Balance improvements in older women: effects of exercise training.

The prevention of falls and the substantial morbidity associated with fall-related injuries will become increasingly important for preserving the health and independence of older women. Fall risk has been shown to increase with reduced lower-extremity joint moments, weakness on manual muscle testing, and difficulty arising from a chair.[1-3] Previous studies[4,5] have shown decrements in muscle mass, force production per cross-sectional area of muscle, and isokinetic joint moments in several lower-extremity muscle groups with usual aging. This weakness may be an important and potentially reversible component of instability during routine daily activities, predisposing the elderly to falls. Although some researchers[6,7] have demonstrated that older adults are capable of significant improvements in lower-extremity force measures with resistance training, the effect of this type of training on balance measures has not been reported.

Postural sway increases with usual aging. Cross-sectional studies have used force platforms, which record the center of pressure center of reaction force), to estimate body sway.[8] Older persons have slightly higher measures of sway in double stance when compared with younger subjects.[9] Increased postural sway increased the risk of falls in studies of community-dwelling elderly.[10-13]

Differences in sway with age are accentuated during testing in single-stance postures. Single-stance sway measures are threefold greater in older subjects than in younger subjects.[10] Single-stance time (time able to stand on one foot) declines with age,[14] and in advanced age, few persons are able to stand on one leg for more than several seconds.[11] Single-stance time was correlated with a summary hip isometric moment score.[15] Two intervention studies[l6,17] demonstrated that balance training programs that included performing single-stance postures increased single-stance time.

Because the majority of falls occur during walking or turning or on stairs,[18,19] we hypothesized that improved single-stance control may lower the risk of falls and fall-associated injuries. Two previous exercise studies emphasizing postural control demonstrated no improvements in single-stance20 or double-stance balance.[21] The exercise interventions in these studies did not include lower-extremity resistance training. Single-stance balance requires appropriate contraction of several lower-extremity muscle groups in addition to adequate vestibular function and proprioception. Our study tested the hypothesis that a vigorous and frequent program of lower-extremity strengthening, walking, and postural control exercises would improve single-stance balance in healthy older women, compared with a program of postural control exercises performed once a week.

Studies have used different measures derived from the force-plate data. Force-platform data are used to estimate the movement of the center of mass of the standing subject from the movement of center of force (COF) on the platform.[8] Two intervention studies[20,21] have used the mean velocity of COF movement during stance. The measure reported in this study is the mean displacement (DISP) from the average position of the COF during each trial, and is similar to the outcome used in a recent cross-sectional study.[22] The DISP measure assumes that the mean coordinates of the COF obtained from 8 seconds of sampling represents the subject's true center of balance" or neutral position. A longer period of sampling might give a more stable estimate of the subject's neutral position, but few subjects are able to stand in single stance for very long, and the effect of fatigue of the muscles controlling the hip and foot might affect the neutral position of the COF.

Method

Study Design and Recruitment

A randomized control trial design was used to test whether an exercise program that was performed frequently and included resistance training, brisk walking, and flexibility and postural control exercises (combined training) would be superior to a program that was performed once a week and included only flexibility and postural control exercises (flexibility training).

Subjects were recruited by mail from a population of more than 1,300 female retirees or spouses of male retirees of a large Hartford (Conn) insurance company. Subjects completed a medical history screening and physical examination and performed a symptom-limited exercise stress test using a modified Balke protocol.[23] Individuals with the following diagnoses were excluded: coronary or carotid artery disease, neurologic disease, postural hypotension (> 15 mm Hg systolic pressure or > 10 mm Hg diastolic pressure at 3 minutes), malignancy (excluding skin), rheumatoid arthritis, hip or knee joint replacement, or obesity (body mass index >30.5 kg/[m.sup.2]). Because the effects of exercise on bone density were tested in a separate part of our study, subjects taking medications that are known to affect bone density (furosemide, prednisone, estrogen) were also excluded. Subjects who were regularly exercising more than 2 hours per week were excluded. individuals with symptomatic lower-extremity arthritis were allowed to participate if passive range of motion was unrestricted. Subjects gave written informed consent prior to participation in the study.

A total of 114 women responded (8% response rate), of whom 30 women did not wish to participate or did not complete the screening process. Of the remaining 84 potential subjects, 46 women were excluded for medical reasons, and 38 women completed the screening process in three groups. Subjects were randomized to either a combined training or flexibility training group using a randomization rate of 0.55 or 0.45, respectively, based on the assumption that the combined training group would have a higher attrition rate. The flexibility training group was considered to be a control group for the combined training group. Balance measurements were obtained on the first two groups recruited n=30). Twenty-five of the 30 subjects in the first two groups volunteered to participate in balance assessments, which required traveling to a gait laboratory in a neighboring town for balance measurement. Repeat measurements were obtained on 21 subjects (12 combined training, 9 flexibility training). Four subjects dropped out or were unable to return for repeat testing for personal reasons.

Balance Measurement

Postural sway measurements were obtained from an AMTI OR-6 force platform* at the Newington Children's Hospital Gait Laboratory (Newington, Conn). The platform uses strain gauges to measure three orthogonal force components. The force moments were used to calculate the position of the COF in the x and y dimensions. The force platform output was amplified and digitized [dagger] at 100 Hz per channel. We used AMTI software to calculate the DISP m centimeters) for each trial. This measure combines displacement in the sagittal plane anteriorposterior dimension of the foot) and the frontal plane (lateral dimension of the foot). The Appendix contains the equation that was used to calculate this measure.

Baseline measurements were obtained before the exercise program began. Twenty-one subjects had follow-up measurements of postural sway after 6 months of exercise. Subjects wore flat-heel, rubber-soled shoes, and four conditions were tested in sequential order for each subject. Three trials of each condition were recorded, and the average displacement for the three trials is reported.

There were two double-stance postures. For double stance, the subjects placed their feet together, with their arms held at their sides. Testing was performed with eyes open (EO) and with eyes closed (EC). There were two single-stance conditions. in single-stance conditions, subjects shifted their weight to the dominant foot (as determined by subject's selfreported dominant hand) before raising the opposite lower extremity. In upright single stance (USS), the nondominant lower extremity was held in 0 degrees of extension at the hip, and the knee was flexed 80 to 90 degrees. in forward-leaning single stance (FSS), subjects flexed their nondominant hip 10 to 15 degrees with the knee fully extended and leaned forward, placing pressure on the anterior foot. An acrylic panel 5 mm in height was placed under the heel to assist in displacing the force to the anterior part of the foot.

The average displacement of the COF during three trials for each condition is reported. To create a summary variable for single stance, the displacements of the six trials in the two single-stance conditions (USS, FSS) were averaged. This summary single-stance measure was the primary outcome variable for the study.

To minimize learning effects while on the platform, subjects practiced the postures three times during the initial screening examination and again on the day of testing. Subjects attempted to maintain each posture for 13 seconds, with COF data recorded for the final 8 seconds in each stance. If a subject lost her balance or placed the suspended foot on the floor, data from the attempt, were not saved, and she was given an a additional attempt. The mean of three successful attempts is reported for each condition. Visual reference was provided by a landscape poster placed 1 m in front of the platform.

Force Measurement

Force measurement, using single maximal repetitions (1-RMs), were obtained on muscle variable-resistance machines for knee extension [double dagger] and a sitting leg press [section] on separate days, with minimum of 2 days between measurements. Subjects performed five lifts with low resistance to warm up, and then resistance was increased, based on the subjects' perceived difficulty (rated 1-4 for easy to very hard), until the subjects could not complete a lift (failure). Subjects rested at least 1 minute between lifts. The protocol goal was to reach the 1-RM in four or five attempts, but occasionally six or seven lifts were required before a failure occurred. Verbal encouragement before and during each lift was given to the subjects.

To reduce the effect of repeated testing on force measures, subjects per- formed two baseline 1-RM tests, separated by at least 1 week, with the second test recorded as the baseline force measure. Testing the reliability of force measures was not pan of the study design.

Knee extension 1-RMs were performed in a sitting position, with the back supported and the hip flexed at 90 degrees at an angular velocity of about 45 degrees. A knee extension attempt was considered successful if the subject was able to extend the bar from 90 to 10 degrees of knee flexion, using a range-limiting device at 10 degrees to increase the reproducibility of the measure. Knee extension force results are reported in units of extension moment newton-meters), not corrected for lower leg mass. The maximum machine moment at the axis of rotation (torque) occurs at between 40 and 50 degrees of knee flexion; at full knee extension, the torque is 85% of maximum.

For the sitting leg press, a back support (60 [degrees] from the horizontal) was adjusted by moving horizontally to compensate for differences in the subjects' thigh length. The starting position was set at a knee flexion angle of 90 degrees, with hip flexion of approximately 105 degrees and ankle dorsiflexion of approximately 5 degrees, performed at an angular velocity at the hip of 30 [degrees]/s. Subjects were instructed to lift the crossbar or move the footplate slowly and to complete a movement. Verbal encouragement was given during all attempts. Maximal force at the footplate, which occurred at the initiation of the movement, was recorded in newtons. Force was determined by measuring the air pressure in the cylinder (in pounds per square inch) x crosssectional area of the cylinder (in square inches) x 4.4482 N/lb. A straingauge dynamometer was used to confirm the accuracy of the calibration of air pressure over the range of forces found in the study.

Aerobic Capacity

An estimate of fitness was determined from the treadmill time on a symptom-limited stress test using a modified Balke protocol.[23] Maximum treadmill speed was 1.4 [m [multiplied by] s.sup.-1] (3.3 mph) and is reported as metabolic equivalents (1 MET=3.5 mL [O.sub.2 [multiplied by] kg.sup.-1 [multiplied by] min.sup.-1).

Intervention - combined

Treatment Group

Resistance training. The musclestrengthening exercises used variable resistance machines providing knee and hip extension [double dagger] and a sitting leg press. [section] Three sets of bilateral exercise were performed in knee extension and sitting leg press, with a resistance set at about 70% of 1-RM. Subjects exercised to volitional fatigue, which was defined as the maximum number of lifts a subject could perform before failure of the movement occurred. Subjects performed between 10 and 14 repetitions per set, with 2 to 3 minutes of rest between sets. Knee extension exercises were performed in a sitting position, with the back supported, at an angular velocity of about 45 [degrees]/s, with a 1-second pause at full extension. The sitting leg press exercise was performed as in the 1-RM testing, with the back support position adjusted for subject leg length. The exercise involved combined hip extension (from about 105 [degrees] to 60 [degrees] of flexion), knee extension (from 90 [degrees] to O [degrees] of flexion), and ankle plantar flexion (from about 5 [degrees] of dorsiflexion to 15 [degrees] of plantar flexion), performed at an angular velocity at the hip of 30 [degrees]/s.

The resistance settings were greatest in the second set, and were 5% to 7% lower in the third set to permit completion of 10 to 14 repetitions. Resistance was adjusted weekly by the exercise leader, based on a review of the subject's performance. Resistance was increased when a subject could consistently complete more than 12 repetitions at a given resistance. A 1-RM test was performed every 8 to 10 weeks, and further adjustments in the resistance setting were made to keep the resistance at 70% of 1-RM for leg press and knee extension.

Standing hip extensions were performed unilaterally, subjects stabilized their position by holding onto a horizontal bar just below shoulder level. Two sets of 20 repetitions each were performed at a resistance that permitted maintenance of "good form," defined as the completion of the movement in a smooth, nonjerking fashion with maintenance of lumbar lordosis. The lower resistance in the hip extension protocol was made to decrease the risk of lower back injury.

There was close supervision of the resistance exercises by the exercise leaders. One leader supervised the knee extension and hip extension machine, and the other supervised the sitting leg press. Frequent feedback and instruction on correct form were given to each subject.

Walking. Subjects walked for 20 minutes on a 145-m indoor track, with a heart rate goal of 70% of each subject's maximum heart rate recorded at the baseline exercise treadmill test. After the seventh or eighth month, most subjects had reached a maximal walking velocity and were completing the walk with a heart rate below 70% of maximal heart rate. The number of laps walked, walk time, and preexercise and postexercise pulse rate were recorded at each session initially by the exercise leaders and after 2 or 3 months by the subject, with periodic checks by the exercise leaders.

Flexibility and balance. Neck and shoulder girdle flexibility, trunk lateral rotation, and thoracic extension exercises were performed in a standing position. Static stretch of the hip adductor and hamstring muscles was performed in a sitting position on a mat, with hips abducted and laterally externally) rotated, with attention to limit lumbar and thoracic flexion while leaning forward. Subjects were taught to rise from a prone or supine position to a standing position. From the supine position, a "log-roll" turn (legs flexed at hip and knee and limiting trunk rotation) brought subjects to a side-lying position, and then to a quadrupedal position and a half-kneel position with one knee on the mat and the opposite foot on the mat. A vertical rise completed the movement, with attention to knee/thigh aligmnent first ray of foot, patella, and anterior superior iliac spine in same plane) and upright torso. The exercise progressed from using a chair for stability to rising without a chair support, placing the hand on the thigh for stability during vertical rise. Subjects performed mock housecleaning tasks in a "lunge" posture, similar to that used in fencing, which emphasized movement at the knee and hip without spine flexion.

Following the stretching exercises, subjects performed low-resistance hip abduction exercises. Abduction exercises were performed in a side-lying position on a mat, with the dependent leg flexed at the knee and hip. Subjects were taught to keep the thigh at full (O [degrees]) extension and neutral rota-

(*) AMTI Inc. 141 California St, Newton, MA 02158. [dagger] Model DT 2801A, Data Translation Inc, 100 Locke Dr, Marlborough, MA 0175-1192. [double dagger] Eagle knee extension, multi-hip resistance machines, Cybex, Div of Lumex Inc, 2100 Smithtown Ave, Ronkonkoma, NY 11779. [sections] Keiser Sports Health Equipment,411 S West Ave, Fresno, CA 93706-9952. tion (O[degree]) through the ROM of the exercise (about - 10[degrees] to 35[degrees]). Two sets of contractions were performed, initially with no weights, to fatigue or a maximum of 20 repetitions. Ankle weights (1-2 lb) were added when subjects could perform two sets of 20 repetitions each, maintaining good form as defined above. Pelvic tilt exercises were taught in the supine position with knees and hips flexed and feet on the mat. Following pelvic that exercises, bilateral hip extensions, or "bridging" exercises, were performed while remaining in a supine position. Legs were aligned so that there was no genu varus or valgus, and feet were flat on mat, a shoulder distance apart. The buttocks were raised slowly about 10 to 15 cm by extension at the hips and pelvic tilt, and then slowly lowered. When two sets of 12 repetitions each were performed with good form, unilateral exercises were performed, with the nonexercised leg held in the hooklying position.

Simplified tai chi exercises, which involve slow and controlled movements of the body, were performed.[24] Correct trunk and lower-body alignment in the standing position was stressed, and subjects were taught to focus on how weight was distributed on the feet. Subjects exercised in a large room with floor-length mirrors, which provided visual feedback to their posture during the movements. Slow, controlled forward and backward steps were performed as a warm-up. Other movements included lateral and anterior-posterior weight shifts; single-stance postures, and turning and pivoting with weight on the heel or forefoot, while keeping the torso upright and the knees slightly flexed. Arm movements were added after subjects were proficient in leg movements.

Intervention - Flexibility

Training Group

The flexibility training group performed no exercise for the first 12 weeks. After week 13, the group exercised once weekly for 30 minutes and performed the same flexibility and balance exercises as the combined training group. They did not use ankle weights for hip abduction exercises, however, and they performed bilateral (not unilateral) bridging exercises. Subjects were not permitted to participate in any other organized exercise program, but were permitted to perform the exercises at home. All exercise sessions for both groups were led by a physical therapist, a master's degree level exercise scientist, or a physician.

Compliance

Compliance with the exercise program was monitored by regular attendance taken by exercise supervisors, who led the exercise sessions, observed the resistance training, and recorded resistance and repetitions performed. Subjects were taught to obtain pulse rates, and they recorded the number of laps walked, preexercise and postexercise pulse rate, and walk time. Accuracy of pulse recordings were checked by an exercise leader monthly. Pulse sensor wristwatches were used by subjects who had difficulty determining their pulse rate. Mean attendance was 80% for both groups, and all subjects attended more than 50% of the exercise sessions.

Statistical Analysis

Comparability of the combined training and flexibility training groups at baseline was assessed by independent t tests, The null hypothesis was that the improvement in DISP in single stance of the combined training group would not differ from that of the flexibility training group. The primary outcome variable was the mean DISP of the COF during the six trials of the two single-stance conditions (USS and FSS). A two-tailed paired t test was used to determine whether each group had improved (reduced) the DISP during balance testing.

Repeated-measures analysis of variance (ANOVA) was tested for the effect of group assignment (combined training or flexibility training) on single-stance balance outcomes. The same analytic strategy was used for the muscle force measures.

To test for within-session learning, the results from trials 1 to 3 for each condition were analyzed by ANOVA. Pearson Product-moment Correlation Coefficients determined the relationships between single-stance and double-stance measures and between muscle force or fitness measures and balance measures. All statistical analysis was performed using SYSTAT 4.0 software.[parallel] The level for significance of all results was set at .05.

Result

Baseline

The baseline characteristics of the two groups were similar for aH variables tested (Tab. 1).

[TABULAR OMITTED]

Balance Measures

The results for single stance are from 12 subjects in the combined training group and 9 subjects in the flexibility training group. Three subjects had a single failure during single-stance testing at baseline, but completed three successful single-stance tests and were included in the analysis.

The summary single-stance measure (the mean DISP of the six singlestance trials) improved 18% in the combined training group at posttest, from 0.83 [+ or -] 0.19 cm to 0.67 [+ or -] 0.08 cm (4,32% 95% CI, P = .023). This measure did not significantly improve (5%, -9,19% 95% CI) in the flexibility training group (0.77 [+ or -] 0.08 cm to 0.73 [+ or -] 0.10 cm, P = .3). The repeated measures ANOVA found a trend for the improvement in the combined training group to be greater than in the flexibility training group, which did not reach statistical significance (F = 2.2; df = 1,19; P = .15).

The combined training group's improvement was similar in both singlestance conditions. In FSS, DISP was reduced from 0.84 [+ or -] 0.21 cm to 0.66 [+ or -] 0.14 cm (P<.05), and USS DISP tended to be lower (from 0.83 [+ or -] 0.26 cm to 0.70 [+ or -] 0.13 cm, P = .11). The flexibility training group measurements tended to be lower in FSS 0.81 [+ or -] 0.10 cm to 0.73 [+ or -] 0.10 cm, P = .23) and were unchanged in USS. The DISP in both double-stance measures was unchanged for both training groups.

To test whether subjects were able to lean forward in the FSS condition, the mean position of the COF was compared with the position in USS. In FSS, the mean position of the COF was 3.2 [+ or -] 2.2 cm anterior to the mean position in USS, suggesting that subjects were successful in advancing the COF to the anterior portion of their foot in FSS.

To test for within-session improvement, the ANOVA revealed that there was no within-session improvement in any balance condition at baseline or follow-up. For example, in USS at baseline, DISP was 0.74 [+ or -] 0.04 cm for attempt 1, 0.76 [+ or -] 0.03 for attempt 2, and 0.80 [+ or -] 0.09 cm for attempt 3 (ANOVA F = 0.2; df = 2,61; P>.8). The stability of balance measures with repeated attempts within a testing session suggests either (1) subjects were sufficiently trained to the task prior to testing that no further improvement occurred with repeated attempts or (2) stance measures are stable with repeated attempts over a short period of time.

To test for a relationship between double-stance measures and singlestance measures, within-subject correlations were analyzed at baseline. There was a significant within-subject correlation between EO and EC (r = .46, P<.01). Correlations between single-stance measures (USS and FSS) were also strong (r = .67, P<.001). However, there was no relationship between double-stance and singlestance measures for any measure tested. That is, good performance (low DISP) in double-stance conditions did not predict good performance in single-stance conditions.

Force

After 5 months, muscle force in the combined training group had increased in both measures (Tab. 2). Knee extension 1-RM increased from 101 to 126 N [multiplied by] m. The flexibility training group did not demonstrate increased 1-RM knee extension force. The repeated-measures ANOVA demonstrated combined training increased knee extension force more than the flexibility training (F = 17.3, P>.001). Sitting leg press measures significantly improved at repeat testing (F = 58.9, P<.001 ANOVA). The combined training group increased from 568 [+ or -] 132 N to 688 [+ or -] 124 N (P = .001), and the control group increased from 624 [+ or -] 136 N to 717 [+ or -] 110 N (P<.05). The repeated measures ANOVA revealed that there was no difference in the improvement between combined training and flexibility training groups (F = 0.36, P>.5).

[TABULAR OMITTED]

Relationship Between Balance

and Force or Fitness Measures

At baseline, sitting leg press force was inversely correlated with single-stance DISP (r = .70, P = <.001 in USS). This relationship remained significant after leg press force (in newtons) was corrected for body mass (in kilograms) (expressed as N/kg body mass). However, at follow-up testing, gains in force production in leg press were not associated with balance improvements, whether the correlation was tested on the sample as a whole or by treatment group (P>.5). Treadmill time on the symptom-limited exercise test, which provides an estimate of aerobic capacity, was not correlated with any balance measure.

Discussion

Balance

This is the first report of an intervention trial that improved force-plate measures of static balance in neurologically intact older persons. The results support the hypothesis that an exercise program emphasizing postural control, moderate resistance training, and walking improves single-stance balance. The improvement in single-stance DISP was greater in the combined training group than in the flexibility training group (17% reduction compared with 5%). The range of improvements following training was wide, however, and the differences in reduction in DISP of COF between treatment groups did not reach statistical significance. These findings, therefore, should be interpreted as preliminary and will require confirmation in a larger study with a true nonexercise control group. A study with 25 subjects per group would be required to determine, with a power (1-[beta]) of 0.8, whether the true difference in improvement between the two exercise programs is the same as that found in this study (ie, 12%).

The Figure suggests that the greatest improvement occurred in the subjects in the combined training group, with the worst balance at baseline. The improvements in single-stance balance are unlikely to be due to a repeated testing effect. To minimize repeated testing effects, subjects practiced each posture at least six times on two different occasions prior to baseline testing, and the postures tested on the force platform were not similar to those used in the balance exercises in the protocol. There was also. no evidence of improvement in measures with repeated trials during baseline or follow-up testing sessions. Finally, the time between measures was many months. Although it is possible that the subjects may have "practiced" the postures specifically to improve their performance, most of the subjects had forgotten the postures by the date of repeat measures and had to be taught the postures again. Lichtenstein[20] also found that single-stance measures were highly reproducible at repeat testing after 3 months.

The improvement in single stance in our study contrasts with the results of an intervention study by Lichtenstein,[20] who measured single stance on a force plate. The subjects in that study were older, the intervention was shorter, and few of the subjects were able to complete 8 seconds of single-stance balance testing.

Previous studies[16,17] have demonstrated increases in single-stance time following interventions that trained single stance. Single-stance time declines with age[14] and is a marker for poor balance. The reported distributions of single-stance times are not normal, but have a long tail of high values.[16,17] All but one of the subjects in our study could maintain single-stance postures for a minimum of 13 seconds, which is greater than the mean single-stance time in other studies[20,21] and is comparable to or greater than the mean single-stance time in a recent study of older men (mean age = 71 years, range = 20-90).[15] Our study did not test the "time to failure," but tested postural sway measured by the distribution (DISP) of the COF from the mean position of the COF during an 8-second trial. Increases in single-stance time can be due to (1) tolerance of instability, (2) increased resistance to fatigue of the gluteus medius muscle, or (3) improved balance. Force platforrn measures of COF movement eliminate the first factor tolerance of instability). A shorter sampling time reduces the contribution of the second factor (muscle fatigue).

The lack of change in double-stance measures in our study is consistent with the results of two prior studies,[20,21] which also demonstrated no change following balance training programs. Double-stance postures are not challenging to healthy older persons and would not be expected to improve. The absence of a correlation between double-stance and single-stance measures suggests that both postures measure different aspects of balance function. Double-stance measures reflect the integrity of the proprioceptors, muscle stretch receptors, vestibular system, vision, and motor control of postural muscles, but they do not require substantial strength or activation of muscles.[25] In contrast, single stance requires active contraction of several muscle groups, particularly the ipsilateral hip adductor and gluteus medius muscles, in addition to the systems involved in double-stance. The stability of the double-stance measures suggests that there is reproducibility of the platform- and software-derived measurements and we believe indicates that the improvement in single-stance balance found in our study is not due to measurement artifact.

Force

This study demonstrated improvements in both measures of force production in the combined training group and improvement in sitting leg press force in the flexibility training group. The improvement in sitting leg press force in the flexibility training group may be due to increased familiarity with the machine as well as possible effects of the flexibility exercise program, which included hip extension ("bridging") exercises.

The moderate improvements in knee extension force in this study are greater than those shown in a study of older women who trained with low-resistance exercises.26 Knee extension training using ankle weights 0.75 kg) did not improve knee extension force in women (mean age = 71 years) after 25 weeks of training. In contrast, Frontera et al[6] and Fiaterone et al[7] reported much larger improvements (107% and 174%, respectively) in knee extension 1-RM after resistance training. These interventions used three sets of eight unilateral extensions at a slow angular velocity (30[degrees]/s) at 80% of 1-RM, with a 10-second rest between repetitions. These studies did not estimate the joint moment or machine moment, but reported results as pounds or kilograms lifted, which makes comparison between studies difficult. The absolute increase in 1-RM lifted was 20 kg and 11.6 kg, respectively. To attempt a crude comparison, if we assume the average lever arm of our machine was 0.30 M, our absolute increase in weight lifted was about 8.3 kg. Our study used bilateral training at a resistance of about 70% of 1-RM, with a faster rate of contraction and no rest between contractions. Unilateral heavy-resistance training appears to be more effective than the present protocol in achieving rapid, large 1-RM knee extension force gains.

Balance and Force Development

The lack of correlation between sitting leg press force and balance at follow-up suggests that control of postural muscles rather than muscle force development may be the critical factor in single-stance balance. In the untrained state, sitting leg press forces appear to be important, but may be only a proxy for another factor, such as the frequency of demanding physical activities that maintain lower-extremity force. Following exercise and postural training, muscle control rather than muscle force development may be the critical variable determining single-stance balance. Thus, lower-extremity force is probably a necessary, but not sufficient, condition for the maintenance of single-stance balance.

Few functionally impaired persons are able to stand on one leg for more than a few seconds.[14] Single-stance measures, although useful in our study of healthy subjects, are unlikely to be useful in assessing balance in very old or functionally impaired populations. Hassan et al,[27] however, have developed a protocol that corrects sway measures for trials in which there is a loss of balance. If this procedure is accurate, short-duration sampling times could be used for persons unable to stand on one foot for more than a few seconds. Single-stance measures have not been used in a prospective studies of falls, and no inferences should be drawn on reductions of the risk of falls from this study.

Conclusions

The women involved in this study were healthy and willing to travel for 30 to 60 minutes to attend the exercise program. The selection of active, high-level functioning women with good baseline balance would tend to limit the improvement possible from an exercise intervention. Similar or greater improvements may be achievable in a less active group of subjects, if they could be induced to volunteer and complete an exercise program.

References

[1] 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. [2] Tinetti ME, Speechley M, Ginter S. Risk factor's for falls among elderly persons living in the community. N Engl J Med. 1988;319:1701-1707. [3] Cummings SR, Nevitt MC, Browner WS, et al. Hip and wrist fractures are due to different types of falls, not different types of osteoporosis. J Bone Miner Res. 1989;4:S10o. Abstract. [4] Klitgaard H, Mantoni M, Schiaffino S, et al. Function, morphology and protein expression of ageing skeletal muscle: a cross-sectional study of elderly men with different training backgrounds. Acta Physiol Scand. 1990;140: 41-54. [5] Frontera WR, Hughes VA, Lutz KJ, Evans WJ. A cross-sectional study of muscle strength and mass in 45- to 78-year-old men and women, J Appl Physiol. 1991;71:644-650. [6] Frontera WR, Meredith CN, O'Reilly KP, et al. Strength conditioning in older men: skeletal muscle hypertrophy and improved function. J Appl Pbysiol. 1988;64:1038-1044. [7] Fiatarone MA, Marks EC, Ryan ND, et al. High intensity strength training in nonagenarians: effects on skeletal muscle. JAMA 1990;263: 3029-3034. [8] Shimba T. An estimation of center of gravity from force platform data. J Biomech. 1984;17: 53-60. [9] Era P, Heikkinen E. Postural sway during standing and unexpected disturbance of balance in random samples of men of different ages. J Gerontol. 1985;40:287-295. [10] Campbell AJ, Borrie MJ, Spears GF. Risk factors for falls in a community based prospective study of people 70 years and older. J Gerontol. 1989;44:M112-M117. [11] Lord SR, Clark RD, Webster IW. Physiological factors associated with falls in an elderly population. J Am Geriatr Soc. 1991;39: 1194-1200. [12] Brockelhurst JC, Robertson D, James-Groom P. Clinical correlates of sway in older age-sensory modalities. Age Aging. 1982;11: 1-10. [13] Fernie GB, Gryfe CI, Holliday PJ, Llewellyn A. The relationship of postural sway in standing to the incidence of falls in geriatric subjects. Aging. 1982;11:11-16. [14] Bohannon RW, Larkin PA, Cook AC, et al. Decrease in timed balance test scores with aging. Phys Ther. 1984;64:1067-1070. [15] Iverson BD, Gossman MR, Shaddeau SA, Turner, Jr ME. Balance performance, force production, and activity levels in noninstitutionalized men 60 to 90 years of age. Phys Ther. 1990;70:348-355. [16] Ledin T, Knronhed AC, Moller C, et al. Effects of balance training in elderly evaluated by clinical tests and dynamic posturography. Journal of Vestibular Research. 1991;1:123-128. [17] Johansson G, Jarnlo G-B. Balance training in 70-year-old women. Physiotherapy Theory and Practice. 1991;7:121-125. [18] Nevitt MC, Cummings SR, Kidd S, et al. Risk factors for recurrent nonsyncopal falls. JAMA. 1989;261:2263-2268. [19] Nevitt MC, Cummings SR, Hudes ES. Risk factors for injurious falls: a prospective study. J Gerontol. 1991;46:M164-M170. [20] Lichtenstein MJ. Exercise and balance in aged women: a pilot controlled clinical trial. Arch Phys Med Rehabil. 1989;70:138-143. [21] Crilly RG, Willems DA, Trenholm KJ, et al. Effect of exercise on postural sway in the elderly. Gerontology. 1989;35:137-143. [22] Maki BE, Holliday PJ, Topper AK. Fear of falling and postural performance in the elderly. J Gerontol. 1991;46:M123-M131. [23] Blair SN, ed. Guidelines for Exercise Testing and Prescription. 3rd ed. Philadelphia, Pa: Lea & Febiger; 1986. [24] Kauz H. Tai Chi Handbook. New York, NY: Doubleday Dolphin Books; 1974:15-104. [25] Patla A, Frank J, Winter DA. Assessment of balance control in the elderly: major issues. Physiotherapy Canada. 1990;42:89-97. [26] Agre JC, Pierce LE, Raab DM, et al. Light resistance and stretching exercise in elderly women: effect upon strength. Arch Phys Med Rehabil. 1988;69:273-276. [27] Hassan SS, Lichtenstein MJ, Schiavi RG. Effect of loss of balance on biomechanics platform measures of sway: influence of stance and a method for adjustment. J Biomech. 1990; 23:783-789.