Between 1983 and 1995, 546 Blalock-Taussig shunt procedures were performed in 472 patients: 128 (23.0%) were classical shunts, 90 of them on the same side as to the aortic arch, and 418 (77.0%) were modified shunts, 182 on the same side of the arch. At the time of surgery, 78 patients were aged below one week, 270 from one week to 12 months, and 198 patients were over one year of age. The mean pre-operative arterial saturation (71.7%±16.5%) was significantly increased to 83%±17.9% imme-diately after the procedure (p=0.017). The overall hospital mortality rate was 2.9% (16/546), with rates of 2.3% (3/128) for the classical, and 3.1% (13/418) for the modified shunts (p= not significant). The rate was significantly higher, however, for classical shunts when the pulmonary arterial diameter was less than 4 mm (15.4% versus zero; p=0.047), though this relationship was reversed for modified shunts (zero versus 3.6%; p=0.338). Early mortality was significantly influenced by the age at surgery, 5/78 (6.4%) in patients aged below 1 week, 3.7% between 1 week and 1 year, and 0.5% over 1 year (p=0.019). Early mortality was also significantly increased in patients weighing 3kg or less, 8/156 (5.1%), versus 3/303 (1.0%), p=0.037. Overall, 51 shunts failed (9–3%), 10 early and 41 late. Early failure was significantly increased in patients weighing 3kg or less, 8/156 (5.1%) versus 3/303 (1.0%), p=0.016. The overall early failure rate was 1.4% (3/215) when heparin was administered intra-operatively and for 48 hours postoperatively, in contrast to an early failure rate of 3.4% (7/203) when heparin was not used (p = 0.294). Overall rates of failure during follow-up were 9.1% (17/188) in heparinized patients versus!3.6% (24/177), (p=0.173) in non-heparinized patients. Failure of classical shunts was 10.2% (13/128), compared with 6.7% (28/418) for modified shunts (p=0.195). Failure was more common overall if the pulmonary arterial diameter was less than 4 mm, 14.7% (9/61), as opposed to 8.7% (26/300) when the diameter was 4 mm or greater, (p=0.144). Administration of aspirin during follow-up after the modified shunt procedure reduced failure from 11 % (18/163) to 6.7% (10/150), p=0.176. Classical or modified Blalock-Taussig shunts, either on the same side or opposite to the aortic arch, can be performed on patients of any age with minimum postoperative complications and low operative mortality. The use of intra- and post-operative heparin appears to reduce the overall rate of failure, and the administration of aspirin during follow-up appears to reduce failure of modified Blalock-Taussig shunts.