Aims/hypothesis. Currently, three categories of measures are used to assess cardiovascular autonomic dysfunction: measures of the Ewing-test, measures of heart-rate variability, and measures of baroreflex sensitivity. We studied the determinants of these measures obtained from cardiovascular autonomic function tests in the Hoorn Study. Methods. The study group (n = 631) consisted of a glucose-tolerance-stratified sample from a 50- to 75- year-old group of people. Cardiac cycle duration (RR interval) and continuous finger arterial pressure were measured under three conditions: during (a) spontaneous breathing, (b) six deep breaths over one minute, and (c) an active change in position from lying to standing. From these readings, ten measures of autonomic function were assessed (three Ewing, six heart-rate variability and one baroreflex sensitivity). As possible determinants we considered age, sex, glucose tolerance, cardiovascular disease, use of anti- hypertensive drugs, anthropometric factors, metabolic factors and lifestyle factors. Results. Multivariate analysis showed that eight of ten cardiovascular autonomic function measures were most strongly associated with glucose tolerance. Furthermore, measures were moderately associated with age, sex, waist-to-hip ratio, use of anti-hypertensive drugs, and insulin. The measures were weakly associated With coronary artery disease but not with lipids. The strongest determinants seemed to differ between subjects with and without diabetes: in the non-diabetic subjects the most strongly associated were age and use of anti-hypertensive drugs and in subjects with diabetes, insulin. No consistent differences in association between the three categories of measures were observed. Conclusion/interpretation. The strongest determinants of autonomic function were age, presence of diabetes and use of anti-hypertensive drugs.