Page 1
Geographic Distribution of Staphylococcus aureus
Causing Invasive Infections in Europe:
A Molecular-Epidemiological Analysis
Hajo Grundmann1,2*, David M. Aanensen3, Cees C. van den Wijngaard1, Brian G. Spratt3, Dag Harmsen4,
Alexander W. Friedrich5, the European Staphylococcal Reference Laboratory Working Group"
1National Institute for Public Health and the Environment, Bilthoven, The Netherlands, 2Department of Medical Microbiology, University Medical Centre, Groningen, The
Netherlands, 3Department of Infectious Disease Epidemiology, Imperial College London, London, United Kingdom, 4Department of Periodontology, University Hospital
Mu¨nster, Germany, 5 Institute of Hygiene, University Hospital Mu¨nster, Germany
Abstract
Background: Staphylococcus aureus is one of the most important human pathogens and methicillin-resistant variants
(MRSAs) are a major cause of hospital and community-acquired infection. We aimed to map the geographic distribution of
the dominant clones that cause invasive infections in Europe.
Methods and Findings: In each country, staphylococcal reference laboratories secured the participation of a sufficient
number of hospital laboratories to achieve national geo-demographic representation. Participating laboratories collected
successive methicillin-susceptible (MSSA) and MRSA isolates from patients with invasive S. aureus infection using an agreed
protocol. All isolates were sent to the respective national reference laboratories and characterised by quality-controlled
sequence typing of the variable region of the staphylococcal spa gene (spa typing), and data were uploaded to a central
database. Relevant genetic and phenotypic information was assembled for interactive interrogation by a purpose-built
Web-based mapping application. Between September 2006 and February 2007, 357 laboratories serving 450 hospitals in 26
countries collected 2,890 MSSA and MRSA isolates from patients with invasive S. aureus infection. A wide geographical
distribution of spa types was found with some prevalent in all European countries. MSSA were more diverse than MRSA.
Genetic diversity of MRSA differed considerably between countries with dominant MRSA spa types forming distinctive
geographical clusters. We provide evidence that a network approach consisting of decentralised typing and visualisation of
aggregated data using an interactive mapping tool can provide important information on the dynamics of MRSA
populations such as early signalling of emerging strains, cross border spread, and importation by travel.
Conclusions: In contrast to MSSA, MRSA spa types have a predominantly regional distribution in Europe. This finding is
indicative of the selection and spread of a limited number of clones within health care networks, suggesting that control
efforts aimed at interrupting the spread within and between health care institutions may not only be feasible but ultimately
successful and should therefore be strongly encouraged.
Please see later in the article for the Editors’ Summary.
Citation: Grundmann H, Aanensen DM, van den Wijngaard CC, Spratt BG, Harmsen D, et al. (2010) Geographic Distribution of Staphylococcus aureus Causing
Invasive Infections in Europe: A Molecular-Epidemiological Analysis. PLoS Med 7(1): e1000215. doi:10.1371/journal.pmed.1000215
Academic Editor: Henry F. Chambers, University of California San Francisco and San Francisco General Hospital, United States of America
Received July 2, 2009; Accepted December 4, 2009; Published January 12, 2010
Copyright: � 2010 Grundmann et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: All contributions from the staphylococcal reference laboratories were funded by national sources. Data collection, management, and analysis were
funded by the Dutch Ministry of Welfare and Sports and the Directorate General for Health, Consumer Protection of the European Commission under the
Agreement Number -2003212 and the Wellcome Trust under grant number 030662. SeqNet.org was partially funded by the Interregio Program of the European
Union IIIA fund 2-EUR-V-1-96. The funders had no role in the design of the project, decision to publish, or preparation of the manuscript.
Competing Interests: DH is one of the developers of the Ridom StaphType software and the SpaServer mentioned in the manuscript. The client software is
distributed and sold by the company Ridom GmbH, which is partially owned by him. All other authors have declared that no competing interests exist.
Abbreviations: CA-MRSA, community-acquired methicillin-resistant Staphylococcus aureus; CI, confidence interval; CO-MRSA, community-onset methicillin-
resistant Staphylococcus aureus; EARSS, European Antimicrobial Resistance Surveillance System; MLST, multilocus sequence typing; MRSA, methicillin-resistant
Staphylococcus aureus; MSSA, methicillin-sensitive Staphylococcus aureus; PVL, Panton-Valentine Leukocidin; SRL, S. aureus Reference Laboratories; ST, sequence
type
* E-mail: Hajo.Grundmann@rivm.nl
" Membership of the European Staphylococcal Reference Laboratory Working Group is provided in the Acknowledgments.
PLoS Medicine | www.plosmedicine.org 1 January 2010 | Volume 7 | Issue 1 | e1000215
Page 2
Introduction
Staphylococcus aureus is the main cause of purulent infection
in humans [1]. S. aureus has the potential for local as well as
disseminated infection and can cause lesions in all tissues and
anatomical sites. Infections can be either acquired in the
community or in association with health care. The position of
S. aureus as one of the most important human pathogens is largely
due to its virulence potential and ubiquitous occurrence as a
coloniser in humans, domestic animals, and livestock [2]. Between
25% and 35% of healthy human individuals carry
S. aureus on the skin or mucous membranes [3]. Any injury that
compromises epithelial integrity, trauma, medical or surgical
interventions, as well as viral infections, can lead to tissue invasion
[4–6]. It is assumed that severity and outcome depend largely on
the virulence of the introduced strain and the immune repertoire
of the host [7,8]. Occasionally, S. aureus acquires enhanced
virulence and antimicrobial resistance through horizontal DNA
transfer and maintains these mobile genetic elements in a
predominantly clonal genomic background. Thus, clones of S.
aureus are relatively stable and mainly diversify by the accumula-
tion of single nucleotide substitutions in the absence of frequent
interstrain recombination. It is therefore possible to discern
different clones and clonal lineages by molecular typing [9]. This
method allows several important observations to be made
regarding the evolution, epidemiology, and spread of clones with
particular public health importance, such as hospital-, community-
, and livestock-associated methicillin-resistant S. aureus (MRSA).
For MRSA, this surveillance is particularly important because it
appears that certain clones have disseminated over wide
geographical regions and are threatening major improvements in
curative and public health medicine [10]. A geographically
detailed description of this expansion on a continent-wide scale
has been inadequate, however, due to the lack of appropriate
surveys and agreement on a consistent application of standardized
molecular typing approaches. At the same time, little is known
about the population structure and geographical abundance of
methicillin-susceptible S. aureus (MSSA), which provides the
genetic reservoir from which MRSA emerge.
The present study was designed to fill these knowledge gaps and
to provide (i) the first representative and contemporaneous
snapshot of the genetic population structure of S. aureus (based
on spa typing) that cause invasive infection in the European region;
(ii) insights into the geographic distribution of different clones
among MSSA and MRSA on a continent-wide scale; and (iii) a
public Web-based mapping tool supplying interactive access and
an intuitive illustration of the results generated by this large-scale
typing initiative. The current study was also set up to establish the
logistics for future collaborative studies that will continue to
improve crucial knowledge for clinicians and diagnostic laborato-
ries about the geographic and temporal dynamics of the MSSA/
MRSA clones and their epidemic patterns in neighbouring
geographical areas. Lastly, the study is intended to additionally
strengthen the role of the S. aureus Reference Laboratories (SRLs)
by exposing and communicating potentially important public
health threats to health professionals and the general public.
Methods
spa Typing
Epidemiological typing uses highly discriminatory genetic
markers that characterize human pathogens allowing the identi-
fication of isolates that are closely related due to recent common
ancestry. The spa locus of S. aureus codes for protein A, a species-
specific gene product known for its IgG binding capacity. This
locus is highly polymorphic due to an internal variable region of
short tandem repeats, which vary not only in numbers but also
because of nucleotide substitutions within individual repeat units
[11]. DNA sequences of the spa gene therefore provide portable,
unambiguous, and biologically meaningful molecular typing data,
which have demonstrated their utility for epidemiological purposes
such as transmission and outbreak investigations at various
geographical levels [12,13].
Capacity Building
During three workshops organised for technical personnel from
28 European SRLs, participants received hands-on training in spa
sequence typing and spa sequence analysis according to a standard
protocol using a purpose-designed software tool, StaphType,
developed by Ridom GmbH [13]. Proficiency testing was carried
out by mailing each SRL five well-characterized S. aureus isolates
and five sequence chromatograms (trace files) of known spa types
as described previously [14,15]. All laboratories participating in
the structured survey described here fulfilled quantifiable quality
criteria, which consisted of an unambiguous base-calling of all
sequenced nucleotides for both forward and reverse sequencing
runs of the test panel.
Structured Survey
A protocol was drawn up and circulated for comment to all
SRLs and agreed upon in April 2006. Following this protocol,
European SRLs were asked to identify approximately 20
laboratories that serve hospitals and that are geographically and
demographically representative of their country, and secure their
participation. These laboratories were chosen from those that
regularly participate in the European Antimicrobial Resistance
Surveillance System (EARSS). For 6 mo, from September 2006
until February 2007, these participants were asked to submit the
first five successive MSSA isolates and the first five successive
MRSA isolates from individual patients with invasive infection. An
invasive infection was defined as a localised or systemic
inflammatory response to the presence of S. aureus at otherwise
sterile anatomical sites. Isolates were dispatched by the partici-
pating laboratories to the SRLs accompanied by additional
information, including the EARSS laboratory identifier, the
sample number, the date of isolation, origin of clinical specimen,
demographic detail (such as age and gender), epidemiological
context (hospital-acquired when symptoms developed more than
48 h after admission or as community-onset otherwise), isoxazo-
lylpenicillin- (i.e., oxacillin) or cefoxitin-resistance, and all-cause
mortality 14 d after isolation of the first invasive isolate. SRLs
confirmed MRSA by mecA PCR or determination of minimum
inhibitory concentration for oxacillin together with PBP2a
agglutination. Additional data were uploaded to the database
and Web application if available. These consisted of staphylococ-
cal cassette chromosome mec (SCCmec) typing, and identification of
luk-PV genes, indicative of Panton-Valentine Leukocidin (PVL).
All SRLs preserved the isolates in strain collections and performed
spa sequence typing according to the standard protocol, uploaded
the sequence information, and made this available by synchroni-
sation with the central Ridom SpaServer (www.spaserver.ridom.
de) curated by SeqNet.org at the Institute of Hygiene, University
Hospital Mu¨nster, Germany [15,16]. Submitted sequences were
quality controlled by comparison with accompanying chromato-
grams (trace files) and excluded if stringent quality criteria for
excellent sequencing data were not fulfilled. spa types were
grouped into spa complexes if a single genetic event such as single
insertions, single deletions, or single nucleotide polymorphism
Geographic Distribution of S. aureus
PLoS Medicine | www.plosmedicine.org 2 January 2010 | Volume 7 | Issue 1 | e1000215
Page 3
could account for the observed sequence divergence. In the
following the designation of spa types indicated by lower case ‘‘t’’
follow the nomenclature used by the spa server and multilocus
sequence types are reported as sequence type (ST) according to
convention [17]. Finally, the SCCmec type is also added to a string
consisting of spa type/ST/SCCmec all in parenthesis, e.g., (t032/
ST22/SCCmecIV).
Epidemiological and typing data were communicated in parallel
to a central SQL database at the National Institute for Public
Health and the Environment (RIVM) of the Netherlands. For each
participating laboratory, SRLs also provided the postal address
and indicated their administrative region (such as region, province,
state, department, or NUTS-2 level) if the location of laboratories
were to be aggregated on a higher administrative geographical
level for display on the interactive mapping tool (which was done
only for Austria, Belgium, Czech Republic, and Poland). All data
were anonymous and collected in accordance with the European
Parliament and Council decision for the epidemiological surveil-
lance and control of communicable disease in the European
community [18,19]. Ethical approval and informed consent were
thus not required.
Data Analysis, Geographical Illustration, and Cluster
Identification
All data were inspected for inconsistencies and analysed on a
country-by-country basis and returned to SRLs for feedback,
clarification of inconsistencies and final approval in February 2008.
After final approval, data were analysed using Stata version 10 and
SAS version 9.1 (SAS Institute Inc.) using chi-square test for
proportions and Student’s t-test for continuous variables. The index
of diversity is an unbiased measure of the probability of drawing two
different spa types given the distribution of spa types in the sample.
The 95% confidence intervals (CIs) were calculated as described
previously [20]. Postal address and location of all sampling
laboratories were converted into decimal cartesian coordinates
using the geocoding facility at www.spatialepidemiology.net [21].
Pairwise distances of laboratories that isolated MSSA and MRSA
with identical spa types were calculated and distance matrices
were summarised and compared by nonparametric tests. The
Web application SRL-Maps (http://www.spatialepidemiology.net/
srl-maps) was developed to interrogate the data on the basis of
mapping of laboratory locations and on centroids of administrative
regions (when laboratory results were aggregated at the level of
administrative region).
A spatial scan statistic was used to identify the geographic
distribution of specific spa types at two levels: (i) country-specific
frequencies that take into account all spa types within national
boundaries and (ii) regional clusters of varying size independent of
national boundaries. To identify spa types with higher than
expected occurrence in any of the participating countries, the
observed number of all spa types isolated within each country was
compared with the number expected under the assumption of a
European-wide random distribution. For the identification of
regional clusters, circular windows around all laboratory locations
were projected. For each location, the radius of the window was
varied from 0 to 1,000 km. In this way, a finite number of distinct
windows was created. For each window, the observed number of
isolates with a specific spa type was compared with the expected
number under the assumption of a random distribution. 10,000
random distributions were obtained by varying the composition of
local spa types at all laboratory locations consistent with
cumulative spa-specific frequencies using Monte Carlo Simula-
tions. The test statistic was calculated by log-likelihood ratio test,
whereby countries or windows where the observed spa type
frequencies differed from those obtained by simulation were
considered to contain significant clusters with an alpha error of
p,0.0001. The scan-statistic was executed with SaTScan software
using SAS macros [22,23].
Results
Summary Statistics
26 SRLs from 26 countries satisfactorily fulfilled the proficiency
criteria for spa sequence typing and contributed data for final
analysis. Between September 2006 and February 2007, 357
laboratories serving 450 hospitals (Figure 1) collected 2,890
successive MSSA and MRSA isolates from patients with invasive
Figure 1. Locations of participating laboratories.
doi:10.1371/journal.pmed.1000215.g001
Geographic Distribution of S. aureus
PLoS Medicine | www.plosmedicine.org 3 January 2010 | Volume 7 | Issue 1 | e1000215
Page 4
S. aureus infection (2,603 from blood cultures, 90.1%; 17 from
cerebrospinal fluid, 0.6%; and 270 from puncture fluids of other
normally sterile anatomical sites, 9.3%). Final inspection of data
revealed missing information for gender (28 isolates, 1%), age (54
isolates, 1.9%), sampling dates (74 isolates, 2.6%), epidemiological
context (community onset or hospital acquisition, 568 isolates
19.7%), all-cause mortality 14 d after S. aureus isolation (1052
isolates, 36.4%), and spa type (40 isolates, 1.4%). Table 1 gives a
summary overview of the number of participating laboratories,
isolates, and spa types submitted by country. Many laboratories
were unable to collect five invasive MRSA isolates within the
sampling period because of a low MRSA incidence in the hospitals
they serve. Therefore, the combined collection consisted of
two-thirds MSSA (1,923; 66.5%) and one-third MRSA (967;
33.5%, Table 2). Patients with invasive MRSA infections were
older (Figure 2) with a median age of 69 y compared to a median
age of 63 y in MSSA patients (p,0.001). Moreover, MRSA
patients had higher all-cause mortality (20.8%) 14 d after the first
isolation of S. aureus than MSSA patients (13.2%, p,0.0001). More
males (1,765/2,862; 61.7%) than females had invasive S. aureus
infections. The proportion of MRSA compared to MSSA did not
differ between the sexes (p=0.2). Of the 231 MRSA that were
reported as community-onset (CO-MRSA), 226 (97.8%) were
tested for the presence of PVL-associated genes luk-PV and ten
(4.4%) were positive. Of the 585 hospital-acquired MRSA (HA-
MRSA), 551 (94.2%) were tested for PVL and six (1.1%) were
positive. The difference was significant (p=0.009).
spa Typing
A total of 660 spa types were reported (Table 2). Among all spa
types, 565 and 155 were assigned to MSSA and MRSA,
respectively, of which, 505 were exclusively identified as MSSA
and 95 for MRSA alone. 60 spa types contained both MSSA and
MRSA. 27 of the MSSA (1.4%) and 13 of the MRSA (1.3%)
isolates were nontypeable. Table 3 shows the rank order of the
most frequent spa types isolated during the survey and Table 4 the
three most frequent spa types by country. MRSA was less diverse
than MSSA. Only five spa types accounted for almost half (48.1%)
of all MRSA isolates, whereas the same proportion of MSSA
isolates comprised the 26 most frequent MSSA spa types. Estimates
of the genetic diversity differed significantly with an index of
diversity for MSSA of 0.985 (95% CI 0.983–0.987) and 0.940
(95% CI 0.933–0.947) for MRSA. While MSSA diversity ranged
between 0.934 in Latvia and 1.0 in Iceland (unpublished data),
Table 1. Summary overview of participating laboratories, hospitals, number of invasive isolates of MSSA and MRSA, and spa types
by country.
Country n Labs n Hospitals n Isolates MSSA MRSAa n spa Types MSSA n spa Types MRSA n Not Typeable Percent Not Typed
Austria 18 48 174 120 54 70 19 1 0.6
Belgium 22 22 195 107 88 65 25 1 0.5
Bulgaria 8 8 54 29 25 23 11 0 0.0
Croatia 11 11 88 50 38 27 13 6 6.8
Cyprus 1 1 16 9 7 8 5 0 0.0
Czech Republic 20 20 145 94 51 64 9 0 0.0
Denmark 14 30 112 108 4 72 2 0 0.0
Finland 5 5 22 15 7 14 7 0 0.0
France 23 23 225 114 111 75 27 0 0.0
Germany 27 27 180 98 82 56 20 1 0.6
Greece 3 3 35 20 15 12 6 6 17.1
Hungary 10 13 110 66 44 35 9 2 1.8
Iceland 1 1 5 5 0 5 0 0 0.0
Ireland 22 22 169 85 84 55 26 0 0.0
Italy 19 19 147 80 67 53 15 0 0.0
Latvia 11 12 43 38 5 20 1 0 0.0
Malta 1 1 15 3 12 2 5 3 20.0
Netherlands 18 21 204 195 9 98 9 6 2.9
Norway 11 20 55 55 0 38 0 1 1.8
Poland 23 23 179 132 47 42 14 0 0.0
Portugal 12 12 88 48 40 36 13 0 0.0
Romania 10 10 36 25 11 18 3 0 0.0
Slovenia 11 12 58 48 10 29 3 2 3.4
Spain 21 21 204 113 91 58 19 1 0.9
Sweden 20 47 200 195 5 90 5 3 1.5
UK 15 18 131 71 60 51 18 7 5.3
Total 357 450 2,890 1,923 967 565 155 40 1.4
aThe number of MRSA isolates does not reflect a prevalence or occurrence in particular countries as the protocol requested submission of the first five MSSA and MRSA
isolates.
doi:10.1371/journal.pmed.1000215.t001
Geographic Distribution of S. aureus
PLoS Medicine | www.plosmedicine.org 4 January 2010 | Volume 7 | Issue 1 | e1000215
Page 5
MRSA diversity between countries was more heterogeneous
ranging from 0.62 in Romania to 0.91 in Austria (Figure 3),
indicating the presence of few dominant MRSA spa types in
several countries. Accordingly, MRSA showed a higher degree of
geographic clustering as the average distance between laboratories
that isolated the same spa type was significantly smaller than for
MSSA (Table 2). No correlation between genetic diversity of
MRSA and overall proportion of MRSA among S. aureus blood
stream infections at country level as reported to the EARSS
database for 2007 was found (r=20.09, p=0.75) [24], indicating
that single successful spa types cannot explain the variance in the
proportion of MRSA causing S. aureus blood stream infections
observed in Europe.
Clustering of spa Types at Country and Regional Level
In 17 countries, 22 spa types were found with frequencies that
were unexpected when applying the hypothesis of a random
distribution, indicative of local epidemics (Table 5). Most (86.9%)
of these were MRSA. In ten countries two spa types coexisted with
unexpected frequencies. In four of them, these two spa types
showed a close genetic relationship and belonged to the same spa
complex whereby a single genetic event could account for the
sequence divergence between the types. There was also a frequent
regional coincidence with neighbouring countries sharing identical
epidemic spa types. The Czech Republic and Germany shared spa
type t003 (t003/ST225/SCCmecII), Bulgaria and Romania shared
t030 (t030/ST239/SCCmecIII), the UK and Ireland t032 (t032/
ST22/SCCmecIV), Italy and Croatia shared t041 (t041/ST228/
SCCmecI), and strain t067 (t067/ST5125/SCCmecIV) whose
dominance in Spain was first identified through this initiative
[25], was also found in southern France. The notion of regional
spread was supported by the cluster statistic that projected
windows beyond national boundaries for this dataset (Table 6).
The degree of unexpectedness, which is an indication of the
significance of each cluster, is expressed by the log likelihood ratio
(LLR). The majority of regional clusters extended beyond national
boundaries and 74% of all isolates that occurred in these clusters
were MRSA. The most significant cluster was identified in Spain
and consisted of spa type t067 (t067/ST5&125/SCCmecIV). A
northern Balkan/Adriatic cluster consisting of spa type t041 (t041/
ST228/SCCmecI) was found in Austria, Hungary, Slovenia,
Croatia, and northern and central Italy. In Britain and Ireland,
t032 (t032/ST22/SCCmecIV), known as epidemic MRSA 15
(EMRSA-15), was the dominant type and represented the third
most significant cluster. An additional cluster of spa type t032,
albeit less significant and much smaller, was located in the
Brandenburg area of Germany. Central Germany, the Czech
Republic, and western Poland were included in a large regional
cluster of spa type t003 (t003/ST225/SCCmecII), which was
geographically centred in Saxony and had a radius of approx-
imately 400 km corresponding to the German hospitals partici-
pating in the study. Figure 4 provides a geographical illustration of
these clusters. The largest cluster in size as well as in numbers
(radius 930 km, 119 isolates) consisted of spa type t008 and was
centred in southern France. This cluster consisted of ST8 and
contained different subclones as it included both MSSA and
MRSA, and MRSA isolates exhibited two different SCCmec
elements (SCCmecI and IV). A smaller cluster, ranking in sixth
position in terms of significance, was located in Flanders on the
Belgian-Dutch border and consisted of spa type t740 (t740/ST45/
SCCmecIV). Interestingly, regional spa clusters with overlapping
geographical range were frequently made up of spa types that
belonged to the same spa complex, a clear indication that local
spread is accompanied by local evolution of the rapidly evolving
spa locus. Clusters with the smallest size (0 km) included those
submitted by single laboratories most likely reflecting single
hospital outbreaks. Three regional clusters consisted of MSSA
alone. They were located in Latvia (t435/ST425), Poland (t127/
ST1), and Denmark (t230/ST45), indicating that regional spread
of S. aureus is not limited to MRSA alone.
SRL-Maps
The Web application SRL-Maps (http://www.spatialepidemiology.
net/srl-maps) provides a community tool for the interrogation of the
geographic distribution of different spa types. All laboratory and
regional locations across Europe are represented as placemarks on
a Google map. Clicking on a placemark displays, below the map,
all spa types identified at that location (and their frequency)
along with the number of isolates (and number of geographic
locations) found elsewhere (if any) for each of these
Table 2. Summary statistics of S. aureus isolated in 26 European countries.
Statistics na MSSA MRSA Total/Overall p-Value*
Frequency (%) 2,890 1,923 (66.5) 967 (33.5) 2,890 (100%) —
Median age (IQR) 2,836 63 (46–75) 69 (55–78) 66 (49–76) ,0.0001
Male gender (%) 2,862 1,159 (60.8) 606 (63.3) 1,765 (61.7) 0.2
All-cause mortality after 14 d (%) 1,838 153 (13.2) 141(20.8) 294 (16.0) ,0.0001
Hospital acquisition (%) 2,322 777 (51.6) 585 (71.7) 1,362 (58.7) ,0.0001
N spa types 2,850 565 155 660b —
N not typeable 2,850 27 (1.4) 13 (1.3) 40 (1.4) 0.9
Index of diversity (95% CI) 2,850 0.985 (0.983–0.987) 0.940 (0.933–0.947) 0.983 (0.982–0.984) ,0.05c
Mean distance in kilometres between laboratories
that isolated identical spa types (95% CI)
1,614d 1,046.2 (1109.5–983.0) 786.8 (975.7–597.9) — 0.03d
*p-Value for the comparison of MSSA versus MRSA.
aNumber of isolates for which data were available.
bTotal number of spa types includes 60 spa types that contain both MSSA and MRSA.
cDeduced from non-overlapping 95% confidence intervals.
dIncludes only MRSA and MSSA with more than ten isolates per spa type.
IQR, interquartile range.
doi:10.1371/journal.pmed.1000215.t002
Geographic Distribution of S. aureus
PLoS Medicine | www.plosmedicine.org 5 January 2010 | Volume 7 | Issue 1 | e1000215
End of preview.
