Infectious Diseases in Critical Care Medicine

CLINICAL PRESENTATION AND DIAGNOSTIC TESTING
Establishing a Diagnosis of VAP
There is no single pathognomonic test that ensures or excludes the presence of VAP. Wide
spectrum antimicrobial therapy should be started if there is reasonable suspicion, and this can
then be adjusted once the results of microbiological tests become available (26,179,180).
The American Thoracic Society and the Infectious Disease Society of America (24)
recently defined VAP as the presence of new or progressing lung infiltrates plus clinical
evidence that the infiltrates are of an infectious origin. The presence of infection is determined
on the basis of two or more of the following data: fever greater than 38 8 C or hypothermia,
leukocytosis or leukopenia, purulent secretions, and reduced oxygenation (181). Diagnosis of
VAP requires an abnormal chest radiograph. In the absence of demonstrable pulmonary
infiltrates, a diagnosis of infective tracheobronchitis is pursued (182).
Unfortunately, radiographic data from chest X Rays show low sensitivity and specificity
for diagnosing VAP (169,183–185). Radiological infiltrates are difficult to define and difficult to
distinguish from other frequent conditions in this patient population. Moreover, they correlate
poorly with CT data and postmortem criteria. Lung infiltrates are also provoked by other
causes such as atelectasis, pulmonary edema, pleural effusion, pulmonary hemorrhage, lung
infarction, and ARDS (186). In a study comparing the use of portable chest X Rays and CT
scans, 26% of infiltrates detected by CT were missed by the portable chest X Rays, particularly
those located in lower lobes (187). This also occurs when we compare any gold standards such
as the postmortem examination (181,185) and bronchoscopic examination (185,188–190).
CT has shown a sensitivity and a specificity of 53% and 63%, respectively, for the diagnosis
of VAP (191). Ground glass infiltrates appeared to have a higher specificity, but were found in
only 45% of patients. Added to these limitations, we find interobserver variability in interpreting
radiological observations (192). To date, multi-detector row CT with its excellent contrast
resolution is the most sensitive modality for evaluating lung parenchyma infections (193).
The sensitivity of the use of other clinical data increases if only one criterion is considered
sufficient, but this occurs at the expense of specificity, leading to significantly more antibiotic
treatment (181). For patients with ARDS, suspicion of pneumonia should be high, and even
one of the clinical criteria described should prompt further diagnostic testing (194).
When clinical diagnoses of nosocomial pneumonia were compared with histopathologic
diagnoses made at autopsy, pneumonia was diagnosed correctly in less than two-thirds of
cases (195).
The Clinical Pulmonary Infection Score (CPIS) described by Pugin et al. (196) is a
multifactorial system used to make a diagnosis of VAP. This method is based on assigning
points to clinical, radiological, and physiological variables. In the original report, a score of
6 points was found to correlate well with a diagnosis of VAP. However, in subsequent studies,
the sensitivity and specificity of the CPIS score proved to be not much improved over the
subjective clinical approach unless the score included microbiological information (rapid Gram
stain or culture results) (Table 4) (197). Nonetheless, a clinical score of6 is good at identifying a
subset of patients who either do not require antimicrobial therapy for VAP or, when antibiotics
are prescribed, are amenable to a short course (three days) of treatment, provided the patient
remains clinically stable and with a nonincreased score three days later (198).

Table 4 Modified Clinical Pulmonary Infection Score

Points

Criterion 0 1 2

Temperature 36.5 8 C–38.4 8 C 38.5 8 C–38.9 8 C  368 C– 398 C

Blood leukocytes (/ml) 4,000–11,000 <4,000 or>11,000 <4,000 or>11,000þbands

(>500)

Tracheal secretions Rare Abundant Abundant and purulent

Chest X-ray infiltrates None Diffuse Localized

PaO 2 /FiO 2 >240 or ARDS <240 and no ARDS

Microbiology Negative Positive

Nosocomial Pneumonia in Critical Care 187