In ventilated preterm infants the flow sensor contributes significantly to the total apparatus dead space, which may impair gas exchange. The aim of the study was to quantify to which extent a dead space reduced Kolobow tube (KB) without flow sensor improves the gas exchange compared with a conventional ventilator circuit with flow sensor [Babylog 8000 (BL)]. In a cross-over trial in 14 tracheotomized, surfactant-depleted (saline lavage) and mechanically ventilated newborn piglets (age <12 h; body weight 705-1200 g) BL and KB was applied alternately for 15 min and blood gases were recorded. The inner diameter of the endotracheal tube was 3.6 mm and the apparatus dead space of BL and KB including the endotracheal tube were 3.0 and 1.34 mL. Despite a 50 % apparatus dead space reduction with KB compared to BL statistically significant improvements were only observed for body weights <900 g. In this weight group median paCO2 was decreased by 5 mmHg (p < 0.01), whereas the improvement decreased with decreasing baseline paCO2. Furthermore, median paO2 was increased by 4 mmHg (p < 0.05) and O2 saturation was increased by 2.5 % (p < 0.05). No significant changes were seen in the circulatory parameters. In very small, ventilated lungs the use of KB improved the gas exchange; however, the improvement was moderate and does not justify the waiving of volume monitoring.