Content uploaded by Xavier Bertrand
Author content
All content in this area was uploaded by Xavier Bertrand on Feb 04, 2016
Content may be subject to copyright.
MRSA and Antimicrobial Use •CID 2003:36 (15 April) •971
MAJOR ARTICLE
Relationship between Spread
of Methicillin-Resistant Staphylococcus
aureus and Antimicrobial Use
in a French University Hospital
Arno A. Muller,
1
Fre´de´ric Mauny,
2
Maud Bertin,
1
Christian Cornette,
3
Jose´-Maria Lopez-Lozano,
4
Jean Franc¸ois Viel,
2
Daniel R. Talon,
1
and Xavier Bertrand
1
1
Service d’Hygie` ne Hospitalie`re et d’Epide´miologie Mole´ culaire,
2
De´ partement d’Information Me´dicale, and
3
Pharmacie Centrale, Centre Hospitalier
Universitaire Jean Minjoz, Besanc¸ on, France; and
4
Preventive Medicine Unit, Hospital Vega Baja, Alicante, Spain
The objective of our study was to determine whether antibiotic pressure in the units of a teaching hospital
affects the acquisition of methicillin-resistant Staphylococcus aureus (MRSA), independently of the other
collective risk factors previously shown to be involved (MRSA colonization pressure, type of hospitalization
unit, and care workload). The average incidence of acquisition of MRSA during the 1-year study period was
0.31 cases per 1000 days of hospitalization, and the use of ineffective antimicrobials reached 504.54 daily
defined doses (DDDs) per 1000 days of hospitalization. Univariate analysis showed that acquisition of MRSA
was significantly correlated with the use of all antimicrobials, as well as correlated with the use of each class
of antimicrobial and with colonization pressure. Multivariate analysis with a Poisson regression model showed
that the use of antimicrobials was associated with the incidence of acquisition of MRSA, independently of the
other variables studied, but it did not allow us to determine the hierarchy of the different antimicrobial classes
with respect to the effect.
Methicillin-resistant Staphylococcus aureus (MRSA) is a
major nosocomial pathogen that causes severe mor-
bidity and mortality in hospitals in France and world-
wide [1–4]. Control of the spread of this pathogen is
an enormous challenge for clinicians, infection control
physicians, and hospital administrators [5]. It is now
well established that colonized and infected inpatients
are the major reservoir of this pathogen and that the
transient carriage of MRSA on the hands of hospital
personnel is the most common mechanism of patient-
to-patient transmission [6, 7]. After acquisition, MRSA
Received 11 October 2002; accepted 9 January 2003; electronically published
4 April 2003.
Reprints or correspondence: Dr. Xavier Bertrand, Service d’Hygie` ne Hospitalie`re
et d’Epide´miologie Mole´ culaire, CHU Jean Minjoz, 25030 Besanc¸ on, France
(daniel.talon@ufc-chu.univ-fcomte.fr).
Clinical Infectious Diseases 2003;36:971–8
2003 by the Infectious Diseases Society of America. All rights reserved.
1058-4838/2003/3608-0005$15.00
strains multiply on the contaminated tissue and may
then colonize and possibly infect the patient. This pro-
gression to symptomatic infection is promoted by the
existence of a site of entry, such as a wound or an
indwelling venous or urinary catheter.
Many studies have identified individual risk factors
for MRSA infection. These factors can be divided into
3 categories: (1) those related to the number of potential
reservoirs and the number of opportunities for cross-
transmission, (2) those associated with the immuno-
logical status of the patients, and (3) those related to
the antibiotics used to treat the patients [7–12]. These
case-control studies only identified associations at the
level of individual patients and did not, therefore, take
into account the wider picture and the complexity of
the problem [13]. Indeed, the use of antimicrobials to
treat individuals has an ecological impact on all of the
patients hospitalized in the same unit [14]. Only a few
studies have taken these collective approaches into
by guest on February 2, 2016http://cid.oxfordjournals.org/Downloaded from
972 •CID 2003:36 (15 April) •Muller et al.
account. These studies tend to be fragmented, and they consider
the influence of colonization pressure [15] or antibiotic pres-
sure [16] but never both simultaneously. Moreover, these stud-
ies also have attempted to determine whether antibiotic pres-
sure was a risk factor for MRSA acquisition by comparing data
from different hospitals or by using the data collected in a
single type of unit, such as intensive care units [15, 17]. The
objective of our study was to determine whether antibiotic
pressure within the units of a teaching hospital affects the in-
cidence of acquisition of MRSA, independently of the other
collective risk factors previously shown to be involved.
PATIENTS AND METHODS
Setting and study period. The Besanc¸ on Hospital (Besanc¸on,
France) is a university-affiliated hospital with 1228 acute care
beds and is divided into 59 units (35 medical units, 21 surgical
units, and 3 intensive care units). Specialty services include
cardiothoracic surgery and organ and bone marrow transplan-
tation. Approximately 50,000 inpatients are admitted per year,
for a total of 350,000 patient days. Data were collected from 1
October 2000 through 30 September 2001.
MRSA control program. In 1994, our infection control
committee made the control of MRSA a major priority. An
MRSA-control program was progressively implemented in all
high-risk units. Since 1997, 14 of the 59 units (corresponding
to all of the high-risk departments, including the adult intensive
care units and the septic surgical unit) have been following this
program. The control strategy is based on screening nasal fluid
samples from all patients for MRSA at the time of admission
to the unit. In the first 48 h after admission, before the results
of the screening cultures are available, patients are considered
to be positive for MRSA and kept in isolation. Patients who
are actually positive for MRSA are then given nasally admin-
istered mupirocin. All MRSA-positive patients are kept in in-
dividual rooms or in cohorts. Special precautions are taken to
prevent cross-contamination, including the use of disposable
gowns and gloves, the use of an antiseptic soap for hand wash-
ing, and the implementation of strict environmental hygiene
measures. These procedures are also applied to patients in low-
risk departments (who are not screened at the time of admission
to the unit) whose clinical samples test positive for MRSA. The
program did not include any restrictions on antibiotic use.
MRSA-positive patients. A patient was classified as being
“MRSA positive” if they provided a clinical specimen that tested
positive for MRSA during the study period and if the patient
was not known to have tested positive for MRSA during the
previous 3 years, on the basis of the mean duration of MRSA
carriage (figure 1) [18]. MRSA-positive patients were classified
as having “imported MRSA carriage” if they tested positive for
MRSA within 48 h after admission to the hospital or as having
“acquired MRSA carriage” if they tested negative for MRSA
during the first 48 h after admission to the hospital. Patients
with acquired MRSA carriage were further classified as having
“endogenous acquired MRSA carriage” (the NAC group) on
the basis of a previous MRSA-positive surveillance culture re-
sult, “exogenous acquired MRSA carriage” (the XAC group)
on the basis of a previous MRSA-negative surveillance culture
result, or “undetermined acquired MRSA carriage” (the UAC
group) if there was no surveillance culture result available. The
mean duration of hospitalization before acquisition was 18 days
for the XAC group, 16 days for the UAC group, and 10 days
for the NAC group. The UAC group was combined with the
XAC group, and we considered that this combined group con-
tained the correct number of patients with XAC (which we
labelled as the “real acquired case” group). This combination
was based on the mean delay before acquisition and on the
distribution of cases of acquired MRSA carriage within the
hospital units participating in the MRSA-control program. In-
deed, in this type of unit, the XAC group accounted for 80%
of the total number of cases of acquired MRSA carriage. We
also combined the imported MRSA carriage group and the
NAC group (which we labelled the “real imported case” group).
The incidence of MRSA carriage was expressed as cases per
1000 days of hospitalization.
MRSA colonization pressure. The colonization pressure
was calculated for each hospital unit as the ratio of the number
of MRSA-positive patient–days (whatever the type of specimen
[clinical or screening] used to identify MRSA carriage) to the
total number of patient-days. Patients previously known to be
MRSA positive, patients with MRSA carriage (i.e., patients for
whom only the screening culture tested positive for MRSA),
and patients in the real imported cases group were considered
to have had MRSA carriage throughout their hospital stay. Pa-
tients in the real acquired cases group were considered to have
had MRSA carriage from the date of obtainment of the sample
that tested positive until discharge from the hospital or death.
We distinguished 2 types of colonization pressure, according
to the nature of the cases of MRSA carriage: “CP1,” which
labelled the colonization pressure exerted by the patients with
real imported cases, the patients with MRSA carriage, and the
patients previously known to be MRSA positive; and “CP2,”
which labelled the colonization pressure exerted by the patients
with real acquired cases (figure 1). The colonization pressure
values were calculated for each hospital unit and expressed as
days per 1000 days of hospitalization.
Antibiotic selective pressure. The quantities of each an-
timicrobial delivered to each hospitalization unit during the
study period were determined from the pharmacy informa-
tion system. Data on the amounts of antimicrobials used that
were expressed in grams and international units were con-
verted to express use as defined daily doses (DDDs), following
by guest on February 2, 2016http://cid.oxfordjournals.org/Downloaded from
MRSA and Antimicrobial Use •CID 2003:36 (15 April) •973
Figure 1. Classification of patients infected with methicillin-resistant Staphylococcus aureus
the recommendations of the World Health Organization [19].
Antimicrobials were grouped into 6 classes: aminoglycosides,
b-lactams, fluoroquinolones, glycopeptides, macrolides, and
other; the b-lactams were further subdivided into penicillins
and cephalosporins. For the analysis, antimicrobials were
classed as being effective or ineffective against MRSA. Gly-
copeptides, gentamicin, rifampicin, pristinamycin, and fusidic
acid were considered to be effective. Indeed, 85% of the MRSA
isolates displayed the same phenotypic characteristics (sus-
ceptibility to antibiotics). All other antimicrobials were con-
sidered to be ineffective. Antimicrobial use was finally ex-
pressed as the number of DDDs per 1000 patient-days for
each unit.
Characterization of hospitalization units and care work-
load. Data on the number of patient-days for each unit dur-
ing the study period was obtained from the hospital admission
department. The hospital units were classified into 3 types:
medicine units ( ), surgery units ( ), and intensivenp37 np19
care units ( ). The units were further divided into 2groupsnp3
according to whether they participated in the MRSA control
program. The nursing workload was assessed according to the
nursing care requirements of the Project for Research in Nurs-
ing (PRN) [20]. The PRN is a Canadian information system
for the management of nursing staff in hospitals; it has been
validated [21]. Outside Canada, PRN is used in France, Spain,
and the United Kingdom. It estimates the quantity of nursing
care required by each patient during a 24-h period. Annual
averages were used to quantify the nursing workload for each
hospital unit. This program measures the temporal nursing care
workload per patient and per 24-h period by computer.
Statistical analysis. The study approach was ecological,
focused on the hospitalization units. Thus, the data were all
aggregated at this level. Statistical analysis was conducted to
explore the relationships between the number of acquired cases
of MRSA (expressed as an annual incidence rate) and several
variables for each unit: the type of hospital unit (medicine,
surgery, or intensive care), participation in the MRSA control
program, care workload, antimicrobial pressure exerted by use
by guest on February 2, 2016http://cid.oxfordjournals.org/Downloaded from
974 •CID 2003:36 (15 April) •Muller et al.
Table 1. Nonparametric correlation between the incidence of
real acquired cases (RACs) of methicillin-resistant Staphylococ-
cus aureus (MRSA) carriage and colonization pressure, antimi-
crobial selective pressure, and care workload.
Variable
Association with the
incidence of RACs
of MRSA carriage
Correlation
coefficient P
a
Colonization pressure
b
0.60 !.001
Global antimicrobial selective pressure,
by class of antimicrobial 0.56 !.001
Aminoglycosides 0.30 !.05
b-Lactams 0.53 !.001
Fluoroquinolones 0.57 !.001
Macrolides 0.32 !.05
Other 0.50 !.001
Care workload 0.12 .43
a
Determined by Spearman test
b
Colonization pressure exerted by the patients with real imported cases of
MRSA carriage, the patients with MRSA carriage, and the patients previously
known to be MRSA positive (“CP1”; see Methods, “MRSA colonization
pressure”).
Table 2. Poisson-regression modelling of incidence of real acquired
cases (RACs) of methicillin-resistant Staphylococcus aureus carriage: re-
sults of the reference model.
Variable, class
a
Association with the incidence
of RACs of MRSA carriage
Correlation
coefficient SD P
b
Rate ratio (95% CI)
Constant ⫺8.707 — — —
Colonization pressure
c
0.023 0.003 !.001 1.02 (1.02–1.03)
Type of hospital unit
Medicine 0.167 0.214 !.001 1.18 (0.78–1.80)
Intensive care 1.275 0.326 3.58 (1.89–6.78)
a
Medicine and intensive care units versus surgery unit used as a reference.
b
Determined by likelihood ratio test.
c
Colonization pressure exerted by the patients with real imported cases of MRSA
carriage, the patients with MRSA carriage, and the patients previously known to be MRSA
positive (“CP1”; see Methods, “MRSA colonization pressure”).
of ineffective antimicrobials, and colonization pressure exerted
by imported MRSA carriage (i.e., CP1), which are recognized
as potential risk factors for the acquisition of MRSA carriage
[15–17]. In our model, the number of acquired cases of MRSA
was directly linked, by construction in our model, to CP2. Thus,
we retained CP1 as an indicator of colonization pressure. Fur-
thermore, this choice was supported by the results of previous
studies that showed MRSA acquisition is mainly dependent on
CP1 [17, 19].
Association between variables and the incidence of MRSA
carriage were tested in univariate analysis with the Spearman,
Kruskall-Wallis, and Mann-Whitney tests. We retained those
variables that seemed to be statistically associated with an in-
cidence of MRSA carriage with a threshold Pvalue of .20.
Subsequently, Poisson regression was used in a 2-stage multi-
variate analysis. First, excluding antimicrobial pressure, a ref-
erence model was built by introducing the variables that were
found to be significant at in the univariate analysis.P!.20
Second, antimicrobial variables were separately introduced into
the reference model. At this stage, and for each antimicrobial
class, the hospitalization units were grouped and classified as
“weak consumers” (this was used as the base group), “medium
consumers,” or “high consumers” of the specified class of an-
timicrobials; it was assumed that there were an equal number
of hospital units in each of the 3 categories. Because of the use
of these classifications, no hypotheses have to be made about
the relationships between MRSA carriage and antimicrobial
pressure, and some statistical relationships could be explored
(linear, threshold or plateau effect). Likelihood-ratio tests were
used to compare the fit of nested models and to provide a test
of significance for the last term added to the model. Use of
interaction terms of each significant variable in the final models
did not significantly improve the models. was consid-P⭐.05
ered to be significant. Analyses were performed with the Systat
software (SPSS) and Egret software (Cytel Software), version
2.0.
RESULTS
During the study period, 234 cases of MRSA carriage were
identified. Of these 234 cases, 124 were classified as realacquired
cases, which corresponds to an annual incidence of 0.31 cases
per 1000 days of hospitalization in the entire hospital and to
annual incidences ranging from 0.0 to 1.63 cases per 1000 days
of hospitalization in the different hospital units. The mean
MRSA colonization pressure was 20 days per 1000 days of
by guest on February 2, 2016http://cid.oxfordjournals.org/Downloaded from
MRSA and Antimicrobial Use •CID 2003:36 (15 April) •975
Table 3. Results of Poisson-regression modelling of real acquired cases
of methicillin-resistant Staphylococcus aureus carriage: results of the 7
derived models, each including the reference model variables and 1 an-
timicrobial class.
Class of drug,
class of hospital unit
a
Correlation
coefficient SD P
b
Rate ratio (95% CI)
Aminoglycosides
MC 0.580 0.258 1.79 (1.08–2.96)
HC 0.597 0.275 .046 1.82 (1.06–3.11)
b-Lactams
MC 1.867 0.403 6.47 (2.93–14.24)
HC 1.894 0.414 !.001 6.65 (2.96–14.95)
Penicillins
MC 1.485 0.373 4.41 (2.12–9.17)
HC 1.582 0.372 !.001 4.87 (2.34–10.10)
Cephalosporins
MC 0.910 0.318 2.48 (1.33–4.63)
HC 1.352 0.294 !.001 3.86 (2.17–6.88)
Fluoroquinolones
MC 0.930 0.337 2.53 (1.31–4.91)
HC 1.582 0.315 !.001 4.86 (2.62–9.02)
Macrolides
MC 1.082 0.335 2.95 (1.53–5.69)
HC 1.079 0.355 .004 2.94 (1.47–5.90)
Other
MC 1.207 0.316 3.34 (1.80–6.21)
HC 1.376 0.329 !.001 3.96 (2.08–7.54)
NOTE. For definitions of models and hospital unit classes, see Methods, “Statistical
analysis.” HC, high consumer of stated drug class; MC, medium consumer of stated
drug class.
a
HC or MC units versus weak consumer units used as a reference.
b
Determined by likelihood ratio test.
hospitalization, varying from 12 to 29 days per 1000 days of
hospitalization during the study period. CP1 accounted for 51%
of the total colonization pressure. The mean CP1 was 10.26
days per 1000 days of hospitalization.
The average antimicrobial use observed was 563.22 DDDs
per 1000 patient days. Ineffective antimicrobials accounted for
89.6% of all antimicrobial treatments (504.54 DDDs per 1000
patient days). The distribution among antibiotic classes of all
the antimicrobials used was as follows: 74.7% of DDD for b-
lactams (62.1% for penicillins and 12.6% for cephalosporins),
14.6% for fluoroquinolones, 4.4% for aminoglycosides, 1,4%
for macrolides, and 4.9% for other antibiotics.
Neither CP1 nor antimicrobial selective pressure were as-
sociated with the type of unit ( and , respec-Pp.85 Pp.22
tively). The incidence of acquired MRSA carriage in hospital
units was not related to the unit’s participation in the MRSA
control program ( ) or to the type of hospital unitPp.28
( ). CP1 and antibiotic selective pressure values werePp.06
associated with the incidence of acquisition of MRSA carriage
(table 1). Moreover, CP1 and antibiotic selective pressure (both
global and that exerted by each class) were correlated with each
other (data not shown). Multivariate analysis yielded areference
model that included CP1 and the type of hospitalization unit
(table 2). Results for the derived models, including antimicro-
bial use, are shown in table 3. Antimicrobial use significantly
influenced the rate of acquisition of MRSA. The incidences of
acquired MRSA carriage were significantly higher in the me-
dium and high consumer units than in the weak consumer
units. Rate ratios ranged from 1.8 to 6.6, depending on the
type of consumer and the antimicrobial class (figure 2). The
highest risks were observed for b-lactams (especially penicillins)
and fluoroquinolones.
DISCUSSION
The findings of this study confirm those of a small number of
previous studies that have found a relationship between anti-
microbial use and the frequency of MRSA acquisition [16, 22,
by guest on February 2, 2016http://cid.oxfordjournals.org/Downloaded from
976 •CID 2003:36 (15 April) •Muller et al.
Figure 2. Rate ratios for the incidence of acquisition of MRSA carriage, as detemined by multivariate analysis, for selected antimicrobials in 3
classes of hospital units stratified by category of antimicrobial usage (i.e., “weak, “medium,” and “high consumers” of antimicrobials).
23]. It also shows that this relationship persists when MRSA
colonization pressure and the type of hospital unit are taken
into account, and that it is observed in hospital units other
than intensive care units. To date, no studies havedemonstrated
the effect of antimicrobial use on MRSA spread when consid-
ering other identified risk factors, such as MRSA colonization
pressure and the type of hospital unit. Moreover, we consider
that the incidence of real acquired cases of MRSA carriage is
the best indicator to measure the risk of acquiring MRSA. Other
indirect indicators, such as frequency of resistance within the
species or total cases of acquired MRSA carriage, are less
specific.
Three hypotheses can explain the lack of association between
MRSA acquisition and the care workload. First, the care work-
load does not influence the rate of MRSA acquisition. Second,
our indicator did not correctly represent the care workload.
Third, this indicator is not adequate in our type of study and
would be more accurate in another study design.
The literature provides indirect evidence of a relationship
between antimicrobial use and the emergence of resistance in
hospitals. According to McGowan [24], this evidence can be
divided into 4 categories: (1) evidence of consistent associa-
tions, (2) evidence of dose-response effects, (3) evidence of
concomitant variation, and (4) evidence from a plausible bio-
logical model. A consistent association established at the level
of the hospital unit is particularly strong for b-lactams and
fluoroquinolones. The positive correlation between evidence for
use of these different antimicrobial families is associated with
a similar effect on the incidence of MRSA acquisition. This
correlation can be explained by the frequent use of antimicro-
bial combinations for the treatment of bacterial infections and
by the fact that our study did not take the effect of time into
account, such that data for the different treatments form an
indivisible whole. Nevertheless, the stratification of the hospital
units into antimicrobial-consumer classes allowed us to elim-
inate the scale effect of each antimicrobial class from the data
for total antimicrobial use and to characterize the impact of
high antimicrobial use compared with weak antimicrobial use
for each antimicrobial class. Despite our use of this analysis,
we could not directly determine a hierarchy among the different
antimicrobial classes. The importance of b-lactams and fluor-
oquinolones that we observed is consistent with the findings
of Loulergues et al. [22], who compared data for different sur-
gical units in a hospital, and those of Crowcroft et al. [16] and
Monnet et al. [25], who performed interhospital comparisons.
Crowcroft et al. [16] and Monnet et al. [25] observed a positive
correlation between the frequency of resistance among Staph-
ylococcus aureus isolates among the species and antimicrobial
use; in the study of Monnet et al. [25] (ICARE project), the
correlation was with use of carboxy- or ureido-penicillins, and,
in the study of Crowcroft et al. [16] (which was performed in
Belgian hospitals), the correlation was with use of ceftazidime,
cefsulodin, amoxicillin plus clavulanic acid, and quinolones.
Our study revealed different dose-response effects. Unlike
the 2 studies mentioned above [16, 25], we tested the hypothesis
of a linear dose-effect relationship. Given the rate ratios (table
3 and figure 2), the hypothesis of a linear effect seems to be
true for fluoroquinolones and cephalosporins. For the other
antimicrobial classes, particularly the penicillins, a plateau effect
is more likely [26].
The third criterion that can be used to establish causality is
concomitant variation—that is, an increase in antimicrobial use
followed by an increase in resistance. The suspected cause (a
change in antimicrobial use) must take place before the effect
(a change in the level of antimicrobial resistance). In our study,
the MRSA acquisition events were rare, and the study was
performed on the scale of the hospital unit; therefore, we had
to analyze the relationship on the basis of aggregated yearly
by guest on February 2, 2016http://cid.oxfordjournals.org/Downloaded from
MRSA and Antimicrobial Use •CID 2003:36 (15 April) •977
data, rather than taking into account the temporal criteria. We
plan to improve our analysis by using time series analysis as
statistical tool. Unlike classical statistical methods that assume
that the observed data are independent random variables, time
series analysis takes into account relationships between con-
secutive observations [26]. This technique was developed by
Lopez-Lozano and Monnet and colleagues [23, 27] to inves-
tigate the relationship between the emergence of antimicrobial
resistance and the use of antimicrobials.
McGowan [24] proposed a biological model to explain the
relationship between antimicrobial use and the emergence of
resistance. At the level of the individual patient, antimicrobial
treatment leads to a large modification in the endogenous flora.
The usual result is that susceptible strains are replaced by re-
sistant ones. At the collective level, antimicrobial use in a hos-
pital unit tends to maintain the presence of multidrug-resistant
organisms in inpatients, health care workers, and the environ-
ment. In cases in which basic infection-control practices are
inconsistently applied, these pathogens are implicated in the
majority of infections. Antimicrobials such b-lactams and fluor-
oquinolones, which are ineffective against MRSA and have ex-
cellent tissue diffusion, could promote the acquisition of MRSA
by increasing the “receptiveness” of the patients and thereby
allowing the progression towards colonization and infection.
MRSA acquisition depends on 2 major and independent
determinants: colonization pressure and antimicrobial selective
pressure. Previous studies have reported that the role of each
of these 2 factors may vary depending on the epidemiological
situation [28, 29]. In the case of colonization with distinct
multiple clones of MRSA, antimicrobial pressure plays the ma-
jor role; in the case of colonization with a single dominant
clone of MRSA, colonization pressure plays a major role. Our
results are not consistent with this model. Indeed, in our hos-
pital, 180% of the cases of MRSA colonization/infection are
caused by a single clone (data not shown). However, our study
demonstrates that the selective pressure caused by use of an-
timicrobials is an independent risk factor for MRSA acquisition.
Our results are consistent with those of several recent studies
that support a causal relationship between antimicrobial use
and MRSA acquisition [13, 14, 30–33]. The dissemination of
epidemic clones does not necessarily require antimicrobial se-
lective pressure. However, the results of these recent studies
and our results suggest that antimicrobials contribute to MRSA
spread. Furthermore, 14 hospitals in countries with very low
incidences of MRSA, particularly Nordic countries, use the least
amount of antimicrobials in Europe [34]. Additional research
is needed to understand fully the relationship between anti-
microbial use and MRSA acquisition. However, there is evi-
dence supporting the implementation of programs to control
and to improve prescription practices when infection control
alone fails to control the spread of MRSA.
References
1. Aubry-Damon H, Legrand P, Brun-Buisson C, Astier A, Soussy CJ,
Leclerq R. Reemergence of gentamicin-susceptible strains of methicil-
lin-resistant Staphylococcus aureus: roles of an infection control pro-
gram and changes in aminoglycoside use. Clin Infect Dis 1997;25:
647–53.
2. Bertrand X, Thouverez M, Talon D. Antibiotic susceptibility and ge-
notypic characterization of methicillin-resistant Staphylococcus aureus
strains in eastern France. J Hosp Infect 2000; 46:280–7.
3. Boyce JM. Increasing prevalence of methicillin-resistant Staphylococcus
aureus in the United States. Infect Control Hosp Epidemiol 1990;11:
639–42.
4. Voss A, Milatovic D, Wallrauch-Schwarz C, Rosdahl VT, Braveny I.
Methicillin-resistant Staphylococcus aureus in Europe. Eur J Clin Mi-
crobiol Infect Dis 1994; 13:50–5.
5. Goldmann DA, Huskins WC. Control of nosocomial antimicrobial-
resistant bacteria: a strategic priority for hospitals worldwide. Clin In-
fect Dis 1997; 24(Suppl 1):S139–45.
6. Boyce JM, Potter-Bynoe G, Chenevert C, King T. Environmental con-
tamination due to methicillin-resistant Staphylococcus aureus possible
infection control implications. Infect Control Hosp Epidemiol 1997;
18:622–7.
7. Thompson RL, Cabezudo I, Wenzel RP. Epidemiology of nosocomial
infections caused by methicillin-resistant Staphylococcus aureus. Ann
Intern Med 1982; 97:309–17.
8. Asensio A, Guerrero A, Querada C, Liza’n M, Martinez-Ferrer M.
Colonization and infection with methicillin-resistant Staphylococcusau-
reus: associated factors and eradication. Infect Control Hosp Epidemiol
1996; 17:20–8.
9. Ayliffe GAJ. The progressive intercontinental spread methicillin-resis-
tant Staphylococcus aureus. Clin Infect Dis 1997; 24:S74–9.
10. Humphreys H, Duckworth G. Methicillin-resistant Staphylococcus au-
reus (MRSA): a reappraisal of control measures in the light of changing
circumstances. J Hosp Infect 1997; 36:167–70.
11. Hershow RC, Khayr WF, Smith NL. A comparison of clinical virulence
of nosocomially acquired methicillin-resistant and methicillin-sensitive
Staphylococcus aureus infections in a university hospital. Infect Control
Hosp Epidemiol 1992; 13:587–93.
12. Peacock JE, Marsik FJ, Wenzel RP. Methicillin-resistant Staphylococcus
aureus: introduction and spread within a hospital. Ann Intern Med
1980;93:526–32.
13. Monnet DL. Methicillin-resistant Staphylococcus aureus and its rela-
tionship to antimicrobial use: possible implications for control. Infect
Control Hosp Epidemiol 1998; 19:552–9.
14. Gould IM A review of the role of antibiotic policies in the control of
antibiotic resistance. J Antimicrob Chemother 1999; 43:459–65.
15. Merrer J, Santoli F, Appe´re´-De Vecchi C, Tran B, De Jonghe B, Outin
H. “Colonization pressure” and risk of acquisition of methicillin-
resistant Staphylococcus aureus in a medical intensive care unit. Infect
Control Hosp Epidemiol 2000; 21:718–23.
16. Crowcroft NS, Ronveaux O, Monnet DL, Mertens R. Methicillin-
resistant Staphylococcus aureus and antimicrobial use in belgian hos-
pitals. Infect Control Hosp Epidemiol 1999; 20:31–6.
17. Se´bille V, Chevret S, Valleron AJ. Modeling the spread of resistant
nosocomial pathogens in an intensive-care unit. Infect Control Hosp
Epidemiol 1997; 18:84–92.
18. Scanvic A, Denic L, Gaillon S, Giry P, Andremont A, LucetJC. Duration
of colonization by methicillin-resistant Staphylococcus aureus after hos-
pital discharge and risk factors for prolonged carriage. Clin Infect Dis
2001; 32:1393–8.
19. Natsch S, Hekster YA, de Jong R, Heerdink ER, Herings RM, van der
Meer JW. Application of the ATC/DDD methodology to monitor an-
tibiotic drug use. Eur J Clin Microbiol Infect Dis 1998; 17:20–4.
20. Tilquin C, Saulnier D, Roussel B. PRN 87: measuring the level of
nursing care required. Montreal: EROS (Equipe de Recherche Ope´r-
ationnelle en Sante´), 1989:175.
by guest on February 2, 2016http://cid.oxfordjournals.org/Downloaded from
978 •CID 2003:36 (15 April) •Muller et al.
21. O’Brien-Pallas L, Leatt P, Deber R, Till JA. A comparison of workload
estimates using three methods of patients classification. Can J Nurs
Adm 1989; 2:16–23.
22. Loulergue J, Audurier A, DeLarbre JM, De Gialluly C. Changes in
microbial ecology and use of cloxacillin. J Hosp Infect 1994;27:275–83.
23. Lopez-Lozano JM, Monnet DL, Yague A, et al. Modelling and fore-
casting antimicrobial resistance and its dynamic relation-ship to an-
timicrobial use: a time series analysis. Int J Antimicrob Agents 2000;
14:21–31.
24. McGowan JE Jr. Antibiotic resistance in hospital organisms and its
relation to antibiotic use. Rev Infect Dis 1983;5:1033–48.
25. Monnet DL, Archibald LK, Phillips L, Tenover FC, McGowan JE Jr,
Gaynes RP, the Intensice Care Antimicrobial Resistance Epidemiology
Project and the National Nosocomial Infections Surveillance System
Hospitals. Antimicrobial use and resistance in eight US hospitals: com-
plexities of analysis and modeling. Infect Control Hosp Epidemiol
1998; 19:388–94.
26. Levy SB. Balancing the drug-resistance equation. Trends Microbiol
1994; 2:341–2.
27. Monnet DL, Lopez-Lozano JM, Campillos P, Burgos A, Yague A, Gon-
zalo N. Making sense of antimicrobial use and resistance surveillance
data: application of ARIMA and transfer functions models. Clin Mi-
crobiol Infect 2001; 7:S29–36.
28. Franck MO, Sorensen SJ, Carr JA, McComb JS, Harstein AI. Decrease
in selected nosocomial infections following implementation of an an-
timicrobial prescribing improvement program. Infect Control Hosp
Epidemiol 1996; 17:36.
29. Harstein AI, LeMonte AM, Iwamoto PKL. DNA typing and control of
methicillinresistant Staphylococcus aureus at two affiliated hospitals.
Infect Control Hosp Epidemiol 1997; 18:42–8.
30. Hill DA, Herford T, Parratt D. Antibiotic usage and methicillin-resis-
tant Staphylococcus aureus: an analysis of causality. J Antimicrob Che-
mother 1998; 42:676–7.
31. Landman D, Chockalingam M, Quale JM. Reduction in the incidence
of methicillin-resistant Staphylococcus aureus and ceftazidime-resistant
Klebsiella pneumoniae following changes in a hospital antibiotic for-
mulary. Clin Infect Dis 1999; 28:1062–6.
32. Onorato M, Borucki MJ, Baillargeon G, et al. Risk factors for colo-
nization or infection due to methicillin-resistant Staphylococcus aureus
in HIV-positive patients: a retrospective case-control study. Infect Con-
trol Hosp Epidemiol 1999; 20:26–30.
33. Soriano A, Martinez JA, Mensa J, et al. Pathogenic significance of
methicillin resistance for patients with Staphylococcus aureus bacter-
emia. Clin Infect Dis 2000; 30:368–73.
34. Cars O, Mo¨ lstad S, Melander A. Variation in antibiotic use in the
European Union. Lancet 2001; 357:1851–3.
by guest on February 2, 2016http://cid.oxfordjournals.org/Downloaded from
