Passive Ankle Stiffness in Subjects With Diabetes and Peripheral Neuropathy Versus an Age-Matched Comparison Group.

Several musculoskeletal factors contribute to the production and control of human movement, including passive muscle stiffness, active muscle tension, and joint motion. Passive stiffness can be defined as the resistance to elongation or, in physics terms, the change in tension per unit change in length.[1] Active tension is generated when the muscle receives input at the neuromuscular junction (eg, during a voluntary or reflexive contraction), and it is attributed to structures within the contractile element of the muscle.[2-4] Passive tension generation occurs when a passive (not contracting) muscle is lengthened, and it is believed to originate from the series elastic and parallel elastic elements of muscle (eg, the tendon, crossbridge attachments, structural proteins within the myofibril, connective tissue around the muscle fibers and fascicles).[3,4]

Both active tension and passive tension contribute to the overall (total) tension produced by a muscle; therefore, both types of tension may contribute to movement. Although the effect of active tension production on movement has been studied under many different conditions,[5-13] the effect of passive tension has not been studied extensively. Under conditions of decreased active tension production, such as disease or disuse, passive tension may make a greater contribution to total tension production than in healthy conditions.[14,15] For example, the net ankle joint moment during the stance phase of gait is due to both active and passive components of the plantar flexor muscles.[16] In the absence of appropriate active tension, it is possible that a considerable portion of the ankle moment during gait may come from passive structures.[14,15]

Several researchers[17,18] have attempted to quantify the passive tension (and passive stiffness) of plantar flexor muscles in humans by measuring passive ankle torque and assuming that the major contributor to this torque is the plantar flexor muscle group. Gajdosik et al[17] described a method of assessing plantar flexor extensibility (the inverse of stiffness) in a study investigating the effect of age on passive and active torque variables in women. Although they reported no difference in actual passive extensibility (change in ankle joint angle per change in passive torque) of the plantar flexors, they observed a left shift of the peak passive and active torque values in the older age group (peak active and passive torque values occurred at a relatively more plantarflexed joint angle).

Chesworth and Vandervoort[19] attempted to quantify passive stiffness variables in subjects who had their foot and ankle casted after an ankle fracture. Using a method similar to that of Gajdosik et al,[17] they generated passive torque versus angle curves for the casted and noncasted ankles of each subject, and they determined the passive torque at a specific joint angle (0 [degrees] of dorsiflexion) and passive stiffness (slope of the torque curve) at the same angle. The results showed no difference in passive torque at 0 degrees of dorsiflexion between the casted and noncasted ankles. There was a difference in passive stiffness, however, at 0 degrees of dorsiflexion. The authors suggested that the angle at which the ankles were casted (0 [degrees] of dorsiflexion) may have been sufficient to lengthen the plantar flexors during immobilization, such that the effects of immobilization on passive torque at that angle were prevented. They noted, however, that the casted ankles demonstrated less dorsiflexion motion.

Although researchers have examined passive ankle torque and stiffness in elderly people with no known pathology[17] and young patients postfracture,[19] few studies have quantified passive muscle stiffness in patients with notable pathology and impairments such as decreased range of motion and decreased force. Sinkjaer and Magnussen[18] measured passive and reflex-mediated stiffness in the plantar flexor muscles of patients with hemiplegia and found an increase in passive stiffness of the paretic limb compared with the contralateral limb. The amplitude of the stretch, however, was relatively small (4 [degrees]). Passive torque and stiffness at end-range dorsiflexion were not examined.

Another group of patients with known pathology and impairments consists of patients with diabetes mellitus (DM) and peripheral neuropathy (PN). Mueller et al[20] and Andersen and colleagues[21,22] found that subjects with DM and PN had decreased plantar flexor muscle peak torque compared with control subjects, and several authors[20,23,24] have documented decreased ankle joint motion in this population. Furthermore, some subjects with DM and PN have been shown to have altered gait characteristics[20] and postural instability,[25] and we have observed that they often complain of stiffness during daily activities.

Research has provided evidence that the ultrastructure of collagen, a component of the elastic elements of muscle, is altered with long-term DM,[26] With sustained high glucose concentrations, a chemical reaction appears to occur between the free amino group of structural proteins and glucose, forming irreversible products called advanced glycosylated end-products (AGEs).[26] These AGEs tend to accumulate on long-lived structural proteins, such as collagen and basement membrane proteins, leading to cross-linking between AGE molecules, and with other unmodified proteins through covalent trapping.[26]

In light of the diminished peak torque and ankle joint motion, altered gait characteristics, and physiological changes in connective tissue that occur with DM and PN, we believe passive stiffness also may be altered in this population, which has important implications for movement. For example, increased plantar flexor stiffness may limit dorsiflexion motion, leading to abnormal gait characteristics, but it may also positively influence gait by contributing to supportive and propulsive forces during the stance phase. Furthermore, quantifying stiffness may provide insights into the management of patients with DM and PN and other neuromuscular or musculoskeletal conditions. For example, increased plantar flexor stiffness may contribute to increased plantar pressure during gait, a phenomenon associated with recurrent plantar ulceration in patients with DM and PN.[27] Minimizing plantar flexor stiffness or teaching patients to walk within a "low stiffness" ankle range of motion may be indicated. The purpose of our study was to quantify and compare variables associated with passive stiffness in subjects with DM and PN and an age-matched comparison group. We hypothesized that subjects with DM and PN would have greater plantar flexor stiffness (change in passive torque per unit change in joint angle) and greater passive torque at 5 degrees of dorsiflexion than subjects in the comparison group. We also hypothesized that subjects with DM and PN would have less ankle motion (dorsiflexion range of motion), resulting in a "shift" of the passive torque versus angle curve in the direction of plantar flexion.

Method

Subjects

Two groups of subjects were examined: 17 subjects with DM and PN and 17 age-matched subjects in a comparison group. We chose to study subjects with DM and PN (not DM without PN or PN without DM) because we believed that subjects with DM and PN, when combined with subjects in a comparison group, would provide a wide range of values for muscle stiffness and joint motion. Furthermore, it was not the purpose of our study to parse out the contributions from DM and PN separately. Subject groups (mean values) were matched by age, sex, height, weight, and body mass index. In addition, subjects with DM and PN had the following deficits compared with subjects in the comparison group: decreased dorsiflexion range of motion (measured with a goniometer), decreased peak concentric plantar flexor torque, and decreased walking speed. These deficits, however, were not mandatory inclusion criteria. Subject characteristics are detailed in Table 1. Subjects with DM and PN were recruited through the Diabetic Foot Centers at Barnes/Jewish Hospital, Christian Hospital Northeast, and St John's Mercy Medical Center, St Louis, Mo. Inclusion criteria for this group consisted of a history of DM (of any duration), ability to lie supine, ability to walk independently without an assistive device, and the presence of peripheral sensory. neuropathy. Peripheral neuropathy was characterized by the loss of protective sensation on the plantar surface of the foot, measured using Semmes-Weinstein monofilaments.[28] Exclusion criteria consisted of a current or previous severe orthopedic injury to the lower extremity (eg, ankle fracture, joint fusion, joint replacement) or a current or previous neurological insult (eg, cerebrovascular accident).

Table 1. Subject Characteristics for Both Groups(a)

                                  Subjects With DM and PN
                                  (n = 17)

Variable                          [bar]X      SD         Range

Age (y)                            58.5      11.7       38 to 77
Mass (kg)                          89.7      23.1     56.2 to 152
Body mass index (kg/[m.sup.2])     30.3       6.6     21.6 to 49.5
Height (cm)                       171.4      11.5    147.3 to 191.8
Dorsiflexion range of
 motion([degrees])(c)               0.5       4.8     - 10 to 8
Peak concentric plantar
 flexor torque
  (N [multiplied by] m/kg)(d)       0.44      0.15    0.22 to 0.72
Walking speed (m/min)              53.37     16.63    27.6 to 84.7
Duration of DM (y)                 17.97      9.89

                                   Number    Number
History of foot ulcers               13
Sex
  Male                               10        10
  Female                              7         7

                                   Comparison Group
                                   (n=17)

Variable                          [bar]X      SD          Range

Age (y)                            62.7       6.2       50 to 73
Mass (kg)                          83.7      15.8     55.3 to 113.9
Body mass index (kg/[m.sup.2])     28.0       3.5     22.3 to 35.3
Height (cm)                       172.5      11.4    154.9 to 190.5
Dorsiflexion range of
 motion([degrees])(c)               6.7       4.8       -2 to 16
Peak concentric plantar
 flexor torque
  (N [multiplied by] m/kg)(d)      0.69      0.19      0.45 to 1.1
Walking speed (m/min)             68.89      9.31     50.06 to 86.5
Duration of DM (y)

History of foot ulcers
Sex
  Male
  Female

Variable                               p(b)

Age (y)                                .20
Mass (kg)                              .39
Body mass index (kg/[m.sup.2])         .20
Height (cm)                            .78
Dorsiflexion range of
 motion([degrees])(c)                  .001(e)
Peak concentric plantar
 flexor torque
  (N [multiplied by] m/kg)(d)         <.001(e)
Walking speed (m/min)                  .002
Duration of DM (y)

History of foot ulcers
Sex
  Male
  Female

(a) There were no differences in age, height, mass, or body mass index. Clinical deficits in subjects with diabetes mellitus (DM) and peripheral neuropathy (PN) included decreased peak concentric plantar flexor torque, decreased dorsiflexion range of motion, and decreased walking speed.

(b) Student t test, df=32.

(c) Maximal dorsiflexion angle measured with a goniometer.

(d) Peak concentric plantar flexor torque (60 [degrees]/s) is measured on a Kin-Com dynamometer.

(e) Significant group differences, [Alpha]=.006.

Subjects in the comparison group were recruited from family members of the subjects with DM and PN, from employees of the medical center, and from the community. Inclusion criteria for the comparison group were the ability to lie supine and the ability to walk independently without an assistive device. Exclusion criteria were the same as those of the subjects with DM and PN, in addition to a history of DM and involvement in a regular aerobic exercise program ([is greater than or equal to] 2 times per week). No attempt was made to assess the activity level of the subjects of either group. Because none of the subjects with DM and PN were involved in regular aerobic exercise, however, we added this exclusion criterion for the subjects in the comparison group.

Procedure

Each subject was tested in a single session. The procedures were explained thoroughly, and all subjects read and signed an institutional review board-approved informed consent statement prior to testing. A brief medical history was obtained, and demographic data were recorded, including date of birth, sex, height, weight, body mass index, general health status, and activity level. Subjects with DM and PN were asked specific questions regarding the history, duration, and their control of diabetes.

Range of motion and peak torque. To characterize the clinical status of our subjects, measurements of dorsiflexion range of motion and plantar flexor peak torque were obtained. All measurements were taken on one lower extremity for each subject. No attempt was made to randomize the selection of right versus left side, but there was no reason to believe there would be an inherent "side" difference in the measures used in this study. Maximal dorsiflexion range of motion was measured with a masked goniometer (1 [degrees] increments), with the subject lying prone in knee extension. After 3 to 5 practice trials, the subject was asked to pull his or her toes up and hold the position while the examiner aligned the stationary arm of the goniometer parallel to the fibula. The movable arm of the goniometer was positioned parallel to the sole of the foot, and the axis fell approximately over the lateral calcaneus. The goniometer was then given to a second examiner to read. Intrarater reliability was determined by having the examiner obtain 2 dorsiflexion measurements, approximately 30 seconds apart, while the subjects remained in the prone position. With a sample size of 34 (17 subjects with DM and 17 subjects without DM), the intraclass correlation coefficient (ICC [2,1]) was .95. We acknowledge the possibility of inflated ICC values due to the method of obtaining the repeated measurements.

Plantar flexor peak torque was estimated using an isokinetic device. The set-up procedures are described in the "Stiffness Measurements" section. Once positioned properly, subjects were asked to push as hard possible from maximal dorsiflexion into plantar flexion (determination of range is described in the "Stiffness Measurements" section). Concentric peak torque values from 3 trials of maximal plantar flexion were recorded and averaged for each subject. Intrarater reliability for concentric peak torque measurements was obtained using the 3 maximal plantar-flexion trials. Subjects were not removed from the device, but there was approximately 30 seconds of rest between trials. Using 34 subjects (17 with DM and 17 without DM), the ICC (2,1) value was .97.

Sensation testing. Sensory testing using Semmes-Weinstein monofilaments was performed on each subject to assist in the confirmation of PN in the subjects with DM. This procedure involved using a 5.07 monofilament and testing several sites on the plantar surface of both feet as described by Mueller et al.[23] These sites, tested approximately 3 to 5 times, were (1) the first, third, and fifth toes and the metatarsal heads, (2) the medial and lateral midfoot, and(3) the heel.[23] Subjects lacked protective sensation if they could not detect the 5.07 monofilament on 80% of the trials.[28]

Stiffness measurements. In order to verify that the stiffness measurements were truly passive (no active muscle contractions elicited), electromyographic (EMG) monitoring (GCS-67 Multichannel Electromyographic System(*)) was used throughout the procedure. The anterior and posterior surfaces of the shank (same lower extremity used in the previous procedures) were cleaned with alcohol to reduce skin impedance. Surface electrodes, with attached preamplifiers, were applied over the belly of the anterior tibialis, gastrocnemius, and soleus muscle (distal to the gastrocnemius muscle belly and lateral to the Achilles tendon). The EMG amplifier gain settings ranged from 500 to 10,000 and were adjusted based on the EMG signal output viewed on an analog oscilloscope. The raw signal was collected and high-pass filtered at 40 Hz, creating a frequency response of 40 to 4,000 Hz.

A Kin-Cam dynamometer (Kinetic Communicator Exercise System III-500H, software version 4.06)([dagger]) with attached ankle apparatus was used to assess passive plantar flexor torque. The Kin-Cam is a hydraulically driven, computer-controlled device that monitors force through a strain gauge bridge transducer located on a rigid mechanical lever arm. The Kin-Cam was set in the isokinetic mode for ankle plantar flexion, and the gravity-correction procedure was performed on the empty ankle apparatus according to the manufacturer's instructions. The limb was not included in the gravity correction because we assumed the weight of the foot (approximately 1.5% of body weight[16]) to be negligible. Each subject was positioned supine on the Kin-Cam bench with the knee in 10 degrees of flexion (maintained by placing a rolled towel under the knee). Ten degrees of knee flexion was chosen because it approximates the maximal knee extension angle during gait[29,30] and we believe it is a better position in which to assess stiffness. The foot was then placed in the ankle apparatus and positioned, by visual approximation, such that the point midway between the lateral and medial malleolus in the sagittal plane was aligned with the axis of rotation of the Kin-Com. The subject's foot, ankle, and thigh were secured with straps. To check for a discrepancy between the subject's ankle angle and the Kin-Com angle, the height of the dynamometer was adjusted until the subject's shank was parallel to the floor, and the Kin-Com foot-plate was placed in a vertical position (confirmed with a level), which produced a reading of 0 degrees. The subject's ankle angle was measured with a goniometer to verify a neutral dorsiflexion position (0 [degrees] [+ or -] 1 [degrees]).

Once proper joint alignment was achieved, the examiner passively moved the subject's ankle into dorsiflexion and plantar flexion approximately 10 times to allow the calf muscles to relax. The examiner determined the maximal dorsiflexion angle by the presence of a firm end-feel (the point at which firm resistance to motion was detected, even when additional force was applied). Subject complaints (eg, pain, severe stretching), increased EMG activity, and limb movement in the apparatus were monitored during the procedure. If increased EMG activity was viewed on the oscilloscope, the subject was instructed to relax, and the procedure was repeated. For 3 subjects (2 subjects with DM and PN and 1 subject in the comparison group), EMG activity could not be ablated at end-range. As a result, it was necessary to use a lesser maximal dorsiflexion angle ([bar]X=6.7 [degrees]) for these subjects. If any heel movement was detected in the apparatus, the subject's foot was repositioned and straps were adjusted. Once the maximal dorsiflexion angle was established, the subject's ankle was moved into plantar flexion as far as possible within the constraints of the apparatus, and this plantar-flexion angle became the starting position for the passive test. No specific maximal plantar-flexion angle was used, as all subjects were able to reach an angle sufficient to slacken the plantar flexor muscles. Intrarater reliability of maximal dorsiflexion angle measurements was established in a group of 10 subjects of mixed sex, age, activity level, and disease status (subjects with DM and PN versus subjects in the comparison group). The examiner, who was blinded to the Kin-Com readings, obtained 3 measurements of maximal dorsiflexion angle. Subjects were not removed from the apparatus, but a 30-second rest period was provided between measurements. The resulting ICC (2,1) value was .98.

The Kin-Com ankle apparatus moved the ankle joint from the starting position of maximal plantar flexion into maximal dorsiflexion (average range: 43.8 [degrees] for subjects with DM and PN and 53.4 [degrees] for subjects in the comparison group), while the subject's muscles remained passive. Three trials of torque and angle data were collected at a speed of 60 [degrees]/s, chosen because it approximates the average ankle joint velocity during gait.[16] In addition, pilot work indicated that passive torque curves at 60 [degrees]/s. did not appear different from curves generated at 5 [degrees]/s (ie, there was no evidence of a velocity effect on passive torque) (GB Salsich and colleagues, unpublished research). Similar findings were noted by Lamontagne et al.[31]

During the passive test, EMG signals were viewed on the oscilloscope to check for deviations from the subject's baseline signals, which were visually observed on the oscilloscope prior to ankle movement. In addition, the passive torque curves were viewed throughout the procedure to check the consistency of their shape (overlay), and a coefficient of variation (CV) of the whole curve (ensemble average)[29] was calculated for each subject, using the 3 recorded passive torque curves. The average CV was 0.09 (range=0.05-0.20) for subjects with DM and PN and 0.07 (range=0.3-0.16) for subjects in the comparison group. The consistency of the passive torque curves indicated that the subjects' muscles remained passive throughout the trials.

Data Reduction

For each subject, ASCII data from the 3 passive torque trials were imported into a spreadsheet for manipulation. The 3 passive torque trials were averaged, and the resulting torque versus angle curves were normalized by body mass and plotted. From the averaged passive curve, the following variables were obtained: maximal dorsiflexion angle, initial angle (defined as the angle at which passive torque is 0 N [multiplied by] m) (Fig. 1), plantar flexor muscle excursion (the difference between the maximal dorsiflexion angle and the initial angle), peak passive torque, passive torque at 5 degrees of dorsiflexion, and stiffness. Although 10 degrees of dorsiflexion has been reported to be necessary for normal gait[32] and, therefore, would have been a more "critical" angle at which to measure passive torque, 5 degrees of dorsiflexion was used because it was found to be the mean maximal dorsiflexion angle in a group of subjects with DM and PN.[23]

[Figure 1 ILLUSTRATION OMITTED]

Stiffness (slope of the passive torque curve) was calculated separately for the first and second halves of the passive torque curve, using the initial angle and maximal dorsiflexion angle to determine the total range (Fig. 1). Negative passive torque values were not considered in the slope calculations, as these values were attributed to torque from the dorsiflexor muscle group. Calculating 2 slopes provided a better means of characterizing the passive torque curve, due to its curvilinear nature. This method, similar to that used by Chleboun et al[33] to characterize passive elbow stiffness, was preferred to higher-order curve-fitting techniques, which can result in coefficients that are difficult to interpret.

Intrarater reliability of all passive torque variables was determined from the 3 passive torque trials. Subjects were not removed from the device, but there was approximately 30 seconds of rest between trials. The ICCs (2,1) were .97 (initial angle), .97 (peak passive torque), .97 (passive torque at 5 [degrees] of dorsiflexion), .87 (slope of the first half of the passive torque curve), and .95 (slope of the second half of the passive torque curve).

Data Analysis

To determine whether mean differences in passive torque variables occurred between the 2 subject groups, Student t tests were performed on the following variables: slope of the first half of the passive torque curve, slope of the second half of the passive torque curve, mean passive torque at 5 degrees of dorsiflexion, and mean peak passive torque. In addition, t tests were used to compare the mean values of maximal dorsiflexion angle (Kin-Com measurement), initial angle, and plantar flexor muscle excursion. The [Alpha] level was set at .007 for each test (.05/7), in order to protect against a Type I error.[34]

Results

Table 2 shows the mean ([+ or -] SD) ankle joint angle, passive torque, and passive stiffness values for both subject groups. As we hypothesized, the subjects with DM and PN demonstrated a decreased maximal dorsiflexion angle compared with the subjects in the comparison group (10.8 [degrees] [+ or -] 5.2 [degrees] versus 17.6 [degrees] [+ or -] 4.0 [degrees], P [is less than] .001). Despite this finding, there was no apparent "shift" in the passive torque curve toward plantar flexion, as there was no difference in the initial angle (angle at 0 N [multiplied by] m) between groups (subjects with DM and PN: -33.6 [degrees] [+ or -] 6.3 [degrees]; subjects in comparison group: -36.4 [degrees] [+ or -] 4.6 [degrees]). Plantar flexor muscle excursion (the difference between the initial angle and the maximal dorsiflexion angle) was less in the subjects with DM and PN compared with the subjects in the comparison group (43.8 [degrees] [+ or -] 9.7 [degrees] versus 53.4 [degrees] [+ or -] 5.7 [degrees], P=.001).

Table 2. Mean Angle and Passive Torque Values for Subjects With Diabetes Mellitus (DM) and Peripheral Neuropathy (PN) and Subjects in the Comparison Group

                              Subjects With DM and PN
                              (n=17)

Variable                      [bar]X      SD          Range

Maximal dorsiflexion
 angle ([degrees])             10.8       5.2        0 to 18.1
Initial angle
 ([degrees])(b)               -33.6       6.3     -44.8 to -20.2
Plantar flexor excursion
 ([degrees])(c)                43.8       9.7      26.8 to 56.0
Stiffness 1 (N
 [multiplied by] m/kg/
 degree)(d)                     0.002     0.001    0.0009 to 0.004
Stiffness 2 (N
 [multiplied by] m/kg/
 degree)(e)                     0.008     0.002    0.006 to 0.015
Positive torque at 5
 [degrees] of dorsiflexion
 (N [multiplied by] m/kg)       0.158     0.036    0.11 to 0.23
Peak passive torque (N
 [multiplied by] m/kg)          0.222     0.071    0.12 to 0.44

                               Comparison Group
                               (n=17)

Variable                       [bar]X      SD           Range
Maximal dorsiflexion
 angle ([degrees])              17.6       4.0       10.6 to 25.0
Initial angle
 ([degrees])(b)                -36.4       4.6      -43.8 to -28.1
Plantar flexor excursion
 ([degrees])(c)                 53.4       5.7       43.5 to 64.5
Stiffness 1 (N
 [multiplied by] m/kg/
 degree)(d)                      0.002     0.001    0.0006 to 0.004
Stiffness 2 (N
 [multiplied by] m/kg/
 degree)(e)                      0.009     0.002    0.005 to 0.013
Positive torque at 5
 [degrees] of dorsiflexion
 (N [multiplied by] m/kg)        0.149     0.036    0.08 to 0.20
Peak passive torque (N
 [multiplied by] m/kg)           0.303     0.071    0.16 to 0.40

Variable                         P

Maximal dorsiflexion
 angle ([degrees])               <.001(f)
Initial angle
 ([degrees])(b)                  .16
Plantar flexor excursion
 ([degrees])(c)                  .001(f)
Stiffness 1 (N
 [multiplied by] m/kg/
 degree)(d)                      .33
Stiffness 2 (N
 [multiplied by] m/kg/
 degree)(e)                      .17
Positive torque at 5
 [degrees] of dorsiflexion
 (N [multiplied by] m/kg)        .44
Peak passive torque (N
 [multiplied by] m/kg)           .002(f)

(a) Student t test, df=32.

(b) Angle at onset of positive passive torque (0 N [multiplied by] m).

(c) Maximal dorsiflexion angle minus initial angle (Kin-Com measurements).

(d) Slope of first half of passive torque versus angle curve.

(e) Slope of second half of passive torque versus angle curve.

(f) Significant group differences, [Alpha] =.007.

Another unexpected finding was that there was no evidence of increased stiffness in the subjects with DM and PN. The mean slope values for the first half of the torque curves were identical in both groups (0.002 [+ o r-] 0.001 N [multiplied by] m/ kg/degree), and the mean slope values for the second half of the torque curves were not notably different (subjects with DM and PN: 0.008 [+ or -] 0.002 N [multiplied by] m/kg/degree; subjects in comparison group: 0.009 [+ or -] 0.002 N [multiplied by] m/kg/degree). In addition, there was no difference in the amount of passive torque at 5 degrees of dorsiflexion (subjects with DM and PN: 0.158 [+ or -] 0.036 N [multiplied by] m/kg; subjects in comparison group: 0.149 [+ or -] 0.036 N [multiplied by] m/kg), but the subjects with DM and PN demonstrated less peak passive torque than did the subjects in the comparison group (0.222 [+ or -] 0.071 N [multiplied] m/kg versus 0.303 [+ or -] 0.071 N [multiplied by] m/kg; P=.002). Figure 2 shows a plot of the mean passive torque versus angle curves for both groups (ensemble averaged curve, n=17). We believe there were similarities in shape, slope, and variability between the curves of the subjects with DM and PN and the curves of the subjects in the comparison group.

[Figure 2 ILLUSTRATION OMITTED]

Because 3 subjects required positioning at a lesser maximal dorsiflexion angle than was available (due to increased EMG activity), the data analysis procedures were repeated without these subjects. Although the actual mean values for the stiffness and range of motion variables were slightly altered, the significant and non-significant variables remained the same. As a result, data from these 3 subjects were kept in the analysis.

Discussion

Ankle Angle Measurements

As we hypothesized, the subjects with DM and PN had less dorsiflexion range of motion than did the subjects in the comparison group, as reflected in the Kin-Com maximal dorsiflexion angle (10.8 [degrees] [+ or -] 5.2 [degrees] versus 17.6 [degrees] [+ or -] 4.0 [degrees], P [is less than] .001). This finding supports our finding of decreased dorsiflexion range of motion, as measured with a goniometer (Tab. 1).

There are several possible explanations as to why the subjects with DM and PN demonstrated a decreased maximal dorsiflexion angle compared with the subjects in the comparison group. Patients with DM and PN often have a decreased ability to generate muscle force (active torque production) associated with motor neuropathy,[7,21,22] and our subject population appeared to follow this pattern. Table 1 shows that the subjects with DM and PN had approximately 36% less concentric plantar flexor peak torque compared with subjects in the comparison group. In weight-bearing conditions such as the stance phase of gait, decreased plantar flexor muscle active torque production may result in a loss of stability, especially at greater dorsiflexion angles, where the body's center of mass is more anterior to the ankle joint. In order to maintain stability, a person might limit the amount of dorsiflexion, decreasing the force required by the plantar flexor muscles to maintain equilibrium. In addition, a loss of dorsiflexor muscle force may result in decreased dorsiflexion motion in non-weight-bearing conditions such as the swing phase of gait.

Another factor that may have contributed to the observed decreased dorsiflexion motion is the sensory loss associated with PN. Subjects with DM and PN have not only demonstrated loss of light touch and pressure-detection ability[23,25,35] and decreased vibratory sense,[24,25,35] they also have been shown to have diminished movement perception at the ankle joint.[25,35] In addition, each of these sensory deficits has been found to be related to decreased stability in subjects with DM and PN.[25] Although we measured only light touch and pressure, it is likely that our subjects with DM and PN had additional sensory deficits.

Lastly, the accumulation of AGEs in collagen may have played a role in limiting maximal dorsiflexion. Several authors[24,36,37] have described limited joint mobility in patients with diabetes and attributed this syndrome to changes in the ultrastructure of collagen in various periarticular tissues.

It is possible that all of these factors (ie, loss of plantar flexor force, loss of dorsiflexor force, diminished sensation, and accumulation of AGEs) lead to a limitation in the amount of dorsiflexion range of motion used by people with DM and PN. Over time, it is not unlikely that a limitation in dorsiflexion could result in the development of plantar flexion contractures. In this study, the initial angle (angle at 0 N [multiplied by] m) was not different in the 2 groups (subjects with DM and PN: -33.6 [degrees] [+ or -] 6.3 [degrees]; subjects in comparison group: -36.4 [degrees] [+ or -] 4.6 [degrees]), indicating that subjects with DM and PN did not have a complete shift in the passive torque curve toward plantar flexion. Instead, they appeared to have decreased excursion (length change capability) of the plantar flexor muscle group, as reflected in the difference between the initial angle and the maximal dorsiflexion angle between the 2 groups (subjects with DM and PN: 43.8 [degrees] [+ or -] 9.7 [degrees]; subjects in comparison group: 53.4 [degrees] [+ or -] 5.7 [degrees], P=0.001). Although the ankle joints of our subjects with DM and PN were not "immobilized," these subjects demonstrated a decreased maximal dorsiflexion angle, possibly from working in a restricted range of motion over time. Williams et al[38] found that hamster diaphragm muscles that were forced to work in a shortened range of motion demonstrated a loss of sarcomeres, similar to the loss of sarcomeres that occurs during cast immobilization of a muscle in a shortened position.

A puzzling result, however, was that there was less peak passive torque in the subjects with DM and PN than in the subjects in the comparison group (0.22 [+ or -] 0.07 N [multiplied by] m/kg versus 0.30 [+ or -] 0.07 N [multiplied by] m/kg; P=.002). If a plantar flexor contracture had developed, we would expect at least as much peak passive torque in this subject group, similar to the findings of Tardieu et al[39] in their study of the effects of immobilization on cat soleus muscles. There is some recent evidence, however, in support of our findings. Brown et al[40] reported a decrease in peak passive tension in the soleus and peroneus longus muscles of rats that had undergone a period of hind-limb unweighting (a model of reduced muscle use). In addition, these muscles had reduced excursion (length change from resting length). The authors reported that actual tearing of the unweighted muscles limited the amount of length change that could be induced.[40] Similarly, Gajdosik et al[41] reported decreased peak passive plantar flexor torque and decreased angular change (synonymous with our definition of plantar flexor excursion) in older women compared with younger women.

Our finding of less peak passive torque in subjects with DM and PN suggests that passive elastic stiffness was not the main factor determining the maximal dorsiflexion angle. As described in our "Method" section, we noted subject complaints, EMG activity, and lower-extremity movement during the maximal dorsiflexion angle measurement. Most of our subjects were limited by the sensation of end-range (excessive calf stretching). Three subjects, however, were limited by excessive EMG activity (2 subjects with DM and PN and 1 subject in the comparison group). Our belief was that the examiner's effort was consistent across all subjects such that there was no subjective examiner bias limiting dorsiflexion in the subjects with DM and PN. Instead, it appeared that the subjects with increased EMG activity were "protecting" their plantar flexor muscles from further lengthening.

Chesworth and Vandervoort[19] described similar phenomena in their study of stiffness in casted versus noncasted ankles. Their results showed that the casted ankles had less dorsiflexor motion and excessive plantar flexor EMG activity compared with the noncasted ankles. The authors interpreted these findings as evidence of a protective mechanism elicited by the nervous system to prevent the immobilized muscle from being injured as it was lengthened.

Stiffness Measurements

Contrary to our hypothesis, the subjects with DM and PN did not demonstrate greater passive stiffness in the plantar flexor muscle group than did the subjects in the comparison group, as measured by any of the passive torque variables (Tab. 2). These findings are similar to those of Gajdosik et al,[41] who studied the effect of age on passive plantar flexor stiffness in women. They reported no difference in stiffness over the first half of the range of motion between age groups (young: 0.26 N [multiplied by] m/ degree, middle-aged: 0.30 N [multiplied by] m/degree, old: 0.29 N [multiplied by] m/degree) and no difference in stiffness over the second half of the range of motion, except between the oldest age group and the youngest age group (young: 0.74 N [multiplied by] m/degree, middle-aged: 0.69 N [multiplied by] m/degree, old: 0.59 N [multiplied by] m/degree). Not only are our findings similar to those of Gajdosik et al, but our actual stiffness values are similar as well, considering that our groups contained male subjects. Had we used Gajdosik and colleagues' method of determining initial angle and had we not normalized torque by body mass, our stiffness 1 values would have been 0.33 N [multiplied by] m/degree for the subjects in the comparison group and 0.28 N [multiplied by] m/degree for the subjects with DM and PN and our stiffness 2 values would have been 0.92 N [multiplied by] m/degree for the subjects in the comparison group and 0.79 N [multiplied by] m/degree for the subjects with DM and PN.

One possible explanation for our finding of no difference in stiffness is that the changes in collagen that have been associated with diabetes might not affect the elastic stiffness of the plantar flexor muscle group. Several researchers[42,43] have suggested that structures containing collagen within the muscle tendon unit (perimysium, endomysium) contribute to passive stiffness mostly at end-range (long sarcomere lengths) and that, within the physiological range of muscle length change, passive stiffness can be attributed to structures within the myofibril (eg, structural proteins such as titin). Thus, changes in collagen ultra-structure (from AGEs) would most likely have little effect on passive stiffness measured throughout the range of motion, as was done in our study.

Furthermore, if the major sources of passive tension in skeletal muscle were myofibrillar structures, passive stiffness would be directly proportional to the amount of these structures present. Muscle size, therefore, would be positively correlated with passive stiffness, and recent research has provided evidence of this. Chleboun et al[33] examined passive stiffness in the elbow flexor muscles and found a positive correlation between elbow flexor muscle volume and passive stiffness (r =.92). Given this finding, it would be likely that muscle atrophy would be associated with a decrease in passive stiffness, and subjects with DM and PN have been shown to have decreased muscle cross-sectional area associated with muscle atrophy. Andersen et al[22] found a 43% decrease in the cross-sectional area of the plantar flexor muscle group that corresponded to a 45% reduction in muscle force in subjects with DM and PN compared with control subjects. Consequently, the combination of muscle atrophy (which would decrease passive stiffness) and collagen cross-linking (which would increase passive stiffness) may have resulted in no net change in passive stiffness in the plantar flexor muscles of the subjects with DM and PN.

Implications and Limitations of the Study

The results of our study indicate that in subjects with DM and PN, decreased dorsiflexion range of motion occurred without a change in plantar flexor muscle stiffness. In light of this finding, clinicians should be cautious when using the terms "stiffness" and "decreased range of motion" interchangeably. Patients with DM and PN often describe joint stiffness, as do patients with a wide variety of diseases and injuries, and clinicians typically believe their interventions will decrease stiffness. The results of our study suggest that what appears to be muscle stiffness may actually be decreased joint motion due to loss of muscle excursion (shortness). Therefore, treatment techniques such as muscle stretching, although possibly increasing range of motion, most likely will have no effect on passive muscle stiffness. The findings of Halbertsma et al[44] support this clinical implication. The authors reported no change in passive hamstring muscle stiffness after a stretching protocol, despite an increase in hip flexion range of motion. Understanding treatment rationale is critical to effective patient management.

Although the results of our study provide clinical implications, there are several limitations in the study design. One limitation was in the subject groups chosen for analysis. Although the subjects with DM and PN had less dorsiflexion range of motion than the subjects in the comparison group, the groups did not differ in stiffness. Because we did not study subjects with DM who did not have PN, we cannot dissociate the effect of the disease from the effect of neuropathy. It was not the purpose of our study, however, to parse out the effects of specific components of diabetes on muscle stiffness, but rather to determine whether differences in stiffness could be detected in subjects with known pathology and impairments (eg, decreased range of motion, decreased muscle force) compared with a comparison group.

Ankle joint stiffness is not synonymous with plantar flexor muscle stiffness. Although it is possible that tissues such as joint capsule, skin, ligaments, and cartilage contribute to passive ankle torque, we believe their contributions are minor throughout the range of motion used in our study. Gillette and Fell[45] examined the contributions to passive ankle tension from muscle, skin, and joint tissues in rats after a period of hind-limb suspension. The authors attributed 75% of passive ankle joint tension (present at 45 [degrees] of dorsiflexion) to the plantar flexor muscle tendon units and 25% to joint structures.

Another limitation involves the number of subjects we studied (n=17 per group). The lack of difference in stiffness between groups may have been a result of the sample size, and increasing the number of subjects may have made it easier to detect differences (ie, greater statistical power). However, we believe the similarities of the passive torque curves in both groups indicate that lack of power was not the issue. The mean passive plantar flexor torque versus ankle joint angle curve for the subjects with DM and PN was nearly identical to that of the comparison group (Fig. 2).

Further studies are needed to determine whether passive stiffness is altered in other patient populations and to determine whether passive stiffness relates to other clinical measures (active muscle tension and joint range of motion). In addition, studies are needed to determine whether passive stiffness affects functional activities such as gait. Knowledge of these relationships would provide insight into mechanisms used by the body to compensate for specific tension or range of motion deficits as well as how specific impairments affect function. Knowledge of these relationships also would provide insight into treatment implications for patients with various diseases or disuse conditions.

Conclusions

The results of this study indicate that, compared with individuals without known pathology, individuals with DM and PN have decreased dorsiflexion range of motion and decreased plantar flexor muscle excursion in the absence of increased passive plantar flexor muscle stiffness. Clinicians should exercise caution when using terminology to describe joint or muscle stiffness. Further studies are needed to determine the relationships among passive muscle stiffness, active muscle tension, and joint motion, and how these relationships influence gait characteristics.

(*) Therapeutics Unlimited Inc, 2835 Friendship St, Iowa City, IA 52240.

([dagger]) Chattanooga Group Inc, 4717 Adams Rd, PO Box 489, Hixson, TN 37343.

References

[1] Beer FP, Johnston ER. Vector Mechanics for Engineers: Statics. 5th ed. New York, NY: McGraw-Hill Book Co; 1988.

[2] Hill AV. The heat of shortening and the dynamic constants of muscle. Proc R Soc Lond B. 1938;126:136-195.

[3] Zajac FE. Muscle and tendon: properties, models, scaling, and application to biomechanics and motor control. Crit Rev Biomed Eng. 1989;17:359-411.

[4] Cole GK, van den Bogert AJ, Herzog W, Gerritsen KGM. Modeling of force production in skeletal muscle undergoing stretch. J Biomech. 1996;29:1091-1104.

[5] Bassey EJ, Bendall MJ, Pearson M. Muscle strength in the triceps surae and objectively measured customary walking activity in men and women over 65 years of age. Clin Sci (Colch). 1988;74:85-89.

[6] Salsich GB, Mueller MJ. Relationships between measures of function, strength and walking speed in patients with diabetes and transmetatarsal amputation. Clin Rehabil. 1997;11:60- 67.

[7] Mueller MJ, Minor SD, Schaaf JA, et al. Relationship of plantarflexor peak torque and dorsiflexion range of motion to kinetic variables during walking. Phys Ther. 1995;75:684-693.

[8] Brown M, Sinacore DR, Host HH. The relationship of strength to function in the older adult. J Gerontol A Biol Sci Med Sci. 1995;50:55-59.

[9] Nakamura R, Watanabe S, Handa T, Morohashi I. The relationship between walking speed and muscle strength for knee extension in hemiparetic stroke patients: a follow-up study. Tohoku J Exp Med. 1988;154:111-113.

[10] Powers CM, Boyd LA, Fontaine CA, Perry J. The influence of lower-extremity muscle force on gait characteristics in individuals with below-knee amputations secondary to vascular disease. Phys Ther. 1996; 76:369 -377.

[11] Potter JM, Evans AL, Duncan G. Gait speed and activities of daily living function in geriatric patients. Arch Phys Med Rehabil. 1995;76: 997-999.

[12] Ensrud KE, Nevitt MC, Yunis C, et al. Correlates of impaired function in older women. J Am Geriatr Soc. 1994;42:481-489.

[13] Perry J, Mulroy SJ, Renwick SE. The relationship of lower extremity strength and gait parameters in patients with post-polio syndrome. Arch Phys Med Rehabil. 1993;74:165-169.

[14] Siegler S, Moskowitz GD, Freedman W. Passive and active components of the internal moment developed about the ankle joint during human ambulation. J Biomech. 1984;17:647-652.

[15] Tardieu C, Lespargot A, Tabary C, Bret MD. Toe-walking in children with cerebral palsy: contributions of contracture and excessive contraction of triceps surae muscle. Phys Ther. 1989;69:656-662.

[16] Winter DA. Biomechanics and Motor Control of Human Movement. 2nd ed. New York, NY: John Wiley & Sons Inc; 1990.

[17] Gajdosik RL, Vander Linden DW, Williams AK. Influence of age on concentric isokinetic torque and passive extensibility variables of the calf muscles of women. Eur J Appl Physiol. 1996;74:279-286.

[18] Sinkjaer T, Magnussen I. Passive, intrinsic and reflex-mediated stiffness in the ankle extensors of hemiparetic patients. Brain. 1994; 117 (pt 2) :355-363.

[19] Chesworth BM, Vandervoort AA. Comparison of passive stiffness variables and range of motion in uninvolved and involved ankle joints of patients following ankle fractures. Phys Ther. 1995;75:253-261.

[20] Mueller MJ, Minor SD, Sahrmann SA, et al. Differences in the gait characteristics of patients with diabetes and peripheral neuropathy compared with age-matched controls. Phys Ther. 1994;74:299-313.

[21] Andersen H, Poulsen PL, Mogensen CE, Jakobsen J. Isokinetic muscle strength in long-term IDDM patients in relation to diabetic complications. Diabetes. 1996;45:440-445.

[22] Andersen H, Gadeberg PC, Brock B, Jakobsen J. Muscular atrophy in diabetic neuropathy: a stereological magnetic resonance imaging study. Diabetologia. 1997;40:1062-1069.

[23] Mueller MJ, Diamond JE, Delitto A, Sinacore DR. Insensitivity, limited joint mobility, and plantar ulcers in patients with diabetes mellitus. Phys Ther. 1989;69:453-462.

[24] Andersen H, Mogensen PH. Disordered mobility of large joints in association with neuropathy in patients with long-standing insulin-dependent diabetes mellitus. Diabet Med. 1997;14:221-227.

[25] Simoneau GG, Ulbrecht JS, Derr JA, et al. Postural instability in patients with diabetic sensory neuropathy. Diabetes Care. 1994;17: 1411-1421.

[26] Vlassara H. Advanced nonenzymatic tissue glycosylation: cell-mediated interactions implicated in the complications associated with diabetes and aging. Blood Purif. 1990;8:223-232.

[27] Brand PW. The diabetic foot. In: Ellenberg M, Rifkin H, eds. Diabetes Mellitus: Theory and Practice. 3rd ed. New Hyde Park, NY: Medical Examination Publishing Co Inc; 1983:829-849.

[28] Birke JA, Sims DS Jr, Buford WL. Walking casts: effect on plantar foot pressures. J Rehabil Res Dev. 1985;22:18-22.

[29] Winter DA. Kinematic and kinetic patterns in human gait: variability and compensating effects. Human Movement Science. 1984;3:51-76.

[30] Winter DA, Patla AE, Frank JS, Walt SE. Biomechanical walking pattern changes in the fit and healthy elderly. Phys Ther. 1990;70: 340 -347.

[31] Lamontagne A, Malouin F, Richards CL. Viscoelastic behavior of plantar flexor muscle-tendon unit at rest. J Orthop Sports Phys Ther. 1997;26:244-252.

[32] Perry J. Gait Analysis: Normal and Pathological Function. Thorofare, NJ: Slack Inc; 1992.

[33] Chleboun GS, Howell JN, Conaster RR, Giesey JJ. The relationship between elbow flexor volume and angular stiffness at the elbow. Clinical Biomechanics. 1997; 12:383-392.

[34] Hays WL. Statistics. 5th ed. Fort Worth, Tex: Harcourt Brace College Publishers; 1994.

[35] Simoneau GG, Derr JA, Ulbrecht JS, et al. Diabetic sensory neuropathy effect on ankle joint movement perception. Arch Phys Med Rehabil. 1996;77:453-460.

[36] Schulte L, Roberts MS, Zimmerman C, et al. A quantitative assessment of limited joint mobility in patients with diabetes: goniometric analysis of upper extremity passive range of motion. Arthritis Rheum. 1993;36:1429 -1443.

[37] Starkman HS, Gleason RE, Rand LI, et al. Limited joint mobility (LJM) of the hand in patients with diabetes mellitus: relation to chronic complications. Ann Rheum Dis. 1986;45:130-135.

[38] Williams PE, Catanese T, Lucey EG, Goldspink G. The importance of stretch and contractile activity in the prevention of connective tissue accumulation in muscle. J Anat. 1988;158:109-114.

[39] Tardieu C, Tabary JC, 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.

[40] Brown M, Fisher JS, Salsich GB. Stiffness and muscle function with age and reduced muscle use. J Orthop Res. 1999;17:409-414.

[41] Gajdosik RL, Vander Linden DW, Williams AK. Influence of age on length and passive elastic stiffness characteristics of the calf muscle-tendon unit of women. Phys Ther. 1999;79:827-838.

[42] Magid A, Law DJ. Myofibrils bear most of the resting tension in frog skeletal muscle. Science. 1985;230:1280-1282.

[43] Wang K, McCarter R, Wright J, et al. Regulation of skeletal muscle stiffness and elasticity by titin isoforms: a test of the segmental extension model of resting tension. Proc Natl Acad Sci U S A. 1991;88: 7101-7105.

[44] Halbertsma JP, van Bolhuis AI, Goeken LN. Sport stretching: effect on passive muscle stiffness of short hamstrings. Arch Phys Med Rehabil. 1996;77:688-692.

[45] Gillette PD, Fell RD. Passive tension in rat hindlimb during suspension unloading and recovery: muscle/joint contributions. J Appl Physiol. 1996;81:724-730.

GB Salsich, PT, PhD, is Postdoctoral Fellow, Department of Biokinesiology and Physical Therapy, University of Southern California, 1540 E Alcazar St, CHP-155, Los Angeles, CA 90089 (USA) ([e-mail]). Address all correspondence to Dr Salsich.

MJ Mueller, PT, PhD, is Associate Professor, Program in Physical Therapy, Washington University School of Medicine, St Louis, Mo.

SA Sahrmann, PT, PhD, FAPTA, is Professor and Associate Director for Doctoral Studies, Program in Physical Therapy, Washington University School of Medicine.

All authors provided concept/research design, fund procurement, and consultation (including review of manuscript before submission). Dr Salsich and Dr Mueller provided project management and subjects. Dr Salsich provided writing and data collection and analysis. Dr. Mueller provided facilities/equipment and institutional liaisons. Richard Gajdosik, PT, PhD, assisted with the development of the plantar flexor stiffness measurement methods. Scott D Minor, PT, PhD, contributed to study design and methods. Mary Hastings, PT, MHS, assisted with data collection and analysis.

This study was completed in partial fulfillment of the requirements for Dr Salsich's doctoral degree in the Interdisciplinary Program in Movement Science, Washington University School of Medicine.

This study was approved by the Washington University School of Medicine Human Studies Committee.

This study was supported by a postprofessional doctoral scholarship from the American Physical Therapy Association and by National Center for Medical Rehabilitation Research grants 2T32HD07434-04A1 and 1RO1HD36802-01 (Dr Mueller).

This article was submitted March 15, 1999, and was accepted January 3, 2000.