Relationship between trunk muscle torque and bone mineral content of the lumbar spine and hip in healthy postmenopausal women.

Relationship Between Trunk Muscle Torque and Bone Mineral Content of the Lumbar Spine and Hip in Healthy Postmenopausal Women Bone provides the framework upon which other structures of the body are supported. Healthy bone is well suited to this support role because its compressive strength is greater than that of even the best reinforced concrete. (1) Bone tissue is also adaptable, continuously undergoing change in response to imposed mechanical stress. The remodelling of bone occurs through an interplay between osteoclasts, which cause resorption cavities, and osteoblasts, which deposit new bone. This dynamic remodeling activity provices the body with a way of responding to changes in stress, so than an individual exposed to increased stresses will normally exhibit an accelerated rate of bone deposition. (2)

Maximum bone mineral density (BMD) usually occurs early in the third decade of life. Osteoclastic and osteoblastic activity then achieves a balance that maintains the maximum BMD for about another 10 years. Beyond the 30- to 40-year-old age range, the osteoblastic activity fails to completely balance osteoclastic activity, resulting in an increase in the absorption of bone. (3) The increased resorption results in an increased size of the resorption cavities and ultimately in decreased bone mass. The bone loss for women over 40 years of age is greater than the loss experienced by men, although typical rate of loss contonues to be debated. Hansson and Roos (4) report a linear bone loss of approximately 1% a year from maturity to death for women, and Krolner and Nielsen (5) found a bone-loss rate of jearly 6% a year for several years immediately after menopause. Ultimately, the noted loss in bone mass slows as a woman increases in age. (5) The loss in bone mineral content is accounted for by a variety of factors, including the effects of gender, age, diet, physical activity, and differences in hormonal functions. (6) Calcium intake, racial heritage (eg, northern Europeans tend to have low BMD levels, and blacks tend to have high BMD levels), a family history of osteoporosis, cigarette smoking, excessive alcohol intake, and excessive coffee or protein intake also appear to be risk factors associated with increased bone loss. (7) Other factors have been associated with bone loss are menopause, reduced calcitonin, reduced adrenal androgens, and declining calcium absorption. (8)

Whereas the loss of BMD with aging is a normal physiological process, this loss is a particular problem for women. (3,6) Osteoporosis, a reduction in bone mass per unit volume to a level that may lead to fracture, Affects 40% of women be age 60 years. (3,7) In the United States, at least eight million white women develop osteoporosis annually. (7) The scope of this problem has prompted investigation into factors that might slow or reverse the progression of this bone loss.

One promising area of investigation is the potential effect of exercise on bone mineral content. (9-13) Increased physical activity has been associated with an inhibition or reversal of bone loss. Power-lifters' mid-lumbar spine bone mineral content has been shown to be closely related to the training dosage, measured in tons lifted annually. (13) Krolner et al (19) demonstrated that, following an 8-month exercise program, the bone mineral content of the lumbar vertebrae of a group of women, aged 50 to 72 years, increased by 3.5%, whereas that of a control group decreased by 2.7%. Jacobson et al [12] have shown that BMD values were higher among athletic women, aged 55 to 70 years, than among the age-matched control group. Smith et al (14) demonstrated the forearm exercise over a 3-year period resulted in an increase in the bone mineral content of the distal radius of 2.29%, whereas a control group had a decrease of 3.29%. These studies suggest that exercise may be one intervention that can be used to retard the rapid bone loss observed in postmenopausal women.

The hypothesis that has evolved proposes that exercise stresses bone, stimulates its synthesis, and hence decreases bone loss with aging. Weight-bearing stresses are implicated in that when they are not present, such as during prolonged bed rest or the weightless conditions of space, a significant increase in the rate of bone loss occurs. (15-17) Research with athletes has demonstrated that an increase in BMD occurs primarily in the areas in which the forces are applied, a finding that supports the role of bone stimulation via muscle action. (13,18)

Although muscle action has been implicated as a factor in the stimulation of bone synthesis, few studies have attempted to directly measure muscle performance and correlate it with the status of an individual's skeletal health. Sinaki and colleagues (19,20) attempted to measure this association in two separate studies. In both studies, maximum torque values were obtained isometrically. Sinaki et al [19] found no correlation between BMD in the radius and torque values based on a maximum elbow flexion contraction and grip force. They concluded that bone mineral content and torque appear to change independently of each other as age increases. A more recent study by Sinaki and Offord, [20] however, revealed a significant correlation between back extensor torque and the bone mineral content of the lumbar spine.

In view of both the apparent lack of conclusive results and equivocal information concerning the role that strength has in relation to maintaining or reducing the rate of bone loss with aging, further research into this area is needed. Our research dealt with postmenopausal women, because as a group they are at high risk for significantly decreased BMD with aging. The purpose of the study was to determine the relationship between trunk muscle strength and the BMD of the lumbar spine and proximal femur in postmenopausal women. Torque and work measurements were used to assess trunk muscle strength. We hypothesized that BMD would be related to the forces normally generated and that individuals able to generated larger torque and work values would demonstrate enhanced BMD of the lower lumbar spine and proximal femur. This study expands on previous research by examining trunk muscle performance on three types of muscle contractions: isometric and constant-velocity concentric and eccentric contractions. This type of information is needed because instrumentation capable of assessing all three modes is available and routinely used by clinicians and researchers.

Method

Subjects

Trunk torque and work assessment and BMD assessment of the lumbar spine and hip were performed on 61 healthy, white, postmenopausal women, aged 38 to 73 years. Three of the women subsequently developed low back pain following either the orientation seesion or the test session and were excluded from the study. Two additional subjects were excluded because of other medical problems. The findings of this study, therefore, are based on an analysis of the data of the remaining 56 women. Sixteen of the women had previously undergone either a hysterectomy or an oophorectomy, or both procedures, whereas natural onset of menopause was reported by the rest of the subjects. Subjects were recruited by an advertisement in the local newspaper. Criteria used to determine eligibility for participation in the study included 1) physician consent; 2) no incidence of back pain lasting for more than 3 days for at least 1 year prior to initiation of this research; 3) no history of renal, cardiac, neuromuscular, or musculoskeletal dysfunction within the previous year; 4) no current involvement in a weight-training program for abdominal muscles or back extensors; 5) no obesity that would preclude the ability to use trunk testing equipment; and 6) at least 1 year postmenipause, as reported by the subject. The average age at menopause was 45 years (SD = 7, range = 20-55). The average number of years postmenopause was 10 (SD = 6, range = 1-28). Seventeen of the women were receiving estrogen therapy, which is known to affect BMD, and they were followed as a subgroup to control for the potential confounding influence of the medication. No other medications known to affect BMD were taken by any of the subjects. All subjects were informed of the nature and risks of the procedures to be used and signed informed consent forms. Further descriptive statistics are given in Table 1.

Procedure

Bone mineral density. Bone mineral density values (in grams per square centimeter) were determined by dual-photon absorptiometry (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. [21] The four measurement sites evaluated were L2-4, the femoral neck, the Ward's triangle, and the greater trochanter. Each site at the proximal femur provides important information about structural strength. The Ward's triangle is a trabecular bone region within the femoral neck. The femoral neck and greater trochanter values also reflect structural strength of the proximal femur, with the femoral neck supporting the weight of the trunk and the greater trochanter serving as the major distal site of gluteus medius muscle insertion. Reduction in bone mass is one of the most important factors leading to increased frequency of hip fractures in postmenopausal women. [22] Coefficient-of-variation measures yielded a reproducibility for dual-photon absorptiometry of approximately 1.3% for the lumbar spine and 2.2% for the femoral neck. [23] The accuracy of this technique, as determined by ash weight, is approximately 5% to 6% for the lumbar spine and 3% to 4% for the more uniform femoral neck. [24] Previous research has shown that bone strength values correlate highly with dual-photon absorptiometry values. [25]

Kin-Com[R] Trunk Testing Unit. The Kin-Com[R] Trunk Testing Unit(*2) was used to measure the subjects' torque during isometric contractions and during constant-velocity measurements with concentric and eccentric contractions. Measurements obtained with this device have been shown to be reliable, with 94% of the kinetic trunk variables tested in a test-retest situation (peak torque, torque rise, torque decay, torque time integral, and work), providing Pearson product-moment correlation (r) values greater than .90. [26] The manufacturer contends that the Kin-Com[R] capable of measuring forces up to 2,100 n at the point of application, with a resolution of [+ or -] 4 N for force, [+ or -] 1 degree for position, and [+ or -] 2.5% for velocity. Torque values were monitored on-line through the Kin-Com[R]-dedicated computer, and data were collected off-line with a PDP 11/34 computer. (*3) The methodological setup used in this study has been described in detail previously. [26] Kin-Com[R] software enabled a full calibration check at the beginning of each day of exercise testing. The Figure shows the Kin-Com[R] Trunk Testing Unit in a typical configuration for a trunk extension test.

Kinetic variables. Maximal voluntary trunk flexor and extensor torque and work tests were performed under eccentric, concentric, and isometric contractions. Trunk strength was assessed from peak torque values over the initial three contractions generated about an estimated axis at the L5-S1 vertebral location. Work was calculated from the integral of the torque curve generated during the isokinetic flexor or extensor contractions.

Several days prior to testing, each subject was instructed in the trunk muscle-testing procedure and had the opportunity to practice each exercise mode (eccentric and concentric during constant-velocity movements at 20[degrees]sec and isometric). Practice continued with submaxial repetitions until the subject demonstrated an understanding of the procedure by exhibiting smooth force curves. The subject then performed three maximal contractions of each specific mode of exercise in order to become familiar with the amount of effort required during the actual testing.

Within 1 week of the initial instruction, subjects returned for the actual testing session. Subjects were verbally instructed again in each exercise mode. General stretching exercises were performed prior to testing. Warm-up with the Kin-Com[R] system consisted of three submaxial and three near-maximal contractions, followed by a brief rest period. Testing began when the examiner (JSH) observed smooth force curves that were similar in size and shape on the Kin-Com[R] monitor and the subjects expressed that they were rested and felt confident in the exercise procedures.

The subjects were trained and tested in the sitting position with the pelvis and lower extremities stabilized. The axis of the Kin-Com[R] was aligned with the estimated location of the L5-S1 disk space. Based on the work of Herring (TB Herring, unpublished data, 1982), we considered the L5-S1 disk space as the functional axis of spinal rotation. The L5-S1 disk space was estimated following bi-iliac measurements on each subject with calipers. Research with dry adult pelvis specimens has demonstrated that if the bi-iliac measurements are multiplied by 0.15 for women, an estimate can be made of the distance from the surface of the back anteriorly to the L5-S1 interspace (TB Herring, unpublished data, 1982). The force-application pad was placed at the level of the upper manubrium for the flexion test and near the fourth thoracic vertebra for the extension test. The trunk flexors were evaluated isometrically in an upright sitting position for three repetitions of a 3-second maximal effort, with 1- ti 2-second rest periods.

Standardized verbal encouragement was given during the test in an effort to facilitate the response. The torque measurements (eccentric and concentric) were obtained from a movement envelope of 35 degrees (10[degrees] of extension to 25[degrees] of flexion, with 0[degrees] coinciding with the upright sitting position). The constant angular velocity for the isokinetic tests was 20[degrees]/sec. Trunk range of motion (ROM) and isokinetic speed selection were based on patient comfort considerations and previous research. [26,27] Ten maximal repetitions (a concentric contraction followed by an eccentric contraction was considered one repetition) were performed, with a 1-second rest allowed at the end of each arc of motion. Verbal encouragement of "push, push, push" or "pull, pull, pull" was given as appropriate for each contraction. The Kin-Com[R] unit was then repositioned for the trunk extensors, and the measurement procedure was repeated for the extensor muscle group. The correlation and regression analysis was based on the mean values obtained from the initial three maximal contractions of each specific exercise trial. The initial three is isokinetic contractions were used in the analysis to permit equal sample weighting with the three isometric contractions. Pilot work associated with the development of the protocol in this study and in previous studies (RW White, unpublished data, 1987; MA ANderson, unpublished data, 1987), [26] which included a systematic warm-up procedure, also demonstrated that the first 3 of 10 contractions generally yielded the highest peak torque measurements. The potential effects of fatigue over a 10-contraction test and the electromyographic data obtained during these tests will be discussed in a subsequent article.

In addition to explicit instruction requesting maximal effort in all exercise modes, exertion was assessed from the subjects with a 20-point exertion rating scale developed by Borg. [28] The Borg Index is a subjective rating of the intensity of exertion and is linearly related to heart rate. [28] This scale has also been shown to closely follow trends observed for heart rate during exertion with the Kin-Com[R] Trunk Testing Unit. [26]

Data Reduction, Transmission,

Storage, and Analysis

A PDP 11/34 computer calculated torque and work values, and these data were retained in hard-copy form and electronically transmitted to an IBM 4381 mainframe computer. (*4) These values were analyzed with Statistical Analysis System (SAS) software programs. (*5) Descriptive statistics, including means and standard deviations, were obtained on each of the variables analyzed. The associations between strength measurements and BMD measurements were assessed with Pearson product-moment correlation coefficients (r) and with a linear multiple-regression technique (backward-regression technique). All variables listed in Table 2 began in the statistical model described by the equation:

Y = [B.sub.0] + [B.sub.1.X.sub.1] + [B.sub.2.X.sub.2] + ... [B.sub.i.X.sub.i] + E

Those variables not providing a statistically significant contribution based on a .10 level of probability were removed from the model. Two-tailed probability values at the .05 level were considered evidence of a statistically significant relationship.

Results

Bone Mineral Density

The BMD of the four sites sampled was expressed in grams per square centimeter. The mean L2-4 BMD value obtained was 1.13 g/[cm.sup.2], and the mean proximal femur BMD values were 0.78, 0.64, and 0.67 g/[cm.sup.2] for the femoral neck, Ward's triangle, and greater trochanter, respectively. These obtained values were similar to the expected values for women in this age group (age-adjusted mean BMD values [expressed as percentage of expected value] = 90.7%-101.7%). Mean BMD values with their respective standard deviations are given in Table 3.

In addition to the mean values for the entire group, mean values were also determined for two subgroups of women of this population on the basis of their use or nonuse of estrogen supplements. This additional categorization was performed because BMD has been shown to be strongly influenced by hormonal levels, particularly estrogen levels. The estrogen group had lower, but not statistically different, BMD values compared with the non-estrogen group, when evaluated with estrogen as a covariate in the multiple-regression analysis. The variables compared in this analysis included body weight and torque and work values used in the regression models shown in Table 4. This analysis suggests the two subgroups may have coincident regression equations; therefore, both subgroups were considered functionally as one group in this analysis.

Kinetic Variables

Torque values for concentric, eccentric, and isometric contractions of the trunk flexors and extensors are presented in Table 5. The trunk extensor values were consistently higher than those obtained for the trunk flexors. In addition, eccentric contraction values were higher than either isometric or concentric contraction values across both muscle groups tested. Isometric torque values tended to be greater than the concentric torque |alues. Calculated work values are also presented in Table 5.

Mean exertion levels, as measured by the Borg Index, suggest the subjects performed near-maximal efforts for both the trunk flexors and extensors. A value of 15 on the Borg Index indicates hard effort, 17 is very hard effort, and 20 is maximal effort. [28] The subjects' mean exertion values were 16.1 for isometric trunk flexion and 16.7 for extension.

Trunk muscle torque values, anthropometric measurements, and BMD values were examined with Pearson's product-moment correlation procedure to determine whether any of the four BMD sites assessed had a linear relationship with these variables. Correlation values ranged from -- .21 to .52 overall and from .09 to .47 for kinetic variables (Tab. 2). The subjects' weight exhibited the strongest simple correlation with the four BMD sites, whereas age was negatively correlated with BMD. Most of the kinetic variables (67.5%) demonstrated a significant simple linear relationship with BMD. This relationship was evidenced by probability values that were sufficiently extreme, that at the .05 level, a relationship significantly different from zero existed between the variables tested and BMD.

Although many of the Pearson product-moment correlation values obtained from simple correlation analysis were significant at the .05 or the .01 level, no clear trend was demonstrated between BMD and torque values, suggesting the need for further analysis. The follow-up analysis used was a backward-regression technique. Although some differences existed among the variables examined in the linear multiple-regression equation, depending on the site of BMD assessed, some consistent patterns developed. Torque values for eccentric and concentric contractions tended to be better predictors of BMD and yielded higher regression values than when the isometric values were considered in the model. Height, age, and isomeric torque measurements were not retained in any of the derived models. Weight sas an important variable in 75% of models obtained, but its influence did not override the importance of other torque variables. Significant partial correlation coefficients were found between the torque measure and BMD even after controlling for weight. The obtained linear multiple-regression models explained from 18% to 44% of the variability over the four sites considered. The best-fit linear multiple-regression models are presented in Table 4.

Discussion

The correlation and linear multiple-regression results suggest a positive relationship between trunk muscle strength and BMD in the spine and proximal femur. Although no signle strength measure emerged as strongly predictive of BMD in the four regions studied, a combination of strength and intrinsic anthropometric measures provided evidence of an improved relationship between strength and BMD. The low correlations obtained between age and BMD were likely due to the large age range of this sample and the high degree of variability associated with BMD across ages. The positive relationship between trunk torque and work values and BMD values suggests that women who are ablt to generate high torque values will also demonstrate high BMD values. High bone mineral content levels are desirable, because they are an indirect measure of bone's ultimate compression strength and they have been related to the incidence of fractures. [25]

The correlation coefficients between those variables measuring trunk strength and BMD ranged from .09 to .47 (Tab. 2). When a single strength variable was considered, trunk extensor values, generally had a higher correlation with BMD than did trunk flexion values, and the isokinetic extension strength values demonstrated higher correlations than the isometric extension values in every case (Tab. 2). Despite the lower correlation of flexion torque values than extensor torque values to BMD, trunk flexor strength contributed independent information when considered in the multiple-regression model.

The significant relationships between torque and work values and BMD values should be viewed in the context of the entire individual. As stated previously, BMD is potential affected by a myriad of factors, including diet, hormonal levels, race, calcium intake, heredity, and smoking. [6, 7] Despite the many factors that can influence BMD, it is noteworthy that torque, work, and weight values emerged as having a positive relationship, explaining from 18% to 44% of the variability over the four sites assessed (Tab. 4). Therefore, even though cause and effect cannot be inferred from a correlational study of this type and even though other factors are certainly associated with BMD, the variability in BMD explained by trunk strength suggests that this factor is potentially important.

These findings seem to be in agreement with those of Sinaki and Offord, [20] who found a significant positive correlation of .34 between trunk extensor force and the BMD of the lumbar spine. The strength measurements in their study were performed isometrically with a strain-gauge dynamometer with the patient positioned prone, in contrast to our research in which the subject was in a sitting position and asked to perform isometric, concentric, and eccentric contractions. Our isometric correlation values for the lumbar spine BMD are similar to those reported previously. The correlation values for eccentric and concentric contractions reported in this study are higher than previously cited values. Concentric and eccentric peak trunk extension torque values produced correlations of .46 and .47, respectively, when compared with the lumbar spine BMD values. Sinaki and Offord [20] reported that when examining regression models, age and height were the two most predictive variables and other variables added nothing further to define the relationship. In contrast, our research on post-menopausal women demonstrated torque to be an important predictive variable and age and height to be of lesser value. Positioning and methodological differences may have accounted for some of these observed differences.

If strength or the tensile force developed by a muscle directly influences bone production, as suggested by Sinaki and Offord, [20] then it might be expected that the BMD values at muscular attachment sites would demonstrate this effect. Of the four sites tested in our study, the two sites that have direct muscular attachment had the highest multiple-regression correlation values following the backward-regression procedure. The lumbar spine is the proximal attachment site for the psoas major muscle, and the greater trochanter is the distal attachment site for the gluteus medius muscle. Therefore, even though the exact nature of the stress postulated to be responsible for the observed differences in BMD across women of varying strength levels cannot be determined from this research, it appears that contractile forces, in addition to weight-bearing stresses, could influence BMD, particularly in areas where the influencing muscles are immediately contiguous to the bones in which BMD is being assessed.

The possible importance of testing the strength of the muscles immediately adjacent to the bones being examined is illustrated by the conflicting results between our research and that of Chow et al. [29] In their study of 31 postmenopausal women, aged 50 to 59 years, Chow et al [29] found no significant correlation between muscle force production and the calcium bone index. They used a leg-press maneuver, with the knees flexed 110 degrees, to test leg muscle performance and a bench-press maneuver to test arm muscle performance. The calcium bone index of the trunk and proximal thigh were then examined. Much of the force generated to accomplish these tasks, particularly the bench-press maneuver, occurred in muscle groups distant from the bony site examined. Additionally, a maneuver such as a bench press or a leg press tests only that weight that an individual can lift through the weakest part of the ROM being examined. The Kin-Com[R] Trunk Testing Unit, however, permitted maximal torque assessment at all points in the ROM across eccentric and concentric modes of contraction. The procedures used in our study examined muscle groups located in the anterior trunk and lumbar region, and the assessment technique allowed the determination of maximum achievable torque values. Therefore, these methodological differences in location and testing account for the observed differences in results between our study and that of Chow et al. [29]

The force production of the back extensors has also been found to have a significant positive correlation with physical activity. [20] This relationship is noteworthy, because increased physical activity levels have been associated with improved BMD. [11] Although torque values rather than physical activity were assessed in our study, the finding of a positive relationship between strength and BMD is in agreement with research that has examined the effect of increased activity levels on BMD. Aloia et al, [11] Smith et al, [14] and Chow et al [29] all found that bone loss in aging women could be retarded or reversed through increased physical activity. In research with turkeys, where periods of hormonally mediated bone loss could be controlled, Lanyon et al [30] found that normal functional activity inhibits resorption and thus exerts a protective and conservative effect on bone mass. Although our study was not designed to examine the mechanisms that may be associated with the relationships between strength, physical activity, and BMD, it appears that strength, like physical activity, may be an important variable related to BMD values.

Previous research has not examined the relationship between trunk flexor torque and BMD. The results of our study, however, suggest that such a relationship does exist. For the osteoporotic individual with a compression fracture, our results support Lukert's recommendation that "once acute pain has subsided the patient should be instructed in extension exercises and exercises to strengthen abdominal muscle." [3(p12)] Aisenbrey [18] and Sinaki, [31] however, do not recommend flexion exercises for the individual with osteoporosis, because they speculate that the increased compression forces on the vertebrae can result in compression fractures. Although our study was performed on healthy subjects rather than osteoporotic individuals, the findings suggest that it is important to consider the trunk flexors in trunk muscle strengthening programs. Trunk muscle strengthening prior to the onset of menopause may help maintain the BMD of the lower spine. In osteoporotic individuals, however, flexion exercises should be approached cautiously, at least in the early stages of engaging in a new exercise lrogram, because of the potential of additional compression fractures associated with their use.

In terms of the individual at risk of developing osteoporosis, it should also be pointed out that although fractures can occur in a number of sites, the most commonly affected areas are those that have a predominance of trabecular bone. The disproportionate loss of trabecular bone could be caused by the relatively higher osteoclastic and osteoblastic activity in trabecular bone as compared with cortical bone. Reinbold et al [32] have pointed out that trabecular bone has a turnover rate that is approximately eight times that of cortical bone, and, as a consequence, it is more responsive to metabolic stimuli and may exhibit greater changes in response to disease and aging. The higher turnover rate, in part, accounts for the frequent finding of fractures of the spine, hip, or distal forearm in patients with osteoporosis. Because we examined areas that were predominantly trabecular bone, the findings of our study also suggest that maintaining good trunk muscle strength may be beneficial to patients at risk for osteoporosis.

It is noteworthy that, although the Kin-Com[R] Trunk Testing Unit is in use in many clinical setting, we have been unable to find other reports of the use of this equipment for testing women in this age group. Two limitations associated with this research, therefore, are the lack of population-specific torque and work values for comparison and the lack of reports of subjects' tolerance to this type of testing. During the initial evaluation session and subsequent test sessions, residual muscle soreness lasting several days was occasionally reported, and, in the case of four of the five individuals excluded from the study, complaints associated with the musculoskeletal system persisted beyond what was expected.

Extensive testing of younger populations with this same protocol has not resulted in complaints other than transient muscle soreness. The incidence of injury in this study may have been due to the age and generally reduced activity level of this particular population, coupled with the intensive nature of this form of trunk muscle strength testing. Therefore, the subject's age, health, and previous activity level should be considered in terms of potential risks to the musculoskeletal system prior to this type of testing.

Although a relationship between trunk muscle torque and BMD was evident in this study, further research is needed to clarify the specific mechanisms responsible for this relationship. Additionally, longitudinal studies are needed to determine whether exercise can be used to retard bone loss in postmenopausal women, and, if so, what specific exercise prescription should be used. Further studies determining whether previous activity levels directly influence BMD also need to be undertaken. Ongoing research to answer some of these questions is being conducted in our laboratory.

(*1) Lunar Radiation Corp, 303 3 Beltline Hwy, Madison, WI 53713.

(*2) Chattecx Corp, 101 Memorial Dr, PO Box 4287, Chattanooga, TN 37405.

(*3) Digital Equipment Corp, 146 Main St, Maynard, MA 01754.

(*5) International Business Machines Corp, 1000 NW 51st St, Boca Raton, FL 33432.

(*5) SAS Institute Inc, Cary, NC 27511.

References

[1] Guyton A. Textbook of Human Physiology. 5th ed. Philadelphia, PA: WB Saunders Co; 1976:1057.

[2] Eisenberg E, Gordon G. Skeletal dynamics in man measured by nonradioactive strontium. J Clin Invest. 1961;40:1809-1825.

[3] Lukert B. Diagnosis and management of osteoporosis. Compr Ther. December 1984;10:8-14.

[4] Hansson T, Roos B. Age changes in the bone mineral of the lumbar spine in normal women. Calcif Tissu Int. 1986;38:249-251.

[5] Krolner B, Nielsen S. Bone mineral content of the lumbar spine in normal and osteoporotic women: cross-sectional and longitudinal studies. Clin Sci. 1982;62:329-336.

[6] Dawson-Hughes B, Li X. Preventing osteoporosis: evidence for diet and exercise. Geriatrics. 1987;42:76-82.

[7] Chesnut C. Treatment of postmenopausal osteoporosis. Compr Ther. July 1984;10:41-47.

[8] Francis RM. The aging spine. In: Hukins DWL, Nelson MA, eds. Osteoporosis in the Elderly. Oxford, England: Oxford University Lress; 1987:18-39.

[9] Krolner B, Toft B, Nielsen S, Tondevold E. Physical exercise as a prophylaxis against involutional vertebral bone loss: a controlled trial. Clin Sci. 1983;64:541-546.

[10] Stillman R, Lohman T, Slaughter M, Massey B. Physical activity and bone mineral content in women aged 30 to 85 years. Med Sci Sports Exerc. 1986;18:576-580.

[11] Aloia J, Stanton H, Ostuni J, et al. Prevention of involutional bone loss by exercise. Ann Intern Med. 1978;89:356-358.

[12] Jacobson P, Beaver W, Grubb S, et al. Bone density in women: college athletes and older athletic women. J Orthop Res. 1984;2:328-332.

[13] Granhed H, Jonson R, Hansson T. The loads on the lumbar spine during extreme weight lifting. Spine. 1987;12:146-149.

[14] Smith E, Reddan W, Smith P. Physical activity and calcium modalities for bone mineral increase in aged women. Med Sci Sports Exerc. 1981;13:60-64.

[15] Krolner B, Toft B. Vertebral bone loss: an unheeded side effect of therapeutic bed rest. Clin. Sci. 1983;64:537-540.

[16] Mack P, Vogt F. Roentgenographic bone density changes in astronauts during representative Apollo space flight. American Journal of Roentgenology, Radium Therapy, and Nuclear Medicine. 1971;113:621-633.

[17] Hansson T, Sandstrom J, Roos B, et al. The bone mineral content of the lumbar spine in patients with chronic low back pain. Spine. 1985;10:158-160.

[18] Aisenberry JA. Exercise in the prevention and management of osteoporosis. Phys Ther. 1987;67:1100-1104.

[19] Sinaki M, Opitz J, Wahner H. Bone mineral content: relationship to muscle strength in normal subjects. Arch Phys Med Rehabil. 1974;55:508-512.

[20] Sinaki M, Offord K. Physical activity in post-menopausal women: effect on back muscle strength and bone mineral density of the spine. Arch Phys Med Rehabil. 1988;69:277-280.

[21] User's Manual: Lunar DP3 Dual-Photon Scanner. Madison, Wis: Lunar Radiation Corp; 1985:7.

[22] Consensus conference. Osteoporosis. JAMA. 1984;252:799-802.

[23] Dunn W, Wahner H, Riggs B. Measurement of bone mineral content in human vertebrae and hip by dual-photon absorptiometry. Radiology. 1980;136:485-487.

[24] Wahner H, Dunn W, Mazess et al. Dual-photon Gal-153 absorptiometry of bone. Radiology. 1985;156:203-206.

[25] Hansson T, Roos B, Nachemson A. The bone mineral content and ultimate compressive strength of lumbar vertebrae. Spine. 1980;5:45-55.

[26] Smidt G, Blanpied P, White R. Exploration of mechanical and electromyographic responses of abdominal and back extensor muscles to high density exercise. Spine. 1989;14:815-830.

[27] Smidt GL, Blanpied P, Anderson M, White R. Comparison of clinical and objective methods of assessing trunk muscle strength: an experimental approach. Spine. 1987;12:1020-1024.

[28] Bar-Or O, Skinner JS, Buskirk ER, Borg G. Physiological and perceptual indicators of physical stress in 41- to 60-year-old men who vary in conditioning level and in body fatness. Med Sci Sports. 1972;4:96-100.

[29] Chow R, Harrison J, Brown C, Hajek V. Physical fitness effect on bone mass in post-menopausal women. Arch Phys Med Rehabil. 1986;67:231-234.

[3] Lanyon L, Rubin C, Baust G. Modulation of bone loss during calcium insufficiency by controlled dynamic loading. Calcif Tissue Int. 1986;38:209-216.

[31] Sinaki M. Postmenopausal spinal osteoporosis: physical therapy and rehabilitation principles. Mayo Clin Proc. 1982;57:699-703.

[32] Reinbold W, Genant H, Reiser U, et al. Bone mineral content in early postmenopausal and postmenopausal osteoporotic women: comparison of measurement methods. Radiology. 1986;160:469-478.

J Halle, MS, PT, is a doctoral student in the physical therapy track in the Department of Exercise Science, The University of Iowa, 2600 Steindler Bldg, Iowa City, IA 52242. Address all correspondence to MAJ Halle at Physical Therapy Branch, Academy of Health Sciences, US Army, Fort Sam Houston, TX 78234-6100 (USA).

G Smidt, PhD, PT, FAPTA, is Professor, Physical Therapy Graduate Program, College of Medicine, The University of Iowa.

K O'Dwyer, MA, PT, is Interim Physical Therapy Director, Hand and Orthopedic Rehabilitation Associates, Raleigh, NC 27702.

S-Y Lin, MA, PT, is a physical therapy faculty designee in Taiwan.

Ms O'Dwyer and Ms Lin were master's degree students, Physical Therapy Education, College of Medicine, The University of Iowa, when this study was conducted.

This study was supported by grant RR59 from the General Clincal Research Centers Program, Division of Research Resources, NIH, and by a seed grant on aging from The University of Iowa.

This study was approved by the Human Subjects Review Committee, The University of Iowa.

The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the US Department of the Army or the US Department of Defense.