Lower-Extremity Muscle Force and Balance Performance in Adults Aged 65 Years and Older.

The ability to maintain control of posture is critical for the successful performance of most daily activities. Postural control is defined as the maintenance of the body's center of gravity within its base of support during stance or voluntary movements and in response to external perturbations.[1] Visual, vestibular, and somatosensory signals are sent to the central nervous system, which can adjust body sway and posture by integrating this information and by controlling skeletal muscles to appropriately generate joint torques and adjust joint angles. Impairment in any component of the postural control system can lead to instability and falls in older people.[2]

Postural control is complex, and no single comprehensive measure is available that tests all aspects of the postural control system. Measures that are sensitive and easily administered in the clinical setting and that yield reliable and valid measurements are needed in order to predict who is at risk of falling and to evaluate the effectiveness of interventions in improving postural control in older adults. Three commonly used tools for the measurement of balance impairment are the Berg Balance Scale (BBS), the Functional Reach Test (FRT), and the Timed Get Up & Go Test (GUG). The BBS was developed to measure balance impairments in elderly people and people with neurological disorders. The scale consists of 14 common functional activities that are scored from 0 to 4, where 0 indicates an inability to perform the task and 4 indicates the task was performed correctly and independently.[3] The FRT was designed to measure the limits of stability in an anterior direction.[4] The maximal distance that subjects can reach forward horizontally while maintaining a fixed base of support is measured. The GUG was designed as a quick measure of basic balance and mobility skill in elderly people. The time taken for subjects to rise from an armchair, walk 3 m, turn, and return to the chair is measured in seconds.[5,6]

These tools were all designed to provide measurements of balance in older adults with balance impairment. All 3 tests measure the global output of the postural control system but do not provide an indication of underlying impairments. The tests incorporate different tasks, and although used for the same purpose, it is unlikely that they measure the same components of postural control.

In both cross-sectional and longitudinal studies,[7,8] lower-extremity muscle weakness has been identified as a risk factor contributing to falls in older people. Balance and muscle force deteriorate with aging,[1,9-12] and it has been suggested that a decrease in the ability to generate force in the lower-extremity muscles contributes to balance impairment.[1] The relationship between impairments in muscle force generation and balance, however, has not been extensively investigated.[13,14] Wolfson et al[14] found that torques generated by the ankle muscles were reduced in older adults who were identified as having the greatest balance impairment on the Sensory Organization Test. Iverson et al,[13] however, reported a relationship between hip muscle force and single-limb stance time and between hip muscle force and sharpened Romberg Test scores. The degree to which impairment in one component of postural control (muscle force) is predictive of scores on commonly used functional measures of balance has not been reported. The purpose of this study was to determine the extent to which measures of lower-extremity muscle force are predictive of scores on the BBS, the FRT, and the GUG in community-dwelling older people. A secondary objective was to determine whether muscle force values differed between a subgroup identified as having had one or more unexplained falls in the year previous to the study and those who had not fallen and whether force values were predictive of fall status.

Method

Subjects

Participants in the study were community-dwelling volunteers who were 65 years of age or older. Participants were ambulatory and did not have any medical history of low back or lower-extremity pathology, any diagnosed vestibular or central nervous system pathology, postural hypotension, cognitive impairment severe enough to interfere with the ability to follow instructions, or any other medical conditions that may have affected their ability to participate in the study. Two participants (nonfallers) used a cane for ambulation in the community. Volunteers were recruited through advertisements posted in seniors centers and placed in the local newspaper.

Seventy-four volunteers were initially screened over the telephone. Twenty-four people were excluded, either after telephone screening or following the history-taking session and examination, because they did not meet the selection criteria or were unable to participate in the study. Five of these individuals were away during the testing period, and 8 individuals were not interested in participating after having details of the study explained to them. One potential participant had a vestibular disorder, 4 people had Meniere disease, 2 people had a previous cerebrovascular accident, and 3 people were ill during the testing period. Data from 1 subject was not used due to equipment failure during testing. Thus, data were obtained for 50 subjects between 65 and 91 years of age (25 men with a mean age of 72.8 years [SD=6.08] and 25 women with a mean age of 76.9 years [SD=5.52]) who met the inclusion and exclusion criteria. Seven women and 4 men reported one or more falls in the 12-month period prior to the date of recruitment into the study that were not related to a known intrinsic event (eg, acute medical illness) or overwhelming hazard (eg, slip on ice). Seven of these subjects reported falling once in the 12 months previous to the study, and 4 subjects had 2 or more falls. There were 18 women and 21 men in the group reporting no falls.

The mean age, height, weight, and scores on the 3 balance measures for subjects who did and did not report falls are shown in Table 1. There was no difference between the subjects who fell and the subjects who did not fall for any of these variables. Data from subjects who fell, however, were omitted from the regression analysis to determine predictors of scores on the 3 balance measures in order to improve homogeneity of the sample.

Table 1. Descriptive Characteristics of Subjects and Balance Measures(a)

              Subjects With Falls (n=11)

              [bar] X   SD      Range

Age (y)        73.00     5.44     65-81
Height (cm)   169.64    11.24    153-192
Weight (kg)    79.36    14.06     66-103
BBS            49.00     3.85     44-55
GUG (s)        13.15     3.08   7.47-18.05
FRT (cm)       27.77     7.08     13-40

              Subjects Without Falls (n=39)

              [bar] X   SD      Range

Age (y)        75.33     6.26     65-91
Height (cm)   167.33    13.32    123-189
Weight (kg)    74.28    17.30     38-115
BBS            49.87     7.08     26-56
GUG (s)        11.13     3.70   6.08-23.33
FRT (cm)       29.33     8.06      3-43

              Independent
              t Test

              P

Age (y)       NS
Height (cm)   NS
Weight (kg)   NS
BBS           NS
GUG (s)       NS
FRT (cm)      NS

(a) NS=no significant difference between groups (independent t test), BBS=Berg Balance Scale (maximal score=56); GUG=Timed Get Up & Go Test; FRT=Functional Reach Test.

Measurements

One investigator was responsible for obtaining all outcome measurements (MED). For safety reasons, an assistant was always present during balance testing. The 3 balance measures were the BBS, the FRT, and the GUG. The 14 tasks scored on the BBS are (1) sit-to-stand, (2) standing unsupported, (3) sitting unsupported, (4) stand-to-sit, (5) transfers, (6) standing with eyes closed, (7) standing with feet together, (8) reaching forward with an outstretched arm, (9) retrieving an object from the floor, (10) turning the trunk with feet fixed, (11) turning 360 degrees, (12) stool stepping, (13) tandem standing, and (14) standing on one limb. The maximal possible score on this test is 56. Reliability for this measure was previously established in a study involving 14 subjects, aged 65 years and over, with varying degrees of balance impairment who were scored by 5 physical therapists with experience in geriatric rehabilitation.[3] Intraclass correlation coefficients (ICCs) for interrater and intrarater reliability were .98 and .99, respectively.[3]

Functional reach was measured as the maximal distance that subjects could reach forward horizontally while maintaining a fixed base of support. The distance was measured (in centimeters) on a tape measure fixed to a wall. Test-retest reliability for this measure was established in 128 subjects between the ages of 20 and 87 years (ICC=.92).[4] The ICC for interobserver measures was .98. For the GUG, the time taken for subjects to rise from an armchair, walk 3 m, turn, and return to the chair was measured (in seconds)[5,6] Reliability for this test was previously demonstrated in a study of 2 frail, community-dwelling subjects over 70 years of age (intrarater ICC=.99, interrater ICC=.99).[5]

Lower-extremity muscle force was evaluated using a handheld dynamometer.(*) In order to improve the stabilization of the handheld dynamometer, a stabilization frame was designed and manufactured. This frame was used for testing all muscle groups tested in supine and long-sitting positions. The frame could be attached to the plinth in 1 of 2 positions. There were guide ropes connected to the handheld dynamometer and strung through adjustable cam cleats on the stabilization frame in order to provide additional support and stabilization to the researcher when holding the dynamometer (Figure).

[Figure ILLUSTRATION OMITTED]

All tests done were "make" tests where the dynamometer was held stationary by the examiner and ropes while the subjects exerted a maximum force against it. All muscle groups were tested in midrange of joint motion. The dynamometer was always held perpendicular to the limb segment, and, where appropriate, the subject positioning and dynamometer placement were done as described by Bohannon.[15] After the subject was positioned and stabilization was achieved, the subject was asked to flex and extend his or her entire lower extremity; to abduct, adduct, and medially (internally) and laterally (externally) rotate the hip; and to dorsiflex and plantar flex the ankle and invert and evert the subtalar joints actively as a warm-up. One practice trial was given prior to testing for each movement. The mean of the 3 measurements was used for data analysis. Each trial lasted 4 to 5 seconds so the subjects could be instructed to increase their force to maximum over a few seconds' time. Instructions were standardized for each test, and the digital display on the dynamometer was covered to minimize experimenter bias. The handheld dynamometer was connected to a strip chart recorder to provide a printout of the force output.

The force of the hip flexor and extensor muscles was tested with each subject positioned supine with the hip flexed to 90 degrees (Figure). The subject was positioned on the testing plinth with stabilization straps placed over the trunk and pelvis as well as around the contralateral midthigh. A research assistant helped stabilize the subject further, if necessary, to prevent substitution and compensation when performing the movements. The subject's leg was supported by the researcher. The lower extremity not being tested was supported with a pillow under the knee. For flexion, the end piece of the dynamometer was applied to the anterior surface of the thigh just proximal to the condyles of the femur. For extension, the end piece of the dynamometer was applied to the posterior surface of the thigh just proximal to the knee, superior to the popliteal fossa.

Hip abductor muscle force was tested with the subject positioned supine with the hip in neutral extension, abduction, and rotation and with the knee extended. The stabilization technique was the same as for the testing of the hip flexor and extensor muscles. The end piece of the dynamometer was applied to the lateral surface of the thigh just proximal to the lateral condyle of the femur. Hip adductor muscle force was tested with the subject positioned supine with the hip in neutral extension, in 30 degrees of abduction, and in neutral rotation and with the knee extended. The end piece of the dynamometer was applied to the medial surface of the thigh just proximal to the medial condyle of the femur.

Knee extensor and flexor muscle forces were tested with the subject positioned supine with the hip flexed to 45 degrees and the knee flexed to 90 degrees. The stabilization method was the same as for the testing of the hip flexor and extensor muscles. The knee was placed over a padded portion of the stabilization frame for testing the extensor muscles and over a knee roll for testing the flexor muscles, with the foot over the end of the plinth. The end piece of the dynamometer was applied to the anterior surface of the distal tibia for extension and to the posterior surface of the distal calf, just proximal to the malleoli, for flexion.

Ankle plantar-flexor muscle force was tested with the subject positioned supine with the hip and knee extended and the ankle in neutral dorsiflexion. Ankle dorsiflexor force was tested with the subject in a long-sitting position with the hip flexed between 70 and 80 degrees and the knee extended. The end piece of the dynamometer was applied to the plantar surface of the foot, proximal to the metatarsophalangeal joints, for measurement of plantar-flexor muscle force and to the dorsal surface of the foot, just proximal to the metatarsophalangeal joints, for measurement of dorsiflexor muscle force.

Subtalar invertor and evertor muscle force measurements were obtained with the subject in a long-sitting position with the hip flexed to between 70 and 80 degrees, the knee extended, and the ankle in neutral dorsiflexion. The end piece of the dynamometer was applied to the medial surface of the foot, proximal to the metatarsophalangeal joints, for testing of invertor muscle force and to the lateral surface of the foot, just proximal to the metatarsophalangeal joints, for testing the force of the evertor muscles.

Hip medial and lateral rotation were tested with the subject in a sitting position with the hip and knee flexed to 90 degrees and with the hip in neutral rotation. The subjects were stabilized when sitting on the testing plinth with straps over the pelvis and their thighs fully supported on the plinth. The end piece of the dynamometer was applied to the lateral surface of the fibula, just proximal to the lateral malleolus, and to the medial surface of the tibia,just proximal to the medial malleolus of the ankle, for testing of hip medial and lateral rotator muscle force, respectively.

Reliability of Force Measurements

Although some handheld dynamometer force measurements have been found to be reliable for a number of muscle groups in some types of patients and, in particular, in older subjects,[16-18] we believed it is important that reliability be determined for each rater.[15] In order to determine the intrarater reliability of the force measurements, a subset of 11 subjects (5 men and 6 women) were tested twice, 7 days apart. These subjects had a mean age of 75.46 years (SD=5.09, range=65-81), a mean weight of 76.5 kg (SD=8.88, range=65-93), and a mean height of 168.46 cm (SD=11.52, range=150-182). All of these subjects volunteered to return for additional testing and tended to be the more active individuals in the study. None reported falls during the previous year. All of the muscle groups were tested in random order and then retested 1 week later in the same order. Intrarater reliability was established for force measurements using ICCs (2,1).[19] Variances for the calculation of the ICCs were obtained from the results of a 2-way analysis of variance There was no difference in force measurements of any muscle groups between the 2 days, and intrarater ICC values ranged from 70 to .96. All muscle force measurements, therefore, were considered to have acceptable reliability (Tab. 2).

Table 2. Intratester Reliability of Muscle Force Measurements Obtained With a Handheld Dynamometer

              Muscle Force (kg)

              Day 1

Variable(a)   [bar] X   SD     Range

HIPFLEX       14.49     3.90    9.10-20.72
HIPEXT        31.09     4.95   23.86-39.20
HIPABD        10.21     5.09    6.70-24.72
HIPADD        13.74     3.57    9.10-20.94
HIPMR         10.92     3.12    7.38-16.08
HIPLR          7.99     2.52    4.12-12.02
KNEEFLEX      12.79     3.12    9.16-18.94
KNEEEXT       18.41     4.85   10.82-27.81
DF            13.22     3.09    7.38-18.48
PF            28.04     9.02   14.48-38.68
EV             8.05     1.95    5.15-10.53
INV            8.69     2.02    5.32-12.30

              Day 2

              [bar] X   SD      Range        ICC(2,1)

HIPFLEX       13.77     4.90    7.04-22.89   .95
HIPEXT        33.09     3.23   28.50-40.69   .81
HIPABD        11.20     4.34    6.47-20.83   .80
HIPADD        13.41     3.17    8.13-18.65   .71
HIPMR         10.56     3.64    5.04-16.25   .96
HIPLR          7.74     2.18    4.81-10.93   .94
KNEEFLEX      12.75     3.44    7.78-17.85   .94
KNEEEXT       20.46     3.93   14.53-26.09   .70
DF            14.18     4.10   10.99-24.03   .82
PF            26.21     6.46   17.17-40.51   .90
EV             7.74     2.13    4.69-12.02   .90
INV            8.59     2.04    5.95-12.93   .93

(a) HIPFLEX = hip flexion, HIPEXT = hip extension, HIPABD = hip abduction, HIPADD = hip adduction, HIPMR = hip medial rotation, HIPLR = hip lateral rotation, KNEEFLEX = knee flexion, KNEEEXT = knee extension, DF = dorsiflexion, PF = plantar flexion, EV = eversion, INV = inversion.

Data Analysis

A SYSTAT computer statistical package([dagger]) was used to manage and analyze the data. All lower-extremity muscle force measurements were divided by body weight (kilograms/kilograms of body weight) in order to normalize the force data. Descriptive statistics (mean, median, and standard deviation) were calculated for all variables. Histograms and stem-and-leaf plots were constructed for all of the raw data to determine the distribution of the values for each variable. All of the data that were not normally distributed were nonlinearly transformed to ensure that the assumptions of normal distribution and homogeneity of variance were met. Hip extensor, knee flexor and extensor, and ankle evertor force measurements were normally distributed. Log transformation was carried out for the ankle plantar-flexor force measurements and the BBS scores. A square root transformation resulted in a normal distribution for GUG, FRT, and hip flexor, adductor, abductor, medial and lateral rotator, and ankle dorsiflexor and invertor force data, and this transformation was used for analyses involving these measurements.

A forward stepwise multiple regression analysis (alpha to enter and remove of .05 and a minimum tolerance of .01) was applied to the force measurements to determine which measurements were important predictors of the BBS, FRT, and GUG results and of fall status. The transformed data were used for both dependent and independent variables in the regression analysis. Parametric tests (independent t tests with Bonferroni correction) were applied to the transformed force data to determine differences in force measurements between subjects who reported falls and subjects who did not report falls. With the Bonferroni correction, a probability value of [is less than] .004 was required for statistical significance.

Results

Results of the regression analysis are shown in Table 3. Of the 12 muscle groups tested, only the force of the muscles about the ankle was predictive of scores on the balance tests. The ankle dorsiflexor and subtalar evertor muscle force values were the independent variables left in the equation to predict the score on the BBS (P [is less than] .001), accounting for 58% of the score on this measure. The ankle plantar-flexor and subtalar invertor muscle groups were predictors of performance on the GUG (P [is less than] .001) and accounted for 48.4% of the score on this measure. Ankle plantar flexors were the only group that contributed to the score on the FRT (P [is less than] .05, [R.sup.2]=.13). Ankle dorsiflexor force was the only force variable that predicted fall status (P=.003, [R.sup.2]=.17).

Table 3. Results of Forward Stepwise Regression Analysis of Muscle Force Measurements as Predictors of Balance(a)

Measure       Predictor  Variable   [R.sup.2]     P

Fall status            DF              .17      .003
(n=50)
BBS                    DF, EV          .58       .00
(n=39)
FRT                    PF              .13      .025
(n=39)
GUG                    PF, IN          .48       .00
(n=39)

(a) [Alpha] to enter and remove = .05, minimum tolerance = .01.BBS = Berg Balance Scale, FRT = Functional Reach Test, GUG = Timed Get Up & Go Test, DF = dorsiflexion, PF = plantar flexion, EV = eversion, IN = inversion.

Mean muscle force measurements (not normalized) for subjects who reported falls and subjects who did not report falls are presented in Table 4. Mean force measurements were not different in the group reporting falls, with the exception of the ankle dorsiflexor muscles ([bar] X=6.36 kg [SD=1.75] and [bar] X=9.49 kg [SD=4.21] in subjects with and without falls, respectively) and the hip extensor muscles ([bar] X=20.70 kg [SD=6.60] and [bar] X=24.75 kg [SD=8.57] in subjects with and without falls, respectively) (P [is less than] .004).

Table 4. Muscle Force Measurements (in Kilograms)

Subjects With Falls (n=11)

Variable(a)   [bar] X   SD     Range

HIPFLEX       10.26     3.89    5.87-18.29
HIPEXT        20.70     6.60   13.95-34.26
HIPABD         7.88     2.44    4.25-11.98
HIPADD         8.94     2.79    4.41-12.97
HIPMR          7.61     2.92    4.43-12.12
HIPLR          6.57     1.96    4.42-9.76
KNEEFLEX       9.84     3.53    4.92-15.43
KNEEEXT       15.53     8.65    7.90-33.68
DF             6.36     1.75    3.35-9.51
PF            15.70     5.21    9.62-26.04
INV            5.93     1.96    3.55-8.90
EV             5.68     1.93    3.44-9.33

              Subjects Without Falls (n=39)

Variable(a)   [bar] X   SD     Range        P

HIPFLEX       12.96     5.72   5.63-28.70   NS(b)
HIPEXT        24.75     8.57   8.62-39.36   <.004
HIPABD         8.89     3.85   3.61-19.38   NS(b)
HIPADD        10.75     4.46   3.29-24.71   NS(b)
HIPMR          8.18     3.49   3.69-20.94   NS(b)
HIPLR          7.12     2.81   3.51-17.06   NS(b)
KNEEFLEX      11.34     4.92   3.85-27.88   NS
KNEEEXT       16.82     6.80   5.41-38.87   NS
DF             9.49     4.21   3.59-19.33   <.004(B)
PF            19.61     9.06   6.25-37.75   NS(b)
INV            6.90     2.80   2.27-12.20   NS(b)
EV             6.34     2.92   3.08-18.04   NS

(a) HIPFLEX = hip flexion, HIPEXT = hip extension, HIPABD = hip abduction, HIPADD = hip adduction, HIPMR = hip medial rotation, HIPLR = hip lateral rotation, KNEEFLEX = knee flexion, KNEEEXT = knee extension, DF = dorsiflexion, PF = plantar flexion INV = inversion EV = eversion.

(b) Independent t test on transformed data.

Discussion

We examined the extent to which lower-extremity muscle force measurements were predictive of scores on 3 commonly used measures of balance. The force-generating capacity of the ankle muscles was found to be predictive of scores on all 3 measures. Ankle dorsiflexor and evertor muscle force accounted for 58% of the score on the BBS, and ankle dorsiflexor force was a predictor of

fall status. Previous researchers have also identified the ankle dorsiflexor muscles as being important in postural control. Whipple et al[20] compared peak torque and power of the knee flexors and extensors and ankle dorsiflexors and plantar flexors between nursing home residents identified as fallers and a comparison group identified as nonfallers. Both peak torque and power were lower in the group identified as fallers, with the greatest difference found in the ankle muscles. Dorsiflexion force was the most impaired of the variables measured. Similarly, Studenski et al[21] reported weakness of the ankle dorsiflexors and plantar flexors in 10 adults over the age of 60 years who had unexplained falls compared with a control group of 24 nonfallers. The electromyographic response latency of the tibialis anterior muscle to anterior platform perturbation was prolonged in the subjects who fell, whereas the response latency of the gastrocnemius muscle did not differ between the groups.

MacRae et al[22] measured the force of 7 lower-extremity muscle groups (hip flexors, abductors, and adductors; knee flexors and extensors; and ankle dorsiflexors and plantar flexors) in adults aged 60 to 89 years in a study designed to determine risk factors for falls. As in our study, of the muscle groups tested, the ankle dorsiflexors were found to be the best predictor of fall status. Similarly, Lord et al,[7] in a 1-year prospective study of adults aged 50 to 97 years living in supportive accommodation, reported that force of the ankle dorsiflexors was one of several variables that helped differentiate between individuals who experienced multiple falls and individuals who had not fallen.

During gait, the ankle dorsiflexors are involved, together with the hip and knee flexors, in lifting the lower limb during the swing phase to allow sufficient clearance of the toes over the ground to prevent tripping. Human subjects respond to postural disturbances through movement primarily at the ankles and hips (ankle and hip strategy). The ankle strategy requires sufficient ankle range of motion and force in the ankle muscles and is most effective when perturbations to equilibrium are slow and small and the support surface is firm and wide.[1] The ankle muscles provide proprioceptive information and correct for postural sway by controlling the net ankle moment, thus regulating the body's center of gravity and keeping the center of mass located about the foot.[23,24] Wolfson et al[25] found that a crucial aspect of a balance response to a destabilizing force is ankle dorsiflexion because it stops the backward movement produced by the destabilizing force, lifts the forefoot, and helps to create an anteriorly directed counter-moment that helps re-equilibrate the body (ankle strategy). In our study, the force of the ankle dorsiflexors and evertors were predictors of the score on the BBS ([R.sup.2]=.58). This finding supports the ankle strategy theory of postural control for static balance because the majority of components on the BBS involve measurement of performance in maintaining a position.

Ankle plantar-flexor and subtalar invertor force contributed to the score on the GUG, which has a major ambulation component. This finding supports previous research by Bendall et al[26] in which ankle plantar-flexor force was found to be a predictor of gait speed in a group of 125 men and women aged 65 to 90 years. Ankle plantar-flexor force, measured isometrically, accounted for 13% of the gait speed value in both men and women.[26] The ankle plantar flexors contribute to the support moment in the stance phase of gait and the plantar-flexor moment of the push-off phase of the gait cycle, resulting in a high level of plantar-flexor activity with each step.[27] Thus, it is not surprising that force of this muscle group contributed to the prediction of the score on a test that includes a measure of gait speed.

Ankle plantar-flexor force also contributed to prediction of the score on the FRT. Little of the variability in this score, however, was explained by force of this muscle group ([R.sup.2]=.13). This relationship can be explained by the eccentric control required by the ankle plantar flexors for maximal performance on this test. The FRT, however, uses complex upper-body as well as lower-body strategies and, thus, may be more related to upper-body flexibility and force measures rather than to lower-extremity muscle force measures.[28]

The 11 subjects who reported unexplained falls in the year prior to the study did not differ from the subjects who did not report a fall in age, height, weight, or scores on the 3 functional balance measures. Results indicate that there was less balance impairment in this group compared with groups identified as "fallers" in previous studies.[20,21] This finding may be attributed to the manner in which the subjects who fell were identified in this study. The subjects were community-dwelling volunteers with no known impairments. Fall status was determined during the initial interview, during which subjects were asked about falls in the previous year and those subjects with one or more unexplained falls were identified. A difference in balance measurements would be expected had people who fell been recruited directly through physicians' offices or community care centers for elderly people.

When large external perturbations are applied in stance, the control of the body is thought to be primarily the responsibility of the hip muscles (hip strategy) as opposed to the ankle muscles when perturbations are smaller.[1] The hip flexor and extensor muscles act during the single-limb support portion of the stance phase of gait to control the angular acceleration of the head, arms, and trunk.[24] Thus, weakness of these muscles may contribute to instability during gait and dynamic activities.

We used a retrospective design; therefore, causal relationships between muscle force measures and balance could not be determined. Subjects were community-dwelling volunteers, and balance measurements did not differ between subjects who reported falls and subjects who did not report falls. Some of the subjects who had not experienced a fall performed poorly on some or all of the balance measures. That they had not fallen in the previous year may have been a result of carefully restricting their activities, thus avoiding the risk of falling due to their poor balance. There may have been a self-selection bias because all of the subjects were volunteers who had seen advertisements in either the newspaper or on posters in their community and could differ from subjects recruited through fall prevention programs and medical clinics.

Conclusions

Ankle muscle force measures contributed to the prediction of scores on the BBS, the FRT, and the GUG. Dorsiflexor and invertor force contributed to the score on the BBS, which includes several measures of a subject's ability to maintain a position. Plantar-flexor and invertor muscle force measures contributed to the score on the GUG, which has a gait component. Ankle plantar-flexor force also contributed to the score on FRT, possibly related to the eccentric control required by this muscle group to control the forward displacement of the body's center of gravity. The only differences in lower-extremity muscle force between the fallers and nonfallers were found in the ankle dorsiflexor and hip extensor muscles. Only ankle dorsiflexion force contributed to the prediction of fall status. The results of this study support previous studies suggesting that the force-generating capability of the distal musculature is important in the maintenance of balance in older adults.

(*) Penny and Giles, Biometrics Division, Blackwood Gwent, NP2 1YD United Kingdom.

([dagger]) SYSTAT Inc, 1800 Sherman Ave, Evanston, IL 60201.

References

[1] Horak FB, Shupert CL, Mirka A. Components of postural dyscontrol in the elderly: a review. Neurobiol Aging. 1989;10:727-738.

[2] Kaufmann T. Impact of aging-related musculoskeletal and postural changes on falls. Topics in Geriatric Rehabilitation. 1990;5:34-43.

[3] Berg KO, Wood-Dauphinee SL, Williams JI, et al. Measuring balance in the elderly: preliminary development of an instrument. Physiotherapy Canada. 1989;41:304-311.

[4] Duncan PW, Weiner DK, Chandler J, Studenski S. Functional reach: a new clinical measure of balance. J Gerontol. 1990;45:M192-M197.

[5] Podsiadlo D, Richardson S. The timed "Up & Go": a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc. 1991;39: 142-148.

[6] Mathias S, Nayak USL, Isaacs B. Balance in elderly patients: the "get-up and go" test. Arch Phys Med Rehabil. 1986;67:387-389.

[7] Lord SR, Clark RD, Webster IW. Physiological factors associated with falls in an elderly population. J Am Geriatr Soc. 1991;39:1194-1200.

[8] Robbins AS, Rubenstein LZ, Josephson KR, et al. Predictors of falls among elderly people: results of two population-based studies. Arch Intern Med. 1989;149:1628-1633.

[9] King MB, Judge JO, Wolfson L. Functional base of support decreases with age. J Gerontol. 1994;49:M258-M263.

[10] Aniansson A, Zetterberg C, Hedberg M, Henriksson KG. Impaired muscle function with aging: a background factor in the incidence of fractures of the proximal end of the femur. Clin Orthop. 1984;191: 193-201.

[11] Cahalan T, Johnson ME, Liu S, Chao EY. Quantitative measurements of hip strength in different age groups. Clin Orthop. 1989;246: 136-145.

[12] Larsson L, Grimby G, Karlsson J. Muscle strength and speed of movement in relation to age and muscle morphology. J Appl Physiol. 1979;46:451-456.

[13] Iverson BD, Gossman MR, Shaddeau SA, Turner ME Jr. Balance performance, force production, and activity levels in noninstitutionalized men 60 to 90 years of age. Phys Ther. 1990;70:348-355.

[14] 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;50:64-67.

[15] Bohannon RW. Muscle strength testing with hand held dynamometry. In: Amundsen L, ed. Muscle Strength Testing: Instrumented and Non-instrumented Systems. New York, NY: Churchill Livingstone Inc; 1990:69-112.

[16] Bohannon RW. Knee extension torque in stroke patients: comparison of measurements obtained with a hand-held and Cybex dynamometer. Physiotherapy Canada. 1990;42:284-287.

[17] Reed RL, Den Hartog R, Yochum K, et al. A comparison of hand-held isometric strength measurement with isokinetic muscle strength measurement in the elderly. J Am Geriatr Soc. 1993;41:53-56.

[18] Wadsworth CT, Nielsen DH, Corcoran DS, et al. Interrater reliability of hand-held dynamometry: effects of rater gender, body weight, and grip strength. J Orthop Sports Phys Ther. 1992;16:74-81.

[19] Krebs DE. Intraclass correlation coefficients: use and calculation [computer communication]. Phys Ther. 1984;64:1581-1582.

[20] Whipple RH, Wolfson LI, Amerman PM. The relationship of knee and ankle weakness to falls in nursing home residents: an isokinetic study. J Am Geriatr Soc. 1987;35:13-20.

[21] Studenski S, Duncan PW, Chandler J. Postural responses and effector factors in persons with unexplained falls: results and methodological issues. J Am Geriatr Soc. 1991;39:229-234.

[22] MacRae PG, Lacourse M, Moldavon R. Physical performance measures that predict fall status in community-dwelling older adults. J Orthop Sports Phys Ther. 1992;16:123-128.

[23] Kuo AD, Zajac FE. A biomechanical analysis of muscle strength as limiting factor in standing posture. J Biomech. 1993;26(suppl 1): 137-150.

[24] Winter DA, Patla AE, Frank JS. Assessment of balance control in humans. Med Prog Technol. 1990;16:31-51.

[25] Wolfson LI, Whipple R, Amerman P, Kleinberg A. Stressing the postural response: a quantitative method for testing balance. J Am Geriatr Soc. 1986;34:845-850.

[26] Bendall MJ, Bassey EJ, Pearson MB. Factors affecting walking speed of elderly people. Age Ageing. 1989;18:327-332.

[27] Olney SJ, Culham EG. Changes in posture and gait. In: Pickles B, Simpson JM, Cott C, Vandervoort AA, eds. Physiotherapy With Older People. Toronto, Ontario, Canada: WB Saunders Co Ltd; 1995:81-94.

[28] Light KE, Rose DK, Purser JL. The Functional Reach Test for balance: strategies of elderly subjects with and without disequilibrium. Physical & Occupational Therapy in Geriatrics. 1996;14:39-52.

ME Daubney, PT, MSc, was a Master of Science degree candidate, School of Rehabilitation Therapy, Queen's University, Kingston, Ontario, Canada, at the time of this study. This study was completed in partial fulfillment of the requirements of Ms Daubney's Master of Science degree in rehabilitation.

EG Culham, PT, PhD, is Associate Professor, Physical Therapy Program, School of Rehabilitation Therapy, Queen's University, Kingston, Ontario, Canada K7L 3N6 ([e-mail]). Address all correspondence to Dr Culham.

Concept and research design, writing, data analysis, project management, and facilities and equipment were provided by Daubney and Culham; data collection, by Daubney; and institutional liaisons, by Culham. Sandra Olney, PT, PhD, and Elizabeth Tata, PT, MSc, at the School of Rehabilitation Therapy, Queen's University, Kingston, Ontario, Canada, contributed to concept and research design and gave input regarding the methodology and implementation of this study. DrJ Terry Smith, Queen's University Statlab, contributed to data analysis and gave statistical advice.

This study was approved by the Queen's University Health Science Human Research Ethics Board.

This project was partially funded by the Gwen Keough Memorial Scholarship and the Queen's Graduate Student Award awarded to Ms Daubney. The results of this research were presented at the Canadian Physiotherapy Association Congress, June 1997, Winnipeg, Manitoba, Canada.

This article was submitted February 19, 1998, and was accepted August 12, 1999.