suggested by univariate analysis include duration of hospitalization (6,9), oropharyngeal colonization with gram-negative bacilli (10), obesity (9), antibiotic therapy (6), reflux esophagitis (124), and previous pneumonia (124).
(123), neuromuscular disease (121), and male gender (119). Additional risk factors suggested by univariate analysis include duration of hospitalization (6,9), oropharyngeal colonization with gram-negative bacilli (10), obesity (9), antibiotic therapy (6), reflux esophagitis (124), and previous pneumonia (124).
TABLE 22-3 Risk Factors for ICU-Acquired and Ventilator-Associated Pneumonia | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The risk factors associated with the development of VAP have been determined in studies using multivariate analysis techniques, Cox regression techniques, and casecontrol designs, or have been suggested on the basis of reviews. Determination of risk factors for VAP has several clinical implications. They offer prognostic information about the probability of developing VAP, they help to reveal the pathogenesis of VAP, and they may provide possible targets for preventive strategies. By risk stratification, one can determine which patients may benefit most
from pneumonia prophylaxis (125). Risk factor analyses for ICU-acquired pneumonia (i.e., pneumonia diagnosed in ICU patients with or without MV) have clearly identified MV to be the most important risk factor (85,89,91, 92, 93). In general, the risk factors that have been identified can be divided into three groups: (a) risk factors that are well known (intubation, duration of MV, etc.) but very difficult to modify and that offer no or only limited possibilities for prevention, (b) risk factors that seem to play a role in the pathogenesis of VAP and have stimulated the development of a number of preventive strategies, and (c) risk factors that have been identified only incidentally or need further investigations to assess their significance and the possibility for prevention. Several risk factor analyses identified previous antibiotic use to be significantly associated with the development of VAP (94,95,99,100,107,112). In contrast, antibiotics conferred protection for VAP in a risk factor analysis (79), and the absence of prior antibiotic treatment was a risk factor for VAP caused by Haemophilus influenzae (113). Lately, attention has been drawn to the association between the mode of MV, ventilator-induced lung injury, and inflammation or infection (126,127).
from pneumonia prophylaxis (125). Risk factor analyses for ICU-acquired pneumonia (i.e., pneumonia diagnosed in ICU patients with or without MV) have clearly identified MV to be the most important risk factor (85,89,91, 92, 93). In general, the risk factors that have been identified can be divided into three groups: (a) risk factors that are well known (intubation, duration of MV, etc.) but very difficult to modify and that offer no or only limited possibilities for prevention, (b) risk factors that seem to play a role in the pathogenesis of VAP and have stimulated the development of a number of preventive strategies, and (c) risk factors that have been identified only incidentally or need further investigations to assess their significance and the possibility for prevention. Several risk factor analyses identified previous antibiotic use to be significantly associated with the development of VAP (94,95,99,100,107,112). In contrast, antibiotics conferred protection for VAP in a risk factor analysis (79), and the absence of prior antibiotic treatment was a risk factor for VAP caused by Haemophilus influenzae (113). Lately, attention has been drawn to the association between the mode of MV, ventilator-induced lung injury, and inflammation or infection (126,127).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
