OBJECTIVE--To study electrocardiographic changes in infants and children in whom topical cooling was used during surgical repair of congenital heart defects. DESIGN--A retrospective study of all patients who had surgical repair of congenital heart disease during cold blood cardioplegia and topical cooling from January to August 1990. Eleven patients (group 1) had topical cooling with ice and 15 (group 2) with cold saline. PATIENTS--All 36 paediatric patients operated on during this period. All the available electrocardiographic records were analysed. Ten patients in whom reliable records were not available were excluded. Twenty six patients entered in this retrospective study. INTERVENTIONS--Topical cooling with ice or with a slush of cold saline. MAIN OUTCOME MEASURE--Recordings from all the precordial leads were examined and scored as the sum of the maximum ST elevation (mV) in each precordial lead. The score obtained for each electrocardiogram was recorded together with the timing of the electrocardiogram (preoperative, arrival in intensive care unit immediately after surgery, postoperatively in the intensive care unit, and at discharge). RESULTS--There were no differences between the two groups in terms of demographic data, diagnosis, duration of ischaemia, and postoperative myocardial performance. There was temporary ST elevation during the first 48 postoperative hours in all the children in group 1 but in only seven of the 15 children in group 2 (Fisher's test, p less than 0.005). The mean (SD) score for maximum ST elevation was 1.34 (0.83) mV in group 1 and 0.52 (0.64) mV in group 2 (Student's t test, p less than 0.01). CONCLUSIONS--These temporary electrocardiographic changes in the presence of adequate myocardial performance were attributed to epicardial damage induced by hypothermicosmotic injury. The use of ice for topical cooling may damage the epicardium in children.