When it comes to diagnosing ventilator-associated pneumonia, there is little consensus in the literature because of different interpretation of the clinical signs and symptoms suggestive of lung infection, variable differentiation between colonization and infection of the lower respiratory tract, and poorly specified prior antibiotic use in the ICU. As clinical signs are nonspecific and subjective, no single clinical criterion has been specifically diagnostic for VAP. Fever, tachycardia, and leukocytosis are all relatively nonspecific. Chest radiographs are most helpful when normal, as they can rule out pneumonia. Air bronchograms or alveolar opacities on a chest radiograph in patients without acute respiratory distress syndrome (ARDS) are most
correlated with pneumonia and have a 68% diagnostic accuracy. When chest radiographs have been compared with subsequent autopsy findings, a localized infiltrate was 87% sensitive for VAP but only 25% specific. Furthermore, approximately 10% of the cases of pneumonia failed to have a new or worsening infiltrate at all.

The clinical pulmonary infection score (CPIS) was developed in an attempt to find more reliable criteria for the diagnosis of VAP. The CPIS combines clinical, radiographic, physiological, and micro-biologic data into a single numerical result. The score is derived by awarding 0, 1, or 2 points for each of seven variables: temperature; white blood count; volume and quality of tracheal suctions; oxygenation; chest radiograph findings; and semiquantitative culture of tracheal aspirate. When the CPIS exceeds 6, good correlation with the presence of pneumonia, as defined by quantitative cultures of bronchoscopic and nonbronchoscopic lavage specimens, has been found. When compared with postmortem quantitative lung cultures as the reference standard, the CPIS has a sensitivity of 77% and a specificity of 42% for VAP.

Increased specificity at the cost of sensitivity over the CPIS may be obtained by using diagnostic criteria of a radiographic infiltrate and at least two of the three following clinical features of VAP: fever greater than 38C; leukocytosis or leukopenia; and purulent tracheal secretions. Use of these criteria resulted in a 69% sensitivity and a 75% specificity for VAP, which represents the most accurate clinical criteria for starting empiric antibiotic therapy.

Once the diagnosis of VAP is being considered, a lower respiratory tract sample for culture to guide therapy should be collected before antibiotic administration. If cultures are obtained after initiation of antibiotics, a 40% false negative rate ensues. If there is a high pretest probability of pneumonia, or in the 10% of patients with evidence of sepsis, prompt therapy is required, regardless of whether bacteria are found on microscopic examination of lower respiratory tract samples.