ABSTRACT
Enterobacteriaceae producing the Ambler class D OXA-48 carbapenemase, combined with additional resistance mechanisms, such as permeability defects or cocarriage of class A, B, or C β-lactamases, can become highly resistant to most β-lactams currently in use, including carbapenems. A total of 45,872 Enterobacteriaceae clinical isolates collected in 39 countries as part of the International Network for Optimal Resistance Monitoring (INFORM) global surveillance study in 2012 to 2015 were tested for susceptibility to β-lactams and comparator agents using the Clinical and Laboratory Standards Institute broth microdilution methodology and screened for the presence of β-lactamases. The blaOXA-48 and blaOXA-48-like genes were detected in 333 isolates across 14 species of Enterobacteriaceae collected in 20 countries across the globe. Few agents tested were effective in vitro against the overall collection of OXA-48-producers (n = 265), with tigecycline (MIC90, 2 µg/ml; 92.5% susceptible), ceftazidime-avibactam (MIC90, 4 µg/ml; 92.5% susceptible), and aztreonam-avibactam (MIC90, 0.5 µg/ml; 99.6% of isolates with MIC ≤8 µg/ml) demonstrating the greatest activity. Similarly, colistin (MIC90, 1 µg/ml; 94.2% susceptible), tigecycline (MIC90, 2 µg/ml; 92.6% susceptible), ceftazidime-avibactam (MIC90, >128 µg/ml; 89.7% susceptible), and aztreonam-avibactam (MIC90, 4 µg/ml; 100% of isolates with MIC ≤8 µg/ml) were most active against OXA-48-like-positive isolates (n = 68). The in vitro activity of ceftazidime-avibactam was improved against the subset of metallo-β-lactamase (MBL)-negative, OXA-48- and OXA-48-like-positive isolates (99.2% and 100% susceptible, respectively). The data reported here support the continued investigation of ceftazidime-avibactam and aztreonam-avibactam for the treatment of infections caused by carbapenem-resistant Enterobacteriaceae carrying OXA-48 and OXA-48-like β-lactamases in combination with serine- or metallo-β-lactamases.
INTRODUCTION
Carbapenem-resistant Enterobacteriaceae cause a high health care burden in hospital and community settings and are globally recognized as threats to the continued utility and efficacy of our existing antimicrobial armamentarium (1, 2). Resistance to carbapenems can be mediated by the production of carbapenemases of Ambler class A (KPC, GES, NMC/IMI, and SME), class B (metallo-β-lactamases [MBL], NDM, VIM, IMP, SPM, and GIM), or class D (OXA-23, OXA-24, OXA-48, OXA-58, and related enzymes), or the hyperproduction of class C cephalosporinases (AmpC) or class A extended-spectrum β-lactamases (ESBLs) coupled with defects in membrane permeability; each of these resistance mechanisms poses a different challenge to treatment (3–8).
OXA-48 hydrolyzes oxacillin, penicillins, and cephalothin, displays poor activity against extended-spectrum cephalosporins (e.g., cefotaxime, ceftazidime, cefepime, and cefpirome) and monobactams (aztreonam), and is poorly inhibited by classic β-lactamase inhibitors, such as clavulanic acid and tazobactam (7, 9). Although it demonstrates greater catalytic activity (kcat/Km) for imipenem than other OXA-type β-lactamases (7), OXA-48 is a weak carbapenemase that often does not confer resistance to carbapenems unless combined with permeability defects (9–13), making it difficult to detect in clinical settings (5, 8). blaOXA-48 is located within a composite Tn1999-type transposon (9) inserted in a self-conjugative, epidemic, broad-host-range IncL/M-type plasmid (pOXA-48a) (11, 14) with high transfer efficiency among Enterobacteriaceae species (15), and it has also been found integrated into the chromosome of Escherichia coli isolates (16, 17). pOXA-48a carries no additional resistance markers; however, as many as 80% of OXA-48-positive isolates are reported to coproduce ESBLs, either through acquisition of additional resistance plasmids or via cocarriage of CTX-M-15 in the mosaic Tn1999.4 transposon (18, 19). In addition, coproduction of OXA-48 and NDM- and VIM-type MBLs has been reported (20–22). OXA-48-producing Enterobacteriaceae were first identified in Turkey and later in North African countries (Morocco, Algeria, Tunisia, Egypt, and Libya), and these areas are considered the main reservoirs of this resistance mechanism (11, 23, 24). However, OXA-48-positive isolates soon spread to countries in the Mediterranean area and have now disseminated throughout much of Europe (5, 6, 25).
Closely related OXA-48-like β-lactamases have also been described. The majority of these display susceptibility profiles similar to that of OXA-48; however, a small number of variants with a four-amino-acid deletion near the enzyme active site (e.g., OXA-163) are able to hydrolyze extended-spectrum cephalosporins and possess attenuated carbapenemase activity (5, 6, 13, 26–30). The genetic contexts and plasmid replicons associated with OXA-48-like β-lactamases differ from those of OXA-48, as does the geographic distribution of these enzymes, though some, notably OXA-181, appear to be disseminating rapidly (5, 6, 26, 27, 29–33).
Avibactam is a non-β-lactam–β-lactamase inhibitor that reacts with the active-site serine of class A, class C, and some class D β-lactamases. Avibactam is not a traditional suicide inhibitor but rather forms a reversible acyl enzyme that is stable to hydrolysis by most serine β-lactamases (34–37). The spectrum of inhibition of avibactam includes class A penicillinases, ESBLs, including CTX-M-type, inhibitor-resistant TEM variants, and carbapenemases, including KPC. It also inhibits both plasmid and chromosomally mediated class C enzymes, as well as some class D OXA-type enzymes, including OXA-48 (34, 35, 37–40). It does not inhibit class B metallo-β-lactamases that have a catalytic zinc atom in the active site (41). When paired with an extended-spectrum cephalosporin, such as ceftazidime, or aztreonam, avibactam protects its partner β-lactam from inactivation by members of these β-lactamase classes, thereby restoring their efficacy. Ceftazidime-avibactam was recently approved and is highly effective against carbapenem-resistant organisms that do not produce MBLs (42–44). Aztreonam-avibactam is being developed for use against infections caused by MBL-producing Enterobacteriaceae. Although aztreonam is stable to MBL hydrolysis (4), MBL-positive isolates often cocarry aztreonam-hydrolyzing serine β-lactamases. By combining aztreonam with avibactam, the hydrolysis of aztreonam is prevented, thereby restoring its activity (45).
This study reports the distribution of OXA-48- and OXA-48-like-positive Enterobacteriaceae isolates collected from 2012 to 2015 as part of the International Network for Optimal Resistance Monitoring (INFORM) global surveillance program, and it evaluates the in vitro activities of ceftazidime-avibactam, aztreonam-avibactam, and comparator agents against these isolates.
RESULTS
A total of 45,872 Enterobacteriaceae isolates were collected in 39 countries participating in the INFORM global surveillance program in 2012 to 2015 (inclusive). In this collection, 11,408 isolates met the molecular screening criteria, and of these, 265 (0.6% of all Enterobacteriaceae) isolates collected from hospitals in 16 countries carried blaOXA-48 (Fig. 1 and Table 1). Klebsiella pneumoniae was the most commonly identified OXA-48-positive species (n = 212 [80.0%]), followed by Escherichia coli (n = 14 [5.3%]) and Enterobacter cloacae (n = 13 [4.9%]). The remaining ∼10% of OXA-48-positive isolates were scattered in prevalence among 11 additional species (6 Klebsiella oxytoca, 5 Citrobacter freundii, 3 Raoultella planticola, 2 each Enterobacter kobei, Morganella morganii, Raoultella ornithinolytica, and Serratia marcescens, and 1 each Citrobacter braakii, Citrobacter koseri, Enterobacter asburiae, and Proteus mirabilis). Isolates carrying blaOXA-48 were found in all four regions surveyed (Europe, Asia-Pacific, the Middle East-Africa region, and Latin America; isolates from the United States were not included in the study). A total of 93.6% of the isolates (n = 248) were collected in Europe, predominantly in Turkey (6.3% of the isolates collected from this country), Russia (3.1% of the collected isolates), Romania (5.2% of the collected isolates), and Spain (1.2% of the collected isolates); in contrast, the percentage of OXA-48 among all Enterobacteriaceae isolates was <1% for most other countries surveyed (Table 1; see also Table S1 in the supplemental material). Smaller numbers of isolates were collected in the Middle East-Africa region (n = 14 of 265 [5.3%]), Asia-Pacific (n = 2 of 265 [0.8%]), and Latin America (n = 1 of 265 [0.4%]) (Table 1). It should be noted that isolates collected from patients in China and Hong Kong, none of which carried blaOXA-48, were only obtained in 2012 to 2013 and 2012 to 2014, respectively, due to export restrictions of bacterial pathogens imposed in later years. In all countries in which at least two OXA-48-positive isolates were identified, isolates were collected at more than one participating institution, except for Italy. OXA-48-positive pathogens were collected from all infection sources (1.2% of bloodstream isolates, 0.9% of respiratory tract isolates, 0.5% of urinary tract isolates, 0.5% of skin and soft tissue isolates, and 0.3% of intra-abdominal isolates), and 87.5% of isolates were collected from patients more than 48 h after hospital admission, suggesting nosocomial acquisition (Table 2). The corresponding data for 68 additional isolates of K. pneumoniae, E. coli, and E. cloacae carrying OXA-48-like β-lactamases (OXA-181, n = 25; OXA-163, n = 13; OXA-232, n = 12; OXA-244, n = 12; OXA-162, n = 3; OXA-370, n = 2; and OXA-439, n = 1) also found during the course of this study are described in Tables S2 and S3.
Distribution of OXA-48-positive Enterobacteriaceae collected in 2012 to 2015. Countries in which OXA-48-positive isolates were detected were Belgium, Brazil, Denmark, France, Germany, Hungary, Israel, Italy, Kuwait, Nigeria, Romania, Russia, Spain, Thailand, Turkey, and the United Kingdom. Participating countries in which OXA-48-positive isolates were not detected were Argentina, Australia, Austria, Chile, China, Hong Kong, Colombia, the Czech Republic, Greece, Ireland, Japan, Kenya, Malaysia, Mexico, Netherlands, Philippines, Poland, Portugal, South Africa, South Korea, Sweden, Taiwan, and Venezuela. Countries in which OXA-48-like-positive isolates were detected were Argentina (OXA-163 and OXA-439), Belgium (OXA-181 and OXA-232), Brazil (OXA-370), Greece (OXA-163), Italy (OXA-162), Kenya (OXA-181), Kuwait (OXA-232), Mexico (OXA-163 and OXA-232), Romania (OXA-162 and OXA-163), Russia (OXA-244), Thailand (OXA-181 and OXA-232), and Turkey (OXA-162 and OXA-181).
Geographic and species distribution of 265 OXA-48 positive isolates collected as part of the INFORM global surveillance program in 2012 to 2015a
Body source distribution of 265 OXA-48-positive Enterobacteriaceae isolates collected in 2012 to 2015a
Thirty isolates (11.5%) carried OXA-48 alone or with an original-spectrum β-lactamase (OSBL; includes TEM-1, TEM-2, SHV-1, and SHV-11) that does not impact the susceptibility of either ceftazidime or aztreonam alone. The majority of the isolates expressing OXA-48 also carried additional β-lactamases, including plasmid-encoded ESBLs, AmpC, and MBLs, as well as presumed intrinsic chromosomally encoded AmpC enzymes (Fig. 2 and Tables 3 and 4). A total of 88.7% of OXA-48-positive isolates (n = 235) carried additional β-lactamases capable of hydrolyzing expanded-spectrum cephalosporins or aztreonam (Fig. 2 and Table 3) (46, 47). The most common ESBL encountered was CTX-M-15 (158 isolates), followed by CTX-M-14 (24 isolates) and CTX-M-3 (8 isolates). Twenty-seven isolates of Citrobacter spp., Enterobacter spp., M. morganii, and S. marcescens presumed to harbor intrinsic chromosomally encoded AmpC β-lactamases were OXA-48 positive, whereas acquired DHA-type and CMY-type enzymes were found in 7 isolates of E. coli, K. pneumoniae, and E. cloacae. Notably, 23 isolates carried OXA-48 and an MBL. Of these, 11 K. pneumoniae isolates collected at two medical centers in Romania carried OXA-48, NDM-1, and CTX-M-15, whereas four isolates (3 K. pneumoniae and 1 E. cloacae) collected at one center in Romania carried OXA-48 and NDM-1. Six additional isolates (5 C. freundii and 1 E. cloacae) collected at one medical center in Turkey carried OXA-48 and VIM-31, and two E. cloacae isolates collected at one center in Kuwait carried OXA-48, VIM-4, and CMY-4; one of the two E. cloacae isolates also carried an SHV-12 ESBL (Table 3). No isolates cocarrying OXA-48 and the KPC carbapenemase were identified during this study. The β-lactamase carriage of isolates positive for OXA-48-like variants is described in Table S4; of these, 10 isolates carried additional carbapenemases (NDM-1 and OXA-232, n = 7; KPC-2 and OXA-163, n = 2; KPC-2 and OXA-370, n = 1).
Cocarriage of OXA-48 and other β-lactamases in 265 OXA-48-positive Enterobacteriaceae collected in 2012 to 2015. OSBL, original-spectrum β-lactamase unable to hydrolyze extended-spectrum cephalosporins (includes TEM-1, TEM-2, SHV-1, and SHV-11); ESBL, extended-spectrum β-lactamase; AmpC, chromosomal- or plasmid-encoded Ambler class C cephalosporinase; MBL, metallo-β-lactamase. β-Lactamases detected in isolates cocarrying OXA-48 and ESBLs (n) were CTX-M-15 (135), CTX-M-14 (19), CTX-M-3 (8), CTX-M-55 (7), CTX-M-28 (4), CTX-M-24 (2), CTX-M-27 (2), CTX-M-type (2), CTX-M-14 and CTX-M-27 (1), CTX-M-15 and CTX-M-14 (2), and CTX-M-15 and SHV-12 (1). Six isolates were assumed to carry the intrinsic ESBL common to K. oxytoca with CTX-M-14 and CTX-M-27 (1) or without (5) additional acquired ESBLs. β-Lactamases detected in isolates cocarrying OXA-48 and AmpC β-lactamases (n) were intrinsic AmpC β-lactamases common to Citrobacter spp. (2), Enterobacter spp. (7), M. morganii (1), and S. marcescens (1). β-Lactamases detected in isolates cocarrying OXA-48 and AmpC β-lactamases and ESBLs (n) were DHA-1 and CTX-M-15 (3), CMY-6 and CTX-M-15 (1), CMY-42 and CTX-M-15 (1), the intrinsic AmpC common to Enterobacter spp. and CTX-M-15 (3) or CTX-M-9 (1) or SHV-12 (1), the intrinsic AmpC common to M. morganii and CTX-M-15 and CTX-M-14 (1), and the intrinsic AmpC common to S. marcescens and CTX-M-22 (1). β-Lactamases detected in isolates cocarrying OXA-48 and MBLs with or without additional β-lactamases (n) were NDM-1 (3), NDM-1 and CTX-M-15 (11), VIM-31 and the intrinsic AmpC common to Citrobacter spp. (5), NDM-1 (1) or VIM-31 (1) or VIM-4 and CMY-4 (1) or VIM-4, CMY-4, and SHV-12 (1), and the intrinsic AmpC common to Enterobacter spp.
Cocarriage of OXA-48 and other β-lactamases in Enterobacteriaceae isolates collected in 2012 to 2015a
The in vitro activities of ceftazidime-avibactam, aztreonam-avibactam, and comparator agents were determined against the overall collection of Enterobacteriaceae isolates and subsets of OXA-48-positive isolates cocarrying Ambler class A, B, and C β-lactamases (Table 4). The activities of most tested β-lactams, including ceftazidime, aztreonam, cefepime, meropenem, imipenem, and piperacillin-tazobactam, were greatly reduced against the overall collection of OXA-48-positive isolates and against most subsets of isolates cocarrying additional β-lactamases. The percentages of susceptibility to meropenem ranged from 0 to 36.4% and were consistently higher than susceptibilities to imipenem (range, 0 to 11.1% susceptible) among all OXA-48-positive isolates and among subsets of isolates that did not carry MBLs, in agreement with the known hydrolytic properties of OXA-48. The percentages of susceptibility to ceftazidime, aztreonam, and cefepime were <22% for all OXA-48-positive isolates (reflective of the high proportion of isolates cocarrying ESBLs) as well as for ESBL-containing subsets. Percentages of susceptibility to these expanded-spectrum cephalosporins and aztreonam were higher for subsets of isolates carrying OXA-48 alone or in combination with OSBLs that are not expected to have significant activity against these agents or AmpC β-lactamases (73.9 to 100% susceptible). A combination of avibactam with ceftazidime or aztreonam increased the activities of these two β-lactams against all OXA-48-positive isolates by at least 64- and 512-fold, respectively. The MIC90 values for ceftazidime-avibactam and ceftazidime were 4 and >128 µg/ml, respectively, against all OXA-48-positive isolates and 1 and >128 µg/ml, respectively, against the set of OXA-48-positive, MBL-negative isolates (n = 242), whereas the MIC90 values for aztreonam-avibactam and aztreonam were 0.5 and >128 µg/ml, respectively, against all and MBL-negative, OXA-48-positive isolates. Notably, the MIC90 of aztreonam-avibactam against OXA-48-positive isolates coproducing MBLs (n = 23) was 0.5 µg/ml and, as expected, ceftazidime-avibactam showed greatly reduced activity against these isolates (MIC90, >128 µg/ml; 21.7% susceptible). Both aztreonam-avibactam and ceftazidime-avibactam retained good activity against subsets of isolates expressing multiple classes of β-lactamases, including those that produced ESBLs and AmpC β-lactamases in addition to OXA-48. For these strains, the MIC90 values for aztreonam-avibactam and ceftazidime-avibactam were 8 and 4 µg/ml, respectively. The activities of agents from other drug classes (amikacin, tigecycline, and colistin) against OXA-48-positive, MBL-negative isolates were comparable to or reduced compared to the activity of ceftazidime-avibactam (76.1 to 92.6% susceptible) and exceeded it against MBL-positive isolates (34.8 to 100% susceptible). Amikacin, tigecycline, and colistin were less active than aztreonam-avibactam, based on MIC90 values, against all OXA-48-producing subsets except for isolates cocarrying AmpC and ESBL β-lactamases (MIC90, 2 µg/ml and 8 µg/ml for tigecycline and aztreonam-avibactam, respectively) (Table 4).
In vitro activity of antimicrobial agents tested against OXA-48-producing isolates collected in 2012 to 2015
Only two of the OXA-48-positive, MBL-negative isolates were not susceptible to ceftazidime-avibactam (MIC, >8 µg/ml). One carbapenem-nonsusceptible K. pneumoniae isolate carrying OXA-48 and an SHV-OSBL, collected in Hungary, had a ceftazidime-avibactam MIC of 16 µg/ml and aztreonam-avibactam MIC of 0.5 µg/ml and was susceptible to amikacin and tigecycline. The second isolate, a carbapenem-nonsusceptible S. marcescens isolate collected in Russia, carried OXA-48, CTX-M-22, and an SHV-OSBL, had a ceftazidime-avibactam MIC of 64 µg/ml and aztreonam-avibactam MIC of 16 µg/ml, and was resistant to amikacin and tigecycline. The susceptibilities of these two isolates to colistin were not determined. All except one of the OXA-48-positive isolates had an MIC of ≤8 µg/ml to aztreonam-avibactam. This isolate, an S. marcescens strain cocarrying OXA-48, CTX-M-22, and an SHV-OSBL, was also resistant to ceftazidime-avibactam and other classes of agents, as described above.
The activities of ceftazidime-avibactam and aztreonam-avibactam were also evaluated against the OXA-48-like-producing isolates collected during the course of the study (Table 5). Ceftazidime-avibactam MIC90 values against all and MBL-negative isolates were >128 µg/ml (with 89.7% of the isolates testing as susceptible) and 4 µg/ml (with 100% of the isolates testing as susceptible), respectively, and ranged from 1 µg/ml to 4 µg/ml against subsets of MBL-negative isolates carrying individual OXA sequence variants. In comparison, aztreonam-avibactam MIC90s were 4 µg/ml against all and MBL-negative OXA-48-like-positive isolates and ranged from 0.25 µg/ml to 8 µg/ml against individual OXA-48-like variants. All MBL-negative isolates had MIC values of ≤8 µg/ml for both agents.
In vitro activity of antimicrobial agents tested against isolates producing OXA-48-like β-lactamases collected in 2012 to 2015
Higher MIC90 values for both ceftazidime-avibactam and aztreonam-avibactam were observed against 25 OXA-181-producing isolates (4 µg/ml and 8 µg/ml, respectively). This subset of isolates was composed of seven K. pneumoniae isolates with ceftazidime-avibactam MICs of ≤2 µg/ml and aztreonam-avibactam MICs of ≤0.5 µg/ml collected in three countries and 18 E. coli isolates with ceftazidime-avibactam MICs ranging from 1 to 8 µg/ml and aztreonam-avibactam MICs from 4 to 8 µg/ml, 17 of which were collected from one hospital in Turkey over an 11-week period in 2015 (Table S5). Sequencing of the gene encoding penicillin binding protein 3 (pbp3) in a selection of these E. coli isolates revealed a duplication resulting in the insertion of the four-amino-acid sequence YRIN after residue 333 in PBP3 (Table S5). In light of this finding, the pbp3 gene was sequenced in five additional isolates, with four isolates having aztreonam-avibactam MICs of 4 to 16 µg/ml carrying OXA-48 (n = 3) or OXA-163 (n = 1) and one isolate with an MIC of 2 µg/ml carrying OXA-232 (Table S6). The PBP3 YRIN insertion was found in one OXA-48-positive E. coli isolate, which was collected from the same hospital in Turkey as the OXA-181-positive E. coli isolates described in Table S5. No insertions or major sequence alterations in pbp3 were found in the other four isolates screened (two K. pneumoniae, one E. cloacae, and one S. marcescens isolate; Tables S6 and S7).
DISCUSSION
In this report, the incidence, distribution, and antimicrobial susceptibilities of a large set of blaOXA-48-carrying clinical isolates of Enterobacteriaceae collected globally were determined during a recent 4-year period. Although the INFORM surveillance program was not structured to determine the prevalence of organisms carrying specific antimicrobial resistance mechanisms, the findings of this study are in general agreement with those reported by others. The majority (239/265 [90.2%]) of OXA-48-positive isolates were found to be K. pneumoniae, E. coli, and E. cloacae, with dissemination to the greatest number of additional species observed among isolates from Turkey, one of the main reservoirs of blaOXA-48 (5, 6). A total of 64.9% (172/265) of all OXA-48-positive isolates were collected in countries in which endemic (Turkey), interregional (Belgium, France, Romania, and Spain), or regional (Germany and Italy) spread of OXA-48 was recently reported (25, 48, 49). An additional 27.5% (73/265) of isolates were collected from multiple medical centers in Russia; though limited information regarding the epidemiology of carbapenemase-producing Enterobacteriaceae is available for this country, presumed interspecies and interpatient spread of blaOXA-48 was reported during a 2-year period in a hospital intensive care unit located in Moscow (50). Several outbreaks of organisms producing OXA-48 have been noted in European hospitals, most notably in Belgium and the Netherlands where, in general, resistance remains fairly low (51, 52). In this study, a small percentage (0.8%) of the isolates from Belgium harbored OXA-48. These results confirm those of a previous survey in Belgium that showed a consistent low-level incidence of OXA-48 (53). None of the isolates from the Netherlands were positive for OXA-48 in the present study, which correlates with the reported successful infection control measures taken to stop the outbreak in that country (54). Isolates carrying OXA-48-like β-lactamases were found mostly in countries where they had been previously reported (OXA-162, Turkey [55]; OXA-163, Argentina [26]; OXA-181, Belgium, Romania, and Thailand [6, 56, 57]; OXA-232, Mexico and Thailand [56, 58]; OXA-244, Russia [50]; and OXA-370, Brazil [31]). The detection of OXA-162, OXA-163, and OXA-232 in previously undocumented areas may reflect travel by infected individuals or dissemination by other routes.
It should be noted that most likely, this study underreported the number of OXA-48- and OXA-48-like-producing isolates collected in 2012 to 2015 for several reasons. OXA-48 displays higher catalytic activity for imipenem and ertapenem than for meropenem, and isolates that do not harbor additional permeability defects can test as susceptible to one or more, or even all, carbapenems (4, 5, 9–12, 19). Ertapenem, the carbapenem most susceptible to OXA-48 activity, was tested against isolates collected in 2012 and 2013 and removed from the panel in later years. Additionally, the β-lactamase screening criteria were altered in 2015 to exclude the characterization of isolates that tested as imipenem or doripenem nonsusceptible but meropenem susceptible. Although the majority of OXA-48- and OXA-48-like-producers cocarry an ESBL and isolates resistant to ceftazidime (MIC, >8 µg/ml) were also screened for β-lactamase genes, some meropenem-susceptible isolates that did not cocarry ceftazidime-hydrolyzing β-lactamases could also have been omitted from this analysis (5, 12, 19). Finally, due to its insensitivity to inhibition by clavulanic acid, OXA-48-positive isolates rarely test as phenotypically ESBL positive by combination testing. Another limitation is that sequence typing and/or plasmid typing to determine strain relatedness were not assessed.
Ceftazidime-avibactam and aztreonam-avibactam showed good activity against OXA-48- and OXA-48-like producing isolates. Comparable susceptibility results have been reported by others. Livermore et al. obtained MIC90 values of 1 µg/ml (range, 0.06 to 1 µg/ml) for ceftazidime-avibactam and 0.5 µg/ml (range, ≤0.03 to 0.5 µg/ml) for aztreonam-avibactam against 19 isolates of OXA-48-positive K. pneumoniae (41). Vasoo et al. also tested both agents and reported MIC90 values of 4 µg/ml (range, ≤0.06 to >512 µg/ml) for ceftazidime-avibactam and 1 µg/ml (range, ≤0.06 to 1 µg/ml) for aztreonam-avibactam against 14 isolates (12 K. pneumoniae complex, 1 E. coli, and 1 C. koseri) carrying OXA-48 (n = 8), OXA-181 (n = 4), and OXA-232 (n = 2) β-lactamases (59). Sader et al. obtained an aztreonam-avibactam MIC90 value of 0.25 µg/ml (range, ≤0.03 to 0.5 µg/ml) against 57 isolates carrying OXA-48-like β-lactamases (OXA-48, n = 55; OXA-244, n = 1; OXA-370, n = 1) (60), and Haidar et al. reported ceftazidime-avibactam MICs of 1 µg/ml against two OXA-48-positive K. pneumoniae isolates (61). Two MBL-negative OXA-48-positive isolates collected as part of this study had MICs of >8 µg/ml to aztreonam-avibactam and/or ceftazidime-avibactam; these isolates may harbor additional resistance mechanisms, such as amino acid substitutions in cocarried ESBLs or AmpC β-lactamases (62, 63) or changes in drug uptake that have been previously reported to impact the activity of β-lactam–avibactam combinations (64). Twenty-one isolates carrying OXA-48, OXA-181 (17 of 18 isolates possibly representing a clonal outbreak), and OXA-163 had aztreonam-avibactam MICs of 4 to 8 µg/ml. The reduced activity is unlikely to be associated with the OXA enzymes and more likely to be due to insertions in PBP3 which are known to cause reduced activity for aztreonam-avibactam (65). Of these 21 isolates, nine E. coli isolates carrying OXA-181 or OXA-48 were found to possess the four-amino-acid-long YRIN insertion in PBP3. The insertion of YRIN or YRIK after position 333, near the β-lactam binding pocket of E. coli PBP3, was reported to decrease susceptibility to ceftazidime by 4- to 8-fold and to aztreonam by 32-fold and resulted in MIC values for aztreonam-avibactam and ceftazidime-avibactam of 4 to 16 µg/ml and 2 to 8 µg/ml, respectively (65). The insertion of a different four-residue sequence (TIPY inserted after position 344) into the PBP3 of E. coli was proposed to account for MIC values of 8 µg/ml observed for ceftaroline-avibactam, ceftazidime-avibactam, and aztreonam-avibactam in a different study (66). No major disruptions in the sequence of pbp3 were identified in the K. pneumoniae, E. cloacae, and S. marcescens isolates sequenced as part of this study, which were assumed to harbor different resistance mechanisms.
It is important to know the local prevalences of OXA-48 and OXA-48-like β-lactamases among clinical isolates so that infections caused by isolates with a higher MIC value for a carbapenem, even if the MIC does not cross the threshold of “resistant,” should be considered for treatment options other than carbapenems. Therapeutic options available for treating infections caused by carbapenem-resistant Enterobacteriaceae are often limited to older agents, often with impaired safety profiles, such as aminoglycosides, tigecycline, fosfomycin, and polymyxins (8); in this light, recent reports of colistin-resistant OXA-48-producing isolates are very concerning (67–70). This study also showed an increased incidence of colistin resistance in OXA-48-producing isolates, with 21.3% of the isolates being nonsusceptible to colistin, compared to only 17% for all Enterobacteriaceae. This trend is in line with an earlier analysis of clinical isolates from the INFORM study (71). Both ceftazidime-avibactam and aztreonam-avibactam retained activity against all of the OXA-48 and OXA-48-like expressing isolates that contained only that enzyme or coproduced it along with other serine-based ESBLs or AmpC enzymes. Aztreonam-avibactam showed good activity against isolates that also harbored a metallo-β-lactamase. These results confirm that these β-lactam–β-lactamase inhibitor combinations may provide utility in their intended purpose of combating β-lactamase-mediated resistance, including that caused by OXA-48 and OXA-48-like enzymes, β-lactamases that have now become a global problem.
MATERIALS AND METHODS
Nonduplicate clinical isolates were collected from patients with intra-abdominal (n = 8,955), urinary tract (n = 13,166), lower respiratory tract (n = 10,009), skin and soft tissue (n = 11,370), bloodstream (n = 2,243; added as an acceptable specimen source in 2014), and other (n = 129) infections from 182 medical centers located in 39 countries in Europe (24,826 isolates collected by 97 laboratories in 19 countries), Asia-Pacific (9,149 isolates collected by 42 laboratories in 9 countries), the Middle East-Africa region (4,232 isolates collected by 17 laboratories in 5 countries), and Latin America (7,665 isolates collected by 26 laboratories in 6 countries) in 2012 to 2015. Investigators were asked to collect a predefined number of isolates of selected bacterial species, regardless of antimicrobial susceptibility; only one isolate was collected per patient. Basic demographic data were collected but were not linked to patient identity or specific antibiotic therapy. The total number of Enterobacteriaceae isolates collected in each country is listed in Table S1.
Isolates were sent to a central laboratory, where the species identification was confirmed by matrix-assisted laser desorption ionization–time of flight mass spectroscopy (MALDI Biotyper; Bruker Daltonics, Billerica, MA, USA). Susceptibility testing was performed by broth microdilution using custom frozen panels prepared in-house. Panel preparation and isolate and quality control testing followed Clinical and Laboratory Standards Institute (CLSI) guidelines (72, 73). Ceftazidime and aztreonam were tested in combination with a fixed concentration of 4 µg/ml avibactam. MICs were interpreted using current CLSI breakpoints for all drugs (73), except tigecycline, for which breakpoints set by the United States Food and Drug Administration (US FDA) (74) were used; colistin, for which current breakpoints set by the European Committee on Antimicrobial Susceptibility Testing (EUCAST) (75) were used; and aztreonam-avibactam, for which clinical breakpoints have not been assigned.
All Enterobacteriaceae isolates that tested as nonsusceptible to meropenem (MICs, >1 µg/ml; isolates collected in all years), doripenem or imipenem (MICs, >1 µg/ml; isolates collected in 2012 to 2014), or ertapenem (MICs, >0.5 µg/ml; isolates collected in 2012 to 2013), as well as E. coli, K. pneumoniae, K. oxytoca, and P. mirabilis isolates resistant to ceftazidime (MICs, >8 µg/ml) and/or that were positive for extended-spectrum β-lactamase (ESBL) activity by combination clavulanic acid testing (72), were screened for the presence of genes encoding OXA-48-like and other β-lactamases (KPC, NDM, IMP, VIM, SPM, GIM, TEM, SHV, CTX-M, VEB, PER, GES, ACC, ACT, CMY, DHA, FOX, MIR, and MOX) by multiplex PCR, followed by amplification of detected genes using flanking primers and Sanger sequencing of both DNA strands, as described previously (76). Multiplex primers OXA-48-F (5′-GCTTGATCGCCCTCGATT-3′) and OXA-48-R2 (5′-GATTTGCTSSGTRGCCGAAA-3′) (77) were employed to detect genes encoding OXA-48-like β-lactamases.
ACKNOWLEDGMENTS
We gratefully acknowledge the contributions of the study investigators, laboratory personnel, and all members of the AstraZeneca INFORM Surveillance program.
This study was sponsored by AstraZeneca Pharmaceuticals LP, which also included compensation fees for manuscript preparation. AstraZeneca’s rights to ceftazidime-avibactam and aztreonam-avibactam were acquired by Pfizer in December 2016. K.M.K., P.A.B., B.L.M.D.J., and G.G.S. wrote and edited the manuscript.
K.M.K. and D.F.S. are employees of IHMA. P.A.B. is the owner of Antimicrobial Development Specialists, LLC. B.L.M.D.J. and G.G.S. are employees of Pfizer. None of the IHMA authors has a personal financial interest in the sponsor of this paper (AstraZeneca Pharmaceuticals). P.A.B., B.L.M.D.J., and G.G.S. were employees of and shareholders in AstraZeneca at the time of the study.
All authors provided analysis input and have read and approved the final manuscript.
FOOTNOTES
- Received 26 March 2018.
- Returned for modification 6 June 2018.
- Accepted 7 September 2018.
- Accepted manuscript posted online 24 September 2018.
Supplemental material for this article may be found at https://doi.org/10.1128/AAC.00592-18.
- Copyright © 2018 American Society for Microbiology.
