INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

MEASURES OF HEART RATE variability (or heart period variability) are used as probes into the dynamics of the cardiovascular control system. Results from spectral and nonlinear analyses (1, 12, 14, 15) are often interpreted in terms of long-term, deterministic controls that are presumed to underlie heart rate variability. However, analyses with nonlinear predictability, correlation dimension, and information scaling have shown that heart period sequences do not have the characteristics of signals with long-term determinism (6, 7, 18). What then might be the origin of heart period variability?

One possibility is that heart period variability is due to linear concatenation of temporally localized events. In this paradigm, heart period sequences are simply sequences of transient and characteristically structured changes in heart period, much like sentences are composed of strings of words. There would then be a lexicon of recurrent, similarly shaped transient structures, like words, and each word would have a specific physiological basis. We coin the word "lexon" to refer to the meaningful transient structures in heart period sequences. The first lexon candidate is the brief burst of tachycardia that occurs at the initiation of exercise (3, 4, 15, 24). It is associated with transient hypotension. Bursts have a monoexponential decrease in heart period of ~250 ms over 8-12 s, followed by a 10- to 14-s recovery period to baseline. They are detected easily on ambulatory electrocardiogram (ECG), and these detected bursts closely resemble bursts induced in the laboratory. This easy recognition has made the burst the first isolated event on ambulatory ECG whose temporally localized physiology might be understood. Investigations of the physiology of bursts and related transient tachycardic structures have been reported sporadically over the past 30 yr. Two hypotheses have arisen (3, 4, 8, 21): 1) bursts might be a tachycardia reflexively induced by the initiation of muscle contraction, and 2) they might be simple baroreceptor-mediated responses to the transient hypotension that occurs at the initiation of exercise.

In this report, we assessed two physiological aspects of bursts. First, we determined whether muscle workload or duration altered burst morphology. Second, we determined whether burst sinus tachycardia followed or coincided with the hypotension that accompanies bursts. For bursts to be baroreceptor-mediated responses to hypotension, they should follow the blood pressure drop by at least a single beat, rather than precede it.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experimental protocols. Ten subjects (6 male) with mean age 38 yr (range 17-69 yr) reported to the lab in the postabsorptive state, having refrained from smoking and caffeine for a minimum of 2 h before the testing. After 10 min of quiet semisupine rest at a 30° upright position on a stationary bicycle, the subjects completed two dynamic exercise protocols. In the first protocol, subjects completed three exercise bouts of one, three, and five revolutions against a resistance of 0 W at a rate of 60 rpm. The exercise bouts were randomly assigned, and after the initial exercise bout, subsequent bouts were separated by an additional 2 min of quiet rest while the subject remained seated. After an additional 5 min of quiet rest, subjects then performed a second exercise protocol that involved cycling against a resistance of 25, 50, and 75 W for five revolutions at a rate of 60 rpm. The exercise loads were randomly assigned, and after the initial exercise bout, subsequent bouts were separated by an additional 2 min of quiet rest. Before and during each exercise bout, the subjects were reminded to continue breathing evenly.

Data acquisition. Subjects were instrumented with ECG leads and the Finapress noninvasive blood pressure monitor (Ohmeda, Madison, WI) finger cuff. Both outputs were digitized at 200 Hz in CVSoft (Odessa Software, Calgary, Alberta, Canada) on a personal computer. The files were delineated and manually overread in CVSoft. As a convention, we used the time of the first R wave of each R-R interval as the time of that beat's heart period value. The blood pressure measurement at this time was recorded as the beat's diastolic pressure, and the subsequent systolic pressure was recorded as the beat's systolic blood pressure. Mean arterial pressure (MAP) values for each beat were calculated as follows: MAP = (systolic pressure)/3 + 2 × (diastolic pressure)/3 (15). The resulting heart period, diastolic, systolic, and MAP sequences were stored in Matlab (The Mathworks, Natick, MA).

Heart period morphological parameters of bursts. The three general heart period features of a burst are its duration, magnitude, and shape (15, 17). All the heart period morphological parameters are depicted in Fig. 1 and defined in Table 1. The major magnitude parameters are the heart period values at baseline and at the trough of the burst as well as the difference between them. Three of the four duration parameters are the time intervals between the initial beat, the minimum heart period beat, and final beat of a burst. Respiratory sinus arrhythmia was evident in the interburst intervals of most subjects, making delineation of the initial and final beats subject to imprecision. To circumvent this, we also delineated the time between the first beat in a burst to exceed 50% of the tachycardic magnitude, and the first beat to exceed 50% of the subsequent heart period recovery. The burst shape parameters depended on the observation that the burst's initial rapid heart period decay resembled a monoexponential decay function. We fit the heart period decay to a monoexponential curve and obtained a theoretical asymptote heart period, a decay rate constant, and a correlation coefficient quantifying the fit of the model to the heart period data.

View larger version (29K): [in this window] [in a new window]   Fig. 1.   Illustration of parameters used to describe heart period and blood pressure features of bursts. Solid diamonds represent beats included in burst, and open diamonds indicate beats fitted by monoexponential decay model (thick-lined curve). HP, heart period; MAP, mean arterial pressure; t, time. See Table 1 for full definition of parameters.

Blood pressure morphological parameters of bursts. Figure 1 also presents the MAP parameters. The MAP baseline value is the 20-s mean of the MAP values preceding the burst. We defined the magnitude of the transient hypertension that usually occurs at burst initiation (MAP_hyper) as the peak MAP increase from baseline occurring immediately after exercise initiation and before hypotension. The time from the initial beat to the peak MAP is defined as t_hyper. We defined the magnitude of hypotension (MAP_hypo) as the drop in MAP from the baseline to the lowest MAP value recorded in the 30 s immediately after the onset of exercise. We define t_hypo as the time from the start beat to this trough MAP beat.

Statistical analysis. We first determined whether the variable was normally distributed or skewed. Means (±SD) and medians were calculated for normally distributed continuous variables, whereas median values and their associated 25% and 75% quartile values are reported for non-normal distributions. Analyses of variance with repeated measures were performed for normally distributed populations. Post hoc significance tests of grouping are performed using Fisher's protected least significant difference tests. Kruskal-Wallistests were performed on grouped variables from non-normal distributions.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effect of work duration on bursts. Subjects cycled for 1, 3, and 5 s against a workload of 50 W. Bursts were observed in 28 of 30 bouts of zero-load cycling. Figure 2 displays the heart period sequences for two representative burst responses and the two failed attempts. There is a large-magnitude drop in heart period that follows exercise initiation, an apparently monoexponential decay in heart period, and a subsequent recovery toward baseline heart period values. Table 2 compares the values of the burst variables induced by the three work durations. No significant differences in measures of burst magnitude or shape occurred among the three groups. In contrast, two of the timing parameters did depend on work duration. These were the total durations of the bursts, t_burst (P = 0.033), and the times during which the heart period decrease was at least 50% of its maximal value, t_tachy (P = 0.0066). Thus, although the morphology of bursts is similar across a range of the work durations, burst duration does depend on work-load duration.

View larger version (30K): [in this window] [in a new window]   Fig. 2.   Top panels: heart period sequences spanning 2 failed attempts to induce bursts. Dashed line represents time of exercise initiation. These results were not used in analysis because a clear tachycardic response did not occur. Bottom panels: 2 heart period sequence recordings showing typical "burst" response to exercise initiation.

Effect of workload on bursts. Subjects cycled for 5 s against variable workloads of 0, 25, 50, and 75 W. Bursts were observed in 40 of 40 bouts of cycling. Table 3 compares the values of the burst variables induced by the four workloads. No significant differences in measures of burst duration, shape, or amplitude were seen. Thus burst morphology is remarkably similar across a range of the workloads.

MAP changes during bursts. Figure 1 and Table 4 show the typical MAP changes accompanying a burst. Fifty-six of sixty bursts were accompanied by initial hypertension rather than hypotension. All bursts were also accompanied by hypotension, which in 56 of 60 bursts immediately followed the hypertension. MAP dropped 12.5 ± 6.8 mmHg, reaching trough blood pressure after 14.5 ± 4.3 s.

Relationship between heart period and blood pressure changes. If the sinus tachycardia of bursts is mediated by arterial baroreceptors as a response to the hypotension that occurs with exercise initiation, then the burst tachycardia should lag behind the burst hypotension by at least a single beat. To test this hypothesis, we examined the relative timing of the hypotension and tachycardia of the bursts. We used the onset and peak effect times for comparing the timings of hypotension and tachycardia. Figure 3 illustrates the typical timing relationship between the tachycardia and the hypotension. Table 5 reports that in 56 of 58 bursts, the onset of tachycardia did not follow the onset of hypotension but preceded it. In 2 of 58 bursts, the tachycardia and the hypotension began simultaneously. With an average of all bursts, the onset of tachycardia preceded the onset of hypotension by 4.6 ± 2.2 s (P < 0.0001).

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Representative simultaneous heart period and blood pressure sequences in response to semisupine cycling for 5 s against a workload of 50 W. Under arterial baroreceptor model (see text), tachycardia onset should follow hypotension onset. Tachycardia peak should also follow hypotension trough.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This work shows that the morphologies of heart rate bursts induced by exercise initiation are generally similar, regardless of load and duration of the work. Furthermore, the relative timing of the heart rate burst compared with the associated hypotensive transient is evidence that the tachycardia is not mediated by arterial baroreceptors; rather, the heart rate burst appears to be a near-quantal response to the initiation of exertion. This has implications both for our understanding of the response to the initiation of exertion and of the structure of heart period variability.

Physiology of exercise initiation. Transient hypotension and sinus tachycardia lasting ~20-25 s (3, 4, 24) accompany large muscle activation. The hypotension might be due to cardiopulmonary reflex-mediated vasodilation, local metabolic vasodilation, central command-mediated cholinergic vasodilation, or muscle pump activation, possibly via flow-mediated vasodilation. Its time course and resistance to autonomic neuropathies and autonomic drugs suggest that it is due to muscle pump activation (2-5, 8, 21, 23). The recovery phase may be mediated by arterial baroreceptors. Toska and Eriksen (23) observed a reduction in total peripheral compliance after ~12 s of dynamic leg exercise and suggested that this was because of systemic vasoconstriction mediated by arterial baroreceptor activation.

Physiology of heart rate bursts. In our subjects, the heart rate burst consistently preceded hypotension. Although it is possible that later phases of bursts are modulated by arterial baroreceptors (11, 23), we conclude that arterial baroreceptors do not cause the first phase of heart rate bursts. A second possible cause is central command. This too seems unlikely given that both voluntary and involuntary activations of forearm adductor muscles induce similar small bursts of heart rate lasting ~5 s (2, 5). Small muscle activation seems to cause heart rate accelerations similar to bursts induced by major muscle activation. Because externally stimulated contraction of these muscles induces transient sinus tachycardia, central command is unlikely.

Work duration, workload, and burst morphology. We reasoned that if muscle mechanoreceptors are the primary sensory organ, then within the burst there might be a graded response to graded workload or duration. This has not been systematically addressed before, although unilateral and bilateral limb muscle activations cause similar increases in heart rate (2, 5). Most of the morphological parameters of the exercise-induced bursts were independent of dynamic exercise workload and duration. Although there were significant differences in some measures of burst duration with exercise duration, they were seen only between the 1-s exercise bouts and the longer bouts. This might be because of the small sample size or a threshold effect. The shape of the initial heart rate acceleration was not changed by variations in workload and duration. Thus burst morphology is generally independent of the exercise load or duration that induces it. Whatever the origin, the mechanism of burst tachycardia is likely due to withdrawal of vagal tone, given that atropine but not propranolol prevents it (4, 24). For this reason, we refer to its first phase as heart period decay, in recognition of the possible role of the decay of synaptic acetylcholine in the genesis of the tachycardia.

Origin of heart rate bursts. We speculate that the burst is a universal response to the initiation of exercise from relative quiescence. Bursts occur at the initiation of supine and upright cycling and standing up (3, 4, 15, 24), and we have easily induced bursts with sitting up, standing up, initiation of walking, and single leg presses (unpublished data). They do not occur when workload is abruptly increased during dynamic exercise. The great similarity of burst morphology in published reports, and in response to variations in workload and duration, suggests that bursts may be a near-quantal heart response to the initiation of exercise from rest. The muscle mechanoreceptors may be the afferent sensors, but the heart rate response is not graded but rather nearly quantal.

The burst lexon. The heart rate burst at exercise initiation is the first heart period lexon to be described (15, 17). It is inducible in all healthy subjects, it is detectable on ambulatory ECG of all healthy subjects, the detected bursts resemble the induced bursts in three or more features, and the provocative factors associated with the detected bursts are plausibly related to the factors that induced the burst under controlled circumstances (15). Thus a burst is a discrete, imperfectly reproducible heart period event that is related to known physiological behavior and is easily detectable on ambulatory heart period recordings.

Perspectives on the lexical approach. The lexical approach builds easily on earlier findings in this field. Statistical analysis (9), time-frequency decomposition studies (10), and the broad spectral widths of both the low-frequency band (related to hemodynamic control) and the high-frequency band (related to respiration) all argued that the origins of these band widths are transient and therefore temporally localized (1, 12, 14, 19). Similarly, the sources of the ultralow frequency band are temporally localized around the times of going to bed and getting up (16).

Limitations. We only examined the response to the initiation of cycling, and not the effects of aging (13, 22), disease (22), or drugs (4) on burst morphology. Indeed, burst magnitude may fall with advancing age (13). We did not assess the impact of other physiological factors such as exertion status or orthostatic stress.