The outcome of 121 children who underwent palliative cardiac surgery at the Red Cross War Memorial Children's Hospital over a 5 year period, 1980 1984, was retrospectively examined. 79 children had systemic artery to pulmonary artery shunt operations (SPS), 40 had pulmonary artery bands (PAB) and 2 had surgical septectomies. SPS was most often done for children with Tetralogy of Fallot (TOF, 26 cases) or complex univentricular hearts with right ventricular outflow tract obstruction (27 cases). PAB was done chiefly for ventricular septa! defects, alone (VSD, 8 cases) or with coarctation of the aorta (9 cases). Children were referred from a wide area with 63 cases being referred from other major centres and foreign countries. Overall, 36 children died (30 % mortality): 5 died at surgery, 6 within 48 hours of surgery, a further 5 within 31 days; and 20 died after 31 days. SPS and PAB had the same early mortality rates ( 13 % ) • SPS had higher late and overall mortality rates (20 and 33 %) than PAB (10 and 23 %). Age at operation was found to be the most significant determinant of the overall mortality rate: children less than six months had a mortality of 42 % and those over 6 months, 13 % • The children were grouped into those with lesions which were probably correctable and those that were unlikely to be so, based on diagnosis and age at surgery: those with correctable lesions had a lower overall mortality (22 %) than those with uncorrectable lesions (43 %). Where the surgery was performed as an emergency, there was a higher overall and early mortality (55 and 35 respectively), compared to those operations which were performed electively ( 25 and 9 % ) • The presence of other medical conditions, for example congenital abnormalities and infections, was also a determinant of death (44 % mortality if other medical condition present, 26 % if absent). sex, population group, home address and type of surgery performed did not significantly affect mortality when examined by multivariate analysis. Using routine methods of follow up, it was initially thought that 17 % of all patients (22 % of survivors) were lost to follow up. An important determinant of this was the referral centre. 31 % of cases from other major centres and 20 % of foreign cases were lost, as compared to 8 % of cases from smaller towns near Cape Town and 2 % of children from Cape Town. Population group (35 % Blacks, 14 % Coloureds and 7 % Whites were lost), and palliative operation (23 % SPS, and 5 % PAB lost) were also significant determinants. It was possible to trace 12 of the 20 children who were thought to be lost to follow. 8 had died, 3 were still awaiting correction and 1 was traced and received corrective surgery. The records of the children who underwent cardiac surgery in 1987 were also analysed. There was no difference in the demographic characteristics of either group, and the early mortality was the same. This study shows that the outcome after palliative cardiac surgery is poor, with a high mortality rate and children often being lost to follow up. The decision to palliate rather than to correct a congenital heart defect must be made after balancing these risks with those of early correction for the particular surgical team. Should palliative surgery be undertaken, careful follow up is essential to ensure that complications of palliation do not set in and that corrective surgery is done at the optimal time.