The family Enterobacteriaceae comprises a wide array of gram-negative bacilli whose reservoirs include soil, water, plants, and the gastrointestinal tracts of humans and animals. As a group, Enterobacteriaceae are the most frequent bacterial isolates recovered from inpatient and outpatient clinical specimens (1). In 2006 to 2007, Enterobacteriaceae accounted for 21% of pathogens isolated from all infection sites in the Centers for Disease Control and Prevention (CDC) National Healthcare Safety Network (NHSN) Surveillance System, the successor to the CDC’s National Healthcare-associated Infections Surveillance (NNIS) System (2).
OVERVIEW
Microbiologically, all members of the Enterobacteriaceae are facultative anaerobes that, with few exceptions, ferment glucose, reduce nitrate to nitrite, and are oxidase negative (3). Several approaches to classifying the Enterobacteriaceae have been used over the years, including phenotypic subgroupings (4), DNA-relatedness studies (5), and a combination of the two methods (6). A summary of a current classification is presented in Table 34-1(7).
For identification of aerobic gram-negative bacilli, many hospital microbiology laboratories now use automated rapid identification systems rather than conventional biochemical testing (3). Of particular importance to infection control is the ability to determine in the microbiology laboratory whether healthcare-associated infections are due to the spread of a single species. This requires the ability to type strains by classic or newer molecular methods (Table 34-2) (8). Pulsed-field gel electrophoresis (PFGE) has been the most widely used method of genotyping, and for small sets of isolates empiric guidelines have been formulated to interpret chromosomal DNA restriction patterns produced by this method (9) (see Chapter 102). These guidelines have been validated for some species (10). Currently though, PCR-based (multiple-locus variable number tandem repeat Analysis [MLVA]) and sequence-based methods (multiple-locus sequence typing [MLST] and whole genome sequencing) have become the standard.
As the use of invasive devices, broad-spectrum antibiotics, and immunosuppressive agents has increased in hospitals, the Enterobacteriaceae, particularly Escherichia coli, have become somewhat less prevalent, and gram-positive microorganisms, especially staphylococci and enterococci, more prevalent as causes of healthcareassociated infection. In 1999 and in 2006 to 2007 CDC data, the major device-associated infections, that is, centralline-associated bloodstream infections (CLABSIs), catheter-associated urinary tract infections (CAUTIs), and ventilator-associated pneumonia (VAP), were each caused most often by the same four pathogens, respectively (2,11); the one exception was that for VAP, in the 2006 to 2007 data, Acinetobacter baumannii tied Enterobacter spp. as the third most frequent pathogen (2).
To illustrate the changing overall role of Enterobacteriaceae in the pathogenesis of healthcare-associated infections, Table 34-3 presents NNIS data from 1980 to 1982 and 1990 to 1996 and NHSN data from 2006 to 2007. Comparisons between NNIS and NHSN need to be made with caution due to differences between the two systems (e.g., NHSN includes units outside of intensive care, has no minimum hospital size requirement for participation, and has a greatly expanded hospital base due to mandated use in many states for public reporting purposes). The percentage of pathogens recovered from healthcare-associated infections that were Enterobacteriaceae declined from 42% in 1980 to 1982 to 29% in 1990 to 1996, and continued to trend down to 21% for 2006 to 2007, primarily because of less frequent recovery of E. coli. This trend is apparent and continues in all major infection sites. For example, Enterobacteriaceae accounted for 18% of the 14,424 isolates causing bloodstream infections (BSIs) in the 1990 to 1996 NNIS data, but accounted for only 10% of the 21,943 isolates causing BSIs in the 1992 to 1999 data (11) and for 12.4% of 11,428 isolates from CLABSIs in 2006 to 2007 (2). Selected data from other recent multicenter healthcareassociated surveillance systems are shown in Table 34-4.
Although the overall percentage of healthcare-associated infections due to the Enterobacteriaceae has declined, Enterobacteriaceae remain important healthcare-associated pathogens. They have been implicated in almost a third (34%) of all healthcare-associated urinary tract infections (UTIs), in nearly a fifth (18%) of all SSIs, in up to 12% of all BSIs, and in 23% of all VAP. Overall, E. coli, Klebsiella pneumoniae, Enterobacter spp., and K. oxytoca were the most common healthcare-associated pathogens from the family Enterobacteriaceae and together accounted for about one-fifth of all healthcare-associated isolates in 2006 to 2007 (2).
TABLE 34-1 Aerobic Gram-Negative Bacilli: Enterobacteriaceae (Pertinent Characteristics: Ferment Sugars; Oxidase Negative; Most Reduce Nitrate to Nitrite)
Current Name
Synonym
Current Name
Synonym
Budvicia aquatica
Leclercia adecarboxylata
Escherichia adecarboxylata
Buttiauxella noackiae
CDC enteric group 59
CDC enteric group 41
Cedecea davisae
CDC enteric group 15
Leminorella grimontii
CDC enteric group 57
Cedecea lapagei
Leminorella richardii
Cedecea neteri
Cedecea sp. 4
Moellerella wisconsensis
CDC enteric group 46
Cedecea sp. 3
Morganella morganii ssp. morganii
Proteus morganii
Cedecea sp. 5
Citrobacter amalonaticus
Levinea amalonatica
Morganella morganii ssp. sibonii
Proteus morganii
Citrobacter braakii
Citrobacter freundii
Citrobacter diversus
Pantoea agglomerans
Enterobacter agglomerans
Citrobacter farmeri
Citrobacter amalonaticus biogroup 1
Pantoea dispersa Photorhabdus luminescens
Xenohabdus luminescens
Citrobacter freundii
Colobactrum freundii
Pragia fontium
Citrobacter gillenii
Citrobacter genomospecies 10
Proteus mirabilis
Citrobacter freundii
Proteus penneri
Proteus vulgaris biogroup 1
Citrobacter koseri
Citrobacter diversus
Proteus vulgaris
Proteus vulgaris biogroup 1
Levinea malonatica
Providencia alcalifaciens
Proteus inconstans
Citrobacter murliniae
Citrobacter genomospecies 11
Providencia rettgeri
Proteus rettgeri
Citrobacter freundii
Providencia rustigianii
Providencia alcalifaciens biogroup 3
Citrobacter rodentium
Citrobacter genomospecies 9
Providencia stuartii
Proteus inconstans
Citrobacter freundii
Rahnella aquatilis
Citrobacter sedlakii
Citrobacter genomospecies 8
Salmonella bongori
Salmonella subgroup 5
Citrobacter werkmanii
Citrobacter freundii
Citrobacter genomospecies 7
Salmonella choleraesus ssp. arizonae
Salmonella subgroup3a
Citrobacter youngae
Citrobacter freundii
Citrobacter genomospecies 5
Salmonella choleraesuis ssp. choleraesius
Salmonella subgroup 1
Edwardsiella hoshinae
Citrobacter freundii
Salmonella choleraesuis ssp. diarizonae
Salmonella subgroup 3b
Edwardsiella tarda
Enterobacter aerogenes
Aerobacter aerogenes
Salmonella choleraesuis ssp. houtenae
Salmonella subgroup 4
Enterobacter agglomerans group
Salmonella choleraesuis ssp. indica
Salmonella subgroup 6
Enterobacter amnigenus
Enterobacter asburiae
CDC enteric group 17
Salmonella choleraesuis ssp. salamae
Salmonella subgroup 2
Enterobacter cancerogenus
Enterobacter taylorae
Serratia ficaria
Erwinia cancerogena
Serratia fonticola
CDC enteric group 19
Serratia grimesii
Serratia liquefaciens
Enterobacter cloacae
Serratia liquefaciens
Enterobacter liquefaciens
Enterobacer gergoviae
Serratia marcescens
Enterobacter hormaechei
CDC enteric group 75
Serratia odoriferae
Enterobacter intermedius
Enterobacter intermedium
Serrattia plymuthica
Enterobacter kobei
Enterobacter sakazakii
Serratia proteamaculans ssp. proteamaculans
Serratia liquefaciens
Erwinia persicinus
Escherichia blattae
Serratia proteamaculans ssp. quinovora
Serratia liquefaciens
Escherichia coli
Serratia rubidaea
Escherichia fergusonii
CDC entric group 10
Shigella boydii
Shigella biogroup C
Escherichia hermannii
CDC enteric group 11
Shigella dysenteriae
Shigella biogroup A
Escherichia vulnerius
CDC enteric group 1
Shigella flexneri
Shigella biogroup B
Ewingella americana
CDC enteric group 40
Shigella sonnei
Shigella biogroup D
Hafnia alvei
Enterobacter hafniae
Tatumella ptyseos
CDC group EF-9
Klebsiella ornithinolytica
Klebsiella oxytoca ornithine Positive
Trabulsiella guamensis
Yersinia aldovae
CDC enteric group 90
Klebsiella oxytoca
Yersinia bercovieri
Yersinia enterocolitica biogroup 3b
Klebsiella planticola
Klebsiella travisanii
Yersinia enterocolitica
Pasteurella enterocolitica
Klebsiella pneumoniae ssp. ozaenae
Klebsiella ozaenae
Yersinia frederiksenii
Yersinia intermedia
Klebsiella pneumoniae ssp. pneumoniae
Klebsiella pneumoniae
Yersinia kristensenii
Yersinia mollaretii
Yersinia enterocolitica biogroup 3a
Klebsiella pneumoniae ssp. rhinoscleromatis
Klebsiella rhinoscleromatis
Yersinia pestis
Yersinia pseudotuberculosis
Pasteurella pestis
Pasteurella pseudotuberculosis
Klebsiella terrigena
Yersinia rohdei
Kluyvera ascorbata
CDC enteric group 8
Yokenella regensburgei
Koserella trabulsii CDC enteric group 45
Kluyvera cryocrescens
Kluyvera georgiana
CDC enteric group 36/37
Kluyvera species group 3
Note: Diagnostic laboratories may report Salmonella serovars by name, for example, Salmonella typhi or Salmonella serovar typhi.
CDC, Centers for Disease Control and Prevention.
Enterobacteriaceae discussed in the text are highlighted.
(Adapted from Bruckner DA, Colonna P, Bearson BL. Nomenclature for aerobic and facultative bacteria. Clin Infect Dis 1999;29:713-723.)
Though Enterobacteriaceae comprise a slightly smaller portion of healthcare-associated infections than in the past, the alarming increase in antimicrobial resistance, particularly to carbapenem antibiotics and the presence of resistance cassettes on transmissible genetic elements, compels healthcare professionals to understand the pathogenesis, infection control, and preventive measures to limit their spread.
PATHOGEN-SPECIFIC FACTORS IN THE PATHOGENESIS OF HEALTHCARE-ASSOCIATED INFECTIONS CAUSED BY ENTEROBACTERIACEAE
Multiple factors are involved in the pathogenesis of infection caused by Enterobacteriaceae. As discussed below, a variety of pathogen-specific factors, device-related factors, and host factors act together to determine the likelihood of infection. The virulence of the microorganism (i.e., the ability to invade and cause disease) relates to both pathogen factors and to the immune status of the patient.
Adhesion
Bacterial adhesion is a highly specific phenomenon that leads to attachment of bacteria to mucosal surfaces and, thus, to colonization and potentially to bacterial overgrowth and tissue invasion. Adhesins may also function as invasins, promote biofilm formation, and transmit signals to epithelial cells leading to inflammation (19,20). Among Enterobacteriaceae, adhesion is mediated by both fimbrial and nonfimbrial adhesins (Table 34-5) that are encoded on plasmids and on the bacterial genome, forming “pathogenicity islands” (21). The locus for enterocyte effacement on the chromosome encodes the virulence types necessary for attachment and effacement of E. coli to enterocytes (22). The E. coli fimbrial adhesins are among the most studied and the best characterized of the bacterial adhesins.
P fimbriae anchor bacteria to uroepithelial cells (29) and are found in strains that cause pyelonephritis in adults and children (30, 31and32). The symbol P was chosen because P-fimbriated E. coli were a frequent cause of pyelonephritis and because glycolipids were receptors for P fimbriae and antigens in the P blood group system (33). Compared to non-P-fimbriated strains of E. coli, isolates with P fimbriae can adhere to specific receptors on human colonic epithelial cells (leading to colonization), spread more easily to the urinary tract, have a better ability to persist in kidneys and bladders, and enhance the inflammatory response (33).
A relation between adherence and virulence has been demonstrated. Among E. coli strains from patients with different forms of UTIs, in vitro adherence to uroepithelial cells was found in 80% of the patients with pyelonephritis, 40% to 50% of the patients with acute cystitis, and 20% of the patients with asymptomatic bacteriuria (33). Studies of bacteremia secondary to urosepsis have shown that E. coli strains that cause urosepsis in healthy patients almost always have P fimbriae, whereas E. coli urosepsis in immunocompromised patients is less often due to such P-fimbriated strains (30, 31and32). In a study of fimbrial types found in respiratory isolates from intensive care unit (ICU) patients with presumed healthcare-associated pneumonia, P fimbriae were found in approximately half of the E. coli respiratory isolates (34). This rate is higher than the rate of P fimbriation commonly found in fecal isolates (14-16%) and, thus, raises the question of how the presence of this adhesin may be advantageous to strains causing pulmonary infection (34). Another study looked at the role of the papG class II gene, a P-fimbrial structural allele causing uroepithelial attachment of E. coli, and the pathogenesis of E. coli bacteremia in upper UTIs and ascending cholangitis. The authors found a significant difference between the presence of the virulence factor, papG class II, in bacteremic patients with upper UTI compared to bacteremic patients with ascending cholangitis and to controls (35).
TABLE 34-2 Characteristics of Bacterial Typing Systems
Proportion of Discriminatory Typing System
Strains Typeable
Reproducibility
Power
Phenotypic methods
Biotyping
All
Fair
Poor
Antimicrobial susceptibility testing
All
Fair
Poor
Serotyping
Most
Good
Fair
Bacteriophage typing
Most
Fair
Poor
Immunoblotting
All
Excellent
Good
Multilocus enzyme electrophoresis
All
Excellent
Good
Genotypic methods
Plasmid profile analysis
Most
Fair
Fair
Restriction endonuclease analysis
All
Good
Fair
Ribotyping
All
Excellent
Fair
PFGE
All
Good
Excellent
Polymerase chain reaction restriction
All
Excellent
Good digests
Arbitrarily primed polymerase
All
Good
Good chain reaction
MLVA
All
Excellent
Excellent
MLST
All
Excellent
Excellent
Note: These judgments represent the views of Maslow et al. (10); many systems remain incompletely evaluated, and characteristics may vary when the systems are applied to different species. (Modified from Maslow JN, Mulligan ME, Arbeit RD. Molecular epidemiology: application of contemporary techniques to the typing of microorganisms. Clin Infect Dis 1993;17:153-164.)
Cranberry juice consumption may offer protection against P-fimbriated strains of E. coli by the action of cranberry proanthocyanidin (condensed tannin), which inhibits P-fimbriated E. coli from adhering to uroepithelial cells (36). A randomized controlled study with 50 women per arm compared drinking 50 mL of cranberry juice concentrate daily for 6 months to drinking 100 mL of lactobacillus GG 5 days a week for 1 year, compared to controls. The authors found an absolute risk reduction for recurrent UTI of 20% in the group that drank cranberry juice (37).
S fimbriae are present on many E. coli strains that cause infant meningitis. The presence of binding sites for S fimbriae on blood vessels in the central nervous system (CNS) and on epithelial cells of the choroid plexus and of the ventricle of the infant rat brain provides a model for the pathogenesis of neonatal E. coli meningitis (38). The binding affinity of S fimbriae for vascular endothelium and epithelium of the choroid plexus and the ventricles decreases after the neonatal period in rats, paralleling the decrease in susceptibility to E. coli meningitis (39). S fimbriae also allow E. coli to bind to intact endothelial cells (38,40) and thus may be an important virulence factor for septicemia. When isolates of E. coli that caused a variety of invasive bacterial infections were compared to fecal isolates in healthy children, P fimbriae and S fimbriae were predominant in E. coli isolates causing invasive disease (41).
Type I pili have been associated with uropathogenic E. coli (UPEC). Type I pili facilitate entry of UPEC into bladder epithelial cells, with subsequent exfoliation (42). The ability of UPEC to invade and persist in bladder epithelial cells has been suggested as an explanation of recurrent UTIs.
Type IV pili have been identified in enteropathogenic E. coli, which frequently cause childhood diarrhea in developing countries. Type IV pili are known as bundle-forming pili (BFP) and are critical to the full virulence of these bacteria (43). Mutants without these pili could not attach to epithelial cells in vitro and were relatively benign when fed to human volunteers. These pili facilitate bacterial bundling into ropelike filaments that attach to epithelial cells; subsequently, the clumped bacteria disperse to cause infection (43).
In addition to Type 1 pili and P fimbriae, the Dr adhesion family of UPEC has been associated with pyelonephritis in pregnant women (44). The Dr adhesin family includes a fimbrial adhesin and nonfimbrial adhesins and is termed Dr for the blood group antigen, a common receptor for this family of adhesins (45). Dr adhesins are associated with bacterial persistence in the urinary tract and facilitate invasion of the bladder and kidney tissue and binding to decay accelerating factor, a host protective protein that prevents autologous complement mediated damage (46,47). Currently, it is unknown if this group of adhesins is expressed during UPEC-associated CAUTIs.
TABLE 34-3 Distribution of Selected Enterobacteriaceae and Other Pathogens Isolated from all Major Infection Sites, CDC
Autotransporter (AT) proteins are another group of adhesins associated with UTI virulence (48). Antigen 43 (Ag 43), a bacterial surface AT protein, has been identified in UPEC and enteropathogenic E. coli (48), though not specifically in CAUTI. Antigen 43 confers characteristic surface properties such as autoaggregation and promotes bacterial biofilm formation, and the Ag 43a variant recently has been shown to promote long-term persistence in the urinary bladder in mouse models of UTI (49). These autotranporters may be future targets for novel vaccines against gramnegative pathogens (50).
The role of adhesins in the pathogenesis of infection caused by other Enterobacteriaceae is not well characterized. Type 1, 3, and 6 fimbriae have been found in Klebsiella, but their function as virulence factors remains largely unknown (51,52). The majority of respiratory isolates of K. pneumoniae and K. oxytoca from ICU patients with presumed healthcare-associated pneumonia have been shown to express type 3 fimbriae and a mannose-resistant, Klebsiella-like (MR-K) hemagglutinin (34). Multidrug-resistant K. pneumoniae strains from a variety of healthcare-associated infections have been found to colonize the human intestinal tract through a plasmid-encoded 29,000-dalton surface protein (51) that facilitates adherence to gastrointestinal epithelium. Type 3 fimbriae also are commonly found in Klebsiella isolates associated with human UTIs (51). An MR-K fimbria has been isolated in Providencia stuartii and appears to be related to adherence to genitourinary catheters (53). Cell adhesins that allow attachment to exfoliated uroepithelium (54, 55and56) have been found in Proteus spp. as well. Nonfimbrial adhesive factors also are being characterized in the Enterobacteriaceae (57,58). An R-plasmid encoded nonfimbrial adhesive factor has been isolated from strains of K. pneumoniae responsible for a variety of healthcare-associated infections (57).
Capsules
The bacterial capsule, which is well characterized for Klebsiella spp., E. coli, and Salmonella typhi, can partly protect the microorganisms against the bactericidal effect of serum and against phagocytosis (57,59,60). However, most of the Enterobacteriaceae do not possess a substantial bacterial capsule and do not have serum resistance. In a prospective observational study from six United States university teaching hospitals evaluating the incidence and the risk factors for the development of endocarditis in bacteremic patients with prosthetic cardiac valves, a significant proportion of cases of new endocarditis were due to gram-negative aerobic bacilli, often when a portal of entry was found (61). This study suggests that the previous hypothesis that endocarditis was unlikely in the presence of gram-negative bacteremia, presumably because gram-negative bacilli are serum susceptible or if a portal of entry is identified, may not be correct.
TABLE 34-4 Examples of Multicenter Surveillance Studies of Healthcare-Associated Enterobacteriaceae
aAverage number of centers participating each year.
A capsule, when present, can also directly suppress the host immune response (62). In invasive E. coli disease in children, K1 and K5 capsules are found most commonly (41). It has been suggested that these capsules are more virulent, because they are structurally similar to human antigens, and therefore may be spared by or elude specific host defense mechanisms. The size of the capsule and the rate of capsule polysaccharide production appear to influence bacterial virulence (62).
Iron Chelators
The ability of some gram-negative bacteria to acquire iron for growth becomes an important factor in many gramnegative infections. Almost all the iron in the human body is bound to various proteins such as hemoglobin, myoglobin, and transferrin, thereby limiting the availability of free iron for utilization by bacteria. Some Enterobacteriaceae contain low molecular weight, high-affinity iron chelators called siderophores. The chelator permits the bacteria to scavenge iron from the host for growth purposes.
Aerobactin is an iron-chelating bacterial siderophore associated with increased virulence in E. coli (63) and Klebsiella (64,52). In Enterobacteriaceae, the catechol enterobactin is the most commonly occurring iron-chelating siderophore but does not appear to be associated with increased bacterial virulence (64,65) possibly because enterobactin is more antigenic than aerobactin and causes a strong antibody response in the host that diminishes enterobactin’s ability to take up iron (65).
Yersinia enterocolitica 1B, Y. pseudotuberculosis, and Y. pestis have been found to contain chromosomal gene sets designated high-pathogenicity islands (HPIs) that are involved in the synthesis, transport, and regulation of the siderophore yersiniabactin. This HPI has also been found in other genera including E. coli, Klebsiella, Citrobacter, and Enterobacter (66, 67and68).Y. enterocolitica has increased virulence in patients receiving desferrioxamine therapy, presumably because Yersinia can use desferrioxamine to meet some of its growth requirements more effectively in these patients (69,70).
Another method by which bacteria may acquire iron is hemolysis. Hemolysins are cytotoxic proteins encoded by chromosomal or plasmid genes. The chromosomal localization seems to be predominant for E. coli causing extraintestinal infections, whereas hemolysins are usually carried on plasmid genes in E. coli strains from veterinary sources. Hemolysins are cytotoxic for erythrocytes, and in vitro for polymorphonuclear leukocytes, monocytes, and isolated renal tubular cells. These proteins contribute to virulence in intraperitoneal infection models, but their role in ascending UTIs is uncertain. Hemolysin production is frequent in pyelonephritic clones of E. coli, but does not enhance bacterial persistence in kidneys and bladders (33,41).
Percent of Total Isolates
Escherichia
Klebsiella
Enterobacter
Serratia
Proteus
Morganella
Citrobacter
19
17
14
6
4
NR
2
19
6
3
NR
2
NR
NR
35
27
19
7
8
NR
5
35
24
13
8
12
NR
8
13
6
4
3
NR
NR
NR
8
4
3
2
2
NR
NR
12
8
5
NR
3
NR
NR
16
3
3
NR
3
NR
NR
9
6
4
3
NR
NR
NR
10
9
7
3
2
NR
2
13
7
5
2
1
NR
1
10
7
5
NR
NR
NR
NR
Other Pathogen Factors and Tropisms
Other virulence factors, such as bacterial motility (71); the ability to grow in alkaline pH; the ability to colonize skin (especially hands) of healthcare workers; the ability to produce urease, which catalyzes hydrolysis of urea in the urine and increases urinary pH (72); and the ability to produce biofilms (73), contribute to the ability of various members of the Enterobacteriaceae to produce disease, especially in healthcare settings. Enterobacteriaceae liberate numerous toxins, endotoxin being one of the most lethal, which contribute to bacterial virulence (Table 34-5). The role of endotoxins in healthcare-associated infection is no different from their role in community-acquired infection. Finally, some virulence factors have been associated with worsened patient outcome, although precise mechanisms of tissue injury are unknown. For example, a minor outer membrane protein (molecular weight of 32,000) is found more often in strains of Citrobacter koseri causing neonatal meningitis and abscess than in strains of C. koseri from other body sites (74). Evidence from an infant rat model suggests that strains of C. koseri with this outer membrane protein can produce more extensive histopathologic changes within the brain (75).
Infections by several species of Enterobacteriaceae have been associated with specific devices, materials, and/or procedures because of increased device affinity or specific tropisms. For example, Proteus mirabilis is an urease-producing bacterium that has been associated with bacteriuria and obstructed urinary catheters in patients with long-term indwelling bladder catheters (76). Urease catalyzes the hydrolysis of urea in the urine, thus alkalinizing it; this permits the formation of struvite and carbonate-apatite stones or sludge or concretions within the catheter lumen, leading to catheter obstruction. Other members of the Enterobacteriaceae family, including Morganella morganii, K. pneumoniae, P. vulgaris, and P. stuartii, also produce urease. Although no association between bacteriuria and catheter obstruction has been demonstrated for these microorganisms (76), some, such as P. stuartii, are very commonly associated with long-term bladder catheterization. An MR-K hemagglutinin has been identified in P. stuartii that increases adherence to catheter material (53).
TABLE 34-5 Examples of Pathogen-Specific Virulence Factors in Enterobacteriaceae
CFA I, II, colonization factor antigen I, II; HC, hemorrhagic colitis; HUS, hemolytic-uremic syndrome; LT, long-term; MR-K, mannose-resistant Klebsiella-like hemagglutinin; UTI, urinary tract infection.
The cell-surface characteristics of K. pneumoniae may also play a role in increased adherence to ventriculoperitoneal shunts. For example, when a multiresistant strain of K. pneumoniae was compared to its spontaneous in vitro antibiotic-susceptible derivative, the derivative was more adherent to the surface of ventriculoperitoneal catheters (58). Genetic studies suggested that the absence of a plasmid-mediated outer membrane protein led to increased adherence to the ventriculoperitoneal shunt surface.
Enterobacter sakazakii has been associated with several neonatal outbreaks and sporadic cases of sepsis, meningitis, and diarrhea (77, 78, 79, 80and81). No environmental source for E. sakazakii has been identified (81). Most outbreaks have been associated with either intrinsic or extrinsic contamination of powdered milk substitute. E. sakazakii has been isolated from powdered infant formula produced in 13 different countries by multiple manufacturers (79,82), suggesting that this microorganism has the propensity to contaminate such products. In vitro studies show that E. sakazakii survives better than E. cloacae in infant formula (77).
Y. enterocolitica has a well-documented association with blood transfusion-related sepsis (83,84). Apparently, blood donors with asymptomatic gastrointestinal infection with Y. enterocolitica and transient bacteremia at the time of blood donation are the most common source of such cases (85). The environment of cold stored red blood cells favors the growth of Y. enterocolitica more than the growth of other, more likely contaminants (e.g., skin flora from donors) since Y. enterocolitica survives better than most bacteria at refrigeration temperatures. In addition, progressive hemolysis of stored blood may provide an ongoing supply of iron for Yersinia’s growth. Virulent strains of Yersinia can also grow in calcium-free media such as that produced by citrate chelation of red blood cells for storage. Serotype 0:3 has accounted for the majority of cases of transfusion-related Yersinia sepsis. This serotype shows a persistent resistance to the bactericidal effect of serum at cold temperatures and has a growth-response curve that is directly related to the iron content of the culture medium (86).
Enterobacter spp. and Citrobacter spp. thrive in aqueous environments and may cause healthcare-associated bacteremia by their ability to grow in infusion fluids (87).Enterobacter spp. can fix nitrogen, allowing for replication in nitrogen-deficient fluids, and have been shown to have more rapid replication than E. coli, Klebsiella, Pseudomonas, or Proteus (88) in dextrose-containing solutions.
Examples of Genetics of Some Virulence Factors
R-plasmids, commonly found in Enterobacteriaceae, are also associated with bacterial virulence. They carry genes encoding virulence factors, such as adhesive factors (57), enterotoxins, and hemolysins (89). Plasmids code for outer membrane proteins for various Enterobacteriaceae. In Y. enterocolitica, a 70-kb plasmid codes proteins of the outer membrane that are associated with resistance to complement-mediated opsonization, to neutrophil phagocytosis, and to bactericidal activity of human serum (90). Similar plasmids have been isolated in Y. pseudotuberculosis and Y. pestis. Two soluble plasmid-mediated antigens, V and W, have been isolated from virulent strains of Y. pestis, Y. pseudotuberculosis, and Y. enterocolitica (90). Because plasmids are also important determinants of antimicrobial resistance, they may allow pathogens to link drug resistance and virulence determinants, which may be transferred together to other species.
Temporal Evolution of Antibiotic Resistance
The increasing prevalence of antibiotic-resistant Enterobacteriaceae has contributed to the difficulty in treating healthcare-associated infections (2,91,92). Antibiotic resistance often is related to excessive or widespread use of a particular antibiotic (93). For example, aminoglycoside and cephalosporin resistance in Klebsiella has been correlated with exposure to and intensity of use of these drugs (94,95). Mechanisms of antibiotic resistance in the Enterobacteriaceae include enzyme production that can inactivate or modify the drug (e.g., β-lactamase production), diminished permeability of antibiotics, and altered antibiotic target sites. Bacteria may acquire these mechanisms of resistance spontaneously via chromosomal mutation or via transfer of plasmids or transposable genetic elements from other bacteria (96). Genes that determine resistance to different classes of antibiotics may occur on a single plasmid so that use of one antimicrobial can lead to resistance to other classes of antibiotics.
Common mechanisms of β-lactam antibiotic resistance include chromosomal mutation, which is frequent and offers a bacterial survival advantage during antibiotic therapy. This is the mechanism of resistance found in Enterobacteriaceae, such as Enterobacter, Serratia, indole-positive Proteus, and Citrobacter, which carry chromosomal genes that encode a type 1 β-lactamase (96,97). These bacteria can undergo single-step mutations to constitutive high-level β-lactamase production. Thus, initially susceptible strains of Enterobacter and other Enterobacteriaceae may develop spontaneous resistant mutants to broad-spectrum cephalosporins during 20% to 50% of courses of therapy (98). These AmpC enzymes may be present on plasmids and therefore, transmissible to other species such as E. coli and Klebsiella (99,100). The AmpC β-lactamase hydrolyzes cephamycins and oxyimino-β-lactamases (or third-generation cephalosporins) (101). AmpC enzymes are not inhibited by β-lactamase inhibitors such as clauvulanic acid, a characteristic that distinguishes them from extended-spectrum β-lactamases (ESBLs) (102).
Enterobacteriaceae have responded to the widespread use of β-lactam antibiotics with inactivation of these drugs by a variety of β-lactamases (such as TEM-1, TEM-2, and SHV-1 β-lactamases), which are typically plasmid-encoded (96). These resistances were overcome by the pharmaceutical industry development of second-and third-generation cephalosporins and combinations of β-lactam antibiotics with β-lactamase inhibitors. However, in 1982, the first Enterobacteriaceae with resistance to broad-spectrum cephalosporins, such as cefuroxime, cefotaxime, and ceftazidime, were isolated in Europe (103,104). These ESBLs arose by point mutations, which arose in the face of widespread use of antibiotics. ESBLs differ in only one or a few amino acids from the original TEM-1, TEM-2, and SHV-1 β-lactamases and are also plasmid-mediated (96). Hundreds of ESBLs have been identified (101,105, 106, 107and108). Although the first ESBLs were reported from Europe, they have spread to most continents during the last two decades (109,110). Some of these ESBLs are highlighted in Table 34-6. Klebsiella spp. and E. coli have been the primary carriers of ESBLs in North America and Europe (111,112). ESBL profiles differ dramatically between the United States and Canada and between North America and Europe (111,112). Large surveillance studies have shown various rates of ESBL production depending on geographic location, on whether surveillance relied only on phenotypic data or used confirmatory microbiologic testing, and on which patient care areas were assessed (i.e., ICU vs. other wards) (Table 34-7).
While the healthcare-associated pathogen distribution remains similar (11,116), increasing antibiotic resistance among Enterobacteriaceae infections is a major cause of concern in healthcare facilities. CDC reported that from 1998 to 2002 to 2003, there was a 47% increase in the rate of resistance to third-generation cephalosporins among K. pneumoniae recovered from healthcare-associated infections in ICU patients (91). In addition, comparing 1986-2003 to 2006-2007, among pathogens causing device-associated healthcare-associated infections (HAIs) (2), resistance to extended-spectrum cephalosporins increased among healthcare-associated E. coli (6% vs. 6-11%) and K. pneumoniae (21% vs. 21-27%) isolates, and resistance to carbapenems increased among Pseudomonas aeruginosa (21% vs. 25%). Among all HAIs, while resistant gram-positive species represent the four most common antimicrobialresistant pathogens, P. aeruginosa isolates resistant to fluoroquinolones, carbapenems, and β-lactam/β-lactamase inhibitor combinations represent the fourth, fifth, and sixth most common resistant pathogens, respectively. K. pneumoniae resistant to third-generation cephalosporins, E. coli resistant to fluoroquinolones, and A. baumannii resistant to carbapenems also are prominent causes of CLABSI, CAUTI, and VAP (Table 34-7).
TABLE 34-6 Selected β-Lactamases of Enterobacteriaceae
gRef (372) % resistant to ceftazidime, ceftriaxone or cefotaxime; ref (92) % resistant to ceftriaxone or cefotaxime and remainder of studies, % resistant to ceftriaxone.
