INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

HEART RATE VARIABILITY (HRV) is influenced by a variety of physiological stimuli and has been used to try to understand some aspects of the cardiovascular control mechanisms. Almost all attention has been focused on various ways of analyzing the R-R interval or heart rate. However, there is evidence that specific branches of the autonomic nervous system may influence different parts of the cardiac cycle to differing degrees. For example, the Q-T interval is prolonged in sleep, independent of heart rate (4), suggesting that cardiac cycle length (R-R interval) and repolarization (Q-T interval) are independently controlled by the autonomic system.

Power spectral analysis (PSA) has been commonly used to quantify the relative dominance of vagal and sympathetic influence on the heart (6, 9). Although the interpretation of PSA of HRV in various clinical situations is not fully understood, it is widely agreed that the area under the high-frequency (HF) peak (0.15-0.4 Hz) is predominantly related to respiratory modulation of the cardiac vagal input (13) and the relative low-frequency (LF; 0.04-0.15 Hz)-to-HF ratio (LF/HF) is a reasonable measure of the relative effects of the sympathetic and parasympathetic outflows to the sinoatrial (SA) node in healthy volunteers (5). In the supine position there is a vagal predominance, whereas the upright position results in vagal inhibition and sympathetic predominance. To our knowledge, the PSA of the P-R and R-T variability has not been studied previously in healthy volunteers. Therefore, this study investigates the feasibility of determining the PSA of R-R, R-T, and P-R intervals and reports on the effect of change in posture.

	METHODS

Top Abstract Introduction Methods Results Discussion References

The study was approved by the regional ethics committee, and fully informed consent was obtained from all participants. Ten healthy volunteers (4 male and 6 female) between 26 and 35 years of age (mean age 29.7) participated in the study. A three-lead electrocardiogram (ECG) was recorded from the subject at the same time as a respiratory trace was recorded. This was achieved with an inductance plethysmograph (Respitrace; Studely Data Systems, Oxford, UK). Each data record was 300 s in duration. The subjects were initially supine and were cued by a computer bleep to breathe regularly at 5-s intervals. To determine the broadband response in terms of an impulse response, a second recording was then taken after the computer program had been changed to cue irregular breathing having a random Poisson distribution as described by Berger et al. (mean 5 s, range 1-15 s; Ref. 2). A further two similar recordings were then repeated with the subject in an upright position. Sampling frequency was 200 Hz. Data were collected with an analog-to-digital converter (Strobes APC; Wellington) and stored on a personal computer for further analysis.

Waveform and spectral analysis. The data were analyzed off-line using purpose-written software written in C++ language and Matlab (Matlab 4.2, The Mathworks). The QRS complex was detected (7), and then the peak of the R wave was precisely located using local quadratic interpolation of the surrounding five data points. On the basis of previous data collected at a sampling frequency of 1,000 Hz, this method of interpolation gave an estimated accuracy of ±1 ms. To estimate the P-R and R-T intervals, the peaks of the P wave and T wave were identified as those two peaks on each side of the R peaks within a certain time window. The width of the window for the P peak was one-third of the R-R interval, and that for the T peak was one-fourth of the R-R interval. For ease of data acquisition we followed Sarma et al. (11) in using the peaks of the ECG waves to define the P-R and R-T intervals. Thus our usage of P-R and R-T intervals is not identical to the commonly used definitions, which define the P-R interval as being from the start of the P wave to the start of the Q wave and the Q-T interval as being from the start of the Q wave to the end of the T wave.

Statistical analysis. To provide some internal validation of our results, we used three different methods to estimate the effect of vagal modulation of heart rate. First, when the subjects were breathing regularly, the power in the HF wave band was used as a measure of vagal tone. Second, the lung volume-to-heart rate impulse response was derived from the random breathing part of the study using the method of Yana et al. (14). By "impulse response," we mean that a short pulse of respiratory "energy" is introduced to observe the system's return to equilibrium. The impulse-response function is used to describe the relationship between input (lung volume) and output (heart rate) in the time domain. Its equivalent in the frequency domain is the transfer function. The impulse response is commonly modeled using one (or two) exponential functions, thus enabling the total effect of breathing on the heart rate to be understood intuitively in terms of a certain amplitude of heart rate response per unit change in lung volume and a corresponding time constant to describe the time for the heart rate to return to a homeostatic steady state.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Overall, the amount of power in the P-R and R-T spectra was two to three orders of magnitude less than that in the R-R spectrum (P < 0.001). The impulse response of breathing on the R-R interval (23.0 ± 4.9 ms · ml¹ · s¹) was 20 times the P-R (0.94 ± 0.29 ms · ml¹ · s¹) or R-T (1.19 ± 0.09 ms · ml¹ · s¹) intervals (P < 0.001).

In the P-R and R-T spectra the proportion of LF power (LF/HF) was approximately one-third of that in the R-R spectrum (P < 0.001). When the subjects were breathing regularly, we were able to show a clear respiratory peak in the power spectra of all three intervals (see Fig. 1).

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Typical power spectra of R-R, P-R, and R-T intervals showing clear respiration-related high-frequency peaks at 0.2 Hz in individuals in supine and upright postures. f, Frequency.

Effects of posture. The mean durations of all three intervals were significantly reduced when posture was changed to the upright position. The R-R interval decreased the most (932 to 787 ms, P < 0.001) compared with the P-R (139 to 130 ms, P = 0.002) and R-T (268 to 251 ms, P = 0.007) intervals.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

This study demonstrated that it is possible to use power spectral techniques to analyze autonomic influences on the P-R interval and ventricular repolarization. It must be acknowledged that, by measuring the P-R interval from the peak of the P wave to the peak of the R wave, we are not accurately measuring the pure atrioventricular conduction time but are including effects of variation in P wave morphology and early bundle of His depolarization. Nevertheless, AV conduction is generally considered to be the most important determinant of the P-R interval (3), and we clearly demonstrated a respiratory related pattern rather than random noise, even though the P-R variability lay in the range of ~2-20 ms, which is at the limits of our resolution of the waveforms.

The second aspect of our study demonstrated the functional specificity in indexes of activity of different branches of the autonomic supply to the heart. This complements the work by Randall and co-workers (10) who have previously demonstrated anatomically discrete parasympathetic and sympathetic pathways to the SA and AV nodes in dogs. In addition to observing the effects of localized pharmacological blockade of sympathetic and parasympathetic influence, they also showed phasic cardiorespiratory electrical activity in the vagal ganglia leading to the SA node.

In agreement with previous studies, the effect of standing caused shortening of all three intervals and a marked decrease in indexes of respiration-induced vagal influence on the SA node (as shown by the R-R interval HRV in the HF band, LF/HF, and the impulse-response function). In contrast, there was little change, or even an increase, in respiratory modulation on the AV node as estimated by the P-R interval HRV indexes and ventricular repolarization (R-T interval HRV indexes). Although further work is needed to separate effects due to the local cardiac responses from those caused by the changes in autonomic neural outflow, our observations suggest that standing results in: 1) an inhibition of respiratory vagal modulation of the SA node and, possibly, an increase in sympathetic tone, which shortens all three intervals and increases the heart rate; and 2) relative preservation of the respiratory modulation of the AV-node conduction time and ventricular repolarization time.

Recent work by Piepoli et al. (8) suggests that respiration-induced discharges from the arterial baroreceptors form the afferent loop of a reflex arc that acts via the cardiac vagus on the SA node to cause most of the observed respiratory sinus arrythmia. From our study, because the respiration-related variability of the P-R and R-T intervals did not change with standing, one may speculate that vagal activity to the AV node and ventricular myocardium may be relatively independent of baroreceptor input.

Also of interest was the strong vagal influence on the R-T variability. Using beta blockade, Sarma et al. (11) suggested that the Q-T interval was relatively insensitive to vagal activity; however, we found clear respiratory-related PSA peaks in most subjects. The differences in results may be due to the fact that we studied healthy volunteers, whereas Sarma et al. were studying patients suffering from angina, which is known to be associated with a sympathetic predominance as well as inhibited cardiac end-organ sensitivity to neural influences.

Thus, in summary, the effects of standing were an observed reduction in vagal influence on the HRV of the R-R interval and maintenance of lung volume-related vagal modulation of the P-R and R-T intervals. Standing did not produce a significant change in the LF and VLF spectral power of any interval.