Between January 1985 and March 1990, 66 children with the tetralogy of Fallot underwent 85 cardiac catheterization procedures. The mean age at first procedure was 2.5 years (range 0.1–;14.4 years) and the mean weight was 10.4 kg ( range 2.4–36.0 kg). Diagnostic cardiac catheterization was performed in 60 procedures and balloon dilatation in 25. Hypercyanotic spells had occurred prior to 24 (28%) of the procedures (all the patients being on propranolol) and a systemic-to-pulmonary arterial shunt had been constructed before 28 (33%) procedures. Of the procedures, 54 (64%) were performed under local and 31(36%) under general anesthesia. The pulmonary trunk was entered antegradely in 52 procedures, retrogradely through a shunt in 6 and not entered in 27. Balloon dilatation was performed under general anesthesia on 25 occasions. No procedure was abandoned because of a cyanotic spell. Nine (11%) spells occurred during 86 procedures, one of the procedures being postponed because ofa spell occurring after premedication, the procedure and not, therefore, continuing to catheterization. Five spells occurred before the catheter was positioned in the heart, 2 spells occurred during catheterization. Of the spells, eight occurred during procedures in children who had not had previous shunts. Antegrade entry into the pulmonary trunk in the group with shunts was associated with no spells compared with 6/38 (18%) in the group not having undergone surgery (p<0.l). In the group not undergoing surgery, when the pulmonary trunk was not entered, 1 (5%) spell occurred during 19 procedures compared with 6/38 (18%) when the pulmonary trunk was entered (p<0.4). The only clinically important factor which significantly influenced the incidence of spells was the use of general anesthesia, which was associated with 6/31 (19%) spells compared with local anesthesia which was associated with 2/54 (4%) spells (p<0.026).