Carbapenems are usually the drug of choice for serious Acinetobacter infections; nevertheless their utility is increasingly compromised by the rapid emergence of resistance (Peleg et al., 2008; Doi et al., 2015). Acinetobacter spp. isolates (n = 21) from the UMMC, which is located in the capital city of Kuala Lumpur, and collected in 1987 showed imipenem resistance rates of only 4.8% but a decade after that, imipenem resistance rates have increased to 36.4% for isolates collected between 1996 and 1998 (n = 88) (Figure 2) (Misbah et al., 2004). The first NSAR data in 2003 showed that the national resistance rate for meropenem was slightly below 30% and this was also reflected in a study of isolates from HUSM, located in the northeastern state of Kelantan, from 2003–2004 (Deris et al., 2009). However, by 2008, the NSAR data showed that the resistance rates for meropenem as well as imipenem have reached 50%. Nevertheless, there has not been any drastic increase in the nationwide carbapenem resistance rates from 2008–2016 which has stayed around 50–60%. Several studies on A. baumannii isolates from individual hospitals showed carbapenem resistance rates higher than the national average: ICU isolates from the UMMC collected from 2006–2009 showed very high resistance rates for imipenem at 96.5% and meropenem at 98.2% (Kong et al., 2011), as did isolates from several ward in Hospital Selayang (located also in Kuala Lumpur) in 2010 with a 92.5% resistance rate for meropenem whereas the imipenem resistance rate was lower at 67.5% (Nazmul et al., 2012). Likewise, A. baumannii isolates collected in 2010 and 2011 from various ward in HSA in the southern state of Johor, displayed resistance rates of 88% for both imipenem and meropenem (Dhanoa et al., 2015). Resistance rates of >70% were also reported for isolates from UKMMC (located south of Kuala Lumpur) in 2010–2011 (Biglari et al., 2015, 2017) and HSNZ (located in the east coast state of Terengganu) in 2011(Lean et al., 2014).

The national A. baumannii resistance rates for the extended-spectrum cephalosporins of the third generation, ceftazidime, and the fourth generation, cefepime, were around 30% in 2003 but increased to around 50% between 2005 and 2009 (Figure 3). The resistance rates for both ceftazidime and cefepime remained within the 50–60% range throughout 2010–2014. From 2015 onward NSAR only reported rates for ceftazidine, which maintained between 55 and 60%. Reports of strains that were isolated from individual hospitals showed higher resistance rates for ceftazidime and cefepime when compared to the national average: strains from HSA in 2010 and 2011(Dhanoa et al., 2015) showed resistance rates of nearly 90% whereas strains from UKMMC from 2010 and 2011 (Biglari et al., 2015) and HSNZ in 2011 (Lean et al., 2014) showed resistance rates of around 70%. Ceftazidime resistance rates for A. baumannii isolates from Hospital Selayang in 2010 (Nazmul et al., 2012) were closer to the national resistance rate of 58% for that year, as was the resistance rate for cefepime of isolates from UMMC in 2008–2009 (51%) although the resistance rate for ceftazidime was about 10% higher than the national resistance rate for that period of time (Dhabaan et al., 2012). In stark contrast, all 170 isolates obtained from the ICU of UMMC in 2006–2009 were resistant to ceftazidime and cefepime (Kong et al., 2011). Very high ceftazidime resistance rates had earlier been reported for Acinetobacter spp. isolates from UMMC that were isolated in 1987 (81%) and between 1996 and 1998 (97.7%) (Misbah et al., 2004).

The resistance rates for another third generation extended-spectrum cephalosporin, cefotaxime, were consistently higher than ceftazidime and cefepime (Figure 3). NSAR first reported the national resistance rates for cefotaxime in 2007 and this was already at 75.4%. An earlier study from HUSM from 2003–2004 showed an even higher cefotaxime resistance rate at 88% (Deris et al., 2009) and this reached 94.7% in strains isolated from the same hospital between 2005 and 2009 (Ariffin et al., 2012). The national resistance rates for cefotaxime remained above 70% for 2009–2012 but dipped slightly below 70% in 2013–2014. Cefotaxime resistance rates for UKMMC in 2010–2011 (Biglari et al., 2015) and HSNZ in 2011 (Lean et al., 2014) were similar to the national resistance rate at that time frame (i.e., around 70%). Interestingly, cefotaxime resistance for Acinetobacter spp. isolates from UMMC from 1987 was even higher at 81% and this further increased to 97.7% in isolates obtained from 1996–1998 (Misbah et al., 2004). No data for cefotaxime were available in the NSAR reports for 2015 and 2016.

No NSAR data is also available for the fourth extended-spectrum cephalosporin that was listed in the CLSI and the ECDC-CDC guidelines, i.e., ceftriaxone. However, data from Acinetobacter spp. isolates obtained from UMMC in 1987 showed a high resistance rate of 90.5% and this further increased to 97.7% for isolates in 1996–1998 (Misbah et al., 2004). By the following decade, a 100% resistance rate to ceftriaxone was reported for Acinetobacter isolates from the UMMC ICU (collected from 2006–2009) (Kong et al., 2011).

The NSAR report from 2003 showed a nationwide gentamicin resistance rate of 39.1% and an amikacin resistance rate that is four-fold lower at 8.8%. Resistance rates steadily increased and by 2008, the resistance rates for both aminoglycosides were similar although the rates for amikacin were around 2–5% lower than that of gentamicin (Figure 4). Throughout this period, gentamicin resistance rates increased from 39.1% in 2003 to about 50% in 2010 and remained around that level until the latest NSAR report for 2016. When looking at the aminoglycoside resistance data from individual hospitals as reported by other groups of researchers, the resistance rates for gentamicin were generally higher than for amikacin as shown in the NSAR data (Figure 4). However, isolates from three hospitals showed around 20% higher resistance rates than the NSAR data: UKMMC in 2010–2011 (70.2% for gentamicin) (Biglari et al., 2015), HSNZ in 2011 (66.7% for gentamicin, 57.4% for amikacin) (Lean et al., 2014) and HSA in 2011–2012 (79.5% for gentamicin, 72.4% for amikacin) (Dhanoa et al., 2015). A random sample of 42 A. baumannii isolates from various hospitals in Malaysia taken from 2008–2009 yielded a gentamicin resistance rate of 76.2% (Kim et al., 2013), which is also above the national resistance rate as reported by NSAR, although for this particular study, the isolates chosen were all carbapenem resistant.

Only ciprofloxacin from the fluoroquinolone group of antimicrobials has been used to assess the antimicrobial susceptibility rates for Acinetobacter spp. in Malaysia. The NSAR data showed that ciprofloxacin resistance rates increased from about 20% in 2003 to around 50% in 2008 with rates remaining around 50–55% until the latest report for 2016. Results from individual hospitals more or less reflected the national trend with the exception of UKMMC in 2010–2011 which showed a resistance rate of 79.6% (Biglari et al., 2017), HSNZ in 2011 with a rate of 66.1% (Lean et al., 2014) and HSA in 2011–2012 with a rate of 84.1% (Dhanoa et al., 2015). ICU isolates from UMMC (2006–2009) showed highest ciprofloxacin resistance rates at 99.4% (Kong et al., 2011).

NSAR reported Acinetobacter spp. resistance rates for ampicillin and piperacillin from 2007 to 2014. The Malaysian Acinetobacter isolates displayed very high resistance rates for ampicillin, which averaged at 89.2% whereas piperacillin showed a lower average resistance rate of 55.6% within the 7-year surveillance period.

The national resistance rate of Acinetobacter spp. toward the combination of piperacillin/tazobactam was relatively low (at 19.2%) in 2003 but this gradually increased to 55.8% by 2008 (Figure 5). NSAR data showed that from 2008 to 2016, the national resistance rates for piperacillin/tazobactam remained within the 55–60% range. However, reports of strains isolated from individual hospitals showed markedly higher resistance rates, as had been observed for other antimicrobials. Isolates from HSA in 2011 and 2012 (Dhanoa et al., 2015) showed resistance rates of about 90% whereas the resistance rates were lower at around 70% for UKMMC in 2010 and 2011 (Biglari et al., 2015), and HSNZ in 2011 (Lean et al., 2014) (Figure 5).

NSAR data for the combination of ticarcillin/clavulanate was available from 2007–2014 and the national Acinetobacter spp. resistance rates remained around the 40% level with the exception of 2010 when it spiked to 57.6% before decreasing to 47.8% the following year (Figure 5). The national resistance rate for ampicillin/sulbactam was around 40% from 2005 to 2009, thereafter increasing to between 50 and 60% from 2010 to 2016 (Figure 5). Reported resistance rates for the ampicillin/sulbactam combination from individual hospitals were higher, at 84.1% in the HSA A. baumannii isolates obtained in 2011 and 2012 (Dhanoa et al., 2015), and about 70% for the UKMMC isolates between 2010 and 2011 (Biglari et al., 2015) and the HSNZ isolates in 2011 (Lean et al., 2014). Lower resistance rates were generally observed for the sulbactam/cefoperazone combination when compared to ampicillin/sulbactam. When NSAR first reported data for sulbactam/cefoperazone in 2005, the resistance rate was at 14% and remained around that level for 2007–2008. The national sulbactam/cefoperazone resistance rate increased considerably to 33.4% in 2009 and it remained between 40 and 45% from 2010 to 2016 with the notable exception of 2014 where the reported rate was at 25.7%. However, A. baumannii isolates from HSA (in 2011 and 2012) showed a much higher sulbactam/cefoperazone resistance rate of 94.1%, higher than the ampicillin/sulbactam resistance rate of 84.1% (Dhanoa et al., 2015).

There are very few reports on the prevalence of tetracycline resistance in Malaysian Acinetobacter isolates. Lean et al. (2014) reported that out of 54 A. baumannii isolates that were collected from various ward in HSNZ in Terengganu during 2011, 87% were resistant to tetracycline while 61.1% were resistant to doxycycline. Similar high resistance rates for tetracycline were reported (79.1%) for a collection of 43 MDR A. baumannii isolates that were obtained from various hospitals mainly around the town of Ipoh, Malaysia although the year of their collection and the identity of the hospitals were not stated (Kor et al., 2014).

Tigecycline is a semisynthetic antibiotic belonging to the tetracycline-derived glycylcycline family and along with the lipopeptides or polymyxins (i.e., polymyxin B and colistin, or polymyxin E), tigecycline is considered one of the ‘last resort’ drugs for the treatment of Acinetobacter infections (Lim et al., 2011; Doi et al., 2015; Li et al., 2015; Pogue et al., 2015). However, it should be noted that guidelines such as the latest Infectious Diseases Society of America (IDSA) and the American Thoracic Society (ATS) for the management of adults with hospital-acquired pneumonia and ventilator-associated pneumonia (HAP/VAP) strongly recommends against the use of tigecycline in Acinetobacter infections (Kalil et al., 2016). Latest systematic reviews and meta-analyses also disfavor the use of a tigecycline-based regimen for the treatment of MDR A. baumannii infections, despite its lower nephrotoxicity compared with colistin (Ni et al., 2016; Kengkla et al., 2017). NSAR only reported tigecycline resistance rates for A. baumannii blood isolates from 2013–2016 with fairly constant rates of 15–18% for the 4 year period. An earlier study from the UMMC with isolates obtained from 2008–2009 indicated a 5% intermediate susceptibility to tigecycline for their clinical isolates but a much higher percentage (60%) of intermediate susceptibility for hospital environmental isolates (Dhabaan et al., 2012), which is surprising and a cause for concern. On the other hand, Kor et al. (2014) had reported a 58.1% tigecycline resistance rate on their collection of 43 MDR A. baumannii from various hospitals in Ipoh but their susceptibility testing for tigecycline was performed using the Kirby-Bauer disk diffusion assay for which no standard breakpoints were available. The 2008–2009 UMMC isolates were assessed for tigecycline susceptibility using both E-test and broth microdilution, and the MIC breakpoints from the United States Food and Drug Administration (FDA) were used for their interpretation of tigecycline susceptibility (Dhabaan et al., 2012), a move which was recently supported (Nicolau et al., 2015) in the absence of any CLSI guidelines for tigecycline until now (CLSI, 2017). Broth microdilution is recommended for determining tigecycline MIC values as a report had shown that tigecycline MICs varied greatly according to the in vitro testing methods used with Etest giving significantly elevated MICs and were thus, deemed inaccurate (Marchaim et al., 2014).

NSAR only reported A. baumannii resistance rates for colistin from 2015 onward where rates were low at 0.8% in 2015 and all isolates were susceptible in 2016. Data for the other polymyxin, polymyxin B, was only reported for blood isolates of A. baumannii from 2013–2016 with a resistance rate of 1.4% in 2013 and all isolates susceptible in 2014–2016. In stark contrast, Lean et al. (2014) had reported an alarmingly high resistance rate of 25.9% for polymyxin B in HSNZ. So far, this is the only peer-reviewed, published report of polymyxin resistant A. baumannii from Malaysia. The UMMC study on A. baumannii isolates obtained from 2008 and 2009 did not detect any polymyxin resistance (Dhabaan et al., 2012), as were isolates obtained from the UMMC ICU earlier (between 2006 and 2009) (Kong et al., 2011). Likewise, no polymyxin-resistant isolates were found in the 2011–2012 HSA study (Dhanoa et al., 2015) and the 2010– 2011 UKMMC study (Biglari et al., 2013).

Carbapenem resistance in Acinetobacter spp. is now increasingly reported worldwide and is usually mediated by enzymatic inactivation (via carbapenemases), active efflux of drugs and target site modification (i.e., altered penicillin-binding proteins) (Zarrilli et al., 2009). More than 210 β-lactamases belonging to 16 families have been identified in Acinetobacter spp. (Zhao and Hu, 2012) with class D β-lactamases being the most widespread carbapenemase in A. baumannii (Zarrilli et al., 2009; Bush, 2013). Class B metallo-β-lactamases (MBL; IMP-, VIM-, SIM- and NDM-types) have also been sporadically reported worldwide in A. baumannii, being able to hydrolyze carbapenems and other β-lactams, except aztreonam, and resistant to clinically available β-lactamase inhibitors (Zhao and Hu, 2012). Several insertion sequence (IS) elements such as ISAba1, ISAba2, ISAba3 and IS18, have been found to increase the expression of class D β-lactamase genes (including bla_OXA-23-like and bla_OXA-58-like genes) when they are inserted immediately upstream due to the presence of an outward-directing promoter at the ends of these IS elements (Zarrilli et al., 2009; Hsu et al., 2017). The A. baumannnii chromosome also encodes an intrinsic bla_OXA-51-like gene that is weakly expressed but does not confer resistance to carbapenems. However, it has been demonstrated that insertion of an ISAba1 element upstream of the gene conferred carbapenem resistance (Turton et al., 2006).

There are very few papers that have investigated the possible carbapenem resistance mechanisms in Acinetobacter spp. isolates from Malaysia. The bla_OXA-23 gene appeared to be the predominant acquired carbapenemase in the Malaysian A. baumannii isolates, which is not surprising as bla_OXA-23 is the most common cause of carbapenem resistance in A. baumannii, and the most widely spread acquired OXA carbapenemase worldwide (Kamolvit et al., 2015). The prevalence of the bla_OXA-23 gene was 75.9% in the 2011 A. baumannii HSNZ isolates (Lean et al., 2014) and 82% in the 2010–2011 UKMMC isolates (Biglari et al., 2015, 2017). In an earlier study, nearly 95% of carbapenem-resistant Acinetobacter spp. isolated in 2003–2004 from UMMC, were positive for bla_OXA-23 (Wong et al., 2009). However, almost half of the UKMMC isolates that contained the ISAba1-bla_OXA-51-like configuration were susceptible to carbapenems, leading the authors to conclude that ISAba1 may not upregulate the expression of the intrinsic bla_OXA-51-like gene and mediate carbapenem resistance (Biglari et al., 2015), as had been previously proposed (Turton et al., 2006). No bla_OXA-24-like and bla_OXA-58-like genes were detected so far in the Malaysian A. baumannii isolates (Biglari et al., 2015; Lean et al., 2014) although these class D β-lactamases have been found elsewhere, particularly in European isolates (D’Andrea et al., 2009; Merino et al., 2010; Novovic et al., 2015; Chatterjee et al., 2016). Among the Class B MBLs, only bla_IMP has been reported albeit only in 9.9% of the UKMMC A. baumannii isolates (Biglari et al., 2015) and 5.1% in the carbapenem-resistant 2003–2004 UMMC Acinetobacter spp. isolates (Wong et al., 2009), whereas neither bla_IMP nor bla_{V IM} was found in the HSNZ A. baumannii isolates from 2011 (Lean et al., 2014). Southern hybridization localized the bla_IMP-4 gene in an A. calcoaceticus isolate from UMMC to a class 1 integron on an approximately 35 kb plasmid (Wong et al., 2009). Interestingly, genome sequencing of an A. pittii isolated in 2014 from a hospital in the state of Perak (in Peninsular Malaysia) led to the discovery of bla_NDM-1 and bla_OXA-58 co-residing in the isolate (Ang et al., 2016). The bla_NDM-1 gene was found within a 10,038 bp composite transposon which resided on a 140 kb megaplasmid whereas the bla_OXA-58 gene was located on a 35 kb plasmid. Metallo-β-lactamase production in this A. pittii strain was validated by testing with the Etest MBL kit from BioMériux (Ang et al., 2016).

Acinetobacter spp. are known to encode Acinetobacter-specific AmpC cephalosporinases in the chromosome, designated ADCs (Hujer et al., 2005). More than 45 variants of ADCs (ADC-1 to ADC-56) have been categorized for the genus Acinetobacter with many more that remain uncategorized (Zhao and Hu, 2012). In cephalosporin-resistant A. baumannii isolates from UKMMC, the bla_ADC gene was present in 93.7% of the isolates and in most of these bla_ADC-positive isolates, ISAba1 was detected upstream of the bla_ADC gene (Biglari et al., 2015). ADCs are normally expressed at low levels and are not inducible (Hujer et al., 2005) but the insertion of ISAba1 upstream often leads to the overexpression of these cephalosporinases (Héritier et al., 2006). The specific ADC type was, however, not determined for the UKMMC isolates. The bla_ADC sequence from 3 cephalosporin-resistant A. baumannii from HSNZ isolated in 2011 (i.e., AC12, AC29 and AC30) were found to be a hitherto uncategorized ADC (with R80S and G246S mutations in reference to ADC-7) (Lean et al., 2015, 2016). However, these bla_ADC genes were characterized as belonging to the ampC allele 20 in a recent paper reporting on the re-curation of the A. baumannii-encoded ampC genes in a new database hosted at http://pubmlst.org/abaumannii (Karah et al., 2017). These bla_ADC genes from A. baumannii AC12, AC29 and AC30 were cloned into a pET30a expression vector and expressed in Escherichia coli BL21, leading to the recombinant E. coli strains displaying resistance to ceftazidime, cefepime, aztreonam and even imipenem (Lean et al., 2016). This suggests that the ADC from these isolates were indeed extended-spectrum Acinetobacter-derived AmpC (ESAC) as ADCs typically hydrolyze penicillins, narrow- and extended-spectrum cephalosporins but not carbapenems and zwitterionic cephalosporins such as cefepime (Rodríguez-Martínez et al., 2010; Lean et al., 2016).

The main mechanisms of fluroquinolone resistance are mutations that alter the target sites DNA gyrase (encoded by gyrA and gyrB) and DNA topoisomerase IV (encoded by parC and parE) (Jacoby, 2005). Ciprofloxacin-resistant A. baumannii isolates from UKMMC and A. baumannii AC12, AC29 and AC30 from HSNZ all displayed the characteristic serine-to-leucine substitution at position 83 for GyrA and position 80 for ParC (Lean et al., 2015, 2016; Biglari et al., 2017), mutations which have been implicated in fluoroquinolone resistance in Acinetobacter (Wisplinghoff et al., 2000; Fournier et al., 2006).

Resistance to polymyxins (polymxin B and colistin) in A. baumannii is mediated by multiple factors but is mainly due to modification of the LPS moieties that form the outer membrane layer of the cell (Olaitan et al., 2014; Jeannot et al., 2017; Poirel et al., 2017). In some polymyxin-resistant A. baumannii, phosphoethanolamine is enzymatically added to the lipid A of LPS (Arroyo et al., 2011) whereas in other resistant isolates, the LPS part of the outer membrane is completely absent due to mutations in the genes involved in LPS biosynthesis (Moffatt et al., 2010, 2011; Henry et al., 2012). These LPS alterations decrease the net negative charge, preventing the binding of the cationic polymyxin molecules to the bacterial surface (Jeannot et al., 2017; Poirel et al., 2017). PmrAB is a two-component regulatory system that regulates the expression of the genes involved in LPS modification; some mutations in pmrAB resulted in polymyxin resistance due to constitutive upregulation of the LPS modification pathway (Arroyo et al., 2011; Park et al., 2011; Lim et al., 2015; Dahdouh et al., 2017). Investigations into the polymyxin-resistant A. baumannii isolates from HSNZ in 2011 indicated a P102H mutation in the pmrA gene in all resistant isolates and several point mutations in the lpxC, lpxD and lpsB genes involved in LPS biosynthesis (Lean et al., 2014). Further experimental studies on two of these polymyxin-resistant isolates, A. baumannii AC12 and AC30, indicated upregulation of the pmrB gene as well as possible impairment (but not total loss) of the LPS (Lean et al., 2016). These mutations are intrinsic, and not transmissible, and are likely the result of selective pressure (Jeannot et al., 2017; Poirel et al., 2017). However, the recent discovery of the transmissible polymyxin-resistant genes, mcr-1, mcr-1.2, and mcr-2 (which encode phosphoethanolamine transferases) in Enterobacteriaceae (Liu et al., 2015; Giamarellou, 2016) raised the alarming possibility of its spread to Acinetobacter spp. and other bacteria. Although no reports of mcr-positive Acinetobacter spp. have emerged until now, it is likely just a matter of time as the mcr genes are carried on transmissible plasmids (Malhotra-Kumar et al., 2016; Jeannot et al., 2017). A recent report highlighted this when it was shown that laboratory transformation of an mcr-1-encoded recombinant plasmid into several strains of A. baumannii led to the development of colistin resistance in these strains (Liu et al., 2017).