Molecular epidemiology and clinical significance of carbapenemase genes in carbapenem-resistant Acinetobacter baumannii isolates in southern Poland

Introduction

Acinetobacter spp. are nonfermenting gram-negative opportunistic pathogens that can colonize the skin, respiratory system, and gastrointestinal tract. They are responsible for many health care–associated infections, including ventilator-associated pneumonia and bloodstream infections (BSIs), especially in patients hospitalized in intensive care units (ICUs).¹ Their ability to evade the host immune system significantly complicates efforts to manage and treat these infections.² Since 2017, carbapenem-resistant Acinetobacter baumannii (CRAB) has consistently remained a critical priority pathogen and urgent threat to the public according to the World Health Organization and the United States (US) Centers for Disease Control and Prevention (CDC).^3,4

Importantly, the epidemiology of Acinetobacter spp. can vary greatly by region. In the European Union / European Economic Area, approximately one-third of all Acinetobacter spp. isolates between 2017 and 2021 were carbapenem-resistant.⁵ After the peak of the COVID-19 pandemic in 2021, the number of reported Acinetobacter spp. cases resistant to carbapenems, fluoroquinolones, and aminoglycosides more than doubled, as compared with the 2018−2019 average. The greatest increases in Acinetobacter spp. antimicrobial resistant (AMR) cases were reported by countries that already had high AMR prevalence, such as Poland.⁵

In the US, incidence of CRAB has decreased from 3.3 per 10 000 hospitalizations in 2012 to 2.47 in 2017.⁴ The decrease is generally attributed to improved infection control and antibiotic stewardship practices that have become requirements for hospital accreditation in the US.⁶ There is concern this progress has been hampered by the COVID-19 pandemic, which led to increased antibiotic use and breakdown of infection control practices leading to reports of CRAB outbreaks.⁷ The global impact of CRAB is in part due to the ability of its resistance genes to spread via mobile genetic elements, as well as widespread antibiotic use.² This has a negative impact on patient outcomes, as the mortality rate of A. baumannii infections ranges from 45% to 70%. One of its main predictors is carbapenem resistance,¹ and carbapenems are frequently an empiric treatment choice for infections in areas such as Poland, where the prevalence of extensively drug-resistant organisms was 22.6% in the ICUs and 14.8% in non-ICUs,⁸ leading to delays in optimal antibiotic coverage for CRAB infections. This is further emphasized by CRAB being the fifth leading cause of death from resistant pathogens worldwide.⁹

In terms of resistance, A. baumannii employs a variety of mechanisms to reduce the effectiveness of antibiotic therapy. They include the use of efflux pumps, changes to antibiotic active sites, reduction of cell membrane permeability, and, probably the most common, the use of enzymes to inactivate antibiotics, particularly β-lactamases.¹⁰ A. baumannii has 2 intrinsic types of β-lactamases: AmpC-type cephalosporinases (no effect on extended-spectrum cephalosporin efficacy) and oxacillinases, mainly represented by the OXA-51-like variants^11,12 and belonging to the OXAAb enzyme group, including OXA-66.¹³ In addition to the intrinsic β-lactamases, other β-lactamases have also been identified in A. baumannii as a source of resistance to carbapenems, including OXA-23, OXA-40, OXA-48, and OXA-58.¹⁴ The mechanisms of resistance related to carbapenemase production can be classified into 3 groups: Ambler class A (plasmid-mediated KPC and chromosomal IMI, SME, GES, and NMC-A genes), Ambler class B (metallo-β-lactamases whose genes are located on integrons and plasmids [NDM, IMP, VIM, GIM, SPM, and SIM genes]), and Ambler class D (plasmid-mediated oxacillinases, eg, OXA-48 or OXA-23).¹⁵ Prevalence of the resistance genes is geographically-specific, which results in difficulty generalizing local studies to other regions, and highlights the importance of local and regional studies to guide locally-relevant optimal treatment approaches.

Treatment of CRAB is difficult both in terms of diagnosis and antibiotic therapy. Rapid diagnostics (eg, based on nucleic acid amplification tests) to confirm infections with CRAB are often lacking in areas with high CRAB prevalence, which may result in delayed administration of appropriate antibiotics and negatively affect patient outcomes.¹⁶ The most common CRAB-related infection is pneumonia,¹⁷ and there are specific pharmacokinetic / pharmacodynamic properties of antibiotics used against CRAB that result in difficulty achieving appropriate drug levels in lung tissues.^18,19 In addition, minimum inhibitory concentration (MIC) breakpoints for antibiotics against CRAB differ by organization^20,21 (the European Committee on Antimicrobial Susceptibility Testing [EUCAST] recommendations apply in Poland) or do not exist, posing challenges for antibiotic choice and dosing. In other infections, especially those involving medical devices, such as endotracheal or tracheostomy tubes, CRAB’s ability to form biofilms can prevent antibiotics from penetrating and reaching the pathogen.^22,23

In this study, we evaluated the resistance profiles of CRAB in 4 hospitals in southern Poland to characterize the prevalence of various CRAB β-lactamases and to inform regional empiric therapies for these difficult-to-treat infections. We also commented on the utility of commercial rapid tests in CRAB surveillance upon hospital admission. The resistance genes identified by these tests are not commonly detected in CRAB. Consequently, there is a potential risk of missing CRAB infections or colonization when these tests are utilized.

Patients and methods

This laboratory-based study was carried out at the University Hospital in Kraków, St. Luke’s Provincial Hospital in Tarnów, the Provincial Hospital No. 5 in Sosnowiec, and the District Hospital in Bochnia.

Ethical approval was waived by the Bioethics Committee of the Jagiellonian University in view of the retrospective nature of the study, all procedures being performed as part of routine care, and the analysis not including any participant-identifying data. All data analyzed during this study were anonymized prior to the analysis. As a result, no informed consent was required from the participants.

The samples were collected between July and December 2022. A total number of admissions was 70 859. During the study, a multiplex polymerase chain reaction (PCR) (GeneXpert System, Cepheid, Sunnyvale, California, United States) was used for screening of the carbapenemase genes (types: bla_VIM, bla_NDM, bla_IMP, bla_KPC, bla_OXA-48) to confirm the presence of carbapenemase-producing Enterobacterale (CPE) isolates on admission to the hospital in the cases of suspected CPE colonization (eg, antibiotic therapy, previous hospitalization, stay in long-term care facilities). A bacterial health care–associated infection (HAI) in adult patients was defined as a symptomatic infection diagnosed (or recognized) over 48 hours since the hospital admission. HAI cases were analyzed retrospectively using definitions from the Healthcare-Associated Infections Surveillance Network,²⁴ which included BSIs, pneumonia, and urinary tract infections. As recommended by the European Centre for Disease Prevention and Control, bacterial diagnostic testing and its interpretation were performed according to the specimen type.

Some of the specimens were analyzed quantitatively; these included urine and positive quantitative cultures from minimally or possibly contaminated lower respiratory tract (LRT) specimens. Interpretation of quantitative results in mono- and polymicrobial cultures depended on the relative quantity of each microorganism. For a urine culture to be positive, the sample was required to have no more than 2 organism species and at least 10⁵ colony forming units (CFUs) per milliliter. For an LRT culture to be positive, bronchoalveolar lavage (BAL) was required to have at least 10⁴ CFU/ml or at least 5% of BAL-obtained cells needed to contain intracellular bacteria on direct microscopic examination, specimens from the LRT collected with a Wimberley protected brush were required to have at least 10³ CFU/ml, distal protected aspirate needed at least 10³ CFU/ml, and endotracheal aspirate needed at least 10⁶ CFU/ml.²⁴

Microbiologic samples were obtained from the sites of infection. Only laboratory-confirmed A. baumannii HAI cases based on culture growth were qualified for the analysis; only the first isolate from each HAI case was analyzed. During the study period, 120 HAI cases (eg, pneumonia) were identified as caused by A. baumannii. As many as 107 out of the 120 isolates (89.2%) were identified as CRAB, and 82 CRAB isolates were analyzed (25 isolates were lost to follow-up for reasons independent from the authors).

The isolated organisms were identified by matrix-assisted laser desorption / ionization time-of-flight mass spectrometry (Bruker, Billerica, Massachusetts, United States or VITEK MS, bioMérieux, Craponne, France). Antibiotic susceptibility tests were performed by the hospital diagnostic laboratories using the automatic VITEK 2 method (bioMérieux), Phoenix M50 system (Becton Dickinson, Sparks, Maryland, United States), disc diffusion (Oxoid, Basingstoke, United Kingdom), MIC test strips (Liofilchem, Roseto degli Abruzzi, Italy), or broth microdilution in the case of colistin (MIC stripped plates, Diagnostics, Galanta, Slovakia), according to the EUCAST guidelines.²⁰

The GeneMATRIX bacterial and yeast genomic DNA purification kit (EURx, Gdańsk, Poland) was used to extract genomic DNA from the CRAB isolates following the manufacturer’s protocol. The concentration and purity of the isolated DNA were assessed using a Nano Drop Lite spectrophotometer (Thermo Fisher Scientific, Waltham, Massachusetts, United States). DNA extracted from pure cultures was stored at –20 °C for further analyses.

The carbapenemase genes were identified using multiplex PCR (bla_VIM, bla_OXA-48, bla_OXA-23, bla_KPC, bla_NDM, bla_OXA-40, bla_OXA-58, bla_IMP, bla_GIM, bla_GES, bla_OXA-51, bla_IMI, and bla_VIM)²⁵ and real-time PCR (bla_OXA-66 and bla_TEM).²⁶ PCR amplification was performed using the Color OptiTaq PCR Master Mix (EURx) in a final volume of 25 μl and a final primer concentration of 0.1 μl for each primer. Bacterial DNA served as a template. For the multiplex PCR, the reaction was conducted with an initial denaturation step of 3 minutes at 94 °C, followed by 30 cycles of 30 seconds at 94 °C, 15 seconds at 58 °C, and 1 minute at 72 °C for amplification. A final extension step of 5 minutes at 72 °C was performed. The PCR products were analyzed via gel electrophoresis. For the real-time PCR, cycling was carried out at 50 °C for 2  minutes, and then 95 °C for 2  minutes, followed by 40 cycles at 95 °C for 15  seconds, 55 °C for 15  seconds, and 72 °C for 1  minute (primer data are described in Supplementary material, Table S1). The datasets analyzed during the current study are available from the corresponding author upon reasonable request.

Statistical analysis

In the statistical analysis, relative and absolute frequencies were used for nominal variables (described as number and percentage), and mean values with SD were used for quantitative variables. For data with non-normal distribution, median with interquartile range (IQR) was used. For independent samples with nominal variables, the χ² test was employed. The Yates correction was used for 2 × 2 Tables of nominal variables (when the expected values were <⁠5) and the Fisher exact test was used for 2 × (n) Tables (when the expected values were <⁠5 and there was a small overall sample size [n <⁠40]). For continuous data with normal distribution, the t test was used. For continuous variables without normal distribution and independent samples, the Mann–Whitney test was performed.

The analyses were carried out with the International Business Machines Corporation Statistical Package for the Social Sciences, version 29 (IBM Corporation, Armonk, New York, United States). In all analyses, the significance level was set at a P value below 0.05.

Results

In total, 82 CRAB isolates collected from the patients with A. baumannii infection were analyzed. Median (IQR) age of the patients was 65 (ICU, 65 [63–73]; non-ICU, 68 [52–80]) years (Table 1). Sources of infection included LRT (n = 38), bloodstream (n = 23), urinary tract (n = 13), and others (n = 8; including 6 wounds and 2 cerebrospinal fluid sources). Fifty-two CRAB isolates were collected from ICU patients and the most common CRAB sources in the ICU were LRT infections (n = 28; 54%) and BSIs (n = 14; 27%). In the non-ICU setting, 30 CRAB isolates were collected, and the most common CRAB sources were also LRT infections (n = 10; 33%) and BSIs (n = 9; 30%).

Table 1. Patient demographics and epidemiology of Acinetobacter baumannii hospital-acquired infections in southern Poland between June and December 2022

Parameter	ICU (n = 52)	non-ICU (n = 30)	P value
Demographic data
Age, y, median (IQR)	65 (63–73)	68 (52–80)	0.44
Sex, n (%)
Male	42 (80.8)	17 (56.7)	0.04
Female	10 (19.2)	13 (43.3)
Admissions, n (%)
Tarnów	199 (21.4)	14 114 (20.1)	<⁠0.001
Sosnowiec	210 (22.6)	13 323 (19.1)
Kraków	441 (47.5)	31 398 (44.9)
Bochnia	79 (8.5)	11 095 (15.9)
Non-CRAB infections, n (%)
Tarnów	1 (25)	0 (0)	0.15
Sosnowiec	2 (50)	9 (100)
Kraków	0 (0)	0 (0)
Bochnia	1 (25)	0 (0)
CRAB infectionsa, n (%)
Tarnów	12 (21.4)	20 (39.2)	<⁠0.001
Sosnowiec	9 (16.1)	28 (54.9)
Kraków	29 (51.8)	1 (2)
Bochnia	6 (10.7)	2 (4)
Prevalence of CRAB, %
Tarnów	92.3	100	0.73
Sosnowiec	81.8	75.7
Kraków	100	100
Bochnia	85.7	100
Incidence rate of CRAB infections (10 000 admissions)
Tarnów	603	14.2	<⁠0.001
Sosnowiec	428.6	21
Kraków	657.6	0.3
Bochnia	759.5	1.8
a CRAB infections by hospital type are provided for 82 out of the 107 isolates identified as CRAB. Abbreviations: CRAB, carbapenem-resistant Acinetobacter baumannii; ICU, intensive care unit; IQR, interquartile range

Depending on the hospital, the incidence rate of CRAB infections ranged from 428.6 to 759.5 per 10 000 admissions in ICUs, and from 0.3 to 21 per 10 000 admissions in non-ICUs. The proportion of CRAB isolates (in particular units) ranged from 81.8% to 100% in the ICUs and 75.7% to 100% in the non-ICUs (Table 1).

The most prevalent bla_OXA genes were bla_OXA-66-1 (95%), bla_OXA-40 (71%), bla_OXA-23 (24%), and bla_OXA-51 (12%). The bla_NDM genes were detected in 2% of the isolates, and bla_TEM-1 genes were detected in 41% of the isolates. No bla_OXA-48, bla_OXA-58, bla_VIM, bla_KPC, bla_IMI, bla_GES, bla_GIM, or bla_IMP genes were detected (Table 2). The bla_OXA-23 gene was found more often in men than in women (32.2% vs 4.3%; P = 0.019).

Table 2. β-Lactamase gene prevalence in 82 carbapenem-resistant Acinetobacter baumannii isolates from 4 hospitals in southern Poland between June and December 2022

Genea	ICU (n = 52), n (%)	Non-ICU (n = 30), n (%)		P value
blaOXA-23	15 (28.8)		5 (16.7)	0.33
blaOXA-40	33 (63.5)		25 (83.3)	0.1
blaOXA-51	10 (19.2)		0	0.03
blaOXA-66-1	52 (100)		26 (86.7)	0.03
blaNDM	2 (3.8)		0	0.73
blaTEM	17 (32.7)		17 (56.7)	0.06
a There were no carbapenem-resistant Acinetobacter baumannii isolates with the blaOXA-48, blaOXA-58, blaGES, blaGIM, blaIMI, blaIMP, blaKPC, and blaVIM genes. Abbreviations: see Table 1

Only 4 isolates had a single gene encoding carbapenemases, while the other had multiple β-lactamase genes. As many as 43 isolates had 2 genes, 29 had 3, and 5 had 4. Frequency of the detected gene patterns is shown in Table 3 (additional data are provided in Supplementary material, Table S2).

Table 3. Prevalence of detected gene patterns in 82 carbapenemase-producing Acinetobacter baumannii isolates from health care–associated infections at 4 hospitals in southern Poland between June and December 2022

Gene pattern	Number of strains	%
blaOXA-40, blaOXA-66-1	27	32.9
blaOXA-40, blaOXA-66-1, blaTEM	22	26.8
blaOXA-23, blaOXA-66-1	8	9.8
blaOXA-23, blaTEM	4	4.9
blaOXA-66-1	4	4.9
blaOXA-40, blaOXA-66-1, blaOXA-51	2	2.4
blaOXA-40, blaOXA-23, blaOXA-66-1	2	2.4
blaOXA-23, blaOXA-66-1, blaOXA-51	2	2.4
blaOXA-40, blaOXA-66-1, blaOXA-51, blaTEM	2	2.4
blaOXA-66-1, blaTEM	2	2.4
blaOXA-40, blaOXA-23, blaOXA-66-1, blaTEM	2	2.4
blaOXA-40, blaOXA-23, blaOXA-66-1, blaOXA-51	1	1.2
blaOXA-23, blaOXA-66-1, blaOXA-51, blaNDM, blaTEM	1	1.2
blaOXA-66-1, blaOXA-51	1	1.2
blaOXA-66-1, blaOXA-51, blaTEM	1	1.2
blaOXA-66-1, blaNDM	1	1.2

CRAB antibiotic susceptibility was the highest for cefiderocol (100%), colistin (96%), tigecycline (77%), gentamicin (51%), and ampicillin / sulbactam (36%). Less than 4% susceptibility was found in the case of ciprofloxacin, levofloxacin, meropenem, imipenem, trimethoprim / sulfamethoxazole, piperacillin / tazobactam, piperacillin, and imipenem / relebactam (Table 4) (additional data are provided in Supplementary material, Table S3). Statistical analysis did not reveal significant differences among the sources of infection (blood, urine, or LRT), and susceptibility of CRAB to individual antibiotics.

Table 4. Antibacterial susceptibility^a of 82 carbapenemase-producing Acinetobacter baumannii isolates collected from patients with health care–associated infections at 4 hospitals in southern Poland between June and December 2022 by the sampling source

Antibiotics	Susceptibility, %
	Total (n = 82)b	Respiratory tract (n = 38)	Bloodstream (n = 23)	Urine (n = 13)
β-Lactams, penicillins
Ampicillin / sulbactam	52.4	52.6	65.2	30.8
Piperacillin	0	0	0	0
Piperacillin / tazobactam	0	0	0	0
Carbapenems
Imipenem	1.2	0	0	0
Imipenem / relebactam	0	0	0	0
Meropenem	2.4	0	4.3	0
Aminoglycosides
Amikacin	24.4	26.3	21.7	23.1
Gentamicin	51.2	60.5	39.1	38.5
Tobramycin	19.5	23.7	13	15.4
Fluoroquinolones
Ciprofloxacin	4.9	5.3	0	7.7
Levofloxacin	3.7	2.6	4.3	7.7
Other antibiotics
Cefiderocol	100	100	100	100
Colistin	96.3	97.4	91.3	100
Minocycline	39	47.4	30.4	30.8
Tigecycline	76.8	78.9	73.9	69.2
Trimethoprim / sulfamethoxazole	2.4	2.6	4.3	0
a Antibiotic susceptibility results were interpreted using the European Committee on Antimicrobial Susceptibility Testing guidelines.20 b Includes 8 other sources

Discussion

The observed prevalence and incidence of CRAB in the studied hospitals significantly exceeded expected values. In addition, CRAB resistance was high against most antibiotics, and the antibiotics with better susceptibility profiles are generally associated with toxicities (eg, colistin). This highlights an urgent need to prevent CRAB infections, including avoiding the spread of CRAB, and the need for new treatments against CRAB infections.

Most A. baumannii isolates harbored multiple carbapenemase genes highlighting their ability to use many resistance mechanisms. The most common combination was bla_OXA-66-1 with bla_OXA-40. The bla_OXA-66-1 gene (part of the bla_OXA-51-like gene family) is intrinsic to A. baumannii.¹² At appropriate levels of expression, it has been associated with high levels of imipenem resistance in A. baumannii,²⁷ but there is evidence that the degree of carbapenem resistance can be associated with the expression of other genes and resistance mechanisms within the pathogen.²⁸ In our sample of carbapenem-resistant organisms, bla_OXA-66-1 was found with another β-lactamase gene in all but 4 organisms, indicating possible reliance on other enzymes for carbapenem resistance. The bla_OXA-40 gene prevalence has previously been described in southern Poland for the years 2005–2010, when 51 out of 104 isolates (49%) harbored this gene.²⁹ Its enzymatic product is capable of hydrolyzing carbapenems as well as other β-lactams.²⁸ As many as 58% of our isolates (54.2%) contained the bla_OXA-40 genes, thus showing approximately a 5% point increase in its prevalence. This is a worrying phenomenon, which indicates that the prevalence of bla_OXA-40 in Poland is approximately 3-fold higher than in the rest of Europe.³⁰ Almost half of the studied isolates (41%), harbored bla_TEM genes. The TEM genes typically confer resistance to penicillins and early cephalosporins but have expanded their activity to include resistance against second-, third- and fourth-generation cephalosporins, monobactam, and β-lactamase inhibitors.³¹ The TEM genes are found at high frequencies in hospitals and clinics around the world,³² and often co-occur with other chromosomal (AmpC) or plasmid-mediated (SHV, OXA, CTX-M) β-lactamase encoding genes.³³ The 24% prevalence of bla_OXA-23 in our sample is 3-fold lower than in the rest of Europe, where bla_OXA-23 genes were identified in 74.5% of analyzed A. baumannii isolates.³⁰ Of note, commercially available rapid diagnostic tests only cover 5 enzymes (or their genes) (VIM, NDM, IMP, KPC, and OXA-48). Unfortunately, these genes / enzymes are not the major driver of A. baumannii resistance (as noted in our study), complicating efforts to detect the resistance mechanisms in A. baumannii.

A. baumannii is inherently difficult to treat, as it exhibits a variety of resistance mechanisms. In CRAB, the resistance to carbapenems is often associated with resistance to other categories of antibiotics, such as fluoroquinolones.³⁴ Thus, the aim of this study was to provide clinicians with insight on empiric antibiotic regimens in Poland. There is no clear answer which antibiotic or which combination of antibiotics is a superior choice in treatment of CRAB infections. Yet, even with this uncertainty, the guidelines of both the European Society of Clinical Microbiology and Infectious Diseases (ESCMID) and the Infectious Diseases Society of America (IDSA) are mostly consistent in their CRAB treatment recommendations; they recommend the use of a combination therapy for treating CRAB infections.^35,36 This increases the likelihood that one of the agents is effective against CRAB, given the lack of efficacy of single antibiotic regimens. Many CRAB infections occur in the ICU, where severely ill patients are at high risk for these infections and increased antibiotic use predisposes them to infections with resistant pathogens.³⁷ The ICU setting often involves advanced life support interventions, such as continuous renal replacement therapy and extracorporeal membrane oxygenation, further complicating antibiotic treatment and dosing considerations due to diffusion, convection, and ultrafiltration mechanisms that alter antimicrobial concentrations in tissues.³⁸

The IDSA guidelines recommend the use of sulbactam as the basis in most treatment regimens,³⁵ regardless of the susceptibility profile of ampicillin / sulbactam.³⁹ Only 36% of our CRAB isolates were susceptible to ampicillin / sulbactam, so in accordance with the IDSA guidelines, most isolates would need to be treated with high-dose ampicillin / sulbactam (9 g of sulbactam component) over an extended infusion time due to ampicillin / sulbactam resistance.³⁵

ESCMID, contrary to IDSA, recommends ampicillin / sulbactam only when susceptibility is confirmed. Polymyxin (colistin) or high-dose tigecycline, if an isolate is susceptible in vitro, can be used when resistance to ampicillin / sulbactam occurs.³⁶ Our isolates exhibited relatively high susceptibility to these antibiotics (colistin 96% and tigecycline 77%). Unfortunately, their use is not free from issues. Colistin-related toxicities include kidney injury and poor lung penetration,^16,40 and the most recent Clinical and Laboratory Standards Institute update on colistin breakpoints does not include reference MIC values,²¹ though EUCAST does.²⁰ The use of tigecycline against CRAB infections needs to be considered in the context of limited clinical data (ESCMID suggests its use in complicated intra-abdominal infections and in skin and soft tissue infections)³⁶. Cefiderocol is an option against CRAB, especially for pneumonia,⁴¹ and has been suggested in combination with sulbactam.¹⁶ Our isolates were equally susceptible to cefiderocol but there were issues with laboratory techniques and obtaining accurate cefiderocol MICs.⁴² Studies have not found an association between the bla_OXA genes and cefiderocol resistance,⁴³ which is reflected in 100% sensitivity of our isolates to cefiderocol. Fluoroquinolones are not recommended by IDSA, and would not be useful given low susceptibility of our isolates.³⁵ The ESCMID guidelines do not refer to fluoroquinolones in CRAB treatment.³⁶ Thus, the combination antibiotic regimen of choice depends on multiple factors that include the site of infection, pharmacokinetic / pharmacodynamic considerations, and the pathogen’s susceptibility profile. For example, Shields et al¹⁶ recommend ampicillin / sulbactam with tigecycline or cefiderocol for pneumonia and ampicillin / sulbactam with polymyxin B or cefiderocol for BSIs.

The use of meropenem in combination with other antibiotics has fallen out of favor. There were 2 clinical trials comparing colistin and colistin plus meropenem,^44,45 which showed no difference between the colistin arm and the combination of meropenem with colistin in patients with CRAB. This suggests the use of meropenem did not add to the efficacy of colistin even when dosed appropriately and over an extended infusion period. Thus, the IDSA guidelines do not recommend the use of carbapenems in CRAB infections.³⁵ ESCMID allows for administration of high-dose extended-infusion of carbapenem when the MIC is equal to or below 8 µg/l but as part of a combination therapy. It is worth noting that this is a good practice statement and expert opinion.³⁶

New treatments against CRAB are thus urgently needed. Some of the most recently introduced agents are sulbactam / durlobactam. They were studied, in combination with imipenem / cilastatin, in a clinical noninferiority trial⁴⁶ that showed higher mortality outcomes for colistin with imipenem / cilastatin than sulbactam / durlobactam with imipenem / cilastatin. The active component in sulbactam / durlobactam combination is sulbactam; durlobactam was added to inactivate OXA carbapenemases, which allows sulbactam to reach its target. The sulbactam / durlobactam combination has been recently approved by the US Food and Drug Administration against A. baumannii pneumonia but further data are needed to determine whether it should be used as monotherapy or in combination with other active antibiotics.⁴⁷ Due to novelty of this therapy, we were unable to include sulbactam / durlobactam in our study, and this combination has not yet been approved for clinical use in Europe. The CDC indicates that other β-lactam / β-lactam inhibitors, meropenem / vaborbactam, can be effective in resistant gram-negative infections, but specifically in infections caused by CPE that express KPC and OXA-48–like enzymes, not in CRAB infections.⁴⁸

Given difficult treatment of CRAB infections, prevention is a key component in their surveillance. Screening for CRAB can decrease prevalence of CRAB infections in an ICU by reducing transmission between patients when early isolation is implemented.⁴⁹ It is important to consider which body site to utilize for CRAB screening, since the sites harbor isolates of different sensitivity.⁵⁰ The skin swab isolates show the highest sensitivity (92% sensitivity from skin alone and 99% sensitivity from a combination of buccal mucosal and skin samples). Isolates collected from other sites can show a high rate of false-negative results (eg, 50% sensitivity in detecting CRAB colonization from the rectal site).⁵⁰ The prevalence of CRAB colonization in our hospitals is unknown since screening is not routinely performed.

Carbapenem resistance markers found on routine early detection tests are common in Enterobacterales but not in A. baumannii, limiting utility of these methods in CRAB detection. These early-detection tests include phenotype-based methods detecting the activity of selected carbapenemases (eg, rapid colorimetric methods), immunochromatographic tests, and molecular methods (mainly based on polymerase chain reaction). The main limitation preventing their use is the small number of detected carbapenemases (only KPC, NDM, OXA-48, IMP, and VIM, or the genes encoding them). All of them are characteristic for the Enterobacterales order⁵¹; almost none of them occurred in the A. baumannii isolates we tested. Only 5% of the tested isolates (4 of 82) had at least 1 gene encoding these carbapenemases. Other tests, detecting OXA-23, OXA-40, and OXA-58 dedicated to A. baumannii, can be used for screening, but this still does not fully solve the surveillance problem of CRAB. Hence, infection control and prevention of carbapenemase-producing microorganisms, including CRAB infections is limited to nonspecific methods, such as hand hygiene.

Limitations

Poland does not have an effective national system of surveillance for AMR,⁵² which limits comparability of the current data with the rest of the country and restricts the possibility of properly assessing temporal trends. Such systems are crucial to understand the epidemiology of CRAB and to evaluate the effect of interventions to slow its spread. Our samples came from 4 hospitals of different type but from the same geographic area, which limits generalizability of this study to other parts of Poland. These limitations provide opportunities for further research into the epidemiology of CRAB in Poland and support the need for a national plan to better understand and curb antimicrobial resistance.

Conclusions

CRAB is an urgent public health threat. It is widespread throughout the world and is a major pathogen in Polish hospitals. Understanding its epidemiology, resistance mechanisms, and antibiotic susceptibility is crucial to implement focused infection control and antibiotic stewardship techniques and to help inform empiric and directed antibiotic treatments. This study provides up-to-date data on CRAB epidemiology in southern Poland, and interprets international treatment guidelines in the regional context. It indicates that southern Poland’s CRAB prevalence is higher, as compared with the rest of Europe, and is probably rising. This is concerning, since the antibiotic susceptibility profiles established in this study showed that the isolates are highly resistant to most used antibiotics, regardless of the infection site, leaving patients with less effective and more toxic treatment options. The genes involved in CRAB resistance are also not detected by routine surveillance tests, further complicating surveillance efforts. Thus, our study highlights the need for effective AMR prevention initiatives, better diagnostic surveillance tests, and improved treatment options against this difficult-to-treat pathogen.