Insufflation of carbon dioxide into the chest wound is used in open-heart surgery to de-air the heart and great vessels. In a cardiothoracic wound model, we compared the degree of air displacement achieved by a new insufflation device, a gas-diffuser, with that of a thin open-ended tube during steady-state and with carbon dioxide flows of 2.5, 5, 7.5, and 10 L/min. We also studied air displacement at the start of and after discontinuation of carbon dioxide insufflation with the gas-diffuser and evaluated the influence of an open pleura. During steady state, the gas-diffuser produced efficient air displacement in the wound cavity model at carbon dioxide flows of > or = 5 L/min (< or = 0.65% remaining air), whereas the open-ended tube was inefficient (> or = 82% remaining air) at all studied carbon dioxide flows (P < 0.001). An open pleural cavity prolonged the time needed to obtain a high degree of air displacement in the wound cavity (P = 0.001). Carbon dioxide insufflation of the cardiothoracic wound cavity should be initiated at a carbon dioxide flow of 10 L/min at least 1 min before the incision of the heart and great vessels and should be continued at a carbon dioxide flow of at least 5 L/min until surgical closure.