INTRODUCTION — Acinetobacter is a gram-negative coccobacillus that has emerged from an organism of questionable pathogenicity to an infectious agent of importance to hospitals worldwide [1]. The organism has the ability to accumulate diverse mechanisms of resistance, leading to the emergence of strains that are resistant to all commercially available antibiotics [2].

Acinetobacter baumannii is one of the ESCAPE organisms, a group of clinically important, predominantly health care-associated organisms that have the potential for substantial antimicrobial resistance [3,4]. Other ESCAPE organisms are Enterococcus faecium, Staphylococcus aureus, Clostridioides difficile, Pseudomonas aeruginosa, and Enterobacteriaceae [4]. In addition, carbapenem-resistant A. baumanii is one of the critical-priority pathogens on the World Health Organization priority list of antibiotic-resistant bacteria for effective drug development [5].

The treatment and prevention of Acinetobacter infection will be reviewed here. The clinical features, epidemiology, microbiology, and pathogenesis of Acinetobacter infection are discussed separately. (See "Acinetobacter infection: Epidemiology, microbiology, pathogenesis, clinical features, and diagnosis".)

ANTIMICROBIAL RESISTANCE — Acinetobacter has the ability to develop resistance through several diverse mechanisms, which has led to emergence of strains that are resistant to all commercially available antibiotics [2].

Definitions — In 2011, a joint initiative by the European and United States Centers for Disease Control and Prevention (ECDC and CDC) proposed specific definitions for characterizing drug resistance in organisms that cause many health care-associated infections [6]. For Acinetobacter, the following definitions were established based on the extent of resistance to antibiotics that would otherwise serve as treatments for Acinetobacter (ie, cephalosporins, fluoroquinolones, and carbapenems) (table 1):

●Multidrug-resistant: isolate is non-susceptible to at least one agent in three or more antibiotic classes

●Extensively drug-resistant: isolate is non-susceptible to at least one agent in all but two or fewer antibiotic classes

●Pandrug-resistant: isolate is non-susceptible to all agents

Prior to this publication, there was no standard definition for the term "multidrug resistance," which partly accounts for the significant heterogeneity of clinical studies evaluating various regimens for drug resistant Acinetobacter infections.

Prevalence — Since the 1980s, resistant strains have become increasingly common causes of nosocomial infections globally [7-12]. In a 2009 report of surveillance data from more than 100 centers worldwide (Meropenem Yearly Susceptibility Test Information Collection; MYSTIC), 61 percent of Acinetobacter isolates were resistant to ceftazidime and 67 percent were resistant to ciprofloxacin [13]. These results are significantly worse than those published in 2007 from the same reporting system (34 and 40 percent resistance, respectively) [8]. Carbapenem and tobramycin susceptibility has decreased notably in this short span of time (92 versus 59 percent for tobramycin, and 86 to 92 versus 46 to 52 percent for carbapenems). The emergence of resistance among Acinetobacter strains has also been demonstrated by analysis of The Surveillance Network, an electronic passive surveillance database that collects information from clinical laboratories across the United States [14]. Multidrug resistance, defined as non-susceptibility to at least one agent in three or more antibiotic classes (excluding fluoroquinolones), increased from 21 percent during 2003 to 2005 to 35 percent during 2009 to 2012 [15]. Resistance to carbapenems doubled over this time from 21 to 48 percent.

Other studies have also demonstrated emergent carbapenem-resistant strains worldwide, with exceedingly high rates of carbapenem resistance in some locations [16-28]. As an example, all of the 51 A. baumannii strains isolated from patients in the intensive care unit (ICU) of an Italian tertiary care hospital in 2017 were carbapenem-resistant due to the bla OXA-23 and the bla OXA-24/40 genes [28].

In the countries of the Arab League, reported prevalence of carbapenem resistance among A. baumannii isolates is also quite high, ranging from 36 to 100 percent [29]. In some of those countries, production of OXA-23 is the main mechanism of resistance of Acinetobacter species [30]. (See 'Mechanisms' below.)

Heavy use of third generation cephalosporins, aztreonam, and imipenem has contributed to the problem of carbapenem resistance [18,31]. There are also reports of discordant susceptibility to carbapenems, in which an isolate is susceptible to imipenem but resistant to meropenem, and vice versa [32,33].

Although polymyxins, such as colistin, usually have in vitro activity against Acinetobacter [34,35], resistance to polymyxins has been observed [36,37]. In one surveillance report, resistance was detected among 2.7 percent of clinical isolates in Europe and 1.7 percent in North and Latin America [36]. In the study of The Surveillance Network described above, colistin resistance in isolates in the United States was 6.9 percent during 2009 to 2012 [15]. Heteroresistance, in which there is a resistant subpopulation detected within an otherwise susceptible population, has also been observed, although varying standards to detect these subpopulations has rendered assessment of the rate of heteroresistance to colistin difficult [38]. In one study, colistin-resistant strains paradoxically demonstrated enhanced susceptibility to most antibiotics tested and a reduced ability to form biofilm compared with colistin-susceptible strains [39].

The rising prevalence of antimicrobial resistance among A. baumannii isolates has influenced the epidemiology of serious hospital-acquired infections. In a systematic review, carbapenem-resistant and multidrug-resistant A. baumannii accounted for 65 and 59 percent, respectively, of all hospital-acquired infections among intensive care unit patients in Southeast Asia [40]. In some countries in the Arab League, A. baumannii is the most common cause of ventilator-associated pneumonia [41].

Risk factors — Independent risk factors for colonization or infection with resistant strains of Acinetobacter include the following [42-45]:

●Prior colonization with methicillin-resistant S. aureus (MRSA)

●Prior beta-lactam use, particularly carbapenems

●Prior fluoroquinolone use

●Bedridden status

●Current or prior intensive care unit admission

●Presence of a central venous catheter

●Recent surgery

●Mechanical ventilation

●Hemodialysis

●Malignancy

●Glucocorticoid therapy

Prognosis — Patients with infection due to resistant strains appear to have higher mortality than patients with infection due to susceptible strains [46]. In a systematic review of observational studies that included over 2500 patients with either carbapenem susceptible or resistant Acinetobacter infection, the overall mortality rate was 33 percent, and carbapenem resistance was associated with a greater risk of death (pooled odds ratio 2.22, 95% CI 1.66-2.98). Of note, patients with carbapenem-resistant infection were more likely to have severe underlying illness or receive inappropriate empiric antibiotic therapy, which are likely confounding variables that contribute to the excess mortality.

In a subsequent retrospective study of 205 patients with carbapenem-resistant A. baumannii bacteremia on inappropriate therapy, the 28-day mortality rate was 70 percent, and most patients who died did so within the first five days of a positive blood culture [47]. A sequential organ failure assessment (SOFA) score of ≥10 was the strongest predictor of mortality; other independent risk factors included immunocompromised state, vasopressor use, pneumonia, and diabetes mellitus.

In another retrospective cohort study that included 153 abdominal solid organ transplant recipients with a carbapenem-resistant gram-negative bacterial infection, A. baumannii was the most commonly isolated pathogen (n = 47) [48]. Eleven patients with A. baumannii infection died. Overall, mechanical ventilation, septic shock, and platelet count <50,000/microL were independent risk factors for mortality.

Mechanisms — Acinetobacter species are capable of accumulating multiple antibiotic resistance genes, leading to the development of multidrug-resistant or extensively drug-resistant strains [49,50]. Frequently expressed resistance mechanisms in nosocomial strains of Acinetobacter include beta-lactamases, alterations in cell-wall channels (porins) and efflux pumps:

●AmpC beta-lactamases are chromosomally encoded cephalosporinases intrinsic to all Acinetobacter baumannii. Usually, such beta-lactamases have a low level of expression that does not cause clinically appreciable resistance; however, the addition of a promoter insertion sequence ISAba1 next the ampC gene increases beta-lactamase production, causing resistance to cephalosporins [51].

The most troubling clinical resistance mechanism has been the acquisition of beta-lactamases in Acinetobacter, including serine and metallo-beta-lactamases, which confer resistance to carbapenems [52]. Acquired extended-spectrum beta-lactamase carriage occurs in Acinetobacter but is not as widespread as in K. pneumoniae or Escherichia coli [53]. (See "Extended-spectrum beta-lactamases" and "Overview of carbapenemase-producing gram-negative bacilli".)

●Porin channels in A. baumannii are poorly characterized; it is known that reduced expression or mutations of bacterial porin proteins can hinder passage of beta-lactam antibiotics into the periplasmic space, leading to antibiotic resistance.

●Overexpression of bacterial efflux pumps can decrease the concentration of beta-lactam antibiotics in the periplasmic space. To cause clinical resistance in Acinetobacter, efflux pumps usually act in association with overexpression of AmpC beta-lactamases or carbapenemases. Efflux pumps can remove beta-lactam antibiotics as well as quinolones, tetracyclines, chloramphenicol, and tigecycline [54].

A. baumannii can become resistant to quinolones through mutations in the genes gyrA and parC, and can become resistant to aminoglycosides by expressing aminoglycoside-modifying enzymes [52].

The mechanism of resistance of Acinetobacter to colistin appears to be associated with a mutation in the genes encoding the PmrA and B proteins; additional regulatory factors remain to be determined [55].

Heteroresistance, characterized by resistant subpopulations within a single strain, has been described in Acinetobacter strains [56]. (See "Overview of antibacterial susceptibility testing", section on 'Heteroresistance'.)

ANTIBIOTIC EFFICACY AND SAFETY — Inherent and acquired resistance limits the number of antimicrobial options for Acinetobacter. Very few trials have evaluated the efficacy and safety of different antimicrobial regimens for Acinetobacter infections. Thus, most support for the use of various antibiotics for Acinetobacter infections is based upon in vitro data and observational series. Most of these studies, however, are limited by their small sample sizes, variability in the severity of disease and comorbidities in the included patients, and the lack of a comparator group.

First line agents for susceptible organisms — When infections are caused by antibiotic-susceptible Acinetobacter isolates, there may be several therapeutic options, including a broad-spectrum cephalosporin (ceftazidime or cefepime), a combination beta-lactam/beta-lactamase inhibitor (ie, one that includes sulbactam), or a carbapenem (eg, imipenem or meropenem). Dosing is summarized separately (table 2).

Carbapenems are highly bactericidal against susceptible strains of Acinetobacter [57]. The clinical cure rates with imipenem for ventilator-associated pneumonia due to Acinetobacter range from 57 to 83 percent in small series [54-56]. Because isolates that are susceptible to imipenem may be resistant to meropenem, and vice versa, susceptibility to the specific carbapenem should be confirmed prior to its use.

The beta-lactamase inhibitor sulbactam also has excellent bactericidal activity against Acinetobacter isolates [57]. In the United States, sulbactam is available only in combination with ampicillin [58-60]. Data from observational studies suggest that ampicillin-sulbactam is comparable in efficacy to imipenem [58,60]. Even strains that are resistant to carbapenems sometimes retain in vitro susceptibility to sulbactam [61], and in one small observational series, ampicillin-sulbactam appeared to be effective in patients with imipenem-resistant ventilator-associated pneumonia [62]. The adverse effects of beta-lactams are discussed in detail elsewhere. (See "Beta-lactam antibiotics: Mechanisms of action and resistance and adverse effects", section on 'Adverse effects'.)

Emergence of resistance during therapy has been observed with ampicillin-sulbactam, cephalosporins, and carbapenems when used as single agents [9,63]. For this reason, these agents are sometimes used in combination with an antipseudomonal fluoroquinolone or an aminoglycoside [32,33]. However, there are no clinical data demonstrating that combination therapy either reduces the risk of emergent resistance during therapy or improves clinical outcomes in cases of Acinetobacter infections. (See 'Combination therapy' below.)

Alternative agents for resistant organisms — In the setting of resistance to the above agents, therapeutic options are limited. Polymyxins are the main therapeutic options for extensively drug-resistant Acinetobacter. Certain tetracyclines (minocycline and tigecycline) may also have a role. Dosing is summarized separately (table 2).

Polymyxins — Polymyxins (polymyxin B and colistin [polymyxin E]) are the most commonly used agents for Acinetobacter isolates resistant to first-line agents. The dose depends on the formulation of colistin available, which varies by geographic region. (See "Polymyxins: An overview", section on 'Intravenous administration'.)

There are no randomized trials addressing their efficacy, largely because they are reserved for use in the setting of highly resistant organisms. Colistin has been used with some success for the treatment of Acinetobacter pneumonia, bacteremia, and meningitis [64,65]. In a meta-analysis of six studies (359 patients) evaluating treatment of ventilator-associated pneumonia due mainly to A. baumannii but also P. aeruginosa, clinical improvement rates, 28-day mortality, and ICU lengths of stay with intravenous colistin were similar to those observed with a comparator agent (carbapenem or high-dose ampicillin-sulbactam) [66]. Among nine studies (178 patients) that did not include a comparator treatment, the pooled clinical response rate for intravenous colistin was 66 percent. However, one small series of 20 cases of nosocomial pneumonia that was not included in the analysis reported a success rate of only 25 percent [64]. Despite the limitations of the meta-analysis, including heterogeneity of the study design and patient populations evaluated, we believe the available data suggest that intravenous colistin (or polymyxin B) remains an effective option for patients with colistin-sensitive Acinetobacter that is resistant to other agents.

Nephrotoxicity is the most notorious adverse effect associated with systemic colistin and has been reported in up to 36 percent of patients [67], although in the meta-analysis above, intravenous colistin use was not associated with excess renal dysfunction compared with the other agents evaluated [66]. Neurotoxicity is another important side effect but consists mainly of paresthesias and is relatively uncommon. Colistin dosing depends on the available preparation and should be adjusted in patients with impaired renal function. Polymyxin B is associated with lower rates of nephrotoxicity than colistin and does not need to be dose-adjusted for renal function. (See "Polymyxins: An overview", section on 'Adverse reactions'.)

Minocycline — Many resistant strains of A. baumannii are susceptible in vitro to minocycline, which can be given intravenously, and limited clinical experience suggests favorable outcomes with its use.

In a review of retrospective case series in which minocycline was used for multi- or extensively resistant A. baumannii infections, predominantly ventilator-associated pneumonia, but also skin and soft tissue infections, successful clinical and microbiologic outcomes were reported for most patients [68]. As an example, in a retrospective study of patients with ventilator-associated, carbapenem-resistant A. baumannii pneumonia, the clinical response rate was 80 percent among the 19 who were treated with minocycline [69]. Although approximately two-thirds of them received minocycline in combination with at least one other agent, response rates were similar between those who received minocycline monotherapy and combination therapy. In another retrospective study of 55 individuals with a multidrug-resistant, but minocycline susceptible A. baumannii infection (mainly pneumonia and bacteremia), almost all received at least one other agent, and the clinical response rate was 73 percent [70]. The infection-related mortality rate was 25 percent.

In vitro susceptibility to minocycline can be inferred from susceptibility to tetracycline, although minocycline retains activity against some tetracycline-resistant strains. Among nearly 5500 A. baumannii strains collected from medical centers worldwide from 2007 to 2011, 79 percent were susceptible in vitro to minocycline (minimum inhibitory concentration [MIC] ≤4 µg/mL) compared with 30 and 60 percent to tetracycline and doxycycline, respectively [71].

Overall, both intravenous and oral minocycline are well-tolerated, but like other agents in the tetracycline class can cause photosensitivity and gastrointestinal side effects [68].

Tigecycline — Tigecycline has activity against some multidrug- and extensively drug-resistant strains of A. baumannii, although resistance has been reported and clinical experience is limited [36,54,61,72-77]. In general, however, tigecycline should not be used in circumstances in which other effective antibiotic choices are available.

Tigecycline was comparable to colistin in one retrospective study from Korea [78]. Among 70 critically ill patients with multidrug-resistant A. baumannii pneumonia, clinical success rates were similar (47 and 48 percent) between those who received tigecycline- versus colistin-based therapy (both mono- and combination therapy regimens). In a separate retrospective study of 21 patients with carbapenem-resistant Acinetobacter infections, tigecycline was used as monotherapy in seven patients and as part of a combination therapy in 14 patients. Most patients had surgical site infections followed by ventilator-associated pneumonia. A favorable response was attained in 81 percent of cases; ventilator-associated pneumonia was associated with worse outcomes [79]. In contrast, in a multicenter study from Argentina, 73 patients with ventilator-associated pneumonia due to Acinetobacter were treated with tigecycline, with a success rate of around 70 percent [80]. Notably, in the first study, clinical response did not predict microbiological response, highlighting the difficulty in eradicating this microorganism [79].

A higher dose of tigecycline may be an option. A retrospective study from Italy suggested that tigecycline was well tolerated at a higher than standard dose in critically ill patients (many of whom had ventilator-associated pneumonia) with multidrug-resistant gram-negative infections, including Acinetobacter infections [81]. The higher dose was associated with better outcomes than standard dosing.

There has been concern with potential adverse outcomes with tigecycline. In a meta-analysis of studies evaluating tigecycline for multidrug-resistant A. baumannii infections, there were no differences in all-cause mortality and clinical response with tigecycline versus the comparators, but tigecycline was associated with a lower microbial eradication rate and a trend for longer hospitalization [82]. In previous studies of tigecycline for various infections, it had been associated with an increased risk of all-cause mortality compared with other agents, most clearly among patients with hospital-acquired pneumonia [83,84]. In addition, tigecycline rapidly enters tissues following administration, which results in low serum levels; thus it may not be appropriate for cases of Acinetobacter bacteremia [57].

Other agents

●Cefiderocol – Most isolates of extensively drug-resistant A. baumannii, including those possessing OXA-type beta-lactamases, remain susceptible to cefiderocol [85]. However, clinical experience is limited. Furthermore, in a randomized trial comparing cefiderocol to best-available therapy for carbapenem-resistant gram-negative infections (including 54 patients with A. baumannii infections), there was a trend toward higher all-cause mortality with cefiderocol by the end of the study (34 versus 18 percent), despite comparable clinical cure rates [86]. The mortality difference was more pronounced among patients with Acinetobacter spp infections (50 versus 18 percent).

●Eravacycline – This broad-spectrum tetracycline is active in vitro against most carbapenem-resistant A. baumannii, at MICs that are lower than with tigecycline [87,88]. However, clinical experience is extremely limited.

The novel beta-lactam-beta-lactamase combinations (eg, ceftazidime-avibactam, meropenem-vaborbactam, imipenem-cilastatin-relebactam) do not have activity against carbapenem-resistant A. baumannii [89].

Combination therapy — Combination antimicrobial therapy is frequently used in Acinetobacter infections as a strategy to increase the likelihood of adequate empiric antibiotic coverage before drug susceptibility testing results are known, to decrease the risk of emergent resistance, and to improve outcomes in multidrug or extensively drug-resistant infections, but there are no definitive clinical data to support its use for these purposes. Nevertheless, because of the excess mortality rate associated with inappropriate empiric antibiotic therapy and with drug-resistant infections, we use a combination antimicrobial regimen for empiric therapy of Acinetobacter infections when local rates of resistance to the chosen antibiotic are high and for directed therapy in the setting of infection with extensively drug-resistant isolates. The approach to combination regimen selection is discussed elsewhere. (See 'Empiric therapy' below and 'Resistant isolates' below.)

Some, but not all [90], clinical evidence suggests that combination therapy is associated with improved outcomes for drug-resistant Acinetobacter infections [91-94].

●In a retrospective study that included over 300 adults with Acinetobacter ventilator associated pneumonia, empiric monotherapy (as opposed to combination therapy) was independently associated with 30-day mortality [91]. Among those with imipenem-resistant infections, 30-day mortality rates were higher with monotherapy than with combination therapy (67 versus 52 percent). In another retrospective study that included 83 critically ill patients with extensively drug-resistant Acinetobacter infections, therapy with polymyxin B plus another agent was associated with a lower 30-day mortality rate than with polymyxin B monotherapy (42 versus 68 percent) [93].

●However, in a multicenter trial, 210 patients with serious infections caused by Acinetobacter isolates susceptible only to colistin were randomly assigned to receive colistin alone or colistin with intravenous rifampicin [90]. Combination therapy did not decrease the 30-day mortality rate (43.3 versus 42.9 percent with colistin alone) or the infection-related death rate (21.5 versus 26.6 percent) despite increasing the likelihood of microbiological cure (60.6 versus 44.8 percent). The trial results contrast with prior in vitro studies, which had demonstrated synergistic or additive effects by combining polymyxins with rifampin [95].

Very few studies have been performed to evaluate the effect of combination therapy on the emergence of resistance during treatment for Acinetobacter, and there is no evidence that it decreases the risk [96]. In a meta-analysis of eight trials that compared combination beta-lactam and aminoglycoside therapy with beta-lactam monotherapy for the treatment of infections caused by various organisms, combination therapy did not decrease the rate of emergent resistance (odds ratio 0.9, 95% CI 0.56-1.47) [97]. Among Acinetobacter infections specifically, there was no difference in the development of resistance between combination and monotherapy treated cases (0 of 11 versus 1 of 22 infections, respectively).

Other combinations that have favorable effects on multi- and extensively drug-resistant isolates in vitro, or in animal models include a carbapenem with colistin, tigecycline with colistin [38], vancomycin with colistin, minocycline with colistin, and meropenem and fosfomycin with colistin [38,98-100]. However, there are limited clinical data evaluating these. With the emergence of polymyxin-resistant bacteria, well-designed studies to understand the optimal clinical use of such combinations are essential.

Other strategies for resistant organisms — Carbapenems have been used in extended-infusion dosing (eg, infused over 3 to 4 hours) for the treatment of serious infections caused by multidrug-resistant gram-negative organisms, but specific data on extended infusion therapy in the setting of multidrug-resistant Acinetobacter infections is scarce [101-103]. Existing data suggest that it is a safe and cost-effective strategy in serious infections, and may be associated with improved outcomes in some settings.

In a small trial of thirty patients in an intensive care unit with hospital-acquired pneumonia due to drug-resistant Acinetobacter, clinical efficacy and relapse rates were similar among patients randomly assigned to receive 500 mg meropenem infused over three hours every six hours versus 1 g infused over one hour every eight hours [103]. However, the percent of time that antibiotic levels remained above the minimum inhibitory concentration was higher in the prolonged infusion group (75 versus 54 percent), and the overall cost of antibiotics was less with prolonged infusion.

There are more data on use of extended infusion of carbapenems for serious hospital acquired pneumonia due to drug-resistant organisms. In a trial of 531 patients with ventilator-associated pneumonia, clinical cure rates were similar among patients randomly assigned to receive an extended infusion of doripenem (a carbapenem that is not widely available) or standard dosing of imipenem (68 versus 64 percent clinical cures, respectively) [102]. However, among older and more severely ill patients, clinical cure rates were higher with doripenem. In a small study of 42 patients with life-threatening pneumonia, extended infusion of meropenem was associated with lower mortality compared with standard administration [101]. Use of extended infusion dosing for the treatment of hospital acquired pneumonia is discussed elsewhere. (See "Treatment of hospital-acquired and ventilator-associated pneumonia in adults", section on 'Prolonged infusions'.)

New antimicrobial agents targeting multidrug-resistant A. baumannii are under development; certain agents are available in some countries, although clinical data informing their use for A. baumannii remain very limited [104]. Phage therapy has been promising as a novel therapeutic approach, but clinical data are limited to a few case reports [105,106].

GENERAL APPROACH TO ANTIMICROBIAL SELECTION

Empiric therapy — Empiric antibiotic therapy for Acinetobacter, before results of antimicrobial susceptibility testing are available, should be selected based on local susceptibility patterns. In general, it should consist of a broad spectrum cephalosporin, a combination beta-lactam/beta-lactamase inhibitor (eg, a combination including sulbactam), or a carbapenem. An additional agent may be warranted if local resistance rates to the chosen antibiotic class are high (eg, greater than 10 to 15 percent). Dosing is summarized in the Table (table 2).

When rates of resistance to the selected antimicrobial agent are low (ie, below 10 to 15 percent), monotherapy is likely adequate as there are no data to clearly demonstrate that combination therapy improves outcomes through synergistic effect. However, when rates of resistance are higher, it is reasonable to use one of the agents above in combination with an antipseudomonal fluoroquinolone, an aminoglycoside, or colistin to improve the likelihood of administering an antibiotic agent that retains activity. While there are no clear clinical data to support this practice for Acinetobacter infections, many experts favor empiric combination therapy for serious infections with these and other potentially resistant gram-negative organisms because of the increased mortality associated with inappropriate empiric therapy. (See "Gram-negative bacillary bacteremia in adults", section on 'Indications and rationale for combination therapy' and "Treatment of hospital-acquired and ventilator-associated pneumonia in adults", section on 'Gram-negative pathogens'.)

Other considerations in choosing empiric therapy include prior colonization and infection with Acinetobacter, in which case an empiric regimen should be chosen to be effective against prior isolates, and recent receipt of antibiotics, which may warrant selection of a different class of antibiotics.

Directed therapy — Once results of antimicrobial susceptibility testing are available, a regimen can be chosen from among the active agents. (See 'Isolates susceptible to first line agents' below and 'Resistant isolates' below.)

Additional management issues specific to particular infections, including the use of inhaled or intrathecal antibiotics, are discussed below. (See 'Disease specific considerations' below.)

Isolates susceptible to first line agents — If results of antimicrobial susceptibility testing reveal susceptibility to beta-lactams or carbapenems, an agent from one of these classes can be chosen as monotherapy. We favor choosing the agent with the narrowest spectrum of activity. Other considerations in selection of a regimen include patient drug allergies or intolerance, need to cover additional infections, and hospital formulary.

With any of these agents, there is the risk of resistance emerging during therapy. However, there are no data to demonstrate that adding a second agent limits this risk and we do not routinely use a second agent for this purpose in patients who have infections with susceptible strains. (See 'Combination therapy' above.)

For cases of Acinetobacter central nervous system infections, variable penetration of antibiotics into the cerebrospinal fluid further limits the selection of antibiotics. (See 'Meningitis' below.)

Resistant isolates — In the setting of resistance to first line agents, therapeutic options are generally limited to polymyxins (colistin [polymyxin E] and polymyxin B), minocycline, and tigecycline. We generally use polymyxins, for which there is the most clinical experience in treating extensively drug-resistant Acinetobacter. Furthermore, tigecycline may not reach adequate levels in the serum, urinary tract, or CNS to successfully treat infections in these compartments [53,107]. Susceptibility testing for these agents should be performed as well prior to their use given the possibility of resistance. (See 'Alternative agents for resistant organisms' above.)

We generally favor using a second agent, such as a carbapenem, minocycline, tigecycline, or rifampin, in addition to polymyxins for serious infections (eg, bacteremia, pneumonia, critical illness) with resistant isolates. There are no definitive clinical data that demonstrate improved outcomes with combination versus monotherapy, and some randomized trials have suggested that certain combinations (colistin and rifampin or colistin and meropenem) resulted in comparable clinical outcomes as monotherapy with colistin [90,108]. Nevertheless, infections with multidrug-resistant Acinetobacter are associated with high mortality rates, and we are concerned that the use of a single agent is not adequate, particularly since resistance can develop during therapy, leaving no therapeutic alternatives. (See 'Combination therapy' above.)

In cases of multi- or extensively drug-resistant Acinetobacter infections, management should be individualized and consultation with an expert in the management of such infections is advised.

Monitoring — Because of the possibility of emergent resistance to antibiotics during therapy, continued monitoring of the patient is important. If there is clinical decline following initial improvement on therapy, repeat cultures to evaluate for growth of resistant Acinetobacter isolates is warranted. Therapy can then be changed according to new susceptibility testing results.

DISEASE SPECIFIC CONSIDERATIONS

Pneumonia — The initial approach to empiric and directed antimicrobial therapy of pneumonia caused by Acinetobacter is the same as that for Acinetobacter infections in general, as described above (see 'General approach to antimicrobial selection' above). Additional considerations include the possible use of adjunctive inhaled antibiotics.

Inhaled colistin may be beneficial in select patients [66,109-111], although not all studies suggest a benefit [112]. We favor use of inhaled colistin among patients with severe pneumonia due to Acinetobacter that is resistant to beta-lactams and carbapenems (ie, sensitive to colistin only), since intravenous colistin yields low lung concentration. If other options are available, we avoid the use of colistin, whether intravenous or inhaled. Among three studies evaluating inhaled colistin as adjunctive therapy to intravenous antibiotics for ventilator-associated pneumonia with drug-resistant gram-negative bacilli, predominantly A. baumannii, the pooled response rate was 80 percent [66]. Although in one series of 17 patients with Acinetobacter pneumonia, clinical improvement with inhaled colistin without active systemic antibiotics was observed in 57 percent of cases [109], we favor using inhaled colistin in such patients only with concomitant administration of intravenous antibiotics. The main adverse effect of inhaled colistin is bronchoconstriction [67]. The optimal dose of inhaled colistin is uncertain and ranges from 75 to 150 mg colistin base activity (2.25 to 4.5 million international units CMS) twice daily. Higher doses, up to 5 million international units colistimethate sodium (approximately 167 mg colistin base) every eight hours, have also been used for ventilator-associated pneumonia with Acinetobacter [113].

The duration of therapy is similar to that for other causes of pneumonia and is discussed separately. (See "Treatment of community-acquired pneumonia in adults who require hospitalization", section on 'Duration of therapy' and "Treatment of hospital-acquired and ventilator-associated pneumonia in adults", section on 'Duration'.)

Although there are reports of successful treatment of multidrug-resistant respiratory infections with inhaled and systemic polymyxin B [114], we do not typically use inhaled polymyxin B because of the risk of bronchospasm. (See "Polymyxins: An overview", section on 'Inhaled administration'.)

Bloodstream infection — The initial approach to empiric and directed antimicrobial therapy of bloodstream infections caused by Acinetobacter is the same as that for Acinetobacter infections in general, as described above (see 'General approach to antimicrobial selection' above). However, because tigecycline rapidly enters tissues following administration, resulting in low serum levels, it may not be an appropriate choice for extensively drug-resistant Acinetobacter bacteremia.

Additionally, bloodstream infections with drug-resistant isolates are associated with particularly poor outcomes, regardless of therapy. In one study, treatment with colistin did not reduce mortality in patients with bacteremia due to multidrug-resistant Acinetobacter compared to mortality rates prior to the availability of colistin in that institution [115].

Additional considerations depend on other features of the bloodstream infection. If the bacteremia is associated with an intravascular catheter, that device should be removed [116]. The duration of therapy is typically 10 to 14 days. (See "Intravascular non-hemodialysis catheter-related infection: Treatment".).

If endocarditis accompanies Acinetobacter bacteremia, the treatment issues are similar to endocarditis due to other gram-negative organisms. (See "Antimicrobial therapy of left-sided native valve endocarditis", section on 'Other gram-negative organisms' and "Antimicrobial therapy of prosthetic valve endocarditis", section on 'Other gram-negative organisms' and "Antimicrobial therapy of left-sided native valve endocarditis", section on 'Clinical response to initial therapy' and "Surgery for left-sided native valve infective endocarditis" and "Surgery for prosthetic valve endocarditis".)

Meningitis — The initial approach to empiric and directed antimicrobial therapy of meningitis caused by Acinetobacter is similar to that for Acinetobacter infections in general, as described above (see 'General approach to antimicrobial selection' above). However, variable cerebrospinal fluid (CSF) penetration of antibiotic agents further limits the therapeutic choices for Acinetobacter central nervous system (CNS) infections. Additional considerations include the possible use of intrathecal antibiotics for drug-resistant isolates and the removal of CNS devices, if present [117].

Of the first line agents, the carbapenems most reliably enter into the CSF, particularly when inflammation of the meninges is minimal [118]. Because of the association between high dose imipenem and seizures, meropenem is the most appropriate choice of the carbapenems. For CNS infections, the meropenem dose should be 2 g every eight hours. If susceptible, ceftazidime or cefepime can also be used at meningeal doses.

For carbapenem-resistant isolates, polymyxins have been used with some success [118]. When colistin is administered intravenously, there is moderate penetration of inflamed meninges and spinal fluid levels reach approximately 25 percent of serum levels [64,119]. For this reason, we also use intrathecal or intraventricular colistin in the setting of central nervous system infections with drug-resistant Acinetobacter.

Colistin can be administered intraventricularly or intrathecally, usually in conjunction with an active intravenous agent, if possible [107,120-123]. The dose range of intraventricular or intrathecal colistin therapy has varied widely from 0.75 to 7.5 mg colistin base activity (25,000 to 250,000 international units colistimethate sodium) total daily dose [120]. Complications of intraventricular and intrathecal colistin include aseptic chemical meningitis or ventriculitis; dose reduction is required as the CSF white blood cell count increases [120]. (See "Gram-negative bacillary meningitis: Treatment", section on 'Intrathecal and intraventricular therapy'.)

If intrathecal or intraventricular colistin is being considered for treatment of an Acinetobacter central nervous system infection, consultation with an expert in the management of such infections is advised.

The treatment of Acinetobacter meningitis is usually at least three weeks. The response should be assessed clinically and with repeat CSF cultures.

Skin, soft tissue, and bone infection — The initial approach to empiric and directed antimicrobial therapy of skin, soft tissue, and bone infections caused by Acinetobacter is the same as that for Acinetobacter infections in general, as described above (see 'General approach to antimicrobial selection' above). In addition, debridement of affected tissue, particularly in the case of osteomyelitis, may be necessary for optimal control of the infection.

The usual duration of therapy for skin and soft tissue infections is 10 to 14 days or until local signs of infection have resolved. Patients with osteomyelitis should be treated for 4 to 6 weeks following surgical debridement. (See "Nonvertebral osteomyelitis in adults: Treatment".)

Urinary tract infection — The initial approach to empiric and directed antimicrobial therapy of urinary tract infections caused by Acinetobacter is the same as that for Acinetobacter infections in general, as described above. (See 'General approach to antimicrobial selection' above.)

Because Acinetobacter readily colonizes the urinary tract, particularly in the presence of an indwelling catheter, treatment for infection should only be initiated if a positive culture is accompanied by pyuria and systemic signs or symptoms in the absence of another source of infection.

Additionally, tigecycline has poor excretion in the urinary tract. Given this, along with the observed increased risk of mortality with tigecycline use [83,84], this agent should only be used when no other options are available.

If present, any urinary catheter should be removed. The duration of therapy is typically 10 to 14 days. (See "Catheter-associated urinary tract infection in adults", section on 'Treatment'.)

Other infections — Acinetobacter infection of the eye can include corneal ulcers, endophthalmitis, periorbital cellulitis, and infection after penetrating trauma [124-129]. Treatment often consists of topical, subconjunctival, or intravitreal ophthalmic antibiotic preparations, guided by susceptibility results. The duration of treatment is generally similar to that for infections caused by other gram-negative bacilli and depends on the site of infection. (See "Bacterial endophthalmitis" and "Orbital cellulitis".)

Acinetobacter can cause nosocomial sinusitis in patients admitted to the intensive care unit; mechanical ventilation is the most important predisposing factor [130,131]. Treatment of Acinetobacter sinusitis consists of nasal tube removal and sinus drainage and lavage.

Acinetobacter peritonitis has been described in patients undergoing peritoneal dialysis [132-134]. The management of peritonitis in patients undergoing peritoneal dialysis is discussed separately. (See "Microbiology and therapy of peritonitis in peritoneal dialysis".)

PREVENTION AND CONTROL — The goals for control of multidrug-resistant Acinetobacter are early recognition, aggressive control of spread, and preventing establishment of endemic strains. General principles of infection control, as well as strategies to prevent health care-associated infections, are essential and are discussed in detail elsewhere. (See "Infections and antimicrobial resistance in the intensive care unit: Epidemiology and prevention".)

With regards to environmental cleansing, disinfection is particularly important because of the ability of Acinetobacter to survive on inanimate surfaces and contaminate other surfaces that contact it [135]. As an example, in a study of 199 interactions between health care personnel and patients colonized with multidrug-resistant Acinetobacter, 39 percent resulted in contamination of gloves and/or gowns of the health care personnel, a more frequent rate than that observed for multidrug-resistant Pseudomonas [136]. Multidrug-resistant Acinetobacter remains largely susceptible to disinfectants and antiseptics; occasional reports of failure are more likely to represent failure of personnel to follow cleaning procedures than disinfectant resistance [137].

Control is most successful when a common source is identified and eliminated [1,138]. Aggressive and monitored cleaning of environmental reservoirs (using ultraviolet markers of cleaning efficacy) is also important [138]. Hydrogen peroxide vapor was found to be an effective decontamination method [139]. When neither common sources nor environmental reservoirs are identified, control depends on active surveillance, contact isolation, health care worker compliance with hand hygiene, and aseptic care of vascular catheters and endotracheal tubes [138,140]. (See "Infection prevention: Precautions for preventing transmission of infection".)

SUMMARY AND RECOMMENDATIONS

●Acinetobacter has the ability to develop resistance through several diverse mechanisms, leading to the emergence worldwide of drug-resistant strains, which are more difficult to treat and are associated with a higher mortality than susceptible strains. Health care exposures, including prior antibiotic receipt (particularly carbapenems and fluoroquinolones), are associated with colonization and infection due to drug-resistant isolates. (See 'Antimicrobial resistance' above.)

●Most support for the use of various antibiotics for Acinetobacter infections is based upon in vitro data and observational series. Very few trials have evaluated the efficacy and safety of different antimicrobial regimens for Acinetobacter infections. When infections are caused by antibiotic-susceptible Acinetobacter isolates, there may be several therapeutic options, including a broad-spectrum cephalosporin (ceftazidime or cefepime), a combination beta-lactam/beta-lactamase inhibitor (ie, one that includes sulbactam), or a carbapenem (eg, imipenem or meropenem). In the setting of resistance to the above agents, therapeutic options are polymyxins and possibly tigecycline. Dosing is summarized separately (table 2). (See 'Antibiotic efficacy and safety' above.)

●Empiric antibiotic therapy for Acinetobacter, before results of antimicrobial susceptibility testing are available, should be selected based on local susceptibility patterns. In general, it should consist of a broad spectrum cephalosporin, a combination beta-lactam/beta-lactamase inhibitor (eg, a combination including sulbactam), or a carbapenem. For empiric therapy of patients with Acinetobacter infection in a location where resistance to the chosen antibiotic is high, we suggest addition of a second agent pending susceptibility results (Grade 2C). An antipseudomonal fluoroquinolone, an aminoglycoside, or colistin are second agent options. (See 'Empiric therapy' above.)

●Once results of antimicrobial susceptibility testing are available, a regimen can be chosen from among the active agents. We favor choosing the agent with the narrowest spectrum of activity. For patients with infections due to extensively drug-resistant Acinetobacter, therapeutic options are generally limited to polymyxins (colistin and polymyxin B) and possibly certain tetracyclines (minocycline and tigecycline). For such patients, we suggest using a second agent in addition to one of these (Grade 2C). (See 'Directed therapy' above.)

●Because of the possibility of emergent resistance to antibiotics during therapy, continued monitoring of the patient for clinical worsening following initial improvement is important. (See 'Monitoring' above.)

●Adequate management of Acinetobacter infections also includes removal of associated foreign material, such as urinary or venous catheters. In patients who have Acinetobacter pneumonia resistant to beta-lactams and carbapenems and thus receive an alternate intravenous antibiotic, we suggest inhaled colistin as adjunctive therapy (Grade 2C). Variable cerebrospinal fluid (CSF) penetration of antibiotic agents further limits the therapeutic choices for Acinetobacter central nervous system (CNS) infections, for which higher doses of antibiotics is generally warranted. Additional considerations include the possible use of intrathecal antibiotics for drug-resistant isolates and the removal of CNS devices, if present. (See 'Disease specific considerations' above.)

●Prevention of drug-resistant Acinetobacter depends on early recognition, aggressive control of spread, and preventing establishment of endemic strains. Drug-resistant Acinetobacter remains largely susceptible to disinfectants and antiseptics. (See 'Prevention and control' above and "Infection prevention: Precautions for preventing transmission of infection".)