Enterococci Resistant to Multiple Antimicrobial Agents, Including Vancomycin: Establishment of Endemicity in a University Medical Center

Enterococci, particularly Enterococcus faecium, have always had a high intrinsic level of resistance to antimicrobial agents [1, 2]. It was first recognized in 1971 that enterococci could be optimally killed by taking advantage of the synergistic effect obtained when ampicillin or vancomycin is combined with an aminoglycoside [3-5]. However, by the 1980s, high-level resistance to gentamicin and other aminoglycosides was being seen with increasing frequency [1, 6, 7]. This discovery was followed in 1988 by the clinical isolation of an enterococcal strain that was resistant to vancomycin [8, 9]. From 1989 to 1993, the percentage of enterococcal isolates with vancomycin resistance reported to the National Nosocomial Infections Surveillance System of the Centers for Disease Control and Prevention (CDC) increased from 0.3% to 7.9% [10]. Although various resistance patterns have been reported, vancomycin-resistant E. faecium also commonly have ampicillin and high-level aminoglycoside resistance [1, 2]. This results in a bacterial strain that may be untreatable with currently available antimicrobial agents. Although vancomycin-resistant enterococci have been identified in more than 33 states, infections have occurred most frequently in hospitals in New York, Pennsylvania, and Maryland [10-12]. Patients infected or colonized with this organism were first detected in the Baltimore area in 1989; the first episode at our institution was reported in December of that year [13]. The appearance of these highly resistant strains prompted us to do a series of studies to assess the extent of colonization and infection with vancomycin-resistant enterococci in our patient population, to define risk factors for acquisition, and to evaluate the effect of interventions on rates of colonization and infection. Methods Data were collected from a university hospital (557 beds), cancer center (62 beds), and adjacent shock trauma center (138 beds). The university hospital serves both as a state and regional tertiary referral center and the primary source of care for the immediate West Baltimore community. During the study period (May 1992 to June 1994), 59 196 patients were admitted to these institutions. Beginning in May 1992, basic demographic and clinical data were collected on all patients from whom vancomycin-resistant enterococci were isolated. Standard CDC definitions were used to differentiate active infection from colonization [14]. Data on blood isolation of methicillin-resistant Staphylococcus aureus, coagulase-negative staphylococci, and vancomycin-susceptible enterococci were obtained by review of data from the hospital microbiology laboratory. Computerized pharmacy databases were used to determine the number of doses administered of selected antimicrobial agents, including vancomycin, ciprofloxacin, ceftriaxone, and ceftazidime. Point-Prevalence Surveys To determine the rate of the colonization of patient stool with vancomycin-resistant enterococci, we did a series of point-prevalence surveys. For each survey, we selected a random 30% sample of all persons who were on the inpatient census as of midnight on the dates selected for study (25 July, 16 August, and 27 September 1993). During the 3 days after each randomization date, efforts were made to collect a stool sample for culture from each patient selected in the randomization. Stool samples were also obtained from normal, healthy community volunteers who had been recruited as outpatients for unrelated vaccine trial studies. Case-Control Studies Two casecontrol studies were done in October 1993. In the first study, patients on the surgical intensive care and surgical intermediate care units from whom vancomycin-resistant enterococci had been isolated between May 1992 and October 1993 (case-patients) were compared with an unmatched random sample of patients who had been hospitalized on the same units during the same time period (controls). Controls could only be patients who had been hospitalized for at least 7 days (the minimum number of days between hospital admission and the isolation of vancomycin-resistant enterococci from a case-patient). In the second study, all adult patients with urine cultures positive for nosocomially acquired vancomycin-resistant enterococci who were identified between May 1992 and October 1993 (case-patients) were compared with an unmatched random sample of patients with urine cultures positive for nosocomially acquired vancomycin-susceptible enterococci (controls). Both studies included colonized and infected case-patients. Age; sex; diagnoses; and data on instrumentation, antimicrobial therapy, and severity of illness (measured by the Acute Physiologic and Chronic Health Evaluation II [APACHE II] score [15]) were obtained by chart review. For case-patients, data reflected events occurring before the positive culture was collected. For controls, data represented either the entire time the patient was in the intensive care or intermediate care unit (surgical intensive care unit study) or the hospital stay until the positive culture for vancomycin-sensitive enterococci (urinary tract infection study); the midpoint of the intensive care unit or hospital stay was designated as the day of positive culture for comparison with case-patients. Observational Studies Patients with vancomycin-resistant enterococci were placed in multidrug-resistant organism isolation, similar to the CDC's recommended contact isolation [16]. From 1 November to 14 November, compliance with isolation procedures was seen on four wards: two general medical and surgical wards, the surgical intensive care unit, and the surgical intermediate care unit. For twenty 30-minute periods, each room was observed and the occupation, duties, and patient or environment contacts of each person entering the room were recorded. We noted whether each patient's room had a sign indicating isolation for a multidrug-resistant organism, whether gowns and gloves were readily accessible outside the room, whether each person washed his or her hands before and after the visit, and whether gowns and gloves were worn in situations requiring their use. As part of these studies, hand cultures were obtained from health care workers on these four wards by having volunteers rub a sterile hand wipe moistened with 0.02% polysorbate 80 detergent over their hands for at least 30 seconds [17]. Selected environmental cultures were obtained by using sterile swabs moistened with sterile saline. Environmental and hand-wipe specimens were processed at the CDC and were cultured by using the membrane filter technique [17, 18]. Membrane filters were then cultured on m-Enterococcus agar [18]. Interventions Infection Control Measures Beginning in October 1993, routine admission and weekly stool surveillance cultures for vancomycin-resistant enterococci were obtained from all patients in the surgical intensive care and intermediate care units. Patients colonized or infected with vancomycin-resistant enterococci in these units were spatially segregated, and nurses caring for the patients were placed in cohorts. Nurses caring for culture-positive patients in the cancer center were also placed in cohorts. Restrictions on Vancomycin Use Beginning in December 1993, vancomycin use was restricted to 1) treatment of documented, culture-proven infections with organisms not susceptible to alternative agents [that is, methicillin-resistant S. aureus or coagulase-negative staphylococci]; 2) treatment of patients in whom there was a high index of suspicion of methicillin-resistant staphylococcal infections on the basis of history, surveillance cultures, or Gram stain; 3) treatment of Clostridium difficile infections in patients who had not responded to an initial course of therapy with metronidazole; 4) treatment of gram-positive infections in patients with documented, severe allergy to -lactam antimicrobial agents; 5) single doses of vancomycin just before hospital discharge in anephric patients with gram-positive infections; and 6) surgical prophylaxis as part of a previously approved investigational protocol for recipients of solid organ transplants. For all other uses, the patient's attending physician was required to obtain permission from the infectious disease attending physician on service. Compliance with this policy was monitored by the Department of Pharmacy Services. Vancomycin was not dispensed unless an appropriate indication was documented or the infectious disease attending physician approved use of the drug. If inappropriate therapy was identified, recommendations were made for the use of alternative agents. Microbiological Analysis Enterococci were identified by using standard microbiological methods, including hydrolyzing esculin and growth in 6.5% NaCl [19]. Identification of the organisms to the species level was done using the conventional test scheme defined by Facklam and Collins [20]. Antimicrobial susceptibilities to ampicillin, vancomycin, streptomycin, and gentamicin were determined by using the E-test quantitative minimum inhibitory concentration procedure (AB Biodisk, Solna, Sweden). Quality control strains of E. faecalis (ATCC 29212, 51299) were used to ensure the potency of each antimicrobial agent tested. With the exception of vancomycin, susceptibility interpretations followed the guidelines proposed by the National Committee for Clinical Laboratory Standards [21, 22]. For this study, vancomycin resistance was defined as any enterococcal isolate with a minimum inhibitory concentration to vancomycin of at least 16 g/mL. Vancomycin resistance was confirmed by hybridization to specific gene probes (see below). We used trypticase soy agar supplemented with 5% sheep blood (BBL, Cockeysville, Maryland) to isolate enterococci from urine, wounds, and sterile body fluids. For stool samples collected as part of the point-prevalence studies and for routine stool surveillance for vancomycin-resistant enterococci, sa