Metabolic demand and muscle mechanical tension are closely coupled during exercise, making their respective drives to the circulatory response difficult to establish. This coupling being altered in eccentric cycling, we implemented an experimental design featuring eccentric vs. concentric constant-load cycling bouts to gain insights into the control of the exercise-induced circulatory response in humans. Heart rate (HR), stroke volume (SV), cardiac output (Q̇), oxygen uptake (V̇o2), and electromyographic (EMG) activity of quadriceps muscles were measured in 11 subjects during heavy concentric (heavy CON: 270 ± 13 W; V̇o2 = 3.59 ± 0.20 l/min), heavy eccentric (heavy ECC: 270 ± 13 W, V̇o2 = 1.17 ± 0.15 l/min), and light concentric (light CON: 70 ± 9 W, V̇o2 = 1.14 ± 0.12 l/min) cycle bouts. Using a reductionist approach, the circulatory responses observed between heavy CON vs. light CON (difference in V̇o2 and power output) was ascribed either to metabolic demand, as estimated from heavy CON vs. heavy ECC (similar power output, different V̇o2), or to muscle mechanical tension, as estimated from heavy ECC vs. light CON (similar V̇o2, different power output). 74% of the Q̇ response was determined by the metabolic demand, also accounting for 65% and 84% of HR and SV responses, respectively. Consequently, muscle mechanical tension determined 26%, 35%, and 16% of the Q̇, HR, and SV responses, respectively. Q̇ was significantly related to V̇o2 (r2 = 0.83) and EMG activity (r2 = 0.82; both P < 0.001). These results suggest that the exercise-induced circulatory response is mainly under metabolic control and support the idea that the level of muscle activation plays a role in the cardiovascular regulation during cycle exercise in humans.
- metabolic demand
- muscle mechanical tension
- heart rate
- stroke volume
- circulatory control
because of its pivotal role in setting systemic oxygen (O2) delivery and pulmonary oxygen uptake (V̇o2), cardiac output (Q̇) is considered a determining factor of aerobic exercise capacity in health (18) and many disease states (46). It has been postulated that metabolic demand and muscle mechanical tension are neurally integrated in the exercising organism to determine the appropriate Q̇ response, either via the “central command” process and/or via the so-called “exercise pressor reflex” (40). As both mechanisms are probably not mutually exclusive (42), the Q̇ response to exercise may be controlled by a combination of direct “cardiovascular outflow” from the higher brain centers as well as by activation of metabolic and mechanical afferent neurons originating in the exercising limbs (9, 12, 33, 45). Therefore, metabolic perturbations and/or muscle tension-related mechanisms are considered important inputs underlying Q̇ adjustments during exercise (25). However, the respective contributions of the metabolic vs. mechanical control in setting the exercise-induced circulatory response during whole-body exercise remains unclear.
During exercise, metabolic demand and muscle mechanical tension vary simultaneously in a workload-dependent manner. In traditional (i.e., concentric) cycle exercise, Q̇ adjustments are expected to couple the systemic oxygen delivery to the oxygen demand of muscle activation (39). In eccentric (i.e., forced lengthening of activated muscle) cycling, Q̇ and electromyographic (EMG) activity of the vastus lateralis muscle are lower than during concentric cycling despite similar muscle mechanical tension (i.e., same power output and pedal frequency) (4, 11, 43). These findings are in agreement with the lower metabolic demand of eccentric cycling, as established from V̇o2 and blood lactate measurements (1, 2, 11, 37). However, when compared at identical levels of metabolic demand (i.e., similar V̇o2), higher |Ap Q̇ and greater EMG activity are observed in eccentric vs. concentric cycling (4, 11, 43). Therefore, Q̇ responses seem closely related to the changes in EMG activity, suggesting that metabolic demand and muscle mechanical tension may act in synergy with motor command (i.e., muscle activity) to set the circulatory response to exercise. However, the extent to which EMG activity is associated with the Q̇ response remains to be investigated in exercising humans.
Therefore, the first aim of this study was to establish the respective contribution of metabolic demand and muscle mechanical tension as the main determinants of the Q̇ response to exercise. A second aim was to determine the relationship between EMG activity and the Q̇ response to exercise, whatever the magnitude of metabolic demand and/or muscle mechanical tension. To achieve these goals, a comparison of concentric vs. eccentric cycle exercises, inducing target levels of metabolic demand and muscle mechanical tension, as well as recruiting large muscle masses, was thought to shed light on the metabolic and mechanical control of Q̇.
Eleven healthy men [aged 28 ± 6 (sd) years; height, 180 ± 5 cm; body mass, 71 ± 8 kg; maximal oxygen uptake, 3.65 ± 0.59 l/min; peak power, 322 ± 17 W] participated in this study, which had local institutional and ethics approval. All subjects displayed no history of muscle, tendinous, or articular problems and gave voluntary written consent to participate.
During the 2 wk preceding the study, all subjects underwent 3 or 4 familiarization sessions, separated by 2 or 3 days recovery, to acquire the specific coordination of eccentric cycling and minimize muscle, tendinous, or articular problems (11, 37). Thereafter, they performed an incremental concentric cycle ergometer test to exhaustion (30 W/min) at 60 rpm to determine maximal oxygen uptake (V̇o2 max) and peak power.
To address the aim of the study, the subjects completed three constant-load cycling bouts, preceded by 4 min of very light cycling (15 W) and separated by 24 h of recovery: 1) heavy concentric (heavy CON; 270 ± 13 W); 2) heavy eccentric (heavy ECC; 270 ± 13 W); and 3) light concentric (light CON; 70 ± 9 W) designed to elicit the same V̇o2 as in the heavy ECC (∼1.15 ± 0.15 l/min). Each test was performed in a random order with a pedal frequency of 60 rpm. The duration of each trial was fixed at 6 min to ensure a steady-state V̇o2 in heavy ECC and light CON (37). These experimental conditions allowed for the identification of the individual and combined effects of the metabolic and mechanical loading in the exercise-induced circulatory responses (see Fig. 1). Similar power output and pedal frequency, as occurred in the heavy CON and heavy ECC trials, required the subject to exert identical mechanical torque on the pedals, thereby providing experimental condition of similar muscle mechanical tension but different metabolic demand.
All exercises were performed on an adapted ergometer in the semirecumbent position. A commercially available ergometer (Recline XT, TechnoGym, Gambettola, Italy) was modified with a 2.2-kW asynchronous electric motor (MasterDrive Simovert Vector Control, Siemens, Erlangen, Germany) (11) and controlled by a computer using a specific custom software (Labview FDS 5.1, National Instruments, Austin, TX). Power output was verified independently using a calibration device comprising a torque meter and a tachometer. The accuracy of the power output measurements was ±5 W (between 0 and 1,000 W), in both concentric and eccentric cycling. During eccentric cycling, the power output has to be self-adjusted through visual feedback on a computer screen. The seat and handlebar positions were adjusted to each subject's morphology and kept constant for all tests.
V̇o2 and carbon dioxide output were measured with a breath-by-breath open-circuit metabolic cart (Ergocard with Exp'Air software version 1.26.35, Medi-Soft, Dinant, Belgium). Stroke volume (SV) was measured by bioimpedance, whereas heart rate (HR) was simultaneously estimated from the electrocardiogram first derivative (Physio Flow, Manatec type PF05L1, Paris, France). The accuracy of this bioimpedance device has previously been established against the direct Fick method over a wide range of Q̇ values during rest and exercise (5 to 25 l/min) in patients (8) and normal healthy subjects (38). Calibration of the impedance device was done using a procedure based on 24 consecutive heartbeats recorded with the subject resting on the ergometer (8). Systolic and diastolic arterial blood pressures were measured manually from the right arm by the auscultatory method using an inflatable cuff. Mean arterial pressure was calculated as [(2 × diastolic blood pressure) + systolic blood pressure]/3. Systemic vascular conductance was established from the ratio between Q̇ and mean arterial pressure.
37). The exercise value of muscle activity was subsequently normalized to the value recorded at the end of baseline (norRMS). Lastly, normalized values of EMG activity obtained from the three muscles were averaged to generate an index of the whole quadriceps muscle activity.
During each trial, blood samples were drawn from a forearm vein at rest, at baseline and at the end of exercise. Plasma lactate concentration was determined by an enzymatic method (ABL 700 series, Radiometer, Bronshoj, Denmark), whereas plasma catecholamine concentrations (epinephrine and norepinephrine) were determined by high-performance liquid chromatography (Gilson, Middleton, WI) coupled with an electrochemical detection (Coulochem, ESA Biosciences, Chelmsford, MA).
A two-way (trial-by-time) repeated-measures ANOVA was employed to test significant differences between (heavy CON, heavy ECC, and light CON) and within (rest, baseline, exercise) trials. Significant pairwise differences were localized with Newman-Keuls post hoc tests. Relationships between variables were established using a multiple regression procedure in which the subjects were treated as a categorical variable and entered as a predictor parameter, providing a correlation coefficient for within-subject changes (5). P < 0.05 was considered as statistically significant. All results are expressed as means ± SE.
Establishment of Experimental Conditions
Resting, baseline, and exercise circulatory, metabolic and EMG activity values in each experimental condition are indicated in Table 1. No differences were observed at rest or at baseline. As designed, power output was the same in heavy CON and heavy ECC (270 ± 13 W) but was ∼fourfold lower in light CON (70 ± 9 W). Simultaneously, V̇o2 was similar in heavy ECC and light CON (1.17 ± 0.15 vs. 1.14 ± 0.12 l/min, respectively) but ∼threefold greater in heavy CON (3.59 ± 0.2 l/min). Correspondingly, plasma lactate concentration was the same in heavy ECC and light CON (1.6 ± 0.2 vs. 1.3 ± 0.1 mmol/l respectively) but was ∼5.5-fold higher in heavy CON (8.0 ± 0.8 mmol/l). Therefore, V̇o2, plasma lactate concentration, and power output were successfully manipulated across trials to investigate the individual and combined contributions of metabolic demand and muscle mechanical tension in the circulatory responses to exercise (Fig. 1).
Contribution of Metabolic and Mechanical Loading to the Circulatory Response to Exercise
Combined effect of metabolic and mechanical loading (heavy CON vs. light CON).
As a result of the combined three- to fourfold increase in V̇o2 and power output, Q̇ was elevated by 12.4 ± 0.9 l/min in heavy CON due to a ∼twofold higher HR and a 23% greater SV. Mean arterial pressure, systemic vascular conductance, and EMG activity were higher in heavy CON, as were plasma lactate, epinephrine, and norepinephrine concentrations (Table 1).
Individual effect of metabolic loading (heavy CON vs. heavy ECC).
In line with the approximately threefold higher metabolic demand, Q̇ was 8.9 ± 1.1 l/min higher in heavy CON, associated with a ∼1.5-fold higher HR and a 13% greater SV. Therefore, Q̇, HR and SV were significantly related to the changes in V̇o2, with r2 values being 0.83, 0.84, and 0.50, respectively (all P < 0.001; Fig. 2). Correspondingly, mean arterial pressure and systemic vascular conductance were also higher in heavy CON, as were plasma lactate, epinephrine, and norepinephrine concentrations and EMG activity (Table 1 and Fig. 3).
Individual effect of mechanical loading (heavy ECC vs. light CON).
At the same V̇o2, Q̇ was 3.4 ± 0.7 l/min higher in heavy ECC, through a 23% higher HR and a 11% greater SV. A similar pattern of response appeared for mean arterial pressure, systemic vascular conductance and EMG activity (Table 1). Moreover, EMG activity was closely related to Q̇ (r2 = 0.82, P < 0.001; Fig. 4). Plasma epinephrine concentration was ∼60% greater in the heavy ECC trial, whereas plasma lactate and norepinephrine concentrations were not different.
Relative metabolic and mechanical contributions to the exercise-induced circulatory response.
The individual effect of metabolic demand was estimated to account for 74 ± 7%, 66 ± 7%, and 84 ± 16% of the overall Q̇, HR, and SV responses, respectively, observed in heavy CON vs. light CON, with the remaining 16 to 34% being attributed to the individual effect of muscle mechanical tension (Fig. 5).
This study used eccentric and concentric cycling models to estimate the respective contribution of the metabolic demand and muscle mechanical tension to the Q̇ response to cycle exercise. Additionally, this work also aimed to investigate the relationship between the Q̇ response and EMG activity of the quadriceps muscles. A main finding is that Q̇ is mostly under metabolic control with a smaller but still significant regulatory influence attributed to a mechanical-related drive. An additional observation is that the Q̇ response appeared closely related to the quadriceps EMG activity, over a wide range of metabolic and/or mechanical stimulations. Collectively, these findings suggest that the exercise-induced Q̇ response is mainly controlled through changes in metabolic demand and support the idea that the level of EMG activity plays a role in this process.
Metabolic Contribution to the Circulatory Control
Heavy CON and heavy ECC cycle bouts were performed at similar pedal frequency and power output (270 ± 13 W), leading to equivalent mechanical torque and muscle tension, regardless of the type of muscle contraction. Therefore, the differences in circulatory responses between heavy CON and heavy ECC can be ascribed to the large difference in metabolic demand, as indicated by the differences in V̇o2 (∼2.4 l/min) and plasma lactate concentration (∼5.6 mmol/l). This large metabolic demand per se was estimated to account for 74% of the Q̇ response (Fig. 5), and this percentage was found consistent between both the reductionist computation and the correlation analysis, where Q̇ and V̇o2 appeared highly related (r2 = 0.83). This major metabolic control of the Q̇ response agrees with previous findings obtained at the muscle (22, 23) and/or systemic level (6, 19), showing that Q̇ and V̇o2 are strongly related at least up to 80% of peak power (34). Therefore, our results provide for the first time a quantitative estimation of the influence of metabolic demand per se on systemic circulatory responses to high intensity but submaximal exercise (∼85% peak power).
Such a tight coupling between Q̇ and metabolic demand was achieved through concomitant changes in HR and SV. Indeed, 66% of HR adjustments, as well as 84% of the SV response were related to the metabolic demand. Simultaneously, systemic vascular conductance was also found elevated in heavy CON vs. heavy ECC, allowing for the large Q̇ difference (8.9 ± 1.1 l/min), while attenuating the perturbations in mean arterial pressure (12 ± 5 mmHg). The exact pathways underlying the metabolic control of Q̇ during exercise are not fully understood. With increasing metabolic demand, accumulation of by-product compounds occur in exercising muscles, as suggested by the marked elevation of plasma lactate concentration in heavy CON vs. heavy ECC. Subsequent diffusion within the interstitium milieu might be sufficient to activate the metabosensitive nerve endings of type III and IV afferent fibers, thereby contributing to increase muscle sympathetic nerve activity (41). In good agreement with this possibility, the plasma norepinephrine response was found similarly elevated after individual metabolic loading (heavy CON vs. heavy ECC) than after combined metabolic and mechanical loading (heavy CON vs. light CON). An enhanced catecholamine outflow exerts chronotropic and inotropic effects on the heart, possibly contributing to the metabolically induced elevation of HR and Q̇. Conversely, this enhanced sympathetic drive is mainly vasoconstrictive within the working muscles, theoretically impairing muscle blood flow. However, at submaximal exercise, this vasoconstrictive influence is counteracted by the release of local factors (i.e., H+, Pi, K+, prostaglandins, adenosine, nitric oxide, ATP), thereby still allowing vasodilation, as suggested by the elevated systemic vascular conductance observed in heavy CON vs. heavy ECC. Nevertheless, such a regulatory influence of metabolite accumulation may be attenuated near maximal exercise, where Q̇ can plateau or even decline (34). Additional metabolite accumulation, such as ATP infusion in the femoral artery, does not seem to prevent this Q̇ limitation during maximal cycling exercise (7), pointing toward a possible impairment of the metabolic circulatory control at very high intensity of exercise.
Mechanical Contribution to the Circulatory Control
Heavy ECC and light CON cycle bouts were designed to elicit the same absolute V̇o2 (∼1.1 l/min). On the basis of the similar V̇o2 and concomitant equivalent plasma lactate values, both trials were performed at near-identical metabolic demand. In this condition, differences in Q̇ responses arise from the large difference in power output (i.e., 200 W) and therefore in mechanical torque and muscle tension. Previous studies investigating the mechanical control of the circulatory response in exercising humans have mainly used passive cycling models, which restrict the mechanical stimulation to low levels of muscle tension (35, 47). Conversely, the original approach employed in the present study generates large differences in muscle tension, while keeping similar metabolic demand.
Muscle mechanical tension is not usually considered to be a primary determinant of local and systemic blood flow response, but much of this knowledge originates from studies where metabolic demand was the dependent variable, whereas mechanical load was kept constant (22, 23). Our present results suggest that muscle mechanical tension, although not the main governor, still exerts a potent driving effect, accounting for 26% of the Q̇ adjustments during exercise. Whether the excess Q̇ response observed in heavy ECC vs. light CON is directed toward active muscles and/or inactive tissues (i.e., cutaneous territories) could not be determined in this experiment. However, as heavy ECC induced a lower arteriovenous O2 difference than light CON (8.7 ± 0.7 vs. 11 ± 0.7 ml/100 ml respectively, P < 0.05), one possible mechanism is that the excess Q̇ was used for nonenergetic purposes (i.e., thermoregulation). Alternatively, the reduced arteriovenous O2 difference may also result from a greater blood flow/O2 uptake ratio within the active muscles, thereby decreasing O2 extraction.
This “delinking” of Q̇ from metabolic demand stems from the higher HR and greater SV observed in heavy ECC vs. light CON. The reductionist computation indicates that 34% of the HR response to exercise is related to the driving effect of the mechanical tension generated within the exercising limbs. Large muscle tensions are likely to activate peripheral mechanoreceptors, which constitute about 10% of the myelinated nerve fibers in leg muscles (20), leading to sympathetic activation (17). The specific increase in plasma epinephrine in heavy ECC vs. light CON, with no change in plasma norepinephrine supports a mechanically induced sympathetic activation, as previously reported in animals and humans (32, 36, 44). The small, but remaining mechanical effect, on SV (16%) may be related to exercise-induced intramuscular pressure oscillations (3), possibly facilitating venous return to the heart (i.e., muscle pump) (30). Interestingly, the systemic vascular conductance was also elevated in heavy ECC vs. light CON, in line with a role of the mechanical deformation of resistance vessels and vascular beds in the circulatory response to exercise (21), possibly through myogenic vasodilation (39).
Neural Integration of Metabolic and Mechanical Stimulations: A Role for EMG Activity?
A new finding of the present study is that the Q̇ response to exercise is related to EMG activity during whole-body exercise in humans (r2 = 0.82), over a wide range of metabolic demand and/or muscle mechanical tension. Although eccentric and concentric muscle contractions differ in their respective pattern of fiber activation (13), the maximal level of EMG activity, and therefore the “reserve” of EMG activity, has been shown to be similar in both types of work (28). Consequently, despite the limitations associated with EMG recordings (10), EMG measurements have been analyzed previously in the time domain (as in the present work) to provide information on the global level of muscle activation during concentric and eccentric cycling (4, 37).
The present relationship between Q̇ and EMG activity fits well with the metabolic and mechanical drives to Q̇ adjustments. First, this observation suggests that Q̇ is regulated in proportion to the active muscle mass and, therefore, in relation to the energetic requirements of the recruited muscle fibers. A relationship between EMG activity and circulatory responses has previously been reported at the muscle level, where the changes in the perfusion of vastus lateralis and vastus medialis muscles observed during knee-extensor exercises significantly correlated with alterations in their respective level of EMG activity (29). In this way, the present results fully support the notion that Q̇ is mostly under metabolic control. Second, when Q̇ was higher in heavy ECC vs. light CON despite similar V̇o2, EMG activity was also significantly higher. Why EMG activity is higher in eccentric vs. concentric cycling at similar V̇o2 remains unclear. Historically, it has been suggested that fewer muscle fibers are recruited and that the recruited muscle fibers also consume less oxygen in eccentric vs. concentric cycling at similar power output (4). This implies that a greater EMG activity is required to reach a similar V̇o2 in eccentric vs. traditional concentric cycling. Whatever the mechanism, the higher EMG activity reported for heavy ECC vs. light CON may contribute to part of the Q̇ response to exercise (26%) that is not accounted for by the metabolic demand. Therefore, factors related to the level of EMG activity (i.e., muscle tension) may assist the metabolic regulation to provide a fine tuning to the Q̇ response to exercise. Then, metabolic and/or mechanical signals could both be part of an integrated sensory input, eliciting circulatory adjustments during cycle exercise in humans.
Along with metabolically or mechanically related afferent signals from the exercising muscles, the tight relationship between Q̇ response and EMG activity may also point to a feedforward cardiovascular regulation, as previously suggested by studies investigating the role of “central command” (16, 31). It should be emphasized that the “central command” was never quantified in the experimental conditions tested in the present study. Consequently, a central process may also have played an independent role in the regulation of the Q̇ response. Although a functional role of the “central command” in the cardiovascular regulation is well established for static exercise (24), experiments using dynamic exercises in humans have reported more conflicting results. For instance, electrically induced cycling has been used to bypass the “central command” during cycling exercise (27), leaving alone the muscle feedback regulation. In this case, HR and Q̇ responses are not affected, suggesting little functional role for a “central command” process in the cardiovascular regulation. Conversely, the normal HR response to cycling exercise is maintained when preventing afferent nerve traffic using partial (15) or complete epidural anesthesia (14), suggesting that the “central command” can contribute to the cardiovascular adjustments. Similarly, the similar HR and Q̇ responses to exercise in hypoxia with or without epidural anesthesia (26) demonstrate that the excess cardiovascular response in hypoxia is not mediated by muscle afferent signals, thereby providing indirect support for a functional role of the “central command” in cardiovascular regulation. The above and our present findings suggest that EMG activity may be involved in the control of the cardiovascular response to exercise. However, the intimate mechanisms underlying the “central command” as well as its relative contribution vs. the metabolic/mechanical afferent cardiovascular regulation deserve further investigations.
In conclusion, although metabolic demand is the main driving factor for Q̇ adjustment during cycle exercise, this study shows the importance of nonmetabolic stressors, related to muscle mechanical tension in the exercising limbs. The mechanisms underlying the control of the circulatory responses to exercise remain to be established but could involve a key contribution of EMG activity, supporting the idea that the circulatory responses are adjusted to the level of muscle activation during cycle exercise in humans.
This research was funded by the Clinical Research Department of the Hôpitaux Civil de Strasbourg.
Special thanks are given to the subjects in this study for their enthusiastic participation, to the laboratory staff for daily support, as well as to Sophie Bayer for her help in the coordination of the biological measurements. The excellent engineering assistance of J. P. Speich has been greatly appreciated for the conception and development of the eccentric cycle ergometer. The authors are also grateful to Prof. Janet Peacock for her strong statistical support and to Sara Horne for English revisions.
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 © 2007 the American Physiological Society