- Email: [email protected]
Epidemiology and control of vancomycin-resistant enterococci in an adult and children's hospital
E p i d e m i o l o g y and control of v a n c o m y c i n - r e s i s t a n t e n t e r o c o c c i in an adult and children's hospital i
Victor Tucci, MPH Mary Ann Haran, RN, BSN Henry D. Isenberg, PhD New Hyde Park, New York
Background: The incidence of vancomycin-resistant enterococci (VRE) has reached endemic proportions in many medical centers. To initiate an effective infection control program, an understanding of the epidemiologic attributes of the genus in medical facilities is imperative. Methods: We studied 138 consecutive cases of VRE from April through December 1995. We created a database to analyze the risk factors for patients in both an adult hospital and a children's hospital and screened all specimens, submitted for routine microbiologic analysis, for VRE. Results: One hundred twenty-three cases (89%) occurred in the adult acute care hospital, and 15 (11%) occurred in the children's hospital. Eighty patients (58%) were colonized with VRE, and 58 (42%) had an infection with VRE. Eighty-three percent of all the cases of VRE were nosocomially acquired. The majority of cases occurred in the medical service. Urine was the most important clinical specimen infected or colonized. Prior use of an antibiotic, other than vancomycin, was the most important risk factor for all nosocomial cases, followed by prior vancomycin use for surgical patients and residence in a unit with other patients infected with VRE for the medical service. Direct admission from another hospital was the most important risk factor for community-acquired cases. Special microbiologic screening of cultures yielded 48% of all VRE identified. Enterococcus faecium was the predominant resistant isolate recovered. Conclusions: The control of VRE in the hospital setting is difficult for several reasons. Almost half of all patients carrying VRE would not have been identified without special microbiologic screening efforts, as would patients, admitted from the community, who are already colonized with VRE. Controlling antibiotic use both in the hospital and the community is basic for controlling these organisms. Continuous education of all staff about VRE and other nosocomially significant organisms is the key to controlling the spread of these bacteria. (AJIC Am J Infect Control 1997;25:371-6)
Infection tients with (VRE) have wide since of VRE in
and colonization of hospitalized pavancomycin-resistant enterococci shown a dramatic increase nation1989.1 W e o b s e r v e d o u r f i r s t c a s e the oncology unit of the 150-bed
From the Epidemiology Section of the Microbiology Division, Pathology Department, Long Island Jewish Medical Center, The Long Island Campus of the Albert Einstein College of Medicine. Reprint requests: Henry D. Isenberg, PhD, Long Island Jewish Medical Center 270-05 76th Ave., New Hyde Park, NY 11040. Copyright © 1997 by the Association for Professionals in Infection Control and Epidemiology, Inc. 0196-6553/97 $5.00 + 0
17/46/77975
Schneider Children's Hospital (SCH) during M a r c h 1990. I n 1991 w e r e p o r t e d o n o u r e f f o r t s t o control a cluster epidemic in this oncology unit. 2 Use of contact isolation of patients with cultures positive for VRE, screening stool and urine cultures for VRE, and limiting the use of vancomycin proved to be effective control measures. In the 550-bed adult hospital, Long Island Jewish Medical Center (LIJMC), we recorded only seven c a s e s o f V R E d u r i n g 1991. S i n c e t h e n , t h e i n c i dence of VRE has reached endemic proportions, and treatment options for clinicians have dwind l e d r a p i d l y . D u r i n g 1995 i n t e n s e i n f e c t i o n c o n trol activity was instituted according to the guidelines issued by Hospital Infection Control
371
AJIC
372 Tucci et al.
October 1997
70% I
64%
Medicine
Surgery
N
3~
I
..........
I
....................
I
Fig. 1. Distribution of VRE by service.
Practices Advisory Committee. 3 To initiate an effective infection control program for VRE, an understanding of the epidemiologic attributes of the genus in hospitalized patients is imperative. The data presented here are specific to our teaching hospital. It behooves individual institutions to collect and analyze their own data to understand the distribution of organisms such as VRE and to develop control measures to interrupt transmission. In April 1995, we established a prospective database of all patients infected or colonized with VRE. This report proffers our findings of 138 cases observed in both the children's and adult hospitals from April through December 1995 as illustration of one medical center's effort to control the nosocomial spread of VRE. METHODS
To a c c o m m o d a t e the Hospital Infection Control Advisory Committee guidelines, the Microbiology Division initiated screening of all cultures, including stool and low-count urine cultures, for VRE by including a selective enterococcal agar (BBL; Becton Dickinson Microbiology Systems, Cockeysville, Md.) with 10 gg/ml vancomycin. All VREs were identified to species level, and a complete antibiotic susceptibility profile was obtained by using standard microbiologic methods. 4 The Epidemiology Department staff surveyed all microbiology reports daily to identify and record VRE cases. A case report was generated by re-
viewing the patient's medical record, interviewing appropriate health care staff, and listing all pertinent information in a customized database consisting of admission date, service, diagnosis, procedures, locations, treatments including antibiotics, culture results, use of Foley catheters, central lines, intubation, and whether the organism was acquired nosocomially. In addition, the charge nurse of each affected unit was notified in person and in writing about each patient with cultures positive for VRE and was provided with infection control suggestions (e.g., placing the patient in a single room if possible and instituting contact isolation). We defined infection versus colonization according to the National Nosocomial Infection Study: "Infection is the presence of an organism(s) in body tissues, or fluids accompanied by a clinically adverse effect (either locally or systemically) on the host. Colonization is the persistence of organisms on skin, in body tissues, or in body fluids but without a clinically adverse effect. ''5 RESULTS AND DISCUSSION
During the 9 months of the study, VRE was detected in 138 patients. One hundred twenty-three patients (89%) were in the adult hospital (LIJMC) and 15 (11%) were in SCH. Of the total number, 80 (58%) were colonized and 58 (42%) were infected. Although 52% of the cases were detected by routine analysis of submitted specimens, 48% were
AJ|C
Tucci et
Volume 25, Number 5
al. 373
5000%
40.06%
io ~ 3o.0o%
20.0(~
10.00%
0.00% bl~:d
wounds
~tne mspiralo~ line assoc 51 Inf~t~l (S'/l~mlat~). 63 oolonlzed (72 I ~ a t ~ )
feces
Fig, 2, Distribution of nosocomially transmitted cases by body site.
70%
50%
j
40%-
11)%
0% blood
wounds
urine ~piratoex ~ mmo 7 Infected (8 I~olat~). 11 colonized (24 bmlams|
~
Fig. 3. Distribution of community-acquired cases by body sites.
identified only by the special screening of feces, sputum, and low-count urine specimens for VRE. Nosocomial acquisition of VRE was demonstrated in 114 of 138 patients (83%), whereas 24 (17%) cases of VRE were identified on admission. Fig. 1 shows the distribution of VRE by service. The majority of the cases occurred in the adult medical service (64%), followed by 24% in the adult surgical service. Pediatric patients accounted for 8%; 3% were noted in the separate neonatal (premature) unit. Only 1% of the cases were encountered in the obstetrics and gynecology unit.
The distribution by body sites of the nosocomially transmitted cases (114) is shown in Fig. 2. Urine was the primary clinical source of both infection and colonization, followed by wounds in the infected patients. Feces and central line tips were the next frequently colonized sites. The distribution by sites of the community-acquired cases (24) is shown in Fig. 3. Again urine was most frequently involved for infected and colonized cases. The respiratory tract and feces were also important colonized sites in this category. As the figures indicate, enterococci were frequently isolated from more than one body site of the same patient.
AJIC
3 7 4 Tucci et al.
October 1997
80%.
70% -
6O%
& 20% -
10%
0%
on unit
pci~ yahoo I m r g ¢ l ~ ~II C I ~
ph~r antlb
pdor Icu pcu
mcMIk~ll1~ 7 t ¢ ~ 1 1 0
p~o¢ abd surgery
OIIIwt ~ m c m
Fig. 4. Risk factors studied for nosocomial cases, vanco, Vancomycin; abd, abdominal.
o
i
f
from nsg I'~rle
an~hlr hl:ll~ld
i~t h~rm 111m ~ k l l .
dlech UJ > 1 mort
Jl4 trod
Fig. 5. Risk factors for all nonnosocomial cases, nsg, Nursing; pvt, private; disch, discharged from.
The risk factors studied for the nosocomial cases (Fig. 4) were (1) individual in a unit at the same time as other patients with cultures positive for VRE, (2) prior vancomycin use, (3) prior use of other antibiotics, (4) prior stay in the intensive care unit (ICU) or progressive care unit (PCU), and (5) prior abdominal surgery. The risk factors for the nosocomial cases were analyzed for medical, surgical, and all cases. Fig. 4 also displays the important risk factors for surgical patients in decreasing order: (1) prior use of an antibiotic other than vancomycin, (2) prior stay in the ICU or PCU, (3) prior vancomycin use, followed by (4) residence in a unit where other patients had VRE,
and finally, (5) prior abdominal surgery. The risk factors important for medical patients were (1) prior antibiotic use, followed by (2) residence in a unit with other patients with cultures positive for VRE, (3) prior vancomycin use, (4) prior stay in the ICU or PCU. The use of multiple antibiotic drugs is known to select the more resistant endogenous organisms or encourage resistant hospital microorganisms to become part of the patient's microbiota. As Fig. 4 indicates, antibiotic use was the most important risk factor for all nosocomially acquired VRE. The risk factors for all nonnosocomial cases (Fig. 5) in decreasing order of importance were:
AJIC
Volume 25, Number 5
T~CCi et
aI. 375
70-
60
50
|° 30-
20
wound
u~
rmp.
n mm'Om'yand 30 m ~
line moooc
~
I ~
Fig. 6. Distribution of VRE in the ICU-PCU population for both medical and surgical patients.
(1) admission directly from another hospital, (2) admission from h o m e in the c o m m u n i t y (domicile), followed by (3) admission from a nursing home, and (4) readmission after discharge from LIJMC in excess of 1 month. A total of 165 VRE isolates were recovered from 138 patients. Species distribution was as follows: 130 (79%) were Enterococcus faecium, 14 (8.5%) were Enterococcus faecalis, 9 (5.5%) were Enterococcus durans, 1 (0.6%) was Enterococcus avium, and 11 (6.7%) were not identified to species level. Control of microbial spread among hospitalized patients requires careful epidemiologic analysis. Many factors have contributed to VRE emergence (e.g., inappropriate use of antibiotics not only in hospitals b u t also in the community, u n k n o w n level of colonization of patients with VRE, lack of understanding or refusal to understand the significance of VRE recovered from hospitalized patients, and absence of consistent infection control practices). At LIJMC, other important factors were considered and include the increase in methicillin-resistant Staphylococcus aureus and 6l a c t a m - r e s i s t a n t coagulase-negative staphylococci in the same population, requiring the use of vancomycin, and the increase in the n u m b e r of patients with Clostridium difficile intoxication who, until recently, were treated primarily with oral vancomycin (unpublished data). The VRE problem in our medical center is primarily in the adult patients, whereas the initial introduction of the organism was in our children's hospital. The initial group of enterococci from
SCH belonged to a single clone, but we n o w deal with multiclonal representatives of the genus. 2,6 Most likely, the initial episode involved the introduction of one VRE into a separate oncology unit in SCH where better adherence to infection control practices in a closed special unit, greater involvement of the professional staff, and better control of antibiotic use were readily achieved. At LIJMC the isolation of VRE was primarily from medical patients (Fig. 1), and urine was the most frequent site of infection or colonization (Fig. 2). With different strains of VRE gaining access to the facility, environmental contamination and cross-contamination from patient to patient play a major role in the transmission of VRE, as demonstrated by VRE acquisition during residence in a unit with other patients infected or colonized with VRE. Of course, enterococci constitute part of the n o r m a l h u m a n intestinal microbiota and can contaminate a variety of articles in patients' rooms (e.g., bed rails and commodes). 7 Improved housekeeping and strict adherence to universal precautions should decrease the spread of not only VRE but other important nosocomially acquired organisms. In addition, 26 of 36 (72%) of patients with urine cultures designated as infected had an indwelling Foley catheter before the positive culture, and 13 of 32 (40%) of those designated as colonized had been catheterized. This again points to the role of health care workers and prosthetic devices in the transmission of and complication by VRE and other nosocomially significant microorganisms.
AJIC
376 Tucci et al.
Prior use of vancomycin has been identified by a n u m b e r of authors as an important risk factor for acquiring g R g . 3,s In our survey, vancomycin use was only the third most important risk factor, after therapy with other antibiotics and residence in a unit with other patients infected or colonized with VRE. However, we had restricted use of vancomycin at the time of the survey, which might have affected the hierarchy of significance. The probability that multiple strains of VRE are involved is based on a previous study that demonstrated at least 15 different molecular clones of VRE in our institution. 6 Prior stay in the ICU or PCU was demonstrated to be an important risk factor primarily for surgical patients (Fig. 4). We evaluated the distribution of VRE in the ICU-PCU population for both medical and surgical patients (Fig. 6). Urine, wounds, and central lines were equally important sites for surgical patients; whereas urine was by far the most important site of infection or colonization for the medical patients in the IeUs. These differences emphasize that patients in the same unit can have different sites involved when exposed to the particular practices of different services. Infection control practices should be tailored to take account of these differences. The control of VRE in the hospital setting is made more difficult because almost half of the patients carrying VRE would not have been identified in the absence of microbiologic surveillance of clinical specimens for VRE. In our survey 79 of 165 (48%) of all VRE isolates were recovered in this m a n n e r . Eradication of VRE from hospitalized adult patients is difficult for several reasons. It m a y be red u n d a n t to point out that enterococci are normal residents of the h u m a n intestine. Selective pressures such as nonmedical use of antibiotics in the c o m m u n i t y m a y have shifted the ratio of E. faecalls and E. faecium reported at 87% to 9%, respectively, 9 in favor of E. faecium; and our results suggest that any prior antibiotic use tends to select VRE, leading to the admission of patients colonized and/or infected with VRE. Control of VRE in this population can be accomplished only if there is hospital-wide cooperation of all services and management. In our institution we have made the control of VRE a Quality Management issue. It is well known that improper use of antibiotics selects various resistant microorganisms that can complicate the recovery of patients. Our results clearly demonstrate that antibiotic use is the most important risk factor for acquiring VRE. We at-
October 1997
tempt to control antibiotic use by limiting the length of time between ordering antibiotics and by having the Infectious Disease Service approve the use of m a n y antibiotic drugs including vancomycin. This program has had limited success because of lack of sufficient personnel to follow up all approval requests thoroughly to determine whether the antibiotic use was appropriate on the basis of culture results or clinical presentation. Continuous education of staff about VRE and other nosocomially significant organisms is the key to controlling the spread of these bacteria. Notification of the charge nurse in person, by the epidemiologist who provided infection control suggestions, increased the awareness of all staff members. In addition, meetings were held with housekeeping and patient support managers to educate them about VRE and other nosocomial significant organisms and discuss methods tO improve cleaning of patients' rooms and equipment. The challenge will be to continue this task while shrinking health care dollars are forcing hospitals to cut costs by reducing the n u m b e r s of employees. This strategy could result in compromising the control of nosocomial disease. References 1. Center of Disease Control and Prevention. Nosocomial enterococci resistant to vancomycin--United States, 19891993. MMWR 1993;42:597-9. 2. Rubin LG, Tucci V, Cercenado E, Eliopoulos G, Isenberg HD. Vancomycin-resistant Emerococcus fiaecium in hospitalized children. Infect Control Hosp Epidemiol 1992; 13:700-5. 3. Centers for Disease Control and Prevention. Recommendations for preventing the spread of vancomycin resistance. Recommendations of the Hospital Infection Control Practices Advisory Committee (HIPAC). MMWR 1995; 44(RR-12). 4. Isenberg HD. Clinical microbiology procedure handbook. Washington (DC): American Society for Microbiology; 1992. 5. Center for Infectious Disease, Centers for Disease Control. Guidelines for the prevention and control of nosocomial infections. Atlanta (GA): Centers for Disease Control; Feb 1981. 6. Lam S, Singer C, Tucci V, Morthland V, Pfaller M, Isenberg HD. The challenge of vancomycin-resistant enterococci: a clinical and epidemiologic Study. AJIC Am J Infect Control 1995;23:170-80. 7. Yamaguchi E, Valena F, Smith S, Simmons A, Eng R. Colonization pattern of vancomycin-resistant Enterococcus faecium. AJIC Am J Infect Control 1994;22:202-6. 8. Karanfil L, Murphy M, Josephson A, Gaynes R, Mandel L, Hill B, et al. Cluster of vancomycin-resistant Enterococcus faecium in an intensive care unit. Infect Control Hosp Epidemiol 1992;13:195-200. 9. Ruoff KL, De La Maza L, Murtagh MJ, Spargo JD, Ferraro MJ. Species identities of enterococci isolated from clinical specimens. J Clin Microbiol 1990;28:435-7.
Victor Tucci, MPH Mary Ann Haran, RN, BSN Henry D. Isenberg, PhD New Hyde Park, New York
Background: The incidence of vancomycin-resistant enterococci (VRE) has reached endemic proportions in many medical centers. To initiate an effective infection control program, an understanding of the epidemiologic attributes of the genus in medical facilities is imperative. Methods: We studied 138 consecutive cases of VRE from April through December 1995. We created a database to analyze the risk factors for patients in both an adult hospital and a children's hospital and screened all specimens, submitted for routine microbiologic analysis, for VRE. Results: One hundred twenty-three cases (89%) occurred in the adult acute care hospital, and 15 (11%) occurred in the children's hospital. Eighty patients (58%) were colonized with VRE, and 58 (42%) had an infection with VRE. Eighty-three percent of all the cases of VRE were nosocomially acquired. The majority of cases occurred in the medical service. Urine was the most important clinical specimen infected or colonized. Prior use of an antibiotic, other than vancomycin, was the most important risk factor for all nosocomial cases, followed by prior vancomycin use for surgical patients and residence in a unit with other patients infected with VRE for the medical service. Direct admission from another hospital was the most important risk factor for community-acquired cases. Special microbiologic screening of cultures yielded 48% of all VRE identified. Enterococcus faecium was the predominant resistant isolate recovered. Conclusions: The control of VRE in the hospital setting is difficult for several reasons. Almost half of all patients carrying VRE would not have been identified without special microbiologic screening efforts, as would patients, admitted from the community, who are already colonized with VRE. Controlling antibiotic use both in the hospital and the community is basic for controlling these organisms. Continuous education of all staff about VRE and other nosocomially significant organisms is the key to controlling the spread of these bacteria. (AJIC Am J Infect Control 1997;25:371-6)
Infection tients with (VRE) have wide since of VRE in
and colonization of hospitalized pavancomycin-resistant enterococci shown a dramatic increase nation1989.1 W e o b s e r v e d o u r f i r s t c a s e the oncology unit of the 150-bed
From the Epidemiology Section of the Microbiology Division, Pathology Department, Long Island Jewish Medical Center, The Long Island Campus of the Albert Einstein College of Medicine. Reprint requests: Henry D. Isenberg, PhD, Long Island Jewish Medical Center 270-05 76th Ave., New Hyde Park, NY 11040. Copyright © 1997 by the Association for Professionals in Infection Control and Epidemiology, Inc. 0196-6553/97 $5.00 + 0
17/46/77975
Schneider Children's Hospital (SCH) during M a r c h 1990. I n 1991 w e r e p o r t e d o n o u r e f f o r t s t o control a cluster epidemic in this oncology unit. 2 Use of contact isolation of patients with cultures positive for VRE, screening stool and urine cultures for VRE, and limiting the use of vancomycin proved to be effective control measures. In the 550-bed adult hospital, Long Island Jewish Medical Center (LIJMC), we recorded only seven c a s e s o f V R E d u r i n g 1991. S i n c e t h e n , t h e i n c i dence of VRE has reached endemic proportions, and treatment options for clinicians have dwind l e d r a p i d l y . D u r i n g 1995 i n t e n s e i n f e c t i o n c o n trol activity was instituted according to the guidelines issued by Hospital Infection Control
371
AJIC
372 Tucci et al.
October 1997
70% I
64%
Medicine
Surgery
N
3~
I
..........
I
....................
I
Fig. 1. Distribution of VRE by service.
Practices Advisory Committee. 3 To initiate an effective infection control program for VRE, an understanding of the epidemiologic attributes of the genus in hospitalized patients is imperative. The data presented here are specific to our teaching hospital. It behooves individual institutions to collect and analyze their own data to understand the distribution of organisms such as VRE and to develop control measures to interrupt transmission. In April 1995, we established a prospective database of all patients infected or colonized with VRE. This report proffers our findings of 138 cases observed in both the children's and adult hospitals from April through December 1995 as illustration of one medical center's effort to control the nosocomial spread of VRE. METHODS
To a c c o m m o d a t e the Hospital Infection Control Advisory Committee guidelines, the Microbiology Division initiated screening of all cultures, including stool and low-count urine cultures, for VRE by including a selective enterococcal agar (BBL; Becton Dickinson Microbiology Systems, Cockeysville, Md.) with 10 gg/ml vancomycin. All VREs were identified to species level, and a complete antibiotic susceptibility profile was obtained by using standard microbiologic methods. 4 The Epidemiology Department staff surveyed all microbiology reports daily to identify and record VRE cases. A case report was generated by re-
viewing the patient's medical record, interviewing appropriate health care staff, and listing all pertinent information in a customized database consisting of admission date, service, diagnosis, procedures, locations, treatments including antibiotics, culture results, use of Foley catheters, central lines, intubation, and whether the organism was acquired nosocomially. In addition, the charge nurse of each affected unit was notified in person and in writing about each patient with cultures positive for VRE and was provided with infection control suggestions (e.g., placing the patient in a single room if possible and instituting contact isolation). We defined infection versus colonization according to the National Nosocomial Infection Study: "Infection is the presence of an organism(s) in body tissues, or fluids accompanied by a clinically adverse effect (either locally or systemically) on the host. Colonization is the persistence of organisms on skin, in body tissues, or in body fluids but without a clinically adverse effect. ''5 RESULTS AND DISCUSSION
During the 9 months of the study, VRE was detected in 138 patients. One hundred twenty-three patients (89%) were in the adult hospital (LIJMC) and 15 (11%) were in SCH. Of the total number, 80 (58%) were colonized and 58 (42%) were infected. Although 52% of the cases were detected by routine analysis of submitted specimens, 48% were
AJ|C
Tucci et
Volume 25, Number 5
al. 373
5000%
40.06%
io ~ 3o.0o%
20.0(~
10.00%
0.00% bl~:d
wounds
~tne mspiralo~ line assoc 51 Inf~t~l (S'/l~mlat~). 63 oolonlzed (72 I ~ a t ~ )
feces
Fig, 2, Distribution of nosocomially transmitted cases by body site.
70%
50%
j
40%-
11)%
0% blood
wounds
urine ~piratoex ~ mmo 7 Infected (8 I~olat~). 11 colonized (24 bmlams|
~
Fig. 3. Distribution of community-acquired cases by body sites.
identified only by the special screening of feces, sputum, and low-count urine specimens for VRE. Nosocomial acquisition of VRE was demonstrated in 114 of 138 patients (83%), whereas 24 (17%) cases of VRE were identified on admission. Fig. 1 shows the distribution of VRE by service. The majority of the cases occurred in the adult medical service (64%), followed by 24% in the adult surgical service. Pediatric patients accounted for 8%; 3% were noted in the separate neonatal (premature) unit. Only 1% of the cases were encountered in the obstetrics and gynecology unit.
The distribution by body sites of the nosocomially transmitted cases (114) is shown in Fig. 2. Urine was the primary clinical source of both infection and colonization, followed by wounds in the infected patients. Feces and central line tips were the next frequently colonized sites. The distribution by sites of the community-acquired cases (24) is shown in Fig. 3. Again urine was most frequently involved for infected and colonized cases. The respiratory tract and feces were also important colonized sites in this category. As the figures indicate, enterococci were frequently isolated from more than one body site of the same patient.
AJIC
3 7 4 Tucci et al.
October 1997
80%.
70% -
6O%
& 20% -
10%
0%
on unit
pci~ yahoo I m r g ¢ l ~ ~II C I ~
ph~r antlb
pdor Icu pcu
mcMIk~ll1~ 7 t ¢ ~ 1 1 0
p~o¢ abd surgery
OIIIwt ~ m c m
Fig. 4. Risk factors studied for nosocomial cases, vanco, Vancomycin; abd, abdominal.
o
i
f
from nsg I'~rle
an~hlr hl:ll~ld
i~t h~rm 111m ~ k l l .
dlech UJ > 1 mort
Jl4 trod
Fig. 5. Risk factors for all nonnosocomial cases, nsg, Nursing; pvt, private; disch, discharged from.
The risk factors studied for the nosocomial cases (Fig. 4) were (1) individual in a unit at the same time as other patients with cultures positive for VRE, (2) prior vancomycin use, (3) prior use of other antibiotics, (4) prior stay in the intensive care unit (ICU) or progressive care unit (PCU), and (5) prior abdominal surgery. The risk factors for the nosocomial cases were analyzed for medical, surgical, and all cases. Fig. 4 also displays the important risk factors for surgical patients in decreasing order: (1) prior use of an antibiotic other than vancomycin, (2) prior stay in the ICU or PCU, (3) prior vancomycin use, followed by (4) residence in a unit where other patients had VRE,
and finally, (5) prior abdominal surgery. The risk factors important for medical patients were (1) prior antibiotic use, followed by (2) residence in a unit with other patients with cultures positive for VRE, (3) prior vancomycin use, (4) prior stay in the ICU or PCU. The use of multiple antibiotic drugs is known to select the more resistant endogenous organisms or encourage resistant hospital microorganisms to become part of the patient's microbiota. As Fig. 4 indicates, antibiotic use was the most important risk factor for all nosocomially acquired VRE. The risk factors for all nonnosocomial cases (Fig. 5) in decreasing order of importance were:
AJIC
Volume 25, Number 5
T~CCi et
aI. 375
70-
60
50
|° 30-
20
wound
u~
rmp.
n mm'Om'yand 30 m ~
line moooc
~
I ~
Fig. 6. Distribution of VRE in the ICU-PCU population for both medical and surgical patients.
(1) admission directly from another hospital, (2) admission from h o m e in the c o m m u n i t y (domicile), followed by (3) admission from a nursing home, and (4) readmission after discharge from LIJMC in excess of 1 month. A total of 165 VRE isolates were recovered from 138 patients. Species distribution was as follows: 130 (79%) were Enterococcus faecium, 14 (8.5%) were Enterococcus faecalis, 9 (5.5%) were Enterococcus durans, 1 (0.6%) was Enterococcus avium, and 11 (6.7%) were not identified to species level. Control of microbial spread among hospitalized patients requires careful epidemiologic analysis. Many factors have contributed to VRE emergence (e.g., inappropriate use of antibiotics not only in hospitals b u t also in the community, u n k n o w n level of colonization of patients with VRE, lack of understanding or refusal to understand the significance of VRE recovered from hospitalized patients, and absence of consistent infection control practices). At LIJMC, other important factors were considered and include the increase in methicillin-resistant Staphylococcus aureus and 6l a c t a m - r e s i s t a n t coagulase-negative staphylococci in the same population, requiring the use of vancomycin, and the increase in the n u m b e r of patients with Clostridium difficile intoxication who, until recently, were treated primarily with oral vancomycin (unpublished data). The VRE problem in our medical center is primarily in the adult patients, whereas the initial introduction of the organism was in our children's hospital. The initial group of enterococci from
SCH belonged to a single clone, but we n o w deal with multiclonal representatives of the genus. 2,6 Most likely, the initial episode involved the introduction of one VRE into a separate oncology unit in SCH where better adherence to infection control practices in a closed special unit, greater involvement of the professional staff, and better control of antibiotic use were readily achieved. At LIJMC the isolation of VRE was primarily from medical patients (Fig. 1), and urine was the most frequent site of infection or colonization (Fig. 2). With different strains of VRE gaining access to the facility, environmental contamination and cross-contamination from patient to patient play a major role in the transmission of VRE, as demonstrated by VRE acquisition during residence in a unit with other patients infected or colonized with VRE. Of course, enterococci constitute part of the n o r m a l h u m a n intestinal microbiota and can contaminate a variety of articles in patients' rooms (e.g., bed rails and commodes). 7 Improved housekeeping and strict adherence to universal precautions should decrease the spread of not only VRE but other important nosocomially acquired organisms. In addition, 26 of 36 (72%) of patients with urine cultures designated as infected had an indwelling Foley catheter before the positive culture, and 13 of 32 (40%) of those designated as colonized had been catheterized. This again points to the role of health care workers and prosthetic devices in the transmission of and complication by VRE and other nosocomially significant microorganisms.
AJIC
376 Tucci et al.
Prior use of vancomycin has been identified by a n u m b e r of authors as an important risk factor for acquiring g R g . 3,s In our survey, vancomycin use was only the third most important risk factor, after therapy with other antibiotics and residence in a unit with other patients infected or colonized with VRE. However, we had restricted use of vancomycin at the time of the survey, which might have affected the hierarchy of significance. The probability that multiple strains of VRE are involved is based on a previous study that demonstrated at least 15 different molecular clones of VRE in our institution. 6 Prior stay in the ICU or PCU was demonstrated to be an important risk factor primarily for surgical patients (Fig. 4). We evaluated the distribution of VRE in the ICU-PCU population for both medical and surgical patients (Fig. 6). Urine, wounds, and central lines were equally important sites for surgical patients; whereas urine was by far the most important site of infection or colonization for the medical patients in the IeUs. These differences emphasize that patients in the same unit can have different sites involved when exposed to the particular practices of different services. Infection control practices should be tailored to take account of these differences. The control of VRE in the hospital setting is made more difficult because almost half of the patients carrying VRE would not have been identified in the absence of microbiologic surveillance of clinical specimens for VRE. In our survey 79 of 165 (48%) of all VRE isolates were recovered in this m a n n e r . Eradication of VRE from hospitalized adult patients is difficult for several reasons. It m a y be red u n d a n t to point out that enterococci are normal residents of the h u m a n intestine. Selective pressures such as nonmedical use of antibiotics in the c o m m u n i t y m a y have shifted the ratio of E. faecalls and E. faecium reported at 87% to 9%, respectively, 9 in favor of E. faecium; and our results suggest that any prior antibiotic use tends to select VRE, leading to the admission of patients colonized and/or infected with VRE. Control of VRE in this population can be accomplished only if there is hospital-wide cooperation of all services and management. In our institution we have made the control of VRE a Quality Management issue. It is well known that improper use of antibiotics selects various resistant microorganisms that can complicate the recovery of patients. Our results clearly demonstrate that antibiotic use is the most important risk factor for acquiring VRE. We at-
October 1997
tempt to control antibiotic use by limiting the length of time between ordering antibiotics and by having the Infectious Disease Service approve the use of m a n y antibiotic drugs including vancomycin. This program has had limited success because of lack of sufficient personnel to follow up all approval requests thoroughly to determine whether the antibiotic use was appropriate on the basis of culture results or clinical presentation. Continuous education of staff about VRE and other nosocomially significant organisms is the key to controlling the spread of these bacteria. Notification of the charge nurse in person, by the epidemiologist who provided infection control suggestions, increased the awareness of all staff members. In addition, meetings were held with housekeeping and patient support managers to educate them about VRE and other nosocomial significant organisms and discuss methods tO improve cleaning of patients' rooms and equipment. The challenge will be to continue this task while shrinking health care dollars are forcing hospitals to cut costs by reducing the n u m b e r s of employees. This strategy could result in compromising the control of nosocomial disease. References 1. Center of Disease Control and Prevention. Nosocomial enterococci resistant to vancomycin--United States, 19891993. MMWR 1993;42:597-9. 2. Rubin LG, Tucci V, Cercenado E, Eliopoulos G, Isenberg HD. Vancomycin-resistant Emerococcus fiaecium in hospitalized children. Infect Control Hosp Epidemiol 1992; 13:700-5. 3. Centers for Disease Control and Prevention. Recommendations for preventing the spread of vancomycin resistance. Recommendations of the Hospital Infection Control Practices Advisory Committee (HIPAC). MMWR 1995; 44(RR-12). 4. Isenberg HD. Clinical microbiology procedure handbook. Washington (DC): American Society for Microbiology; 1992. 5. Center for Infectious Disease, Centers for Disease Control. Guidelines for the prevention and control of nosocomial infections. Atlanta (GA): Centers for Disease Control; Feb 1981. 6. Lam S, Singer C, Tucci V, Morthland V, Pfaller M, Isenberg HD. The challenge of vancomycin-resistant enterococci: a clinical and epidemiologic Study. AJIC Am J Infect Control 1995;23:170-80. 7. Yamaguchi E, Valena F, Smith S, Simmons A, Eng R. Colonization pattern of vancomycin-resistant Enterococcus faecium. AJIC Am J Infect Control 1994;22:202-6. 8. Karanfil L, Murphy M, Josephson A, Gaynes R, Mandel L, Hill B, et al. Cluster of vancomycin-resistant Enterococcus faecium in an intensive care unit. Infect Control Hosp Epidemiol 1992;13:195-200. 9. Ruoff KL, De La Maza L, Murtagh MJ, Spargo JD, Ferraro MJ. Species identities of enterococci isolated from clinical specimens. J Clin Microbiol 1990;28:435-7.