Descriptive analysis of targeted carbapenemase genes and antibiotic susceptibility profiles among carbapenem-resistant Acinetobacter baumannii tested in the Antimicrobial Resistance Laboratory Network—United States, 2017–2020

ABSTRACT Acinetobacter baumannii is a Gram-negative bacillus that can cause severe and difficult-to-treat healthcare-associated infections. A. baumannii can harbor mobile genetic elements carrying genes that produce carbapenemase enzymes, further limiting therapeutic options for infections. In the United States, the Antimicrobial Resistance Laboratory Network (AR Lab Network) conducts sentinel surveillance of carbapenem-resistant Acinetobacter baumannii (CRAB). Participating clinical laboratories sent CRAB isolates to the AR Lab Network for characterization, including antimicrobial susceptibility testing and molecular detection of class A (Klebsiella pneumoniae carbapenemase), class B (Active-on-Imipenem, New Delhi metallo-β-lactamase, and Verona integron-encoded metallo-β-lactamase), and class D (Oxacillinase, blaOXA-23-like, blaOXA-24/40-like, blaOXA-48-like, and blaOXA-58-like) carbapenemase genes. During 2017‒2020, 6,026 CRAB isolates from 45 states were tested for targeted carbapenemase genes; 1% (64 of 5,481) of CRAB tested for targeted class A and class B genes were positive, but 83% (3,351 of 4,041) of CRAB tested for targeted class D genes were positive. The number of CRAB isolates carrying a class A or B gene increased from 2 of 312 (<1%) tested in 2017 to 26 of 1,708 (2%) tested in 2020. Eighty-three percent (2,355 of 2,846) of CRAB with at least one of the targeted carbapenemase genes and 54% (271 of 500) of CRAB without were categorized as extensively drug resistant; 95% (42 of 44) of isolates carrying more than one targeted gene had difficult-to-treat susceptibility profiles. CRAB isolates carrying targeted carbapenemase genes present an emerging public health threat in the United States, and their rapid detection is crucial to improving patient safety. IMPORTANCE The Centers for Disease Control and Prevention has classified CRAB as an urgent public health threat. In this paper, we used a collection of >6,000 contemporary clinical isolates to evaluate the phenotypic and genotypic properties of CRAB detected in the United States. We describe the frequency of specific carbapenemase genes detected, antimicrobial susceptibility profiles, and the distribution of CRAB isolates categorized as multidrug resistant, extensively drug-resistant, or difficult to treat. We further discuss the proportion of isolates showing susceptibility to Food and Drug Administration-approved agents. Of note, 84% of CRAB tested harbored at least one class A, B, or D carbapenemase genes targeted for detection and 83% of these carbapenemase gene-positive CRAB were categorized as extensively drug resistant. Fifty-four percent of CRAB isolates without any of these carbapenemase genes detected were still extensively drug-resistant, indicating that infections caused by CRAB are highly resistant and pose a significant risk to patient safety regardless of the presence of one of these carbapenemase genes.


IMPORTANCE
The Centers for Disease Control and Prevention has classified CRAB as an urgent public health threat.In this paper, we used a collection of >6,000 contemporary clinical isolates to evaluate the phenotypic and genotypic properties of CRAB detected in the United States.We describe the frequency of specific carbapenemase genes detected, antimicrobial susceptibility profiles, and the distribution of CRAB isolates categorized as multidrug resistant, extensively drug-resistant, or difficult to treat.We further discuss the proportion of isolates showing susceptibility to Food and Drug Administration-approved agents.Of note, 84% of CRAB tested harbored at least one class A, B, or D carbapene mase genes targeted for detection and 83% of these carbapenemase gene-positive CRAB were categorized as extensively drug resistant.Fifty-four percent of CRAB isolates without any of these carbapenemase genes detected were still extensively drug-resist ant, indicating that infections caused by CRAB are highly resistant and pose a significant risk to patient safety regardless of the presence of one of these carbapenemase genes.
O ver the last two decades, there have been increasing worldwide reports of Acinetobacter baumannii causing healthcare-associated infections (HAIs) and outbreaks (1)(2)(3).The global dissemination of A. baumannii as a cause of HAIs is in part attributed to its ability to persist in the environment for long periods of time and acquire resistance genes that diminish the effects of antimicrobials used against it (4).Together, these attributes contributed to the spread of multidrug-resistant A. baumannii strains.
Infections caused by A. baumannii can be severe, difficult to treat, and associated with high mortality rates (5)(6)(7).The Centers for Disease Control and Prevention's 2019 Antibiotic Resistance Threats Report identified carbapenem-resistant Acinetobacter baumannii (CRAB) as an urgent public health threat (8).Carbapenems have often been the antibiotics of choice for treating A. baumannii infections, but data reported to the National Healthcare Safety Network in 2020 demonstrated that 40% and 39.3% of A. baumannii HAI isolates in the United States were carbapenem-and multidrug-resistant, respectively (9)(10)(11).
One of the most concerning mechanisms conferring carbapenem resistance to A. baumannii and other Gram-negative bacteria is the presence of mobile carbapene mase genes, encoded by genetic elements that can easily spread among bacteria and consequently, between patients and facilities.Carbapenemase enzymes encoded by genes can be classified into molecular classes such as class A [e.g., Klebsiella pneumoniae carbapenemase (bla KPC )], class B [e.g., Active-on-Imipenem (bla IMP ), New Delhi metalloβ-lactamase (bla NDM ), and Verona integron-encoded metallo-β-lactamase (bla VIM )], and class D [e.g., Oxacillinase (bla OXA-23-like , bla OXA-24/40-like , bla OXA-48-like , and bla OXA-58- like )].Varying frequencies of these carbapenemases genes in A. baumannii have been reported globally, with class D carbapenemase genes most often detected (12)(13)(14)(15).However, there is a lack of data describing CRAB antimicrobial susceptibility profiles by class and the frequencies of specific class D subgroups detected among CRAB from across the United States.Isolates harboring these genes further limit treatment options for infections caused by organisms that are already highly resistant to many available therapeutics; therefore, rapid detection and prevention is essential to reduce ongoing transmission.
Using a large set of contemporary isolates collected nationwide from January 2017 through December 2020 as part of the Antimicrobial Resistance Laboratory Network (AR Lab Network), we describe the frequency of targeted carbapenemase genes and the antibiotic susceptibility profiles among this unique collection of CRAB isolates.

MATERIALS AND METHODS
The AR Lab Network comprises public health laboratories (PHLs) in all 50 states plus several large cities and Puerto Rico; seven of these PHLs also serve as regional laborato ries (16).Each regional laboratory conducted sentinel surveillance of CRAB isolates for their respective region (Table 1).All PHLs in the Network were asked to recruit at least one clinical laboratory from their jurisdiction to submit all CRAB isolates detected to their regional laboratory for characterization.
Since 2017, the AR Lab Network has focused its routine testing of Enterobacterales, Pseudomonas aeruginosa, and A. baumannii on detection of five carbapenemase genes, bla KPC , bla IMP , bla NDM , bla OXA-48-like , and bla VIM by real-time PCR (RT-PCR) protocols.Centers for Disease Control and Prevention (CDC)-developed or in-house RT-PCR assays for bla IMP were implemented in all regional laboratories in 2018.From January 2019, additional CDC-developed RT-PCR assays to detect the class D carbapenemase genes common among Acinetobacter species (bla OXA-23-like , bla OXA-24/40-like , and bla OXA-58-like ) were also validated and employed for A. baumannii isolate testing in regional labs.These assays use the following primers and probes: bla OXA-23-like , forward 5-GAC ACT AGG AGA AGC CAT GAA G, reverse 5-CAG CAT TAC CGA AAC CAA TAC G, and probe FAM-CCA GTC TAT CAG GAA CTT GCG CGA-BHQ1; bla OXA-24/40-like , forward 5-GAT GAC CTT GCA CAT AAC CG, reverse 5-CAG TCA ACC AAC CTA CCT GTG, and probe HEX-AGT AAC ACC CAT TCC CCA TCC ACT TTT-BHQ1; bla OXA-58-like , forward 5-AAG ATT TTA CTT TGG GCG AAG C, reverse 5-CAA CTT CCG TGC CTA TTT GC, and probe CalRd610a-TGG ACC AAT ACG ACG TGC CAA TTC T-BHQ2; and bla VIM , forward 5-GAA AAA CAC AGC GGC ACT TCT, reverse 5-CAC GCG TTA CRaG GAA GTC CAA, and probe HEX-CGG AGA TTG ARaA AGC A-BHQ1.The RT-PCR conditions for these assays were similar to previously published CDC-developed RT-PCR assays (16).
Three regional laboratories (Mountain, Central, and Northeast) performed RT-PCR assays for all eight gene targets (through additional prospective or retrospective testing) to cover the entire evaluation period of 2017-2020.Some regional laboratories also used whole-genome sequencing (WGS) to enhance molecular detection of these targets.WGS methods may vary by regional laboratory, but guidance from CDC was to follow PulseNet sequencing protocols (18).Carbapenemase gene-positive CRAB (CP-CRAB) was defined as RT-PCR or WGS positive for at least one targeted carbapenemase gene.
were defined as being not susceptible to cephalosporins, fluroquinolones, carbapenems, and β-lactam combination agents (20).Regional laboratories submitted laboratory testing results of CRAB isolates tested from January 2017 through December 2020.Data were compiled and analysed at CDC using statistical analysis software (SAS v.9.4,Cary, NC).Because not all isolates were tested against all gene targets and antimicrobial agents uniformly, denominators for frequencies were calculated using only those isolates tested for the specified genes and antimicrobial agents.

RESULTS
During 2017 through 2020, the AR Lab Network tested 6,026 CRAB isolates submitted by 45 of the 50 U.S. states (Table 1).The number of CRAB isolates tested increased annually from 312 in 2017 to 2,084 in 2020 (Table 2).Class A or class B carbapenemase genes, either alongside or without any class D carbapenemase genes, were detected in 2 of 306 (<1%) CRAB isolates from 2017 and in 26 of 1,708 (2%) CRAB isolates from 2020.Across all years evaluated, bla KPC (n = 15) and bla NDM (n = 49) were the only targeted class A or B genes detected among all CRAB isolates tested.
In 2019, when all regional laboratories were also routinely testing CRAB for additional class D carbapenemase genes, 79% (1,409 of 1,779) of all CRAB isolates tested harbored at least one class D gene (Table 3).The group and frequency of class D genes detected varied, with bla OXA-23-like detected in 47% (839 of 1779) and bla OXA-24/40-like detected in 30% (531 of 1779) of CRAB isolates tested.Detection of targeted class D genes in CRAB also varied by region, ranging from 41% (33 of 81) in the West region to 86% (378 of 440) in the Mountain region.The highest proportion of bla OXA-23-like CRAB was detected in the Mountain region (63%, 279 of 440); the highest proportion of bla OXA-24/40-like CRAB was detected in the Central region (66%, 141 of 214).No isolates harboring bla IMP ,

No
a CRAB, carbapenem-resistant Acinetobacter baumannii.b All regional labs performed testing for bla KPC , bla NDM , and bla VIM for all years of the evaluation period and for bla IMP starting in 2018.In January 2019, RT-PCR assays to detect additional class D carbapenemase genes (bla OXA-23-like , bla OXA-24/40-like , and bla OXA-58-like ) were also included for A. baumannii isolate testing.Three regional laboratories (Mountain, Central, and Northeast) performed assays for all eight gene targets (through additional prospective or retrospective testing) to cover the entire evaluation period of 2017-2020.
Table 4 shows the distribution of intermediate and resistant CLSI interpretative criteria for CRAB and CP-CRAB isolates tested from 2017 to 2020.Among CRAB with or without targeted carbapenemase genes detected, resistance to the following drugs was the least common; polymyxin B (3%, 16 of 479), colistin (5%, 24 of 499), and minocycline (13%, 65 of 483) for CRAB isolates without targeted carbapenemase genes, and polymyxin B (4%, 90 of 2,316), colistin (5%, 152 of 2,840), and minocycline (21%, 537 of 2,606) for CRAB isolates with at least one targeted carbapenemase gene.The presence of any one targeted carbapenemase gene increased the percentage of isolates that were resistant to any given antimicrobial class tested, except for the polymyxins, where no change in susceptibility was observed.For instance, imipenem resistance was observed among 56% (281 of 500) and 99% (2860 of 2,888) of isolates without and with a single targeted carbapenemase gene, respectively.Isolates harboring multiple carbapenemase genes had higher percentage (76%, 50 of 66) of isolates resistant to the aminoglycoside amikacin compared with isolates harboring a single carbapenemase gene (53%, 1486 of 2,823).A similar trend was noted for the polymyxins.Among all CRAB tested, the MIC range for tigecycline was 0.25 to ≥8 µg/mL but the MIC 50 (defined as the MIC required to inhibit the growth of 50% of the organisms) was lower (0.5 µg/mL) among isolates without any carbapenemase genes than among CP-CRAB (1 µg/mL).
Among isolates with no targeted carbapenemase genes, 84% (420 of 500) were MDR and 54% (271 of 500) were XDR (Table 5) (19).The proportion of isolates categorized as MDR and XDR increased to 100% (2,833 of 2,846) and 83% (2,355 of 2,846), respectively, when at least one targeted gene was present.Among all CP-CRAB, the proportion of XDR isolates was higher among isolates with multiple targeted carbapenemase genes (95%, 63 of 66) than among isolates harboring only one (82%, 2,292 of 2,780).Eighty-four percent (1,115 of 1,330) of the CRAB tested were DTR, but the frequency of DTR ranged from 75% (102 of 136) among CRAB without any targeted genes to 95% (42 of 44) among CP-CRAB harboring more than one targeted gene.The percentage XDR and DTR was slightly higher among bla OXA-24/40-like isolates than bla OXA23-like (Table S1).

DISCUSSION
During the years 2017-2020, although there was a slight increase in the number of CRAB isolates found to carry class A or B targeted carbapenemase genes, the overarching frequencies of these genes among all CRAB tested was still quite low at less than 1%.Although previous literature has indicated that class D carbapenemase genes are common in CRAB isolates, there is a paucity of data providing a more nuanced char acterization of how common these genes are and whether their frequency differs by geography or subgroups.Data presented in this paper revealed that the frequency of class D genes detected differed by group and region.The overall frequency of com mon class D genes was 79%, but the frequency of specific subgroups (bla OXA-23-like , bla OXA-24/40-like , and bla OXA-58-like ) varied from 0% to 47%.bla OXA-23-like genes were the most often identified; bla OXA-24/40-like genes were less common, and bla OXA-58-like and bla OXA-48-like genes were absent among the 6,026 isolates tested for this evaluation.Their geographic distribution also varied, with bla OXA-23-like detected most frequently in the Mountain region and bla OXA-24/40-like most frequently detected in the Midwest.In addition, 84% of CRAB isolates harbored at least one of the aforementioned class A, B, or D carbapenemase genes, and 83% of these CP-CRABs were categorized as XDR, yet 54% of CRAB without any of these genes were still categorized as XDR.
Class A and class D genes   These findings from AR Lab Network testing support previous reports indicating class D carbapenemase groups are frequently detected in, and associated with, carbapenemresistant A. baumannii (3,14,15,21).However, bla OXA-23-like , which was the most frequently detected gene among the CRAB isolates tested in the AR Lab Network, has also been detected recently in Proteus mirabilis (22)(23)(24).Although these bla OXA-23-likepositive P. mirabilis isolates were found to be susceptible to many antimicrobial agents, including carbapenems, the clinical and public health implications of A. baumannii serving as potential reservoirs for the spread of class D carbapenemase genes to other Gram-negative bacteria, including but not limited to P. mirabilis, are not yet understood.
Consistent with previous studies, the CRAB isolates described herein were highly drug resistant (25,26).More than half of the CRAB tested were not susceptible to at least seven classes of antimicrobial agents tested, and this percentage further increased among isolates harboring at least one targeted carbapenemase gene.In addition, the majority of CRAB isolates, whether or not they harbored a targeted gene, were not susceptible to first-line agents (i.e., DTR), highlighting how highly resistant CRAB infections can be.Although the presence of metallo-β-lactamase genes, such as bla NDM , are known to produce carbapenemase enzymes that confer resistance to almost all available β-lactam antimicrobial agents, our findings indicate that other classes of carbapenemases were also common among isolates with XDR and DTR phenotypes (27).The presence of more than one targeted carbapenemase gene further augmented resistance to the antimicrobial agents tested.These carbapenemase genes, on mobile genetic elements, are typically co-located with additional genetic determinants of resistance, which may explain the observed impact on several classes of antimicrobi als.Nevertheless, polymyxins, aminoglycosides, and tetracyclines retained some activity against CRAB isolates with and without the targeted carbapenemase genes.Antimicro bial agents such as colistin and amikacin have been associated with side effects such as nephrotoxicity and neurotoxicity, and colistin susceptibility testing can be challenging, which may further limit its utility (28).Cefiderocol, has shown potent in vitro activity against highly resistant A. baumannii isolates; however, similar to colistin, there may be some difficulty in performing cefiderocol susceptibility testing, and clinical safety and effectiveness are still being investigated (29,30).The newly Food and Drug Adminis tration-approved drug, sulbactam-durlobactam, was specifically developed for treating CRAB infections and has shown similar activity to colistin, with a more favorable safety profile (31,32).Taken together, both CRAB and CP-CRAB present difficult challenges to patient safety, given their increasing spread and resistance to almost all available therapeutics (33,34).
The data summarized have several limitations.Firstly, because the A. baumannii isolates tested were submitted as part of a sentinel surveillance program and cover age of testing varied between regions and jurisdictions, the isolates included in this evaluation do not represent all clinical isolates of CRAB circulating in the United States; therefore, any increase in CP-CRAB detected over the evaluation period does not reflect an increase in prevalence.Secondly, staggered deployment, validation, and implemen tation of RT-PCR assays to detect additional class D carbapenemase genes limits our understanding of relative frequencies of detection, by region.Thirdly, data reported to the AR Lab Network includes data from specimens collected for clinical diagnosis, surveillance, and outbreak investigation and the proportion of specimens that were positive for any of the targeted carbapenemase genes may not reflect their true relative frequencies, even among these sentinel facilities.Finally, additional access to WGS to characterize antimicrobial resistance genes and gene variants and susceptibility testing of cefiderocol could have improved our molecular and phenotypic characterization of these CRAB isolates.
The presence of mobile carbapenemase genes in CRAB that confer cross-resistance to many antibiotic agents and classes represents a significant clinical and public health concern.Using a large collection of contemporary isolates from the AR Lab Network, we described the landscape of targeted class D carbapenemase genes circulating in the United States and revealed that their relative frequencies vary substantially, depending on the specific gene group and geography.In addition, CRAB isolates harboring class A and class B carbapenemase genes were rare.Among the CRAB isolates tested, irrespec tive of whether they were carrying a targeted carbapenemase gene, were often MDR, XDR, and DTR, yet the frequency of these findings was amplified among isolates that harbored at least one targeted carbapenemase gene, indicating that CRAB infections are highly resistant and pose a significant risk to patient safety, regardless of the presence or absence of these carbapenemase genes.Given the limited treatment options available for treating CRAB infections, and the possibility of these targeted carbapenemase genes spreading among and from A. baumannii, rapid detection, treatment, and prevention of CRAB are essential to improving patient outcomes and reducing onward transmission of antibiotic resistance.

8 TABLE 4 a
Antimicrobial susceptibility a,b profiles of CRAB isolates with and without targeted c carbapenemase genes-Antimicrobial Resistance Laboratory Network, 2017-2020 k Data include antimicrobial susceptibility testing results from commercial broth microdilution panels and disk diffusion and use the CLSI interpretive criteria (M100 31st ed).

TABLE 1
Regional laboratory jurisdiction assignments in the Antimicrobial Resistance Laboratory Network a

TABLE 2
Targeted carbapenemase genes detected in CRAB a isolates-Antimicrobial Resistance Laboratory Network, 2017-2020 b

TABLE 3
Regional distribution of CRAB, a by targeted carbapenemase gene(s) detected-Antimicrobial Resistance Laboratory Network, 2019

TABLE 4
Antimicrobial susceptibility a,b profiles of CRAB isolates with and without targeted c carbapenemase genes-Antimicrobial Resistance Laboratory Network, 2017-2020 k

TABLE 4
Antimicrobial susceptibility a,b profiles of CRAB isolates with and without targeted c carbapenemase genes-Antimicrobial Resistance Laboratory Network, 2017-2020 k Due to limited doripenem concentration ranges on the Sensititre GNX2F panel (0.12-2.0 µg/mL g/mL), the interpretations presented are not-susceptible (MIC >2 µg/mL) and susceptible (MIC ≤2 µg/mL).Includes all CRAB isolates harboring any targeted carbapenemase gene, as well as those isolates from the positive for one targeted carbapenemase gene and positive for >1 targeted carbapenemase gene categories.Isolates were not uniformly tested against additional class D carbapenemase genes; therefore, the presence of those genes cannot be excluded.Includes CRAB with more than one targeted carbapenemase gene.Two isolates harbored three targeted carbapenemase genes (bla b No percentages were calculated for categories with <30 isolates tested.c Targeted carbapenemase gene-positive CRAB was defined as RT-PCR or WGS positive for at least one of the following carbapenemase genes tested: bla KPC , bla NDM , bla d e No intermediate category is available using CLSI M100 interpretative criteria.f Isolates harbored no targeted carbapenemase genes if they were tested against all gene targets (bla KPC , bla NDM , bla VIM , or bla g h Includes CRAB with only one targeted carbapenemase gene.i j k

TABLE 5
Antimicrobial resistance a,b phenotypes of CRAB isolates with and without targeted c carbapene mase genes-Antimicrobial Resistance Laboratory Network, 2017-2020 k Data include antimicrobial susceptibility testing results from commercial broth microdilution panels, and disk diffusion and use the CLSI interpretive criteria (M100 31st edition).CLSI, Clinical and Laboratory Standards Institute; DTR, difficult-to-treat resistance (not susceptible to all reported carbapenems, β-lactams, fluoroquinolones, and piperacillin-tazobactam); MDR, multidrug resistant (not susceptible in ≥1 agent in ≥3 categories); XDR, extensively drug resistant (not susceptible to ≥1 agent in all but ≤2 categories).
(20)No percentages were calculated for categories with <30 isolates tested.cTargetedcarbapenemasegene-positiveCRAB was defined as RT-PCR or WGS positive for at least one of the following carbapenemase genes tested: bla KPC , bla NDM , bla VIM , bla IMP , bla OXA-23-like , bla OXA-24/40-like , bla OXA-48- like , or bla OXA-58-like .dMultidrug-resistant(notsusceptible in ≥1 antimicrobial agent in ≥3 antimicrobial classes) and extensively drug resistant (not susceptible to ≥1 antimicrobial agent in all but ≤2 antimicrobial classes) definitions used CLSI interpretative criteria except for polymyxins, where EUCAST breakpoints were used and were based on definitions set by Magiorakos et al.(19).Difficult-to-treat resistance (not susceptible to all reported carbapenems, β-lactams, fluoroquinolones, and piperacillin-tazobactam) definitions used CLSI interpretative criteria and were based on definitions set by Kadri et al.(20).e Only isolates with antimicrobial susceptibility data for at least seven antimicrobial classes were included for analysis.f Isolates harbored no targeted carbapenemase genes if they were tested against all gene targets (bla KPC , bla NDM , bla VIM , or bla IMP bla OXA-23-like , bla OXA-24/40-like , bla OXA-48-like , or bla OXA-58-like ) by RT-PCR or WGS and were negative for all.g Includes all CRAB isolates harboring any targeted carbapenemase gene, as well as those isolates from the positive for one targeted carbapenemase gene and positive for >1 targeted carbapenemase gene categories.h Includes CRAB with only one targeted carbapenemase gene.i Isolates were not uniformly tested against additional class D carbapenemase genes; therefore, presence of those genes cannot be excluded.j Includes CRAB with more than one targeted carbapenemase gene.Two isolates harbored three targeted carbapenemase genes (bla NDM /bla OXA-23-like /bla OXA-24/40-like ).k Abbreviations: CRAB, carbapenem-resistant Acinetobacter baumannii;