key: cord-0006333-qgn68avp
authors: Andriesse, Gunnar I.; Verhoef, Jan
title: Nosocomial Pneumonia: Rationalizing the Approach to Empirical Therapy
date: 2012-08-23
journal: Treat Respir Med
DOI: 10.2165/00151829-200605010-00002
sha: 5b939ec4f8d611cfb7dcd6268ce5d10862fe8997
doc_id: 6333
cord_uid: qgn68avp

Nosocomial pneumonia or hospital-acquired pneumonia (HAP) causes considerable morbidity and mortality. It is the second most common nosocomial infection and the leading cause of death from hospital-acquired infections. In 1996 the American Thoracic Society (ATS) published guidelines for empirical therapy of HAP. This review focuses on the literature that has appeared since the ATS statement. Early diagnosis of HAP and its etiology is crucial in guiding empirical therapy. Since 1996, it has become clear that differentiating mere colonization from etiologic pathogens infecting the lower respiratory tract is best achieved by employing bronchoalveolar lavage (BAL) or protected specimen brush (PSB) in combination with quantitative culture and detection of intracellular microorganisms. Endotracheal aspirate and non-bronchoscopic BAL/PSB in combination with quantitative culture provide a good alternative in patients suspected of ventilator-associated pneumonia. Since culture results take 2–3 days, initial therapy of HAP is by definition empirical. Epidemiologic studies have identified the most frequently involved pathogens: Enterobacteriaceae, Haemophilus influenzae, Streptococcus pneumoniae and Staphylococcus aureus (‘core pathogens’). Empirical therapy covering only the ‘core pathogens’ will suffice in patients without risk factors for resistant microorganisms. Studies that have appeared since the ATS statement issued in 1996, demonstrate several new risk factors for HAP with multiresistant pathogens. In patients with risk factors, empirical therapy should consist of antibacterials with a broader spectrum. The most important risk factors for resistant microorganisms are late onset of HAP (≥5 days after admission), recent use of antibacterial therapy, and mechanical ventilation. Multiresistant bacteria of specific interest are methicillin-resistant S. aureus (MRSA), Pseudomonas aeruginosa, Acinetobacter calcoaceticus- baumannii, Stenotrophomonas maltophilia and extended-spectrum β-lactamase (ESBL)-producing Enterobacteriaceae. Each of these organisms has its specific susceptibility pattern, demanding appropriate antibacterial treatment. To further improve outcomes, specific therapeutic options for multiresistant pathogens and pharmacological factors are discussed. Antibacterials developed since 1996 or antibacterials with renewed interest (linezolid, quinupristin/dalfopristin, teicoplanin, meropenem, new fluoroquinolones, and fourth-generation cephalosporins) are discussed in the light of developing resistance. Since the ATS statement, many reports have shown increasing incidences of resistant microorganisms. Therefore, one of the most important conclusions from this review is that empirical therapy for HAP should not be based on general guidelines alone, but that local epidemiology should be taken into account and used in the formulation of local guidelines.

interest are methicillin-resistant S. aureus (MRSA), Pseudomonas aeruginosa, Acinetobacter calcoaceticus-baumannii, Stenotrophomonas maltophilia and extended-spectrum β-lactamase (ESBL)-producing Enterobacteriaceae. Each of these organisms has its specific susceptibility pattern, demanding appropriate antibacterial treatment. To further improve outcomes, specific therapeutic options for multiresistant pathogens and pharmacological factors are discussed. Antibacterials developed since 1996 or antibacterials with renewed interest (linezolid, quinupristin/dalfopristin, teicoplanin, meropenem, new fluoroquinolones, and fourth-generation cephalosporins) are discussed in the light of developing resistance.

Since the ATS statement, many reports have shown increasing incidences of resistant microorganisms. Therefore, one of the most important conclusions from this review is that empirical therapy for HAP should not be based on general guidelines alone, but that local epidemiology should be taken into account and used in the formulation of local guidelines.

Nosocomial pneumonia or hospital-acquired pneumonia (HAP) (pulmonary edema). Consequently, in only one-third of all ICU causes considerable morbidity and mortality. It is the second most patients with pulmonary infiltrate is pneumonia thought to be the common nosocomial infection and the leading cause of death from underlying etiology. [9, 10] On the other hand, the absence of pulmohospital-acquired infections. [1] Because diagnostic criteria differ nary infiltrates on the chest radiograph does not exclude pneumoconsiderably between various studies, only estimates of morbidity nia. [11] The clinical pulmonary infection score (CPIS) can be and mortality rates are available. The overall incidence of HAP is helpful in supporting the clinician in identifying patients with estimated to be between five and ten cases per 1000 hospital HAP. The score is calculated from temperature readings, leukoadmissions, increasing 6-to 20-fold in mechanically ventilated cyte counts, purulence of tracheal secretions, oxygenation (partial patients. [2] [3] [4] Up to 28% of patients who develop pneumonia in a pressure of oxygen in arterial blood [PaO2]/fraction of inspired general ward require transfer to an intensive care unit (ICU). [5] oxygen [FIO 2 ] ratio), aspects of pulmonary radiography, progres-Crude mortality rates for HAP range from 10% to 50%, with sion of pulmonary infiltrate, and quantitative culture of tracheal highest risks for ventilator-associated pneumonia (VAP). Estiaspirate (see also section 3.2 and section 4). mates of mortality rates directly attributable to HAP are up to By consensus HAP is defined as pneumonia occurring ≥48 30%. [6] [7] [8] hours after hospital admission but excluding any infection that is Because HAP constitutes a major clinical entity, the American incubating at the time of admission. [4] Empirical antimicrobial Thoracic Society (ATS) published a consensus statement in 1996 therapy for HAP is based on epidemiological studies and designed describing the current ideas of that time about empirical antimicrospecifically to cover microorganisms causing HAP. Therefore, bial therapy. Since 1996, a large body of research on diagnosis and successful empirical therapy greatly relies on an accurate diagnotreatment of HAP has appeared in the literature, especially regardsis of HAP and its etiology. ing the use of bronchoscopic techniques in identifying etiologic

In the pathogenesis of HAP, microaspiration of a small quantity pathogens of HAP in individual patients, which has increased our of oropharyngeal secretion, previously colonized with potential knowledge of the etiology of HAP. That knowledge forms the pathogenic bacteria, is the most common route of entry of pathobasis for the design of empirical therapy. The aim of this review is gens into the lower respiratory tract. Bacteria colonizing the to further rationalize the approach to empirical therapy and focus oropharyngeal epithelial lining of hospitalized patients may be on relevant research that has appeared since the ATS statement.

part of the patient's endogenous flora, or may originate from other patients, hospital personnel, or environmental sources. [12] [13] [14] [15] The 1. Diagnosis Before Empirical Therapy risk of colonization of the oropharynx and upper respiratory tract In general, pneumonia is diagnosed by the presence of a new by yeast or potentially pathogenic bacteria increases with the lung infiltrate plus evidence that this infiltrate is of infectious duration of hospital stay and is found in up to approximately 90% origin. An infection is suspected if fever occurs together with of intubated patients. [16] [17] [18] Moreover, the risk that this colonization purulent sputum and leukocytosis. However, several noninfectious encompasses resistant pathogens also increases. Especially in critcauses can mimic pneumonia and should be ruled out: atelectasis, ically ill patients or patients on mechanical ventilation, once acute respiratory distress syndrome (ARDS), adverse drug reac-microaspiration has occurred, secretions are insufficiently elimitions, pulmonary thromboembolism, pulmonary hemorrhage, pul-nated from the lower respiratory tract, which can lead to the monary fibrosis, lung carcinoma and congestive heart failure development of pneumonia. [19] As a result, it can be very difficult to differentiate colonizing pathogens in respiratory specimens benefit of the diagnostic procedure (e.g. refractory cases of HAP). from pathogens that are involved in active invasive infection.

Alternatively, a diagnostic BAL or PSB in patients currently receiving antibacterial therapy can be performed, but the results Microbiology of HAP is crucial for making the diagnosis and should be interpreted with caution. [32] [33] [34] initiating optimal therapy. Culture and Gram stain examination of expectorated sputum from patients with community-acquired Although bronchoscopically obtained quantitative BAL or PSB pneumonia (CAP) can identify the etiologic pathogen with reasonseem to offer maximal diagnostic reliability, in several studies able sensitivity and specificity. Most bacterial pathogens of CAP other less invasive techniques have shown comparable sensitivity are easily differentiated from the normal oropharyngeal flora that and specificity in diagnosing HAP. [35, 36] In patients with an endoinevitably is found in most cultures of expectorated sputum. In tracheal tube, techniques that can be considered as alternatives to contrast, culture of expectorated sputum from hospitalized patients bronchoscopic techniques are endotracheal aspirate and non-bronsuspected of having bacterial HAP is frequently contaminated by choscopic BAL/PSB in combination with quantitative culture. A colonizing microorganisms of the upper and lower respiratory major advantage is that these techniques can be employed by nontract. Although colonization of the upper respiratory tract by bronchoscopists and are less expensive. However, a major disadenteric Gram-negative bacteria is recognized as a risk factor for vantage of these techniques is the potential sampling error as a developing HAP, [20] [21] [22] [23] a positive sputum culture does not differenresult of the blind technique without airway visualization. When tiate between colonization and the etiologic pathogen of HAP in using endotracheal aspirates, mere tracheal colonization is differany given patient. In non-intubated and non-critically ill patients a entiated from positive cultures resulting from pneumonia by sputum culture can, at most, identify suspected pathogens and means of a certain cut-off value of the number of microorganisms their resistance pattern, but lacks sensitivity and specificity to per volume. However, if a cut-off of 10 6 cfu/mL is used many identify the etiologic microorganism of HAP. Since colonization patients may not be identified (false-negative sampling). [37] On the of the upper respiratory tract is most frequently found in critically other hand, diagnosing HAP with a lower cut-off value would ill or intubated patients, and because HAP in these groups is a liferesult in unnecessary treatment of patients without HAP (false threatening infection, a more reliable diagnostic sampling is positives). Similarly, during non-bronchoscopic BAL or PSB the needed in these patients.

catheter is inserted blindly into the respiratory tract with the risk Several bronchoscopic techniques for collection of secretions that microbiologic samples are obtained from unaffected segments from the lower respiratory tract in patients suspected of VAP have of the lung, yielding false-negative cultures in patients with been under evaluation for many years now. Since the 1996 ATS HAP. [38] statement many reports have appeared about the diagnostic value

The diagnostic technique for HAP that is of best value in of bronchoscopic techniques for endobronchial microbiological clinical decision-making probably depends mostly on the local sampling. Although consensus about which technique can be best situation. For bronchoscopic BAL or PSB to be of superior quality employed to diagnose VAP is lacking, both protected specimen a hospital needs experienced bronchoscopists, a sufficiently brush (PSB) and bronchoalveolar lavage (BAL) in combination equipped microbiology laboratory, the appropriate patient populawith quantitative culture have shown good results in diagnosing tion and, above all, physicians who are willing to respond to the VAP. [24] Although most studies using invasive techniques in diagoutcome: stop antibacterial treatment when confronted with a nosing HAP do so in patients suspected of VAP, bronchoscopic negative culture. Each hospital should devise a diagnostic protocol techniques may also play a crucial role in adapting antimicrobial for HAP that is the most accurate and the most practical, knowing therapy in non-mechanically ventilated patients suspected of the pitfalls of the chosen technique. HAP. [25] Quantitative culturing can help to differentiate HAP from non-Quantitative cultures were usually defined as bacterial growth infectious lung infiltrates, but definite results cannot be expected of the etiologic pathogen of >10 3 cfu/mL for PSB or >10 3-4 cfu/ before 48 hours after sampling. Alternatively, microscopic exami-mL for BAL. [26] [27] [28] [29] [30] [31] Also, in a meta-analysis including 26 studies, nation of Gram-stained samples of BAL can give early clues of PSB and BAL were both found to be reliable techniques to HAP. The detection of intracellular organisms in the diagnose bacterial HAP. [32] However, no technique can reliably polymorphonuclear leukocytes (PMNLs) and/or macrophages diagnose HAP for a patient already receiving antibacterial therof BAL has been correlated to VAP with variable sensitiviapy. [32] [33] [34] One study has demonstrated that antibacterial therapy ty. [24, 34, [38] [39] [40] [41] [42] [43] Most studies used a cut-off of ≥5% of the cells should be discontinued for at least 48 hours to obtain reliable positive for intracellular microorganisms. Although the Gram specimens by PSB, BAL, protected BAL, or endobronchial aspirastain can be used for rapid diagnosis of VAP, adjustment of tion. [33] The risk of discontinuation of therapy should outweigh the empirical therapy should be postponed until definite results of the with early onset (table I) ; (ii) mild to moderate HAP with risk quantitative culture are determined. [28, 44] factors, onset any time (table II); and (iii) severe HAP with risk  factors, early onset, or severe HAP with late onset (table III) . All Besides culturing of respiratory specimens, blood cultures and treatment groups are based on the different risk factors associated antigen tests should also be considered. Although only a minority with certain pathogens. of HAP patients produce positive blood cultures, these can help identify the causative pathogen in HAP. Similarly, pneumococcal Enterobacteriaceae, Haemophilus influenzae, S. pneumoniae urinary antigen tests can be helpful in cases of Streptococcus and Staphylococcus aureus are considered to be core pathogens; pneumoniae HAP. [45, 46] However, although both techniques may these microorganisms are most frequently isolated and should be be supportive in identifying the causative microorganism of HAP, covered by empirical therapy in any patient suspected of having they do not support the diagnosis of pneumonia.

HAP (table I) . However, when choosing empirical therapy for individual patients, risk factors linked to the emergence of mul-2. Pathogens Associated with Hospital-Acquired tiresistant bacteria need to be screened for. If a patient is at risk for Pneumonia (HAP) certain resistant microorganisms, empirical therapy needs to be adjusted to achieve adequate antimicrobial activity (tables II and Since microbiological identification of possible pathogens and III). their susceptibility pattern takes 2-3 days, initial antimicrobial Drug-resistant organisms that are of major concern in choosing therapy is, by definition, empirical. Even if respiratory culture empirical therapy of HAP are methicillin-resistant S. aureus results are available from before the onset of HAP, these are of (MRSA), Pseudomonas aeruginosa, Acinetobacter calcoaceticuslimited value in guiding initial antimicrobial therapy decisions for patients with suspected VAP; in one study all the organisms baumannii, Stenotrophomonas (Xanthomonas) maltophilia and ultimately responsible for pneumonia were recovered from only extended-spectrum β-lactamase (ESBL)-producing Enterobacter-35% of the specimens taken a median of 8 days before onset of iaceae. Each of these organisms has its specific susceptibility pneumonia. [47] pattern, demanding appropriate antibacterial treatment. Several studies have demonstrated that inappropriateness of initial empiri-The choice of empirical antimicrobial therapy is a balance between broad-spectrum antibacterial treatment with activity cal antibacterial treatment is significantly associated with an inagainst a wide range of etiologic bacteria, and antibacterial treatcreased mortality. [48] Consequently, initial empirical therapy ment with very selective activity, which minimizes adverse reacshould cover all drug-resistant bacteria that can be expected in tions and the development of antibacterial resistance. To achieve association with certain risk factors. However, most risk factors this, empirical therapy is aimed at the most frequently isolated associated with developing HAP with resistant organisms are bacteria in HAP. However, certain patients are at risk for HAP nonspecific and do not inform the clinician as to what specific with resistant pathogens. Consequently, if risk factors for resistant resistant organism the patient is at risk from. Based on epidemiopathogenic organisms exist, extended-spectrum antibacterial therlogic data describing organism-related prevalences, a risk profile apy is recommended.

can be outlined and empirical therapy can be adjusted accordingly. The incidence of specific resistant organisms varies greatly from In 1996 the ATS published guidelines for empirical therapy of country to country and even from one hospital to another. [49] For HAP in a consensus statement. [4] Based on the literature of that example, S. maltophilia infections are infrequent, except during time, the most likely pattern of bacterial pathogens of HAP was outbreaks. As a result, optimal empirical therapy can only be assessed. The statement defines three treatment groups: (i) mild to achieved if up-to-date local epidemiology is taken into account. moderate HAP without risk factors, onset any time, or severe HAP Adequacy of initial antimicrobial therapy is not the only deter-most HAPs occurring before that time are caused by one of the mining factor for outcome. Several studies have shown that mulcore pathogens. [4, 52] tidrug-resistant microorganisms are associated with higher levels

Since 1996 several publications have appeared on microbioloof mortality. However, only a few studies have adjusted for gy of early-versus late-onset HAP, all applying bronchoscopic comorbidity or the severity of underlying disease. Rello et al. [50] sampling techniques (BAL/PSB) in combination with quantitative found that even if initial antimicrobial therapy was active against cultures to identify 'real' cases of pneumonia. However, most P. aeruginosa, this organism was associated with an excess of studies were performed in patients admitted to an ICU and on mortality that could not be attributed to the severity of the underlymechanical ventilation. In an ICU study by George et al. [18] of 28 ing disease alone. Similarly, in a prospective case-control study patients with VAP, late-onset VAP was defined as pneumonia Bercault and Boulain [51] identified 92 cases of HAP with sensitive, developing after >5 days of mechanical ventilation. S. pneumoniae and 43 with multiresistant, etiologic pathogens. The latter group and Haemophilus spp. predominated in early-onset VAP and was significantly and independently associated with an increased Pseudomonas spp. and MRSA in late-onset VAP. MRSA was only mortality. Thus, multiresistant organisms are more than just resisfound in late-onset cases of VAP. However, prior use of antant to certain antibacterials and, consequently, identifying risk tibacterials can lead to selection of resistant strains. This was not factors associated with their presence is crucial in selecting the corrected for and may have influenced the outcome of this study. most vigorous empirical therapy.

In another prospective study of 135 episodes of VAP, the use of Risk factors for resistant pathogenic organisms are: (i) the broad-spectrum antibacterials prior to developing VAP was corlength of hospital stay before the occurrence of HAP; (ii) prior rected for; results indicated that drug-resistant bacteria were indeantibacterial therapy; (iii) severity of the pneumonia; (iv) presence pendently associated with the duration of mechanical ventilaof co-existing illness; (v) microorganism-specific risk factors; and tion. [53] Multiresistant bacteria (non-fermenting enteric Gram-neg-(vi) unsuspected resistant pathogens (table IV) . ative bacteria and/or MRSA) were not found before day 7 of mechanical ventilation. In a study by Rello et al., [49] 9 of 89

episodes of VAP occurring before day 7 of mechanical ventilation were caused by multiresistant bacteria. However, P. aeruginosa was found in three patients, all experiencing COPD. COPD is a Probably the most important risk factor for resistant etiologic known risk factor for VAP with P. aeruginosa. Therefore, empiripathogens of HAP is prolonged stay in a hospital. Analogous to the cal therapy for patients with COPD should include antiongoing process of colonization of the respiratory tract with potenpseudomonal coverage, especially in patients with a long-term tial pathogens, HAP occurring later during hospital admission is history of COPD with recurrent use of antibacterials. The other six correlated to an increased risk for resistant pathogens obtained episodes of VAP were caused by Acinetobacter baumannii. These from the hospital environment. [18] Consequently, the ATS advises cases were all found in two specific hospitals with a high prevadifferentiation between early-onset and late-onset HAP. The ATS lence of these bacteria. Again, this emphasizes that each hospital defines early-onset HAP as pneumonia occurring within 5 days of should be aware of local epidemiology of multiresistant bacteria. admission and late-onset HAP as pneumonia occurring ≥5 days

In hospitals with a high prevalence, empirical therapy should be after admission. The fifth day is taken as the 'cut-off' because adapted accordingly, especially in patients with severe HAP and is associated with HAP caused by multiresistant bacteria. A few those at increased risk of resistant bacteria. studies that appeared prior to the ATS statement suggested that In a prospective comparative analysis of 3668 ICU patients, HAP developing after antibacterial treatment was more likely to Ibrahim et al. [54] identified 235 patients with early-onset HAP (≤96 be caused by multiresistant pathogens. [58] [59] [60] In the study by Rello hours of ICU admission) and 185 patients with late-onset HAP. et al. [60] in 1993, from analysis of 129 consecutive episodes of P. aeruginosa was the only pathogen found significantly more VAP it was concluded that prior use of antibacterials was associatfrequently in patients with late-onset HAP compared with earlyed with a significantly greater mortality. Further logistic regresonset HAP. However, P. aeruginosa was isolated from patients sion analysis established that this was only independently related with early-onset HAP in greater numbers than expected and differto the presence of multiresistant pathogens. Although confounding ences between both groups were small: 25% and 38% in early-and factors were adjusted for, this study included many patients with a late-onset HAP, respectively. P. aeruginosa was found more frehistory of COPD, which itself is associated with multiresistant quently in patients with early-onset HAP than would be expected microorganisms and repeated use of antibacterials. based on the ATS statement. The authors discussed whether anti-Several studies appearing after the ATS statement confirm pseudomonal coverage should also be considered in some cases. these earlier reports. Trouillet et al. [53] conducted a study in 1998, However, this study defined the onset in relation to time of ICU analyzing 135 consecutive episodes of VAP. In this study prior admission (early-onset HAP was pneumonia occurring within the antibacterial use was also identified as an independent variable first 96 hours of ICU admission), in contrast to the ATS, which associated with VAP caused by potentially resistant bacteria. defined the onset in relation to the time of hospital admission. [54] In Moreover, not only the presence or absence of antibacterial therthis study more than 50% of early-onset HAP was caused by apy before the onset of pneumonia, but also the specific use of P. aeruginosa, MRSA, S. maltophilia, Enterobacter spp. and Acinetobacter spp. The high incidence of multiresistant bacteria broad-spectrum antibacterial agents such as third-generation was because of hospitalization prior to ICU admission. These cephalosporins, imipenem, or fluoroquinolones, was independentstudies demonstrate that the definition of early-and late-onset ly related to antimicrobial resistance of VAP. Moreover, in 2002 HAP should not be based exclusively on time of admission to an Trouillet et al. [61] demonstrated by multivariate analysis that not ICU. Since colonization of the respiratory tract commences at only was the occurrence of P. aeruginosa linked to previous admission to the hospital, this also defines the time of onset.

antibacterial use, but that the occurrence of piperacillin-resistant Although mechanical ventilation is a risk factor for the developstrains of P. aeruginosa could also be linked to prior antibacterial ment of HAP, it also seems to enhance the colonization process of use, especially that of fluoroquinolone. Since piperacillin is a resistant pathogens. [55] Therefore, the time a patient has been on major empirical antipseudomonal drug, in patients having remechanical ventilation should also be taken in account when ceived fluoroquinolones prior to the development of HAP, empiriassessing the risk for HAP with resistant microorganisms.

cal therapy for HAP with risk factors should not include piperacillin.

[61] Moreover, after having received fluoroquinolones prior to 2.2 Prior Antibacterial Therapy the onset of HAP, these antibacterials should also be avoided in empirical therapy since the existence of fluoroquinolone-resistant Prior use of antibacterials has previously been linked to the P. aeruginosa is to be expected. development of HAP. [56, 57] Moreover, prior antibacterial treatment The role of previous use of antibacterials becomes even more Analogous to CAP, severe HAP is defined by: (i) the need for pronounced during outbreaks with multiresistant bacteria. Husni et admission to an ICU; (ii) respiratory failure (need for mechanical al. [62] showed that during an outbreak with multiresistant ventilation or >35% oxygen); (iii) severe abnormalities on the Acinetobacter spp., prior use of ceftazidime was significantly chest radiography (progression, multilobarity or cavitation of the more frequent in patients who developed HAP with this microorpneumonia); or (iv) severe sepsis with signs of shock (hypotenganism compared with patients without HAP. [62] Similarly, during sion, oliguria/anuria, acute renal failure requiring dialysis). [4] Sean outbreak with MRSA, patients who developed HAP with vere sepsis should be defined by clinical parameters and not by the MRSA had received prior antibacterials significantly more often presence of bacteremia. Although blood cultures should always be than patients who developed HAP with methicillin-susceptible drawn in patients with serious infections, the presence of bacter-S. aureus (MSSA) infection. [ Rello et al. [50] showed that crude mortality in patients with tibacterial therapy in neutropenic patients if they become febrile nosocomial pneumonia appeared to be related to the degree of for any reason. [69] Consequently, especially in neutropenic patients organ dysfunction at diagnosis rather than to any characteristics of who remain febrile and develop pulmonary symptoms during the the pneumonia. The presence of resistant pathogens seemed to be course of broad-spectrum antimicrobial therapy, fungal pneumomore important than the severity of the pneumonia. Even if the nia should be considered. [70] Although Pneumocystis carinii (P. initial antibacterial therapy was active against P. aeruginosa, this jirovecii) is best known for infecting HIV-infected patients with organism was associated with an excess of mortality that could not low CD4 counts (<200/mm 3 ), patients with hematological maligbe attributed to the severity of the underlying disease alone. [50] nancies, solid tumors, collagen-vascular diseases and transplant In contrast to what is suggested by the ATS statement, there is recipients are also at risk. [71] Particularly in at-risk patients who insufficient evidence to establish severity of HAP as an indepenhave a protracted hospital stay, specimens obtained by bronchosdent risk factor for resistant pathogens. However, since patients copy should also be tested for P. carinii. with severe pneumonia are at increased risk of mortality, adequacy of initial antibacterial therapy may be of utmost importance.

Therefore, in patients with severe HAP of early onset, initial antibacterial treatment with a spectrum against local resistant Although no single risk factor can accurately predict the occurpathogens of high prevalence should be considered.

rence of specific resistant pathogens, the ATS statement does mention certain circumstances that increase the risk of certain

types of resistant bacteria involved in HAP. Since 1996 few studies have focused on microorganism-specific risk factors. As mentioned in sections 2.1 and 2.2, COPD is a known risk Although previous antibacterial treatment and prolonged hospifactor for VAP with P. aeruginosa as a pathogen. In addition, tal stay are risk factors for HAP caused by P. aeruginosa (as patients with any kind of structural lung diseases (e.g. bronchiectadiscussed in sections 2.1 and 2.2), corticosteroid therapy, malnusis and cystic fibrosis) are at risk for multiresistant pathogens, trition, structural lung disease (bronchiectasis, cystic fibrosis) and especially P. aeruginosa. The fact that P. aeruginosa is found in a mechanical ventilation may also increase the risk for P. aerugilarge percentage of these patient groups is possibly explained by nosa. [4] The ATS statement based the risk factors for P. aerugithe fact that these microorganisms survive best in humid environnosa on a study by Niederman, [22] which did not investigate cases ments, which are found in mucus-retaining bronchiectic pockets.

of HAP but cases of colonization with P. aeruginosa. However, In patients with cystic fibrosis, once P. aeruginosa or Burkholderrecent data confirm that patients with COPD are at increased risk ia cepacia have established in the airways it is almost impossible for P. aeruginosa HAP, and that this is the strongest risk factor to eradicate them; consequently, 30-40% of patients with cystic after prior use of antibacterials and prolonged hospital stay. [72] fibrosis will have long-term pseudomonal infection. [66] Therefore, Similarly, in 2002 Trouillet et al. [61] identified the presence of empirical therapy for patients with COPD should include antiunderlying fatal medical conditions as an independent factor assopseudomonal coverage, especially in patients with a long-term ciated with HAP with P. aeruginosa. In 2000, Akca et al. [73] history of structural lung disease with recurrent use of antibacteridemonstrated in 33 cases of VAP by P. aeruginosa that earlyals.

onset HAP can also be caused by these resistant bacteria and that In immunocompromised hosts, specific resistant bacteria may this was significantly associated with emergency intubation, aspiplay a role in developing HAP. S. maltophilia has recently ration and a Glasgow Coma score of ≤9. However, early-onset emerged as an important nosocomial pathogen in immunocom-VAP was defined by the time of intubation and, as a result, promised cancer patients and transplant recipients. Risk analysis prolonged hospital stay may still be the leading risk factor. Furhas shown that mechanically ventilated ICU patients receiving thermore, culture data were based on tracheal aspirates and could antibacterials, especially carbapenems, are at increased risk of merely have resulted from colonization of the tube. colonization/infection. [67] Similarly, Legionella spp. have a predi-

The use of mechanical ventilation is associated with P. aerugilection for infecting immunocompromised patients, and transplant nosa and Acinetobacter spp. in particular. One pathogenic mecharecipients have the highest risk. Moreover, Legionella spp. have nism described is the difference in adherence by different bacterial been the most common cause of nosocomial pneumonia among species to different catheter surfaces. In contrast to urinary cathetransplant recipients at selected medical centers. [68] Neutropenic ters, the role of biofilm in respiratory tube adherence and pathopatients (PMNLs <500/mm 3 ), in particular, are at risk of fungal genesis of VAP is less clear. [74] Nevertheless, water condensate in infections. Most guidelines recommend broad-spectrum an-the endotracheal tube may favor certain bacterial species such as patients admitted to the hospital or residence in long-term care P. aeruginosa and Acinetobacter spp.

facilities were thought to be at risk of acquiring ESBL-producing Klebsiella or E. coli. However, Einhorn et al. [91] showed that in In a study of 148 episodes of VAP, Baraibar et al. [75] demon-14% of all ESBL cases in Chicago, Illinois, USA, in 2002, the strated that A. baumannii was independently associated with neuinfection was acquired in the community by patients who resided rosurgery, head trauma, ARDS and aspiration. In contrast to at home. Although ESBL-producing bacteria constitute a major P. aeruginosa or Enterobacteriaceae, in this study A. baumannii therapeutic problem most reports could not demonstrate an inseemed not to be related to co-morbid illness, severity of disease, creased mortality in patients who developed infections with ESor exposure to antibacterial therapy. However, this study was of BL-producing organisms compared with patients with infections relatively small sample size and in 16% of all HAP episodes the with non-ESBL-producing organisms. [89, 92] Consequently, during bacterial diagnosis remained uncertain, despite invasive bronchoan outbreak with an ESBL-producing microorganism, patients scopic techniques. The underlying mechanism of these findings with HAP are at risk, especially those who have received prior could not be explained from these results.

antibacterial therapy, and those with late-onset and ventilator-In addition to P. aeruginosa, S. maltophilia has also been associated HAP. identified as a high-risk pathogen; VAP with S. maltophilia is associated with increased length of ICU stay and mortality. In

In the ATS guidelines thoracoabdominal surgery and witnessed addition, S. maltophilia carries intrinsic resistance to most antibacaspiration are pointed out as risk factors for developing HAP with terials. [76] In a multivariate analysis, patients with tracheostomy, anaerobic microorganisms. This was based on a study by Bartlett cefepime exposure and severe trauma with lung contusion were et al. [93] performed in 1986 on cultures of pleural effusions, blood significantly more at risk for S. maltophilia HAP. [77] and nonquantitative cultures of tracheal aspirates. This study needs Since S. aureus is one of the major pathogens of HAP, MRSA careful interpretation because based on today's knowledge of is one of the major concerns in the design of empirical treatment bronchoscopic sampling, the anaerobes found in 35% of cases of protocols. Although debated for a long time, a recent meta-analy-HAP may well have been the result of sampling of high airway sis has provided evidence that MRSA is associated with a significolonization. [94] Because of the technical difficulties and relatively cant increase in mortality in comparison with similar infections high costs, most microbiological laboratories do not routinely with MSSA. [78] Recent data indicate that the MRSA incidence is employ anaerobic culture techniques for respiratory specimens. increasing despite recommendations for isolation precautions. [79] However, sometimes even simple measures can improve the diag-No type of infection has a predilection for MRSA; however, in nosis of anaerobic infections; PSB samples should be transported contrast to MSSA, prior antibacterials (especially levofloxacin and in thioglycolate instead of saline. [95] Overall, the role of anaerobes macrolides), previous hospitalization, enteral feeding, surgery and in HAP may be underexposed in the literature. However, a few length of stay before culture were independently associated with studies have appeared since the ATS statement, using broncho-MRSA infection. [80] During a 5-year period, Pujol et al. [81] studied scopic sampling techniques and anaerobic culturing. The main all VAP cases caused by MRSA and found that MRSA caused anaerobic strains isolated were Prevotella melaninogenica (36%), exclusively late-onset VAP, while MSSA caused both early-onset Fusobacterium nucleatum (17%), and Veillonella parvula (12%). and late-onset VAP. [81] In patients with HAP who have had prior VAP with anaerobes occurred significantly more often in patients antibacterial treatment and who are admitted to a hospital with a who were orotracheally intubated than those nasotracheally intuhigh prevalence of MRSA, have a long hospital stay and are on bated and significantly more frequently in early-onset VAP than mechanical ventilation, empirical treatment should have activity late-onset VAP. Multivariate analysis demonstrated that the presagainst MRSA.

ence of altered levels of consciousness, higher disease severity, Since the ATS statement in 1996, an increasing number of and admission to the medical ICU were the factors independently reports have appeared on outbreaks with ESBL-producing Kleb-predisposing to the development of VAP with anaerobes. [96] In siella spp. or Escherichia coli. [82] [83] [84] [85] Non-ESBL-producing Kleb-contrast, Marik and Careau [97] performed a similar study and found siella spp. and E. coli are generally susceptible to most β-lactam only one anaerobic microorganism in 75 patients with HAP. Also, antibacterials. In contrast, ESBL-producing strains are resistant to in 12 patients with observed aspiration, no anaerobes were recovmost β-lactams with the exception of carbapenems. Furthermore, ered from the bronchoscopically obtained samples. Eight years recently even imipenem-resistant Klebsiella strains have been after the ATS statement, too few studies have appeared to establish observed. [86, 87] Intubation, previous antibacterial therapy, and cen-the exact role of anaerobes in HAP. However, their presence as cotral venous catheter insertion have been identified as risk factors pathogens should always be suspected, especially in intubated for infection with ESBL-producing strains. [88] [89] [90] Up to now, only patients or after aspiration. Analogous to CAP, patients with HAP and periodontal disease, bronchiectasis or bronchial stenosis (by 3. Antibacterial Treatment of HAP tumor, stenosis or foreign body) may also be at increased risk for anaerobic pulmonary infection. [70] In contrast to HAP, the role of

Once the clinical diagnosis of HAP is made and risk factors pneumonia and lung abscess. [70, 94] On the other hand, anaerobes have been assessed, empirical treatment guidelines can be obare usually of low virulence and many patients commonly recover tained using the algorithms that the ATS published in 1996 (see from HAP without receiving specific anti-anaerobic therapy.

tables I, II and III). These antibacterial recommendations were based on well designed, controlled clinical trials whenever possi-

ble. [4] When sufficient data were lacking, the spectrum of antimicrobial activity and pharmacokinetic data were taken into account. Although relatively rare, nosocomial legionnaires' disease can All guidelines for empirical therapy are focused on the initial occur in outbreaks as well as in single patients. As is stated in the treatment of patients with HAP; as soon as possible pathogens are ATS guidelines, empirical treatment for nosocomial pneumonia identified and antimicrobial susceptibility patterns are available, does not include specific antibacterials against Legionella empirical therapy should be re-evaluated. Obviously, if susceptipneumophila. As a result, mortality from legionnaires' disease is bility testing of possible pathogens demonstrates resistance, empirical therapy should be changed to antibacterials with an effechigh in patients who receive inappropriate antibacterial therapy.

tive spectrum of activity. On the other hand, if microbiological test Two reports by Chang et al. [98] and Goetz et al. [99] demonstrate that results show susceptible pathogens (e.g. H. influenzae or S. routine environmental cultures play a major role in stimulating the pneumoniae), empirical therapy must be changed to antibacterials application of Legionella laboratory testing. In both studies posiwith a narrower spectrum. [104] This should be done not only for tive water-system samples subsequently identified unsuspected reasons of suppressing global development of antibacterial resispatients with nosocomial legionnaires' disease. [99] As this contance, but also to minimize the risk of serious adverse effects of firms the ATS guideline, clinicians and microbiologists should broad-spectrum antibacterial therapy such as pseudomembranous indeed bear in mind the need to also test for legionnaires' disease colitis and selection of, and super infection with, multiresistant in patients with severe or culture-negative HAP. This is especially bacteria. [105] true if patients are at risk for legionnaires' disease during high-

The major goal of empirical therapy guidelines is to ensure that dose corticosteroid treatment, malignancy, renal failure, neutropeinitial antibacterial therapy has sufficient activity against the unnia or cytotoxic chemotherapy. [4] In 2000, a study by Stout et known pathogen causing HAP. Inappropriateness of initial al. [100] revealed that long-term care residents are at risk of acquirantibacterial treatment is associated with an increased mortaliing nosocomial legionnaires' disease in the presence of a coloty. [56, 106] Recent studies, based on modern sampling techniques, nized water system. Consequently, nosocomial legionnaires' disalso show that adequacy of initial antibacterial treatment, based on ease should be suspected if a sudden flare-up of incidence is susceptibility tests, is of great importance for clinical efficacy. Studying 119 nosocomial infections in four ICUs, Zaidi et al. [107] noted. [12, 101, 102] Therefore, urine antigen tests for L. pneumophila found that the major risk factors for mortality were inadequate type 1 should be readily available if even the slightest possibility antibacterial treatment and development of VAP. Similarly, Luna of legionnaires' disease exists (see also section 4). et al. [64] performed bronchoscopic sampling and multivariate anal-Other more rare bacterial causes of HAP with microorganisms ysis on 162 cases of VAP and found that inadequacy of initial requiring specific antibacterials are Mycoplasma pneumoniae, antimicrobial therapy and age >50 years were the only factors Chlamydia pneumoniae and Mycobacterium tuberculosis. [98, 103] associated with mortality. Although it is probably appropriate to These pathogenic microorganisms are frequently identified in employ general guidelines for empirical therapy for HAP, it cancommunity-acquired respiratory tract infections but rarely in not be stressed enough that local epidemiologic surveys should be HAP. These microorganisms can only be identified by serologic performed to monitor for possible resistant pathogens. Similarly, testing or specific culture techniques that are not routinely used in during outbreaks of resistant strains, the local empirical antibacterlaboratory testing for HAP. However, if a patient is admitted to the ial policy should be re-evaluated. hospital for a long period, the risk of infection with these patho-Part of the adequacy of initial therapy is also determined by gens may increase, especially if the patient is cared for in the time management. In one study, in almost 31% of all patients vicinity of an index case with M. tuberculosis, for example. meeting the diagnostic criteria for VAP, the initial appropriate antibacterial treatment was delayed for 24 hours. [108] In logistic bacterial resistance to these recommended β-lactams is bacterial regression analysis, delayed antibacterial treatment was identified production of β-lactamases that can hydrolyze β-lactam antibacteas an independent risk factor associated with increased mortality. rials to inactive compounds. Different types of β-lactamases are The most common reason for deferral of therapy was a delay in produced by different bacteria. Second-generation cephalosporins writing the antibacterial orders! are resistant to the β-lactamases produced by certain strains of H. In the ATS algorithm, the first step is to define the severity of influenzae, Klebsiella, E. coli and S. aureus. Although these βillness as either mild to moderate or severe, analogous to treatment lactamases can hydrolyze amoxicillin, ticarcillin or piperacillin, guidelines for CAP. [109] In patients with mild to moderate illness when combined with a β-lactamase inhibitor, hydrolysis is prerecommended empirical therapy is directed against the core pathovented and the antibacterial retains its activity. However, βgens (table I), independent of length of hospital admission. Howlactamases produced by certain strains of Enterobacter, Serratia, ever, if specific risk factors for infection are present, specific and Citrobacter (class C, type AmpC) are also capable of hydroantibacterials should be added to the 'core empirical therapy' lyzing amoxicillin despite the presence of clavulanic acid. Para- (table II) . Patients with severe HAP will fall into the description of doxically, clavulanic acid is a stronger inducer of AmpC βtables I or III. The treatment will depend mostly on whether the lactamases, which it cannot inhibit, than sulbactam and patient developed HAP early, within 5 days of hospitalization tazobactam. However, the real concern lies with strains that pro-( and aminoglycosides as alternatives. [85, 111] These so called 'derepressed' mutants have become very prevalent, with incidences

Europe. [112] [113] [114] The risk of selecting derepressed mutants during Under mean general epidemiologic circumstances, according to therapy is approximately 20% when third-generation ATS guidelines in patients with mild to moderate HAP without cephalosporins are used to treat bacteremia caused by Enterrisk factors and onset anytime, and in patients with severe HAP obacter spp. [115] and is probably higher in pneumonia. [112] Howevwithout risk factors and of early onset, adequacy of empirical er, induction does not usually take place within a couple of days of antibacterial therapy is achieved using 'core antibacterials' (table therapy. Therefore, if after a few days culture results indicate I). The definition of 'core antibacterials' is that they have activity Enterobacter spp., Citrobacter spp. or Serratia spp. as the pathoagainst enteric Gram-negative bacteria, H. influenzae, S. gen, in severe cases a change of therapy should be considered. If pneumoniae and MSSA. Since P. aeruginosa is only seldom found local epidemiologic data demonstrates high incidences of one of in this patient group, 'core antibacterials' need not have antithese species, β-lactamase production may already have been pseudomonal activity. First-choice antimicrobial therapy for these induced and henceforth empirical therapy should altogether be core pathogens would consist of either a cephalosporin or a βswitched to therapy as is suggested in table III. lactam/β-lactamase inhibitor combination. cephalosporins of the second generation (e.g. cefotetan, cefoxitin, cefuroxime) or non-

pseudomonal third generation (e.g. ceftriaxone, cefotaxime, ceftizoxime) are recommended. β-Lactam/β-lactamase inhibitor If patients with mild to moderate HAP are at risk for resistant combinations are amoxicillin/clavulanic acid, ticarcillin/clavulanpathogens, independent of the time of onset, empirical therapy can ic acid, ampicillin/sulbactam and piperacillin/tazobactam. In a be adapted according to table II. Although the role of anaerobes in study by Speich et al. [110] the efficacy of piperacillin/tazobactam, HAP is not clear, under some circumstances (described in section the most recently developed β-lactam/β-lactamase inhibitor com-2.6 and table IV) specific anaerobic antimicrobial therapy should bination, was compared with that of amoxicillin/clavulanic acid in be considered. In a relatively small study in patients with HAP, the the treatment of severe pneumonia. [110] The agents proved to be most frequently isolated anaerobic bacteria were Prevotella spp., equally and highly efficacious treatments. However, only a minorwhich were more frequently resistant to cefotaxime (37%), ity of pneumonias in this study was of nosocomial origin and ceftazidime (50%), and ciprofloxacin (32%) than usually reported consequently Enterobacter spp., Proteus spp. and Serratia spp.

in the literature. [116] Sixty-six percent of these strains produced βwere not encountered as core pathogens.

lactamases. From these results it could be concluded that patients who had received empirical anti-anaerobic antimicrobial therapy β-Lactam antibacterials form the basis of empirical therapy for had a significantly better outcome after 10 days. [116] Moreover, HAP without risk factors (table I). The major mechanism of since penicillin-resistant anaerobic organisms, usually Bacter-failure, neutropenia or cytotoxic chemotherapy are also at risk for oides spp., can also be encountered in infections of the lower legionnaires' disease, especially if the hospital water system is respiratory tract, [117] specific anaerobic treatment may be war-known to be infected. [4] Especially in severe cases of legionnaires' ranted. Of the core antibacterials (table I) only cefoxitin and disease, empirical treatment with core antibacterials has insufficefotetan (second-generation cephalosporins) [118] and β-lactam/βcient activity against L. pneumophila. If there is a risk for legionlactamase inhibitor combinations possess sufficient activity naires' disease, specific empirical treatment should be considered. against anaerobes including Bacteroides fragilis. Therefore, if a Instead of erythromycin plus rifampicin, as is stated by the ATS, patient is suspected of HAP with possible involvement of anaer-recent literature suggests that fluoroquinolones or newer obes, first-choice empirical therapy should consist of these drugs macrolides (e.g. azithromycin) should now be considered firstor any combination with clindamycin or metronidazole. However, choice therapy for legionnaires' disease. [128] [129] [130] Rifampin several case-controlled studies have suggested that metronidazole (rifampicin) may have additional efficacy in severe cases of HAP. monotherapy of anaerobic pulmonary infections is less effective

than clindamycin as a single drug. [119] [120] [121] Since clindamycin is also with Late Onset active against many Gram-positive microorganisms (e.g. S.

In 1996 the ATS recommended that patients with severe HAP pneumoniae, S. aureus, and β-hemolytic streptococci) this would with risk factors and early onset, and patients with severe HAP be the first choice in patients with HAP with increased risk for with late onset should be treated according to table III. In these anaerobic pneumonia. However, clindamycin has limited or no patients, next to the core pathogens, the main concern is the high activity against some strains of B. fragilis. Therefore, in selected prevalence of P. aeruginosa and other multiresistant bacteria. cases of HAP with a high suspicion of anaerobic microorganisms Particularly in patients with mechanical ventilation, HAP caused (e.g. observed aspiration or post-obstruction pneumonia) combiby these microorganisms is associated with increased mortalination therapy with clindamycin and metronidazole can be considty. [131, 132] Because bactericidal synergy against Pseudomonas and ered.

Acinetobacter spp. has been shown when carbenicillin and an In hospitals where MRSA is not highly prevalent, empirical aminoglycoside are combined, the use of an antipseudomonal βtherapy consisting of a second-generation (cefalotin or cefazolin) lactam (piperacillin, ticarcillin, ceftazidime, or imipenem) in comor third-generation cephalosporin or a β-lactam/β-lactamase inbination with an aminoglycoside remains the preferred therapeutic hibitor will have sufficient activity against S. aureus. However, in approach where possible. [133] Similarly, when considering time-North and Latin America, respectively, 44% and 46% of all kill curves in vitro, the combination of ciprofloxacin plus piper-S. aureus isolates from patients with pneumonia consisted of acillin plus tazobactam achieved greater killing than other combi-MRSA. [122, 123] Similar rates are found in Europe and Japan. [79, 124] nations or monotherapy against P. aeruginosa. [134] Whether syner-Therefore, if MRSA is highly prevalent and the patient is at risk of gistic activity against P. aeruginosa can be achieved or not, there HAP with S. aureus, empirical therapy should consist of a core is another reason for combining two different antibacterial categoantibacterial plus vancomycin. [125] From earlier studies quinuprisries; when P. aeruginosa is implicated, monotherapy, even with tin/dalfopristin has been considered to be an option for unresponbroad-spectrum antibacterials, is associated with a rapid increase sive MRSA infections, where few proven treatment options exin resistance and a high rate of clinical failure. Therefore, for ist. [126] In cases of patient allergy or intolerance to glycopeptides, pseudomonal HAP, combination therapy consisting of an in one study quinupristin/dalfopristin proved to be equally safe, antipseudomonal β-lactam plus an aminoglycoside or a fluoroquibut somewhat less efficacious than vancomycin. [127] Other alternanolone (e.g. ciprofloxacin) is advised. [135] tives are teicoplanin and linezolid (see section 3.1.4). However, During outbreaks with ESBL-producing microorganisms, vancomycin is not the drug of choice for patients infected with carbapenems are the first-choice empirical therapy for patients MSSA. In a study by Gonzalez et al., [63] mortality was significantsuspected of HAP with late onset or other risk factors, depending ly higher among MSSA-infected patients treated with vancomycin on the characteristics of the strains involved. Although β-lactam/ than among those treated with cloxacillin (47% vs none). Hence, if β-lactamase inhibitor combinations can be used for some suscepticulture results show susceptible S. aureus as a single pathogen, ble strains, even for the most potent (piperacillin/tazobactam), therapy should preferably be switched to flucloxacillin or nafcillin resistance in Europe has risen significantly from 31% to 63% over monotherapy.

a period of 4 years. [136] Fluoroquinolones can be used as an According to ATS guidelines, patients receiving high-dose alternative empirical treatment, but one should realize that both in corticosteroid therapy are at risk for legionnaires' disease. [4] How-Europe and the US up to 31% of ESBL-producing isolates are also ever, based on recent literature, patients with malignancy, renal ciprofloxacin-resistant. [83, 136] When fluoroquinolones are used in cases of high ESBL risk, adding an aminoglycoside to the empiri-tion. [144, 145] Similar to imipenem, meropenem has shown good efficacy as monotherapy in patients with HAP. [146] cal treatment may improve activity against possible ciprofloxacinresistant strains. [111] Recent epidemiologic studies have shown that Pseudomonas spp. also retain good susceptibility to fourth-generation

cephalosporins. Moreover, multiresistant Enterobacteriaceae also show high rates of susceptibility to cefepime and cefpirome. [137, 147] Antibacterials developed since 1996 or antibacterials with re-This may explain why empirical therapy for VAP with cefepime or newed interest for the treatment of HAP are linezolid, quinupriscefpirome was relatively superior to ceftazidime in some studtin/dalfopristin, teicoplanin, meropenem, new fluoroquinolones ies. [148, 149] Although ceftazidime remains a superior antiand fourth-generation cephalosporins.

pseudomonal drug, especially in hospitals with high incidences of Since the ATS statement a new class of antibacterials has been cases with ceftazidime-resistant P. aeruginosa, fourth-generation developed: oxazolidinones. Oxazolidinones having activity cephalosporins (cefepime and cefpirome) can be considered as against Gram-positive bacteria are of most interest in the treatment empirical therapy in cases of severe HAP. of HAP with MRSA. In a large epidemiological study involving

The number of HAP cases with S. maltophilia is increasing and the ICUs of 25 European university hospitals (SENTRY antimiconstitutes a major therapeutic problem. [67] S. maltophilia is resiscrobial surveillance program) resistance to oxacillin was found in tant against many antibacterials used as empirical therapy. Conse-39% of all collected S. aureus strains. However, all these isolates quently, prior antibacterial treatment is a risk factor for HAP with were fully susceptible to linezolid and vancomycin. [137] Two stud-S. maltophilia. [77] In one study antibacterial susceptibility testing ies have examined the use of linezolid as an alternative treatment revealed that isolates were most sensitive to sulfamethoxazole to vancomycin: in both studies patients received linezolid plus (80%), chloramphenicol (75.5%) and ceftazidime (64.5%). [150] In aztreonam or vancomycin plus aztreonam as empirical treatment contrast, a recent report has shown that, in vitro, more strains are for HAP. [138, 139] Clinical cure and microbiological eradication rates sensitive to minocycline, doxycycline and moxifloxacin. More were equivalent between treatment groups. Similar results were than 70% of strains were resistant against ceftazidime, cefepime, documented for HAP patients with MRSA. [138, 140] Linezolid is piperacillin, ticarcillin and aztreonam. Only 25% of all strains well tolerated and has a major advantage over vancomycin in that were resistant against trimethoprim/sulfamethoxazole. Thus, in oral formulations are also available. [141] Another possible alternapatients highly suspected of HAP with S. maltophilia (e.g. during tive treatment for HAP with MRSA is quinupristin/dalfoprisan outbreak), it is now recommended to initiate empirical treattin. [142] Similar to linezolid, quinupristin/dalfopristin also has in ment with trimethoprim/sulfamethoxazole and ticarcillin/clavuvitro activity against most MRSA strains, comparable to telanic acid in combination. [151] icoplanin and vancomycin. [143] In one study by Fagon et al., [127] In the search for new antimicrobial therapies for Acinetobacter 31% of MRSA cases treated with quinepristin/dalfopristin were spp., unconventional antibacterial treatment strategies have been clinically successful. Nonetheless, there was 44% success in the tested. Wolff et al. [152] studied the efficacy of β-lactams/βvancomycin-treated group. More studies will be needed to exlactamase inhibitors and rifampin in a mouse model for A. amine the clinical potentials of these drugs as empirical therapy for baumannii pneumonia. The best survival rates (≥80%), even when HAP with a high risk of MRSA. Teicoplanin is an effective and mice were infected with a multiresistant strain, were obtained with safe alternative for vancomycin in the treatment of resistant Gramregimens containing rifampin and sulbactam. This suggests that positive infections. However, the use of teicoplanin in the treatnon-classical combinations of β-lactam/β-lactamase inhibitor and ment of HAP has not been investigated. Consequently, teicoplanin rifampin should be considered for the treatment of nosocomial should only be considered in selected cases of MRSA where other pneumonia caused by multiresistant A. baumannii. [152] This was alternative treatment is contraindicated.

confirmed by Montero et al. [153] using an experimental pneumonia mouse model with carbapenem-resistant strains: colistin appeared Meropenem is the second carbapenem since the development far less potent in reducing lung bacterial counts, clearance of of imipenem/cilastatin. It shows enhanced Gram-negative activity bacteremia, and survival than imipenem, sulbactam, rifampin and relative to imipenem/cilastatin and often retains activity against tobramycin. [153] From another mouse model study, doxycycline strains resistant to third-generation cephalosporins and imipenem/ plus amikacin was suggested as an alternative to imipenem in the cilastatin (including P. aeruginosa). Furthermore, in contrast to therapy of A. baumannii pneumonia. [154] imipenem/cilastatin it has far less epileptogenic and nephrotoxic activity, making it especially suitable for treatment in patients with Of all the new fluoroquinolones that have been developed since underlying central nervous system pathology or renal dysfuncthe ATS statement, levofloxacin and moxifloxacin have been of particular interest concerning the treatment of CAP. They offer of pulmonary infiltrates and quantitative culture of tracheal aspiexcellent activity against Gram-negative bacilli and improved rates. Singh et al. [159] used this score to discriminate all ICU Gram-positive activity (e.g. against S. pneumoniae and S. aureus) patients in whom VAP was considered unlikely (CPIS score ≤6) over ciprofloxacin. Furthermore, these agents may result in cost from those that were likely to have VAP (CPIS score >6). Patients savings especially in situations where, because of their potent with CPIS ≤6 were randomized to receive either ciprofloxacin for broad-spectrum activity and excellent bioavailability, they may be 3 days or standard care (antibacterials for 10-21 days). After 3 used orally in place of intravenous antibacterials. However, there days patients in the ciprofloxacin group were re-evaluated; all is only limited experience with levofloxacin and moxifloxacin in patients with CPIS >6 received further treatment for pneumonia, the treatment of HAP. In one large study (n = 438) with most but all patients with CPIS ≤6 at 3 days had antibacterial treatment patients on mechanical ventilation, levofloxacin monotherapy stopped. Antibacterials were continued beyond 3 days in 90% (38 proved equally efficacious as imipenem/cilastatin in both microbi-of 42) of the patients in the standard therapy group compared with ologic eradication rate and clinical success rate. [155] The diagnosis 28% (11 of 39) in the experimental therapy group (p = 0.0001). In of HAP was based on clinical symptoms, radiographic findings patients in whom CPIS remained ≤6 at the 3-day evaluation point, and respiratory cultures. However, patients in this study were also antibacterials were still continued in 96% (24 of 25) in the stanincluded based on a positive sputum culture, making the diagnosis dard therapy group but in none (0 of 25) of the patients in the of HAP less reliable. Furthermore, patients with severe disease experimental therapy group (p = 0.0001). Mortality and length of were excluded and the mean APACHE II scores in both groups ICU stay did not differ despite a shorter duration and lower cost of were relatively low. Another point of concern is raised by recent antimicrobial therapy in the experimental group. More surprisingreports that have shown failure of treatment with levofloxacin in ly, antimicrobial resistance, or superinfections, or both, developed several cases of pneumococcal pneumonia. [156, 157] Therefore, the in 14% (5 of 37) of patients in the experimental group versus 38% use of levofloxacin and moxifloxacin in HAP should be restricted (14 of 37) of patients in the standard therapy group (p = 0.017). to selected cases pending the results of future investigations.

Penetration of antibacterials to the site of infection is important in achieving antibacterial concentrations beyond the minimal in-3.2 Duration of Antibacterial Therapy hibiting concentration of the pathogen involved. The ratio between drug concentration in respiratory secretions and serum, the frac-At the time of the ATS statement, few data were available to tional penetration, is about 10-20% for β-lactam antibacterials, support solid recommendations on duration of antibacterial ther-20-40% for aminoglycosides and 50-100% for the fluoroquiapy for HAP. A major step towards tailor-made antimicrobial nolones. [161, 162] Macrolides, tetracyclines and trimethoprim also therapy would be identifying clinical parameters that can indicate show good penetration into bronchial secretions. [163] However, when it is safe to stop antibacterial treatment in any given patient.

controversy remains about which parameter is the most relevant Few studies have investigated this issue. Dennesen et al. [158] studderivative of tissue concentration: antibacterial concentration in ied the time course of several infectious parameters in patients bronchial secretions or alveolar lining fluid. [164] Administering the with VAP after the start of appropriate antibacterial treatment.

correct dosage of antibacterials is of utmost importance in the They found that most improvements in temperature, leukocyte empirical therapy of HAP. Although most antibacterials are marcounts and oxygenation (PaO 2/FIO2 ratio) were observed within keted as 'one dose for all', each patient may require an individualthe first 6 days of antibacterial treatment. Although H. influenzae ized dosage. To ensure good initial antibacterial tissue concentraand S. pneumoniae were eradicated from tracheal aspirates, Entertions, initial antibacterial therapy of HAP should be administered obacteriaceae, S. aureus and P. aeruginosa persisted despite antiintravenously. Switching to oral administration should not be microbial susceptibility. [158] Therefore, eradication of the latter considered before the first signs of clinical improvement. In conmicroorganisms from the lower respiratory tract does not provide trast to research on HAP, several reports on treatment of CAP have a useful means of monitoring the clinical effect of antimicrobial appeared since 1996 investigating an early switch from intravetreatment.

nous to oral antibacterials. [165] [166] [167] [168] [169] Although patients with CAP In another study, Singh et al. [159] used a clinical pulmonary have different microorganisms and a dissimilar pathogenesis at infection score (CPIS) by Pugin et al. [160] to determine the likeli-play, certain parallels can be made. It appears that hospitalized hood that any given patient's clinical findings were caused by (non-ICU) patients with mild CAP can be treated safely with only pneumonia. The score is calculated from temperature readings, a short course (≤ 3 days) of intravenous antibacterials and that a leukocyte counts, purulence of tracheal secretions, oxygenation subsequent treatment of 7 days with oral antibacterials is sufficient (PaO 2/FIO2 ratio), aspects of pulmonary radiography, progression in most cases.

Although insufficient data are available to make solid recom-the possibility of other noninfectious processes. [171] In cases of mendations on the duration of treatment in patients with HAP, raised suspicion of resistant pathogens, fresh endobronchial specifrom the aforementioned studies several points of attention can be mens for culture should be obtained. made. Because no gold standard exists on making the diagnosis of Certain factors may play a role in the failure of treatment. HAP and because physicians are unwilling to risk missing a Although a bacterial etiology of HAP is the most frequent, viruses treatable infection, a large proportion of patients diagnosed with are also potential pathogens. [172] [173] [174] Clinically, viral pneumonias HAP will either receive too many antibacterials or do not need are difficult to differentiate from bacterial pneumonias. Since viral antibacterial treatment at all. Furthermore, the total duration of agents are not routinely tested for, the incidence of viral pneumoantibacterial treatment and the duration of intravenous antibacterinias is almost certainly underestimated. [173] Influenza and respiraal treatment can probably be shortened in many cases of mild tory syncytial virus infections contribute substantially to the mor-HAP, especially in cases of HAP caused by S. pneumoniae or bidity and mortality associated with viral pneumonia, especially in H. influenzae, and in patients showing normalization of temperayoung children and the elderly. [174, 175] Other viruses associated ture, leukocyte counts and oxygenation.

with HAP are parainfluenza virus, adenovirus, varicella-zoster virus, cytomegalovirus, herpes virus and measles. [176] Patients at 4. Failure of Antibacterial Treatment of HAP risk for developing serious pneumonia with these viruses are neonates and immunocompromised patients. Extra suspicions for Even with optimal and adequate empirical antibacterial therapy a viral etiology of HAP should be raised during and following the HAP remains a disease with high mortality. In most patients, clear annual community outbreaks of influenza and RSV.

[177] clinical improvement should not be expected within 24 hours after Other microorganisms that should be excluded in patients with the start of antibacterial treatment. Dennesen et al. [158] have shown treatment failure or deterioration despite antibacterial treatment that leukocyte counts, temperature and oxygenation normalize, on are L. pneumophila and fungi. Especially in severe pneumonia, average, 6 days from the start of antibacterials. However, the first negative urine antigen tests for L. pneumophila can almost rule out signs of improvement of clinical parameters were found within 48 (nosocomially acquired) legionnaires' disease. In rare cases of hours. Consequently, failure of treatment cannot be established infections with type 2 L. pneumophila, and in mild cases, antigen within 24 hours after start of therapy. However, in severe cases of tests lack sensitivity and can give a (false) negative result. [178, 179] VAP or in patients on mechanical ventilation because of HAP, Patients with neutropenia or patients receiving intensive treatment who experience ongoing deterioration of clinical parameters, with corticosteroids (e.g. patients with COPD) are at risk for more switching antibacterial treatment to a more broad-spectrum regiacute presentations of Aspergillus spp. infections. However, even men should first be considered after 24 hours. This is especially in non-immunosuppressed, non-neutropenic patients, severe VAP true in patients for whom microbiological results remain inconcluresulting from Aspergillus spp. has been reported. [180] Consequentsive. In mild to moderate cases, changing antibacterial therapy ly, especially in high-risk patients with extended hospital stay and should not be considered until at least 48 hours have passed since failure of antibacterial treatment, cultures for fungi of BAL or PSB the start of therapy. The CPIS score can help to monitor the specimens, and serum fungal antigen tests should be considered. development of VAP (or patients with HAP on mechanical ventilation) after start of therapy. Luna et al. [170] have shown that in patients who received adequate antibacterial treatment (no resis-5. Conclusions tant microorganisms were cultured) and recovered from VAP, the CPIS score was significantly improved at day 3. [170] In contrast, Both the ATS statement and the many reports on the diagnosis patients who had inadequate antibacterial treatment and did not and treatment of HAP that have appeared since then, have focused survive VAP, did not display improved CPIS scores at day 3. In on patients admitted to ICUs and/or those on mechanical ventilatwo-way analysis of variance with repeated measures, clinical tion. Although this subgroup consists of patients with severe improvement or worsening was most accurately depicted by the nosocomial pneumonia, more research on non-ICU patients is PaO 2/FIO2 ratio. In another study with 298 patients with HAP, needed to explore whether the same guidelines apply to less severe multivariate analysis revealed that six variables were associated cases of HAP. From the large body of literature on the diagnosis of with decreased likelihood of clinical success; >65 years of age, HAP it can be concluded that bronchoscopic techniques, in combichronic lung disease, diabetes mellitus, mechanical ventilation for nation with quantitative cultures, are superior to other more con->5 days, multilobar pneumonia, and bacteremic pneumonia. [127] servative techniques to obtain respiratory samples and to establish Patients who fail to respond, or experience clinical deterioration, early diagnosis. Detection of intracellular microorganisms in lowshould be re-examined carefully, and thought should be given to er respiratory tract specimens has been demonstrated to be a 

Epidemiology of hospital-acquired pneumonia

Analysis of charges associated with

Infect diagnosis of nosocomial pneumonia: can routine bronchoscopy be justified?

Epidemiology and risk factors for nosocomial pneumonia: emphasis 27

Hospital-acquired pneumonia in adults: diagnosis assessment of severity initial scopic methods for microbiological specimen sampling

Usefulness of quantitative

Nosocomial pneumonia on general medical cultures of BAL fluid for diagnosing nosocomial pneumonia in ventilated and surgical wards in a tertiary-care hospital

Using bronchoalveolar lavage to distin

Risk factors for nosocomial pneumonia: guish nosocomial pneumonia from systemic inflammatory response syndrome: focus on prophylaxis

Mortality of nosocomial pneumonia in a cohort study evaluating attributable mortality and hospital stay

Specificity of endotracheal aspiration, mortality and morbidity

Pulmonary infiltrates in the surgical ICU: ed patients

Protected specimen brush or bronchoalveo

Causes of fever and pulmonary meta-analysis

Diagnosing ventilator-associated pneumonia: monia

The chest radiograph in critically ill surgical patients is inaccurate in predicting ventilator-associated pneumonia

diagnosis of ventilator-associated pneumonia

The effects of four antimicrobial drug regimens on sputum tor-associated pneumonia: executive summary

Nosocomial pneumonia in patients receiving

Am J Respir Crit Care Med continuous mechanical ventilation: prospective analysis of 52 episodes with use

Role of quantitative cultures of endotracheal aspirates in the diagnosis of nosocomial pneumonia

Pneumonia in acute respiratory distress 1230-5 syndrome: a prospective evaluation of bilateral bronchoscopic sampling

Value of intracellular bacteria detection versus piperacillin-susceptible organisms

Risk factors for an outbreak of multi-378-84 drug-resistant Acinetobacter nosocomial pneumonia among intubated patients

tion of protected specimen brush samples in nosocomial pneumonia

Role of bronchoalveolar lavage in mechanicaland methicillin-susceptible organisms

Usefulness of microscopic examination pneumonia

Prediction of clinical severity and

Comparison of direct examination of three physiology score with systemic inflammatory mediators. Am J Respir Crit Care types of bronchoscopy specimens used to diagnose nosocomial pneumonia

The treatment of respiratory pseudomonas infection in

Gram stain of bronchoalveolar cystic fibrosis: what drug and which way?

lavage fluid in the early diagnosis of ventilator-associated pneumonia. Br J 67. Spencer RC. The emergence of epidemic, multiple-antibiotic-resistant Steno

Rapid diagnosis of bacteremic pneumocepacia

Legionella: a major opportunistic pathogen in transplant urinary antigen test: a prospective, controlled clinical evaluation

Management of fever in patients with cancer and treatment-induced

Evaluation of the immunochroneutropenia

s principles and urinary antigen in a prospective study of community-acquired pneumonia in practice of infectious diseases

Role of serial routine microbiologic culture 71. Walzer PD. Pneumocystis carinii pneumonia in patients without human immuresults in the initial management of ventilator-associated pneumonia

Risks and routes for ventilator-associated

Mortality associated with late-onset pneumonia with Pseudomonas aeruginosa

the intensive care unit: results of a multi-center cohort study

Risk factors for early-onset

Variations in etiology of ventilatorpneumonia in critical care patients: selected multiresistant versus nonresistant associated pneumonia across four treatment sites: implications for antimicrobial bacteria

Structural features of tracheal tube biofilm

Survival in patients with nosocomial pneumonia: formed during prolonged mechanical ventilation

Mortality rate attributable to ventilator-associated

nosocomial pneumonia in an adult intensive care unit: a prospective case

Nosocomial pneumonia likely caused by control study

Stenotrophomonas maltophilia in two patients with polymyositis

A study of the etiologies and treatment of nosocomial

Ventilator-associated pneumonia caused patients

by potentially drug-resistant bacteria

Comparison of mortality associated with methicillin-resistant and methicillin-susceptible Staphylococcus aure

A comparative analysis of patients with us bacteremia: a meta-analysis

Surveillance of methicillin-resistant

Aetiology of pneumonia following isolated Staphylococcus aureus (MRSA) and Enterobacteriaceae producing extendedclosed head injury

Nosocomial pneumonia: a multivariate analysis of risk and prognosis

Risk factors associated with nosocomial methicillin

Nosocomial pneumoresistant Staphylococcus aureus (MRSA) infection including previous use of nia in patients undergoing heart surgery

Clinical and epidemiological findings in 104. Kollef MH. Inadequate antimicrobial treatment: an important determinant of mechanically-ventilated patients with methicillin-resistant Staphylococcus auoutcome for hospitalized patients

Clinical and economic evaluation of subse

Nosocomial infections in the ICU: the growing quent infection following intravenous ciprofloxacin or imipenem therapy in importance of antibiotic-resistant pathogens

Incidence, risk, and prognosis factors of resistance and its relationship to extended-spectrum beta-lactamase production nosocomial pneumonia in mechanically ventilated patients. Am Rev Respir Dis in Klebsiella pneumoniae isolates causing bacteremia

Extended-spectrum beta-lactamaseresistance: risk factors for mortality in the intensive care unit

Beta-lactamase-mediated resistance and opportunities for its conof appropriate antibiotic treatment for ventilator-associated pneumonia. Chest trol

Guidelines for the management of taxonomy, typing methods, and pathogenicity factors

Clinical characteristics and molecular epidemiology associated with imipenem-resistant Klebsiella pneumoniae. Clin 110

Spread of extended-spectrum beta-lactamaseproducing Klebsiella pneumoniae: are beta-lactamase inhibitors of therapeutic 111. Bouza E, Cercenado E. Klebsiella and Enterobacter: antibiotic resistance and value?

Clinical significance and impact on 112. Sanders CC, Sanders WE. β-Lactam resistance in gram-negative bacteria: global mortality of extended-spectrum beta lactamase-producing Enterobacteriaceae trends and clinical impact

Outbreak of a multiresistant selected nosocomial gram-negative bacteria in the United States. J Infect Dis Klebsiella pneumoniae strain in an intensive care unit: antibiotic use as risk

Enterobacter bacteraemia: clinical features (1): 14-20 and emergence of antibiotic resistance during therapy

anaerobic bacteria in ICU patients: therapeutic considerations and outcome

Bacteriology of hospital-acquired pneumonia

Anaerobic bacterial infections of the lung

Treatment of aspiration pneumonia and primary lung 118. Chow AW, Bednorz D. Comparative in vitro activity of newer cephalosporins abscess: penicillin G vs clindamycin

Effect of enriched thioglycolate on direct 119. Schreiner A. Use of clindamycin in lower respiratory tract infections. Scand J examination of respiratory specimens and guiding initial empirical therapy in Infect Dis

Metronidazole vs clindamycin treatment of anaerobic pulmonary

Incidence of anaerobes in ventilator-associated 1424-7 pneumonia with use of a protected specimen brush

Pathogen of occurrence and pneumonia and aspiration pneumonia: a prospective study

Respiratory tract pathogens isolated from patients implications

Nosocomial legionnaires' disease discovered rence and antimicrobial susceptibility profile: results from the SENTRY Antiin community hospitals following cultures of the water system: seek and ye microbial Surveillance Program

Staphylococcus aureus strains heterogeneously and intermediately resistant to vancomycin are not disseminated throughout Japanese hospitals

by keeping the circulating hot water temperature above 55 degrees C: experience from a 10-year surveillance programme in a district general hospital. J 125. Fagon JY, Maillet JM, Novara A. Hospital-acquired pneumonia: methicillin resis-Hosp Infect

Microbial etiology of acute pneumonia 126. Lamb HM, Figgitt DP, Faulds D. Quinupristin/dalfopristin: a review of its use in in hospitalized patients

pneumonia: prospective randomized comparison of quinupristin/dalfopristin versus vancomycin. Nosocomial Pneumonia Group

Bacteraemia due to Stenomycin, dirithromycin, quinupristin/dalfopristin and erythromycin against Letrophomonas maltophilia: an analysis of 45 episodes

Stenotrophomonas maltophilia and Burkholderia

Mandell, Douglas, and levofloxacin, and erythromycin against Legionella species by broth microdilu-Bennett's principles and practice of infectious diseases. Philadelphia (PA): tion and intracellular susceptibility testing in HL-60 cells

beta-lactams, beta-lactamase inhibitors, and rifampin against Acinetobacter baumannii in a mouse pneumonia model

Ventilator-associated pneumonia due

Mortality due to ventilator-associated aminoglycosides, and rifampin as monotherapy in a mouse model of pneumonia pneumonia or colonization with Pseudomonas or Acinetobacter species: assesscaused by multiresistant Acinetobacter baumannii. Antimicrob Agents ment by quantitative culture of samples obtained by a protected specimen

Problem pathogens (Pseudomonas aeruginosa and amikacin in monotherapy and in combination in Acinetobacter baumannii Acinetobacter)

Levofloxacin compared with imipenem/ piperacillin-tazobactam, and ciprofloxacin alone and in combination against cilastatin followed by ciprofloxacin in adult patients with nosocomial pneumo-Staphylococcus aureus, Klebsiella pneumoniae, Enterobacter cloacae, and nia: a multicenter, prospective, randomized, open-label study

Hospital-acquired pneumonia: risk factors, microbiology, and treat-156

Antimicrobial resistance amongst Klebsiella spp. 747-50 collected from intensive care units in Southern and Western Europe in 157

Resolution of infectious tance of gram-negative and gram-positive bacteria from patients in intensive parameters after antimicrobial therapy in patients with ventilator-associated care units of 25 European university hospitals participating in the European arm pneumonia

Short-course empiric antibiotic therapy for patients with pulmonary infiltrates in the intensive care unit: a proposed

PNU-100766) versus solution for indiscriminate antibiotic prescription. Am J Respir Crit Care Med vancomycin in the treatment of hospitalized patients with nosocomial pneumo-2000

Diagnosis of ventilator-associated pneumonia by bacteriologic analysis of bronchoscopic and nonbronchoscopic 'blind

double-blind, multicenter study of linezolid versus vancomycin in the treatment of patients with nosocomial pneumonia

Systemic antibiotic treatment of nosocomial pneumonia

Linezolid versus vancomycin for the treatment of methicillin-resistant Staphylococcus aureus infections

Worldwide assessment of linezolid's clinical safety and tolerability: comparator-controlled phase III studies

Drug treatment of pneumonia in the hospital: what are the timicrob

Penetration of netilmicin in the resistant staphylococci infections: the role of quinupristin/dalfopristin. J Anlower respiratory tract after once-daily dosing

A prospective randomized study of

In vitro activity of linezolid, quinupristin-dalfopristin, vancomycin, teinpatient iv. antibiotics for community-acquired pneumonia: the optimal duraicoplanin, moxifloxacin and mupirocin against methicillin-resistant Staphylotion of therapy

therapy for community-acquired pneumonia: duration of therapy

Meropenem, a new carbapenem antibiotic. Pharmacothercomes, and cost analysis

Efficacy and 145. Craig WA. The pharmacology of meropenem, a new carbapenem antibiotic. Clin safety of oral and early-switch therapy for community-acquired pneumonia: a

Efficacy of meropenem as monotherapy in the treatment of ventilator-168

tested against over 2773 bacterial isolates from hospitalized patients with pneumonia: I--results from the SENTRY Antimicrobial Surveillance Program 169

Resolution of ventilator-associated the treatment of nosocomial pneumonias in ventilated patients. Multicenter pneumonia: prospective evaluation of the clinical pulmonary infection score as group study

an early clinical predictor of outcome

Med Clin North admitted during an epidemic season

Impact of management guidelines on the outcome of update of epidemiology and risk factors

A recurrent outbreak of nosocomial

Nosocomial viral pneumonia in the intensive care legionnaires' disease detected by urinary antigen testing: evidence for longunit

CDC guidelines focus on prevention of nosocomial pneumonia

Nosocomial ventilator-associated pneumonias

Hospital-acquired pneumonia in children: the role of respiratory viruses. caused by Aspergillus fumigatus in non-immunosuppressed, non-neutropenic

Nosocomial pneumonia in pediatric patients: practical problems and rational solutions