Antimicrobial resistance of Acinetobacter spp. in Europe

Authors

  • M. Van Looveren,

    1. Department of Medical Microbiology, University Hospital Antwerp, UA, Antwerp, Belgium
    Search for more papers by this author
  • H. Goossens,

    1. Department of Medical Microbiology, University Hospital Antwerp, UA, Antwerp, Belgium
    2. ESCMID Study Group for Antimicrobial Resistance Surveillance (ESGARS)
    3. Department of Medical Microbiology, Leiden University Medical Center, The Netherlands
    Search for more papers by this author
  • the ARPAC Steering Group

    Search for more papers by this author
    • *

      Members of the ARPAC Steering Group are: F. Baquero (Madrid, Spain), J. Bruce (Aberdeen, UK), B. Cookson (London, UK), G. Cornaglia (Verona, Italy), L. Dijkshoorn (Leiden, The Netherlands), H. Goossens (Antwerp, Belgium), I. Gould (Aberdeen, UK), G. Kahlmeter (Vaxjo, Sweden), V. Krcmery (Bratislava, Slovak Republic), D. Monnet (ad hoc) (Copenhagen, Denmark), F. MacKenzie (Aberdeen, UK), J. Mollison (Aberdeen, UK), M. Struelens (Brussels, Belgium), K. Towner (Nottingham, UK), P. J. van den Broek (Leiden, The Netherlands), J. van der Meer (Nijmegen, The Netherlands), M. Van Looveren (Antwerp, Belgium), J. Vila (Barcelona, Spain), A. Voss (ad hoc) (Nijmegen, The Netherlands) and D. Wagner (Brussels, Belgium).


Corresponding author and reprint requests: M. Van Looveren, Department of Medical Microbiology, Campus Drie Eiken, University of Antwerp, Universiteitsplein 1, B-2610 Wilrijk, Belgium
E-mail: marleen.vanlooveren@ua.ac.be

Abstract

Bacteria of the genus Acinetobacter are ubiquitous in nature. These organisms were invariably susceptible to many antibiotics in the 1970s. Since that time, acinetobacters have emerged as multiresistant opportunistic nosocomial pathogens. The taxonomy of the genus Acinetobacter underwent extensive revision in the mid-1980s, and at least 32 named and unnamed species have now been described. Of these, Acinetobacter baumannii and the closely related unnamed genomic species 3 and 13 sensu Tjernberg and Ursing (13TU) are the most relevant clinically. Multiresistant strains of these species causing bacteraemia, pneumonia, meningitis, urinary tract infections and surgical wound infections have been isolated from hospitalised patients worldwide. This review provides an overview of the antimicrobial susceptibilities of Acinetobacter spp. in Europe, as well as the main mechanisms of antimicrobial resistance, and summarises the remaining treatment options for multiresistant Acinetobacter infections.

Search strategy and selection criteria

Data for this review were obtained through searches of Medline and Pubmed, from references cited in relevant articles, through searches of abstracts and posters presented at different national and international meetings, from the worldwide web, and from surveillance studies, associated with the introduction of new agents, conducted by the pharmaceutical industry. Search terms used were ‘Acinetobacter’, ‘antimicrobial resistance’, ‘antibiotics’, ‘carbapenems’, ‘aminoglycosides’, ‘quinolones’ and ‘treatment’. Only studies published in English, French, German or Dutch were reviewed. The first phase of the literature search resulted in the identification of > 3500 references, but initial review of the abstracts revealed that many articles were duplicated across the different search strategies. After exclusion of duplicate references, 2158 studies published between 1963 and 2003 remained. Of these, 267 were downloaded for detailed review.

The acinetobacter genus and clinically important species

Acinetobacter spp. are glucose-non-fermentative, non-fastidious, strictly aerobic Gram-negative coccobacilli, usually occurring in diploid formation, or in chains of variable length. They are non-motile, catalase-positive and oxidase-negative [1]. Since 1986, the taxonomy of the genus Acinetobacter has undergone extensive revision. The original single species named Acinetobacter calcoaceticus has been abandoned, and at least 32 genomic species have now been proposed, of which 17 have been assigned species names. Correct identification of acinetobacters to the species level by the use of phenotypic methods is problematic [2,3]. In particular, discrimination between Acinetobacter baumannii and the genetically closely related unnamed genomic species 3 and 13 sensu Tjernberg and Ursing (13TU) is difficult. Genotypic methods, such as amplified ribosomal DNA restriction analysis [4] and analysis of whole genome fingerprints obtained by selective amplification of restriction fragments [5,6], have been useful for precise species identification when tested against libraries of well-validated strains, but these methods are not currently suitable for use in routine clinical diagnostic microbiology laboratories.

Studies using well-validated identification methods have shown clearly that most clinical isolates are strains of A. baumannii[7], and that this species, and to a lesser extent the closely related unnamed genomic species 3 and 13TU, are responsible for most infections and hospital outbreaks involving Acinetobacter spp. [3,4,6,8–11]. Other species, such as Acinetobacter junii, have been implicated only occasionally in outbreaks of nosocomial infection [3].

It is worth emphasising that the species names mentioned in the present review are those used by the authors of the original papers, and should be considered with some caution, given the problems of phenotypic identification in most of the studies reviewed. In articles published before the mid-1980s, the taxonomic status of the organisms is even more obscure, since, at that time, the genus Acinetobacter included only one all-encompassing species. Because of these considerations, acinetobacters are denoted as Acinetobacter spp. in the current review if their taxonomic status at the species level is uncertain.

Nosocomial infections caused by acinetobacter spp.

Acinetobacter spp. are recognised as important opportunistic pathogens mainly in immunocompromised patients [12]. Their contribution to nosocomial infection has increased over the past three decades, and many outbreaks of hospital infection involving acinetobacters have been reported worldwide. According to data from the National Nosocomial Infections Surveillance system, Acinetobacter spp. were isolated in 1% of all nosocomial infections from 1990 to 1992 [13]. However, the true frequency of nosocomial infection caused by Acinetobacter spp. is difficult to assess because isolation of Acinetobacter spp. in clinical specimens may reflect colonisation rather than infection [14].

Although prevalent in nature and regarded generally as commensals of human skin and the human respiratory tract, acinetobacters have also been implicated as the cause of serious infectious diseases such as pneumonia, urinary tract infection, endocarditis, wound infection, meningitis and septicaemia, involving mostly patients with impaired host defences [12]. Acinetobacter spp. have emerged as particularly important organisms in intensive care units (ICUs), and this is probably related, at least in part, to the increasingly invasive diagnostic and therapeutic procedures used in hospital ICUs in recent years [12]. Risk factors for acquisition of Acinetobacter spp. include hospitalisation, poor general medical status of patients, mechanical ventilation, cardiovascular or respiratory failure, previous infection or antimicrobial therapy, and the presence of central venous or urinary catheters [15].

Epidemiology of acinetobacter spp.

Acinetobacters are ubiquitous in nature; they can be recovered easily from soil or water, and have also been found frequently in animal and human hosts [16]. Several studies during the 1960s and 1970s reported isolation of these organisms from the skin of healthy individuals at rates of 0.8–20% for glucose-acidifying acinetobacters (Herellea vaginicola c.q. Acinetobacter anitratus), and 0–33.6% for glucose-non-acidifying acinetobacters (Mima polymorpha c.q. Acinetobacter lwoffii) [17–20]. Skin colonisation of patients plays an important role in the subsequent contamination of the hands of hospital staff during contacts, thereby contributing to the spread of the organisms [21]. High colonisation rates of the skin, throat, respiratory system or digestive tract, of various degrees of importance, have been documented in several outbreaks. However, general conclusions on the clinical significance of skin and mucosal Acinetobacter carriage are difficult to draw if the organisms are not identified correctly to the species level.

The epidemiology of Acinetobacter at the local institutional level can be investigated without reference to the taxonomy if the organisms are typed using appropriate methods. Cell-envelope protein profiling has shown that multiple body sites of patients can be colonised with clinically relevant strains for days to weeks, even if the organisms were not detected in clinical specimens of patients [22,23]. Two recent studies in Germany and the UK, using accurate identification methods, have reported carriage of acinetobacters on the body surface in patients and/or healthy individuals [24,25]. Overall colonisation rates for Acinetobacter spp. were > 40% for volunteers, and 75% for patients, when multiple body sites were sampled. Interestingly, A. baumannii and the unnamed genomic species 13TU, which predominate in hospital infections, were rarely isolated in the German study, while the unnamed genomic species 3 represented 11% of the total number of isolates, but was not detected in the UK study. In a similar study in Hong Kong [26], A. baumannii and spp. 3 and 13TU had a relatively high prevalence, both in patients and in healthy individuals, with a surprising level of > 35% for sp. 3 in community individuals and student nurses.

The reservoirs from which epidemic strains are imported into hospitals have not yet been elucidated. A recent study in New York, which compared A. baumannii strains from patients in two hospitals with isolates from the hands of community members, showed that strains in the community were distinct from those in the hospitals [27]. The authors concluded that the reservoir for epidemic strains was in the hospital itself.

Clonal spread of epidemic strains

A study comparing outbreak and non-outbreak A. baumannii isolates allowed delineation of two groups of genetically highly related, multiresistant strains among outbreak isolates from different northwestern European cities [4]. It was hypothesised that these groups represented two clonal lineages, the occurrence of which has since been reported also in the Czech Republic [10]. A third clone of widespread multiresistant strains from hospitals in other European countries was described by van Dessel et al.[28]. Most recently, the spread of another A. baumannii clone among 24 hospitals in the London (UK) region has been detected (J. Turton et al., unpublished results). The possible relatedness of the above clones to an amikacin-resistant strain that has spread among eight hospitals in Spain has not yet been investigated [29]. Overall, the findings from these studies indicate that several clones are responsible for many outbreaks in Europe.

The emergence of resistance

Acinetobacter spp. have become resistant to almost all antimicrobial agents that are currently available, including the aminoglycosides, quinolones and broad-spectrum β-lactams. Most strains are resistant to cephalosporins, while resistance to carbapenems is being reported increasingly [30,31]. Important differences in antimicrobial susceptibility exist between A. baumannii and other species in the genus, with A. baumannii being the most resistant species [32–34]. Difficulty in eradicating these bacteria has allowed them to colonise niches left vacant following the eradication of more susceptible microbes.

This review provides an overview of the evolution of the susceptibilities of Acinetobacter spp. to the different antimicrobial classes. However, it should be stressed that comparison of susceptibility rates between studies is rendered difficult, not only because of variations in local epidemiology, antibiotic selection pressure and patient populations, but also because of differences in study protocols. Thirty-one studies concerning the antimicrobial surveillance of Acinetobacter spp. in Europe were identified. An overview of these studies is given in Table 1. Duplicates were excluded and internal quality control through the use of standard control strains was performed in most of the studies. One study used the breakpoints recommended by the British Society for Antimicrobial Chemotherapy (BSAC), another did not specify the breakpoints used, and the remaining studies used the breakpoints recommended by the National Committee for Clinical Laboratory Standards. The identification methods used for the Acinetobacter strains were heterogeneous; however, as mentioned above, species identification using a correct method is an important factor to be taken into account when comparing different studies.

Table 1.  Summary of surveillance papers reviewed
AuthorsCountrySettingType of isolates or hospital departmentIsolation periodMethodBreakpointsExternal quality assuranceInternal quality controlExclusion of duplicatesStrain identificationNo. of isolates testedAntimicrobial agents tested (no.)
  1. NCCLS, National Committee for Clinical Laboratory Standards.

Aksaray et al.[85]Turkey8 hospitalsICU1997EtestNCCLSNot mentionedYesNoAcinetobacter spp.164Amikacin, amoxycillin–clavulanate, cefepime, cefodizime, cefotaxime, ceftazidime, ceftriaxone, cefuroxime, ciprofloxacin, gentamicin, imipenem, piperacillin–tazobactam (12)
Betriu et al.[80]Spain12 medical centresClinical isolates2001Agar dilutionNCCLSNot mentionedYesYesA. baumannii 64Ampicillin–sulbactam, cefepime, gentamicin, imipenem, levofloxacin, piperacillin–tazobactam, polymyxin B, sulbactam, tigecycline (9)
Buirma et al.[66]Netherlands8 hospitalsICU or medium-care surgical unit1990Broth microdilutionNCCLSNot mentionedYesNoAcinetobacter spp. 11Amikacin, ampicillin, amoxycillin–clavulanate, aztreonam, cefazolin, cefotaxime, ceftazidime, ceftriaxone, cefuroxime, ciprofloxacin, gentamicin, imipenem, mezlocillin, piperacillin, ticarcillin–clavulanate, tobramycin (16)
Fluit et al.[93] (SENTRY)14 European countries25 university hospitalsBlood isolates1997–1999Broth microdilutionNCCLSNot mentionedYesYesAcinetobacter spp.247Amikacin, amoxycillin–clavulanate, ampicillin, aztreonam, cefazolin, cefepime, cefoxitin, ceftazidime, ceftriaxone, cefuroxime, ciprofloxacin, gatifloxacin, gentamicin, imipenem, levofloxacin, meropenem, ofloxacin, piperacillin, piperacillin–tazobactam, sparfloxacin, tetracycline, ticarcillin, ticarcillin–clavulanate, tobramycin, trovafloxacin (25)
Garcia-Arata et al.[76]SpainTeaching hospitalClinical isolates1990–1994Agar dilutionNCCLSNot mentionedYesYesA. calcoaceticus–A. baumannii complex177Amikacin amoxycillin–clavulanate, ampicillin, ampicillin–sulbactam, ceftazidime, gentamicin, imipenem, meropenem, ofloxacin, piperacillin, piperacillin–tazobactam, sulbactam, ticarcillin, tobramycin (14)
Glupczynski et al.[67]Belgium18 hospitalsICU1994–1995EtestNCCLSNot mentionedYesYesAcinetobacter spp. 23Amikacin, amoxycillin–clavulanate, aztreonam, ceftazidime, ceftriaxone, cefuroxime, ciprofloxacin, gentamicin, imipenem, piperacillin, piperacillin–tazobactam, ticarcillin–clavulanate (12)
Glupczynski et al.[71]Belgium26 hospitalsICU1996–1997 and 1998–1999EtestNCCLSNot mentionedYesYesAcinetobacter spp.65 (1996–1997); 34 (1998–1999)Amoxycillin–clavulanate, aztreonam, ceftazidime, ceftriaxone, cefuroxime, ciprofloxacin, gentamicin, imipenem, piperacillin, piperacillin–tazobactam (10)
Günseren et al.[84]Turkey8 hospitalsICU1996EtestNCCLSNot mentionedYesNoAcinetobacter spp. 80Amikacin, amoxycillin–clavulanate, cefepime, cefodizime, cefotaxime, ceftazidime, ceftazidime–clavulanate, ceftriaxone, cefuroxime, ciprofloxacin, gentamicin, imipenem, piperacillin–tazobactam (13)
Hanberger et al.[68]Belgium, France, Portugal, Spain, Sweden118 hospitalsICU1994–1995EtestNCCLSNot mentionedYesYesAcinetobacter spp. 29Amikacin, ceftazidime, ceftriaxone, ciprofloxacin, gentamicin, imipenem, piperacillin, piperacillin–tazobactam (8)
Henwood et al.[72]UK54 diagnostic laboratoriesClinical isolates2000Agar dilution and Etest (confirmation)BSACNot mentionedNot mentionedYesA. baumannii complex and other genomic groups595Amikacin, cefotaxime, ceftazidime, ciprofloxacin, colistin, gentamicin, imipenem, meropenem, minocycline, piperacillin, piperacillin–tazobactam, rifampicin, sulbactam, tetracycline, tigecycline (15)
Hoban et al.[111]16 European countries61 laboratoriesClinical isolates1997–1999Broth microdilutionNCCLSNot mentionedYesYesA. baumannii368Ciprofloxacin, gemifloxacin, levofloxacin, ofloxacin (4)
Hostacka and Klokocnikova [89]SlovakiaHospitalClinical isolatesNot specifiedBroth microdilutionNot specifiedNot mentionedNot mentionedNot mentionedA. baumannii, A. lwoffii, A. calcoaceticus, A. haemolyticus 50Amikacin, ampicillin, ampicillin–sulbactam, cefotaxime, ceftazidime, cefuroxime, ciprofloxacin, co-trimoxazole, gentamicin, meropenem, netilmicin, piperacillin, piperacillin–tazobactam (13)
Jarlier et al.[75]France39 teaching hospitalsICU1991Broth microdilutionNCCLSNot mentionedYesYesA. baumannii268Aztreonam, cefotaxime, ceftazidime, ciprofloxacin, gentamicin, imipenem, piperacillin, tobramycin (8)
Jones et al. [94] (SENTRY)12 European countries20 university hospitalsSkin and soft tissue infections1997Broth microdilutionNCCLSNot mentionedYesYesAcinetobacter spp. 41Amikacin, amoxycillin–clavulanate, ampicillin-amoxycillin, aztreonam, cefazolin, cefepime, cefotaxime–ceftriaxone, cefoxitin, ceftazidime, cefuroxime, ciprofloxacin, gatifloxacin, gentamicin, imipenem, levofloxacin, meropenem, ofloxacin, piperacillin, piperacillin–tazobactam, sparfloxacin, tetracycline, ticarcillin, ticarcillin–clavulanate, tobramycin, trimethoprim– sulphamethoxazole, trovafloxacin (26)
Kocazeybek [87]Turkey4 hospitalsSurgical ICU1999Broth microdilution and EtestNCCLSNot mentionedYesNoA. baumannii 32Amikacin, amoxycillin–clavulanate, ampicillin, ampicillin–sulbactam, aztreonam, cefazolin, cefoperazone, cefotaxime, cefotetan, ceftazidime, ceftriaxone, cefuroxime, ciprofloxacin, gentamicin, imipenem, piperacillin, tetracycline, ticarcillin, ticarcillin-clavulanate, tobramycin, trimethoprim-sulphamethoxazole (21)
Maniatis et al. [83]Greece9 tertiary care hospitalsICU1998Broth microdilutionNCCLSNot mentionedYesYesA. baumannii121Amikacin, ampicillin–sulbactam, aztreonam, ceftazidime, ciprofloxacin, gentamicin, imipenem, netilmicin, piperacillin, ticarcillin–clavulanate, tobramycin (11)
Martín- Lozano et al. [79]SpainUniversity hospitalBlood isolates/ bacteraemia1997–1999Broth microdilution and Etest (colistin)NCCLSNot mentionedNot mentionedYesA. baumannii109Amikacin, ampicillin, ampicillin–sulbactam, aztreonam, cefotaxime, ceftazidime, ciprofloxacin, colistin, gentamicin, imipenem, tetracycline, trimethoprim– sulphamethoxazole (12)
Naaber et al. [91]EstoniaUniversity hospitalICU1995 and 1998Disk diffusionNCCLSNot mentionedNot mentionedYesAcinetobacter spp.Not specified, only %Amikacin, aztreonam, cefepime, ceftazidime, ciprofloxacin, gentamicin, imipenem (7)
Patzer et al. [90]PolandHospitalICU1997–2000Agar dilutionNCCLSNot mentionedNot mentionedYesAcinetobacter spp.  32Cefepime, cefotaxime, ceftazidime, ciprofloxacin, gentamicin, imipenem, meropenem, piperacillin–tazobactam, tobramycin (9)
Pfaller et al. [86]Turkey9 university hospitalsClinical isolates1997EtestNCCLSNot mentionedYesNot mentionedAcinetobacter spp.  80Aztreonam, cefepime, cefoperazone–sulbactam, cefotaxime, ceftazidime, imipenem, ticarcillin–clavulanate (7)
Ruiz et al. [77]SpainUniversity hospitalClinical isolates1991–1996Broth microdilutionNCCLSNot mentionedYesYesA. calcoaceticus–A. baumannii complex1532Amikacin, ampicillin, ampicillin–sulbactam, aztreonam, cefotaxime, ceftazidime, ciprofloxacin, gentamicin, imipenem, ofloxacin, piperacillin, tetracycline, ticarcillin, trimethoprim–sulphamethoxazole, tobramycin (15)
Schmitz et al. [110] (SENTRY)12 European countries20 university hospitalsBlood isolates, pneumonia, wound, and urinary tract infections1997–1998Broth microdilutionNCCLSNot mentionedYesYesAcinetobacter spp. 279Ciprofloxacin, gatifloxacin, levofloxacin, ofloxacin, sparfloxacin, trovafloxacin (6)
Schmitz et al. [101] (SENTRY)12 European countries20 university hospitalsBlood isolates, pneumonia, wound, and urinary tract infections1997–1998Broth microdilutionNCCLSNot mentionedYesYesAcinetobacter spp. 279Amikacin, gentamicin, tobramycin (3)
Seifert et al. [32]Germany11 hospitalsBlood cultures, central venous catheters, cerebrospinal fluidPeriod of 18 monthsBroth microdilutionNCCLSNot mentionedYesNot mentionedA. baumannii, A. haemolyticus, A. johnsonni, A. junii A. lwoffii and other Acinetobacter spp. 180Amikacin, amoxycillin-clavulanate, ampicillin, aztreonam, cefazolin, cefotaxime, cefoxitin, ceftazidime, ceftriaxone, cefuroxime, ciprofloxacin, gentamicin, imipenem, mezlocillin, piperacillin, tobramycin (16)
Shah et al. [65]Germany10 hospitalsICU1990Broth microdilutionNCCLSNot mentionedYesYesAcinetobacter spp.  23Amikacin, ampicillin, amoxycillin–clavulanate, aztreonam, cefazolin, cefotaxime, ceftazidime, ceftriaxone, cefuroxime, ciprofloxacin, gentamicin, imipenem, mezlocillin, piperacillin, ticarcillin–clavulanate, tobramycin (16)
Stratchounski et al. [92]Russia10 hospitalsICU1995–1996EtestNCCLSYesYesYesAcinetobacter spp. 77Amikacin, amoxycillin–clavulanate, cefotaxime, ceftazidime, ceftriaxone, cefuroxime, ciprofloxacin, co-trimoxazole, gentamicin, imipenem, piperacillin, piperacillin–tazobactam (12)
Tambic et al. [88]Croatia22 microbiology laboratoriesClinical isolates1999Disk diffusionNCCLSYesYesYesAcinetobacter spp.Not specified, only %Amikacin, ampicillin–sulbactam, imipenem, netilmicin (4 + several other antibiotics that are not specified)
Turner et al. [74] (MYSTIC)12 European countries37 hospital centresMajority ICU1997–2000Agar dilutionNCCLSNot mentionedYesNot mentionedA. baumannii, A. calcoaceticus var. lwoffii and other Acinetobacter spp. 635Ceftazidime, ciprofloxacin, imipenem, meropenem, piperacillin–tazobactam, tobramycin (6)
Verbist [69]Belgium16 hospitalsICU1990Broth microdilutionNCCLSNot mentionedYesNoAcinetobacter spp. 70Amikacin, ampicillin, amoxycillin–clavulanate, aztreonam, cefazolin, cefotaxime, ceftazidime, ceftriaxone, cefuroxime, ciprofloxacin, gentamicin, imipenem, mezlocillin, piperacillin, ticarcillin–clavulanate, tobramycin (16)
Verbist and Glupczynski [70]Belgium28 hospitalsClinical isolates1992EtestNCCLSNot mentionedYesYesAcinetobacter spp.111Aztreonam, cefepime, cefotaxime, ceftazidime, imipenem (5)
Vila et al. [34]Spain3 hospitalsClinical isolatesNot specifiedAgar dilutionNCCLSNot mentionedNot mentionedNot mentionedA. baumannii 54Amikacin, amoxycillin, amoxycillin–clavulanate, ampicillin, ampicillin–sulbactam, aztreonam, cefotaxime, ceftazidime, ceftizoxime, ceftriaxone, ciprofloxacin, chloramphenicol, doxycycline, enoxacin, gentamicin, imipenem, isepamicin, netilmicin, norfloxacin, ofloxacin, piperacillin, ticarcillin, ticarcillin–clavulanate, tobramycin, trimethoprim– sulphamethoxazole (25)

Β-Lactams

Acinetobacter is resistant to most β-lactam antibiotics, particularly penicillins and cephalosporins, especially in ICU patients [32–35]. Ceftazidime, piperacillin and carbapenems are among the β-lactam antibiotics most active against A. baumannii.

The main mechanism of resistance to β-lactam antibiotics in Acinetobacter spp. is the production of β-lactamases encoded either by the chromosome or by plasmids [36]. In addition, the low permeability of the outer-membrane of Acinetobacter, resulting from the small outer-membrane pore size and/or limited porin production [37], as well as alterations in the affinity of penicillin-binding proteins (PBPs), have been implicated in the resistance of Acinetobacter to these antibiotics. Danes et al.[38] described the distribution of β-lactamases in a collection of epidemiologically unrelated A. baumannii clinical isolates. The results suggested that over-expression of the chromosomal cephalosporinase AmpC could play an important role in resistance to β-lactam antibiotics.

Resistance to ampicillin, carboxypenicillin and ureidopenicillin has been attributed to the production of TEM-1 [34,39], TEM-2 [40], OXA-21 [38,41] or OXA-37 [42]β-lactamase. Most of the TEM and SHV types of extended-spectrum β-lactamases have not yet been detected in A. baumannii. The non-TEM, non-SHV extended-spectrum β-lactamases PER-1 and VEB-1 are the only extended-spectrum β-lactamases reported to date in A. baumannii. PER-1 has been detected in Turkish and French isolates [43–45]. An epidemiological survey performed in Turkey in 1996 identified the spread of PER-1-positive A. baumannii isolates [44], and infections with these isolates positive for PER-1 have been associated with a higher risk of mortality [45]. Isolates of A. baumannii producing VEB-1 were associated with a French outbreak in 2001 [46].

Carbapenems have become the preferred treatment for serious Acinetobacter infections in many centres, and have retained better activity than other antimicrobial agents; however, the number of reports of carbapenem resistance is growing steadily and is a cause for concern. In the last few years, carbapenem-resistant Acinetobacter isolates have been reported worldwide [47,48]. In northern Europe, carbapenem-resistant strains of A. baumannii have been mostly sporadic, but in some southern European countries, including parts of Spain, they are endemic [49,50]. In 2001, the International Network for the Study and Prevention of Emerging Antimicrobial Resistance (INSPEAR) defined the emergence of carbapenem resistance in Acinetobacter as a ‘global sentinel event’, warranting prompt epidemiological and microbiological interventions [51].

Several mechanisms of carbapenem resistance have been described in A. baumannii, including loss of outer-membrane proteins [52] and altered PBPs [53]. In addition, isolates of A. baumannii can acquire carbapenemases, including class B metallo-β-lactamases [54,55] and class A and D β-lactamases [56–59]. A combination of several mechanisms may be present in the same isolate, as has been described in other Gram-negative bacteria [49]. In general, the prevalence of carbapenemases is still relatively limited compared to the prevalence of other β-lactamases [60].

The first known A. baumannii isolate with a carbapenem-hydrolysing β-lactamase was collected in 1985 in Scotland, and was initially designated ARI-1 (subsequently renamed OXA-23) [30]. This enzyme hydrolyses imipenem and also confers resistance to penicillins, but not to second- and third-generation cephalosporins. A later study of this isolate identified a plasmid location for the OXA-23 gene [61]. Acinetobacter isolates producing carbapenemases have now been reported from at least 12 countries, including Belgium, France, Italy, Spain and the UK [56]. Some of these carbapenemases are IMP- or VIM-type metallo-β-lactamases belonging to class B [55,62,63], but most carbapenem-resistant acinetobacters produce zinc-independent β-lactamases of class D [56]. Sequenced carbapenemases from acinetobacters of this latter class include OXA-23 (ARI-1) [59], OXA-24 [57], OXA-25, OXA-26, OXA-27 [56] and OXA-40 [64]. The OXA-type β-lactamases all have relatively weak activity against carbapenems.

In a study from Germany in the early 1990s, imipenem was found to be the most active agent against A. baumannii. All 180 Acinetobacter spp. isolates tested were fully susceptible to imipenem. Amoxycillin–clavulanate showed moderate activity, whereas ampicillin, broad-spectrum penicillins and cephalosporins were less active. Similarly, in another report dating from 1991, 23 Acinetobacter spp. isolates were obtained from ICU patients in ten German hospitals. Ceftazidime and imipenem were the most active β-lactam antibiotics, with 96% of the isolates remaining susceptible. Susceptibilities to piperacillin and cefotaxime were 65% and 61%, respectively [65]. All 11 Acinetobacter spp. strains isolated in 1990 from patients in eight Dutch hospitals were susceptible to imipenem, and ten (91%) of the strains were susceptible to ceftazidime, ceftriaxone and amoxycillin–clavulanate [66].

Acinetobacter spp. collected between June 1994 and June 1995 from ICUs in five European countries showed susceptibilities to ceftazidime of 82% in Belgium, 30% in France, 19% in Portugal, 24% in Spain and 100% in Sweden. Susceptibilities to imipenem were 88% in Belgium, 91% in France, 95% in Portugal, 84% in Spain and 81% in Sweden [67,68]. In 1990, 70 Acinetobacter spp. from ICU patients in 16 Belgian hospitals showed susceptibilities to imipenem, ceftazidime and ceftriaxone of 93%, 86% and 74%, respectively [69].

Between September and October 1992, 111 Acinetobacter spp. were collected in 28 Belgian hospitals. Ceftazidime was the most active agent (78% susceptible), followed by cefepime (74%), cefotaxime (66%), piperacillin (56%) and aztreonam (47%) [70]. During 1996–1997 and 1998–1999, 41 and 11 Acinetobacter spp., respectively, were collected in the ICUs of Belgian hospitals. Imipenem (90% and 89%) and ceftazidime (80% and 100%) were the most active agents. Some important changes in resistance rates took place between both collections. However, the small number of isolates, as well as the lack of complete identification to the species level for most isolates, precluded any comparison between both collection periods [71].

In the UK, 13 (2.2%) of 595 Acinetobacter spp. isolated during 2000 from routine clinical specimens at 54 sentinel laboratories were carbapenem-resistant (BSAC breakpoint, MIC ≥ 8 mg/L) [72]. An allele of blaIMP was detected in one of these isolates, but the other 12 isolates either had carbapenemase-independent resistance, or undetectable carbapenemase activity combined with other resistance mechanisms. Routine surveillance data obtained in 2001 from England and Wales similarly showed only 1% resistance to imipenem in bacteraemia isolates of Acinetobacter spp. [73]. These results are in sharp contrast to those of the MYSTIC (Meropenem Yearly Susceptibility Test Information Collection) study reported below [74]; however, only isolates of A. baumannii were investigated in the MYSTIC study.

Of 268 A. baumannii isolates from French teaching hospitals in 1991, 21% were susceptible to piperacillin and cefotaxime, 20% to aztreonam, 71% to ceftazidime, and 100% to imipenem [75]. Similar results were reported with 177 A. calcoaceticus–A. baumannii complex strains isolated in 1990–1994 from patients admitted to a Spanish teaching hospital. Imipenem and meropenem (99% susceptibility) were the most active agents tested, with 97% of isolates being susceptible to ampicillin–sulbactam. Only 25% of the isolates were susceptible to ceftazidime [76].

In a Spanish study published in 1993, almost all 54 A. baumannii isolates tested were resistant to both ampicillin and amoxycillin. Addition of the β-lactamase inhibitor sulbactam increased the percentage of strains susceptible to ampicillin to 52%, while the addition of clavulanic acid to amoxycillin or ticarcillin did not significantly change the percentage of susceptible strains. However, the enhanced activity in the presence of sulbactam may be attributable to the activity of sulbactam alone. More than 50% of the isolates were resistant to piperacillin, cefotaxime, ticarcillin and ceftazidime [34].

Other Spanish studies have also documented significant levels of resistance. In one study, ceftazidime resistance increased from 57.4% in 1991 to 86.8% in 1996, while imipenem resistance increased from 1.3% to 80.0%[77]. In another study, seven (21%) of 34 multiresistant clinical isolates recovered during 1990–1995 were resistant to imipenem [78]. In a third study during the period 1997–1999, isolates from patients with nosocomial A. baumannii bacteraemia were studied. Of 109 isolates, 71% were resistant to cefotaxime, 66% to ceftazidime, and 34% to imipenem [79]. In a fourth study, of 64 A. baumannii isolates obtained from 12 Spanish medical centres in 2001, 37.5% were resistant to cefepime, and 28.1% to imipenem [80]. Finally, in a hospital in Seville, all A. baumannii isolates in blood were susceptible to imipenem in 1991, whereas 50% were resistant to imipenem in 2000 [81].

In Greece, all A. baumannii isolates in the Greek WHONET study in 1996 were fully susceptible to imipenem (http://mednet.gr/whonet) [82]. During a 4-month period from January to April 1998, 121 clinical isolates of A. baumannii were collected from patients hospitalised in the ICUs of nine Greek tertiary care hospitals. High rates of resistance to β-lactam antibiotics, such as aztreonam (93.4%) and ceftazidime (95.9%), were detected, but imipenem-resistant acinetobacters were not isolated. However, after the study period, a few imipenem-resistant acinetobacters emerged in a large Greek hospital that had participated in the study [83].

In Turkey, a surveillance study in 1996 of Gram-negative bacteria isolated from ICUs in eight hospitals found that 5% of 80 Acinetobacter spp. isolates were susceptible to cefotaxime and ceftriaxone, 7.5% to ceftazidime, 11.2% to cefepime, and 71.2% to imipenem [84]. When the study was repeated in 1997, no antibacterial agent other than imipenem was effective against Acinetobacter spp., and of the 164 isolates investigated, only 49.3% were susceptible to imipenem [85]. Also in Turkey, a multicentre evaluation in 1997 of seven broad-spectrum β-lactams found that the 80 Acinetobacter spp. isolates investigated were generally not susceptible to ceftazidime, cefotaxime, aztreonam or ticarcillin–clavulanate (range 17.2–29.3% susceptible), but were more susceptible to both imipenem (85.0%) and cefoperazone–sulbactam (73.8%) [86], although the latter finding could be caused by the intrinsic activity of sulbactam against Acinetobacter spp.

A third Turkish study in 1999 of 32 A. baumannii isolates from the ICUs of four different hospitals found that 5.6% were susceptible to cefotaxime, 3.1% to ceftriaxone, 20.6% to ceftazidime, and 100% to imipenem [87].

In Croatia, imipenem resistance rates of Acinetobacter spp. in 1999 varied between 0% and 8% in 22 Croatian microbiology laboratories representing the major geographical regions of the country, with 18% of isolates resistant to ampicillin–sulbactam [88]. In a Slovakian study published in 2002, 46 (92%) of 50 clinical Acinetobacter spp. isolates (A. baumannii, A. lwoffii, A. calcoaceticus, A. haemolyticus) were resistant to ampicillin, 90% to cefuroxime, 58% to piperacillin, 50% to cefotaxime, 42% to ceftazidime, 38% to piperacillin–tazobactam, and 16% to ampicillin–sulbactam. None of the isolates was resistant to meropenem [89].

In Poland, all 32 isolates investigated during 1997–2000 remained susceptible to imipenem and meropenem (MIC90 0.5-2 mg/L) [90], while a 1995 study in the ICUs of an Estonian university hospital found that 56% of Acinetobacter spp. isolates were susceptible to ceftazidime, decreasing to 43% in 1998. In 1998, 47% of isolates remained susceptible to cefepime, with a small increase in susceptibility to imipenem from 93% in 1995 to 99% in 1998 [91]. In ten Russian hospitals between September 1995 and May 1996, 77 Acinetobacter spp. strains isolated from ICU patients showed high resistance rates to β-lactam antibiotics, ranging between 73% and 96%. Imipenem was the most active agent, with all isolates remaining susceptible [92].

In the 1997–1998 European arm of the SENTRY antimicrobial surveillance programme, imipenem and meropenem were the most active drugs against Acinetobacter spp. isolated from blood, with 80.2% and 78.1%, respectively, of the 247 Acinetobacter spp. isolates remaining susceptible. Of the cephalosporins tested, ceftazidime and cefepime were the most active, although only 51.4% and 62.8%, respectively, of the isolates remained susceptible to these agents [93]. In 1997, the antimicrobial susceptibilities of 41 Acinetobacter spp. isolates associated with skin and soft tissue infections were determined in the SENTRY study. Imipenem was the most active β-lactam compound tested, with 90.2% of the isolates being susceptible. Susceptibilities to meropenem, ceftazidime and cefepime were 85.4%, 41.5% and 48.8%, respectively [94].

Among the 490 A. baumannii isolates from patients with serious infections in 37 European hospitals participating in the 1997–2000 MYSTIC programme, the two carbapenems showed the greatest clinically useful activity. Susceptibilities of A. baumannii to meropenem were very high (97–100%) in all countries except Italy (70%), Turkey (66%) and the UK (77%). A similar pattern was seen for imipenem (93–100%), except for Italy (78%), Turkey (62%) and the UK (78%). The 51 isolates of A. calcoaceticus var. lwoffii were more susceptible than A. baumannii. Susceptibilities of A. calcoaceticus var. lwoffii to meropenem and imipenem were 100% in all countries except Turkey, where there was 87% susceptibility to meropenem, and 73% susceptibility to imipenem. For the 94 isolates of other Acinetobacter spp., high (91–100%) susceptibilities to meropenem and imipenem were seen in all countries that obtained isolates from this group, except Italy (88%) and Turkey (27%) [74].

Data from the 1999–2000 ESAR (European Surveillance of Antibiotic Resistance) study showed no resistance of Acinetobacter to carbapenems in Scotland (0/1316 isolates), 1.51% resistance in Germany (6/398 isolates), 0.4% resistance in Slovakia (10/250 isolates), and 0.56% resistance in Poland (3/535 isolates) (http://www.esbic.de/esbic/esar/results2001/esar-body116htm).

Aminoglycosides

Aminoglycosides are used widely for the treatment of Acinetobacter infections, but increasing numbers of highly resistant strains have been reported since the late 1970s.

The most frequent cause of resistance to aminoglycosides in Acinetobacter spp. is the modification of hydroxyl or amino groups of the antibiotic by aminoglycoside-modifying enzymes. All three types of aminoglycoside-modifying enzymes (acetylases, adenylases and phosphotransferases) have been detected in clinical isolates of Acinetobacter spp. [40,95]. However, geographical variations have been observed. For example, the gene for AAC(3)-Ia was found frequently in isolates of Acinetobacter spp. from Belgium (36 of 45 strains) [96], but in only two of 54 strains from Spain [34]. Various adenylating enzymes have been described in Acinetobacter spp. In Spain, 15% of the 54 clinical A. baumannii isolates studied contained ANT(3′′)9, which modifies streptomycin and spectinomycin [34]. The phosphotransferase found most frequently in Acinetobacter spp. is APH(3′)VI. In 1990, Lambert et al.[97] described the dissemination of the aph6 gene in France, and APH(3′)VI was detected in 15 of the 54 clinical A. baumannii isolates studied in Spain [34]. In addition, it was demonstrated that the spread of amikacin resistance in A. baumannii isolated in Spain was associated with an epidemic strain carrying the aph(3′)-VIa gene [29]. A. haemolyticus and related species are intrinsically resistant to aminoglycosides through synthesis of the chromosomally encoded specific N-acetyltransferase AAC(6′) [98, 99]. Other mechanisms of resistance to aminoglycosides in Acinetobacter spp. include alterations of the target ribosomal protein, and ineffective transportation of the antibiotic to the interior of bacteria [34]. In 2001, Magnet et al.[100] described a resistance-nodulation-cell division-type efflux pump involved in aminoglycoside resistance in a multiresistant A. baumannii isolate from a patient with urinary tract infection. This efflux pump was also shown to affect the level of susceptibility to other drugs, including fluoroquinolones, tetracyclines, chloramphenicol and trimethoprim.

Resistance to aminoglycosides is relatively common in clinical isolates of Acinetobacter spp. In a study performed in Germany in 1990, only 57% of 23 isolates from ICU patients were susceptible to gentamicin, and 78% were susceptible to tobramycin [65]. In The Netherlands, 11 Acinetobacter spp. isolated in 1990 from patients of eight hospitals were all susceptible to amikacin, with 72% susceptible to tobramycin, and 45% to gentamicin [66]. In 1994–1995, the percentages of gentamicin-susceptible Acinetobacter spp. isolates from ICU patients were 82% in Belgium, 34% in France, 36% in Portugal, 19% in Spain and 100% in Sweden. For amikacin, the figures were 85% in Belgium, 64% in France, 90% in Portugal and 49% in Spain [67,68].

Of 70 Acinetobacter spp. isolated in 1990 from Belgian ICU patients, 57% were susceptible to amikacin, 49% to tobramycin and 43% to gentamicin [69]. Between 1996 and 1999, 83% of 41 Acinetobacter spp. isolates from ICUs of Belgian hospitals in 1996–1997 were susceptible to gentamicin, compared with 56% of 11 isolates in 1998–1999 [71]. A French study in 1991 found that 19% and 28%, respectively, of 268 A. baumannii isolates from ICU patients were susceptible to gentamicin and tobramycin [75].

In a Spanish study during the early 1990s, 50% of 54 A. baumannii isolates tested were susceptible to tobramycin, 33% to gentamicin, 66% to netilmicin, and 72% to amikacin and isepamicin [34]. Of 177 A. calcoaceticus–A. baumanni complex isolates from patients admitted to a Spanish teaching hospital between 1990 and 1994, 94% were susceptible to amikacin [76]. Between 1991 and 1996, an increase in aminoglycoside resistance among clinical isolates of Acinetobacter spp. was noticed in Spain, rising from 33.0% to 71.8% for tobramycin, and from 21.0% to 83.7% for amikacin [77].

The Greek System for Surveillance of Antimicrobial Resistance reported in 1996 that 41.5% and 75.6%, respectively, of A. baumannii ICU isolates were resistant to netilmicin and gentamicin, while resistance rates of 51.6% and 58.4%, respectively, were reported in hospital wards (http://mednet.gr/whonet) [82]. In 1998, 121 A. baumannii isolates from ICU patients of nine hospitals in Greece showed resistance levels of 87.6% and 56.2%, respectively, to gentamicin and netilmicin, while amikacin retained activity against 70.2% of the isolates [83].

In Turkey, only 8.7% of 80 isolates from ICUs in 1996 were susceptible to gentamicin and only 29.1% to amikacin [84]. In 1997, of 164 isolates of Acinetobacter spp., 17.1% were susceptible to gentamicin and 34.8% to amikacin [85]. Of 32 A. baumannii isolates from Turkish ICUs in 1999, 62.5% were susceptible to amikacin and 15.6% to gentamicin [87]. In Croatia in 1999, 25% of isolates were resistant to amikacin and 26% to netilmicin [88], while in a Slovakian study of 2002, 58% of 50 Acinetobacter spp. isolates (A. baumannii, A. lwoffii, A. calcoaceticus, A. haemolyticus) were resistant to gentamicin, 44% to amikacin and 24% to netilmicin [89]. In Estonia, 27% of isolates of Acinetobacter spp. from ICUs were susceptible to gentamicin in 1995, decreasing to 19% in 1998, while susceptibility to amikacin decreased from 95% in 1995 to 60% in 1998 [91]. In Russia, 91% of 77 Acinetobacter spp. isolates in 1995–1996 were resistant to gentamicin, while only 7% were resistant to amikacin [92].

In the 1997–1998 European SENTRY study, tobramycin showed the greatest activity of the tested aminoglycosides against Acinetobacter spp., with 60.2% of 279 strains susceptible. Susceptibilities to amikacin and gentamicin were 58.1% and 43.4%, respectively [101]. Of 247 Acinetobacter spp. isolates from blood cultures, 62.4% were susceptible to tobramycin, 59.1% to amikacin and 48.6% to gentamicin [93]. During a 3-month period in 1997, 41 Acinetobacter spp. isolates associated with skin and soft tissue infections were isolated in 20 European hospitals. Susceptibilities to amikacin, gentamicin and tobramycin were 46.3%, 34.2% and 43.9%, respectively [94].

Quinolones

Until 1988, quinolones had good activity against Acinetobacter strains [102] compared to expanded-spectrum cephalosporins and aminoglycosides. However, resistance to these antibiotics has emerged rapidly in clinical isolates [32–34]. Resistance of A. baumannii to the fluoroquinolones has been attributed to changes in the structure of DNA gyrase or topoisomerase IV, caused by mutations in the gyrA or parC genes, respectively, which lower the affinity of the drug for the enzyme–DNA complex [103–106]. A second mechanism of resistance involves mutations of chromosomally-encoded drug-influx and -efflux systems that determine intracellular drug accumulation [103,105,107]. These mutations result either in reduced production of specific outer-membrane proteins which mediate quinolone influx, or over-expression of some efflux system(s), leading to active drug expulsion. Two studies suggesting the involvement of efflux pumps in the acquisition of resistance to quinolones in A. baumannii have been published [108,109].

In Germany, 96% of Acinetobacter spp. isolates from ICU patients were susceptible to ciprofloxacin [65], and all 11 Acinetobacter spp. isolated in 1990 from patients of eight Dutch hospitals were susceptible to ciprofloxacin [66]. In 1994–1995, susceptibilities to ciprofloxacin in isolates from ICU patients were 82% in Belgium, 22% in France, 25% in Portugal, 19% in Spain and 81% in Sweden [67, 68]. In Belgium, 51% of the 70 Acinetobacter spp. isolates from ICU patients in 1990 were susceptible to ciprofloxacin [69], while 76% of the 41 Acinetobacter spp. isolated in 1997 from Belgian ICUs were susceptible to ciprofloxacin, compared with 56% of the 11 isolates in 1999 [71]. Of the 268 A. baumannii isolated from the ICUs of 39 French teaching hospitals in 1991, 18% were susceptible to ciprofloxacin [75].

In Spain, Vila et al.[34] found ciprofloxacin (70%) and ofloxacin (72%) to be more active against clinical isolates of A. baumannii than norfloxacin (18%), but in a separate study, ciprofloxacin resistance in clinical isolates of Acinetobacter increased in Spain from 54.4% in 1991 to 90.4% in 1996 [77].

In 1996 in Greece, 76.6% of the A. baumannii isolates from wards, and 92.4% of the ICU isolates, were resistant to ciprofloxacin (http://mednet.gr/whonet) [82]. Of 121 A. baumannii isolates collected in 1998 from Greek ICUs, 92.6% were resistant to ciprofloxacin [83]. In 1996 in Turkey, 26.4% of Acinetobacter spp. isolates from ICUs were susceptible to ciprofloxacin [84] compared with 32.9% in 1997 [85], while a separate study in 1999 found that 31.3% of 32 A. baumannii isolates from ICUs were susceptible to ciprofloxacin [87].

In a Slovakian study published in 2002, 68% of the 50 tested Acinetobacter spp. isolates (A. baumannii, A. lwoffii, A. calcoaceticus, A. haemolyticus) were resistant to ciprofloxacin [89], while 81.3% of 32 Acinetobacter spp. isolates collected from children in a Polish ICU between 1997 and 2000 were susceptible to ciprofloxacin [90]. In 1995, 67% of Acinetobacter spp. isolates from Estonian ICUs were susceptible to ciprofloxacin, decreasing to 33% in 1998 [91]. Of 77 Acinetobacter spp. strains isolated in Russia from patients with ICU-acquired infections, 53% were resistant to ciprofloxacin [92].

Of the 279 clinical Acinetobacter spp. isolates from 20 European university hospitals participating in the 1997–1998 SENTRY study, 45.2%, 46.6% and 47.3% were susceptible to ciprofloxacin, ofloxacin and levofloxacin, respectively. Gatifloxacin and trovafloxacin showed the best in-vitro activities against Acinetobacter spp. [110]. Quinolones showed poor activity against Acinetobacter spp. from blood cultures. Only 50.6%, 52.6% and 54.7% of the 247 isolates showed in-vitro susceptibility to ciprofloxacin, ofloxacin and levofloxacin, respectively. Similar resistance rates were seen throughout the different European centres [93]. Of the 41 Acinetobacter spp. isolates associated with skin and soft tissue infections, 41.5%, 46.3% and 48.8% were susceptible to ciprofloxacin, ofloxacin and levofloxacin, respectively [94].

Between 1997 and 1999, 368 A. baumannii isolates were collected from 16 European countries. Susceptibilities to gemifloxacin, ciprofloxacin, levofloxacin and ofloxacin were 53.8%, 49.7%, 61.7% and 51.4%, respectively. The isolates of A. anitratus, A. calcoaceticus, A. haemolyticus and A. lwoffii investigated were less resistant than those of A. baumannii[111].

Other antibiotics

A. baumannii has a high degree of resistance to both chloramphenicol and trimethoprim–sulphamethoxazole, but little is known about the genetic basis of resistance to these compounds in these bacteria. Devaud et al.[40] found that chloramphenicol resistance involves the synthesis of chloramphenicol acetyltransferase I (CAT1). The CAT1 gene has been associated with both chromosomal and plasmid DNA in a clinical Acinetobacter isolate, suggesting that the CAT1 gene might be transposon-encoded, thereby improving its survival potential by being located in both replicons [112]. However, in another study, CAT1 activity was not detected, suggesting that resistance could result from a change in permeability to the antibiotic or a mutation in the target protein [34]. Similarly, Goldstein et al.[39] studied a multiresistant strain of A. calcoaceticus var. anitratus and found that resistance to chloramphenicol was not associated with CAT production. In a Spanish study published in 1993, all 54 A. baumanni isolates tested were resistant to chloramphenicol [34]. In other bacteria, resistance to sulphonamides is normally caused by the acquisition of plasmids encoding resistant versions of the target protein, dihydropteroate synthase. Similarly, high-level trimethoprim resistance is generally caused by the acquisition of plasmid DNA carrying a dhfr gene encoding a dihydrofolate reductase with low affinity for trimethoprim [113]. No specific studies on mechanisms of resistance to these antibiotics in Acinetobacter spp. have been published.

A Spanish study on the evolution of resistance among clinical isolates of Acinetobacter found that 41.1% and 88.9% of isolates were resistant to trimethoprim-sulphamethoxazole in 1991 and 1996, respectively [77]. Of 109 isolates in Spain between 1997 and 1999 from patients with nosocomial A. baumannii bacteraemia, 85% were resistant to trimethoprim-sulphamethoxazole [79]. In a Spanish study published in 1993, 63% of 54 A. baumannii isolates tested were susceptible to trimethoprim–sulphamethoxazole [34]. In 1999, 43.8% of 32 A. baumannii isolates from the ICUs of four different hospitals in Turkey were susceptible to trimethoprim–sulphamethoxazole [87], while in a Slovakian study published in 2002, 58% of 50 Acinetobacter spp. isolates (A. baumannii, A. lwoffii, A. calcoaceticus, A. haemolyticus) were resistant to trimethoprim–sulphamethoxazole [89]. Of the 77 Acinetobacter spp. isolates from patients with ICU-acquired infections in ten Russian hospitals, 88% were resistant to trimethoprim–sulphamethoxazole [92].

Tetracycline acts by binding to the 30S ribosomal subunit, resulting in the inhibition of protein synthesis [114]. Tetracycline-resistant bacteria generally express one of two different resistance mechanisms: an efflux pump or a ribosomal protection system. Different tetracycline resistance determinant classes have been recognised and classified, with classes A–E being detected most frequently among Gram-negative bacteria. As most tetracycline resistance genes have been found on plasmids or transposons, acquisition of resistance is generally assumed to be mediated mainly by gene transfer [115]. Guardabassi et al.[116] found the TetA and TetB determinants in clinical and aquatic strains of A. baumannii. A transposon containing the tetA determinant was characterised partially by Ribera et al.[117], who also detected the presence of the TetM determinant in a clinical isolate of A. baumannii[118].

Different publications have reported excellent activity of doxycycline or minocycline, but not tetracycline, against Acinetobacter spp. [34,119]. This may result from the fact that TetA, the major tetracycline resistance determinant, confers resistance to tetracycline, but not to minocycline. In a Spanish study of the early 1990s, 98% of 54 A. baumannii isolates tested were susceptible to doxycycline [34]. Of 109 A. baumannii isolates tested in Spain between 1997 and 1999, 85% were resistant to tetracycline [79]. Tigecycline, a new glycylcycline, was evaluated in the UK with 595 Acinetobacter spp. isolated during 2000 from routine clinical specimens at 54 sentinel laboratories. Tigecycline was found to be less active than minocycline, but both agents overcame most tetracycline resistance [72]. In the 1997–1998 SENTRY study, 247 Acinetobacter spp. were isolated from blood cultures, of which 51% were susceptible to tetracycline [93]. Of 41 Acinetobacter spp. isolates in 1997 that were associated with skin and soft tissue infections, 43.9% were susceptible to tetracycline [94].

Genetics of resistance

Acinetobacter is a genus that appears to have a propensity to develop antibiotic resistance extremely rapidly, perhaps as a consequence of its long-term evolutionary exposure to antibiotic-producing organisms in soil [12]. Furthermore, Acinetobacter spp., and A. baumannii in particular, are intrinsically resistant to many of the antimicrobial agents used most commonly [120].

A major contributing factor in the emergence of resistant Acinetobacter spp. is the acquisition and transfer of antibiotic resistance on plasmids, transposons and integrons. Several studies have reported that > 80% of Acinetobacter spp. isolates carry multiple indigenous plasmids of various molecular sizes [121,122], although another report found plasmids in only 28% of the clinical isolates of A. baumannii analysed [123]. In a recent Dutch study, plasmids were detected in 42% (20/48) of Acinetobacter isolates [124].

Transposons probably play an important role, in conjunction with integrons [125], in ensuring that particular novel genes can become established in a new gene pool. Different studies have reported chromosomally-located transposons carrying multiple antibiotic resistance genes in clinical isolates of Acinetobacter spp. [126].

The presence of integrons in Acinetobacter spp. has been well-described, as has their relatively high frequency of carriage in epidemic strains [124,127]. Three main classes of integrons have been described. Class 1 integrons (mostly associated with the sul1 gene) include the gene encoding the Int1 integrase (intI1). Class 2 integrons (related to transposon Tn7 and its close relatives) have a defective intI gene (intI2*) with partial homology to intI1. Class 3 integrons encode the IntI3 integrase, showing 60.9% homology with the amino-acid sequence of the IntI1 integrase [128]. Most investigators have found predominantly class 1 integrons in Acinetobacter spp. [124,129,130]. However, Gonzalez et al.[131] found predominantly class 2 integrons in A. baumannii isolates from Chilean hospitals. Possibly, the isolates from the latter study were more genetically related. Among A. baumannii integrons, a high prevalence of genes encoding aminoglycoside-modifying enzymes and β-lactamases have been found [41,129,130,132]. Seward and Towner [129] found similar integrons in genotypically distinct Acinetobacter spp. isolates from different locations worldwide. This is in agreement with the findings of Gombac et al.[127], who found that integron structures with the same variable region can be retrieved from genotypically distinguishable strains, and with the findings of Ribera et al.[133], who demonstrated that non-related A. baumannii isolates from different geographical areas are able to acquire common integrons.

Characterisation of integron structure in epidemiologically unrelated strains of A. baumannii suggests, based on integron structure similarity, that inter-species transfer may have occurred from Enterobacteriaceae [127]. Similarly, in a Spanish study, the integrons carried genes that were identical or closely related to genes found previously in integrons from organisms such as Pseudomonas aeruginosa, suggesting the potential transfer of genetic material between A. baumannii and P. aeruginosa[134]. On the other hand, related strains possessing unrelated integrons have also been found [134].

Treatment options for carbapenem-resistant acinetobacter spp.

Very few of the major antibiotics are now reliably effective for the treatment of severe nosocomial Acinetobacter infections, particularly in patients confined to ICUs. Until recently, the recommended drugs for therapy were extended-spectrum penicillins, broad-spectrum cephalosporins or carbapenems, combined with an aminoglycoside [12]. However, increasing resistance to these antimicrobial agents necessitates a critical appraisal of the remaining antibiotic treatment options.

Sulbactam is a synthetic β-lactam molecule, with structural, chemical and pharmacokinetic properties similar to those of the aminopenicillins. A feature that distinguishes sulbactam from other available β-lactamase inhibitors is its direct antimicrobial activity against Bacteroides fragilis and Acinetobacter spp., organisms against which most cephalosporins display little or no activity [135]. Binding of sulbactam to PBP2 of these organisms results in intrinsic antibacterial activity [136].

Most studies have investigated only the ampicillin–sulbactam combination, since sulbactam alone is not available commercially in many countries. In 1993, ampicillin–sulbactam was used for the treatment of ten patients with infections caused by imipenem-resistant A. calcoaceticus, nine of whom improved clinically [137]. In 1996, a prospective observational study followed 79 patients with A. baumannii bacteraemia. Ampicillin–sulbactam was used in eight patients, with a cure rate of 88%[138]. In 1997, a series of patients with multiresistant A. baumannii meningitis who were treated with ampicillin–sulbactam was reported. Eight cases of nosocomial meningitis were treated with ampicillin–sulbactam 2 g + 1 g every 6 h (seven patients) or 2 g + 1 g every 8 h (one patient). All isolates were resistant to gentamicin, ceftazidime and ciprofloxacin, while seven of eight were imipenem-resistant. Six patients were cured, while two died of meningitis [139]. Corbella et al.[140] treated 42 patients with non-life-threatening multiresistant A. baumannii infections, including seven bacteraemias, with sulbactam alone and in combination with ampicillin (1 g every 8 h); 39 improved or were cured with no major adverse effects. In this study, killing curves showed that sulbactam was bacteriostatic, and no synergy was observed between ampicillin and sulbactam. The authors suggested a role for sulbactam in non-life-threatening infections caused by A. baumannii. A retrospective analysis (1987–1999) compared treatment outcomes of 48 patients with A. baumannii bacteraemia treated with imipenem–cilastatin or ampicillin–sulbactam for ≥ 72 h. Ampicillin–sulbactam was at least as effective as imipenem–cilastatin and was a cost-effective alternative for treatment [141].

Unfortunately, emergence of resistance to sulbactam has been noted in imipenem-resistant strains of A. baumannii, leaving the polymyxins (colistimethate and polymyxin B) as the only treatment alternative [142]. Colistin is rapidly bactericidal and exerts its effects by acting as a cationic detergent, causing disruption of the integrity of the bacterial cell membrane, with leakage of intracellular contents and cell death [143]. Resistance to colistin has been postulated to occur via decreased affinity of lipopolysaccharides for colistin [143]. Colistin was used in the 1960s and 1970s, but was abandoned because of adverse side effects, including nephrotoxicity, neuromuscular blockade and neurotoxicity [144], and because of the emergence of newer and safer antimicrobials. Levin et al.[144] reported the outcomes following treatment with colistin of 60 nosocomial infections caused by multiresistant A. baumannii and P. aeruginosa. There was a good outcome for 35 (58%) patients, while three patients died within the first 48 h of treatment. The poorest results were observed in cases of pneumonia, where only five (25%) of 20 patients had a good outcome. The main adverse effect of treatment was renal failure (27% in patients with initial normal renal function, and 58% in patients with initial abnormal renal function); nevertheless, colistin can be recommended for the treatment of severe infections caused by multiresistant A. baumannii. In 2002, Jimenez-Mejias et al.[145] reported a case of meningitis caused by multiresistant A. baumannii which was treated successfully with intravenous colistin sulphomethate sodium (5 mg/kg/day). Colistin penetrated the cerebrospinal fluid at one-quarter the serum levels without adverse effects. In 1994, Go et al.[146] reported a nosocomial outbreak of infections caused by imipenem-resistant A. baumannii in a New York hospital. Infection and colonisation were eliminated by intensive infection control measures, and wound irrigation with polymyxin B.

Conclusions

Acinetobacter spp. are important nosocomial pathogens, capable of rapid adaptation to the hospital environment. There is no doubt that these organisms will pose continuing problems, which is disturbing because of the extent of their antibiotic resistance profiles. Although antimicrobial resistance of Acinetobacter spp. appears to be increasing across Europe, it is difficult to estimate accurately the extent of this emerging problem, in part because the published susceptibility data are based on different methods, and also because of population bias and clonal variation. A reference method for susceptibility testing and MIC breakpoints should be established to better monitor trends of resistance. Surveillance of antimicrobial resistance, the study of resistance mechanisms, the development of new drugs, and the prevention of the spread of multiresistant strains, are all important measures required to control the impact of these multiresistant bacteria.

Acknowledgements

This work was supported by a grant from the European Commission (QLK2-2001-00915).

Ancillary