| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Physiology (2002), 544.1, pp. 137-147
© Copyright 2002 The Physiological Society
DOI: 10.1113/jphysiol.2002.025254
 |
ABSTRACT |
|---|
 TopAbstract IntroductionMethodsResultsDiscussionReferences |
|---|
Most athletes stretch as part of their training regimen and it is commonly believed that this practice prevents muscle injury. We tested this belief using an animal model, in situ mouse extensor digitorum longus (EDL) muscle. One lower hindlimb was slowly stretched for 1 min on alternate days for 12 days; the other leg served as a control. The mouse was lightly anaesthetized during the stretching protocol (isofluorane). Both legs were tested in situ by measuring maximum isometric force and maximum work before and after an eccentric contraction that was designed to cause a contraction-induced injury. The difference between a contraction before and after (i.e. the deficit) was used as a measure of damage caused by the eccentric contraction. There was a threshold for force deficit at a peak to peak eccentric excursion amplitude of 19.5 % (i.e. Lo ± 9.75 %, where Lo is muscle length at peak isometric force). There was a significant increase in force deficit, work deficit, and curve shift with an increase in eccentric excursion amplitude above the threshold. There was no statistical difference in the force deficit, work deficit, or curve shift between the stretched leg and the control leg (P > 0.05). A routine stretching programme, at least at the intensities employed in this experiment, did not prevent contraction-induced injury in the in situ mouse EDL muscle.
(Resubmitted 8 April 2002; accepted after revision 8 July 2002; first published online 26 July 2002)
Corresponding author E. D. Stevens: Department of Zoology, University of Guelph, Guelph, Ontario, Canada N1G 2W1. Email: dstevens{at}uoguelph.ca
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Passive stretching is widely accepted in athletic circles as a means of injury prevention. However, recent literature suggests that there is little clinical evidence to support this notion (see Shrier, 1999, for a meta-analysis and critical review). In previous experiments, using both in vitro and in situ mouse extensor digitorum longus (EDL) preparations, we showed that a slow passive stretch applied immediately prior to the experimental trial did not reduce or prevent contraction-induced muscle injury (Black, 2000; Black & Stevens, 2000). The focus of the present study concerns the efficacy of a routine stretching programme in simulating regular flexibility training because some have argued that it is the routine stretching programme that is important in injury prevention.
Pope et al. (1998) showed a definite correlation between lack of flexibility and an increase in the risk of lower limb injury and, as pointed out in the review by Corbin & Noble (1980), 'there is general agreement among researchers that flexibility can be improved with regular flexibility training'. There does not seem to be agreement on what constitutes regular flexibility training, but one of the most cited reviews on this issue states 'when planning a stretching program it is important to remember that increasing flexibility is a gradual process. It will take several weeks before benefits occur. Athletes who do not begin to stretch until the beginning of their season will receive minimal benefits, if any, that season' (Beaulieu, 1981). It has been shown that repeated passive stretching in humans can alter the viscoelastic properties of skeletal muscle and reduce passive tension (Magnusson et al. 1996; Taylor et al. 1997) and may prevent injury by lengthening soft tissues and increasing joint range of motion (Krivickas & Feinberg, 1996; McNair & Stanley, 1996; Magnusson et al. 1996).
Although it has been demonstrated that the nature of muscle damage is similar in both human and animal models (Lieber et al. 1991), there is a paucity of literature which quantitatively assesses the relationship between stretching and injury. To our knowledge there have been no animal studies on the relationship between a routine stretching programme over a period of days and injury prevention. We hypothesized that there may be some accommodation of muscle when it is stretched repeatedly over time (days) and that this accommodation may reduce or prevent contraction-induced muscle injury. To this end, we developed a protocol to simulate a routine stretching programme where EDL muscles of mice were stretched to a constant force every second day for 2 weeks prior to damage testing.
|
METHODS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Experimental animals
Twenty-nine female CD-1 mice, approximately 6 weeks of age were obtained from Charles River Breeding Laboratories and were housed at the Central Animal Facility (CAF) at the University of Guelph. Mice were cared for and used in compliance with Canadian Council on Animal Care (CCAC) guidelines.
In vivo routine muscle stretching programme
The extensor digitorum longus and other foot extensor muscles were stretched by extending a foot of the mouse into plantar flexion. This was accomplished using a specially designed apparatus (Fig. 1). The rationale behind this design was that the rotary platform would provide a constant moment-arm to apply torque to the ankle joint and allow the muscles that move the foot to be stretched to a constant force. The foot platform was rotated by slowly moving the force transducer backwards (away from the mouse) to stretch the lower hindlimb extensor muscles, including the EDL.
 |
View larger version [in this window] [in a new window] |
|
|
Figure 1. Diagram of the apparatus used to stretch the mouse hindlimb during the 12 day routine stretching programme The force transducer (FT) was mounted on a movable stage. Moving the force transducer backward rotated a pulley (P) and its axle causing the foot cradle (FC), situated on a rotary platform (RP) to rotate up and stretch the foot extensors. A spring (SP) between the force transducer and pulley maintained constant force on the rotary platform during the stretch. The mouse was positioned on a platform (ST) between two bolsters (BB) and a velcro strip was used to secure the anaesthetic tube (AT). A movable metal bar (RB, 1.8 mm o.d.), positioned just above the ankle, was used to secure the stretched leg. The force transducer measured the force both to move the foot cradle and to plantarflex the foot. | ||
The mice were anaesthetized briefly for each stretch. Anaesthesia was induced with inhalant isofluorane (5 % at 2.5 l O2 min-1), until the animal was motionless and unresponsive to touch; this took approximately 1 min. Anaesthesia was maintained with isofluorane at 2.5 % for the duration of the stretch protocol, which was approximately 2 min (Fig. 2A).
 |
View larger version [in this window] [in a new window] |
|
|
Figure 2. Protocol during the 2 week routine stretching programme A, timing of events during one stretch during the 12 day in vivo routine stretching programme. A randomly chosen hindlimb of a mouse was stretched using the apparatus described in Fig. 1. The same limb was used throughout the 12 days of stretching. Isofluorane was used to induce anaesthesia: 5 % in O2 initially, 2.5 % for maintenance. B, results from a separate experiment to estimate the force applied to the EDL muscle-tendon unit during the 12 day in vivo stretching protocol. Results are from 3 mice in which the foot extensors were stretched exactly as during the in vivo protocol except that the force was measured at the proximal tendon of the EDL. Passive force was measured during a number of trials at different stretch intensities. The heavy vertical line indicates the mean stretch intensity during the in vivo protocols in the main experiment. These results show that the average force on the EDL muscle-tendon unit was about 46 mN over rest force or about 70 mN total force. Average rest force during the main experiment was 24 mN and developed tetanic force was 279 mN. | ||
The leg to be stretched was randomly selected prior to the experiment, and each mouse received the stretching protocol every other day for 12 days. The anaesthetized mouse was placed in a prone position on the stretching apparatus, with its snout in the anaesthesia tube. The mouse platform had two bolsters to keep the body centred with respect to the foot cradle, and to keep the hindlimbs aligned in the proper plane of motion. The foot to be stretched was placed in the appropriate groove on the rotating foot platform and positioned such that the restraining bar could be lowered into place behind the calcaneus to hold the heel in place. The muscles were stretched to the appropriate force by adjusting the manipulator, and were held in the extended position for 60 s (Fig. 2A). The leg was stretched slowly so as to avoid evoking phasic stretch receptors; torque during the 60 s hold phase of the stretch was 1.35 
10-5 N m. Then the tension was slowly released to the relaxed position and the mouse was allowed to recover from the anaesthetic.
A separate experiment was carried out on three mice to estimate the force applied to the EDL muscle-tendon unit during the 12 day in vivo stretching protocol. The foot extensors were stretched exactly as during the in vivo protocol except that the force was measured at the proximal tendon of the EDL. That is, the foot was slowly stretched, held stretched for 60 s then slowly released (Fig. 2A). Force during passive stretch was measured during a number of trials at different stretch intensities (Fig. 2B). These trials allowed us to relate the torque on the foot to actual force on the EDL muscle-tendon unit. The heavy vertical line in Fig. 2B demarcates the average torque that was used in the 12 day routine stretching programme (1.38 10-5 N m). The average force on the EDL muscle-tendon unit associated with this average torque value was about 46 mN above rest force (24 mN) or about 70 mN total force. These forces during the 12 day routine stretching programme can be compared with average rest force during the main in situ experiment which was 24 mN and average developed tetanic force which was 279 mN.
In situ extensor digitorum longus muscle preparation
Mice were tested 3 days after their last in vivo stretching protocol. The in situ preparation was similar to that described by Brooks & Faulkner (1990). Mice were anaesthetized using sodium pentobarbital (80 mg kg-1 I.P.) and additional anaesthetic was administered as needed. A small incision (3-5 mm long) was made above the ankle to expose the distal EDL tendon. A 5-0 silk suture was secured to the distal EDL tendon, just below the myotendinous junction. The approximate length of the muscle was measured with digital calipers. Muscle length was taken as the distance from the suture tie to the knee joint. The accuracy of this measurement was within ± 3 % of the actual muscle length. The distal EDL tendon was severed just proximal to the superior transverse ligament. Care was taken to avoid cutting blood vessels. The severed distal tendon was folded back on itself and secured with the same 5-0 silk suture.
The mouse was placed on its side on a heated (37 °C) Plexiglass platform, then its uppermost hindlimb was straightened and placed on a slightly raised platform. The knee was extended and the leg was secured to the platform by pinning the knee at the distal end of the femur and taping the foot with surgical tape. The distal EDL tendon was tied with 5-0 silk suture to the lever arm of a servomotor (Cambridge Technologies, Aurora, ON, Canada). The servomotor was used to effect length changes and measure muscle force during the experiment. Two stainless steel pin electrodes were placed on either side of the peroneal nerve using established anatomical landmarks. Stimulation and muscle length change were controlled by a computer, and position and force information were recorded on a Nicolet Pro Model 420 Digital storage oscilloscope. The incision site was moistened with warmed physiological saline (37 °C) throughout the experiment.
Contraction protocol
Prior to each experiment, stimulus voltage and muscle length were adjusted for maximal isometric tetanic force (Po). With voltage set at 1.2 maximal level, length was adjusted in 0.2 mm increments to produce a preliminary length-force curve. Lo was defined as the length at which Po (the maximum isometric force value) was generated.
Stimulus train duration was 51 ms and pulse duration was 0.4 ms. The stimulus train consisted of 12 pulses at 180 Hz with a 'doublet' inserted at the start of the train to increase the rate of force development (Stevens, 1996b). The stimulus train duty cycle was 100 % 51/250 = 20.4 %. Each contraction protocol lasted 1.25 s and consisted of five consecutive cycles at 4 Hz in the following order: single passive cycle (length change with no stimulation), two active cycles (stimulation during length change, i.e. working contractions), a single stimulation with no length change (i.e. isometric contraction) and a final passive cycle (Fig. 3A).
 |
View larger version [in this window] [in a new window] |
|
|
Figure 3. Contraction protocol during the in situ trials A, representative length and force traces in situ to show the five consecutive cycles that comprised the contraction protocol. The length change was produced by a servomotor oscillating at 4 Hz with an excursion amplitude of 10 %; beginning at Lo and oscillating between 95 % and 105 % of Lo. The horizontal dashed line represents muscle length at Lo. The force trace was produced by the EDL muscle. The muscle was stimulated at a phase of 20 % during two length-change cycles to produce two consecutive concentric working contractions (shortening occurred from 25 % to 75 % of the imposed length-change cycle). It was also stimulated during a period of no length change to produce a single isometric contraction. The symbols on the length trace represent the period of stimulation and began at phase of 20 %. B, examples of length-force curves before and after an eccentric contraction protocol with an excursion amplitude of 40 %. Top curves show developed force (peak isometric force minus resting force) and bottom curves show passive or resting force between contractions. There was a 45 s rest period between each measurement. The fitted curves are second order polynomials; the heavy vertical lines indicate the muscle length at peak force. The force deficit in this preparation was 13 % and the curve shift was 2.6 %. Force during the eccentric contraction which caused the force deficit was 1.4 times the isometric force before the eccentric contraction. Each point on the length-force curves represents a measurement taken from the isometric portion of the contraction protocol (cycle 4; panel A). | ||
During working contractions, muscles were subjected to imposed sinusoidal cyclic length changes at a frequency of 4 Hz using the work-loop method (Josephson, 1993). The work-loop method of length change and of activation probably simulates what happens during normal locomotion in vivo more accurately than the ramp stretches that are sometimes used. A working contraction was a contraction during which the muscle length was changed as it developed force, and hence did work (force displacement). Two types of working contraction were used: concentric contractions to measure work (the muscle was stimulated during the shortening portion of the imposed length change) and eccentric contractions to cause damage (the muscle was stimulated during the lengthening portion of the imposed length-change cycle). The term excursion amplitude is used to describe the magnitude of the imposed length change and usually is expressed as a percentage of Lo. The term phase is used to describe the timing of the start of the stimulus train relative to the imposed length change. We defined a phase of 0 % as the start of each imposed length-change cycle at the initial muscle length. The muscle was lengthened from phase = 0 % to phase = 25 %, shortening occurred from phase 25 to 75 %, and from phase 75 to 100 % the muscle was lengthened back to its initial length.
For all concentric working contractions, peak to peak excursion amplitude for the imposed length change was 10 % from a minimum of Lo - 5 % to a maximum of Lo + 5 %. For example, a muscle with Lo = 14 mm was 13.3 mm long at the minimum of the imposed length change and was 14.7 mm long at the maximum of the imposed length change. The muscle was stimulated at a phase of 20 % for concentric contractions. Thus, the stimulus train for concentric contractions began just before the start of shortening, causing the muscle to contract during the shortening part of the imposed length-change cycle. This phase of stimulation was used for concentric contractions because it gives maximum net positive work per cycle and maximum power (Stevens, 1996a).
An eccentric contraction protocol was used to cause contraction-induced injury and consisted of two consecutive eccentric contractions. In this case the muscle was stimulated during the lengthening portion of the imposed length-change cycle. The stimulus train started at a phase of 0 % for eccentric contractions, the centre of the lengthening part of the imposed length-change cycle. This phase was used for eccentric contractions because it causes maximal damage, as measured by force deficit or work deficit in this preparation (Stevens, 1996a). The stimulus train duration (51 ms), stimulus voltage, and stimulus pulse duration were the same for eccentric contractions as for concentric contractions.
Experimental trial
During an experimental trial, each muscle was allowed to rest for 45 s between each contraction protocol. We chose 45 s because we had carried out a preliminary experiment and tested intervals of 30, 45, 60 and 90 s between contraction protocols used to generate a length-force curve. All aspects of this preliminary experiment were identical to those described in the present study. There was no significant difference in mean peak force or mean peak work (ANOVA, n = 13, P > 0.05), suggesting that any of these rest durations from 30 to 90 s was adequate for metabolic recovery. We chose 45 s rest to ensure adequate time for metabolic recovery and adequate time to measure and record the data.
An experimental trial consisted of construction of an initial length-force curve, then an eccentric contraction protocol to cause muscle damage, and finally the construction of a second length-force curve to assess the amount of damage (Fig. 3B). Muscle length was set to Lo - 1.5 mm prior to generating each length-force relationship. In generating the initial length-force curve, as the length approached the approximate Lo, length increments were reduced to 0.1 mm. After Lo was reached, only two additional points (i.e. 0.2 mm beyond Lo) were recorded to prevent over-stretching the preparation. Then Lo was measured using digital calipers. A contraction protocol (as shown in Fig. 3A) was carried out at each length when generating each length-force relationship. Thus, at the maximum length (0.2 mm beyond Lo), the maximum the muscle was stretched was 5 % of Lo + 0.2 mm. For example, a muscle with Lo = 14 mm would be stretched to a maximum of 14.2 mm when generating the length-force relationship and at this initial length it would be 13.5 mm long at the minimum of the imposed length change and 14.9 mm long at the maximum of the imposed length change during the passive cycles and during concentric working contractions. This constitutes a peak stretch of 6 % relative to the initial length (0.9 mm/14 mm) and thus the muscle would not be over-stretched during the contraction protocol when generating the length-force relationship. Brooks et al. (1995) reported that the threshold for damage from a passive stretch in the mouse in situ EDL preparation occurs at an excursion amplitude greater than 13.2 % of Lo.
Each muscle was subjected to one eccentric contraction protocol in an attempt to effect a contraction-induced injury. The peak to peak magnitude of the excursion amplitude during the eccentric contraction protocol varied from 0 to 40 % of Lo, in increments of 4 %. This was a paired experiment, so both the control leg and the stretched leg were exposed to the same magnitude of excursion amplitude during the eccentric contraction protocol. An eccentric excursion amplitude of 40 % means that the muscle experienced an imposed length change from a minimum of Lo - 20 % to a maximum of Lo + 20 %. For example, if the excursion amplitude was 40 %, then a muscle with Lo = 14 mm was 11.2 mm long at the minimum of the imposed length change and 16.8 mm long at the maximum of the imposed length change.
A final length-force curve was generated to measure the shift of the length-force curve resulting from the eccentric contraction and to ensure that we measured maximum force. The order of the treatments (left leg, right leg) and the order of excursion amplitudes were randomized. The mechanical testing was done without prior knowledge of which leg had been subjected to the in vivo preventive routine stretching programme.
Measurements and calculations
Force deficit and work deficit were used to quantify the damage the muscle sustained, after the eccentric contraction protocol. Force deficit and work deficit were calculated as the difference between pre-eccentric contraction and post-eccentric contraction measures of active force and net work per length-change cycle, respectively. They were expressed as a percentage of the pre-treatment value. A positive force deficit means that the eccentric contraction resulted in a decrease in force. A negative force deficit means that force was enhanced by the eccentric contraction. The force level used to calculate the deficit for the post-eccentric contraction was that at the peak of the length-force curve after the eccentric contraction; that is, it was the maximum force that could be achieved after the eccentric contraction.
After the final in situ force measurement, the mouse was killed with an overdose of anaesthetic and its EDL muscle was removed. All of the tendon was removed leaving only muscle which was oven-dried overnight, and weighed to the nearest 0.1 mg.
The values of Lo and Po were used to calculate normalized force (Pnorm, force per cross-sectional area, in mN mm-2):
Pnorm to muscle length = Po Lo d/m,
and
Pnorm to fibre length = Po Lf d/m,
where Po represents maximal tetanic force (mN), Lo is optimal muscle length (mm), Lf is optimal fibre length (mm) (Lf = Lo 0.44; McCully & Faulkner, 1985), d is density of mammalian muscle (1.06 g cm-3; Mendez & Keys, 1960) and m is wet mass of muscle in grams (calculated as dry weight 100/25; Brooks & Faulkner, 1988).
Statistics
For all analyses, an observation was considered significant if P < 0.05. Student's paired t test was used to test for differences between the stretched and control leg for all variables listed in Table 1. Length-force shift was analysed using Student's two-tailed paired t test; the length at peak force before the eccentric contraction was compared with that after the eccentric contraction.
Brooks et al. (1995), using the same preparation that we used, showed that there is a threshold for injury after eccentric contractions. We used a wide range of excursion amplitudes of eccentric contraction (0-40 %) and predicted that our force deficit results also would show a threshold. To ascertain whether our data exhibited the threshold described by Brooks et al. (1995), we performed a breakpoint regression analysis (Yeager & Ultsch, 1989). An F test was used to compare the fit of the linear regression (one straight line) with the breakpoint regression (two straight lines). Our analysis was concordant with that of Brooks et al. (1995) in that it showed that there was a threshold at about 20 % peak to peak eccentric excursion amplitude. Both the linear and breakpoint regressions were performed in GraphPad Prism version 3.02 (GraphPad Software, San Diego, CA, USA).
In order to be sure that the eccentric contractions did cause damage at the larger excursion amplitudes, we did a regression analysis of the force deficit, work deficit and curve shift data against eccentric excursion amplitude using excursion amplitudes above the threshold determined with the above breakpoint analysis (Neter et al. 1996).
To test whether or not stretch was effective in preventing injury (i.e. in reducing the force deficit) we took advantage of the fact that this was a paired experiment; every mouse had one control leg and one leg that had been stretched every other day for 12 days. We calculated the difference in the force deficit between the two legs and used a one-sample t test to test if this difference was equal to zero. We also tested the slope of the regression line of this difference against eccentric excursion amplitude using an F test. We also did these statistical tests on work deficit and curve shift.
As failure to show a significant difference is not the same as successfully showing no difference, we needed to establish the power of our test. We did not have an estimate of population variance to perform an a priori power analysis, and so performed a post-hoc estimation of the minimum detectable difference (Zar, 1996).
|
RESULTS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
There were no obvious changes in routine behaviour associated with the stretching protocol. Changes in routine behaviour refer to the fact that the person doing the stretching trials did not notice any change in the way the mice walked or behaved in their mouse cages even after 12 stretching trials. The point of this observation is that the stretching trial in itself did not appear to cause any injury; it was not possible to identify which leg had been stretched by observing the walking behaviour of the mice.
There were no significant differences between the stretched and control legs during the in situ tests in any of the parameters tested (paired t test, P > 0.05, Table 1). In particular, there were no differences in the passive or viscoelastic properties between the stretched and control legs; there were no differences in rest force before or after the eccentric contraction, or net work during a passive cycle before or after the eccentric contraction. Net work during a passive cycle is negative work and is an estimate of the viscoelastic properties of muscle (Syme, 1990).
We asked three questions regarding the main results.
(1) Was there a threshold for damage? Breakpoint regression provided a significantly better fit to the force deficit data than did linear regression (F2,48 = 3.723, P = 0.031); there was a threshold for damage at an eccentric excursion amplitude of 19.49 % (Fig. 4). The R2 for the breakpoint regression (two lines) was 0.49 whereas the R2 for the simple regression (one line) was 0.41.
 |
View larger version [in this window] [in a new window] |
|
|
Figure 4. Breakpoint regression analysis of the relationships between excursion amplitude of the eccentric contraction and force deficit
| ||
(2) Did the eccentric contractions cause significant damage and was the magnitude of the damage correlated with the magnitude of the eccentric contraction? We tested the data above 20 % excursion amplitude because there was a significant breakpoint at 19.5 % of Lo. A linear regression of the force deficit data for excursions of 20 % and greater had a slope significantly greater than zero (F1,5 = 46.87, P < 0.0001; Fig. 5A). The mean force deficit was 9.02 ± 1.97 at eccentric excursion amplitudes above 20 %. A linear regression of the work deficit data for excursions of 20 % and greater had a slope significantly greater than zero (F1,25 = 10.05, P < 0.004; Fig. 5B). A linear regression of the curve shift data for excursions of 20 % and greater had a slope significantly greater than zero (F1,25 = 19.22, P < 0.0001; Fig. 5C). Thus eccentric contractions greater than 20 % caused a significant force deficit, work deficit and curve shift that increased in magnitude with an increase in amplitude of the eccentric contraction.
 |
View larger version [in this window] [in a new window] |
|
|
Figure 5. Regression analysis to show that eccentric contractions above the breakpoint caused damage and that the magnitude of the damage increased with the magnitude of the eccentric excursion amplitude
| ||
(3) Did the stretching protocol prevent the deficit? There was no significant difference between the force deficit in the control leg and the stretched leg. That is, the force deficit difference (Control - Stretched) was not significantly greater than zero (t = -0.59, P = 0.72). The mean force deficit difference (Control - Stretched) was -0.04 ± 1.92 %. A regression of the force deficit difference against excursion amplitude did not have a slope different from zero using only the data above the threshold (F1,12 = 0.02; P = 0.90) or using all the data (F1,24 = 0.01; P = 0.93; Fig. 6). If the stretch was effective at preventing a force deficit we would predict that the injury in the control leg would be greater than that in the stretched leg and the difference (Control - Stretched) would have increased with excursion amplitude because damage increased significantly with excursion amplitude.
 |
View larger version [in this window] [in a new window] |
|
|
Figure 6 The effect of the stretching protocol was assessed by calculating the difference between the effect on the control leg from that on the stretched leg in each mouse. In each panel the symbols represent the difference (i.e. Control - Stretched) for each mouse. The dashed lines are the regression lines for eccentric excursion amplitudes from 0 to 40 % and are not constrained; the intercepts are: 0.296 for force deficit, -1.960 for work deficit, and -0.169 for curve shift. The continuous lines are the regression lines for data above threshold and were constrained to pass through (20,0) because we showed there was a threshold (see 'Statistics' in Methods). None of the regression lines has a slope statistically different from zero and none of the mean differences is statistically different from zero. That is, there was no difference between the control leg and the stretched leg with respect to force deficit, work deficit, or shift of the length-force curve. | ||
The work deficit difference (Control - Stretched) was not significantly greater than zero (t = -0.62, P = 0.73). The mean work deficit difference (Control - Stretched) was -0.46 ± 4.16 %. A regression of the work deficit difference against excursion amplitude did not have a slope different from zero using only the data above the threshold (F1,12 = 0.19; P = 0.67) or using all the data (F1,24 = 0.00; P = 0.96; Fig. 6). The curve shift difference (Control - Stretched) was not significantly greater than zero (t = -0.41, P = 0.66). The mean curve shift difference (Control - Stretched) was -0.272 ± 0.301 %. A regression of the curve shift difference against excursion amplitude did not have a slope different from zero using only the data above the threshold (F1,12 = 0.44; P = 0.52) or using all the data (F1,24 = 0.05; P = 0.82; Fig. 6).
The minimum detectable differences of our experiments were calculated by the method described by Zar (1996). Given our experimental design, we could detect a difference in force deficit of approximately 7 %. As shown above, the mean force deficit for eccentric excursion amplitudes above the breakpoint was 9 % and thus detectable. The mean force deficits at eccentric excursion amplitudes of 28, 32, 36 and 40 % were all greater than 7 % (11.8 % at 28, 8.3 % at 32, 13.3 % at 36 and 18.5 % at 40 %; or 12.7 % overall at excursions from 28 to 40 %). If the routine stretching programme had provided complete protection against contraction-induced injury, then the regression line for the stretched leg would have had a slope of zero over the complete range of eccentric excursion amplitudes from 0 to 40 %, whereas there would have been no change in the data for the control leg. The experimental data produced a regression line for the stretched leg with slope of 0.90 above the threshold, essentially the same as that for the control leg. If there was partial protection against contraction-induced injury, then the slope of the difference as plotted in Fig. 6 (Control - Stretched) would range from no protection (slope = 0.0) to complete protection (slope ~= 0.9). This analysis showed that given our variance, a slope of 0.58 would be statistically significant at the 95 % confidence level (F1,13 = 4.78, P = 0.049). A slope of 0.49 would be statistically significant at the 90 % confidence level (F1,13 = 3.20, P = 0.099). That is, we would have detected partial protection against contraction-induced injury if the slope of the difference (Control - Stretched) was 0.58 or greater. At a slope of 0.58, the force deficit at eccentric excursion amplitudes from 28 to 40 % in the stretched leg would have been 5.4 % and in the control leg 12.7 %.
In summary, we have shown that there was a threshold or breakpoint, that above this threshold there were significant force deficits that were larger with larger eccentric excursion amplitudes, that there were no statistically significant differences between the control and the stretched leg and that our statistical tests were powerful enough to detect differences between the control and stretched leg if they had been present.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
It is commonly stated in textbooks and reviews that flexibility reduces or prevents injury and that a regular stretching programme increases flexibility. We cite five examples: 'Many clinicians and authors currently advise that stretching before exercise, with or without warm-up, prevents injury' (Shrier, 1999); 'There are indications that training programs employing adequate stretching and warm-up can help to decrease muscle injuries' (Garrett, 1990); 'Stretching regularly will produce an enhanced flexibility...An increase in flexibility through stretching may decrease the incidence of musculotendinous injuries' (Safran et al. 1989); '...Perhaps the most well-known reason for stretching is injury prevention, with some studies reporting a decreased incidence of musculoskeletal injuries...stretch throughout the season and in the off-season to maintain flexibility. Improved flexibility can only be achieved through a long-term stretching program' (Smith, 1994); 'A preseason flexibility program may decrease injuries in college men athletes' (Krivickas & Feinberg, 1996). However, there are very few tests of stretching alone as a means of injury prevention in humans and none that we are aware of in animals.
Stretching could have two types of effect: short-term and long-term. Short-term or acute effects (seconds to minutes) could include events such as a change in tendon length (i.e. a change in its stress-strain properties), a change in the series elastic component (SEC) or parallel elastic component (PEC) elements, or a change in the distribution of sarcomere lengths within the muscle fibres. Long-term or chronic effects (days to weeks) could include any of the above plus a change in the composition (i.e. isoforms) of either the tendon or SEC/PEC of the muscle that could change the viscoelastic properties of the muscle-tendon unit. A number of studies have shown both in animal models and in humans that with repeated stretches there is a change in the viscoelastic properties of muscles. These studies show that with consecutive stretches there is a decrease in stress at any strain for the first three or four stretches (Taylor et al. 1990; Magnusson et al. 1996). However, these effects are transient and do not last longer than 1 h and the long-term effects of stretch on the muscle-tendon unit are unknown. With regard to long-term effects, 'the increased ROM (range of motion) achieved from static stretching is a consequence of increased stretch tolerance on the part of the subject rather than a change in the mechanical or viscoelastic properties of the muscle' (Magnusson et al. 1998). In spite of this evidence to the contrary, as pointed out above, there is a popular belief that stretching increases flexibility and that the increased flexibility does prevent muscle injury. We have previously performed experiments to test if acute stretching immediately prior to exercise could prevent injury and found that this acute type of stretching regime did not prevent muscle injury in mouse EDL muscle either in vitro or in situ (Black, 2000; Black & Stevens, 2001). The present study was designed to test if routine stretching over a longer period as opposed to acute stretching would lead to a protection against contraction-induced injury in mouse EDL. Our 12 day routine stretching programme was not effective in preventing eccentric contraction-induced injury in this muscle preparation.
Physical and mechanical parameters
The 12 day stretching protocol had no measurable effect on the EDL for any mechanical parameters including both force and work deficit, rest force, and passive work. If injury prevention is related to flexibility and if flexibility is altered by a routine stretching programme then we would expect to see changes in the viscoelastic properties of our preparation as a result of the routine stretching programme. The routine stretching programme did not cause a change in net work during a passive work cycle. Net work during a passive cycle is negative work and is a measure of the viscoelastic properties of muscle (Syme, 1990). This is concordant with the observations of Magnusson et al. (1998) showing no change in viscoelastic properties mentioned above and is contrary to the common belief that stretching protects against injury by reducing passive muscle and/or tendon properties (Safran et al. 1989; Smith, 1994; Rosenbaum & Hennig, 1995; Krivickas & Feinberg, 1996; McNair & Stanley, 1996).
Force deficit and damage
The 12 day routine stretching programme did not prevent the force deficit resulting from eccentric contractions. Given our experimental design and our observed variance, we calculated that our experiment could detect complete protection against injury with very high confidence (99.7 %), a difference in force deficit of approximately 7 % with 95 % confidence, and 50 % protection against injury with only 90 % confidence. Our statistical power was not sufficient to reject the alternative hypothesis that the 2 week routine stretching programme halved the extent of injury. Force deficit is associated with structural damage to a muscle's contractile components (McCully & Faulkner, 1985; Brooks et al. 1995; Gibala et al. 1995; Talbot & Morgan, 1996) and therefore our results suggest that our stretching protocol did not prevent injury to the mouse EDL muscle.
In addition, the clinical evidence (Macera et al. 1989; Walter et al. 1989; Brunet et al. 1990; van Mechelen et al. 1993; Pope et al. 1998, 2000) does not support the hypothesis that a routine stretching programme will prevent contraction-induced injury (Shrier, 1999).
We are convinced that our eccentric contraction protocol caused injury and that the force deficit was not due to fatigue. In separate trials we have shown that the force deficit persists for at least 2 h after the contraction-induced injury and thus is not due to fatigue (Black & Stevens, 2001). In addition we have measured serum creatine kinase and examined the histology of the injured muscle 2 days after the injury. Creatine kinase was elevated, and injured muscles (i.e. those with large force deficits) showed regions of extensive macrophage invasion, whereas control muscles showed none.
The breakpoint regression analysis for force deficit revealed a breakpoint at a peak to peak excursion amplitude of about 19 %. This suggests that damage may result from eccentric contractions of greater than 9.5 % strain (i.e. 19 % excursion amplitude). Brooks et al. (1995) presented both histological and force deficit evidence to show that there is a threshold for injury after a single eccentric contraction and that this threshold lies between 8.8 % and 13.2 % excursion beyond Lo. Our observation of a threshold at 9.5 % excursion beyond Lo is concordant with the observations of Brooks et al. (1995). It appears there is a threshold at which eccentric contractions will result in damage and that this threshold corresponds to an excursion of about 10 % Lo in the mouse in situ EDL preparation. Similarly, there is a threshold for injury in rabbit EDL and tibialis anterior after a single eccentric contraction (Hasselman et al. 1995).
Shift of the length-force relationship
The eccentric contraction resulted in a right shift in the length-force relationship in both the stretched and control legs. This kind of shift could be due to: a transient increase in muscle length, an increase in tendon or other passive structure length, or a reorganization of sarcomere lengths within the muscle or sarcomere 'popping' (Morgan, 1990). There were no changes in passive properties associated with stretching treatment during the experiment (Table 1) which argues against a change in passive muscle structures. Rearrangements of sarcomere lengths, due to sarcomere 'popping', is thought to occur after an eccentric contraction which may help to explain the right shift in the length-force relationship (Wood et al. 1993; Jones et al. 1997; Whitehead et al. 1998).
Relevance to sports injuries
Muscle injuries are classified based on their clinical presentation (Rachun et al. 1976; Safran et al. 1989). Type I injury, experienced by almost every person who engages in sport, is characterized by muscle pain 24-48 h after unaccustomed exercise. Type II, also called muscle 'pull', is characterized by acute pain associated with tearing of a few muscle fibres (1st degree) to complete tearing of the muscle and fascia (3rd degree). It is often reported that Type II is the most common injury in competitive athletics. Type II injury invariably results in loss of time from sport and other activity, often impairs performance following a return to competition, and has a high incidence of recurrence (Glick, 1980; Safran et al. 1989; Garrett, 1990; Taylor et al. 1993).
In the present experiments we intentionally kept the eccentric excursion amplitude below a level that would tear the muscle because it was essential that we were able to measure the contractile properties of the muscle after the eccentric contraction; thus the imposed injury was a Type I injury. In our experience, tears (Type II injuries) occur with peak to peak eccentric excursion amplitudes of > 50 % (Lo ± 25 %) in mouse EDL (authors' unpublished observations). In rabbit muscle, EDL tears at Lo + 25 % and tibialis anterior tears at Lo + 38 % (Hasselman et al. 1995).
Type I injury results in an immediate force deficit and shift of the length-force curve related to the structural damage of some parts of some muscle fibres (Armstrong, 1990). It also results in delayed effects: soreness (so called delayed onset muscle soreness or DOMS), swelling, activation of neutrophils and proteolysis of injured fibres (Canon et al. 1991) resulting in the release of muscle-specific proteins (creatine kinase and myoglobin) into plasma, followed by complete regeneration of injured fibres (Bar et al. 1997). This type of injury has been elicited experimentally using eccentric contractions in both humans and animal models in a large number of studies. In vivo it has been induced by downhill running or down-stepping. It is seen commonly after long distance running, skiing, and horseback riding (Faulkner et al. 1993). Type I injuries are important in sport because the force deficit can persist for days to weeks during the recovery phase (Nosaka & Clarkson, 1996). Furthermore, during this recovery phase, performance is impaired and there is an increased risk of acute Type II injury (Garrett, 1990). More importantly, Hasselman et al. (1995) firmly established that there is a continuum of injury caused by eccentric contractions all the way from the DOMS of Type I injury up to the 3rd degree of Type II injury - small strains (in the biophysical sense) cause small strains (in the clinical sense) and large eccentric strains cause 3rd degree strains. Thus our experiments have some bearing on sports injury and the potential protection thought to be provided by a routine static stretching programme.
Conclusions
Based on our results, a 12 day routine stretching programme, at least at the intensities employed in this experiment, was not effective in preventing contraction-induced injury because there were no differences between the stretched and control legs in the average force or work deficits at various excursion amplitudes. Furthermore, it appears that the routine stretching programme had no effect on any mechanical parameter that has been considered to be an important effect of stretching.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
| ARMSTRONG, R. B. (1990). Initial events in exercise-induced muscular injury. Medicine and Science in Sports and Exercise 22, 429-435 | [Medline] |
| BAR, D. P. R., REIJNEVELD, J. C., WOKKE, J. H. J., JACOBS, S. C. J. M. & BOOTSMA, A. L. (1997). Muscle damage induced by exercise: nature, prevention and repair. In Muscle Damage, ed. SALMONS, S., pp. 1-27. Oxford University Press, Oxford | |
| BEAULIEU, J. B. (1981). Developing a stretching program. Physician and Sports Medicine 9, 59-69 | |
| BLACK, J. D. J. (2000). Slow passive stretching does not protect against acute damage associated with contraction-induced injury in mouse EDL muscle. MSc Thesis, University of Guelph, Guelph, Ontario, Canada | |
| BLACK, J. D. J. & STEVENS, E. D. (2001). Passive stretching does not protect against acute contraction-induced injury in mouse EDL muscle. Journal of Muscle Research and Cell Motility 22, 301-310 | [Medline] |
| BROOKS, S. V. & FAULKNER, J. A. (1988). Contractile properties of skeletal muscles from young, adult and aged mice. Journal of Physiology 404, 71-82 | [Abstract] |
| BROOKS, S. V. & FAULKNER, J. A. (1990). Contraction-induced injury: recovery of skeletal muscle in young and old mice. American Journal of Physiology 258, C436-442 | [Medline] |
| BROOKS, S. V., ZERBA, E. & FAULKNER, J. A. (1995). Injury to skeletal muscle after single stretching of passive and maximally activated muscle in mice. Journal of Physiology 488, 459-469 | [Abstract] |
| BRUNET, M. E., COOK, S. D., BRINKER, M. R. & DICKINSON, J. A. (1990). A survey of running injuries in 1505 competitive and recreational runners. Journal of Sports Medicine and Physical Fitness 30, 307-315 | [Medline] |
| CANNON, J. G., MEYDANI, S. N., FIELDING, R. A., FIATARONE, M. A., MEYDANI, M., FARHANGHEHR, M., ORENCOLE, S. F., BLUMBERG, J. B. & EVANS, W. J. (1991). Acute phase response in exercise. American Journal of Physiology 260, R1235-1240 | [Medline] |
| CORBIN, C. B. & NOBLE, L. (1980). Flexibility a major component of physical fitness. Journal of Physical Education and Recreation 51, 23-60 | |
| FAULKNER, J. A., BROOKS, S. V. & OPITECK, J. A. (1993). Injury to skeletal muscle fibers during contractions: conditions of occurrence and prevention. Physical Therapy 73, 911-921 | [Medline] |
| GARRETT, W. E. (1990). Muscle strain injuries: clinical and basic aspects. Medicine and Science in Sports and Exercise 22, 436-443 | [Medline] |
| GIBALA, M. J., MCDOUGALL, J. D., TARNOPOLSKY, M. A., STAUBER, W. T. & ELORRIAGA, A. (1995). Changes in human skeletal muscle ultrastructure and force production after acute resistance exercise. Journal of Applied Physiology 78, 702-708 | [Abstract] |
| GLICK, J. M. (1980). Muscle strains, prevention and treatment. Physician and Sports Medicine 8, 73-77 | |
| HASSELMAN, C. T., BEST, T. M., SEABER, A. V. & GARRETT, W. E. (1995). A threshold and continuum of injury during active stretch of rabbit skeletal muscle. American Journal of Sports Medicine 23, 65-73 | [Abstract] |
| JONES, C., ALLEN, T., TALBOT, J., MORGAN, D. L. & PROSKE, U. (1997). Changes in the mechanical properties of human and amphibian muscle after eccentric exercise. European Journal of Applied Physiology 76, 21-31 | |
| JOSEPHSON, R. K. (1993). Contraction dynamics and power output of skeletal muscle. Annual Review of Physiology 55, 527-546 | [Medline] |
| KRIVICKAS, L. S. & FEINBERG, J. H. (1996). Lower extremity injuries in college athletes: relation between ligamentous laxity and lower extremity muscle tightness. Archives of Physical Medicine and Rehabilitation 77, 1139-1143 | [Medline] |
| LIEBER, R. L., WOODBURN, T. M. & FRIDÉN, J. (1991). Muscle damage induced by eccentric contractions of 25 % strain. Journal of Applied Physiology 70, 2498-2507 | [Abstract] |
| MCCULLY, K. K. & FAULKNER, J. A. (1985). Injury to skeletal muscle fibers of mice following lengthening contractions. Journal of Applied Physiology 59, 119-126 | [Medline] |
| MACERA, C. A., PATE, R. R., POWELL, K. E., JACKSON, K. L., KENDRICK, J. S. & CRAVEN, T. E. (1989). Predicting lower extremity injuries among habitual runners. Archives of Internal Medicine 149, 2565-2568 | [Medline] |
| MCNAIR, P. J. & STANLEY, S. N. (1996). Effect of passive stretching and jogging on the series elastic muscle stiffness and range of motion of the ankle joint. British Journal of Sports Medicine 30, 313-318 | [Abstract] |
| MAGNUSSON, S. P., AAGAARD, P., SIMONSEN, E. B. & BOJSEN-MØLLER, F. (1998). A biomechanical evaluation of cyclic and static stretch in human skeletal muscle. International Journal of Sports Medicine 19, 310-316 | [Medline] |
| MAGNUSSON, S. P., SIMONSEN, E. B., DYHRE-POULSEN, P., AAGAARD, P., MOHR, T. & KJAER, M. (1996). Viscoelastic stress relaxation during static stretch in human skeletal muscle in the absence of EMG activity. Scandinavian Journal of Medicine and Science in Sports 6, 323-328 | [Medline] |
| MENDEZ, J. & KEYS, A. (1960). Density and composition of mammalian muscle. Metabolism 9, 184-188 | |
| MORGAN, D. L. (1990). New insights into the behavior of muscle during active muscle lengthening. Biophysical Journal 57, 209-221 | [Abstract] |
| NETER, J., KUTNER, M. H., NACHTSHEIM, C. J. & WASSERMAN, W. (1996). Applied Linear Models, 3rd edn. Irwin Publishing, Toronto | |
| NOSAKA, K. & CLARKSON, P. M. (1996). Changes in indicators of inflammation after eccentric exercise of the elbow flexor. Medicine and Science in Sports and Exercise 28, 953- 961 | [Medline] |
| POPE, R. P., HERBERT, R. D. & KIRWAN, J. D. (1998). Effect of ankle dorsiflexion range and pre-exercise calf muscle stretching on injury risk in army recruits. Australian Journal of Physiotherapy 44, 165-177 | [Medline] |
| POPE, R. P., HERBERT, R. D., KIRWAN, J. D. & GRAHAM, B. J. (2000). A randomized trial of pre-exercise stretching for prevention of lower-limb injury. Medicine and Science in Sports and Exercise 32, 271-277 | [Medline] |
| RACHUN, A., ALLMAN, F. L., CLARKE, K. S. & BELL, J. A. (1976). Standard Nomenclature of Athletic Injuries, p. 171. American Medical Association, Monroe, WI, USA | |
| ROSENBAUM, D. & HENNIG, E. M. (1995). The influence of stretching and warm-up exercises on Achilles tendon reflex activity. Journal of Sports Medicine 13, 481-490 | |
| SAFRAN, M. R., SEABER, A. V. & GARRETT, W. E. (1989). Warm-up and muscular injury prevention: an update. Sports Medicine 8, 239-249 | [Medline] |
| SHRIER, I. (1999). Stretching before exercise does not reduce the risk of local muscle injury: a critical review of the clinical and basic science literature. Clinical Journal of Sports Medicine 9, 221-227 | |
| SMITH, C. A. (1994). The warm-up procedure: to stretch or not to stretch. A brief review. Journal of Orthopaedic and Sports Physical Therapy 19, 12-17 | [Medline] |
| SOKAL, R. R. & ROHLF, F. J. (1995). Biometry: the Principles and Practice of Statistics in Biological Research. W. H. Freeman, New York | |
| STEVENS, E. D. (1996a). Effect of phase of stimulation on acute damage caused by eccentric contractions in mouse soleus. Journal of Applied Physiology 80, 1958-1962 | [Abstract] |
| STEVENS, E. D. (1996b). The pattern of stimulation influences the amount of oscillatory work done by frog muscle. Journal of Physiology 494, 279-285 | [Abstract] |
| SYME, D. A. (1990). Passive work and viscoelastic properties of isolated rat diaphragm muscle. Journal of Physiology 424, 301-315 | [Abstract] |
| TALBOT, J. A. & MORGAN, D. L. (1996). Quantitative analysis of sarcomere non-uniformities in active muscle following a stretch. Journal of Muscle Research and Cell Motility 17, 261-268 | [Medline] |
| TAYLOR, D. C., BROOKS, D. E. & RYAN, J. B. (1997). Viscoelastic characteristics of muscle: passive stretching versus muscular contractions. Medicine and Science in Sports and Exercise 29, 1619-1624 | [Medline] |
| TAYLOR, D. C., DALTON, J. D., SEABER, A. V. & GARRETT, W. E. JR (1990). Viscoelastic properties of the muscle-tendon units: The biomechanical effects of stretching. American Journal of Sports Medicine 18, 300-309 | [Abstract] |
| TAYLOR, D. C., DALTON, J. D., SEABER, A. V. & GARRETT, W. E. JR (1993). Experimental muscle strain injury. American Journal of Sports Medicine 21, 190-194 | [Abstract] |
| VAN MECHELEN, W., HLOBIL, H., KEMPER, H. C. G., VOORN, W. J. & DE JONGH, H. R. (1993). Prevention of running injuries by warm-up, cool-down, and stretching exercises. American Journal of Sports Medicine 21, 711-719 | [Abstract] |
| WALTER, S. D., HART, L. E., MCINTOSH, J. M. & SUTTON, J. R. (1989). The Ontario cohort study of running-related injuries. Archive of Internal Medicine 149, 2561-2564 | |
| WHITEHEAD, N. P., ALLEN, T. J., MORGAN, D. L. & PROSKE, U. (1998). Damage to human muscle from eccentric exercise after training with concentric exercise. Journal of Physiology 512, 615-620 | [Abstract/Full Text] |
| WOOD, S. A., MORGAN, D. L. & PROSKE, U. (1993). Effects of repeated eccentric contraction on structure and mechanical properties of toad sartorius muscle. Journal of Physiology 265, C792-800 | |
| YEAGER, D. P. & ULTSCH, G. R. (1989). Physiological regulation and conformation: a programme for the determination of critical points. Physiological Zoology 62, 888-907 | |
| ZAR, J. H. (1996). Biostatistical Analysis, p. 135. Prentice-Hall Press, New Jersey |
Acknowledgements
This work was supported by the Natural Science and Research Council of Canada operating grant to E.D.S. We thank Professor Beren Robinson for help with the regression statistics, Dr Dan Hyden for help with the bootstrap statistics and Dr D. Dyson for teaching us how to use the anaesthetic machine. Dr Brian Allen, Professor and Chair of Statistics, provided extensive assistance with the statistical approach. We thank one of the reviewers who forced us to think intensely about these experiments and what they mean.
This article has been cited by other articles:
|
|
 |
|
  N. C. Lockhart and S. V. Brooks Protection from contraction-induced injury provided to skeletal muscles of young and old mice by passive stretch is not due to a decrease in initial mechanical damage. J. Gerontol. A Biol. Sci. Med. Sci., June 1, 2006; 61(6): 527 - 533. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, stretched leg; 
, control leg. Breakpoint regression provided a significantly better fit to the force deficit data than did linear regression (F2,48 = 3.723, P = 0.031); the break occurred at a peak to peak excursion amplitude at 19.49 % (i.e. 9.75 % strain).