Contraction-induced Injury

Contraction-induced injury is defined as morphological damage to small focal groups of sarcomeres as a result of mechanical disruption of the interdigitation of the thick and thin filament arrays or of the Z-lines of single sarcomeres (Faulkner & Brooks, 1995). The injury initiates a cascade of events that produce a more severe secondary injury after two or three days. This involves an inflammatory response, free radical damage, appearance of cytosolic enzymes in the serum, and phagacytosis of elements within the cytosol of damaged sarcomeres (Faulkner & Brooks, 1995) Human beings report pain associated with this phase as delayed onset muscle soreness. Striated muscle is rarely injured when active in isometric, or shortening contractions, but injuries are known to occur when muscle is activated and forcibly stretched (Brooks, Zerba, & Faulkner, 1995) in a so-called 'plyometric' contraction. That stretch can be caused by impact of a body segment with an external surface or object, as well as during certain self-initiated movements such as jumping up onto something. Depending upon the severity of the injury, 1-4 weeks is required for complete recovery of muscle structure and function (Brooks & Faulkner, 1990).

Single-event muscle strain injuries:

Stretch-related injuries in striated muscle can also be caused by a rapid movement associated with a recovery from loss of balance, a slip, trip, throw, catch, landing, or other rapid movement. Over the past fifteen years research in man, as well as in situ and in vitro animal studies, has indicated that stretch-related injuries in striated muscle are caused by a mechanical-mediated event rather than a chemical or metabolic event. Generally, this involves a single rapid stretch to actively contracting muscle (Lieber & Friden, 1993) or a series of repetitive plyometric contractions (Friden, Sjostrom, & Ekblom, 1983) (Newham, Jones, & Edwards, 1983a) (McCulley & Faulkner, 1985). At present, in animal experiments the muscle injury is best characterized by the resulting deficit in maximum force developed by the muscle, rather than by ultrastructural measures ((Newham, McPhail, Mills, & Edwards, 1983b) cited by (Brooks & Faulkner, 1996)). For single stretches the threshold for injury in mice has been estimated as a work input of 150 J/Kg muscle, corresponding to stretches in excess of 20% L₀ (optimal muscle length) strain, say, at a rate of 2Lf/s (muscle fiber lengths per second). Experiments in rodents suggest that one of the factors best predicting the injury is the combination of the muscle strain and average force, or work input to the muscle, and its initial length (Macpherson, Schork, & Faulkner, 1996) (Hunter & Faulkner, 1997). Evidence from single permiabilized rat soleus muscle fibers indicates that sarcomere lengths can be heterogeneous in a contracting muscle and that regions of muscle with the longest sarcomere lengths contained the majority of damaged sarcomeres after injury (Macpherson, Dennis, & Faulkner, 1997), possibly because their length is due to reduced thick and thin filament overlap (Macpherson et al., 1996). Comparative experiments in immature, young adult and old adult mice have shown that when eccentric contractions are used to injure muscles to the same degree, then not only did muscle fibers in old muscle exhibit a greater force deficit (injury) than the younger animals, but that immature and young adult recovered within two weeks, whereas the muscles of old adult mice had not recovered after even two months (Brooks & Faulkner, 1994). To what extent postural muscle activity above a certain threshold can delay such muscle healing and/or prolong symptoms is a question of direct relevance to overuse disorders that might be explored using this type of experimental model.

Muscle Fatigue

Sustained postural muscle activity can create significant musculoskeletal problems. Epidemiological studies in assembly workers, for example, indicate that redesigning the work place in order to reduce average trapezius muscle activity below 2% MVC over an eight hour day can significantly reduce sick leave (for example, (Aaraas, 1987)).

Localized muscle fatigue has been defined as an acute impairment in performance associated with an increase in the perceived effort to produce that force and eventual inability to maintain that force(Chaffin, 1973). Muscle fatigue occurs when physical tasks demand large sustained forces, high power short-term repetitive contractions, and/or low power sustained single or repetitive contractions (Faulkner & Brooks, 1995). A detailed review of the neural and muscular mechanisms, both centrally-and peripherally-mediated, is given by Gandevia et al. (Gandevia, Enoka, McComas, Stuart, & Thomas, 1995). Enoka and Stuart have identified four factors that dictate how and whether fatigue will occur: internal and external factors affecting task dependency, muscle force-fatigue relationship, muscle synergy adaptations to counteract fatigue, and perceived sense of effort (Enoka & Stuart, 1992).

Whereas fatigue in an isometric contraction is caused by a reduction in maximum force, fatigue involving a muscle which is actively shortening or lengthening involves reductions in the force which can be developed as well as reductions in the maximum velocity of shortening (Faulkner & Brooks, 1995). One useful measure of muscle output under such conditions is muscle power (defined as force x velocity). For young human muscles at 35°C, fast muscle fibers can normally develop 225W/kg whereas slow muscle fibers can develop 65W/kg (Faulkner & Brooks, 1995). The ability to sustain power output is a function of fiber composition and the oxidative capacity of the muscle and depends upon the balance between the energy supply and energy output(Brooks & Faulkner, 1991); highly trained individuals can maintain 15% of maximal power for several hours after a warm-up period(Faulkner & Brooks, 1995).

Relatively few investigations have focused on the energy supply to the muscle during low power sustained fatiguing contractions. An important study in human quadriceps (Sjogaard, Kiens, Jorgensen, & Saltin, 1986) demonstrated considerable variability in intramuscular blood pressure during a 1 hour contraction at 5% MVC. During that contraction the blood pressure averaged approximately twice the resting pressure value, indicating an adequate blood supply pressure during the contraction. However, the pressure tracing was characterized by fluctuations perhaps indicating that different parts of the muscle were recruited sequentially to develop and maintain the given force level. After the 1 hour contraction maximum force capability was reduced by 12 % while the perceived exertion increased from 0 to 4.4 on a 0-10 scale. A more recent study (Murthy, Kahan, Hargens, & Rempel, 1997) used non-invasive spectroscopy to measure changes in tissue oxygenation in vivo in the human ECRB muscle. A significant decrease in oxygenation was found after a 1 min. contraction at 10 % or greater MVC effort level. The association between muscle oxygenation levels and fatigue, or between muscle oxygenation levels and subjectively perceived discomfort or pain remains unexplored.

A recent experimental investigation in humans has corroborated earlier animal and human research that muscle fatigue does affect proprioceptive acuity. In this work, the ability to discriminate different arm velocities was significantly and adversely affected at the shoulder, with proprioception in women being significantly more affected by muscle fatigue than men (Pedersen, Lonn, Hellstrom, Djupsjobacka, & Johansson, 1997). Since proprioception is known to be important for motor control, it has been hypothesized that this could be one factor underlying increased coactivation, inefficient muscle use, and an increase in the workload of the affected muscles (Pedersen, 1997), although this has yet to be shown directly.

Finally, it is important to fit the tool to the worker, rather than the worker to the tool. There are many studies showing that ergonomic interventions can reduce loading exposure. One illustrative example shows that merely selecting an appropriate power tool handle size for the size of hand of the worker can significantly reduce grip forces used (Oh & Radwin, 1993). Large-handed individuals using small handles, and small-handed individuals using large handles employed significantly larger grip forces than normal.

Muscle Pain

Increased stiffness, tenderness and muscle pain, particularly in the neck and shoulder regions, are common work-related complaints. These symptoms are often associated with work involving raised arms, repetitive motion tasks, visual control and relatively high levels of mental concentration.

One of the leading theories proposed to explain such disorders is that of Johansson and Sojka (Johansson & Sojka, 1991). Briefly, this theory starts with evidence that muscle pain, inflammation, ischemia, or sustained static muscle contractions are known to lead to the release of KCL, lactic acid, arachidonic acid, bradykinin, serotinin and histamine in the affected muscle. These substances, in turn, are known to excite chemosensitive Group III and IV afferents which, more recently, have been shown to have a potent effect on gamma-muscle spindle systems in cat limb and neck (splenius and trapezius) muscles, and the response of those spindles to stretch (Pedersen, 1997). Although not yet demonstrated directly by experiment, it is postulated that because spindles are known to play a central role in reflex-mediated stiffness, that this increased spindle output could be the genesis of the increased muscle stiffness, tension and pain symptoms reported as myalgia.

The Johannson-Sojka theory identifies specific neural pathways to explain the circlus vitiosus of Travell et al. who theorized that pain could lead to increased muscle tone which in turn lead to more pain (Travell, Rinzler, & Herman, 1942). Thus the Johansson -Sojka theory emphasizes fusimotor hyperactivity rather than tonic alpha motoneurone hyperactivity. The Travell tonic muscle hyperactivity theory received little support from Lund et al (Lund, Donga, Widmer, & Stohler, 1991) who have pointed out that it is rarely observed in the clinic for example in facial pain syndomes, nor was it observed in healthy volunteers in a recent placebo-controlled trial (Stohler, Zhang, & Lund, 1996). The significance of the increased fusimotor activity, on the other hand, is that it would lead to increased muscle stiffness and thence greater alpha motoneurone recruitment whenever that muscle is used at work, particularly whenever it is stretched actively or passively, or vibrated. Evidence of increased muscle activity during work in patients with pain and/or myalgia comes from EMG studies in Norwegian workers performing stereotyped work (Veierstad, Westgaard, & Anderen, 1990). The Johansson-Sojka model also provide a neurophysiological explanation by which the myalgia could spread to contralateral muscles as discussed and detailed by Pedersen(Pedersen, 1997).

The outcome of experimental studies conducted on patients with intermittent/chronic pain of possible muscular origin in order to ascertain underlying mechanisms are often affected by confounding factors. For this reason, a number of experimental models for studying acute muscle pain in healthy volunteers have been developed in order to better study underlying pain mechanisms. These models include electrical stimulation of the target muscle, intramuscular infusion of chemical irritants into the target muscle, and the use of fatiguing muscle contractions that include the target muscle. Because the pain derived from electrical stimulation, and delayed muscle soreness from fatiguing contractions is variable and difficult to control, one of the more successful muscle pain models has involved the infusion of hypertonic saline into the muscle. One of the advantages of such experiments in man is that it allows for the subjective rating of perceived discomfort and pain, an important dimension that is missing in animal pain experiments.

The hypertonic saline infusion model has been used in sham-controlled studies of healthy young subjects to test the Travell hypothesis that tonic muscle pain of an intensity equivalent to that reported by the majority of pain patients causes an increase in the postural myoelectric activity of the target muscle. In one case the hypothesis was tested in the jaw musculature (Stohler et al., 1996) because of its relevance to temporomandibular pain syndromes; in the other case it was tested in the neck (sternocleidomastoid) muscle (Ashton-Miller, McGlashen, Herzenberg, & Stohler, 1990) because of its involvement in neck and shoulder disorders. The resulting pain develops over a few minutes as a deep aching or burning sensation, often with referred pain, and the magnitude of the pain can be quantified subjectively using a visual analog scale. The pain intensity can be adjusted automatically over many minutes (Zhang, Ashton-Miller, & Stohler, 1993). In both sham-controlled repeated-measures experiments the development of tonic muscle pain was accompanied by reports of referred pain, and a small but significant increase in myoelectric activity in the affected muscle; however, according to rigid scientific criteria this was not greater than the sham pain used as a control in the Stohler experiment (Stohler et al., 1996).

Several points from these studies are worthy of discussion. First, the results provide little support for the tonic muscle hyperactivity (Travell) pain model which assumes a re-inforcing link between pain and muscle hyperactivity. Secondly, it is noteworthy that the jaw muscles are significantly more sensitive to pain than the sternocleidomastoid muscle because 50 times the quantity of hypertonic saline is needed to induce a similar intensity of pain. Whether the mechanism underlying this difference in pain sensitivity is due to a difference in the number of pain muscle receptors, a difference in their central projections, and/or a difference in the gain of those projections is presently unclear. The observation that different muscles have different pain thresholds deserves further research given that overuse symptoms seem to affect some muscles more than others. Thirdly, acute pain at an intensity of 5 on a VAS scale of 0-10 caused autonomic changes such as sweating and blanching in these healthy subjects. Since pain-related autonomic changes such as sweating and blanching are rarely, if ever, observed in patients with work-related overuse syndromes, yet pain scores of 5 on a 10 point VAS scale are commonly reported by them, these patients may have an average level of pain that is an order of magnitude less than that fostered in the above acute pain model. Fifth, the epidemiology of work-related disorders seems to point to an over-representation of females (see accompanying Workshop reviews). Although gender differences have been observed in tendon and ligament (Hart et al., 1998), there could also be a gender difference in the mechanisms underlying human response to pain. Data supporting such a hypothesis suggest that women did indeed perceive significantly greater pain from a standardized noxious cutaneous stimulus than men (Paulson, Minoshima, Morrow, & Casey, 1998).

Aging Effects in Muscle

The effects of age on muscle were extensively reviewed at the NIH Workshop on Sarcopenia (Holloszy, 1995). With age comes a ~20% decrease in human muscle mass, a ~20% decrease in maximum isometric specific force (Faulkner & Brooks, 1995), and a ~35% decrease in the maximal rate of developing force and power (Thelen, Schultz, Alexander, & Ashton-Miller, 1996c). This latter reduction is not due to differences in muscle recruitment strategies (Thelen, Ashton-Miller, Schultz, & Alexander, 1996b), but rather due to a change in the contractility of the muscle itself, and corroborates similar findings in animal experiments. The muscles of old animals similarly demonstrate a loss in sustained power of up to 50% which translates into an equivalent loss in endurance, a fact that is reflected in the loss of athletic performance in humans between the age of 30 and 80 years (Faulkner & Brooks, 1995). In practical terms this translates into a marked decrease in the ability to sustain power over repeated contractions in older individuals. The decline in the ability of muscles to generate force rapidly is one of the most important factors that causes elderly to have difficulty making rapid movements in time-critical situations requiring high strength (Thelen, Wojcik, Schultz, Ashton-Miller, & Alexander, 1997), particularly in elderly women (Wojcyk, Thelen, Schultz, Ashton-Miller, & Alexander, 1998) (Schultz, Ashton-Miller, & Alexander, 1997). Thus, the need to accelerate a tool of a given mass at a given acceleration will require more effort, and hence induce fatigue faster in women than in men, unless hand tools are scaled to accommodate gender differences. The same would be true of older workers vis-à-vis younger workers.

For a given provocation, the magnitude of the initial single stretch injury in rodents has been shown to be greater in muscle from older animals than in younger animals (Brooks & Faulkner, 1996). Older rodent muscle subjected to the same injuring plyometric contraction, also heals significantly more slowly than young muscle (Brooks & Faulkner, 1990), a fact that might have significance when considering age differences in overuse syndromes. These effects help explain why older athletes seem to require greater rest intervals between training sessions in order to avoid overuse problems. And the same effects could partially explain why workers in physically demanding jobs tend to change to less demanding jobs with age.

Effect of Exercise on Muscle

A large literature demonstrates that, at any age, exercise intervention programs are effective for improving muscle strength (Frontera, Meredith, O'Reilly, Knuttgen, & Evans, 1988) (Skelton, Young, Greig, & Malbut, 1995) and endurance (Overend, Cunningham, Paterson, & Smith, 1992) by from ~5 to 30% depending upon the particular muscles. But training has selective effects: while elder men (70-100 years) with life long endurance training had greater power output than age-matched controls, for example, they did not have greater muscle strength than the controls (Harridge, Magnusson, & Saltin, 1997).

Estimation of Muscle and Tendon Forces

The performance of many physical activities requires humans to generate considerable muscle forces as they interact with equipment or their environment. It is these muscle forces, when combined with the external reaction forces applied to the body by its physical environment, that load the remaining soft tissues and skeleton with arrays of forces than can peak with values many times body weight. In any physical activity accurate estimation of muscle forces is the first step to characterizing the forces experienced by each tissue. In addition to posture, motion, vibration, the muscle force exerted during a task is known to be a risk factor for tendon-related disorders (Armstrong, Fine, Goldstein, Lifshitz, & Silverstein, 1987a)

There are two basic techniques for estimating muscle forces. The first, usually used in animal experiments, involves placing a tendon buckle transducer in situ to directly monitor tendon force during a given activity. Recently, this technique has been used in man to monitor finger muscle force (Dennerlein, Diao, Mote, & Rempel, 1998). Importantly, their results show a nearly three-fold variation in the finger tip-tendon tension ratio among nine subjects. This suggests that intrinsic factors can cause a tendon force to be much higher in some individuals than others performing the same task. A limitation of the method was that the forces in cocontracting muscles could not be monitored concurrently because of practical considerations at the time of surgery.

The second, non-invasive, technique involves the use of mathematical models to predict muscle contraction forces. The art of using biomechanical models has evolved over the last thirty years from using simple quasistatic two-dimensional models to three-dimensional models that directly address the muscle indeterminacy problem using optimization techniques in the wrist (An, Kwak, Chao, & Morrey, 1990). Those models have evolved from using optimization techniques and neural networks to predict the set of muscle forces required to establish static equilibrium while minimizing loads on the spine during lifting tasks (Schultz, Haderspek, Warwick, & Portillo, 1983) (Nussbaum, Chaffin, & Martin, 1995), to the use of myoelectrically-driven estimates of muscle activity to improve muscle force predictions during dynamic tasks (McGill & Norman, 1986) (Thelen, Ashton-Miller, & Schultz, 1996a) and muscle co-contraction patterns (Thelen, Schultz, & Ashton-Miller, 1995). Most recently, the effects of muscle stiffness are being included in order to provide for spine stability in the face of external perturbations during physical exertions (Gardner-Morse, Stokes, & Laible, 1995). In the trunk, the complexity of the muscle slip architecture and the multiplicity of muscles has precluded validation of model predictions by direct measurement of muscle forces; rather those models have had to be validated by comparing predicted spine compression forces with measured intradiscal pressures (Schultz, Andersson, Ortengren, Haderspek, & Nachemson, 1982), or by comparing predicted spine moments with measured spine moments (Thelen et al., 1996a). In simpler musculoskeletal systems such as the cat hind-limb, validation of the predictive strategies is possible by directly comparing predicted and measured muscle forces during activities such as locomotion, with errors found to be in the range 25 to 50% (Prilutsky, Herzog, & Allinger, 1997). Such direct comparisons between predicted and measured forces allow refinements to be made to the prediction method (Dennerlein et al., 1998).

However, such errors should also serve as a caveat. Even in these simpler musculoskeletal systems our ability to predict tissue loading is still far from satisfactory. Our accuracy in more complicated systems such as the spine is unlikely to be better, and may well be worse. Given those uncertain tissue loading estimates, and given additional uncertainties in tissue material properties, attempts to predict tissue strain states are currently imprecise. Inaccuracies in these estimates of tissue strain states is one of the major impediments to identifying cause and effect relationships in tissue overuse syndromes. A case in point is the difficulty of separating the effects of tool vibration on muscle and tendon from the effects of vibration on the grip force used to hold the tool(Armstrong, Fine, Radwin, & Silverstein, 1987b)

Ligament

Ligaments are tensile structures which interconnect bones constraining their relative motions. Excellent reviews of ligament structure and function exist (for example, (Frank et al., 1988)). Although they differ somewhat, all ligaments consist predominantly of Type I and Type III collagen, with small proportions of glycosaminoglycans, elastin, and other substances. Ligaments exhibit both elastic and creep in response to tensile loading. The mechanisms of the creep are not well understood but involve progressive fiber recruitment, as the increasing strain progressively reduces the characteristic crimp (or waviness) in the different collagen fibrils, and causes hydration shifts. While many biomechanical studies have characterized the response of certain ligaments to load, much remains to be done to elucidate the effects of age, gender, the female menstrual cycle, and repair processes on ligament function.

Much of the classic work on ligamentous response to changes in activity and injury concerns human and animal knee ligaments because they are so often involved in sports-related injuries. Measurements of human anterior cruciate ligament strains in vivo have provided valuable basic insights and data for improving rehabilitation programs so that exercises do not reinjure surgically-repaired ligaments (Renstrom, Arms, Stanwyck, Johnson, & Pope, 1986). However, the knee ligaments are rarely the site of work-related disorders, so these studies have limited relevance to the problem at hand. In addition, much of the work on ligaments has concerned responses to surgical insult which, although important, may have limited relevance to repetitive loading disorders.

The vascular response of the ligament to injury has been measured in rabbit medial collateral knee ligaments using colored microspheres. This sham-controlled study showed that by six months post-injury, the blood flow was still well over 20 times its baseline value (Bray, Butterwick, Doschak, & Tyberg, 1996). Equivalent studies for the response of ligament to repetitive loading have not yet been conducted. What is known that ligament healing and remodeling times are generally slow. After 10 months the population of collagen fibrils in a small gap made in the adult rabbit MCL were uniformly smaller than controls. At two years follow-up, even though the collagen fibrils showed evidence of slow and steady remodeling, 90% were still less than control values (Frank, McDonald, & Shrive, 1997). Again, it is unlikely that healing of this size gap in an MCL ligament has relevance to ligament overuse injuries because, according to the cumulative trauma model, trauma is more likely on a microscopic scale.

The retinaculum is a specialized type of ligament which acts as a pulley for the finger and toe tendons to prevent 'bowstringing'. Upper extremity dysfunction, for example, such as that involved in De Quervain's tenosynovitis (Moore, 1997), is associated with hypertrophy of the retinaculum covering the wrist, and is thought to be associated with repetitive and forceful thumb use. The time constants involved in the healing and remodeling of retinaculae are unknown.

Ligaments, retinaculae and joint capsules are highly sentient structures providing afferent information for reflexes involved in the control of movement (see review by Barrack and Skinner (Barrack & Skinner, 1990).

Ligament response to alterations in loading

Ligaments adapt to changes in physical loading given time and these adaptations have been studied extensively in several animal models. The most reliable animal studies utilize sham-operated controls rather than the contralateral extremity as a control, because the experimental procedure often causes abnormal loading of the uninvolved contralateral extremity. As summarized elsewhere (Woo, Wang, Netwon, & Lyon, 1990) the rabbit medial collateral ligament (MCL) showed an approximately 50 % decrease in tensile stiffness after nine weeks of immobilization (Woo et al., 1987). Upon remobilization, the rabbit MCL stiffness had returned to normal values following nine weeks of remobilization, although its ultimate tensile strength still exhibited a 20% deficit at that time point. Such experiments underline the important point that soft tissues take longer than is generally appreciated by the lay person to adapt to changes in loading regimen.

Ligamentous response to repetitive loading

Ligament viscoelasticity is evident under repetitive loading. In man, 1 hour of exercise has been observed to increase wrist laxity by approximately 30% which then returned to baseline after 24 hours (Crisco, Chelikani, Brown, & Wolfe, 1997). Frank et al report that in young rats subjected to a 1 month intensive exercise program, the collagen fibrils in the knee ligaments of the exercised rats demonstrated an increased population of medium and smaller collagen fibrils not found in unexercised controls, and this was presumed evidence of a remodeling response(Frank et al., 1988). Since the overall ligament cross-sectional area had not changed, the reduced diameters were used to help explain the acute loss in elastic modulus reported in ligaments after an intensive exercise regime.

Age-Related Changes in Ligaments

Many studies have documented age-related degradation in the mechanical properties of ligaments. In one of the most reliable studies, a significant decrease in elastic stiffness and ultimate load at failure were found when specimens with a mean age of 76 years were compared with those from donors with a mean age of 35 years (Woo, Hollis, Adams, Lyon, & Takai, 1991). The older specimens exhibited more mid-substance failures. One would anticipate that, for the same applied stress history, microstructural fatigue failures would occur at lower numbers of cycles in older ligament than young ligament, but this remains to be demonstrated.

Tendon

Work-related overuse disorders are most common in the tendons and tendon insertions of the upper extremities. The most commonly-affected tendons are in the hand, the wrist, the forearm, and the humeral epicondyles. Conversely, sport-related tendon disorders are more common in the lower extremities, particularly in those who use running as their main form of conditioning.

The structure, function and nutrition of tendons have been reviewed elsewhere (Gelberman et al., 1988) (Jozsa & Kannus, 1997). Tendons are the tensile mechanical structures that allow muscles to transfer their contractile force to bones, and vice versa, with very little elongation. In certain tasks these forces exceeds 100 N in the finger tendons (Armstrong, Foulke, Joseph, & Goldstein, 1982) and thousands of Newtons in the tendons of the large extremity muscles. Depending upon their function and location, tendon excursions can be considerable with body movements; even finger tendons can exceed 5 cm excursion between full flexion and extension.

The tendon proper consists of the epitenon, endotenon, tendon bundles, and blood vessels. The bundles consist of long, parallel, sometimes spiraling, bundles of collagen fibers separated by mature fibroblasts (tenocytes). The extracellular matrix of healthy tendon is water, collagen (primarily Type I) and glycosaminoglycans and glycoproteins. Loose areolar tissue surrounds the bundles at different hierarchical levels and contains blood vessels and nerves. The epitenon is a fine fibrous and cellular layer adherent to the tendon surface (Gelberman et al., 1988). In high friction areas, where they change direction, tendons have fibrocartilagenous regions (Vogel, 1995) to withstand the compressive stresses as well as synovial sheaths which are comprised of circumferentially arranged fibroblasts and collagen fibers—the sheaths are lined with fibroblasts. In the hand, the tendons are covered by the paratenon (Gelberman et al., 1988). To prevent 'bow-stringing' during movements of the wrists or ankles, the sheaths are held in place by a retinaculum.

The nutritional supply to tendons is complex. The perfusion of the finger tendons, for example, is accomplished by vessels running longitudinally within the tendon, entering the tendon at the proximal synovial reflection, vessels entering through long and short vinculae, and vessels entering through the ossesous insertions(Gelberman et al., 1988). One or more avascular regions can be found in healthy tendons, depending upon the specific tendon, and these are presumably nourished by diffusion much like the inner intervertebral disc or meniscus.

Tendons exhibit non-linear viscoelastic behavior (Goldstein, Armstrong, Chaffin, & Matthews, 1987) and this can affect the behavior of the muscle-tendon unit (see below), especially under conditions of repetitive, vibratory or impact loading. Because of their viscoelasticity, tendons essentially act as low-pass mechanical filters between muscle and bone, their filtering characteristic dependent on their structure, composition, size and length.

Tendon disorders can occur at the myotendinous junction (Garret, Dahners, Maynard, & Tidball, 1988), in the tendon proper and at their bony insertion, the osseotendinous junction. The structure and function of the tendon and its response to injury have been reviewed in considerable detail by Gelberman et al. (Gelberman et al., 1988) (Jozsa & Kannus, 1997).

Physicians in sports medicine have suggested tendon disorders fall into four main categories (Leadbetter, 1992), although there is, however, no consensus that this is a completely satisfactory classification system (Jozsa & Kannus, 1997).

I.

paratenonitis (replacing the terms tenosynovitis, tenovaginitis, peritendinitis), an inflammation of the paratenon (whether synovial or non synovial); histological findings include inflammatory cells in paratenon or peritendinous areolar tissue. Clinical signs and symptoms include swelling, pain, crepitation, local tenderness, warmth, dysfunction.

II.

paratenonitis with tendinosis (replacing the term tendinitis), paratenon inflammation together with intratendinous degeneration; histological findings include same as above, with loss of tendon collage, fiber disorientation, scattered vascular ingrowth, but no prominent intratendinous inflammation. Clinical signs include same as above with frequently palpable tendon nodule, swelling, and inflammatory signs.

III.

tendinosis (replacing the term tendinitis), intratendinous degeneration due to atrophy (aging, microtrauma, vascular compromise, etc.); histological findings include inflammatory intratendinous degeneration with fiber disorientation, hypocellularity, scattered vascular ingrowth, occasional necrosis, and/or calcification. Clinical signs include an asymptomatic tendon nodule, which might be point tender. Swelling of tendon sheath is absent.

IV.

tendinitis (replacing the terms tendon strain or tear), symptomatic degeneration of the tendon with vascular disruption and inflammatory repair response; histological findings show three subgroups with variable signs ranging from pure inflammation with acute hemorrhage and tear, to inflammation sure imposed upon pre-existing degeneration, to calcification and tendinosis changes in chronic conditions. In chronic condition there may be: 1) Interstitial microinjury, 2) central tendon necrosis, 3) frank partial rupture, and 4) acute complete rupture. Clinical signs and symptoms are inflammatory and proportional to vascular disruption, hematoma, or atrophyrelated cell necrosis. The duration of symptoms defines three sub-groups: 1. Acute (<2 wks), 2. Subacute (4-6 wks), 3. Chronic (>6 weeks).

Leadbetter's histological findings in adult athletes with overuse tendon injuries requiring surgical treatment of the Achilles tendon, posterior tibial, finger flexor, lateral elbow extensors, medial elbow flexor, patellar and tricpes tendons showed (Jozsa & Kannus, 1997): (1) tenocyte hyperplasia, (2) a blastlike change in morphology from normal tenocyte appearance, (3) prominent small vessel ingrowth with accompanying mesenchymal cells, (4) paravascular collections of histiocytes or macrophage-like cells, (5) endothelial hyperplasia and microvascular thrombosis, (6) collagen fiber disorganization with mixed reparative and degenerative changes, (7) microtears and collagen separations (Leadbetter, 1992). Inflammatory cell populations were prominent in the paratenon and tenosynovium, and reparative cells were found to be present despite the degeneration.

Tenosynovitis in upper extremity tendons most commonly occurs at the (1) extensor pollicis brevis and abductor pollicis longus at the radial styloid, (2) finger flexors at the wrist (beneath the palmar carpal ligament and flexor retinaculum), (3) flexor pollicis longus at the MCP joint, (4) extensor carpi ulnaris at the wrist beneath the extensor retinaculum, (5) flexor carpi radialis at the wrist beneath the flexor carpal ligament and the extension of the flexor retinaculum, (6) proximal long head of the biceps brachii at the intertubercular groove beneath the tranverse ligament (Jozsa & Kannus, 1997). Acute tenosynovitis can take two chronic forms that can co-exist to variable degrees: (1) stenosing tenosynovitis (marked fibrous hyperplasia which tend to hamper gliding of the tendon within its sheath) and exudative-hypertrophic tenosynovitis (synovial hypertrophy and excessive fibrin exudate in the paratendinous space, as well as adhesions between the synovial sheath and the tendon).

Although the most popular imaging modalities using ultrasonography and magnetic resonance imaging for tendon have steadily improved over the last decade to the point that they are used to aid diagnosis and treatment of tendon disorders, their resolution is still inadequate for quantitative in vivo studies of underlying mechanisms.

While tendons are primarily loaded by the uniaxial tensile forces generated by or transferred to muscle, they and their related paratendinous structures are also loaded in the transverse direction by reaction forces where they pass over adjacent hard and soft structures such as bursae, bony pulleys and retinaculae, especially in awkward postures or at the end of range of motion(Armstrong, Castelli, Evans, & Dias-Perez, 1984). The combination of those tensile and transverse forces result in shear forces which can affect gliding of the tendon in its sheath or over its pulley. The resulting frictional forces between the tendon and the paratendinous structures have been implicated in tendinitis. Wilson and Goodship, for example, have quantified the temperature at the tendon core as being 5.4 °C above the tendon surface temperature in the superficial digital flexor tendons of exercising thoroughbred races horses (Wilson & Goodship, 1994). Since this temperature is sufficient to cause fibroblast death in vitro, they suggest this is a possible contributor to tendon core degeneration of the horse digital tendon and possibly human Achilles tendon.

Friction between the tendon and its pulley systems is a possible cause of surface degeneration in tendons. Such friction is usually countered by synovial cells and related structures whose response to mechanical stimulation has been reviewed by Schumacher (Schumacher, 1995). Elegant in vitro biomechanical studies by Uchiyama et al have shown frictional force to be proportional to the magnitude of axial tension in the tendon (determined in the hand by the force used to grip or use a hand tool), the coefficient of friction between the tendon and its substrate (which can vary several fold between different tendons (PL vs. FDL), and the change in the angle of the tendon as it passes around a tendon pulley (Uchiyama, Coert, Berglund, Amadio, & An, 1995b). Awkward hand and wrist postures can increase this pulley angle thereby increasing friction forces on the tendon. At the highest loads, extrasynovial hand tendon had approximately four times as much gliding resistance of an intrasynovial hand tendon (Uchiyama, Amadio, Coert, Berglund, & An, 1995a). Since tendon friction forces will probably be highest during plyometric contractions, it might be instructive to examine whether tasks involving plyometric contractions have higher rates of tendon dysfunction. The transverse reaction force from the pulley will tend to flatten and deform the tendon cross-sectional profile. The in vivo response of tendon and its bearing surfaces to fluctuations in this form of transverse loading post-injury is presently unknown. Tendons are also loaded by intermittent hydrostatic pressure in the carpal tunnel (Rempel, Keit, Smutz, & Hargens, 1997). The tissue response of tendon to these types of non-axial loads is only beginning to attract attention.

Cumulative Strain

One theory for tendon involvement in repetitive work-related disorders states that tendon strain can accumulate over the day's repeated loading. In a classic investigation of the cumulative stress-strain behavior of tendon under cyclic loading, an in vitro study of 25 human flexor digitorum profundus tendons from four females and three males (5572 years) was conducted using step and cyclic loadings (Goldstein et al., 1987). Under the physiologic loads applied, elastic strains ranging from 0.2 to 1.8% were recorded using clip transducers. The stresses at the tendon-sheath and retinacular interfaces were significant for wrist positions deviated from neutral. Significantly lower (~50%) strains were found in the female tendons loaded at the same stress level. Why the female tendons were stiffer is unclear. Under repetitive loading with an 8 s loading phase and 2 sec recovery time, a significant (~40%) increase in strain occurred between the 100 and 400^th loading cycle, demonstrating viscoelastic creep behavior. When the duty cycle was reversed (2 sec loading, 8 sec recovery) the strain accumulation was significantly less.

For this theory to gain acceptance, experiments are needed to provide direct in vivo evidence of a causal link between excessive cumulative strain and the development of histological changes consistent with tendinitis in that region of the tendon. At present, there are no figures defining what that threshold for excessive strain is in different types of tendon.

Animal models of tendon response to exercise

Many studies of tendon response to exercise (Gillis, Meagher, & Poole, 1993) (Michna & Hartmann, 1989) (Patterson-Kane, Wilson, Firth, Parry, & Goodship, 1997), though informative, involved immature animals and thus have limited relevance to work-related disorders in the adult population. In studies reviewed by Archambault (Archambault, Wiley, & Bray, 1995), young mice exposed to an intense 10 week exercise protocol initially displayed many new collagen fibrils in their flexor digitorum tendons. However, by the end of the protocol, the morphological parameters were not much different than controls, except that the exercised tendons had the largest fibrils, a broad distribution of fiber diameters and the closest packing density. In young thoroughbred horses subjected to a life-long (18 month) training program, smaller fibrils were again found in the superficial digital flexor tendon (which is particularly prone to training injury). These were taken as evidence of repeated microtrauma in a region of the tendon already weakened by the training regimen(Patterson-Kane et al., 1997). The same authors point out that approximately 30% of thoroughbreds are lost to racing because of developing tendinitis in this tendon, principally because of overuse due to excessive gallop training. In one study of pig digital extensor tendons exercised for 1 year, again in relatively young animals, Woo et al. found increases in biomechanical properties such as elastic stiffness (20%), tenson ultimate strength, ultimate strain, cross-sectional area and collagen content (Woo, Ritter, & Amiel, 1980), but a corresponding study of digital flexor studies failed to find similar increases—rather there was a slight decrease in tendon material properties, or collagen content (Woo, Gomez, & Amiel, 1981).

These and related studies indicate that tendon shows a modest, slow and variable adaptational response to increased levels of activity, suggesting that it is easy to increase loading demands more rapidly than the tendon remodeling process can keep up with. On the other hand, much larger changes are caused by immobilization which significantly and adversely affects overall mechanical properties because of alterations in collage turnover, decreased GAG content, decreased water content, and so on.

Animal models for inducing tendinosis.

The first experimental evidence that repetitive loading could cause changes resembling peritendinitis crepitans was provided by Rais using a rabbit hindlimb (Rais, 1961). The duration of the hyperactivity was the best predictor of the severity of the tissue abnormality. A promising model for inducing paratenonitis with tendinosis was developed in the anesthetized caged rabbit (6-9 month; 2.5-4 kg) Achilles tendon(Backman, Boquist, Friden, Lorentzon, & Toolanen, 1990). 13 rabbits were induced to kick forcefully using surface electrical stimulation of the passively lengthened triceps surae. This was achieved in a non-weight-bearing posture using a device which cyclically rotated the ankle through 20° dorsiflexion and 35° plantarflexion. The animals were exercised for 5-6 weeks at a rate of 150 flexion and extensions/min for 2 h, three times a week. The contralateral limb served as a control. Following four weeks, all animals had irregular thickening over the Achilles tendon, and most animals demonstrated palpable nodules 0.5-1 cm above the insertion to the calcaneous. The tendons of the exercised legs demonstrated degenerative changes of varying severity and distribution, most often in the central tendon region. These included alterations in the structure of the tendon, the presence of inflammatory cells, an increased number of capillaries, as well as edema and fibrosis in the paratenon. The paper provides evidence supporting the hypothesis that repetitive loading can cause histological evidence of tendonitis/osis. Limitations of the model include the use of high rates of loading, but over much shorter durations than would be the case for the human work week. In addition, Achilles tendinitis is not part of a common occupational syndrome (except for professional runners). Additional limitations include the fact that no sham-operated controls were used, only the contralateral limb was used as control. Moreover, no quantitative histological change scores were given and no gender comparison were made, even though both genders were used. Sample size was inadequate to test a gender-related hypothesis. Despite limitations, however, these papers provide evidence that repetitive loading can induce histological changes associated with repetitive stress disorders.

Measurements of Carpal Tunnel Pressure

Although large grip and pinch forces have been associated with an increased risk for developing carpal tunnel syndrome (Armstrong & Chaffin, 1979) (Silverstein, Fine, & Armstrong, 1987), the pathomechanics remain uncertain. A recent study has reported on the association between carpal tunnel hydrostatic pressures in vivo and controlled fingertip loading in man (Rempel et al., 1997). That study showed an early increase in carpal tunnel hydrostatic pressure when light finger forces were exerted, followed by smaller pressure increase with additional loading. Though carpal tunnel hydrostatic pressure was proportional to finger tip force, it was essentially independent of wrist posture. The paper thus is not in agreement with earlier studies that showed that carpal tunnel pressure increased with both pronation and wrist flexion(Smith, Sonstegard, & Anderson, 1977). One explanation could be that the Rempel study monitored hydrostatic pressures, whereas earlier studies monitored hydrostatic pressure and contact stress.

Muscle-Tendon Unit

The behavior of the muscle-tendon unit under repeated passive stretching has been examined in situ (Taylor, Dalton, Seaber, & Garrett, 1990). In that acute study of rabbit EDL and TA stretched to 10% past L₀, viscoelastic behavior was observed in that the unit increased in length by some 3.5% after cyclic stretching and the peak force required decreased by some 17% of ten stretches. Most of the reduction, however, occurred in the first four loading cycles. A mild rate effect was found in that peak loads increased by 30% when loading rate was increased by three orders of magnitude. There is no evidence per se that stretching alters the risk for developing repetitive stress disorders. However, if progressive shortening of the muscle -tendon unit were to reduce the range of joint motion to the point that tendons now received continuous loading due to reduced hand or wrist range of motion, then that could alter carpal tunnel loading. Considerable research has focussed on how tendon properties modulate muscle output to the skeleton, and vice versa.

Tendon/ligament

1.

At least two experimental studies provide evidence that repetitive loading of tendon can induce histological changes similar to those found in repetitive strain disorders. There is a need for further studies of how the various characteristics of mechanical loading (i.e., Extrinsic factors, Table 1) interact to cause tendon dysfunction and, in particular, prevent healing and adequate remodeling.

2.

While cumulative microstrain is a plausible hypothesis for tendon and ligament injury in over-use syndromes, direct tests of this hypothesis are needed in vivo.

3.

Results from many controlled animal model studies pertaining to exercise and injury responses of ligament and tendon, while informative as basic mechanism, have limited applicability to considering the problem of overuse disorders in industry. This is because many of the experimental studies employed immature animals. The responses of mature and older animals may be more relevant because of slowed healing and remodeling times associated with aging.

4.

While experimental models that allow study of healing and remodeling responses following surgically-induced defects in knee ligament and Achilles tendon do provide insights into basic healing and remodeling mechanisms, they probably have little relevance for upper extremity work-related disorders because of differences in the structures involved, and the extent and type of injury involved. Better models are needed.

Muscle

5.

Recent studies of muscle clearly demonstrate that plyometric contractions can lead to muscle injury and loss of function. There is a need for studies to determine whether repetitive strain disorders might be initiated by single forceful events which trigger the injury, but might be maintained by repeated loading that prevents adequate healing/remodeling.

6.

The most recent decade of research has produced new insights into the mechanisms underlying myalgia. Direct experimental tests of some of the most salient hypotheses are needed.

7.

There is evidence that women respond to certain types of pain differently than men do. There is a need to test this hypothesis directly with respect to muscle and tendon pain.