Secondary Logo

Journal Logo

CME Article


Kerschan, Katharina MD2; Alacamlioglu, Yesim MD; Kollmitzer, Josef PhD; Wöber, Christian MD; Kaider, Alexandra MSC; Hartard, Manfred MD; Ghanem, Abdel-Halim MD; Preisinger, Elisabeth MD

Author Information
American Journal of Physical Medicine & Rehabilitation: July 1998 - Volume 77 - Issue 4 - p 326-332
  • Free


Objectives: Upon completion of this article, the reader should be able to(1) identify the risk factors of falling and fractures; (2) recognize which parameters may be used for follow-up in patients with osteoporosis and patients with the risk of osteoporosis; (3) contrast the advantages and disadvantages of different training regimens with the aim of preventing fractures. Level: Comprehensive

The Association of Academic Physiatrists is accredited by the Accreditation Council for Continuing Medical Education to sponsor continuing medical education for physicians.

The Association of Academic Physiatrists designates this continuing medical education activity for a maximum of one credit hour in Category 1 of the Physician's Recognition Award of the American Medical Association. Each physician should claim only those hours of credit that he/she actually spent in the educational activity.

*Disclosure statements have been obtained regarding the authors' relationships with financial supporters of this activity. There is no apparent conflict of interest related to the context of participation of the authors of this article.

According to World Health Organization criteria, an estimated 30% of postmenopausal white women in the United States are thought to have osteoporosis, with similar prevalence rates reported in some other populations.1 Roughly one-third of elderly people fall each year,2 and fractures often lead to permanent or long-term immobility and dependency. Therefore, osteoporosis is a major community health problem and the incidence of fractures has been reported to be on the increase in this population. Osteoporotic fractures may occur even during minimal stress. The most serious fractures are those of the hip, and the axial skeleton, which substantially contribute to morbidity, pain, disability, and health care costs.

An elderly person's risk of fracture depends on the reduction in bone mass and the risk of falling because of reduced muscle strength and coordination.3 Aitken and coworders4 postulated that the primary cause of hip fracture in the elderly was an increased likelihood of falling rather than a reduced bone mass. Chow and colleagues5 stated that, besides other factors, poor balance, incoordination, and poor muscle strength or tone may cause a person to fall. Control of balance and mobility requires reception and input of information from multiple sources (vestibular, visual, proprioceptive, and kinesthetic sensation). At least some decline in physiologic balance and mobility control occurs in most humans; static posturography is an objective method of measuring quiet standing balance. In older people sway increases, and the frequency of falls rises as sway increases.6 Measurements of balance were shown to be strong predictors of falling in the elderly.7 Poor performance on testing neuromuscular function, such as gait speed, was associated with an increased risk of hip fracture.8

Exercise is essential for maintaining functional independence in older adults, because it sustains and improves muscle strength, coordination, and balance and may thus reduce the likelihood of fracture-risk falls.9 In women with osteoporosis, regular training significantly relieves musculoskeletal pain and increases general well-being.3 Strengthening and stretching of muscles as well as exercises correcting posture seem to be necessary to avoid faulty posture and pain and to promote motor function. Compared with drug therapy, an adapted individual training program has several advantages: it has no adverse side effects, is less costly which is especially important in view of exploding health care costs, and is more enjoyable.10 Such a training program should be easy to learn, easy to perform, and easy to review at regular intervals.

The risk of fracture in young women with low gonadal hormone levels emphasizes the caution that must be observed in training elderly postmenopausal women.11 In view of the potential hazardous consequences of letting untrained osteoporotic individuals perform vigorous weight-bearing exercises on their own, the exercise program tested in this study focused on improving flexibility, muscle strength, and coordination. In a 3-yr follow-up period,12 this exercise program had been shown to delay bone loss in postmenopausal women exercising on a regular basis. The aim of the present cross-sectional analysis was to evaluate the long-term effects of this home-based exercise program on fracture rates and factors responsible for the risk of falling, such as disability, gait velocity, muscle strength, and body sway in elderly women.



All subjects were recruited from the outpatient clinic of the Department of Physical Medicine and Rehabilitation, University of Vienna. Women who had been included in intervention or control groups from 1985 to 1991 were called in. All subjects were informed about the purpose of this cross-sectional investigation and consented to participate.

Five to 10 yr ago, 246 women fulfilling the following inclusion criteria had visited the osteoporosis laboratory for the first time: The women were Caucasian, between 45 and 75 yr of age without malabsorption or chronic disease, nonsmokers, and had had a sedentary lifestyle. At that time they had been postmenopausal for at least 1 yr, with or without estrogen replacement therapy. The postmenopausal status had been confirmed by serum levels of the follicle-stimulating hormone, and secondary osteoporosis had been excluded by analyzing laboratory parameters. Independent of the group (training group (T) or control group (C)) subjects had been stratified into those who received medication and those who did not. All women were advised to ingest a regular mixed diet of at least 4200 kJ daily and to have a daily intake of at least 1200 mg of calcium with their food.


The exercise program included a warm-up period (brisk walking, modest jogging), stretching exercises (hamstring, gastrocnemius, iliopsoas, pectoralis, and external rotation muscles of the hips), and exercises directed toward improving muscle strength, faulty posture, and coordination. While subjects sitting on big gymnastic balls, pelvic tilting and hopping were performed. Muscles were strengthened, and coordination was promoted with exercises in the closed chain against resistance of elastic bands (Thera Band, The Hygenic Corp., Akron, OH) or imaginary resistance.

During the initial phase, the exercises were supervised 20 times by a physical therapist. The sessions took place twice a week and lasted 45 min each. Women were instructed to perform the exercises at least three times a week for 20 min on a regular basis. At half-yearly intervals, the subjects could come for exercise control and correction in five supervised training sessions.

Assessment and Follow-up

During the follow-up period, the women were called in regularly for blood and urine analysis, densitometry, and a physical examination. Five to 10 yr after their first visit to the outpatient clinic, the following data were obtained: history, physical examination, walking test, measurement of grip strength, knee extension and knee flexion strength, static posturography, biochemistry, and x-ray of the spine. All tested variables are listed in Table 1.

Tested variables

In keeping with the usual practice in the outpatient clinic, patients were interviewed about their usual frequency of training per week during the last year and duration of training per unit. Leisure time activities (e.g., walking and bicycling) were recorded in addition.

Pain Disability Index (PDI)13

The women filled in a visual analog scale evaluating pain related disability during various activities (housework, leisure activities, social life, work, sexual life, and maintenance activities).

Walking Test

Walking velocities were evaluated using the Alge Comet Sports timing. The women had to walk a distance of 6 m at (1) normal, self-chosen speed and (2) maximum speed. Of three trials, the mean self-chosen gait velocity and mean maximal speed were calculated and corrected for height and gravity.14 The coefficient of variation for the nonnormalized gait velocity of healthy elderly women varies between 0.09 and 0.14.15 Because the entire test was visually controlled accidental tripping was excluded.

Muscle Strength (Isometric, Isokinetic)

Isometric grip strength was measured with the Jamar dynamometer, a device with an error no more than 1.5% in force readings across a range of 0 to 200 pounds.16 Both the dominant and nondominant hands were tested three times consecutively, with a 30-s recovery period between the measurements. For testing, the subject was seated with the shoulder in neutral position and adduction and the elbow unsupported in 90° flexion. The maximum of three consecutive tests in each side was taken for statistical analysis. Quadriceps and hamstring muscles (knee flexion and extension) were evaluated on a Cybex 6000 dynamometer: isometric force in 45° knee flexion and extension (maximum of 3 repetitions with 30-s breaks), isokinetic force with alternating angular velocities of 60°/s (3 repetitions) and 240°/s (20 repetitions), with a break of 30 s in between. The Cybex 6000 isokinetic dynamometer has a high reliability in measuring isokinetic concentric and eccentric variables.17

Posturography was used as a body sway test to determine balance and was carried out in a quiet, well lighted, comfortably heated room. The subjects were barefoot and were told to stand on the posturographic platform upright in as stable a position as possible, arms at the sides and feet 4 cm apart. The result of each measurement was a sway direction diagram characterized by four sway parameters: sway path (SP), sway area(SA), anteroposterior sway (APS), and lateral sway (LATS). Eight measurements were performed, two in each of four conditions: eyes open, eyes closed during stance on a firm surface, and eyes open and closed during stance on foam. Full descriptions of the apparatus, procedures, and reliability were reported previously.18 The better of the two measurements of each condition was used for statistical analysis. The effect of the exercise program on body sway was analyzed; it was hypothesized that compliant exercisers would show less body sway than noncompliant exercisers and the control group.

Additionally, functional assessment of the range of motion of the limb joints, strength of abdominal, back, and gluteal muscles, and flexibility was performed. Blood and urine samples (complete blood count, sedimentation rate, total thyroxine, parathyroid hormone, electrolytes, total protein, alkaline phosphatase, creatinine, and routine urinalysis) were taken to rule out secondary osteoporosis. The women also underwent radiography of the entire vertebral column. Follow-up x-rays mainly focused on the possible presence of vertebral fractures.


An intention-to-treat analysis was performed to compare changes in variables, both in controls and in the intervention group. The unpaired t test was used for normally distributed continuous variables, and Wilcoxon's rank sum test for abnormally distributed continuous variables. Nominal scaled variables were analyzed by cross-tabulation and the chi-squared test. The null hypothesis was rejected if P was equal to or lower than 0.05. In addition, we defined a group (T1) with high compliance in performing the exercise program regularly three times a week for at least 20 min and a further group (T2) whose members stopped exercising, performed the program irregularly or for less than 1 h/wk. To compare continuous variables between these two intervention groups and controls, analyses of variance were performed. The Kruskal-Wallis test was used for abnormally distributed variables. Twenty age-adjusted matched pairs of the control group and compliant exercisers were compared by the unpaired t test. Logistic regression analyses were performed to evaluate the influence of the group membership, the corrected self-chosen gait velocity, age, and menopausal age on the PDI (0 v 1-10 v >10). Both univariate and multiple models were performed. The multiple logistic regression model evaluated the influence of each factor after adjustment for all the other factors included in the model. Statistical analysis was done using the SAS software package.


After a follow-up period of 7.7 ± 1.1 (mean ± SD) yr, 127 of 246 contacted women answered the first call and visited the outpatient clinic. Among those who came for follow-up two women developed Parkinson's disease and one breast cancer. All three were excluded from statistical analysis. Of the remaining 124 women 67 had been allocated to an exercise group and 57 to a control group. Three subjects refused some of the examinations, and 16 strength tests were not carried out because of hypertension or cardiac risk. A total of nine women failed to perform parts of the examinations because of other problems: six women had other limiting illnesses(operation in the eye, orthopedic problems, transient ischemic attack, etc.), two lacked the time to undergo some of the tests because of their jobs, and one simply failed to show up (Table 2).

Reasons for nontesting

Intention-to-Treat Analysis

Anthropometric data and other baseline characteristics are summarized in Table 3. Intergroup differences in age were statistically significant but not clinically relevant. Menopausal age, height, body weight, and body mass index showed no differences. Similarly, at follow-up there were no significant differences between the two groups in terms of height, body weight, and body mass index (Table 4). In the controls, fewer individuals tended to take bone-specific medication. In both groups walking aids were rarely used. Blood and urine analysis showed no intergroup differences.

Baseline characteristics
Characteristics at time of investigation

Three-Group Comparison

Table 5 shows the main results of our crosssectional investigation. Twenty-four women (36%) of the exercise group reported to perform their home program on a regular basis. Therefore, the intervention group was further divided into the regularly exercising T1 group (n = 24) and the less compliant T2 group (n = 43). Intergroup differences were found with regard to age and the number of years since menopause. In terms of PDI, back pain, self-chosen gait velocity hand dynamometry as well as lower limb extensor and flexor isometric and isokinetic (60 and 240°) strength, muscle shortening, and range of motion of upper and lower limb joints, no statistically significant differences were observed between the groups.

Cross-sectional data: results of main variables

The longitudinal data showed that height loss, body weight, and body mass index changes were not significant. In each group approximately onethird of the women had fractures of the peripheral skeleton and/or fractures of the vertebral bodies during the follow-up period.

Matched-Pair Analysis

Statistical analysis of 20 age-adjusted matched pairs is shown inTable 6. Compliant exercisers had significantly stronger abdominal muscles than the controls. For back strength, only a slight tendency but no significant difference was observed. Significantly more exercisers took more bone-specific medication than the control subjects.

Matched pairs

Multiple Logistic Regression

Regression analysis showed a significant influence (P = 0.005) of the self-chosen gait velocity on the PDI. The higher the corrected self-chosen gait velocity of a woman, the better the chance of achieving a PDI of zero (see Fig. 1). This significant influence was confirmed by multiple regression (P = 0.01), showing that women with higher corrected self-chosen gait velocity achieve a better PDI than those with lower gait velocity, given the same values in the factors group membership, age, and menopausal age. Although women of the intervention groups tended to have a lower PDI, no statistically significant influence of group membership on the PDI was detected by the logistic regression analysis.

Figure 1
Figure 1:
Association between corrected self-chosen gait velocity and pain disability index.


Persons who performed an unvarying homebased exercise program for 5 to 10 yr did not show significant differences in pain induced disability, muscle strength, postural stability, and fracture rate compared to a group of nonexercisers of the same age. Although no intergroup differences could be found, a negative association between pain induced disability and corrected self-chosen gait velocity existed. Therefore, gait velocity can be regarded as an objective parameter of functional impairment. Compared with irregular exercisers and nonexercisers, regular exercisers had a slightly higher self-chosen gait velocity.

In view of the relatively long observation period of almost 8 yr, the rate of 36% (about 20% when the dropout rate is included) still exercising on a regular basis at home was relatively high for elderly women who had never performed any previous physical training. Compliance was assessed solely on the basis of self-reported data. In the 3-yr follow-up period, the exercise compliance was 48%12; this rate of compliance is comparable to that for drug therapy. Berman et al.19 found that 59% of 2106 women entirely discontinued their estrogen replacement therapy and that 89.6% had slipped below the 80% compliance level after 2 yr.

Vertebral deformities cause substantial pain, disability, or loss of height only if vertebral ratios fall four standard deviations below the normal mean.20 The patients in this study were healthy and hardly functionally impaired, able to live at home and to perform housework. The previous longitudinal study21 showed a significant improvement in locomotor system-related pain following the exercise program. This cross-sectional analysis revealed that back pain was more common in nonexercisers. Logistic regression analysis did not show significant differences, but there was a trend toward a lower PDI among patients exercising regularly. Given a maximum score of 70, women of all groups had a low PDI score. Therefore, the pain disability index does not seem appropriate for functionally largely nondisabled persons; it seems to be only suitable for the more disabled.

It is well known that bone density declines with increasing age. After menopause, the set points of minimum effective strain increase.22 A steep increase in the relative risk of fracture with reduced bone mass is observed below 75 yr of age. Above this age the increase in risk is relatively small, and neuromuscular responses that protect the skeleton against trauma may be more important than bone mass.23 The fact that most of the hip fractures result from a fall, and the fact that some elderly women who experience falls limit their daily activities because they fear falling again, prompt a search for factors minimizing the risk of falling. What are the risk factors of falling? Balance and gait impairment, incoordination, poor muscle strength or tone, a small calf circumference, and a relatively poor health rating besides other nonfunctional risk factors were cited.5,7,8,24

Self-chosen gait velocity of regular exercisers was slightly higher than that of irregular exercisers or nonexercisers. The reason for the non-significance of the difference may be that our subjects were approximately 12 yr younger than those examined by Dargent-Molina and coworkers,24 who found a significant association between hip fracture and neuromuscular impairment reflected by gait speed in women with a mean age of 80.5 yr. Dalsky25 recommended that only persons at imminent risk for fracture were in need of improving balance and gait patterns. Obviously, they expected this group of patients to benefit most from training balance and gait patterns. Women participating in the present study were mostly nondisabled. Range of motion testing confirmed their functional independence.

No intergroup differences in hand grip or lower limb strength testing could be found. Clinical testing of the abdominal muscles showed a significantly higher grade of strength in the regularly exercising group; back muscles did not show such a difference. However, on critical review, these results cannot be regarded as hard facts because they were merely obtained from clinical testing. Static posturography also did not show any group-specific differences, indicating that the exercise program did not have any effect on static body sway. Following the hypothesis that muscle strength and body sway were independent predictors of fracture incidence,6,7 subjects of both groups had similar risk factors of falling. This is supported by the fact that the fracture rate during the observation period did not differ between the two groups. For all participants, independent of group membership, fracture rates and height reductions during the observation period were low. Altogether, PDI and fracture rates were too low in both groups to warrant any conclusions. An elderly person's inability to regain balance rapidly and avoid an actual fall, and his/her inability to correct balance once stumbled,6 may be the reasons why dynamic posturography yields more information than static posturography.26

Compliant exercisers took more bone-specific medication, especially fluorides and calcium supplements than noncompliant or control persons. This suggests that compliant exercisers generally took more care of their health. However, it cannot be stated with certainty whether those who exercise regularly are in better health as a result of exercising regularly, or whether their better health allows them to exercise regularly. Taking functional joint-testing into account, all women were fairly well mobile. The whole collective seems to be uniform, because functional parameters, anthropometric data, and usual leisure activities were similar for all groups.

The study may be biased by the fact that exercise compliance was solely assessed on the basis of patients' reports. As the women had been questioned regularly about their exercising habits at each follow-up (half yearly or yearly intervals), they were already familiar with this question. The data from this interview were compared with previous data. No remarkable differences between answers at different check-ups were found. The high drop-out rate is, no doubt, another weakness of this study. Approximately one-half of the 246 persons invited by letter did not show up at the outpatient clinic for the follow-up examination after 7.7 yr. The reasons were change of the address, death, various diseases causing disability, or refusal to undergo reexamination.


We suggest that in healthy, functionally largely nondisabled postmenopausal women aged between 60 and 70 yr, an unvarying training program does not yield enough force to show intergroup differences in pain, disability, body sway, and muscle strength. Persons with lower pain induced disability were more agile. Corrected self-chosen gait velocity seems to be an objective parameter of the patients' disability status. Further studies are needed to define more appropriate exercise programs for a comprehensive improvement of functional outcome in a population at high risk of osteoporosis.


We thank T. Saradeth, M.D., for help in data processing and Mr. Grafendorfer for data retrieval. The National Bank fund is gratefully acknowledged for financial support. Statistical Reviewer: Alexandra Kaider, MSC, Department of Medical Computer Sciences, 1090 Vienna, Währinger Gürtel 18-20, Austria.


1. Melton LJ III: Perspectives: how many women have osteoporosis now? J Bone Miner Res 1995;10:175-177.
2. Cummings SR, Nevitt MC: Falls (editorial). N Engl J Med 1994;331:872-873.
3. Hartard M, Haber P, Hieva D, et al.: Präventiver und therapeutischer Wert von Krafttraining bei Osteoporose. Sportorthopäd Sporttraumatol 1995;11:160-166.
4. Aitken JM: Relevance of osteoporosis in women with fracture of the femoral neck. Br Med J 1984;288:597-601.
5. Chow R, Harrison J, Dornan J: Prevention and rehabilitation of osteoporosis program: exercise and osteoporosis. Int J Rehabil Res 1989;12:49-56.
6. Overstall PW, Exton-Smith AN, Imms FJ, et al.: Falls in the elderly related to postural imbalance. Br Med J 1977;1:261-264.
7. Nguyen T, Sambrook P, Kelly P, et al.: Prediction of osteoporotic fractures by postural instability and bone density. Br Med J 1993;307:1111-1115.
8. Cummings SR, Nevitt MC, Browner WS, et al.: Risk factors for hip fracture in white women. N Engl J Med 1995;332:767-773.
9. Suominen H: Physical activity and bone. Ann Chir Gynaecol 1988;77:184-188.
10. Simkin A, Ayalon J, Leichter I: Increased trabecular bone density due to bone-loading exercises in postmenopausal osteoporotic women. Calcif Tissue 1987;40:59-63.
11.Gutin B, Kasper MJ: Can vigorous exercise play a role in osteoporosis prevention? A review. Osteoporosis Int 1992;2:55-69.
12.Preisinger E, Alacamlioglu Y, Pils K, Saradeth T, Schneider B: Therapeutic exercise in the prevention of bone loss: a controlled trial with women after menopause. Am J Phys Med Rehabil 1995;74:120-123.
13. Tait RC, Pollard CA, Margolis RB, Duckro PN, Krause SJ: The pain disability index: psychometric validity data. Arch Phys Med Rehabil 1987;68:438-441.
14.Wagenaar RC, Beek WJ: Hemiplegic gait: a kinematic analysis using walking speed as a basis. J Biomech 1992;25:1007-1015.
15.Öberg T, Karsznia A, Öberg K: Basic gait parameters: reference data for normal subjects, 10-79 years of age. J Rehabil Res Dev 1993;30:210-223.
16. Niebuhr BR, Marion R: Detecting sincerity of effort when measuring grip strength. Am J Phys Med 1987;66:16-24.
17.Li RC, Wu Y, Maffulli N, Chan KM, Chan LJ: Eccentric and concentric isokinetic knee flexion and extension: a reliability study using the Cybex 6000 dynamometer.Br J Sports Med 1996;30:156-160.
18. Wöber C, Oder W, Kollegger H, et al.: Posturographic measurement of body sway in survivors of severe closed head injury. Arch Phys Med Rehabil 1993;74:1151-1156.
19. Berman RS, Epstein RS, Lydick EG: Compliance of women in taking estrogen replacement therapy. J Women's Health 1996;5:213-220.
20. Ettinger B, Black D, Nevitt M, et al.: Contribution of vertebral deformities to chronic back pain and disability. J Bone Min Res 1992;7:449-456.
21. Preisinger E, Alacamlioglu Y, Pils K, et al.: Exercise therapy for osteoporosis: results of a randomised controlled trial. Br J Sports Med 1996;30:209-212.
22. Frost HM: A new direction for osteoporosis research: a review and proposal. Bone 1991;12:429-437.
23. Cooper C, Barker DJ, Morris J, Briggs RS: Osteoporosis, falls, and age in fracture of the proximal femur. Br Med J 1987;295:13-15.
24. Dargent-Molina P, Favier F, Grandjean H, et al.: Fall-related factors and risk of hip fracture: the EPIDOS prospective study. Lancet 1996;348:145-149.
25. Dalsky GP: The role of exercise in the prevention of osteoporosis. Compr Ther 1989;15:30-37.
26. Baloh RW, Spain S, Socotch TM, Jacobson KM, Bell T: Posturography and balance in older people. J Am Geriatr Soc 1995;43:638-644.

How to Obtain CME Category 1 Credits

To obtain CME Category 1 credit, this educational activity must be completed and postmarked by December 31, 1999. Participants may read the articles and take the exams issue by issue or wait to study several issues together. After reading the three CME Articles in this issue, participants may complete the Self-Assessment Exam by answering the questions on the CME Answering Sheet and the Evaluation pages, which appear later in this section. Send the completed forms to: CME Department, Association of Academic Physiatrists, 5987 E. 71st Street, Suite 112, Indianapolis, IN 46220. Documentation can be received at the AAP National Office at any time throughout the year, and accurate records will be maintained for each participant. CME certificates are issued only once a year in January for the total number of credits earned during the prior year.


Postmenopausal Women; Home-Based Exercise Program; Prevention of Fractures; Gait Velocity; Posturography; Muscle Strength; Disability; Pain

© Williams & Wilkins 1998. All Rights Reserved.