Confirming the Etiology
Although the presence of clinical signs should raise a suspicion of VAP, confirming the
diagnosis is much more difficult, since clinical variables are of no use for defining the
microbiologic etiology of pneumonia. All patients suspected of having VAP should undergo
lower respiratory tract sampling, with subsequent microscopic analysis and culture of the
specimen (24). For an etiologic diagnosis of VAP, a quantitative or semiquantitative lower
respiratory tract culture is needed. The threshold bacterial count depends on the type of
specimen collected (more or less dilution of the original respiratory secretions), the collection
method, and the sampling time (whether there has been a recent change or not in antimicrobial
therapy) (24). Growth below the threshold is assumed to be due to colonization or
contamination. This type of information has been used as a basis for decisions about whether
to start antibiotic therapy, which pathogens are responsible for infection, which antimicrobial
agents to use, and whether to continue therapy (199,200).
Today, the most common methods of sampling the lower respiratory tract are
endotracheal aspirates (EA), protected specimen brush (PSB) samples, and bronchoalveolar
lavage (BAL). No single method is considered better than any other, including bronchoscopic
versus non-bronchoscopic sampling (182,201–207). The evidence indicates that bronchoscopic
sampling does not improve mortality, length of hospital stay, duration of mechanical
ventilation, or length of ICU stay (208–210). However, it may lead to a narrower antimicrobial
regimen or more rapid de-escalation of antimicrobial therapy (208,211–213).
To adequately process a sample and interpret the results, it is essential that the laboratory
is informed of the type of sample submitted (24). Nonetheless, in a survey of different
sampling techniques, Ruiz et al. (214) found no differences in rates of diagnoses, changing of
antimicrobial treatment due to etiologic findings, length of ICU stay and of mechanical
ventilation, and crude 30-d- or adjusted mortality.
Quantitative cultures have been found especially useful for diagnosing VAP in patients
with a low or equivocal clinical suspicion of infection (206,215). Fagon et al. performed a
multicenter, randomized, uncontrolled trial to evaluate the effects on clinical outcome and
antibiotic use of the two approaches, “clinical” versus “bacteriological,” to diagnose VAP and
select the initial treatment for this condition (211). These authors concluded that the invasive
management strategy was significantly associated with fewer deaths at 14 days, earlier
improvement of organ dysfunction, and a reduced use of antibiotics.
Blood cultures are not very useful for diagnosing VAP (216,217). Overall, their sensitivity
is less than 25%, and, when positive, the organisms detected could largely correspond to an
extrapulmonary source, even if VAP is also present (218). Blood cultures are mainly useful for
diagnosing extrapulmonary infections or for detecting respiratory pathogens in patients with
borderlinerespiratory sample cultures (218–220).
Value of Rapid Gram Stain
A reliable EA Gram stain can be used to decide upon initial empirical antimicrobial therapy
(221) and infrequently gives rise to inappropriate treatment (26,211). In our experience, in
patients in whom VAP is suspected, Gram staining of an EA sample and timely reporting of
the results is a method of extraordinary value for the etiologic diagnosis of VAP and for early
decision making about treatment. At our center, the diagnostic performance of an EA Gram
stain when there is suspicion of VAP is as follows: sensitivity, (Sen) 91%; specificity (Spec),
61%; positive predictive value (PPV), 50.5%; negative predictive value (NPV), 94%; accuracy (A),
70%; positive likelihood ratio, 2.3; negative likelihood ratio, 0.14; post-test probability for a
negative result, 6%. In other words, a negative result will reasonably exclude a diagnosis of
VAP. The medical literature, however, is filled with varying data on the Sen (57% to 95%), Spec
(48% to 87%), PPV (47% to 78%), NPV (69% to 96%), and accuracy (60% to 88%) of the Gram
stain technique in managing the patient with VAP (200,222–226). Blot et al. (225) assessed the
value of the Gram stain in patients with suspected VAP used on respiratory secretions taken
both by endotracheal aspiration and using a plugged telescoping catheter. Used on EA, the
method showed a high Sen for diagnosing microbiologically proven VAP (91%) and a high
NPV (94%) in patients without a recent antibiotic treatment change. On plugged telescoping
catheter samples, the Gram stain showed a high Spec (95%) but lower Sen (67%). These authors
188 Bouza and Burillo