Comparison of passive stiffness variables and range of motion in uninvolved and involved ankle joints of patients following ankle fractures.

Adams[1] has noted that patients with ankle fracture undergo three phases of treatment--reduction, immobilization, and rehabilitation--and that physical therapists should be involved in the rehabililation component of treatment. A common residual problem encountered dufing this time is ankle stiffness.[2] Physical therapists know that immobilization produces joint contfucture,[3] and although joint stiffness sometimes refers to a subjective complaint, it can be quantified.[3,4] By measuring resistive torque with a strain gauge and the range of motion (ROM) with a potentiometer, a graph of torque versus ROM can be generated. From this torque-versus-displacement curve, variables describing joint stiffness can be identified. Figure 1 illustrates three such variables relevant to this study: passive torque (in newton-meters) and the slope of this curve, passive elastic stiffness at a specific degree of dorsiflexion DF), and energy loss (hysteresis) (in newton-meter-degrees) during the cyclical motion of DF and plantar flexion (PF). The primary concept behind the measurement of joint stiffness in human subjects is that resistive torque is quantified as a function of joint ROM and the corresponding length of the surrounding soft tissue structures. This is analogous to measuring passive tension as a function of muscle, tendon, or ligamentous length in studies of animals.

[Figure 1: ILLUSTRATION OMITTED] Research using animal models has provided some ifformation regarding the efects of inunobdization on the mechanical behavior of soft tissue. Woo et al[5] Studied the biomechanical and biochemical changes that occurred in rabbit knees immobilized for 9 weeks. In periarticular tissue, the amounts of water and glycosaminoglycans were reduced, with the latter correlating well with resistive torque values (r=.78). Testing under anesthesia and through a variety of joint ROMS showed the resistive torque values to be 6 to 10 times higher in the experimental (immobilized) limbs. When dissected joints were tested, the torque values for the immobilized knees were 2.5 to 4 times higher than for the control knees. Although the hysteresis was larger for the experimental knees, both the hysteresis and resistive torque values dropped dramatically after the first cycle of stretching through flexion and extension. This finding prompted the authors to suggest that intennolecular crosslinking and adhesions between gross structures interfered with joint extensibflity, especiafly during the first cycle of stretching.

The cause of contracture formation is not fully understood. Binkley[3] has reviewed the structure and mechanics of ligaments and tendons as they relate to contracture formation and has noted that immobilized ligamentous tissue reaches its failure point at lower force levels than nonimmobilized tissue. With lower slope values, isolated ligamentous stiffness may not necessarily contribute to the production of joint contractures following immobilization.

Studies on the effect of immobilization on contractile tissue have been reviewed[6,7] and have demonstrated that different positions of immobilization will result in different adaptive changes.[8-11] Tabary et al[8] described a series of studies in which the hind limbs of cats were immobilized in casts for 4 weeks. The animals were separated into two groups: a group in which the animals' hind limbs were not immobilized (controls) and a group in which the hind limbs were immobilized in full PF or full DF. Soleus muscles immobilized in DF (ie, a lengthened position) exhibited longer muscle fiber lengths, a 19% increase in the number of sarcomeres, and a decrease in sarcomere length. The passive length-tension curves of muscles from the experimental limbs were no different than those of muscles from the controls. Muscles immobilizontractile and noncontractile soft tissue structures controlling the ankle joint do exist. Although research on animals may be rigorously controlled, anesthesia was used in all of the studies described in the literature cited.[5,8-10] Physical therapists treat joint stiffness in fully conscious humans. Therefore, the study of joint stiffness in patients would help to clarify the link between practice and findings on animal studies. Anesthetics suppress spinal reflex activity contributing to muscle relaxation.[15] This factor could have affected the reflex contribution to the resistance that develops during passive stretching. Research on the effects of immobilization on the stretch reflex,[16,17] Suggests that increases in the resistance to passive stretching may result from an increase in the sensitivity of reflex mechanisms.

Following cast removal in patients with ankle fractures, physical therapists frequently encounter increased resistance to passive stretching into DF, even though immobilized ligaments may demonstrate less stiffness.[3] Because passive stretching excites the stretch reflex,[18] analyzing passive stiffness in the absence of electromyographic (EMG) activity may help to explain the nature of the mechanical changes resulting from immobilization. The purpose of this study, therefore, was to compare, in the absence of EMG activity, several variables of passive stiffness and the range of passive DF in the ankles of patients referred for physical therapy following cast removal post-ankle fracture. The research hypothesis was that the passive stiffness values would be greater and the passive DF ROM would be less in the casted (fractured) ankles.

Method

Subjects

Following physician approval, a total of 36 patients with ankle fractures were recruited after cast removal from the outpatient orthopedic clinic of a local acute care hospital. To be included in the study, each patient was judged by the attending orthopedic surgeon to be clinically stable, with satisfactory alignment on radiographic examination. Only patients with fractures of the medial and lateral malleoli[19] were included. As severity of ankle fracture could also affect the measured stiffness, due to the extent of bone and soft tissue damage, patients were divided into two groups. The less severe fracture group consisted of patients with open- or closed-reduction, laterul malleolar fractures (n = 19). The more severe fracture[19] group consisted of patients with open-or closed-reduction, bimalleolar and trimalleolar fractures, with or without dislocation (n = 11). The descriptive data for both groups are presented in Table 1. Thirteen subjects had their fractured ankles repaired by open surgery, of which most (n=10) were in the more severe fracture group. The remainder (n= 17) underwent closed reduction, with the majority (n= 16) in the less severe fracture group.

[TABULAR DATA 1 OMITTED]

Patients with fractures of the midshaft or the distal tibia or fibula that did not involve the malleolar portion of the bone were excluded. Although in-man[20] originally reported that movement of the laterul malleolus during ankle DF is barely perceptible, others[21,22] have suggested that movement between the tibia and fibula does occur. More recently,[23] it has been suggested that the controversy regarding distal tibiofibular fixation and its effect on DF ROM has not been clarified because of variation in the functional anatomy of the ankle joint. In patients with surgical fixation of the distal tibia and fibula, those with limited DF ROM may need more movement between the tibia and fibula. Because the controversy regarding the effect of this fixation on DF ROM has not been resolved, patients with fixation of the distal tibia to the fibula (eg, use of a diastasis screw) were excluded from the study. As reflex activity of the triceps surae muscle could affect passive torque values, patients who could not relax their triceps surae musculature (three patients) and those with a history of central nervous system lesions were also excluded. Due to testing limitations of the ankle stiffness analysis system, patients who were not capable of reaching 0 degrees of DF in the absence of triceps surae muscle EMG activity were excluded (three patients).

All fractures had been immobilized by the attending orthopedic surgeon with the ankle in or near the neutral position (ie, 0[degrees] of DF).[24] Although the position of immobilization was not prospectively controlled, confirmation that the ankle had been immobilized in 0 degrees of DF was obtained from the medical record at the time of recruitment.

All subjects were tested within 4 days of cast removal, after giving informed consent and prior to commencing outpatient physical therapy. The mean length of time between cast removal and testing was 1.4 days (SD=0.9) for the less severe fracture group and 1.7 days (SD = 1.4) for the more severe fracture group. At the time of cast removal, the subjects were instructed in active ROM exercises for the ankle and advised to commence partial weight bearing with crutches as tolerated.

Measurement System

The ankle stiffness analysis system used in this study has been described previously.[25] This analysis system was custom-made (University of Western Ontario Mechanical Shop) and relied on a computer-controlled electric stepping motor(*) to control both the angular velocity and displacement of the ankle footplate. A strain gauge([dagger]) consisting of small wire filaments mounted on the axle that rotated the footplate recorded the resistive torque. Angular displacement was measured with a film potentiometer[double dagger] attached to the axis of rotation. After analog-to-digital conversion,([sections]) custom-made software([parallel]) was used to acquire and store the output data from the strain gauge and potentiometer on the laboratory microcomputer.([sharp])

Preliminary static testing of known torques within the physiological range demonstrated excellent precision of the strain gauge (1-20 N.m, intraclass correlation coefficient [ICC, equation 2,1][26]=.989). Repeated testing was performed on the same day with a small group of subjects (n=10). All subjects were using crutches with a partial weight-bearing restriction after cast removal. Many subjects, because they had traveled a considerable length of time to have their cast removed and visit the laboratory for testing, were not asked to return for a second round of testing within 24 hours. Passive stiffness variables, therefore, were quantified on two occasions within a 30-minute period during the same testing session to assess reliability. Reliability values (ICC[2,1])[26] for passive torque at 0 degrees of DF, passive elastic stiffness at 0 degrees of DF, and energy loss were .849,.903, and .909. These findings were consistent with those of our previous day-to-day reliability investigations of the computerized method of measuring passive ankle stiffness.[27,28] All stiffness measurements were obtained in the absence of triceps surae muscle EMG activity. Electromyographic activity of the triceps surae musculature was monitored with a Dantec 15C01 EMG amplifier(**) (upper frequency limit: 1 khz, lower frequency limit: 20 Hz, sensitivity: 20 [mu]V/division).

Subject Evaluation Procedure

Subjects received verbal and written explanations regarding the nature of the testing protocol. After screening for adequate DF ROM, basic descriptive information was obtained from the subjects. Electromyographic surface electrodes([double dagger]) (9 mm) were placed on the skin over the central portion of the soleus muscle after preparation with conductive gel. Each subject's ankle was positioned on the footplate of the ankle stiffness analysis system, with the subject resting comfortably on an adjustable treatment table([sharp]) (Fig. 2).

The two-point kneeling position was used to minimize the possibility that knee position might affect the stiffness and ankle ROM measurements due to the pull of the gastrocnemius muscle. To facilitate the subjects' ability to relax, a large, firm foam cushion was positioned under the subjects' abdomen and the rest of the upper body for support. The hips were positioned in approximately 90 degrees of flexion. The subjects' axis of rotation for ankle DF was defined as a line passing between the distal tips of the malleoli.[20] Therefore, as noted in previous publications,[27,28] the rotatory axis of the footplate was visually aligned with a transverse line halfway between the distal extent of the medial and lateral malleoli. The footplate design allowed for adjustments in both the transverse and frontal planes. We believe that the resistance to passive movement at the subtalar joint and the remaining intertarsal joints could not be evaluated separately from the talocrural joint and therefore would contribute to the passive ankle stiffness measures. The hindfoot was secured in an adjustable posterior heel cup with a strap around the dorsum of the foot over the area of the talar head. A distal strap secured the midtarsal joints and the metatarsals. The starting position for all testing was 10 degrees of PF.

Maximum passive DF ROM (in degrees) was then determined by an independent evaluator who did not know the subjects' group assignment or which ankle had been casted. Three trials of manually controlled passive DF were performed with the ankle secured in the footplate. This procedure was done very slowly to avoid producing reflex activity. Maximum passive DF ROM was defined as the point in DF range just 1 degree before EMG activity of the triceps surae musculature was elicited. This value was programmed into the computer-controlled measurement system for the testing procedure to generate the torque-versus-displacement curve.

Consistency of determining the onset of EMG activity between days was examined in a limited manner with one blinded observer using the method error (ME) statistic.[29] This statistic is calculated using the standard deviation of the difference scores (Sd). Thus, ME=[S.sub.d]/[squreroot of] 2. Using one patient with an ankle fracture as a subject, 10 determinations were made of the point in the range at which EMG activity occurred, on 2 separate days. The ME, expressed as a coefficient of variation (C[V.sub.ME]=[2ME/[X.sub.1]+[X.sub.2]] X 100, where [X.sub.1] and [X.sub.2] were the mean ROM for the 2 days),[29] was 7.33% between days.

From the starting position of 10 degrees of PF, the ankle was cycled passively through the test ROM at a slow (6[degrees]/s) velocity. Previously, it has been shown that no EMG activity occurs during stiffness testing of healthy ankles at velocities of less than 20[degrees]/s.[30] Thus, our slow testing velocity was another means of minimizing the occurrence of reflex activity of the triceps surae musculature. The passive cyclical movement of the ankle joint continued uninterrupted for six cycles of DF and PF.

Analysis

Stiffness variables were calculated by averaging the last five cycles of the test ROM. This calculation allowed for a better signal-to-noise ratio, improving the accuracy of fitting the analog signal for the resistance curve with the fourth-order polynomial equation. The first cycle was eliminated because it has been considered to be an acclimatization cycle[28,30] and differs considerably from the second and subsequent hysteresis curves.[5,31] The resultant torque-versus-displacement curve was computer generated and has been described in our earlier work.[28] From this curve, the following variables describing joint stiffness were obtained: passive torque (in newtonmeters) and passive elastic stiffness (in newton-meters per degree) into DF at 5-degree intervals of DF, and energy loss (in newton-meter-degrees) during the DF-PF test cycles (Fig. 1).

Differences in maximum passive DF ROM and the stiffness variables between the casted and noncasted ankles were evaluated separately according to grouping factors of limb involvement and fracture seventy with the analysis of variance design (df=28, n=30). The Newman-Keuls statistic was used for post hoc comparisons.[32] Because changes in energy loss may in part be explained by differences in DF ROM, this variable was also evaluated after normalizing for maximum passive DF ROM.

Results

Figure 3 illustrates passive torque into DF at 5-degree intervals for the casted and noncasted ankles. The value in parentheses above each ankle position indicates the number of subjects who were capable of achieving that angular displacement in the absence of triceps surae muscle EMG activity. Only 7 subjects were capable of reaching beyond 5 degrees of passive DF without evoking any triceps surae muscle EMG activity. We believe that such patients often show increased resistance to DF stretching at the position of immobilization. Because of this and because the sample size was larger (n=30) at 0 degrees of DF, statistical calculations regarding passive torque and passive elastic stiffness were performed with the values that were obtained at this position.

[Figure 3: ILLUSTRATION OMITTED]

Table 2 presents the means and standard deviations for passive torque and passive elastic stiffness into DF at 0 degrees of DF. With passive torque, there was no difference between the casted and noncasted ankles (F = 1.495, P>.05) and no interaction with fracture severity (F = 2.400, P > .05). With passive elastic stiffness, a difference was found between the casted and noncasted ankles (F = 6.245, P < .05), as well as an interaction with fracture severity (F = 6.532, P < .05). Post hoc testing revealed that the difference existed only between the more severe ankle fractures and their matched noncasted counterpart (P > .05). There was a difference in energy loss (Tab. 3) between the casted and noncasted ankles (F = 46.223, P < .001) and no interaction with fracture severity (F = 3.525, P > .05). Even after normalizing for DF ROM, energy loss per degree of DF for the casted ankles (more severe: X = 0.51 N.m, SD = 0.21 N.m; less severe: X = 0.45 N.m, SD = 0.18 N.m) was different than for their noncasted counterparts (more severe: X = 0.65 N.m, SD=0.31 N.m; less severe: X = 0.83 N.m, SD = 0.30 N.m; F = 16.024, P < .001). There was no interaction with fracture severity (F = 3.412, P > .05). This finding suggests that the difference between the ankles was not solely a function of the available ROM.

Table 3 illustrates that, consistent with the common clinical finding that patients with fractured ankles have a limitation in DF ROM, there was a difference in maximum passive DF ROM between the casted ankles and their matched noncasted counterparts (F = 95.553, P < .001). There was no interaction with fracture severity (F = 0.365, P > .05).

[TABULAR DATA OMITTED]

Discussion

Passive Torque Into Dorsiflexion

Although no specific reference was made to the passive tension at the position of immobilization, length-tension curves from animals casted with their muscles in a lengthened position (ie, in full DF)[8,10] have been found to be no different than length-tension curves in control limbs. with patients, the ankle is not casted in full PF or DF, but at 0 degrees of DF. Our passive torque findings at 0 degrees of DF in the casted ankles were no different than in the noncasted ankles. Therefore, at 0 degrees of DF, it appears that the patients' calf muscles were lengthened enough to yield an effect on postimmobilization resistance to passive DF similar to that of animals casted in full DF.

Although subjects were tested in a two-point kneeling position to minimize the effect of length-associated changes in the gastrocnemius muscle on the stiffness variables, Tabary et al[11] have demonstrated that active shortening by electrical stimulation of the muscle can cause increases in hypoextensibility if the muscle activity is produced when the muscle is allowed to shorten. Patients with an ankle fracture are usually non-weight bearing during the initial period of cast immobilization. Therefore, even though the soleus muscle may be relatively lengthened with immobilization in 0 degrees of DF, the knee would likely be held in a more flexed position, causing an active shortening effect via the gastrocnemius muscle. This effect could contribute to increases in passive tension, even though the ankle had been immobilized 0 degrees of DF, which should cause a lengthening effect on postimmobilization stiffness. In this way, two opposing immobilization-induced effects on passive torque may be present at the same time. If there is no increase in passive torque at 0 degrees of DF, these two effects may counteract each other.

Passive Elastic Stiffnew Into Dorsiflexion

The severely fractured ankles had greater passive elastic stiffness values compared with the noncasted ankles. increased passive elastic stiffness indicates that the rate of tension development is increasing as the ankle moves further into DF. Therefore, if the ankle could have been moved further into passive DF in the absence of triceps surae muscle EMG activity, differences in passive torque may have been found. This is similar to findings in animal research where the slope of the passive tension curves were reported to be much steeper[8-10] in the group in which the muscles of the immobilized limb were in a shortened position.

The changes in passive tension during stretch may be due to muscle spindle activity. Jozsa et al,[33] for example, found that rat ankles immobilized with the muscles in a shortened position exhibited capsular thickening and fibrosis in the muscle spindle, with capsular fibrosis and decreased elasticity and fluid content in the Golgi tendon organ. Maier et al[16] found that muscle spindle discharge rates were more sensitive to passive stretching in immobilized muscles. The authors believed this finding to be due to the increased stretching load borne by the spindles. Because the muscle spindles were presumed to lie in a more direct alignment with the atrophied muscle, Maier and colleagues felt that there was more translation of tension and length changes to the spindle when the muscle belly was stretched. Because the atrophy noted in extrafusal fibers was greater than the atrophy noted in the intrafusal fibers, there would be a greater proportion of total muscle tension placed on the intrafusal fibers. In addition, Maier et al believed that increased collagenous tissue might have reduced the elasticity at the polar attachments of the spindle as well as contributing to the increased stretch sensitivity.

We believe that our method was an important factor in our passive elastic stiffness and passive torque findings. The limit of passive DF was defined as the point in the range at which EMG activity began. Increased sensitivity of muscle spindles may explain why the ankles of fewer subjects could be passively dorsiflexed beyond 0 degrees of DF, because the spindles would be stimulated more readily at increasing values of passive elastic stiffness. This explanation is consistent with Bertoft and Westerberg's[17] finding of immobilization-induced changes in the sensitivity of the stretch reflex in a small sample of patients.

Energy Loss

Energy Loss is a measure of the energy that was absorbed by the ankle joint and its surrounding soft tissue structures during the cyclical motion through DF and PF.[30] Hufschmidt and Mauritz,[30] who studied subjects with and without neurological impairment, suggested that two antagonistic effects may play a role in the development of contractures. In one scenario, increased stiffness of the series elastic structures (actin-myosin bridges and Z bands) would increase the stretch on the sarcomeres and therefore increase the plastic resistance. Because the plastic forces resisting stretch contribute to energy loss,[30] this variable would be increased. Alternatively, any increased stiffness of the parallel elastic structures (connective tissue in the muscle belly) would decrease the plastic resistance to stretching by protecting the sarcomeres from being stretched at the cost of the series elastic elements. Immobilization-induced changes found in muscle connective tissue by Tardieu et al[9] are consistent with this second theory.

Although our study did not address the independent roles of these two effects, our observations of small energy loss values in the casted ankles are also consistent with the second theory. Therefore, it seems that a protective mechanism was present in the casted ankles, a mechanism that effectively limited the amount of maximum passive DF that could be obtained in the absence of EMG activity.

Maximum Passive Dorsiflexion Range of Motion

The difference in maximum passive DF ROM between the fractured and contralateral ankles explains why it is common clinically to find that many patients with ankle fructures have limitations in ankle DF. A paradox seems to be present, because this variable was different between the two ankles, but no difference was found in passive torque at 0 degrees of DF--the position of immobilization. We believe that the reflex length-tension relationship was altered in the fractured ankles. This alteration may be due to abnormal length-tension relationships of the triceps surae musculature secondary to immobilization. Because maximum passive DF ROM was defined as the angular displacement obtained just before the onset of triceps surae muscle EMG activity, marked decreases in DF ROM could therefore be expected.

Conclusions

No difference in passive torque was found between the fractured and contralateral ankles at 0 degrees of DF. There was a small difference in passive elastic stiffness at 0 degrees of DF, although this difference was of questionable meaningfulness. A large difference existed between the fractured and uninvolved ankles in both energy loss and maximum passive DF ROM. Because these findings were noted in the absence of triceps surae muscle EMG activity, they support the theory that a reflex-based protective mechanism prevents overstretching,[8,9,30] which in our study was overstretching into passive DF. Physical therapists, therefore, should be aware of the contribution that neuromuscular mechanisms play in the production of "passive" stretching of joint contractures.

The effects of immobilizing the muscles of various species of animals suggest that altered length-tension relationships and neuromuscular mechanisms play a role in the development of joint stiffness. Our results on humans who had their ankles in a cast are consistent with both effects. Further research is needed to determine whether treatment protocols directed toward the contractile soft tissue structures would be an effective strategy to decrease postimmobilization stiffness. Attention should be directed toward the time course of treatment gains in passive ROM, soft tissue mechanics, and normalization of stretch reflexes.

Acknowledgments

We thank the orthopedic surgeons and their team at University Hospital, London, Ontario, Canada, for their assistance. We are also grateful to Antoine Helewa and Donna Tierney for their contribution to this project and to Bob Kager and Hugh Jack for technical support.

(*) M112-FF-401 SLO-SYN Synchronous Stepping Motor, Superior Electric Co, Bristol, CT 06010. ([dagger]) Series EA-13-062TV-350 Strain Gauge, Measurements Group Inc, Raleigh, NC 27611. ([double dagger]) Model 205 Potentiometer, Vernitron Corp, Deer Park, NY 11729. ([sections]) DT2821 A/D Conversion Board, Data Translation Inc, Marlborough, MA 01752. ([parallel]) CY Software, Toronto, Ontario, Canada M8Y 3R3. (#) IBM PC/AT, International Business Machines Corp, Boca Raton, FL 33432. (**) DISA Dantec Elektronic, Skovlunde, Denmark DK-2740. ([double dagger]) Beckman Instruments Canada Inc, 1095 Tristar Dr, Mississauga, Ontario, Canada L5T 1W5. ([double dagger]) Cardon Mobilization Table R28535, Cardon Rehabilitation Products, 3206 Wharton Way, Mississauga, Ontario, Canada L4X 2C1.

References

[1] Adams JC. Outline of Fractures. 6th ed. Edinburgh, Scotland: Churchill Livingstone; 1972:29-44. [2] Salter RB. Textbook of Disorders and Injuries of the Musculoskeletal System. Baltimore, Md: Williams & Wilkins; 1970:507. [3] Binkley J. Overview of ligament and tendon structure and mechanics: implications for clinical practice. Physiotherapy Canada. 1989;41:24-30. [4] Wright V. Stiffness: a review of its measurement and physiological importance. Physiotherapy. 1973;59:107-111. [5] Woo SL-Y, Matthews JV, Akeson WH, et al. Connective tissue response to immobility: correlative study of biomechanical and biochemical measurements of normal and immobilized rabbit knees. Arthritis Rheum. 1975;18:257-264. [6] Gossman MR, Sahrmann SA, Rose SJ. Review of length-associated changes in muscle: experimental evidence and clinical implications. Phys Ther. 1982;62:1799-1808. [7] Herbert R. The passive mechanical properties of muscle and their adaptations to altered patterns of use. Australian Journal of Physiotherapy. 1988;34:141-149. [8] Tabary J-C, Tabary C, Tardieu C, et al. Physiological and structural changes in the cat's soleus muscle due to immobilization at different lengths by plaster casts. J Physiol (Lond). 1972;224:132-144. [9] Tardieu C, Tabary J-C, Tabary C, Tardieu G. Adaptation of connective tissue length to immobilization in the lengthened and shortened positions in cat soleus muscle. J Physiol (Paris). 1982;78:214-220, [10] Williams PE, Goldspink G. Changes in sarcomere length and physiological properties in immobilized muscle. J Anat. 1978;127:459-468. [11] Tabary J-C, Tardieu C, Tardieu G, Tabary C. Experimental rapid sarcomere loss with concomitant hypoextensibility. Muscle Nerve. 1981;4:198-203. [12] Lieber RL. Skeletal Muscle Structure and Function. Baltimore, Md: Williams & Wilkins; 1992:58. [13] Higuchi H, Suzuki T, Kimura S, et al. Localization and elasticity of connectin (titin) filaments in skinned frog muscle fibres subjected to partial depolymerization of thick filaments. J Muscle Res Cell Motil. 1992; 13: 285-294. [14] Lieber RL, Bodine-Fowler SC. Skeletal muscle mechanics: implications for rehabilitation. Phys Ther. 1993;73:844-856. [15] Trevor AJ, Miller RD. General anesthetics. In: Katzung BG, ed. Basic and Clinical Pharmacology. 3rd ed. East Norwalk, Conn: Appleton & Lange; 1987:279-288. [16] Maier A, Eldred E, Edgerton VR. The effects of muscle atrophy and hypertrophy. Exp Neurol. 1972;37:100-123. [17] Bertoft ES, Westerberg CE. Gait disturbance after immobilization of one leg. Acta Orthop Scand. 1987;58:320-321. [18] Rothwell JC. Control of Human Voluntary Movement. Rockville, Md: Aspen Publishers Inc; 1987:126-157. [19] Muller ME. The comprehensive classification of fractures of long bones. In: Muller ME, Allgower M, Schneider R, Willenegger H, eds. Manual of Internal Fixation: Techniques Recommended by the AO-ASIF Group. 3rd ed. Berlin, Federal Republic of Germany: Springer-Verlag; 1991:118-150, [20] Inman VT. The Joints of the Ankle. Baltimore, Md: Williams & Wilkins; 1976:24. [21] Gray H; Warwick R, Williams PL, eds. Gray's Anatomy. 35th ed. Edinburgh, Scotland: Longman Group; 1978:462.

[22] Kapandji IA. The Physiology of the joints, Vol 2: Lower Limb. Edinburgh, Scotland: Churchill Livingstone; 1977:152. [23] Inman VT; Stiehl JB, ed. Inman's Joints of the Ankle. 2nd ed. Baltimore, Md: Williams Wilkins; 1991:18. [24] Hamilton WC. Malleolar fractures and dislocations of the ankle: treatment, In: Hamilton WC, ed. Traumatic Disorders of the Ankle. New York, NY: Springer-Verlag New York Inc; 1984:81-99. [25] Chesworth BM, Vandervoort AA, Koval JJ. A pilot study to compare the subjective and objective evaluation of passive ankle joint stiffness. Physiotherapy Canada. 1991;43:13-18. [26] Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull. 1979;86:420-428. [27] Chesworth BM, Vandervoort AA. Reliability of a torque motor system for measurement of passive ankle joint stiffness in control subjects. Physiotherapy Canada. 1988;40:300-303. [28] Chesworth BM, Vandervoort AA. Age and passive ankle stiffness in healthy women. Phys Ther. 1989;9:217-224. [29] Portney LG, Watkins MP. Foundations of Clinical Research: Applications to Practice. East Norwalk, Conn: Appleton & Lange; 1993:525. [30] Hufschmidt A, Mauritz K-H. Chronic transformation of muscle in spasticity: a peripheral contribution to increased tone. J Neurol Neurosurg Psychiatry. 1985;48:676-685. [31] Tindale P. Force/displacement curves of the knee. In: Dalziel BA, Snowsill JC, eds. Proceedings of the Fifth Biennial Conference, Manipulative Therapists Association Of Australia,. November 25-28, 1987; Melbourne, Victoria, Australia. 1987:271-296. [32] Winer BJ. Statistical Principles in Experimental Design. 2nd ed. New York, NY: McGraw-Hill Book Co; 1971. [33] Jozsa L, Kvist M, Kannus P, Jarvinen M. The effect of tenotomy and immobilization on muscle spindles and tendon organs of the rat calf muscles: a histochemical and morphometrical study. Acta Neuropathol (Berl). 1988; 76:465-470.

B Chesworth, MCISCPT, is Coordinator, Clinical Studies, Physiotherapy Department, Victoria Hospital, 800 Commissioners Rd E, London, Ontario, Canada N6A 4G5. He was Physical Therapy Clinical Associate, University Hospital/University of Western Ontario Physical Therapy Clinic, University of Western Ontario, London, Ontario, Canada N6G 1H1, during data collection and the preparation of this report. Address all correspondence to Mr Chesworth at Physiotherapy Department, Victoria Hospital, 800 Commissioners Rd E, London, Ontatio, Canada N6A 4G5.

A Vandervoort, Phd, is Associate Professor, Department of Physical Therapy, Faculty of Applied Health Sciences, University of Western Ontario.

This study was supported in part by grants from the Physiotherapy Foundation of Canada and the National Health Research Development Program, Health and Welfare Canada.

This article was submitted May 2, 1994, and was accepted December 8, 1994.