Differences in the gait characteristics of patients with diabetes and peripheral neuropathy compared with age-matched controls.

Patients with diabetes and peripheral neuropathy have a high incidence of injuries during walking[1-3] and a low level of perceived safety.[1] Cavanagh et al[1] report that subjects in a group of patients with diabetes and peripheral neuropathy were 15 times more likely to report an injury (fall, fracture, sprained ankle, or cuts and bruises) during walking or standing than subjects in a control group of patients with diabetes but no peripheral neuropathy. in addition, peripheral neuropathy is a risk factor for developing plantar ulcers.[2,3] Most of these ulcers are thought to develop during walking.[2,3] Analysis of the walking patterns of patients with diabetes and peripheral neuropathy may lead to a better understanding of the mechanics of gait and to treatments to reduce the number of injuries in this patient population.

There are several methods to study biomechanical aspects of walking in these patients. Research investigating human gait has traditionally focused on kinematic variables (eg, joint angles, stride length, walking velocity), a description of motion without regard to forces causing movement. In order to include the forces that cause movement, investigators need a kinetic model.[4,5] Understanding kinetic variables, such as forces, moments, and power, may enhance the ability of investigators to detect and alter movement patterns.[5,6]

The link-segment biomechanical model can be used to estimate kinetic variables of human movement.[7,8] The model's basic assumption is that the body is composed of a system of rigid links connected by smooth pin joints. Given simultaneous input of kinematic variables, ground reaction forces, and anthropometric data, the link-segment model has been used to estimate various kinetic variables acting at lower-extremity joints during gait including joint reaction forces, moments, and power.[8,9]

The kinematic pattern of joint motion is predictable for normal walking.[8,10] The moments that influence these motions, however, may be variable, depending on the strategy used to achieve the kinematic result. During the stance phase of walking, only movement about the ankle has a consistently recognizable moment pattern.[4,7] Investigators[4,7] have described a small dorsiflexor moment followed by a large plantar-flexor moment. The moment pattern at the hip is less consistent than the moment pattern at the ankle, but usually shows an extensor moment followed by a flexor moment. The moment pattern at the knee is the least consistent lower-extremity joint moment.4 This inconsistency is expected because moments at the ankle and hip can control movement at the knee when the foot is fixed on the ground.[4,5]

Despite inconsistency of the moment patterns at the hip and the knee, the summation of the ankle plantar-flexor moment, the knee extensor moment, and the hip extensor moment is relatively consistent and has been termed the support moment.[4,8] This support moment has a shape similar to the vertical ground reaction force during stance. The contribution from each joint may vary, but the overall pattern remains similar, Winter[6] believes this joint variation is a reflection of indeterminacy in the central nervous system, providing considerable flexibility during walking.

Power is the kinetic variable that reflects the rate of work performed at a given joint. It can be calculated by taking the product of the moment and the joint angular velocity.[7,8] Power will be positive when the moment and angular velocity have the same direction. A positive power usually indicates that the muscle is generating mechanical energy (concentric contraction). A negative power usually indicates that the muscle is absorbing mechanical energy (eccentric contraction).[8] Although muscle power varies during walking, the power generated by the plantar flexors at push-off is typically the largest power burst recorded during the gait cycle. In normal walking, this power burst is critical for propelling the leg and body forward for the next step.[5]

The patient with diabetes mellitus (DM) and peripheral neuropathy offers a good clinical model for investigating the strategies that patients develop to compensate for the effects of decreased ankle function. Limited joint mobility has been documented in the ankle and feet of patients with DM.[2,11] Peripheral neuropathy and inactivity can also lead to a loss of function in foot and ankle musculature,[12] although quantitative descriptions of these motor deficits are lacking. If patients with peripheral neuropathy are unable to generate sufficient moments about the ankle during walking, we would expect them to take shorter steps and walk more slowly than individuals without peripheral neuropathy.

The purpose of this study was to compare (1) the gait characteristics, (2) the plantar-flexor peak torques, and (3) the ankle range of motion (ROM) of subjects with diabetes and peripheral neuropathy with those of age-matched controls. Compared with age-matched controls, we hypothesized that subjects with DM and peripheral neuropathy would show the following specific differences: (1) decreased ankle motion, ankle moment, ankle power, walking velocity, and step length during walking and (2) decreased plantar-flexor peak torque and ankle ROM measured non-weight bearing.

Method

Subjects

Patients of the Diabetic Foot Center and the Irene Walter Johnson Institute of Rehabilitation at Barnes Hospital, St Louis, Mo, were recruited for the study. Criteria for patient selection were that the patient had been diagnosed with DM, have a history of a neuropathic ulcer, be able to walk independently without pain or an assistive device (eg, walker or crutches), and have no history of a transmetatarsal amputation. Patients with a transmetatarsal amputation were excluded because this amputation may cause changes in walking patterns. History of a neuropathic ulcer was the clinical method used to document existence of severe peripheral neuropathy.[12] Age-matched nondiabetic employees and outpatients in well elderly and cardiac rehabilitation programs at Washington University were recruited to serve as a control group. Eligibility criteria for nondiabetic subjects were no history of diabetes, absence of significant neurological diseases and orthopedic problems, and ability to ambulate independently without pain or the use of an assistive device. All screening and data collection were conducted by the primary investigator (MJM)

Table 1 provides a description of the subjects. Ten patients (8 men, 2 women) with DM and peripheral neuropathy and a mean age of 57.7 years (SD=14.5, range=35-75) were designated the DM group. Ten subjects (8 men, 2 women) with no history of diabetes and a mean age of 56.8 years (SD=11.3, range=37-68) constituted the NODM group. One subject in the DM group required a polypropylene ankle-foot orthosis to ambulate. The mean duration of diabetes in the DM group was 20.8 years (SD=6.09, range=14-35). There was no significant difference (P>.05) in age, weight, height (t test), or sex (chi square) between the groups.

[TABULAR DATA 1 OMITTED]

Procedure

The procedures of the study were explained to all subjects. All subjects signed an approved informed consent statement prior to testing, A brief history was taken, and demographic data were recorded including birth date, sex, and significant medical problems, Patients were questioned regarding history of diabetes, type of diabetes (non-insulin-dependent or insulin-dependent DM), reported duration of diabetes, and history of plantar ulceration.

All measurements were taken a minimum of two times during the same testing session to determine test-retest reliability. Knowledge of results was not available to the tester, except for ankle ROM and plantar-flexion peak torque. All intraclass correlation coefficients ICC[2,K])[13] of measures used in this study were [greater than or equal to].83 (Tab. 2). The fixed-effects model was chosen because two measures were taken by one tester, with results to be used for the purposes of this study only.

To measure ankle ROM, patients were positioned prone on a firm treatment table, with the tested foot and ankle extended over the end. A standard plastic goniometer with the scale marked in 2-degree increments was used. The axis of the goniometer was placed over the lateral malleolus; one arm of the goniometer was aligned with the fifth metatarsal head, and the other arm was aligned with the head of the fibula.[14] The tester passively dorsiflexed and plantar flexed the foot at the ankle, and the subject was then asked to actively move the foot at the ankle. The greater angle of the active or passive motion was used. The plantar-flexion and dorsiflexion angles were added to yield a measure of ankle motion. The intrarater ICC of .97 was similar to the value of .96 determined in a previous study assessing reliability of ankle goniometric measures in a group of patients with diabetes.[14]

Plantar-flexor peak torque was measured next. Subjects were placed supine on a LIDO[R] Active isokinetic table* with their knees extended and their right foot secured to the foot plate by two VELCRO[R]straps.[dagger] The axis of the dynamometer was aligned with the axis of the ankle joint (at approximately the level of the distal lateral malleolus). To stabilize the leg further during testing, a strap was secured over the right thigh. Subjects were tested at a speed of 60[degrees]/s. This speed was chosen because 60[degrees]/s is a comparable angular speed observed at the ankle joint during the stance phase of walking.[8] Subjects were allowed several submaximal practice contractions to become acquainted with the resistance and speed of movement. To measure plantar-flexor peak torque, the foot was placed in a position of maximum dorsiflexion and the patient was instructed to push as hard and as fast as possible through the full ROM. This measurement was repeated a total of four times. The highest value was taken as the plantar-flexor peak torque.[15]

Table 2. Reliability Coefficients
([ICC.sup.a]) for Intrarater Agreement for
Variables Used in This Study
Measure                            Icc
Ankle [ROM.sup.b] (goniometer)     .97
Plantar-flexor peak torque         .99
Walking velocity                   .98
Cycle time                         .95
Cadence                            .95
Stride length                      .98
Ankle ROM (video)                  .95
Ankle moment                       .84
Knee moment                        .83
Hip moment                         .93
[GRF,.sup.c] medial-lateral        .87
GRF, anterior-posterior            .98
GRF, vertical                      .94
Ankle power                        .84

[a]ICC=intraclass correlation coefficient (equation 2,k).[13] [b]ROM=range of motion. [c]=ground reaction force.

Gait Testing

Subjects wore shorts and shoes with heels of 2.54 cm (1 in) or less. Retroreflective markers were placed on the skin over the head of the fifth metatarsal, the ankle (distal lateral malleolus), the knee at the lateral femoral condyle, the greater trochanter of the hip, and the mid-superior iliac crest directly superior to the greater trochanter of the hip. A reflective ball (1 cm in diameter) was placed on the walkway to mark the x axis of the force platform.

Footswitches were used to determine initial contact (heel-strike) and toe-off during ambulation and to verify synchronization of video data and force platform data. The thin on-off foot-switches were secured to the plantar surface of the forefoot (at the first metatarsal head) and the rear foot (at the posterior lateral heel) bilaterally with double-sided adhesive tape and wrapped with paper tape. Foot-switches were connected via a thin, flexible cable to a light-emitting diode (LED) box and the interface box. The LED box was placed below the walkway in the camera's field of view.

The ExpertVision system[double dagger] was used for video analysis and consisted of the following components: a video camera[sections] and lens, a spotlight, a video monitor, a proprietary video processor (VP-110),[double dagger] and a videocassette recorder (VCR).[parallel] The video camera, sampling at 60 fields per second, and the attached spotlight were placed 6 m to the side of the walkway perpendicular to the x axis of the force platform. The raw coordinate data were smoothed using a fourth-order Butterworth filter with a cutoff at 6 Hz.[16] The effective frequency was 60 Hz.

Data were collected as subjects walked across a 6.8-m elevated walkway. The 60- x 120-cm AMTI LG6-2-1 force platform,[#] sampling at 60 Hz, was embedded at floor level midway down the walkway. Subjects were allowed to practice walking on the walkway with the footswitches secured to their feet. They were instructed to walk normally toward the opposite end of the walkway. Once the subjects were accustomed to walking on the walk-way, data were collected. Ten trials were recorded for each subject. Subjects were allowed to rest as needed.

Data Analysis

A standard two-dimensional link-segment model of the lower limb was developed for the foot, shank, and thigh segments. Data from kinematic, ground reaction force, and anthropometric variables were combined with standard link-segment equations to calculate moments of force at the ankle, knee, and hip.[8] Power at each joint was calculated by determining the product of the moment and the angular velocity.[8] Values were calculated for every 0.0167 second (60-Hz sampling frequency) of the stance phase. Analysis was conducted on the peak values, which occurred during terminal stance (70%-100% of stance). Terminal stance was chosen because we believed ankle strength and mobility would have their greatest effect during this portion of the gait cycle. For each kinematic and kinetic variable of walking, a mean of three trials was used to enhance reliability.[17] As shown in Table 2, ICC values for all measures ranged from .83 to .99, demonstrating high reliability. Kinetic variables were divided by body mass to reduce intersubject variability.[8,9] Independent t tests were performed to determine differences between the DM and NODM groups for each variable. The alpha level was set at .05 for a priori hypotheses (ie, group differences expected in ankle motion, ankle moment, ankle power, walking velocity, and stride length measured during walking, and plantar-flexor peak torque and ankle motion measured non-weight bearing).[18] To protect against a Type I error, a Bonferroni correction factor was used for all remaining variables tested for significant differences (Tab. 3).[18]

Results

Angle position, angle velocity, moments, and power at the ankle, knee, and hip were plotted against the normalized percentage of stance phase for each subject. Zero percent represented heel-strike, and 100% represented toe-off. A series of graphs illustrating a representative DM group subject compared with a representative NODM group subject are shown in Figures 1 through 4. The graphs illustrate the mean ([plus or minus]1 standard deviation) of three trials for the stance phase.

Table 3 contains a summary of the comparison of multiple variables between groups. During walking, the DM group showed less ankle joint motion (t=3.95, P=.001), lower peak ankle moment (t=2.33, P=.032), lower peak ankle power (t=2.70, P=.015), slower walking velocity (t=2.35, P=.031), and shorter stride length (t=3.34, P=.004) than the NODM group, In non-weight-bearing measures, the DM group showed less ankle ROM (t=4.34, P<.001), and plantar-flexor peak torque (t=4.02, P=.001) than the NODM group. These results are consistent with our hypotheses. The DM group also showed lower ground reaction forces in the anterior-posterior direction (t=2.70, P=.001) than the NODM group.

[TABULAR DATA 3 OMITTED]

Discussion

Compared with NODM group subjects, DM group subjects showed slower walking velocity, shorter stride length, less ankle motion, lower peak ankle moments, and lower peak ankle power (Tab. 3). We believe these changes were a result of the significantly decreased plantar-flexor peak torque and ankle mobility in the DM group compared with the NODM group (Tab. 3). Decreased plantar-flexor strength appeared to diminish the ability of the plantar-flexor muscles to push off and generate plantar-flexor moments or power during terminal stance (ie, 70%-100% of normalized stance). Decreased push-off resulted in shorter steps and a slower walking velocity in the DM group subjects compared with the NODM group subjects. These findings are consistent with our hypotheses.

Consistent with other reports on normal walking,[4,7,8] most NODM group subjects showed peak moments and power at the ankle during terminal stance phase (ie, 70%-100%), as shown in Figures 3 and 4. There was a significant decrease in the peak ankle moment and power in the DM group compared with the NODM group, but no difference for peak hip moments, knee moments, or power (Tab. 3). This finding is reasonable, because most of the musculoskeletal limitations as a result of peripheral neuropathy are in the foot and ankle. As demonstrated by the moment and power profiles in Figures 3 and 4, many DM group subjects had greater hip moments and power than ankle moments and power during terminal stance, opposite to the pattern observed in the NODM group. These findings suggest that the DM group subjects pulled their leg forward using hip flexor muscles (hip strategy), rather than pushing the leg and superincumbent body weight forward with plantar-flexor muscles (ankle strategy).

Although joint moments and power can be assessed only using kinetic analysis, the ankle and hip strategies can be described using kinematic data. As demonstrated in Figures 1 and 2, the DM group subject using a hip strategy shows less plantar-flexion angular excursion and angular velocity at terminal stance than the NODM group subject using an ankle strategy. The DM group subject appears to walk flat-footed, with minimal push-off into plantar flexion. In addition, the hip strategy uses hip flexion earlier in the gait cycle compared with the ankle strategy (Fig. 1, 82% versus 90%). Earlier hip flexion is required because the subject must pull the leg forward with hip flexor muscles rather than pushing the leg forward using plantar-flexor muscles. Other kinematic observations of a hip strategy include a shorter step length and a slower walking velocity (Tab. 3).

The hip and ankle strategies seen in the DM and NODM group subjects during walking are similar in concept to those proposed by Horak et al[19] for postural control. These authors describe hip and ankle strategies that are used to maintain upright stance following perturbations. A hip strategy demonstrates primary movement at the hip and is normally used in response to large, fast perturbations and when the support surface is compliant or smaller than the feet. An ankle strategy is normally used in response to slow, small perturbations on a firm, wide surface capable of resisting ankle moments. The hip and ankle strategies seen in the DM and NODM group subjects of this study during walking may be a dynamic representation of these types of strategies. Further research is needed to determine the relationship between these strategies in standing and walking.

Stein et al[20] note, however, that a movement performed to accomplish a specific goal (ie, task) can be composed of a combination, or "blend," of several strategies, There are many tasks that use a combination of movement strategies. In this study, one subtask was to advance the lower extremity as part of the larger task of walking. A subject could (1) pull the leg forward using a hip strategy, (2) push the leg forward using an ankle strategy, or (3) push and pull the leg forward using a blend of the two strategies. Many NODM group subjects appeared to use predominantly an ankle strategy. The DM group subjects with a moderate amount of ankle weakness, and some NODM group subjects, used a blend of the two strategies. The DM group subjects with profound ankle weakness (plantar-flexor peak torque = < 50 N[multiplied by]m) seemed to be limited to the hip strategy. Perhaps this limitation in available movement strategies contribute to their reported high incidence of falls and injuries during walking.

Although not explicitly stated, several studies allude to the notion of a hip strategy used by various patient populations. Olney et al[9] describe how children with hemiplegic cerebral palsy develop greater power at the hip than at the ankle. Some studies[21,22] have documented the rather obvious need for patients with lower-extremity amputations fitted with prostheses to generate greater moments and power at the hip than at the ankle. Marshall and Nade[23] reported greater-than-normal joint moments at the knee and hip to compensate for arthrodesis of the subtalar joint. Winter et al[16] reported that elderly subjects showed decreased walking velocity, decreased stride length, and a trend toward decreased ankle power at push-off compared with younger subjects. These changes are consistent with a shift from an ankle strategy to a hip strategy.

Although the normal walking pattern used by healthy young adults appears to be predominantly an ankle strategy, they can temporarily adopt a hip strategy. An example of a healthy young subject using a hip strategy would include the person walking across ice. When walking on ice, the subject is unable to generate significant forces in the anterior-posterior direction. Because of the absence of friction, a person is unable to push off using the plantar-flexor muscles and must temporarily adopt the safer, less destabilizing hip strategy. Another example is the person walking in very stiff boots, such as those used for skiing. In this scenario, subjects are unable to generate plantar-flexor power because of limited ankle mobility and must rely on their hip musculature to advance the leg.

We believe the differences in gait characteristics between the DM and NODM groups in this study were a result of decreased strength and mobility at the ankle in the DM group. The differences cannot be attributed to age, sex, weight, or height because there were no significant differences between the groups for these variables (Tab. 1). Although walking velocity,[24,25] sensory loss, or other complications of DM can have an effect on kinematic and kinetic variables, we do not believe these variables were the primary factors influencing gait characteristics in this study. This study was part of a larger project that analyzed the contribution of multiple subject characteristics to kinetic variables.[26] Hierarchical, multiple regression analyses on all 20 subjects were performed, with peak ankle moments and peak ankle power as dependent variables and plantar-flexor peak torque, dorsiflexion ROM, group (ie, DM or NODM), and walking velocity entered as independent variables. Although plantar-flexor peak torque and dorsiflexion ROM made a significant contribution when entered into the regression equation (P<.01), walking velocity and group added no unique variance (P>.05).[26] These findings support the hypothesis that ankle plantar-flexor peak torque (strength) and mobility, rather than walking velocity, sensory loss, or other complications of diabetes, were the primary factors contributing to the changes in gait characteristics of the DM group. Additional research, with specific measures of sensation, is needed to determine how this complication of diabetes affects posture and movement.

Clinical Implications

As noted previously, there is a high incidence of falls, fractures, and neuropathic ulcers during walking in patients with diabetes and peripheral neuropathy.[1-3] Based on the results of this study, speculation can be made regarding appropriate treatment to reduce these complications.

This group of patients with a long duration of disease (bar X=21 years) showed significantly less strength than their age-matched controls (Tab. 3). Increasing plantar-flexor strength via progressive resistance exercise may increase plantar-flexor ankle moments, and thus stability, during walking. Although the ability to increase the strength of the aged is documented,[27] increasing the strength of muscle with severe peripheral neuropathy is doubtful.

Because peripheral neuropathy affects primarily the distal musculature (ie, the foot and ankle), the most advantageous exercise might focus on the hip and knee. Perhaps patients could be trained to use their hip and knee muscles more efficiently to compensate for the decreased function at the ankle. Such a strategy would seem to emphasize hip extensor and knee flexor moments at early stance phase and emphasize hip flexor and knee extensor moments at late stance phase (ie, push-off). Theoretically, this strategy would supply necessary stability throughout stance and necessary mobility at the end of stance to pull the lower extremity forward (from the hip flexor moment). Additional research is needed to test the hypothesis that this type of treatment strategy will enhance stability and reduce the number of injuries during walking.

The results of this research suggest that emphasizing a hip strategy for patients with peripheral neuropathy may place lower pressures under the metatarsal heads and help to prevent the occurrence of plantar ulcers. Lower pressures under the metatarsal heads may occur for several reasons. First, lower ground reaction forces are developed in the anterior-posterior direction using the hip strategy (Tab. 3). These shear forces are believed to be particularly damaging to the insensitive foot.3 Second, and consistent with the suggestions of Brand,[3] we hypothesize that ground reaction forces would be spread over a greater area of the plantar foot because there is minimal push-off using the hip strategy. Rather than all forces concentrated at the forefoot, forces may be distributed over the entire plantar surface of the foot during a longer period of terminal stance. This distribution of forces would result in lower plantar pressures at the forefoot, the most frequent site of ulceration[3] Recently, this laboratory tested the hypothesis that training subjects with DM and a forefoot plantar ulcer to walk emphasizing a hip strategy will decrease peak forefoot pressures. The results of this study and other work[28] provide evidence that training subjects to walk flat-footed with diminished push-off and shorter steps can reduce peak forefoot pressures substantially. Additional research is needed to determine whether this decrease in peak forefoot plantar pressures results in a decreased incidence of ulceration.

For some patients, orthotic devices can be used successfully at the foot and ankle to facilitate stability.[29] The one DM group subject who used a polypropylene ankle-foot orthosis in this study had zero dorsiflexor and plantar-flexor strength, as evidenced by manual muscle testing and isokinetic testing. Despite zero ankle strength, he was able to generate a plantar-flexor moment of 865 N[multiplied by]m (1.21 N[multiplied by]m/kg) during walking. The exact contribution of the orthotic or soft tissue limitations to the ankle moment is impossible to determine from this study. In the presence of severe weakness of the plantar-flexor muscles, limited dorsiflexion ROM due to soft tissue shortening[30] or use of an orthotic device may increase the ability to generate a plantar-flexor moment (and hence stability) during walking. Additional research using the link-segment model to quantify the kinetics of walking is required to determine the optimal balance between these variables with patients having a variety of disabilities. This study provides further evidence for the advantages of using the link-segment model to study the mechanical factors of walking. Despite the benefits of using the link-segment model, certain limitations should be considered when interpreting the results of this study. Walking is performed in three dimensions, whereas this model considered movement and forces only in the sagittal plane at the lower extremities. Joints were considered hinge joints with motion occurring in the sagittal plane. This is a reasonable assumption, particularly for the hip and knee, which show most of their movement during walking in the sagittal plane.[31] Movement at the ankle is also primarily in the sagittal plane, but compensations can be made in the frontal and transverse planes to compensate for limitations in ankle motion.[32] The DM group subjects may have shown interesting compensations at the trunk or in the frontal and transverse planes that were not recorded in this study. Because mechanical changes in the frontal plane have been reported to be small even in the presence of gait pathologies,[33] we do not consider this a serious limitation.

The link-segment model provides insight to the net moments or power acting at a joint but says little about what is happening in individual muscle groups. If a co-contraction is occurring at a given joint, the analysis yields the net effect of the agonist and antagonist muscles.[8] In addition, the contribution of active or passive structures to a joint moment is impossible to discriminate unless the muscle is known to be silent or incapable of functioning, The contribution of passive tension development from soft tissue limitation can be greatly increased at a joint's end ROM. Many of the DM group subjects appeared to be functioning in their dorsiflexion end ROM during terminal stance. Because of their significant strength deficits, the passive stretching of the soft tissue around the ankle joint appeared to be making a contribution to the plantar-flexor moment. Although there is support for the hypothesis that soft tissues around a joint can make a significant contribution to the ankle plantar-flexor moment,[30] additional research is needed to determine the specific contribution of passive and active structures in various patient populations.

Conclusion

Patients with DM and peripheral neuropathy demonstrated decreased ankle plantar-flexor strength and ankle mobility compared with age-matched controls. These deficits appeared to contribute to limited ankle motion, ankle moments, ankle power, stride length, and velocity during walking compared with age-matched controls. The DM group subjects appeared to pull their leg forward using hip flexor muscles (hip strategy) rather than pushing the leg forward using plantar flexor muscles (ankle strategy). The results of these kinetic analyses can be used to design treatment strategies to enhance the mechanical factors of walking.

Acknowledgments

We acknowledge Paul SG Stein, Phd, Professor of Biology, Washington University, for his suggestions to clarify the descriptions of walking strategies.

References

[1] Cavanagh PR, Derr JA, Ulbrecht JS, et al. Problems with gait and posture in neuropathic patients with insulin-dependent diabetes mellitus. Diabetic Med. 1992;9:469-474. [2] 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. [3] 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. [4] Winter DA. Overall principle of lower limb support during stance phase of gait. J Biomech. 1980;13:923-927, [5] Winter DA. Concerning the scientific basis for the diagnosis of pathological gait and for rehabilitation protocols. Physiotherapy Canada. 1985;37:245-252. [6] Winter DA. Biomechanics of normal and pathological gait: implications for understanding human locomotor control. Journal of Motor Behavior, 1989;21:337-355. [7] Elftman H. Forces and energy changes in the leg during walking. Am J Physiol. 1939; 125: 339-356. [8] Winter DA. Biomechanics and Motor Control of Human Movement, 2nd ed. New York, NY: John Wiley & Sons Inc; 1990, [9] Olney SJ, MacPhail HE, Hedden DM, Boyce WF. Work and power in hemiplegic cerebral palsy gait. Phys Ther. 1990;70:431-438. [10] Murray M, Kory R, Cladisen B. Walking patterns of healthy old men. J Gerontol. 1969;24: 169-178, [11] Delbridge L, Perry P, Marr S, et al. Limited joint mobility in the diabetic foot: relationship to neuropathic ulceration. Diabetic Med. 1988; 5:333-337. [12] Proceedings of a concensus development conference on standardized measures in diabetic neuropathy: clinical measures. Diabetes Care. 1992; 15:1081-1083. [13] Lahey MA, Downey RG, Saal FE. Intraclass correlations: There's more than meets the eye. Psychol Bull. 1983;93:586-595. [14] Diamond JE, Mueller MJ, Delitto A. Reliability of a diabetic foot evaluation. Phys Ther, 1989;69:797-802, [15] Fugl-Meyer AR, Gustafsson L, Burstedt Y. Isokinetic and static plantar flexion characteristics. Eur J Appl Physiol, 1980;45:221-234. [16] Winter DA, Patla AE, Frank JS, Walt SE. Biomechanical walking pattern changes in the fit and healthy elderly. Phys Ther. 1990;70:340-347. [17] Mueller MJ, Norton BJ. Reliability of kinematic measurements of rear-foot motion. Phys Ther. 1992;72:731-737. [18] Rosenthal R, Rosnow RL. Contrast Analysis: Focused Comparisons in the Analysis of Variance. New York, NY: Cambridge University Press; 1985, [19] Horak FB, Shupert CL, Mirka A. Components of postural dyscontrol in the elderly: a review. Neurobiol Aging. 1989;10:727-738. [20] Stein PSG, Mortin LI, Robertson GA. The forms of a task and their blends. In: Grillner S, Stein PSG, Stuart DG, et al, eds. Neurobiology of Vertebrate Locomotion, London, England: Macmillan Press; 1986:201-216. [21] Lewallen R, Dyck G, Quanbury A, et al. Gait kinematics in below-knee child amputees: a force plate analysis. J Pediatr Orthop. 1986;6: 291-298. [22] Winter DA, Sienko SE. Biomechanics of below-knee amputee gait. J Biomech. 1988;2 1: 361-367 [23] Marshall R, Nade S. Effects of arthrodeses on walking: kinematic and kinetic studies of subtalar and knee arthrodesis. Clin Biomech. 1991;6:51-59. [24] Andriacchi TP. Walking speed as a basis for normal and abnormal gait measurements. J Biomech. 1977;10:261-268. [25] Winter DA. Foot trajectory in human gait: a precise and multifactorial motor control task. Phys Ther. 1992;72:45-56. [26] Mueller MJ. Effect of Ankle Strength and Mobility on Kinetics of Gait. St Louis, Mo: Washington University; 1992. Doctoral dissertation. [27] Frontera WR, Meredith CN, O'Reilly KP, Evans WJ. Strength training and determinants of V[O.sub.2]max in older men. J Appl Physiol. 1990; 68:329-333. [28] Hongsheng Z, Wertsch JJ, Harris GF, et al. Foot pressure distribution during walking and shuffling. Arch Phys Med Rehabil. 1991;72:390-397. [29] Butler PB, Thompson N, Major RE. Improvement in walking performance of children with cerebral palsy: preliminary Results. Dev Med Child Neurol. 1992;34:567-576. [30] 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. [31] Ounpuu S, Gage JR, Davis RB. Three-dimensional lower extremity joint kinetics in normal pediatric gait. J Pediatr Orthop. 1991; 11:341-349, [32] Mann RA. Biomechanics of the foot. In: American Academy of Orthopedic Surgeons. Atlas of Orthotics: Biomechanics Principles and Applications. St Louis, Mo: CV Mosby Co; 1975. [33] Mansour JM, Lesh MD, Nowak MD, Simon SR. A three-dimensional multi-segmental analysis of the energetics of normal and pathological human gait. J Biomech. 1982; 15:51-59.

Invited Commentaries

Following are two commentaries on "Differences in the Gait Characteristics of Patients With Diabetes and Peripheral Neuropathy Compared With Age-Matcbed Controls."

The authors should be congratulated for their innovative and timely research investigation. The primary thrust of previous research endeavors involving patients with diabetes and neuropathy has been focused on etiological factors leading to foot ulceration as well as appropriate treatment protocols to heal these ulcers. The authors of this investigation have gone beyond just considering the obvious problem of foot ulcerations and have examined the effect of the devastating clinical features of diabetes mellitus on the movement pattern of the entire lower extremity during the walking cycle. The concept of considering how the diabetic disease process affects not only the foot but the movement pattern of the entire lower extremity is fundamental in order to determine the most optimal treatment approach for these patients. No doubt, this investigation will establish a standard by which future studies will evaluate the effect of diabetes mellitus on the lower extremity and the foot.

Despite the tremendous importance of the authors' contribution to the existing body of knowledge on the effects of diabetes on the movement pattern of the lower extremity, I have two major concerns regarding the conclusions stated by the authors: (1) the proposed rationale for differences in the walking patterns between the diabetic and control groups and (2) the recommendation by the authors to instruct patients with diabetes in the use of a hip strategy.

The authors state that the differences in the walking patterns for the diabetic group subjects, in comparison with the nondiabetic group subjects, were the result of decreased plantar-flexor strength and limited mobility at the ankle, They further note that walking velocity, sensory loss, and other complications associated with diabetes mellitus were not the primary factors influencing the walking patterns in this study. Thus, plantar-flexor weakness and limited joint mobility were interpreted by the authors to be the two primary factors that caused the diabetic group subjects to use a hip strategy during walking.

An important question that must be considered is what caused this reduction in strength and joint mobility. Although the authors allude to diabetes as the cause, the dilemma with this interpretation of the findings is the composition of the diabetic group. The authors' criterion for ensuring that the diabetic group subjects selected had a severe peripheral neuropathy was a history of a neuropathic ulcer. This criterion would suggest that the subjects with diabetes who participated in this study must each have undergone a period of rehabilitation in order to heal their previous ulceration. Although the types of treatment used to heal the diabetic group subjects' ulcers were not stated, one would assume that the management program could have included whirlpool, total contact casting, bed rest, and only partial weight-bearing activities.

Although the use of total contact casting has been shown to have only a minimal effect on decreasing mobility of the foot joints,[1] patients undergoing the process of healing a foot ulceration are often just not as active as they normally would be, especially if other systemic complications associated with diabetes are present, In severe cases of ulceration, the patient with diabetes may be subjected to marked limitations in the amount of weight-bearing activities permitted on the involved extremity, These limitations in activity, whether self-imposed or prescribed, during the time period required for healing these ulcers could be a major factor in causing the reduction in strength and joint mobility observed. Thus, the question remains as to whether the plantar-flexor muscle weakness and limited joint mobility were caused by complications associated with diabetes or instead by disuse of the lower extremity associated with the process of healing the previous ulceration. Future investigations should consider the addition of a second group of subjects with diabetes and neuropathy but no previous history of ulceration in an attempt to control for possible changes in strength and joint mobility associated with ulceration management.

The authors also discussed how the decreased plantar-flexor strength diminished the ability of the plantar-flexor muscles to push off as well as generate power during terminal stance. There has always been considerable debate as to whether the plantar flexors push off or propel the leg and body forward during the walking cycle. Sutherland et al[2] concluded from their study of five subjects, before and after tibial nerve block, that the plantar flexors do not propel the body forward. However, they did note that maximum step length could not be achieved without the stabilizing effect of the plantar flexors. They also found that the pattern of center of pressure, as determined from the force platform, illustrated distinct differences between pre- and post-tibial nerve block trials, The analysis of the center of pressure patterns in future investigations may add further insight in regard to the effect of plantar-flexor weakness associated with peripheral neuropathy. In their discussion of the clinical implications of their study, the authors recommended that a hip strategy be emphasized for patients with peripheral neuropathy to reduce metatarsal head plantar pressures by distributing ground reaction forces over a greater surface area. They further state that training patients to take shorter steps and to walk flat-footed with a diminished push-off might help to decrease the incidence of plantar ulcers. Although the location of the ulcerations was not provided by the authors, one would assume that the primary location of the ulcerations for the subjects with diabetes in this study was in the metatarsal head region.

The results of this investigation, however, indicate that the diabetic group subjects were already utilizing a hip strategy prior to data collection. Thus, one could question why the patients with diabetes in this investigation ever developed ulcerations considering the fact that they were already using a hip strategy for walking. The diabetic group subjects perhaps were instructed to use this type of walking pattern after their ulceration was healed, as part of their management program. These subjects may have just modified their walking pattern to incorporate the hip strategy independently as a result of the complications associated with diabetes. Regardless of how the patients with diabetes in this study adapted their walking pattern, this problem further illustrates the need for an additional group of subjects with diabetes and neuropathy but no previous history of ulceration.

Finally, to suggest that the use of a hip strategy as a mechanism by which patients with diabetes could possibly reduce the risk of plantar ulceration during walking is at best speculation. I agree with the authors that additional research is necessary to assess the effects of using a hip, knee, or ankle strategy during walking on the magnitude and duration of plantar pressures. Future studies evaluating the walking pattern of patients with diabetes must incorporate the objective assessment of plantar pressures in relation to the measurement of joint moments and power obtained from the link-segment model. Only with this type of multifactor analysis will the clinician fully understand the ramifications of using various strategies for walking in an attempt to reduce foot plantar pressures in the patient with diabetes.

Again, I congratulate Dr Mueller and colleagues for conducting an excellent and pioneering investigation.

References

[1] Diamond JE, Mueller MJ, Delitto A. Effect of total contact cast immobilization on subtalar and talocrural joint motion in patients with diabetes mellitus. Phys Ther. 1993;73:310-315. [2] Sutherland DH, Cooper L, Daniel D. The role of the ankle plantar flexors in normal walking. J Bone Joint Surg [Am]. 1980;62: 354-363.

The search to understand the neural control of human locomotion has been advanced by a number of pathological models and, as Mueller and co-workers point out in this article, the patient with diabetes and neuropathy has a number of features that may allow this understanding to progress to deeper levels. This is a view that our group shares enthusiastically.[1,2]

Peripheral neuropathy, which may occur in 25% to 50% of patients who have had the disease for more than 10 years, affects all divisions of the peripheral nervous system (sensory, motor, and autonomic).[3] The prevailing view of the control of locomotion is that central pattern generators provide the fundamental control input, which is modulated by peripheral feedback to accomplish the desired result.[4,5] Because somatosensory feedback is impaired in a patient with neuropathy, the gain of the feedback loops will be diminished and thus one might expect an altered control process and possibly a different output.

Although autonomic neuropathy probably has little or no effect on motor control, motor neuropathy can lead to muscle atrophy and subsequent weakness. If there is differential weakness on one side of a joint, then deformity may result in contractures, causing change in the resting position of the joint. There is no convincing evidence, however, that peripheral neuropathy is responsible for limited joint mobility (LJM). Peripheral neuropathy is but one of the many long-term complications of diabetes, and there is increasing evidence of genetic predisposition to specific complications or groups of complications.6 There is also a search for underlying mechanisms common to several complications, and Brownlee and associates[7,8] have advanced the view that glycosylation of proteins may be such a unifying mechanism. Glycosylation of connective tissues certainly seems to be responsible for the LJM that has been observed in children with insulin-dependent diabetes mellitus who have no clinical evidence of neuropathy.[9]

This brief review of the effects of diabetes on the musculoskeletal system is intended to serve as an indication of both the promise and the dangers of using patients with diabetes as a model for fundamental studies of motor control. Exclusion criteria become an issue of critical importance in such studies in which differences in the severity or distribution of neuropathy, the effects of central nervous system medications, the long-term history of glycemic control, the staging of the neuropathy, and other factors can all complicate the interpretation of experimental findings.

Mueller et al have chosen to study patients with diabetes principally because "the patient with diabetes mellitus (DM) and peripheral neuropathy offers a good clinical model for investigating strategies that patients develop to compensate for the effects of decreased ankle function." The multifaceted nature of the effects of diabetes and of peripheral neuropathy described show how complex such an interpretation might be. For example, suppose two patients had the same restriction of ankle motion, but one was neuropathic and one was not. Would the absence of somatosensory feedback in the patient with neuropathy cause unique compensatory strategies? Because LJM is not associated with neuropathy, it would seem that a more powerful approach would be to study the effects of LJM on gait in patients without neuropathy, to study the effects of neuropathy on gait in patients matched on the basis of LJM, and finally, to study the effects of neuropathy on compensation for LJM. The approach used in this study allows for no such separation of possible causative factors.

On the topic of exclusion criteria, mention should be made of the use of a history of plantar ulceration as a determining factor for neuropathy. Although it is true that all patients selected on this basis will probably have peripheral neuropathy, the severity, distribution, and staging of the neuropathy has the potential to be very different within this group of patients (although the long mean duration of disease reduces this risk). Staging is important because the onset of neuropathy is chronic rather than acute, and it is likely that compensation for lack of somatosensory feedback occurs as neuropathy progresses. It might also have been reasonable to exclude the subject with an ankle-foot orthosis from the study because the orthosis is highly likely to affect the generation of joint moments during gait.

The methodology used in the experiment was conventional and appropriate, although a sensitivity analysis to indicate the important factors that may have contributed to errors would have been useful, as would more detailed information concerning the exact methods used for the estimation of segmental masses, moments of inertia, and center-of-mass position. The control of walking speed is always a problematic issue in studies of gait (whether to choose free speed or fixed speed), and it proved to be so in this study. The investigators did not control speed of walking, which resulted in a mean walking speed and a mean stride length that were both significantly smaller in the diabetic group compared with the control group. Although this is an interesting result in its own right, it complicates the interpretation of the significantly decreased range of ankle motion used, the lower joint moments and power at the ankle, and the smaller anterior-posterior ground reaction forces in the diabetic group. All of these differences could have been the result of a slower walking speed, and although the authors believe these changes were a result of the significantly decreased plantar-flexor peak torque and ankle mobility," it is not possible to show this conclusively given the present experimental design.

In the presentation of results, it may have been helpful to the reader to have shown average angle, velocity, support moment, and joint power curves as functions of time for the two groups because the data were normalized and it is often difficult to interpret representative graphs in relation to tables containing mean values. In addition, the timing of certain peak events may be as important as the peak values themselves.

The crux of the report is the statement that patients with neuropathy may use biomechanical strategies during the late support phase of gait that are different from those used by persons without neuropathy. Regardless of the exact cause of such differences, this notion deserves further examination. The authors observed that "many DM group subjects had greater hip moments and power than ankle moments and power during terminal stance" and that this finding is "opposite to the pattern observed" in the subjects without diabetes. From these observations, they conclude that "the DM group subjects pulled their leg forward using hip flexor muscles (hip strategy), rather than pushing ... with the plantar-flexor muscles." The values for moments and power in Table 3 represent the values measured "during terminal stance of walking" and allow a perspective on any differences between groups to be obtained. For example, the ratio of ankle to hip moments for the groups as a whole is 1.24 for both groups, implying no difference in the relative roles of the net moment at each joint. The corresponding values for power are 1.63 for the control group and 1.31 for the diabetic group. Thus, although the notion of different joint strategies is intriguing, it does not seem to be supported by the tabular data for the groups as a whole. The authors, however, were able to gain insight from an examination of the individual curves for all subjects, and it is difficult for the reader, in the absence of statistical analysis, to judge the significance of observations made on the basis of the sample curves shown. In considering the interaction of moments at the various joints of the lower extremity, the work of Zajac and colleagues[10] has provided some significant new directions for interpreting coordination of movement, and these ideas may prove useful in future studies of the current problem.

At a clinical level, the authors dismiss the possibility that strength training could have a meaningful effect on muscle strength in patients with peripheral neuropathy and do not mention the possibility of increasing the range of motion at joints with LJM. I believe, however, that both of these interventions deserve serious consideration. Although neuropathy is progressive, it is likely that existing muscle fibers may undergo hypertrophy in a patient with neuropathy much as they would in an individual without neuropathy, allowing the patient to maximize the output of currently available musculature. Similarly, although there is no evidence that glycosylation of connective tissue can be reversed, it is still an open question as to whether the range of motion can be increased at a joint limited because of such changes. In summary, the authors have raised interesting new ideas about the differences between neuropathic and nonneuropathic gait. Future studies need to separate the effects of potentially confounding factors and to explore the variables that may define "strategies" in a more rigorous statistical manner. if such future studies confirm the hypotheses presented by the Mueller et al, this article will be recognized as having broken new and interesting ground.

References

[1] Cavanagh PR, Simoneau GG, Ulbrecht JS. Ulceration, unsteadiness, and uncertainty: the biomechanical consequences of diabetes mellitus. J Biomech. 1993;26(suppl 1):23-40. [2] Cavanagh PR, Simoneau GG, Ulbrecht JS. Posture and gait in patients with diabetic distal symmetrical polyneuropathy. Bebavioral and Brain Sciences. 1992;15:724-725. [3] Greene DA, Sima AAF, Albers JW, Pfeifer NIA. Diabetic neuropathy. In: Rifkin H, Porte D Jr, eds. Ellenberg and Rifkin's Diabetes Mellitus: Theory and Practice. 4th ed. New York, NY: Elsevier Science Publishing Co Inc; 1990:710-755. [4] Patla A. Understanding the control of human locomotion: a prologue. In: Patla A, ed. Adaptability of Human Gait: Implications for the Control of Locomotion Advance in Psychology. Amsterdam, the Netherlands: Elsevier Science Publishers BV; 1991:3-20. [5] Grillner S, Debuc R. Control of locomotion in vertebrates: spinal and surpaspinal mechanisms. In: Waxman SG, ed. Advances in Neurology, Vol. 47: Functional Recovery in Neurological Disease. New York, NY: Raven Press; 1988. [6] Mann JI, Houston AC. Genetic factors in diabetes mellitus. In: Davidson JD, ed. Clinical Diabetes Mellitus: A Problem-Oriented Approach. New York, NY: Thieme Medical Publishers Inc; 1986:46-53. [7] Brownlee M, Vlassara H, Cerami A. Nonenzymatic glycosylation and the pathogenesis of diabetic complications. Ann Intern Med. 1984; 101:527-537. [8] Brownlee M, Cerami A, Vlassara H. Advanced glycosylation end products in tissue and the biochemical basis of diabetic complications. N Engl J Med. 1988;318:1315-1321. [9] Rosenbloom AL, Silverstein JH, Riley WJ, MacLaren NK. Limited joint mobility in childhood diabetics: family studies. Diabetes Care. 1983;6:370-373. [10] Zajac FE. Muscle coordination of movement: a perspective. J Biomech. 1993;26 (suppl 1):109-124.

Author Response

The results of our study identified many differences in the gait characteristics of patients with diabetes mellitus (DM) and peripheral neuropathy compared with age-matched control subjects. The reasons for and significance of these differences are the bases for considerable speculation. We appreciate the opportunity to respond to the thoughtful commentaries of Dr McPoil and Dr Cavanagh and to elaborate on this speculation.

As explained in our report, we believe that the differences in the gait characteristics observed between the DM and control groups resulted primarily from a decrease in ankle strength and mobility in the DM group. We speculated that the plantar-flexor muscle weakness was a result of peripheral neuropathy as a complication of diabetes. Dr Cavanagh comments that other effects of diabetes and differences in walking speed also may have been factors that influenced the differences in gait characteristics. As indicated in the "Discussion" section of our article, we agree that other effects of DM, especially sensory deficits, and walking speed can influence gait characteristics. If these factors did account for differences in gait characteristics in this study, however, one would have expected to see a significant contribution by group (ie, DM versus no DM) and walking velocity in the hierarchical, multiple regression analysis cited in our article. Group added no unique variance in explaining ankle moments. Although sensory deficits or other complications of diabetes may affect gait, according to this statistical analysis, those factors were redundant with the loss of ankle strength and mobility. Likewise, walking velocity added no significant unique variance after plantar-flexor peak torque and ankle mobility were entered into the analysis. Based on these results, we believe the documented limitations in ankle strength and range of motion (ROM) were the primary factors that limited the walking speed and caused the differences in gait characteristics. Studies that control walking speed experimentally are needed to address this issue completely. Although we believe that the primary reason for the plantar-flexor muscle weakness was peripheral neuropathy, we agree with Dr McPoil that other factors, such as inactivity during the wound-healing process, may have contributed to weakness and decreased dorsiflexion ROM. We agree with both commentaries that additional research using multiple groups to assess the unique effects of peripheral neuropathy, history of plantar ulcer, and limited joint mobility would be insightful.

Dr McPoil also questions the utility of instructing patients with diabetes who already are using the hip strategy to emphasize using the hip strategy to minimize the incidence of forefoot plantar ulcers. As indicated in the "Discussion" section of our report, subjects in both groups appeared to use a blend of the hip and ankle strategies to accomplish the subtask of advancing their lower extremity. Although most subjects with DM and peripheral neuropathy appeared to use a predominant hip strategy, most still showed some push-off at the ankle, suggesting a component of the ankle strategy, Most patients with a history of DM and a plantar ulcer seem to be capable of an increased reliance on the hip strategy. In recent research conducted in our laboratory, patients with DM and a history of a plantar ulcer have been trained to emphasize the hip strategy (ie, decrease push-off and shorten stride length, but maintain velocity) and have shown a 27% average decrease in forefoot peak plantar pressures. Because plantar neuropathic wounds are thought to be due primarily to excessively high pressures on an insensate foot,[1] training patients with plantar ulcers to emphasize the hip strategy may be a helpful adjunct treatment to heal or prevent ulceration in addition to the benefits of wearing orthopedic footwear. Certainly, more research is needed to determine whether emphasizing a hip strategy cannot only reduce peak plantar pressures, but also improve the healing rate of forefoot neuropathic ulcers.

Dr Cavanagh questions the existence of a hip strategy used by patients with DM and peripheral neuropathy. The primary purpose of our study was to compare the gait characteristics of patients with DM and peripheral neuropathy with those of age-matched control subjects rather than to define walking strategies. Several aspects of the data analysis, however, suggested the existence of different strategies that can be used during walking. Compared with subjects in the control group, the subjects in the DM group showed decreased ankle strength and ROM measured non-weight-bearing, decreased ground reaction forces in the anterior-posterior direction, and decreased ankle moments and power during walking, but no difference in knee and hip moments and power during walking (Tab. 3). Additionally, as shown in Figure 3, hip flexor moments appeared to begin earlier during the stance phase in some subjects. These results provide some evidence for a difference in walking strategies as described in our article. We also would again emphasize that hip and ankle strategies do not appear to be mutually exclusive categories. Rather, there seems to be a continuum by which individuals rely on either proximal or distal musculature to advance the lower extremity. Because of their deficits in ankle strength and ROM, the subjects with DM and peripheral neuropathy appeared to be more limited to using the hip strategy than did age-matched control subjects. Given the results and hypotheses generated in this study, prospective studies are needed to describe these strategies in healthy and pathological conditions in greater detail.

We agree with Dr Cavanagh that exercise to increase the muscle strength in patients with peripheral neuropathy deserves consideration, particularly in the early stages of neuropathy or after prolonged immobilization. We remain skeptical, however, that strength can be significantly improved in muscle with severe neuropathy. Additionally, although increasing plantar-flexor muscle strength may enable patients to take a more vigorous push-off and longer stride length, this vigorous push-off places more stresses on the forefoot.[2] Increased stresses would not be indicated in this group of patients who are susceptible to plantar ulcers. Rather, efforts might best be directed toward training proximal musculature and emphasizing the hip strategy as discussed in our report.

This study has identified a number of differences in the gait characteristics of patients with DM and peripheral neuropathy compared with those of age-matched control subjects. The study and accompanying dialogue also have raised a number of interesting speculations regarding ankle and hip walking strategies that may be used by healthy and patient groups. These strategies may be useful in training patients to compensate for various musculoskeletal and neurological limitations. Additional research is needed to clarify the reasons for the differences in gait characteristics and to define the walking strategies in healthy and pathological conditions.

References

[1] Boulton AJM, Betts RP, Franks CI, et al. Abnormalities of foot pressure in early diabetic neuropathy. Diabetic Med. 1987;4:225-228. [2] Hongsheng Z, Wertsch JJ, Harris GF, et al. Foot pressure distribution during walking and shuffling. Arch Phys Med Rehabil. 1991;72:390-397.

MJ Mueller, PhD, PT, is Assistant Professor, Program in Physical Therapy, Washington University School of Medicine, 660 S Euclid, Box 8083, St Louis, MO 63110 (USA). Address all correspondence to Dr Mueller.

SD Minor, PhD, PT, is Assistant Professor, Program in Physical Therapy, Washington University School of Medicine.

SA Sahrmann, PhD, PT, FAPTA, is associate Professor, Program in Physical Therapy, Washington University School of Medicine.

JA Schaaf, PhD, is Visiting Research Associate and Lecturer, Department of Mechanical and Aeronaut Engineering, University of California, Davis, CA 95615. He was Hunter Assistant Professor of Mechanical Design, Mechanical Engineering, Washington University, at the time of this study.

MJ Strube, PhD, is Professor of Psychology, Washington University.

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

This study was supported by a grant from the Foundation for Physical Therapy Inc.

This article was submitted February 17, 1993, and was accepted October 27, 1993.