The effect of a halothane/nitrous oxide/oxygen mask induction on arterial hemoglobin saturation (SaO2) was studied in 25 patients with cyanotic congenital heart disease. During this induction the SaO2 increased from 80% +/- 2 (mean +/- SEM) in awake patients breathing room air to 86% +/- 1 after induction of anesthesia, with an average increase of 6.8% +/- 2 (absolute) (N = 25; P < .005). Patients were retrospectively divided into two groups, A and B. Group A patients had pulmonic infundibular stenosis (PIS) (N = 13) with the potential for variable pulmonary outflow tract obstruction. These patients had awake SaO, values of 76% +/- 3 that increased significantly after induction to 86% +/- 1, with an average increase of 10% +/- 4 (absolute) (P < .01). Group B patients did not have PIS; they had an awake SaO2 of 83% +/- 2, which rose insignificantly after induction to 87% +/- 1, with an average increase of 4% +/- 2 (absolute) (P > .1). Thus, the significant increase in SaO2 during induction for the entire group was primarily due to the large response of patients with PIS. No correlation was found between SaO2 and systolic blood pressure for the entire study population as well as for groups A and B separately. The lack of correlation between systolic blood pressure and SaO2 may indicate that relaxation of pulmonic infundibular stenosis is an important mechanism explaining the observed increase in SaO2 during anesthetic induction in group A (PIS) patients. Six patients had clinically important decreases in SaO2 during the induction of anesthesia, and airway obstruction was usually the cause. Anesthetic induction with halothane/nitrous oxide/oxygen leads to an increase in SaO2 in patients with many forms of cyanotic heart disease, but especially those with the potential for variable pulmonary outflow tract obstruction.