The appropriate treatment of ventilator-associated pneumonia (VAP) must be based on accurate diagnosis, which can be done by microbiological examination of the samples obtained from the respiratory tract by nonbronchoscopic bronchoalveolar lavages (NB-BAL). This study was designed to determine the effectiveness of NB-BAL in diagnosing VAP in newborns. Two hundred and seven NB-BAL samples were obtained from 145 intubated neonates for microbiologic and cytologic evaluation of the distal airway. The NB-BAL samples were processed for microscopic quantification of the polymorphonuclear cells (PMN) containing intracellular bacteria (ICB) and quantitative culture (positive threshold, 10(5) cfu/ml). VAP was defined as a new, progressive, or persistent (>24 hrs) infiltrate on the chest radiograph, with two or more of the following criteria: (a) macroscopically purulent tracheal secretions, (b) fever or hypothermia, (c) leukocytosis or leukopenia, and (d) worsening of respiratory status with a Pa O2/F IO2 ratio of <240. Colonization was defined as mechanical ventilation for more than seven days, no signs of infection, and isolation of the same bacteria species in two previously obtained NB-BAL samples. Of the 145 neonates, 40 (27.5%) were infected and 12 (8.3%) were colonized. Forty-four patients (30%) developed VAP according to diagnostic categories based on clinical and radiologic criteria. Forty newborns with VAP (90%) had positive NB-BAL culture. The sensitivity, specificity, and positive and negative predictive values of NB-BAL fluid culture for VAP diagnosis were 90%, 90%, 70%, and 97%, respectively. The percentage of ICB was significantly higher in newborns with VAP. The presence of ICB in 2% or more on Giemsa-stained smears corresponded to a sensitivity of 94%, specificity of 83%, positive predictive value of 94%, and negative predictive value of 83%. The sensitivity and specificity of combination of ICB and NB-BAL quantitative culture in diagnostic samples were 94% and 90%, respectively. The positive and negative predictive values were 71% and 98%. In our study, the presence of leukocytes in the NB-BAL fluid smear of infants with VAP was higher than that of the colonized babies (84%, 26%). This difference was statistically significant (p < 0.0001). The sensitivity and specificity of PMNs in NB-BAL fluid for the diagnosis were 86% and 75%, respectively, and the positive and negative predictive values were 89% and 69%. We conclude that NB-BAL lavage is well tolerated and clinically useful in mechanically ventilated newborns. These results suggest that NB-BAL fluid microscopic examination and cultures can offer a sensitive and specific means to diagnose VAP in newborns and may provide relevant information about the causative pathogens.