Professional Documents
Culture Documents
5
Copyright © 1998 by The Johns Hopkins University School of Hygiene and Public Hearth Printed In U.S.A
AU rights reserved
John A. Jernigan,1 Maureen G. Titus,1 Dieter H. M. Grbschel,1 Sandra I. Getchell-White,1 and Barry M. Farr1
Contact isolation has been recommended by the Centers for Disease Control and Prevention for the
prevention of nosocomial transmission of methicillin-resistant Staphylococcus aureus (MRSA), but there are
few data which prospectively quantitate the effectiveness of contact isolation for this purpose. During an
outbreak of MRSA in a neonatal intensive care unit between July 18, 1991 and January 30, 1992, weekly
surveillance cultures were performed on all patients. Sixteen of 331 admissions became colonized with MRSA,
and 3 (19%) developed infections: bacteremia, conjunctivitis, and dialysis catheter site infection. The isolates
from all 16 patients were submitted to plasmid profile analysis and restriction enzyme analysis of whole cell
DNA. All of the patients had identical chromosomal patterns and plasmid profiles, which differed from control
isolates from other wards, indicating that the outbreak resulted from spread of a unique strain. None of 144
personnel who were cultured after recent contact with newly colonized patients during the outbreak were
found to carry MRSA, which suggests that patients were the reservoir for transmission rather than caregivers.
The most probable source for each individual transmission was determined based on proximity in time and
space and shared exposure to caregivers. The rate of transmission of MRSA from patients on contact isolation
was significantly lower (0.009 transmissions per day on isolation) than the rate for patients not on isolation
(0.140 transmissions per day unisolated, relative risk = 15.6, 95% confidence interval 5.3-45.6, p < 0.0001).
The authors conclude that the risk of nosocomial transmission of MRSA was reduced 16-fold by contact
isolation during the outbreak in this neonatal intensive care unit. These data confirm the results of previous
studies which have suggested that contact isolation was effective in controlling the epidemic spread of
methicillin-resistant Staphylococcus aureus. Am J Epidemiol 1996;143:496-504.
cross infection; infection control; intensive care units, neonatal; methicillin resistance; Staphylococcus
aureus
Staphylococcus aureus resistant to methicillin infection in patients colonized with MRSA compared
(MRSA) has become an increasingly important noso- with those colonized with MSSA (4). Some investiga-
comial pathogen in US hospitals. Data from the Na- tors have reported that increasing rates of MRSA
tional Nosocomial Infection Surveillance System infection have increased the total endemic rate of 5.
(NNIS) show that the proportion of 5. aureus isolates aureus infections rather than merely replacing an
resistant to methicillin has increased from 2 percent to equal number of methicillin-sensitive 5. aureus infec-
29 percent in NNIS hospitals during the last 15 years tions (5-7), although other investigators have not con-
(1). MRSA has spread rapidly in many hospitals, caus- firmed this finding (8, 9). There is concern that van-
ing substantial morbidity and mortality (2). Although comycin use may increase substantially as MRSA
methicillin-resistant and methicillin-sensitive strains becomes more prevalent, resulting in increased selec-
of S. aureus (MSSA) have had similar virulence in tion pressure for vancomycin-resistant organisms (10).
animal models (3), one prospective study demon- For these reasons, many believe that efforts to control
strated a significantly greater risk of staphylococcal MRSA are warranted (2, 10, 11). A recent survey of
US hospital epidemiologists (2) found that 91 percent
Received for publication August 11,1994, and in final form June
29, 1995. of the epidemiologists used MRSA control measures
Abbreviations: DNA, deoxyribonucletc add; MRSA, methicillin- of some kind in their hospitals.
resistant Staphylococcus aurmis; MSSA, methicillin-sensitive
Staphylococcus aumus; NICU, neonatal intensive care unit; NNIS,
Although barrier precautions are often included in
National Nosocomial Infection Surveillance System. recommended control measures for MRSA (2, 10-
1
University of Virginia Health Sciences Center, Chariottesville, 12), and are used by 76 percent of US hospitals (2), the
VA.
Reprint requests to Dr. Barry M. Farr, University of Virginia Health data on their effectiveness are conflicting. While some
Sciences Center, Box 473, Chariottesvtlle, VA 22908. studies have found a decrease in the incidence of
496
Contact Isolation Effectiveness for Methicillin-resistant Staphylococcus aureus 497
endemic MRSA infection/colonization after adopting veillance cultures for detecting colonization were ob-
barrier isolation procedures (8, 9, 13, 14), others have tained from the nose, axilla, and groin and, also, from
failed to demonstrate a change in incidence using the the site of any percutaneous device or skin wound.
same or similar measures (15-17). To our knowledge,
no study has directly compared the rate of transmis- Outbreak control measures
sion of MRSA from unisolated patients with the rate of
transmission from patients who have been placed in When the outbreak began, four additional control
contact isolation. measures were instituted:
During a 7-month outbreak of MRSA in a neonatal 1. Weekly surveillance cultures were obtained of
intensive care unit in Charlottesville, Virginia, pro- the nares, groin, axilla, and wounds (if present)
spective culturing and epidemiologic analysis allowed on all patients in the unit not previously known to
measurement of transmission rates from unisolated be colonized or infected with MRSA.
patients for comparison with the transmission rates 2. Staff compliance with control measures, includ-
from patients who were in contact isolation. ing diligent hand washing with chlorhexidine
soap, was repeatedly encouraged through discus-
MATERIALS AND METHODS sions with unit personnel and memoranda.
Description of the hospital and neonatal intensive 3. For selected patients, attempts at eradication of
care unit colonization were made. Eradication regimens
were selected by the patient's primary physician.
The University of Virginia Hospital is a 700-bed Eradication was defined by three consecutive
hospital with nine intensive care units. The hospital daily cultures (nares, groin, axilla, wound, and
provides primary and tertiary care, and admits over any other site known to be previously colonized)
28,000 patients annually. The neonatal intensive care beginning at least 72 hours after discontinuation
unit (NICU) has 33 beds and admits approximately of antibiotics. Culture surveillance of previously
700 neonates each year. The unit consists of four pods colonized patients was continued until hospital
containing 6-10 beds each, and an isolation room discharge with four consecutive weekly cultures,
containing an anteroom and two beds. One nurse gen- followed by monthly cultures.
erally cares for two infants during a single shift. When 4. Surveillance cultures were obtained from person-
a nurse leaves the unit during the shift, the nurse for an nel who had had contact with new cases during
adjacent patient will observe and provide emergent the 2 weeks preceding identification of a case by
care if needed. The pods are separated by partitions conversion of surveillance cultures from negative
interrupted by walkways through which personnel to positive. Cultures were obtained from the na-
pass from one pod to the other. res and any visible skin lesions of personnel.
Using prospective weekly culture results, the ap- vals for simple proportions were determined using
proximate day of acquisition of MRSA was deter- published tables of exact confidence limits for bino-
mined for each patient. Because surveillance cultures mial distributions (26).
were obtained every 7 days, patients acquired MRSA
at some point between day 1 and day 7 following their RESULTS
last negative culture. Acquisition was estimated to
have occurred at the midpoint of that interval, i.e., day Description of the outbreak
3.5. For patients exposed to a known source at some In the 5 years preceding this outbreak, only four
time following their last negative culture, the date of sporadic nosocomial cases of MRSA colonization or
acquisition was estimated as the midpoint between infection had been identified in the neonatal intensive
first exposure to the source and their first positive care unit. The outbreak occurred during the 7-month
culture. The end of the period of colonization within period between July 18, 1991 and January 30, 1992
the unit was defined as the date of discharge from the (figure 1). The index case was identified through an
unit or, if eradication attempts were successful, the eye culture of an infant with purulent conjunctivitis on
midpoint between initiation of eradication therapy and July 18, 1991. This patient had recently traveled out of
the initiation of cultures that documented eradication. the NICU for a surgical procedure in the operating
room and was being followed by a non-NICU surgical
Microbiology consultation team. Over the next week, four additional
cases were identified that were epidemiologically
Staphylococcus aureus was identified by standard
linked to the index case (figure 2). Outbreak control
laboratory procedures (22). The determination of me-
measures, including weekly surveillance cultures of all
thicillin-resistance was performed according to the
patients in the unit who were not known to be previ-
methods recommended by the National Committee on
Clinical Laboratory Standards for disk diffusion test- ously colonized, were begun on July 30, 1991. Over
ing (23), and the use of an oxacillin salt agar screening the next 6 months, 11 new colonizations or infections
plate (24). were detected. By October 28, 71 days into the epi-
demic, there was only one remaining case in the NICU
isolation room. This case remained the only reservoir
Statistical methods in the unit for the ensuing 5 weeks before the next
Rates were compared using the large-sample test for transmission was detected. Surveillance cultures at the
comparison of incidence rates (25). Confidence inter- end of that 5-week period revealed MRSA coloniza-
7 * = Index Case
4 --
o
o
o
II 2
2
1 4-
o M A 4 44 4 5 1 A
July Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun
1991 1992
FIGURE 1. Bar graph showing the prevalence of methidllin-resistant Stephytococcus aureus (MRSA) colonization during an outbreak in a
neonatal intensive care unit, Charlottesville, Virginia, 1991-1992. The arrows indicate the incidence of new cases (see key).
Transferred on 7/20
(7/24) T (7/24)
W22)
7/18-8/19
Isolation
uu uuuu
9/16-10/1
UU UUUU
10/28-12/2
Itotatkm ,~-\.~'
UU UUUU
12/13-1/7
(1/7)
UU _ U U UU UUUU
1/18- 1/30
Isolation
nn -" Bed not containing MRSA patient
Transmission from an unisotated infant at beginning of time period
Transmission from an isolated Infant Bed containing MRSA patient
Intraunft patient transfer at beginning of time period
FIGURE 2. Methicillin-resistant Staphylococcus aureus (MRSA) transmissions depicted in space and time in an outbreak in a neonatal
intensive care unit, Chariottesville, Virginia, 1991-1992. Five floor plans of the neonatal intensive care unit are shown. Each floor plan
corresponds to a time period during which transmissions occurred. Arrows indicate the donor and recipient of each transmission. Dates
indicate when cultures were recognized to be growing MRSA.
tion of three new babies located a considerable dis- respiratory distress syndrome (25 percent), and intra-
tance from the isolation room. The only personnel who ventricular hemorrhage (19 percent). Thirteen of the
had shared care for the isolated baby and the others cases were colonized and three were infected. In ad-
were the attending and resident physicians. One of the dition to the index case (conjunctivitis), the infections
first-year residents, who reportedly had more manual included a Tenkhoff dialysis catheter site infection and
contact with the babies than the other physicians, had one bacteremia. All three infections were successfully
been primarily responsible for case 8 in the isolation treated with vancomycin. The most common sites of
room and also for new cases 12 and 13 (figure 2). At colonization were nares (88 percent), umbilicus (56
least one colonized or infected infant remained in the percent), groin (50 percent), and axilla (31 percent).
unit until April 27, 1992 (figure 1). Weekly surveil-
lance cultures of every infant in the unit was continued
for another 2 weeks, and no new cases were identified Results of personnel cultures
during that period. None of the 144 personnel cultures taken during the
Five months after the final transmission of the out- outbreak were positive for MRSA. During the inves-
break and 6 weeks after discharge of the final patient tigation of the two cases which appeared after the
reservoir from the NICU, two new cases of MRSA resolution of the outbreak, two (1.1 percent) of 181
colonization with the outbreak strain were identified in personnel were found to be colonized.
adjacent beds within the unit. One of these patients
had been admitted to the unit following discharge of
the last patient reservoir, and the patient developed Outcome of eradication regimens
MRSA infection of a surgical wound. The other pa- Ten patients received therapy to eradicate coloniza-
tient had been present in the unit during the period tion and had evaluable outcome data. Eradication was
when there were other patient reservoirs present, but successful in seven of the 10 patients. The regimens
was culture-negative for MRSA for the duration of used and duration of follow-up are detailed in table 1.
weekly surveillance of the unit. Because no patient
reservoir existed in the unit at the time of acquisition,
extensive surveillance culturing of personnel was re- Molecular typing
peated. Two nurses were found to be colonized with All 16 isolates had identical restriction fragment
MRSA. One of these nurses (Nurse A) was epidemi- profiles for total cellular DNA which clearly differed
ologically linked to the new cases. Nurse A had from control MRSA isolates obtained from other
worked with the last colonized infant in the isolation wards (figure 3). All 16 isolates contained a single
room of the unit. She was found to be colonized with plasmid of identical molecular weight, which differed
the outbreak strain in the groin and at the site of an significantly from the control isolates. The isolates
eczematous lesion on her hand. This nurse had been from the two patients and two personnel which were
cultured during the outbreak and was negative at that discovered following resolution of the outbreak had
time. The other nurse (Nurse B), who had not worked DNA restriction fragment length and plasmid profiles
with any of the previous MRSA cases, worked with identical to the original outbreak isolates and different
the two new cases after Nurse A and prior to their from controls.
being isolated; she was colonized in the nares only.
The nurses' colonization was eradicated, as shown in
table 1. The infants were placed on contact isolation Epidemiologic studies
and discharged within 3 weeks, and weekly surveil-
There were 629.5 patient-days of MRSA coloniza-
lance cultures of all patients were continued for an
tion during the outbreak. A total of 558 patient-days of
additional 3 weeks following their discharge with no
colonization were spent in contact isolation, and 71.5
positive results. No further cases of MRSA coloniza-
patient-days of colonization were not spent in contact
tion or infection were observed in the ensuing 44
isolation (table 2). In all, there were 15 transmissions,
months.
giving a total rate of 0.17 transmissions per colonized
patient-week during the outbreak.
Description of the patients
The results of epidemiologic tracing of the source of
During the 7 months, 16 (4.8 percent) of 331 neo- the transmission by two independent observers were
nates admitted to the unit acquired MRSA. The most 100 percent concordant. Unisolated infants were
common diagnoses among patients with MRSA infec- judged to be the source of 10 transmissions, while
tion or colonization were prematurity (87 percent), isolated infants were believed to be responsible for
patent ductus arteriosus (44 percent), pneumonitis/ five transmissions (table 2).
Patient Follow-up
Sites colonized Regimen (duration) Outcome
no. (days)
1 Nares Nasal muplrocin (14 days), Eradicated 14
vancomycin (7 days),
rifampln (4 days)
The relative risk of transmission from unisolated have examined the effect of isolation procedures in
patients was 15.6 compared with patients on contact controlling nosocomial transmission of MRSA, the
isolation (table 3). results have not been consistent (13-17, 27). Some
studies found a decrease in the incidence of endemic
DISCUSSION MRSA infection/colonization after adopting barrier
The majority of the published literature supports the isolation procedures, including one study conducted at
concept that colonized patients are the major reservoir the University of Virginia Hospital more than a decade
for MRSA and that most transmissions occur via the ago (8, 9, 13, 14). Others have failed to demonstrate a
hands of hospital personnel. Although several studies change in incidence with the use of the same or similar
1 Unknown source 24 2
2 Unlsolated 63 3
3 Unisolated 8 3
4 Unisolated 13 2
5 Unisolated 9 2
6 Unisolated 46 5
7 Unisolated 41 9
8 Unisolated 116 5.5
9 Unisolated 11 5
10 Contact Isolation 21 5.5
11 Unisolated 16 5.5
12 Contact Isolation 42 4
13 Contact Isolation 8 5.5
14 Unisolated 44 3.5
15 Contact Isolation 7 5.5
16 Contact isolation 89 5.5
Source of transmission
FIGURE 3. Electrophoresis patterns of a representative sample of
methicillin-reslstant Staphylococcus aureus (MRSA) isolates from Isolated Unisolated
an outbreak in a neonatal intensive care unit, Charlottesville, Vir-
Transmissions 5 10
ginia, 1991-1992, when subjected to restriction endonuclease di-
gestion. Lanes 1-5 are outbreak Isolates. The control lane is an Patient-days 558 71.5
MRSA isolate obtained from a different hospital ward. Rate of transmission 0.0090 0.140*
• Relative risk = 15.6, 95% confidence interval 5.3-45.6,
p < 0.0001.
22. Kloos WE, Lambe DW. Staphylococcus. In: Balows A, Hau- for bacteria that grow aerobically. (NCCLS document M7-
sler WJ, Hermann KL, et al, eds. Manual of clinical microbi- A3). Villanova, PA: NCCLS, 1993:17.
ology. Washington, DC: American Society of Microbiology, 25. Rosner B. Fundamentals of biostatistics. Boston: PWS-Kent
1991:222-37. Publishing Co, 1990:366.
23. National Committee on Clinical Laboratory Standards (NC- 26. Documenta Geigy: scientific tables. Basle, Switzerland: JR
CLS). Performance standard for antimicrobial disk suscepti- Geigy, 1970:85.
bility tests. (NCCLS document M2-A5). Villanova, PA: 27. Ribner BS, Landry MN, Gholson GL. Strict versus modified
NNCLS, 1993:14. isolation for prevention of nosocomial transmission of methi-
24. National Committee on Clinical Laboratory Standards (NC- cillin-resistant Staphylococcus aureus. Infect Control 1986;7:
CLS). Methods for dilution antimicrobial susceptibility tests 317-20.