Relationship of extremity muscle torque and bone mineral density in postmenopausal women.

Key Words: Bone and bones; Menopause; Muscle peformance, measurement; Osteoporosis; Torsion.

Osteoporosis is a clinical condition in which progressive loss of bone predisposes the skeleton to fractures. An estimated 350,000 of the 1 million fractures that occur in the United States annually in women over 45 years of age could be prevented by the elimination of osteoporosis. [1] The expense of osteoporotic fracture care is staggering. Fractures of the femoral neck alone may exceed the cost of $1 billion annually. [1]

Normal bone density is maintained by a balance between osteoclastic and osteoblastic activity. Osteoporosis is marked by a negative imbalance of these processes. An increase in osteoclastic activity can cause increased bone resorption. if this normal process is not compensated for by an increase in osteoblastic (ie, boneforming) activity, there is a corresponding decrease in bone mineral density (BMD). In other cases of osteoporosis, a decrease in osteoblastic activity will lead to a subsequent loss of BMD.

There is, however, a normal decline of BMD with age. This decline is more pronounced in women than in men. Average bone loss is 1 % per year after the age of 35 years. At the onset of menopause, the rate is accelerated to as high as 7% per year, which is believed to be due to a decrease in estrogen levels that accompanies menopause. [2] Furthermore, there is a reduction in the sensitivity of mineralized bone to parathyroid hormone. This hormone is responsible for maintaining calcium levels in the plasma within a narrow range. Excessive release of parathyroid hormone results in demineralization attributable to the breakdown of hydroxyapatite crystals. [3] Huppert has proposed that osteoclastic activity is facilitated, which results in decreased BMD. [4] The rate of loss of both cortical and trabecular bone is decreased by estrogen and accelerated in its absence. [4] Estrogen supplements are often used to treat this problem in postmenopausal women.

Bone mineral density has also been shown to decrease in response to factors such as bed rest, immobilization, and decreased weight bearing. Issekutz et al found increased calcium excretion after 30 days of total bed rest, even if supine bed exercise or eight hours of sitting per day were allowed. However, if a harness was applied such that forces equal in magnitude to those observed in weight bearing could be applied to the longitudinal axis of the body, then calcium levels remained normal. [5] Krolner and Toft found an average bone mineral decrease of 0.9% per week after 27 days of bed rest. [6] Reambulation resulted in bone mineral gain, and the restoration of lumbar spine BMD was nearly complete after four months. They reasoned that bone mineral content changes simply reflect an adaptive response to decreased (eg, immobilization) and increased (eg, gravity and exercise) degrees of vertebral strain. [6] Smith and Raab noted that immobilization produced bone loss that was specific to the immobilized area. [4] They also noted that bone loss is more rapid in areas of the body that are normally weight bearing. Postinjury decrease in BMD of the tibia has been reported. [7] Cessation of weight bearing and a limited activity level were cited as the probable mechanisms. [7] Bone demineralization was also demonstrated in both the foot and hand of astronauts during orbital flights. [8] These studies point to the importance of weight bearing and physical activity in the mechanical stimulation of bone growth and remodeling.

Bone mineral density has been shown to increase with weight bearing and exercise. In an exercise program consisting of two, 1-hour sessions per week over an eight-month period, Krolner et al showed that lumbar spine BMD of an exercise group increased by 3.5%, whereas that of a control group decreased by 2.7%. [9] Smith and Redden followed 29 postmenopausal women in a nursing home for three years. Bone mineral density, determined by photon absorptiometry, increased 4.2% in the exercise group and decreased 2.5% in the control group. [10] Aloia et al found significant differences in total body calcium, measured with total body neutron activation analysis, in an exercise group as compared with a control group. These postmenopausal women performed the President's Council on Physical Fitness program for one hour three times per week over a one-year period. [11] Simkin et al, using loads applied only by muscle activity and body weight in exercises, concluded that trabecular bone tissue in the distal radius of postmenopausal women responds favorably to dynamic and diverse bone-stressing exercises. [12] A number of reports have discussed the high BMD found in physically trained individuals. Nilsson and Westlin measured a higher BMD in the distal femur of athletes compared with nonathletes. [13] In an animal model, Burr reported that normal loading through the lower extremity for 20 minutes per day over a three-week period provided a stimulus for increasing BMD in the respective limb and proximally to the lower spine. [14] In response to exercise and weightbearing studies, Chamay and Tschantz concluded that bone hypertrophy will occur in response to increased weight bearing and muscular stress. [15]

The literature concerning the relationships between muscle strength and BMD documents differing results. Sinaki et al found a decrease in both BMD of the radius and strength of the elbow flexors with increasing age to be independent of one another.[16] Sinaki and Offord, however, found a significant positive correlation between L2-L4 BMD and back extensor strength. [17] Gillespie found that severing the anterior root of the spinal nerves of kittens resulted in muscular atrophy as well as a 20% decrease in the weight of bone in the affected area. [18] Doyle et al found a significant correlation between the ash weight of the third lumbar vertebral body and the weight of the left psoas muscles when body weight, height, and age were taken into account. [19] Nilsson and Westlin hypothesized that

  The weight of a muscle reflects the
  forces it exerts on bones to which it is
  attached and an increase or decrease
  in muscle weight results in a corre
-  sponding increase or decrease of
  bone. [13](P179)

Pogrund et al demonstrated an association between osteoporosis and the psoas muscles in both men and women. [20] A decrease in muscle width appeared to begin about 10 years before bone loss became apparent. [20] Nearly a century ago, Wolff recognized the relationship of bone formation to the functional forces acting on it. [21] Readily available clinical indexes for quantifying these mechanical stresses and BMD are currently available with strength tests and dualphoton absorptiometry. However, there is a lack of data regarding the association between extremity strength and bone density at the hip and spine, two areas of the axial skeleton where osteoporotic fractures are common. If a strong association exists, it may assist in identification and prevention of osteoporotic changes before they result in fracture.

The purposes of this study were 1) to examine the correlations between the torque produced by four extremity muscle groups and the BMD of the lumbar spine and hip in postmenopausal women and 2) to develop a prediction model for BMD using muscle strength and body weight.

Strong correlations were expected between the torques and the BMD of the muscle's attachment site. Because these muscles are responsible for supporting and stabilizing the body during weight bearing, the torque they can generate was expected to correlate with the BMD of those skeletal regions stressed during weight bearing.

Method

Subjects

Following approval of the study by the Human Subjects Committee of The University of Iowa, 56 postmenopausal women voluntarily participated in this study. In all subjects, menopause was a result of oophorectomy or natural onset. The subjects entered the study in response to an advertisement or by referral from their physiclan. Following physician approval, subjects were oriented to the procedures and potential benefits and risks of the study. An informed consent form was then signed by each subject. Criteria for exclusion from the study were a continuous three-day episode of low back pain in the previous year, prior severe low back dysfunction, or orthopedic pathology of the right upper or lower extremity. The subjects had a mean age of 55.4 years (s = 7.1), a mean height of 162.6 cm (s = 6.4), and a mean weight of 65.9 kg (s = 9.3). Seventeen subjects were currently taking supplemental estrogen, whereas the remaining 39 subjects were not. Table 1 reports the subject characteristics, grouped according to estrogen use.

Materials

The BMD values (in grams per square centimeter) were determined using the Lunar DP3 Dual Photon Scanner.* The technique used was established at the University of Wisconsin using a 153-gadolinium radionuclide source that emits 44 and 100 keV. [22] Bone mineral density was evaluated at four sites: 1) the L2-L4 vertebral bodies, 2) the femoral neck, 3) Ward's triangle, and 4) the greater trochanter of the right femur.

Muscle performance values were measured in the right extremity and calculated as force (in Newtons) or torque (in newton-meters) using the Spark Muscle Examination and Exercise Dosimeter (MEED) 3000 system. The MEED 3000 system consists of a transducer unit that houses the strain gauge and preamplification electronic components. The transducer unit is connected to the control unit. The control unit consists of a central processing unit, a keyboard for variable functions, and a digital display. The transducer unit is composed of specific couplers for use with different attachments, including cuff attachments and a grip attachment for extremity testing. The transducer unit was calibrated using known weights, which were incremented to and released from a maximum of 550 N.

The Spark ECOL Ergonomic Chair + (Fig. 1), used in knee extension and hip flexion testing, consists of a hard seating surface with a seat belt to provide stabilization. other features include a retractable foot platform and multiple eyelet connectors to allow for testing in several positions. An adjustable trolley connector allows for testing of knee extension strength. An adjustable chain allowed the angle of the line of force for strength measurement and the long axis of the body segment to be 90 degrees.

Procedure

Each subject was oriented, gave informed consent, and had both the right femur and lumbar spine BMD scan performed during the first visit. The subject was positioned supine with the hips and knees flexed to 90 degrees to ensure flattening of the lordotic curve during the lumbar spine BMD scan. The supine position with feet in a vertical orientation was used for the hip BMD scan. Total time for the BMD scan was approximately one hour.

* Lunar Radiation Corp, 916 Williamson St, Madison, WI 53703.

+ Spark Instruments and Academics Inc, PO Box 5123, Coralville, IA 52241. Torque was measured during a second visit. Digital displays of each trial and a three-trial mean were monitored on the MEED 3000 system. The mean of three trials served as the force value in the torque calculation. The moment arms were measured from surface correlates of the joint's center of rotation to the midpoint of the cuff attached to the extremity to be tested.

A consistent set of commands were given for each test trial. A cadence of "Ready, set, push, push, push, push" constituted a single trial. The subject was instructed to apply a small amount of preload to the dynamometer-cuff system at "set" and then to contract the tested muscle group maximally for 3 or 4 seconds during the trial. A test consisted of two practice trials (setf-determined 50% maximal effort), followed by three maximal-effort trials. To minimize fatigue, a 30-second rest period followed each trial.

The torques tested were those generated by the hip abductors, the knee extensors, the hip flexors, and the muscles used in gripping. Tests were performed in a randomized order to eliminate fatigue effects. Hip abductor torque was tested with the subject positioned supine on a plinth (Fig. 2). The knees were actively held in extension. The cuff attachments were applied with the distal edge of the cuff 2.54 cm proximal to the distal end of the lateral malleolus. The transducer unit was connected between the cuff attachments. The moment arm was measured from the proximal tip of the greater trochanter to the middle of the cuff. The subject was then instructed to abduct the right leg while an investigator (CLZ) manually stabilized the left lower extremity.

Knee extensor torque was measured with the subject seated in the ergonomic chair (Fig. 1). Knee angles were maintained between 70 and 80 degrees of flexion. The cuff of the adjustable cuff attachment was applied just proximal to the subject's lateral malleolus, and the other end of the device was attached to the ergonomic chair. The transducer unit was connected between the two ends of the cuff attachment. The moment arm was measured from the femoral epicondyle to the center of the cuff on the distal leg. The subject was stabilized using the seat-belt attachment and provided self-stabilization by gripping the front edge of the chair. The subject was instructed to extend the knee.

Hip flexor torque also was measured with the subject seated in the ergonomic chair (Fig. 3). A cuff attachment was positioned so that the line of pull was at a 90-degree angle to the long axis of the femur and was applied to the distal femur proximal to the femoral epicondyle. Hip angles were maintained between 85 and 95 degrees of flexion. The moment arm was measured from the proximal tip of the greater trochanter to the middle of the cuff. Stabilization was achieved with the same method used for knee extension. The subject was instructed to flex the hip.

Grip measurements were taken with the subject seated in the ergonomic chair (Fig. 4). The upper extremity was positioned in neutral adduction and flexion, with the elbow flexed to 90 degrees and the forearm in a midprone position. The wrist was maintained in a neutral position. The subject was instructed to squeeze the handgrip attachment of the MEED 3000 system.

Because use of the MEED 3000 system has not been reported in the research literature, the ability of the instrument to measure forces was determined, as were interrater reliability and intersession reliability. The calibration trials for the MEED 3000 system were performed over a range of 0 to 500 N of force. Interrater and intersession data were obtained on the same day. Data Analysis

An analysis using Bonferroni adjusted t tests was performed to compare mean height, weight, and age of the estrogen and nonestrogen groups. This analysis was performed as a check for group similarity. Interrater reliability and intersession reliability for the strength-testing method were assessed by use of the Pearson product-moment correlation. A data set consisting of the measured weight, height, age, BMD, and strengths was used for the correlational and regression analyses. The BMD and muscle strength correlations were tested using Pearson product-moment correlations. Prediction models-a badward-elimination regression model and a five-variable regression model-were also used. The backward-elimination regression used a cutoff level for significance of .01 for a variable to remain in the model.

Results

Subjects

Descriptive data on the subjects are shown in Table 1. No significant difference was found between the means of the two groups, which allowed analysis of the group as a whole.

Evaluation of Muscle Examination and Exercise Dosimeter (MEED) 3000 System

Values for the interrater and intersession correlations are presented in Table 2. The correlations for interrater reliability ranged from .85 for mean hip flexor torque to .98 for mean grip force. Intersession reliability ranged from .67 for mean hip abductor torque to .90 for mean hip flexor torque.

Bone Mineral Density and Torque Values

The BMD values for each site tested are presented in Table 3. These values are representative of the expected values for women in this age group.[23] Mean percentages of age-adjusted BMD varied from 90.7% to 101.7% over the four sites sampled. Mean torques and grip force and their standard deviations are shown in Table 3.

Relationship Between Bone Mineral Density and Torque

Correlations between BMD, torque, height, weight, and age were calculated. The BMD and muscle torque correlations are presented in Table 4. A significant positive correlation was found between hip flexor torque and all four of the BMD sites. The following significant positive correlations were also noted: hip abductor torque with femoral neck and greater trochanter BMDs; knee extensor torque with L2-L4, femoral neck, and greater trochanter BMDs; and grip force with L2-L4 BMD. The remaining correlations were not significant.

Significant positive correlations were noted between body weight and BMD and between body weight and strength. Thus, weight-normalized correlations between BMD and strength were calculated and are presented in Table 4. Significant correlations following weight normalization were found for grip force and L2-L4 BMD, hip flexor torque and L2-L4 BMD, and hip flexor torque and greater trochanter BMD.

The initial backward-elimination regression model to determine BMD contained all strength measures along with height, weight, and age. The final models from the backward-elimination regression are presented in Table 5. Hip flexor torque was included in two of the models, and weight was included in three models. Height was included in only one of the models. No other variables were included in any models. The coefficients of determination for the models ranged from .1037 to .3366, and all models were significant at the .01 level.

A five-variable regression model containing the four strength values and weight was used to determine the relationship between strength and BMD. The result,,, of the five-variable regression are presented in Table 5.

Discussion

We believe that, within the scope of this study, the measurements of extremity torque were reasonably reliable from an interrater perspective. The lower intersession correlations may be attributable to variation in subject effort. Perhaps the examiners could have been better trained.

The lumbar spine and hip were selected for BMD analysis based on previous research and the high incidence of osteoporotic fractures in these skeletal regions. Hip abduction, hip flexion, and knee extension were selected for strength measurements based on Wolfs law, weight bearing, and antigravity function in relation to the BMD sites selected.

The results of this study provide insight into the different relationships between extremity muscle strength and BMD. For example, we found a significant relationship between hip flexor torque and all four BMD measures. The strongest correlation for hip flexor torque occurred with the L2-L4 BMD, which could be attributed to the attachment sites of the psoas muscles in the lumbar spine. Attachment site may similarly account for the significant correlation between hip abductor torque and greater trochanter BMD. These findings are in accordance with previous literature concerning the effect of muscular loads imposed on bone. [3,13,15,18]

Another significant correlation shown in Table 4 was that between knee extensor torque and femoral neck BMD. The weight-bearing function of an antigravity muscle group, such as the quadriceps femoris musculature, may explain this result. The concept of antigravity function can also be used to address the correlations between knee extensor torque and L2-L4 BMD and between knee extensor torque and greater trochanter BMD. The relationship between hip abductor torque and femoral neck BMD and between hip abductor torque and greater trochanter BMD may also be explained in this manner. Other studies have supported this positive relationship between BMD and weight bearing. [5-7]

The BMD of Ward's triangle showed poor to fair correlations with all measured torque values. As Ward's triangle is located in an area that is not stressed specifically by muscle attachment or weight bearing, perhaps these results should be expected. Grip force correlations with the lower extremity BMD were in the poor to fair range. Therefore, a relationship between BMD and a general indicator of strength-grip force-was not found.

Weight normalization of torque and BMD values produced a decrease in all muscle torque and BMD correlations except those between grip force and all BMD measures and those between all torque measures and L2-L4 BMD. The minimal weight-bearing function of the upper extremity may explain the lack of effect of weight normalization on grip force relationships. The decreased degree of weight bearing in the spine relative to that in the femur may account for the effect of weight normalization on correlations between strength and BMD in the spine and femur.

All backward-elimination regression models for the nonspine BMDs included body weight as a significant factor. it may be concluded that weight is an essential element in the prediction of BMD in the femur. The only model that did not include weight was the L2-L4 model, which may be a result of the relative weightbearing functions of the spine and femur, as stated earlier.

The poor correlations previously cited reveal some apparent inconsistencies, which may be explained by the great number of factors affecting muscle strength and BMD. Some of the factors affecting BMD include smoking, alcohol use, activity level, genetics, number of pregnancies, and the age of menopausal onset. Body size and activity level will tend to affect muscle force output. [10] Ideally, therefore, these confounding factors must be considered when trying to relate a single variable (ie, torque) to BMD. A longitudinal study of similar design considering changes in torque and BMD would be valuable in addressing these associations.

Clinical implications

Noting the significant correlations between the hip flexor torque and BMD of the lumbar spine and proximal femur, torque of the hip flexors may play an important role in the treatment of patients at risk of developing osteoporosis in the lumbar spine or hip.

The five-variable regression model does provide a simple model for the clinician to obtain estimates of a patient's BMD while performing strength assessment. Even though these models explain only 16% to 34% of the BMD, they do provide an indicator of an individual's bone density status.

Conclusions

Based on the results of this study, hip flexor torque in postmenopausal women is significantly related to BMD in the lumbar vertebrae and hip. There is also a significant relationship between torque and BMD of anatomically related areas in the spine and hip of postmenopausal women. The five-variable regression model does give some insight into the BMD status of the lumbar spine and hip. The variables selected, however, do not allow prediction of BMD from the muscle strength measures used in this study.

(Tables and other figures omitted)

References

1 Chesnut CH III, Charles H: Treatment of post-menopausal osteoporosis. Compr Ther 10(7):41-47,1984

2 Hannson T, Roos B: Age changes in the bone mineral density of the lumbar spine in normal women. Calcif Tissue Int 38:328-332, 1984

3 Smith E, Raab D: Osteoporosis and physical activity. Acta Med Scand [Suppl] 711:149-156, 1986

4 Huppert LC: Hormonal replacement therapy: Benefits, risks, and doses. Med Clin North Am 71:27, 1987

5 Issekutz B, Blizzard JJ, Birkhead NC, et al: Effect of prolonged bedrest on urinary calcium output. J Appl Physiol 21:1013-1020, 1966

6 Krolner B, Toft B: Vertebral bone loss: An unheeded side effect of therapeutic bedrest. Clin Sci 64:537-540, 1983

7 Stein H: Bone density in the lower tibia of normal and post-injury limbs. Clin Orthop 174:181-187,1983

8 Mack PB, LaChance PA, Vose GP, et al: Bone demineralization of foot and hand of Gemini-Titan IV, V, and VII astronauts during orbital flights. Am J Roentgenol 100:503-511, 1967

9 Krolner B, Toft B, Nielsen S, et al: Physical exercise as prophylaxis against involutional vertebral bone loss: Controlled trial. Clin Sci 64:541-546, 1983

10 Smith EL, Redden W: Physical activity: Modality for bone accretion in the aged. Am J Roentgenol 126:1297, 1976

11 Aloia JF, Cohn SH, Ostuni JA, et al: Prevention of involutional bone loss by exercise. Ann Intern Med 89:356-358, 1978

12 Simkin A, Ayalon J, Leichter 1: Increased trabecular bone density due to bone-loading exercises in postmenopausal osteoporotic women. Calcif Tissue Int 40:59-63, 1987

13 Nilsson BE, Westlin NE: Bone density in athletes. Clin Orthop 77:179, 1971

14 Burr DB: Lower extremity loading and vertebral bone formation. Spine 8:681-686, 1983

15 Chamay A, Tschantz P: Mechanical influences in bone modeling: Experimental research on Wolff's law. J Biomech 5:173-180, 1972

16 Sinaki M, Optiz JL, Wahner HW: Bone mineral content: Relationship to muscle strength in normal subjects. Arch Phys Med Rehabil 55:508-512, 1974

17 Sinaki M, Offord KP; Physical activity in postmenopausal women: Effect on back muscle strength and bone mineral density of the spine. Arch Phys Med Rehabil 69:277-280, 1988

18 Gillespie JA: The nature of bone changes associated with nerve injuries and disuse. j Bone joint Surg [Am] 36:464-473, 1954

19 Doyle F, Brown J, LaChance C: Relation between bone mass and muscle weight; A comparative review. Lancet 1:391-393, 1970

20 Pogrund H, Bloom RA, Weinberg H: Relationship of psoas width to osteoporosis. Acta Orthop Scand 57:208-210, 1986

21 Wolf J, referred to in Bassett CAL: Effect of force on skeletal tissues. in Downey JA, Darling RC (eds): Physiological Basis of Rehabilitative Medicine. Philadelphia, PA, W B Saunders Co, 1971, p 283

22 User's Manual: Lunar DP3 Dual Photon Scanner. Madison, \VI, Lunar Radiation Corp, p 7

23 Aniansson A, Grimby G, Rundgren A. Isometric and isokinetic quadriceps muscle strength in 70-year-old men and women. Scand J Rehabil Med 12:161-168, 1980

Table 2. Reliability of Torque Mea
-surements (N = 8)
                   Interrater    intersession
Measure             r)            r)
Hip abductor
  torque           .96           .67
Knee extensor
  torque           .97           .73
Hip flexor
  torque           .85           .90
Grip force          98           .73
  (Tables 1, 3, 4, 5, and other figures omitted)