We tested the hypothesis that vagal withdrawal plays a role in the rapid (phase I) cardiopulmonary response to exercise. To this aim, in five men (24.6 ± 3.4 yr, 82.1 ± 13.7 kg, maximal aerobic power 330 ± 67 W), we determined beat-by-beat cardiac output (Q̇), oxygen delivery (Q̇aO2), and breath-by-breath lung oxygen uptake (V̇o2) at light exercise (50 and 100 W) in normoxia and acute hypoxia (fraction of inspired O2 = 0.11), because the latter reduces resting vagal activity. We computed Q̇ from stroke volume (Qst, by model flow) and heart rate (fH, electrocardiography), and Q̇aO2 from Q̇ and arterial O2 concentration. Double exponentials were fitted to the data. In hypoxia compared with normoxia, steady-state fH and Q̇ were higher, and Qst and V̇o2 were unchanged. Q̇aO2 was unchanged at rest and lower at exercise. During transients, amplitude of phase I (A1) for V̇o2 was unchanged. For fH, Q̇ and Q̇aO2, A1 was lower. Phase I time constant (τ1) for Q̇aO2 and V̇o2 was unchanged. The same was the case for Q̇ at 100 W and for fH at 50 W. Qst kinetics were unaffected. In conclusion, the results do not fully support the hypothesis that vagal withdrawal determines phase I, because it was not completely suppressed. Although we can attribute the decrease in A1 of fH to a diminished degree of vagal withdrawal in hypoxia, this is not so for Qst. Thus the dual origin of the phase I of Q̇ and Q̇aO2, neural (vagal) and mechanical (venous return increase by muscle pump action), would rather be confirmed.
- cardiovascular response
although our knowledge of the central (neural) control of the cardiovascular system at the exercise steady state is quite well established (19, 42, 57), how the circulatory readjustments upon exercise onset occur and match the increase in pulmonary oxygen uptake (V̇o2) is less understood, as are the mechanisms underlying this matching. The kinetics of V̇o2 at exercise onset were seen for a long time as reflecting essentially the metabolic adaptations in the working muscles (15, 31, 33). Some authors, however, soon identified two components of the V̇o2 kinetics: 1) a rapid, almost immediate phase (phase I) (5, 54, 55), which they attributed to an immediate increase in cardiac output (Q̇) at exercise start; and 2) a subsequent slower phase (phase II), to which they restricted the influence of muscle metabolic adjustments. The strongest support to this view came from the demonstration that the kinetics of Q̇ (12, 13, 16, 60) and arterial O2 flow (Q̇aO2) (27) are very rapid.
The concept of a close correspondence between V̇o2 and muscle O2 consumption was further undermined by the recent demonstration that, upon the onset of light exercise, the V̇o2 kinetics are faster than the kinetics of muscle O2 consumption estimated from the monoexponential decrease in phosphocreatine concentration (7, 14, 43). This would imply dissociation of the kinetics of V̇o2 and muscle O2 consumption, which should respond to different control mechanisms.
Our postulate is that the V̇o2 kinetics are dictated by the regulation of the systemic cardiovascular response to exercise, whereas the metabolic regulatory processes dictate only the kinetics of muscle O2 consumption. In this context, Fagraeus and Linnarsson (18) proposed that the rapid heart rate (fH) changes in exercise transients “are mediated through a withdrawal of vagal tone” (termed “vagal withdrawal” from this point), which can be defined as a quasi-immediate inhibition of vagal action on the sinus node at exercise start. In fact, they showed that the rapid phase of the fH kinetics was cancelled out under vagal blockade, whereas β-adrenergic blockade with propranolol did not affect it. In this study, we tested the hypothesis that vagal withdrawal also plays a major role in determining phase I kinetics of Q̇, Q̇aO2, and V̇o2. If this is so, then in acute normobaric hypoxia, wherein reduced vagal activity (8, 26) and increased sympathetic activity at rest have been postulated (21, 26, 58, 59), phase I would be either absent or at least less intense compared with normoxia.
With this hypothesis in mind, the aim of this study was to perform simultaneous determinations of the phase I kinetics of fH, Q̇, Q̇aO2, and V̇o2 upon exercise onset in normoxia and acute normobaric hypoxia. Such an experiment was never carried out in the past, to the best of our knowledge.
Five healthy, nonsmoking young male subjects took part in the experiments. They were 24.6 ± 3.4 yr old, 1.79 ± 0.09 m tall, and weighed 82.1 ± 13.7 kg. Their maximal O2 consumption and maximal aerobic mechanical power in normoxia were 4.42 ± 0.62 l/min and 330 ± 67 W, respectively. The corresponding values in hypoxia were 3.41 ± 0.83 l/min and 255 ± 78 W, respectively. All subjects were preliminarily informed of all procedures and risks associated with the experimental testing. Informed consent was obtained from each volunteer, who was aware of his right to withdraw from the study at any time without jeopardy. The study was conducted in accordance with the Declaration of Helsinki. The protocol was approved by the Comités d'Ethique des Hôpitaux Universitaires Genevois (Switzerland). The experiments were carried out at Geneva, Switzerland.
V̇o2 was determined on a breath-by-breath basis. The time course of O2 and CO2 partial pressures throughout the respiratory cycles were continuously monitored with a mass spectrometer (Balzers Prisma, Balzers, Liechtenstein) calibrated against gas mixtures of known composition. The inspiratory and expiratory ventilations were measured by an ultrasonic flowmeter (Spiroson; Ecomedics, Duernten, Switzerland) calibrated with a 3-liter syringe. The alignment of traces was corrected for the time delay between the flowmeter and the mass spectrometer. Breath-by-breath V̇o2 and CO2 output (V̇co2) were then computed off-line by means of a modified version of Grønlund's algorithm (9). Software purposely written under the Labview developing environment (Labview 5.0; National Instruments, Austin, TX) was used. The characteristics and physiological implications of Grønlund's algorithm are widely discussed elsewhere (9, 11, 27).
fH and arterial oxygen saturation (SaO2) were continuously measured using electrocardiography (Elmed ETM 2000; Heiligenhaus, Germany) and fingertip infrared oximetry (Ohmeda 2350; Finapres, Englewood, CO), respectively. In hypoxia, SaO2 data were corrected for time delay between lungs and fingertip (30). Continuous recordings of arterial pulse pressure were obtained at a fingertip of the right arm by means of a noninvasive cuff pressure recorder (Portapres; FMS, Amsterdam, The Netherlands). Beat-by-beat mean arterial pressure (P̄) was computed as the integral mean of each pressure profile using the Beatscope software package (FMS).
The stroke volume of the heart (Qst) was determined on a beat-by-beat basis by means of the model flow method (53), applied off-line to the pulse pressure profiles, again using the Beatscope software package. Beat-by-beat Q̇ was computed as the product of single-beat Qst times the corresponding single-beat fh. Correction for the inaccuracy of the method was applied as previously described (2, 27, 50). To this purpose, steady-state Q̇ values also were obtained by means of the open-circuit acetylene method (4) using a procedure that was previously described (27). Individual correction factors at rest and at each workload were calculated as previously described (27) and also applied during dynamic states with rapid changes in Q̇ (51).
Exercise was carried out on an electrically braked cycle ergometer (Ergometrics 800-S; Ergoline, Bitz, Germany). The pedaling frequency was recorded, and its sudden increase at the exercise onset and decrease at the exercise offset were used as markers to identify precisely the start and the end of exercise. The electromechanical characteristics of the ergometer were such as to permit workload application in <50 ms. All the signals were digitalized in parallel by a 16-channel analog-to-digital converter (MP100; Biopac Systems, Goleta, CA) and stored on a computer. The acquisition rate was 100 Hz.
Blood hemoglobin concentration ([Hb]) was measured using a photometric technique (HemoCue, Ängelholm, Sweden) on 10-μl blood samples from a peripheral venous line inserted in the left forearm. Blood lactate concentration ([La]b) was measured using an electroenzymatic method (Eppendorf EBIO 6666, Erlangen, Germany) on 20-μl blood samples from the same venous line. Arterial blood gas composition was measured with microelectrodes (Instrumentation Laboratory Synthesis 10, Lexington, MA) on 300-μl blood samples taken from an arterial catheter inserted in the left radial artery.
Experiments were first performed in normoxia and then in acute normobaric hypoxia (fraction of inspired O2, 0.11; inspired O2 partial pressure, 80 mmHg). In hypoxia, inspired gas was administered from high-pressure gas cylinders via an 80-liter Douglas bag buffer. The fraction of inspired O2 was monitored on the inspiratory line, close to the mouth. The gas flow from the cylinders was continuously adjusted to the subject's ventilation. Experiments in hypoxia were preceded by a 10-min period for gas store equilibration. The experimental protocol started with the performance of blood sampling and the measurement of acetylene Q̇ at rest, and then 2 min of quiet resting recordings were allowed, after which the exercise at 50 W started, for a duration of 10 min. Arterial blood gas composition and [La]b were measured at minute 5 and at the end of exercise. At minute 7, the measurement of Q̇ with the acetylene technique was initiated. The 50-W exercise was followed by a 10-min recovery, during which [La]b was measured at minutes 2, 4, and 6, and arterial blood gas composition was determined at minutes 5 and 10. The 100-W exercise was then carried out, for a 10-min duration, and with the same timing of events as at 50 W. A 10-min recovery followed, with the same characteristics as the previous one. The overall duration of this protocol was about 60 min, during which [Hb] was systematically measured at 1-min intervals.
Each subject repeated this protocol four times, in both normoxia and hypoxia. At each repetition, the performance of blood sampling for [Hb] determination was shifted by 15 s, as previously described (27), to obtain, after superposition of the four tests, an overall description of the changes in [Hb] on a 15-s time basis.
The superimposed time course of [Hb] was smoothed by a four-sample mobile mean, to account for interrepetition variability, and interpolated by means of a 6th degree polynomial, as previously described (27). The continuous SaO2 traces from the four repetitions were temporally aligned and superimposed by means of an ensemble average procedure. The resulting overall SaO2 trace was then interpolated by means of a 6th degree polynomial. The resulting functions, describing the time course of [Hb] and SaO2, were used to compute the time course of arterial O2 concentration (CaO2, ml/l) on an equivalent beat-by-beat time scale, established after the pulse pressure profile traces, as in a previous study (27).
The beat-by-beat fH, Qst, P̄, and Q̇ values from the four repetitions of each subject were aligned temporally by setting the time of exercise start as time 0 for the analysis of on kinetics. The data were then averaged on a beat-by-beat basis to obtain a single averaged, superimposed time series for each parameter and subject. Beat-by-beat Q̇aO2 was then calculated as (1)
Beat-by-beat total peripheral resistance (Rp) was calculated by dividing each P̄ value by the corresponding Q̇ value, on the assumption that the pressure in the right atrium can be neglected as a determinant of peripheral resistance.
Based on the conclusions arrived at in a previous study (27), the kinetics of V̇o2, Q̇, and Q̇aO2 were described by means of a two-phase model, whereby an exponential increase in flow (phase II) is preceded by a faster flow increase in the first seconds of exercise (phase I), which Barstow and Molé (5) also treated as an exponential. To compute the characteristic parameters of the exponential equations describing phase I, the four repetitions were interpolated to 1-s intervals (28) and then aligned temporally, as described above, and averaged to obtain a single superimposed time series (27). Since the tested hypothesis concerns specifically a phenomenon that takes place during phase I, in the results we report only the phase I parameters, neglecting the phase II parameters, to which the tested hypothesis does not pertain.
Data are means and standard deviations of the values obtained for each parameter from the average superimposed files of each subject, to account for interindividual variability. The effects of exercise intensity and hypoxia on the investigated parameters were analyzed separately using a one-tailed t-test for paired observations. Bonferroni correction was then applied. The parameters of the models were estimated by utilizing a weighted nonlinear least squares procedure (10), implemented under Labview (version 5.0; National Instruments, Austin TX). Initial guesses of the parameters of the model were entered after visual inspection of the data. The effects of exercise intensity and hypoxia on these values were investigated using a one-tailed t-test for paired observations. The results were considered significant if P < 0.025.
The [Hb], SaO2, and CaO2 values at rest and at the exercise steady state are reported in Table 1, together with arterial blood pH, Po2, and Pco2. The SaO2 and CaO2 values in hypoxia were lower than the corresponding values in normoxia. Arterial blood pH was higher in hypoxia than in normoxia and was unaffected by the exercise intensity in both conditions. Po2 and Pco2 were both lower in hypoxia than in normoxia. In the latter, they did not vary at exercise. In hypoxia, Po2 decreased during exercise (P < 0.025) and Pco2 tended to decrease (not significant, NS). In normoxia, [La]b was 1.3 ± 0.3 mM at rest and did not change at exercise. In hypoxia, [La]b was 2.0 ± 0.5 mM at rest and 2.3 ± 0.5 mM at 50 W (NS). At 100 W, [La]b increased to 3.7 ± 1.3 mM at minute 5 and 4.5 ± 1.7 mM at minute 10.
The mean values for Q̇, fH, Qst, P̄, Rp, Q̇aO2, and V̇o2, obtained at rest and at the exercise steady state at both powers, are reported in Table 2. At all metabolic powers, fH was higher in hypoxia than in normoxia. Qst was the same in hypoxia as in normoxia so that Q̇ resulted systematically higher in hypoxia than in normoxia. At rest, P̄ was lower in hypoxia than in normoxia, but this difference disappeared at exercise. As a consequence, systematically lower Rp values were found in hypoxia than in normoxia (NS at rest, P < 0.025 at 50 and 100 W). Q̇aO2 was not significantly different from normoxia at rest. At 50 and 100 W, however, Q̇aO2 turned out lower in hypoxia than in normoxia. V̇o2 was the same in hypoxia as in normoxia.
The time courses of fH, Qst, Q̇, P̄, and Rp upon the onset of 50- and 100-W exercise are shown in Fig. 1. Beat-by-beat data collected in the 15 s that preceded and in the 45 s that followed the start of exercise are shown in Fig. 1, to draw attention to phase I events. In normoxia, a steady state for fH appeared as soon as phase I was completed at 50 W, whereas a clear slower phase II increase was evident at 100 W. In hypoxia, the relative contribution of phase I to the fH response was less than in normoxia. The time course of Qst was the same in hypoxia as in normoxia. Thus the initial change of Q̇ in hypoxia, compared with normoxia, followed essentially the same patterns as for fH. The increase in P̄ was modest and slow, in both normoxia and hypoxia. Conversely, Rp underwent a sudden dramatic decrease, the amplitude of which was smaller in hypoxia than in normoxia.
The evolution of beat-by-beat fH as a function of beat-by-beat P̄ is shown in Fig. 2. For both normoxia and hypoxia, the resting values are located on the lower left side of the plot, and the exercise steady-state values are located on the upper right side. However, the resting values in hypoxia were displaced upward and leftward with respect to those in normoxia, as were the exercise values. In normoxia, at both workloads, the pattern of displacement of the baroreflex operational point from rest to exercise was dictated by the rapid increase in fH, as demonstrated by the small number of points required to attain the cluster of the fH and P̄ values at exercise. Similar patterns were observed in hypoxia, although 1) the size of the displacement of baroreflex operational point was larger than in normoxia, and 2) the number of beats required to complete this displacement in hypoxia (within 30 and 60 beats at 50 and 100 W, respectively) was greater (slower increase) than in normoxia (within 20 and 45 beats at 50 and 100 W, respectively).
In normoxia, since SaO2 was unchanged, the evolution of CaO2 followed the changes in [Hb]. In hypoxia, CaO2 underwent larger changes than in normoxia, which were dictated not only by the changes in [Hb] but also by the decrease in SaO2 in the exercise transient. In hypoxia, a steady CaO2 level lower than at rest was attained within 2 min.
The time courses of Q̇aO2 and V̇o2 upon the onset of 50- and 100-W exercise are reported in Fig. 3. The rate of readjustment of V̇o2 followed the same trend in hypoxia as in normoxia. In normoxia, it was slower than that of Q̇aO2. This difference disappeared in hypoxia, because the rate of readjustment of Q̇aO2 was slower in hypoxia than in normoxia.
The characteristic parameters describing the Q̇, Q̇aO2, V̇o2, and fH kinetics during phase I are presented in Table 3. For V̇o2, amplitude of phase I (A1) was the same in hypoxia as in normoxia. For Q̇ and Q̇aO2, A1 was significantly lower in hypoxia than in normoxia. In both normoxia and hypoxia, A1 was the same at 100 W as at 50 W. For Q̇aO2 and V̇o2, τ1 was unaffected by hypoxia. For Q̇, τ1 was shorter in hypoxia than in normoxia at 50 W but not at 100 W. For fH, A1 was lower in hypoxia than in normoxia at 50 W but not at 100 W. In normoxia, A1 was the same at 50 W as at 100 W, as it was in hypoxia. τ1 was higher in hypoxia than in normoxia at 100 W but not at 50 W. In hypoxia, τ1 was significantly greater at 100 W than at 50 W.
This study was carried out to test the hypothesis that vagal withdrawal determines the phase I kinetics of Q̇, Q̇aO2, and V̇o2 at exercise onset. This hypothesis implies that phase I would be either absent or eventually less intense compared with normoxia. The main finding of this study is that the amplitude of the phase I (A1) of the kinetics of Q̇ and Q̇aO2 at exercise onset was smaller in acute normobaric hypoxia, wherein reduced vagal activity (8, 26) and increased sympathetic activity at rest have been postulated (21, 26, 58, 59), than in normoxia, whereas its time constant τ1 was unchanged. No differences appeared concerning the phase I of V̇o2 kinetics. In hypoxia, the reductions in A1 are coherent with the concept of a lesser effect of postulated vagal withdrawal at exercise onset.
The increased sympathetic activity to the heart in acute hypoxia (21, 22, 44, 45), perhaps through peripheral chemoreceptor stimulation (22), may be sufficient to explain the higher fH in hypoxia than in normoxia, both at rest and at any given work level. Since Qst is unaffected by acute hypoxia, the increase in fH entails a corresponding increase in Q̇, as already demonstrated in several studies (1, 23, 49). The present data (see Table 2) are in full agreement with this picture.
A larger sympathetic activity, if directed to peripheral vessels as well, might also imply peripheral vasoconstriction and, hence, a greater Rp in hypoxia than in normoxia. However, this was not so in the present study, since Rp (Table 2) was lower in hypoxia than in normoxia, whether at rest or at the two investigated workloads, because a higher Q̇ was associated with an unchanged P̄. Rp was rarely looked at in hypoxia in the past, yet it was possible to compute it from some studies (1, 23, 49). The obtained data are coherent with those of the present study. Moreover, increased peripheral sympathetic activation in hypoxia, though providing a potent peripheral vasoconstriction stimulus, is not accompanied by increased leg vascular resistance at rest, which was rather found to be reduced compared with normoxia (22).
This apparent contradiction may be explained by admitting either of these three hypotheses : 1) hypoxemia reduces the sensitivity and increases the activation threshold of vascular sympathetic receptors (sympatholysis); 2) hypoxemia superimposes a vasodilating stimulus in peripheral circulation; or 3) the intensity and quality of sympathetic output may differ among various target organs in hypoxia. The first hypothesis was recently contradicted by the demonstration that the vascular response to tyramine is not reduced in hypoxia (56). The two other hypotheses were supported by the observation of β2-mediated vasodilatation in resting skeletal muscle in hypoxia due to increased adrenaline release (52). Peripheral O2 sensing mechanisms may be implied in this effect. For instance, according to Stamler et al. (48), the conformation of the reduced hemoglobin determines the rise of nitric oxide (NO) in blood with consequent vasodilatation. In contradiction to this, however, Weisbrod et al. (52) failed to show a reduction of peripheral hypoxic vasodilatation after NO synthase blockade. Other SaO2-related mechanisms were postulated, which would imply ATP-mediated vasodilatation (32). A clear picture of the events that lead to decreased Rp is still far from being established.
Phase I kinetics.
The hypothesis that vagal withdrawal determines phase I relies essentially on observations made on fH upon exercise onset in normoxia, the kinetics of which were similar in this and previous studies (6, 18, 37, 46). In fact, the fast component of fH kinetics 1) was cancelled out under vagal blockade (18) and 2) was not found in heart transplant recipients, whose hearts are denervated (3, 20, 40). To extend this hypothesis to explain phase I kinetics of Q̇, Q̇aO2, and V̇o2, we should be able to demonstrate that when vagal tone is attenuated, as is the case in acute normobaric hypoxia (26), the phase I should either be reduced or disappear for all these parameters. Indeed, hypoxia reduced A1 significantly, for Q̇ and Q̇aO2 (Table 3) at both 50 and 100 W and for fH at 50 W, but did not extinguish it. On the other hand, hypoxia acted very little on τ1, whose very low values were essentially invariant; we noticed only a reduction at 50 W for Q̇ and an increase at 100 W for fH. This would mean that 1) the τ1 values in hypoxia also are compatible with a very rapid neural phenomenon, vagal withdrawal, as originally proposed by Fagraeus and Linnarsson (18) for normoxia; 2) vagal withdrawal has a smaller amplitude in hypoxia than in normoxia because of lesser vagal activation in the former; and 3) the patterns in the time domain of vagal withdrawal at exercise onset are fixed and invariant.
However, the occurrence of phase I in hypoxia with a smaller amplitude, instead of a full suppression of it, may imply that 1) hypoxia did not fully suppress vagal activity at rest so that some degree of vagal withdrawal still took place at exercise onset, or 2) other mechanisms than vagal withdrawal participate in phase I. The former may indeed be the case for fH. The latter is suggested by the apparent lack of changes in the Qst kinetics in hypoxia with respect to normoxia (Fig. 1). In the absence of a clear predetermined model for the Qst kinetics, whereby we refrained from fitting parameters through Qst data, we evaluated the contribution of Qst to Q̇ A1, as follows. Since Q̇ is the product of fh times Qst, the absolute Q̇ value at the peak of phase I is equal to (2) where Q̇, fH, and Qst are the resting values and ΔQ̇, ΔfH, and ΔQst are the corresponding increments during phase I, namely, the respective A1 values. Solution of this equation for ΔQst thus provides an estimate of the Qst amplitude during phase I that is necessary to sustain the observed increase in Q̇. At 50 W, ΔQst was 23.9 ± 10.5 and 14.7 ± 7.1 ml in normoxia and hypoxia, respectively. The corresponding ΔQst values at 100 W were 33.1 ± 9.3 and 24.2 ± 21.2 ml. At both powers, although affected by a large scatter, ΔQst did not differ in hypoxia from normoxia (P > 0.1 in both cases), suggesting that, contrary to fH, the alleged amplitude of Qst in phase I may not vary in hypoxia with respect to normoxia. This being the case, then 1) if indeed the reduction of the A1 of fH in hypoxia is due to lesser vagal withdrawal, then the same should be the case for the reduction of the A1 of Q̇; and 2) the Qst changes during an exercise transient are independent of mechanisms related to vagal withdrawal. Concerning the latter, in supine posture, a condition in which central blood volume is increased (24, 29, 47), phase I of Q̇ is not evident (25, 29), although resting vagal activation is greater supine than upright (25, 38). A higher central blood volume would reduce the amount of blood suddenly displaced from the periphery to the heart by muscle pump action, and thus the size of the immediate increase in venous return, thus preventing an efficient Frank-Starling mechanism. In the present study exercise was carried out in upright posture only, so it is likely that the increase in venous return due to muscle pump action would be the same in hypoxia as in normoxia, whence equivalent Qst kinetics.
In both normoxia and hypoxia, the τ1 of V̇o2 kinetics was extremely rapid and functionally instantaneous, indicating a practically immediate upward translation of V̇o2 that appears since the first breath. The τ1 of V̇o2 did not differ between powers and can be considered equal to those of Q̇ and Q̇aO2 (Table 3), given that the minimal functional time window in which V̇o2 can be determined is one breathing cycle. This suggests that the phase I changes in V̇o2 are imposed by the corresponding phase I changes in Q̇. Because of a delay between muscle O2 consumption and V̇o2, we can assume that during the first seconds of exercise, arterial-venous O2 difference (CaO2-CvO2) remains equal to that at rest (5). On this basis, the Fick principle allows a prediction of the expected V̇o2 increase in phase I as a consequence of the observed Q̇ increase. In normoxia, A1 of Q̇ was on average 4.57 l/min (see Table 3). For an average resting CaO2-CvO2 of 79 ml/l, we would expect an immediate V̇o2 increase upon exercise start of 0.36 l/min, compared with a measured A1 of V̇o2 of 0.39 l/min (Table 3). By analogy, in hypoxia, A1 of Q̇ was on average 2.52 l/min (Table 3), and the resting CaO2-CvO2 was 70 ml/l. Thus the expected V̇o2 increase would be 0.18 l/min, compared with a measured A1 of V̇o2 of 0.35 l/min (Table 3). Despite this sizeable discrepancy, the two values are not significantly different, probably because of the relatively large coefficient of variation of the data in hypoxia. Nevertheless, the results of this analysis suggest that A1 of V̇o2 may be a direct consequence of the rapid Q̇ increase during phase I, in agreement with the so-called cardiodynamic hypothesis of lung V̇o2 transients (55).
At rest in normoxia, P̄ was 90 mmHg and fH was 74.5 min-1. Let us assume that these values set the operating point of the average baroreflex curve of present subjects and that the operating point of resting subjects in normoxia corresponds to the centering point of the baroreflex curve (42). Assume also that the maximal gain and the operating range of the baroreflex curve are as previously reported (36). On this basis, we can construct a resting baroreflex response curve for the present subjects, which is reported in Fig. 4. If we then add the average resting value observed in hypoxia to that curve, we can see that the subjects operated in hypoxia on the same baroreflex curve as in normoxia, with a displacement of the operating point along the curve toward the threshold. This is a result of the decrease in P̄ induced by peripheral vasodilation, to which the subjects responded with an increase in fh, supporting the notion that peripheral vascular changes play a significant role in the baroreflex response of resting humans (17, 35). If indeed hypoxemia induces vasodilation in peripheral circulation via β2-sympathetic stimulation, then we can propose a role for peripheral chemoreflexes in the displacement of the baroreflex operating point in hypoxia.
Exercise displaces the baroreflex curve upward and rightward, without changes in gain [baroreflex resetting (42)]. This phenomenon is part of the rapid cardiovascular response upon exercise start. Baroreflex resetting includes a fast phase (Fig. 2), completed within a few heartbeats, within the duration of phase I, which goes on in parallel with the rapid changes in Rp that take place at exercise start. However, the dynamics of baroreflex resetting implied a larger number of heartbeats in hypoxia than in normoxia. If vagal withdrawal sets phase I of fH, then it will also contribute to the rapid upward shift of the baroreflex curve, but its role in baroreflex resetting will be less important in hypoxia than in normoxia.
Figure 4 also reports baroreflex curves for 50- and 100-W exercises, which are shifted upward as much as described in previous studies (34, 35, 39, 41). When the 50- and 100-W steady-state P̄ and fH values were added to Fig. 4, they were further displaced upward and rightward with respect to the predicted exercise baroreflex curve, suggesting that in the present study, baroreflex resetting was somewhat more intense than previously reported. Moreover, at both 50 and 100 W, the segment relating the experimental point in normoxia to the experimental point in hypoxia had a greater slope than the expected baroreflex gain, suggesting that the latter experimental point may lie on a different baroreflex curve than the former in both cases. A 50-W exercise provides a higher power relative to the maximum in hypoxia than in normoxia because of the decrease in maximal O2 uptake in hypoxia, whereby implying a greater role of the chemical component of the exercise pressor reflex (42). Yet we cannot distinguish the relative roles of central command or of the exercise pressor reflex in determining baroreflex resetting from the present data (42).
We conclude that in hypoxia, with respect to normoxia, 1) phase I is not completely suppressed for fH, although the A1 for fH is decreased, likely because the degree of vagal withdrawal is less; 2) since phase I is partly maintained, arterial baroreflex resetting continues to be very rapid, taking place essentially within phase I; 3) ΔQst is unchanged, because the increase in venous return due to muscle pump action is unchanged; 4) the coupling of a decreased A1 for fH with an unchanged ΔQst would produce the observed decrease in A1 of Q̇; and 5) the same factors that determine the phase I of Q̇ are also responsible for the phase I of V̇o2. Thus the present results provide only partial support to the hypothesis that vagal withdrawal determines the phase I kinetics of cardiovascular O2 flow and lung O2 uptake, because the suppression of phase I was incomplete. Although we can attribute the decrease in A1 of fH to a diminished degree of vagal withdrawal in hypoxia, this is not so for Qst. Under these circumstances, the dual origin of the phase I of Q̇ and Q̇aO2, neural (vagal) and mechanical (increase in venous return by muscle pump action), would rather be confirmed.
This study was supported by Swiss National Science Foundation Grants 3200-061780 and 3200B0-114033 (to G. Ferretti) and Italian Space Agency Contract DCMC-1B133 (to C. Capelli).
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.
- Copyright © 2008 the American Physiological Society