Background: Cardiac surgery is associated with an important risk of central or peripheral organ damage, attributed in part to air embolism from incompletely deaired cardiac chambers. Insufflation of carbon dioxide into the thoracic cavity is widely used for organ protection in cardiac surgery. Methods: In patients operated on through a sternotomy, the gas was insufflated through a standard cardioplegia line (group I, n = 10) or a Jackson-Pratt drain (group II, n = 10), with flow rates of 2, 4, and 6 L/min. In patients undergoing mitral valve surgery through a right anterolateral minithoracotomy, application through a gas port (group III, n = 10) was compared with application through a Veress needle (group IV, n = 10). In groups I and IV measurements were repeated with a gauze sponge to divert the gas stream. Results: At a flow of 2 L/min, carbon dioxide levels in the thoracic cavity reached 52% ± 30% in group I and increased to 81% ± 27% when a gauze sponge was used. In group II a level of 91% ± 5% was achieved. In minimally invasive procedures carbon dioxide levels reached 92% ± 6% in group III and 60% ± 25% in group IV without a gauze sponge and 97% ± 2% in group IV with a gauze sponge. Increasing flow rates from 2 to 6 L/min decreased carbon dioxide levels in the thoracic cavity. Arterial blood gas analysis did not reveal critical levels of partial pressure of carbon dioxide at any time. Conclusions: For optimized carbon dioxide concentrations during cardiac procedures, jet effects in the thoracic cavity have to be avoided. The highest levels were achieved with infusion lines covered by a gauze sponge or a perforated drain for conventional operations and a sponge-covered Veress needle or a gas port for minimally invasive approaches.J Thorac Cardiovasc Surg 2002;124:387-91