Gait Problems in Diabetic Neuropathic Patients - Archives of Physical ...

849

Gait Problems in Diabetic Neuropathic Patients Richard Courtemanche, MSc, Normand Teasdale, Phi), Pierre Boucher, DC, MicheUe Fleury, Phi), Yves Lajoie, Phi), Chantal Bard, PhD ABSTRACT. Courtemanche R, Teasdale N, Boucher P, Fleury M, Lajoie Y, Bard C. Gait problems in diabetic neuropathic patients. Arch Phys Med Rehabil 1996;77:849-55. Objective: To examine whether a reduced peripheral sensibility caused by diabetic neuropathy increases the attentional demands necessary for controlling and regulating gait. Design: Nonrandomized control trial. Setting: University motor performance laboratory. Subjects: Twelve diabetic patients with peripheral nenropathy and 7 control subjects, all volunteers. Interventions: All subjects first performed a control seated reaction time task. For the walking task, auditory stimuli were randomly presented in the third, fourth, or fifth walking cycle on left foot toe off or on left foot heel contact. The subject's task was to respond verbally as fast as possible to the auditory stimulus, while maintaining progression. Main Outcome Measures: Simple reaction times and kinematics of the gait pattern (cycle amplitude, cycle duration, cycle speed, cadence and percentage of time spent in the single support phase) were evaluated. Results: For the walking task, diabetic neuropathic patients had a smaller cycle amplitude, cycle speed, and percentage of time spent in the single support phase than control subjects. Also, reaction times while walking were higher for diabetic neuropathic patients than for control subjects. Conclusions: Diabetic neuropathic patients show a less destabilizing and more conservative gait than control subjects. The increased attentional demands in gait for the diabetic neuropathic patients, along with their more conservative gait pattern, suggest that a lack of proprioception from the legs affects the control of gait. Diminished sensory information makes gait control more cognitively dependent in diabetic neuropathic persons than in control subjects.

© 1996 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation ALKING is a highly practiced activity, often considered "automated." This attribute has emerged from several experiments demonstrating, with the animal model, that the spinal cord (1) produces an oscillatory pattern responsible for

W

From the Laboratoire de Performance Motrice Humaine, Universit6 Laval, Quebec, Canada. Submitted for publication June 28, 1995. Accepted in revised form February 13, 1996. Supported by various grants from the Natural Sciences and Engineering Research Council of Canada and Quebec's Fonds pour la Formation de Chercheurs et l'Aide ~i la Recherche. P. Boucher and R. Courtemanche received support from the Association Diab~te Quebec. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors or upon any organization with which the authors are associated. R. Courtemanche is now at the Department of Physiology, Universit6 de Montreal. Y. Lajoie is now at the School of Physical Education, Laurentian University, Sudbury, Ontario. Reprint requests to Normand Teasdale, PhD, Laboratoire de Performance Motrice Humaine, Universit6 Laval, Qu6bec, Canada, G1K 7P4. © 1996 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation 0003-9993/96/770%357753.00/0

the automatic production of elaborate locomotor synergies ~'2and (2) is capable of functionally adaptive responses when afferent inputs are available)-5 When the locomotor task requires clearing an obstacle, walking on irregular terrain, and changing direction, these low level systems are not sufficient to insure appropriate gait regulation and several sources of sensory information are also used to implement changes adapted to the environmental context. Several recent proposals suggest that the integrity of higher cortical/central factors is of significance for posture and gait control. 6~° Highly cognitive tasks, such as mental arithmetic, have been shown to interfere with the act of walking. 6'11 Deficits in the ability to divide attention and/or to allocate properly the resources between concurrent tasks 8,~z't3 provide a possible explanation for these observations. Attention is often considered as an extension of sensory processes) 4 The testing of attention traditionally relies on three basic underlying assumptions: (1) there is a limited central information processing capacity, (2) performing a task requires part of the limited processing capacity within the central nervous system, and (3) if two tasks share the processing capacity, the performance in one or both tasks can be disturbed if the limited central processing capacity is exceeded. ~'~5'~6 Using this general approach, several authors have demonstrated that walking requires more cognitive processing than simple sitting or upright standing posture/v-~9 Lajoie et all 9 also showed that the attentional demands for controlling and regulating gait vary within a walking cycle. In these studies, the single support phase, which involves limb oscillation and requires adequate foot trajectory and placement, demands more attention than the double support phase that could be used as a restabilizing phase. These important findings demonstrate that even though walking is a highly practiced task, it requires some cognitive processing even when performed in a constant and predictable environment. A pathology of the proprioceptive system, like diabetic neuropathy, could affect walking not only because there is reduced peripheral sensory information available to local peripheral systems but also because reduced information puts a burden on higher cortical centers involved in the processing of sensory information. The most common type of neuropathy in diabetes is a distal symmetric polyneuropathy, which is characterized mostly by sensory and autonomic manifestations, z° Its effects on the motor systems could result from the loss of large fibre proprioceptive and somatic feedback from receptors in the legs and feet and/or from muscular weakness due to motor neuropathy.Z~ The symptoms are a lessened sensibility to touch and vibration and a diminished sense of limb position; a clear reduction in the amplitude and velocity of the peripheral nerve action potentials has been shown) z Recent studies have shown marked decrements in postural stability in diabetic neuropathic patients compared with a control population, z3-z5 Diabetic neuropathic patients also feel less safe while standing and walking, 21 and their gait is characterized by a slower velocity and a shorter stride length than the gait of control subjects, z6,z7 The present experiment was aimed at (1) examining the gait kinematics of patients suffering from a reduced peripheral sensibility caused by diabetic neuropathy and (2) determining whether the reduced peripheral sensibility increases the attentional demands necessary for controlling and Arch Phys Med Rehabil Vol 77, September 1996

850

GAIT AND DIABETIC NEUROPATHY, Courtemanche

regulating gait. Parts of this work have been published in abstract form. e:

METHODS Subjects Subjects were recruited through advertisements in local newspapers. A first advertisement asked for diabetic neuropathic persons; a second asked for healthy nondiabetic and nonneuropathic control subjects. Forty-six persons responded to both advertisements (27 diabetic and 19 control subjects). A first screening for medical, orthopedic or neurological conditions (other than diabetes and diabetic sensory neuropathy) that could affect postural stability was done by telephone. From this screening, 17 diabetic neuropathic patients and 12 control subjects were selected to participate; 17 persons were rejected for a variety of problems (eg, walking with a cane, feet ulcers, visual problems, etc). The 29 qualified subjects gave their informed consent to participate. A neurological history was taken following the 1992 San Antonio Conference recommendations. 28 To exclude nondiabetic neuropathies, the history included questions about family history of nondiabetic peripheral nerve disease and the presence of toxic, metabolic, mechanical, and vascular causes of nerve disease. Following this examination, 2 subjects were excluded from the study, one because of probable alcoholic neuropathy and another because of monocular blindness.

Scoring System for Quantifying the Polyneuropathy The polyneuropathy was quantified using a scoring system recently developed by Valk and colleagues, e The scoring system includes 4 levels of neuropathy: normal, mild, moderate, and severe. It consists of clinical testing of (1) sensory modalities (light touch, vibration, pain), (2) anatomic level below which light touch sensation is impaired, (3) muscle strength, and (4) ankle jerk. Pinprick and light touch sense of the dorsum of the foot, vibration sense of the ankle, and position sense of the first toe are scored separately as follows: 0, normal; 2, impaired in comparison with proximal sensation; and 4, absent. Light touch sense is also related to the anatomic level below which it is impaired and is scored as follows: 1, toe; 2, midfoot; 3, ankle; 4, midcalf; and 5, knee. Strength of the extensor hallucis longus and gastrocnemius muscles are scored separately as follows: 0, normal; 2, impaired; and 4, absent. Ankle jerks are scored as follows: 0, normal; 2, impaired in comparison with other reflexes (knee jerk); and 4, absent. The outcome of the total score varies between 0 and 33. A total score of 0 is graded as no polyneuropathy, l to 9 as mild polyneuropathy, 10 to 18 as moderate polyneuropathy, and 19 to 33 as severe polyneuropathy. Valk 29 showed that this score correlates well with scores obtained from neurophysiological examination (the scores were obtained from 78 patients; r = 0.7, p < .001).

Nerve Conduction Velocity Motor conduction velocity of the tibial and peroneal nerves and sensory conduction velocity of the sural nerve were tested by the same experimenter. The stimulation technique is described by Hugon~°: the cathode was a 1-cm-diameter stainless steel sphere placed over the stimulated nerve; the anode was a 20-cm 2 tin plate placed across the limb (eg, patella for stimulation of the tibial nerve). According to Hugon, 3° this arrangement allows reduction of the stimulus artefact. The stimulator unit was a Grass $88. a For patients' protection, the electrical stimulus was administered percutaneously through an isolation unit (Grass Stimulus Isolation Unit SIU5Aa). Prespaced Ag-AgC1

Arch Phys Med Rehabil Vol 77, September 1996

surface recording electrodes were connected to an amplifier (Bortec Electronics Inc.b). Standard techniques of supramaximal percutaneous nerve stimulation and surface recording were employed. 3~ Skin temperature was maintained at 34°C with an electric heating pad and monitored at midcalf with a thermistor thermometer. The tibial nerve was stimulated proximally at the popliteal fossa along the flexor crease of the knee and distally behind and proximal to the medial malleohis. Recordings were made over the abductor halhicis. The deep peroneal nerve was stimulated proximally behind and proximal to the fibular head and distally at the ankle between the extensor digitorum longus and extensor hallucis longus tendons. Recordings were made over the belly of the extensor digitorum brevis muscle. The method of testing the sural nerve was antidromic. The nerve was stimulated slightly distal to the lower border of the bellies of the gastrocnemius, approximately at the junction of the middle and lower thirds of the foreleg. Recordings were made between the lateral malleolus and the Achilles tendon at the malleolar level. Motor nerve conduction velocities were calculated using proximal and distal onset latencies. Sensory conduction velocities were calculated by dividing the distance by the onset latency.

Group Descriptions The clinical score was used solely to select neuropathic patients and control subjects for the gait experiment. Specifically, neuropathic patients were selected if their score was greater than 0 on the Valk scale. Five of 17 neuropathic patients were excluded from the study because they obtained a score of 0 on the Valk scale (3 patients had painful neuropathy and 2 patients had autonomic neuropathy). The neuropathic group (n = 12) comprised eight type I and four type II diabetic patients. Control subjects were selected if they scored 0 on the Valk scale (n = 7). Five of 12 control subjects were excluded because they obtained a score greater than 0 on the Valk scale. All these persons had diminished vibration sense at the ankle. A general description of the subjects, total Valk score, and nerve conduction velocities is provided in table 1. Table 2 presents the details of the clinical examination of sensory functions, muscle strength, and tendon reflexes for the neuropathic patients.

Experimental Tasks and Apparatus Control seated task. All subjects first performed a control seated reaction time (RT) task. This task served to establish the baseline speed of cognitive processing for each subject. For this purpose, binaural auditory stimuli were delivered through headphones (1.5KHz tone, 200msec duration). Electrically conductive thimbles were placed on the subjects' index and thumb fingertips; a break in the contact between the fingers sent a digital pulse to the computer. The subject produced a bilateral thumb-index extension-flexion response to each stimulus. RT was defined as the time interval between stimulus presentation and response onset. Walking task. The walking task was performed on a nonslippery steel-covered pathway 8m long. The subjects' shoes were instrumented with conductive material under the heel and toes of each foot. The different contacts with the pathway were digitally coded to provide accurate temporal values corresponding to the onset and offset of right and left single support and double support phases. Left and right foot displacements were recorded by means of small monofilament wires attached to the rear of each shoe. Each wire was wrapped around a plastic wheel (30-cm circumference) fixed onto a gear-box system having a 4.6:1 ratio. A 10-turn high-precision potentiometer was mounted on the shaft of the rotating axis and provided a voltage

851


Table 1: Valk Score and Nerve Conduction Studies for the Diabetic Neuropathic Patients and Control Subjects Subject No. Neuropathic patients 1 2 3 4 5 6 7 8 9 10 11 12 Mean (SD) Control subjects 1 2 3 4 5 6 7 Mean (SD)

Sex

Age

F F M M M F M M M M M M

59 59 72 77 66 62 48 62 62 56 66 62 62.5 (7.4)

M F M F M M M

50 58 63 62 59 66 66 60.6 (5.6)

Diabetes Type

Time of Diagnosis (y)

Valk Score

7 11 8 7 6 12 2 6 13 13 20 12 9,75 (4.7)

4 4 6 6 8 9 12 13 13 18 21 28 12.0 (7.4)

Peroneal MNCV (m/sec)

Tibial MNCV (m/sec)

51.8 38.5 33.2 33.7 41.5 23.8 35.3 34.6 36.2 27.4

42.8 43.7 33.6 35.4 39.4 26.0 40.3 40.8 44.0 28.6 32.4

NA NA 35.6 (7.6)

0 0 0 0 0 0 0 0

Sural SNCV (m/sec)

NA 37.0 (6.2)

55.3 40.8 16.9 14.0 43.9 NA 29.2 40.0 25.0 45.3 18.5 NA 32.9 (14.1)

49.5 48.3 42.4 45.9 50.6 43.2 50.3 47.2 (3.4)

42.7 61.4 51.4 42.0 46.6 38.7 43,5 46.6 (7.6)

42.6 47.0 42.7 45.8 47.5 45.6 46.6 45.4 (2.0)

Abbreviations: MNCV, motor nerve conduction velocity; SNCV, sensory nerve conduction velocity; NA, not available (no action potential could be recorded).

proportional to the distance covered. The wires were maintained stable with a constant resistance spring (.66N) fixed into the rotating mechanisms. This low resistance served only to prevent the wires from shivering; it did not affect the walking behavior and was not perceptible by the subjects. The system provided a resolution of 3ram. Subjects also wore a helmet instrumented with a microphone. Signals that issued from foot contact ruptures, potentiometer recordings, and microphone were sent to a computer via an A/D converter and sampled at a rate of 500Hz. The data were stored on disk and analyzed on screen via a custom-designed data analysis system allowing the temporal measures to be derived. A piezoelectric buzzer, located on a desk near the center of the walking platform, was used to produce a sound of lkHz during 50msec. Subjects responded verbally and as fast as possible by saying " T O P " to the auditory stimulus. The analog signal from the microphone was used to signify the onset of the verbal response. RT was defined as the time interval between presentation of the sound and the first deviation from the background noise of the microphone signal. Procedures Initially, subjects performed the control seated RT task. Four trials of 30sec were collected. For each trial, 13 auditory stimuli were randomly presented (interstimuli delay varying from .7 to 1.5sec). An identical series of 4 trials also was collected at the

end of the experiment to examine potential learning or fatigue effects. For the walking experiment, a dual-task paradigm, consisting of 46 trials, was used to determine the attentional demands of walking. Subjects were instructed that walking was the primary task and that they should not modify their gait to respond faster to the auditory stimuli (secondary task). They were asked to walk at their preferred pace. They were initially submitted to a familiarization period (8 trials without stimulus). Afterwards, auditory stimuli were randomly presented either in the third, fourth, or fifth walking cycle on left foot toe off (ie, at the onset of the single support condition) or on left foot heel contact (ie, at the onset of the double support condition). Subjects thus responded to one stimulus per trial. The subjects' task was to respond verbally as fast as possible to the auditory stimulus, while maintaining progression. Five trials per experimental condition were given, for a total of 30 trials (ie, 3 cycles × 2 stance conditions). In addition, 8 trials without stimulus were randomly presented to prevent subjects from anticipating the presence of the auditory stimulus. The walking platform permitted subjects to complete 6 walking cycles. The 3rd and 4th cycles were conserved for data analysis to avoid any contamination due to the acceleration (lst and 2nd cycles) and the deceleration (5th and 6th cycles). For each trial, cycle amplitude, cycle duration, cycle speed, cadence,

Table 2: Clinical Examination of Sensory Functions, Muscle Strength, and Tendon Reflexes for the Neuropathic Patients Subject No.

1 2 3 4 5 6 7 8 9 10 11 12

Pain (Pinprick)

Vibration Sense

0 0 0 0

2 2 2 2

0

0

0 4 2 4 4 4 4

0 4 2 2 4 4 2

Position Sense

Light Touch Sense

Light Touch Sense (AnatomicPosition)

Force (ExtensorHallucis Longus)

Force (Gastrocnemius)

Ankle Jerk

Total Valk Score

2 0 2 2 0 0 0 0 0 0 2 4

0 0 0 0 2 4 0 2 4 4 4 4

0 0 0 0 I 1 0 1 1 2 1 5

0 0 0 0

0 0 0 0

0 2 2 2

4 4 6 6

I

0

4

8

0 0 1 1 1 1 4

0 0 1 1 1 1 1

4 4 4 0 2 4 4

9 12 13 13 18 21 28


852


Table 3: FValues and Significance Levels, for Each Group and for All 5 Kinematic Variables, of the Separate One-Way ANOVA Describing the Effects of the Probe-RT on the Gait Pattern

Diabetic neuropathic patients Control subjects

Cycle Amplitude

Cycle Duration

Cycle Speed

Cadence

% of Time in Single Support

F(2,22) = 1.85 F(2,12) = .01

F(2,22) = .48 F(2,12) = .03

F(2,22) = .37 F(2,12) = .74

F(2,22) = .48 F(2,12) = .08

F(2,22) = .43 F(2,12) = .22

A N O V A w a s n o t s i g n i f i c a n t (p > .05).

and percentage of time spent in the single support phase were measured. Cycle amplitude is the length (m) travelled between successive heel contacts of the same foot; cycle duration (see) is the time elapsed between successive heel contacts of the same foot; cycle speed (m/see) is cycle amplitude divided by cycle duration; cadence is the number of cycles produced within 1 minute; finally, the percentage of time spent in the single support phase is the time spent on one limb during one cycle divided by the cycle duration, times 100.

Validation of the Data To verify if RTs were affected by learning or fatigue, mean RTs (average of 13 RTs/trial × 4 trials) obtained for the initial seated trials were compared with those collected at the end of the experimental session. Data were submitted to a Group (2) by Session (2) analysis of variance (ANOVA). The main effects of Group (F(1,16) = .32, p > .05), Session (F(1,16) = 2.76, p > .05), and the interaction of Group by Session (F(1,16) = .03, p > .05) were not significant. Therefore, RTs for the 2 seated sessions were collapsed. The RTs observed during the control seated condition can be taken as a basic index of the speed to process information. To ensure that RTs were a valid index of the attentional demands necessary for controlling and regulating gait, it was essential to demonstrate that subjects did not modify their gait pattern when the auditory stimulus was introduced (ie, that subjects did not stop or slow their gait to respond more rapidly to the auditory stimulus). To validate this prerequisite, we compared the gait patterns when the auditory stimuli were given in single support, double support, and when no stimulus was given. One-way ANOVA was conducted for each group and for all 5 kinematic variables analyzed (ie, cycle amplitude, cycle duration, cycle speed, cadence, percentage of time spent in the single support phase). Comparison of the gait patterns show that subjects did not modify their gait pattern when an auditory stimulus was presented. Results of the ANOVA are summarized in table 3. Statistical Analysis For the analysis of the gait pattern, mean values and standard deviations for the diabetic neuropathic patients and control subjects for each kinematic variable (cycle amplitude, cycle duration, cycle speed, cadence and percentage of time spent in the single support phase) were computed. Data for each of these variables were submitted to a separate one-way ANOVA with group as a factor. To evaluate attentional demands during gait, RTs collected during the control seated task (average of 13 RTs/trial × 8 trials) and when subjects were in a single support or double support phase of the walking task (average of 1 RT/trial × 15 trials/condition) were submitted to a Group (2) by Condition (3) ANOVA, with repeated measures on the last factor. RESULTS

Gait Pattern of Diabetic Neuropathic Patients and Control Subjects Mean values and standard deviations for the diabetic neuropathic patients and control subjects for each kinematic variable Arch Phys Med Rehabil Vol 77, September 1996

(cycle amplitude, cycle duration, cycle speed, cadence and percentage of time spent in the single support phase) are presented in figure 1. The ANOVA showed that, compared to control subjects, the diabetic neuropathie patients had a shorter cycle amplitude (1.23m for the diabetic neuropathic patients vs 1.43m for the control subjects; F(1,17) = 9.86p < .01), a slower cycle speed (1.06m/see for the diabetic neuropathic patients vs 1.32m/ sec for the control subjects; F(1,17) = 6.25, p < .05), and a smaller percentage of time spent in the single support phase (65.17% for the diabetic neuropathic patients vs 70.14% for control subjects; F(l,17) = 7.84, p < .05). Further, the diabetic neuropathic patients had a longer cycle duration (1.183see for the diabetic neuropathic patients vs 1.105see for control subjects) and a slower cadence (102.5steps/rain for the diabetic neuropathic patients vs 109.8steps/min for control subjects). These latter differences, however, failed to reach significance level (F(1,17) = 1.84, p > .05 for cycle duration, and F(1,17) = 1.74, p > .05 for cadence). The above results (smaller cycle amplitude, slower speed and greater proportion of time in single support than control subjects) suggest that diabetic neuropathic patients adopt a less destabilizing and more conservative gait pattern than control subjects.

Reaction Times Reaction times for the control seated condition and the walking conditions are presented in figure 2. The analysis revealed significant main effects of Group (F(1,16) = 6.75; p < .05), Condition (F(2,32) = 91.44; p < .001), and a significant interaction of Group by Condition (F(2,32) = 5.22, p < .05). Figure 2 clearly shows that, for both groups, walking yielded longer RTs than the control seated condition. More important, an orthogonal decomposition of the interaction showed that, while both groups had similar control seated RTs (200msec for diabetic neuropathic patients vs 202msec for control subjects; F(1,16) = .02, p > .05), the increase of probe-RTs when walking was more important for the diabetic neuropathic patients than for control subjects (increase of 193msec for the diabetic neuropathic patients vs 121msec for control subjects; F(1,16) = 5.94, p < .05). For both groups, the RTs for the single support and double support phases were identical (F(1,16) = 5.94, p < .05). DISCUSSION Compared with age-matched control subjects, the gait of diabetic neuropathic patients was characterized by marked decrements in speed and stride length, as well as a greater proportion of time spent in double support. Values obtained for our control subjects compare well with normative data previously reported for a population of the same age. For example, Winter et a132 reported a stride length of 1.43m (1.39m for our control subjects) and a cadence of 109.8steps/rain (110.5steps/min for our control subjects). These observations differ from a phenomenological observation of an 81-year-old diabetes mellitus patient with absent ankle reflexes and vibratory sensation in the feetY This patient was reported to have a gait pattern similar to that of aged-matched control subjects. Mueller and associates,26 however, recently showed that diabetic neuropathic patients


1.75

853

13o

1.5

1.4

120

~ 1.5 Q.

Fig 1. Means and between-subject standard deviations for cycle amplitude, cycle duration, cycle speed, cadence, and percentage of time in the single support phase for the diabetic neuropathic patients ([3} and control subjects (U}.

,~110 =o ~, 1.25 O

0

1.1

500 450 4OO 350

300

rr

250

/

2OO 150

Controlsitting

I 90 !

with a history of foot ulceration had a slower walking velocity and a shorter stride length than age-matched control subjects. As for the present experiment, the diabetic neuropathic and the control groups showed similar cadence and cycle duration. These gait differences between diabetic neuropathic patients and control subjects are reminiscent of differences observed between young and healthy older adults. When asked to do so, older adults are able to walk as fast as control subjects. They adopt a slower walking speed that presumably serves chiefly to produce a more secure gait and an energy-efficient speed of progression. 34'35 Compared with control subjects, diabetic sensory neuropathy may also increase the need for producing a more secure gait. In support of this argument, it has been shown that diabetic neuropathic patients feel less safe than control subjects while standing and walking. 2~ Further, risk avoidance, fear of falling, and lack of confidence are considered important factors that influence postural control and gait behavior. 36'37 These results presented here suggest that the peripheral neuropathy required the patients to allocate a greater proportion of the attentional resources available to the walking task. The attentional demands for walking were more important for the diabetic neuropathic patients than for control subjects. It follows that the gait adaptations described above may also serve to reduce the attentional demands necessary to control the more destabilizing gait associated with a decreased peripheral sensory information. Diabetic neuropathic patients would also make greater use of attentional processes throughout the gait cycle, as indicated by the longer RTs. The gait adaptations could then

:=

O 100

/

Singlesupport

Doublesupport

Fig 2. Reaction times for the diabetic neuropathic patients (E3) and control subjects (11) when performing the control seated task and when probes were given in the single support and double support phases of the walking task. Data are means and between-subject standard deviations.

be the result of a loss of automaticity for the walking task. Indeed, Mulder and coworkers 6 reported that healthy elderly persons, contrary to young persons, significantly reduce their gait speed when asked to perform a concurrent mental task. They suggest that the decreasing peripheral sensibility associated with normal aging yields an increased cognitive regulation and/or a decreased or defective automaticity in the sensory integration processes. These authors have proposed that similar processes occur after structural damage to the peripheral sensorimotor system. 38 Hence, a deterioration of the peripheral sensory systems could potentiate gait and balance problems because of increasing attentional demands for the postural tasks. Ahernatively, Mueller 26 proposed that a decreased strength and mobility at the ankle could explain differences in gait between diabetes mellitus patients with a neuropathic ulcer and control subjects. This interpretation, however, was disputed through invited commentaries on the Mueller a r t i c l e . 39'4° Specifically, it is argued that patients undergoing the process of healing a foot ulceration are often less active than they normally would be and may be subjected to marked limitations in the amount of weight-bearing activities permitted on the involved extremity. This suggests that disuse of the lower extremity associated with the process of healing the previous ulceration may explain the muscle weakness and limited joint mobility rather than complication associated with diabetes. In the present experiment, none of the neuropathic patients had a foot ulceration and their ankle strength did not differ significantly from that of control subjects. Results from a recent experiment by Bergin et a125 also argue against a strength hypothesis for explaining gait differences between diabetic neuropathic patients and control subjects. In that experiment, muscle strength did not correlate with body sway, whereas vibration perception did, suggesting that a main source of unsteadiness in these patients is not a decreased muscle strength but a loss of proprioceptive information. 25 The results we obtained confirm that when compared to a seated condition, walking requires additional attentional resources, j7'19 More important, a decreased peripheral sensory information, due to diabetic neuropathy, required these patients to allocate more attention to walking than control subjects. Indeed, both groups had similar RTs in a control seated condition; when walking, however, the diabetic neuropathic patients responded with a greater delay than control subjects to the auditory probe (increase of 193msec for the diabetic neuropathic patients vs 121msec for the control subjects). Whether it is in a static or in a dynamic condition, like walking, there is cumulative evidence that decreased peripheral sensory information due to normal aging (which is characterized by a gradual and continuous degeneration of the peripheral sensory system4~) requires a greater allocation of the attentional resources to the control

Arch Phys Med Rehabil Vo177, September 1996

854


of postural stability. Recently, Boucher and associates 24reported an important relationship between the presence and severity of sensory neuropathy (evaluated with a clinical scale and measures of nerve conduction velocity in the lower limbs) and postural stability. A reduced peripheral sensibility due to symmetrical distal diabetic neuropathy may result in a decreased automaticity in the sensory integration processes. The implication is that complex attention-demanding environments are liable to stress postural control, particularly in those, like diabetic neuropathic patients, who have preexisting impaired postural control. 6,9,36 Overall, our study highlights the important role of proprioception in the control of human gait. A proprioceptive deficit, as documented for the diabetic neuropathic patients, yields a slower and more conservative gait pattern as well as an increase in attentional demands when walking. When confronted with more challenging motor tasks (eg, walking on rough grounds), neuropathic persons may be at higher risk of falling. Particular attention to balance evaluation and locomotion should be considered in diabetic neuropathic patients. These findings also point to a growing body of evidence suggesting that the potential influence of higher cortical/central factors on postural control has been underemphasized. 36 Acknowledgment: Special thanks to Benoit Genest and Gilles Bouchard for programming and technical assistance, respectively, and to Mtlanie Hamelin for help in data collection and analysis. This project was supported by various grants from the Natural Sciences and Engineering Research Council of Canada and Qutbec' s Fonds pour la Formation de Chercheurs et l'Aide ~ la Recherche. P. Boucher and R. Courtemanche received support from the Association Diabbte Qutbec. References 1. Grillner S. Control of locomotion in bipeds, tetrapods and fish. In: Brooks VB, editor. Handbook of physiology: the nervous system II. Bethesda (MD): American Physiological Society, 1981:1179236. 2. Cohen AH, Rossignol S, Grillner S, editors. Neural control of rhythmic movements in vertebrates. New York: Wiley, 1988. 3. Forssberg H. Stumbling corrective reaction: a phase-dependent compensatory reaction during locomotion. J Neurophysiol 1979;42: 936-53. 4. Berger W, Dietz V, Quintern J. Corrective reactions to stumbling in man: neuronal coordination of bilateral leg muscle activity during gait. J Physiol (Lond) 1984; 357:109-25. 5. Grillner S, Rossignol S. On the initiation of the swing phase in chronic spinal cats. Brain Res 1978; 146:269-77. 6. Mulder T, Berndt H, Pauwels J, Nienhuis B. Sensorimotor adaptability in the elderly and disabled. In: Stelmach GE, HiSmberg V, editors. Sensorimotor impairments in the elderly. Dordrecht: Kluwer Academic Publishers, 1993:413-26. 7. Shumway-Cook A, Baldwin M, Kerns K, Woollacott M. The effect of cognitive demands on postural control in elderly fallers and nonfallers. Society for Neuroscience Abstracts 1993; 19:990. 8. Teasdale N, Bard C, Dadouchi F, Fleury M, LaRue J, Stelmach GE. Posture and elderly persons: evidence for deficits in the central integrative mechanisms. In: Stelmach GE, Requin J, editors. Tutorials in motor behavior II. Amsterdam: North Holland, 1992:917-31. 9. Teasdale N, Bard C, Larue J, Fleury M. On the cognitive penetrability of posture control. Exp Aging Res 1993; 19:1-13. 10. Duysens J, Tax AAM, Vanderdoelen B, Trippel M, Dietz V. Selective activation of human soleus or gastrocncmius in reflex responses during walking and running. Exp Brain Res 1991;87:193-204. 11. Kahneman D. Attention and effort. Englewood Cliffs (NJ): PrenticeHall, 1973. 12. Baron A, Myerson J, Hale S. An integrated analysis of the structure and function of behavior: Aging and the cost of dividing attention. In: Davey G, Cullen C, editors. Human operant conditioning and behavior modification. New York: Wiley, 1988:139-66. 13. Craik FIM, Byrd M. Aging and cognitive deficits: The role of


14. 15. 16. 17. 18.

19. 20. 21. 22. 23. 24. 25.

26.

27. 28. 29. 30. 31. 32. 33. 34. 35.

36. 37. 38.

attentional resources. In: Craik FIM, Trehub SE, editors. Advances in the study of communication and affect: Aging and cognitive processes, vol 8. New York: Plenum, 1982:191-211. Cohen RA, Sparling-Cohen YA. Response selection and the executive control of attention. In: Cohen RA, editor. The neuropsychology of attention. New York: Plenum, 1993:49-73. Parasuraman R. Sustained attention: a multifactorial approach. In: Long J, Baddeley A, editors. Attention and performance IX. Hillsdale (NJ): Lawrence Erlbaum, 1981 ;493-511. Cohen RA. The Neuropsychology of attention. New York: Plenum, 1993. Bardy BG, Laurent M. Visual cues and attention demands in locomotor positioning. Percept Motor Skills 1991;72:915-26. Girouard Y. L'attention et l'acquisition de l'habilet6 motrice. In: Nadeau CH, Halliwell WR, Newell KM, Roberts GC, editors. Psychology of motor behavior and sport. Champaign (IL): Human Kinetics Publishers, 1980:535-52. Lajoie Y, Teasdale N, Bard C, Fleury M. Attentional demands for static and dynamic equilibrium. Exp Brain Res 1993;97:139-44. Thomas PK. Diabetic neuropathy: models, mechanisms and mayhem. Can J Neurol Sci 1992; 19:1-7. Cavanagh PR, Derr JA, Ulbrecht JS, Maser RE, Orchard TJ. Problems with gait and posture in neuropathic patients with insulindependent diabetes mellitus. Diabet Med 1992;9:469-74. Thomas PK, Brown MJ. Diabetic polyneuropathy. In: Dyck PJ, Thomas PK, Asbury AK, Winegrad AI, Porte D Jr, editors. Diabetic neuropathy. Philadelphia: Saunders, 1987:56-65. Simoneau GG, Ulbrecht JS, Derr JA, Becker MB, Cavanagh PR. Postural instability in patients with diabetic sensory neuropathy. Diabetes Care 1994; 17:1411-21. Boucber P, Teasdale N, Courtemanche R, Bard C, Fleury M. Postural stability in diabetic polyneuropathy. Diabetes Care 1995; 18: 638-45. Bergin PS, Bronstein AM, Murray NMF, Sancovic S, Zeppenfeld K. Body sway and vibration perception thresholds in normal aging and in patients with polyneuropathy. J Neurol Neurosurg Psychiatry 1995;58:335-40. Mueller MJ, Minor SD, Sahrmann SA, Schaaf JA, Strube MJ. Differences in the gait characteristics of patients with diabetes and peripheral neuropathy compared with age-matched controls. Phys Ther 1994;74:299-313. Courtemanche R, Teasdale N, Boucher P, Fleury M, Bard C. Gait and postural control in diabetic neuropathics. Society for Neuroscience Abstracts 1994;20:794. San Antonio Conference on Diabetic Neuropathy. Proceedings of a consensus development conference on standardized measures in diabetic neuropathy. Neurology 1992;42:1823-39. Valk GD, Nauta JJP, Strijers RLM, Bertelsman FW. Clinical examination versus neurophysiological examination in the diagnosis of diabetic polyneuropathy. Diabet Med 1992;9:716-21. Hugon M. Methodology of the Hoffmann reflex in man. In: Desmedt JE, editor. New developments in electromyography and clinical neurophysiology, vol 3. Basel: Karger, 1973:277-93. Ma DM, Liveson JA. Nerve conduction handbook. Philadelphia: Davis, 1985. Winter DA, Patla AE, Frank JS, Walt SE. Biomechanical walking pattern changes in the fit and healthy elderly. Phys Ther 1990;70: 340-7. Elble RJ, Sienko Thomas S, Higgins C, Colliver J. Stride-dependent changes in gait of older people. J Neurol 1991;238:1-5. Ferrandez AM, Paiihous J, Dump M. Slowness in elderly gait. Exp Aging Res 1990; 16:79-89. Larish DD, Martin PE, Mungiole M. Characteristic patterns of gait in the healthy old. In: Joseph JA, editor. Central determinants of age-related declines in motor function, vol 515. New York: The New York Academy of Sciences, 1988:18-32. Alexander NB. Postural control in older adults. J Am Geriatr Soc 1994; 42:93-108. Tinetti ME, Powell L. Fear of falling and low self-efficacy: a cause of dependence in elderly persons. J Gerontol 1993;48(Special Issue):35-48. Geurts ACH, Mulder TW, Nienhuis B, Rijken RAJ. Postural reorganization following lower limb amputation. Scand J Rehabil 1992; 24:83-90.


39. McPoil TG. Differences in the gait characteristics of patients with diabetes and peripheral neuropathy compared with age-matched controls: invited commentary. Phys Ther 1994;74:309-10. 40. Cavanagh PR. Differences in the gait characteristics of patients with diabetes and peripheral neuropathy compared with age-matched controls: invited commentary. Phys Ther 1994;74:310-2. 4l. Calne DB. Normal aging and the nervous system. In: Andres R,

855

Bierman L, Hazard WR, editors. Principles of geriatric medecine. New York: McGraw-Hill, 1985:23l-5.

Suppliers a. Grass Instruments Co., 101 Old Colony Ave, PO Box 516, Quincy, MA 02269. b. Bortec Electronics Inc., 7172 Sierra Morena Boulevard SW, Calgary, Alberta T3H 3G6 Canada.

Arch Phys Med Rehabil Vo177, September 1996