Respiration Physiology. 84 (1991) 49-60 Elsevier
49
RESP 01760
Recovery from fatigue of human diaphragm and limb muscles D.K. McKenzie and S.C. Gandevia Department of Clinical Neurophysiology, Institute of Neurological Sciences and Department of Respiratory Medicine, The Prince Henry & Prince of Wales Hospitals and School of Medicine, University of New South Wales, Sydney, Australia. (Accepted 19 December 1990) Abstract. This study was designed to compare the recovery from fatigue of human inspiratory and limb muscles using repeated maximal static contractions. Series of 18 maximal contractions of 10 see duration were performed with a duty cycle of 50% for maximal inspiratory efforts (against a shutter at FRC), and with duty cycles of 5%, 10%, 20% and 50% for the elbow flexors in repeated studies on 6 subjects. The peak inspiratory pressure at the end of the series declined to 86.7% _+5.3% (mean + S.D.) of its initial value; maximal force ofthe elbow flexors declined to 83.5% + 7.0% (5% duty cycle), 80.0% + 5.5% (10% duty cycle), 70.0% + 9.3% (20% duty cycle), and 66.4% + 8.0% (50% duty cycle). Thus, the elbow flexors required approximately a 10-fold reduction in duty cycle to maintain over a series of contractions a force generating capacity comparable to that of the diaphragm. A small degree of 'central' fatigue developed progressively during all series of contractions but did not correlate with duty cycle. Fatigue-induced changes in twitch contraction properties varied with changes in duty cycle. Our major conclusions are that the human diaphragm has a marked capacity to recover from fatigue and that this may have been underestimated in previous studies from this and other laboratories.
Muscle fatigue; Respiratory muscle
Of the few direct comparisons of the endurance capacity of human respiratory and limb muscles, an endurance test based on repeated maximal static voluntary contractions showed that the relative decline in force over repeated contractions was greater for the flexors and extensors of the elbow joint than for the inspiratory muscles (Gandevia etal., 1983; cf Bellemare and Grassino, 1982). By comparison of separate data obtained for the human diaphragm and quadriceps femoris a relative enhancement of diaphragmatic endurance can also be observed (Bellemare and Bigland-Ritchie, 1987; Vollestad et al., 1988). Subsequent studies revealed that final force of the diaphragm during repeated maximal static expulsive efforts was intermediate between that during inspiratory efforts (without elevation of abdominal pressure) and that ofthe elbow flexor Correspondence to: S.C. Gandevia, Dept. of Clinical Neurophysiology, Institute of Neurological Sciences, The Prince Henry Hospital, P.O. Box 233, Matraville, N.S.W. 2036, Sydney, Australia. 0034-5687/91/$03.50 © 1991 Elsevier Science Publishers B.V.
50
D.K. McKENZIE AND S.C. GANDEVIA
muscles (Gandevia and McKenzie, 1988a). It was postulated that the difference in diaphragmatic endurance between the two manoeuvres reflected impairment of diaphragmatic perfusion during expulsive efforts due to the marked elevation of abdominal pressure (Pab) (Gandevia and McKenzie, 1988a; see also Buchler et al., 1985). In our standard endurance test the relative force sustained after 18 maximal contractions (10 sec contraction, 10 sec rest, i. e. 50% duty cycle) declines to about 65% for the elbow flexors but to only 90 ~o for the inspiratory muscles (e.g. Gandevia et al., 1983; Gandevia and McKenzie, 1988a). At a simplistic level this may suggest that diaphragmatic endurance was 40 ~o greater than that of the limb muscles. The present study was designed to evaluate this projection using a different approach. We aimed to determine how much the duty cycle of the contractions of the limb muscle needed to be reduced (by increasing the rest interval between contractions) for the final force produced by the limb muscle to equal that of the diaphragm performing inspiratory and expulsive manoeuvres. This would provide an alternative measure of the relative endurance capacity of the diaphragm. A second reason to adopt this strategy was that with the standard endurance test we have documented small enhancements in the final inspiratory force produced by severe asthmatics and patients with chronic airflow limitation compared with control groups (McKenzie et ai., 1986; Newell et al., 1989). However, the functional implication of these increases in performance was not revealed. A sensitive form of twitch interpolation was used in the present study to quantitate the extent to which maximal voluntary contractions may produce submaximal force due to inadequate voluntary drive (Gandevia and McKenzie, 1988b). The results show that the elbow muscles require a 10-fold reduction in duty cycle in order to sustain force as well as the diaphragm during inspiratory manoeuvres.
Methods Experiments were performed on six healthy adult volunteers (4 males and 2 females, aged 21 to 38 years). All subjects were familiar with performing maximal voluntary efforts with respiratory and limb muscles but only two were aware of the hypotheses being tested. Each subject performed one endurance test of the inspiratory muscles (18 maximal contractions of 10 sec duration with rest intervals of 10 sec, 50~o duty cycle) and four endurance tests (with different duty cycles; see below) of the elbow flexors. The five study days were separated by at least 48 hours and the different tests were presented in pseudo-random order. All subjects gave informed consent and the studies were approved by the institutional ethics committee. Unless otherwise qualified the term endurance is used to refer to the peak or mean force produced at the end of a sequence of maximal voluntary contractions. Limb muscle protocol. Maximal contractions of the elbow flexors were performed with the subject seated at a table and the fully supinated forearm fixed to a vertical isometric
RECOVERY FROM FATIGUE
51
myograph and the elbow flexed at 90 ° (see Gandevia et al., 1983; McKenzie and Gandevia, 1987 for details). Torque was monitored continuously on an oscilloscope and by a microcomputer sampling at 100 Hz. It was also displayed to the subject on an array of light emitting diodes. During maximal efforts the subject was continually and loudly exhorted to keep as many lights illuminated as possible. However, to remove feedback about the absolute decline in force, the input gain to the device was changed between contractions. Prior to each endurance test the subject performed at least three brief maximal voluntary efforts ('trial' maxima) separated by intervals of 1 min. The subject then performed 18 maximal voluntary isometric contractions each lasting 10 sec and separated by rest intervals which varied between tests (10 sec, 40 sec, 100 sec and 200 sec corresponding to duty cycles of 50%, 20%, 10% and 5.0%, respectively). For each contraction the peak force (which always occurred in first 2 sec) and the average force sustained over the 10 sec contraction were calculated on-line. Both measures of force were normalized to the highest values achieved in the first two contractions of the series and expressed as a percentage.
Twitch interpolation. To assess the extent of voluntary activation and thus quantitate the degree of'central fatigue', electrical stimuli were delivered to the elbow flexors during maximal efforts early and late in the series of contractions. If such stimuli, delivered during an attempted maximal voluntary contraction, evoke a force increment, then voluntary drive to the muscle was submaximal (e.g. Belanger and McComas, 1981). After the arm was positioned in the myograph 8 mm square electrodes (aluminium foil wrapped in gauze, soaked in saline and covered with electrode paste) were taped over the motor point of biceps brachii and the distal biceps tendon. To prevent desiccation of the electrodes, the upper arm was covered with plastic wrap. The stability of this electrode arrangement was tested in two control studies. First, the amplitude of the twitch responses to supramaximal electrical stimuli (see below) was stable (range + 10% of initial response) over 2 hours. Second, the amplitude of these responses was monitored for up to 3 hours after an endurance test until the response returned to at least 90% of the pre-fatigue value. Prior to each endurance test the stimulus intensity required for a maximal response of relaxed, rested muscle to a single electrical shock was determined and then increased 15-20% for the remainder of the experiment (fig. 1A). The amplitude and contraction time (stimulus to peak) for these (unpotentiated) responses to single stimuli (100 #sec duration) and pairs of stimuli (10 msec interval) were measured (fig. 1B) using a digital oscilloscope equipped with cursors. The subject then performed three brief (3-5 sec) maximal efforts with interpolation of paired stimuli at maximal voluntary force levels to assess the degree of voluntary activation. As shown in fig. 2, the small time-locked fluctuations of force evoked by the stimuli were captured using a 'sample-and-hold' amplifier which zeroed the voluntary force at the time of stimulation and amplified the subsequent response 10 times for display on a digital oscilloscope. Immediately after the maximal efforts (which 'potentiated' the relaxed twitch), responses to single and paired
52
D.K. McKENZIE AND S.C. GANDEVIA
D ~
contraction
//-~-,.~ "~'~,~j num;er 13
j
] 100ms
/
4
j 100ms
Fig. 1. Data from a typical subject to show the behaviour ofthe twitch responses. All stimuli delivered with the muscle relaxed (at arrows). (A) 2 superimposed twitch responses for the rested elbow flexors. The stimulus intensity was increased 20% between them. (B) the response to a paired stimulus (10 msec interval) is superimposed on the smaller response to a single stimulus. (C) single twitch responses of relaxed muscle produced before and after potentiation by a maximal voluntary contraction. (D) single twitch responses obtained during an endurance run with a duty cycle of 50% (see Methods). The twitch contraction time and the half-relaxation times are joined by a dotted line for responses obtained after the control and the sixteenth maximal contraction. Both parameters oftwitch speed increase as twitch amplitude decreases due to fatigue. Vertical calibration: 5 Nm.
stimuli were obtained (fig. IC). The degree of failure of voluntary activation of the (stimulated) muscle was quantitated as the ratio of the twitch response (to twin stimuli) during maximal effort to the potentiated twitch response (to twin stimuli) obtained immediately after (see fig. 2B,C). It was expressed as a percentage and is referred to as the 'failure index'. For each subject the lower two of the three values were retained for analysis of group data. The subjects then performed the endurance series of 18 maximal efforts. Voluntary activation was assessed using twitch interpolation with twin stimuli (see above) during the 3rd, 6th, 15th and 18th contractions (e.g. fig. 2C). For each subject the values for the failure of voluntary activation from the two early tests and the two late tests were averaged for analysis of the group data. The potentiated resting twitch responses to single stimuli were measured after the 1st, 4th, 7th, 10th, 13th and 16th contractions (fig. 1D).
lnspiratory muscle protocol.
Respiratory measurements were similar to those reported previously (McKenzie and Gandevia, 1987). In brief, measurements of thoracic gas volume (FRC, TLC) were made using a Boyle's law method and normality of airway function was confirmed with maximal expiratory flow-volume manoeuvres. The relationship between maximal static inspiratory pressure and absolute lung volume was deter-
53
RECOVERY FROM FATIGUE
B (enlarged)
t m
20 Nm Trial MVC , ls
t-~Trial MVC /
't
/
/x .ai° '100ms'
C
t !
Contraction3
I
100 ms
Contraction 18
Fig. 2. Typical data from one subject for the study of the elbow flexors (duty cycle 50%). (A)a trial maximal voluntary contraction with the response to twin stimuli (10 ms interval) indicated by the arrow during, and after the contraction. These records are shown at a low gain and on a slow time-base. In panels B and C the records are shown at higher magnification. (B) the response to stimuli interpolated during the maximal contraction (in panel A, box at left) is shown at high gain and superimposed on a control twitch response (twin stimuli; Panel A, box at right) at rest. The response to interpolated stimuli is shown at increased gain ( × 2) relative to the control. (C) control twitch and the responses to interpolated stimuli for a contraction early (No. 3) and late (No. 18) in the endurance sequence. The co-,trol twitch responses were obtained immediately after the respectivv maximal voluntary contraction. Note the increase in twitch force to the stimuli interpolated during these contractions (i.e. failure of voluntary activation).
mined with the subject seated in a body plethysmograph. Subjects then performed a series of 18 maximal inspiratory efforts of 10 see duration and with a 50% duty cycle. Efforts commenced at FRC (measured by exhalation from TLC and by Boyle's law). Subjects were provided with visual feedback of airway pressure and they received encouragement as for the studies of elbow flexors. Absolute lung volume, airway pressure and time were sampled at 100 Hz. Data was measured in the same way as for limb contractions except that pressure measurements were corrected for variation in the absolute lung volume at which the contractions commenced using each individual's previously obtained maximal static pressure-volume curve (McKenzie and Gandevia, 1987). Unless stated otherwise, the mean value and the standard deviation of the mean are reported. Student's t-tests and analysis of variance were used to compare endurance of the two muscle groups and the influence of different duty cycles. Statistics.
54
D.K. McKENZIE AND S.C. GANDEVIA
Results
Maximal force and resting twitch properties. Maximal inspiratory pressure attained at FRC ranged from - 110 to - 136 cmH20 (mean + SD; - 123 + 10 cmH20) for the male subjects and was - 69 and - 109 cmH20 for the two females. Maximal torque of the elbow flexors ranged from 52 to 73 Nm in the male subjects and was 34 and 38 Nm in the females. These values are within the normal ranges established in this laboratory. The amplitude of the twitch response to single supramaximal stimuli ranged from 12 to 20% (mean 15.8% + 2.5) of the maximal voluntary force (MVC) while that for paired stimuli (10 ms interval) ranged from 18 to 28% (mean 23.3% + 3.6). These twitch responses had been potentiated by three brief maximal voluntary contractions. The contraction time for potentiated twitches (stimulus to peak) ranged from 61 to 87 msec (mean 74 + 7.5 msec). Coefficients of variation within individuals for measurements made in the four separate tests were less than 5% for MVC and less than 12% for twitch amplitude and contraction time.
Inspiratory muscle performance.
Over the series of 18 maximal inspiratory efforts performed at FRC with a duty cycle of ~ 50%, force declined to 86.7 + 5.3% of its initial maximal value for peak inspiratory pressure and to 90.2 + 6.2% for the mean pressure sustained throughout contractions. Figure 3 illustrates that this performance is similar to that of the elbow flexors contracting with a duty cycle of 5 % (i. e. a 20-fold increase in the rest intervals between contractions, see above).
Endurance of elbow flexors.
With a duty cycle of 50~o the relative force produced by the elbow flexors declined rapidly over the series of maximal contractions (see fig. 4). The peak force attained in the last 2 contractions (mean)was 66.4 +_ 8.0% ofthe highest force achieved in the first two contractions. As the duty cycle for contractions of the PEAK
AVERAGE
100.
100 •
90
90
80
80
70
70,
Foroe
(%max)
60
o Elbowflexors5% DC • Inspiratory50% OC
50
60
50 Contraction number
Contraction number
Fig. 3. Data from the group of subjects for the elbow flexors (5?/o duty cycle: open circles) and the inspiratory muscles (50% duty cycle: closed circles). Mean _+ SEM is plotted for the peak forces (left) and for the average sustained force (right).
RECOVERY FROM FATIGUE PEAK
55
AVERAGE
100.
100
90
90
80
80
Force
(°/,max) 70
60
70
05% • 10% --20%
60
0505, 50
50
~nlTa~ion number
ConlfaGfionnumber
Fig. 4. Mean data ( + SEM) for the group of subjects obtained with the 4 duty cycles for the contraction of the elbow flexors. Data are shown for the peak force achieved in each contraction (left) and the average force sustained in the contraction (right). For clarity the SEM is plotted only for the 5% and 50% duty cycle data but was of similar magnitude for the other duty cycles.
elbow flexors decreased there was an increase in the relative peak force sustained over the series of contractions from 66.4 +_ 8.0% (50% duty cycle) to 70.0 _+ 9.3% (20% duty cycle), 80.0 + 5.5~ (10% duty cycle) and 83.5 _+ 7.070 (5% duty cycle). Performance differed significantly in the tests with different duty cycle (ANOVA, P < 0.01). To assess whether variation in central fatigue could have influenced the results for the elbow flexors, twitch interpolation (see Methods) was used to quantitate the failure of voluntary drive (fig. 5). In the brief maximal trial contractions prior to the endurance test, subjects activated by voluntary effort at least 98% of the stimulated muscle mass
12 g 10
~ 5% 10% g 20% 50%
CONTROL
EARLY
LATE
Fig. 5. Data for the 'failare index' obtained during the control trial maximal contractions of the elbow flexors, and then early (contraction nos. 3 and 6) and late (nos. 15 and 18) in the sequence of 18 contractions. Means ( _+SEM) are plotted for the 4 different duty cycles (5%, 10%, 20% and 50%). The failure index is derived by expressing the response to interpolated stimuli as a percentage of the relevant resting twitch response. The ability "£o obtain optimal force output from the muscle was high during the initial control comractions but deteriorated by the end of the contraction sequence.
56
D,K. McKENZIE AND S.C. GANDEVIA Twitch time
Twitch force 120
100.
eo.
11o
\
7O
Time
(%max)
5O
90
0 5% • 10°/., 4O
,'20% u50%
80
3O
ConVaction number
Contraction number
Fig. 6. Data from the group of subjects for the twitch force (left) and the twitch contraction time (right). Mean responses are plotted (+ SEM for 5% and 50% duty cycle sequences). Twitch contraction force declined most for the sequence with a 50% duty cycle, and was associated with an increase in twitch contraction time. The 5 % and 10% duty cycle runs were associated with less reduction in twitch force and a slight reduction in twitch contraction time.
(range 95-100%). All 6 subjects were capable of full voluntary activation (i. e. no force increase to the interpolated electrical stimulus)on at least one trial maximal contraction during the complete series of tests and two subjects achieved this on each experimental day. Over the 18 maximal voluntary contractions, there was a trend for the degree of central fatigue to increase with time (figs. 2 and 5). However, there was no correlation with duty cycle. By the end of the endurance tests subjects failed on average to activate 4-10% of the stimulated muscle involved in production of elbow flexor torque. Twitch properties during fatigue. Over all series of contractions the amplitude of the potentiated response to single stimuli to elbow flexors decreased progressively (fig. 6, left; see also fig. I D). The decline in twitch force was relatively greater than the decline in maximal voluntary force and this relationship did not appear to be influenced by changes in duty cycle. By contrast, the decline in amplitude of the response to paired stimuli (10 msec interval) more closely paralleled the decline in voluntary force. Twitch contraction time increased about 20% during contractions at a duty cycle of 50% and about 5% for a duty cycle of 20% ; however, it decreased about 5% for duty cycles of 10% and 5% (fig. 6, right).
Discussion
The capacity for endurance of the diaphragm in vitro has often been noted (e.g. Faulkner et al., 1979; Gandevia et al., 1983; Maxwell et al., 1983) although there have been few direct comparisons of the performance ofhuman limb and respiratory muscles
RECOVERY FROM FATIGUE
57
in vivo. Using an endurance test based on repeated maximal static contractions with a duty cycle of 50~o, we have documented previously (and confirmed here) that after 6 mins the final force produced by the inspiratory muscles (relative to the initial force) was about 40~0 greater than that for the elbow flexors in normal subjects (Gandevia et al., 1983; McKenzie and Gandevia, 1987; Gandevia and McKenzie 1988a), and patients with airway disease (McKenzie and Gandevia, 1986; Newell et al., 1989). However, the present study shows for the first time that the elbow flexors require a 10-fold reduction in duty cycle for maximal static contractions to sustain a force comparable to that of the diaphragm. This present result could not be accounted for by variation in the degree of central fatigue with changing duty cycle and is unlikely to reflect a difference in the extent of voluntary activation of the elbow flexors and diaphragm (see below). Critique of methods. The present comparison of inspiratory and limb muscle performance relies on the assumption that the conditions of contraction were similar for the two muscle groups. Based on twitch interpolation the diaphragm can be fully activated in briefinspiratory and expuisive manoeuvres (Bellemare and Bigland-Ritchie, 1984; Gandevia and McKenzie, 1985; Gandevia et al., 1990), as can most limb muscles (e.g. Belanger and McComas, 1981; Bellemare et al., 1983; Gandevia and McKenzie, 1988b). The well-known difference in maximal transdiaphragmatic pressure between inspiratory and expulsive manoeuvres reflects differences between the manoeuvres in the static and dynamic properties of the diaphragm rather than variable activation of the phrenic motoneurone pool (Gandevia et ai., 1990). Four of the 6 subjects in the present study had previously demonstrated that they were capable (in some trials) of extinction of the twitch response to bilateral supramaximal stimulation of the phrenic nerves at inspiratory pressures within 5~o of those achieved here (Gandevia et aL, 1990). The enhancement of diaphragmatic performance compared with elbow flexors could reflect less central fatigue of the diaphragm over the series of contractions. However, Bellemare and Bigland-Ritchie (1987) have provided evidence from twitch interpolation to suggest that in prolonged tasks (in which abdominal pressure increased) the diaphragm is more prone than limb muscles to central fatigue. In a limited number of experiments, we observed a small increase in central fatique of the diaphragm (over 6 min) comparable to that documented for the elbow flexors in the present study (Gandevia and McKenzie, 1988a). The progressive increase in central fatigue quantitated here for elbow flexors has been noted previously for a single duty cycle (Lloyd et al., 1990). The trend for the increase in central fatigue to be smaller for the 50~,, duty cycle may reflect the shorter total duration of the test. Subjects became uncomfortable sitting with the arm restrained for the long periods required to study the other duty cycles. With the 5% and 10~o duty cycles the wrist clamp was loosened during some of the long rest intervals but this was not possible for the 20 % duty cycle. Influence offatigue on twitch amplitude and contraction time. The influence of duty cycle on fatigue-induced changes in twitch characteristics has not been reported previously.
58
D.K. McKENZIE AND S.C. GANDEVIA
A progressive increase of twitch contraction time was observed with duty cycles of 50Yo and 20~o but with duty cycles of 10~o and 5 9/0 there was significant shortening of twitch contraction time. There appears to be a critical duration of the aerobic rest interval between contractions below which twitch prolongation occurs. If the dynamics of the twitch contraction reflect key aspects of force generation at an intracellular level, our data highlight the differences in fatigue produced simply by changing the rest interval between contractions. The changes in twitch time are unlikely to be explained by increases in muscle temperature. We confirm that twitch amplitude declines at a greater rate than force in maximal voluntary contractions with a high duty cycle, and that contraction and relaxation times increase (e.g. Edwards et al., 1977). The mechanisms responisble for the slowing of the twitch contraction and relaxation times in human muscle fatigue remain to be fully elucidated but are likely to involve changes in intracellular pH and inorganic phosphate (e.g. Dawson et al., 1980; Cady et al., 1989). The accentuated decline in twitch amplitude has been attributed to a type of failure of excitation-contraction coupling which can be overcome by high-frequency stimulation or voluntary effort (Edwards et al., 1977). In the present study the relationship between the decline in twitch amplitude and the decline in maximal voluntary force appeared relatively independent of changes in duty cycle suggesting that the mechanism is long lasting and independent of changes in contractile speed. Conclusions and implications for previous studies. Based on these studies we conclude that the human diaphragm has a higher endurance capacity than the elbow flexors and that this difference may have been underestimated in previous studies from this laboratory (Gandevia etal., 1983; Gandevia and McKenzie, 1988a). The mammalian
90
Insp I
80 Final force (% max) 70 Limb 60
50
0
10
20
30
40
50
Duty cycle (%) Fig. 7. Relationship between final force of the elbow flexors after 18 maximal efforts of 10 see duration and 50~ duty cycle (filled circles). A logarithmic curve was fitted to the data (r - 0.97). For comparison, data for the diaphragm during inspiratory efforts (present study, open circles) and expulsive efforts (Gandevia and McKenzie, 1988a, open squares) with a duty cycle of 500~, have been added. The dashed line indicates the final force of elbow flexors with a duty cycle of 5 %.
RECOVERY FROM FATIGUE
59
diaphragm is well adapted for aerobic work: the upper limit of its blood flow is two to four times that available to most limb muscles (e.g. Faulkner et al., 1979; Reid and Johnson, 1983); it has a relatively high capillary density, mitochondrial volume density, and maximal oxygen consumption (e.g. Hoppeler et al., 1981), although there is no agreement about the relative oxidative capacity of its muscle fibres (e.g. Faulkner et al., 1979; Metzger and Fitts, 1986; Sieck et al., 1986). The present results also allow three earlier studies of diaphragmatic endurance to be reinterpreted. When the diaphragm performed comparable series of maximal static expulsive contractions with a duty cycle of 50~o its final force was about 80% of initial force, intermediate between that during inspiratory efforts and that of the elbow flexors (Gandevia and McKenzie, 1988a). The non-linear relationship between duty cycle and final maximal force (fig. 7) means that the elbow flexors require a five-fold reduction in duty cycle to match the force generating capacity of the diaphragm during repetitive, presumably ischaemic contractions (Buchler et al., 1985). We have also noted small, but statistically significant, enhancements of diaphragmatic endurance in patients with asthma (McKenzie and Gandevia, 1986) and chronic airflow limitation (Newell et al., 1989) compared with matched control groups. These small differences in relative force at the end of series of contractions (5-10%), presumably due to muscle training, were originally considered to be of doubtful physiological significance. However, these small differences may reflect substantial adaptive changes in the muscle in terms of recovery from fatigue.
Acknowledgements. This study was supported by the Asthma Foundation of New South Wales and the National Health and Medical Research Council of Australia. We are most grateful to Mr R. B. German and Mr J. P. Hales for expert technical assistance and to Dr B. Bigland-Ritchie for comments on the manuscript.
References Bolanger, A.Y. and A.J. McComas (1981). Extent of motor unit activation during effort. J. Appl. Physiol. 51: 1131-1135. Bellemare, F. and A. Grassino (1982). Effect of pressure and timing of a contraction on human diaphragm fatigue. J. AppL Physiol. 53:1190-1195. Beilemare, F., J.J. Woods, R.S. Johansson and B. Bigland-Ritchie (1983). Motor unit discharge rates in maximal voluntary contractions of three human muscles. J. Neurophysioi. 50: 1380-1392. Bellemare, F. and B. Bigland-Ritchie (1984). ~ ssessment of human diaphragm strength and activation using phrenic nerve stimulation. Resp. Physiol 58: 263-277. Bellemare, F. and B. Bigland-Ritchie (1987). Central components of diaphragmatic fatigue assessed by phrenic nerve stimulation. J. AppL Physiol. 62: 1307-1316. Buchler, B., S. Magder, H. Katsardis, Y. )amines and C. Roussos (1985). Effects of pleural pressure and abdominal pressure on diaphragmatic blood flow. J. Appl. Physiol. 58: 691-697. Cady, E. B., H. Eishove, D.A. Jones and A. Moll (1989). The metabolic causes of slow relaxation in fatigued human skeletal muscle. J. Physiol. (Londen) 418: 327-337. Dawson, M.J., D.G. Gadian and D.R. Wilkie (1980). Mechanical relaxation rate and metabolism s:udied in fatiguing muscle by phosphorus magnetic resonance. J. Physiol. (London) 299: 465-484. Edwards, R.H.T., D.K. Hill, D.A. Jones and P.A. Merton (1977). Fatigue of long duration in human skeletal muscle after exercise. J. Physiol. (London) 272: 769-778.
60
D.K. McKENZIE AND S.C. GANDEVIA
Faulkner, J.A., L.C. Maxwell, G.F. Ruff and T.P. White (1979). The diaphragm as a muscle: contractile properties. Am. Rev. Resp. Dis. 119: Suppl. 89-92. Gandevia, S.C., D.K. McKenzie and I.R. Neering (1983). Endurance properties of respiratory and limb muscles. Respir. Physiol. 53: 47-61. Gandevia, S.C. and D.K. McKenzie (1985). Activation of the human diaphragm during maximal static efforts. J. Physiol. (London) 367: 45-56. Gandevia, S.C. and D.K. McKenzie (1988a). Human diaphragmatic endurance during different maximal respiratory efforts. J. Physiol. (London) 395: 625-638. Gandevia, S.C. and D.K. McKenzie (1988b). Activation of human muscles at short muscle lengths during maximal static efforts. J. Physiol. (London) 407: 599-613. Gandevia, S.C., D.K. McKenzie and B.L. Plassman (1990). Activation of human respiratory muscles during different voluntary manoeuvres, J. Physiol. (London) 428: 387-403. Hoppeler, H., O. Mathieu, R. Krauer, H. Ciaasen, R. B. Armstrong and E. R. Weibei (1981a). Design of the mammalian respiratory system. IV. Distribution of mitochondria and capillaries in various muscles. Respir. Physiol. 44:87-111. Lloyd, A.R., S.C. Gandevia and J.P. Hales (1990). Muscle performance, voluntary activation, twitch properties and perceived effort in normal subjects and patients with chronic fatigue syndrome. Brain 113, in press. Maxwell, L.C., R J. M. McCarter, T.J. Kuehl and J. U Robotham (1983). Development of histochemical and functional properties of baboon respiratory muscles. J. Appl. Physiol. 54: 551-561. McKenzie, D.K. and S.C. Gandevia (1986). Strength and endurance of inspiratory, expiratory, and limb muscles in asthma. Am. Rev. Respir. Dis. 134: 999-1004. McKenzie, D.K. and S.C. Gandevia (1987). Influence of muscle length on human inspiratory and limb muscle endurance. Respir. Physiol. 67: 171-183. Metzger, J.M. and R.H. Fitts (1986). Contractile and biochemical properties of diaphragm: effects of exercise training and fatigue. J. Appl. Physiol. 60: 1752-1758. Newell, S.Z., D.K. McKenzie and S.C. Gandevia (1989). inspiratory and skeletal muscle strength, endurance and diaphragmatic activation in patients with chronic airflow limitation. Thorax 44: 903-912. Reid, M.B. and R.L. Johnson (1983). Efficiency, maximal blood flow, and aerobic work capacity of the canine diaphragm. J. Appl. Physiol. 54: 763-772. Sleek, G.C., R. D, Sacks, C.E. Blanco and V.R. Edgerton (1986). SDH activity and oross-seotional area of muscle fibers in cat diaphragm. J. Appi, Physiol. 60: 1284-1292. Vollostad, N.K., O.M, Sejersted, R. Bahr, J.L Woods and B. Bigland-Ritchie (1988). Motor drive and metabolic responses during repeated submaximal contractions in humans. J. Appl. Physiol. 64: 1421-1427,
49
RESP 01760
Recovery from fatigue of human diaphragm and limb muscles D.K. McKenzie and S.C. Gandevia Department of Clinical Neurophysiology, Institute of Neurological Sciences and Department of Respiratory Medicine, The Prince Henry & Prince of Wales Hospitals and School of Medicine, University of New South Wales, Sydney, Australia. (Accepted 19 December 1990) Abstract. This study was designed to compare the recovery from fatigue of human inspiratory and limb muscles using repeated maximal static contractions. Series of 18 maximal contractions of 10 see duration were performed with a duty cycle of 50% for maximal inspiratory efforts (against a shutter at FRC), and with duty cycles of 5%, 10%, 20% and 50% for the elbow flexors in repeated studies on 6 subjects. The peak inspiratory pressure at the end of the series declined to 86.7% _+5.3% (mean + S.D.) of its initial value; maximal force ofthe elbow flexors declined to 83.5% + 7.0% (5% duty cycle), 80.0% + 5.5% (10% duty cycle), 70.0% + 9.3% (20% duty cycle), and 66.4% + 8.0% (50% duty cycle). Thus, the elbow flexors required approximately a 10-fold reduction in duty cycle to maintain over a series of contractions a force generating capacity comparable to that of the diaphragm. A small degree of 'central' fatigue developed progressively during all series of contractions but did not correlate with duty cycle. Fatigue-induced changes in twitch contraction properties varied with changes in duty cycle. Our major conclusions are that the human diaphragm has a marked capacity to recover from fatigue and that this may have been underestimated in previous studies from this and other laboratories.
Muscle fatigue; Respiratory muscle
Of the few direct comparisons of the endurance capacity of human respiratory and limb muscles, an endurance test based on repeated maximal static voluntary contractions showed that the relative decline in force over repeated contractions was greater for the flexors and extensors of the elbow joint than for the inspiratory muscles (Gandevia etal., 1983; cf Bellemare and Grassino, 1982). By comparison of separate data obtained for the human diaphragm and quadriceps femoris a relative enhancement of diaphragmatic endurance can also be observed (Bellemare and Bigland-Ritchie, 1987; Vollestad et al., 1988). Subsequent studies revealed that final force of the diaphragm during repeated maximal static expulsive efforts was intermediate between that during inspiratory efforts (without elevation of abdominal pressure) and that ofthe elbow flexor Correspondence to: S.C. Gandevia, Dept. of Clinical Neurophysiology, Institute of Neurological Sciences, The Prince Henry Hospital, P.O. Box 233, Matraville, N.S.W. 2036, Sydney, Australia. 0034-5687/91/$03.50 © 1991 Elsevier Science Publishers B.V.
50
D.K. McKENZIE AND S.C. GANDEVIA
muscles (Gandevia and McKenzie, 1988a). It was postulated that the difference in diaphragmatic endurance between the two manoeuvres reflected impairment of diaphragmatic perfusion during expulsive efforts due to the marked elevation of abdominal pressure (Pab) (Gandevia and McKenzie, 1988a; see also Buchler et al., 1985). In our standard endurance test the relative force sustained after 18 maximal contractions (10 sec contraction, 10 sec rest, i. e. 50% duty cycle) declines to about 65% for the elbow flexors but to only 90 ~o for the inspiratory muscles (e.g. Gandevia et al., 1983; Gandevia and McKenzie, 1988a). At a simplistic level this may suggest that diaphragmatic endurance was 40 ~o greater than that of the limb muscles. The present study was designed to evaluate this projection using a different approach. We aimed to determine how much the duty cycle of the contractions of the limb muscle needed to be reduced (by increasing the rest interval between contractions) for the final force produced by the limb muscle to equal that of the diaphragm performing inspiratory and expulsive manoeuvres. This would provide an alternative measure of the relative endurance capacity of the diaphragm. A second reason to adopt this strategy was that with the standard endurance test we have documented small enhancements in the final inspiratory force produced by severe asthmatics and patients with chronic airflow limitation compared with control groups (McKenzie et ai., 1986; Newell et al., 1989). However, the functional implication of these increases in performance was not revealed. A sensitive form of twitch interpolation was used in the present study to quantitate the extent to which maximal voluntary contractions may produce submaximal force due to inadequate voluntary drive (Gandevia and McKenzie, 1988b). The results show that the elbow muscles require a 10-fold reduction in duty cycle in order to sustain force as well as the diaphragm during inspiratory manoeuvres.
Methods Experiments were performed on six healthy adult volunteers (4 males and 2 females, aged 21 to 38 years). All subjects were familiar with performing maximal voluntary efforts with respiratory and limb muscles but only two were aware of the hypotheses being tested. Each subject performed one endurance test of the inspiratory muscles (18 maximal contractions of 10 sec duration with rest intervals of 10 sec, 50~o duty cycle) and four endurance tests (with different duty cycles; see below) of the elbow flexors. The five study days were separated by at least 48 hours and the different tests were presented in pseudo-random order. All subjects gave informed consent and the studies were approved by the institutional ethics committee. Unless otherwise qualified the term endurance is used to refer to the peak or mean force produced at the end of a sequence of maximal voluntary contractions. Limb muscle protocol. Maximal contractions of the elbow flexors were performed with the subject seated at a table and the fully supinated forearm fixed to a vertical isometric
RECOVERY FROM FATIGUE
51
myograph and the elbow flexed at 90 ° (see Gandevia et al., 1983; McKenzie and Gandevia, 1987 for details). Torque was monitored continuously on an oscilloscope and by a microcomputer sampling at 100 Hz. It was also displayed to the subject on an array of light emitting diodes. During maximal efforts the subject was continually and loudly exhorted to keep as many lights illuminated as possible. However, to remove feedback about the absolute decline in force, the input gain to the device was changed between contractions. Prior to each endurance test the subject performed at least three brief maximal voluntary efforts ('trial' maxima) separated by intervals of 1 min. The subject then performed 18 maximal voluntary isometric contractions each lasting 10 sec and separated by rest intervals which varied between tests (10 sec, 40 sec, 100 sec and 200 sec corresponding to duty cycles of 50%, 20%, 10% and 5.0%, respectively). For each contraction the peak force (which always occurred in first 2 sec) and the average force sustained over the 10 sec contraction were calculated on-line. Both measures of force were normalized to the highest values achieved in the first two contractions of the series and expressed as a percentage.
Twitch interpolation. To assess the extent of voluntary activation and thus quantitate the degree of'central fatigue', electrical stimuli were delivered to the elbow flexors during maximal efforts early and late in the series of contractions. If such stimuli, delivered during an attempted maximal voluntary contraction, evoke a force increment, then voluntary drive to the muscle was submaximal (e.g. Belanger and McComas, 1981). After the arm was positioned in the myograph 8 mm square electrodes (aluminium foil wrapped in gauze, soaked in saline and covered with electrode paste) were taped over the motor point of biceps brachii and the distal biceps tendon. To prevent desiccation of the electrodes, the upper arm was covered with plastic wrap. The stability of this electrode arrangement was tested in two control studies. First, the amplitude of the twitch responses to supramaximal electrical stimuli (see below) was stable (range + 10% of initial response) over 2 hours. Second, the amplitude of these responses was monitored for up to 3 hours after an endurance test until the response returned to at least 90% of the pre-fatigue value. Prior to each endurance test the stimulus intensity required for a maximal response of relaxed, rested muscle to a single electrical shock was determined and then increased 15-20% for the remainder of the experiment (fig. 1A). The amplitude and contraction time (stimulus to peak) for these (unpotentiated) responses to single stimuli (100 #sec duration) and pairs of stimuli (10 msec interval) were measured (fig. 1B) using a digital oscilloscope equipped with cursors. The subject then performed three brief (3-5 sec) maximal efforts with interpolation of paired stimuli at maximal voluntary force levels to assess the degree of voluntary activation. As shown in fig. 2, the small time-locked fluctuations of force evoked by the stimuli were captured using a 'sample-and-hold' amplifier which zeroed the voluntary force at the time of stimulation and amplified the subsequent response 10 times for display on a digital oscilloscope. Immediately after the maximal efforts (which 'potentiated' the relaxed twitch), responses to single and paired
52
D.K. McKENZIE AND S.C. GANDEVIA
D ~
contraction
//-~-,.~ "~'~,~j num;er 13
j
] 100ms
/
4
j 100ms
Fig. 1. Data from a typical subject to show the behaviour ofthe twitch responses. All stimuli delivered with the muscle relaxed (at arrows). (A) 2 superimposed twitch responses for the rested elbow flexors. The stimulus intensity was increased 20% between them. (B) the response to a paired stimulus (10 msec interval) is superimposed on the smaller response to a single stimulus. (C) single twitch responses of relaxed muscle produced before and after potentiation by a maximal voluntary contraction. (D) single twitch responses obtained during an endurance run with a duty cycle of 50% (see Methods). The twitch contraction time and the half-relaxation times are joined by a dotted line for responses obtained after the control and the sixteenth maximal contraction. Both parameters oftwitch speed increase as twitch amplitude decreases due to fatigue. Vertical calibration: 5 Nm.
stimuli were obtained (fig. IC). The degree of failure of voluntary activation of the (stimulated) muscle was quantitated as the ratio of the twitch response (to twin stimuli) during maximal effort to the potentiated twitch response (to twin stimuli) obtained immediately after (see fig. 2B,C). It was expressed as a percentage and is referred to as the 'failure index'. For each subject the lower two of the three values were retained for analysis of group data. The subjects then performed the endurance series of 18 maximal efforts. Voluntary activation was assessed using twitch interpolation with twin stimuli (see above) during the 3rd, 6th, 15th and 18th contractions (e.g. fig. 2C). For each subject the values for the failure of voluntary activation from the two early tests and the two late tests were averaged for analysis of the group data. The potentiated resting twitch responses to single stimuli were measured after the 1st, 4th, 7th, 10th, 13th and 16th contractions (fig. 1D).
lnspiratory muscle protocol.
Respiratory measurements were similar to those reported previously (McKenzie and Gandevia, 1987). In brief, measurements of thoracic gas volume (FRC, TLC) were made using a Boyle's law method and normality of airway function was confirmed with maximal expiratory flow-volume manoeuvres. The relationship between maximal static inspiratory pressure and absolute lung volume was deter-
53
RECOVERY FROM FATIGUE
B (enlarged)
t m
20 Nm Trial MVC , ls
t-~Trial MVC /
't
/
/x .ai° '100ms'
C
t !
Contraction3
I
100 ms
Contraction 18
Fig. 2. Typical data from one subject for the study of the elbow flexors (duty cycle 50%). (A)a trial maximal voluntary contraction with the response to twin stimuli (10 ms interval) indicated by the arrow during, and after the contraction. These records are shown at a low gain and on a slow time-base. In panels B and C the records are shown at higher magnification. (B) the response to stimuli interpolated during the maximal contraction (in panel A, box at left) is shown at high gain and superimposed on a control twitch response (twin stimuli; Panel A, box at right) at rest. The response to interpolated stimuli is shown at increased gain ( × 2) relative to the control. (C) control twitch and the responses to interpolated stimuli for a contraction early (No. 3) and late (No. 18) in the endurance sequence. The co-,trol twitch responses were obtained immediately after the respectivv maximal voluntary contraction. Note the increase in twitch force to the stimuli interpolated during these contractions (i.e. failure of voluntary activation).
mined with the subject seated in a body plethysmograph. Subjects then performed a series of 18 maximal inspiratory efforts of 10 see duration and with a 50% duty cycle. Efforts commenced at FRC (measured by exhalation from TLC and by Boyle's law). Subjects were provided with visual feedback of airway pressure and they received encouragement as for the studies of elbow flexors. Absolute lung volume, airway pressure and time were sampled at 100 Hz. Data was measured in the same way as for limb contractions except that pressure measurements were corrected for variation in the absolute lung volume at which the contractions commenced using each individual's previously obtained maximal static pressure-volume curve (McKenzie and Gandevia, 1987). Unless stated otherwise, the mean value and the standard deviation of the mean are reported. Student's t-tests and analysis of variance were used to compare endurance of the two muscle groups and the influence of different duty cycles. Statistics.
54
D.K. McKENZIE AND S.C. GANDEVIA
Results
Maximal force and resting twitch properties. Maximal inspiratory pressure attained at FRC ranged from - 110 to - 136 cmH20 (mean + SD; - 123 + 10 cmH20) for the male subjects and was - 69 and - 109 cmH20 for the two females. Maximal torque of the elbow flexors ranged from 52 to 73 Nm in the male subjects and was 34 and 38 Nm in the females. These values are within the normal ranges established in this laboratory. The amplitude of the twitch response to single supramaximal stimuli ranged from 12 to 20% (mean 15.8% + 2.5) of the maximal voluntary force (MVC) while that for paired stimuli (10 ms interval) ranged from 18 to 28% (mean 23.3% + 3.6). These twitch responses had been potentiated by three brief maximal voluntary contractions. The contraction time for potentiated twitches (stimulus to peak) ranged from 61 to 87 msec (mean 74 + 7.5 msec). Coefficients of variation within individuals for measurements made in the four separate tests were less than 5% for MVC and less than 12% for twitch amplitude and contraction time.
Inspiratory muscle performance.
Over the series of 18 maximal inspiratory efforts performed at FRC with a duty cycle of ~ 50%, force declined to 86.7 + 5.3% of its initial maximal value for peak inspiratory pressure and to 90.2 + 6.2% for the mean pressure sustained throughout contractions. Figure 3 illustrates that this performance is similar to that of the elbow flexors contracting with a duty cycle of 5 % (i. e. a 20-fold increase in the rest intervals between contractions, see above).
Endurance of elbow flexors.
With a duty cycle of 50~o the relative force produced by the elbow flexors declined rapidly over the series of maximal contractions (see fig. 4). The peak force attained in the last 2 contractions (mean)was 66.4 +_ 8.0% ofthe highest force achieved in the first two contractions. As the duty cycle for contractions of the PEAK
AVERAGE
100.
100 •
90
90
80
80
70
70,
Foroe
(%max)
60
o Elbowflexors5% DC • Inspiratory50% OC
50
60
50 Contraction number
Contraction number
Fig. 3. Data from the group of subjects for the elbow flexors (5?/o duty cycle: open circles) and the inspiratory muscles (50% duty cycle: closed circles). Mean _+ SEM is plotted for the peak forces (left) and for the average sustained force (right).
RECOVERY FROM FATIGUE PEAK
55
AVERAGE
100.
100
90
90
80
80
Force
(°/,max) 70
60
70
05% • 10% --20%
60
0505, 50
50
~nlTa~ion number
ConlfaGfionnumber
Fig. 4. Mean data ( + SEM) for the group of subjects obtained with the 4 duty cycles for the contraction of the elbow flexors. Data are shown for the peak force achieved in each contraction (left) and the average force sustained in the contraction (right). For clarity the SEM is plotted only for the 5% and 50% duty cycle data but was of similar magnitude for the other duty cycles.
elbow flexors decreased there was an increase in the relative peak force sustained over the series of contractions from 66.4 +_ 8.0% (50% duty cycle) to 70.0 _+ 9.3% (20% duty cycle), 80.0 + 5.5~ (10% duty cycle) and 83.5 _+ 7.070 (5% duty cycle). Performance differed significantly in the tests with different duty cycle (ANOVA, P < 0.01). To assess whether variation in central fatigue could have influenced the results for the elbow flexors, twitch interpolation (see Methods) was used to quantitate the failure of voluntary drive (fig. 5). In the brief maximal trial contractions prior to the endurance test, subjects activated by voluntary effort at least 98% of the stimulated muscle mass
12 g 10
~ 5% 10% g 20% 50%
CONTROL
EARLY
LATE
Fig. 5. Data for the 'failare index' obtained during the control trial maximal contractions of the elbow flexors, and then early (contraction nos. 3 and 6) and late (nos. 15 and 18) in the sequence of 18 contractions. Means ( _+SEM) are plotted for the 4 different duty cycles (5%, 10%, 20% and 50%). The failure index is derived by expressing the response to interpolated stimuli as a percentage of the relevant resting twitch response. The ability "£o obtain optimal force output from the muscle was high during the initial control comractions but deteriorated by the end of the contraction sequence.
56
D,K. McKENZIE AND S.C. GANDEVIA Twitch time
Twitch force 120
100.
eo.
11o
\
7O
Time
(%max)
5O
90
0 5% • 10°/., 4O
,'20% u50%
80
3O
ConVaction number
Contraction number
Fig. 6. Data from the group of subjects for the twitch force (left) and the twitch contraction time (right). Mean responses are plotted (+ SEM for 5% and 50% duty cycle sequences). Twitch contraction force declined most for the sequence with a 50% duty cycle, and was associated with an increase in twitch contraction time. The 5 % and 10% duty cycle runs were associated with less reduction in twitch force and a slight reduction in twitch contraction time.
(range 95-100%). All 6 subjects were capable of full voluntary activation (i. e. no force increase to the interpolated electrical stimulus)on at least one trial maximal contraction during the complete series of tests and two subjects achieved this on each experimental day. Over the 18 maximal voluntary contractions, there was a trend for the degree of central fatigue to increase with time (figs. 2 and 5). However, there was no correlation with duty cycle. By the end of the endurance tests subjects failed on average to activate 4-10% of the stimulated muscle involved in production of elbow flexor torque. Twitch properties during fatigue. Over all series of contractions the amplitude of the potentiated response to single stimuli to elbow flexors decreased progressively (fig. 6, left; see also fig. I D). The decline in twitch force was relatively greater than the decline in maximal voluntary force and this relationship did not appear to be influenced by changes in duty cycle. By contrast, the decline in amplitude of the response to paired stimuli (10 msec interval) more closely paralleled the decline in voluntary force. Twitch contraction time increased about 20% during contractions at a duty cycle of 50% and about 5% for a duty cycle of 20% ; however, it decreased about 5% for duty cycles of 10% and 5% (fig. 6, right).
Discussion
The capacity for endurance of the diaphragm in vitro has often been noted (e.g. Faulkner et al., 1979; Gandevia et al., 1983; Maxwell et al., 1983) although there have been few direct comparisons of the performance ofhuman limb and respiratory muscles
RECOVERY FROM FATIGUE
57
in vivo. Using an endurance test based on repeated maximal static contractions with a duty cycle of 50~o, we have documented previously (and confirmed here) that after 6 mins the final force produced by the inspiratory muscles (relative to the initial force) was about 40~0 greater than that for the elbow flexors in normal subjects (Gandevia et al., 1983; McKenzie and Gandevia, 1987; Gandevia and McKenzie 1988a), and patients with airway disease (McKenzie and Gandevia, 1986; Newell et al., 1989). However, the present study shows for the first time that the elbow flexors require a 10-fold reduction in duty cycle for maximal static contractions to sustain a force comparable to that of the diaphragm. This present result could not be accounted for by variation in the degree of central fatigue with changing duty cycle and is unlikely to reflect a difference in the extent of voluntary activation of the elbow flexors and diaphragm (see below). Critique of methods. The present comparison of inspiratory and limb muscle performance relies on the assumption that the conditions of contraction were similar for the two muscle groups. Based on twitch interpolation the diaphragm can be fully activated in briefinspiratory and expuisive manoeuvres (Bellemare and Bigland-Ritchie, 1984; Gandevia and McKenzie, 1985; Gandevia et al., 1990), as can most limb muscles (e.g. Belanger and McComas, 1981; Bellemare et al., 1983; Gandevia and McKenzie, 1988b). The well-known difference in maximal transdiaphragmatic pressure between inspiratory and expulsive manoeuvres reflects differences between the manoeuvres in the static and dynamic properties of the diaphragm rather than variable activation of the phrenic motoneurone pool (Gandevia et ai., 1990). Four of the 6 subjects in the present study had previously demonstrated that they were capable (in some trials) of extinction of the twitch response to bilateral supramaximal stimulation of the phrenic nerves at inspiratory pressures within 5~o of those achieved here (Gandevia et aL, 1990). The enhancement of diaphragmatic performance compared with elbow flexors could reflect less central fatigue of the diaphragm over the series of contractions. However, Bellemare and Bigland-Ritchie (1987) have provided evidence from twitch interpolation to suggest that in prolonged tasks (in which abdominal pressure increased) the diaphragm is more prone than limb muscles to central fatigue. In a limited number of experiments, we observed a small increase in central fatique of the diaphragm (over 6 min) comparable to that documented for the elbow flexors in the present study (Gandevia and McKenzie, 1988a). The progressive increase in central fatigue quantitated here for elbow flexors has been noted previously for a single duty cycle (Lloyd et al., 1990). The trend for the increase in central fatigue to be smaller for the 50~,, duty cycle may reflect the shorter total duration of the test. Subjects became uncomfortable sitting with the arm restrained for the long periods required to study the other duty cycles. With the 5% and 10~o duty cycles the wrist clamp was loosened during some of the long rest intervals but this was not possible for the 20 % duty cycle. Influence offatigue on twitch amplitude and contraction time. The influence of duty cycle on fatigue-induced changes in twitch characteristics has not been reported previously.
58
D.K. McKENZIE AND S.C. GANDEVIA
A progressive increase of twitch contraction time was observed with duty cycles of 50Yo and 20~o but with duty cycles of 10~o and 5 9/0 there was significant shortening of twitch contraction time. There appears to be a critical duration of the aerobic rest interval between contractions below which twitch prolongation occurs. If the dynamics of the twitch contraction reflect key aspects of force generation at an intracellular level, our data highlight the differences in fatigue produced simply by changing the rest interval between contractions. The changes in twitch time are unlikely to be explained by increases in muscle temperature. We confirm that twitch amplitude declines at a greater rate than force in maximal voluntary contractions with a high duty cycle, and that contraction and relaxation times increase (e.g. Edwards et al., 1977). The mechanisms responisble for the slowing of the twitch contraction and relaxation times in human muscle fatigue remain to be fully elucidated but are likely to involve changes in intracellular pH and inorganic phosphate (e.g. Dawson et al., 1980; Cady et al., 1989). The accentuated decline in twitch amplitude has been attributed to a type of failure of excitation-contraction coupling which can be overcome by high-frequency stimulation or voluntary effort (Edwards et al., 1977). In the present study the relationship between the decline in twitch amplitude and the decline in maximal voluntary force appeared relatively independent of changes in duty cycle suggesting that the mechanism is long lasting and independent of changes in contractile speed. Conclusions and implications for previous studies. Based on these studies we conclude that the human diaphragm has a higher endurance capacity than the elbow flexors and that this difference may have been underestimated in previous studies from this laboratory (Gandevia etal., 1983; Gandevia and McKenzie, 1988a). The mammalian
90
Insp I
80 Final force (% max) 70 Limb 60
50
0
10
20
30
40
50
Duty cycle (%) Fig. 7. Relationship between final force of the elbow flexors after 18 maximal efforts of 10 see duration and 50~ duty cycle (filled circles). A logarithmic curve was fitted to the data (r - 0.97). For comparison, data for the diaphragm during inspiratory efforts (present study, open circles) and expulsive efforts (Gandevia and McKenzie, 1988a, open squares) with a duty cycle of 500~, have been added. The dashed line indicates the final force of elbow flexors with a duty cycle of 5 %.
RECOVERY FROM FATIGUE
59
diaphragm is well adapted for aerobic work: the upper limit of its blood flow is two to four times that available to most limb muscles (e.g. Faulkner et al., 1979; Reid and Johnson, 1983); it has a relatively high capillary density, mitochondrial volume density, and maximal oxygen consumption (e.g. Hoppeler et al., 1981), although there is no agreement about the relative oxidative capacity of its muscle fibres (e.g. Faulkner et al., 1979; Metzger and Fitts, 1986; Sieck et al., 1986). The present results also allow three earlier studies of diaphragmatic endurance to be reinterpreted. When the diaphragm performed comparable series of maximal static expulsive contractions with a duty cycle of 50~o its final force was about 80% of initial force, intermediate between that during inspiratory efforts and that of the elbow flexors (Gandevia and McKenzie, 1988a). The non-linear relationship between duty cycle and final maximal force (fig. 7) means that the elbow flexors require a five-fold reduction in duty cycle to match the force generating capacity of the diaphragm during repetitive, presumably ischaemic contractions (Buchler et al., 1985). We have also noted small, but statistically significant, enhancements of diaphragmatic endurance in patients with asthma (McKenzie and Gandevia, 1986) and chronic airflow limitation (Newell et al., 1989) compared with matched control groups. These small differences in relative force at the end of series of contractions (5-10%), presumably due to muscle training, were originally considered to be of doubtful physiological significance. However, these small differences may reflect substantial adaptive changes in the muscle in terms of recovery from fatigue.
Acknowledgements. This study was supported by the Asthma Foundation of New South Wales and the National Health and Medical Research Council of Australia. We are most grateful to Mr R. B. German and Mr J. P. Hales for expert technical assistance and to Dr B. Bigland-Ritchie for comments on the manuscript.
References Bolanger, A.Y. and A.J. McComas (1981). Extent of motor unit activation during effort. J. Appl. Physiol. 51: 1131-1135. Bellemare, F. and A. Grassino (1982). Effect of pressure and timing of a contraction on human diaphragm fatigue. J. AppL Physiol. 53:1190-1195. Beilemare, F., J.J. Woods, R.S. Johansson and B. Bigland-Ritchie (1983). Motor unit discharge rates in maximal voluntary contractions of three human muscles. J. Neurophysioi. 50: 1380-1392. Bellemare, F. and B. Bigland-Ritchie (1984). ~ ssessment of human diaphragm strength and activation using phrenic nerve stimulation. Resp. Physiol 58: 263-277. Bellemare, F. and B. Bigland-Ritchie (1987). Central components of diaphragmatic fatigue assessed by phrenic nerve stimulation. J. AppL Physiol. 62: 1307-1316. Buchler, B., S. Magder, H. Katsardis, Y. )amines and C. Roussos (1985). Effects of pleural pressure and abdominal pressure on diaphragmatic blood flow. J. Appl. Physiol. 58: 691-697. Cady, E. B., H. Eishove, D.A. Jones and A. Moll (1989). The metabolic causes of slow relaxation in fatigued human skeletal muscle. J. Physiol. (Londen) 418: 327-337. Dawson, M.J., D.G. Gadian and D.R. Wilkie (1980). Mechanical relaxation rate and metabolism s:udied in fatiguing muscle by phosphorus magnetic resonance. J. Physiol. (London) 299: 465-484. Edwards, R.H.T., D.K. Hill, D.A. Jones and P.A. Merton (1977). Fatigue of long duration in human skeletal muscle after exercise. J. Physiol. (London) 272: 769-778.
60
D.K. McKENZIE AND S.C. GANDEVIA
Faulkner, J.A., L.C. Maxwell, G.F. Ruff and T.P. White (1979). The diaphragm as a muscle: contractile properties. Am. Rev. Resp. Dis. 119: Suppl. 89-92. Gandevia, S.C., D.K. McKenzie and I.R. Neering (1983). Endurance properties of respiratory and limb muscles. Respir. Physiol. 53: 47-61. Gandevia, S.C. and D.K. McKenzie (1985). Activation of the human diaphragm during maximal static efforts. J. Physiol. (London) 367: 45-56. Gandevia, S.C. and D.K. McKenzie (1988a). Human diaphragmatic endurance during different maximal respiratory efforts. J. Physiol. (London) 395: 625-638. Gandevia, S.C. and D.K. McKenzie (1988b). Activation of human muscles at short muscle lengths during maximal static efforts. J. Physiol. (London) 407: 599-613. Gandevia, S.C., D.K. McKenzie and B.L. Plassman (1990). Activation of human respiratory muscles during different voluntary manoeuvres, J. Physiol. (London) 428: 387-403. Hoppeler, H., O. Mathieu, R. Krauer, H. Ciaasen, R. B. Armstrong and E. R. Weibei (1981a). Design of the mammalian respiratory system. IV. Distribution of mitochondria and capillaries in various muscles. Respir. Physiol. 44:87-111. Lloyd, A.R., S.C. Gandevia and J.P. Hales (1990). Muscle performance, voluntary activation, twitch properties and perceived effort in normal subjects and patients with chronic fatigue syndrome. Brain 113, in press. Maxwell, L.C., R J. M. McCarter, T.J. Kuehl and J. U Robotham (1983). Development of histochemical and functional properties of baboon respiratory muscles. J. Appl. Physiol. 54: 551-561. McKenzie, D.K. and S.C. Gandevia (1986). Strength and endurance of inspiratory, expiratory, and limb muscles in asthma. Am. Rev. Respir. Dis. 134: 999-1004. McKenzie, D.K. and S.C. Gandevia (1987). Influence of muscle length on human inspiratory and limb muscle endurance. Respir. Physiol. 67: 171-183. Metzger, J.M. and R.H. Fitts (1986). Contractile and biochemical properties of diaphragm: effects of exercise training and fatigue. J. Appl. Physiol. 60: 1752-1758. Newell, S.Z., D.K. McKenzie and S.C. Gandevia (1989). inspiratory and skeletal muscle strength, endurance and diaphragmatic activation in patients with chronic airflow limitation. Thorax 44: 903-912. Reid, M.B. and R.L. Johnson (1983). Efficiency, maximal blood flow, and aerobic work capacity of the canine diaphragm. J. Appl. Physiol. 54: 763-772. Sleek, G.C., R. D, Sacks, C.E. Blanco and V.R. Edgerton (1986). SDH activity and oross-seotional area of muscle fibers in cat diaphragm. J. Appi, Physiol. 60: 1284-1292. Vollostad, N.K., O.M, Sejersted, R. Bahr, J.L Woods and B. Bigland-Ritchie (1988). Motor drive and metabolic responses during repeated submaximal contractions in humans. J. Appl. Physiol. 64: 1421-1427,
