Because animal studies have demonstrated that mechanical ventilation at high volume and pressure can be deleterious to the lungs, limitation of airway pressure, allowing hypercapnia if necessary, is already used for ventilation of acute respiratory distress syndrome (ARDS). Whether a systematic and more drastic reduction is necessary is debatable. A multicenter randomized study was undertaken to compare a strategy aimed at limiting the end-inspiratory plateau pressure to 25 cm H2O, using tidal volume (VT) below 10 ml/kg of body weight, versus a more conventional ventilatory approach (with regard to current practice) using VT at 10 ml/kg or above and close to normal PaCO2. Both arms used a similar level of positive end-expiratory pressure. A total of 116 patients with ARDS and no organ failure other than the lung were enrolled over 32 mo in 25 centers. The two groups were similar at inclusion. Patients in the two arms were ventilated with different VT (7.1 +/- 1.3 versus 10.3 +/- 1.7 ml/kg at Day 1, p < 0.001) and plateau pressures (25.7 +/- 5. 0 versus 31.7 +/- 6.6 cm H2O at Day 1, p < 0.001), resulting in different PaCO2 (59.5 +/- 15.0 versus 41.3 +/- 7.6 mm Hg, p < 0.001) and pH (7.28 +/- 0.09 versus 7.4 +/- 0.09, p < 0.001), but a similar level of oxygenation. The new approach did not reduce mortality at Day 60 (46.6% versus 37.9% in control subjects, p = 0.38), the duration of mechanical ventilation (23.1 +/- 20.2 versus 21.4 +/- 16. 3 d, p = 0.85), the incidence of pneumothorax (14% versus 12%, p = 0. 78), or the secondary occurrence of multiple organ failure (41% versus 41%, p = 1). We conclude that no benefit could be observed with reduced VT titrated to reach plateau pressures around 25 cm H2O compared with a more conventional approach in which normocapnia was achieved with plateau pressures already below 35 cm H2O.