The predictive value of blood pressure (BP), heart rate (HR), and catecholamines in terms of any subsequent development of cardiovascular disease was investigated. Systolic (S) and diastolic (D) BP, HR, epinephrine (E) and norepinephrine (NE) were measured three times a year in 1980, 1984, and 1989 on 20 clinically healthy subjects, 18 patients with 'essential hypertension', and 22 patients with angina pectoris. Of the 22 patients in the latter group, 15 died during a 2-year follow-up (1990)(1991). Each individual data series was analyzed by single cosinor to assess the circannual variation. Results were summarized by population-mean cosinor for each group. Parameter tests were used to compare the circannual rhythm characteristics among the different patient groups. A circannual rhythm was invariably demonstrated on a group basis (P < 0.05). Differences in MESOR and/or circannual amplitude were found among the different groups. In particular, patients with angina pectoris who will die within the 2-year follow-up differ in terms of their E and NE from all other patient groups, a difference already detected at the beginning of the study, more than 10 years before they die. A similar separation is not achieved in terms of BP or HR.