Prospective comparative study on etiologic diagnosis of ventilator-associated pneumonia

Ventilator-associated pneumonia (VAP) is a pulmonary infection highly prevalent and lethal. The clinical diagnostic is the most used, even though it is not specific, and the collection of the biological material, fundamental for the handling of these patients, does not have a definite gold standard method. The present study aimed to evaluate different methods, invasive or noninvasive, on VAP etiologic diagnosis. It is a prospective, observational study carried on in a tertiary hospital. Thirty (30) patients with a clinical VAP suspicion were submitted to the endotracheal aspirate (ETA), bronchoscopic bronchoalveolar lavage (BAL) and bronchoalveolar lavage non-bronchoscopic (BALNB), being patients its own controller. The measures of accuracy, sensibility, specificity, positive predictive value (PPV) and negative predicted value (NPV) for the ETA were 83%, 85%, 50%, 96% and 20% respectively. For the BALNB, the following were observed for accuracy, sensibility, specificity, PPV and PNV: 50%, 50%, 50%, 93% and 7% respectively. In relation to the BAL, the accuracy, sensibility, specificity, PPV and PNV found were 63%, 60%, 100%, 100% and 15%. The agreement between the isolated microorganisms in the BAL with the ones obtained with the BALNB and ETA was good, with kappa coefficient of 0.63 and 0.64 respectively, being the Gram negative bacteria the main etiological agents. Comparing invasive and non-invasive methods, we concluded that the ETA represents the best option to the etiologic confirmation of the VAP.