CHAPTER 45: ENTEROCOCCAL INFECTIONS

Enterococci have been recognized as potential human pathogens for more than a century, but only in recent years have these organisms acquired prominence as important causes of nosocomial infections. The ability of enterococci to survive and/or disseminate in the hospital environment and to acquire antibiotic resistance determinants makes the treatment of some enterococcal infections in critically ill patients a difficult challenge. Enterococci were first mentioned in the French literature in 1899; the “entérocoque” was found in the human gastrointestinal tract and was noted to have the potential to produce significant disease. Indeed, the first pathologic description of an enterococcal infection dates to the same year. A clinical isolate from a patient who died as a consequence of endocarditis was initially designated Micrococcus zymogenes, was later named Streptococcus faecalis subspecies zymogenes, and would now be classified as Enterococcus faecalis. The ability of this isolate to cause severe disease in both rabbits and mice illustrated its potential lethality in the appropriate settings.

Enterococci are gram-positive organisms. In clinical specimens, they are usually observed as single cells, diplococci, or short chains (Fig. 45-1), although long chains are noted with some strains. Enterococci were originally classified as streptococci because organisms of the two genera share many morphologic and phenotypic characteristics, including a generally negative catalase reaction. Only DNA hybridization studies and later 16S rRNA sequencing clearly demonstrated that enterococci should be grouped as a genus distinct from the streptococci. Nonetheless, unlike the majority of streptococci, enterococci hydrolyze esculin in the presence of 40% bile salts and grow at high salt concentrations (e.g., 6.5%) and at high temperatures (46°C). Enterococci are usually reported by the clinical laboratory to be nonhemolytic on the basis of their inability to lyse the ovine or bovine red blood cells (RBCs) commonly used in agar plates; however, some strains of E. faecalis do lyse RBCs from humans, horses, and rabbits. The majority of clinically relevant enterococcal species hydrolyze pyrrolidonyl-β-naphthylamide (PYR); this characteristic is helpful in differentiating enterococci from organisms of the Streptococcus gallolyticus group (formerly known as S. bovis), which includes S. gallolyticus, Streptococcus pasteurianus, and Streptococcus infantarius, and from Leuconostoc species. Although at least 18 species of enterococci have been isolated from human infections, the overwhelming majority of cases are caused by two species, E. faecalis and Enterococcus faecium. Less frequently isolated species include Enterococcus gallinarum, Enterococcus durans, Enterococcus hirae, and Enterococcus avium.

Enterococci are normal inhabitants of the large bowel of human adults, although they usually make up <1% of the culturable intestinal microflora. In the healthy human gastrointestinal tract, enterococci are typical symbionts that coexist with other gastrointestinal bacteria; in fact, the utility of certain enterococcal strains as probiotics in the treatment of diarrhea suggests their possible role in maintaining the homeostatic equilibrium of the bowel. Enterococci are intrinsically resistant to a variety of commonly used antibacterial drugs. One of the most important factors that disrupts this equilibrium and promotes increased gastrointestinal colonization by enterococci is the administration of antimicrobial agents. In particular, antibiotics that are excreted in the bile and have broad-spectrum activity (e.g., certain cephalosporins that target anaerobes and gram-negative bacteria) are usually associated with the recovery of higher numbers of enterococci from feces. This increased colonization appears to be due not only to the simple enterococcal replacement in a “biologic niche” after the eradication of competing components of the flora, but also (at least in mice) to the suppression—upon reduction of the gram-negative microflora by antibiotics—of important immunologic signals (e.g., by the lectin RegIIIγ) that help keep enterococcal counts low in the normal human bowel. Several studies have shown that higher levels of gastrointestinal colonization are a critical factor in the pathogenesis of enterococcal infections. However, the mechanisms by which enterococci successfully colonize the bowel and gain access to the lymphatics and/or bloodstream remain incompletely understood.

Several vertebrate, worm, and insect models have been developed to study the role of possible pathogenic determinants in both E. faecalis and E. faecium. Three main groups of virulence factors may increase the ability of enterococci to colonize the gastrointestinal tract and/or cause disease. The first group, enterococcal secreted factors, are molecules released outside the bacterial cell that contribute to the process of infection. The best-studied of these molecules include enterococcal hemolysin/cytolysin and two enterococcal proteases (gelatinase and serine protease). Enterococcal cytolysin is a heterodimeric toxin produced by some strains of E. faecalis that is capable of lysing human RBCs as well as polymorphonuclear leukocytes and macrophages. E. faecalis gelatinase and serine protease are thought to mediate virulence by several mechanisms, including the degradation of host tissues and the modification of critical components of the immune system. Mutants lacking the genes corresponding to these proteins are highly attenuated in experimental peritonitis, endocarditis, and endophthalmitis.

A second group of virulence factors, enterococcal surface components, are thought to contribute to bacterial attachment to extracellular matrix molecules in the human host. Several molecules on the surface of enterococci have been characterized and shown to play a role in the pathogenesis of enterococcal infections. Among the characterized adhesins is aggregation substance of E. faecalis, which mediates the attachment of bacterial cells to each other, thereby facilitating conjugative plasmid exchange. Several lines of evidence indicate that aggregation substance and enterococcal cytolysin act synergistically to increase the virulence potential of E. faecalis strains in experimental endocarditis. The surface protein adhesin of collagen of E. faecalis (Ace) and its E. faecium homologue (Acm) recognize adhesive matrix molecules (MSCRAMMs) involved in bacterial attachment to host proteins such as collagen, fibronectin, and fibrinogen; both Ace and Acm are important in the pathogenesis of experimental endocarditis. Pili of gram-positive bacteria have been shown to be important mediators of attachment to and invasion of host tissues and are considered potential targets for immunotherapy. Both E. faecalis and E. faecium have surface pili. Mutants of E. faecalis lacking pili are attenuated in biofilm production, experimental endocarditis, and urinary tract infections (UTIs). Other surface proteins that share structural homology with MSCRAMMs and appear to play a role in enterococcal attachment to the host and in virulence include the E. faecalis surface protein Esp and its E. faecium homologue Esp_fm, the second collagen adhesin of E. faecium (Scm), the surface proteins of E. faecium (Fms), SgrA (which binds to components of the basal lamina), and EcbA (which binds to collagen type V). Additional surface components apparently associated with pathogenicity include the Elr protein (a protein from the WxL family) and polysaccharides, which are thought to interfere with phagocytosis of the organism by host immune cells. Some E. faecalis strains appear to harbor at least three distinct classes of capsular polysaccharide; some of these polysaccharides play a role in virulence and are potential targets for immunotherapy.

The third group of virulence factors has not been well characterized but consists of the E. faecalis stress protein Gls24, which has been associated with enterococcal resistance to bile salts and appears to be important in the pathogenesis of endocarditis, and the hyl_Efm-containing plasmids of E. faecium, which are transferable between strains and increase gastrointestinal colonization by E. faecium. In mouse peritonitis, acquisition of these plasmids increased the lethality of a commensal strain of E. faecium. Recently, a gene encoding a regulator of oxidative stress (AsrR) has been identified as an important virulence factor of E. faecium.

The ability to sequence bacterial genomes has increased our understanding of bacterial diversity, evolution, pathogenesis, and mechanisms of antibiotic resistance. The genome sequences of more than 560 enterococcal strains are currently available, and some have been entirely closed and annotated. Sequence analysis has shown that the genetic diversity of enterococci is related in large part to the acquisition of exogenous DNA and the mobilization of large chromosomal regions, resulting in recombination of the “core” genomes. In addition, analyses indicate that E. faecium harbors a malleable accessory genome incorporating a substantial content of exogenous elements, including DNA from phages. Indeed, a hospital-associated E. faecium clade that contains most clinical and outbreak-associated strains is the predominant genetic lineage circulating in hospitals around the world. This clade appears to be evolving rapidly, and genomic comparisons suggest that this lineage emerged 75 years ago—a time point that coincides with the introduction of antimicrobial drugs—and evolved from animal strains, not from human commensal isolates. An initial genomic separation within E. faecium appears to have occurred ~3000 years ago, simultaneous with urbanization and domestication of animals. This genomic information provides new clues with regard to the evolution of enterococci from commensal organisms to important nosocomial pathogens.

According to the National Healthcare Safety Network of the Centers for Disease Control and Prevention, enterococci are the second most common organisms (after staphylococci) isolated from hospital-associated infections in the United States. Although E. faecalis remains the predominant species recovered from nosocomial infections, the isolation of E. faecium has increased substantially in the past 20 years. In fact, E. faecium is now almost as common as E. faecalis as an etiologic agent of hospital-associated infections. This point is important, because E. faecium is by far the most resistant and challenging enterococcal species to treat; indeed, more than 80% of E. faecium isolates recovered in U.S. hospitals are resistant to vancomycin, and more than 90% are resistant to ampicillin (historically the most effective β-lactam agent against enterococci). Resistance to vancomycin and ampicillin in E. faecalis isolates is much less common (~7% and ~4%, respectively).

The dynamics of enterococcal transmission and dissemination in the hospital environment have been extensively studied, with a focus on vancomycin-resistant enterococci (VRE). These studies have revealed that VRE colonization of the gastrointestinal tract is a critical step in the natural history of enterococcal disease and that a substantial proportion of patients colonized with VRE remain colonized for prolonged periods (sometimes >1 year) and are more likely to develop an Enterococcus-related illness (e.g., bacteremia). The most important factors associated with VRE colonization and persistence in the gut include prolonged hospitalization; long courses of antibiotic therapy; hospitalization in long-term-care facilities, surgical units, and/or intensive care units; organ transplantation; renal failure (particularly in patients undergoing hemodialysis) and/or diabetes; high Acute Physiology and Chronic Health Evaluation (APACHE) scores; and physical proximity to patients infected or colonized with VRE or these patients’ rooms. Once a patient becomes colonized with VRE, several key factors are involved in the organisms’ dissemination in the hospital environment. VRE can survive exposure to heat and certain disinfectants and have been found on numerous inanimate objects in the hospital, including bed rails, medical equipment, doorknobs, gloves, telephones, and computer keyboards. Thus health care workers and the environment play pivotal roles in enterococcal transmission from patient to patient, and infection control measures are crucial in breaking the chain of transmission. Moreover, two meta-analyses have found that, independent of the patient’s clinical status, VRE infection increases the risk of death over that among individuals infected with a glycopeptide-susceptible enterococcal strain.

The epidemiology of enterococcal disease and the emergence of VRE have followed slightly different trends in other parts of the world than in the United States. In Europe, the emergence of VRE in the mid-1980s was seen primarily in isolates recovered from animals and healthy humans rather than from hospitalized patients. The presence of VRE was associated with the use of the glycopeptide avoparcin as a growth promoter in animal feeds; this association prompted the European Union to ban the use of this compound in animal husbandry in 1996. However, after an initial decrease in the isolation of VRE from animals and humans, the prevalence of hospital-associated VRE infections has slowly increased in certain European countries, with important regional differences. For example, rates of vancomycin resistance among E. faecium clinical isolates in Europe are highest in Greece, the United Kingdom, and Portugal (10–30%), whereas rates in the Scandinavian countries and the Netherlands are <1%. These regional differences have been attributed in part to the implementation of aggressive “search-and-destroy” policies of infection control in countries such as the Netherlands; these policies have kept the frequency of nosocomial methicillin-resistant Staphylococcus aureus (MRSA) and VRE very low. Despite regional differences, rates of VRE continue to be much lower in Europe than in the United States. The reasons are not totally understood, although it has been postulated that this difference is related to the higher levels of human antibiotic use in the United States. Rates of enterococcal resistance to vancomycin in some Latin American countries are also lower (~4%) than those in the United States. Conversely, in Asia, rates of vancomycin resistance among enterococci appear to be similar to those in U.S. hospitals.

As mentioned above, genomic analyses of vancomycin-resistant E. faecium in different parts of the world suggest that the ­emergence and dissemination of these organisms in the hospital environment worldwide are due to the success of a unique hospital-associated genetic clade that acquired the genes responsible for ­vancomycin resistance as well as other antibiotic resistance determinants.

URINARY TRACT INFECTION AND PROSTATITIS

Enterococci are well-known causes of nosocomial UTI—the most common infection caused by these organisms (Chap. 33). Enterococcal UTIs are usually associated with indwelling catheters, instrumentation, or anatomic abnormalities of the genitourinary tract, and it is often challenging to differentiate between true infection and colonization (particularly in patients with chronic indwelling catheters). The presence of leukocytes in the urine in conjunction with systemic manifestations (e.g., fever) or local signs and symptoms of infection with no other explanation and a positive urine culture (≥10⁵ colony-forming units [CFU]/mL) suggests the diagnosis. Moreover, enterococcal UTIs often occur in critically ill patients whose comorbidities may obscure the diagnosis. In many cases, removal of the indwelling catheter may suffice to eradicate the organism without specific antimicrobial therapy. In rare circumstances, UTIs caused by enterococci may run a complicated course, with the development of pyelonephritis and perinephric abscesses that may be a portal of entry for bloodstream infections (see below). Enterococci are also known causes of chronic prostatitis, particularly in patients whose urinary tract has been manipulated surgically or endoscopically. These infections can be difficult to treat because the agents most potent against enterococci (i.e., aminopenicillins and glycopeptides) penetrate prostatic tissue poorly. Chronic prostatic infection can be a source of recurrent enterococcal bacteremia.

BACTEREMIA AND ENDOCARDITIS

Bacteremia without endocarditis is one of the most common presentations of enterococcal disease. Intravascular catheters and other devices are commonly associated with these bacteremic episodes (Chap. 17). Other well-known sources of enterococcal bacteremia include the gastrointestinal and hepatobiliary tracts; pelvic and intraabdominal foci; and, less frequently, wound infections, UTIs, and bone infections. In the United States, enterococci are ranked second (after coagulase-negative staphylococci) as etiologic agents of central line–associated bacteremia. Patients with enterococcal bacteremia usually have comorbidities and have been in the hospital for prolonged periods; they commonly have received several courses of antibiotics. Several studies indicate that the isolation of E. faecium from the blood may lead to worse outcomes and higher mortality rates than when other enterococcal species are isolated; this finding may be related to the higher prevalence of vancomycin and ampicillin resistance in E. faecium than in other enterococcal species, with the consequent reduction of therapeutic options. In many cases (usually when the gastrointestinal tract is the source), enterococcal bacteremia may be polymicrobial, with gram-negative organisms isolated at the same time. In addition, several cases have now been documented in which enterococcal bacteremia was associated with Strongyloides stercoralis hyperinfection syndrome in immunocompromised patients.

Enterococci are important causes of community- and health care–associated endocarditis, ranking second after staphylococci in the latter infections. The presumed initial source of bacteremia leading to endocarditis is the gastrointestinal or genitourinary tract—e.g., in patients who have malignant and inflammatory conditions of the gut or have undergone procedures in which these tracts are manipulated. The affected patients tend to be male and elderly and to have other debilitating diseases and heart conditions. Both prosthetic and native valves can be involved; mitral and aortic valves are affected most often. Community-associated endocarditis (usually caused by E. faecalis) also occurs in patients with no apparent risk factors or cardiac abnormalities. Endocarditis in women of childbearing age has been well described. The typical presentation of enterococcal endocarditis is a subacute course of fever, weight loss, malaise, and cardiac murmur; typical stigmata of endocarditis (e.g., petechiae, Osler’s nodes, Roth’s spots) are found in only a minority of patients. Atypical manifestations include arthralgias and manifestations of metastatic disease (splenic abscesses, hiccups, pain in the left flank, pleural effusion, and spondylodiscitis). Embolic complications are variable and can affect the brain. Heart failure is a common complication of enterococcal endocarditis, and valve replacement may be critical in curing this infection, particularly when multidrug-resistant organisms or major complications are involved. The duration of therapy is usually 4–6 weeks, with more prolonged courses suggested for multidrug-resistant isolates in the absence of valvular replacement.

MENINGITIS

Enterococcal meningitis is an uncommon disease (accounting for only ~4% of meningitis cases) that is usually associated with neurosurgical interventions and conditions such as shunts, central nervous system (CNS) trauma, and cerebrospinal fluid (CSF) leakage. In some instances—usually in patients with a debilitating condition, such as cardiovascular or congenital heart disease, chronic renal failure, malignancy, receipt of immunosuppressive therapy, or HIV/AIDS—presumed hematogenous seeding of the meninges is seen in infections such as endocarditis or bacteremia. Fever and changes in mental status are common, whereas overt meningeal signs are less so. CSF findings are consistent with bacterial infection—i.e., pleocytosis with a predominance of polymorphonuclear leukocytes (average, ~500/μL), an elevated serum protein level (usually >100 mg/dL), and a decreased glucose concentration (average, 28 mg/dL). Gram’s staining yields a positive result in about half of cases, with a high rate of organism recovery from CSF cultures; the most common species isolated are E. faecalis and E. faecium. Complications include hydrocephalus, brain abscesses, and stroke. As mentioned before for bacteremia, an association with Strongyloides hyperinfection has also been documented.

INTRAABDOMINAL, PELVIC, AND SOFT TISSUE INFECTIONS

As mentioned earlier, enterococci are part of the commensal flora of the gastrointestinal tract and can produce spontaneous peritonitis in cirrhotic individuals and in patients undergoing chronic ambulatory peritoneal dialysis (Chap. 29). These organisms are commonly found (usually along with other bacteria, including enteric gram-negative species and anaerobes) in clinical samples from intraabdominal and pelvic collections. The presence of enterococci in intraabdominal infections is sometimes considered to be of little clinical relevance. Several studies have shown that the role of enterococci in intraabdominal infections originating in the community and involving previously healthy patients is minor, because surgery and broad-spectrum antimicrobial drugs that do not target enterococci are often sufficient to treat these infections successfully. In the last few decades, however, these organisms have become prominent as a cause of intraabdominal infections in hospitalized patients because of the emergence and spread of vancomycin resistance among enterococci and the increase in rates of nosocomial infections due to multidrug-resistant E. faecium isolates. In fact, several studies have now documented treatment failures due to enterococci, with consequently increased rates of postoperative complications and death among patients with intraabdominal infections. Thus, anti-enterococcal therapy is recommended for nosocomial peritonitis in immunocompromised and severely ill patients who have had a prolonged hospital stay, have undergone multiple procedures, have persistent abdominal sepsis and collections, or have risk factors for the development of endocarditis (e.g., prosthetic or damaged heart valves). Conversely, specific treatment for enterococci in the first episode of intraabdominal infections originating in the community and affecting previously healthy patients with no important cardiac risk factors for endocarditis does not appear to be beneficial.

Enterococci are commonly isolated from soft tissue infections (Chap. 26), particularly those involving surgical wounds (Chap. 17). In fact, these organisms rank third as agents of nosocomial surgical-site infections, with E. faecalis the most frequently isolated species. The clinical relevance of enterococci in some of these infections—as in intraabdominal infections—is a matter of debate; differentiating between colonization and true infection is sometimes challenging, although in some cases enterococci have been recovered from lung, liver, and skin abscesses. Diabetic foot and decubitus ulcers are often colonized with enterococci and may be the portal of entry for bone infections.

OTHER INFECTIONS

Enterococci are well-known causes of neonatal infections, including sepsis (mostly late-onset), bacteremia, meningitis, pneumonia, and UTI. Outbreaks of enterococcal sepsis in neonatal units have been well documented. Risk factors for enterococcal disease in newborns include prematurity, low birth weight, indwelling devices, and abdominal surgery. Enterococci have also been described as etiologic agents of bone and joint infections, including vertebral osteomyelitis, usually in patients with underlying conditions such as diabetes or endocarditis. Similarly, enterococci have been isolated from bone infections in patients who have undergone arthroplasty or reconstruction of fractures with the placement of hardware. Because enterococci can produce a biofilm that is likely to alter the efficacy of otherwise active anti-enterococcal agents, treatment of infections that involve foreign material is challenging, and removal of the hardware may be necessary to eradicate the infection. Rare cases of enterococcal pneumonia, lung abscess, and spontaneous empyema have been described.

TREATMENT