data changes clinical or economic outcomes. Thus, many
physicians continue to use endotracheal specimens and
other clinical features in diagnosing VAP.
Endotracheal Aspiration in the
Diagnosis of Ventilator-
Associated Pneumonia*
Analysis
Deborah Cook, MD, FCCP; and Lionel Mandell, MD, FRCPC
(CHEST 2000; 117:195S–197S)
The most acceptable reference standards or “gold stan-
dards” usually include biopsy or autopsy reports. Since
these often are not feasible standards, alternatives are
usually used. They include cultures of pleural and blood
specimens, long-term follow-up to exclude other diag-
noses, and quantitative cultures of specimens obtained
through bronchoscopic techniques. The use of broncho-
scopic specimens makes interpretation of the reference
standard particularly difficult, since the test properties of
BAL and protected-specimen brush sampling still are
being evaluated. Tables 9–11 cover the reference stan-
dards used in these studies.
We calculated the sensitivity and specificity of a test
result on an endotracheal specimen according to formulas
presented earlier in this report. In parentheses in Tables
9–11, we indicated the sensitivity and specificity estimates
of the authors. The calculation of likelihood ratios requires
knowledge of the number of patients with and without
pneumonia, as determined by the reference standard,
separated according to the culture results from endotra-
cheal specimens. Most studies did not report data on all
four groups of patients, so we could not calculate a 2 ϫ 2
table and did not include likelihood ratios for these
studies. Due to heterogeneous study designs and results,
we did not statistically pool data in the form of a meta-
analysis.
Abbreviations: EA ϭ endotracheal aspiration; VAP ϭ ventila-
tor-associated pneumonia
he time-honored method of identifying bacterial patho-
Tgens that are potentially responsible for nosocomial
pneumonia in patients receiving mechanical ventilation is
the microscopic examination of specimens obtained by
endotracheal aspiration (EA). This technique is the sim-
plest noninvasive means of obtaining respiratory secre-
tions from patients receiving mechanical ventilation; it is
readily performed at the bedside and requires minimal
training by health-care providers. This section focuses on
clinical studies evaluating diagnostic procedures using
endotracheal specimens (ie, cytologic examination, anti-
body coating, elastin fibers, Gram’s stain, and culture) in
immunocompetent adults with suspected VAP.
The cytologic examination of specimens containing a
large number of leukocytes and a paucity of epithelial cells
is likely to produce the most valid representation of
infectious organisms. A test for the detection of the
presence of antibody coating on bacteria has been devel-
oped in an attempt to distinguish organisms that are
colonizing the lower respiratory tract from those that
actually are infecting it. The test is based on the premise
that an infection will elicit an antibody response in the host
and that this response will be detectable on the microor-
ganism. In addition, using 40% potassium hydroxide to
detect elastin fibers has been promoted as a rapid and
inexpensive way to demonstrate the destruction of lung
parenchymal tissue that is caused by pneumonia.
However, an analysis of endotracheal specimens ob-
tained by aspiration has been diagnostically inadequate.
Several qualitative articles have reviewed the use of
endotracheal specimens to diagnose VAP. Among the
challenges for investigators and clinicians are the follow-
ing: distinguishing upper from lower respiratory tract
infection; distinguishing infection from colonization and
contamination; standardizing aspiration collection meth-
ods and microbiological techniques; and interpreting test
properties in light of the host’s immune status, the patho-
genic load, and the effect of prior antimicrobial therapy.
Newer bronchoscopic methods for diagnosing VAP
have become the focus of recent investigations, confer-
ences, and professional documents. Invasive approaches
have not necessarily been adopted by clinicians,2 at least in
part because of procedural access, cost, and the absence of
compelling evidence that treatment based on the derived
Results
The comprehensive literature search described earlier2
yielded 12 relevant citations2,13–15,18,23,91–96 published from
1985 to 1995. Nine studies evaluated cultures from EA,
two studies evaluated antibody coating, and three studies
evaluated elastin fibers. Study characteristics and the results
of Gram’s stain and aspiration culture appear in Table 9. The
same data are recorded for antibody coating in Table 10, and
for elastin fibers in Table 11. The original articles present the
details of the design and results of the studies.
Most studies were prospective. Several stated that
patients were enrolled consecutively. Most patients re-
ceived mechanical ventilator support. Some studies pro-
filed patients according to the duration of ventilator
support. Most patients were receiving antibiotics at the
time of testing. The methods of analyzing endotracheal
specimens were recorded in all studies. In no studies were
test results interpreted by investigators blinded to the
results of other tests. In one study,17 the reference stan-
dard was interpreted by an investigator blinded to the
results of the test under evaluation. Most studies focused
on sensitivity by enrolling patients with suspected VAP. A
valuable study by Torres et al23 determined specificity in
patients without suspected VAP.
*From the Department of Medicine (Dr. Cook), St. Joseph’s
Hospital, Hamilton, Ontario, Canada; and the Department of
Medicine (Dr. Mandell), Henderson General Hospital, Hamil-
ton, Ontario, Canada.
Most investigators acknowledged the difficulty with
choosing a reference standard for VAP. For example, one
CHEST / 117 / 4 / APRIL, 2000 SUPPLEMENT
195S
Boyer
