Comparison of baroreflex sensitivity estimated from ECG R-R and inter-systolic intervals obtained by finger plethysmography and radial tonometry.

Spontaneous BRS estimates may considerable vary according to the technique of blood pressure and heart rate assessment. To optimise and standardise BRS estimation for clinical use we evaluated possible differences between spontaneous BRS indices estimated from either finger plethysmography or radial tonometry. Forty-five healthy volunteers underwent simultaneous recordings of electrocardiogram, finger plethysmography and radial tonometry in supine position and during 60° head-up tilt. BRS was computed by spectral analysis from either R-R time series and/or arterial pressure pulse. Radial tonometry generated higher mean BRS estimates than finger plethysmography. The difference decreased upon postural change from supine to upright. In the upright position, BRS estimates based on R-R interval proved to be generally lower compared to BRS indices estimated from arterial pressure pulse. The ratio of low-to-high-frequency power of inter-systolic interval and systolic blood pressure from tonometry was lower than that from plethysmography in supine and approximated in upright position. Spectral parameters of inter-systolic interval and R-R interval did not differ in supine but diverged in upright position. Changes of spectral parameters were most pronounced in R-R interval. Arterial pressure pulse is adequate for estimation of BRS under resting conditions but it may distort BRS estimates under physical load. We, therefore, recommend using an ECG signal for BRS estimation especially in non-stationary conditions.

Trigonometric regressive spectral analysis: an innovative tool for evaluating the autonomic nervous system.

Biological rhythms, describing the temporal variation of biological processes, are a characteristic feature of complex systems. The analysis of biological rhythms can provide important insights into the pathophysiology of different diseases, especially, in cardiovascular medicine. In the field of the autonomic nervous system, heart rate variability (HRV) and baroreflex sensitivity (BRS) describe important fluctuations of blood pressure and heart rate which are often analyzed by Fourier transformation. However, these parameters are stochastic with overlaying rhythmical structures. R-R intervals as independent variables of time are not equidistant. That is why the trigonometric regressive spectral (TRS) analysis--reviewed in this paper--was introduced, considering both the statistical and rhythmical features of such time series. The data segments required for TRS analysis can be as short as 20 s allowing for dynamic evaluation of heart rate and blood pressure interaction over longer periods. Beyond HRV, TRS also estimates BRS based on linear regression analyses of coherent heart rate and blood pressure oscillations. An additional advantage is that all oscillations are analyzed by the same (maximal) number of R-R intervals thereby providing a high number of individual BRS values. This ensures a high confidence level of BRS determination which, along with short recording periods, may be of profound clinical relevance. The dynamic assessment of heart rate and blood pressure spectra by TRS allows a more precise evaluation of cardiovascular modulation under different settings as has already been demonstrated in different clinical studies.

Determination of baroreflex sensitivity during the modified Oxford maneuver by trigonometric regressive spectral analysis.

BACKGROUND: Differences in spontaneous and drug-induced baroreflex sensitivity (BRS) have been attributed to its different operating ranges. The current study attempted to compare BRS estimates during cardiovascular steady-state and pharmacologically stimulation using an innovative algorithm for dynamic determination of baroreflex gain. METHODOLOGY/PRINCIPAL FINDINGS: Forty-five volunteers underwent the modified Oxford maneuver in supine and 60° tilted position with blood pressure and heart rate being continuously recorded. Drug-induced BRS-estimates were calculated from data obtained by bolus injections of nitroprusside and phenylephrine. Spontaneous indices were derived from data obtained during rest (stationary) and under pharmacological stimulation (non-stationary) using the algorithm of trigonometric regressive spectral analysis (TRS). Spontaneous and drug-induced BRS values were significantly correlated and display directionally similar changes under different situations. Using the Bland-Altman method, systematic differences between spontaneous and drug-induced estimates were found and revealed that the discrepancy can be as large as the gain itself. Fixed bias was not evident with ordinary least products regression. The correlation and agreement between the estimates increased significantly when BRS was calculated by TRS in non-stationary mode during the drug injection period. TRS-BRS significantly increased during phenylephrine and decreased under nitroprusside. CONCLUSIONS/SIGNIFICANCE: The TRS analysis provides a reliable, non-invasive assessment of human BRS not only under static steady state conditions, but also during pharmacological perturbation of the cardiovascular system.

Trigonometric regressive spectral analysis reliably maps dynamic changes in baroreflex sensitivity and autonomic tone: the effect of gender and age.

BACKGROUND: The assessment of baroreflex sensitivity (BRS) has emerged as prognostic tool in cardiology. Although available computer-assisted methods, measuring spontaneous fluctuations of heart rate and blood pressure in the time and frequency domain are easily applicable, they do not allow for quantification of BRS during cardiovascular adaption processes. This, however, seems an essential criterion for clinical application. We evaluated a novel algorithm based on trigonometric regression regarding its ability to map dynamic changes in BRS and autonomic tone during cardiovascular provocation in relation to gender and age. METHODOLOGY/PRINCIPAL FINDINGS: We continuously recorded systemic arterial pressure, electrocardiogram and respiration in 23 young subjects (25+/-2 years) and 22 middle-aged subjects (56+/-4 years) during cardiovascular autonomic testing (metronomic breathing, Valsalva manoeuvre, head-up tilt). Baroreflex- and spectral analysis was performed using the algorithm of trigonometric regressive spectral analysis. There was an age-related decline in spontaneous BRS and high frequency oscillations of RR intervals. Changes in autonomic tone evoked by cardiovascular provocation were observed as shifts in the ratio of low to high frequency oscillations of RR intervals and blood pressure. Respiration at 0.1 Hz elicited an increase in BRS while head-up tilt and Valsalva manoeuvre resulted in a downregulation of BRS. The extent of autonomic adaption was in general more pronounced in young individuals and declined stronger with age in women than in men. CONCLUSIONS/SIGNIFICANCE: The trigonometric regressive spectral analysis reliably maps age- and gender-related differences in baroreflex- and autonomic function and is able to describe adaption processes of baroreceptor circuit during cardiovascular stimulation. Hence, this novel algorithm may be a useful screening tool to detect abnormalities in cardiovascular adaption processes even when resting values appear to be normal.

Influence of ECG sampling frequency on spectral analysis of RR intervals and baroreflex sensitivity using the EUROBAVAR data set.

To evaluate the impact of different ECG sampling frequencies on parameters of spectral and baroreflex analysis. Spectral and baroreflex analyses were performed in the EUROBAVAR data set (46 recordings of 23 persons) using the original ECG sampling frequency of 500 Hz and - simulated - sampling frequencies of 200 and 100 Hz. For this analysis, the technique of trigonometric regressive spectral (TRS) analysis was used. In the standing position, there were no statistically significant differences in baroreflex sensitivity and frequency bands ranging from VLF to HF using 100 Hz instead of the original 500 Hz. Only the UHF band (>0.40 Hz) was significantly different. In the supine position, similar results could be described for 100 Hz, although there were slight, but significant (P < 0.05) changes in baroreflex sensitivity of around 1 ms/mmHg at the simulated 100 Hz. Using a simulated 200 Hz instead of a 500 Hz sampling frequency had no significant impact on the spectral and baroreflex parameters. The probability to demonstrate an impact of different ECG sampling frequencies was higher in people with pathologically decreased variability of RR intervals. In most of the cases, it is sufficient for spectral and baroreflex analysis by TRS to use data with an ECG sampling frequency of 100 Hz in comparison to 500 Hz. Only if there is a pathologically decreased variability of RR intervals in patients, spectral and baroreflex parameters could be significantly influenced by lower ECG sampling frequencies of up to 100 Hz, but only to a minor degree.