Objective: To evaluate the diagnostic value of procalcitonin (PCT) in ventilator-associated pneumonia (VAP).
Methods: Prospective study was performed. All adult patients who were admitted into the intensive care unit (ICU) of West China Hospital of Sichuan University between June 1st and October 1st and were clinically suspected of having developed VAP after 48 hours of mechanical ventilation were enrolled. Patients who had active infection or lung cancer at ICU admission or developed extrapulmonary infection during the study period were excluded. PCT and C-reactive protein (CRP) levels were measured and the clinical pulmonary infection score (CPIS) was calculated at study entry and on the day of VAP suspicion.
Results: In total, 49 suspected episodes of VAP in 31 cases were microbiologically confirmed in 23 and refuted in 26. Median PCT levels were 0.68 microg/L (interquartile range 0.28, 2.31) and 0.18 microg/L (0.06, 0.28) respectively in patients with and without VAP on the suspicion day (P<0.01). Using 0.31 microg/L as the best cutoff, PCT had 73.9% sensitivity and 80.8% specificity. The CPIS resulted in higher sensitivity (95.7%) but lower specificity (53.8%) when the cutoff of CPIS > or = 5. CRP had the lowest sensitivity and specificity (56.5%, 61.5%) when the cutoff of CRP was 109.5 mg/L. A CPIS > or = 5 combined with serum levels of PCT > or = 0.31 microg/L did not improve the sensitivity (69.6%), but resulted in 88.5% specificity.
Conclusion: PCT had better specificity than CRP and CPIS in early diagnosis of VAP. Combined values of PCT and CPIS below the cut-off points could effectively exclude false-positive diagnosis of VAP.