In recent decades, antibiotic-resistant bacteria have become an urgent public health problem. This fact has increased attention to the extended-spectrum β-lactamase (ESBLs) group of enzymes in strains of the order Enterobacterales. These enzymes can be acquired via plasmid transfer, favoring the resistance of their host strains to β-lactam antibiotics and, in particular, the third-generation cephalosporins [1,2,3]. ESBLs are a diverse group of β-lactamases, and their classification is based on their substrate profile and genetic relatedness. The most common types of ESBLs include CTX-M enzymes, as well as some variants of TEM and SHV. The CTX-M enzymes, in particular, have emerged as the most prevalent in recent years, with the CTX-M-15 variant being widely distributed [4]. ESBLs confer resistance by hydrolyzing third-generation cephalosporins but generally do not affect carbapenems [5].

Strains producing ESBLs together with other mechanisms of resistance for non-beta-lactams have increased, what facilitate he use carbapenems for treatment. However, the emergence of carbapenemases remains a significant public health concern [6,7]. ESBL-encoding genes are often localized on mobile genetic elements, such as plasmids, facilitating horizontal gene transfer between bacterial genera and species [8]. This ability to spread across diverse bacterial species and genera complicates treatment and increases the likelihood of outbreaks. The transfer of these genes can occur within a single bacterial species and between Enterobacterales and non-Enterobacterales members, such as Pseudomonas or Acinetobacter, further exacerbating the challenge of multidrug resistance [9,10].

According to the 2023 annual epidemiological report from the European Centre for Disease Prevention and Control (ECDC) in Portugal, 50.7% of invasive Klebsiella pneumoniae isolates and 17.2% of Escherichia coli isolates were resistant to third-generation cephalosporins, while 13.1% of K. pneumoniae and 0.5% of E. coli were resistant to carbapenems, reaching higher values since 2006 for K. pneumoniae and 2001 for E. coli (https://atlas.ecdc.europa.eu/; accessed on 14 January 2025) [11]. The emergence and subsequent surveillance of strains producing ESBLs are necessary for good hospital management and patient therapy, particularly for those with a clinical diagnosis of septicemia.

Klebsiella pneumoniae is one of the species that have been most frequently isolated from clinical specimens and is a cause of sepsis, meningitis, and bacteremia. Many K. pneumoniae strains are resistant to a broad set of antimicrobial drugs, and their capacity to produce ESBLs is an essential mechanism of bacterial resistance, which increases in both the hospital and community environment [12,13,14].

Two previous studies have been conducted in a hospital located in North Portugal, which showed the predominance of CTX-M-15 among ESBL-producing K. pneumoniae isolates alongside SHV-type β-lactamases like SHV-27 and SHV-12 in isolates obtained during 2016–2018 from different origins, including bloodstream infections (BSIs) [15,16]. Both studies reported carbapenem-resistant isolates, and some K. pneumoniae strains showed a cooccurrence of ESBL and carbapenemase-encoding genes, highlighting the alarming levels of multidrug resistance within this species [15,16]. A recent study was also performed in the same hospital with K. pneumoniae isolates recovered from different origins during a nine-month period in 2021–2022, analyzing their content in β-lactamases by PCR [17]. The present study is focused on the characterization by whole-genome sequencing (WGS) of a collection of 14 ESBL-producing K. pneumoniae isolates recovered from BSIs during 2021–2022 in the same hospital in order to deepen in the molecular characterization of K. pneumoniae isolates that circulate in this hospital in invasive infections.

The purpose of the present study was to analyze the WGS of the ESBL-producing K. pneumoniae isolates recovered at the hospital level (hospital center of Trás-os-Montes and Alto Douro) of BSIs during a period of nine months. K. pneumoniae is frequently resistant to a wide range of antibiotics, including different β-lactams by β-lactamases [21]. In detail, infections caused by ESBL-producing Enterobacterales represent a worldwide issue concerning public health, especially given their association with poor outcomes, growing community-onset, and high ecological treatment costs [22]. Although the ESBL family is heterogeneous, the global pandemic of plasmids carrying CTX-M-type genes, which started mainly in the 2000s, is the primary driver of ESBL dissemination and has replaced other ESBL enzymes [23].

For this study, 14 ESBL-producing K. pneumoniae isolates recovered from BSIs from a Portuguese hospital were analyzed. One previous study was conducted on K. pneumoniae isolates of BSIs in the same hospital, but the study was performed four years before [15], and another one was conducted in isolates from the same period as the present study [17]. However, neither of these studies included the characterization of isolates by WGS. Concerning K. pneumoniae BSIs and the significant impact of the latter infections in the clinical setting drove our interest in studying these infections. BSIs due to K. pneumoniae are associated with higher mortality rates than infections at other body sites [24]. Consequently, filling this gap in the literature is imperative to improve patient outcomes and provide evidence-based treatment options for such serious infections. A study by Gorrie et al. conducted a year-long prospective surveillance study of K. pneumoniae clinical isolates in hospital patients and reported that 10% of patients had K. pneumoniae isolated from disseminated infections (bloodstream and/or cerebral spinal fluid), a similar percentage to ours [25]. This consistency across studies suggests that a range of infections caused by K. pneumoniae progress to BSIs, suggesting that bacterial or host factors may contribute to the potential to disseminate K. pneumoniae.

Regarding ESBL production, the reported ESBL production rates in BSIs caused by Enterobacterales are highly variable. The European Centre for Disease Prevention and Control (ECDC), in its 2023 annual epidemiological report, revealed that, in Portugal, the percentage of invasive K. pneumoniae isolates resistant to third-generation cephalosporins was 50.7% (https://atlas.ecdc.europa.eu/; accessed on 14 January 2025) [11]. This ECDC value represents the highest recorded percentage for third-generation cephalosporins since surveillance began in 2006 in Portugal and is a little higher than the one found in this study (43.7%). This upward trend in resistance poses a substantial challenge to public health, as it limits the effectiveness of commonly used antibiotics and complicates treatment options for infections caused by K. pneumoniae [26]. In Ethiopia, 67.3% of BSI isolates were ESBL producers [27]. A study conducted in Egypt reported lower percentages of ESBL-producing isolates than the previous study at 48.93% [28]. Also, a study in India demonstrated that 43.5% of Gram-negative isolates were β-lactamase producers [29]. The high prevalence of CTX-M among K. pneumoniae isolates is an important finding for applying rapid diagnostic methods. For example, rapid immunochromatographic or molecular methods capable of detecting CTX-M could be implemented in areas with a high prevalence of CTX-M ESBL [30]. These rates of ESBL occurrence represent a worryingly high prevalence amongst BSIs. This paper also underlines once more the call to strengthen surveillance and detection mechanisms in hospitals towards BSIs caused by ESBL-producing Enterobacterales.

Concerning the analysis of the whole genomes, three different sequence types were detected among the 14 ESBL-producing K. pneumoniae sequenced isolates: ST15 (57.1%), ST307 (28.6%), and ST34 (14.3%). Various studies have described different STs among the BSI K. pneumoniae isolates, where ST15 frequently emerges alongside other common types, such as ST11 and ST23 [31,32,33,34]. The prominence of ST15 in our isolates is notable, as it is associated with multidrug resistance and often implicated in healthcare-associated infections, indicating a potentially high transmission and adaptation rate in clinical settings. However, less commonly found, both ST34 and ST307 have also been reported [35,36,37,38]. The identification of these sequence types in our study highlights the diversity and adaptability of K. pneumoniae strains contributing to BSIs.

Regarding β-lactamase genes, we verified a predominance of the bla_CTX-M-15 gene, which was also predominant in the previous studies conducted by our group [15]. Still, in some strains, this gene was associated with a carbapenemase gene (bla_KPC or bla_OXA-48). It is interesting to show that, in four years, the genetic lineages seemed to differ in both studies. While our study reported isolates belonging to ST15, ST307, and ST34, the previous one reported isolates belonging to ST348, ST11, and ST15. ESBLs encoded by bla_CTX-M-15 are one of the most commonly found among K. pneumoniae isolates causing BSIs. Multiple studies have reported a high prevalence of this gene; for example, in South Korea, a study conducted between 2017 and 2019 verified that the dominant ESBL type in K. pneumoniae blood isolates was CTX-M-group 1 (78.5%), which included CTX-M-15 (n = 407) and CTX-M-55 (n = 5), while CTX-M-group 9 was less frequently identified (14.0%) but included CTX-M-14 (n = 60) and CTX-M-27 (n = 8) [39]. Also, in Italy, among the K. pneumoniae isolates, all encoded CTX-M-15 [40], and in Algeria, CTX-M-15 ESBL-K. pneumoniae strains infected 8.2% of neonates [41]. Regarding carbapenemase-producing enzymes, we detected the bla_KPC-3 gene (in addition to bla_CTX-M-15) in three isolates analyzed by WGS. The bla_KPC gene is frequent in BSI isolates of K. pneumoniae. In this sense, a prevalence of this gene of 33.3% has been identified in China [42] and 67.4% in carbapenem-resistant Enterobacterales isolates in India [43]. In Brazil, the prevalence of the bla_KPC gene in carbapenem-resistant K. pneumoniae BSI isolates was 90.4% [44]. It has been shown that the presence of bla_KPC genes increases the total mortality in BSI patients [42,44]. In the case of Enterobacterales, KPC-2 and -3 are the most commonly detected variants [35,39,45,46]. In Portugal, CTX-M-15 and KPC-3 are predominant, which is in line with our results. In the last five years, multiple studies have reported the presence of these enzymes, in particular KPC-3, among Enterobacterales isolates, underscoring their significance in antimicrobial resistance profiles within the region [15,16,47,48,49,50,51,52]. However, the genetic lineages identified in our study appear to diverge from those previously reported in Portugal. While ST15 is frequently encountered in Portuguese clinical bacterial isolates, the other two sequence types have been less commonly documented in healthcare settings [15,16,48,50,52,53,54,55]. The persistence of these resistance mechanisms reflects local and global trends in antimicrobial resistance and highlights the need for ongoing surveillance and stringent infection control measures.

Besides ESBLs, several other AMR genes were identified in BSI K. pneumoniae isolates, including those conferring resistance to aminoglycoside, fluoroquinolone, trimethoprim-sulfamethoxazole, tetracycline, and chloramphenicol. Among the commonly detected genes, those that encode aminoglycoside resistance included aac(3)-IIe, aph(3″)-Ib, aph(6)-Id, and aadA16; fluoroquinolones resistance included qnrB1, qnrB6, and co-resistance to fluoroquinolones aac(6′)-Ib-cr genes; trimethoprim-sulfamethoxazole included sul1, sul2, dfrA14, and dfrA27; tetracyclines included tetD genes; and chloramphenicol included catA1, catA2, and catB3. Also, genes for the efflux pumps, such as oqxA, oqxA10, oqxB19, and oqxB20, which further facilitated the ability of bacteria to efflux a wide range of antibiotics, were detected. Other studies have also identified several genes that confer resistance to these antibiotic classes, for example, a study conducted in the Eastern Democratic Republic of Congo reported similar results by finding that the K. pneumoniae isolates carried at least one gene linked to resistance to aminoglycosides and/or quinolones genes (aac(3)-IId, aac(3)-IIe, aadA1, aadA16, aadA2, aph(3′)-Ia, aph(3″)-Ib, aph(6)-Id, aac(6′)-Ib-cr, or aac(6′)-Ib-cr5), as well as to phenicols (catA1, catA2, and catB); sulfonamides (sul1 and sul2); tetracycline (tetA or tetD); trimethoprim (dfrA1, dfrA12, dfrA14, dfrA15, dfrA25, and dfrA27); plasmid-mediated quinolone resistance genes (qnrB, qnrB1, qnrB2, qnrB6, and qnrS1); and also efflux pump genes (oqxA, oqxB, oxqB5, oqxB19, oqxB21, and/or oqxB32) [56]. The diverse AMR genes that underpin its complex resistance mechanisms stress the need for intensive surveillance and targeted antimicrobial strategies in controlling this pathogen.

Regarding detected plasmids, we have identified either Inc plasmids or Col plasmids. Among the Inc plasmids, IncFIB(K), IncFIA(HI1), and IncR were the most detected ones, and regarding Col plasmids, we have detected Col440I, Col440II, and ColRNAI. Plasmid incompatibility groups, especially Inc-type plasmids, are known to play a significant role in the spread of antimicrobial resistance genes. IncF and IncR plasmids, for example, are frequently associated with resistance genes and have been widely documented in K. pneumoniae BSI isolates [57]. The high prevalence of IncFIB(K) and IncR plasmids is consistent with reports linking these plasmids to multidrug-resistant strains and their capacity to carry multiple resistance genes, thus enhancing the potential for resistance dissemination within hospital settings. While Inc plasmids are generally more common than Col plasmids in K. pneumoniae, Col-type plasmids like Col440I and ColRNAI can also contribute to survival under antibiotic pressure by facilitating resistance gene transfer, albeit on a smaller scale [58,59,60,61]. Their presence suggests that Col plasmids might play a secondary yet complementary role in the persistence and adaptability of K. pneumoniae in clinical environments. Understanding the roles these plasmids play in resistance gene transmission is crucial, as they represent key targets for surveillance and control measures aimed at reducing the spread of resistant strains in healthcare settings.

Regarding virulence genes, the detection of ybtP and ybtQ virulence genes was made in six isolates. These genes belong to the ybtPQXS operon, making part of a yersiniabactin biosynthesis and transport system. This operon is critical for acquiring and transporting iron through a siderophore-mediated mechanism, where yersiniabactin acts as the siderophore [62,63,64]. We also detected the temperature-sensitive hemagglutinin gene (tsh) in one isolate. This virulence factor promotes adherence to host cells, a critical first step in colonization and infection, and degrades the host proteins, which may facilitate tissue invasion and nutrient acquisition [65,66]. Apart from the virulence-associated genes identified, we also found in our isolates other genes that are not categorized under the traditional virulence factors but may considerably enable bacteria to survive against various environmental stresses. For instance, the product of the fieF gene is a ferrous iron efflux pump that helps bacteria regulate and export excessive iron ions. This function helps reduce metal-induced stress, supporting bacterial growth and contributing to the bacteria’s ability to survive in metal-rich environments, including during host infection [67,68]. We also found eight genes related to heat resistance: hsp20, psi-GI, kefB-GI, trxLHR, hdeD-GI, yfdX1, yfdX2, and shsP. These same eight genes have also been reported in a study performed on E. coli from cattle [69]. These heat resistance genes are crucial factors for bacterial survival in extreme environments. For example, the hsp20 gene encodes for a small-sized heat shock protein that protects bacterial cells under various stressful conditions, such as heat shock, osmotic shock, and oxidative stress [70,71]. Another example is the locus of heat resistance (LHR), a genomic island that confers exceptional heat resistance in Enterobacterales, including E. coli and Salmonella. It includes genes such as yfdX1, yfdX2, and hdeD, which are necessary for high-level heat resistance [72,73,74]. Altogether, these genes confer an impressive armamentarium on bacteria to resist stressing conditions and raise their survival capability outside their host and in environmental reservoirs. This indirectly contributes to the persistence and spread of antibiotic-resistant bacteria, as such genes support their populations in persisting and expanding under the variability of environmental conditions, challenges by antibiotics, or host immune defenses. These survival genes are important in understanding bacterial resilience, though virulence-related, because they can enable the long-term persistence of pathogenic and antibiotic-resistant strains in the ecosystem.