James H. Williams, Jr.
BACKGROUND AND ETIOLOGY
The Gram-negative bacilli (GNB) Pseudomonas aeruginosa, Escherichia coli, and organisms of the Proteus, Serratia, Enterobacter, and Acinetobacter species are most commonly associated with nosocomial (hospital-acquired) pneumonia (HAP), including ventilator-associated pneumonia (VAP). GNB are associated with less than 20% of pneumonias among ambulatory patients with community-acquired pneumonia (CAP). However, these GNB are more commonly recovered from the airways of debilitated, institutionalized patients with pneumonia, who are included among those defined as health care–associated pneumonia (HCAP). GNB are associated with as many as half the deaths from bacterial pneumonia in these patients, and are common pathogens identified in the airways of hospitalized patients.
Predisposing factors for GNB pneumonia vary with the population at risk. In the community, chronic bronchitis, bronchiectasis, alcoholism, diabetes, altered mental status, and neutropenia appear to be the major risk factors. Prior antibiotic selection pressure contributes to the emergence of these organisms. In the hospital, GNB pneumonia most often occurs with prolonged intubation, including tracheostomy. Even without intubation, prolonged hospital stay (particularly in the ICU), recent thoracic or abdominal surgery, advanced age, and severe underlying illness all are risk factors. During acute and chronic illnesses, patients more often have relatively lower levels of some micronutrients (e.g., selenium, zinc), potentially resulting in altered host responses, while more frankly immune-compromised hosts are at particular risk for adverse outcomes from these infections.
Although contaminated respiratory equipment has caused occasional outbreaks, particularly those due to Serratia and Pseudomonas species, these outbreaks are uncommon with the use of disposable equipment and aseptic techniques. Medical staff can facilitate colonization of patients with potentially resistant organisms by careless hygiene, which can be diminished by careful cleansing of hands before and after patient contact and avoiding contact with commonly shared fomites (e.g., stethoscopes, door handles, bed controls, etc.) without cleansing. GNB pneumonias may result from bacteremia introduced by bladder catheters, intravenous catheters, or infections in the abdomen or elsewhere. However, GNB causing HCAP/HAP/VAP are more commonly delivered to the lungs via the airways.
Colonization of the upper airways, including the pharynx and nasal sinuses, by GNB generally precedes pneumonia. Nasal tubes increase retention of secretions in the nasal sinuses and drain into the posterior pharynx. Selection of GNB is encouraged in a hospital by a number of additional factors commonly encountered in the ICU, particularly antibiotic selection pressure, increased adherence of GNB to the airway epithelium, and retained secretions in seriously ill patients. Reflux of gastric contents into the posterior pharynx also can contribute, particularly in the supine position and with larger gastric volume. Suppression of gastric acidity selectively promotes GNB proliferation in the stomach, although the magnitude of this effect on the development of GNB HAP/VAP has been debated. Instillation of medications and nutrition via nasogastric tubes likely enhances the risk of aspiration, while passing feeding tubes beyond the stomach (postpyloric position) may reduce risk.
Access of nasopharyngeal flora to the lower airways is facilitated by a number of factors inherent to ICU patients, particularly intubation. Translaryngeal intubation mechanically holds open the epiglottis and vocal cords. Although cuffed endotracheal tubes (ETTs) diminish the rate at which large volumes can enter the lower airways, the reservoir of secretions above the cuff continue to ooze down around the cuff, which is kept at low pressure to avoid tracheal necrosis. Efforts to diminish the size of this subglottic secretion pool with specially designed ETTs have produced variable results, perhaps reflecting in part the tenacious character of and limited access to subglottic secretions. Bacteria adherent to the ETT provide an additional nidus for infection, and silver impregnated tubes may limit recovery of organisms in culture that would suggest VAP. However, these approaches have not been reliably demonstrated to translate into a reduction in overall mortality.
The normal reflex clearance of airway secretions is attenuated by many factors as well. Endotracheal and tracheostomy tubes create a smaller lumen through which to expectorate. The effectiveness of expectoration is also inhibited by CNS depression (e.g., narcotics, sedatives, metabolic instability, CNS lesions), local reflex depression (e.g., topical anesthetic, learned tolerance of foreign nasotracheal or nasogastric tubes), and pain (particularly chest and abdominal surgery). Mucociliary activity can be decreased (e.g., alcohol, chronic inflammation, metabolic disorders), and phagocytic activity can be impaired (e.g., immunocompromised patients, alcohol, overwhelmed reserves).
Tracheostomy has advantages for patients requiring prolonged intubation, including stabilization of airway access, patient comfort, and less physical interruption of airway closure during swallowing. However, tracheostomy also delays reestablishment of normal airway architecture during recovery. After weaning from ventilator support, clearance of secretions is still inhibited by diminished ability to generate a high positive pressure for cough and by flow limitations of the tube. Airway protection is also impaired by limiting the normally generated positive airway pressure during swallowing, by potentially hindering tracheal lift for closure of the epiglottis, and by applying pressure to the upper esophagus through the membranous posterior tracheal surface. These factors, along with the underlying problems of these patients, lead to a high incidence of recurrent pneumonia, often with GNB in this setting.
CLINICAL PRESENTATIONS
The clinical features of GNB infections are intertwined with the underlying diseases with which they are usually associated. The classic descriptions of GNB pneumonias focus on community-acquired cases, uncomplicated by adult respiratory distress syndrome (ARDS), heart failure, or fluid imbalance and, therefore, incompletely represent the spectrum of nosocomial GNB pneumonias. Regardless, they provide useful comparisons of pathologic responses in relatively fit individuals. In contrast, immunocompromised patients may exhibit relatively few signs or symptoms and less evidence of infiltrate on chest film. Even more difficult are patients with underlying acute or chronic lung injury who have similar signs and symptoms and areas of increased lung tissue density from prior injury. The presence of GNB organisms in such patients may reflect either colonization or acute infection, making diagnosis of acute GNB pneumonia more challenging.
P. aeruginosa frequently colonizes the skin or mucosa of patients as well as the hospital environment (soap, liquid media, and hospital staff). It can colonize or infect tracheostomy sites, burns, wounds, and the urinary tract as well as the lower airways of patients with chronic bronchitis/bronchiectasis. Mucoid strains often emerge in the airways of patients with cystic fibrosis. Pneumonia is usually acquired via the airway and tends to be more prominent in dependent lung zones, whereas hematogenous infections may lead to more widespread changes. Pathologically, severe focal necrosis may be seen with nodular infarcts and vessel wall necrosis leading to hemorrhage and formation of small cavities. Purulent pleural effusions are more often found at autopsy. Clinically, patients often appear toxic, presenting with chills, fever, and dyspnea; sputum often is copious and can be blood-tinged. Pleuritic chest pain is less common. Ecthyma gangrenosum is an uncommon cutaneous maculopapular eruption representing infection and necrosis in vessel walls and may present as hemorrhagic bullae, ulcers, or nodular lesions. Though historically linked to bacteremia with these organisms, it can be seen with other infections. Radiographically, consolidation in dependent areas is most common and it is classically associated with abscesses varying in size from 2 to 11 cm. Small effusions also may be present. Bilateral patchy infiltrates or bilateral nodules are occasionally seen with hematogenous infection.
E. coli pneumonia may follow aspiration or hematogenous dissemination from urinary tract or gastrointestinal infections. Pathologically, a diffuse, hemorrhagic pneumonia is often present, but abscess formation is less common. Clinically, patients often appear toxic, with fever, dyspnea, productive cough, and, more often, pleuritic chest pain. Classically, one may see a relative bradycardia for the degree of temperature elevation and a paucity of signs of parenchymal consolidation. The chest roentgenogram usually demonstrates patchy bronchopneumonia, often in the lower lobes. Pleural effusion may be present.
Proteus species are less common causes of respiratory tract infection, frequently associated with altered consciousness, potentially leading to aspiration. Pathologically, the pneumonia is hemorrhagic and associated with small abscesses. Clinically, patients usually appear less toxic, although chills, fever, dyspnea, productive cough, and pleuritic chest pain may be present. The chest roentgenogram demonstrates dense infiltrates, more often in the dependent segments of the upper lobes and superior segment of the lower lobes, and volume contraction may be seen. Pleural effusion is less common.
Serratia species occasionally cause pneumonia. Clustered cases have been linked in the past to contaminated respiratory equipment. Pathologically, diffuse bronchopneumonia can occur with small (2–3 mm) abscesses. Patients typically are toxic, with fever, chills, and productive cough. Pseudohemoptysis, the production of sputum tainted with a red pigment produced by some strains, is classically described but is uncommon. The chest radiograph often demonstrates diffuse, patchy, bronchopneumonia similar to Pseudomonas pneumonia, although abscess formation has been reported less frequently. Pleural effusion and empyema may occur.
Enterobacter pneumonia is less well characterized than the other GNB pneumonias. In one small series, symptoms included fever, dyspnea, and cough productive of yellow sputum, but pleuritic pain was uncommon. Chest radiographs most often demonstrate bilateral broncho-pneumonia, but abscesses and empyema formation are uncommon. The emergence of drug-resistant strains has increased the frequency and seriousness of infection from these organisms.
Acinetobacter species have emerged more recently as multidrug-resistant (MDR) organisms associated with HAP/VAP, likely in response to the prolific use of broad-spectrum antibiotics. Colonization of hospitalized patients with these organisms has been observed with increasing frequency. The presence of MDR Acinetobacter species in airway cultures of febrile intubated patients with pulmonary infiltrates presents diagnostic and therapeutic dilemmas. The chest roentgenogram may demonstrate multilobar infiltrates, occasionally with signs of necrosis (cavitation) or effusion.
DIAGNOSIS
The diagnosis of Gram-negative pneumonia by examination of airway secretions is problematic because of the frequency with which GNB colonize the airways of patients at risk, many of whom have infiltrates on chest radiographs for other reasons or infiltrates obscured behind the diaphragms and mediastinal shadows on portable anteroposterior (AP) films. Demonstration of numerous GNB and neutrophils on smears of airway secretions collected via expectoration or suctioning provides presumptive evidence of infection, particularly with intracellular organisms indicating evidence of a host response; however, this still may reflect bronchitis rather than pneumonia. Attempts to reach beyond the upper airway with bronchoscopic brushing or lavage are complicated by the fact that the upper airways are traversed in the process, contaminating the sampling channel during suctioning to maintain visualization. Fewer organisms sometimes recovered with these techniques may simply reflect a smaller or diluted sample. Therefore, while these samples provide a basis for a presumptive diagnosis, invasion by these organisms is more firmly supported by positive cultures from blood or pleural fluid. While routine use in this setting is debated, bronchoscopy can provide important access to samples in patients who cannot provide adequate lower airway samples and sometimes demonstrates organisms not reported in cultures of upper airway samples laden with other bacteria. Demonstrating a clear predominance of one organism increases confidence of its role in a respiratory infection, but recovery of multiple organisms does not preclude a role for one or more in an apparent pneumonia. Occasionally, if the patient is deteriorating while undergoing treatment, demonstration of tissue infection via lung biopsy may be warranted.
TREATMENT
Decisions about when and how to treat HCAP/HAP/VAP are sometimes difficult. Published guidelines provide useful direction, but clinical judgment is still required in individual cases. Issues related to diagnostic uncertainty complicate confidence in choosing antibiotics, and initiating antibiotics immediately adds selection pressure for emergence of drug-resistant strains. However, the high mortality associated with HAP/VAP with GNB is worsened by delay in starting antibiotics to which the organism is sensitive, particularly among those admitted to the ICU. This argues for aggressive, broad-spectrum empiric coverage initially, and against limited coverage while waiting for results of cultures, particularly among those who are critically ill.
The initial drug or drugs chosen should be based on current resistance patterns in the hospital and organisms likely to be present based on the patient’s presentation and risk profile. For example, Pseudomonas is more likely among patients with VAP and those previously receiving antibiotics, particularly those with recurrent respiratory tract infections associated with bronchiectasis. To limit antibiotic selection pressure, current opinion favors a “de-escalation” strategy, starting with relatively broad coverage, particularly in critically ill patients, and then narrowing subsequent coverage based on the results of cultures. However, subsequent de-escalation can be worrisome when a patient is improving slowly on an initial regimen, particularly if there is uncertainty regarding sensitivity of cultures and in very sick patients with limited ability to tolerate deterioration. Recognizing that repetitive use of standardized regimens induces resistance, some rotate every few months the classes of drugs selected empirically. Underlying all such strategies is a desire to reserve certain classes of drugs for use with subsequent infections in both individuals and the institution and to reduce the emergence of multidrug-resistance profiles, although evidence of outcome benefit from specific approaches remains limited and continued investigation is needed.
When Pseudomonas species are demonstrated or suspected, two classes of drugs have been used simultaneously because drug resistance often emerges in these organisms. This may be particularly important with prior exposure to antipseudomonal antibiotics of the same class. Traditionally, a high dose semisynthetic penicillin (e.g., piperacillin) has been combined with an aminoglycoside (e.g., tobramycin, gentamycin, amikacin). In recent years, high doses of fluoroquinolones (e.g., ciprofloxacin, possibly levofloxacin) are often substituted for aminoglycosides, particularly among older, sicker patients at increased risk of ototoxicity and nephrotoxicity. Of note, while adding a β-lactamase inhibitor (e.g., piperacillin-tazobactam—pip/tazo) enhances piperacillin efficacy against many organisms, this is not true for Pseudomonas species. Therefore, the pip/tazo combination is usually recommended at higher doses (4.5 g every 6 hours) for Pseudomonas
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
