158 AJTCCM VOL. 29 NO. 4 2023
ORIGINAL RESEARCH: ARTICLES
Background. Ventilator-associated pneumonia (VAP) has an estimated incidence of 10 - 41.5 events per 1 000 ventilator days in developing
countries, and carries high mortality. Little is known about the incidence and outcomes of VAP in Johannesburg, South Africa.
Objectives. To describe VAP in a tertiary public hospital in Johannesburg, assess the microbiological pathogens associated with VAP (both
early and late), and outline the outcomes of these patients.
Methods. e study was a retrospective record review of patients admitted to the Helen Joseph Hospital intensive care unit (ICU) between
March 2013 and January 2016.
Results. VAP developed in 24/842 ventilated patients (2.9%; 95% condence interval (CI) 1.8 - 4.2), with an incidence of 23 events per 1000
ventilator days, during the study period. Of these patients, one-third (29.2%) died and 70.8% were discharged from the ICU. Late-onset VAP
(onset ≥5 days aer intubation, incidence 45.8%) was associated with higher mortality (54.6%) than early-onset VAP (onset within 4 days
aer intubation, incidence 54.2% and mortality 7.7%). Commonly isolated organisms were Klebsiella pneumoniae, Acinetobacter baumannii
and Pseudomonas aeruginosa. ere was a trend towards an increased risk of multidrug-resistant organisms with late-onset VAP (adjusted
relative risk 2.26; 95% CI 0.92 - 5.57; p=0.077) and airway access through a tracheostomy (relative risk 1.68; 95% CI 0.78 - 3.57).
Conclusion. e study showed a low to moderate incidence of VAP of 23 events per 1 000 ventilator days. A tracheostomy and late-onset
VAP were associated with infection by drug-resistant organisms. e mortality rate was 29.2% in this setting, with a seven-fold increase in
mortality with late-onset VAP.
Keywords. Ventilator-associated pneumonia, VAP, early onset, late onset, South Africa.
Afr J Thoracic Crit Care Med 2023;29(4):e154. https://doi.org/10.7196/AJTCCM.2023.v29i4.154
Ventilator-associated pneumonia (VAP) has an estimated incidence of
8 - 28% (10 - 41.5 per 1 000 ventilator days) in low- and middle-income
countries (LMICs).[1-3] VAP carries high mortality, with rates ranging
from 24% to 50%, and as high as 76% in LMICs.[4,5] Internationally,
it has been dicult to delineate the trueprevalence of VAP owing to
multiple challenges in the diagnosis of this condition, central to which
is the lack of a gold standard for diagnosis and therefore no standardised
diagnostic criteria,[3] although the Johanson criteria are the most
widely accepted. e Clinical Pulmonology Infection Score (CPIS), the
US Centers for Disease Control criteria and the HospitalsinEurope
Ventilator-associated pneumonia in an academic intensive care
unit in Johannesburg, South Africa
S Mazwi,1 MB ChB, Dip HIV Man (SA), FCP (SA), MMed (Int Med);
S A van Blydenstein,2 MB BCh, DCH (SA), FCP (SA), MMed (Int Med), Cert Pulmonology (SA);
M Mukansi,3 MB ChB, MMed (Int Med), FCP (SA), Cert Pulmonology (SA), FCCP, MBL
1 Division of Internal Medicine, Chris Hani Baragwanath Academic Hospital and Faculty of Health Sciences, University of the Witwatersrand, Johannesburg,
South Africa
2 Division of Pulmonology, Chris Hani Baragwanath Academic Hospital and Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, SouthAfrica
3 Division of Critical Care, Helen Joseph Hospital and Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa
Corresponding author: S Mazwi (mazwezwe.mazwi41@gmail.com)
Study synopsis
What the study adds. is study helps to improve understanding of the incidence of ventilator-associated pneumonia in South Africa,
alow- to middle-income country, and the commonly encountered causative pathogens. It indicates the importance of a short intensive care
unit (ICU) stay as a target outcome for prevention of nosocomial infections and other complications.
Implications of the ndings.
e study:
reinforces the importance of preventive mesures in the ICU and keeping up to date with the evidence in the eld
highlights the importance of knowing local microbial resistance patterns in order to develop precise antibiograms
shows the need for research in ICU care for people of advanced age, and the impact that admission rationing has on our ICU populations.
AJTCCM VOL. 29 NO. 4 2023 159
ORIGINAL RESEARCH: ARTICLES
Link for Infection Control through Surveillance criteriahave been
used in clinical, research and public health environments.[1,4,6] Features
of VAP overlap with other common conditions in the intensive care
unit (ICU) environment such as acute respiratory distress syndrome
(ARDS), atelectasis, other ICU infections, and ventilator-associated
tracheobronchitis (VAT). VAT has the potential to develop into
VAP in 30% of cases.[3] Patient characteristics such as advanced
age (>65 years), male sex and smoking, increased mechanical
ventilation time and prolonged mechanical ventilation, disorders of
consciousness and head trauma, burns, comorbidities (coronary heart
disease, diabetes, pre-existing pulmonary disease/chronic obstructive
pulmonary disease, HIVinfection, and multiple organ system failure),
prior antibiotic therapy, invasive operations and gene polymorphisms
are currently the internationally recognised independent risk factors
associated with VAP.[7,8] Non-modiable treatment-related risk factors
include the necessity for neurosurgery, monitoring of intracranial
pressure, reintubation, and transportation out of the ICU.[8]
e well-established risk factors for multidrug-resistant (MDR)
organisms are prior intravenous antibiotic use (within 90 days), septic
shock, ARDS preceding VAP (a high index of suspicion is required to
make a diagnosis in this situation), ≥5 days of hospitalisation prior
to VAP, and acute dialysis prior to VAP.[1,5,9,10] e pathogenesis of
pneumonia stems from invasion of the lower respiratory tract and
lung parenchyma by micro-organisms.[3,4]
The two most important contributors to VAP are biofilm
establishment within the tube lumen (endotracheal tube/
tracheostomy), and microaspiration of secretions, particularly
subglottic and above the tube cu.
e presence of the tube in the trachea alters the natural defence
mechanisms, thus allowing microaspiration and for the aspirated
particles to pass into the lower respiratory tract. e tube biolm
is pushed further down the respiratory tract by the ventilator
cycles and serves as a nidus for infection. Host factors, particularly
immunosuppression (which could be multifactorial with critical
illness), play a major role.[6]
e common micro-organisms isolated in VAP are Gram-negative
bacilli, accounting for 60% of cases in studies in the developed world
and 41 - 92% in the developing world.[2,5,6,11] Of these, Pseudomonas
aeruginosa is the leading organism, followed by Acinetobacter species,
Proteus mirabilis, Escherichia coli, Klebsiella species and Haemophilus
inuenzae.[11]
Gram-positive cocci account for a substantial proportion of
cases, with Staphylococcus aureus accounting for 20% of cases in the
developed world. Gram-positive cocci in total make up 6 - 58% of
cases in LMICs. VAP occurring ≥5 days aer intubation is most likely
to be associated with resistant organisms, i.e. carbapenem-resistant
Enterobacteriaceae, vancomycin-resistant Enterococcus, methicillin-
resistant S. aureus (MRSA), and Pseudomonas and Acinetobacter
species, and with prior exposure to antibiotics.[9,11-14]
Prevention of VAP is based on trying to mitigate the modiable
risk factors and intervene in the main pathogenic factors mentioned
above. Historically, care bundles had four care interventions, namely
daily sedation holds, bed head elevation, gastric ulcer prophylaxis
and oral care. e version of care bundles commonly employed was
updated in 2010 and includes oral hygiene with adequate antiseptic,
subglottic aspiration, and endotracheal tube cu monitoring.[4,15-17]
Supplementary to the bundles of care are appropriate humidication
of inspired air, deep-vein thrombosis prophylaxis, suctioning of
secretions, and appropriate tubing management (such as avoiding
unnecessary circuit tubing changes).[17,18] ese bundles have shown
some eectiveness in VAP prevention. Several studies have shown that
nursing care education is key to VAP prevention.[17] e 2016 clinical
practice guidelines from the Infectious Diseases Society of America
and the American oracic Society recommend that each ICU should
develop its own antibiogram to refer to, based on local antimicrobial
resistance patterns.[4]
A paucity of data pertaining to the prevalence of VAP in South
Africa (SA), including information on the most common organisms
causing this disease, hampers guideline development, especially for
the adult population.[19] e aim of the present study was to describe
VAP in a tertiary public hospital ICU in Johannesburg. Specically,
we sought to: (i) determine the incidence of ventilator-associated
pneumonia in the ICU; (ii) assess the microbiological pathogens
associated with both early-onset VAP (dened as occurring within
4days after intubation) and late-onset VAP; (iii) determine the
outcome of patients diagnosed with VAP in the ICU; (iv) estimate
the time from admission (intubation and ventilation) to development
of VAP; and (v) determine factors associated with MDR organisms.
Methods
Design and study population
A retrospective record review was performed of patients admitted
to the Helen Joseph Hospital ICU in Johannesburg between March
2013 and January 2016, a period of 1 006 days. Data were collected
following approval by the University of the Witwatersrand Human
Research Ethics Committee (ref. no. M190124). To be eligible for
inclusion in the study, patients had to be ≥18 years old. Patients who
developed pneumonia within 48 hours of admission, patients admitted
with a diagnosis of pneumonia, patients with ARDS, and mechanically
ventilated patients who died within 48 hours of admission were
excluded.
Setting
e 10-bed multidisciplinary ICU comprises 10 isolation cubicles and
is separated into two 5-bed sections. e nurse-to-patient ratio is 1:1
and the doctor-to-patient ratio 1:2. e unit has two washbasins at the
entrance of each section, with handwashing detergent at each bedside
trolley and at the unit entrance. e infection control team supervises
the handwashing.
Data collection
All patients admitted to the ICU were recorded in a patient register with
their corresponding diagnoses, both provisional and denitive. Patients
who were diagnosed with VAP by the end of their ICU stay, as entered
on the Helen Joseph Hospital ICU patient database, were identied and
their records were retrieved from the hospital records. ese records
were further ltered to those that tted the denitions below.
VAP was dened as pneumonia occurring in mechanically ventilated
patients ≥48 hours after initiation of intubation and ventilation.
Multidrug resistance was dened as resistance to at least one agent in
three or more antimicrobial classes, as per National Health Laboratory
Service microbiological testing.
160 AJTCCM VOL. 29 NO. 4 2023
ORIGINAL RESEARCH: ARTICLES
Data collected included demographics, clinical data, measures taken
to prevent hospital-acquired infections, investigations performed
(blood, radiology, tracheal aspirate), disease severity at the onset of
VAP, and the management approach employed.
Data from medical records were captured in Excel 2013 (Microso,
USA) and exported into Stata 14.2 (StataCorp, USA) for analysis.
Descriptive statistics were used to describe individuals admitted to
the ICU who developed VAP. Medians and interquartile ranges (IQRs)
were used for continuous variables, while frequencies and percentages
were used for categorical data. Methods of analysis for the specic
objectives are described below.
1. e incidence of VAP in the ICU was determined as number
of patients who developed VAP divided by the total number
of ICU admission days. is was presented as the number of
patients who developed VAP per 1000 ICU admission days,
with a 95% condence interval (CI). e number who developed
VAP was obtained from the database of VAP patients, while
aggregatedICUdata were used to obtain the number admitted
to the ICU.
2. To assess the microbiological pathogens associated with VAP, both
early and late, the proportions of VAP patients who had dierent
organisms detected on tracheal aspirates and blood specimens
were determined as frequencies and proportions and presented by
early or late VAP status.
3. The outcome of patients diagnosed with VAP in the ICU was
determined as proportions of all patients with VAP, and also by
early or late VAP status.
4. The time from admission (intubation and ventilation) to
development of VAP was determined as the number of days
between admission and VAP diagnosis, and presented as medians
with IQRs.
5. Factors associated with the detection of MDR organisms from
tracheal aspirates or blood specimens were determined by
univariable and multivariable binomial regression analysis.
Variables that had a p-value <0.2 in univariable analysis were
included in the multivariable analysis. e estimate of the relative
risks (RRs) associated with the dierent factors was reported with
a 95% CI.
Results
Incidence of VAP
During the study period, a total of 1185 patients were admitted to the
ICU. Of the admitted patients, 842 (71.1%) were ventilated, and of
those who were ventilated, 24 (2.9%; 95% CI 1.8 - 4.2) had a diagnosis
of VAP by the end of their ICU stay. ese cases of VAP occurred over
a period of 1 006 admission days, which equates to 23 events per 1 000
days, assuming that there was at least one ventilated patient in the ICU
every day during this period.
Description of patients who developed VAP and
frequently occurring factors in patients who
developed VAP
Of the 24 patients who developed VAP, 18 (75.0%) were male
(Table1). e median (IQR) age was 40 (29 - 62) years. e majority
of the patients who developed VAP were admitted from casualty and
medical wards, with diagnoses of trauma or medical conditions on
admission to the ICU. Table2 compares surgical v. medical conditions
on admission. The median (IQR) length of stay in the ICU was
12(9-15) days, with a median time on antibiotics of 7 (7 - 8) days.
Acute Physiology and Chronic Health Evaluation (APACHE II) scores
were available for 11/24 (45.6%) of the patients, of whom 5 (45.5%)
had scores ≥24 and 6 (54.5%) scores <24.
The median (IQR) time from intubation and ventilation to
VAP diagnosis was 3 (2 - 5.5) days. An analysis of factors revealed
that advanced age was not a signicant risk factor in our cohort,
while male gender, prolonged ventilation time and the presence of
comorbidities, including being HIV positive, were common risk
factors.
Table1. Demographic and clinical characteristics of patients
who developed VAP during their ICU admission (N=24)
Characteristic n (%)*
Age (years), median (IQR) 40 (29 - 62)
Male 18 (75.0)
One or more comorbidities 7 (29.2)
Admission source
Emergency/casualty 12 (50.0)
Medical inpatients 8 (33.3)
Surgical inpatients 2 (8.3)
Other 2 (8.3)
Diagnoses at ICU admission
Medical 11 (45.8)
Trauma 9 (37.5)
Surgical 3 (12.5)
Missing 1 (4.2)
Airway access at admission
Endotracheal tube 20 (83.3)
Tracheostomy 4 (16.7)
Indication for ventilation
Respiratory failure 8 (33.3)
Airway protection 13 (54.2)
Anaesthesia 3 (12.5)
Head of bed elevated 24 (100)
Adequate sedation provided 23 (95.8)
Oral care provided 24 (100)
Antacid provided 24 (100)
HIV status
Positive 7 (29.2)
Negative 9 (37.5)
Unknown 8 (33.3)
White blood cell count (× 109 cells/L),
median(IQR) 12.8 (8.8 - 17.8)
Length of stay in ICU (days), median (IQR) 12 (9 - 15)
Time on antibiotics (days), median (IQR) 7 (7 - 8)
APACHE II score
<24 6 (25.0)
≥24 5 (20.8)
Missing/unavailable 13 (54.2)
VAP = ventilator-associated pneumonia; ICU = intensive care unit; IQR = interquartile range;
APACHE II = Acute Physiology and Chronic Health Evaluation.
*Except where otherwise indicated.
AJTCCM VOL. 29 NO. 4 2023 161
ORIGINAL RESEARCH: ARTICLES
Microbiological pathogens associated with early and
late VAP
e most frequently isolated pathogens were the Gram-negative
bacilli (P. aeruginosa, Klebsiella pneumoniae and Acinetobacter
baumannii), comprising 56.6% of all isolates, with both blood
cultures and tracheal aspirates being positive for similar organisms in
many patients. e other Gram-negative bacilli isolated were E. coli,
Stenotrophomonas maltophilia, Citrobacter koseri and P.mirabilis,
which made up 13.3% of all positive tracheal aspirates, bringing the
total percentage of Gram-negative bacilli isolated to 69.9%. Table3
shows the distribution of micro-organisms causing early-onset VAP
v. late-onset VAP.
In univariate analysis, there was a trend towards an increased
risk of MDR organisms with airway access through a tracheostomy
(RR .68; 95% CI 0.78 - 3.57), being admitted to the ICU during the
period 2013/2014 (RR 1.96; 95% CI 0.86 - 4.49), and late-onset VAP
(RR2.36; 95% 0.95 - 5.88). In the adjusted analysis, the trend towards
an increased risk of MDR organisms remained with late-onset VAP
(adjusted RR (aRR) 2.26; 95% CI 0.92 - 5.57; p=0.077) and admission
during the period 2013/2014 (aRR 1.89; 95% CI 0.81 - 4.42; p=0.140).
Airway access via a tracheostomy is still considered an independent
risk factor despite the multivariate analysis outcomes shown in Table4.
Outcomes of patients with VAP
Of the patients who had VAP, 29.2% died, while the remainder were
discharged from the ICU. Late-onset VAP was associated with much
higher mortality (54.6%) compared with early-onset VAP (7.7%)
(p=0.018, χ² test).
Table2. Medical v. surgical conditions associated with VAP in the ICU
Diagnosis Medical Trauma or surgery Total
Acute exacerbation of COPD 1 1
Acute weakness (GBS/myasthenia) 2 2
Bowel obstruction (sigmoid stricture) 1 1
Head injury 1 1
Heart failure 1 1
Hyperglycaemic hyperosmolar state 1 1
Iatrogenic pulmonary oedema 1 1
Meningitis 1 1
Motor vehicle accident 1 1
Myocardial infarction 1 1
Organophosphate poisoning 3 3
Perforated peptic ulcer 1 1
Polytrauma 2 2
Postoperative 1 1
Stab chest 1 1
Traumatic brain injury 4 4
Not recorded 1 1
VAP = ventilator-associated pneumonia; ICU = intensive care unit; COPD = chronic obstructive pulmonary disease; GBS = Guillain-Barré syndrome.
Table3. Micro-organisms associated with early-onset v. late-onset VAP
Organism All patients (N=24), n (%) Late VAP (n=11), n (%) Early VAP (n=13), n (%)
Acinetobacter baumannii 6 (20.0) 4 (23.5) 2 (15.4)
Candida albicans 1 (3.3) 0 1 (7.7)
Citrobacter koseri 1 (3.3) 1 (5.9) 0
Escherichia coli 1 (3.3) 0 1 (7.7)
Enterobacter 1 (3.3) 1 (5.9) 0
Haemophilus inuenzae 3 (10.0) 0 3 (23.1)
Klebsiella pneumoniae 6 (20.0) 5 (29.4) 1 (7.7)
Mycobacterium tuberculosis 1 (3.3) 1 (5.9) 0
Proteus mirabilis 1 (3.3) 0 1 (7.7)
Pseudomonas aeruginosa 5 (16.7) 4 (23.5) 1 (7.7)
Staphylococcus aureus 3 (10.0) 1 (5.9) 2 (15.4)
Stenotrophomonas maltophilia 1 (3.3) 0 1 (7.7)
Total micro-organisms isolated 30 17 13
VAP = ventilator-associated pneumonia.
162 AJTCCM VOL. 29 NO. 4 2023
ORIGINAL RESEARCH: ARTICLES
Discussion
In this review of VAP in the public sector, there were 1 185 admitted
patients (1.17 admissions per day), and 71.1% of these were ventilated
(0.84 ventilated admissions per day). e ICU has <14 beds, and
would generally be classied as a small ICU.
e overall average length of stay in this unit during the study
period was 11 days for patients with VAP, which is similar to the
13-day average observed in other parts of the world.[20] e ICU in the
study has a rapid turnover as a result of high demand, with occupancy
at any given time ~90%.
e absolute percentage of patients who were diagnosed with VAP
in the unit was 2.9% of all ventilated patients. e incidence was
23events per 1 000 ventilator days, assuming that there was at least
one ventilated patient in the unit every day. e estimated incidence
of VAP in the developing world varies widely from 10 to 41.5 episodes
per 1 000 ventilator days,[2] so our gure falls on the low end of the
average, although even lower numbers have been reported in ICUs
with a heterogeneous population like ours.[21] Our relatively low
incidence could be attributed to a few possible factors. ere is strict
adherence in the unit to care bundles to reduce complications such
as VAP. e diagnosis of VAP at the time of the study was largely
based on the CPIS, which has a high inter-observer variability in its
calculation, limiting its diagnostic impact.[6,19,22] Subtle cases may
therefore have been missed.
e possibility of underdiagnosing VAP is also a consideration,
especially as postmortem studies evaluating the eectiveness of
clinical criteria for VAP diagnosis show that these criteria have a
69% sensitivity.[22] Studies to determine whether care bundles do
indeed prevent VAP are equivocal for the most part.[16] e main
factor resulting in the success of care bundles is dedication of the
nursing teams, as these are the professionals who ensure continuous
adherence to the interventions.[17] As shown in Table1, the minimum
requirements of head elevation, adequate sedation, antacids and
oral care, which was the recommended bundle at the time of the
study, were adhered to at least 95.8% of the time. Inhaled antibiotics
are not used in the ICU studied, so we cannot comment on the
eectivenessof this strategy in the prevention and management
ofVAP.
Male sex (75.0% of patients) was a dominant characteristic. We
identified comorbidities in 29.2% of the cohort, excluding HIV
infection. ese comorbidities diered so widely that none of them can
be specically identied as associated with the development of VAP.
Table4. Factors associated with the identication of an MDR organism (N=24)
Variabl e
n/N (%) with MDR
organism
Univariable
RR (95% CI) p-value
Multivariable
aRR (95% CI) p-value
Age ≥50 years 0.429
No 4/10 (40.0) 1.00 - -
Yes 8/14 (57.1) 1.42 (0.59 - 3.52) - -
Sex 0.999
Female 3/6 (50.0) 1.00 - -
Male 9/18 (50.0) 1.00 (0.40 - 2.57) - -
Has comorbidity 0.668
No 9/17 (52.9) 1.00 - -
Yes 3/7 (42.9) 0.81 (0.31 - 2.17) - -
HIV positive 0.151*
No 6/9 (66.6) 0.93 - -
Yes 5/7 (71.4) 1.07 (0.82 - 3.68) - -
Unknown 3/8 (37.5) - -
Admission period 0.112* 0.140
2015/16 5/14 (35.7) 1.00 1.00
2013/14 7/10 (70.0) 1.96 (0.86 - 4.49) 1.89 (0.81 - 4.42)
Airway access 0.188* 0.456
ETT 9/20 (45.0) 1.00 1.00
Tracheostomy 3/4 (75.0) 1.68 (0.78 - 3.57) 0.75 (0.36 - 1.58)
Ventilation indication 0.692
Other 5/11 (45.5) 1.00 - -
Airway protection 7/13 (53.9) 1.18 (0.52 - 2.74) - -
Diagnosis at ICU admission 0.686
Other 7/13 (53.9) 1.00 - -
Medical 5/11 (45.5) 0.84 (0.37 - 1.95) - -
Onset of VAP 0.064* 0.077
Early 4/13 (30.8) 1.00 1.00
Late 8/11 (72.7) 2.36 (0.95 - 5.88) 2.26 (0.92 - 5.57)
MDR = multidrug resistant; RR = relative risk; CI = condence interval; aRR = adjusted relative risk; ETT = endotracheal tube; ICU = intensive care unit; VAP = ventilator-associated pneumonia.
*Signicant (p<0.2).
AJTCCM VOL. 29 NO. 4 2023 163
ORIGINAL RESEARCH: ARTICLES
Immunosuppression has been described as an independent risk factor,
and 29.2% of the patients were HIV positive.[8] e average length of
ICU stay in this patient group was 11 days, with up to 6 ventilator days
before VAP diagnosis, each day of ventilation increasing the possibility
of developing VAP.[8]
Advanced age did not play a role in our cohort, as the median
age was 40 years. A signicant contributor to this nding may be
that 37.5% of the patients were trauma patients, who tend to be
relatively young. A previous SA study found no association between
age and the development of VAP, with a mean age of 55 years.[19]
Inthe context of an LMIC with a limited number of critical care
beds, the concern whether ICU admission rationing plays a role in
the number of patients of advanced age in our ICU could be an area
of further research.[23]
e present study shows that longer duration of admission was
associated with late VAP and therefore with infections with MDR
organisms (Fig.1). e average time to late VAP diagnosis was 6 days.
e prevalence of early and late VAP was 54.2% and 45.8%, respectively.
However, it must be kept in mind that the leading indication for a
tracheostomy in ICU patients is prolonged intubation, for a period
of at least 10 days. ese patients had therefore been admitted for
longer than 10 days more than the time it took to develop MDR VAP
in association with a tracheostomy.
Gram-negative bacilli were isolated more frequently in the late-
onset VAP group than in the early-onset group. Over three-quarters
(83.3%) of the VAP patients had an organism isolated on respiratory
specimens (tracheal aspirates), while 45.8% were blood culture
positive. is nding is in keeping with observations throughout the
world, including previous ndings in the SA context, with Gram-
negative bacilli making up ~60% of all isolates in VAP patients.[5,6,11,19]
The most common organism isolated in early-onset VAP was
H.inuenzae (23.0%), consistent with ndings around the world.[24]
In the present study, H. inuenzae was isolated in early-onset VAP
only. is nding is in accordance with the knowledge that prior
antibiotic exposure is not a risk factor for respiratory infection with
H. inuenzae. e majority of patients with early-onset VAP would be
expected to have had no prior antibiotic exposure. Rello etal.[24] noted
in 1992 that patients with H. inuenzae infection were co-infected
with Gram-positive cocci, which was conrmed in the present study,
where two-thirds of the patients with H. inuenzae infection were
co-infected with S. aureus.
S. aureus also made up 10.0% of all organisms isolated, two-thirds
of which were from tracheal aspirates. A third of these isolates were
MRSA. In LMICs, S. aureus accounts for between 6% and 58% of
VAP cases, with high-income countries (HICs) recording an average
of 20%. e ndings of the present study fall within the expected
range for HICs.[2,11] e isolates from early-onset VAP samples were
methicillin sensitive, attesting to the fact that such patients would have
had limited healthcare and antibiotic exposure prior to the diagnosis
of VAP. MRSA was associated with late-onset VAP.
The least common infections were Mycobacterium tuberculosis
and fungal, comprising 1 case (3.3%) each. The mycobacterial
infection occured in the setting of HIV. It is important to note that
M. tuberculosis is not a typical causative organism for VAP, and it may
represent the reactivation of latent TB as a result of immune paresis
from acute illness. Candida albicans was the single fungal organism
isolated, and this low prevalence is comparable to ndings elsewhere
in the world, with gures of 0.9% in the USA and up to 7% in LMICs
being reported.[11]
In the present cohort, many patients with MDR organisms had
number of days in the hospital prior to VAP and possible previous
intravenous antibiotics as risk factors, as these patients were in the
late VAP group. As shown in Fig.1, the average number of days for
the diagnosis of late VAP was 6 days. In addition to the above, airway
access via a tracheostomy was identied as another possible risk
factor for MDR organisms. ese ndings have been discussed above.
ere has been conicting evidence as to whether tracheostomies are
protective against or predispose to VAP.[23,25]
VAP carries high all-cause mortality, with rates ranging from 24%
to 50%, and the present study is within this range at 29.2%, not very
different from a previous SA report of 37.5% mortality.[1,4,19] The
outcome of patients who were discharged from the ICU once they
were transferred to the general ward is beyond the scope of this study.
Late-onset VAP is associated with seven times higher mortality than
early VAP. is is probably due to two issues, one being that a long
duration of stay in the ICU of 6 days (average days to late VAP) is
an indirect indication of severe illness. Secondly, as shown above,
more resistant organisms were isolated in the late VAP group, making
managing these patients even more dicult and highlighting the
importance of dierentiating early from late VAP.
is study was a retrospective review, and as such accurate record
keeping is a concern, as demonstrated by a large number of data points
missing, particularly with regard to disease severity. e absence of a
control arm for the assessment of risk factor prevalence meant that
only the prominent characteristics in this cohort known to be risk
factors could be highlighted.
Conclusion
is study showed a low to moderate prevalence of VAP of 23 events
per 1 000 ventilator days. Late-onset VAP and airway access via
tracheostomy were associated with infection with drug-resistant
organisms. e study also showed a VAP mortality of 29.2%, with
Late-onset VAP Early-onset VAP
20
18
16
14
12
10
8
6
4
2
0
Fig.1. Time from intubation to diagnosis in early- and late-onset VAP.
e line in the shaded box for late-onset VAP and the line between the
dots for early-onset VAP depict the medians (6 and 2 days).
164 AJTCCM VOL. 29 NO. 4 2023
ORIGINAL RESEARCH: ARTICLES
the majority of deaths associated with late VAP, highlighting the
importance of implementing preventive measures. A causative
organism was isolated on culture in most cases, from tracheal aspirates
or blood or from both, with Gram-negative bacilli being the most
common. Clinicians are encouraged to develop antibiograms for their
ICU units to guide empirical treatment.
Declaration. e research for this study was done in partial fullment of
the requirements for SM’s MMed (Int Med) degree at the University of
the Witwatersrand.
Acknowledgements. e authors thank Ms Tendesai Kufa for assistance
with statistical analysis.
Author contributions. SM: writing up of the manuscript, data collection,
statistical analysis, review of the manuscript. SAvB: writing up of the
manuscript, review of the manuscript. MM: writing up of the manuscript,
review of the manuscript.
Funding.None.
Conicts of interest.None.
1. Chastre J, Fagon J-Y. Ventilator-associated pneumonia. Am J Respir Crit Care Med
2002;165(7):867-903. https://doi.org/10.1164/ajrccm.165.7.2105078
2. Arabi Y, Al-Shirawi N, Memish Z, Anzueto A. Ventilator-associated pneumonia in
adults in developing countries: A systematic review. Int J Infect Dis 2008;12(5):505-
512. https://doi.org/10.1016/j.ijid.2008.02.010
3. Wagh H, Acharya D. Ventilator associated pneumonia – an overview. Br J Med Pract
2009;2(2):16-19.
4. Kalil AC, Metersky ML, Klompas M, etal. Management of adults with hospital-
acquired and ventilator-associated pneumonia: 2016 clinical practice guidelines by
the Infectious Diseases Society of America and the American oracic Society. Clin
Infect Dis 2016;63(5):e61-e111. https://doi.org/10.1093/cid/ciw353
5. But A, Yetkin MA, Kanyilmaz D, etal. Analysis of epidemiology and risk factors for
mortality in ventilator-associated pneumonia attacks in intensive care unit patients.
Turk J Med Sci 2017;47(3):812-816. https://doi.org/10.3906/sag-1601-38
6. Gunasekera P, Gratrix A. Ventilator-associated pneumonia. BJA Educ 2016;16(6):198-
202. https://doi.org/10.1093/bjaed/mkv046
7. Wu D, Wu C, Zhang S, Zhong Y. Risk factors of ventilator-associated pneumonia
in critically ill patients. Front Pharmacol 2019;10:482. https://doi.org/10.3389/
fphar.2019.00482
8. Bonten MJ, Kollef MH, Hall JB. Risk factors for ventilator-associated pneumonia:
From epidemiology to patient management. Clin Infect Dis 2004;38(8):1141-1149.
https://doi.org/10.1086/383039
9. Rello J, Borgatta B, Lisboa T. Risk factors for Pseudomonas aeruginosa pneumonia in
the early twenty-rst century. Intensive Care Med 2013;39(12):2204-2206. https://doi.
org/10.1007/s00134-013-3046-1
10. Singer M, Deutschman CS, Seymour CW, etal. e ird International Consensus
Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA 2016;315(8):801-810.
https://doi.org/10.1001/jama.2016.0287
11. Cilloniz C, Martin-Loeches I, Garcia-Vidal C, San Jose A, Torres A. Microbial
etiology of pneumonia: Epidemiology, diagnosis and resistance patterns. Int J Mol Sci
2016;17(12):2120. https://doi.org/10.3390/ijms17122120
12. Mallick UK, Faruq MO, Ahsan AA, etal. Spectrum of early onset and late onset
ventilator associated pneumonia (VAP) in a tertiary care hospital of Bangladesh:
A prospective cohort study. Bangladesh Crit Care J 2015;3(1):9-13. https://doi.
org/10.3329/bccj.v3i1.24095
13. Restrepo MI, Peterson J, Fernandez JF, Qin Z, Fisher AC, Nicholson SC. Comparison
of the bacterial etiology of early-onset and late-onset ventilator-associated pneumonia
in subjects enrolled in 2 large clinical studies. Respir Care 2013;58(7):1220-1225.
https://doi.org/10.4187/respcare.02173
14. Golia S, Sangeetha KT, Vasudha CL. Microbial profile of early and late onset
ventilator associated pneumonia in the intensive care unit of a tertiary care hospital
in Bangalore, India. J Clin Diagn Res 2013;7(11):2462-2466. https://doi.org/10.7860/
JCDR/2013/6344.3580
15. Celis R, Torres A, Gatell JM, Almela M, Rodríguez-Roisin R, Agustí-Vidal A.
Nosocomial pneumonia: A multivariate analysis of risk and prognosis. Chest
1988;93(2):318-324. https://doi.org/10.1378/chest.93.2.318
16. Camporota L, Brett S. Care bundles: Implementing evidence or common sense? Crit
Care 2011;15(3):159. https://doi.org/10.1186/cc10232
17. Rello J, Afonso E, Lisboa T, etal. A care bundle approach for prevention of ventilator-
associated pneumonia. Clin Microbiol Infect 2013;19(4):363-369. https://doi.
org/10.1111/j.1469-0691.2012.03808.x
18. Torres A, Niederman MS, Chastre J, etal. International ERS/ESICM/ESCMID/ALAT
guidelines for the management of hospital-acquired pneumonia and ventilator-
associated pneumonia: Guidelines for the management of hospital-acquired
pneumonia (HAP)/ventilator-associated pneumonia (VAP) of the European
Respiratory Society (ERS), European Society of Intensive Care Medicine (ESICM),
European Society of Clinical Microbiology and Infectious Diseases (ESCMID) and
Asociación Latinoamericana del Tórax (ALAT). Eur Respir J 2017;50(3):1700582.
https://doi.org/10.1183/13993003.00582-2017
19. Behari AA, Kalafatis N. Incidence and outcome of ventilator-associated pneumonia
in Inkosi Albert Luthuli and King Edward VIII Hospital surgical intensive care units.
South Afr J Crit Care 2015;31(1):16-18. https://doi.org/10.7196/SAJCC227
20. Mathai AS, Phillips A, Isaac R. Ventilator-associated pneumonia: A persistent
healthcare problem in Indian intensive care units! Lung India 2016;33(5):512-516.
https://doi.org/10.4103/0970-2113.188971
21. Ali HS, Khan FY, George S, Shaikh N, Al-Ajmi J. Epidemiology and outcome of
ventilator-associated pneumonia in a heterogeneous ICU population in Qatar. Biomed
Res Int 2016;2016:8231787. https://doi.org/10.1155/2016/8231787
22. Kalanuria AA, Zai W, Mirski M. Ventilator-associated pneumonia in the ICU. Crit
Care 2014;18(2):208. https://doi.org/10.1186/cc13775
23. Othman HA, Gamil NM, Elgazzar AE, Fouad TA. Ventilator associated pneumonia,
incidence and risk factors in emergency intensive care unit Zagazig university
hospitals. Egypt J Chest Dis Tuberc 2017;66(4):703-708. https://doi.org/10.1016/j.
ejcdt.2017.08.004
24. Rello J, Ricart M, Ausina V, Net A, Prats G. Pneumonia due to Haemophilus inuenzae
among mechanically ventilated patients: Incidence, outcome, and risk factors. Chest
1992;102(5):1562-1565. https://doi.org/10.1378/chest.102.5.1562
25. Nseir S, Pompeo CD, Jozefowicz E, etal. Relationship between tracheotomy and
ventilator-associated pneumonia: A case-control study. Eur Respir J 2007;30(2):314-
320. https://doi.org/10.1183/09031936.06.00024906
Submited 28 April 2021. Accepted 10 September 2023. Published 27 November 2023.