Robin Patel MD
Essentials of Diagnosis
Enterococcal infections also occur in hospitalized patients or in patients undergoing peritoneal or hemodialysis, and the organisms causing such infections often appear to be exogenously acquired. There is clear-cut evidence for spread of strains of enterococci between patients and even for dissemination of such strains from one institution to another. Strains of enterococci causing nosocomial infections have been found on the hands of medical personnel and on environmental surfaces in hospitals and in nursing homes. Most likely, enterococci from patients or hospital personnel first colonize the gastrointestinal tract (or occasionally the skin and groin or other contiguous areas) before causing infections in other patients. Devices such as electronic rectal thermometers may also aid in the spread of enterococci, especially antibiotic-resistant enterococci.
Enterococci account for 12% of nosocomial bacterial infections. The most common nosocomial enterococcal infections involve the urinary tract (62% of cases) and are often associated with urinary tract instrumentation or structural abnormalities. Wound infections are the next most common presentation of nosocomial enterococcal infections, usually involving abdominal or pelvic sites (25% of cases). Bacteremia (10% of cases) is the third most commonly reported nosocomial enterococcal infection.
Risk factors for acquiring nosocomial enterococcal infection or colonization, especially those caused by vancomycin-resistant enterococci (VRE), include being critically ill; having severe underlying disease; being immunosuppressed (especially patients on oncology or transplant wards); having renal insufficiency; undergoing intra-abdominal or cardiothoracic surgical or other invasive procedures; having an indwelling urinary or central venous catheter; having a prolonged hospital or intensive care unit stay; intrahospital transfer between floors; multi-antimicrobial agent, third-generation cephalosporin, anti-anaerobic antimicrobial or vancomycin therapy; proximity to infected patients; care by colonized staff; and the receipt of selective bowel decontamination, sucralfate, or enteral feedings.
Enterococci are characterized by intrinsic resistance to many antibiotics (eg, oxacillin, clindamycin, cephalosporins, and aminoglycosides) and the capacity to acquire resistance to many others. The uniformly poor activity of β-lactams, particularly cephalosporins, against enterococci results from reduced affinity of the penicillin-binding proteins. Unlike most streptococci, enterococci are inhibited but not killed by apparently active penicillins (ampicillin, mezlocillin, penicillin, piperacillin), carbapenems (imipenem, meropenem), and glycopeptides (vancomycin or teicoplanin). Cephalosporins are not clinically useful against enterococci.
Enterococci are also naturally resistant to clinically achievable levels of aminoglycosides used alone. For treating serious enterococcal infections, combination therapy is recommended, generally penicillin, ampicillin, or vancomycin with an aminoglycoside, usually gentamicin or streptomycin. The rationale for this approach is to weaken the cell wall with the cell wall-active agent, thereby facilitating entry of the then bactericidal aminoglycoside. Such combinations have been shown to be synergistic, as long as the organism is susceptible to the cell wall-active agent and does not demonstrate high-level resistance to the aminoglycoside (gentamicin minimum inhibitory concentration < 500 µg/mL, streptomycin minimum inhibitory concentration < 2000 µg/mL).
Acquired resistance in enterococci can develop by genetic mutation or by acquiring altered DNA from a resistant organism. Enterococci have been reported to have acquired high-level resistance to aminoglycosides, cell wall-active agents (including penicillins and vancomycin), chloramphenicol, clindamycin, erythromycin, and the newer quinolones. E faecium isolates exhibit more antibiotic resistance than do other species.
Most enterococci with high-level resistance to aminoglycosides contain aminoglycoside-modifying enzymes. When high-level resistance to aminoglycosides is a concern, only gentamicin and streptomycin minimum inhibitory concentrations need to be tested in the clinical microbiology laboratory because the enzyme that neutralizes gentamicin also modifies tobramycin, amikacin, kanamycin, and netilmicin but not streptomycin. In addition to enzymatic resistance, ribosomal modification has been described as a second mechanism for high-level streptomycin resistance.
Acquired resistance to ampicillin and penicillin is generally caused by altered penicillin-binding proteins. This confers higher levels of resistance than the intrinsic resistance present in enterococci. This mechanism confers resistance to all β-lactams (including the carbapenems). β-lactamase production among enterococci is rare and has been reported in a few hospitals.
In the background of increasing resistance of enterococci to penicillins and aminoglycosides, vancomycin resistance emerged in the late 1980s. In vancomycin-resistant cells, an abnormal pentapeptide peptidoglycan precursor, with which vancomycin interacts minimally, is synthesized and incorporated into the cell wall of the organism. Resistance to the glycopeptide antibiotic vancomycin in enterococci, as understood to date, is phenotypically and genotypically heterogeneous. Three glycopeptide resistance phenotypes, VanA, VanB, and VanC, have been described in enterococci; they can be distinguished by the level and inducibility of resistance to vancomycin and teicoplanin. VanA-type glycopeptide resistance is characterized by acquired inducible resistance to both vancomycin and teicoplanin, and it is transferable. VanB-type glycopeptide resistance is characterized by acquired inducible resistance to various concentrations of vancomycin but not to teicoplanin, and it is also transferable. VanC-type glycopeptide resistance is characterized by low-level vancomycin resistance but teicoplanin susceptibility and has been described as an intrinsic property of most isolates of E gallinarum, E casseliflavus, and E flavescens. VanD-, VanE- and VanG- type glycopeptide resistance have been recently described and do not appear to be common.
Although enterococci are found in one-fifth of intra-abdominal infections, their exact role in polymicrobial infection is controversial. In cases of intra-abdominal infections, selective therapy against Escherichia coli and Bacteroides fragilis, which has minimal in vitro activity against enterococci, has been found to be sufficient to reduce enterococcal counts. Nevertheless, in animal models of experimental polymicrobial intra-abdominal infection, enterococci have been found to enhance abscess formation, weight loss, and mortality. Similarly, clinical reports have indicated the emergence of enterococcal abscesses and bacteremia after treatment of intra-abdominal sepsis with antimicrobial agents that lack significant in vitro enterococcal activity. A recent multicenter study of intra-abdominal infection has found that the presence of enterococci in the initial cultures, in addition to serious underlying disease, independently predicts treatment failure with broad-spectrum antimicrobial regimens that lack specific enterococcal activity.
URINARY TRACT INFECTION
Urinary tract infections, including uncomplicated cystitis, pyelonephritis, prostatitis, and perinephric abscess, are the most common type of clinical infections produced by enterococci (Box 49-1). Most enterococcal urinary tract infections are nosocomial and are associated with urinary catheterization or instrumentation.
BACTEREMIA & ENDOCARDITIS
Nosocomial enterococcal bacteremias are commonly polymicrobial. Portals of entry for enterococcal bacteremia include the urinary tract, intra-abdominal or pelvic sources, wounds (especially burns, decubitus ulcers, and diabetic foot infections), intravascular catheters, and the biliary tree. Metastatic infections other than endocarditis are rare in enterococcal bacteremia.
Enterococci account for ~5–10% of all cases of infective endocarditis (see Chapter 11). Most cases are caused by E faecalis, but E faecium, E casseliflavus, E durans, E gallinarum, and E raffinosus have also been isolated from patients with endocarditis. Only ~2% of cases of enterococcal bacteremia are associated with endocarditis, and this is much more common in patients whose bacteremia is community acquired versus those with nosocomial enterococcal bacteremia. Enterococcal endocarditis is a disease of older patients, with males outnumbering females in most series. Most cases occur in patients with underlying valvular heart disease or prosthetic valves, although enterococci are capable of causing infections of anatomically normal valves. An association between enterococcal endocarditis and urinary tract infection or urinary instrumentation in older men and abortion or childbirth in younger women has been suggested. Enterococci usually produce left-sided endocarditis with more frequent involvement of the mitral than the aortic valve. The typical clinical course of enterococcal endocarditis is that of subacute bacterial endocarditis. There is a suggestion that the relapse rate is higher in patients who have had symptoms of endocarditis for ≥ 3 months before treatment, which has implications for length of therapy (see below).
BOX 49-1 Enterococcal Infections
INTRA-ABDOMINAL & PELVIC INFECTION
Enterococci are frequently found as part of mixed aerobic and anaerobic flora in intra-abdominal and pelvic infections. As discussed above, the exact role of enterococci in these mixed aerobic and anaerobic intra-abdominal and pelvic infections is unclear. Enterococci can cause spontaneous peritonitis in patients with nephrotic syndrome or cirrhosis and can cause peritonitis in patients undergoing chronic ambulatory peritoneal dialysis. Monomicrobial enterococcal peritonitis is also occasionally seen as a complication of abdominal surgery or trauma. Enterococci can produce abscesses and bacteremia as a complication of endometritis, Cesarean section, and acute salpingitis.
WOUND & SOFT TISSUE INFECTION
Enterococci may be isolated from mixed cultures with gram-negative bacilli and anaerobes in surgical wound infections, decubitus ulcers, and diabetic foot infections; the significance of enterococci in these settings is difficult to assess. Enterococcal wound colonization and sepsis have been described in burn patients whose wounds have been covered with porcine xenografts presumably contaminated with enterococci. Enterococci occasionally cause chronic osteomyelitis.
Enterococci rarely cause meningitis. Most cases occur in patients who have anatomic defects of the central nervous system or who have had previous neurosurgery or head trauma. Rarely, however, meningitis may be a complication of high-grade bacteremia such as that seen in patients with enterococcal endocarditis. Meningitis may also be seen in association with enterococcal bacteremia in patients with human immunodeficiency virus infection and acute leukemia and in neonates and transplant recipients.
Enterococci may cause neonatal sepsis characterized by fever, lethargy, and respiratory difficulty accompanied by bacteremia, meningitis, or both. Early onset bacteremia in otherwise normal neonates may be seen, as well as nosocomial bacteremia, meningitis, or both, which has been described in premature or low-birth-weight neonates who have nasogastric tubes and intravascular devices.
The diagnosis of enterococcal infection is made by isolating enterococci from typically sterile sites (eg, urine, blood, intra-abdominal or pelvic fluid, or spinal fluid). As discussed above, community acquisition of enterococcal bacteremia suggests the possible presence of endocarditis. The diagnosis of endocarditis, in the appropriate clinical context (see Chapter 11), may be further confirmed by the presence of clinical findings consistent with endocarditis as well as by the use of echocardiography. The diagnosis of intra-abdominal and pelvic infections, in the appropriate clinical context, can be assisted by the use of ultrasound or computed tomographic imaging with drainage, Gram stain, and culture of any fluid collection(s) present.
The diagnosis of enterococcal meningitis can be confirmed by the isolation of enterococci from the spinal fluid. Typically, enterococcal meningitis is associated with a low cerebrospinal fluid leukocyte count (< 200/mL), although this is not always the case.
The isolation of enterococci from respiratory secretions is of questionable significance, although there are very rare well-documented cases of enterococcal pneumonia and even lung abscesses in patients with severe and debilitating diseases.
Penicillin or ampicillin remains the antibiotic of choice for treating enterococcal infections such as urinary tract infections, peritonitis, and wound infections (ie, infections that do not require bactericidal treatment) (Box 49-2). Vancomycin is the alternative agent for patients allergic to penicillin or for organisms with high-level penicillin resistance that are β-lactamase negative and vancomycin susceptible.
Combination therapy with a cell wall-active agent such as penicillin, ampicillin, or vancomycin, along with an aminoglycoside, is essential for the treatment of enterococcal endocarditis and probably for enterococcal meningitis as well. The situation is not as clear-cut for enterococcal bacteremia alone; there is no consensus as to whether combination therapy is required in this setting.
For the treatment of enterococcal endocarditis, combinations of cell wall-active agents such as penicillin, ampicillin, or vancomycin with aminoglycosides (usually streptomycin or gentamicin) are required (see Chapter 11 for specific treatment recommendations). In most cases, 4 weeks of combination therapy appear to be adequate, with a 6-week regimen reserved for patients who have had symptoms for > 3 months before starting treatment, for patients with prosthetic valves, or for patients who have relapsed after previous shorter courses of therapy.
Most strains of enterococci are susceptible to nitrofurantoin, and this agent has been successfully used to treat uncomplicated enterococcal infections limited to the urinary tract. The quinolones such as ciprofloxacin, ofloxacin, levofloxacin, gatifloxacin, enoxacin, and norfloxacin have in vitro activity against enterococci and may be useful for treating some enterococcal urinary tract infections, but their effectiveness for enterococcal infections in general has not been convincingly demonstrated, and increasing resistance has been demonstrated. Tetracyclines and chloramphenicol may exhibit in vitro activity against some strains of enterococci, but they are only bacteriostatic against enterococci, and clinical failures of chloramphenicol and tetracyclines have been reported.
The emergence of multiply resistant enterococci now greatly complicates therapeutic choices. For patients who have endocarditis caused by enterococci with high-level gentamicin resistance, it is useful to test for high-level streptomycin resistance because some highly gentamicin-resistant strains are synergistically killed by cell wall-active agents plus streptomycin. For endocarditis due to strains with both high-level streptomycin and gentamicin resistance, no combination will consistently produce synergism. These patients may be treated with prolonged courses (8–12 weeks) of intravenous ampicillin given by continuous infusion, although this approach should be considered experimental at this time. Surgical excision of infected valves may be required in such cases.
Infections caused by enterococci with high-level penicillin resistance and β-lactamase negativity should be treated with vancomycin. β-lactamase–producing, penicillin- and ampicillin-resistant enterococci remain susceptible to β-lactam–β-lactamase inhibitor combinations such as ampicillin-sulbactam and amoxicillin-clavulanate. VRE that remain susceptible to penicillin or ampicillin may be treated with penicillin or ampicillin. Infections caused by organisms with both high-level penicillin resistance and vancomycin resistance should be treated with either quinupristin-dalfopristin or linezolid. Notably, E faecium is more susceptible to quinupristin-dalfopristin than is E faecalis. Neither quinupristin-dalfopristin nor linezolid have bactericidal activity against VRE resulting in enormous challenges in the management of VRE endocarditis.
Prevention & Control
Prophylaxis for Infective Endocarditis. Antimicrobial prophylaxis for enterococcal endocarditis is recommended for patients with certain cardiac lesions predisposing to endocarditis who undergo invasive procedures with an increased risk of enterococcal bacteremia. Details of patients, procedures, and regimens are outlined in Chapter 11.
BOX 49-2 Treatment of Enterococcal Infectionsg1
Prevention of Vancomycin Resistance. From 1989 to 1993, the percentage of nosocomial enterococcal infections reported to the Centers for Disease Control and Prevention's National Nosocomial Infection Surveillance System that were caused by VRE increased from 0.3 to 7.9%. This overall increase primarily reflected the 34-fold increase in the percentage of VRE infections in patients in intensive care units (ie, from 0.4 to 13.6%), although a trend toward an increased percentage of VRE infections in non-intensive care patients was also noted.
Laboratory experiments have achieved the transfer of high-level vancomycin resistance from enterococci to Staphylococcus aureus, and reduced susceptibility of S aureus to vancomycin has recently been described in clinical S aureus isolates, although the mechanism of the latter resistance differs from that in enterococci. Vancomycin resistance has also been transferred by conjugation or transformation from enterococci to Streptococcus sanguis, Lactococcus lactis, Streptococcus pyogenes, and Listeria monocytogenes. The vanA gene has been found in vancomycin-resistant clinical isolates of Oerskovia turbata and Arcanobacterium haemolyticum (typically these organisms are vancomycin susceptible) isolated from the stools of two patients during an outbreak of VRE infection in London. The vanA gene has been identified in a Bacillus circulans clinical isolate. The vanB gene has recently been found in a vancomycin-resistant isolate of Streptococcus bovis isolated from a stool swab collected from a patient on admission as surveillance for VRE. The potential for emergence of vancomycin resistance in clinical isolates of Staphylococcus epidermidis, Streptococcus pneumoniae, viridans streptococci, and Corynebacterium spp. as a result of transfer of vancomycin resistance genes from enterococci is also a public health concern.
Table 49-1. Recommendations for preventing the spread of vancomycin resistance: prudent vancomycin use.
In November 1994, the Hospital Infection Control Practices Advisory Committee of the Centers for Disease Control and Prevention issued the following recommendations for preventing and controlling the spread of vancomycin resistance with a special focus on VRE. It was recommended that each hospital through collaboration of its quality improvement and infection control programs; pharmacy and therapeutics committee; microbiology laboratory; clinical departments; and nursing, administrative, and housekeeping services, develop a comprehensive institution-specific strategic plan to detect, prevent, and control infection and colonization with VRE. The first aspect of this plan incorporated prudent vancomycin use. (Vancomycin has been reported as a risk factor for infection and colonization with VRE.) (Table 49-1.)
BOX 49-3 Prevention of VRE Transmission
In addition, education programs concerning the epidemiology of VRE and the potential impact of this pathogen on the cost and outcome of patient care were recommended for hospital staff including attending and consulting physicians; medical residents and students; pharmacy, nursing, and laboratory personnel; and other direct patient care providers. Specific recommendations were also issued for the microbiology laboratory regarding the detection, reporting, and control of VRE. In addition, recommendations were given regarding how to proceed when VRE are isolated from a clinical specimen, how to screen for detection of VRE, and how to prevent transmission of VRE once identified (Box 49-3). Recommendations for screening for roommates found to be infected or colonized with VRE were also issued.
Centers for Disease Control and Infection: Nosocomial enterococci resistant to vancomycin—United States, 1989–1993. Morbid Mortal Weekly Rep 1993; 42: 597–99.
Hospital Infection Control Practices Committee. 1995 Recommendations for preventing the spread of vancomycin resistance. J Infect Control Hosp Epidemiol 1995;6: 105–13.