BACKGROUND: Our purpose was to determine the repeatability, after 2 weeks, of frequency domain measures of heart rate variability (HRV) during simple cardiovascular reflex tests. METHODS: Twenty healthy volunteers aged 29.3 +/- 2.5 years were assessed twice (at weeks 0 and 2). Continuous electrocardiogram and minute-to-minute blood pressure were recorded during spontaneous and metronome-paced breathing (0.2 Hz). Under paced breathing, two tests were performed: 1) active change of posture: 5 min supine position, 5 min seated upright, and 5 min standing up, and 2) cold pressor test: the right hand was immersed in cold water (5 degrees C) for 2 min. RESULTS: Paced breathing elicited a significant increase of the high-frequency (HF) component of HRV. This effect was repeatable on 95% of the subjects. Active change of posture induced a significant increase of the heart rate with an increase of the low-frequency/high-frequency ratio of HRV. Although repeatability was better for the response to being seated upright than for the response to being standing up, it was always higher than 90%. The cold pressor test induced a significant increase of the heart rate and blood pressure, but with variable changes on the HRV measurements (either a decrease or an increase). Repeatability of responses was evident for 95-100% of the subjects. Although repeatability of HRV measurements in the same subject during the tests was higher than 95%, coefficients of repeatability reflected large differences among the subjects. CONCLUSIONS: The results suggest that it is not advisable to use isolated HRV changes to interpret the response to simple cardiovascular reflex tests in groups of healthy subjects.