205

A Focused Exercise Regimen Improves Clinical Measures of

Balance in Patients With Peripheral Neuropathy
James K. Richardson, MD, David Sandman, BS, Steve Vela, BS
ABSTRACT. Richardson JK, Sandman D, Vela S. A focused exercise regimen improves clinical measures of balance
in patients with peripheral neuropathy. Arch Phys Med Rehabil
2001;82:205-9.
Objective: To determine the effect of a specific exercise
regimen on clinical measures of postural stability and confidence in a population with peripheral neuropathy (PN).
Design: Prospective, controlled, single blind study.
Setting: Outpatient clinic of a university hospital.
Participants: Twenty subjects with diabetes mellitus and
electrodiagnostically confirmed PN.
Intervention: Ten subjects underwent a 3-week intervention
exercise regimen designed to increase rapidly available distal
strength and balance. The other 10 subjects performed a control
exercise regimen.
Main Outcome Measures: Unipedal stance time, functional
reach, tandem stance time, and score on the activities-specific
balance and confidence (ABC) scale.
Results: The intervention subjects, but not the control subjects, showed significant improvement in all 3 clinical measures of balance and nonsignificant improvement on the ABC
scale.
Conclusion: A brief, specific exercise regimen improved
clinical measures of balance in patients with diabetic PN.
Further studies are needed to determine if this result translates
into a lower fall frequency in this high-risk population.
Key Words: Balance; Diabetes mellitus; Exercise; Peripheral nervous system diseases; Rehabilitation.
2001 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and
Rehabilitation
N PREVIOUS WORK, we found that older persons with
Iforperipheral
neuropathy (PN) are at a markedly increased risk
falls when compared with older persons with healthy pe1,2

ripheral nerves. This postural instability was confirmed in the

laboratory setting. Subjects with PN balanced less reliably on 1
foot for 3 seconds than did matched control subjects without
PN.3 In addition, a decreased unipedal stance time among
persons with PN has, in 2 separate studies, been associated with
a history of falls over the previous year.2,4 It has also been
noted that diabetic subjects with PN, as identified by decreased

From the Department of Physical Medicine and Rehabilitation (Richardson), University of Michigan (Sandman, Vela), Ann Arbor, MI.
Accepted in revised form May 23, 2000.
Supported by the University of Michigan Department of Physical Medicine, Public
Health Service (grant no. AG-08808), and the University of Michigan Geriatrics
Research and Training Center.
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.
Reprint requests to James K. Richardson, Dept of Physical Medicine and Rehabilitation, 1500 E Medical Center Dr, D5200, University of Michigan Health Systems,
Ann Arbor, MI 48109-0718, e-mail: jkrich@umich.edu.
0003-9993/01/8201-6014$35.00/0
doi:10.1053/apmr.2001.19742

distal vibratory thresholds, were 15 times more likely to report

an injury from a fall than a control diabetic group.5 Furthermore, PN is common among older persons, particularly those
with diabetes mellitus. Among people over 60 with type 2
diabetes mellitus, the prevalence of PN is greater than 50%.6
Taken together, the data suggest that PN is common among
older persons and markedly increases their fall risk.
We have identified specific, distal sensory and motor impairments in older persons with PN that appear to underlie their
postural instability. The sensory impairments identified are
increases in ankle inversion and eversion proprioceptive
thresholds, which are about 5 times greater in older persons
with PN compared with older persons without PN (1.46 vs
0.3).7 As a result, older persons with PN are likely less able to
perceive ground irregularities and subtle shifts in their centers
of mass, and are, therefore, predisposed to falls. More recently,
we8 identified a distal motor impairment about the ankle in
women who had PN and diabetes. These women showed a
significantly decreased ankle rate of torque development compared with the age-matched women with diabetes but no PN
(78.2 50.8N m/s vs 152.7 54.6N m/s, p .016). A
second outcome was the ability to recover balance on 1 foot
when released from a lateral leaning posture (quantified as a
percentage of foot width). Three of the 6 women with diabetes
but no PN were able to recover from a 5% lean, whereas none
of the women with PN was able to recover her balance ( p
.083). These findings suggest that older women with diabetes
and PN have impaired ability to rapidly develop torque at the
ankle, which has an impact on balance.
It is not known if an exercise regimen will improve the
balance impairments identified in older persons with PN.
Therefore, our primary hypothesis was that older persons with
PN who perform an exercise regimen designed to increase
rapidly available ankle strength would show improved balance,
as reflected by increased functional reach, as well as tandem
and unipedal stance times, compared with those who perform a
control exercise regimen. Our secondary hypothesis was that
subjects who performed the intervention exercises would show
greater confidence, on a validated scale, in their mobility skills
(vs the control group).
METHODS
Subjects
The study was approved by the institutions review board.
All subjects gave written and verbal consent.
Inclusion criteria included: (1) being between 50 and 80
years old; (2) a known history of diabetes mellitus treated by
diet, oral hypoglycemic, or insulin therapy; (3) lower extremity
symptoms consistent with PN; (4) ability to walk household
distances without assistance or an assistive device (though
subjects may use a cane intermittently in the community); (5)
willingness to participate in the study; (6) strength of ankle
dorsiflexors, invertors, and evertors at least antigravity (grade 3
or greater by manual muscle testing); and (7) conclusive electrodiagnostic evidence of a diffuse, primarily axonal, peripheral polyneuropathy as evidenced by:
Arch Phys Med Rehabil Vol 82, February 2001

206

NEUROPATHY, BALANCE, AND EXERCISE, Richardson

(a) sural response: absent or decreased amplitude (6V)

with a normal or minimally prolonged distal latency
(5ms) stimulating 14cm from the recording site posterior to the lateral malleolus.9 If the sural response was
absent bilaterally, the motor responses were not performed.
(b) peroneal or tibial responses: absent or decreased in amplitude (2mV for peroneal, 3mV for tibial) with a
normal distal latency (6.2ms stimulating 9 and 8cm
from recording sites over the extensor digitorum brevis
and abductor hallicus muscles, respectively).9
Exclusion criteria included: (1) a history or evidence on
physical examination of significant central nervous system
dysfunction (ie, hemiparesis, myelopathy, cerebellar ataxia);
(2) significant musculoskeletal deformity (ie, amputation, scoliosis, abnormality of range of motion [ROM]) that would
prevent participation (90 of humeral abduction, inability to
grip, 10 of combined ankle inversion/eversion); (3) lower
extremity arthritis or pain that limits standing or weight-bearing exercise; (4) electrodiagnostic evidence of any diagnosis
other than PN; (5) a history or evidence on physical examination of vestibular dysfunction; (6) a history of angina or anginaequivalent symptoms (ie, nausea, diaphoresis, shortness of
breath with exercise); (7) symptomatic postural hypotension
(postural lightheadedness that interferes with standing for
5min); and (8) a history or evidence on physical examination of
plantar skin pressure ulcer.
The subjects also were evaluated by using the Michigan
Diabetes Neuropathy Score (MDNS).10 This is a 46-point scale
(0 46, with higher score reflecting more severe PN) that has
been shown to correlate well with more extensive neuropathy
staging scales. The scale includes muscle stretch reflexes at the
biceps, triceps, patella, and Achilles; pinprick sensation at the
great toe; ability to perceive the touch of a 10-gram monofilament; ability to perceive a 128Hz tuning fork at the great toe;
strength of hand dorsal interossei; great toe extension; and
ankle dorsiflexion. The first 10 subjects recruited were placed
in the intervention group, and the next 10 subjects were placed
in the control group.
Intervention Exercises
The exercise interventions, performed daily on a firm surface
for 3 weeks, included:
(1) Warm up (open chain active ankle ROM exercises).
Subjects wrote the alphabet in the air with each foot by
moving the ankle.
(2) Bipedal toe raises and heel raises (lifting the forefoot as
one does to balance on a heel). Subjects did these as
quickly as possible, using support as necessary. Subjects
started with 1 set of 10 and increased by 1 set every 5
exercise sessions for a total of 3 sets.
(3) Bipedal inversion and eversion. In this exercise, subjects center of mass was shifted laterally as subjects
strengthened ankle invertors and evertors via closed
chain exercises. The goal was to do so without using the
upper extremities, but support was used as necessary.
Subjects started with 10 repetitions in each direction and
increased to 2 sets of 10 repetitions after 5 exercise
sessions.
(4) Unipedal toe raises and heel raises. Again, subjects
attempted to perform this quickly even if that was not
possible. Subjects started with 5 repetitions of each
exercise and increased to 10 repetitions after 5 exercises
and then to 2 sets of 10 after 10 exercise sessions.
(5) Unipedal inversion and eversion. Subjects inverted and
everted the foot while standing on it to challenge balance
Arch Phys Med Rehabil Vol 82, February 2001

and to create a closed chain exercise of the ankle invertors and evertors. It was anticipated that most subjects
would find this task challenging and so they used their
hands for balance when needed. Subjects started with 1
set of 5 repetitions in each direction and increased to 10
repetitions after 5 exercise sessions.
(6) Wall slides. Subjects started with bipedal slides with
knee flexion maximum of about 45. They performed 3
sets of 10. After 5 exercise sessions the first set was
performed on each foot.
(7) Unipedal balance for time. Three tries on each foot.
Control Exercises
The control exercise regimen was performed in a seated
position. Subjects performed neck flexion and rotation stretching with eyes open and then closed. They then used a resistance
band to perform strengthening exercises for the scapular abductors, shoulder external rotators, and elbow flexors. The
exercises were performed 5 or more times (if tolerated) per
week for 3 weeks. Control exercises were suggested at a
slightly decreased frequency because of concern that they
might lead to an overuse injury, which was not felt to be an
ethical risk from a control intervention.
Outcomes
All subjects underwent 3 trials of tandem stance, functional
reach, and unipedal stance before and after their exercise
programs. Tandem stance, functional reach, and unipedal
stance were performed and graded as described elsewhere.2,11,12 In addition, all subjects filled out the activitiesspecific balance confidence (ABC) scale13 before and after the
exercise regimen. The ABC scale lists 16 activities (eg, walking up and down stairs, walking on an icy sidewalk) and
subjects describe their degree of confidence in performing each
activity, on a scale from 0% (no confidence) to 100% (complete confidence). All subjects were evaluated before and after
their 3-week exercise programs by the same examiner (DS).
Statistical Analysis
A paired, 2-tailed t test was used to detect significant
changes in tandem stance, functional reach, and unipedal
stance. A p value of less than .05 was considered significant
and a p value of .05 or greater and less than .10 was considered
a trend. A 2-tailed t test was also used for evaluating the
responses to the 16 activities on the ABC scale; however, to
compensate for making multiple comparisons, p less than .0125
was considered significant and p .125 or greater and less than
.025 was considered a trend.
RESULTS
Nine of the 10 intervention subjects and 7 of the 10 control
subjects completed the study. The intervention subject dropped
out because of foot-ankle pain, which was attributed to the
exercise regimen aggravating an underlying arthritis. One of
the 3 control subjects developed an illness and 2 dropped out
without specifying a reason. Subject characteristics of gender
and age are listed in table 1. There was a trend toward an
increased MDNS score, representing more severe PN, among
the intervention subjects compared with the control subjects
(table 1). There was no significant difference between intervention and control subjects sural, peroneal motor, or tibial
motor response amplitudes.
The 2 groups showed grossly similar baseline values for
tandem stance and functional reach. The shorter unipedal
stance time at baseline in the intervention group, compared

207

NEUROPATHY, BALANCE, AND EXERCISE, Richardson

Table 1: Clinical Characteristics of Subjects

Age (yr) (mean SD)

Gender (% men)
MDNS score (SD)*
Response amplitudes
Sural (V)
Peroneal motor (mV)
Tibial motor (mV)

Intervention Group
(n 9)

Control Group
(n 7)

64.0 6.3
8 (89%)
18.6 5.3

63.3 7.6
4 (57%)
11.9 3.0

NS
NS
.060

.25 .79
.57 .63
.18 .33

.15 .47
.33 .43
.50 1.1

.34
.11
.27

Abbreviation: NS, not signicant.

* MDNS scores: higher score more severe neuropathy (maximum 46).

Represents a trend toward more severe neuropathy in the intervention group versus the control group.

with the control group ( p .11), was consistent with the trend
toward an increased MDNS score in the intervention group;
however, none of the baseline differences was significant. The
intervention subjects showed a significant improvement in all 3
outcomes after the exercise regimen (table 2). In contrast, the
control subjects showed insignificant improvements in tandem
stance and functional reach, and an insignificant decrease in
unipedal stance time (table 2).
The intervention and control groups showed nearly identical
initial composite confidence scores (table 3). Although the
intervention group reported improved confidence, as reflected
by a greater composite ABC score postintervention, the difference was not significant. There was no significant change in the
control group confidence scores. When confidence scores for
isolated activities were considered, there was a trend toward
the intervention subjects reporting greater confidence in climbing and descending stairs after the intervention exercises (table
3). The intervention group also reported a 10% or greater
improvement in confidence for 3 other activities (bending over
to pick up an object, standing on tip toes to reach overhead,
walking on icy sidewalks), but the changes were not significant. The control group showed no significant changes or
trends toward a change after their exercise program. There was
a change of 10% toward greater confidence among the control
group for bending over to pick up an object.
DISCUSSION
The data from this study showed that an exercise regimen
designed to increase rapidly available ankle strength improved
3 commonly used clinical measures of balance (functional
reach, tandem stance, unipedal stance) among older persons
with mild to moderate PN. The study further showed that these
improvements develop in a relatively short period of time and
that the exercise regimen is well tolerated. The data also
suggested, but did not confirm, that the exercise regimen is
associated with improvements in subjects confidence in their
abilities to perform daily tasks that challenge balance.
There was evidence to suggest that exercise is a reasonable
intervention, even in those with predominantly sensory PN.
Aside from our study, which showed impaired ankle rate of
torque development among subjects with clinically normal
ankle strength,8 other work suggests that patients with clinically mild or sensory-only PN likely have motor impairments
about the ankle. In an electrophysiologic study14 of subjects
selected for clinical evidence of sensory-only PN, most subjects (70%) had abnormalities during needle electromyography
of the anterior tibialis or medial gastrocnemius muscles, both
of which provide torque to the ankle. Those investigators14

concluded that subclinical motor involvement is often detected on electrophysiologic studies in patients . . . who have
only sensory signs. Also, there is evidence in animal models15
and in humans16 that the number of type II muscle fibers not
only decreases with age, but that there is a preferential loss of
type II motor units in the setting of denervation, particularly
distal denervation.17 Therefore, it appears likely that a generalized PN, even among patients with predominantly sensory
findings, is associated with motor deficits characterized by
decreased rapidly available torque at the ankle. This impairment, though important for postural stability under challenging
circumstances, is likely difficult or impossible to detect on
routine physical examination and, therefore, remains subclinical.
Because functional reach and tandem stance were not part of
the intervention exercise regimen, improvement in these outcomes suggests that some change occurred beyond a practice
effect. It is possible that the intervention subjects increased
their ankle strength in response to the exercise regimen. Brown
et al18 found that type II muscle fiber concentration significantly increased in older persons undergoing rigorous strengthening programs. The intervention applied in our study was
designed to increase rapidly available torque about the ankle19
and recruit type II motor units.20 Brown18 tested the biceps, a
proximal upper extremity muscle, and, therefore, their findings
may not apply to distal lower extremity muscles, which were
the target of strengthening in our study. Another mechanism of
strengthening is possible. Early strength gains appear to be
related to neural changespossibly improved synchronization
of motor unitsrather than muscle hypertrophy.21 Strengthening has, therefore, been found to occur in response to exercise
in diseases such as hereditary motor and sensory neuropathy
types I and II22 and postpoliomyelitis syndrome,23 which decrease available motor units in a manner similar to that of a
generalized PN. Given the brevity of the exercise intervention,
any strengthening that occurred in the intervention group was
more likely related to a synchronization of motor units rather
than muscle hypertrophy.
An isolated improvement in the strength of the ankle musculature would likely be sufficient to lead to the improvements
noted in this study. Others have found that increased muscle
strength among older subjects was an independent predictor of
a decreased risk for loss of balance during a difficult test of
balance that reduced proprioceptive input, a condition that
mimics the patient with PN.24 Wolfson et al25 emphasized the
strong association between falls/loss of balance and decreased
ankle strength among nursing home residents. However, an
improvement in ankle muscle strength, and, therefore, muscle
tension, may also improve ankle proprioceptive thresholds. In

Table 2: Change in Clinical Measures of Balance

Pre- and Postexercise

Intervention Group
Tandem stance (s)
Functional reach (in)
Unipedal stance (s)
Control Group
Tandem stance (s)
Functional reach (in)
Unipedal stance (s)

Preexercise

Postexercise

p*

17.5 13.4
10.5 2.1
5.4 4.7

23.5 10.9
11.5 2.2
11.6 10.2

.004
.0012
.0014

19.0 11.8
11.3 3.6
9.3 8.6

22.0 12.0
11.9 2.8
7.9 5.9

.13
.23
.33

NOTE. Values presented as mean SD.

* Two-tailed t test.

Arch Phys Med Rehabil Vol 82, February 2001

208

NEUROPATHY, BALANCE, AND EXERCISE, Richardson

Table 3: Change in Subject Responses on the ABC Scale
Intervention Group

Question 2*
Question 3
Question 5
Question 16
Mean of all questions

Control Group

Before

After

Before

After

70 28
81 24
78 26
51 38
80 21

83 16
93 9
90 13
67 32
88 11

.022
.093
.093
.048
.14

76 14
76 29
77 30
60 36
80 21

81 21
86 18
76 28
62 32
80 20

.36
.11
.36
.45
.64

NOTE. Values presented as %, in mean SD.

* 100% complete condence for activity; 0% no condence for activity. See text for activities.

Represents a trend toward intervention subjects reporting greater condence after the exercise regimen.

another study, perception of ankle plantar- and dorsiflexion

improved markedly when the calf musculature was tested under tension compared with when the musculature was relaxed.26
Among the present studys strengths is the careful subject
selection, using electrodiagnostic criteria to confirm the presence of PN, and the use of a control group with a control
exercise regimen. In addition, the outcomes are thought to have
clinical meaning; impairments in 2 of the 3 outcomes unipedal stance and functional reach have been associated with
injurious falls.12,27 Our study was well controlled, with the
control subjects receiving attention equivalent to the intervention group. In addition, the control subjects showed their greatest improvement on the ABC scale for a task involving gross
head motion, similar to some of their exercises, suggesting the
possibility that those subjects found the control exercises
meaningful.
Our studys greatest limitation was its design: it was not
double-blind. Although the tests were administered as objectively as possible, bias is possible. This seems less likely for
unipedal stance, which requires little judgment on the examiners part. The decreased MDNS score and increased unipedal
stance time of the control subjects1 preexercise suggests that
they had less severe PN than the intervention subjects. However, the control subjects still had much room to improve
particularly in tandem and unipedal stanceso that a ceiling
effect does not appear to be an explanation for their lack of
improvement. Although the subjects were carefully selected
and had similar baseline characteristics, the relatively small
numbers diminish the strength of the conclusions. A larger
study is planned to confirm the findings of our work. A last
concern is that most subjects were men. Therefore, translating
these results to women should be performed with caution.
CONCLUSION
A brief, intense exercise regimen designed to improve distal
lower extremity strength was well tolerated and improved 3
clinical parameters of balance in a group of older persons with
PN. Although increased confidence or distal lower extremity
strength may be responsible for these findings, the study provided no clear insight into the mechanism of the improvements
seen in the intervention group. Whether the improvements in
clinical balance noted in the intervention subjects translate into
decreased fall risk in daily life is unknown. However, given the
minimal risk from the intervention exercises and the magnitude
of the benefit from preventing falls in the population studied, it
appears reasonable for clinicians to consider prescribing these
exercises for their patients with postural instability caused
by PN.
Arch Phys Med Rehabil Vol 82, February 2001

References
1. Richardson JK, Ching C, Hurvitz EA. The relationship between
electromyographically documented peripheral neuropathy and
falls. J Am Geriatr Soc 1992;40:1008-12.
2. Richardson JK, Hurvitz EA. Peripheral neuropathy: a true risk
factor for falls. J Gerontol A Biol Sci Med Sci 1995;50:M211-5.
3. Richardson JK, Ashton-Miller JA, Lee SG, Jacobs K. Moderate
peripheral neuropathy impairs weight transfer and unipedal balance in the elderly. Arch Phys Med Rehabil 1996;77:1152-6.
4. Hurvitz EA, Richardson JK, Werner RA, Ruhl A, Dixon M.
Unipedal stance testing as an indicator of fall risk among older
outpatients. Arch Phys Med Rehabil 2000;81:587-91.
5. Cavanagh PR, Derr JA, Ulbrecht JS, Maser RE, Orchard TJ.
Problems with gait and posture in neuropathic patients with insulin-dependent diabetes mellitus. Diabetic Med 1992;9:469-74.
6. Young M, Boultin A, Maclead A, Williams D, Sonksen P. A
multicentre study of the prevalence of diabetic peripheral neuropathy in the United Kingdom hospital clinic population. Diabetologia 1993;36:150-4.
7. Van de Bosh C, Gilsing MG, Lee SG, Richardson JK, AshtonMiller JA. Peripheral neuropathy effect on ankle inversion and
eversion thresholds. Arch Phys Med Rehabil 1995;76:850-6.
8. Gutierrez MS, Helber MB, Dealva D, Ashton-Miller, Richardson
JK. Mild diabetic neuropathy affects ankle motor function. Diabetes Care. In press.
9. Liveson JA, Ma DM, editors. Laboratory reference for clinical
neurophysiology. Philadelphia: FA Davis; 1992.
10. Feldman EL, Stevens MJ, Thomas PK, Brown MB, Canal N,
Greene DA. A practical two-step quantitative clinical and eletrophysiological assessment for the diagnosis and staging of diabetic
neuropathy. Diabetes Care 1994;17:1281-9.
11. Berg KO, Wood-Dauphinee SL, Williams JI, Maki B. Measuring
balance in the elderly: validation of an instrument. Can J Public
Health 1992;83(Suppl 2):S7-11.
12. Duncan PW, Studenski S, Chandler J, Prescott B. Functional
reach: predictive validity in a sample of elderly male veterans. J
Gerontol A Biol Sci Med Sci 1992;47:M93-8.
13. Powell LE, Myers AM. The activities-specific balance confidence
(ABC) scale. J Gerontol A Biol Sci Med Sci 1995;50:M28-34.
14. Bril V, Werb MR, Greene DA, Sima AAF. Single-fiber electromyography in diabetic peripheral polyneuropathy. Muscle Nerve
1996;19:2-9.
15. Kadhiresan VA, Hassett CA, Faulkner JA. Properties of single
motor units in medial gastrocnemius muscles of adult and old rats.
J Physiol 1996;493(Pt 2):43-52.
16. Hakkinen K, Kraemer WJ, Kallinen M, Linnamo V, Pastinem
UM, Newton RU. Bilateral and unilateral neuromuscular function
and muscle cross-sectional area in middle-aged and elderly men
and women. J Gerontol A Biol Sci Med Sci 1996;51:21-9.
17. Bishop DL, Milton RL. The effects of denervation location on
fiber type mix in self-reinnervated mouse soleus muscles. Exp
Neurol 1997;147:151-8.
18. Brown AB, McCartney N, Sale DG. Positive adaptations to
weight-lifting training in the elderly. J Appl Physiol 1990;69:
1725-33.

NEUROPATHY, BALANCE, AND EXERCISE, Richardson

19. Behm DG, Sale DG. Intended rather than actual movement velocity determines velocity specific training response. J Appl
Physiol 1993;74:359-68.
20. Vandervoort AA, Sale DG, Moroz J. Comparison of motor unit
activation during unilateral and bilateral leg extension. J Appl
Physiol 1984;56:46-51.
21. Moritani T, de Vries HA. Neural factors versus hypertrophy in the
time course of muscle strength gain. Am J Phys Med 1979;58:
115-30.
22. Lindeman E, Leffers P, Spaans F, Drukker J, Reulen J, Ierchkoffs
M, et al. Strength training in patients with myotonic dystrohy and
hereditary motor and sensory neuropathy: a randomized clinical
trial. Arch Phys Med Rehabil 1995;76:612-20.
23. Agre JC, Rodriguez AA, Harmon RL, Swiggum ER, Franke TM.
Strengthening exercise can improve muscle function in post-polio

24.
25.
26.
27.

209

subjects without detectable adverse effect upon the surviving

motor units or muscle [abstract]. Arch Phys Med Rehabil 1995;
76:1036.
Judge JO, King MB, Whipple R, Clive J, Wolfson LI. Dynamic
balance in older persons: effects of reduced visual and proprioceptive input. J Gerontol A Biol Sci Med Sci 1995;50:M263-70.
Wolfson L, Judge J, Whipple R, King M. Strength is a major
factor in balance, gait, and the occurrence of falls. J Gerontol A
Biol Sci Med Sci 1995;50:64-7.
Refshauge KM, Fitzpatrick RC. Perception of movement at the
human ankle: effects of leg position. J Physiol 1995;488(Pt 1):
243-8.
Vellas BJ, Wayne SJ, Romero L, Baumgartner RN, Rubenstein
LZ, Garry PJ. One-leg balance is an important predictor of injurious falls in older persons. J Am Geriatr Soc 1997;45:735-8.

Arch Phys Med Rehabil Vol 82, February 2001