INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

RESPIRATORY SINUS ARRHYTHMIA (RSA), the variations in heart rate associated with respiration, is widely used as an index of cardiac vagal tone (2, 3, 8, 16, 17). At the same time, RSA is a manifestation of cardiorespiratory interactions (4, 19, 21) that contributes to CO₂ elimination by reducing physiological dead space and intrapulmonary shunt, i.e., matching the distribution of pulmonary blood flow to lung volume during each respiratory cycle (13). An increased demand for CO₂ elimination may therefore enhance RSA to facilitate pulmonary gas exchange. We tested the hypothesis that CO₂ directly increases RSA magnitude in conscious humans even when changes in tidal volume (V_T) and breathing frequency (F_B), which may indirectly affect RSA (8, 9, 14), are prevented.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects. Seven healthy nonsmoking volunteers (6 male and 1 female), aged 20-23 yr, were studied following approval by our institutional ethics committee. None had a history of cardiopulmonary disease or was taking any medication. All subjects gave written informed consent.

Measurements. Experiments were performed at a room temperature of 23-25°C, and at least 3 h after the subjects ate their last meal and drank caffeine-containing beverages. Subjects were studied in the sitting position and breathed gas from a spirometer through a face mask with a nonrebreathing valve (Fig. 1). The electrocardiogram (ECG; model 78342A, Hewlett-Packard; Wilmington, DE), airway PCO₂ (Capnomac, Datex; Helsinki, Finland), and expiratory flow (model SC 520, Vacumed; Ventura, CA) were measured and digitized at 1 kHz on a personal computer by an analog-to-digital converter (model DI200, Dataq Instruments; Akron, OH). Breath-by-breath V_T and F_B were measured by integrating the expiratory flow signal. Oxygen saturation (Sp_O2) was measured using a pulse oximeter (model 251 Pulse Oximeter, Datex) with a probe on the left ear and recorded every 30 s.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Schematic of breathing circuit used for both determining the minute ventilation at an end-tidal PCO2 (PETCO2) of 50 mmHg and maintaining tidal volume (VT) constant during measurements. For details, see text. Because gas flow was continuous and the ratio of inspiratory to expiratory duration was 1:1, each VT was composed of the gas flow during the actual inspiration and that to the spirometer during the previous expiration. Thus the difference in spirometer volume between end-inspiration and end-expiration represented only half of the VT. FICO2, fraction of inspired CO2; E, minute ventilation.

Protocol. Data were obtained at three different levels of PET_CO2 (30, 40, and 50 mmHg), all at constant V_T, F_B, and fraction of inspired O₂ (FI_O2) (0.21). We prevented the ventilation of the subjects from varying according to the PCO₂ as follows. Before the experiment, the spontaneous minute ventilation (E) at a PET_CO2 of 50 mmHg was determined in each subject at a fixed F_B of 0.25 Hz (15/min). During the rest of the experiment, E was maintained at this level and F_B was maintained at 0.25 Hz across all PCO₂. F_B was kept constant by having subjects breathe in synchrony with a computer-generated signal at the same frequency (inspiratory-to-expiratory duration ratio = 1:1) heard over headphones. V_T was kept constant by having each subject maintain the end-inspiratory and end-expiratory positions of the spirometer bell constant while gas flow into the spirometer was set equal to the individual's E previously determined at a PET_CO2 of 50 mmHg. The three target levels of PET_CO2 were obtained by manipulating FI_CO2 between 0 and 0.06. All subjects practiced the maneuvers at various levels of PET_CO2 before actual measurements until they were able to perform them without discomfort. The order of conditions (PET_CO2) was randomized among subjects and data were collected for at least 3 min after stabilization at each condition.

Assessment of RSA magnitude. Fluctuations in heart period, including those caused by RSA, were assessed quantitatively by power spectrum analysis of the R-R intervals of the ECG. All R wave peaks were detected automatically with a fast peak-detection algorithm. The ECG waveforms, with markers indicating the positions of detected R wave peaks, were visually inspected for ectopic beats and artifacts. Any errors in R wave detection were edited manually, although no subject showed ectopic beats or artifacts. R-R interval sequences thus obtained were interpolated with a cubic spline function and resampled at 1 Hz to obtain time series with equidistantly spaced data points. Using a previously reported method (11, 12), we measured powers of autoregressive spectral components over the frequency range of 0.04-0.15 Hz and at 0.24-0.26 Hz, referred to as the low-frequency (LF) and high-frequency (HF) components, respectively (3). In the present study, a peak in the power spectrum was observed in the HF range in all subjects. The mean amplitude (square root of doubled power) of the HF component was used as a quantitative assessment of RSA (RSA magnitude). Mean R-R interval and the LF-to-HF power ratio (LF/HF) were used as indexes of cardiac autonomic tone and autonomic function, respectively.

Statistical analysis. One-way repeated-measures analysis of variance (ANOVA), followed by multiple comparisons with the Bonferroni correction, was used to evaluate the effects of PET_CO2 on variables. Data are presented as means ± SE. A P value < 0.05 was considered to indicate significance.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All subjects were able to maintain F_B and V_T constant during measurements at the three different levels of PET_CO2 (Table 1). Figure 2 shows the results obtained from a representative subject. Despite increasing levels of PET_CO2 (the top border of the airway PCO₂ trace), F_B and V_T did not change. At all three levels of PET_CO2, the power spectra of R-R interval showed a narrow peak at 0.25 Hz of the HF component. The area under the peak corresponding to the magnitude of RSA increased with the level of PET_CO2, as expected from the increasing magnitude of R-R interval fluctuation.

View larger version (31K): [in this window] [in a new window]   Fig. 2.   Changes in R-R interval, airway PCO2, VT, and autoregressive power spectra of R-R interval at PETCO2s of 30, 40, and 50 mmHg in a representative subject. The magnitudes of respiratory sinus arrhythmia (RSA) measured as the mean amplitude of the high-frequency (HF) component (0.24-0.26 Hz) were 65, 104, and 123 ms, respectively. The mean amplitudes of the low-frequency (LF) component (0.04-0.15 Hz) were 57, 58, and 122 ms, respectively. PSD, power spectral density.

Despite constant F_B and V_T, the magnitude of RSA increased with PET_CO2 in six of seven subjects and repeated-measures ANOVA detected significant differences at the three levels of PET_CO2 (P < 0.001 for all, Fig. 3). Mean R-R interval also increased with PET_CO2, but a significant difference was detected only between PET_CO2 of 30 and 40 mmHg (P < 0.05). No difference was observed in LF amplitude (42 ± 3, 42 ± 5, and 54 ± 11 ms for PET_CO2s of 30, 40, and 50 mmHg) and the changes in LF/HF did not reach significance.

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Effect of PETCO2 on the magnitude of RSA, mean R-R interval, and LF-to-HF ratio (LF/HF) of R-R interval power spectral components. Left, values in individual subjects; right, average values ± SE for the 7 subjects. Overall significances of the effect of PETCO2, according to repeated-measures ANOVA, are P < 0.001, P < 0.001, and not significant for the magnitude of RSA, mean R-R interval, and LF/HF, respectively. *P < 0.05 vs. 40 mmHg (multiple comparisons by the Bonferroni method).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Increases in PET_CO2 from 30 to 50 mmHg directly increased RSA magnitude when V_T and F_B were fixed. On the other hand, Sp_O2 and mean R-R interval between a PET_CO2 of 40 and 50 mmHg did not change. Given the potential role of RSA in enhancing pulmonary gas exchange (13, 22), our results are consistent with the idea that increased demand for pulmonary CO₂ elimination directly enhances gas exchange via increases in RSA magnitude.

Despite many studies of RSA, there are few studies of the effects of CO₂ alone, particularly in conscious animals or humans. In unanesthetized trained dogs, Yasuma and Hayano (22) reported that hypercapnia (PET_CO2 up to 54 mmHg) increases RSA magnitude by 62% with no concomitant changes in mean R-R interval. Our present observations are consistent with their findings. However, because their dogs increased V_T and F_B 3.7- and 1.2-fold, respectively, in response to hypercapnia, and because RSA magnitude increases with increasing V_T and decreasing F_B (8, 11, 14), their results provide no evidence for a direct effect of CO₂ on RSA magnitude. Yen et al. (23) reported that vagal outflow to the heart increases with hypercapnia and decreases with hypocapnia in unanesthetized decerebrate cats, although they did not report changes in RSA. Our present observations seem consistent with theirs, provided RSA magnitude reflects the tonic level of vagal activity (12, 16). In fact, the significant decreases in both RSA magnitude and mean R-R interval we observed at a PET_CO2 of 30 mmHg are consistent with withdrawal of vagal tone. Moreover, even the increased RSA magnitude in the absence of changes in mean R-R interval at a PET_CO2 of 50 mmHg may be compatible with increased vagal tone if the vagal effect on R-R interval was offset by a concomitant increase in sympathetic activity in response to the hypercapnia.

The most likely mechanism responsible for increased RSA magnitude in our study is chemostimulation that enhances respiratory modulation of vagal outflow. RSA arises primarily as a result of respiratory modulations of vagal outflow to the heart, mainly due to inhibition of vagal cardiac preganglionic neurons by medullary inspiratory neurons (10, 20) and phasic inspiratory gating by pulmonary stretch receptor input of excitatory inputs to these neurons (1, 5). Stimulation of carotid chemoreceptors by increased Pa_CO2 has primarily an excitatory effect on vagal preganglionic neurons to the heart, the so-called primary response, in the expiratory phase (6, 18). Hence, in the present study, CO₂ likely increased vagal outflow during expiration. In contrast, during the inspiratory phase, vagal outflow would have been low because of the CO₂-induced increase in respiratory drive, that is, greater inhibition by inspiratory neurons. In addition, gating of vago-excitatory inputs by slowly adapting pulmonary stretch receptors may be greater at increased Pa_CO2 (9). The net effect will depend on the degree to which the primary excitatory effect is offset by the secondary inhibitory effect from both medullary inspiratory neurons and pulmonary stretch receptors (18). Taken together, increased CO₂ enhances respiratory modulation of vagal outflow, even in the absence of changes in the overt ventilatory (V_T and F_B) responses to increases in CO₂. These putative mechanisms responsible for the CO₂-induced increase in RSA magnitude may also explain why changes in RSA magnitude do not always parallel changes in vagal tone, that is, the net effect of excitatory and inhibitory inputs.

Regulation of RSA magnitude by Pa_CO2 could complement CO₂-modulated changes in airway smooth muscle tone in controlling dead space. Hypercapnia both decreases tracheal diameter in anesthetized dogs (7) and causes bronchoconstriction in decerebrate cats (15). In anesthetized dogs, Hayano et al. (13) also showed that RSA reduces physiological dead space, that is, the alveolar dead space, by matching perfusion to ventilation during each respiratory cycle. Our results, in combination with these earlier findings, suggest that changes in Pa_CO2 are accommodated by dynamically altering both anatomical and alveolar dead space to modify the efficiency of CO₂ elimination.

In conclusion, increases in RSA magnitude with hypercapnia were not, in conscious humans, associated with changes in Sp_O2 or, at PET_CO2s between 40 and 50 mmHg, mean R-R interval. Changes in RSA magnitude did, however, parallel increases in PET_CO2 in the absence of changes in V_T and F_B. Our results are consistent with our hypothesis that increased CO₂ directly increases RSA magnitude even when changes in V_T and F_B are prevented. Given the potential role of RSA in improving pulmonary gas exchange, our results suggest that increased PCO₂ directly enhances RSA to facilitate CO₂ elimination.