The prosthetic dead space makes a significant contribution to the total dead space in low-birth-weight premature newborns receiving artificial ventilation in response to respiratory distress. Use of an endotracheal tube with capillaries molded into the tube wall enables washout of the dead space without insertion of a tracheal catheter. In 10 premature newborns (mean gestational age, 27.5 +/- 2.2 wk; mean weight, 890 +/- 260 g) receiving continuous positive-pressure ventilation (Paw = 12.7 +/- 1.8 cm H2O; FIO2 = 39 +/- 17%), tracheal gas insufflation (TGI) for CO2 washout was conducted using this technique. The flow of tracheal insufflation (0.5 L/min) was derived from the inspiratory line of the ventilator circuit and blown into the trachea. Intratracheal pressures showed little or no TGI-related modification ( < 1 cm H2O). A control system enabled TGI discontinuation in the event of a pressure rise. At constant ventilation pressure, PaCO2 decreased by 12.1 +/- 5.9 mm Hg (delta PaCO2 = -26 +/- 12%) under TGI, whereas PaO2 remained unchanged. While maintaining PaCO2 constant, peak inspiratory pressure (PIP) was decreased by 5.4 +/- 1.7 cm H2O (delta PIP = -22.0 +/- 8.3%). TGI showed immediate efficacy (PCO2 reduction of at least 5 mm Hg) in nine of the 10 newborns who then received chronic TGI (14 to 138 h). TGI appears to be an effective method, suitable for long-term clinical application, enabling a reduction in the aggressive nature of conventional ventilation.