Carbapenem-Resistant Gram-Negative Bacilli Characterization in a Tertiary Care Center from El Bajio, Mexico

Carbapenem-resistant Gram-negative bacilli (CR-GNB) are a major public health concern. We aimed to evaluate the prevalence of CR-GNB and the frequency of carbapenemase-encoding genes in a tertiary referral center from El Bajio, Mexico. A cross-sectional study was conducted between January and October 2022; Gram-negative bacilli (GNB) were screened for in vitro resistance to at least one carbapenem. CR-GNB were further analyzed for carbapenemase-production through phenotypical methods and by real-time PCR for the following genes: blaKPC, blaGES, blaNDM, blaVIM, blaIMP, and blaOXA-48. In total, 37 out of 508 GNB were carbapenem-resistant (7.3%, 95% CI 5.2–9.9). Non-fermenters had higher rates of carbapenem resistance than Enterobacterales (32.5% vs. 2.6%; OR 18.3, 95% CI 8.5–39, p < 0.0001), and Enterobacter cloacae showed higher carbapenem resistance than other Enterobacterales (27% vs. 1.4%; OR 25.9, 95% CI 6.9–95, p < 0.0001). Only 15 (40.5%) CR-GNB had a carbapenemase-encoding gene; Enterobacterales were more likely to have a carbapenemase-encoding gene than non-fermenters (63.6% vs. 30.8%, p = 0.08); blaNDM-1 and blaNDM-5 were the main genes found in Enterobacterales; and blaIMP-75 was the most common for Pseudomonas aeruginosa. The mcr-2 gene was harbored in one polymyxin-resistant E. cloacae. In our setting, NDM was the most common carbapenemase; however, less than half of the CR-GNB showed a carbapenemase-encoding gene.


Introduction
Antimicrobial resistance (AMR) rates have become a major concern in global health; if the trend continues this way, AMR could become the leading cause of death by 2050 [1].Carbapenem-resistant Enterobacterales (CRE), carbapenem-resistant Pseudomonas aeruginosa (CRPA), and carbapenem-resistant Acinetobacter baumannii (CRAB) represent high-priority pathogens [2].The main problem lies in their resistance mechanisms, as they delay appropriate antibiotic therapy, increase hospital length of stay, and worsen patients' survival [3].
The phenomenon of multidrug-resistant (MDR) bacteria has been specially attributed to the indiscriminate use of antibiotics and the presence of antimicrobial resistance genes.Carbapenemases are described as the most frequent mechanism by which carbapenem resistance has increased [4,5]; nevertheless, it is not the only one.Efflux pumps and porin mutations are other mechanisms described in non-carbapenemase-producing CRE, CRPA and CRAB [5,6].
The rate of carbapenem resistance is considerably higher for non-fermenters (often >60%) than fermenters (often <10%) due to intrinsic factors, including reduced outer membrane permeability and efficient efflux pumps, even in wild-type strains [8]; however, prevalence of carbapenemase-encoding genes varies widely among studies, ranging from 4 to 32% for CRPA, with predominance of bla VIM , bla IMP , and bla GES, between 45 and 90% for CRE, with clear dominance of bla KPC and bla NDM , and up to 80 to 100% for CRAB, the main genes found being bla OXA-like [9].Global distribution of carbapenemases has shown a predominance of bla KPC in North America and Europe; bla NDM and bla IMP in some regions of Asia; bla VIM in Europe; and bla OXA-48 in the Middle East and Africa [5].In Mexico, epidemiological data have been expanding in recent years, showing bla NDM as the most frequent gene in both K. pneumoniae and E. coli, followed by bla KPC and bla VIM ; among P. aeruginosa, bla VIM and bla IMP are the main encoded genes, whereas A. baumannii has a predominance of bla OXA-24 , bla OXA-40 , and bla OXA-23 [10,11].
As information on carbapenem resistance varies by geographical situation and is constantly evolving, regional surveillance is fundamental for the implementation of adequate infection control measures and provide good antibiotic use policies.We aimed to estimate the prevalence of carbapenem-resistant Gram-negative bacilli (CR-GNB) and characterize the presence of carbapenemase-encoding genes in a tertiary regional referral center from El Bajío, Mexico.

Carbapenemase Production
First, carbapenemase-producing strains were phenotypically identified by the modified carbapenem inactivation method (mCIM) in 15 of the 37 CR-GNB samples (40.5%); of the latter, twelve were inhibited by the ethylenediaminetetraacetic acid (EDTA) carbapenem inactivation method (eCIM), suggesting Ambler class B carbapenemases, and the remaining three samples were negative to eCIM.Secondly, the NG-test CARBA-5 ® identified the involved protein in 12 of the 37 CR-GNB samples (32.4%).Lately, carbapenemase-encoding genes were identified by PCR; fifteen isolates (40.5%) were positive for one carbapenemaseencoding gene; and, out of them, seven strains (46.7%) had the bla NDM gene, four (26.7%) the bla IMP gene, three (20%) the bla GES gene, and one (6.6%) the bla VIM gene.No bla OXA-like or bla KPC genes were detected.There were no co-producer genes.Enterobacterales showed a higher rate of carbapenemase-encoding genes than non-fermenters (63.6% vs. 30.8%);however, this result did not reach statistical significance (p = 0.08).

Discussion
Globally, many regions have not implemented active surveillance programs to monitor CR-GNB.Our work is the first to characterize carbapenemase genes in the Bajio region of Mexico, expanding the knowledge of the mechanisms of resistance in our country.We found an overall carbapenem-resistance rate in Gram-negative bacteria of 7.3% (95% CI, 5.2-9.9),being 2.6% (95% CI 1.4-3.5)for Enterobacterales and 32.5% (95% CI 22.4-43.9)for non-fermenters, which is comparable with data from Mexico but lower than other international reports [4,9,12].
The Antimicrobial Surveillance Program SENTRY in Latin America showed that 4.3% of Enterobacterales strains were CRE, with the highest rates in Brazil (9%) and Argentina (6.3%) and the lowest in Chile (0.4%) and Mexico (0.7%) [14].However, these numbers have evolved in recent years, and our study showed a CRE rate of 2.6%.The GLASS report in 2021 mentioned a worldwide prevalence of CRE up to 10-15% [15].However, it should be noted that Mexico was not part of the antimicrobial resistance surveillance programs included in this study.On the other hand, the prevalence of carbapenem resistance among non-fermenters in Latin America has been described as high as 66% for P. aeruginosa and up to 90% for A. baumannii [16], discordant to the findings of our study.
Carbapenem resistance is directly influenced by the culture site, being higher in respiratory infections than in bloodstream infections [17].This was concordant with our analysis, where the highest proportion of CR-GNB were isolated from lower respiratory tract samples, whereas the lowest proportion was for urine cultures.
Given the high lethality rate associated with CRE, CRPA, and CRAB, assessment of the hospital-or ward-level risk of CR-GNB is crucial, especially by defining the specific resistance mechanisms [3].However, few laboratories can perform routine phenotypic or genotypic testing for carbapenemases; thus, determining the local epidemiology allows for a better choice of treatment [18].The latter becomes relevant due to the possible use of the new, although not widely available, β-lactam/β-lactamase inhibitors, such as ceftazidime/avibactam, meropenem/vaborbactam, and imipenem/relebactam, against KPC-producing strains [3,18,19].Nevertheless, the emergence of resistance to ceftazidime/avibactam has been recently described in KPC-3-producing strains [20]; ceftazidime/avibactam or cefiderocol as monotherapy are options against OXA-producing strains [18,21,22].For MBL-producing strains, there are more limited options, especially the combination of ceftazidime/avibactam with aztreonam or cefiderocol as monotherapy, whereas tigecycline, eravacycline, and polymyxins remain as alternative but more toxic options, regardless of the presence of carbapenemases [18,23].This becomes a challenge for Mexico since it is a country with a predominance of MBLs [10,11].
The most frequent types of carbapenemase-encoding genes in our study were bla NDM in Enterobacterales and bla IMP in P. aeruginosa; however, no bla OXA or bla KPC were identified.Worldwide, bla NDM and bla KPC are the main carbapenemases found in Enterobacterales;, whereas bla OXA , bla IMP , bla VIM, and bla GES are primarily found in P. aeruginosa and A. baumannii [5,[24][25][26].Few studies in Mexico have evaluated the presence of carbapenemaseencoding genes in CRE, CRPA, and CRAB.For CRPA, bla IMP and bla VIM , followed by bla GES , are the most commonly described [10,11,27,28], in line with our results.Meanwhile, for CRE, some Mexican authors have described bla NDM-1 as the most associated gene; however, bla OXA-48 , bla OXA-181 , bla OXA-232, and bla KPC-2 have also been reported [10,11,[29][30][31].In addition to bla NDM-1 , we identified isolates carrying the bla NDM-5 and bla NDM-7 genes.For CRAB, the most frequent encoding genes described were bla OXA-24 and bla OXA-40 , followed by bla OXA-23 , bla VIM, and bla NDM [11].It is noteworthy that we had a high proportion of isolates with no carbapenemase-encoding genes (60%) in contrast to previously reported studies [4,12,25,32], suggesting the presence of non-enzymatic mechanisms or carbapenemases other than the ones evaluated, closer to those described by Garza-Gonzalez et al. [11], where 47% was negative for the presences of the carbapenemases tested.
The use of point-of-care tests, such as NG-test CARBA-5 ® , is a good complementary option to guide antimicrobial therapy since they have shown excellent performances compared with molecular techniques and provide results within 15-30 min.However, it should be considered that they do not detect all types of carbapenemases, as seen with bla GES in our sample, and may not perform as well for P. aeruginosa [33].
In addition to allowing better guidance of antibiotic therapy, distinguishment between carbapenemase-mediated and non-enzymatic mechanisms is important since the former has been associated with higher odds of dying at 14 days and can spread more efficiently between patients, requiring an efficient equipment disinfection policy, combined with timely isolation of the cohort to reduce the spread of these strains [9,34].However, as noncarbapenemase-producing strains not only target carbapenems and β-lactam antibiotics, they usually are associated with MDR.Mutations in AmpC and its regulatory genes are commonly attributed to P. aeruginosa resistance to new β-lactam-β-lactamase inhibitor combinations [35], whereas plasmid-encoded AmpC is most commonly found in resistant Enterobacterales [36].Among efflux-pump systems, AcrAB-TolC RND is the most commonly described [37].Finally, porin mutations (especially in OmpC and OmpF) block the entry of carbapenems into the cell [38].
Polymyxin susceptibility remains in almost all cases, except for one E. cloacae with the presence of mcr-2 gene.Despite their high toxicities, polymyxins continue as a last-line pharmacological group in cases of Gram-negative MDR infections.The mcr genes are of great concern due to their high potential for horizontal propagation.Among the variants described, mcr-1 is the most prevalent within Mexico [10] and globally [39]; thus far, mcr-2 has been described in human specimens, especially in Asia [40,41].However, to the best of our knowledge, this is the first report of mcr-2 in a Mexican patient.

Design
A cross-sectional study was carried out between 1 January and 31 October 2022 at the Hospital Regional de Alta Especialidad del Bajío, a 184-bed tertiary regional referral center located in León, Guanajuato, Mexico.

Sampling
Isolates from clinically relevant sites samples, such as blood, urine, respiratory, cerebrospinal fluid, peritoneal, pleural, and deep-tissue cultures with Gram-negative bacterial growth, were included.All data were collected in the microbiology laboratory at the Hospital Regional de Alta Especialidad del Bajío.Identification was performed with VITEK-2 (Biomérieux, Marcy l' Etoile, France), and susceptibility tests were performed by microdilution broth method following the Clinical and Laboratory Standards Institute (CLSI) 2018 M07 recommendations [42] and breakpoints according to M100 Ed. 32 (2022) criteria [43].Duplicated isolates, i.e., more than one isolate per patient in the same hospitalization, were identified and discharged of the database.Bacteria with resistance to at least one carbapenem (meropenem, imipenem, doripenem, and ertapenem; this last one was not considered for Pseudomonas aeruginosa) were selected for analysis.Bacteria from single colonies were stored at −20 • C in brain heart infusion broth with 15% of glycerol until used.
An internal protocol (HRAEB-POE-013) for type B biological substances (UN 3373) was used to send the samples to the infectious disease laboratory at the Instituto Nacional de Rehabilitación Luis Guillermo Ibarra Ibarra in Mexico City.

Tests
We performed two phenotypic tests to screen for carbapenemase-producing isolates.First, we performed mCIM (modified carbapenemase inactivation method).To identify metallo ß-lactamase-producing strains, we used the mCIM variant, with EDTA (ethylenediaminetetraacetic acid) as a chelator (eCIM) [42].Once the producer strain was identified, the next step was to identify the protein involved through NG-test CARBA-5 ® [33] (KPC, OXA-48, VIM, IMP, and NDM).(eCIM).All these isolates were further analyzed by a polymerase chain reaction (PCR) to identify the following carbapenemase-encoding genes: bla KPC , bla GES , bla NDM , bla VIM , bla IMP , and bla OXA-48 .An Enterobacter cloacae was shown to be resistant to colistin; therefore, it was decided to be screened for mobile colistin resistance (mcr) genes.The gene sequences used are listed in Table 5.
Table 5. Gene sequences used for polymerase chain reaction.Bacterial DNA was extracted by thermal shock (95 • C/20 min) using PBS 1×.Next, the tubes were centrifuged (10,000 rpm/20 min), and the supernatants were separated in new tubes.PCR conditions were the following: 10× PCR Buffer (New England BioLabs, Ipswich, MA, USA) containing 2 mM of MgCl 2 , 0.2 mM of each of the dNTPs (Invitrogen, Waltham, MA, USA), 10 pmol of each oligonucleotide (Table 5), 1.5 U of Taq polymerase (BioLabs, USA), and 3 uL of DNA.

Gen
The amplification was carried out using the following program: 1 cycle of 5 min at 95 • C, 35 cycles of 50 s at 95 • C, 60 s at 55 • C, and 50 s at 68 • C, followed by 1 final cycle of 10 min at 68 • C (Veriti, Applied Biosystems, Waltham, MA, USA); Tm for bla IMP was 52 • C, Tm bla oxa-48 , bla VIM , bla GES , bla NDM , and bla KPC was 55 • C; Tm for mcr-1 and mcr-2 was 54 • C.
PCR products were separated in a 1% agarose gel at 120 volts for 40 min, and SYBR-Green (Invitrogen, USA) was used such as DNA intercalant.Bands were visualized in Gel Doc XR+ (BioRad, Hercules, CA, USA) with Image Lab Software version 6.1 (BioRad, Hercules, CA, USA).

Sequencing
Prior to the sequencing process, re-amplification and labeling using the BigDye ® Direct Cycle Sequencing Kit (Thermo Scientific, Waltham, MA, USA) were performed.Sequencing was performed using the two oligonucleotides (described for the amplification).A 3730xl DNA Analyzer (Applied Biosystems, USA) was used for sequencing, using POP-7 (Thermo Scientific, USA).GenBank was used to compare the sequences obtained with those of reference organisms in this database [44].Identification was corroborated by performing independent sequencing for each primer.

Statistical Analysis
SPSS software version 25.0 was used for statistical analysis.A descriptive analysis of the distribution of all Gram-negative strains at the hospital was performed, as well as an analysis of the subpopulation of carbapenem-resistant Gram-negative strains.Groups were established according to the site where the sample was obtained and the microorganism isolated.The absolute and relative frequencies for each group and their relationships with the antibiotic resistance profile and the presence of carbapenemases were described.A comparative analysis between groups was performed.p-values were obtained using the chi-square test (x 2 ) or Fisher's exact test for qualitative variables, according to their expected values.For quantitative variables, the Student's t test or Mann-Whitney U test was used according to their distribution, which was evaluated by the Shapiro-Wilk test.The confidence interval will be 95%, with a p-value required for statistical significance <0.05.

Limitations
Despite being a regional referral center, our results may not be applicable in primary care settings, as our patients were more likely to have previous hospitalizations and antibiotic exposures due to their underlying conditions.We did not evaluate all the described carbapenemase-encoding genes.Also, porin and efflux-pump mutations were not assessed; therefore, we could not attribute with certainty the mechanism of resistance in some strains.As we did not evaluate our population throughout time, future research should focus in this direction to analyze the CR-GNB trends in our country.

Conclusions
In the region of El Bajio, Mexico, the prevalences of carbapenem-resistant strains are 2.6% for Enterobacterales and 32.5% for non-fermenters.Less than half of CR-GNB showed a carbapenemase-encoding gene.bla NDM was the most frequent carbapenemase-encoding gene for Enterobacterales, whereas bla IMP-75 was for P. aeruginosa.Continuous surveillance is necessary to provide adequate control measures and improve antibiotic stewardship.E. cloacae in our setting will be of special attention since we reported the first Mexican isolate harboring the mcr-2 gene.Funding: This research received no external funding.The APC was covered by the Hospital Regional de Alta Especialidad del Bajío.

Institutional Review Board Statement:
The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of the Hospital Regional de Alta Especialidad del Bajío (reference number HRAEB/CB/007/2021) in November 2021.
Informed Consent Statement: Patient consent was waived because no intervention was involved, and no patients' identifying information was taken.

Table 1 .
Origin and bacterial differences between carbapenem-resistant and susceptible isolates.

Table 2 .
Frequency of carbapenemase resistance in Gram-negative according to culture site.
Prevalence of carbapenem-resistant bacilli by culture site and species.

Table 4 .
Antimicrobial resistance of the carbapenem-resistant isolates.