The effects of age and gender on heart rate variability as measured by spectral and time domain analysis of 24 h ECG recordings were evaluated in 101 healthy subjects, 49 men and 52 women (20-69 years of age). In the frequency domain, total power, very low-frequency power, low-frequency power and high-frequency power were negatively correlated to age (P < 0.001 for all variables). Total power decreased by 30% between 20-29 and 60-69 years of age. In the time domain, SDNN-index, the mean of the standard deviations of all normal R-R intervals for all 5 min segments of a 24 h ECG recording, was negatively correlated to age (P < 0.001). Total power, very low-frequency power, low-frequency power and the low-frequency/high-frequency ratio were lower in women (P < 0.05, P < 0.05, P < 0.01 and P < 0.01), although the absolute differences were much smaller than for age. There was a pronounced circadian variation; at night total power increased in all age groups (P < 0.01). The results show that age, and to a lesser degree gender, are important determinants of heart rate variability in healthy subjects. Heart rate variability is a valuable tool for risk stratification in cardiovascular disease, but the physiological effects of ageing, with diminishing heart rate variability in older age groups, must also be taken into account.