Spectral analyses of blood pressure and heart rate oscillations are increasingly used to assess the influences of diseases and drugs on the autonomic nervous system. Such influences can only be interpreted in view of the spontaneous variability of these oscillations. We therefore studied the reproducibility of power spectral analyses of blood pressure and heart rate fluctuations measured by a non-invasive finger plethysmographic method in 24 healthy volunteers. 2. Intra-observer reproducibility was assessed from measurements obtained on 3 consecutive days and 1 month later in each subject. Inter-observer reproducibility was assessed by comparing measurements made by two observers on one occasion. 3. There was no significant difference in standard haemodynamic and spectral analysis parameters (low frequency: 60-130 mHz and high frequency: respiration rate +/- 30 mHz) measured on 3 consecutive days and 1 month later in each subject. The standard deviation of differences between systolic blood pressure or heart rate oscillations on different occasions was in the 150-200 and 50-100 mm Hg Hz-1/2 or beats min-1 Hz-1/2 range for low frequency and high frequency oscillations respectively. Similar results were found when inter-observer reproducibility was considered. 4. From these results, we derived a sample-size table giving the number of subjects to be included in studies of cross-over or parallel design in order to detect a non-random difference in spectral analysis parameters.(ABSTRACT TRUNCATED AT 250 WORDS)
