Contemporary multicentre analysis of carbapenemases distribution among clinical carbapenem-resistant Enterobacterales in Lebanon Authors: Jame Tannous, Rindala Saliba, Abdo Mghames, Sima Tokajian, Rana Attieh, Tamima Jisr, Jamil Barhoun, Mohamad Khaled Diab, Diaa Abou Shakra, Jacques Mokhbat, Anwar Al Souheil, Wael Zorkot, Fatima Jaffal, Hasan Salman, Rola Husni, Hind Rahban, Eliane Khoury, Zeina Mohsen, Edmond Abboud, Mona Youssef, Gilbert Karayakoupoglo, Ismail Fawaz, Mohamad Houmani, Dania Abdallah, Rima Moghnieh Published: 2026 ABSTRACT AbstractBackgroundContemporary multicentre data on carbapenemase distribution among carbapenem-resistant Enterobacterales (CRE) in Lebanon remain limited. This study aimed to characterize the distribution of carbapenemase enzyme families among clinical CRE isolates collected from Lebanese hospitals.MethodsA multicentre study was conducted across 14 hospitals representing different Lebanese regions between March and October 2025. CRE isolates were centrally processed using standardized phenotypic microbiological methods and interpreted according to the 2025 European Committee on Antimicrobial Susceptibility Testing breakpoints. Carbapenemase identification was performed using NG-Test® CARBA-5. Agreement between phenotypic prediction and rapid immunochromatographic testing was evaluated.ResultsA total of 59 non-duplicate CRE isolates from individual patients were included. New Delhi metallo-β-lactamase (NDM) was the predominant carbapenemase enzyme family (64.4%), followed by OXA-48–like enzymes (23.7%), with co-production observed in 10.2% of isolates. Among carbapenem-resistant Escherichia coli (34 isolates), NDM predominated (73.5%). In contrast, carbapenem-resistant Klebsiella pneumoniae (15 isolates) more frequently harboured OXA-48–like enzymes (53.3%), although NDM-producing isolates were also identified (33.3%). No NDM-producing isolate was susceptible to ceftazidime–avibactam, whereas OXA-48–like producers retained higher susceptibility (71.4%). Phenotypic carbapenemase prediction showed high agreement with NG-Test® CARBA-5 results (κ = 0.89), including 100% sensitivity for metallo-β-lactamase detection.ConclusionsNDM was the predominant carbapenemase among CRE isolates recovered from participating Lebanese hospitals, while Oxacillinase-48 carbapenemase (OXA-48)–like enzymes remained common. These findings highlight the importance of carbapenemase-specific characterization, as therapeutic options differ according to the underlying resistance mechanism. Rapid carbapenemase detection using NG-Test® CARBA-5, coupled with phenotypic prediction, may support resistance mechanism-guided therapy, antimicrobial stewardship and surveillance efforts. Introduction Infections caused by carbapenem-resistant Enterobacterales (CRE) represent a major global health challenge and are associated with increased mortality, prolonged hospitalization and higher healthcare costs than infections caused by carbapenem-susceptible organisms.1 Carbapenem resistance most commonly results from carbapenemase production, although non-enzymatic mechanisms, such as porin loss and efflux pump overexpression, may also contribute.2–4 Rapid identification of the underlying carbapenemase is essential for guiding appropriate antimicrobial therapy.5 The activity of newer β-lactam/β-lactamase inhibitor combinations depends on the specific resistance mechanism. Current guidelines recommend meropenem–vaborbactam, ceftazidime–avibactam or imipenem–relebactam for infections caused by Klebsiella pneumoniae carbapenemase (KPC)-producing organisms, whereas metallo-β-lactamase (MBL) producers, including New Delhi metallo-β-lactamase (NDM), require alternative therapeutic approaches such as aztreonam–avibactam or cefiderocol.6,7 For OXA-48-like–producing Enterobacterales, ceftazidime–avibactam and cefiderocol remain important treatment options.6,7 These differences highlight the importance of understanding local carbapenemase epidemiology to optimize empiric and targeted therapy. The NG-Test® CARBA-5 is a rapid immunochromatographic assay that detects the five most prevalent carbapenemase families: KPC, NDM, Verona integron-encoded metallo-β-lactamase (VIM), imipenemase metallo-β-lactamase (IMP) and OXA-48 like.3 Multicentre evaluations have demonstrated high agreement with molecular methods, supporting its use as a practical tool for routine carbapenemase detection and mechanism-guided therapeutic decision-making.5,8,9 In Lebanon, CRE have been increasingly reported over the past decade. OXA-48–like enzymes were historically the predominant carbapenemases in tertiary-care centres, whereas more recent reports indicate increasing detection of NDM-producing isolates.10,11 However, these observations are largely derived from independent hospital-based studies, and a contemporary assessment of circulating carbapenemase profiles across Lebanese hospitals remains lacking. In this context, we conducted a multicentre study to characterize the distribution of carbapenemase enzyme families among clinical CRE isolates collected from hospitals across Lebanon using the NG-Test® CARBA-5 assay. We further evaluated their geographic distribution and correlated carbapenemase profiles with antimicrobial susceptibility patterns to provide an updated overview relevant to resistance mechanism-guided therapy. Materials and methods Study design and isolate collection Between March and October 2025, a total of 59 CRE isolates were collected from 14 hospitals across Lebanon (Figure 1). Isolates were submitted during two predefined collection phases, each lasting approximately 4 months, to reduce temporal clustering and minimize overrepresentation of outbreak-related strains. Participating centres were asked to include isolates from both inpatient and outpatient settings to capture a broader representation of CRE epidemiology across different healthcare settings. Each isolate was assigned a unique alphanumeric identifier according to the submitting hospital’s internal coding system. Geographic distribution of clinical Enterobacterales isolates collected from participating hospitals across Lebanese districts (n = 59). The map displays the number of clinical Enterobacterales isolates contributed by participating hospitals within each Lebanese district during the study period. Microbiological identification and antimicrobial susceptibility testing All isolates were transferred to the coordinating laboratory for standardized analysis. Species identification was reconfirmed using routine microbiological methods to ensure consistent classification across centres. Antimicrobial susceptibility testing was performed using the Kirby–Bauer disc diffusion method on Mueller–Hinton agar and interpreted according to the 2025 European Committee on Antimicrobial Susceptibility Testing (EUCAST) clinical breakpoints.12 Susceptibility was evaluated against 24 clinically relevant antimicrobial agents representing β-lactams, aminoglycosides, quinolones and trimethoprim–sulfamethoxazole using commercially available discs (Bio-Rad Laboratories, Hercules, CA, USA). Phenotypic detection of carbapenem resistance mechanisms An isolate was classified as carbapenem resistant if it demonstrated resistance to at least one carbapenem (ertapenem, meropenem or imipenem) according to the 2025 EUCAST clinical breakpoints, in line with European Centre for Disease Prevention and Control (ECDC) surveillance definitions.12,13 Ertapenem non-susceptibility was used as an initial screening trigger because of its higher sensitivity for detecting carbapenemase production and other mechanisms associated with reduced carbapenem susceptibility, including permeability defects combined with Extended-spectrum β-lactamase (ESBL) or AmpC β-lactamase (AmpC) overexpression. All carbapenem-resistance profiles were reconfirmed at the coordinating laboratory prior to carbapenemase testing to ensure standardized interpretation and minimize inter-laboratory variability. Phenotypic characterization of carbapenem-resistance mechanisms was performed using the MASTDISCS® Combi Carba Plus disc system (Mast Group Ltd, Bootle, UK), according to the manufacturer’s instructions. Assays were performed on Mueller–Hinton agar inoculated with a 0.5 McFarland bacterial suspension and incubated at 35–37°C for 18–24 h. The MASTDISCS Combi Carba Plus system comprises five discs: Disc A (faropenem alone), Disc B (faropenem plus MBL inhibitors), Disc C (faropenem plus KPC inhibitor), Disc D (faropenem plus AmpC inhibitor) and Disc E (temocillin plus MBL inhibitors). Results were interpreted according to the manufacturer’s criteria and published validation studies.14 Phenotypic mechanisms were defined as follows: MBL production: zone diameter difference ≥5 mm between Disc B and Disc A, with a difference <5 mm between Disc C or D and Disc A. MBL production: zone diameter difference ≥5 mm between Disc B and Disc A, with a difference <5 mm between Disc C or D and Disc A. KPC production: zone diameter difference ≥5 mm between Disc C and Disc A, with a difference <5 mm between Disc B and Disc A. KPC production: zone diameter difference ≥5 mm between Disc C and Disc A, with a difference <5 mm between Disc B and Disc A. OXA-48–like production: inhibition zone ≤10 mm around Disc E in the absence of MBL or KPC activity. OXA-48–like production: inhibition zone ≤10 mm around Disc E in the absence of MBL or KPC activity. Non-carbapenemase–mediated resistance: negative results across all inhibitor discs, consistent with mechanisms such as AmpC hyperproduction combined with outer-membrane porin loss, with cefoxitin resistance considered supportive of possible AmpC activity. Non-carbapenemase–mediated resistance: negative results across all inhibitor discs, consistent with mechanisms such as AmpC hyperproduction combined with outer-membrane porin loss, with cefoxitin resistance considered supportive of possible AmpC activity. In previously published evaluations against molecularly characterized carbapenemase-producing Enterobacterales isolates, the MASTDISCS Combi Carba Plus system demonstrated a sensitivity of 86% and specificity of 98% for carbapenemase detection.15 Aztreonam–avibactam synergy testing was performed primarily for isolates suspected of producing MBLs based on phenotypic screening results. Aztreonam (30 µg) and avibactam (4 µg) discs were placed 10 mm apart (edge to edge) on Mueller–Hinton agar. Distortion or enhancement of the aztreonam inhibition zone towards the avibactam disc was interpreted according to previously published synergy criteria.16 Consequently, synergy testing was not systematically performed for all isolates. Quality control was performed in parallel with each testing batch using Klebsiella pneumoniae ATCC BAA-1705 (KPC positive), Escherichia coli ATCC BAA-2469 (NDM positive) and Pseudomonas aeruginosa ATCC 27853 as a susceptible control. Zone diameter interpretive criteria were verified against EUCAST quality-control ranges.17 Carbapenemase detection Carbapenemase detection was performed using the NG-Test® CARBA-5 immunochromatographic lateral flow assay (NG Biotech, France), according to the manufacturer’s instructions. The assay detects the five major carbapenemase enzyme families: KPC, NDM, VIM, IMP and OXA-48 like.9 Results were considered valid only when the internal control band was present. Ethical considerations The study was approved by the Institutional Review Board of the Lebanese American University (IRB #: LAUMCRH.RM3.3/Jun/2024). All isolates were obtained as part of routine clinical care. Patient identifiers were removed prior to analysis, and no identifiable patient information was accessed. Statistical analysis Descriptive statistics were used to summarize isolate characteristics. Categorical variables were expressed as frequencies and percentages. Comparisons of antimicrobial resistance patterns between NDM-producing and OXA-48-like–producing isolates were performed using Fisher’s exact test because of small expected cell counts. Agreement between phenotypic carbapenemase prediction and NG-Test® CARBA-5 results was evaluated using contingency table analysis. Carbapenemase results were grouped into three categories: (i) MBL, including NDM and NDM/OXA-48–like co-producers; (ii) OXA-48–like carbapenemase; and (iii) no carbapenemase detected. Overall percent agreement was calculated as the proportion of concordant classifications, and Cohen’s kappa (κ) coefficient was used to assess agreement beyond chance. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) for phenotypic carbapenemase prediction were calculated using NG-Test® CARBA-5 as the operational comparator assay. Exact (Clopper–Pearson) 95% confidence intervals were calculated for diagnostic performance measures. All statistical analyses were performed using GraphPad Prism version 10.2.1 (GraphPad Software, San Diego, CA, USA). A two-tailed P-value of <0.05 was considered statistically significant. Results Demographic and clinical characteristics A total of 59 non-duplicate CRE isolates from individual patients were included. The cohort was predominantly adult (55/59, 93.2%), with a median age of 71 years, and 54.2% of patients were male (Table 1). Most isolates were recovered from inpatient settings (42/59, 71.2%), including 21 (35.6%) from critical care units. The urinary tract was the most common specimen source, followed by skin and soft tissue and respiratory specimens. Isolates were obtained from hospitals across multiple Lebanese districts, with the largest proportion originating from Beirut and its suburbs (Figure 1). Table 1.Demographic and clinical characteristics of patients (n = 59)Characteristics n (%)Age group (years) Adults (median age = 71 years)55 (93.2) Paediatric (<18)4 (6.8) Adults <50 years4 (6.8) Adults ≥50 years51 (86.4)Gender Male32 (54.2) Female27 (45.8)Nationality Lebanese58 (98.3) Other1 (1.7)Care setting Inpatient42 (71.2) Outpatient17 (28.8)Ward type Critical care21 (35.6) Non-critical care21 (35.6)Specimen source Urinary tract31 (52.5) Skin and soft tissue12 (20.3) Respiratory tract8 (13.6) Blood3 (5.1) Screening4 (6.8) Genital tract1 (1.7)Geographic location Beirut/suburbs (five hospitals)26 (44.1) Bekaa (one hospital)4 (6.8) North (two hospitals)10 (16.9) South (three hospitals)6 (10.2) Metn/Keserwan (two hospitals)7 (11.9) Al-Chouf (one hospital)6 (10.2) Demographic and clinical characteristics of patients (n = 59) Species distribution and carbapenemase profiles Escherichia coli was the most common species (34/59, 57.6%), followed by K. pneumoniae (15/59, 25.4%) and Enterobacter spp. (7/59, 11.9%; Figure 2a; Table 2). Distribution of Enterobacterales species and carbapenemase profiles among clinical isolates (n = 59). (a) Species distribution of clinical Enterobacterales isolates. (b) Overall distribution of carbapenemase profiles. (c–e) Species-specific distribution of carbapenemase profiles among E. coli, K. pneumoniae and Enterobacter spp. isolates. NDM, New Delhi metallo-β-lactamase; OXA-48 like, OXA-48–like carbapenemase family. Table 2.Species-specific distribution of carbapenemase profiles and antimicrobial susceptibility patterns among CRE isolates (n = 59)Total (n = 59) (%) E. coli (n = 34) (%) K. pneumoniae (n = 15) (%) Enterobacter spp. (n = 7) (%)Others (n = 3) (%)Carbapenemase profile NDM38 (64.4)25 (73.5)5 (33.3)6 (85.7)2 (66.7) OXA-48 like14 (23.7)6 (17.6)8 (53.3)0 (0)0 (0) NDM + OXA-48 like6 (10.2)3 (8.8)1 (6.7)1 (14.3)1 (33.3) No carbapenemase detected1 (1.7)0 (0)1 (6.7)0 (0)0 (0) Antimicrobial susceptibility β-Lactams Cefoxitin3 (5.1)1 (2.9)2 (13.3)00 Cefotaxime1 (1.7)1 (2.9)000 Ceftazidime1 (1.7)1 (2.9)000 Cefepime2 (3.4)1 (2.9)1 (6.7)00 Piperacillin–tazobactam00000 Ceftazidime–avibactam11 (18.6)5 (14.7)6 (40.0)00 Aztreonam16 (27.1)8 (23.5)6 (40.0)2 (28.6)0Carbapenems Imipenem6 (10.2)3 (8.8)3 (20.0)00 Meropenem3 (5.1)03 (20.0)00 Ertapenem00000Non-β-lactams Gentamicin23 (39.0)17 (50.0)4 (26.7)1 (14.3)1 (33.3) Amikacin32 (54.2)22 (64.7)5 (33.3)4 (57.1)1 (33.3) Ciprofloxacin3 (5.1)2 (5.9)01 (14.3)0 Trimethoprim/sulfamethoxazole3 (5.1)2 (5.9)1 (6.7)00 Species-specific distribution of carbapenemase profiles and antimicrobial susceptibility patterns among CRE isolates (n = 59) NG-Test® CARBA-5 identified NDM as the predominant carbapenemase, detected in 38 of 59 isolates (64.4%), followed by OXA-48–like carbapenemases in 14 (23.7%). Co-production of NDM and OXA-48–like carbapenemases was observed in six isolates (10.2%), while one isolate had no detectable carbapenemase (Figure 2b). NDM predominated among E. coli and Enterobacter spp., whereas OXA-48–like carbapenemases were more frequent among K. pneumoniae (Figure 2c–e; Table 2). Geographic distribution of carbapenemase profiles The distribution of carbapenemase profiles varied across districts (Figure 3). Beirut and its suburbs demonstrated a mixed pattern, including NDM, OXA-48–like carbapenemases and NDM/OXA-48–like co-producers. NDM predominated among isolates recovered from Northern Lebanon, Metn/Keserwan and the Bekaa, whereas OXA-48–like carbapenemases accounted for half of the isolates recovered from Southern Lebanon. Given the limited number of isolates from several districts, these geographic findings should be interpreted cautiously. Distribution of carbapenemase profiles among clinical Enterobacterales isolates collected from participating hospitals across Lebanese districts (n = 59). The graph illustrates the geographic distribution of carbapenemase profiles detected among clinical Enterobacterales isolates submitted by participating hospitals from different Lebanese districts. NDM, New Delhi metallo-β-lactamase; OXA-48 like, OXA-48–like carbapenemase family. Antimicrobial susceptibility patterns Overall susceptibility to carbapenems was low, with no isolate susceptible to ertapenem (Table 2). Susceptibility to ceftazidime–avibactam was observed in 11 isolates (18.6%), while amikacin demonstrated the highest susceptibility rate among non-β-lactam agents (54.2%). When stratified by carbapenemase, no NDM-producing or NDM/OXA-48–like co-producing isolate was susceptible to ceftazidime–avibactam or meropenem (Table 3). In contrast, 71.4% of OXA-48–like producers remained susceptible to ceftazidime–avibactam (P < 0.001). Aztreonam susceptibility was numerically lower among NDM producers than among OXA-48–like producers, although this difference did not reach statistical significance (P = 0.086). Trimethoprim/sulfamethoxazole susceptibility was observed only among OXA-48–like producers (P = 0.016). Table 3.Antimicrobial susceptibility patterns among CRE isolates by carbapenemase profileAntibioticTotal (n = 59) (%)NDM (n = 38) n (%)OXA-48-like (n = 14) n (%)NDM + OXA-48-like (n = 6) n (%) P-valueaCeftazidime–avibactam11 (18.6)0 (0.0)10 (71.4)0 (0.0) <0.001Aztreonam16 (27.1)7 (18.4)6 (42.9)3 (50.0)0.086Ertapenem0 (0.0)0 (0.0)0 (0.0)0 (0.0)—Imipenem6 (10.2)2 (5.3)2 (14.3)1 (16.7)0.291Meropenem3 (5.1)0 (0.0)2 (14.3)0 (0.0)0.069Gentamicin23 (39.0)15 (39.5)7 (50.0)1 (16.7)0.540Amikacin32 (54.2)21 (55.3)9 (64.3)1 (16.7)0.753Ciprofloxacin3 (5.1)2 (5.3)1 (7.1)0 (0.0)0.99Trimethoprim/sulfamethoxazole3 (5.1)0 (0.0)3 (21.4)0 (0.0) 0.016Avibactam–aztreonam synergyb Positive among tested34/36 (94.4)29/31 (93.5)2/2 (100)3/3 (100) Negative among tested2/36 (5.6)2/31 (6.5)0/2 (0.0)0/3 (0.0) Not tested22/58 (37.9)7/38 (18.4)12/14 (85.7)3/6 (50.0) aFisher’s exact test. bAztreonam–avibactam synergy testing was not performed systematically for all isolates and was primarily undertaken in isolates with phenotypic evidence suggestive of metallo-β-lactamase production.Bold values mean statistically significant, P-value < 0.05. Antimicrobial susceptibility patterns among CRE isolates by carbapenemase profile aFisher’s exact test. bAztreonam–avibactam synergy testing was not performed systematically for all isolates and was primarily undertaken in isolates with phenotypic evidence suggestive of metallo-β-lactamase production. Bold values mean statistically significant, P-value < 0.05. Among isolates tested for aztreonam–avibactam synergy, positive synergy was observed in 34 of 36 (94.4%). Testing was not performed systematically and was primarily reserved for isolates with phenotypic findings suggestive of MBL production. Agreement between phenotypic prediction and NG-Test® CARBA-5 High agreement was observed between phenotypic carbapenemase prediction and NG-Test® CARBA-5 results. Overall agreement was 94.9% (56/59), with a Cohen’s κ coefficient of 0.89. The contingency table and diagnostic performance metrics underlying the agreement analysis are provided in Table S1 (available as Supplementary data at JAC-AMR Online). All MBL producers identified by NG-Test® CARBA-5 were correctly classified by phenotypic testing, yielding a sensitivity of 100%, specificity of 80.0%, PPV of 93.6% and NPV of 100%. Three OXA-48–like producers were phenotypically classified as MBL producers. For OXA-48–like carbapenemases, phenotypic prediction demonstrated a sensitivity of 78.6%, specificity of 100%, PPV of 100% and NPV of 93.8%. The single isolate without a detectable carbapenemase was correctly categorized as non-carbapenemase–mediated resistance. For agreement analysis, NDM/OXA-48–like co-producers were categorized as MBL positive. Discussion In this multicentre study of CRE isolates collected from 14 hospitals across Lebanon in 2025, NDM was the predominant carbapenemase, accounting for nearly two-thirds of isolates. OXA-48–like carbapenemases remained an important mechanism, and NDM/OXA-48–like co-production was observed in a subset of isolates. NDM predominated among E. coli and Enterobacter spp., whereas OXA-48–like carbapenemases were more frequent among K. pneumoniae. Importantly, NDM was identified across multiple hospitals and geographic regions. The predominance of NDM was reflected in the low susceptibility of the overall cohort to ceftazidime–avibactam, while OXA-48–like producers retained greater susceptibility to this agent. In addition, phenotypic carbapenemase prediction demonstrated high agreement with NG-Test® CARBA-5, supporting the feasibility of combining phenotypic and rapid immunochromatographic approaches for carbapenemase characterization in routine practice. Evolution of carbapenemase epidemiology in Lebanese hospitals The epidemiology of carbapenemase-producing Enterobacterales in Lebanese hospitals has evolved through distinct phases over the past two decades (Figure 4). The first documented carbapenemase was IMP-1, identified in a single K. pneumoniae isolate in 2007.18 Shortly thereafter, OXA-48 was reported in clinical isolates of K. pneumoniae and E. coli, marking the initial emergence of OXA-48–mediated resistance in the country.19 During this early period, carbapenem resistance was rare and confined to sporadic hospital cases. Evolution of carbapenemase epidemiology among Enterobacterales in Lebanese hospitals, 2007–25. The timeline summarizes key published reports describing the evolution of carbapenemase-producing Enterobacterales in Lebanon, from the initial detection of IMP-1 and OXA-48-like carbapenemases to the establishment of OXA-48 endemicity, followed by increasing diversification and dissemination of NDM-producing isolates. The timeline is based on published studies and intended to provide historical context for the current findings. It does not represent a formal national surveillance dataset. IMP, imipenemase metallo-β-lactamase; NDM, New Delhi metallo-β-lactamase; OXA-48 like, OXA-48–like carbapenemase family. Between 2008 and 2011, hospital-based studies continued to report low rates of carbapenem resistance among ESBL-producing Enterobacterales, with early coexistence of OXA-48 and NDM-1.20 A clear epidemiologic shift occurred in 2012, when a multicentre hospital study demonstrated that OXA-48 accounted for approximately 70% of characterized carbapenemases among carbapenem-non-susceptible Enterobacterales.11 Although overall resistance remained limited, this study established OXA-48 as the dominant carbapenemase in Lebanese hospitals. From 2016 onward, molecular heterogeneity increased. Whole-genome sequencing studies conducted on isolates collected between 2016 and 2020 documented the coexistence of OXA-48, OXA-181 and multiple NDM variants, including NDM-1, NDM-5, NDM-7 and NDM-19.21 While OXA-48 remained frequent, NDM variants collectively represented a substantial proportion of isolates, indicating expansion of MBLs within hospital settings. Species-level analyses suggested that this expansion was not uniform, with NDM-1 predominating among carbapenem-resistant Enterobacter cloacae complex isolates.22 Data from Northern Lebanon further documented the emergence of NDM-5, particularly among carbapenem-resistant E. coli, where it accounted for more than half of detected carbapenemases among isolates collected during 2019–20.10 More recent genomic investigations have reinforced these observations. Whole-genome sequencing of CRE and ceftazidime–avibactam-resistant Enterobacterales collected at a large tertiary-care hospital between 2019 and 2021 demonstrated coexistence of blaNDM and blaOXA-48–like carbapenemases, including multiple NDM variants, among both E. coli and K. pneumoniae isolates.23 Subsequently, ICU-based surveillance from 10 hospitals in southern Lebanon identified NDM-5 as the dominant carbapenemase among carbapenem-resistant E. coli recovered during 2023, whereas K. pneumoniae frequently harboured OXA-48–like enzymes and/or NDM variants.24 Collectively, available evidence suggests a transition from an OXA-48–dominant phase beginning in 2012 towards a more heterogeneous carbapenemase landscape characterized by persistent circulation of OXA-48–like carbapenemases alongside increasing detection of NDM variants. Consistent with these recent molecular and genomic observations, NDM was the predominant carbapenemase in our multicentre cohort. Moreover, NDM-producing isolates were recovered from participating hospitals in Beirut, the North, the South and the Bekaa, indicating their presence across geographically distinct Lebanese hospitals. While the present study was not designed as a formal national surveillance programme and regional sample sizes were modest, it provides an updated multicentre assessment supporting the growing importance of MBL within Lebanese healthcare settings. Contemporary global trends in carbapenemase epidemiology Recent international data indicate that the epidemiology of CRE is increasingly characterized by geographic heterogeneity and diversification of carbapenemase mechanisms rather than dominance of a single enzyme. Globally, a recent meta-analysis of carbapenem-resistant E. coli isolates collected between 2019 and 2023 reported a pooled prevalence of approximately 7%, with marked inter-country variability.25 NDM was the most frequently identified carbapenemase, highlighting the expanding global footprint of MBL-mediated resistance.25 In the USA, surveillance data from 2019 to 2023 demonstrated a 69% increase in carbapenemase-producing CRE incidence, driven largely by expansion of MBLs, particularly NDM, while KPC incidence remained relatively stable.26 By 2023, NDM accounted for an increasing proportion of carbapenem-resistant E. coli and Klebsiella spp., reflecting diversification of the historically KPC-dominated US epidemiology.26 Across Europe, KPC, OXA-48–like enzymes and NDM continue to co-circulate, with increasing reports of MBLs and isolates harbouring multiple carbapenemases.13 Similarly, surveillance data from Africa and the Middle East demonstrate predominance of OXA-48–like enzymes and NDM, with KPC remaining comparatively uncommon.27 More recent regional analyses from Red Sea basin countries have further confirmed sustained co-endemicity of OXA-48–like and NDM carbapenemases and increasing molecular diversity among CRE isolates.28 Within this context, the carbapenemase distribution observed in our study aligns more closely with Mediterranean and Eastern Mediterranean patterns characterized by co-circulation of OXA-48–like enzymes and NDM than with the historically KPC-dominated epidemiology of North America and parts of Europe.13,26–28 Therapeutic implications of the contemporary Lebanese CRE landscape The predominance of NDM-producing isolates in our cohort has important therapeutic implications. MBLs confer resistance to most β-lactams, including ceftazidime–avibactam, whereas OXA-48–like producers frequently remain susceptible to this agent.6,7 Accordingly, susceptibility to ceftazidime–avibactam was observed in 71.4% of OXA-48–like producers in our study but in none of the NDM-producing isolates. These findings highlight the limitations of managing CRE as a homogeneous group and underscore the importance of mechanism-guided therapy once carbapenem resistance has been identified. Beyond carbapenem resistance, many NDM-producing isolates demonstrated resistance to multiple antibiotic classes, reflecting the multidrug-resistant phenotype commonly associated with NDM-carrying plasmids. Nevertheless, a proportion of isolates retained susceptibility to aminoglycosides, particularly amikacin, emphasizing the continued importance of comprehensive susceptibility testing when selecting definitive therapy. As NDM-producing Enterobacterales become increasingly prevalent in Lebanon, therapeutic strategies will increasingly rely on agents active against MBL producers, including aztreonam-based combinations and cefiderocol, whereas ceftazidime–avibactam is likely to retain a greater role against OXA-48–like producers.1,6,7 Accurate carbapenemase detection is therefore central to mechanism-guided treatment. In our study, phenotypic prediction demonstrated high agreement with NG-Test® CARBA-5 results (κ = 0.89), with excellent sensitivity for MBL detection. While phenotypic screening reliably identified MBL activity, differentiation between OXA-48–like and MBL-producing isolates was less precise, resulting in misclassification of a small number of OXA-48–like producers. Although these isolates would likely have remained covered by therapeutic strategies directed against MBL-producing Enterobacterales, accurate differentiation remains clinically relevant because OXA-48–like and MBL-producing isolates differ substantially in their preferred targeted treatment options. Misclassification of OXA-48–like producers as MBLs may therefore result in unnecessary use of broader spectrum or less accessible agents, whereas correct identification may support the use of ceftazidime–avibactam when susceptibility is retained. Accordingly, rapid confirmatory testing may provide clinically relevant information beyond routine susceptibility testing alone by supporting optimization of definitive therapy, antimicrobial stewardship and epidemiological surveillance. Previous validation studies have demonstrated high sensitivity and specificity of NG-Test® CARBA-5 compared with molecular assays and broad detection of clinically relevant OXA-48–like variants, including OXA-181 and OXA-232.5,8,9 Nevertheless, reduced sensitivity has been reported in isolates harbouring multiple carbapenemases, and molecular characterization remains necessary for definitive confirmation.5 In addition, while NG-Test® CARBA-5 reliably identifies targeted carbapenemase families, it does not characterize accompanying resistance mechanisms such as ESBL production, AmpC hyperproduction or porin alterations, which may contribute to the overall resistance phenotype and influence antimicrobial susceptibility patterns. In settings where PCR and whole-genome sequencing are not routinely available, a two-step approach combining phenotypic screening with rapid immunochromatographic confirmation may represent a pragmatic balance between diagnostic accuracy, accessibility and cost. Beyond individual patient management, standardized carbapenemase testing may contribute to strengthening surveillance and antimicrobial stewardship efforts in Lebanon. Although the country adopted a National Action Plan on Antimicrobial Resistance in 2019, coordinated carbapenemase surveillance has not yet been fully implemented.29 Consequently, current understanding of carbapenemase epidemiology remains largely derived from investigator-driven studies and institutional reports rather than a unified surveillance framework. In this context, standardized carbapenemase detection may serve a dual purpose: supporting mechanism-guided clinical decision-making while generating comparable surveillance data across institutions. Such information could improve inter-hospital comparability, support formulary planning aligned with local epidemiology and enhance preparedness for the continued expansion of MBL-producing Enterobacterales within Lebanese healthcare settings. Strengths and limitations This study has several strengths. It provides an updated multicentre assessment of carbapenemase distribution among clinical CRE isolates collected from 14 hospitals across multiple Lebanese regions. Isolates were collected over several months and centrally processed using standardized microbiological methods interpreted according to the 2025 EUCAST criteria, ensuring methodological consistency and minimizing inter-laboratory variability. In addition, the study integrated carbapenemase characterization with antimicrobial susceptibility testing and evaluated agreement between phenotypic prediction and rapid immunochromatographic testing, generating clinically relevant data for mechanism-guided therapy and routine laboratory practice. Several limitations should be acknowledged. Although isolates were collected from multiple hospitals and regions, the study was not designed as a formal national surveillance programme and was not intended to estimate national incidence or prevalence. Regional sample sizes were modest, and geographic comparisons should therefore be interpreted cautiously. Carbapenemase characterization was limited to the enzyme families detectable by NG-Test® CARBA-5, and molecular characterization was not included in the present study. Consequently, additional resistance mechanisms, carbapenemase variants not targeted by the assay and clonal relationships between isolates could not be evaluated. The single carbapenemase-negative K. pneumoniae isolate exhibited an ertapenem resistant but imipenem-, meropenem- and ceftazidime–avibactam-susceptible phenotype, which is compatible with a non-carbapenemase–mediated resistance mechanism, potentially involving ESBL or AmpC activity combined with reduced outer-membrane permeability; however, molecular characterization was not available to confirm the underlying mechanism. Aztreonam–avibactam synergy testing was performed selectively in isolates with suspected MBL production, limiting the generalizability of these findings. Finally, classification according to inpatient or outpatient origin did not allow distinction between healthcare-associated and community-associated acquisition. Conclusions This multicentre study provides an updated assessment of carbapenemase distribution among clinical CRE isolates in Lebanon. NDM was the predominant carbapenemase, while OXA-48–like enzymes remained common and co-production was also observed. These findings highlight the importance of carbapenemase-specific characterization, as therapeutic options differ according to the underlying resistance mechanism. Rapid carbapenemase detection methods such as NG-Test® CARBA-5 may support resistance mechanism-guided therapy, antimicrobial stewardship and surveillance efforts. Supplementary Material