Multicenter Surveillance of Klebsiella pneumoniae: Resistance Shifts, CRKP Prevalence and Hypervirulent Evolution Trajectories in Hebei, China (2019-2023) Authors: Kaiyue Xu, Shiying Fan, Ziye Liu, Cuixin Qiang, Jiansheng Wang, Pu Qin, Ziyuan Mao, Ya Wu, Zhirong Li, Jianhong Zhao Published: 2026 ABSTRACT AbstractObjectiveTo describe the patient characteristics, resistance trends of Klebsiella pneumoniae (Kpn), and the prevalence of carbapenem-resistant Kpn (CRKP) in Hebei (2019–2023), and to explore genetic characteristics of CRKP underlying resistance shifts.MethodsWHONET 5.6 was employed to analyze patient characteristics of Kpn infections, resistance variations and CRKP prevalence using data from all 59 tertiary hospitals participating in Hebei Provincial Antimicrobial Resistance Surveillance System (HBARSS), 2019–2023. A total of 424 CRKP isolates from four hospitals across northern, central, and southern Hebei were subjected to antimicrobial susceptibility testing and whole-genome sequencing. Genomic analyses were performed using Kaptive, Kleborate, and PlasmidFinder.ResultsSurveillance data showed that a total of 114,166 non-duplicate Kpn isolates were identified, including 16,681 CRKP. Kpn infections were more common in males (63.4%) and mainly occurred in elderly (54.1%) and ICU patients (20.2%). Resistance rates to most antimicrobials declined overall, except for minocycline, tigecycline and polymyxin B, which fluctuated. CRKP prevalence fell from 15.3% (2019) to 12.8% (2023), averaging 14.6%. Analysis of 424 clinical CRKP showed that blaKPC-2 ST11-KL64 was the predominant type in both CRKP and hypervirulent CRKP (hv-CRKP), accounting for 40.1% and 69.1%, respectively. Notably, hv-CRKP among CRKP rose from 36.5% (2019) to 69.8% (2023), averaging 55.0%. Within ST11 hv-CRKP, 84.8% of KL64 isolates carried the iucABCDiutA+rmpA+rmpA2+ virulence profile, and 97.6% harbored IncHI1B or IncHI1B(pNDM-MAR)/repB plasmids. In contrast, 92.3% of KL47 isolates carried iucABCDiutA+rmpA−rmpA2+, and 89.7% lacked IncHI1B or IncHI1B(pNDM-MAR)/repB. Furthermore, the proportion of ST11-KL64 hv-CRKP among ST11 CRKP increased from 38.7% in 2019 to 49.2% in 2023, while that of ST11-KL47 rose from 0% to 20.0%.ConclusionMales, the elderly and ICU patients are susceptible populations. Despite declining resistance rate and CRKP prevalence in Hebei Province, China, hv-CRKP increased rapidly. And ST11-KL64 hv-CRKP and ST11-KL47 hv-CRKP showed different evolutionary patterns, with KL47 increasing more prominently.Keywords: Klebsiella pneumoniae, bacterial resistance surveillance, carbapenem resistance, hypervirulence, molecular epidemiology Introduction Klebsiella pneumonia (Kpn) is a common opportunistic pathogen that causes community-acquired and hospital-acquired infections worldwide, which can lead to organ damage and even life-threatening consequences. Data from the China Antimicrobial Resistance Surveillance System (CARSS) indicate that Kpn has become the second most commonly isolated bacterium in clinical settings in China.1 This pathogen can cause infections in multiple organs and sites, including lower respiratory tract infections, urinary tract infections, bloodstream infections, and even meningitis.2 Studies have shown that inappropriate initial antibiotic therapy increases patients’ mortality and accelerates the development of bacterial resistance.3,4 Given China’s vast territory and significant geographical variations, the antibiotic resistance rates of Kpn differ across regions5 A comprehensive understanding of the changing resistance patterns of Kpn in the local region is crucial for promoting the rational use of antibiotics and formulating effective anti-infection strategies within the region. Kpn can be classified into two pathogenic types: classical K. pneumonia (cKP) and hypervirulent K. pneumoniae (hvKP).2 Carbapenem-resistant Klebsiella pneumoniae (CRKP) is frequently associated with hospital-acquired infections and was ranked as a critical priority in the WHO bacterial priority pathogens list, 2024. According to the 2023 China Antimicrobial Surveillance Network (CHINET) data, the resistance rate of Kpn to meropenem has risen steadily from 2.9% in 2005 to 30.0% in 2023, and ST11-KL64 CRKP has been found to exhibit significant clonal spread in China.6,7 hvKP was first identified in Taiwan, China, in the 1980s. Although generally susceptible to antibiotics, it is more pathogenic and invasive than cKP due to its virulence plasmids carrying siderophore-encoding genes, rmpA/rmpA2 and other virulence genes.8 A global meta-analysis reported an overall prevalence of hvKP of 25.7%. The highest prevalence was observed in Africa (27.2%) and Asia (27.0%), whereas the Americas had the lowest (4.9%). In Europe, the prevalence reached 25.3% and showed an increasing trend.9 The global prevalence of hypervirulent CRKP (hv-CRKP) emerged around 2010, predominantly in Asia, particularly in China.10 The primary pathways for the formation of hv-CRKP include hvKP acquiring resistance plasmids, CRKP acquiring virulence plasmids, or Kpn acquiring a single plasmid carrying both virulence and carbapenem resistance genes.11–13 For instance, the acquisition of pK2044-like virulence plasmids by ST11-KL64 strains has been confirmed to correlate with enhanced virulence, higher mortality, and stronger transmissibility in clinical infections.12 IncFIB(K), IncHI1B, and IncHI1B(pNDM-MAR)/repB are common virulence plasmid types in hvKP, typically non-conjugative and unable to move autonomously.14–16 However, the widespread emergence of accessory conjugative plasmids encoding carbapenemases (eg blaKPC-2), coupled with the conserved presence of oriT in virulence plasmids, may facilitate the emergence and transmission of hv-CRKP.15,17–19 Conjugative IncFII plasmids carrying blaKPC-2, especially IncFII(pHN7A8), have become prevalent in ST11 CRKP strains in China.20,21 To date, epidemiological studies on Kpn in Hebei Province have been limited by time and region, lacking a comprehensive analysis of the changes in Kpn resistance and the underlying causes. To address these gaps, this study aimed to analyze the characteristics of Kpn-infected populations, changes in drug resistance rates of Kpn, and the prevalence of CRKP over the past five years in this region, thereby providing evidence for the rational clinical use of antimicrobial agents. Furthermore, based on next-generation sequencing (NGS), a comprehensive molecular epidemiological analysis was conducted on CRKP isolated from multiple centers to explore the genetic characteristics underlying the changes in Kpn resistance rate, thereby offering data support and a theoretical basis for the key prevention and control of CRKP and hv-CRKP. Materials and Methods Analysis of Antimicrobial Resistance Surveillance Data Surveillance data of Kpn isolates were analyzed using WHONET 5.6. Inclusion criteria: antimicrobial susceptibility data of Kpn isolates (the first isolate from each positive patient) and corresponding patient information (sex, age, and department of admission) collected from all 59 tertiary hospitals participating in the Hebei Antimicrobial Resistance Surveillance System (HBARSS) between 2019 and 2023. Exclusion criteria: (1) isolates with antimicrobial susceptibility results that could not be interpreted; (2) cases with incomplete clinical data; and (3) duplicate isolates from the same patient. The gender, age, and departmental distribution of infected patients, as well as the resistance rates of Kpn and the prevalence of CRKP, were analyzed. According to the criteria of CARSS, patients were divided into 4 age groups: neonates (≤28 days), children (29 days-14 years), adults (>14-65 years), and the elderly (>65 years).22 CRKP was defined as Kpn strains resistant to any of ertapenem, imipenem, or meropenem. Breakpoints were referenced to the Clinical and Laboratory Standards Institute (CLSI) M100-34th ed.23 Clinical Strain Collection and Identification A total of 424 CRKP isolates enrolled in this study were randomly selected from the HBARSS Strain Bank. These isolates were collected from 4 hospitals, which were chosen based on geographical location among 59 HBARSS member institutions. Specifically, one hospital is in northern Hebei, one in central Hebei, and two in southern Hebei. The isolates were revived from glycerol stocks and identified using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) (QuanTOF, Intelligene Biosystems Co., Ltd., Qingdao, China). Antimicrobial Susceptibility Test The susceptibility of 424 CRKP to nine antimicrobial agents was tested, including levofloxacin, gentamicin, amikacin, ertapenem, imipenem, meropenem, ceftazidime/avibactam (CZA), tigecycline, and polymyxin B. Antimicrobial susceptibility to levofloxacin, gentamicin, amikacin, imipenem, and meropenem was determined via the agar dilution method. For ertapenem and CZA, the Kirby-Bauer disc diffusion method was employed. Susceptibility test results were interpreted in accordance with the CLSI M100-34th ed.23 For tigecycline and polymyxin B, the Mueller-Hinton (M-H) broth microdilution method was adopted. The results for tigecycline were interpreted based on the standards of the U.S. Food and Drug Administration (FDA), while those for polymyxin B were determined according to the guidelines of the European Committee on Antimicrobial Susceptibility Testing (EUCAST).24 Quality control was performed for each batch of antimicrobial susceptibility testing to ensure the reliability of the results. According to CLSI M100-34th ed, Escherichia coli ATCC 25922 was used as the quality control strain for tigecycline and polymyxin B testing, E. coli ATCC 35218 for ceftazidime/avibactam testing, and Pseudomonas aeruginosa ATCC 27853 for levofloxacin, gentamicin, amikacin, ertapenem, imipenem, and meropenem testing.23 Bacterial DNA Extraction, NGS, and Bioinformatics Analysis Genomic DNA was extracted from strains cultured for 16–18h using the Bacterial Genomic DNA Extraction Kit (TIANLONG, Xi’an, Shanxi, China). The concentration and purity of DNA were determined with Qubit™ 4 Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA) and NanoDrop™ One Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), while DNA integrity was detected by 1% agarose gel electrophoresis. Sequencing libraries were constructed using the TruSeq DNA Sample Preparation Kit (Illumina, San Diego, CA, USA), and genomic sequencing was performed by Personalbio Technology Co., Ltd. (Shanghai, China). After removing adapter contamination with AdapterRemoval software and filtering data using SOAPec software, genomic data assembly was conducted.25,26 Unicycler v0.5.1 was used for assembly to obtain contigs. For the assembled contig sequences, Kaptive v3.1.0 was applied to predict KL and OL types of CRKP strains, and Kleborate v3.2.4 was used to analyze STs, drug-resistant genes and mutations, virulence genes and their lineages, as well as virulence scores.27,28 Plasmid replicons were predicted using PlasmidFinder v2.1.1. According to the definition by Margaret M. C. Lam et al, Kpn strains with a virulence score ≥3 (ie carrying at least the iucABCDiutA gene cluster) were defined as hvKP.28 Statistical Analysis and Visualization Intergroup comparisons were performed using the chi-square test or Fisher’s exact test, with Bonferroni correction applied for further pairwise comparisons The trend chi-square test was used to compare changes in Kpn resistance rates, CRKP prevalence, and drug-resistant genes across different years. A P-value <0.05 was considered statistically significant. The plotting tools used in this study were R language, GraphPad Prism 9.5, and the online plotting platform Chiplot [https://www.chiplot.online/]. Results Epidemiological and Antimicrobial Resistance Characteristics of Kpn Strains in Tertiary Hospitals of Hebei Province, 2019–2023 To investigate the clinical characteristics of patients with Kpn infections, trends in antimicrobial resistance, and the prevalence of CRKP in Hebei Province from 2019 to 2023, data were collected and filtered using predefined inclusion and exclusion criteria. A total of 114,166 Kpn isolates were obtained from 59 tertiary hospitals over the five-year period, of which 16,681 were CRKP. As shown in Table 1, Kpn strains were predominantly isolated from male patients (63.4%), with a male-to-female ratio of 1.7:1. Moreover, the proportion of CRKP among male patients infected with Kpn was significantly higher than that among female patients (χ2=20.489, P<0.001), accounting for 15.0% and 14.0%, respectively. In terms of age, elderly individuals formed the largest group of infected patients (54.1%). Although neonates had the lowest number of Kpn infections (0.8%), they exhibited the highest proportion of CRKP at 17.8%. Cases with Kpn or CRKP detection were mainly distributed in ICU, Neurosurgery Department, Respiratory Medicine Department, General Surgery Department, and Neurology Department. Notably, the ICU accounted for the highest proportion of Kpn isolates (20.2%); however, it did not exhibit the highest proportion of CRKP (29.0%). The Rehabilitation Department showed the highest CRKP detection rate, reaching 29.7%. Table 1General Data of Patients with Kpn and CRKP InfectionItemsNo. of Strains (Proportion %)CRKP Proportion (%)CRKPKpnGender Male10829 (64.9)72,333 (63.4)15.0 Female5852 (35.1)41,833 (36.6)14.0Age Group Neonates168 (1.0)944 (0.8)17.8 Children604 (3.6)4688 (4.1)12.9 Adults6287 (37.7)46,770 (41.0)13.4 The Elderly9622 (57.7)61,764 (54.1)15.6Department ICU6632 (39.8)22,884 (20.2)29.0 Neurosurgery2279 (13.7)12,035 (10.5)18.9 Respiratory Medicine2201 (13.2)13,368 (11.7)16.5 General Surgery603 (3.6)7031 (6.2)8.6 Neurology592 (3.5)6703 (5.9)8.8 Rehabilitation Medicine420 (2.5)1414 (1.2)29.7 Others3954 (23.7)50,731 (44.3)7.8 General Data of Patients with Kpn and CRKP Infection As shown in Table 2, over the past five years, Kpn exhibited low to moderate resistance rates (range: 1.1%-37.4%) to common antibiotics, except for cefazolin; the resistance rate to most antibiotics was below 30.0%. Tigecycline and polymyxin B exhibited excellent in vitro antibacterial activity, with resistance rates of only 1.1% and 2.1%, respectively. Over the 5-year period, except for fluctuating resistance rates to minocycline, polymyxin B, and tigecycline, the resistance rates of Kpn to other agents (eg β-lactams, aminoglycosides, trimethoprim/sulfamethoxazole) showed an overall decreasing trend (P<005), although the resistance rates to most agents increased slightly in 2020 or 2021 (Table 3 and Figure 1). Table 2Resistance Rates, MIC50 and MIC90 of Kpn in Hebei Province, China (2019–2023)Antimicrobial AgentsBreakpoints (μg/mL)MIC (μg/mL)No. of Strains (R%)MIC50MIC90PiperacillinS≤8 R≥32812848,907 (37.4)Amoxicillin/Clavulanic acidS≤8 R≥3283232,838 (22.9)Ampicillin/SulbactamS≤8 R≥3283291,494 (33.3)Ticarcillin/Clavulanic acidS≤16 R≥1281612825,233 (23.8)Piperacillin/TazobactamS≤8 R≥324128111,973 (18.0)Cefazolin (Urine)S≤16 R≥3264645171 (70.7)Cefazolin (non-Urine)S≤2 R≥886441,766 (53.6)CefuroximeS≤8 R≥3286486,848 (35.2)CeftazidimeS≤4 R≥16164110,469 (23.3)CeftriaxoneS≤1 R≥416487,573 (32.7)CefotaximeS≤1 R≥416422,910 (33.4)CefepimeS≤2 R≥16132111,129 (22.4)CefotetanS≤16 R≥6446440,919 (12.3)CefoxitinS≤8 R≥3283251,133 (19.9)AztreonamS≤4 R≥1616477,070 (26.1)ImipenemS≤1 R≥4116111,695 (12.1)MeropenemS≤1 R≥40.251697,320 (11.8)GentamicinS≤2 R≥811693,460 (20.3)TobramycinS≤2 R≥811664,680 (23.6)AmikacinS≤4 R≥162499,974 (8.7)CiprofloxacinS≤0.25 R≥10.25488,354 (34.5)LevofloxacinS≤0.50 R≥20.258100,759 (25.1)Trimethoprim/SulfamethoxazoleS≤2 R≥4116100,488 (27.4)Polymyxin BS≤2 R≥41217,220 (2.1)NitrofurantoinS≤32 R≥1286425630,019 (29.6)ChloramphenicolS≤8 R≥3283235,782 (22.0)TetracyclineS≤4 R≥1621612,969 (29.8)MinocyclineS≤4 R≥1641629,862 (14.3)TigecyclineS≤2 R≥81231,183 (1.1)Note: “R%” means resistance rate (%). Resistance Rates, MIC50 and MIC90 of Kpn in Hebei Province, China (2019–2023) Note: “R%” means resistance rate (%). Table 3Annual Antibiotic Resistance Rates of Kpn Isolated from Tertiary Hospitals in Hebei Province, China (2019–2023)Antimicrobial AgentsNo. of Strains (R%)χ2P-value20192020202120222023Piperacillin8794 (44.6)8459 (41.9)10,203 (37.2)10,906 (32.4)13,666 (34.4)366.114<0.001Amoxicillin/Clavulanic acid4192 (26.4)4506 (23.6)5663 (22.9)9144 (23.7)11,997 (22.4)19.344<0.001Ampicillin/Sulbactam14784 (36.6)15,230 (37.7)18,150 (34.0)20,930 (30.2)28,396 (31.7)241.927<0.001Ticarcillin/Clavulanic acid4397 (24.7)3922 (30.0)4527 (28.2)6855 (20.9)7551 (20.9)95.147<0.001Piperacillin/Tazobactam16317 (19.2)17,740 (21.4)21,753 (19.6)26,428 (17.1)37,240 (16.3)193.656<0.001Cefazolin (Urine)961 (77.4)981 (79.3)1162 (75.6)459 (55.8)1796 (62.5)129.928<0.001Cefazolin (non-Urine)8031 (63.3)7752 (62.7)8271 (55.6)4020 (43.9)15,702 (45.5)1047.515<0.001Cefuroxime12687 (39.3)13,498 (39.0)16,578 (36.1)20,769 (32.6)29,368 (33.7)221.944<0.001Ceftazidime16353 (25.6)17,597 (27.1)21,232 (24.9)25,863 (21.7)36,741 (21.4)263.876<0.001Ceftriaxone13363 (37.6)15,515 (37.8)17,870 (33.9)20,716 (29.4)26,199 (29.7)479.806<0.001Cefotaxime4514 (37.9)3086 (33.5)3567 (32.8)4623 (31.5)8517 (32.6)33.277<0.001Cefepime16732 (24.6)17,802 (25.8)20,822 (23.6)25,895 (21.1)37,230 (20.7)221.965<0.001Cefotetan6800 (14.1)7649 (16.5)8692 (13.5)9040 (9.5)11,290 (9.9)199.140<0.001Cefoxitin7732 (20.0)7108 (22.0)8724 (21.7)12,458 (19.5)18,777 (19.3)14.732<0.001Aztreonam11391 (29.2)12,066 (28.5)15,835 (27.0)18,018 (24.6)24,539 (25.0)115.576<0.001Imipenem16750 (12.9)17,829 (14.7)21,205 (13.9)26,375 (11.3)37,025 (10.6)178.048<0.001Meropenem15111 (13.1)15,443 (14.1)18,940 (13.9)22,784 (10.5)31,436 (10.3)192.251<0.001Gentamicin15489 (24.2)15,425 (24.0)18,473 (21.8)21,335 (16.9)28,873 (17.6)497.787<0.001Tobramycin9157 (26.4)10,165 (28.2)13,070 (25.6)15,179 (20.7)21,131 (21.2)247.540<0.001Amikacin15229 (9.1)16,423 (10.6)19,572 (9.8)24,171 (7.8)31,282 (7.7)14,615.457<0.001Ciprofloxacin14772 (37.0)15,328 (38.9)18,032 (37.2)19,352 (30.6)26,509 (32.1)279.366<0.001Levofloxacin15374 (27.3)16,646 (28.8)19,372 (27.3)22,827 (23.1)33,434 (23.0)254.626<0.001Trimethoprim/Sulfamethoxazole13717 (31.0)15,263 (31.1)18,752 (28.5)24,395 (25.6)35,153 (25.1)313.170<0.001Polymyxin B552 (0.2)985 (1.5)3150 (2.9)5682 (2.3)8534 (2.0)0.1860.666Nitrofurantoin5830 (30.2)5389 (31.4)5951 (33.8)5848 (29.0)8722 (26.5)40.322<0.001Chloramphenicol5509 (23.0)5488 (27.0)6722 (23.4)9028 (19.5)11,546 (20.2)79.898<0.001Tetracycline1814 (31.3)2047 (37.8)2894 (29.9)3176 (27.2)3750 (26.7)52.874<0.001Minocycline3055 (11.5)3925 (14.0)5136 (18.5)7503 (14.4)12,266 (13.4)0.2310.631Tigecycline2908 (0.8)3924 (1.4)5150 (1.2)7133 (1.3)14,085 (0.8)3.8290.050Notes: Chi-square values and P values were calculated via the trend chi-square test. “R%” means resistance rate (%). Annual Antibiotic Resistance Rates of Kpn Isolated from Tertiary Hospitals in Hebei Province, China (2019–2023) Notes: Chi-square values and P values were calculated via the trend chi-square test. “R%” means resistance rate (%). Trends in antimicrobial resistance of Kpn isolated from tertiary hospitals in Hebei Province, 2019–2023. (A) Resistance rates of Kpn to CZ (U), CZ (non-U), PB, MI and TGC. (B) Resistance rates of Kpn to GM, AMC, ATM, PIP, CXM, CRO, CTX and SXT. (C) Resistance rates of Kpn to TZP, CTT, FOX, IPM, MEM, AN and FEP. (D) Resistance rates of Kpn to SAM, CIP, FM, TE, TIM, CAZ, TM, LVX and C. Abbreviations: CZ (U), Cefazolin (Urine); CZ (non-U), Cefazolin (non-Urine); PB, Polymyxin B; MI, Minocycline; TGC, Tigecycline; GM, Gentamicin; AMC, Amoxicillin/Clavulanic acid; ATM, Aztreonam; PIP, Piperacillin; CXM, Cefuroxime; CRO, Ceftriaxone; CTX, Cefotaxime; SXT, Trimethoprim/Sulfamethoxazole; TZP, Piperacillin/Tazobactam; CTT, Cefotetan; FOX, Cefoxitin; IPM, Imipenem; MEM, Meropenem; AN, Amikacin; FEP, Cefepime; SAM, Ampicillin/Sulbactam; CIP, Ciprofloxacin; FM, Nitrofurantoin; TE, Tetracycline; TIM, Ticarcillin/Clavulanic acid; CAZ, Ceftazidime; TM, Tobramycin; LVX, Levofloxacin; C, Chloramphenicol. Over the recent five years, the prevalence of CRKP increased from 15.3% in 2019 to 17.0% in 2020, then decreased continuously to 12.8% in 2023, showing an overall decreasing trend (χ2=176.558, P<0.001) (Figure 2), which was consistent with the trend in drug resistance rates of Kpn to most antibiotics. Moreover, the prevalence of CRKP varied significantly across different regions of Hebei Province, with the lowest in Langfang (5.4%) and the highest in Qinhuangdao (25.5%); the provincial overall prevalence was 14.6%. Except for Qinhuangdao, the proportions of CRKP in Cangzhou (15.9%), Tangshan (18.3%) and Shijiazhuang (20.7%) exceeded the provincial average (Figure S1). Number and prevalence of CRKP isolates from tertiary hospitals in Hebei Province (2019–2023). Collection of Clinical CRKP Isolates To further investigate the causes of changes in Kpn resistance rates in Hebei Province, 424 clinically collected CRKP strains were subjected to antimicrobial susceptibility testing (AST) and NGS. Among these strains, 73 were isolated from a hospital located in the northern region, 88 from one in the central region, and 263 from two hospitals in the southern region (with 165 strains from one hospital and 98 from the other, respectively). The 424 CRKP strains were mainly isolated from clinical specimens including sputum (33.0%), blood (28.5%), and urine (14.4%) (Figure S2). Analysis of Genotypes and Their Epidemiology in 424 CRKP Isolates The results of STs, KL and OL of 424 CRKP are presented in Figure 3A. A total of 34 STs were identified, with ST11, ST15, and ST2237 being the most prevalent. Thirty KL types were detected, among which KL64, KL47, KL19, KL25, KL15, and KL102 were the dominant ones. Excluding three CRKP strains with “Untypeable” OL results (which were not included in the statistical analysis), seven OL types were detected. Among these, OL2α.1, OL13, OL2α.2, OL4, and OL5 were the most common. Genotype distribution and epidemiology of 424 CRKP isolates. (A) Sankey diagram of molecular typing for 424 CRKP isolates. (B) Prevalence of major STs (n > 10) from 2019 to 2023. (C) Prevalence of major KLs (n > 10) from 2019 to 2023. (D) Prevalence of major OLs (n > 10) from 2019 to 2023. As observed in Figure 3A, there was a strong correlation between ST, OL, and KL. The OL types of ST11 CRKP were mainly OL2α.1, OL13, OL4, and OL5, accounting for 58.0%, 26.9%, 7.1%, and 7.1%, respectively. OL2α.2 type was primarily derived from ST15 and ST2237. All KL64 strains belonged to OL2α.1. KL47 exhibited a significant correlation with OL13. All KL19 and KL102 strains were OL2α.2 type. Good consistency was observed within the OL5-KL25 and OL4-KL15 pairs. The prevalence trends of the main STs, KLs, and OLs among 424 CRKP strains from 2019 to 2023 are shown in Figure 3B–D. Among STs, the prevalence of ST11 decreased annually from 83.8% in 2019 to 60% in 2022, then showed an increasing trend in 2023 (75.6%), with an overall prevalence of 73.6%. The proportion of ST15 increased from 4.1% in 2019, peaked at 19.1% in 2021, and then decreased continuously thereafter. Although the prevalence of ST2237 was relatively low, it showed an annual increasing trend starting from 2021 (Figure 3B). KL64 remained the dominant KL type throughout the five-year period, with an overall increasing prevalence trend and an average prevalence of 41.5%. KL47 fluctuated within the range of 12.9%–24.3%, accounting for an average of 19.8%. The average prevalence of all other KL types was below 15% (Figure 3C). As the dominant OL type, OL2α.1 showed a generally steady upward trend in prevalence over the five years, increasing from 35.1% in 2019 to 48.8% in 2023. Due to its strong correlation with KL47, OL13 exhibited fluctuations similar to those of KL47 within the range of 17.1%–25.7%. The average prevalence of all other OL types was below 20% (Figure 3D). Antimicrobial Susceptibility Results of 424 CRKP Isolates AST results of 424 CRKP strains against nine antimicrobial agents are presented in Table 4. First, we compared AST results between the predominantly prevalent ST11 strains and strains of other STs. The results showed that there were statistically significant differences in resistance rates to five of the nine agents, except gentamicin, ertapenem, tigecycline and polymyxin B. Among these, ST11 strains had higher resistance rates to levofloxacin, amikacin, imipenem and meropenem than non-ST11 strains. In contrast, non-ST11 strains exhibited a higher resistance rate to CZA than ST11 strains. Additionally, when comparing resistance levels between hypervirulent strains (virulence score ≥3 evaluated by Kleborate) and non-hypervirulent strains, statistically significant differences in resistance rates were observed for levofloxacin, amikacin, imipenem, CZA and polymyxin B. Specifically, hypervirulent strains had higher resistance rates to levofloxacin, amikacin, imipenem, and polymyxin B than non-hypervirulent strains, whereas their resistance rate to CZA was lower than that of non-hypervirulent strains. Table 4Antimicrobial Susceptibility of 424 CRKP IsolatesAntibioticsCRKP (n=424)No. of Strains (R%)χ2P-valueNo. of strains (R%)χ2P-valueST11(n=312)non-ST11(n=112)hv CRKP(n=233)non-hv CRKP(n=191)Levofloxacin312(100)96(85.7)42.469<0.001230(98.7)178(93.2)8.8030.003Gentamicin262(84.0)92(82.1)0.2010.654195(83.7)159(83.2)0.0150.902Amikacin236(75.6)19(17.0)118.371<0.001183(78.5)72(37.7)73.046<0.001Ertapenem312(100)112(100)--233(100)191(100)--Imipenem309(99.0)101(90.2)17.582<0.001231(99.1)179(93.7)9.6730.002Meropenem311(99.7)107(95.5)7.3910.007231(99.1)187(97.9)0.4340.510Ceftazidime/Avibactam11(3.5)23(20.5)32.330<0.00112(5.2)22(11.5)5.7710.016Tigecycline22(7.1)11(9.8)0.8810.34821(9.0)12(6.3)1.0900.296Polymyxin B17(5.4)4(3.6)0.6170.43217(7.3)4(2.1)6.0330.014Notes: “R%” means resistance rate (%). “-” indicates not applicable. Antimicrobial Susceptibility of 424 CRKP Isolates Notes: “R%” means resistance rate (%). “-” indicates not applicable. Analysis of Drug Resistance Genes and Their Epidemiology in 424 CRKP Isolates Among the antimicrobial resistance (AMR) genes carried by CRKP, the prevalence of different genes varied significantly (Figure 4). For aminoglycoside resistance genes, aadA and rmtB were the most common, with 5-year average detection rates of 51.7% and 55.4%, respectively. The qnrS gene was the most common plasmid-mediated quinolone resistance gene, exhibiting an average prevalence of 40.8%. The chromosomal mutations associated with quinolone resistance were most frequently observed in ParC80I, GyrA83I, and GyrA87G, accounting for 92.2%, 77.8%, and 68.4%, respectively. tet(A) was the most common tetracycline resistance gene, present in 56.8% of the CRKP isolates. The dfrA and sul genes exhibited high prevalence, with rate at 59.4% and 59.0%, respectively. The predominant β-lactam resistance genes were blaKPC, blaCTX-M, and blaTEM-1D, found in 91.0%, 77.8%, and 61.1% of CRKP, respectively. Moreover, mutations in the porin genes OmpK35 and OmpK36 were detected at strikingly high rates of 81.6% and 82.3%, respectively. Heatmap of resistance genes or mutations in 424 CRKP isolates. Abbreviations: AGly, Aminoglycosides; Flq, Fluoroquinolones; Col, Colistin; Tet, tetracycline; Tgc, tigecycline; Fcyn, Fosfomycin; Phe, Phenicols; Rif, Rifampicin; Tmt, Trimethoprim; Sul, Sulfonamides. Furthermore, the prevalence of some acquired AMR genes showed statistically significant variations, namely blaDHA-1, blaSHV-12, blaCTX-M,blaTEM-1D, aph3-Ia, qnrB and dfrA. Except for three genes with low prevalence (blaDHA-1, blaSHV-12 and qnrB), the prevalence of blaCTX-M, blaTEM-1D, aph3-Ia and dfrA showed a decreasing trend starting from 2019 or 2020 (P<0.05). This was consistent with the changes in resistance rates to β-lactams, aminoglycosides and trimethoprim/sulfamethoxazole that we observed earlier in the analysis of Kpn resistance rates in Hebei. As shown in Table 5, among 424 CRKP strains, 403 carried carbapenemase genes, predominantly blaKPC (95.5%), followed by blaNDM (6.0%), with only 1 strain harboring blaIMP-4. Eight strains co-harbored both blaKPC and blaNDM. Among the 21 CRKP isolates negative for carbapenemase production, 15 carried genes encoding ESBL and/or AmpC enzymes and harbored concurrent porin mutations. blaKPC-2 ST11-KL64 was the most prevalent CRKP clone, accounting for 40.1%. blaNDM exhibited distinct distribution across different ST types: a minority (16.7%) was detected in ST11, while the vast majority (83.3%) was present in non-ST11 types. Additionally, we analyzed the prevalence of blaKPC and blaNDM over the five years. blaKPC remained the most prevalent carbapenemase gene, with a five-year prevalence of 91.0%. Its prevalence was the lowest in 2022 (82.9%) and peaked at 95.3% in 2023. The average detection rate of blaNDM was 5.7%, reaching a peak of 11.4% in 2022. (Figure S3). Table 5Distribution of Carbapenemase Genes Among 424 CRKP IsolatesSTKLCarbapenemase Genes (No. of Strains, Proportion%)ST11KL64KPC-2(170, 40.1)KPC-2+NDM-1(1, 0.2)KPC-2+NDM-7(1, 0.2)KPC-2+NDM-13(1, 0.2)KPC-12+NDM-5(1, 0.2)KL47KPC-2(83, 19.6)ND (1, 0.2)KL25KPC-2(29, 6.8)KL15KPC-2(22, 5.2)OtherKPC-2(3, 0.7)ST15KL19KPC-2(45, 10.6)ND (3, 0.7)OtherKPC-2(3, 0.7)NDM-1(1, 0.2)ST2237KL19KPC-2(10, 2.4)ND (1, 0.2)OtherOtherKPC-2(13, 3.1)NDM-5(8, 1.9)NDM-1(6, 1.4)KPC-2+NDM-1(2, 0.5)KPC-2+NDM-5(2, 0.5)NDM-6 (1, 0.2)IMP-4 (1, 0.2)ND (16, 3.8)Note: ST types with fewer than 10 strains and their corresponding KL types are grouped into the “Other” category.Abbreviation: ND, not detected. Distribution of Carbapenemase Genes Among 424 CRKP Isolates Note: ST types with fewer than 10 strains and their corresponding KL types are grouped into the “Other” category. Abbreviation: ND, not detected. Analysis of Virulence Genes and Their Epidemiology in 424 CRKP Isolates Kleborate virulence scoring results showed there were 191 (45.0%) non-hv-CRKP strains and 233 (55.0%) hv-CRKP strains. As presented in Table 6, among all 424 CRKP strains, the most common virulence gene was ybt (89.6%). Except for iucABCDiutA, the most prevalent virulence genes in hv-CRKP were ybt (97.9%) and rmpA2 (97.9%), while clb and iroBCDN were rare. ybt9, iuc1, and rmp1 were the most common virulence gene lineages. Table 6Virulence Genes and Their Lineages in KpnVirulence GenesLineageNo. of Strains (Proportion%)CRKP(n=424)hv-CRKP(n=233)ybtybt 11(0.2)1(0.4)ybt 21(0.2)1(0.4)ybt 41(0.2)NDybt 9311(73.3)212(91.0)ybt 104(0.9)NDybt 1461(14.4)13(5.6)ybt 201(0.2)1(0.4)ND44(10.4)5(2.1)iucABCDiutAiuc 1229(54.0)229(98.3)iuc 31(0.2)1(0.4)iuc 53(0.7)3(1.3)ND191(45.0)-clbclb 21(0.2)1(0.4)clb 31(0.2)1(0.4)ND422(99.5)231(99.1)iroBCDNiro 118(4.2)18(7.7)iro 31(0.2)1(0.4)ND405(95.5)214(91.8)rmpADCrmp1146(34.4)146(62.7)rmp31(0.2)1(0.4)rmp unknown4(0.9)4(1.7)ND273(64.4)82(35.2)rmpA2Detected228(53.8)228(97.9)ND196(46.2)5(2.1)Note: “-” indicates not applicable.Abbreviation: ND, not detected. Virulence Genes and Their Lineages in Kpn Note: “-” indicates not applicable. Abbreviation: ND, not detected. We further analyzed the prevalence of hv-CRKP. In 2019, non-hv-CRKP strains were more common than hv-CRKP, comprising 63.5% and 36.5% of isolates, respectively. After reaching comparable levels in 2020, the prevalence of hv-CRKP increased steadily, while the proportion of non-hv-CRKP declined. By 2023, the proportion of hv-CRKP had risen sharply to 69.8%, representing an increase of 33.3 percentage points over the five-year period, whereas the proportion of non-hv-CRKP fell to 30.2% (Figure 5A). Prevalence of hv-CRKP. (A) Prevalence of hv-CRKP among 424 CRKP. (B)Prevalence of hv-CRKP among ST11 CRKP. The blue line represents the prevalence of hv-CRKP among ST11 CRKP isolates, the yellow line represents the prevalence of ST11-KL64 hv-CRKP among ST11 CRKP isolates, and the green line represents the prevalence of ST11-KL47 hv-CRKP among ST11 CRKP isolates. Among the 233 hv-CRKP strains, ST11 was overwhelmingly dominant, accounting for 213 strains (91.4%), among which KL64 and KL47 were the most prevalent subtypes, followed by ST2237 (4.7%), ST15 (1.3%), and other ST types.blaKPC-2 ST11-KL64 was the most common hv-CRKP type (69.1%) (Table 7). Given that ST11 is the predominant ST type of hv-CRKP, the prevalence of hypervirulent strains among ST11 CRKP was further analyzed (Figure 5B). The proportion of ST11 hv-CRKP among ST11 CRKP increased annually, rising rapidly from 41.9% in 2019 to 81.0% in 2022, and remained stable at 80.0% in 2023. Specifically, the proportion of ST11-KL64 hv-CRKP among ST11 CRKP increased from 38.7% in 2019 to 66.7% in 2022, then rapidly decreased to 49.2% in 2023. In contrast, the proportion of ST11-KL47 hv-CRKP showed a upward trend over five years, further rising from 14.3% in 2022 to 20.0% in 2023. Furthermore, analysis of virulence gene profiles in ST11-KL64 and ST11-KL47 hv-CRKP strains revealed distinct patterns. The vast majority (84.8%) of ST11-KL64 hv-CRKP strains carried the virulence gene profile iucABCDiutA+rmpA+rmpA2+, whereas 92.3% of ST11-KL47 hv-CRKP strains exhibited the profile iucABCDiutA+rmpA−rmpA2+(Table 7). Table 7Molecular Features of Major Hv-CRKP IsolatesST (No. of Strains, Proportion%)KL (No. of Strains, Proportion %)CarbapenemaseGenes (Proportion%)ybt lineage (Proportion%)iuc lineage (Proportion %)rmpA Positive Rate%rmpA2 Positive Rate%Plasmid Profiles (Proportion %)ST11(213,91.4)KL64(165,70.8)KPC-2(69.1)KPC-2+NDM-1(0.4)KPC-2+NDM-7(0.4)KPC-2+NDM-13(0.4)KPC-12+NDM-5(0.4)ybt9(70.8)iuc1(70.8)60.170.8ColRNAI, IncFII(pHN7A8), IncHI1B, IncR(Profile 1)(30.5);ColRNAI, IncFII(pHN7A8), IncHI1B(pNDM-MAR)/repB, IncR (Profile 2)(18.5);Col156, ColRNAI, IncFII(pHN7A8), IncHI1B(pNDM-MAR)/repB, IncR (Profile 7)(7.7);Other(14.2)KL47(39,16.7)KPC-2(16.3)ND(0.4)ybt9(16.7)iuc1(16.7)1.316.7ColRNAI, IncFII, IncFII(pHN7A8), IncR(Profile 9)(6.4);ColRNAI, IncFII(pHN7A8), IncR(Profile 4)(4.7);Other(5.6)KL25(7,3.0)KPC-2(3.0)ybt9(3.0)iuc1(3.0)03.0ColRNAI, IncFII(pHN7A8), IncHI1B, IncR(Profile 1)(2.6);Other(0.4)KL19(2,0.9)KPC-2(0.9)ybt14(0.9)iuc1(0.9)0.90.9ColRNAI, IncFII(K), IncFII(pHN7A8), IncHI1B, IncQ1, IncR(0.9);ST2237(11,4.7)KL19(11,4.7)KPC-2(4.3),ND(0.4)ybt14(4.7)iuc1(4.7)04.7ColRNAI, IncFIB(K)(3.4);Other(1.3)ST15(3,1.3)KL102(3,1.3)KPC-2(1.3)ND(1.3)iuc5(1.3)00ColpVC, IncFIB(AP001918), IncR(1.3)Notes: STs with fewer than 3 isolates in hv-CRKP are not listed. The percentages in the table represent the proportions among all the 233 hv-CRKP strains.Abbreviation: ND, not detected. Molecular Features of Major Hv-CRKP Isolates Notes: STs with fewer than 3 isolates in hv-CRKP are not listed. The percentages in the table represent the proportions among all the 233 hv-CRKP strains. Abbreviation: ND, not detected. Analysis of Plasmid Replicons and Their Epidemiology in 424 CRKP Isolates Using PlasmidFinder, we analyzed plasmid replicons in CRKP and identified 83 plasmid profiles comprising 41 replicon types. Eleven major plasmid profiles (n≥10) are listed in Figure 6, encompassing 11 plasmid replicon types. Among these, Profile 1 (ColRNAI, IncFII(pHN7A8), IncHI1B, IncR), Profile 2 (ColRNAI, IncFII(pHN7A8), IncHI1B(pNDM-MAR)/repB, IncR) and Profile 3 (IncFIB(K), repB(R1701)) were the three most prevalent plasmid profiles; ColRNAI, IncFII(pHN7A8), and IncR were the three most prevalent plasmid replicon types. Notably, the distribution of different profiles differed distinctly between hv-CRKP and non-hv-CRKP (Figure 6). Profile 1, 2, 7, and 9 were predominantly found in hv-CRKP, while Profile 3, 5, 6, 8, 10, and 11 were primarily detected in non-hv-CRKP. And profile 4 was mainly detected in non-hv-CRKP and ST11-KL47 hv-CRKP strains. This observation prompted us to further investigate which plasmid replicon types were associated with hv-CRKP and non-hv-CRKP, respectively. The results showed that the detection rates of ColRNAI, IncFII(pHN7A8), IncR, IncHI1B, IncHI1B(pNDM-MAR)/repB, and Col156 in hv-CRKP were significantly higher than those in non-hv-CRKP (P<0.001). In contrast, the proportions of IncFIB(K), repB(R1701), and IncFII(K) in hv-CRKP were significantly lower than those in non-hv-CRKP (P<0.001), while there was no statistically significant difference in the proportion of IncFII between the two groups (Table 8). Furthermore, we observed that among the 104 CRKP isolates positive for the IncHI1B plasmid replicon, 103 were hv-CRKP and only one was non-hv-CRKP. All 76 isolates positive for IncHI1B(pNDM-MAR)/repB were hv-CRKP. Collectively, 76.8% of hv-CRKP isolates carried either IncHI1B or IncHI1B(pNDM-MAR)/repB plasmid. Interestingly, we also observed that the distribution of hv-CRKP plasmid profiles exhibited preferences among different KL types of ST11. Among ST11-KL64hv-CRKP, plasmid profiles 1, 2, and 7 were the most prevalent, while profiles 4 and 9 predominated in ST11-KL47 hv-CRKP (Table 7 and Figure 6). Further analysis revealed that 97.6% of ST11-KL64 hv-CRKP isolates carried either IncHI1B or IncHI1B(pNDM-MAR)/repB, with only 4 isolates negative for these replicons. In contrast, 89.7% of ST11-KL47 hv-CRKP isolates did not harbor IncHI1B or IncHI1B(pNDM-MAR)/repB. Upset plot of predominant plasmid profiles among 424 CRKP (n>10). Table 8Proportion of Major Plasmid Replicon Types in Hv-CRKP and Non-Hv-CRKP IsolatesPlasmid Replicon TypesNon-hv-CRKP(n=191)hv-CRKP(n=233)χ2P-valueNo.Proportion (%)No.Proportion (%)ColRNAI9549.7422496.14121.282<0.001IncFII(pHN7A8)10052.3621592.2787.573<0.001IncR4322.5120889.27193.657<0.001IncFIB(K)12163.35135.58162.061<0.001IncHI1B10.5210344.21108.190<0.001IncHI1B(pNDM-MAR)42.097632.6263.883<0.001repB00.007833.4878.354<0.001repB(R1701)7539.2720.86104.182<0.001IncFII(K)3719.3773.0030.233<0.001Col15600.002812.0224.576<0.001IncFII52.62156.443.4080.065Notes: Plasmid replicon types in plasmid profiles with more than 10 isolates are listed in the table. Proportion of Major Plasmid Replicon Types in Hv-CRKP and Non-Hv-CRKP Isolates Notes: Plasmid replicon types in plasmid profiles with more than 10 isolates are listed in the table. Given the annual increase in the prevalence of hv-CRKP, we further analyzed the 5-year prevalence of the dominant plasmid replicon types associated with hv-CRKP (Figure 7). ColRNAI, IncFII(pHN7A8), and IncR predominated with an overall upward trend. Col156 was prevalent almost exclusively between 2021 and 2022. The prevalence trends of IncHI1B and IncHI1B(pNDM-MAR)/repB increased alternately. Interestingly, in hv-CRKP strains, the repB and IncHI1B(pNDM-MAR) were almost invariably co-present, resulting in nearly identical prevalence between them. Prevalence trends of predominant plasmid replicons associated with hv-CRKP. Discussion The epidemiology and AMR profiles of Kpn provide critical references for clinicians to initiate appropriate empirical antibiotic therapy. The COVID-19 pandemic profoundly impacted healthcare systems, reshaping patient demographics, clinical practices, and antimicrobial utilization patterns.29,30 During and after this period, changes in AMR rates, genetic characteristics, and evolutionary trends of Kpn isolates in Hebei Province merit further investigation. In this study, using the HBARSS, we performed the first retrospective analysis of the epidemiological features of Kpn isolated from tertiary hospitals over the past five years, stratified by gender, age, department, and region, with a focus on temporal trends in Kpn resistance rates. Combined with molecular epidemiological analyses of clinically isolated strains, we explored the genetic characteristics and evolutionary dynamics of CRKP underlying the observed changes in resistance rates of Kpn. Our findings highlight the need for targeted Kpn surveillance in key populations, departments, and regions, and reveal a more concerning trend of persistent growth in hv-CRKP concurrent with the decline in overall Kpn resistance rates. In this study, more male patients with Kpn infection were observed than females, consistent with previous findings.31 Studies have shown sex hormones clearly contribute to sex differences in immune cell composition and function,32 which may explain the gender disparity in Kpn infection. Regarding age, elderly patients were more susceptible to Kpn, likely due to age-related immunosenescence, comorbidities, and frequent hospitalizations. Although the number of neonates with Kpn infection was relatively small, the proportion of CRKP was the highest. Further analysis revealed that 22.0% of CRKP-infected neonates were admitted to surgical departments such as cardiac surgery. Complex clinical conditions and therapeutic interventions may have increased the risk of CRKP infection. Moreover, the immature immune system of neonates and the presence of multidrug-resistant bacteria in healthcare settings further heightened their infection susceptibility. ICU, neurosurgery, and respiratory medicine were the main departments with Kpn and CRKP prevalence. Notably, despite a low number of Kpn isolates in rehabilitation medicine, the proportion of CRKP was the highest. This may be associated with the high frequency of hospitalizations and prolonged length of stay in rehabilitation patients, most of whom have comorbid cerebrovascular diseases or traumatic sequelae. Strengthened nosocomial infection control and strict hand hygiene practices among healthcare workers are warranted. Geographically, the prevalence of CRKP varied substantially across 11 cities, ranging from 5.4% to 25.5%. Hebei Province, located in northern China, features diverse landforms, uneven population density, and significant regional economic disparities, which may contribute to the variations in CRKP prevalence. Our findings emphasize the need to enhance active surveillance, antimicrobial stewardship, and infection control measures in key regions. To provide valuable data for clinical anti-infective practice, we focused on analyzing the resistance rates and temporal trends of Kpn to common antimicrobial agents over the past five years. Our data showed that Kpn resistance rates to all tested agents were below 30%, except for piperacillin, ampicillin/sulbactam, ciprofloxacin, and first-, second-, and third-generation cephalosporins. Among these, tigecycline and polymyxin B exhibited excellent in vitro antibacterial activity. Regarding trends in drug resistance rates, except for fluctuating resistance to minocycline, tigecycline, and polymyxin B, Kpn showed an overall downward trend in resistance to other common antimicrobial agents, which differed from the trend observed in a tertiary hospital in southwest China during the same period.31 Geographical variations in antimicrobial resistance may be influenced by local antibiotic utilization patterns, healthcare quality, and public health policies; thus, local epidemiological data should be referenced when initiating empirical anti-infective therapy. Furthermore, although a declining trend in the resistance rate of Kpn was observed in this region, the global situation remains far from optimistic. Data from South Africa indicate that carbapenem resistance has increased rapidly, with an estimated annual rise of 8.3% from 2012 to 2024, exceeding 30%. Resistance rates to β-lactam antibiotics, such as amoxicillin and second-generation cephalosporins, reached as high as 69.3% and 70.9%, respectively, without showing any downward trend.33 A study conducted in a Romanian hospital reported that the overall carbapenem resistance rate of Kpn was 44.7% between 2017 and 2023. Resistance to β-lactams, trimethoprim–sulfamethoxazole, and polymyxins demonstrated an overall increasing trend, with polymyxin resistance reaching a striking 28% in 2023.34 Surveillance data from the WHO further revealed that 40% of pathogen–antibiotic combinations showed a statistically significant increase in resistance between 2018 and 2023, with the Kpn–imipenem combination exhibiting the highest annual relative increase (15.3%).35 Collectively, the observed decline in Kpn resistance in this region appears to be a localized phenomenon, whereas the global situation continues to worsen, with pronounced geographic disparities. Antimicrobial resistance in Kpn therefore remains a critical global public health challenge. Consistent with the declining Kpn resistance rates, the prevalence of CRKP also showed a corresponding decrease. Accordingly, we further analyzed the molecular epidemiological characteristics of CRKP in the region, exploring the genetic features and evolutionary dynamics of CRKP underlying the changes in Kpn resistance profiles. A comprehensive molecular epidemiological analysis was performed on 424 clinically collected CRKP strains. As the most prevalent CRKP sequence type in China,36 ST11 accounted for 73.6% of the isolates in this study. ST15 was the second dominant ST and showed a consistent upward trend during the declining prevalence of ST11 from 2019 to 2021. ST15 has been detected in CRKP populations across multiple regions of China, ranking second only to ST11.37,38 Although ST2237 is not a stably prevalent ST compared with ST11 and ST15, it has shown a continuous upward trend since first detected in 2021, with a prevalence rate equal to that of ST15 in 2023. Genetic analysis revealed that all 11 ST2237 strains were hv-CRKP of the KL19:OL2α.2 type, harboring highly similar AMR gene profiles and virulence gene profile (ybt+iucABCDiutA+rmpA2+). These strains were isolated from two hospitals in southern Hebei, suggesting clonal transmission within the region. Currently, ST2237 has been detected in a few regions of China, with no evidence of nationwide prevalence.37,39 Notably, ST2237 strains from Henan Province, located in the southern part of Hebei Province, share the same genetic background (KL19:OL2α.2) and virulence gene profile (ybt+iucABCDiutA+rmpA2+) as those in this study,39 indicating the interregional transmission of this hv-CRKP clone. Serotyping of Kpn is critical for monitoring serotype variations and informing vaccine development strategies. To date, more than ten O serotypes of Kpn have been described,40 whereas the K antigen is highly variable, with at least 79 distinct structures identified to date.41,42 The considerable diversity of these antigens presents a major challenge for achieving broad vaccine coverage, underscoring the importance of ongoing seroepidemiological surveillance of clinical isolates to support effective vaccine design. Kai Zhou et al first documented the serotype shift in ST11 CRKP during 2013–2017, where ST11-KL64 replaced ST11-KL47 as the dominant KL type.43 Our data indicate that KL64 CRKP had an average prevalence of 41.5% and exhibited an overall upward trend over the past five years. In contrast to its declining prevalence or even disappearance in most parts of China,6,44,45 ST11-KL47 CRKP in this study showed fluctuating prevalence at around 20%, suggesting its persistent transmission advantage in this region. Traditional serotyping of Kpn by agglutination is inefficient and has limited resolution. Molecular methods provide a more effective and convenient approach for inferring K and O antigen types. The newly released Kaptive v3.1.0 (April 2025) refines O-type classification by introducing a standardized nomenclature system based on biochemical structures, significantly improving typing accuracy. Kaptive defines OL types according to the LPS biosynthesis gene cluster and associated extra genes. According to the latest typing system, OL2α.1 (in this study, it corresponds to O1/O2v1 in the old system) was the most prevalent OL type, showing a steady increase over the past five years. OL2α.2 (equivalent to O1/O2v2 in this study) was less common, ranking third with an average proportion of 19.6%. Additionally, we detected 91 (21.5%) OL13 CRKP strains, making it the second most prevalent type. This finding is noteworthy as reports on this type remain scarce. Daria Artyszuk et al identified the O-antigen structure of four strains predicted by Kaptive as OL101 as β-Kdop-[→3)-α-L-Rhap-(1→4)-α-D-Glcp-(1→]n, corresponding to the new OL13 serotype46 This result has been incorporated into the updated Kaptive database. When analyzing these ninety-one OL13-type CRKP strains using Kaptive versions prior to v3.1.0, only six were correctly identified as OL13, while 82 were predicted as “unknown(OL101)” and three as “OL101”. This suggests that studies using older Kaptive versions likely underestimated the prevalence of OL13-type Kpn. As one of the most prevalent KL types in ST11, KL47 has been previously reported to exhibit a strong correlation with OL101.6,20,47 This taxonomic change warrants attention regarding its implications Our study identified specific associations between KL and OL types (eg ST11:KL64:OL2α.1, ST11:KL47:OL13, ST15:KL19:OL2α.2, KL15:OL4, KL25:OL5), which is consistent with previous reports.6,48 These results offer a valuable perspective for the development of vaccines based on K and O antigen. We compared antimicrobial susceptibility between ST11 and non-ST11 CRKP, as well as between hv-CRKP and non-hv-CRKP strains, among 424 isolates to determine whether the epidemiologically dominant ST11 CRKP and hv-CRKP also exhibit advantages in resistance phenotypes. ST11 exhibited significantly higher resistance rate to levofloxacin, amikacin, imipenem and meropenem compared with non-ST11 strains. In contrast, resistance rate to CZA was lower in ST11, which is consistent with literature.49,50 Avibactam inhibits class A carbapenemases and some class D carbapenemases but is ineffective against class B metallo-β-lactamases (MBLs). In this study, the carriage of blaNDM differed between ST11 and non-ST11 strains, with the vast majority of this gene detected in non-ST11 strains. Similarly, hv-CRKP showed higher resistance rate to levofloxacin, amikacin, imipenem and polymyxin B, whereas non-hv-CRKP was less susceptible to CZA.51 These findings suggest that ST11 CRKP and hv-CRKP have acquired remarkable resistance advantages during evolution. Additionally, as first-line combination agents for treating CRKP infections, tigecycline and polymyxin B exhibited excellent in vitro antibacterial activity against CRKP, with susceptibility rates of 92.2% and 95.0%, respectively. Global concerns are growing over the spread of multidrug-resistant hvKP, as infections have been reported in many countries. Notably, a prospective study revealed that CRKP isolates from China carried virulence genes at a much higher prevalence than those from other parts of the world.52 To understand the virulence profile of CRKP in Hebei, we analyzed 424 CRKP isolates using Kleborate for key virulence genes (clb, ybt, iucABCDiutA, iro, rmpA, and rmpA2) and their lineages. Results showed that among siderophore genes, the ybt had the highest positivity rate, followed by iucABCDiutA, while iro exhibited the lowest carriage rate. The ybt is often present in the chromosome of Kpn in addition to being carried by virulence plasmids, accounting for its higher prevalence compared to iucABCDiutA and iro. However, despite both iucABCDiutA and iro typically being carried by classic virulence plasmids, their prevalence differs substantially. Studies have indicated that iroBCDN deletion does not attenuate hv-CRKP virulence but may facilitate the transfer of non-conjugative virulence plasmids into CRKP and enhance macrophage survival, providing an evolutionary edge.53 rmpA and rmpA2 are associated with enhanced capsule production and virulence.2 We detected rmpA2 more often than rmpA, but most rmpA2 genes were mutated or truncated. Compared to isolates from Ningbo, China,54 the hv-CRKP isolates in our study exhibited higher frequencies of both rmpA (64.8%) and rmpA2 (97.9%). Despite the declining trends in both Kpn resistance rates and CRKP prevalence in Hebei Province in recent years, virulence genes have continued to accumulate in CRKP. The proportion of hv-CRKP among CRKP isolates increased steadily year by year, reaching a high prevalence of 69.8% in 2023, which represented a total increase of 33.3 percentage points over the five-year period. A nationwide multicenter study showed that the prevalence of hv-CRKP rose from 28.2% in 2016 to 45.7% in 2020.6 However, significant regional differences were observed, with Zhejiang, Jiangsu, and Beijing identified as high-prevalence areas.2 Beyond China, a global upward trend in hv-CRKP has also been observed. GenBank data indicate that hv-CRKP expanded rapidly across multiple STs and continents during the COVID-19 pandemic, with the proportion of ST11 hv-CRKP in Asia increasing from 45.7% (2015–2019) to 53.4% (post-2020).55 Studies have shown that ST11 is the predominant clone of hv-CRKP worldwide. In Asia, particularly in China, it exhibits a clear dominance and is mainly associated with KL64 and KL47 capsular types, which is consistent with our findings.2,56 ST11 belongs to the classical CRKP lineage and can evolve into hv-CRKP through the acquisition of virulence plasmids carrying genes such as rmpA/rmpA2 and siderophore systems, followed by rapid dissemination via clonal expansion.12 In addition, ST23 and ST15 have also been identified among hv-CRKP isolates in China, although their prevalence is much lower than that of ST11.2 In contrast to the high prevalence of hv-CRKP in Asia, regions such as Europe, the Americas, and Africa generally exhibit a sporadic and multiclonal distribution pattern. In these regions, classical hypervirulent clones, including ST23, ST65, and ST86, predominate, and their evolution into hv-CRKP is mainly driven by the acquisition of carbapenem resistance genes.56 These regional differences suggest that the global spread of hv-CRKP is not the result of dissemination of a single clone, but rather arises from independent convergence events between resistance and virulence in strains with diverse genetic backgrounds across different regions. With increasing international transmission, however, the boundaries between these patterns may become increasingly blurred, and cross-transmission and genetic recombination among different lineages are likely to further complicate the molecular epidemiology of hv-CRKP. Among 233 hv-CRKP isolates, ST11 accounted for 91.4%. As the predominant hv-CRKP clone, the proportion of ST11 hv-CRKP among ST11 CRKP nearly doubled over five years (from 41.9% to 80.0%). It is an indisputable fact that ST11-KL64 dominates hv-CRKP in China.2 However, few studies have focused on changes in its prevalence. Our study demonstrated that although the prevalence of ST11-KL64 hv-CRKP among ST11 CRKP isolates rose steadily from 38.7% in 2019 to 66.7% in 2022, it declined sharply to 49.2% in 2023, with an overall increase of 10.5 percentage points in its five-year prevalence. In contrast, the prevalence of ST11-KL47 hv-CRKP surged from 0% in 2019 to 14.1% in 2020, remained stable over the subsequent two years, and further increased to 20.0% in 2023, representing a cumulative rise of 20 percentage points over the five-year period. Studies have shown that the rmpA is barely detectable in ST11-KL47.6,21,47 In this study, the vast majority of ST11-KL64 hv-CRKP harbored rmpA, whereas only a small subset of ST11-KL47 hv-CRKP tested positive for this gene. The virulence gene carriage pattern of ST11-KL47 hv-CRKP (iucABCDiutA+rmpA-rmpA2+) differs from that of ST11-KL64 hv-CRKP (iucABCDiutA+rmpA+rmpA2+), which may indicate that there are differences in their virulence evolution. hvKP and CRKP emerged as distinct lineages in the early stages of Kpn evolution. However, with the widespread dissemination of mobile genetic elements (MGEs) in Kpn, the boundary between these two pathogenic types has become increasingly blurred. Studies have shown that 83.0% of virulence genes in CRKP are located on conjugative or mobilizable plasmids, among which mobilizable plasmids are the primary carriers of acquired virulence genes in Kpn.17 Therefore, this study focused on the plasmid profiles of hv-CRKP in the region to enable targeted surveillance of high-risk clones. Our data showed that the two most prevalent plasmid profiles (Profile 1: ColRNAI, IncFII(pHN7A8), IncHI1B, IncR; Profile 2: ColRNAI, IncFII(pHN7A8), IncHI1B(pNDM-MAR)/repB, IncR) were almost exclusively derived from hv-CRKP, further reflecting the dominant prevalence of hv-CRKP in this region. In contrast to our findings, a study from Neijiang, Sichuan Province, China, reported that the most prevalent plasmid profile among local hv-CRKP was Profile 2.20 Among the 11 predominant plasmid profiles, all exhibited a highly distinct distribution preference except for one exception (Profile 4): Profiles 1, 2, 7, and 9 were almost exclusively detected in hv-CRKP, while Profiles 3, 5, 6, 8, 10, and 11 were nearly restricted to non-hv-CRKP. We hypothesize that this phenomenon is attributed to the presence of virulence plasmids in hv-CRKP. Further analysis revealed that IncHI1B and IncHI1B(pNDM-MAR)/repB were almost exclusively identified in hv-CRKP. Many studies have demonstrated that IncHI1B and IncHI1B(pNDM-MAR)/repB are common virulence plasmid types,14,57–59 indicating that they are the predominant virulence plasmid types in this region. In addition to IncHI1B and IncHI1B(pNDM-MAR)/repB, IncFIB(K) is also a frequently reported virulence plasmid replicon type,16,45 and even the most prevalent one.60 However, in our study, IncFIB(K) was predominantly detected in non-hv-CRKP (90.3%). Furthermore, a study conducted in East China indicated that the most prevalent virulence gene profile—rmpA/rmpA2+iuc1, a finding consistent with our research—was primarily associated with IncFIB/IncHI1B and IncFIB/IncHI1B/IncR replicons, which differed from our results.61 Together, these results point to a region-specific virulence evolution of hv-CRKP in this province. In addition, unlike repB, repB(R1701) was mainly present in non-hv-CRKP (97.4%), which is consistent with the findings of Yi Yuan et al.20 The high prevalence of IncFII(pHN7A8), ColRNAI, and IncR in hv-CRKP drew our attention. Statistical analysis revealed that their detection rates in hv-CRKP were significantly higher than those in non-hv-CRKP. As virulence plasmids, IncHI1B and IncHI1B(pNDM-MAR)/repB are generally non-conjugative due to the lack of a type IV secretion system.14,15 However, their conserved oriT regions enable mobilization: when coexisting with conjugative plasmids, they recruit conjugation-essential components encoded by the latter to initiate the conjugation process and facilitate their transfer.15,17–19 Studies have shown that the combined action of mobile virulence plasmids and IncF-type conjugative plasmids promotes the accumulation of AMR genes and acquired virulence genes in Kpn.15,17 Additionally, CC258 clonal group, such as ST11 and ST258, naturally lack the type I-E CRISPR-Cas system, enabling unimpeded invasion and stable persistence of IncF plasmids in CC258 strains,62 which facilitates the role of IncF-type conjugative plasmids in the formation of ST11 hv-CRKP. Our data showed that the prevalence of IncFII(pHN7A8) exhibited an upward trend, consistent with that of hv-CRKP. Since ColRNAI and IncR are typically non-conjugative plasmid replicons,63,64 we hypothesize that the conjugative plasmid IncFII(pHN7A8) may play a crucial role in driving the prevalence of hv-CRKP in this region. As noted in our aforementioned analysis, ST11-KL64 and ST11-KL47 hv-CRKP exhibit distinct virulence gene carriage patterns and epidemiological trends. Herein, we found that plasmid profiles 1, 2, and 7—harboring IncHI1B or IncHI1B(pNDM-MAR)/repB—were highly prevalent in ST11-KL64 hv-CRKP, whereas plasmid profiles 4 and 9, which lack IncHI1B and IncHI1B(pNDM-MAR)/repB, predominated in ST11-KL47 hv-CRKP. These results indicate significant epidemiological differences in hv-CRKP plasmid profiles across different ST11 KL types, suggesting that the two clones may differ in virulence plasmids and follow distinct pathways of virulence evolution, which may be associated with their divergent epidemiological trends. Fupin Hu et al reported that most KL64 hv-CRKP isolates are associated with the acquisition of pST11-Vir-1-KPC, while most KL47 hv-CRKP isolates are linked to two other plasmids (pST11-Vir-2 and pST11-Vir-3) that only carry the iucA and rmpA2 genes.6 We acknowledge several limitations in this study. First, the retrospectively collected CRKP isolates were only from four tertiary hospitals across three cities; inclusion of isolates from more regions would provide a more comprehensive reflection of the genetic characteristics and evolutionary trends of CRKP in Hebei. Nevertheless, this remains the most comprehensive multicenter five-year longitudinal study on CRKP in the province to date. Second, virulence genes in CRKP isolates were predicted via bioinformatics methods, which do not guarantee a hypervirulent phenotype. Animal infection experiments would help clarify the virulence levels of these potentially hypervirulent strains. Nevertheless, based on systematic five-year bacterial resistance surveillance data and clinical CRKP isolates collected from multiple centers, this study clearly illustrates the evolutionary trend of CRKP toward hypervirulence in this region, thus furnishing critical baseline evidence and reliable data support for subsequent prevention, control, and in-depth research. Conclusion In summary, our data indicate that male, elderly patients, and ICU patients are susceptible populations for Kpn infection. Over the past five years, Kpn AMR rates and CRKP prevalence have generally shown a decreasing trend in Hebei Province, China. However, hv-CRKP has rapidly emerged among CRKP isolates. As the predominant hv-CRKP clones, ST11-KL64 and ST11-KL47 display distinct epidemiological trends, virulence gene profiles, and plasmid profiles, indicating divergent virulence evolutionary pathways. The growing trend of hypervirulence among CRKP isolates in this region underscores the imperative of maintaining heightened vigilance against their clinical dissemination. Acknowledgments We are grateful to Dixia Hu and Xinyu Wang for their valuable assistance in the collection of clinical strains. We would also like to thank Jing Yang, Yanan Niu, and Weigang Wang for providing professional technical guidance during the experimental operations, as well as Mingzhi Guo for his help in data compilation and organization. We are grateful for the contributions of the expert committees, working groups, and member organizations of the Hebei Provincial Antimicrobial Resistance Surveillance System. We would like to acknowledge the assistance of Doubao Generative AI Tools v1.53.7 in the preparation of this manuscript, specifically for the grammatical polishing of a small number of non-critical sentences. Notably, this tool was not involved in any core aspects of the study, including study design, ethical approval application, data collection, data analysis, result interpretation, and drafting of key scientific conclusions. All authors have reviewed, edited, and approved the final version of the manuscript and take full responsibility for its scientific validity, accuracy and integrity. Funding Statement This study was supported by the Medical Science Research Project of Hebei(No. 20250855), Hebei Natural Science Foundation (No.H2025206149, H2025206707), College Students’ Innovative Entrepreneurial Training Plan Program Grant (No. USIP2025023, USIP2024118), Foundation of Hebei Provincial Department of Finance (No.ZF2025084). Data Sharing Statement Aggregated, de-identified data underlying the results are available from the corresponding author, Dr. Jianhong Zhao (zhaojh_2002@hebmu.edu.cn), upon reasonable request and subject to institutional/ethical approvals and a data use agreement. The Whole Genome Shotgun sequencing project for 424 strains of CRKP have been deposited in the GenBank databases under the BioProject ID: PRJNA1390759. Ethics Statement This study was approved by the Ethics Committee of Hebei Provincial Hospital of Chinese Medicine (Approval No. HBZY2024-KY-127-01). Given that the study was based on anonymized retrospective data, the requirement for informed consent was waived by the ethics committee. All procedures involving data processing complied with the ethical principles of the Declaration of Helsinki and relevant institutional regulations. Author Contributions All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work. Disclosure The authors report no conflicts of interest in this work. References Associated Data Data Availability Statement Aggregated, de-identified data underlying the results are available from the corresponding author, Dr. Jianhong Zhao (zhaojh_2002@hebmu.edu.cn), upon reasonable request and subject to institutional/ethical approvals and a data use agreement. The Whole Genome Shotgun sequencing project for 424 strains of CRKP have been deposited in the GenBank databases under the BioProject ID: PRJNA1390759.