The availability of ambulatory ECG monitoring allows identification of transient myocardial ischaemia, the clinical relevance of which is currently being investigated. Ninety-four consecutive patients with ischaemic heart disease and a positive exercise test (greater than or equal to 1 mm ST-segment depression) were studied to evaluate the prevalence of transient myocardial ischaemia (either painless or painful) during 24-h dynamic electrocardiogram (ECG) and the clinical, angiographic and ergometric variables predicting its appearance. Two-hundred-and-eighty-one episodes of transient electocardiographic myocardial ischaemia were recorded in 69 patients (73.4% of all patients). Transient myocardial ischaemia was more frequent, although not significantly so, in patients with diabetes, with previous myocardial infarction, or with multivessel disease. When tested by multivariate analysis, neither the clinical variables nor the severity of coronary artery disease allowed prediction of the occurrence of transient myocardial ischaemia during dynamic ECG. The duration of exercise testing up to the ischaemic threshold (ST-segment depression = 1 mm) and the peak heart rate during exercise were more accurate predictors of transient myocardial ischaemia (P = 0.019 and 0.012 respectively). Patients with transient myocardial ischaemia had a lower ischaemic threshold (355 +/- 175 vs 498 +/- 150 s, mean +/- SD, P = 0.001) despite a lower peak heart rate (129 +/- 18 vs 137 +/- 12 beats min-1, P = 0.047) than patients without transient myocardial ischaemia. In conclusion, exercise testing may help select patients for examination by dynamic ECG.(ABSTRACT TRUNCATED AT 250 WORDS)