The Role of CT Chest in Diagnosing Ventilator Associated Pneumonia

Background: Ventilator Associated Pneumonia (VAP) remains a challenging diagnosis. The role of Computed Tomography (CT) of the chest has not been well defined. We performed a retrospective case-control study that described the findings of CT Chest in patients diagnosed with VAP.

Objective: To identify CT chest findings associated with VAP.

Methods: Cases were defined as patients who were diagnosed with VAP in 2007 and 2008 in medical and surgical intensive care units (ICU), according to the CDC guidelines. Patients were included if CT chest was done within 7 days of the diagnosis or intubation. All patients who were intubated without VAP diagnosis, negative or absent respiratory cultures, and Clinical Pulmonary Infection Scores (CPIS) <7 who had a CT chest within 7 days of intubation were selected from the imaging database as controls. Two board-certified radiologists were blinded to review the CT scans and subsequently classified findings into predefined radiologic descriptions. Agreement between radiologists reading was measured. Statistics were performed with JMP 8.0 (SAS, Cary, NC).

Results: 44 cases of VAP occurred in 2007 and 2008. The median age was 63 years (range 19-88), 68% were male, and 86% were Caucasian. P. aeruginosa (39%), Klebsiella spp (18%), and S. aureus (14%) were the most common pathogens. In 54 controls, median age was 65 years (range 17-94), 56% were male, and 70% were Caucasian. In 63% (34/54) of controls a pulmonary embolism protocol was ordered. Hospital mortality was 48% vs. 44% (p=NS), and median CPIS scores were 7 vs. 4 (p<0.001), amongst cases vs. controls, respectively. Median hospital (43 vs. 20 days, p<0.001) and ICU stay (31 vs. 16 days, p<0.01) were significantly longer in cases when compared to controls. Presence of consolidation or infiltrate on CT chest was not associated with VAP. In contrast, presence of lymphadenopathy (OR 6.3, p=0.004), unilateral pleural effusion (OR 2.9, p=0.03), central nodules (OR 2.7, p=0.05), and ground glass opacities (OR 4.5, p<0.001) were associated with VAP. The presence of none of these findings on the readings of both radiologists had a sensitivity of 95%, and a negative predictive value of 92% for the absence of VAP. However the presence of one or more of these findings had only limited specificity of 45% and positive predictive value of 58% for VAP. Moreover, the agreement for some of these findings was variable (central nodules κ=1, unilateral effusion κ=0.61, ground glass opacities κ=0.51, lymphadenopathy κ=0.44).

Conclusions: Central nodules, unilateral effusion, lymphadenopathy, and ground glass opacities were significantly more common in patients with VAP compared to control patients. These findings suggest that certain CT characteristics may be useful as an adjunct in the diagnosis of VAP. Prospective validation is needed to determine the test characteristics of CT chest in the diagnosis of VAP.