CHEST

Original Research
CRITICAL CARE

The Clinical Impact and Preventability

of Ventilator-Associated Conditions
in Critically Ill Patients Who Are
Mechanically Ventilated
John Muscedere, MD; Tasnim Sinuff, MD, PhD; Daren K. Heyland, MD;
Peter M. Dodek, MD, MHSc; Sean P. Keenan, MD; Gordon Wood, MD;
Xuran Jiang, MSc; Andrew G. Day, MSc; Denny Laporta, MD;
and Michael Klompas, MD, MPH; for the Canadian Critical Care Trials Group

Background: Ventilator-associated conditions (VACs) and infection-related ventilator-associated

complications (iVACs) are the Centers for Disease Control and Preventions new surveillance paradigms for patients who are mechanically ventilated. Little is known regarding the clinical impact
and preventability of VACs and iVACs and their relationship to ventilator-associated pneumonia
(VAP). We evaluated these using data from a large, multicenter, quality-improvement initiative.
Methods: We retrospectively applied definitions for VAC and iVAC to data from a prospective time
series study in which VAP clinical practice guidelines were implemented in 11 North American
ICUs. Each ICU enrolled 30 consecutive patients mechanically ventilated . 48 h during each of
four study periods. Data on clinical outcomes and concordance with prevention recommendations
were collected. VAC, iVAC, and VAP rates over time, the agreement (k statistic) between definitions, associated morbidity/mortality, and independent risk factors for each were determined.
Results: Of 1,320 patients, 139 (10.5%) developed a VAC, 65 (4.9%) developed an iVAC, and
148 (11.2%) developed VAP. The agreement between VAP and VAC was 0.18, and between VAP
and iVAC it was 0.19. Patients who developed a VAC or iVAC had significantly more ventilator days,
hospital days, and antibiotic days and higher hospital mortality than patients who had neither of
these conditions. Increased concordance with VAP prevention guidelines during the study was
associated with decreased VAP and VAC rates but no change in iVAC rates.
Conclusions: VACs and iVACs are associated with significant morbidity and mortality. Although
the agreement between VAC, iVAC, and VAP is poor, a higher adoption of measures to prevent
VAP was associated with lower VAP and VAC rates.
CHEST 2013; 144(5):14531460
Abbreviations: APACHE 5 Acute Physiologic and Chronic Health Evaluation; iVAC 5 infection-related ventilatorassociated complication; LOS 5 length of stay; SAT 5 spontaneous awakening trial; SBT 5 spontaneous breathing trial;
SOFA 5 Sequential Organ Failure Assessment; VAC 5 ventilator-associated condition; VAP 5 ventilator-associated
pneumonia

associated pneumonia (VAP) remains

Ventilator
an important cause of morbidity, mortality, and

increased health-care costs in patients who are mechanically ventilated.1,2 Because VAP is a nosocomially
acquired infection, it is regarded as an important
patient safety measure, and there have been numerous
efforts to promote its prevention.3,4 Reported VAP
rates have been falling, and surveillance data in the
United States up to 2010 reported incidences ranging
from zero to six cases per 1,000 ventilator-days.5 However, there is concern that reported rates are variable

depending on the personnel who collect the data,

the intensity of surveillance, and interpretation of
clinical or laboratory data.6,7 Further, it is recognized
For editorial comment see page 1429
that surveillance of VAP may systematically underestimate its occurrence and that true rates may be much
higher.8,9 Particularly problematic is that VAP rates
as reported from surveillance initiatives, where there

journal.publications.chestnet.org

Downloaded From: http://journal.publications.chestnet.org/ by a Hinari User on 07/30/2016

CHEST / 144 / 5 / NOVEMBER 2013

1453

may be variability and difficulty applying current

definitions, may not correlate with patient-centered
outcomes.10
Moreover, because the definition of VAP is nonspecific and does not correlate with histopathologic
findings of pneumonia, irrespective of the method of
microbiologic confirmation chosen,11,12 it has been
argued that VAP is not an acceptable quality measure
in the ICU.13 The Centers for Disease Control and Prevention have consequently implemented an alternate
surveillance paradigm for patients who are mechanically ventilated, moving from pneumonia to broader
complications in general. These new surveillance definitions were designed to be more objective and more
efficient to collect and are termed ventilator-associated
conditions (VACs) and infection-related ventilatorassociated complications (iVACs), where iVAC is a
subset of VAC.14
Evaluative data regarding these new concepts are
limited. In initial studies, VAC was found to correlate
with clinical outcomes and require less time than VAP
for its determination.15 Further, the objective measures
of respiratory deterioration as measured by sustained
increases in ventilator settings, such as those quantified in VAC, correlate with increased length of stay
(LOS) and hospital mortality.16
Because VAP may be a cause of VAC, and given
the current focus on VAP rates and implementation
of measures to prevent VAP, there is a need for further data regarding the association between VAC,
iVAC, and VAP, their clinical outcomes, and their
responsiveness to measures that are designed to prevent VAP. We sought to determine these relationships
in a large dataset of patients in whom the Canadian
Manuscript received April 8, 2013; revision accepted June 30,
2013.
Affiliations: From the Department of Medicine (Drs Muscedere
and Heyland, Ms Jiang, and Mr Day), Kingston General Hospital,
Queens University, Kingston, ON; the Sunnybrook Research
Institute (Dr Sinuff), Sunnybrook Health Sciences Center and
the Interdepartmental Division of Critical Care Medicine, University of Toronto, Toronto, ON; the Center for Health Evaluation
and Outcome Sciences and Department of Medicine (Dr Dodek),
Providence Health Care and University of British Columbia, Vancouver, BC; the Department of Critical Care Medicine (Dr Keenan),
Fraser Health Authority, BC, and the Department of Medicine
(Dr Keenan), University of British Columbia, Vancouver, BC; the
Vancouver Island Health Authority (Dr Wood), Victoria, BC; and
the Jewish General Hospital (Dr Laporta), McGill University,
Montreal, QC, Canada; the Department of Population Medicine
(Dr Klompas), Harvard Medical School and Harvard Pilgrim
Health Care Institute, Boston, MA; and Brigham and Womens
Hospital (Dr Klompas), Boston, MA.
Funding/Support: The authors have reported to CHEST that
no funding was received for this study.
Correspondence to: John Muscedere, MD, Room 5-411, Angada 4,
Kingston General Hospital, 76 Stuart St, Kingston, ON, K7L 2V7,
Canada; e-mail: muscedej@kgh.kari.net
2013 American College of Chest Physicians. Reproduction
of this article is prohibited without written permission from the
American College of Chest Physicians. See online for more details.
DOI: 10.1378/chest.13-0853

VAP guidelines17,18 were implemented and herein report

the results.

Materials and Methods

This was a retrospective analysis of a prospective, multicenter
study that measured the implementation of VAP clinical practice guidelines over 24 months. The study was conducted in six
academic and five community, medical/surgical/trauma ICUs
(10 Canada, one United States); the average number of beds was
18.5 (SD, 3.7), and 10 had a closed administrative structure.
The complete description of the ICUs and full study details have
been published elsewhere.19,20 Briefly, in an interrupted timeseries design, evidence-based recommendations for the prevention, diagnosis, and treatment of VAP were introduced using a
multifaceted intervention. The effect of these interventions on
concordance with guideline recommendations and clinical outcomes was then assessed.
Study Enrollment
In each ICU, 30 consecutive adult patients who met the inclusion criteria of age . 16 years and who were mechanically ventilated for . 48 h were enrolled during the baseline and three
follow-up data collection periods at 6, 15, and 24 months after the
baseline period between June 1, 2007 and December 1, 2009. For
patients who met the inclusion criteria, there were no exclusion
criteria. A total of 330 patients were enrolled during each data
collection period for a total of 1,320 patients. The median number
of days required for patient enrollment per site over the four
study periods was 55 days (range, 23-140 days).
Data Collection/Outcomes
Patient characteristics collected included age, sex, comorbidities, APACHE (Acute Physiologic and Chronic Health Evaluation)
II21 score at the time of ICU admission, and Sequential Organ
Failure Assessment (SOFA)22 scores at the time of study enrollment (48 h after admission). Clinical outcomes collected included
duration of mechanical ventilation, ICU LOS, hospital LOS, antibiotic use, and ICU and hospital mortality.
Research coordinators collected data by direct observation or
chart review for concordance with each of the 14 guideline recommendations, including 10 prevention recommendations. Concordance was determined as previously described.20,23 The percent
concordance with each prevention recommendation and in aggregate was calculated for each of the four data collection periods.
Although not formally part of the VAP guideline, best practices
for discontinuation of mechanical ventilation were encouraged
throughout the study, and data regarding spontaneous awakening
trials (SATs) and spontaneous breathing trials (SBTs) were collected prospectively.24
Clinically suspected VAP was defined as new or progressive
and persistent infiltrates on a chest radiograph plus two of the
following: abnormal WBC count, presence of fever or hypothermia,
purulent sputum, and deterioration in gas exchange (Table 1).25
Study patients were screened daily for these criteria based on data
that were collected for clinical purposes only; the diagnosis was
then confirmed by the attending physician and the site principal
investigator. In the absence of a reference standard for VAP, all
suspected cases were adjudicated centrally by the study coprincipal investigators (J. M., T. S.) to confirm that the clinical course
was compatible with VAP, based on culture results, antibiotic
response, and lack of other causes to explain the signs and symptoms. Discrepancies were resolved by consensus.

1454

Downloaded From: http://journal.publications.chestnet.org/ by a Hinari User on 07/30/2016

Original Research

Table 1Definitions of VACs, iVACs, and VAP

Syndrome
VAP

VAC

iVAC

Definition
New or progressive and persistent infiltrates on
a chest radiograph plus 2 of the following:
abnormal WBC count (, 4,000 WBC/mL
or . 12,000 WBC/mL), presence of fever or
hypothermia( , 36C or . 38C), purulent sputum,
and deterioration in gas exchange
An increase in daily minimum PEEP . 3 cm H2O
or an increase of the daily minimum Fio2 . 0.20
sustained for 2 calendar days in a patient who
had a baseline period of stability or improvement
on the ventilator, defined by 2 calendar days of
stable or decreasing daily minimum Fio2 or PEEP
An episode of VAC associated with alterations in
WBC count ( 12,000 cells/mL or 4,000 cells/mL)
or temperature (. 38C or , 36C) within
2 calendar days of the start of the VAC
and 4 days of new antibiotics

iVAC 5 infection-related ventilator-associated complication; PEEP 5

positive end expiratory pressure; VAC 5 ventilator-associated condition;
VAP 5 ventilator-associated pneumonia.

VAC was defined as an increase in daily minimum positive end

expiratory pressure (PEEP) 3 cm H2O or an increase of the
daily minimum Fio2 0.20 sustained for 2 calendar days in a
patient who had a baseline period of stability or improvement on
the ventilator, defined by 2 calendar days of stable or decreasing
daily minimum Fio2 or PEEP. iVAC was defined as an episode of
VAC associated with alterations in WBC count ( 12,000 cells/mL
or 4,000 cells/mL) or temperature (. 38C or , 36C) within
2 calendar days of the start of the VAC and 4 days of new antibiotics.15 Although VAP outcomes were collected prospectively,
VAC and iVAC definitions were applied as per guidance from the
CDCs National Healthcare Safety Network to all the patients in
the dataset retrospectively by querying the electronic study database, which contained all the necessary variables to determine if
VAC or iVAC were present.
Data Analysis
Baseline characteristics and clinical outcomes were compared
between patients who had and those who did not have VAP, VAC,
or iVAC. Categorical variables are described as counts and percentages and compared using the Mantel-Haenszel test stratified by ICU. Continuous variables are described as means and
SDs or medians and quartiles for skewed data, and differences
were assessed by linear mixed effects models with ICU as a
random effect. Comparisons between groups were by independent t tests and between periods by one-way analysis of variance.
Patients who died in ICU, hospital, or while ventilated were considered discharged at time of death. Since there is no reference
standard for VAP, VAC and iVAC are proposed as alternative outcome measures rather than as proxies for VAP. Thus, the ability of
VAC to predict VAP is not assessed by measures of diagnostic
accuracy. Rather, the association between VAP, VAC, and iVAC is
cross-tabulated, and the chance-corrected agreement is summarized by the k statistic.26 Two conditional time-dependent Cox
proportional hazards models were constructed to separately estimate the time-dependent effect of VAC and VAP on hospital
mortality while adjusting for the following prespecified covariates:
age, APACHE II score, baseline SOFA score, admission diagnosis, and number of comorbidities. Patients discharged from the
hospital were assumed to have survived longer than any patients

who died in hospital. Two separate multiple conditional logistic

regression models were constructed to identify the effect of concordance with VAP-preventive recommendations on VAC and
iVAC. The original multivariate model included baseline characteristics, percentage of mechanical ventilation days with SATs,
percentage of mechanical ventilation with SBTs, and concordance
for individual preventive measures with the exception of closed
suctioning system, oral route of intubation, and povidone-iodine
mouth wash (since these were all close to 100% or 0%). Backward
selection was then used to arrive at the final result. The Cox and
logistic models were conditional on (ie, stratified by) ICU.
All P values are two-sided without adjustment for multiple
comparisons, and a P value of .05 was considered statistically
significant. A trend toward statistical significance was considered
to be present when the P value was .10 but . .05. Analyses were
performed using SAS V9.2 (SAS Institute Inc). The original study
was approved by the research ethics board of the coordinating
center (DMED-983-06, Queens University, Kingston, Ontario)
and the local research ethics board of each participating hospital.

Results
A total of 1,320 patients were enrolled over the
four study periods. There were no significant differences in baseline patient characteristics across the
four study periods, with the exception that SOFA at
the time of enrollment (48 h) was slightly lower during
the third and fourth data collection periods, respectively
(mean SD, 4.9 3.3, 4.6 3.2, 4.2 3.1, 4.3 3.2 for
each time period; P 5 .04) (Table 2). Overall, enrolled
patients had a high severity of illness, multiple comorbidities, prolonged hospitalization, and high mortality
rates (Table 2).
Of the 1,320 patients, VAC developed in 139 (10.5%),
iVAC in 65 (4.9%), and VAP in 148 (11.2%) (e-Table 1).
The relationships between VAP, VAC, and iVAC
are shown in Table 3 and Figure 1. The agreement
Table 2Baseline Characteristics, Outcomes of All
Patients Enrolled in the Study
Patient Characteristics

Value (N 5 1,320)

Age, mean SD
Men, No. (%)
APACHE II on admission, mean SD
Admission diagnosis, No. (%)
Medical
Surgical: elective
Surgical: emergency
No. of comorbidities, mean SD
SOFA at 48 h, mean SD
Maximum SOFA, mean SD
ICU LOS, median (q1, q3), d
Hospital LOS, median (q1, q3), d
Duration of MV, median (q1, q3), d
Ventilator-free days at 28 d, median (q1, q3)
Mortality-ICU, No. (%)
Mortality-hospital, No. (%)

59.6 17.2
792 (60.0)
23.0 7.5
972 (73.6)
104 (7.9)
244 (18.5)
2.2 1.8
3.7 3.1
7.5 3.9
9.9 (6.0, 16.6)
23.3 (12.6, 44.2)
6.8 (4.1, 11.7)
17.9 (0.7, 22.9)
335 (25.4)
443 (33.6)

APACHE 5 Acute Physiologic and Chronic Health Evaluation; LOS 5

length of stay; MV 5 mechanical ventilation; q 5 quartile; SOFA 5
Sequential Organ Failure Assessment.

journal.publications.chestnet.org

Downloaded From: http://journal.publications.chestnet.org/ by a Hinari User on 07/30/2016

CHEST / 144 / 5 / NOVEMBER 2013

1455

Table 3Relationship Between VAP, VAC, and iVAC

VAP as the Comparison Standard (n 5 148)
Condition as Compared
With VAP
VAC (n 5 139)
iVAC (n 5 65)

Presence of Both VAP

and VAC or iVAC

Sensitivity, %

Specificity, %

Positive Predictive Value, %

Negative Predictive Value, %

39 of 148 (26.4%)
26 of 148 (17.6%)

26
18

91
97

28
40

91
90

See Table 1 legend for expansion of abbreviations.

(k statistic) between VAP and VAC was 0.18 (95% CI,

0.11-0.26), and the agreement between VAP and
iVAC was 0.19 (95% CI, 0.11-0.27). Patients who had
VAC were more likely to also be diagnosed with VAP
(39 of 139 [28.1%]) than patients who did not have
VAC (109 of 1,181 [9.2%], P , .001). Patients who
had iVAC were more likely to have VAP (26 of 65
[40.0%]) than patients who did not (122 of 1,255 [9.7%],
P , .0001). There was a trend toward the diagnosis
of less VAP in patients who had VAC as compared
with those with iVAC (39 of 139 [28.1%] vs 26 of
65 [40.0%], P 5 .1). Of the patients who had VAP,
39 (26.4%) also had VAC, and 26 (17.6%) also had
iVAC. Overall, 1,072 (81.2%) patients did not develop
any of VAP, VAC, or iVAC. When participating centers were ranked on the basis of the incidence of VAC,
iVAC, or VAP, some ICUs had similar rankings on all
three measures, whereas some were significantly different between the three (e-Table 2).
When patients who developed VAC were compared with those who did not, patients who had VAC
had higher SOFA scores at the time of enrollment
and had worse clinical outcomes, including increased
duration of ICU stay, hospital stay, and mechanical
ventilation, and higher hospital mortality (Table 4).
Similarly, compared with patients who did not have

iVAC, patients with iVAC had higher SOFA scores

at the time of enrollment and had worse clinical outcomes, including longer duration of ICU stay, hospital
stay, and mechanical ventilation, and a trend toward
increased hospital mortality (Table 5). In contrast,
compared with patients who did not have VAP, patients
who developed VAP were younger and had fewer
comorbidities but had longer ICU LOS, hospital LOS,
and duration of mechanical ventilation (Table 6).
There was no difference in mortality between patients
who developed VAP and those who did not. In multivariate analysis, both VAC and VAP were associated
with increased mortality (e-Table 3); the hazard ratio
(95% CI) for VAC was 2.09 (1.59-2.75), P , .0001,
and for VAP it was 1.50 (1.09-2.06), P 5 .01.
Although the rates of VAP and VAC decreased
across the four time periods, iVAC rates remained
steady (Fig 2, e-Table 4). Compared with patients
who did not develop VAC, those who did develop VAC
were less likely to be concordant with the oral route
of intubation but were more likely to be concordant
with the recommendations for the frequency of changes
to heat moisture exchangers (every 7 days or if soiled),
frequency of change with heated humidifier (per patient
or if soiled or damaged), and frequency of change of
suctioning system (per patient or if soiled or damaged)

Figure 1. The relationship between VAP, VAC, and iVAC. iVAC 5 infection-related ventilator-associated
complication; VAC 5 ventilator-associated condition; VAP 5 ventilator-associated pneumonia.
1456

Downloaded From: http://journal.publications.chestnet.org/ by a Hinari User on 07/30/2016

Original Research

Table 4Comparison Between the Patients Who Developed VAC and Those Who Did Not
Measure

VAC (n 5 139)

Age, mean SD, y

Women, No. (%)
ICU admission diagnosis, No. (%)
Medical
Surgical: elective
Surgical: emergency
APACHE II, mean SD
Comorbidities, mean SD
SOFA on day of enrollment, mean SD
Days in hospital before ICU admission, median (q1, q3)
ICU LOS, median (q1, q3), d
Hospital LOS, median (q1, q3), d
Duration of MV, median (q1, q3), d
No. days on antibiotics
Hospital mortality, No. (%)

Non-VAC (n 5 1,181)

62.2 16.7
52 (37.4)

59.3 17.2
476 (40.3)

102 (73.4)
14 (10.1)
23 (16.5)
22.7 7.3
2.2 1.8
5.6 3.0
0.7 ( 0.2, 3.1)
18.9 (12.1, 31.6)
31.7 (19.0, 59.9)
15.4 (9.8, 26.6)
15.5 7.3
69 (49.6)

870 (73.7)
90 (7.6)
221 (18.7)
23.1 7.6
2.2 1.8
4.4 3.2
0.4 (0.1, 2.2)
9.0 (5.8, 14.9)
21.8 (12.1, 42.6)
6.2 (3.9, 10.5)
9.0 6.5
374 (31.7)

P Value
.12
.61
.84

.52
.56
.0006
.64
, .0001
, .0001
, .0001
, .0001
, .0001

See Table 1 and 2 legends for expansion of abbreviations.

(e-Table 5). Compared with patients who did not

have iVAC, patients who had iVAC were less likely to
be concordant with the oral route of intubation and
frequency of ventilator circuit changes (per patient
or when soiled or damaged) but were more likely
to be concordant with frequency of change of heated
humidifier, heat moisture exchanger, and suctioning
system. In multivariate analysis for the impact of VAP
preventive measures on the development of VAC
and iVAC, none reached statistical significance. However, for VAC, there was a trend for increased percentage of days with SAT and days with SBT to be
protective (OR [95% CI], 0.93 [0.87-1.00], P 5 .05;
0.97 [0.94-1.01], P 5 .10, respectively). For iVAC,
there was a trend toward increased percentage of
days with SAT and increased frequency of concordance with changes in the suctioning system to be
protective (OR [95% CI]: 0.89 [0.79-1.00], P 5 .05;
0.98 [0.97-1.00], P 5 .08, respectively).

Discussion
In a large dataset of patients who are mechanically
ventilated, in which evidence-based VAP guidelines
were systematically implemented, VAP, VAC, and
iVAC were relatively common and associated with
worse outcomes, including mortality. Of these, VAC
had the strongest association with increased mortality.
Patients who had VAC and iVAC were much more
likely to be diagnosed with VAP. However, a significant number of patients who had VAC and iVAC were
not diagnosed with VAP in spite of rigorous and systematic efforts to detect VAP. Conversely, only a
minority of patients who were diagnosed as having
VAP also met the definition for VAC or iVAC, and
the agreement between VAC, iVAC, and VAP was low.
The low agreement between VAC, iVAC, and VAP
may stem from the nonspecificity of each definition,
because each may be caused by a variety of underlying pathophysiologic processes. The definition for

Table 5Comparison Between Patients Who Developed iVAC and Those Who Did Not
Measure
Age, mean SD, y
Women, No. (%)
ICU admission diagnosis
Medical
Surgical: elective
Surgical: emergency
APACHE II, mean SD
Comorbidities, mean SD
SOFA on day of enrollment, mean SD
Days in hospital before ICU admission, median (q1, q3)
ICU LOS, median (q1, q3), d
Hospital LOS, median (q1, q3), d
Duration of MV, median (q1, q3), d
No. of days on antibiotics
Hospital mortality, No. (%)

iVAC (n 5 65)

Non-iVAC (n 5 1,255)

56.8 18.5
28 (43.1)

59.8 17.1
500 (39.8)

49 (75.4)
3 (4.6)
13 (20.0)
22.8 7.7
1.8 1.8
5.7 3.1
0.5 (0.1, 2.5)
22.0 (13.7, 35.9)
34.6 (21.8, 59.9)
16.9 (11.6, 27.7)
17.8 6.7
29 (44.6)

923 (73.5)
101 (8.0)
231 (18.4)
23.0 7.5
2.2 1.8
4.5 3.2
0.4 (0.1, 2.3)
9.3 (5.9, 15.6)
22.5 (12.4, 43.6)
6.4 (4.0, 10.9)
9.3 6.6
414 (33.0)

P Value
.08
.59
.86

.76
.03
.007
.76
, .0001
.03
, .0001
, .0001
.07

See Table 1 and 2 legends for expansion of abbreviations.

journal.publications.chestnet.org

Downloaded From: http://journal.publications.chestnet.org/ by a Hinari User on 07/30/2016

CHEST / 144 / 5 / NOVEMBER 2013

1457

Table 6Comparison Between the Patients Who Developed VAP and Those Who Did Not
Measure
Age, mean SD, y
Women, No. (%)
ICU admission diagnosis
Medical
Surgical: elective
Surgical: emergency
APACHE II, mean SD
Comorbidities, mean SD
SOFA on day of enrollment, mean SD
Days in hospital before ICU admission, median (q1, q3)
ICU LOS, median (q1, q3), d
Hospital LOS, median (q1, q3), d
Duration of MV, median (q1, q3), d
No. of days on antibiotics
Hospital mortality, No. (%)

VAP (n 5 148)

Non-VAP (n 5 1,172)

54.3 19.0
54 (36.5)

60.3 16.8
474 (40.4)

100 (67.6)
15 (10.1)
33 (22.3)
22.1 7.8
1.8 1.7
4.4 3.0
0.2 (0.1, 1.7)
17.8 (11.7, 27.9)
30.9 (16.0, 55.7)
13.6 (8.7, 23.4)
15.5 7.0
47 (31.8)

872 (74.4)
89 (7.6)
211 (18.0)
23.1 7.5
2.2 1.8
4.5 3.2
0.4 (0.1, 2.3)
9.0 (5.8, 14.9)
22.2 (12.3, 42.4)
6.2 (3.9, 10.6)
9.0 6.5
396 (33.8)

P Value
.0002
.34
.03

.25
.0009
.45
.33
, .0001
.01
, .0001
, .0001
.67

See Table 1 and 2 legends for expansion of abbreviations.

VAC is intentionally broad and can include processes

such as pulmonary edema, acute lung injury/ARDS,
and atelectasis, in addition to pneumonia. Moreover,
patients who meet the definition for iVAC may have
VAC caused by an infectious nonpneumonic process
or a noninfectious pulmonary process coincident with
a nonpneumonic cause of fever, alteration in WBC
count, or need for antibiotics. In theory, all VAP cases
should be a subset of VAC cases, but we found that
only a minority of patients who had VAP also met the
definition for VAC or iVAC. A possible explanation is
that some cases of VAP may not be severe enough to
cause sufficient deterioration in parameters of mechanical ventilation to reach thresholds for the diagnosis
of VAC. Alternatively, VAP that arises in patients
without a 2-day period of stable or improving ventilatory settings would not meet criteria for VAC. In
addition, some patients with early VAP may be mechanically ventilated for , 4 days and, hence, ineligible for
VAC. These may, in turn, be influenced the patients

Figure 2. Rates of VAP, VAC, and iVAC and concordance across

the four data enrollment periods. See Figure 2 legend for expansion of abbreviations.

physiologic reserve present and severity of illness;

thus, differences between VAP and VAC rates may
be dependent on the patient population in which they
are measured. It is also possible that some patients
adjudicated as having VAP had other conditions, such
as ventilator-associated trachea-bronchitis, which would
not be expected to cause significant physiologic deterioration.27 Moreover, discordance between rates of
VAP, VAC, and iVAC may have resulted from difficulties and inaccuracies in data collection. This is likely
minimal in the context of a research study in which
the data are collected by trained research coordinators but would be more pronounced when these definitions are used broadly for surveillance. Electronic
data collection and automated analysis, which is possible for VAC but not VAP, may mitigate this concern.
Although VAC and iVAC may not be specific, their
higher correlation with worse outcomes, ease of data
collection, and objective definitions make them promising options to replace VAP as quality indicators. However, with the minimal amount of overlap between
VAP, VAC, and iVAC, the risk is that with the adoption of VAC and iVAC as quality measures, we will
miss some clinically diagnosed cases of VAP, albeit by
definition it will be cases of VAP without sustained
increases in ventilator settings that meet the VAC
thresholds. This is an important consideration as
we move forward with implementation of quality
improvement metrics for patients who are mechanically ventilated.
Moreover, there is a large body of knowledge demonstrating that preventive measures can reduce the
incidence and associated costs of VAP.28,29 In this study,
although VAP and VAC rates decreased significantly
across the four time intervals, iVAC rates did not change
despite increasing concordance with preventive measures, such as semirecumbency, subglottic secretion

1458

Downloaded From: http://journal.publications.chestnet.org/ by a Hinari User on 07/30/2016

Original Research

drainage, and chlorhexidine mouthwash. The increase

in concordance and eventual degree of concordance
for each of these interventions, however, was relatively
modest (semirecumbency, 29%-41%; subglottic secretion drainage, 36%-58%; and chlorhexidine mouthwash, 16%-50%). This may have limited the capacity
to realize an impact on IVAC rates. Moreover, it is
possible that these interventions may have little
effect on VAC and iVAC rates because they have been
designed solely for the prevention of pulmonary infection. Although some VACs and iVACs may be caused
by infection, many VACs have noninfectious (eg, cardiogenic, noncardiogenic pulmonary edema, atelectasis) causes.15 However, both VACs and iVACs should
be responsive to measures that reduce the time at risk
for complications incurred while invasively mechanically ventilated, and in multivariate analysis, there
was a trend toward a reduced risk of VAC and iVAC
with increased rates of SATs and SBTs.
Strengths of this study include the large database,
the rigorous adjudication of VAP, the measurement
of concordance, as well as the broad inclusion criteria
with few exclusion criteria, which increases the generalizability of our findings. A limitation of this study
is that we did not address the differential feasibility
of implementing VAP, VAC, and iVAC definitions in
our patient population, but this has been reported on
in other studies.15 Further limitations include the lack
of data on the total number of patients from which
this cohort was selected, the observational nature
of this study with the inability to make causal inferences instead of associations, and the relatively modest
increase in concordance with guidelines across the
time span of the study. However, to further investigate the impact of measures that have been shown to
prevent VAP on rates of VAC and iVAC, a randomized
control trial would be required. Future trials of VAPprevention strategies should include VAC and iVAC
as outcomes, or perhaps these interventions should
be designed solely for VACs or iVAC prevention.
In summary, VAP, VAC, and iVAC continue to be
relatively common in critically ill patients who are
mechanically ventilated and are associated with adverse
outcomes. There is some overlap between them, but
the agreement between VAC, iVAC, and VAP is low.
Given the association between VAC and iVAC and
adverse outcomes, they may be useful quality indicators. However, further study is required to determine
the preventability of VAC and iVAC and their relationship to other quality indicators in the ICU, such as
adherence to best practices for ventilator management.
Acknowledgments
Author contributions: Dr Muscedere takes responsibility for
the content of the manuscript, including data and data analysis.

Dr Muscedere: contributed to initial study design and analysis,

interpretation of data, drafting of the submitted article, critical
revisions for intellectual content, and providing final approval
of the version to be published.
Dr Sinuff: contributed to study design and analysis, interpretation
of data, drafting of the submitted article, critical revisions for
intellectual content, and providing final approval of the version to
be published.
Dr Heyland: contributed to study design and analysis, interpretation of data, drafting of the submitted article, critical revisions for
intellectual content, and providing final approval of the version to
be published.
Dr Dodek: contributed to study design and analysis, interpretation of data, drafting of the submitted article, critical revisions for
intellectual content, and providing final approval of the version to
be published.
Dr Keenan: contributed to study design and analysis, interpretation of data, drafting of the submitted article, critical revisions for
intellectual content, and providing final approval of the version to
be published.
Dr Wood: contributed to study design and analysis, interpretation
of data, drafting of the submitted article, critical revisions for
intellectual content, and providing final approval of the version to
be published.
Ms Jiang: contributed to study design and analysis, interpretation
of data, drafting of the submitted article, critical revisions for
intellectual content, and providing final approval of the version to
be published.
Mr Day: contributed to study design and analysis, interpretation
of data, drafting of the submitted article, critical revisions for
intellectual content, and providing final approval of the version to
be published.
Dr Laporta: contributed to study design and analysis, interpretation of data, drafting of the submitted article, critical revisions for
intellectual content, and providing final approval of the version to
be published.
Dr Klompas: contributed to initial study design, analysis, interpretation of data, drafting of the submitted article, critical revisions
for intellectual content, and providing final approval of the version
to be published.
Financial/nonfinancial disclosures: The authors have reported
to CHEST the following conflicts of interest: Dr Klompas has
received grant support for research on VAP surveillance from the
Centers for Disease Control and Prevention. He has also received
honoraria for lectures on VAP surveillance from professional societies and Premier Healthcare Alliance. Drs Muscedere, Sinuff,
Heyland, Dodek, Keenan, Wood, and Laporta; Ms Jiang; and
Mr Day have reported that no potential conflicts of interest exist
with any companies/organizations whose products or services may
be discussed in this article.
Additional information: The e-Tables can be found in the
Supplemental Materials area of the online article.

References
1. Muscedere J, Day A, Heyland DK. Mortality, attributable
mortality, and clinical events as end points for clinical trials
of ventilator-associated pneumonia and hospital-acquired
pneumonia. Clin Infect Dis. 2010;51(suppl 1):S120-S125.
2. Safdar N, Dezfulian C, Collard HR, Saint S. Clinical and economic consequences of ventilator-associated pneumonia: a
systematic review. Crit Care Med. 2005;33(10):2184-2193.
3. Ventilator-associated pneumonia. Canadian Patient Safety
Institute website. http://www.saferhealthcarenow.ca/EN/
Interventions/VAP/Pages/default.aspx. Accessed March 1, 2013.
4. Implement the IHI ventilator bundle. Institute for Healthcare Improvement website. http://www.ihi.org/knowledge/
Pages/Changes/ImplementtheVentilatorBundle.aspx. Accessed
September 13, 2013.
5. Dudeck MA, Horan TC, Peterson KD, et al. National Healthcare Safety Network (NHSN) Report, data summary for

journal.publications.chestnet.org

Downloaded From: http://journal.publications.chestnet.org/ by a Hinari User on 07/30/2016

CHEST / 144 / 5 / NOVEMBER 2013

1459

6.
7.
8.
9.

10.

11.

12.

13.
14.

15.

16.

17.

2010, device-associated module. Am J Infect Control. 2011;

39(10):798-816.
Klompas M. Interobserver variability in ventilator-associated
pneumonia surveillance. Am J Infect Control. 2010;38(3):
237-239.
Novosel TJ, Hodge LA, Weireter LJ, et al. Ventilator-associated
pneumonia: depends on your definition. Am Surg. 2012;
78(8):851-854.
Morrow L, Malesker M, Farrington K. Diagnostic criteria and
intensity of surveillance affect reportable ventilator associated
pneumonia rates [abstract]. Chest. 2006;130(4_suppl):101S.
Drees M, Hausman S, Rogers A, Freeman L, Frosch K,
Wroten K. Underestimating the impact of ventilator-associated
pneumonia by use of surveillance data. Infect Control Hosp
Epidemiol. 2010;31(6):650-652.
van der Kooi TI, de Boer AS, Mannin J, et al. Incidence
and risk factors of device-associated infections and associated mortality at the intensive care in the Dutch surveillance
system. Intensive Care Med. 2007;33(2):271-278.
Kirtland SH, Corley DE, Winterbauer RH, et al. The diagnosis of ventilator-associated pneumonia: a comparison of
histologic, microbiologic, and clinical criteria. Chest. 1997;
112(2):445-457.
Tejerina E, Esteban A, Fernndez-Segoviano P, et al. Accuracy of clinical definitions of ventilator-associated pneumonia:
comparison with autopsy findings. J Crit Care. 2010;25(1):
62-68.
Klompas M, Platt R. Ventilator-associated pneumoniathe
wrong quality measure for benchmarking. Ann Intern Med.
2007;147(11):803-805.
National Healthcare Safety Network: surveillance for ventilator-associated events. Centers for Disease Control and
Prevention website. http://www.cdc.gov/nhsn/acute-carehospital/vae/index.html. Accessed September 13, 2013.
Klompas M, Khan Y, Kleinman K, et al; CDC Prevention
Epicenters Program. Multicenter evaluation of a novel surveillance paradigm for complications of mechanical ventilation. PLoS ONE. 2011;6(3):e18062.
Klompas M, Magill S, Robicsek A, et al; CDC Prevention
Epicenters Program. Objective surveillance definitions for
ventilator-associated pneumonia. Crit Care Med. 2012;40(12):
3154-3161.
Muscedere J, Dodek P, Keenan S, Fowler R, Cook D, Heyland D;
VAP Guidelines Committee and the Canadian Critical Care
Trials Group. Comprehensive evidence-based clinical practice guidelines for ventilator-associated pneumonia: prevention. J Crit Care. 2008;23(1):126-137.

18. Muscedere J, Dodek P, Keenan S, Fowler R, Cook D, Heyland D;

VAP Guidelines Committee and the Canadian Critical Care
Trials Group. Comprehensive evidence-based clinical practice guidelines for ventilator-associated pneumonia: diagnosis
and treatment. J Crit Care. 2008;23(1):138-147.
19. Muscedere J, Sinuff T, Cook D, et al. Active observational
study of the adoption and transfer of clinical practice guidelines through education, for ventilator associated pneumonia:
the ABATE VAP Study [abstract]. Am J Respir Crit Care Med.
2011;183(1_Meeting Abstracts):A3736.
20. Sinuff T, Muscedere J, Cook DJ, Heyland DK; Canadian
Critical Care Trials Group. Implementation of clinical practice guidelines for ventilator associated pneumonia: a multicentre prospective study. Crit Care Med. 2013;41(1):15-23.
21. Knaus WA, Draper EA, Wagner DP, Zimmerman JE.
APACHE II: a severity of disease classification system. Crit
Care Med. 1985;13(10):818-829.
22. Vincent JL, Moreno R, Takala J, et al. The SOFA (Sepsisrelated Organ Failure Assessment) score to describe organ
dysfunction/failure. On behalf of the Working Group on
Sepsis-Related Problems of the European Society of Intensive
Care Medicine. Intensive Care Med. 1996;22(7):707-710.
23. Scott IA, Harper CM. Guideline-discordant care in acute
myocardial infarction: predictors and outcomes. Med J Aust.
2002;177(1):26-31.
24. Girard TD, Kress JP, Fuchs BD, et al. Efficacy and safety
of a paired sedation and ventilator weaning protocol for
mechanically ventilated patients in intensive care (Awakening
and Breathing Controlled trial): a randomised controlled
trial. Lancet. 2008;371(9607):126-134.
25. Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance
definition of health care-associated infection and criteria for
specific types of infections in the acute care setting. Am J
Infect Control. 2008;36(5):309-332.
26. Cohen J. A coefficient of agreement for nominal scales. Educ
Psychol Meas. 1960;20(1):37-46.
27. Nseir S, Di Pompeo C, Pronnier P, et al. Nosocomial tracheobronchitis in mechanically ventilated patients: incidence,
aetiology and outcome. Eur Respir J. 2002;20(6):1483-1489.
28. Lai KK, Baker SP, Fontecchio SA. Impact of a program of
intensive surveillance and interventions targeting ventilated
patients in the reduction of ventilator-associated pneumonia
and its cost-effectiveness. Infect Control Hosp Epidemiol.
2003;24(11):859-863.
29. Zack JE, Garrison T, Trovillion E, et al. Effect of an education
program aimed at reducing the occurrence of ventilatorassociated pneumonia. Crit Care Med. 2002;30(11):2407-2412.

1460

Downloaded From: http://journal.publications.chestnet.org/ by a Hinari User on 07/30/2016

Original Research