We prospectively assessed coronary artery disease and natural history in a series of 104 patients (99 males, mean age 52 +/- 8 years) admitted with recent onset angina (defined as a history of angina of less than 1 month duration). Coronary angiography showed one-vessel disease in 31, two-vessel disease in 22 and three-vessel disease in 14; 37 patients had normal coronary arteries. After a mean follow-up of 36 (range 1 to 52) months, one patient died, 13 sustained a myocardial infarction and 21 underwent surgery. The univariate analysis showed four of 12 clinical features derived from clinical history and data from CCU (exertional angina (P less than 0.001), and recurrence of angina (P less than 0.05)) to be associated with the presence of coronary artery disease. These clinical features were termed clinical risk characteristics. Three-year probability of medical events (death or acute myocardial infarction) for patients with 0-1 clinical risk characteristics was 0 and that of combined events (need for revascularization with or without a preceding medical event) 0.11, whereas patients with 2 or more risk characteristics had probabilities of 0.27 and 0.49, respectively. Multivariate analysis identified the number of clinical risk characteristics as the only independent predictor of medical events (P less than 0.002) and a positive thallium stress test (P less than 0.0001), the number of clinical risk characteristics (P less than 0.002) and the number of involved arteries (P less than 0.002), as independent predictors of combined events.(ABSTRACT TRUNCATED AT 250 WORDS)