Introduction

Manual wheelchair propulsion is an inefficient mode of ambulation, with a gross mechanical efficiency in the range of 2–10%.1, 2 The main mechanical factors associated with the poor efficiency of wheelchair propulsion are: (1) the explicit necessity to stabilize the highly flexible shoulder complex and the hand-wrist system during propulsion;3, 4 (2) the inefficient resultant vector produced by the hand on the wheel during the propulsion phase;2, 5, 6 (3) the discontinuity of hand rim propulsion which requires an amount of energy;7, 8 and (4) the small muscle mass engaged in the propulsion task.9, 10

Hand rim wheelchair propulsion has been extensively studied with respect to external force parameters;2, 5, 11, 12 kinematics13, 14, 15, 16 and myoelectric parameters of arm muscles.4, 17, 18 These studies demonstrated that both the environmental constraint and the level of expertise modify biomechanical and electromyographical responses of a wheelchair user.

However, only a few authors have investigated the propulsion movement and its consequences on oxygen uptake ( ) and muscles activity during fatigue.18, 19, 20, 21 These studies examined movement and muscle activation to develop more specific strength-training programs to avoid upper limb injuries. Within this framework, little is known regarding adjustments of propulsion cycle timing and muscle coordination during fatiguing wheelchair-exercise.

The purpose of this study was to determine the effect of constant-load exercise on , muscle fatigue, and movement cycle timing at two different intensities. We hypothesized that moderate exercise below ventilatory threshold (VT) would not induce fatigue. However, heavy exercise above VT would induce fatigue and thus alter oxygen uptake, muscle activity, and movement cycle timing.

Methods

Subjects

After providing written informed consent, eight able-bodied men (age: 26±4 years; height: 176±4 cm; and weight: 73±7 kg) participated in the study. All participants were right handed and had no prior experience in wheelchair propulsion. All applicable institutional and governmental regulations concerning the ethical use of human volunteers were followed during the course of this research.

Procedure

Each exercise bout was performed on a standard wheelchair (Quickie EX, Sunrise Medical, England), placed on an ergometer (VP100H, H.E.F. Tecmachine, Andrezieux Boutheon, France), at a constant velocity of 5 km/h, with subjects choosing a propulsion frequency.

Subjects first performed incremental exercise to determine and VT. was defined as maximal reached during the incremental exercise.22 VT was determined using the ventilatory equivalents method according to Wasserman et al23 and Vallier et al.24 The initial power output (PO) was set at 28±3 Watts (W) for 2 min and further increased by 10 W for each subsequent 2 min until the subject could no longer maintain the compulsory velocity.

Subjects came back 2 days after the incremental exercise test to perform two constant-load exercise bouts of 6 min duration. The first exercise bout was performed at a PO corresponding to ∼80% of the PO at VT and was referred to as the moderate intensity exercise bout. Following a minimum of 15 min of passive recovery, the subject performed a second exercise bout at a PO corresponding to ∼50% between the PO at VT and the PO at . This exercise bout was referred to as the heavy intensity exercise bout.

Data acquisition

During each exercise bout, oxygen uptake ( ), carbon dioxide production ( ), and ventilation were determined breath-by-breath (K4b2, Cosmed, Roma, Italy) and averaged every 30 s. O2 and CO2 analyzers were calibrated together with signal volume according to manufacturer instructions before each test session, using reference gases of known concentrations and a 3-l calibration syringe.

The wheelchair ergometer system, extensively detailed in Devillard et al,25 simulated inertia of the user-chair system and allowed the measurement of instantaneous force and velocity during exercise at a sampling rate of 200 Hz. Temporal characteristics of the propulsion cycle pattern were calculated using speed and acceleration/deceleration recordings during all tests: cycle time (time to complete one full arm cycle in s), push time (period corresponding to the acceleration of the wheels in s), and recovery time (difference between the cycle time and the push time in s). For both exercise conditions, the mean values of each cycle parameter (ie cycle, push and recovery time) were calculated over the last 15 s before minutes 3 and 6 (moderate exercise) or every minute (heavy exercise).

Electromyographic activity (EMG) was recorded from the biceps brachii, the long head of the triceps brachii, the anterior deltoid, the middle part of the trapezius, the infraspinatus, and the latissimus dorsi muscles of the right shoulder. The rectus abdominis was also studied to evaluate the stabilization of the trunk. Myoelectric signals were recorded using an eight-channel EMG recorder (ME3000P8, Mega Electronics, Kuopio, Finland) with a bandwidth of 8–500 Hz (Common Mode Rejection Ratio=110 dB) at a sampling rate of 1000 Hz. Bipolar surface electrodes were positioned on the contracted muscle belly after appropriate skin preparation with a center-to-center distance of 30 mm, along a line approximately parallel to the direction of the underlying muscle fibers. All EMG cables were taped down to prevent movement artifact. Myoelectric signals were collected during the last 15 s of the third and the sixth minute during the moderate exercise, and at the end of each minute during the heavy exercise. The raw EMG signal was converted into root-mean-square (RMS) with a time constant of 25 ms and into mean power frequency (MPF) using Hamming windowing and 512 points Fast Fourier Transform.

RMS of all muscles studied were summed and averaged to provide an index of the total muscle activity,26 further referred to as the average RMS response. The same procedure was applied to MPF to provide the average MPF response. Each parameter was then normalized to the third minute of exercise.

Activation time was determined for each muscle during the last 15 s (at least eight cycles) of minute 3 and of minute 6 with careful visual inspection during both exercise bouts (method adapted from Wilen et al27). Muscles were considered active when the RMS was above 10% of maximal amplitude for at least 50 ms. Conversely, muscles were considered inactive when the RMS was below 10% of maximal amplitude during 50 ms. Maximal amplitude was defined as the highest RMS value obtained during the exercise. Activation time was then expressed as a percentage of the cycle time and averaged across 15 s at minutes 3 and 6.

Statistical analyses

All data are presented as means±SD unless otherwise specified.

Nonparametric tests were performed when data displayed unequal variances (Bartlett test) or were not normally distributed (Kolmogorov-Smirnov test). The Friedman repeated measures test was used to compare across time and Student–Newman–Keuls post hoc test was performed to determine when became stable across time for each exercise bout. Paired t-tests were used to compare cycle time, push time, recovery time, RMS, MPF, and activation time between the third and the sixth minute during the moderate exercise bout. A repeated measures ANOVA was performed to compare the same parameters across time during the heavy exercise bout. Multiple comparisons were performed using Student–Newman–Keuls post hoc test. The level of significance was set at 0.05 for all tests.

Results

Oxygen uptake

during the incremental exercise test was 2478±486 ml/min (34.49±8.39 ml/min/kg). During the moderate exercise bout (PO=41±6 W), stabilized during the third minute (1042±200 ml/min) (Figure 1). However, no stabilization was observed during the heavy exercise bout (PO=66±6 W) between minutes 3 and 6, with a progressive increase of 437±224 ml/min (+26.62±11.38%). was 2042±453 ml/min at the end of the heavy exercise bout.

Figure 1
figure 1

Oxygen uptake kinetic during moderate (○) and heavy (•) exercise bouts

Full size image

EMG parameters

RMS, MPF, mean RMS, and MPF responses, and activation time of the seven muscles were similar between the third and the sixth minute during the moderate exercise bout. Only the MPF of the anterior deltoid muscle significantly (P<0.05) decreased (−3.12±3.37%) between the third and sixth minute.

During the heavy exercise bout, RMS from all muscles significantly (P<0.05) increased (Figure 2), except for the middle part of the trapezius muscle. The mean RMS response significantly (P<0.05) increased (+39.47±11.68%) between the third and sixth minute, and the mean MPF response decreased (−5.80±4.56%) between the first and third minute only (Figure 3). MPF parameters also significantly (P<0.05) decreased for the triceps brachii and anterior deltoid between the first and third minute and for pectoralis major muscles between the second and the third minute (Figure 4).

Figure 2
figure 2

RMS during the heavy exercise bout

Full size image
Figure 3
figure 3

Mean RMS (•), and MPF (○) responses during the heavy exercise bout

Full size image
Figure 4
figure 4

MPF during the heavy exercise bout

Full size image

Muscle activation time

Muscle activation time did not significantly change during the moderate exercise between the third and sixth minute. During the heavy exercise bout, muscle activation time of the triceps brachii, anterior deltoid, pectoralis major, and rectus abdominis muscles were significantly (P<0.05) longer during the sixth minute than during the third minute (Figure 5).

Figure 5
figure 5

Activation time during the heavy exercise bout at minutes 3 and 6. *P<0.05. (BB: biceps brachii; TB: triceps brachii; TRAP: trapezius, middle part; ISP: infraspinatus; DELT-A: anterior deltoid; PMA: pectoralis major; LD: latissimus dorsi; RA: rectus abdominis)

Full size image

Temporal parameters

No significant difference between the third and sixth minute was observed for cycle (1.62±0.33 s), push (0.36±0.04 s), and recovery (1.26±0.31 s) times during the moderate exercise bout. However, the cycle and recovery times (Figure 6) significantly (P<0.05) decreased between the third and sixth minute during the heavy exercise bout (−10.99±9.15% and −11.68±9.56%, respectively).

Figure 6
figure 6

Cycle (•), push (□), and recovery (○) times during the heavy exercise bout. *Significantly different from minute 3 (P<0.05)

Full size image

Discussion

This study focused on adaptations of oxygen uptake, propulsion cycle timing, and muscle coordination during nonfatiguing and fatiguing wheelchair ergometry. Two exercise conditions were applied: one moderate exercise bout below VT (nonfatiguing), and one heavy exercise bout above VT (fatiguing).

Moderate exercise

Our results show constant after 3 min of moderate exercise, as well as no modification of propulsion cycle timing (characterized by the cycle, push, and recovery times), or of muscle activity and activation time for seven muscles surrounding the shoulder complex and the trunk over the exercise.

However, whereas the RMS of the anterior deltoid remained stable, MPF of the anterior deltoid muscle slowly decreased from minutes 3 to 6 despite low PO values (approximately 40 W), De Groot et al8 have shown that the peak activity of this muscle at a PO of 20 W corresponds to 40% of its maximal voluntary contraction during wheelchair propulsion for able-bodied men. As our study involves a PO approximately twice the PO of the previous study, we suggest that the decrease in MPF observed in our study was a consequence of fatigue. Several phenomena could provoke a reduction in MPF during fatigue, such as reduction of muscle membrane excitability28, 29 and synchronization of active motor units.30, 31 Considering the intensity of the exercise, synchronization of motor units is not a plausible hypothesis because it occurs when motor units are close to exhaustion.30, 32 Therefore, we suggest that the MPF decrease could be due to a decline of the firing frequency of active motor units to preserve muscle membrane excitability. This adaptation has been proposed by previous authors during fatiguing exercises and is referred to as muscle wisdom.33, 34

Heavy exercise

Several changes were observed during the heavy exercise bout, which may be attributed to fatigue. increased by 437±224 ml/min between minutes 3 and 6 (Figure 1). This increase is often observed during heavy square-wave exercise bouts and corresponds to a slow component of , as described by several authors.35, 36, 37, 38 The amplitude of the slow component in our study is consistent with the amplitude observed during arm cranking exercise,22, 39, 40, 41 despite the lower mechanical efficiency9, 42, 43 and the higher cardiorespiratory stress ( , cardiac output…)44 during wheelchair propulsion.

One factor that has been proposed to explain the slow component is recruitment of fast-twitch muscle fibers as a consequence of fatigue.26, 45, 46 One method to identify muscle fatigue is to study the neuromuscular activity over time during exercise.47, 48, 49, 50 Our results show a RMS increase during the exercise bout for seven muscles (Figure 2), which suggests the development of muscle fatigue and the recruitment of additional muscle fibers to maintain the required PO.29, 51 The mean RMS response (Figure 3), as an index of whole neuromuscular activity,26 also suggests a recruitment of additional muscle fibers from the onset of the exercise bout.

Furthermore, mean MPF response (Figure 3) and MPF of the pectoralis major, the anterior deltoid, and the triceps brachii muscles (the prime movers for the push phase) (Figure 4) also decreased during the first minutes. As the MPF reduction occurs only during the first minutes of exercise, we suggest that muscle wisdom (a decline in the firing frequency of active motor units to preserve the muscle membrane excitability) is the most likely explanation for our findings.

Our finding of alterations in the EMG activity of seven muscles supports our second hypothesis that muscle fatigue occurs during wheelchair ergometry. During the push phase, the prime movers of shoulder flexion (pectoralis major, anterior deltoid, and biceps brachii muscles),52, 53 and the triceps brachii muscle are mostly active.18, 54, 55 This involves flexor and extensor muscle coordination around the elbow, leading to an increased tendency to fatigue in individual muscles of the upper extremity.

The infraspinatus muscle, representing the rotator cuff muscles,56 stabilizes the humeral head laterally53 and counteracts the action of the medial rotator muscles on the shoulder.53, 57 The infraspinatus muscle also prevents the superior translation of the humeral head during shoulder abduction, a movement occurring during the recovery phase of the manual wheelchair propulsion cycle.

Previous studies54 have shown that the latissimus dorsi muscle is principally active during the recovery phase. However, it also stabilizes the humeral head during shoulder flexion,53 an action necessary during the push phase. We hypothesize that the activity of the latissimus dorsi during both muscle coordination and the recovery phase can explain the observed fatigue of this muscle.

Rodgers et al21 observed a power shift from the shoulder to the elbow and wrist joints during fatiguing wheelchair exercise. This finding supports the need to study muscle patterns during exercise and, therefore, the effects of fatigue on muscle activation time. An unexpected result of the present study was the observed increase of muscle activation time for the prime movers of the push phase (pectoralis major, anterior deltoid, triceps brachii) (Figure 5), despite a decrease in the duration of the cycle propulsion (Figure 6). This effect was likely due to the impact of fatigue on muscle function. Rodgers et al20 previously observed the same effect. The increase of the relaxation half-time previously observed during fatiguing exercise may have caused the increased activation time during the wheelchair exercise.33, 58, 59 A reduction of calcium pump activity of the sarcoplasmic reticulum59, 60 resulting from an accumulation of inorganic phosphate61 or changes in the action potential shape62 are the most plausible explanations for the decline of the relaxation half-time. These two alterations also contribute to an increase in RMS and a decrease in MPF, electromyographical changes observed in our study.

Furthermore, the increase in activation time during exercise, as fatigue occurs, could induce an increase of the coactivation between agonist and antagonist muscles and, therefore, reduce motor efficiency. This reduction could partly explain the slow component, and warrants further investigation.

The second major finding of the present study relates to the observed changes in propulsion cycle characteristics resulting from fatigue (Figure 6). The temporal characteristics of the propulsion cycle at minutes 3 and 6 were different during the heavy exercise bout. The cycle time decreased because the recovery time declined, while the push time remained constant. The push time as a proportion of the cycle time also increased with fatigue.

Rodgers et al20 observed no temporal characteristic modifications of the propulsion cycle timing over time during exercise bouts at 50 and 75% of for wheelchair users. However, in nonwheelchair users, Rodgers et al21 found the same fatigue effects at 75% of as were observed in the present study. Furthermore, measures of cycle time in the present study (∼1.17 s) were within the 0.8–1.5 s range reported in other studies.63 The push time (24–43% of the cycle time) and recovery time (57–76% of the cycle time) values observed in our study were also similar to previous values.18, 19, 21, 64

Our findings are consistent with the literature and support the belief that fatigue results in modifications of intersegmental coordination. Consequences of these modifications are: (1) an increase of the muscle activation time, which could contribute to the slow component effect but has not yet been investigated; and (2) possible alterations of muscle coordination during exercise, which play a critical role in the development of musculoskeletal injuries.65 Changes in muscle coordination suggest that muscles crossing the joint fatigue at different rates during exhausting activities and produce an unbalanced force distribution around the joint. This unbalanced force distribution could potentially cause unnatural motions of the joint, creating abnormal joint loading. Trying to find a less strenuous propulsion technique could be a valuable perspective to help the wheelchair users and reduce the injury prevalence of this population. However, the mechanisms leading to muscular imbalances remain unclear.

Conclusion

The present study reveals that several effects occur during a heavy intensity bout of wheelchair exercise. The fatiguing exercise induces: (1) a slow component; (2) fatigue of numerous muscles acting around the shoulder joint, particularly muscles involved throughout the propulsion phase; and (3) modifications of propulsion and muscle activation timing. These effects could alter muscle coordination around the joint and increase the risk of injury.