Effects of Immobilization on Plantar-Flexion Torque, Fatigue Resistance, and Functional Ability Following an Ankle Fracture.

Fractures of the ankle mortise remain one of the most common and disabling injuries of the lower extremity. However, optimal treatment for fractures of the ankle malleoli has yet to be determined. Controversy exists in terms of type and duration of immobilization. Recently, there has been a trend toward open reduction-internal fixation (ORIF) and early protected mobilization to minimize the effects of prolonged immobilization on muscle and connective tissue.[1] Yet, in many cases, surgeons still opt for an extended period of immobilization, especially in patients who are at higher risk for reinjury and those with potential medicolegal issues.

One of the most predictable consequences of cast immobilization is loss of lean muscle mass. Studies of animals have shown that cast immobilization can produce a large amount of muscle atrophy, the extent of which is related to the duration of the immobilization period.[2-7] The time course of the loss of muscle mass has been shown to be determined by the half-life of the myofibrillar proteins.[8] As such, muscle atrophies most profoundly in the early phase of immobilization, with initial changes as early as 48 hours.[9,10] During prolonged periods of immobilization, the rate of atrophy progressively decreases. Max et al[3] measured a 30% loss in the gastrocnemius muscle mass in rats after only 3 days of immobilization and a 50% loss at 15 days. Other investigators[2] have reported a 58% loss in muscle mass after 6 weeks of immobilization using the same model.

In studies of humans, a variety of methods have been used to quantify the effect of disuse on muscle morphology. Halkjaer-Kristensen and Ingemann-Hansen,[11] using mathematically corrected girth circumference measures, showed an 11% to 17% loss of lean muscle mass in the thighs of 84 young soccer players after 31 days of immobilization. More recent studies using advanced techniques have shown that not only the amount of atrophy with disuse is more pronounced than would be predicted from anthropometric measures, but it is also muscle specific. Veldhuizen et al,[12] using computerized tomography, found a 21% decrease in the cross-sectional area of the quadriceps femoris muscle after 4 weeks of cast immobilization. Hather et al,[13] using magnetic resonance imaging, found losses of 16% and 7% in the muscle cross-sectional area of the knee extensors and flexors, respectively, and decreases of 16% and 26% in the cross-sectional area of the soleus and gastrocnemius muscles, respectively, after 6 weeks of unilateral lower-limb suspension. We recently demonstrated that the maximal cross-sectional area of the triceps surae muscles is reduced by 20% to 32% in patients with ankle fractures after 8 weeks of cast immobilization.[14] The highest rate of atrophy was measured during the first 2 weeks of immobilization (8.3% per week).

The most evident consequence of immobilization is loss of muscle force. The loss of force, similar to the loss of muscle mass, is a time-dependent process. The decrease in force, however, is not strictly proportional to the loss of muscle mass because neural input[15,16] and metabolic energy stores[17] also play a role in determining the amount of force output. In 1970, Muller[18] demonstrated that upper-extremity force falls precipitously during the first week of immobilization (1%-6% per day). Other authors[12,15,19,20] have reported decreases in force ranging between 40% and 53% during 4 to 6 weeks of cast immobilization.

Of most concern is the fact that the deleterious effects of immobilization do not appear preventable. In a rat model, Widrick and Fitts[21] demonstrated that intermittent weight-bearing and resistive exercises attenuate only 60% of the loss in fiber diameter and muscle force induced by 14 days of non-weight bearing. Stillwell et al[22] studied patients who performed isometric quadriceps femoris muscle contractions while their lower extremity was immobilized in a long leg cast and found no difference in thigh circumference or isometric tension of the quadriceps femoris muscle compared with a cohort of subjects whose lower extremity was immobilized but who did not exercise. Halkjaer-Kristensen and Ingemann-Hansen found no change in the thigh volume[11] or knee extension force[23] of male soccer players who had undergone multiple sessions of voluntary isometric quadriceps femoris muscle contractions or electrical stimulation during 4 to 6 weeks of immobilization versus a similarly injured, nonexercised control group.

Because disuse atrophy does not appear to be preventable, the rehabilitation specialist is left with the daunting task of restoring muscle function after injury. Only a small number of longitudinal studies have documented the recovery of muscle function after either immobilization or unloading (non-weight bearing). Data presented by Berg and colleagues,[24,25] using an unloading model in humans, indicate that the recovery time is dependent on the duration of disuse. The recovery of muscle function following short-term unloading appears to be completed in a shorter time span than the duration of unloading, whereas unloading periods of 4 to 6 weeks result in a recovery period lasting as long as the unloading period or longer. Similar conclusions have been reached following cast immobilization studies in patients. Ingemann-Hansen and Halkjaer-Kristensen[26] studied a large series of soccer players whose injured lower extremity was immobilized for 4 to 6 weeks in a long leg plaster cast after knee ligament injury and reported that the determining factor for the rate of recovery was not the retraining method, but the period of immobilization. Significant recovery of muscle force and endurance (work during 6 minutes of bicycling) was observed after 4 weeks of retraining using a variety of rehabilitation programs, including progressive resistance exercise, one-legged bicycling, isokinetic exercise of 50 [degrees] or 150 [degrees]/s, and maximal voluntary isometric contractions. Subjects also demonstrated a near-normal return of lean thigh volume and oxidative capacity in this time span, regardless of the rehabilitation program. In contrast, cross-sectional studies comparing involved and uninvolved extremities in patients demonstrated functional deficits several years postinjury.[27,28]

The goals of our study were:

1. To examine the effects of cast immobilization following an ankle fracture on ankle plantar-flexor isometric and isokinetic peak torque and fatigue resistance. This goal was achieved via (1) comparisons between the involved and uninvolved lower extremities and (2) a comparison with age- and sex-matched, noninjured comparison subjects.

2. To study these patients longitudinally through a 10-week rehabilitation program (testing at 1, 5, and 10 weeks after immobilization) to determine whether peak torque and fatigue resistance return to baseline levels.

3. To assess the effects of cast immobilization following an ankle fracture on the patients' ability to perform functional tasks such as timed ambulation and stair climbing by comparing their functional performance with that of an age- and sex-matched, noninjured comparison group.

4. To assess the recovery in functional measures over a 10-week rehabilitation period (testing and retesting at 1, 5, and 10 weeks after immobilization).

5. To determine the correlation between plantar-flexion function and functional performance after immobilization.

Materials and Methods

Subjects

Ten patients (3 men, 7 women) and 10 noninjured comparison subjects participated in this study. All patients had sustained a fracture of the ankle malleolus that was treated by ORIF. Following surgery, the patients' involved lower extremity was immobilized in a short leg cast for a total of 8 weeks. The first 4 weeks was spent non-weight bearing, and the last 4 weeks was spent weight bearing as tolerated. The noninjured subjects were age- and sex-matched with the patients. The noninjured subjects and the patients (preinjury) ranged in activity level from sedentary to moderately active, with occupations ranging from desk-type work to those that require a moderate amount of physical activity (eg, physical therapist). About half of the subjects in each group exercised for recreational purposes 1 to 3 times a week. Characteristics of the patients and the noninjured subjects are summarized in Table 1. All participants were informed of the purpose of the investigation and gave their informed consent.

Table 1. Subject Characteristics(a)

                         Age (y)     Weight (kg)

                      [bar]X   SEM   [bar]X   SEM

Patients (n = 10)       35      4     74.1    4.6
Noninjured subjects
  (n = 10)              35      3     72.3    4.1

                      Height (cm)
                                     Involved LE/
                      [bar]X   SEM   Tested LE        Injury

Patients (n = 10)     172.5    5.0    6 L/4 R       1 LM/8 BIM/1
                                                    TRIM
Noninjured subjects
  (n = 10)            167.5    2.5    6 L/4 R

(a) LM = lateral malleolar fracture, BIM = bimalleolar fracture, TRIM = trimalleolar fracture, L = left, R = right, LE = lower extremity.

Experimental Protocol

Ankle plantar-flexor peak torque and fatigue resistance of the lower involved extremity, as well as the functional abilities of each patient, were determined at 1, 5, and 10 weeks post-immobilization. For comparison purposes, ankle-plantar flexor peak torque and fatigue resistance were also assessed in the uninvolved lower extremity (at 1 week after immobilization only) of the patients and the noninjured subjects. The side of the lower extremity tested in the noninjured subjects was matched to the involved side of the patients. The noninjured group also provided a means of comparison for the functional measurements.

Peak Torque and Fatigue Resistance

Isometric and isokinetic ankle plantar-flexion peak torque was measured on a Biodex Isokinetic Dynanometer.(*) The subjects were seated in an upright position on the exercise chair, which was mounted to the floor. Hip flexion angle was approximately 90 to 100 degrees, and the knee position was approximately 0 to 10 degrees of flexion. The axis of the dynamometer was aligned with the lateral malleolus, and the foot was secured to the footplate with a strap placed at the forefoot and ankle. Proximal stabilization was achieved with straps at the chest, hips, and knee. A comfortable range of motion (ROM) was individually determined for each subject.

Isokinetic plantar-flexion torque was measured from the neutral starting position (0 [degrees] of plantar flexion) through the available plantar-flexion ROM at speeds of 30 [degrees], 60 [degrees], 120 [degrees], and 180 [degrees]/s (in random order). Isokinetic peak torque was defined as the highest torque from a set of 5 maximal reciprocal contractions. Subjects performed 2 to 3 submaximal repetitions at increasing intensity as a warm-up at each test speed. A 2-minute rest period was given between tests. In an effort to optimize the reliability of the testing procedure, each test was repeated up to 2 times if the coefficient of variation (CV) among the 3 highest torques was more than 10%. Using this procedure, tests during which subjects did not exert maximal effort were discarded.

Isometric peak torque was assessed at 0 and 10 degrees of plantar flexion. Isometric peak torque was defined as the highest torque during 3 contractions (5-second contractions separated by 30 seconds of rest). The duration of each contraction was set at 5 seconds because, at 1 week after immobilization, torque did not plateau until about 3 seconds into the contraction in the majority of the patients. Similar to the isokinetic tests, if the CV among the 3 contractions exceeded 10%, the testing procedure was repeated after a short rest period (up to 2 times).

Fatigue resistance was determined during 50 successive maximal contractions of the ankle plantar flexors at a rate of 60 [degrees]/s. Fatigue (inverse of fatigue resistance) was defined as the relative (%) decrease in work between the first and last thirds of the exercise period. The total work performed during the 50 isokinetic contractions was also recorded.

Functional Measures

In addition to obtaining peak torque and fatigue data, several functional variables were also evaluated. Ambulation assessment consisted of timed walks of 9.1, 15.2, and 30.5 m (30, 50, and 100 ft) without the aid of an assistive device.[29] For all distances, subjects were timed at their most comfortable walking speed. The 9.1-m walk was also performed at the maximum safe speed,[30] defined as the fastest speed a person can walk without taking unnecessary risks. The subjects were asked to report the pain they experienced during each walking task on a scale of 0 (no pain) through 10 (worst pain imaginable).

In addition to timed walking tests, there were 2 stair-climbing tests: one using the reciprocal technique and one using a self-selected technique. The time to ascend and descend a flight of stairs (10 steps) using each technique was recorded.[31] Both tests were performed without assistance of a handrail or wall.

The ability to perform a single-leg heel-rise (plantar flexion onto the ball of the foot) was also tested, as described by Lunsford and Perry.[32] Subjects stood facing a wall and were asked to perform as many heel-rises as possible. Subjects were allowed to use the wall only to maintain their balance. The test was discontinued if subjects could not complete a heel-rise through their full, available ROM (as compared with the first repetition) or if they used the wall for assistance, flexed their knee, or asked to stop. Using the convention of Di Sabatino, the following nominal grades were assigned based on the number of heel-rises performed: 10 repetitions = "functional," 5 to 9 repetitions = "functionally fair," 1 to 4 repetitions = "functionally poor," and 0 repetitions = "nonfunctional."[33]

Rehabilitation Program

A rehabilitation program focusing on strengthening and ambulation was carried out on a 3-times-per-week basis for a total of 10 weeks. Although the patients began physical therapy immediately after cast removal, strengthening exercises and ambulation were delayed until the second week in an effort to guard against iatrogenic injury.

Patients initially received moist hot packs, applied to the ankle for 15 minutes, to begin their treatment session. The application of moist hot packs was followed by grade 3 or 4[34] anterior and posterior mobilizations of the tibiotalar joint with the ankle in a loose-packed position (10 [degrees] plantar flexed, neutral inversion/eversion).[35] Passive stretching with therapist assistance was used for ankle dorsiflexion with the knee both flexed and extended. Force during passive stretching was modified to patient tolerance. These treatments preceded ambulation and resistance training and were continued until passive ROM for the tibiotalar joint reached approximately 90% of that of the uninvolved lower extremity. Physical therapy sessions also included training on a Biomechanical Ankle Platform System (BAPS). Fifteen repetitions were completed in unilateral stance with the eyes open through all planes of motion at a level deemed appropriate by clinical observation.

Ambulation retraining was completed on a motorized treadmill, with subjects using the handrails for assistance, as necessary. Patients began with 10 minutes of ambulation on a level grade. Each week, 2 minutes and a 1% grade were added until a grade of 8% and a duration of 28 minutes were achieved. If subjects were initially unable to complete the full 10 minutes of exercise, a 2-minute rest period was incorporated. A grade was added to try to encourage dorsiflexion ROM as well as to provide a steadily increasing stimulus for the plantar-flexor muscles.

The resistance training protocol was based on the progressive resistance training principle and was a modification of the protocol used by Frontera et al[36] in older men. Plantar-flexion resistance training was performed on a customized hydraulic apparatus, which provided constant resistance over the entire ROM and allowed for concentric as well as eccentric training (Fig. 1). The resistance protocol was started with 2 sets of 10 repetitions at 40% of the subject's 1 repetition maximum (1RM) as a warm-up and 3 sets of 8 repetitions at 50% of their 1RM as a workout. If patients did not demonstrate an adverse response to resistance training (eg, increased pain, decreased ROM), workout sets were advanced to 80% of 1RM to provide a stimulus of sufficient intensity to increase the force-producing capabilities of the plantar-flexor muscles. The 1RM was determined weekly. Each repetition was performed slowly throughout the available ROM. The entire resistance training protocol was performed both with the knee extended (0 [degrees]) and with the knee flexed (30 [degrees]) in order to train both the gastrocnemius muscle and the soleus muscle.

[Figure 1 ILLUSTRATION OMITTED]

Reliability Assessment

In order to determine the intrasubjegt reliability, as well as possible training effects related to the multiple testing sessions, repeated measurements were performed in 7 noninjured subjects at the same time intervals as the measurements performed in the patients (3 measurements separated by 5 weeks). During this time period, subjects were not allowed to modify their activity level. Intraclass correlation coefficients (ICC [2,1]) were calculated to determine the reliability for each variable. Because it has been established that the actual limits of the ICC do not match the theoretical limits of 0.00 to 1.00 when the subject's variance is low, the CV ([CV.sub.test-retest] = SD/mean) was also determined.[37] We used this statistic to express, in a nonprobabilistic manner, the stability of our measurements.

No training effect was found in the noninjured subjects as a result of the multiple measurements. In addition, repeated measurements showed what we considered good to excellent reliability (ICC = .75-.93) for all muscle tests and functional performance measures, except for peak torque at 180 [degrees]/s, total work, and fatigue resistance, which had moderate reliability. The highest degree of reproducibility was found in the isometric and isokinetic torque measurements at slow speeds (30 [degrees] and 60 [degrees]/s), with ICCs ranging between .88 and .93 and [CV.sub.test-retest] values of 4% to 8%. Intraclass correlation coefficients and [CV.sub.test-retest values for all isometric and isokinetic torque measures are given in Table 2. The [CV.sub.test-retest] values were 22% for fatigue resistance and 17% for total work. The [CV.sub.test-retest] values for the functional variables ranged between 4.5% and 6.1% for all walking tests and between 10.8% and 17.3% for all stair-climbing tests (Tab. 3).

Table 2. Peak Torque: Test-Retest Reliability in Noninjured Subjects(a)

                      ICC (2,1)   CV (%)

ISM 0 [degrees]          .93        3.7
ISM 10 [degrees]         .88        5.8
ISK 30 [degrees]/s       .89        7.4
ISK 60 [degrees]/s       .90        7.8
ISK 120 [degrees]/s      .80        7.8
ISK 180 [degrees]/s      .60       12.2

(a) Intraclass correlation coefficient (ICC [2,1]) and coefficient of variation (CV) based on 3 measurements in each subject (n = 7). ISM = isometric peak torque, ISK = isokinetic peak torque.

Table 3. Functional Variables: Test-Retest Reliability in a Cohort of Noninjured Comparison Subjects(a)

                    CV (%)                        CV (%)

9.1-m walk (comf)    6.1     Stairs up (corr)      17.3
9.1-m walk (max)     5.1     Stairs down (corr)    12.6
15.2-m walk          4.5     Stairs up (any)       12.1
30.5-m walk          5.3     Stairs down (any)     10.8

(a) Coefficient of variation (CV) is based on 3 measurements in each subject (n = 7) (comf = comfortable speed, max = maximum safe speed, corr = using correct reciprocal technique, any = using any technique).

Data Analysis

Comparisons of peak torque and fatigue resistance between the involved and uninvolved lower extremities were performed using a paired t test. Comparisons within the involved lower extremity at different time points of rehabilitation were performed using a one-way analysis of variance for repeated measures. For post hoc analysis, the Bonferroni method was used. All comparisons between the patients and the noninjured subjects were performed using an unpaired t test. To evaluate the relationship between muscle performance and functional ability, a linear regression analysis was used. Statistical significance was established at P [is less than] .05. All values are expressed as mean [+ or -] standard error of the mean (SEM).

Results

Plantar-flexion peak torque was decreased at all angular speeds and positions following 8 weeks of immobilization. At the time of our initial measurement (1 week after immobilization), the isometric peak torque of the involved lower extremity was about half that of the uninvolved lower extremity (P [is less than] .001). Even larger differences were observed when a comparison was made with the noninjured subjects (~68% difference). The isometric and isokinetic peak torque data obtained in the involved and uninvolved lower extremities of the patients and the matched lower extremity of the noninjured subjects are displayed in Figure 2. Large deficits were also noted when the peak torque data were normalized for body weight. Table 4 shows the developed peak torque/body weight (N [multiplied by] m/kg) for the involved limb of the patients and the matched limb of the noninjured subjects.

[Figure 2 ILLUSTRATION OMITTED]

Table 4. Plantar-Flexor Peak Torque/Body Weight (N [multiplied by] m/kg) for Patients (Involved Lower Extremity) and Noninjured Subjects (Matched for Dominance)(a)

                        Noninjured      Patients
                         Subjects

                      [bar]X   SEM    [bar]X   SEM

ISM 0 [degrees]        0.39    0.01    0.16    0.01
ISM 10 [degrees]       0.32    0.02    0.12    0.01
ISK 30 [degrees]/s     0.30    0.02    0.10    0.01
ISK 60 [degrees]/s     0.23    0.01    0.08    0.01
ISK 120 [degrees]/s    O.19    0.01    0.07    0.01
ISK 180 [degrees]/s    0.18    0.01    0.05    0.01

(a) ISM = isometric peak torque, ISK = isokinetic peak torque (unpaired t test, P<.0001, df = 9).

Both isometric and isokinetic plantar-flexion peak torque increased markedly with rehabilitation (Fig. 2). With 5 weeks of rehabilitation, plantar-flexion peak torque of the involved lower extremity increased about 2-fold, such that, at 5 weeks post-immobilization, it was similar to that measured in the uninvolved lower extremity at 1 week post-immobilization. By the completion of 10 weeks of rehabilitation, plantar-flexion torque in the involved lower extremity surpassed the initial torque of the contralateral limb at all positions and speeds (Fig. 2). However, when a comparison was made with the noninjured subjects, no difference could be found between the involved extremity of the patients and the matched limb of the noninjured subjects.

Based on the results of the fatigue test at 1 week after immobilization, the patients' involved lower extremity was more fatigue resistant (P=.004) compared with their uninvolved lower extremity as well as the matched limb of the noninjured subjects (P [is less than] .05). As shown in Figure 3A, the patients' involved lower extremity showed only 25.4% [+ or -] 6.2% fatigue, whereas their uninvolved lower extremity demonstrated 51.8% [+ or -] 4.8% fatigue and the noninjured subjects' matched limb demonstrated 40.6% [+ or -] 3.6% fatigue. With 10 weeks of rehabilitation, muscle fatigue in the involved lower extremity increased to 41.3% [+ or -] 3.6%. Although fatigue resistance decreased with rehabilitation, the total work performed during the fatigue test increased. As shown in Figure 3B, the total work performed during 50 maximal isokinetic contractions at 60 [degrees]/s increased approximately 3-fold with 10 weeks of rehabilitation. The results for all fatigue-related variables are summarized in Figures 3A and 3B.

[Figure 3 ILLUSTRATION OMITTED]

Variables related to function also showed a large difference (P=.0015 to P [is less than] 0001) between the patients and the noninjured subjects at 1 week post-immobilization. The largest difference was observed in the time to descend stairs, with an approximate 4- to 5-fold difference (Figs. 4A and 4B). Patients demonstrated functional improvements throughout the 10 weeks of rehabilitation, although the largest improvements were observed during the first 5 weeks. By the end of the 10 weeks of rehabilitation, no difference could be found between the patients and the noninjured subjects in any of the functional tests (Figs. 4A and 4B).

[Figure 4 ILLUSTRATION OMITTED]

At the time of our initial measurements (1 week after immobilization), 7 of 9 patients were graded as "nonfunctional" in single-leg heel-rises. Two patients were scored as "functional." By 5 weeks after immobilization, 4 of 6 patients were graded as "functional," whereas all subjects were graded as "functional" by the end of the 10 weeks of rehabilitation.

Pain scores during all functional tasks decreased over time (P [is less than] .01). The average pain scores for all walking tests at a comfortable walking speed were 2.9 [+ or -] 0.7 at 1 week after immobilization, 0.8 [+ or -] 0.3 at 5 weeks after immobilization, and 0.6 [+ or -] 0.1 at 10 weeks after immobilization. The pain scores during walking at a maximum safe speed were slightly higher at all time points, with scores of 3.6 [+ or -] 0.8, 1.33 [+ or -] 0.6, and 0.8 [+ or -] 0.4, respectively.

Regression analysis revealed a strong relationship between plantar-flexor peak torque and functional variables. Although isometric and isokinetic torque showed a strong relationship with all functional variables measured in this study, the strongest correlation existed between descending stairs with any technique and isometric torque at 0 and 10 degrees of plantar flexion (Fig. 5). Both measures yielded correlations of r = .90 (P [is less than] .0001). The lowest correlation (r = .49-.58) was found between isometric and isokinetic torque and 9.1-m ambulation at a maximum safe speed. Table 5 provides the regression correlation coefficients between all torque and functional measures.

[Figure 5 ILLUSTRATION OMITTED]

Table 5. Correlation Between Peak Torque and Functional Measures(a)

                                                         Stairs
                      9.1 m    9.1 m   15.2 m   30.5 m     Up
                      (comf)   (max)   (comf)   (comf)   (corr)

ISM 0 [degrees]        .63      .51     .63      .65      .56
ISM 10 [degrees]       .60      .49     .63      .65      .53
ISK 30 [degrees]/s     .63      .59     .63      .64      .74
ISK 60 [degrees]/s     .61      .58     .62      .62      .71
ISK 120 [degrees]/s    .60      .56     .61      .61      .69
ISK 180 [degrees]/s    .57      .49     .57      .57      .58

                      Stairs   Stairs   Stairs   Stairs   Stairs
                       Down    Total      Up      Down    Total
                      (corr)   (corr)   (self)   (self)   (self)

ISM 0 [degrees]        .62      .67      .77      .90      .87
ISM 10 [degrees]       .58      .63      .80      .90      .89
ISK 30 [degrees]/s     .63      .73      .83      .83      .86
ISK 60 [degrees]/s     .64      .74      .83      .82      .85
ISK 120 [degrees]/s    .61      .70      .83      .84      .87
ISK 180 [degrees]/s    .62      .67      .83      .85      .87

(a) ISM = isometric peak torque, ISK = isokinetic peak torque, comf = comfortable speed, max = maximum safe speed, corr = using correct reciprocal technique, self = using self-selected technique (linear regression analysis, P<.05, df = 15-21).

Discussion

The main finding of this study is that 8 weeks of cast immobilization following surgery of the ankle joint causes a large decrease in peak torque and an increase in fatigue resistance, as well as a decrease in functional ability (stair climbing and walking). During 10 weeks of rehabilitation, peak torque and functional ability showed large improvements, whereas the fatigue resistance decreased. The largest improvements in peak torque were observed during the first 5 weeks of rehabilitation. These improvements resulted in similar torque measurements at all speeds and positions in the involved and uninvolved extremities at 5 weeks post-immobilization. Because only 10 patients with a malleolar fracture and 10 noninjured subjects were studied, however, one must be cautious generalizing the findings to all patients with an ankle injury subjected to cast immobilization followed by physical therapy.

In agreement with other studies, we found that immobilization severely affects muscle force. Based on a comparison between the involved and uninvolved lower extremities at 1 week post-immobilization, the patients demonstrated a 45% decrease in isometric peak torque. A comparison with the noninjured subjects showed an even larger deficit (~68%). Similar results were noted using isokinetic testing. Loss of muscle force has been shown to occur most precipitously during the first week of immobilization, with reported rates of 1% to 6% per day.[22] However, force continues to decline at a slower rate throughout the period of immobilization, resulting in deficits as large as 50% after 4 to 6 weeks of immobilization.[12,15,19,20] By comparison, reported reductions in muscle force during unloading average 13% after 10 days[25] and 26% to 46% after 5 weeks.[16,38]

In this study, the recovery in peak torque during rehabilitation occurred more rapidly than we had anticipated. At 5 weeks post-immobilization, isometric peak torque in the involved lower extremity was increased by approximately 70% and equaled that measured in the uninvolved lower extremity at 1 week post-immobilization. By 10 weeks of rehabilitation, peak torque in the involved lower extremity surpassed that of the contralateral lower extremity and was similar to that of the noninjured subjects. Due to the lack of longitudinal studies, no consensus exists as to the rate of recovery of muscle force after immobilization. Similar to our study, Ingemann-Hansen and Halkjaer-Kristensen[26] showed that patients regain between 69% and 92% of isometric and isokinetic knee extension force with 1 month of rehabilitation following an immobilization period of 1 month. In contrast, Seto et al[28] concluded that muscle force does not return within a period as short as 1 month and may not fully return within 5 years. Snyder-Mackler et al[39] showed that, following anterior cruciate ligament (ACL) reconstruction, quadriceps femoris muscle force in the involved lower extremity returns to only 70% of that of the contralateral lower extremity with 4 weeks of rehabilitation. Tegner et al[27] followed patients with chronic conservatively managed ACL injuries and found a 10% quadriceps femoris muscle force deficit immediately after 3 months of resistance training and at a 2-year follow-up. Seto et al[28] reported a remaining 59% to 68% deficit in isokinetic quadriceps femoris muscle force 5 years after endoscopic knee reconstruction. However, data acquired in our study clearly demonstrate the limitation of a bilateral comparison postinjury. We found that, for all angular speeds, the uninvolved lower extremity of the patients demonstrated a lower peak torque compared with the matched limb of the noninjured subjects, indicating adaptations in the contralateral limb during cast immobilization.

Although peak torque showed improvement throughout the 10 weeks of

rehabilitation, our data demonstrated that the largest increase in torque occurred during the first 5 weeks. We believe that this phenomenon is often seen in clinical practice. Force gains within the first month of rehabilitation are commonly thought to be induced by neurologic adaptation[15,16,40] rather than by an increase in muscle mass. However, in a recently published study,[14] we demonstrated hypertrophy of the plantar-flexor muscles within the first 5 weeks of rehabilitation after cast immobilization.

We contend that functionally it may be more important that a muscle demonstrates adequate fatigue resistance rather than a high peak torque during a single contraction. Contrary to the general belief, we found that the plantar-flexor muscles demonstrated an increased fatigue resistance following immobilization. Based on 50 maximal contractions, the patients' involved lower extremity demonstrated only 25% fatigue, whereas their uninvolved lower extremity showed 52% fatigue and the matched lower extremity of the noninjured subjects showed 41% fatigue. Similar to plantar-flexion torque, muscle fatigue patterns evolved toward normal throughout the 10-week period.

There are few studies that have investigated the effect of immobilization or disuse on fatigue resistance in human muscle. Duchateau and Hainaut[15] reported in a case study that 5 weeks of bed rest does not affect the muscle's relative resistance to fatigue. Similar to our study, however, Snyder-Mackler et al[41] demonstrated, using an electrically induced fatigue test, that atrophied quadriceps femoris muscles in patients undergoing ACL reconstruction are more fatigue resistant compared with the contralateral uninvolved muscles. Snyder-Mackler et al postulated that the increase in fatigue resistance with disuse may be due to the selective recruitment of more fatigue-resistant motor units. We conjecture that, as a result of neurological adaptations during disuse, motor units in involved muscles are firing at submaximal rates, providing a metabolic reserve. Alternatively, the increase in fatigue resistance following disuse may be related to a shift in the resting metabolic content. In a recent study using [31]P-magnetic resonance spectroscopy, we found an increase in the basal inorganic phosphate concentration post-immobilization.[14] Elevated inorganic phosphate concentrations have been shown to inhibit actin-myosin cross-bridge cycling and increase resistance to fatigue via a shift in the [Ca.sup.2+]/force curve, providing an alternative mechanism for the observed increase in fatigue resistance with disuse.[17,42,43]

An increase in fatigue resistance during a maximal test does not necessarily mean that the involved lower extremity fatigues less rapidly than the uninvolved lower extremity when performing the same submaximal task. For instance, if subjects were asked to produce a plantarflexion torque equal to 40 N [multiplied by] m, the involved limb would probably fatigue more rapidly than the uninvolved limb because the required torque production is closer to its peak torque. Our data also demonstrated that, even though the fatigue resistance was higher in the patients' involved lower extremity compared with either their uninvolved lower extremity or the matched lower extremity of the noninjured subjects, the total work performed during a maximal fatigue test was lower. The total work performed by the patients' involved lower extremity during 50 maximal isokinetic contractions at 60 [degrees]/s was 235 [+ or -] 39 J, compared with 511 [+ or -] 111 J in their uninvolved lower extremity and 940 [+ or -] 152 J in the matched lower extremity of the noninjured subjects.

We showed that, at 1 week post-immobilization, the patients took longer to complete functional tasks such as ambulation and stair climbing than did the noninjured subjects. For instance, the time to descend a flight of 10 stairs was 4 to 5 times longer in the patients than in the noninjured subjects, whether performed correctly or using a self-selected technique. Differences were also seen in timed ambulation on a level surface. During a 9.1-m maximal safe speed walk, the noninjured subjects ambulated approximately twice as fast as the patients. The patients' performance on both walking and stair-climbing tests returned to control levels by the end of the rehabilitation period. Similar to torque, the greatest rate of improvement was noted during the first 5 weeks of rehabilitation.

Isometric and isokinetic peak torque proved to be good predictors of functional performance in the patients with ankle fracture. Linear regression correlation coefficients between peak torque and functional tasks (stair climbing and walking) ranged between .49 and .90. Previous studies examining the relationship between muscle force and functional indicators have focused on the quadriceps femoris muscle.[31,44,45] Snyder-Mackler et al[39] showed a positive relationship (r = .64) between quadriceps femoris muscle peak torque and knee joint excursion during gait in patients with ACL reconstructions. Bassey et al[32] showed a strong relationship between quadriceps femoris muscle force and walking speed (r = .80), as well as between quadriceps femoris muscle force and the ability to ascend stairs (r = .81) in elderly subjects. We found a linear regression coefficient of.90 between isometric peak torque at 0 and 10 degrees and the time to descend stairs using a self-selected technique. Based on our clinical experience, descending stairs and level-surface ambulation are dependent not only on muscle force but also on dorsiflexion ROM. Although ROM and potentially conflicting impairments such as swelling were not included in our analysis, the high correlation between ankle plantar-flexor torque and descending stairs indicates that plantar-flexion torque may be a key predictor of return to function in patients with ankle fractures.

Despite the complexity of the ankle joint, we presented a reliable testing procedure for isometric and isokinetic torque assessment in the ankle plantar flexors. Intraclass correlation coefficients in this study ranged from .80 to .93 during isometric and isokinetic testing at speeds up to 120 [degrees]/s. Similar high correlation coefficients were reported by Karnofel et al[46] and Andersen.[47] In contrast, Sleivert and Wenger[48] reported much lower ICCs during ankle plantar flexion (ICC = .55-.76), as compared with knee extension (ICC = .64-.94). Because of the complex biomechanics of the ankle joint, involving 3 articulations, the inclusion of additional criteria may be warranted for reliable assessment of peak torque. In this study, each test consisted of 3 to 5 maximal contractions and testing was repeated if the variation (see "Method" section) was more than 10%. Similar criteria for retesting were used by Andersen.[47] In addition, Andersen[47] showed that multiple repetitions are needed for the ankle plantar flexors to reach peak torque.

Although our data confirm that ankle plantar-flexor peak torque can be assessed with a high degree of reliability with an isokinetic dynamometer, we should point out that all reproducibility measurements were performed in noninjured comparison subjects only and not in the group of patients with ankle fracture. We contend that accurate assessment of reproducibility in patients with an acute condition, such as ankle fractures, is difficult, if not impossible. The only way to ensure that there is no change over time as a result of recovery in this patient population is to perform the repeated measures in a short time frame (eg, over consecutive days). However, performing repeated measures in a short time frame in a population that is prone to delayed-onset muscle soreness increases the likelihood that subsequent torque measurements are affected by the previous test. Additionally, frequent testing could adversely affect the patients' medical condition.

The improvements in ankle plantar-flexion torque with rehabilitation were reflected in the patients' ability to perform heel-rises. Heel-rises have historically been used to assess plantar-flexion force due to the difficulty of overpowering these muscles with manual resistance. Heel-rises also have the advantage that they are performed in a upright test position, thereby using the subject's own body weight as resistance. By 5 weeks of rehabilitation, 4 of 6 patients were graded as "functional" on the single-leg heel-rise test at a time when their peak torque was found to be approximately equal to that of the involved lower extremity at 1 week post-immobilization. At 10 weeks of rehabilitation, when the patients' peak torque as well as their performance on functional tests had returned to normal levels, 6 of 6 patients were graded as "functional" on their ability to perform heel-rises. Note that the use of the term "functional" relates to the scale we used, not to our ability to infer how these muscles will be used during daily tasks.

Summary and Conclusions

Eight weeks of cast immobilization following surgery of the ankle joint resulted in a large decrease in peak torque, an increase in fatigue resistance, and a decrease in stair-climbing ability and walking speed. With 10 weeks of rehabilitation, all alterations were reversed. The largest improvement in peak torque was observed during the first 5 weeks, resulting in similar torque measurements at all speeds and positions in the patients' involved and uninvolved limbs. However, comparisons with an age- and sex-matched, noninjured comparison group showed that complete recovery in peak torque, stair climbing, walking, and muscle fatigue resistance was only reached by 10 weeks of rehabilitation.

We demonstrated that ankle plantar-flexion torque is a good predictor of stair-climbing and walking performance in patients with ankle fracture. Based on longitudinal measurements in 10 patients, correlation coefficients between peak torque and time to complete walking and stair-climbing tasks ranged between .49 and .90. The strongest correlation was found between isometric torque and the time to descend stairs using a self-selected technique.

(*) Medical Systems Inc, Brookhaven R&D Plaza, 20 Ramsay Rd, Box 72, Shirley, NY 11967-0702.

References

[1] Tropp H, Norlin R. Ankle performance after ankle fracture: a randomized study of early mobilization. Foot Ankle Int. 1995;16:79-83.

[2] Herbison GJ, Jaweed MM, Ditunno JF. Muscle fiber atrophy after cast immobilization in the rat. Arch Phys Med Rehabil. 1978;59:301-305.

[3] Max SR, Maier RF, Vogelsang L. Lysosomes and disuse atrophy of skeletal muscle. Arch Biochem Biophys. 1971;146:227-232.

[4] Desplanches D, Mayet MH, Sempore B, et al. Effect of spontaneous recovery or retraining after hindlimb suspension on aerobic capacity. J Appl Physiol. 1987;63:1739-1743.

[5] Desplanches D, Mayet MH, Sempore B, Flandrois R. Structural and functional responses to prolonged hindlimb suspension in rat muscle. J Appl Physiol. 1987;63:558-563.

[6] Roy RR, Bello MA, Bouisson P, Edgerton VR. Size and metabolic properties of fibers in rat fast-twitch muscles after hindlimb suspension. J Appl Physiol. 1987;62:2348-2357.

[7] Edgerton VR, Barnard KJ, Peter JB, et al. Properties of the immobilized hind-limb muscles of the Galago senegalensis. Exp Neurol. 1975;46: 115-131.

[8] Booth FW, Criswell DS. Molecular events underlying skeletal muscle atrophy and the development of effective countermeasures. Int J Sports Med. 1997;18(suppl 4):S265-S269.

[9] Goldspink DF. The influence of activity on muscle size and protein turnover. J Physiol (Lond). 1977;264:283-296.

[10] Goldspink DF. The influence of immobilization and stretch on protein turnover of rat skeletal muscle. J Physiol (Lond). 1977;264: 267-282.

[11] Halkjaer-Kristensen J, Ingemann-Hansen T. Wasting of the human quadriceps after knee ligament injuries. Scand J Rehabil Med Suppl. 1985;13:5-11.

[12] Veldhuizen JW, Verstappen FTJ, Vroemen JP, et al. Functional and morphological adaptations following four weeks of knee immobilization. Int J Sports Med. 1993;14:283-287.

[13] Hather BM, Adams GR, Tesh PA, Dudley GA. Skeletal muscle responses to lower limb suspension in humans. J Appl Physiol. 1992;72: 1493-1498.

[14] Vandenborne K, Elliot MA, Walter GA, et al. Longitudinal study of skeletal muscle adaptations during immobilization and rehabilitation. Muscle Nerve. 1998;21:1006-1012.

[15] Duchateau J, Hainaut K. Effects of immobilizatton on contractile properties, recruitment, and firing of human motor units. J Physiol (Lond). 1990;422:55-65.

[16] Duchateau J. Bed rest induces neural and contractile adaptations in triceps surae. Med Sci Sports Exerc. 1995;27:1581-1589.

[17] Kentish JC. The effects of inorganic phosphate and creatine phosphate on force production in skinned muscles from rat ventricle. J Physiol (Lond). 1986;370:585-604.

[18] Muller EA. Influence of training and of inactivity on muscle strength. Arch Phys Med Rehabil. 1970;51:449-462.

[19] Appell HJ. Muscular atrophy following immobilisation: a review. Sports Med. 1990;10:42-58.

[20] MacDougall JD, Elder GC, Sale DG, et al. Effects of strength training and immobilization on human muscle fibres. Eur J Appl Physiol. 1980;43:25-34.

[21] Widrick JJ, Fitts RH. Peak force and maximal shortening velocity of soleus fibers after non-weight-bearing and resistance exercise. J Appl Physiol. 1997;82:189 -195.

[22] Stillwell DM, McLarren GL, Gersten JW. Atrophy of quadriceps muscle due to immobilization of the lower extremity. Arch Phys Med Rehabil. 1967;48:289-295.

[23] Halkjaer-Kristensen J, Ingemann-Hansen T. Wasting of the human quadriceps muscle after knee ligament injuries, IV: dynamic and static muscle function. Scand J Rehabil Med Suppl. 1985;13:29-37.

[24] Berg HE, Dudley GA, Hather B, Tesch PA. Work capacity and metabolic and morphologic characteristics of the human quadriceps muscle in response to unloading. Clin Physiol. 1993; 13:337-347.

[25] Berg HE, Tesch PA. Changes in muscle function in response to 10 days of lower limb unloading in humans. Acta Physiol Scand. 1996;157: 63-70.

[26] Ingemann-Hansen T, Halkjaer-Kristensen J. Physical training of the hypotrophic quadriceps muscle in a man, I: the effects of different training programs on muscular performance. Scand J Rehabil Med Suppl. 1985;13:38-44.

[27] Tegner Y, Lysholm J, Gillquist J, Oberg B. Two-year follow-up of conservative treatment of knee ligament injuries. Acta Orthop Scand. 1984;55:176-180.

[28] Seto JL, Orofino AS, Morrissey MC, et al. Assessment of quadriceps/ hamstring strength, knee ligament stability, functional and sports activity levels five years after anterior cruciate ligament reconstruction. Am J Sports Med. 1988;16:170-180.

[29] Fisher NM, Pendergast DR, Gresham GE, Calkins E. Muscle rehabilitation: its effect on muscular and functional performance of patients with knee osteoarthritis. Arch Phys Med Rehabil. 1991;72:367-374.

[30] Bohannon RW, Walsh S. Nature, reliability, and predictive value of muscle performance measures in patients with hemiparesis following stroke. Arch Phys Med Rehabil. 1993;73:721-725.

[31] Bassey EJ, Fiatarone MA, O'Neill EF, et al. Leg extensor power and functional performance in very old men and women. Clin Sci (Colch). 1992;82:321-327.

[32] Lunsford BR, PerryJ. The standing heel-rise test for ankle plantar flexion: criterion for normal. Phis Ther. 1995;75:694-698.

[33] Di Sabatino D. Ankle and foot region. In: Palmer ML, Epler M, eds. Clinical Assessment Procedures in Physical Therapy. Philadelphia, Pa: JB Lippincott Co; 1990:308-310.

[34] Maitland G. Peripheral Manipulation. 2nd ed. Boston, Mass: Butterworth; 1977.

[35] Kisner C, Colby C. Therapeutic Exercise. 2nd ed. Philadelphia, Pa: FA Davis Co; 1990:10, 199, 202.

[36] Frontera WR, Meredith CN, O'Reilly KP, et al. Strength conditioning in older men: skeletal muscle hypertrophy and improved function. J Appl Physiol. 1988;64:1038-1044.

[37] Portney L, Watkins M. Foundations of Clinical Research: Applications to Practice. Englewood Cliffs, NJ: Prentice Hall; 1993:525.

[38] Leblanc A, Gogia P, Schneider V, et al. Calf muscle area and strength changes after five weeks of horizontal bed rest. Am J Sports Med. 1988; 16:624-629.

[39] Snyder-Mackler L, Delitto A, Bailey SL, Stralka SW. Strength of the quadriceps femoris muscle and functional recovery after reconstruction of the anterior cruciate ligament: a prospective, randomized clinical trial of electrical stimulation. J Bone Joint Surg Am. 1995;77: 1166-1173.

[40] Berg HE, Larsson L, Tesch PA. Lower limb skeletal muscle function after 6 weeks of bed rest. J Appl Physiol. 1997;82:182-188.

[41] Snyder-Mackler L, Binder-Macleod SA, Williams PR. Fatigability of human quadriceps femoris muscle following anterior cruciate ligament reconstruction. Med Sci Sports Exerc. 1993;25:783-789.

[42] Phillips SK, Wiseman RW, Woledge RC, Kushmerick MJ. The effects of metabolic fuel on force production and resting inorganic phosphate levels in mouse skeletal muscle. J Physiol (Lond). 1993;462:135-146.

[43] Cooke R, Franks K, Luciani GB, Pate E. The inhibition of rabbit skeletal muscle contraction by hydrogen ions and phosphate. J Physiol (Lond). 1988;395:77-97.

[44] Fiatarone EA, Marks EC, Ryan ND, et al. High-intensity strength training in nonagenarians: effects on skeletal muscle. JAMA. 1990;263: 3029-3034.

[45] Snyder-Mackler L, De Luca PF, Williams PR, et al. Reflex inhibition of the quadriceps femoris muscle after injury or reconstruction of the anterior cruciate ligament. J Bone Joint Surg Am. 1994;76:555-560.

[46] Karnofel H, Wilkinson K, Lentell G. Reliability of isokinetic muscle testing at the ankle. J Orthop Sports Phys Ther. 1989;11:150-154.

[47] Andersen H. Reliability of isokinetic measurements of ankle dorsal and plantar flexors in normal subjects and in patients with peripheral neuropathy. Arch Phys Meal Rehabil. 1996;77:265-268.

[48] Sleivert GG, Wenger HA. Reliability of measuring isometric and isokinetic peak torque, rate of torque development, integrated electromyography, and tibial nerve conduction velocity. Arch Phys Med Rehabil. 1994;75:1315-1321.

MA Shaffer, PT, MSPT, ATC, is Physical Therapist and Certified Athletic Trainer, Cyclone Sports Medicine, 132 Lied Recreation Facility, Iowa State University, Ames, IA 50011 (USA) ([e-mail]). Mr Shaffer was Staff Physical Therapist, Occupational and Physical Therapy Department, Hospital of the University of Pennsylvania, Philadelphia, Pa, when this study was conducted. Address all correspondence to Mr Shaffer.

E Okereke, PharmD, MD, is Chief, Foot and Ankle Service, Department of Orthopaedic Surgery, Hospital of the University of Pennsylvania, and Assistant Professor of Orthopaedic Surgery, University of Pennsylvania School of Medicine, Philadelphia, Pa.

JL Esterhai Jr, MD, is Associate Professor of Orthopaedic Surgery, University of Pennsylvania School of Medicine.

MA Elliott, BS, is a graduate student in biophysics at the University of Pennsylvania.

GA Walter, PhD, is a postdoctoral researcher at the University of Pennsylvania.

SH Yim, BS, was Research Lab Technician, Department of Rehabilitation Medicine, University of Pennsylvania, when this study was conducted.

K Vandenborne, PT, PhD, is Research Assistant Professor of Physiology, Rehabilitation Medicine, and Radiology, Department of Physiology, University of Pennsylvania.

Mr Shaffer, Dr Okereke, Dr Esterhai, Mr Elliott, Dr Walker, and Dr Vandenborne provided concept/research design, data analysis/interpretation, and consultation (including review of manuscript before submission). Mr Shaffer, Mr Elliott, Dr Walker, and Dr Vandenborne provided writing. Mr Shaffer, Dr Okereke, Dr Esterhai, Mr Yim, and Dr Vandenborne provided data collection. Mr Shaffer, Dr Okereke, and Dr Esterhai provided subjects, and Mr Shaffer and Dr Vandenborne provided facilities/equipment. Mr Shaffer, Mr Kim, and Dr Vandenborne provided project management, and Dr Vandenborne provided fund procurement and institution liaisons. David Lawson, BS, and Alex Swift, BS, assisted with data collection, and Susan Brenneman, PT, assisted with data analysis. John Noh provided secretarial support.

This study was approved by the University of Pennsylvania Human Research Review Board.

This project was funded by a grant from the National Institutes of Health (R29-HD33738).

This article was submitted June 17, 1999, and was accepted April 25, 2000.