BACKGROUND The clinical importance of coronary collaterals in the presence of obstructive coronary artery disease is not clearly defined. METHODS We retrospectively analysed the clinical and angiographic features of 100 patients with > or = 90% luminal diameter stenosis involving at least one major coronary artery. Coronary collaterals were graded 0 to 4 (Nitzberg's classification) and studied to determine their influence on clinical parameters. RESULTS Thirty patients had no collaterals (group I) and 70 showed collaterals (group II). There were no significant differences between groups I and II in age and sex distribution, prevalence of risk factors of coronary artery disease (hypertension, diabetes, smoking, hypercholesterolaemia), duration of symptoms of coronary artery disease and prior myocardial infarction. Groups I and II had similar types (left anterior descending 73% v. 71%; left circumflex 50% v. 50% and right coronary 37% v. 56%) and numbers of arteries involved (one 47% v. 41%; two 47% v. 40%; three 7% v. 19%). Group II had a significantly lower prevalence of rest angina (14% v. 47%, p = 0.002). This difference was also evident when the patients were re-classified according to the extent of flow through the collaterals. Those with good collateralization (Nitzberg grades 3 and 4) had a lower prevalence of rest angina (13%) compared to those with poor collateralization (Nitzberg grades 0 to 2; 35%, p = 0.02). CONCLUSIONS Coronary artery collaterals may reduce the incidence of rest angina in patients with obstructive coronary artery disease.