| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1School of Sport and Exercise Sciences and 2Department of Anaesthesia and Intensive Care, University of Birmingham, Birmingham B15 2TT, United Kingdom
Submitted 17 December 2002 ; accepted in final form 21 April 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
carbon dioxide
The central respiratory rhythm is already detectable before the breakpoint of breath holding from inspiratory muscle activity [recorded indirectly from rhythmic waves of negative esophageal pressure and of diaphragmatic electromyogram (EMG) activity recorded across the esophagus (1, 32, 50, 51)]. From these measurements, however, the central respiratory rhythm is not detectable from the start of breath holding. It might be argued that this is because humans start breath holding by voluntarily stopping their central respiratory rhythm. Alternatively, it might be impossible for them to stop this vital rhythm (after all, humans cannot voluntarily stop their heart from beating). Instead, humans might start breath holding merely by voluntarily suppressing expression of their central respiratory rhythm before it reaches respiratory muscles. These hypotheses can only be distinguished with a potential indicator of the central respiratory rhythm from the start of breath holding. During breath holding, sinus arrhythmia may be especially useful as such an indicator because so many of the rhythmic influences on sinus arrhythmia are absent.
Sinus arrhythmia coincides with the central respiratory rhythm when it
reappears near the break point of breath holding
(1), but the presence of sinus
arrhythmia from the start of breath holds in air is disputed
(6,
17–19,
22,
23,
27,
29,
37,
39,
40,
47–49).
Part of this dispute arises because subjects rarely hold their breath in air
for longer than 1 min and often for only 20–45 s, which is not long
enough to establish precisely what periodicity is present in heart period
variability. We therefore prolonged breath hold times to 
4 min using
preoxygenation (26,
32,
34) to establish its
periodicity unequivocally. We then examined whether there is evidence for any
of the frequency components of sinus arrhythmia behaving as if the central
respiratory rhythm might continue from the start of breath holding.
If the central respiratory rhythm continues from the start of breath holding, we would expect it to start at the subject's eupneic frequency. The frequency components of any sinus arrhythmia can be established objectively, even in the absence of any independent measure of the central respiratory rhythm, by using spectral analysis. We have therefore examined whether any of the frequency components of sinus arrhythmia at the start of breath holding are within each subject's eupneic frequency range.
The characteristic CO2 dependency of whatever causes the central respiratory rhythm (2, 12, 24, 30) uniquely distinguishes it from other oscillators that may also exist in the central nervous system (CNS) (24) that may also oscillate at "respiratory frequencies" (without causing the central respiratory rhythm) and that influence autonomic activity during breath holding. We therefore tested whether any of the frequency components of sinus arrhythmia were reduced by a prior reduction in PCO2 at the start of breath holding and were stimulated by the linear rise in PCO2 throughout breath holding with preoxygenation (34). Preliminary results have been presented (15).
In this paper, we use the term central respiratory rhythm to mean the respiratory output generated by the brain stem. Humans may also inspire voluntarily [especially in hypocapnia (16, 33)] via a cortical projection to the diaphragm that apparently does not synapse in the brain stem (43, 44). This mechanism will not be considered, however, because, as breath holding is voluntary, it is difficult to imagine how subjects could voluntarily suppress breathing while retaining the ability to inspire voluntarily.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Our primary interest was in heart period variability at frequencies at which the central respiratory rhythm might conceivably occur at the start of breath holding and does occur (1, 51) toward the end of breath holding, i.e., between 0.03 and 0.5 Hz (2–30 breaths/min). Spectral analysis is widely used to analyze heart period variability, with the presumption that the underlying waveforms are essentially deterministic and hence determinable. Although the condition of breath holding is not stable, many previous studies (37, 40, 47) have accepted this presumption for breath holding. This appears reasonable (37, 40, 47), particularly when using preoxygenation, which prevents (26) (as we confirm) the drift in mean heart period during breath holding. ECG was converted into a continuous line of instantaneous heart period, and we met the appropriate sampling parameters for fast Fourier transforms and power spectrum analysis by low-pass filtration of the line at 0.5 Hz and sampling it at 8 Hz. Such processing allows consideration of frequencies up to 0.5 Hz without aliasing or distortion. We considered 2-min data periods, generating 64-s blocks of data with a 512-point fast Fourier transform, allocating power to the appropriate 256 frequency bins spanning 0.01–4 Hz with a bin width of 0.0156250 Hz. Data were windowed with a raised cosine function. The bins above 0.5 Hz were empty, so 32 bins describe 0–0.5 Hz. Because the first two bins containing the voltage offset component were ignored, the lowest frequency we resolved was 0.0312500 Hz and the remaining 30 bins describe the frequencies up to 0.5 Hz, i.e., the respiratory frequency range.
We confirmed the validity of our 0.03- to 0.5-Hz power (power0.03 to 0.5 Hz) spectrum analysis by establishing that spectral analysis of a 1-Hz pulse modulated at three separate, specific frequencies allocated power to the correct frequency bins and that increasing the amplitude of the modulation increased the power density only in these bins.
To compare breathing and heart period variability on the same scale, each breath interval in the 2 min preceding the breath hold was converted to a frequency and also allocated to 1 of 30 bins spanning 0.03–0.5 Hz (Fig. 1B). The eupneic frequency range for each subject was therefore indicated by the range of bins containing breaths. This range always spanned at least four bins, and no subject breathed with frequencies outside 0.03–0.5 Hz. For each subject, the power0.3 to 0.5 Hz density in their eupneic frequency range was calculated by summing the power in each bin of their own eupneic frequency range (see Fig. 1B), and the power0.03 to 0.5 Hz was also calculated for the ranges above (bins up to 0.5 Hz) and below (bins down to 0.03 Hz) their own eupneic frequency range. (Our allocation of power to these three frequency ranges is relatively unaffected by increasing the frequency resolution of spectral analysis in that this would only increase the number of bins describing each range; the ranges themselves would be virtually unchanged.)
|
Blood pressure and heart period were averaged in each 2-min period. Means are given with their SEs. Statistical analysis was performed using one-way repeated-measures ANOVA using SPSS general linear modeling for parametric data. Significant F values were found for mean blood pressure (F = 9.6, P < 0.05), and the sources of significance were then investigated using Student's paired t-tests. Statistical analysis was performed using one-way repeated-measures ANOVA using the Kruskal-Wallis test for nonparametric (power density) data. Significant H values were found for power0.03 to 0.5 Hz density within (H = 19.1, P < 0.005), above (H = 15.4, P < 0.05), and below (H = 20.1, P < 0.005) the eupneic frequency range, and the sources of significance were then investigated by analyzing matched pair differences using the Wilcoxon signed-rank test.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Ten subjects breath held with preoxygenation, for a minimum of 3 min (and with a mean duration of 4.0 ± 0.3 min), and breath held longer (32) from hypocapnia, for a minimum of 3.9 min (and with a mean duration of 6.6 ± 0.7 min). The mean PETCO2 at the break point for breath holds from normocapnia was 59 ± 2 mmHg (20 mmHg above their eupneic level). The mean PETCO2 at the break point for breath holds from hypocapnia was 48 ± 3 mmHg (n = 10). There was no detectable hemoglobin desaturation at break point in any condition. The cardiovascular changes we observed during breath holding were similar to those found previously (26, 28). Averaged mean blood pressures were significantly higher at the end of breath holding [by 32 ± 3 mmHg in breath holds from normocapnia (P < 0.001) and by 23 ± 4 mmHg in breath holds from hypocapnia (P < 0.001)] than in eupnea (104 ± 5 mmHg, n = 10). As found previously for breath holds with preoxygenation (26), there was no significant difference in mean heart period at the end of both types of breath hold compared with eupnea (788 ± 35 ms, n = 10) despite the pressure rise.
At the start of breath holding, when so many of the rhythmic influences on sinus arrhythmia are absent [and when rhythmic negative pressure waves are undetectable (1, 32, 50, 51)], minimal sinus arrhythmia is expected. Yet, all subjects showed substantial sinus arrhythmia from the start of breath holding. Figure 1E shows such sinus arrhythmia in the subject with the longest breath hold (6.0 min) from normocapnia. The frequency of the sinus arrhythmia during breath holding appears similar to that of the preceding eupnea (compare Fig. 1, C with E), indicating a possible contribution from central respiratory rhythm. This similarity is objectively confirmed by spectral analysis (compare Fig. 1, D with F), showing substantial power in the eupneic frequency range in both conditions. All subjects had substantial power in sinus arrhythmia in their own eupneic frequency ranges (Fig. 2A) from the start of breath holding. The amount of power in their eupneic ranges in the first 2 min is significantly less than in eupnea itself. This again is consistent with a possible contribution from the central respiratory rhythm because the ways by which it could contribute to sinus arrhythmia are reduced during breath holding [with the absence of pulmonary stretch (7) and the ensuing mechanical sequelae (4, 8, 9, 45)]. We also confirmed previous studies (37, 40, 47) showing that there is also significantly more power0.01 to 0.5 Hz below the eupneic frequency range during breath holding, demonstrating that a number of oscillators may continue to have some influence on sinus arrhythmia at a range of frequencies during breath holding.
We show that, from the start of breath holding, some frequency components have the appropriate sensitivity to decreased PETCO2. Spectral analysis objectively demonstrates (Fig. 2, A and B) that there is significantly less power0.03 to 0.5 Hz both within and below their eupneic frequency ranges in the first 2 min of breath holding from hypocapnia than in this period in breath holds from normocapnia.
PETCO2 rises linearly throughout breath holding with preoxygenation (34). This raises the frequency of the central respiratory rhythm above the eupnoeic frequency range when it is detectable from rhythmic esophageal EMG and negative pressure waves near the end of breath holding (1, 51). We found that all subjects continued to show sinus arrhythmia throughout breath holding, and spectral analysis objectively confirms that the appropriate frequency components of power0.03 to 0.5 Hz in sinus arrhythmia have the appropriate increases with increased PETCO2 between the first and last 2 min of breath holding. Figure 2C shows that there was significantly more power0.03 to 0.5 Hz in frequencies above their eupneic ranges in the last than in the first 2 min of breath holding. Median power0.03 to 0.5 Hz within their eupneic ranges also increased in the last 2 min (Fig. 2A). This increase was not significant over all subjects, but the sensitivity of our comparison is reduced because we had to measure power in 2-min blocks and because some of our data overlap (not all subjects could breath hold for exactly 4 min). Similarly, for breath holds from hypocapnia, median power also increased between the first and last 2 min, but again the increase that our techniques measured was not statistically significant.
Finally, it is known (34) that that the PETCO2 at the break point is lower when breath holding is preceded by hypocapnia that when it is not. Therefore, the central respiratory rhythm at break point from hypocapnia has less chemical drive and must be reduced relative to that at the break point of breath holds from normocapnia. Sinus arrhythmia also indicates this reduction in the central respiratory rhythm. Figure 2A shows that there is significantly less power0.03 to 0.5 Hz in the eupneic frequency range in the last 2 min of breath holds from hypocapnia (mean PETCO2 at break point 48 ± 3 mmHg with 100% O2 saturation) than from normocapnia (mean PETCO2 at break point 59 ± 2 mmHg with 100% O2 saturation).
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
CO2-sensitive sinus arrhythmia from the start of breath holding. At the start of breath holding, when so many of the rhythmic influences on sinus arrhythmia are absent [and when rhythmic negative pressure waves are undetectable (1, 32, 50, 51)], minimal sinus arrhythmia is expected. Yet, all subjects showed sinus arrhythmia from the start (e.g., Fig. 1E) and throughout breath holds with preoxygenation. This extends previous reports of its persistence during breath holds in air (27, 49). Undoubtedly, multiple factors contribute to such sinus arrhythmia. The fact the mean heart period does not change during breath holds with preoxygenation (26) shows that sympathetic tone to the heart does not obviously increase during such breath holds, despite the rise in blood pressure. Nor is there any overt decrease in parasympathetic tone, although neither point excludes the possibility that the balance of such drives is altered in breath holds with preoxygenation.
The central respiratory rhythm is one obvious candidate to consider influencing sinus arrhythmia from the start of breath holding because it makes such a substantial contribution to sinus arrhythmia in eupnea and because it is already known to appear before the break point (1, 32, 50, 51) with sinus arrhythmia coinciding (1). One crucial difficulty, however, is that no independent indicator of the central respiratory rhythm exists to use from the start of breath holding in humans. Until such a measure is found for humans, we can only test whether or not any frequency components of sinus arrhythmia have the two properties expected if the central respiratory rhythm continued. We show that some frequency components of sinus arrhythmia do have both properties.
We would expect the central respiratory rhythm to continue at each subject's eupneic frequency at the start of breath holding if it continues. The frequency of the sinus arrhythmia during breath holding appears similar to that of the preceding eupnea (compare Fig. 1, C with E), and this similarity is objectively confirmed by spectral analysis (compare Fig. 1, D with F). We show that all subjects have substantial power in sinus arrhythmia in their own eupneic frequency ranges (Fig. 2A) from the start of breath holding. When spectrally analyzing heart period variability, it is more usual to look for sinus arrhythmia within the standardized definitions of "low-frequency" or "high-frequency" ranges (36, 46) during breath holding (37, 40, 46, 47, 49). But our hypothesis is to test specifically whether power remains in the eupneic frequency range, so we must consider each subject's eupneic frequency range. Our analysis is also more exacting because it takes into account the fact that different subjects have different eupnoeic frequency ranges (11).
Although the mean eupneic frequency was 0.2 ± 0.02 Hz, some subjects also had some breaths with an instantaneous frequency at 0.1 Hz, a range normally associated with the sympathetic contribution to sinus arrhythmia (11, 46). We must include such subjects because this is part of their eupneic frequency range. In any event, the CNS can only influence heart period variability through the parasympathetic and sympathetic systems. If the central respiratory rhythm produces some breaths at 0.1 Hz, it could contribute to sinus arrhythmia at this frequency via both sympathetic and parasympathetic systems. In such cases, our inclusion of power in the 0.1-Hz bin is perfectly valid. If any of the sympathetic contribution to sinus arrhythmia during breath holding is from the central respiratory rhythm, this too should be CO2 sensitive and will be revealed by our analysis. If none is, the sympathetic component should not change with PCO2 and would not mask any of the parasympathetic contribution that may be CO2 sensitive.
In addition, we show that the absolute power density in their eupneic ranges in the first 2 min is significantly less than in eupnea itself. This too is expected. The ways by which the central respiratory rhythm may contribute to sinus arrhythmia are reduced at the start of breath holding; with the absence of its more indirect contributions related to pulmonary stretch (7) and the ensuing mechanical sequelae (4, 8, 9, 45), any of its remaining influences would be directly onto cardiac preganglionic vagal neurons (25, 31) through modulation of a tonic CO2-dependent drive (42) or of baroreflex function (21, 22, 37) and possibly even by driving dynamic interactions in populations of multiple oscillators (12, 24, 30).
If the central respiratory rhythm continued from the start of breath holding, we would expect it also to remain CO2 dependent. We show the such CO2 sensitivity is apparent even in the first 2 min of breath holding because hypocapnia significantly reduces the power0.03 to 0.5 Hz density in the first 2 min in the eupneic frequency range. Hypocapnia did not abolish power in these ranges, but neither do PETCO2 levels of 23 mmHg stop the central respiratory rhythm (10). One additional means of showing such CO2 sensitivity would be to measure an increase in sinus arrhythmia in breath holds starting from hypercapnia. This is difficult to study, however, because hypercapnia shortens breath hold time (34) and hence reduces the time available to quantify sinus arrhythmia.
Reappearance of the central respiratory rhythm toward the end of breath holding. The central respiratory rhythm is known to reappear before the break point of breath holding because of the reappearance of rhythmic waves of negative esophageal pressure (32, 50, 51) and of esophageal EMG activity from the diaphragm (1). Such waves cannot, however, be used to study sinus arrhythmia from the start of breath holding because they only appear 1.7 ± 0.01 min after the start of breath holding in comparable breath holds [from vital capacity with preoxygenation (32)]. Such negative pressure recordings have nevertheless revealed three important features of breath holding. First, their appearance at or above the eupneic frequency, and their increase in frequency and amplitude toward the break point (1, 51), confirm that the central respiratory rhythm retains its CO2 sensitivity during breath holding. Second, such recordings show that subjects do not produce a rhythm of positive (expiratory) efforts during breath holding. Finally, they show that the effort in suppressing breathing toward the end of breath holding neither causes a sustained increase in intrathoracic pressure nor can explain the gradual increase in blood pressure during breath holding. [Deliberately raising intrathoracic pressure during breath holding reduces cardiac output (38), however, and can cause fainting.]
We can use the fact that PCO2 increases linearly throughout breath holding with preoxygenation (34) to examine whether additional and appropriate increases in the spectral power of sinus arrhythmia occur toward the end of breath holding.
We show there is significantly more power0.03 to 0.5 Hz in frequencies above their eupneic ranges in the last than in the first 2 min of breath holding. A counterproposal might be that only when the central respiratory rhythm causes such negative intrathoracic pressure waves during breath holding can it contribute to sinus arrhythmia. We believe that this proposal is unlikely for several reasons. It does not account for the facts that sinus arrhythmia appears at eupneic frequencies from the start of breath holds and has CO2 dependency before such waves appear. It also would require negative pressure to be one of the principal causes of sinus arrhythmia. It is already known that negative intrathoracic pressure is not the principal cause of sinus arrhythmia in eupnea (7–9, 11, 13, 14, 21, 22, 24, 31, 35, 37, 42). Because it is not, it is difficult to imagine how some of the mechanisms by which the central respiratory rhythm affects sinus arrhythmia [e.g., direct effects of inspiration on cardiac preganglionic vagal neurons (25, 31)] would become ineffective during breath holding, whereas negative pressure was not. Moreover, it is difficult to understand precisely how negative pressure could have such pronounced effects on sinus arrhythmia when there is so little chest expansion during late breath holding to cause pulmonary stretch or to increase venous return or to decrease left ventricular stroke volume. It is more likely that whenever the central respiratory rhythm is present during breath holding, it continues to make a contribution to sinus arrhythmia through the well-described range of mechanisms (7–9, 11, 13, 14, 21, 22, 24, 25, 31, 35, 37, 42). We deliberately did not insert esophageal catheters because their discomfort shortens breath hold duration (51) and because the absence of such waves means they cannot be used to study sinus arrhythmia from the start of breath holding. It is also not possible to get an adequate signal-to-noise ratio from EMG recordings on the chest surface during breath holding to distinguish the activity of particular muscle groups.
There is even a further indicator of CO2 sensitivity at the end of breath holding. Because it is known (34) that PETCO2 at the break point is lower when breath holding is preceded by hypocapnia that when it is not, the central respiratory rhythm at the break point from hypocapnia has less chemical drive and should be reduced relative to that at the break point of breath holds from normocapnia. Our analysis of sinus arrhythmia confirms this reduction, and this also extends previous studies of sinus arrhythmia in breath holds with hypocapnia (39, 47, 49).
Does the central respiratory rhythm continue from the start of breath holding? We show that some sinus arrhythmia is present in the eupneic frequency range from the start of breath holding and is CO2 dependent even before the central respiratory rhythm is detectable independently from esophageal waves of negative pressure or EMG activity. It therefore behaves as if the central respiratory rhythm continues from the start of breath holding, but without a more direct indicator of the central respiratory rhythm in conscious humans there is no definitive means of establishing this. Sinus arrhythmia is the only potential indicator of the central respiratory rhythm available. We carefully explained in the preceding paragraphs the limitations in indicating the central respiratory rhythm from sinus arrhythmia. If these limitations prevent the deduction of the presence of the central respiratory rhythm from the properties of sinus arrhythmia, new and more sensitive techniques need to be devised to indicate if it is present in conscious humans at the start of breath holding. If they do not, we demonstrate that humans cannot stop their central respiratory rhythm voluntarily and therefore breath hold by suppressing its expression before it reaches respiratory muscles. This would have important implications for our understanding of "apnea," although it does not of course preclude other oscillators from also continuing to contribute to sinus arrhythmia throughout breath holding.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. S. Saleh, S. Patel, M. H. Lee, M. I. Boechat, O. Ratib, C. R. Saraiva, and J. P. Finn Contrast-enhanced MR Angiography of the Chest and Abdomen with Use of Controlled Apnea in Children Radiology, June 1, 2007; 243(3): 837 - 846. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J Parkes Breath-holding and its breakpoint Exp Physiol, January 1, 2006; 91(1): 1 - 15. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. E. Cooper, T. H. Clutton-Brock, and M. J. Parkes Contribution of the respiratory rhythm to sinus arrhythmia in normal unanesthetized subjects during positive-pressure mechanical hyperventilation Am J Physiol Heart Circ Physiol, January 1, 2004; 286(1): H402 - H411. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
,