Acinetobacter baumannii: An Emerging and Important Pathogen

Abstract

Objective

To review the clinical significance, management, and control of Acinetobacter infections.

Methods

Literature review.

Results

Acinetobacter infections have become a major cause of hospital-acquired infections worldwide. Acinetobacter is noted for its ability to survive for long periods on hospital surfaces and equipment, its predilection to develop resistance to multiple antibiotics, its affinity to cause serious infections in critically ill patients, and many well described outbreaks attributable to contamination of a common source. The crude ICU mortality is approximately 40%. Rigorous antibiotic stewardship and infection control measures are critical to prevent the spread of multidrug-resistant Acinetobacter infections. There is also a pressing need for new therapeutic options.

Conclusion

Acinetobacter is an emerging pathogen of increasing significance.

Over the last 3 decades, Acinetobacter infections have grown from a limited problem affecting disaster victims and tropical populations to a major cause of hospital-acquired infections worldwide. The most clinically significant pathogen is Acinetobacter baumannii because of the rapidity with which it develops antimicrobial resistance (including strains resistant to all commercially available antibiotics) and the ability of some strains to survive on the surfaces of hospital facilities and equipment for weeks. Acinetobacter can cause a wide variety of clinical syndromes, most prominently respiratory tract infections but also bacteremia, skin and soft tissue infections, urinary tract infections, osteomyelitis, and intracranial infections. The mortality associated with A. baumannii infection in the intensive care unit (ICU) setting can approach 40%. Due to the limited therapeutic options for multidrug-resistant Acinetobacter infections, prevention of health care–associated transmission is critical to prevent morbidity. There is also an urgent need to develop novel therapeutic agents active against multidrug resistant strains.

Taxonomy

Identification of species of the Acinetobacter genus is based mostly on the absence of characteristics common to other gram-negative organisms. The various names it has been assigned since its discovery approximately 100 years ago reflect its relative anonymity: Micrococcus (small), anitratus (nitrate non-reducing), and Acinetobacter (motionless). The genus Acinetobacter tends to grow well on routine solid media, such as sheep blood agar at 37°C. Colonies are 1 to 2 mm, domed, mucoid, and nonpigmented. They may be presumptively identified as aerobic, gram-negative, catalase-positive, oxidase-negative, nonmotile, nonfermenting coccobacilli. Acinetobacter is notorious, however, for variable morphology, differential uptake of Gram stain, and relative lack of reactivity on biochemical assays (Figure). The organism is typically rod-shaped during rapid growth but forms coccobacilli during stationary phase. Uptake of crystal violet varies; hence, the organism is sometimes mis-classified as gram-positive cocci [1]. Differentiation of species within the genus based on phenotype is difficult. Therefore, clinical microbiological laboratories usually divide the genus into “complexes” with A. baumannii–A. calcoaceticus complex (ABC) being the most clinically relevant. Approximately 25 different “genomic species” have been identified based on DNA-DNA hybridization studies. A. baumannii is genomic species number 2 [2].

Figure.

Variable morphology and staining of Acinetobacter species. The organism appears as gram-negative cocci and bacilli in the top figure A (Gram stain of sputum) but as gram-positive bacilli in the bottom figure B (Gram stain of blood). Both specimens came from the same patient during a single episode of illness. (Reprinted with permission from Kortepeter MG, Lemmon JW, Moran KA. A soldier with traumatic brain injury and meningitis. Clin Infect Dis 2005;41:1604–5, 1675–6.)

Epidemiology

The natural reservoir of A. baumannii remains to be determined. In general, Acinetobacter species are unique in their ability to use various metabolic pathways and carbon sources, and thus can survive in diverse environments including animate and inanimate surfaces. Many members of the genus are ubiquitous in soil and water, can colonize human skin, and are frequently found on hospital surfaces and hospitalized patients. Skin carriage of Acinetobacter species has been implicated as a cause of nosocomial outbreaks of infection [3]. However, an important epidemiological study found that most people are typically colonized with Acinetobacter species other than A. baumannii. In this study, samples were obtained from the skin and mucous membranes of 40 hospitalized patients and 40 community- dwelling healthy controls. Species identification was achieved through various methods (gel electrophoresis, ribotyping, and DNA-DNA hybridization). Acinetobacter species were isolated from 43% of community dwellers and 75% of hospitalized patients, particularly those with longer durations of hospitalization. The most frequently isolated species, however, were A. lwoffii (58%), A. johnsonii (20%), A. junii (10%), and Acinetobacter genomic species 3 (6%). A. baumannii was isolated from only 1% and 3% of skin samples from hospitalized patients and community dwellers, respectively [4]. Although A. baumannii is not definitively a normal inhabitant of human skin, it has been found on body lice. A study in France isolated 40 A. baumannii strains from body lice. A. baumannii DNA was later detected in 21% of 622 lice collected worldwide, suggesting that A. baumannii is endemic to human body lice [5].

Acinetobacter species have been implicated in infectious clusters amongst military personnel injured in battles abroad. Between January 2002 and August 2004, multidrug-resistant ABC was isolated from blood samples of 102 veterans of Iraq-Afghanistan combat who were hospitalized in military medical facilities in Iraq [6]. The epidemiologic source of these infections has yet to be definitively identified; possible sources include skin colonization prior to injury, traumatic inoculation with soil during combat, or nosocomial transmission. Studies have found that as many as 17% of healthy U.S.–based soldiers do have skin colonization with ABC; however, their strains differ from those recovered from injured soldiers [7]. A. baumannii has not routinely been isolated from soil and water reservoirs in Iraq [8]. The most widely accepted theory, therefore, is that epidemics of ABC in soldiers wounded abroad are primarily attributable to nosocomial transmission. An investigation of an outbreak of ABC in soldiers from Operation Iraqi Freedom by Scott and colleagues (2007), for example, demonstrated skin colonization with ABC in only 1 of 64 U.S. patients (2%) but 4 of 38 Iraqi patients (11%) [9]. ABC was isolated from only 1 soil sample, and it was not genetically linked to any of the clinical isolates. However, ABC organisms were isolated from patient treatment areas in all 7 field hospitals sampled. Among the 36 ABC isolates, most (68%) were collected in critical care treatment areas (11 from the ICU, 7 from operating rooms, and 7 from emergency departments). In patient care areas, ABC isolates were recovered from operating room equipment, including anesthesia machines (n = 2), operating room tables (n = 1), light fixtures (n = 1), heaters/air conditioners (n = 2), patient beds (n = 12), sinks (n = 8) and tent walls (n = 2). Using pulsed field gel electrophoresis, 5 ABC clusters were identified. Strain typing of clinical isolates in military hospitals located within the United States confirmed suspicions of nosocomial transmission: 3 veterans not deployed in Iraq had infections with isolates genetically related to an isolate recovered from a field hospital in Iraq.

The finding that skin colonization was higher in non–U.S. soldiers than in U.S. soldiers raised the possibility that these individuals may serve as a reservoir or as an initial source for introducing the organism into U.S. field hospitals [9]. A study by Moran et al of culture results in non–U.S. casualities in Iraq supported this hypothesis. They found that gram-negative bacteria were the most frequently isolated pathogens and ABC was the most commonly isolated organism after Klebsiella [10].

The observation that A. baumannii colonizes the skin of those from temperate climates more than those in colder climates [1] accords with the observed seasonal variation of Acinetobacter infections in the United States. Using data from the National Nosocomial Infection Surveillance System, Mc-Donald and colleagues evaluated 3447 Acinetobacter infections in adults and children in ICUs between 1987 and 1996. They found infection rates were 50% higher during July–October compared with other times of year. This seasonal variation is thought to be the result of changes in climate, since warm weather increases the number of Acinetobacter species in the natural environment, and this may affect the hospital environment [11]. Another possible explanation for seasonality is condensation from air conditioning units. For example, in an investigation of a seasonal outbreak of Acinetobacter infections in a neonatal nursery, Acinetobacter isolates were recovered from the air of the nursery and the vents of nursery air conditioners [12].

A review of published Acinetobacter outbreaks between 1977 and 2000 identified 13 outbreaks attributable to a common source. Nine of the 13 outbreaks were attributed to ventilator components, particularly ventilator circuits and resuscitation bags. Sterilization of reusable components was associated with outbreak resolution in the majority of cases [13]. During an investigation for increasing Acinetobacter colonization of pediatric patients in a burn unit, A. baumannii was isolated from the plastic covers of computer keyboards. Gloved and ungloved staff shared the computers and cross-transmission appeared to occur with keyboard use between patients. Improved hand hygiene and scheduled keyboard cleaning reversed this trend [14].

Spread of multidrug-resistant A. baumannii can occur on a national and even international scale. Examples include the spread of the Southeast clone and the Oxa-23 clones 1 and 2 through southeast England [15,16], the dissemination of a multidrug resistant A. baumannii clone throughout Portugal, the interhospital spread of a VEB-1 ESBL-producing A. baumannii clone amongst 55 different medical centers in northern and southeastern France, and the detection of 3 A. baumannii clones (I, II, III) in hospitals located throughout northern and southern Europe. In the United States, a clonal outbreak of resistant A. baumannii was first noted in the mid-1990s in New York City. It occurred on the heels of an outbreak of ESBL-producing Klebsiella pneumoniae and a rise in the use of imipenem [19]. Routine infection surveillance at this hospital detected an imipenem-resistant A. baumannii. Similar outbreaks have occurred across hospitals in New York and other regions of the United States. Lolans et al (2005) described a multicity carbapenem-resistant A. baumannii outbreak in the Midwest. This was the first time that presence of OXA-40 β-lactamase was described in the United States [20]. Since 2005, several institutions (including 5 hospitals and 3 long-term care facilities) have been affected by outbreaks with this clone, involving over 200 patients. Monoclonal multi-institutional outbreaks suggest that movement of personnel, patients, equipment or other shared products may be to blame, underscoring the importance of implementing rigorous infection control procedures.

Mechanisms of Resistance

Acinetobacter has only a limited number of virulence factors and is considered to be an opportunist. Although the cell wall contains lipopolysaccharide, the potential for endotoxigenic cell damage appears limited [21]. Thus, the concerning features of A. baumannii are its prodigious ability to develop resistance and avoid desiccation and the discovery of strains resistant to all current antibiotic classes. Although there are significant differences in the antimicrobial susceptibility profile of A. baumannii, the overall trend is one of increasing resistance since the 1970s [22].

There is no standard definition for multidrug-resistant A. baumannii infection, though it often implies resistance to at least 3 of the 5 classes of drugs that otherwise would be considered for therapy (ie, fluoroquinolones, amino-glycosides, cephalosporins, carbapenems, and ampicillin- sulbactam). Pan-resistance, despite its name, refers to resistance to all β-lactams, fluroquinolones, and aminoglycosides but excludes polymyxins and tigecycline. The major organizations that determine antibiotic susceptibility breakpoints (Clinical and Laboratory Standards Institute and the European Committee on Antimicrobial Susceptibility Testing) have different breakpoints for some of the key antibiotics used for A. baumannii infections. A breakpoint for tigecycline is not provided by either of these organizations.

There are several ways A. baumannii eludes antibiotics. These include efflux pumps, mutations in porins, mutations in antibiotic targets (eg, penicillin-binding proteins, topoisomerases and DNA gyrase), and antibiotic-altering enzymes (β-lactamases, carbapenemases, and aminoglycoside- modifying agents). A. baumannii contains an AmpC β-lactamase, which is a cephalosporinase. Unlike other gram-negative organisms, inducible AmpC expression does not occur. Rather the gene is overly expressed when a promoter insertion sequence (ISAba1) is inserted and triggers the production of the β-lactamase, resulting in treatment failure with most cephalosporins (although cefepime may retain activity) [23]. The upregulation of efflux pumps can move β-lactam antibiotics out of the periplasmic space, and the reduced expression of porins can preclude antibiotics from entering. Efflux pumps also expel quinolones, tetracyclines, tigecycline, trimethoprim, and some disinfectants [24].

Perhaps the most worrisome development is the spread of carbapenem-resistant A. baumannii. This resistance is most often due to the acquisition of mobile genetic elements encoding serine and metallo-β-lactamases. Recently, a transposon based armA (aminoglycoside resistance methylase) gene has been identified in A. baumannii. This gene encodes a 16S rRNA methylase which prevents an amingolycoside from binding to its target site, creating resistance to all aminoglycosides [25]. Fournier et al identified an 86-kb region called the AbaR1 resistance island from an multidrug- resistant isolate of A. baumannii in France. The island contained a cluster of 45 resistance genes, including those coding for VEB-1, AmpC, OXA-10 β-lactamases and various aminoglycoside-modifying enzymes and efflux pumps. The prevalence of this island in other multidrug-resistant A. baumannii isolates has yet to be determined [26].

In addition to resistance, A. baumannii can survive under dry conditions for prolonged periods of time. Jawad et al demonstrated that A. baumannii organisms could survive on glass coverslips at a relative humidity of 31% for an average of 20 days [27]. A follow-up study using the same methodology compared the desiccation properties of 22 outbreak related A. baumannii strains to 17 sporadic strains and found no statistical difference between the two (26.5 days and 27.2 days, respectively). However, outbreak strains were more drug-resistant [28].

Clinical Infections

Acinetobacter is typically associated with health care–associated infections, but community-acquired infections are also well described, most notably an aggressive and often fatal pneumonia. This community-acquired pneumonia has most frequently been reported in tropical Australia and Asia and typically affects adults with compromised immune function (ie, alcoholism, diabetes, renal failure) during the rainy season [29]. The disease is characterized by a fulminant clinical course, secondary bloodstream infections, and a mortality rate of 40% to 60% [30].

The vast majority of cases of Acinetobacter infections, however, occur in hospitalized patients. In the U.S.-wide Surveillance and Control of Pathogens of Epidemiological Importance [SCOPE] study conducted between 1995 and 2002, A. baumannii was the tenth most common organism isolated and was responsible for 1.3% of all monomicrobial nosocomial bloodstream infections (0.6 bloodstream infections per 10,000 admissions). A. baumannii bloodstream infections were more common in ICUs (1.6% of bloodstream infections) compared with non-ICUs (0.9% of bloodstream infections).

Risk factors for nosocomial infection include length of hospital stay, surgery, treatment with broad-spectrum antibiotics, indwelling central intravenous or urinary catheters, admission to a burn unit or ICU, mechanical ventilation, and breaches in infection control practices [31]. Interpreting the significance of A. baumannii isolates from skin, pharynx, GI tract, urethra, conjunctiva, and vagina must be performed carefully since these organisms can colonize both healthy and devitalized tissues in these areas. Most infections occur in tissues with a high fluid content, such as the respiratory tract, peritoneal fluid, and the urinary tract. Indwelling catheters also increase the chance of an isolate signifying an infection rather than colonization. The most common site for A. baumannii infection is the respiratory tract and the most frequent clinical manifestations of Acinetobacter infection are ventilator-associated pneumonia and bloodstream infections.

Nosocomial pneumonias tend to be multilobar and develop later in the hospital stay and can be complicated by effusions and bronchopleural fistulas [20]. Using data from the National Nosocomial Infections Surveillance System (1986–2003), Gaynes et al analyzed over 410,000 bacterial isolates to determine the epidemiology of gram-negative bacilli in ICUs. Although the percentage of pneumonia caused by gram-negative bacilli was constant during the study period, the proportion of ICU pneumonias attributable to Acinetobacter species increased from 4% in 1986 to 7% in 2003 (P < 0.001 for trend).

The crude mortality from A. baumannii bloodstream infection is 34.0% (43.4% in ICU patients versus 16.3% in non-ICU patients). The crude mortality of A. baumannii bloodstream infections in intensive care populations is only exceeded by Pseudomonas aeruginosa and Candida species bacteremias. The mean time from hospital admission to detection of a blood stream infection was the longest for A. baumannii, occurring a mean of 26 days from the time of hospital admission.

The impact of multidrug-resistant Acinetobacter infection on patient outcomes is difficult to assess. It is not certain if the high crude mortality rate associated with A. baumannii infection in the ICU is due to patients’ underlying critical illnesses or whether the organism has significant attributable mortality of its own. Acinetobacter does seem to be consistently associated with prolonged hospital stays. In a retrospective matched controlled study, Acinetobacter acquisition was associated with a 5-day longer stay in the ICU. A second study found that attributable length of stay was even longer in patients with invasive infections: Acinetobacter acquisition alone was associated with prolongation of length of stay by 13 days, whereas invasive infection was associated with prolongation of stay by 23 days [33,34].

In addition to pneumonia and bacteremia, intracranial infections with A. baumannii can occur. Meningitis with A. baumannii is generally described in patients following neurosurgical procedures and head trauma, although it can occur in the absence of these risk factors. In a review by Siegman-Igra et al (1993), 25 patients at a hospital in Tel Aviv were identified with A. baumannii nosocomial meningitis. Although a source was never identified, a reduction in the overall use of antibiotics in the neurosurgical department was associated with a decline in cases [35]. A. baumannii meningitis can present acutely or have a more indolent course. Given its gram-variable staining and morphological range from bacillus to coccobacillus, A. baumannii in the cerebrospinal fluid sometimes is mistaken for more common pathogens associated with meningitis, including Neisseria meningitidis [36].

Acinetobacter is a major pathogen in traumatic wounds and burns. It was first noted to be a significant pathogen in the Korean conflict. This was confirmed in the Vietnam War where it was the most common gram-negative bacillus isolated from traumatic lower extremity infections and the second most common organism isolated from the blood [37]. Returning soldiers from the Iraq and Afghanistan battlefields also have Acinetobacter infections, particularly hospital-acquired multidrug-resistant strains as discussed above. Genitourinary infections have also been reported, though these typically occur in patients with other risk factors for infection such as nephrolithiasis or indwelling catheters [20].

Treatment

Antibiotic-susceptible Acinetobacter isolates are best treated with β-lactams, usually third-generation cephalosporins, β-lactam-β-lactamase combinations, or carbapenems (imipenem and meropenem only; ertapenem has little activity and should not be used). Often these agents are used in combination with an aminoglycoside. Unfortunately, most nosocomial A. baumannii isolates are resistant to many of the aforementioned antibiotics. Carbapenems are often considered first-line agents in the treatment of resistant A. baumannii. Many multidrug-resistant isolates, however, do remain susceptible to sulbactam (available in the United States only as ampicillin-sulbactam). Sulbactam has intrinsic bactericidal activity against A. baumannii by binding to penicillin-binding protein 2 [38]. Sulbactam may retain activity against A. baumannii in the setting of carbapenem resistance and has been shown to be as efficacious as imipenem-cilastatin in treating Acinetobacter-associated ventilator-associated pneumonia [39]. Other options to treat pan-resistant strains include colistin and tigecycline. Intravenous colistin, a polymyxin that has fallen out of routine use due to its associated neuroand nephrotoxicities, has greater activity when combined with rifampin [40]. Inhaled colistin is occasionally employed for ventilator-associated pneumonia although treatment is sometimes limited by bronchospasm. The new glycycline antibiotic tigecycline has in vitro activity against some strains of multidrug-resistant A. baumannii; however, resistance in vivo has been reported to occur within a matter of weeks if not already present prior to initiation of therapy [41].

The antimicrobial availability task force of the IDSA has labeled A. baumannii a problem pathogen. The task force describes A. baumannii as a “prime example of a mismatch between unmet medical needs and the current antimicrobial research and development pipeline [42].” This dismal assessment of the antimicrobial options for MDR A. baumannii underscores the need for effective infection prevention and control.

Infection Prevention and Control

Infection prevention and control should focus on 3 major goals: preventing the development of endemic strains, early detection of multidrug-resistant A. baumannii infection and colonization, and preventing cross-transmission [24]. A rigorous antibiotic stewardship program can decrease the likelihood of developing drug-resistant epidemic strains in the first place [35]. Regular review of laboratory cultures and antibiotic susceptibility patterns can facilitate early detection of multidrug-resistant infections. During epidemic outbreaks, active surveillance to detect colonized patients before they manifest invasive infections can speed timely institution of contact precautions to limit transmission. Additional steps to prevent cross-transmission include fastidious hand hygiene, rigorous environmental cleaning, and alerting receiving facilities to patients’ colonization status prior to interfacility transfers. Significant effort should be devoted to source identification and reservoir control since most epidemics of multidrug-resistant A. baumannii strains have been attributed to environmental contamination, likely due to the length of time A. baumannii can withstand desiccation. Closure of units where outbreaks have occurred for environmental decontamination of potential reservoirs is more common for Acinetobacter infections (22.9%) than for other pathogens (11.7%) [43].