Outbreak of multiresistant Acinetobacter baumannii in a polyvalent intensive care unit: Clinical, epidemiological analysis and PFGE-printing evolution

Abstract

An outbreak of multidrug-resistant Acinetobacter baumannii (MRAB) occurred over the course of a 27-week period in our adult polyvalent intensive care unit (ICU). Twenty-one patients were affected, and 72 strains were identified from different clinical samples. The strains were resistant to all antibiotics except for colistin and ampicillin/sulbactam. Forty-nine MRAB strains collected from 18 patients were analysed by pulsed-field gel electrophoresis (PFGE). This analysis revealed four highly-related PFGE types (genetic similarity index >90%) termed 1, 2, 3 and 4, that were isolated in 13, seven, one, and three patients, respectively. A single PFGE type was identified from five of ten patients with successive isolation of MRAB; in the other five patients, two or three PFGE types were detected. This suggested phased evolution of PFGE types 2, 3 and 4 from PFGE type 1. Global mortality was high (13 patients; 62%). Non-survivors had higher APACHE II scores than survivors on the date that MRAB was isolated (OR = 1.57; 95% CI [1.02, 2.44]). The outbreak was controlled after implementation of an extensive infection control program.

Full text

BRIEF REPORT
Outbreak of multiresistant Acinetobacter baumannii
in a polyvalent intensive care unit: clinical, epidemiological
analysis and PFGE-printing evolution
J. Monterrubio-Villar &C. González-Velasco &
S. Valdezate-Ramos &A. Córdoba-López &
P. Villalón-Panzano &J. A. Saéz-Nieto
Received: 30 April 2009 / Accepted: 5 July 2009 /Published online: 29 July 2009
#Springer-Verlag 2009
Abstract An outbreak of multidrug-resistant Acinetobacter
baumannii (MRAB) occurred over the course of a 27-week
period in our adult polyvalent intensive care unit (ICU).
Twenty-one patients were affected, and 72 strains were
identified from different clinical samples. The strains were
resistant to all antibiotics except for colistin and ampicillin/
sulbactam. Forty-nine MRAB strains collected from 18
patients were analysed by pulsed-field gel electrophoresis
(PFGE). This analysis revealed four highly-related PFGE
types (genetic similarity index >90%) termed 1, 2, 3 and 4,
that were isolated in 13, seven, one, and three patients,
respectively. A single PFGE type was identified from five of
ten patients with successive isolation of MRAB; in the other
five patients, two or three PFGE types were detected. This
suggested phased evolution of PFGE types 2, 3 and 4 from
PFGE type 1. Global mortality was high (13 patients; 62%).
Non-survivors had higher APACHE II scores than survivors
on the date that MRAB was isolated (OR=1.57; 95% CI
[1.02, 2.44]). The outbreak was controlled after implemen-
tation of an extensive infection control program.
In recent years, nosocomial infections caused by Acineto-
bacter baumannii have emerged as a problem worldwide
[1], particularly in patients admitted to intensive care units
(ICUs). The pathogen A. baumannii can survive for long
periods on dry surfaces [2] and on the skin of hospital staff
[3], making it difficult to control. Infection or colonization
in hospital settings have two main patterns: endemic
resistant A. baumannii strains of polyclonal origin [4,5]
and outbreaks caused by a single strain [6,7].
The aim of our study was to analyse an outbreak of
multidrug-resistant A. baumannii (MRAB) in our ICU caused
by several A. baumannii pulsed-field gel electrophoresis
(PFGE) types to characterize the clinical and epidemiological
features of the outbreak and to describe the infection control
measures that were ultimately effective in our hospital.
The MRAB outbreak occurred from July 2005 to
January 2006 in an eight-bed adult mixed ICU in a 297-
bed community hospital. Twenty-one patients hospitalized
in the ICU were infected or colonized by an epidemic
MRAB strain; 72 strains were isolated from clinical
samples. Episodes of colonization or infection were
considered acquired in the ICU if they appeared within
72 h of ICU admission. In July 2005, there were two
MRAB isolates. The first patient was a 40-year-old male
with a respiratory infection and atrial flutter with high
ventricular rate who was ventilated for a long period.
MRAB was cultured from his respiratory secretions after
33 days in the ICU. The number of MRAB isolates
increased to a maximum of 22 in August, decreased to 14
in the next three months and progressively disappeared over
the final two months of the outbreak.
The infected/colonized patients represented approxi-
mately 12.4% of the total population of ICU patients
admitted during the outbreak, and the incidence of A.
baumannii infection/colonization in 2003 and in the first six
J. Monterrubio-Villar (*):A. Córdoba-López
Intensive Care Unit, Hospital Don Benito-Villanueva,
Health Service of Extremadura,
06400 Don Benito, Badajoz, Spain
e-mail: suso1@orangecorreo.es
C. González-Velasco
Microbiology Department, Hospital Don Benito-Villanueva,
Health Service of Extremadura,
06400 Don Benito, Badajoz, Spain
S. Valdezate-Ramos :P. Villalón-Panzano :J. A. Saéz-Nieto
Bacteriology Department, National Centre of Microbiology,
Carlos III Health Institute,
28220 Majadahonda, Madrid, Spain
Eur J Clin Microbiol Infect Dis (2009) 28:1281–1284
DOI 10.1007/s10096-009-0777-6

months of 2004 was 2.4% and 2.2%, respectively. The
strains isolated in 2003–2004 were susceptible to
imipenem, β-lactams and aminoglycosides. There were
no cases of A. baumannii in the ICU in the year prior to
thisoutbreak(fromJuly2004toJuly2005).
The demographics, chronic health status according to
the McCabe and Jackson classification, chronic diseases,
severity of illness as reflected by the APACHE II score,
risk factors for infection/colonization, clinical status,
microbiological data and outcome were analysed and
compared between survivors and non-survivors.
To test for differences between survivors and non-survivors,
we used the Wilcoxon-Mann-Whitney test for quantitative
variables, and Pearson's chi-squared test or Fisher's exact test
for qualitative variables. Statistical significance was set at
P≤0.05. Finally, a multivariate study was performed using
stepwise logistic regression analysis to evaluate the variables
or factors associated with mortality. Independent variables
and factors were tested for correlation with the occurrence of
death using multivariate analysis. The magnitude of the effect
was measured from the corresponding odds ratios (ORs) and
from the 95% confidence intervals (CIs).
MRAB was collected from clinical specimens using
standard methods [8] and isolated by culture with
MacConkey agar plates. Acinetobacter isolates were iden-
tified by standard biochemical reactions using the Wider
system with gram-negative card (Soria Melguizo, Spain)
and the Api NE system (bioMérieux, France). Susceptibility
to the following antimicrobials was determined by the
microdilution method using the Wider commercial system
with the gram-negative card (Soria Melguizo, Spain) and by
the disk diffusion method [9]: ampicillin, ampicillin/
sulbactam, piperacillin/tazobactam, cefepime, ceftazidime,
cefotaxime, aztreonam, imipenem, meropenem, gentamicin,
tobramycin, amikacin, ciprofloxacin, ofloxacin and colistin
[8]. The Clinical and Laboratory Standards Institute (CLSI)
standard breakpoints available for Acinetobacter spp.
organisms were used to determine susceptibility. In the
absence of breakpoints, the Enterobacteriaceae criteria
were used. For analysis, we grouped the ‘intermediate
resistance’antibiotics with those in the ‘resistance’category.
Multidrug resistance was defined as resistance to at least four
separate classes of antibiotics. Escherichia coli ATCC 25922
was used as the control strain [9].
PFGE type characterization was carried out as follows.
Whole-cell DNA from MRAB strains was embedded in
agarose, lysed, and cleaved with the restriction enzyme
ApaI (60 U, Roche Diagnostic, Barcelona. Spain) [10].
The relatedness of the PFGE types was determined using
Infoquest (Applied Maths, Sint-Martens-Latem, Belgium)
to compare the PFGE fingerprints. The percent similarity
between chromosomal fingerprints was scored by the
unweighted pair group method using the averaging
algorithm (UPGMA) and Dice coefficient [11]. A band
position tolerance of 1.0% was used to analyse PFGE
patterns. PFGE types were considered distinct when they
differed by at least one band.
Seventy-two strains of MRAB were isolated from 21
patients resulting in a mean value of 3.4 isolates per
patient. There were 36 clinical samples from the
respiratory tract of 20 patients, 17 isolates from catheter
tips of nine patients, ten positive blood cultures in eight
patients, five isolates from urine in four patients and four
positive wound exudates in three patients. All MRAB
strains showed the same antimicrobial susceptibility
profile, i.e. resistance to all tested antimicrobials with
imipenem MICs values ≥16 μg/ml, except for ampicillin/
sulbactam (<8/4 μg/ml) and colistin (≤2μg/ml).
After MRAB isolation, five patients were treated with
intravenous colistin, aerosolized colistin and ampicillin/
sulbactam; two patients were treated with intravenous
colistin and ampicillin/sulbactam; five patients were treated
with aerosolized colistin and ampicillin/sulbactam; one
patient was treated with intravenous and aerosolized
colistin; one patient was treated with ampicillin/sulbactam;
and seven patients received no specific antibiotic therapy.
With regard to the latter group, five of the seven patients
died. The daily doses of intravenous sulbactam and
aerosolized and intravenous colistin ranged from 2–6g,
1–6 million IU and 2.5–6 million IU, respectively, with
adjustment of the dosage dependent on renal function.
The ICU mortality rate was 57.1% (12 patients) and the
global mortality rate was 61.9% (13 patients). Microbio-
logical evolution was harmless in 57.1% of the total
population and fatal in 42.9%.
Comparing survivors and non-survivors using univariate
analysis, the following variables showed statistical signifi-
cance: the presence of cardiovascular disease (37.5% vs
92.3%, P=0.014), favourable microbiological evolution
(100% vs 30.8%, P=0.005), the APACHE II score on the
date of MRAB isolation (12± 2.39 vs 16.23± 3.94, P=0.025)
and the number of days of hospitalization (75.50±43 vs
41.23±44.84, P=0.045). There were no differences with
respect to specific antimicrobial therapy with either ampicil-
lin/sulbactam or colistin (or both) after MRAB isolation.
These variables were included in a stepwise logistic
regression model that showed that the APACHE II score
on the date of MRAB isolation was independently associated
with mortality (OR= 1.57; 95% CI [1.02, 2.44]).
Molecular characterisation of 49 MRAB strains collected
from 18 patients with infections/colonizations showed four
PFGE types. The genetic similarity coefficients for these
highly-related PFGE types ranged from 90–97% as shown
in the cluster analysis (Fig. 1). PFGE type 1 was the first
type to emerge, followed by PFGE types 2, 3 and 4. The
predominant PFGE type was PFGE type 1, detected in 13
1282 Eur J Clin Microbiol Infect Dis (2009) 28:1281–1284

patients (17 tracheal aspirates, 5 blood cultures, 3 urine
cultures, 1 catheter tip, 1 wound exudate). PFGE type 2, 3
and 4 were detected in only seven, one and three
patients, respectively. The sample distribution for PFGE
type 2 was seven tracheal aspirates, two catheter tips,
one blood culture and one urine culture; PFGE type 3
was isolated from three catheter tips; and PFGE type 4
was isolated from five catheter tips, one tracheal aspirate,
one blood culture and one wound exudate. Of the 18
patients studied, eight patients had a single MRAB
episode, while ten patients experienced repeated MRAB
episodes. A single PFGE type was identified in half of
the ten patients from whom MRAB was successively
isolated. For example, PFGE type 4 was isolated in five
samples from patient number 18 over the course of two-
and-a-half months (Fig. 2). In contrast, two or three
different PFGE types were detected in the other five
patients. In patient number 8, PFGE type 1 was detected
four days after PFGE type 2 was detected; 13 days later,
PFGE type 2 was detected again; PFGE type 1 was again
detected two days after and 48 days later. Finally, PFGE
type 4 emerged three days later (Fig. 2).
We implemented infection control measures in accor-
dance with those recommended previously [4,12]ina
staged process that resulted in the disappearance of new A.
baumannii isolates in the following three years. The first
step included contacting infectious diseases specialists,
cohorting the infected/colonised patients inside the unit,
restricting carbepenem use, informing all ICU staff of the
outbreak, cohorting nursing and reinforcing the need for
hand washing with an alcohol-based gel. The second step
consisted of implementing strict barrier precautions (sterile
gloves and gowns, mask and eye protection), limiting the
number of clinical examinations and number of visits, and
rigorous cleaning of all surfaces with 0.1% hypochlorite.
There are multiple reports of MRAB outbreaks over
the last two decades in different hospital wards, espe-
cially in the ICUs of tertiary-care centres and university
hospitals [13–15]. A. baumannii infections are reported
less frequently in community hospitals [16].
Several reports suggest that A. baumannii infections are
not associated with a poorer prognosis than other infections
or control groups [17,18]. We found a positive, statistically
significant association between the APACHE II score on the
date of MRAB isolation and death. Neither septic shock nor
inappropriate antimicrobial therapy were statistically associ-
ated with a fatal outcome. This may be due to the small
number of patients of our study (n=21) or to the fact that
MRAB infection/colonization did not have any associated
mortality. We found that mortality increased according to
predicted mortality (based on the APACHE II score at ICU
admission; between 17% and 27%), but this difference could
be related to factors other than MRAB infection/coloniza-
tion. It is very difficult to assess the impact of infection on
mortality; even when the best methodologies are used, such
as matched case-control studies, the quality of the matching
process is not always assured.
PFGE typing was a useful tool in characterizing this
outbreak, and was specifically useful for discriminating
among the MRAB strains in the patients over time. All of
the MRAB recovered strains were closely related (genetic
similarity range, 90–97.5%) and were grouped into four
genotypes. PFGE type 1 was the first to be identified and
was the most commonly identified type. Over the 27-week
period, the PFGE type 1 was disseminated over 13 weeks,
while PFGE-types 2, 3 and 4 emerged one, 13 and
18 weeks later, respectively.
In patients from whom MRAB was repeatedly isolated,
some had a single type (five patients), and some had one
PFGE type replaced by other very similar types (5 patients).
This latter could be explained by the infection/colonization
Fig. 1 Genetic relatedness of A. baumanii pulsed-field gel electro-
phoresis (PFGE) types using ApaI-PFGE profiles detected in 18 ICU
patients. PFGE types are coded by time of emergence and similarity.
A percent genetic similarity scale is shown above the dendogram.
a
Patients with a single PFGE type and patients with 1, 2 or 3 PFGE
types;
b
Month/day/year
Fig. 2 Persistence and pulsed-field gel electrophoresis (PFGE) types
evolution in multidrug-resistant Acinetobacter baumannii (MRAB)
strains of ten patients with repetitive isolates, hospitalized in a
polyvalent ICU ward over a 27-week period
Eur J Clin Microbiol Infect Dis (2009) 28:1281–1284 1283

of the hospitalized patients by more than one closely related
MRAB type. Based on the known environmental longevity
of A. baumannii in the hospital setting, such related PFGE
types were most likely variants derived from a common
ancestor. Changes leading to the emergence of new PFGE
types (i.e., evolution) could occur both in the hospital
setting (for example, on contaminated surfaces) and in the
patients themselves.
A. baumannii is an important nosocomial microorganism
because of the diversity of its reservoirs, its capacity to
acquire antibiotic resistance, its resistance to desiccation, its
propensity to cause outbreaks and its epidemiological
complexity. Complete eradication and prevention of dis-
semination may require multiple interventions [19]. Our
extensive infection control program, using a multifaceted
approach based on MRAB PFGE type characterisation and
special attention to hygiene and barrier measures, was
ultimately successful. We may have underestimated the size
of the outbreak because we did not perform active
surveillance culturing. The eradication of MRAB during
the following three years supports the hypothesis of cross-
transmission between patients due to breaches in regular
infection control practices.
In conclusion, we describe an ICU outbreak of MRAB
with high lethality rate and respiratory site predominance
affecting 21 patients. There was a significant difference
between survivors and non-survivors with respect to their
APACHE II scores on the date of A. baumannii isolation;
therapy with specific antimicrobial agents was not associ-
ated with survival. PFGE analysis of 49 MRAB strains
collected from 18 patients yielded four highly-related PFGE
types and suggested phased evolution of the MRAB types
in the hospital setting. An infection control policy based
mainly on hygiene and barrier measures successfully
contained the MRAB outbreak and has thus far prevented
recurrence of the pathogen.
Sources of funding This research was supported by grant MPY
1116/07 from the Instituto de Salud Carlos III.
Conflicts of interest statement The authors declare no conflicts of
interests.
References
1. Villegas MV, Hartstein AI (2003) Acinetobacter outbreaks, 1977–
2000. Infect Control Hosp Epidemiol 24:284–295
2. Wendt C, Dietze B, Dietz E, Rüden H (1997) Survival of
Acinetobacter baumannii on dry surfaces. J Clin Microbiol
35:1394–1397
3. Riley TV, Webb SA, Cadwallader H, Briggs BD, Christiansen L,
Bowman RA (1996) Outbreak of gentamicin-resistant Acinetobacter
baumannii in an intensive care unit: clinical, epidemiological and
microbiological features. Pathology 28:359–363
4. Corbella X, Montero A, Pujol M, Dominguez MA, Ayats J,
Argerich MJ, Garrigosa F, Ariza J, Gudiol F (2000) Emergence
and rapid spread of carbapenem resistance during a large and
sustained hospital outbreak of multiresistant Acinetobacter bau-
mannii. J Clin Microbiol 38:4086–4095
5. Garcia-Garmendia JL, Ortiz-Leyba C, Garnacho-Montero J, Jiménez-
Jiménez FJ, Pérez-Paredes C, Barrero-Almodóvar AE, Gili-Miner M
(2001) Risk factors for Acinetobacter baumannii bacteremia in
critically ill patients: a cohort study. Clin Infect Dis 33:939–946
6. Bou G, Cerveró G, Dominguez MA, Quereda C, Martínez-Beltrán
J (2000) Characterization of a nosocomial outbreak caused by a
multiresistant Acinetobacter baumannii strain with a carbapenem-
hydrolyzing enzyme: high-level carbapenem resistance in A.
baumannii is not solely to the presence of beta-lactamases. J Clin
Microbiol 38:3299–3305
7. del Mar Tomas M, Cartelle M, Pertega S, Beceiro A, Llinares P,
Canle D, Molina F, Villanueva R, Cisneros JM, Bou G (2005)
Hospital outbreak caused by a carbapenem-resistant strain of
Acinetobacter baumannii: patient prognosis and risk-factors for
colonisation and infection. Clin Microbiol Infect 11:540–546
8. Schreckenberger PC, von Graevenitz A (1999) Acinetobacter,
Alcaligenes, Moraxella, Methylobacterium, and other nonfermen-
tative gram-negative rods. In: Murray PR (ed) Manual of clinical
microbiology, 7th edn. ASM Press, Washington DC, pp 539–560
9. Clinical and Laboratory Standards Institute (2008) Performance
standards for antimicrobial susceptibility testing. 18th supplement,
M100-S18. CLSI, Wayne, PA, USA
10. Seifert H, Dolzani L, Bressan R, van der Reijden T, van Strijen B,
Stefanik D, Heersma H, Dijkshoorn L (2005) Standardization and
interlaboratory reproducibility assessment of pulsed-field gel
electrophoresis-generated fingerprints of Acinetobacter bauman-
nii. J Clin Microbiol 43:4328–4335
11. Nei M, Li WH (1979) Mathematical model for studying genetic
variation in terms of restriction endonucleases. Proc Natl Acad Sci
USA 76:5269–5273
12. Garner JS (1996) Guideline for isolation precautions in hospitals.
The hospital infection control practices advisory committee. Infect
Control Hosp Epidemiol 17:53–80
13. Das I, Lambert P, Hill D, Noy M, Bion J, Elliott T (2002)
Carbapenem-resistant Acinetobacter and role of curtains in an
outbreak in intensive care units. J Hosp Infect 50:110–114
14. Denton M, Wilcox MH, Parnell P, Green D, Keer V, Hawkey PM,
Evans I, Murphy P (2004) Role of environmental cleaning in
controlling an outbreak of Acinetobacter baumannii on a
neurosurgical intensive care unit. J Hosp Infect 56:106–110
15. Katragkou A, Kotsiou M, Antachapoulos C, Benos A, Sofianou
D, Tamiolaki M, Roilides E (2006) Acquisition of imipenem-
resistant Acinetobacter baumannii in a pediatric intensive care
unit: a case-control study. Intensive Care Med 32:1384–1391
16. Weernink A, Severin WP, Tjernberg I, Dijkshoorn L (1995)
Pillows, an unexpected source of Acinetobacter. J Hosp Infect
29:189–199
17. Garnacho-Montero J, Ortiz-Leyba C, Fernández-Hinojosa E,
Aldabó-Pallás T, Cayuela A, Marquez-Vácaro JA, Garcia-Curiel
A, Jiménez-Jiménez FJ (2005) Acinetobacter baumannii
ventilator-associated pneumonia: epidemiological and clinical
findings. Intensive Care Med 31:649–655
18. Alvarez-Lerma F, Palomar M, Insausti J, Olaechea P, Cerdá E,
Castillo F, Martínez-Pellús A (2005) Infections caused by
Acinetobacter spp. in critically ill ICU patients. Enferm Infecc
Microbiol Clin 23:533–539
19. Cefai C, Richards J, Gould FK, McPeake P (1990) An outbreak of
Acinetobacter respiratory tract infection resulting from incomplete
disinfection of ventilatory equipment. J Hosp Infect 15:177–182
1284 Eur J Clin Microbiol Infect Dis (2009) 28:1281–1284