I'r t-L~S LONe BEEN KNOWr~ that intemaittent positive pressure breathing (n,PB) decreases cardiac output because the pumping effect of normal inspiration, which augments venous return, is lost when im'B increases mean intrathoraeie pressure? In addition, if IPPB produces hyperventilation, the resultant fall in Paco., (arterial carbon dioxide tension) will further reduce cardiac output ~,~ and facilitate the induction of cardiac arrhythmias. 4-~ These latter effects of acute respiratory alkalosis are magnified in the digitalized patient because of acute secondary hypokalaemia with production of digitalis toxicity. ~-~ Hypocarbia also decreases cerebral blood flow x~ and shifts the oxygen dissociation curve to the left, impairing release of oxygen from haemoglobin to all tissues. These effects of hyperventilation, although potentially serious in all patients, are most threatening in patients with limited cardiac reserve and enhanced myocardial irritability who are undergoing cardiac surgery. However, it is in just these patients that hyperventilation is often necessary to maintain adequate oxygenation because of the association of ventilation-perfusion imbalance and intrinsic pulmonary pathology, secondary to the underlying cardiae disease. This investigation was initiated to seek a simple empirical solution to the problem. METHOD This study was carried out in 110 consecutive patients undergoing open or closed heart surgery for acquired cardiac disease. They varied in age from 17 to 78 years and in weight from 39 to 92 kg. All were premedieated with intramuscular injections of secobarbital and a belladonna drug 1~6 hours prior to induction of anaesthesia. They were placed in semi-Fowler's position on the operating table and arterial and venous cut-downs were performed under local analgesia. Either the brachial or radial artery was used for blood sampling and direct recording of blood pressures. Arterial blood samples were drawn prior to induction of anaesthesia and one hour thereafter. Pao,_,, pH, bicarbonate, Paco2 and base deficit or base excess were determined by the Astrup method. 1~ Following a sleep dose of thiamylal sodium (200 to 400 mg) and oxygen by mask, a single dose of sueeinylcholine (40 to 100 mg) was given to facilitate endotraeheal intubation, Anaesthesia was maintained with a 6-liter flow of nit