Ventilator-associated pneumonia (VAP) is difficult to detect and is often unsuspected during adult respiratory distress syndrome (ARDS). We prospectively evaluated lower respiratory tract (LRT) colonization and infection in 30 patients with severe ARDS (PaO2/FIO2 ratio < 150 mm Hg), using repeated quantitative cultures of plugged telescopic catheter (PTC) specimens taken blindly via the endotracheal tube every 48 to 72 h after onset of ARDS. All patients except one were receiving antibiotics. When VAP was suspected on the presence of clinical criteria for infection, a repeated PTC and, when possible, a bronchoalveolar lavage (BAL) were obtained before any new antimicrobials were administered; samples growing > or = 10(3) cfu/ml (PTC) or > or = 10(4) cfu/ml (BAL) were considered diagnostic of infection. Twenty-four VAP episodes were diagnosed in 18 patients (60% of patients or 4.2/100 ventilator-days) a mean of 9.8+/-5.7 d after onset of ARDS. Eighteen LRT colonization episodes were recorded; 16 of 24 (66%) VAP episodes were preceded (by 2 to 6 d) by LRT colonization with the same organism(s), and only two episodes of colonization were not followed by VAP. We conclude that although VAP is of relatively late-onset during severe ARDS, its incidence is much higher than in other conditions and can be underestimated. Lower airways colonization is consistently followed by infection with the same organisms and precedes VAP in two thirds of episodes. Repeated protected specimens taken blindly may provide a useful means to predict infection and therefore allow early antimicrobial therapy in high-risk patients with diffuse lung injury.