In most ventilators used in anaesthesia tidal volume delivered during mechanical ventilation is different from the tidal volume preset at the respirator on the basis of respirator and circuit compliance and gas compression during inspiration. The error in ventilation due to the compressed volume is especially significant clinically when the tidal volume is very small or when the airway pressure is very high. In newborns and neonates in particular, decreasing lung compliance during a surgical procedure may contribute to marked hypoventilation. We therefore investigated ventilation in newborn piglets during decreasing lung compliance induced by tension pneumothorax. We used the anaesthesia ventilator CICERO (Dräger, Lübeck, Germany) and the SERVO 900 C ventilator (Siemens-Elema, Sweden). MATERIALS AND METHODS. Two anaesthesia ventilators, the CICERO (group I, n = 8) and the SERVO ventilators (group II, n = 8) were investigated following randomized selection in a group of 16 newborn piglets (Table 1). After normoventilation for 60 min a tension pneumothorax at +10 mbar was induced. After 15 min the pneumothorax was increased to +20 mbar and maintained at this level for the rest of the study. When hypercapnia (PaCO2 > 45 mmHg) resulted, the respiratory rate was increased by +10/min after 15 min with pneumothorax at +20 mbar. When hypercapnia continued, the respiratory rate was increased again 25 min and if necessary also 35 min after the induction of pneumothorax at +20 mbar. After normoventilation for 60 min (T1) (Table 2), after 15 min with pneumothorax at +10 mbar (T2) and after 15 min (T3), 25 min (T4), 35 min (T5) and 45 min (T6) with pneumothorax at +20 mbar the following parameters were obtained: central venous (CVP) and mean arterial pressure (MAP), heart rate (HR), arterial (PaCO2) and end-tidal CO2 tension (PetCO2), peak inspiratory pressure (PIP), respiratory frequency (RF) and expiratory tidal (Vtex) and minute volume (VE). RESULTS. In group I the pneumothorax resulted in a significantly smaller increase in PaCO2 (43.3 +/- 6.2 mmHg) than in group II (Fig. 1), and hypercapnia was present in only 3 piglets. Vtex (Fig. 2), VE (Fig. 3) and PIP (Fig. 5) increased significantly, with significantly higher values than in group II, while PetCO2 (Fig. 6) decreased significantly. In group II the pneumothorax was attributed to a significant increase in PaCO2 and a marked hypercapnia in all piglets (PaCO2 61.2 +/- 5.9 mmHg) (Fig. 1). Vtex (Fig. 2) and VE (Fig. 3) remained unchanged, while PIP (Fig. 5) and PetCO2 (Fig. 6) increased. Following the increase in RF (Fig. 4) in all piglets, Vtex and VE increased and PaCO2 and PetCO2 decreased. CONCLUSIONS. During ventilation of neonates with the SERVO ventilator a decrease in lung compliance will cause hypoventilation and hypercapnia. This reflected by an increase in peak inspiratory pressure and can be corrected by increasing the respiratory rate. In contrast, the CICERO is able to preserve ventilation by an internal correction for gas compression, but it does not guarantee normoventilation in all cases. In neither group does the end-tidal PCO2 reflect the true ventilation during decreasing lung compliance, so that arterial blood gas analysis seems to be mandatory for the diagnosis of hypercapnia in such situations.