J. Godard et B. Allaouchiche, VENTILATOR-ASSOCIATED PNEUMONIA - VALUE O F DIAGNOSTIC-TOOLS, Annales francaises d'anesthesie et de reanimation, 13(5), 1994, pp. 699-704
The authors analyzed in published studies providing histologic data th
e value of diagnostic means of nosocomial bronchopneumonia (NBP) in cr
itically ill patients. At least 10 % of patients whose trachea is intu
bated and the lungs mechanically ventilated suffer an histologically-c
onfirmed NBP. Histology of NBP consists of non systematized foci of in
fection, localised in 90 % of cases in the posterior segments of the t
ower pulmonary lobes. In fact, small foci are disseminated in both lun
gs, usually within large areas of non infectious pulmonary lesions and
are often multimicrobial. This may explain why the clinical and radio
logical diagnosis is inaccurate in more than 40 % of cases. On chest X
-ray, aeric bronchogramm and alveolar infiltrates are suggestive but n
on-specific for NBP, as are also fever, leukocytosis and purulent sput
um. Therefore, to confirm the diagnosis of infection and to isolate th
e microorganisms responsible for NBP, additional investigations are re
quired whose interpretation is uneasy. In a simplified manner, bacteri
ological inoculum takes a course parallel to the histological bronchop
ulmonary lesions: no bacteria in the absence of infection, high bacter
ial inoculum in case of confluent NBP, intermediary in case of bronchi
tis, however with exceptions when antibiotherapy has recently been mod
ified. In fact, neither bronchial protected brushing nor broncho-alveo
lar lavage techniques have a perfect sensitivy and specificity. As an
exemple, a diagnostic procedure in assessed in calculating the probabi
lity of an accurate diagnosis of NBP. It is concluded that diagnosis o
f ventilator-associated pneumonia remains still unreliable with the me
ans available today.