INTRODUCTION — Klebsiella pneumoniae is a gram-negative, lactose-fermenting, non-motile, aerobic rod-shaped bacterium. It has been a known human pathogen since it was first isolated in the late nineteenth century by Edwin Klebs.
The microbiology, pathogenesis, and epidemiology of infections due to K. pneumoniae will be reviewed here. The epidemiology, clinical manifestations, diagnosis, and treatment of infections due to this organism are discussed separately. (See "Clinical features, diagnosis, and treatment of Klebsiella pneumoniae infection".)
MICROBIOLOGY — K. pneumoniae has three subspecies with homologous DNAs but different biochemical reactions: K. pneumoniae subsp pneumoniae, K. pneumoniae subsp ozaenae, and K. pneumoniae subsp rhinoscleromatis. For the remainder of this topic review, K. pneumoniae will specifically refer to K. pneumoniae subsp pneumoniae.
The following are the characteristic biochemical reactions of the different organisms:
●K. pneumoniae is lactose fermenting, H2S- and indole-negative, has a positive Voges-Proskauer (VP) reaction, is capable of growth in KCN and using citrate as a sole carbon source, and is incapable of growth at 10ºC.
●The other two subspecies (ozaenae and rhinoscleromatis) are indole-negative with a negative VP reaction.
●Klebsiella oxytoca, which is one of the other species of the genus, is indole-positive, has a positive VP reaction, and is able to grow at 10ºC.
PATHOGENESIS — Five major virulence factors of K. pneumoniae are known to contribute to the pathogenesis of infection: capsular serotype, hypermucoviscosity phenotype, lipopolysaccharide, siderophores, and pili (table 1) [1]. The degree of virulence conferred by K. pneumoniae can be determined by several methods (table 2). A single strain of K. pneumoniae can express an inconsistent degree of virulence in different virulence testing assays.
Capsular serotypes — The capsular polysaccharide antigens of K. pneumoniae can be classified into 77 serotypes. Serotype prevalence varies widely in different regions [2-9]. Worldwide, K2 is the most common capsular serotype of human clinical isolates, usually from urine or sputum specimens [2]. However, in different series, serotypes 2, 21, and 55 were most common in Europe [2], and 21, 2, and 55 were most common in North America [2]; K1 was the predominant capsular serotype in bacteremia, liver abscess, and septic endophthalmitis isolates in Taiwan [3-6] and in liver abscess isolates in Korea [7]; and K54 was the major serotype from urinary, sputum, and blood isolates in Australia [8].
The K1 and K2 isolates are generally more virulent than non-K1/K2 isolates in terms of lethality for a mouse when injected intraperitoneally [10]. The discrepancies in the degree of virulence observed among strains with different capsule types may be explained by two major mechanisms:
●Macrophages with the lectin or mannose receptor can recognize the capsular sugar sequences of mannose-alpha-2/3-mannose or L-rhamnose-alpha-2/3-L-rhamnose, particular of non-K1/K2 antigens. Subsequent macrophage-mediated ingestion and killing of organisms is nonopsonic and known as lectinophagocytosis [11,12]. The K1 or K2 antigens lack these mannose or rhamnose sequences, which can protect these isolates from lectinophagocytosis [13,14].
●Strains of K. pneumoniae with K1 or K2 capsule serotypes are more often hypermucoviscous than non-K1/K2 strains [5]. (See 'RmpA gene' below.)
Some capsular serotypes such as K1, K2, and K25 may contribute to host infection by protecting the organism against phagocytosis by neutrophils [14,15]. In addition, the capsular polysaccharide may play an important role in the development of respiratory tract infection by protecting against phagocytosis by alveolar macrophages [16].
Hypermucoviscosity phenotype — K. pneumoniae strains possess the hypermucoviscosity phenotype if they are capable of producing a mucoviscous exopolysaccharide web. In the microbiology laboratory, these strains grow in sticky colonies on agar plates and are identified by a string test [17]. A positive string test is indicated by a >5 mm viscous string from the colony on an agar plate when stretched by a standard bacteriologic loop (picture 1) [18]. The clinically invasive nature of some K. pneumoniae strains correlates with these microbiologic characteristics.
Strains with the hypermucoviscosity phenotype or with enhanced production of polysaccharide capsule independent of capsular K serotypes are more resistant to complement-mediated serum killing than those without the hypermucoviscosity phenotype or those with reduced capsule production [18,19]. The clinical importance of these observations was illustrated in a study in which hypermucoviscosity correlated with high serum resistance and was present in 52 of 53 strains causing invasive infection (liver abscess) compared with 9 of 52 noninvasive strains [18].
A number of clinical studies have demonstrated a significant association between the hypermucoviscosity phenotype and a destructive tissue abscess syndrome, including abscesses in the liver and/or other sites [5,20-24].
These relationships were illustrated in a review of 151 K. pneumoniae isolates from patients with bacteremia [20]:
●The hypermucoviscosity phenotype was found in 38 percent. The rate was higher with community-acquired compared with nosocomial infections (49 versus 15 percent), a finding that is consistent with the association of the tissue abscess syndrome with community-acquired infections. (See 'Association with primary liver abscess' below.)
●Among 95 isolates from secondary bacteremia (excluding five isolates from mixed infections), a hypermucoviscosity phenotype was found in a significantly higher proportion of isolates from any abscess site (eg, liver, neck) compared with nonabscess sites such as the urinary tract or to isolates from primary bacteremia (85 versus 23 and 25 percent).
The outer layer of the capsule of most K. pneumoniae consists of a network structure of fine fibers derived from capsular polysaccharide [25], known as the exopolysaccharide web. This structure is encoded by the capsular polysaccharide (cps) gene cluster on the chromosome [26] and is positively regulated by plasmid-mediated genes, including rmpA (regulator of mucoid phenotype gene A) [26-28], rmpA2 [29,30] and magA (mucoid associated gene A) [18].
RmpA gene — The capsule of K. pneumoniae is encoded by the capsular polysaccharide (cps) gene cluster [26], which may be regulated by plasmid-born rmpA and rmpA2 genes to enhance extracapsular polysaccharide synthesis and produce the hypermucoviscosity phenotype [26-30].
The rmpA gene was significantly associated with a hypermucoviscosity phenotype, being found in 90 percent of such isolates compared with 22 percent of isolates without a hypermucoviscosity phenotype. On the other hand, the hypermucoviscosity phenotype occurred in only 81 percent of rmpA-positive isolates. These observations suggest that other factors control full expression of rmpA or the hypermucoviscosity phenotype [20]. Non-rmpA factors that might contribute to the hypermucoviscosity phenotype include rmpA2 (as in some rmpA-negative/rmpA2-positive strains), aerobactin iron uptake chelate (Iuc) system, and the transcriptional regulators KvrA and KvrB [31].
Mutation of rmpA gene (insertion or deletion), however, may contribute to the negative hypermucoviscosity phenotype and low virulence in some rmpA-positive isolates [32].
MagA gene — MagA encodes a 43 kD outer membrane protein. The magA gene is located within an operon that is specific to the serotype K1 capsular polysaccharide (cps) gene cluster [33-35], and the operon containing magA is responsible for the K1 capsular serotype [36]. In one report, for example, serotype K1 was present in all 36 magA-positive strains compared to none of 38 magA-negative strains [36]. Restriction of magA to serotype K1 strains has also been noted in other studies [5,37,38]. Further sequencing of the magA flanking region revealed that magA was the serotype K1 allele of the polymerase wzy loci in the cps operon [6,39,40]. As magA was a K1-antigen–specific polymerase, it has been proposed that magA has to be renamed wzyKpK1, the capsular polymerase specific to K. pneumoniae serotype K1 [40]. Other serotypes K2, K5, K20, K54, K57 and a new capsular type KN1 have different alleles at the wzy loci, which are the capsular polymerases specific to their cps gene clusters of different serotypes [6,41,42].
A study using transposon mutagenesis to identify candidate virulence genes demonstrated that magA-positive strains had a mucoviscous exopolysaccharide web, resisted phagocytosis, and produced liver abscesses and meningitis in mice [18]. On the other hand, magA-negative mutants of serotype K1 strains lacked the exopolysaccharide web, became susceptible to phagocytosis, and were not virulent.
Relationship between magA and rmpA — Several studies have evaluated the relationship among magA, rmpA, and the hypermucoviscosity phenotype:
●The hypermucoviscosity phenotype is not confined to magA-positive isolates (ie, serotype K1 strains) [20,38] nor is it found in all magA-positive isolates [38].
●RmpA can be responsible for the hypermucoviscosity phenotype of magA-negative isolates (such as K2 or other non-K1 serotype strains) [20], and rmpA-negative isolates of magA-positive strains may lack the hypermucoviscosity phenotype [20].
●The k2A gene is seen only in serotype K2 strains and corresponds to the magA region in the cps gene cluster in serotype K1 isolates [37]. Thus, k2A in serotype K2 isolates may play a similar pathogenetic role as magA in serotype K1 isolates.
A report of 73 K. pneumoniae isolates from liver abscess evaluated the relationship between rmpA carriage and capsule serotype [5]. RmpA carriage was found in all 49 K1 or K2 isolates compared to 16 of 24 non-K1/K2 isolates, suggesting that K1/K2 isolates are generally more hypermucoviscous than non-K1/K2 isolates. In addition, K1/K2 isolates had more resistance to phagocytosis than rmpA-positive non-K1/K2 isolates. The latter observation is consistent with other studies showing that the mucoid phenotype mediated by the plasmid-encoded rmpA gene is an important virulence factor in addition to the capsule in K1 and K2 K. pneumoniae isolates [28].
In summary, the rmpA gene is a positive regulator of extracapsular polysaccharide synthesis and confers a hypermucoviscosity phenotype in most isolates. In K1 isolates, the magA gene is essential for the formation of the exopolysaccharide web, a process that can be enhanced by rmpA. Capsular serotypes K1 and K2 also may contribute to virulence, independent of rmpA and magA.
Association with primary liver abscess — Primary liver abscess refers to infections that occur in the absence of a history of hepatobiliary disease. Most cases of K. pneumoniae primary liver abscess have been reported in Taiwan and are community-acquired [4-6,18,20,21,36,43-49].
MagA-positivity has been strongly associated with primary community-acquired liver abscess in humans and is less often present in other K. pneumoniae infections [18,36]. The association with magA was illustrated in a report in which 35 of 42 (83 percent) K. pneumoniae strains from patients with primary liver abscess were magA-positive compared with only 1 of 32 strains associated with other types of infection [36]. Similar findings were noted in another study in which magA was significantly more prevalent in invasive strains that caused primary liver abscess compared with noninvasive strains that did not (98 versus 28 percent) [18].
However, lower rates of magA (K1 capsular serotype)-positivity have been noted in other series of patients with liver abscess, with values ranging from 34 to 63 percent [4,6,7,20,21,37], and rmpA-positive isolates have predominated in several series [5,6,20]. The discrepancy in findings may be in part related to inconsistent distinction between primary and secondary liver abscess. As an example, a liver abscess in a patient with a small silent gallstone is more likely to represent a primary infection.
MagA and the hypermucoviscosity phenotype are much less common in K. pneumoniae secondary liver abscess as illustrated by the following observations:
●In a review of 129 patients with liver abscess in which 53 were considered primary and 76 were secondary [6]. The rate of K1 positivity was much higher in the primary abscesses (81 versus 42 percent). As noted above, magA is associated with the K1 capsular serotype [33-36]. Thus, prior studies showing an association between K1 and liver abscess, particularly if complicated by endophthalmitis, may have reflected magA-positive strains.
●In two series, the hypermucoviscosity phenotype was present in 40 of 51 cases of primary liver abscess compared with 9 of 39 cases of cholangitis (78 versus 23 percent) [20,21].
Some studies have suggested that the K1 or K2 capsular serotype is a more important virulence determinant for K. pneumoniae liver abscess than magA or rmpA [5] and that genotype K1 may be particularly associated with the risk of metastatic infection [6]. Such infections can occur in patients without any underlying predisposing medical conditions, which is a reflection of the virulent nature of the K1 strains [6]. In addition, serum resistance assays have shown that K1 isolates were significantly more virulent than K2 or other isolates [6]. Non-K1 or K2 capsular types of K. pneumoniae strains causing primary liver abscess have mainly included K5, K20, K54, K57, and KN1 isolates, which were almost all rmpA-positive and of hypermucoviscosity phenotype [6,41,42,50].
A possible mechanism for the role of the K1 capsular serotype in virulence is that chromosomal regions with additional virulence genes (magA, kfu/PTS [a regulator of iron uptake], and allS) are selectively present on the serotype K1 genome; genotype K1 strains with all three genomic regions are strongly associated with primary liver abscess and metastatic infection [49]. A high prevalence of the plasmid-borne rmpA gene also may contribute to the virulence of K1 strains [5,6]. Another study in Taiwan also determined the prevalence of these virulence factors among 50 K. pneumoniae isolates causing primary liver abscesses. In general, the prevalence of hypermucoviscosity phenotype, plasmid-born rmpA, aerobactin, kfu, and allS genes is higher in K1 isolates than K2 and other non-K1/K2 isolates [50].
Both the K1 capsular serotype and the hypermucoviscosity phenotype are associated with community-acquired rather than nosocomial K. pneumoniae infection, which could explain why isolates from primary liver abscess are almost exclusively acquired in the community [20,44].
Diabetes mellitus is a common risk factor for primary K. pneumoniae liver abscess [4,7,45-47] being present in 78 percent of 134 patients in one report [4]. The data are conflicting as to whether diabetes is [4,45,47] or is not [6] an independent risk factor for metastatic infection. The high prevalence of diabetes or impaired fasting glucose with K. pneumoniae liver abscess is not seen with other causes of liver abscess (eg, 75 versus 5 percent with polymicrobial liver abscess, and 70 versus 33 percent with non-K. pneumoniae liver abscess in two large series) [45,46].
Poor glycemic control reversibly impairs phagocytosis of K1 and K2 strains of K. pneumoniae but not other strains [51]. However, in the series of 177 cases of K. pneumoniae pyogenic liver abscess cited above, diabetes was significantly less common in K1 compared with non-K1 strains (54 versus 70 percent) and the median level of hemoglobin A1c was significantly lower in patients infected with K1 strains, indicating better glycemic control [6]. Similarly, the prevalence of diabetes mellitus was lower in patients with K1/K2 primary liver abscess than with non-K1/K2 liver abscess (53 versus 86 percent) [52]. K1/K2 isolates are virulent strains capable of causing illness among diabetic and non-diabetic hosts. These observations suggest that optimized glucose control may not be sufficient for reducing the risk of K1/K2 liver abscess in diabetic patients, but may be helpful in reducing non-K1/K2 liver abscess, which accounted for 12 (39 percent) of 31 cases of primary liver abscess in diabetic patients [52].
Lipopolysaccharide — K. pneumoniae serum resistance is mainly conferred by the lipopolysaccharide O side chain, which can impede C1q or C3b from binding to the bacterial cell membrane, thereby protecting the organism from subsequent complement-mediated membrane damage and cell death [53-56]. This effect has been demonstrated in mutant strains in which lack of the lipopolysaccharide O side chain is associated with sensitivity to complement-mediating killing [53,54]. The capsule in these strains played no role in serum resistance. However, some clinical K. pneumoniae isolates that lack the lipopolysaccharide O side chain retain resistance to complement-mediating killing if they are heavily encapsulated [19]. (See 'Hypermucoviscosity phenotype' above.)
Klebsiella lipopolysaccharide may also contribute to virulence by other mechanisms. These include increasing lethality in pulmonary infection by enhancing the propensity for bacteremia [57], and acting as an endotoxin, by triggering cytokine pathways leading to the sepsis syndrome and septic shock. (See "Pathophysiology of sepsis".)
Siderophores — Iron is an essential factor in the growth of Enterobacteriaceae such as Klebsiella [58]. Because of the scarcity of iron in the microenvironment, K. pneumoniae and other bacteria have developed multiple mechanisms to enhance iron uptake. These include synthesis of iron chelators (siderophores), such as enterobactin (also called enterochelin) and aerobactin [58-60], the aerobactin receptor to introduce exogenous aerobactin [61], and the kfu iron uptake system [49].
Enterobactin is the main iron uptake system and is synthesized by almost all K. pneumoniae strains [62-64]. In contrast, aerobactin synthesis has been infrequently found (less than 10 percent of strains) in most reports [62-65]. In a study of nine K. pneumoniae isolates with capsular serotype K1 or K2, all produced enterobactin, but only those that produced aerobactin were virulent, defined as an LD50 of less than 10(3) microorganisms [60]. Furthermore, transfer of the aerobactin gene into an avirulent strain enhanced virulence by 100-fold.
Both the aerobactin (iuc gene) and rmpA genes, which as noted above are significantly associated with a hypermucoviscosity phenotype, are located on a 180-kilobase plasmid [28]. This is consistent with the finding in a review of 241 blood isolates of K. pneumoniae that all 15 isolates (6 percent) producing aerobactin were associated with the hypermucoviscosity phenotype [65]. However, 20 percent of the isolates had a hypermucoviscosity phenotype. Thus, although the aerobactin gene was always associated with the rmpA gene, the rmpA gene was not always associated with the aerobactin gene indicating that some rmpA-encoding plasmids do not contain the aerobactin gene.
These observations suggest that the association of aerobactin with virulence could be mediated at least in part by the concurrent presence of the mucoid phenotype. This issue was addressed in knockout experiments in a K2 serotype K. pneumoniae strain [28]. RmpA-negative mutants were 1000 times less virulent than wild-type strains but still 100 times more virulent than the plasmid-less rmpA-negative and aerobactin-negative strains.
The latter observation supports a direct role for aerobactin-mediated iron uptake in virulence in the minority of K. pneumoniae strains that produce aerobactin. Further support for the role of iron uptake in virulence in K. pneumoniae comes from a study in which the kfu iron uptake system was associated with liver abscess formation and distant metastatic infection such as endophthalmitis [49]. In the series of 177 cases of K. pneumoniae liver abscess cited above, kfu was found in 97 percent of K1 strains but in no K2 strains, which may contribute to metastatic infection (eg, endophthalmitis) being relatively rare with K2 strains [6].
Pili (fimbriae) — Strains of K. pneumoniae express two morphologically and functionally distinct filaments: type 1 and the type 3 pili [66]:
●Type 1 pili are heteropolymeric mannose-binding fibers produced by all members of the Enterobacteriaceae family, which mediate adherence to many types of epithelial cells like the bladder epithelium [67]. In a review of 151 K. pneumoniae blood isolates, the type 1 fimbrial adhesion gene (fimH) was present in all but three (98 percent) [20].
●The type 3 fimbrial adhesion protein (MrkD adhesin) plays a central role in the virulence of K. pneumoniae by attaching to host cells, such as of the urogenital, respiratory, and intestinal tracts [68]. This can result in bacterial colonization, subsequent proliferation on the host mucosal surfaces, and clinical infection such as pyelonephritis or pneumonia. Type 3 pili are also required for biofilm formation by K. pneumoniae on plastics and human extracellular matrix, leading to the formation of treatment-resistant biofilm on indwelling plastic devices, such as intravenous and urinary catheters [69,70].
These effects of pili may be counteracted by host responses. As an example, binding of the bacteria to the mucus or epithelial cells of the host may trigger lectinophagocytosis by macrophages, leading to intracellular killing of the bacteria [11,12]. Subsequent invasion of the bacteria into the underlying tissue or the bloodstream may also be hampered by complement-mediated serum killing. Thus, survival and growth of pathogenic bacteria requires the other virulence factors described above.
Virulence of ESBL-producing strains — Dissemination or outbreak of extended-spectrum beta-lactamase (ESBL)-producing K. pneumoniae strains has been reported worldwide. (See "Clinical features, diagnosis, and treatment of Klebsiella pneumoniae infection".)
As noted in the preceding section, fimbriae contribute to the virulence of K. pneumoniae. A specific fimbria, called KPF-28, has been found in the great majority of ESBL-producing K. pneumoniae, and the plasmid encoding the ESBL enzyme is involved in KPF-28 expression [71]. KPF-28 promotes adherence and bowel colonization of ESBL-producing K. pneumoniae, enhancing the likelihood of spread and nosocomial outbreaks.
Except for fimbriae, ESBL-producing K. pneumoniae have a low prevalence of the other Klebsiella virulence factors described above, including the hypermucoviscosity phenotype and aerobactin. This was illustrated in a report of 190 ESBL-producing K. pneumoniae clinical isolates from France: only 7 percent had a hypermucoviscosity phenotype, 4 percent produced aerobactin, and 2 percent had both factors [72]. Additionally, in a study from Taiwan, ESBL-producing K. pneumoniae strains had a lower carriage rate of rmpA gene, but had higher tendency of mutation in rmpA gene, thus resulting in lower prevalence of the hypermucoviscosity phenotype and less virulence of the isolates [32]. (See 'RmpA gene' above.)
The low rate of these virulence factors suggests that ESBL-producing K. pneumoniae might have reduced virulence. The data have been conflicting in other studies as illustrated by the following observations:
●Support for reduced virulence comes from two studies from Taiwan in which ESBL-producing K. pneumoniae were not associated with invasive primary liver abscess [73,74]. The infections were primarily nosocomial [73], which rarely cause tissue abscesses due at least in part to a much lower prevalence of the hypermucoviscosity phenotype (15 versus 49 percent with community-acquired isolates in one report) [20]. (See 'Association with primary liver abscess' above.)
●In a review from five European countries, ESBL-producing K. pneumoniae strains had a significantly higher rate of resistance to the bactericidal effect of serum than non-ESBL-producing strains (30 versus 18 percent) [75]. In a second analysis from the same group, the ESBL-producing strains had a higher rate of capsular serotype K2; serotype K1 was rare [15]. After exclusion of the K1/K2 isolates which, as noted above, are associated with the hypermucoviscosity phenotype, the rate of serum resistance was similar in the ESBL- and non-ESBL-producing strains (about 12 percent).
The cause of serum resistance in these isolates is not well understood since the hypermucoviscosity phenotype and aerobactin are uncommon in Europe [72], a linkage to the lipopolysaccharide O1 serotype is unclear [75], and the ESBL-encoding plasmid was not involved [75]. (See 'Hypermucoviscosity phenotype' above.)
●However, in a study from Taiwan, among non-K1/K2 K. pneumoniae strains, the trend towards more frequent serum resistance among ESBL-producing strains compared with non-ESBL strains was not statistically significant [76].
ESBL production has emerged among hypervirulent K. pneumoniae strains in China [77,78]. In one study from Taiwan, hypervirulent ESBL-producing isolates (those with hypermucoviscosity phenotype, rmpA and rmpA2 genes) were more commonly recovered from diabetic patients and were mainly associated with bacteremia secondary to such infections as pneumonia and urinary tract infections [79]. Nonhypervirulent isolates were more commonly recovered from patients after prolonged hospital stays (>30 days) and were mainly associated with primary bacteremia. The overall in-hospital mortality was 56.3 percent.
K. pneumoniae carbapenemase-producing K. pneumoniae isolates exhibiting hypermucoviscosity phenotype, some of which carry rmpA and rmpA2 genes, have been identified in China, Brazil, the United States, and Taiwan [80-83]. Carbapenem-resistant and hypervirulent K. pneumoniae strains have become a great concern in China, although they remain rare in other parts of the world [84].
SUMMARY
●Klebsiella pneumoniae is a gram-negative, lactose-fermenting, nonmotile, aerobic rod-shaped bacterium. There are three subspecies differentiated by distinct biochemical reactions: K. pneumonia subspecies pneumoniae, ozaenae, and rhinoscleromatis. (See 'Introduction' above and 'Microbiology' above.)
●Five major virulence factors of K. pneumoniae are known to contribute to the pathogenesis of infection. These are the capsular serotype, hypermucoviscosity phenotype, lipopolysaccharide, siderophores, and pili (table 1). (See 'Pathogenesis' above.)
●Isolates with K1 and K2 capsular serotypes are generally more virulent than non-K1/K2 isolates of K. pneumoniae. (See 'Capsular serotypes' above.)
●The hypermucoviscosity phenotype is characterized by the ability to produce a mucoviscous exopolysaccharide web and growth in sticky colonies on agar plates. Clinical studies suggest that formation of destructive tissue abscesses, including liver abscesses, is associated with K. pneumoniae strains with a hypermucoviscosity phenotype. (See 'Hypermucoviscosity phenotype' above and "Invasive liver abscess syndrome caused by Klebsiella pneumoniae".)
●The lipopolysaccharide O side chain can impede complement binding to the K. pneumoniae cell membrane and thus protect the organism from complement-mediated membrane damage and cell death. (See 'Lipopolysaccharide' above.)
●K. pneumoniae has developed multiple mechanisms to enhance iron uptake, including the synthesis of iron chelators, or siderophores. One of these, aerobactin, is associated with high virulence in mouse virulence assays. (See 'Siderophores' above.)
●Pili, or fimbriae, mediate adherence of K. pneumoniae to host cells, including those of the urogenital, respiratory, and intestinal tracts, but their effects may be counteracted by host responses. (See 'Pili (fimbriae)' above.)
●A specific fimbria, called KPF-28, which promotes adherence and bowel colonization, has been found in the great majority of extended-spectrum beta-lactamase-producing K. pneumoniae. Otherwise, these resistant strains generally have a low prevalence of the other K. pneumoniae virulence factors. (See 'Virulence of ESBL-producing strains' above.)
ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Yin-Ching Chuang, MD, who contributed to earlier versions of this topic review.
