claim that a negative Gram stain on an EA sample is of great NPV for a diagnosis of VAP and
allows for the decision to not start antibiotic treatment. A positive Gram stain on a plugged
telescoping catheter sample indicates VAP is highly probable and treatment should be
promptly started. In patients in whom the EA proves positive and the plugged telescoping
catheter sample negative, the best therapeutic approach is not as clear, although the clinician
should probably make the decision to start antimicrobial treatment pending subsequent
adjustment when the culture results are ready. A negative EA culture in a patient without a
recent change in antibiotics (within 72 hours) has strong negative predictive value (94%) for
VAP (225).
Value of Cultures
The etiologic cause of pneumonia can be determined by culture of an EA with initial
microscopic examination (227–229). From a practical standpoint, quantitative culture counts
between 100 and 1000 cfu/mL for PSBsamples and between 1000 and 10,000 cfu/mL for BAL
specimens should be considered probably positive for VAP and are an indication for antibiotic
treatment (230). Counts of100,000 cfu/mL in blind, aspirated, undiluted tracheal secretions
suggest infection rather than colonization (227).
Several technical considerations can affect the results of quantitative cultures and may
explain why the reported accuracy of invasive methods varies so widely. Methodological
issues responsible for the inconsistent results of published studies have been summarized in a
meta-analysis (231). One of the most frequent problems is the dilution of BAL, which could
minimize bacterial counts. This occurs particularly in patients with severe COPD. Knowledge
of the extent of dilution can dramatically increase the value of quantitative cultures. A recently
published study has examined the effect of the dilution factor used for the BAL culture on the
bacterial count (232). The authors compared the concentration of urea in serum and BAL to
determine the dilution factor for the sample and established that 17 additional patients would
have reached the cutoff level after correction for the dilution effect, which varied between
1.8- and 130-fold. These findings stress the implications of the dilutions used in cultures for the
diagnosis and treatment of these patients.
The recent starting or a change in antibiotic therapy is among the main factors causing
false-negative quantitative cultures, especially if the start or change occurs in the preceding
24 to 72 hours (206,233). Thus, all cultures should be obtained prior to treatment. If this is not
possible, then a change in the diagnostic threshold could be useful (179,233). For BAL, the use
of a threshold 10-fold lower than usual may avoid some false-negative results in patients given
antibiotics before testing.
Preemptive Rapid Cultures
The traditional laboratory processing of a respiratory secretion specimen for bacterial isolation
usually takes between three and four days to provide the clinician with a result. After plating
the sample and incubating for 24 to 48 hours, bacterial counts have to be performed and strains
isolated and grown in pure culture. This is followed by microorganism identification and
antimicrobial sensitivity testing, which takes a further 24 hours. To this, we would have to add
the time taken for transmitting information, writing reports, and making therapeutic decisions.
This late information, at least in areas such as blood cultures, clearly helps to improve the
prescription of drugs, optimizes their consumption, and reduces costs, but it has not yet been
possible to establish its impacts on shortening hospital stay or decreasing mortality (234).
Antibiogram procedures require a standardized inoculum and usually start with isolated
bacteria in culture. It is known, however, that antibiograms performed directly on clinical
specimens, i.e., omitting the bacterial-isolation step, can provide preliminary information,
which generally correlates well with that offered by standard procedures. This is presently
undertaken with blood cultures and urine or cerebro spinal fluid (CSF) samples in
circumstances in which the urgency demands. A procedure that is not affected by the
inoculum is the so-called E-test. This method consists of a strip impregnated with increasing
concentrations of an antibiotic. After its diffusion in agar, the strip provides a minimum
inhibitory concentration (MIC) for the particular bacterium and antimicrobial agent tested. We
compared the results of a direct E-test antibiogram, including six antimicrobials commonly
Nosocomial Pneumonia in Critical Care 189