INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

IT HAS BEEN PROPOSED THAT the oxygen uptake (O₂) response to the onset of exercise reflects the regulation of oxygen transport to the tissue (10, 15) and oxygen utilization in muscle tissue (2, 8, 24). Grassi et al. (6) recently reported that peripheral O₂ diffusion does not limit the muscle O₂ response to exercise onset in isolated canine muscle. They used hyperoxia and intra-arterial administration of RSR-13, which induces a rightward shift of the oxygen dissociation curve of Hb, to increase the O₂ diffusion in the muscle tissue. The result of their study clearly indicated that O₂ diffusion is not a limiting factor in in situ dog muscle preparation, but their study did not examine whether O₂ diffusion regulates the O₂ response. Even if we apply their data to humans, it is unclear whether impaired O₂ diffusion impairs the O₂ response. To resolve these issues, it is necessary to determine whether impeding the O₂ diffusion delays the O₂ response at the onset of exercise.

Hyperpnea increases the CO₂ output from the lungs. This excess CO₂ output allows the end-tidal CO₂ partial pressure (PET_CO2) to fall and decreases the arterial CO₂ tension (Pa_CO2), which leads to an increase in the arterial pH. Therefore, hyperpnea brings about hypocapnia and consequent respiratory alkalosis (21, 23). This should induce a leftward shift of the oxygen dissociation curve of hemoglobin (Hb), impeding the unloading of O₂ in a working muscle.

We hypothesized that voluntary hyperpnea slows down the O₂ kinetics at the onset of square-wave exercise by a leftward shift of the oxygen dissociation curve of Hb. To test the role of O₂ unloading in muscle capillaries in the O₂ response, we investigated the effect of respiratory alkalosis induced by voluntary hyperpnea on the O₂ response at the onset of exercise. Our preliminary study revealed that hyperpnea slowed the O₂ response to the exercise onset (n = 8, P < 0.05; unpublished data). However, although the O₂ kinetics are decelerated by the hyperpnea, it is impossible to discriminate between the effects of intrathoracic pressure swing and additional work of respiratory muscle produced by ventilation and that of a leftward shift of the oxygen dissociation curve of Hb. To discriminate between the effects of additional ventilation and changes in PET_CO2, we established two comparisons: 1) normocapnia and hypocapnia induced with hyperpnea to test the effect of O₂ unloading in muscle tissue, and 2) normocapnia with and without hyperpnea to test the effect of additional ventilation. Additionally, the effects of hypercapnia, which may induce the Bohr effect due to increased Pa_CO2, were investigated.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Six healthy males (23.5 ± 2.4 yr, 173.3 ± 5.2 cm, 66.7 ± 5.4 kg, means ± SD) participated in this study. All subjects received an explanation of the study and gave informed consent before participation.

Each subject performed a 20 W/min incremental ramp-exercise test on an electromagnetically braked cycle ergometer (model 232-C, Combi, Japan), to determine the ventilatory threshold (VT) with gas-exchange criteria and maximal oxygen uptake (O_{2 max}). The VT was determined as the O₂ at which the nonlinear increase of carbon dioxide output (CO₂) and minute expiration (E) plotted against O₂ was observed. The O_{2 max} and VT of the subjects were 3.5 ± 0.2 l/min and 2.0 ± 0.2 l/min (means ± SD), respectively.

The subjects performed square-wave exercise protocols on the cycle ergometer. Each protocol consisted of an abrupt work increase for a 6-min period at a work rate corresponding to 80% of the VT of each individual (129 ± 16 W) after an initial 3-min period of 20-W cycling. The subjects kept a constant pedaling frequency of 60 rpm during the cycling. The respiratory rate (FR) was maintained at 30 breaths/min throughout each trial. Four types of trials were conducted. In the control (Con) trial, the tidal volume was not controlled, but the FR was controlled with the inhalation of normal room air (FI_CO2, 0.03%). In the hypocapnia (Hypo) trial, the subjects controlled their E at 60-70 l/min, i.e., hyperpnea, for 2 min before and during 80% VT exercise with room air (FI_CO2 = 0.03%). In the normocapnia (Normo) trial, the subjects controlled their E at 60-70 l/min, i.e., hyperpnea, for 2 min before and during 80% VT exercise with a high-fraction CO₂ gas (FI_CO2 = 3.00%) to prevent the fall of PET_CO2. In the Hypo and Normo trials, the subjects kept their tidal volume at >2 l with feedback from a respiromonitor and instruction from a study staff member. In the hypercapnia (Hyper) trial, the ventilation was eupnea, but the FR was controlled with a high-fraction CO₂ gas inspiration (FI_CO2 = 3.00%). The inspiratory gas for the Normo and Hyper trials contained 3% CO₂ and 21% O₂ and N₂ balance. In all trials, the subject inspired through a Y-shaped valve from a Douglas bag filled with the gas. Each subject completed two repetitions of each trial on different days. The order of the trials was randomized.

The subjects breathed through a face mask connected to a hot wire flowmeter (RM-300; Minato Medical Sciences, Japan) for the measurement of respiratory flow. The flowmeter was calibrated using a 2-l syringe. A small sample (1 ml/s) of respired gas was withdrawn continuously from the mask and analyzed for O₂ and CO₂ with a mass spectrometer (WSMR-1400; Westron, Japan). The mass spectrometer was calibrated with fresh air and precision gas (O₂ 15%, CO₂ 5%). The time delay between the flow and gas concentration signals was calculated to obtain breath-by-breath data. The heart rate (HR) was measured by use of standard bipolar leads (CM5) with an electrocardiogram monitor (OEC-6201; Nihon-Kohden, Japan).

The data of O₂, CO₂, E, HR, FR, and PET_CO2 were interpolated at 1-s intervals and averaged for each subject and trial. To determine the kinetic variables of the increases in O₂ and HR at the onset of exercise, nonlinear least-squares fitting was applied using a computerized regression algorithm to a single component exponential model (15): F(t) = baseline + A{1  exp[(t  TD)/]}; where F(t) represents the O₂ and HR at time t; A is the steady-state increase from 20-W pedaling; and TD and t are the time delay and time constant from the onset of workload, respectively. The TD was set at >0 s for calculation. The mean response time (MRT; sum of TD and t) was used to estimate the response of observed variables. The O₂ response during the first 15 s after the increase of work rate was ignored in the regression to obtain O₂ kinetics at phase II (24). The O₂ response during this period (phase I) depends on the pulmonary blood flow rather than reflecting the time course of the tissue gas exchange (2, 24). The model fitting was applied to the data from 1 min before to 6 min after the work increase.

Values are expressed as means ± SD. A single group repeated-measures ANOVA was applied to compare the responses. When a significant effect was noted, Fisher's protected least squares difference post hoc test was applied. The significance level was set at P < 0.05. These statistical analyses were performed with SAS software (release 6.12, SAS Institute; Cary, NC) at the Osaka University Computation Center.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Representative responses of gas exchange variables and HR in the Con trial are shown in Fig. 1A. O₂, CO₂, E, HR, and PET_CO2 rose to new steady-state levels within 3 min after the work increase. The FR was maintained at 30 breaths/min during the trial. The mean of the steady-state (last 1 min of exercise) E during exercise was 52.2 ± 4 l/min.

View larger version (32K): [in this window] [in a new window]   Fig. 1.   Time courses of cardiorespiratory variables in control (Con; A) and hypocapnia (Hypo; B) trials in 1 subject. Solid vertical line indicates 80% ventilatory threshold (VT) exercise onset; dashed vertical line indicates onset of hyperpnea. In B, time course of Con trial is superimposed. PETCO2, end-tidal CO2 partial pressure; FR, respiratory rate; HR, heart rate; E, expired minute ventilation; CO2, CO2 uptake; O2, O2 uptake; bpm, beats/min.

Figure 1B shows the time course of the individual cardiorespiratory variables in the Hypo trial in the subject shown in Fig. 1A. The FR was maintained at 30 breaths/min during the trial, and E was controlled at ~70 l/min throughout the exercise. O₂ approached the previous baseline levels after the abrupt transient increase at the onset of hyperpnea and thereafter slowly increased to the steady-state level after the exercise increase. CO₂ and HR rose to new steady-state levels after the onset of hyperpnea and then increased after the work increase. The PET_CO2 fell markedly at the start of hyperpnea and reached a steady state before the work increase and then slightly increased after the work increase.

Figure 2A shows the time course of the individual cardiorespiratory variables in the Normo trial in the subject shown in Fig. 1. The FR was maintained at 30 breaths/min during the trial, and E was controlled at ~70 l/min throughout the exercise. CO₂ and HR rose to new steady-state levels after the onset of hyperpnea. PET_CO2 fell markedly at the start of hyperpnea and reached a steady state before the work increase. However, in this case, CO₂ gas was added to maintain the PET_CO2 normal level (i.e., normocapnia).

View larger version (32K): [in this window] [in a new window]   Fig. 2.   Time courses of cardiorespiratory variables in normocapnic (Normo; A) and hypercapnia (Hyper; B) trials in subject shown in Fig. 1. In Normo and Hyper trials, CO2 was added to inspiratory gas (3%). Solid vertical line indicates 80% VT exercise onset; dashed vertical line indicates onset of hyperpnea.

Figure 2B shows the time course of the individual cardiorespiratory variables in the Hyper trial in the same subject. The FR was maintained at 30 breaths/min during the trial. The mean of steady-state E during exercise was 65.6 ± 7.4 l/min. The effect of the CO₂ addition to the inspiratory gas was enough to keep the PET_CO2 higher (i.e., hypercapnia).

The averaged responses of all subjects for E, PET_CO2, HR, and O₂ in these trials are shown in Fig. 3. The time courses were generally similar to the individual examples presented in Figs. 1 and 2. In the Hypo and Normo trials, E was controlled at 60-70 l/min throughout the exercise. E was significantly larger in the Hyper trial than in the Con trial due to the CO₂ added to the inspiratory gas. The averaged values of PET_CO2 for 30 s before the work increase were 38.2 ± 3.2 (Con trial), 21.9 ± 0.9 (Hypo trial), 36.4 ± 0.8 (Normo trial), and 45.0 ± 1.7 mmHg (Hyper trial). In the Hypo trial, PET_CO2 was significantly lower than in the other trials due to hyperpnea. PET_CO2 was maintained at normal levels due to the CO₂ gas in the Normo trial. In the Hyper trial, PET_CO2 was significantly higher than in the other trials due to the CO₂ added to the inspiratory gas. The O₂ kinetics were clearly delayed in the Hypo trial.

View larger version (34K): [in this window] [in a new window]   Fig. 3.   Averaged responses of E, PETCO2, HR, and O2 for all subjects (n = 6). Solid vertical line indicates 80% VT exercise onset. Error bars represent SD displayed each 15 s.

Table 1 summarizes the data for the kinetics variables of O₂ and HR. The hyperpnea and/or CO₂ addition to the inspiratory gas did not affect the baseline and gain of O₂. The hyperpnea significantly increased the baseline of HR and significantly decreased the gain of HR, and, consequently, there was no significant difference in the steady-state values of HR.

View this table: [in this window] [in a new window]   Table 1.   Kinetic variables of O2 and HR responses

The ANOVA revealed a significant effect of trial on the MRTs of O₂. The MRT of O₂ was significantly longer in the Hypo trial than in the Con, Normo, and Hyper trials (P < 0.05) according to the post hoc comparison. In addition, the time constant was significantly longer in the Hypo trial (37.1 ± 11.5 s) than in the Con (21.0 ± 6.5 s), Normo (28.1 ± 7.2 s), and Hyper (23.8 ± 6.3 s) trials. The MRT and time constant were significantly longer in the Normo trial than in the Con trial. The MRT and time constant in the Hyper trial were not significantly different from those in the Con and Normo trials.

The MRT of HR was significantly longer in the Normo and Hyper trials than in the Con trial. There was no significant difference in this variable among the Normo, Hypo, and Hyper trials.

Figure 4 shows the individual values of the MRT of O₂ and HR. All subjects but one had a longer MRT of HR and O₂ in the Hypo trial than in the Con trial (P < 0.05). The effect of lung movement, which may include the drop in PET_CO2, significantly delayed both the MRT of HR and O₂ (Fig 4B, P < 0.05). Figure 4C shows the single effect of the drop in PET_CO2 on the MRT of HR and O₂. The drop significantly delayed the MRT of O₂ in all subjects (P < 0.05) but did not affect the MRT of HR (P > 0.1). The effect of the increase in PET_CO2 significantly delayed the MRT of HR (P < 0.05) but did not affect the MRT of O₂ (P > 0.1, Fig. 4D).

View larger version (42K): [in this window] [in a new window]   Fig. 4.   Comparisons between each pair of trials. Individual values of mean response time (MRT) of O2 and HR are shown. Open circles indicate data of each individual, and large filled circles indicate means ± SD in each trial. A: Con vs. Hypo; B: Con vs. Normo; C: Normo vs. Hypo; and D: Con vs. Hyper. NS, not significant.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The main findings in this study were that 1) the hyperpnea procedure significantly slowed the HR and O₂ kinetics at the onset of the work increase, 2) the drop in PET_CO2 induced an additional slowing of the O₂ response but not of the HR response to the onset of the work increase, and 3) the increase in PET_CO2 did not significantly change the O₂ kinetics at the onset of the work increase. These findings suggest that hypocapnia decelerates O₂ kinetics at the onset of exercise by a leftward shift of the dissociation curve of Hb.

Effect of Hyperpnea Manipulation on Cardiorespiratory Responses to Onset of Exercise (Normo vs. Con)

It is possible that the delay is due to delay of the central and/or peripheral circulatory response. There is evidence that circulation regulates the O₂ response (10, 15). However, evidence against a role for the central circulatory response in the regulation of the O₂ response has been obtained in healthy and heart transplant subjects (7, 9). Grassi et al. (7) reported that changes in the cardiac output response induced by repeated exercise did not affect the O₂ response to the work increase in heart transplant patients. Hayashi et al. (9) suggested that O₂ responses are regulated by local blood distribution rather than by the central circulatory response during the disorder of vagal withdrawal response to the onset of exercise with facial cooling in healthy subjects. Furthermore, Shoemaker et al. (20) and Hughson et al. (11) suggested that an inadequate blood flow delayed the O₂ kinetics. The vigorous movement of respiratory muscle needs O₂, and, consequently, blood flows to these muscles (1). These results support the speculation that the slower O₂ kinetics in the present Normo trial compared with the Con trial were due to an inadequate local blood flow distribution rather than central circulation, although the present results did not clearly confirm this.

There is no significant difference between hyperpnea and normopnea trials for baseline and steady state in the O₂. Previous studies (1, 17) reported higher O₂ during hyperpnea manipulation than eupnea. The rate of increase in O₂ on a 1-liter increase of ventilation ranged from 0.5 to 3 ml/l in terms of increasing ventilation (17). In the present study, hyperpnea manipulation increased E by 30 l/min during 20-W cycling and 10 l/min during 80% VT cycling. The increase of O₂ was estimated to be 90 ml/min during 20-W cycling and 30 ml/min during 80% VT cycling, when the highest value was applied to the estimation. With strict control of breathing, Aaron et al. (1) measured the oxygen cost of hyperpnea. With the use of their regression, the work of ventilation is calculated to be 47 J/min at 60 l/min of E. This 47 J/min is 3 W, which costs ~45 ml/min of O₂ if the efficiency of respiratory muscle is 20%. From these previous studies, oxygen cost for 60 l/min of E is estimated to be much less than 50 ml/min. So it is not strange that such a small difference was not detectable.

Effect of Hypocapnia on Cardiorespiratory Responses to Onset of Exercise (Hypo vs. Normo)

The hyperpnea brought about an increase of CO₂ and a drop in PET_CO2 in the Hypo trial compared with the Normo trial (Figs. 1B and 3). This drop in PET_CO2 confirms a decrease of the Pa_CO2, which leads to an increase in arterial pH. Therefore, the hyperpnea induced hypocapnia and respiratory alkalosis (21, 23). Pa_CO2 has been estimated with good accuracy by using the Jones equation (12). The averaged values of Pa_CO2 calculated from PET_CO2 for 30 s before the work increase by the Jones equation were 39.8 ± 3.3 (Con trial), 25.1 ± 1.2 (Hypo trial) and 38.1 ± 1.1 mmHg (Normo trial). In the Hypo trial, 25 mmHg of Pa_CO2 corresponds to 7.57 of pH, and in the Normo trial, 38 mmHg corresponds to 7.41 of pH according to the acid-base chart for arterial blood (21). In the Hypo trial, the subjects performed the exercise under the hypocapnia and alkalosis condition.

The capacity to release oxygen from the Hb at working muscle tissue is determined by the oxygen dissociation curve of Hb and the tissue PO₂; that is, the capillary-to-tissue PO₂ difference under the condition wherein Pa_O2 is the same. In the present study, the PO₂ in working muscle can be assumed to not be different between the trials, supposing the same amount of O₂ was used at the same workload. The oxygen dissociation curve of Hb was influenced by Pa_CO2, arterial pH, and temperature. The hyperpnea could not influence the muscle temperature. If one supposes that the Pa_O2 in arterial blood was the same among the trials and that the mean capillary PO₂ was 30 mmHg (16), the difference in the saturation between arterial and capillary blood clearly decreases in the Hypo trial. The effects of hypocapnia and respiratory alkalosis shifted the oxygen dissociation curve of Hb leftward and consequently impaired the diffusion gradient for O₂ between the capillary blood and the exercising muscles. It is known that the O₂ half-saturation pressure of Hb (P₅₀) under standard conditions (37°C, pH = 7.4) is 26.6 mmHg (18) and, in the relationship log P₅₀/pH, 0.48 for a pH of 0.1 units is the standard value. According to these values, the P₅₀ is estimated to be 27.7 mmHg in the Normo and 20.4 mmHg in the Hypo trial. We propose that the oxygen diffusion in muscle tissue thus became impaired, which resulted in the decelerated kinetics of O₂ at the onset of exercise. We, therefore, suggest that the O₂ unloading from Hb to the muscle tissue is an important component of the O₂ response to the exercise onset.

Koike et al. (13) reported slowing of O₂ kinetics on inhalation of low concentrations of carbon monoxide. Their study included the effect of a leftward shift of the oxygen dissociation curve and decreased blood O₂ content. It was impossible to discriminate between the effect of Hb content and the dissociation curve of Hb in the previous study. Hypo manipulation in the present study did not include the effect of the content of Hb that is able to bind O₂.

In a modeling study, Cochrane and Hughson (3) reported that the balance between O₂ transport and utilization is very delicate in O₂ kinetics. Their model included the Bohr effect on the oxygen dissociation curve of Hb. However, they did not investigate the effect of changes in Pa_CO2 and pH on the shift in the oxygen dissociation curve of Hb or on O₂ kinetics.

Grassi et al. (6) recently concluded that the enhancement of peripheral O₂ diffusion did not affect the muscle O₂ response at the work onset in isolated canine muscle. They also suggested that a faster O₂ delivery does not affect the O₂ response in isolated muscle (5). Their series of studies in isolated muscle suggested metabolic control of the O₂ response. This does not contradict our results. Their results showed that O₂ diffusion is not a rate-limiting factor under control conditions. Whereas the present results showed that the decreased peripheral O₂ diffusion from Hb did slow the O₂ response to the work increase, this did not imply that the O₂ diffusion is a rate-limiting factor. It must be noted that the present results merely showed the role of O₂ diffusion as a regulator to maintain the O₂ uptake response. This means that the O₂ uptake slows when this regulator does not work properly.

Shiojiri et al. (19) reported that the poor extraction of O₂ and reduced muscle blood flow in exercising muscle at reduced muscle temperatures contributed to the delayed adjustment of O₂. Under the reduced muscle temperature condition, the O₂ extraction was reduced by the temperature-dependent leftward shift of the oxygen dissociation curve of Hb. The present hypothesis that oxygen diffusion in the muscle tissue plays a major role in O₂ kinetics is partly supported by their findings. However, the muscle blood flow was also altered by cold-induced vasoconstriction. The effect of the blood flow distribution in exercising muscle concomitant with a previous manipulation contributes to the adjustment of O₂ at the onset of exercise. Therefore, there is less of a positive basis for a role for oxygen diffusion in the muscle tissue in the O₂ response. In the present study, the O₂ kinetics at the onset of exercise were slowed under the Hypo condition, although the HR kinetics were similar to those observed under the Normo condition. In addition, there is less possibility that the blood flow distribution induced the difference in O₂ kinetics between the Hypo and Normo trials.

Ward et al. (22) reported that E and CO₂ dynamics were slowed considerably after volitional hyperpnea and that the HR dynamics were unaffected, whereas the O₂ dynamics were slowed only slightly. This result is inconsistent with our present findings. We observed that HR kinetics were not affected by hyperpnea, similar to their findings. However, the O₂ kinetics were significantly slowed by hyperpnea. This difference is due to differences in experimental design. In the study by Ward et al., hyperpnea was induced only before exercise, and there was a brief interval between the cessation of hyperpnea and exercise onset. In contrast, hyperpnea was induced throughout the exercise in our study. The hyperpnea during exercise kept the PET_CO2 low for a long time, as shown in Fig. 3. Hypocapnia occurred throughout the exercise. The long duration of hypocapnia contributed to the slowed O₂ kinetics at the onset of exercise.

Limitations. This hypocapnia procedure clearly revealed the effect of O₂ diffusion in muscle tissue on the O₂ response. However, this study has some limitations. First, there might be an effect from some metabolic processes that hypocapnia and changes in pH might alter.

Second, hypocapnia would affect the vascular resistance. Kontos et al. (14) measured arterial and venous pressure during hypocapnia with and without increased ventilation to calculate the vascular resistance. They reported that the decreased Pa_CO2 could increase vascular resistance. Even if vascular resistance had been increased by respiratory alkalosis, the blood flow distribution might not occur in a specific part of the body, because respiratory alkalosis affects the whole body. However, we cannot completely rule out the possibility that the ability to vasodilate appropriately was impaired by the hypocapnia.

Effect of Hypercapnia on Cardiorespiratory Responses to Onset of Exercise (Hyper vs. Normo)

Grassi et al. (6) reported that increased O₂ diffusion by the Bohr effect did not affect the O₂ response in isolated in situ canine muscle. When this increased O₂ diffusion is applied to human subjects, though they stated that the application of the study should be limited to muscles with a high aerobic potential, it is plausible that the Bohr effect induced by hypercapnia did not affect the O₂ response.

In addition, the changes of Pa_CO2 and pH induced by the procedure were smaller in the Hyper trial than the Hypo trial. This would be due to the stronger regulatory system for hypercapnia compared with that for hypocapnia, because the increase in Pa_CO2 can be regulated downward easily by the increase in E. The increase of Pa_CO2 in the present Hyper trial might not be great enough to change the oxygen dissociation curve for speeding O₂ kinetics. It thus is still unclear whether O₂ diffusion increased by the Bohr effect with hypercapnia affects the O₂ response in humans.

In summary, the O₂ kinetics were significantly decelerated by hypocapnia, although the HR kinetics did not change significantly. This finding confirmed that the leftward shift of the oxygen dissociation curve of Hb caused by hypocapnia leads to the slowed kinetics of oxygen diffusion in active muscle, resulting in the decelerated O₂ kinetics at the onset of exercise. These results suggest that the O₂ diffusion in muscle tissue is important for regulating the O₂ response to exercise onset.

Perspectives