After prolonged concentric or isometric contractions of our muscles we become fatigued, but this is short lived. Within 1–2 h we have usually recovered full function. However, after a series of unaccustomed eccentric contractions we not only become fatigued, but our muscles remain weak for days, and they become stiff and sore the day after the exercise. The reason for these additional effects is that eccentric contractions can lead to muscle damage. This review discusses a number of aspects of the damage process and the consequences for muscle properties. A summary of the sequence of events is given in Figure 1.
THE INITIAL EVENT
Although it has been known for many years that eccentric exercise leads to delayed-onset muscle soreness (DOMS), a satisfactory mechanism for the initial event that leads to the soreness has been put forward only relatively recently. For a discussion, see Proske and Morgan (5). The basis for this mechanism is the region of instability on the descending limb of the sarcomere length–tension curve. Briefly, when the myofibrils of a muscle fiber are stretched while contracting, some sarcomeres resist the stretch more than others, perhaps because their myofilament overlap is closer to their optimum value or because the cross-sectional area of the myofibril is slightly greater at that point. As a consequence, weaker sarcomeres take up most of the stretch. If this occurs on the descending limb of the length–tension curve, these sarcomeres get progressively weaker until there is no overlap between the myofilaments. The rising passive tension in elastic elements balances the tension in the remaining unstretched sarcomeres (5).
During a series of eccentric contractions, more and more sarcomeres will become overstretched, beginning with the weakest and including progressively stronger sarcomeres. Each time the muscle relaxes, myofilaments in some overstretched sarcomeres may not reinterdigitate, and the sarcomere becomes disrupted. Overstretched, disrupted sarcomeres lie scattered at random along the length of the myofibril.
Once one or more sarcomeres have become disrupted, the damage may spread longitudinally to adjacent sarcomeres in the myofibril and transversely to adjacent myofibrils. A point will be reached at which the structural distortions produced by the presence of overstretched sarcomeres lead to membrane damage, including membranes of the sarcoplasmic reticulum, transverse tubules, or the sarcolemma. This is accompanied by the uncontrolled movement of Ca2+ into the sarcoplasm, triggering the next stage in the damage process (Fig. 1).
This sequence of events provides a satisfactory explanation for the initial event leading to damage to muscle fibers from eccentric exercise. However, these ideas have not been universally accepted, and a major competing proposal is that the initial event is not mechanical in origin. In the alternative view, the damage is largely the result of excitation–contraction (E-C) uncoupling (8). The supporting evidence for this hypothesis is based on intracellular Ca2+ measurements; one could argue that an increase in Ca2+ is secondary to mechanical changes in the fiber (15). In addition, an E-C coupling mechanism does not readily explain a shift in the length–tension relation of the muscle, a characteristic feature associated with damage from eccentric contractions (5).
An event supplementary to sarcomere disruption is the opening of stretch-activated cation channels as a result of membrane stresses produced by the mechanical changes (Fig. 1). This leads to inward movement of Na+ and Ca2+ into the sarcoplasm (15).
To summarize, the sequence of events would begin with disruption of sarcomeres. Structural distortions triggered by the disruptions lead to membrane damage and interference with E-C coupling. At the same time, the accompanying stresses applied to membranous structures lead to opening of cation channels. All of this produces an increase in sarcoplasmic Ca2+ levels and the Ca2+ triggers proteolysis associated with fiber breakdown and repair.
There have been repeated reports that the damage from eccentric exercise can be reduced by muscle fatigue. Until now, there has been no evidence in support of such views. Experiments in our laboratory (4) have shown that eccentric contractions applied to a muscle whose force output had been reduced by 32% by a previous series of fatiguing concentric contractions produced as much evidence of damage as from the same contractions given to unfatigued muscle. The point emphasizes that fatigue and damage are two quite separate processes, and that neither force levels nor fatigue determines the amount of damage from eccentric contractions (4). It is a common misconception that high stresses during an active stretch are responsible for eccentric damage.
THE FALL IN FORCE
A consequence of a series of eccentric contractions is a fall in force, which is often measured isometrically. In considering the size of the fall, a number of factors must be taken into account. First, if force is measured after a series of eccentric contractions, it is likely that the force decline is caused by both metabolic fatigue and damage. Fatigue effects can be minimized by giving only a small number of eccentric contractions (4) or by waiting for the fatigue effects to subside. Force remains depressed for up to a week after the eccentric contractions, whereas recovery is complete within 1–2 h after concentric contractions (Fig. 2). So, deficits in force measured at 2 h or later after the eccentric contractions are likely to be only caused by the damage.
Second, eccentric contractions lead to a shift of the length–tension relation of the muscle in the direction of longer muscle lengths (see below). If this is not taken into account, and the measurement is made at the original length, it will overestimate the fall in force. It has led to reports of much greater falls in force from eccentric exercise than is really the case. For example, 50 eccentric contractions performed across the optimum length of the medial gastrocnemius muscle of the anesthetized cat led tension to fall by 55% of the control value. This was a measurement made at the original optimum length, determined from the muscle’s length–tension relation. When the length–tension relation was remeasured after the eccentric contraction, optimum length had shifted in the direction of longer muscle lengths by 3.2 mm, representing 16% of the muscle’s working range. Tension measured at the new optimum length was now 39% of control values. That is, not taking the length change into account overestimated the size of the force drop by 41%.
Another effect attributed to eccentric contractions is that force at low stimulation rates is preferentially decreased compared with force at high stimulation rates. This is sometimes referred to as low-frequency fatigue. Such a decrease has been used in the past to argue in support of E-C coupling as a major determinant of eccentric damage. When the change in length–tension relation is taken into account, low-frequency depression is still present, but it is less by approximately 30%.
THE LENGTH–TENSION RELATION
The length–tension relation is a fundamental property of muscle. An important consequence of a series of eccentric contractions is a shift of the optimum length for peak active tension in the direction of longer muscle lengths (Fig. 3). This has now been shown in animals as well as in human subjects (3).
The shift in the length–tension relation is a reliable and useful measure of the amount of damage to the muscle produced by the eccentric contractions. Unlike the force drop, it is independent of levels of fatigue (4). The size of the shift correlates with the amount of damage after the eccentric contractions (3). To explain the shift, it is proposed that after the eccentric contractions, the disrupted sarcomeres lie scattered at random along the myofibril. The presence of these noncontracting sarcomeres in series with still functional sarcomeres increases the series compliance of the fiber. This is a nonlinear increase in compliance (i.e., it increases the fiber rest length), but once the passive fiber is stretched, tension rises more steeply than before, because there are effectively fewer sarcomeres in series being stretched (12). The increased series compliance leads to a shift of the muscle’s optimum length for peak active force in the direction of longer muscle lengths (3). There is no evidence of changes in tendon properties accompanying the increase in compliance in myofibrils.
After a series of eccentric contractions, there is a rise in whole-muscle passive tension. It was originally thought to be the result of changes in the connective tissue matrix, and at one stage it was believed to be associated with muscle swelling. However, more recent studies on humans and animals (12,13) have shown that it is specifically associated with muscle damage and, like the shift in optimum length, and unlike swelling, it is present immediately after the eccentric contractions. This extra passive tension is present in the absence of any recordable electrical activity and, is therefore not nerve mediated (Fig. 4).
When a muscle is subjected to a series of eccentric contractions, the degree of damage is determined by the length at which the contractions are performed (Fig. 3). When the active stretches extend onto the descending limb of the length–tension curve (the region of sarcomere instability (5)), some damage occurs, with the amount increasing with length. In the anesthetized cat, 50 eccentric contractions, given symmetrically about the optimum length, led to both a shift in optimum length and a fall in active tension (12). Accompanying changes in passive tension were measured by recording tension during a slow stretch of the muscle over its full physiological range. After the eccentric contractions, passive tension was higher at most muscle lengths. However, the increases were not uniform, and were greatest in the region of the muscle’s optimum length (Fig. 4). At the optimum length, passive tension had increased by 170% above the preexercise level. A second observation made was that over the period of 1 h after the eccentric contractions, passive tension continued to increase until it was 210% of control values. When the muscle was subjected to several brief, rapid stretches, passive tension collapsed back to near the postexercise value, and then over the next hour redeveloped with a similar time course. The approximate time constant for redevelopment of passive tension was 10 min (12).
Our explanation for the rise in passive tension is that after sarcomere disruption by the eccentric contractions, there is the likelihood of membrane damage, perhaps at the level of the t-tubules or sarcoplasmic reticulum. The consequent uncontrolled release of Ca2+ into the sarcoplasm activates the contractile filaments to develop an injury contracture. Presumably, the contracture will be maintained as long as ATP levels remain high. Sarcomeres within the region of injury shorten, applying stresses to immediately adjacent areas, thereby spreading the contracture. This is a slow process, over time leading to an increase in passive tension.
In other experiments we have shown that the number of regions of disrupted, overstretched sarcomeres, as observed under the electron microscope, become fewer with time after a series of eccentric contractions (3). This observation suggests that the sarcomere disruption process is sometimes reversible, and that some sarcomeres are able to recover normal realignment over time. If so, the series compliance of the muscle should slowly decline after the eccentric contractions, which would contribute to the subsequent increase in passive tension.
Why do rapid stretches reduce the passive tension? We have considered two mechanisms. One possibility is that the stretch may break up regions of contracture in damaged fibers into a larger number of smaller areas, separated by empty sarcolemmal tube. The presence of fiber segments that are devoid of contractile material would reduce the level of passive tension.
A second possibility is that rapid stretches applied to damaged muscle fibers may lower passive tension by producing additional sarcomere disruption within the fiber. The fiber segments in a state of contracture are likely to be stiffer than adjacent, still functional sarcomeres. The stretch would therefore be distributed nonuniformly and would disrupt some of the normal sarcomeres. The resultant increase in series compliance would lower passive tension. An observation consistent with such an interpretation is that after the stretches the muscle had to be stretched further before passive tension began to rise (12).
Why does passive tension redevelop? Conceivably, regions of contracture that are broken up by the stretches may coalesce again, and so reduce the portions of the fiber devoid of contractile material. In addition, series compliance would be expected to fall again with time, and other parts of the damaged fiber may enter a state of contracture.
Does the rise in passive tension have any functional implications? It seems likely that the sensation of stiffness we experience in our muscles after a period of eccentric exercise is caused by the rise in passive tension. A common practice among athletes is to carry out a series of warm-up stretches before a competitive event. Should an individual carry some damage from previous eccentric activity, such stretches are likely to enhance flexibility by lowering passive tension levels.
SORENESS AND SWELLING
A feature of eccentric exercise is that it leaves us stiff and sore the next day. Stiffness has already been discussed. Although there is general agreement about the origin of the soreness and its mechanism, there are some interesting new observations on this subject.
Once the damage process has reached the stage of ruptured membranes and there is a rise in resting intracellular Ca2+, this can trigger proteolysis and facilitate breakdown of the damaged fibers. The accompanying inflammatory process (6) involves invasion of the damaged areas by macrophages and monocytes. Products of the inflammation, histamine, serotonin, substance P, and prostaglandins act to sensitize muscle nociceptors served by Gp III and Gp IV afferent fibers. The inflammation is accompanied by edema, which is presumably responsible for the muscle swelling. The onset of swelling roughly parallels that for soreness, and is present at 24 h after the exercise. Depending on the severity of the exercise, swelling and soreness typically persist for another 3–4 d.
Whereas the conventional view has sensitization of nociceptors as a mechanism for the soreness, DOMS has a number of features that distinguish it from other kinds of muscle soreness. It is not really soreness, but tenderness. There is typically no chronic pain, and pain is experienced only during mechanical stimulation—contracting, stretching, or palpating the muscle. Pain from overt muscle injury or from inflammatory muscle disease is typically chronic.
We have recently made an observation that contributes a new point of view to discussions of the neural basis of DOMS (9). When a mechanical probe is pushed into an unexercised muscle sufficiently firmly, the subject experiences some discomfort. If the probe is vibrated at 80 Hz, this relieves some of the discomfort, an example of the well-known phenomenon, “rubbing it makes it feel better.” When the same thing is performed on an eccentrically exercised muscle with DOMS, the threshold for pain from mechanical stimulation is much lower than before, as expected, but now when the probe is vibrated, pain is exacerbated, not reduced, by the vibration. There has been a switch in the effect of muscle vibration. Sensitized Gp IV afferents are unlikely to be able to respond to high-frequency vibration because of their long refractory period. Whether GpIII afferents can respond at that frequency and, indeed, show some response at up to 120 Hz (9) remains uncertain. It is known, however, that 80 Hz is the optimal frequency for stimulation of the primary endings of muscle spindles. It raises the possibility that DOMS is, in fact, an allodynia, in which changes in processing at the level of the spinal cord allow mechanoreceptors, served by large-diameter afferents, to access the pain pathway.
It is a common experience that after a period of intense exercise, we feel unsteady on our legs and have difficulty in performing finely skilled movements. Such anecdotal observations have led to the suggestion that exercise, particularly eccentric exercise, can disturb proprioception. Consistent with this suggestion, experiments in our laboratory have shown that both the sense of force and the sense of position in elbow flexor muscles of human subjects are disturbed after a period of eccentric exercise (7,10).
It is currently believed that the sense of force may have two components, one involving a peripherally derived signal originating most probably from tendon organs, and a second arising centrally as a sense of effort or of heaviness (2). The sense of position is believed to be provided by afferent signals coming from muscle spindles. Based on these propositions, we considered the possibility that eccentric exercise not only damaged the ordinary muscle fibers, but also disturbed the function of muscle receptors. Any abnormal function might, therefore, explain the disturbance to human proprioception.
In an animal experiment, responses of identified tendon organs in the medial gastrocnemius muscle of the anesthetized cat were studied before and after a series of eccentric contractions. The eccentric contractions produced extensive damage in the muscle, as indicated by the large decline in isometric force and the shift in optimum length. Taking into account that there was also an increase in passive tension after the eccentric contractions, it was found that the population of tendon organs studied was able to faithfully monitor all of these changes, and there was no evidence of any abnormal function.
If eccentric exercise damages the ordinary or extrafusal muscle fibers, it seemed plausible that it might also damage the intrafusal fibers of muscle spindles. In a second series of animal experiments, the effects on muscle spindle responses of a series of eccentric contractions were studied. Again, as for tendon organs, the responses of spindles to stretch and to fusimotor stimulation were normal after the eccentric contractions.
If eccentric exercise does not damage muscle sense organs, what might be responsible for the disturbed proprioception? In experiments on the sense of force, subjects were asked to generate a given level of force in elbow flexors of one arm and match it with their other arm (Fig. 5). Subjects are normally quite good at such a task, achieving matching accuracy to within a few percentage points. Elbow flexors of one arm were then exercised eccentrically, and it was found that in a subsequent matching trial subjects made large errors (Fig. 5). The direction of the errors was consistent with subjects’ matching efforts, not forces. When the exercised arm acted as the reference, achieving the reference level of force required much more effort than before, because the muscle had been fatigued and damaged by the eccentric exercise. It meant that in matching effort, the other unexercised arm generated forces that were far higher that the reference level, yet subjects believed that they were making an accurate match (10).
A similar series of experiments was performed on the sense of position. One forearm was placed at a particular angle by the experimenter. The blindfolded subject had to maintain the unsupported arm at that angle, and then match its perceived position with their other arm. Subjects were able to match arm positions within 1–2° of accuracy. Then one arm was exercised eccentrically, the damage indicators were measured, and the matching experiment was repeated. Here, too, subjects made large and systematic errors. The findings led to the conclusion that subjects use the effort required to hold the arm at a particular angle against the force of gravity as a positional cue. When arm muscles were damaged and fatigued by the eccentric exercise, it meant that maintenance of position by the exercised arm was associated with a much larger effort, leading to matching errors (7).
The general conclusion from these experiments was that with our arms in their normal working space, we rely on the centrally derived sense of effort for information about position and force levels. Proprioception is disturbed after intense exercise because of a disturbance of the effort–force and effort–position relationship.
THE REPEATED BOUT EFFECT
It is a common experience for us to feel stiff and sore the day after having climbed down a mountain. That is because leg muscles undergo eccentric contractions during downhill walking. However, if we walk down the same mountain again 1 wk later, it leaves us much less sore afterwards. The muscles have adapted to the first bout of exercise to provide them with protection against further damage. This is the repeated bout effect, which lasts for 3 wk or longer (1).
It has been proposed that the adaptation process after the damage from the first bout of eccentric exercise involves repair of the damaged fibers and incorporation of additional sarcomeres in series. It is envisaged that the extra sarcomeres are added without changing fiber length, so that sarcomere length is now less for a given fiber length (5). As a consequence, during a stretch across a given portion of the muscle’s working range, the initial sarcomere length will be less, and the stretch will be distributed across a larger number of sarcomeres. The presence of the extra sarcomeres produces a shift of the muscle’s length–tension relation in the direction of longer lengths (1). It is therefore less likely for sarcomeres to be stretched onto the descending limb of their length–tension relation, which is the region of instability and disruption. Although it has been claimed that such a mechanism would be too slow to account for the repeated bout effect, animal experiments on muscle immobilization have shown that additional sarcomeres can be incorporated in fibers within 5 d (14).
The repeated-bout effect can provide the basis for a training program of eccentric exercise designed to provide protection against muscle damage. The aim would be to have a program of increasing intensity that triggered the protective adaptation without leaving the individual temporarily disabled from stiffness and pain.
Consideration of the protection provided by a program of eccentric training has acquired new significance in view of the suggestion that eccentric damage under the conditions of an elite sport can, at times, lead to a muscle strain injury (1). Because the recurrence rates of such injuries are high, rehabilitation from a muscle strain should include a program of mild eccentric exercise to provide protection against reinjury.
When a muscle group is subjected to a regular program of concentric exercise in which the contracting muscle shortens rather than being stretched, there is evidence that the muscle adapts in the opposite direction, that is, it becomes more vulnerable to damage from eccentric exercise (11). The simplest interpretation is that for reasons of economy, and to allow the muscle to generate tension at short lengths, fibers lose some sarcomeres in series. That, in turn, shifts the length–tension relation in the direction of shorter lengths, making the muscle more susceptible to eccentric damage. It is an important consideration for the design of training programs that aim to improve muscle speed and strength. Programs must be arranged so that they include a component of eccentric activity to prevent an increase in vulnerability for eccentric damage. We are left by all of this with the impression that muscle is a plastic tissue, adapting continuously and specifically to the requirements of an activity to achieve peak performance.
This review has focused on the view that a mechanical event is the first step in the damage process after eccentric exercise. The mechanism is based on the instability of the descending limb of the sarcomere length–tension curve. Alternative mechanisms based on E-C coupling dysfunction as the initial event are not able to readily explain the shift in optimum length that accompanies the muscle damage. It is also intuitively less appealing to postulate that, for as yet unexplained reasons, the point of weakness in the muscle is somewhere within the E-C coupling machinery. Finally, a mechanism based on sarcomere disruption is able to explain the repeated-bout effect in terms of the incorporation of additional sarcomeres in muscle fibers. So, eccentric exercise leads to two shifts in the muscle’s length–tension relation. An initial, damage-related shift is followed several days later by an adaptation shift. With the secondary shift comes protection for the muscle against further damage.
It seems at first sight to be a design flaw that eccentric contractions routinely cause damage at the level of sarcomeres. However, such damage may well be the necessary precursor for triggering the adaptation process that follows. Presumably, a similar precursor event exists during adaptation to shortening contractions. All of these changes take place within the muscle itself. They emphasize the plasticity of muscle as it adapts to changing conditions. It is likely that there are also matching alterations in the neural control of the muscle. A specific case in point is the change in central excitability produced by the pain from DOMS (9). An exciting area for future experiments will be to follow the changes within the central nervous system accompanying different kinds of exercise.