epidemiology of nosocomial pneumonia in the intensive care unit of bach mai hospital, 2008-2009

ABBREVIATIONS


A. baumannii
A. baumannii
AFLP
Amplified Fragment Length Polymorphism
CDC
Centers for Disease Control and Prevention
CI
Confidence Interval
COPD
Chronic Obstructive Pulmonary Disease
DNA
Deoxyribo Nucleic Acid
GNB
Gram-Negative Bacteria
HAI
Healthcare - Associated Infection
HAP
Healthcare-Associated Pneumonia
HCFs
Health Care Facilities
HCW
Health Care Worker
HIV
Human Immunodeficiency Virus
IC
Infection Control
ICU
Intensive Care Unit

MDRO
Multidrug-Resistant Organism
MRSA
Methicillin-Resistant Staphylococcus aureus
OR
Odds Ratio
PCR
Polymerase Chain Reaction
PFGE
Pulse-Field Gel Electrophoresis
P.aeruginosa
Pseudomonas aeruginosa
WHO
World Health Organization

1


INTRODUCTION

Healthcare-associated pneumonia (HAP) is defined as pneumonia
that occurs in patients that reside in or have resided in a long-term care
facility, acute-care facility, or other healthcare facility. HAP occurs for at
least 48 hours after admission to the hospital/healthcare facility.
HAP is currently the second most common nosocomial (hospital-
acquired) infection and is the leading infection among those detected in
ICUs. 83% of episodes of HAP were associated with mechanical
ventilation. Microorganisms colonize the stomach, upper airway and
bronchi, and cause infection in the pneumonia. They are often endogenous
(digestive system or nose and throat), but may be exogenous, often from

contaminated respiratory equipment. According to the findings of some
studies conducted in the United States and in the Europe, the mortality rate
in the patients with of HAP caused by common multi-antibiotic resistant
bacteria such as A. baumannii and P.aeruginosa, account for more than
70%; these pathogens are uncommon in non-ICU settings.
HAP is the most common HAI in Vietnamese ICUs and it can extend
hospitalization by an average of 7 to 14 days per patient and increase the
costs of hospitalization. Mechanical ventilation is found the important risk
factor associated with HAP. A. baumannii and P.aeruginosa are the
leading isolated microorganisms contributing to HAI. More than 60%
A.
baumannii and P.aeruginosa
are resistant to commonly used antibiotics.
The emergence and dissemination of resistant organisms is considered as
one of the challenging problems in medical science with clinical,
economical, and public health implications.

There remains a paucity of information on the magnitude, risk
factors, as well as antibiotic resistance and molecular characteristics of
pathogens causing HAP in the Vietnamese ICUs. The lack of such
information represents a challenge for HAP control and prevention. The
objectives of this study were to determine: (1) The incidence of and risk
factors for HAP in ICU of Bach Mai hospital; and (2) Pathogens, antibiotic
resistance and molecular characteristics of common bacteria causing HAP.

SUMMARY OF NEW CONTRIBUTIONS

This is the first thesis in Vietnam aimed to describe comprehensively
the incidence of, risk factors for HAP, as well as antibiotic resistance and
molecular characteristics of pathogens causing HAP in the Intensive care

2


unit of Bach Mai Hospital which is the area with the highest incidence
density of healthcare associated infections among Vietnamese HCFs.
HAP was commonly observed in patients of Intensive care unit, Bach
Mai hospital. The incidence of HAP accounted for 18.9%. The overall
HAP density was 11.6 per 1,000 patient-days. Higher densities found in
patients with tracheotomy and endotracheal intubation. The HAP rates per
1,000 tracheotomy and endotracheal intubation-days were 27.4 and 72.1,
respectively. Risk factors for nosocomial found by logistic regression
analysis were: (1) Chronic respiratory diseases (OR = 1.9; p < 0.001), (2)
Endotracheal intubation (OR = 3.9; p < 0.05), (3) Endotracheal intubation
(OR = 6.3; p < 0.01), (4) Surgical procedure (OR = 2.5; p < 0.05). The
thesis not only confirmed the increased trend in HAP caused by A.
baumannii but also recognized the growth of antibiotic resistances of this
pathogen. Carbapenem is one of most effective antibiotics against A.
baumannii. However, the percentages of A. baumannii isolates resistant to
imipenem and meropenem were 84.9% and 86.8%, respectively, which
were much higher than recent studies conducted in US and Europe.
PFGE analysis revealed that a total of 46 (86.7%) among 53 A.
baumannii isolates belong to 6 major clusters (A, B, C, D, E and F) with
the high similarity index of DNA pattern ranged from 80.0% to 96.1%.
Clone D was the most predominant and had been detected through the
entire 8-month study period. These findings suggest a hypothesis that is
consistent with studies showing the cross-transmission of A. baumannii
between patients. The dissemination of this pathogen is facilitated by poor
compliance with aseptic techniques and contributes to high colonization
rates among hospitalized patients, healthcare workers and frequent
contamination of environments. A. baumannii colonizing might easily

transmitted to other patients via invasive procedures.
The thesis shows scientific evidences on HAP, antibiotic resistance
and molecular characteristics of pathogens causing HAP. These findings
suggest areas for intervention and for developing guideline for prevention
and control of HAP in HCFs.


THESIS STRUCTURE

The thesis consists of 129 pages. Background: 3 pages; Overview: 39
pages; Methods: 20 pages; Results: 33 pages; Discussion: 31 pages;
Conclusions: 2 pages; Recommendations: 1 page; 45 tables ans 18
3


illustrations; 155 references including 37 Vietnamese and 118 English
ones.

Chapter 1. OVERVIEW

1.1. Epidemiology of HAP
1.1.1. HAP in the world: HAP is currently the second most common
nosocomial (hospital-acquired) infection and is the leading infection
among those detected in ICUs. More than 80% of episodes of HAP were
associated with mechanical ventilation. Previous multicenter cohort studies
that consisted of more than 5 mixed ICUs of multidisciplinary hospitals
had reported nearly comparable infection rates for ventilator associated
pneumonia: 14.8 cases per 1,000 ventilator-days in Canadian ICUs, 13.3
cases per 1,000 ventilator-days in German ICUs, 9.4 cases per 1,000
ventilator-days in French ICUs, and 12.6 per 1,000 ventilator-days in

Japan. HAP is the most common HAI in Vietnamese ICUs and it can
extend hospitalization by an average of 7 to 14 days per patient and
increase the mortality rate by from 20% to 30%. The highest rate of death
is found in patients with HAP which was caused by multi-antibiotic
resistant bacteria such as A. baumannii and P.aeruginosa.
1.1.2. HAP in Vietnam: A one day-prevalence study at 36 hospitals
across Vietnam in 2008 showed an overall HAI prevalence of 7.8%. The
most common infection was pneumonia, account for > 60% of the detected
HAIs. HAP rates ranged from 20% to 25% in recent years. Respiratory
invasive procedures (endotracheal intubation and tracheotomy) are the
most important risk factors for HAP
. The increased trend in HAP caused
by multi-antibiotic resistant negative bacteria is recognized. The
percentages of A. baumannii and P.aeruginosa isolates resistant to
imipenem in ICU, Bach Mai hospital increased from < 20% (2002) to >
40% (2006).
1.1.3. Etiology of HAP:
Bacteria have been the most frequently isolated
pathogens. Among microorganisms isolated from oropharyngeal aspiration
and sputum, anaerobes and fungi account for 73% and 4%, respectively.
HAP tends to be associated with multiple organisms and the common
bacteria are gram-negative bacilli. However, MRSA and other Gram-
positive cocci including Streptococcus pneumoniae has increased in
frequency in the last few years. The data collected and reported by
hospitals participating in the National Nosocomial Infections Surveillance
(NNIS) System showed that P.aeruginosa, Enterobacter sp., Klebsiella
4


pneumoniae, Escherichia coli, Serratia marcescens and Proteus sp.

accounted for 50% isolates identified from respiratory specimens of
patients with HAP.
1.1.4.
Source of Infectious Agents: HCWs, patiens, visitors and hospital
environment.

1.1.5. Modes of transmission
1.1.5.1. Contact Transmission:
The most common mode of transmission,
contact transmission is divided into 2 subgroups: direct contact and
indirect contact. Indirect contact is important contributor to the pathogen
transmission involving the transfer of an infectious agent through a
contaminated intermediate object or person such as contaminated hands of
HCWs, patient-care devices, food, water or contaminated infusion.
1.1.5.2. Airborne transmission: Microorganisms from infectious individual
or person who is colonized with bacteria, may be inhaled by susceptible
individuals through the dissemination of either airborne droplet

nuclei or small particles (< 5μm)in the respirable size range containing
infectious agents that remain infective over time and distance (eg, spores of
Aspergillus spp and M. tuberculosis).
1.1.5.3. Droplet transmission: Respiratory droplets (> 5 μm) carrying
infectious pathogens transmit infection when they travel directly from the
respiratory tract of the infectious individual to susceptible mucosal
surfaces of the recipient, generally over short distances.
1.1.6. Susceptible individuals: age 65 years or more, underlying
conditions, especially chronic respiratory diseases, depressed level of
consciousness, thoracic or chest surgery, continuous mechanical
ventilation, use of paralytic agents, severe trauma, upper abdominal
surgery, and recent bronchoscop.

1.1.7. Pathogenesis of HAP: (1) HAP is caused by bacteria from remote
infection sites; (2) HAP is caused by bacteria from adjadcent aer
subdiaphragmatic abscesseas of lungs (subdiaphragmatic abscess,
mediastinal abscess, pleural infection v.v), (3) Aspiration of oropharyngeal
bacteria is the most common initiating event. These bacteria come from
exogenous sources (environment, medical devices, HCWs) or from
endogenous flora (hollow viscera such as respiaratory tract or
gastrointestinal tract).
1.1.8. Risk factors for HAP
1.1.8.1. Host related: age 65 years or more, underlying conditions,
especially chronic respiratory diseases, depressed level of consciousness,
5


thoracic or chest surgery, use of paralytic agents, severe trauma, upper
abdominal surgery, and recent bronchoscopy.
1.1.8.2. Device related: Tracheal intubation, continuous mechanical
ventilation, orogastric or nasogastric tube placement, and frequent (e.g.,
every 24 hours) ventilator circuit changes.
1.1.8.3. Increased colonization: Admission to the ICU, administration of
broad-spectrum antibiotics, prophylaxis for stress ulcer bleeding with
antacids or H2 blocker, exposure to contaminated medical equipment, and
inadequate hand hygiene
1.1.8.4. MDRO: Hospitalization for more than 7 days before the diagnosis
of HAP, transferred from another care facility, ventilation for more than 3
days before the diagnosis of HAP, active malignancy, AIDS, end-stage
liver or renal disease, steroids (e.g., prednisone 10 mg/day or more for
more than 7 days), active chemotherapy or radiotherapy, and
bronchiectasis. Prior antibiotic use for more than 3 days within the
previous 14 days of the diagnosis of HAP also is considered a risk factor

for resistant organisms.
1.1.8.5. Treatment related: No recommendation can be made for routinely
acidifying gastric feeding. It creates favorable conditions for the growth of
bacteria.
1.1.8.6. Others: Inhalation of contaminated aerosols or airborne microbes
can also introduce bacteria into the lower respiratory tract. This can occur
during events such as utility interruptions, remodeling, or construction
where infection control recommendations have not been followed

.
1.1.9. Control and prevention of HAP
1.1.9.1. Prevention of person-to-person transmission of bacteria:
Respiratory equipment must be properly cleaned and sterilized. Ventilator
tubing must be handled appropriately to avoid spilling condensate into the
patient’s airway. Sandard and transmission based precautions must be
applied in patient care.
1.1.9.2. Pneumococcal vaccination: Vaccinate patients at high risk for
severe pneumococcal infections.
1.1.9.3. Precautions for prevention of aspiration: rinsing oral for patients,

positioning the patient supine with the head elevated 30–45 degrees,
suctioning measures including continuous suctioning of subglottic
secretions, and minimizing the use of sedating or paralytic agents.
6


1.1.9.4. Precautions for MDRO: Selection of appropriate therapy is
essential to avoid the detrimental effects of antibiotic overuse and the
production of selective pressure for resistant organisms.
1.1.9.5. Other precautions: Educate HCWs about measures to prevent and

control HAP, conduct surveillance for HAP in ICU patients and provide
feedback on HAP rates and the compliance of HCWs with aseptic
techniques.

1.2. Antibiotic resistance characteristics of bacteria causing HAP
Growing cause of HAP including Acinetobacter baumannii,
P.aeruginosa which are associated with increasing antibiotic resistance in
HCFs. Exposure to any antibiotic active against GNB has been associated
with the emergence of multidrug-resistant. 3 classes of antibiotics have
been most frequently implicated. The use of third-generation
cephalosporins has been implicated in numerous case-control studies. In
addition, Landman found that aggregate use of cephalosporins plus
aztreonam, but not other antibiotic classes, was associated with the
presence of multidrug-resistant (including carbapenem-resistant)
Acinetobacter isolates. Numerous subsequent studies have shown that the
higher rate of multiantibiotic resistance in the strains can be responsible
for oligoclonal outbreaks. Therefore, MDROs illustrate the potential for
outbreak isolates.
Four major mechanisms of antimicrobial resistance include: (1) Drug
inactivation: Occurs when a bacterium produces an enzyme that can
destroy or inactivate the antimicrobial; (2) Alteration in target site: Drug
receptor or target sites may undergo alteration; (3) Decreased permeability
or efflux: Changes in drug permeability or an efflux of drug may be
observed as in the case of P. aeruginosa that has developed resistance to
the carbapenem; (4) Bypass of a metabolic pathway: bacteria may develop
alternative metabolic pathways to bypass the pathway that was inhibited by
the antimicrobial; resistance to trimethoprim-sulfamethoxazole commonly
occurs in this manner.

1.3. Molecular epidemiological profile of infection with multidrug-

resistant A. baumannii and Pseudomonas species
Early studies of panresistance in Pseudomonas aeruginosa showed
little evidence of clonality among these strains. However, numerous
subsequent studies have shown that these strains can be responsible for
7


oligoclonal outbreaks, particularly in ICU, clearly illustrating the potential
for person-to-person spread. It is not clear whether transmission occurred
via common environmental sources or the hands of health care workers.

1.4. Application of Molecular Techniques to the Study of HAI
1.4.1. Characteristics of typing methods

1.4.1.1. PFGE: The principle of this technique is to digest chromosomal
DNA with restriction enzymes, resulting in a series of fragments of
different sizes that form different patterns when analyzed by agarose gel
electrophoresis. By periodically changing the direction of the electrical
field in which the DNA is separated, PFGE allows the separation of DNA
molecules of over 50 kbp in length. In general PFGE is one of most
reproducible and highly discriminatory typing methods available, and it
generally is the method of choice for many hospital epidemiologic
evaluations.
1.4.1.2. Southern blot analysis: The bacterial DNA is digested using a
frequent cutting restriction enzyme, the DNA fragments are separated by
agarose gel electrophoresis, and then the fragments are transferred
(blotted) onto a nitrocellulose or nylon membrane. Next, a labeled
(colorimetric or radioactive) piece of homologous DNA is used to probe
the membrane. The discriminatory power of this method is related to the
copy numbers of the targeted genetic elements in the bacterial genome and

their distribution among the restriction fragments following
electrophoresis.
1.4.1.3. Plasmid Analysis: Typing is performed through the isolation of
plasmid DNA and comparison of the numbers and sizes of the plasmids by
agarose gel electrophoresis.
Evaluation of plasmid content is not generally
useful in delineation of strain relatedness. Plasmid analysis has been
applied in clinical situations to determine the evolution and spread of
antibiotic resistance among isolates with different PFGE profiles or among
different species of organisms within hospitals.
1.4.1.4. Typing Methods Using PCR: This technique is a biochemical in
vitro reaction that permits the synthesis of large quantities of a targeted
nucleic acid sequence. The procedure requires template DNA from the
organism being typed, two complementary oligonucleotide primers that
are designed to flank the sequence on the template DNA to be amplified,
and a heat-stable DNA polymerase. A growing number of organisms have
been studied using this approach.
8


1.4.1.5. AFLP: This is a typing method that utilizes a combination of
restriction enzyme digestion and PCR. The method utilizes the benefits of
restriction fragment length polymorphisms
analysis with the increased
sensitivity of PCR to generate profiles that are reproducible and relatively
easy to interpret and compare to those for other isolates from a nosocomial
outbreak.

1.4.2. Cost-effective application of typing methods in HAI study
Understanding pathogen distribution and relatedness is essential for

determining the epidemiology of HAIs. Molecular techniques can be very
effective in tracking the spread of nosocomial infections due to genetically
related pathogens, which would allow infection control personnel to more
rationally identify potential sources of pathogens and aid infectious disease
physicians in the development of treatment regimens to manage patients
affected by related organisms. In addition, the incorporation of molecular
testing in the infection control program for endemic HAIs is associated
with the ability to enact early interventions following the identification of
pathogen clonality, which could be an early indication of an outbreak.

Conversely, the determination of the unrelatedness of isolates (sporadic
infections), avoids triggering unneeded and costly epidemic investigations.
Cost reduction was also accomplished by earlier recognition of person-to-
person spread of isolates compared to that with traditional surveillance,
thus potentially preventing the spread to additional patients.

Chapter 2. STUDY Population, matERIALS and METHODS

2.1. Study population: Patients admitted to the ICU of Bach Mai hospital
for at least 48 hours
and microbial isolates identified from clinical samples
of suspected/confirmed patients with HAP.
2.2. Study design: Descriptive and molecular study.
2.2.1. Selection of study sample
- Study patients: Patients admitted to the ICU for at least 48 hours were
assessed during the study period from September 2008 to April 2009.
- Bacteria and fungi were isolated from patients with HAP.
2.2.2. Sample size: The caculation was based on WHO formula for
estimating a population proportion with specified relative precision. The
details are as follows:

9


z
2
(1-α/2)
.p.q
n =
p. ε
2

In which: n = Minimum sample size; z
(1-
α
/2)
= Confidence coefficient, with
confidence level of 95% → z
(1-
α
/2)
= 1.96; ε: Relative precision = 0.16, p:
Anticipated population proportion (p = 0.25) was calculated by prospective
study on HAP in ICU, Bach Mai hospital - 2002, q = 1-p. In this case, a
sample size of 450 patients would be needed. The real number of observed
patients in this study was 477 .
2.3. Methods: Data collection techniques include clinical/paraclinical
examination, microbiologic evaluation and PFGE analyses of common
bacteria causing HAP.
2.3.1. Study indicators:

2.3.1.1. HAP characteristics:(1) Incidence and density of HAP; (2)
Etiology of HAP: The distribution of HAP pathogens and their antibiotic
resistant level, (3) Risk factors for HAP and (4) Outcomes of HAP: length
of stay in ICU, patient outcomes, and hospital cost.
2.3.1.2. Molecular characteristics of common bacteria causing HAP:
Number of bacteria, number of clusters, the similarity index of each cluster
and between clusters.
2.3.2. Determination of study indicators
2.3.2.1. HAP surveillance: surveillance team included IC practitioners (an
IC nurse and doctor) and 1 representative physician and nurse from the
ICU who had been trained on the study objectives, surveillance
methodology, HAP definition, clinical sample and data collection.
2.3.2.2. HAP ascertainment: HAIs were diagnosed and ascertained using
surveillance criteria established by the Bach Mai hospital. This criteria was
adjusted from criteria of CDC, USA (1988).
HAI case definitions were
based on the combination of objective clinical findings and supportive data
(eg, radiographs, ultrasound scans, endoscopy findings, and pathology
reports). Isolates from any patients with suspected or confirmed HAIs
were identified, speciated, and tested for susceptibility to a panel of
antimicrobials susceptibility commonly used at the Bach Mai hospital.
2.3.2.3. Determination of risk factors: Based on clinical findings and the
review of nursing and medical charts. Data were collected by recording
study variables (age, gender, primary admission diagnosis,
conditions
associated with increased risk of APACHE II index scores, operation,
10


invasive procedures) in patients diagnosed with HAI and comparing with

similar data for patients without HAIs during the same study period.
2.3.2.4. Microbiologic evaluation: Microbiologic technique was performed
according to standard guidelines of Bach Mai hospital to determine the
HAP pathogens. Samples were sent to Microbiology Department, Bach
Mai hospital during 2 hours after sampling and incubated at 37
o
C under
aerobic conditions. Colony-forming units (CFUs) were subsequently
counted after 48 h. Potential pathogenic CFUs were cultured and identified
at Microbiology Department, Bach Mai hospital.
2.3.2.5. Antimicrobial susceptibility testing
- Paper disc diffusion method: Susceptibilities of bacteria are categorized
into: (1) Sensitive (S), (2) Intermediate (I) and (3) Resistance (R). The
results were interpreted following the Clinical and Laboratory Standards
Institute (NCCLS-National Committee for Clinical Laboratory
Standards).
- Broth microdilution method: Antimicrobial agent are diluted at various
concentrations. The concentration range used may vary with the drug,
the organism tested, and the site of the infection. The antimicrobial
dilutions are in 0.1 volumes in wells of a microdilution tray (usually 96
well trays. The results were interpreted following the Clinical and
Laboratory Standards Institute (NCCLS-National Committee for
Clinical Laboratory Standards). A. baumannii isolates were considered
multidrug-resistant when they showed resistance to more
than 3 of the
following 5 drug classes: cephalosporin, carbapenems, piperacillin-
tazobactam, fluoroquinolone, aminoglycoside and polymixin.
2.3.2.6. Molecular typing: PFGE was performed at the Department of
Infectious Diseases, Research Institute, National Center for Global Health
and Medicine, Tokyo. Images of ethidium-bromide-stained isolates were

converted into TIFF
TM
formats by ChemDoc (BIO-RAD) with which a
cluster analysis was performed using Figerprinting Ⅱ software (Bio-Rad
Laboratories) to construct a dendrogram. Isolates with a similarity index of
more than 80% were considered to be closely related.

2.3.3. Statistical analyses
- Analyses were performed using SPSS 12.0 at the Infection control
Department, Bach Mai hospital.
11


- The differences in proportions were compared by using χ2 tests.
Continuous variables were compared by means of the Student t test. P-
value <0.05 was considered statistically significant.
- To study risk factors, univariate analyses were first performed. All risk
factors with a univariate P value of less than 0.05 were included in a
multivariate analysis using a stepwise forward multivariable logistic
regression model.
- Isolates that differ by three fragments in PFGE analysis were considered
epidemiologically related subtypes of the same strain. Conversely, isolates
differing in the positions of more than three restriction fragments may
represent a more tenuous epidemiologic relation. These isolates were not
included in a cluster.
2.3.4. Study ethics
- Study activities and data collection methods were approved by the
Ethics and Health Research Review Committee of Bach Mai hospital
and Ministry of Health, Vietnam.
- All prospective research participants were fully informed about the

study objectives, research activities. The principle of voluntary
participation requires that people not be coerced into participating in
research. In the case of the patients who were unable to discuss study
issues and poor decision making, their relatives were explained about
the study.
- Study results were timely notified director board and physicians, and
aid them in development appropriate treatments and interventions.

Chapter 3. STUDY RESULTS

3.1. Incidence and risk factors for HAP
3.1.1. Incidence
387 (81,1%)
P a tients w ith H A P
P a tie n ts w ith o ut H A P
90 (18,9%)
Figure 3.1. The incidence of HAP

12


Figure 3.1 shows that from September 2008 to April 2009, a total of
477 patients were included in this study. Of them, 90 acquired HAP. The
overall HAP incidence was 18.9%.


Table 3.11. Incidence density rate of HAP (n = 477)
Exposure factors
No. of
HAP

No. of
exposure days
No. of HAP per/1.000
exposure days
Hospitalization days
90
7.748 11,6
Endotracheal
intubation-days
2.304 39,1
Tracheotomy days 1.249 72,1
Mechanical
ventilator days
3.281 27,4
Table 3.11 shows that HAP densities in patients with tracheotomy,
endotracheal intubation and mechanical ventilator were 72.1, 39.1 and
27.4, respectively. Numbers of HAP per 1.000 hospitalization days were
11.6.

3.1.2. Risk factors for HAP
Table 3.25. Risk factors for HAP per logistic regression analysis
Variables Adjusted OR 95% CI p
APACHE II index of ≥ 13 1.9 1.1 – 3.3 < 0.05
Chronic respiratory diseases 1.9 1.1 – 3.3 < 0.05
Endocrine diseases 2.0 0.9 – 4.3 > 0.05
Surgery 2.5 1.1 – 6.1 < 0.05
Endotracheal intubation 3.9 1.1 – 11.2 < 0.05
Tracheotomy 6.3 3.3 – 11.9 < 0.01
Indwelling urinary catheter 1.4 0.8 – 2.5 > 0.05
Central venous catheter 1.5 0.9 – 2.7 > 0.05

Stomach tube 1.7 0.6 – 4.4 > 0.05
Table 3.25 shows that risk factors for HAP found by logistic
regression analysis including: (1) Chronic respiratory diseases (OR = 1.9;
p < 0.001), (2) Endotracheal intubation (aOR, 3.9; p < 0.05), (3)
Tracheotomy (aOR, 6.3; p < 0.01), (4) Surgical procedure (aOR, 2.5; p <
0.05), and (5) APACHE II index of ≥ 13 (aOR, 1.9, p<0.05).

13


3.2. Pathogens, antibiotic resistance and molecular characteristics of
A. baumannii causing HAP
3.2.1. HAP pathogens
Table 3.32. Distribution of HAP pathogens
Pathogens No. (%)
A. baumannii
59 44.7
Candida spp 26 19.7
P.aeruginosa
23 17.4
Klebsiella pneumonia
7 5.3
Enterococcus species
4 3.0
Burkholderia cepacia
5 3.8
Staphylococcus aureus
2 1.5
Others 6 4.5
Total 132 100

Table 3.32 shows that the 3 most frequently isolated organisms were
A. baumannii (44.7%), Candida spp (19.7%) and P. aeruginosa (17,4%).

3.2.2. Antibiotic resistance of common pathogens
Table 3.33. The percentages of P. aeruginosa isolates resistant to used
antibiotics
Antibiotics No. of isolates
No. of resistant
isolates
(%)
Piperacillin 19 16 84,2
Ticarcilline 19 14 73,7
Levofloxacin 17 12 70,6
Tica + A.clavulanic 19 13 68,4
Gentamycine 23 15 65,2
Meropenem 19 13 68,4
Tobramycine 20 13 65,0
Ciprofloxacin 18 11 61,1
Ceffazidime 20 12 60,0
Amikacin 20 12 60,0
Imipenem 19 11 57,9
Cefepime 19 11 57,9
Aztreonam 17 11 64,7
Table 3.33 shows the high rates of P. aeruginosa isolates resistant to
all tested antibiotics (range 57.9 - 84.2%).

14


Table 3.34. The percentages of A. baumannii isolates resistant to used

antibiotics
Antibiotics No. of isolates
No. of resistant
isolates
(%)
Ampi + Sulbatam 56 53 94.6
Ceftriaxone 57 55 96.5
Piperacillin 55 53 96.4
Ceflazidine 56 54 96.4
Cefotaxime 56 54 96.4
Cefepim 56 54 96.4
Tica + A. clavulanic 54 51 94.4
Co - Trimoxazol 54 51 94.4
Levofloxacin 53 51 96.2
Gentamycine 47 41 87.2
Tobramycine 44 40 90.9
Amikacin 52 47 90.4
Meropenem 54 49 90.7
Imipenem 56 50 89.3
Ciprofloxacin 55 52 94.5
Doxycyclin 56 45 80.4
Table 3.34 shows the high rates of A. baumannii isolates resistant to
all tested antibiotics (range 80.4 – 96.6%).

3.2.3. Molecular characteristics of A. baumannii causing HAP
Table 3.35. DNA pattern similarity of A. baumannii clusters
Cluster
No. of isolates
(n = 46)
Similarity percentage

A 7 (15.2) 96.1
B 7 (15.2) 89.3
C 4 (8.7) 100
D 12 (26.1) 95.9
E 5 (10.9) 100
F 11 (23.9) 80.0
Table 3.35 shows PFGE analysis revealed that a total of 46 of 53
(86.7%) A. baumannii isolates belong to 6 major clusters (A, B, C, D, E
and F) with the high similarity of DNA pattern ranged from 80.0% to
96.1%. Of them, 23 isolates belong to cluster D and F, accounted for 50%
(see Figure 3.1).
15




Figure 3.1. Dendrogram of A. baumanniii isolates (n = 46)



Figure 3.5. Monthly case an the distribution of major clusters

Figure 3.5 shows that cluster D was the most predominant and had
been detected through the entire 8-month study period. Other clusters (A,
B, E, F) survived less than 3 months.

16



3.2.4. Antibiogram of A. baumannii
Table 3.36. Antibiotic resistant levels of A. baumannii isolates (n = 53)

Antibiotics
Breakpoint
for resistance
(mg/L)
Measureme
nt range
(%)
Resistance
Range
(mg/L)
Piperacillin ≥128 4 - 512 52 (98.1) 512 - >512
Piperacillin/taz
obactam
≥128/4 4 - 512 48 (90.6) 32 - 256
Ceftazidime ≥32 4 - 512 52 (98.1) 128 - >512
Imipenem ≥16 4 - 512 45 (84.9) <4 - 32
Meropenem ≥16 4 - 512 46 (86.8) <4 - 32
Ciprofloxacin ≥8 1 - 128 52 (98.1) 32 - >128
Amikacin ≥32 2 - 256 52 (98.1) 32 - >256
Colistin ≥4 0.016 - 256 1 (1.9) 0.125 - 4
Table 3.36 shows that 52 of 53 (98.1%) obtained A. baumannii
isolates were multiantibiotc resistance to piperacillin
, ceftazidime, and

ciprofloxacin. The percentages of A. baumannii isolates resistant to
imipenem (84.9%) and meropenem (86.8%). High-level resistance to
piperacillin (MIC ≥ 512), ceftazidime (MIC, 128 - >512), and
ciprofloxacin (MIC, 32 - >128) were exhibited in the obtained A.
baumannii isolates.

Table 3.43. The association between ceftazidime
resistance and A. baumannii

Clusters No. resistance isolates OR p
A (n = 7) 7/7 - -
B (n = 7) 7/7 - > 0.05
C (n = 4) 4/7 - > 0.05
D (n = 12) 12/12 - > 0.05
E (n = 5) 5/5 - > 0.05
F (n = 11) 11/11 - > 0.05
Table 3.43 shows that there was no difference in the ceftazidime
resistance rate regarding A. baumannii clusters (p > 0.05).


17



Table 3.44. The association between imipenem
resistance and A. baumannii

Clusters No. resistance isolates OR 95% CI p
A (n = 7) 6/7 - -
B (n = 7) 7/7 1.3 0.9 – 18.2 > 0.05

C (n = 4) 4/4 3.5 0 > 0.05
D (n = 12) 12/12 3.5 0 > 0.05
E (n = 5) 3/5 0.6 0 > 0.05
F (n = 11) 9/11 0.9 0.32 – 3.5 > 0.05
Table 3.44 shows that there was no difference in the imipenem
resistance rate regarding A. baumannii clusters (p > 0.05).

Table 3.45. The association between meropenem
resistance and A. baumannii

Clusters
No. resistance
isolates
OR 95% CI p
A (n = 7) 6 /7 - - -
B (n = 7) 7/7 1.3 0.9 – 18.2 > 0.05
C (n = 4) 4/4 3.5 0 > 0.05
D (n = 12) 12/12 3.5 0 > 0.05
E (n = 5) 4/5 3.5 0 > 0.05
F (n = 11) 9/11 0.9 0.6 – 12.9 > 0.05
Table 3.45 shows that there was no difference in the meropenem
resistance rate regarding A. baumannii clusters (p > 0.05).

Chapter 4. DISCUSSION

4.1. Incidence and risk factors for HAP
4.1.1. The incidence of HAP
HAP incidences have accurately described with the
HAP levels. Data
on HAP incidences usually used to compare the difference in HAP levels

between studies. Our study showed the high incidence of HAP (18.9%)
among 477 surveyed patients. This finding is similar to that reported in
some studies of similar design in the ICUs of Vietnam and other
18


developing countries with the overall incidences ranged from 15.0% to
27.0%.
We found that higher rates of HAP observed in patients with
APACHE II index of more than 13 (25.9%) and in those with chronic
admission diseases, which ranged from 28.1% to 60.2%. Particularly,
mechanical ventilation was the important risk factors for HAP. The HAP
rate in patients with mechanical ventilation (27.8%) was much more
higher than those without mechanical ventilation (2.9%). Our results
confirm other previous reports
showing the higher rate of HAP in ICU
patients who have a high risk of infection as a result of the predisposing
factors associated with their underlying conditions, chronic admission
diseases and of the risk associated with the respiratory invasive procedures
that they undergo. The HAP density reported in our study appears to be 2
time higher than that reported in developing countries.
The associations between the use of medical devices and
development of HAP in this study could be explained by current
challenges in IC practices at Bach Mai hospital. The lack of
regulations/guidelines for HAP control and prevention could be related
issue. Most mechanical ventilator devices were not appropriately treated.
The high work load situation in the ICU resulted in limited spaces between
patient beds, which provided favorable condition for cross-transmission in
patient care. Many students were not trained on IC or were not familiar
with guidelines for HAP. Basic IC practices and procedures such as hand

hygiene, glove use and environmental cleaning likely have not become
routine. The surveillance and reporting system for HAP was not
established which result in inappropriate patient isolations.
To address this failure, we believe it is necessary for Bach Mai
hospital to (1) Reduce the high work load situation in the ICU, (2) Provide
comprehensive education programs for healthcare workers, addressing
basic IC issues, (3) Improving the quality of aseptic techniques during
respiratory tract care , and (4) Implement the surveillance of HAP, HCW
compliance as well as timely feedback of surveillance results.

4.1.2. Risk factors for HAP
According to some previous studies, risk factors for HAP including
age 65 years or more, underlying conditions, especially chronic respiratory
diseases, depressed level of consciousness, thoracic or chest surgery. The
associations between invasive procedures and development of HAP were
19


confirmed in many studies conducted in Vietnam and other countries.
Andrew et all reported risk factors for HAP in patients with endotracheal
intubation and tracheotomy ranged from 3 times to 21 times higher than
those without these respiratory invasive procedures. In a report from the
National Nosocomial Infection Surveillance (NNIS)] system, involving
data from 498 998 patients, 83% of episodes of HAP were associated with
mechanical ventilation
. Study on 485 ICU patients at Cho Ray hospital
(2004) also showed that endotracheal intubation and tracheotomy
determined as risk factors for HAP. Our findings are similar to those
reported in published studies. The risk factors were identified by
multivariable logistic regression analyses: Chronic respiratory diseases

(OR=1.9, p<0.05), surgery (OR=2,5, p<0.05), endotracheal intubation
(OR=3,9, p<0,05), tracheotomy (OR=6.3, p<0.05).
These above results revealed the necessary for strict compliance with
IC practices during performing surgery, respiratory invasive procedures
and for removing as soon as possible respiratory devices to reduce HAP.
Our findings also highlighted the need for the simultaneous
implementation of IC interventions (hand hygiene, personal protective
equipment, compliance with aseptic technique) in which hand hygiene is
always considered as the most simple and effective measure in HAP
prevention.

4.2. Pathogens, antibiotic resistance and molecular characteristics of
A. baumannii causing HAP

4.2.1. HAP Pathogens
Recent decades have seen a swing in the pattern of infecting
organisms towards gram-negative infections such as P. aeruginosa,
Acinetobacter spp, E. coli and K. pneumoniae. Fungal pathogens,
especially Candida spp, are becoming increasingly common. In HCFs
where the use of excess antibiotic doses and poor infection prevention and
control practices have been associated with the development of many
strains of multiantibiotic resistant - GNB.
The findings are similar to our
results. A total of 132 pathogens were isolated. The 2 most frequently
isolated organisms were P.aeruginosa
(17.4%) and Acinetobacter
baumanii (44.7%). The next most commonly isolated agent was Candida
species (19,7%).

Several studies have suggested that gram-negative bacteria are most

likely responsible for infection due to their produce of expanded-spectrum
20


β- lactamases (ESBLs) as a result of long-term inappropriate or
insufficient antimicrobial use. Therefore, the identification of pathogens
and their antimicrobial susceptibility patterns is needed for the treatment of
severe infections and for the control of multidrug resistant bacteria.

Furthermore, some surveillances also revealed that GNB such as A.
baumannii and P. aeruginosa can strongly survive in the hospital
environments
, which is a potential for the spread of these bacteria during
hospital outbreaks. These surveillances could explain the recurrence of
hospital outbreaks caused by GNB due to the inappropriate
decontamination of hospital environmental surfaces. So strict compliance
with environmental cleaning procedures in HCFs is an important measure
to eradicate GNB source and/or reservoir.

HAP caused by fungi is becoming increasingly common and it is
considered as one of the challenging problems in HCFs. Fungi can easily
penetrate into body in some favorable circumstances such as patient with
invasive procedures, impairment of immune status and prolonged hospital
stay. Probable reasons for fungi contamination in hospital environments
could include high humidity, poor ventilation in patient rooms, infrequent
and/or inappropriate cleaning of hospital environmental surfaces and of
HCWs’hands. This study showed that Candida spp was the second most
pathogens (20.7%) caused HAP. So HCFs should concentrate on the
methods that could control and prevent fungi contamination in patient
care.

4.2.2. Antibiotic resistance of A. baumannii causing HAP
A. baumannii strains represent the common organism causing HAP
among ICU patients. Recently, they appear in HCFs as an important
pathogen due to accumulation trend of drug resistance mechanism leading
the development of multidrug resistant isolates which are responsible for
hospital outbreaks in various clinical departments. This study not only
confirmed the increased trend in HAP caused by A. baumannii but also
recognized the growth of antibiotic resistances of this pathogen.
52 of 53
(98.1%) obtained A. baumannii isolates were high-level resistance to
piperacillin
(MIC, 512 - >512), ceftazidime (MIC, 128 - >512), and
ciprofloxacin
(MIC, 32 - > 128). Carbapenem is one of most effective
antibiotics against A. baumannii. However, the percentages of A.
baumannii isolates resistant to imipenem (84.9%) and meropenem
(86.8%). These rates, were much higher than recent studies conducted in
21


US and Europe, in which. approximately one-fourth of the A. baumannii
isolates resistant to imipenem .

Inappropriate use of antibiotics is one of main reasons causing the
rapid global emergence of antibiotic resistant A. baumannii strains and can
lead to increased mortality in infected patients with these bacteria. HAIs
caused by A. baumannii, which is associated with a attributable mortality
rate of 25% in hospital wide and of 50% in ICUs only.
Although antibiotic resistance is rising, no new effective antibiotics

against multiantibiotic resistant A. baumannii expected to be in use in the
next several decades. It is necessary to consider the use of colistin in
infected patients with these bacteria. Moreover, the surveillance of HAIs
caused by these multidrug resistant bacteria should seem to be the first
priority in IC activities. Similar surveillance systems should also consider
presenting prevalence rates of multidrug resistant or panresistant
organisms, rather than merely giving rates of resistance to individual
antibiotics, so that the global impact of antibiotic resistance in A.
baumannii can be better understood.

4.2.3. Molecular Characteristics of A. baumannii Causing HAP
In our study, PFGE analysis revealed that a total of 46 of 53 (86.7%)
A. baumannii isolates belong to 6 major clusters (A, B, C, D, E and F) with
the high similarity of DNA pattern ranged from 80.0% to 96.1%. Of them,
23 isolates belong to cluster D and F, accounted for 50%. Cluster D was
the most predominant and had been detected through the entire 8-month
study period. Other clusters survived less than 3 months.
Epidemic investigations using PFGE technique are often used to help
determine the sources of organism in environments. However, in some
cases, increased infection rates of specific pathogens could be due to poor
compliance with IC procedures, which results in the cross-transmission of
A. baumannii between patients.

Our study showed that A. baumannii was the most frequently
pathogen, accounted for more than 44% of the isolated organism. Most
isolated A. baumannii belonging to cluster D and F with the DNA pattern
similarity index of 95.9% and 80.0%, respectively. These findings suggest
a hypothesis that is consistent with studies showing the cross-transmission
of A. baumannii between patients. Probable reasons
for this hypothesis

could include limited resources, inadequate infection control infrastructure
in the ICU, Bach Mai hospital. Most surveyed patients suffered from
severe conditions, chronic admission diseases, impairment of immune
22


status, prolonged antibiotic/immunosuppressive drug use and from risk
associated with the respiratory invasive procedures that they undergo.
However, isolation room for patients with multidrug resistant A.
baumannii was not available. Limited spaces ( < 1 meter) between patient
beds. The ICU always received a large numbers of students who involved
in direct patient care. Most these students had not been trained on Standard
and transmission based precautions. In addition, audits and surveillances
of compliance with aseptic procedures had not become the routine
activities in the ICU due to limited resources. In this situation, colonized or
infected patients with A. baumannii could be the sources for the cross-
transmission of A. baumannii between patients. Further research on the
transmission of these bacteria is needed to
implement effective
interventions, which contribute to reduce the HAP rates caused by A.
baumannii in ICUs.

4.3. Strategies for the prevention and control of HAP
The optimal management of patients with HAP requires close
collaboration among pulmonary and critical care specialists, infectious
disease practitioners, infection control professionals, radiologists, and
hospital microbiologists. This type of collaboration will lead to early
recognition and appropriate management of common source outbreaks and
MDROs.
Education of HCWs regarding HAP pneumonias and infection

prevention is very important. Because colonization with hospital bacteria
can be initiated with the transfer of bacteria from HCWs, simple yet
effective control measures of hand hygiene and appropriate use of barriers
(gloves and gown use) following standard precautions should be
emphasized. Immunizations (e.g., pneumococcus and influenza) for HCWs
and patients are recommended.
Respiratory equipment must be properly cleaned and sterilized.
ventilator tubing must be handled appropriately to avoid spilling
condensate into the patient’s airway. Ventilator circuits should not be
routinely replaced to decrease the risk of pneumonia. Rather, replace
circuits if malfunctioning or if visibly contaminated. Specific instructions
can be found in the CDC guidelines for preventing HAP, 2003. Measures
that can decrease the risk of aspiration include positioning the patient
supine with the head elevated 30–45 degrees, suctioning measures
including continuous suctioning of subglottic secretions, and minimizing
the use of sedating or paralytic agents.
23


HAP control program in HCFs is only effective when the multimodal
and multidisciplinary approach is implemented in combination with efforts
on methods that could improve the behavior of HCWs in patient care.
Institute health care improvement (IHI), USA recommend ICUs of HCFs
for implementing comprehensive program to significantly
reduced their
monthly HAI rates to “zero”. Feedback on the surveillances with IC
procedures at both individual and organizational levels, and involvement
of institutional leaders considered as important strategies to the success in
HAP reduction.

24


CONCLUSIONS

1. Incidence and risk factors for HAP
• The overall HAP incidence was 18.9%. HAP densities in patients
with tracheotomy, endotracheal intubation 72.1 and 39.1,
respectively. Numbers of HAP per 1.000 hospitalization days were
11.6.
• Risk factors for HAP found by logistic regression analysis including:
(1) Chronic respiratory diseases (aOR, 1.9; p < 0.001), (2)
Endotracheal intubation (aOR, 3.9; p < 0.05), (3) Tracheotomy (aOR,
6.3; p < 0.01), (4) Surgical procedure (aOR = 2.5, p < 0.05).
.
2. Pathogens, antibiotic resistance and molecular characteristics of A.
baumannii causing HAP
• The 3 most frequently isolated organisms were A. baumannii
(44.7%), Candida spp (19.7%) and P. aeruginosa (17,4%).
• 52/53 52 of 53 (98.1%) obtained A. baumannii isolates were
multiantibiotc resistance to piperacillin, ceftazidime, and
ciprofloxacin. The percentages of A. baumannii isolates resistant to
imipenem (84.9%) and meropenem (86.8%). High-level resistance to
piperacillin (MIC ≥ 512), ceftazidime (MIC, 128 - >512), and
ciprofloxacin (MIC, 32 - >128) were exhibited in the obtained A.
baumannii isolates.
• PFGE analysis revealed that a total of 46 of 53 (86.7%) A. baumannii
isolates belong to 6 major clusters (A, B, C, D, E and F) with the
high similarity of DNA pattern ranged from 80.0% to 96.1%. Of
them, 23 isolates belong to cluster D and F, accounted for 50%.

• Cluster D was the most predominant and had been detected through
the entire 8-month study period. Other clusters (A, B, C, E, F)
survived less than 3 months.


.