OBJECTIVE: We investigated the relation of the coefficient of variation of the R-R intervals in diabetic patients and healthy subjects, and the cardiovascular responses during exercise. RESEARCH DESIGN AND METHOD: The subjects were 8 males with type 2 diabetes mellitus, and 8 age-matched healthy nondiabetic males (healthy controls). All subjects performed an incremental test on a bicycle ergometer at a work rate of 15 Watt・min-1 to determine the maximum oxygen uptake and to calculate ventilatory threshold. Subsequently, on another day, they all performed a constant-load exercise at a work rate chosen to elicit an oxygen uptake equivalent to 800f their individual the ventilatory threshold. The coefficient of variation of the R-R intervals (CVR-R) was calculated from the electrocardiogram at rest and used as an index of autonomic nerve function. The time constant (tau) and steady state level of the kinetic response for oxygen uptake (VO2) and heart rate (HR) were calculated using a mono-exponential model under constant- load exercise. RESULTS: 1) CVR-R andVO2max were significantly lower in the diabetic patients than in the healthy controls. Tau HR and tauVO2 were significantly longer in the diabetic patients than in the healthy controls. 2) TauVO2 and tau HR were significantly positively correlated in both groups, whereas CVR-R and tau HR, and tauVO2 were significantly negatively correlated in both groups. 3)VO2max and tau HR, and tauVO2 were significantly negatively correlated only in the healthy controls. CONCLUSIONS: These results suggest that CVR-R may reflect the characteristics of cardiovascular responses to exercise. However, though some previous studies have postulated that these characteristics of cardiovascular responses to exercise reflect cardiorespiratory endurance in healthy controls, our study confirmed no such correlation in diabetic patients.