JOURNAL OF BONE AND MINERAL RESEARCH Volume 7, Number 11, 1992 Mary Ann Liebert, Inc., Publishers

Bone Mineral Density, Muscle Strength, and Recreational Exercise in Men CHRISTINE SNOW-HARTER,' ROBERT WHALEN,' KATHY MYBURGH,' SARA ARNAUD,' and ROBERT MARCUS'

ABSTRACT Muscle strength has been shown to predict bone mineral density (BMD) in women. We examined this relationship in 50 healthy men who ranged in age from 28 to 51 years (average 38.3 years). BMD of the lumbar spine, proximal femur, whole body, and tibia were measured by dual-energy x-ray absorptiometry (Hologic QDR lOOOW). Dynamic strength using one repetition maximum was assessed for the biceps, quadriceps, and back extensors and for the hip abductors, adductors, and flexors. Isometric grip strength was measured by dynamometry. Daily walking mileage was assessed by 9 week stepmeter records and kinematic analysis of video filming. Subjects were designated as exercisers and nonexercisers. Exercisers participated in recreational exercise at least two times each week. The results demonstrated that BMD at all sites correlated with back and biceps strength (p c 0.01 t o p = 0.OOOl). Body weight correlated with tibia and whole-body BMD (p < 0.001); age negatively correlated with Ward's triangle BMD (p c 0.01). In stepwise multiple regressions, back strength was the only independent predictor of spine and femoral neck density ( R 2= 0.27). Further, back strength was the most robust predictor of BMD at the trochanter, Ward's triangle, whole body, and tibia, although biceps strength, age, body weight, and leg strength contributed significantly to BMD at these skeletal sites, accounting for 35-52070 of the variance in BMD. Exercisers and nonexercisers were similar for walking (3.97 versus 3.94 miledday), age (37.8 versus 38.5 years), and weight (80.0 versus 77.7 kg). However, BMD and muscle strength were significantly greater in exercisers than in nonexercisers. In conclusion, in young to middle-aged men, (1)muscle strength makes important contributions to bone mineral density; (2) strength of back extensors more powerfully predicts BMD than age, body weight, or strength of other muscle groups; and (3) recreational exercise is a better predictor of BMD than habitual daily walking.

INTRODUCTION (BMD) has been demonstrated to be higher in physically active men than in sedentary controls.('-41 A growing body of evidence suggests that higher BMD may be a function of greater muscle strength. 15-11) Whether the association between strength and BMD is genetic or a function of physical training is yet to be established. Given that mechanical loading stimulates bone accretion,(''-'4) it seems likely that the correlation between strength and BMD is due, at least in part, to the increased magnitude of loading by stronger muscles.

B

ONE MINERAL DENSITY

Much of the research that has assessed muscle strength and BMD has been in women. ( 6 . 7 . 9 , 1 0 ) In two investigations that included men, Bevier and colleagues(') found that isometric back strength correlated with spine density and Block et al. l U 1 reported a strong correlation between paraspinous cross-sectional area and BMD, but not muscle strength. Other investigators have reported results that imply a relationship between muscle strength and BMD. 11.5.11.15.171 For example, Aloia and colleagues(1) found a positive relationship between indices of bone mass and lean body mass in male runners and Doyle et aI.ts) demonstrated a strong correspondence of psoas muscle

'Musculoskeletal Research Laboratory, Aging Study Unit, Geriatrics Research, Education & Clinical Center, VA Medical Center, and Department of Medicine, Stanford University, Palo Alto, California. 'NASA Ames Research Center, Moffett Field, California. 1291

1292

SNOW-HARTER ET AL.

weight to ash density of the third lumbar vertebra. Further, Granhad and colleagues(’5’reported that professional weight lifters had a significantly higher bone mineral content than controls. In fact, the mineral content of the third lumbar vertebra increased as the annual weight lifted increased. In another cross-sectional report, Colletti et al.(’’) reported higher BMD of the lumbar spine, trochanter, and femoral neck in men who had been weight training for at least 1 year compared to controls. It seems that the muscle group attached to the measured skeletal site would confer the most mechanical stress on the bones and therefore best predict BMD at that site. However, the nature of the relationship between muscle and bone has not always been site specific. For example, we found in young women that biceps strength was the most robust predictor of hip density whereas grip strength independently predicted spine density.‘61 Bevier et al. (7) observed that grip strength best predicted spine density in elderly women. Pocock et a1.(9)reported biceps strength as an independent predictor of spine and hip density in women. In these cases, the muscles attached to the sites of bone density measurements were not the independent predictors of BMD. Studies conducted in men have reported a site-specific relationship between back strength and spine mineral d e n ~ i t y , ‘as ~ )well as a strong correlation between paraspinous muscular area and both spine and hip density.(81 In this study, we report data on muscle strength and BMD in a group of healthy, young to middle-aged men. The purposes of this investigation were as follows: (1) to examine the relationship between muscle strength and bone mineral density, (2) to determine the specific nature of the relationship between muscle and bone, and 3) to define the effect of exercise patterns on BMD and muscle strength.

MATERIALS AND METHODS

Subjects A group of 50 healthy white men aged 28-51 years were recruited from the San Francisco Bay area. All had a bone mineral index (BMI, weight/height’) within 10% of the normal range of 24-27(16) and were free of disorders known to alter bone metabolism. The protocol was approved by the Human Subjects Committees of Stanford University and the NASA Ames Research Center. All subjects gave written informed consent.

Bone mineral assessment Bone mineral density (BMD = g/cm’) of the right proximal femur, lumbar spine (L2-4), whole body, and right tibia was evaluated by dual-energy x-ray absorptiometry (Hologic QDR 1O00, Waltham, MA). The tibia value was derived from whole-body data, and analysis was performed by direct edge detection software. Whole-body software permits regional analysis, which allows the operator to outline the tibia in detail and exclude the fibula and other adjacent sites.

In our laboratory, the coefficients of variation on repeated tests of 12 adult men within 7 weeks of each other were as follows: lumbar spine BMD = 0.97%, femoral neck BMD = 0.95%, trochanter BMD = 0.60%, Ward’s triangle BMD = 1.35%, whole-body BMD = 0.40%, and right tibia BMD = 0.85%.

Strength measurements Dynamic strength of the quadriceps, biceps, back extensors, and hip adductors and abductors was assessed using isotonic weight equipment (Universal and Nautilus). One repetition maximum (1 RM) was the end measure and represented the maximum weight a subject could lift one time with acceptable form. “Acceptable form” meant that performance of the exercise isolated primarily the muscle group being tested. Criteria for terminating the test included a change in body position during the lift and/or not completing the full range of motion. When the criteria for acceptable form were not satisfied, the last recorded successful weight was taken as the strength score. Subjects received detailed instructions and performed each exercise several times at very low weight for familiarization and warm-up. The 1 RM test began at a high weight to minimize repetition fatigue. In the majority of cases, each subject required no more than three repetitions to reach maximum. All tests were repeated with incremental weights of 2.5, 5 , and 10 pounds, dependent upon the exercise, until the subject was not able to lift additional weight with acceptable form. A 20 s rest period was given between repetitions, with at least 2 minutes of rest between exercises. The strength test protocol has been described in detail. C 6 ) Coefficients of variation on repeated tests of 20 adults within a 1 week period were as follows: quadriceps = 3.0%, biceps = 6.O%, back extensors = 1.0%, hip adductors = 2.0%, and hip abductors = 2.0%.

Activity histories and daily walking mileage Information with regard to activity history and daily mileage was collected at the NASA Ames Research Center. Each subject completed a 1 year activity history and was then interviewed. Stepmeters (Micronata; Tandy Corp., Fort Worth, TX) were worn 7 days per week for 9 weeks,

VALUES f STANDARD DEVIATION FOR TABLE1. AVERAGE STRENGTH AND BONEMINERAL DENSITY IN 50 HEALTHY MEN

BMD (g/cm’)

Strength (pounds)

Back 198 + Quadriceps 124 f 68 f Biceps Dominant grip 125 f Hip flexor 79 f Hip abductor 87 f Hip adductor 108 f

33 27 15 22 16 19 17

Lumbar spine Femoral neck Trochanter Ward’s Total hip Whole body Right tibia

1.088 f 0.17 0.896 f 0.14 0.767 f 0.11 0.720 f 0.15 1.018 f 0.14 1.171 f 0.08 1.430 f 0.15

1293

BONE DENSITY AND EXERCISE IN MEN except during recreational exercise. The stepmeters registered the number of steps each day, and the subject then recorded this value in his daily activity log. Calibration of the stepmeters was performed 8-10 times during the 9 week study period. Subjects received log books in which they recorded daily activity (sitting, standing, and recreational exercise) and number of steps per day. Daily mileage was computed from the number of steps per day and the stride length. Stride length was calculated from kinematic analysis of video filming (Peak Performance Technology, Colorado Springs, CO). Upon evaluation of exercise patterns over the previous year and the 9 week study period, the group was divided into exercisers (n = 27) and nonexercisers (n = 18). The exercisers participated in physical activity (soccer, basketball, tennis, racquetball, weight training, and jogging) at least two times per week for a minimum of 1 h. The nonexercisers did not participate in regular physical activity. From the original group of 50 subjects, 5 men were omitted from this subgrouping because of unreliable recording of activity.

Data management and analysis Data were stored on a personal computer and analyzed using the Statview I1 software package (Abacus Concepts, Inc., Berkeley, CA). Analyses included standard descriptive statistics, repeated-measures analysis of variance, and simple and multiple stepwise regressions. In the simple regression analyses, we accommodated the use of multiple comparisons by requiring an a level of 0.01 for significance. Correlations were derived from simple regression equations.

RESULTS The 50 subjects were 38.3 f 7.0 years of age. Body weight, height, and BMI were 79.7 f 8.9 kg, 177.9 6.7 cm, and 22.5 =t2.3, respectively. Average values for bone

*

mineral density and muscle strength are presented in Table 1. Normal distributions were found on all strength measures.

Relationships between muscle strength and bone mineral density Lumbar Spine Measurements: Spine BMD was not significantly related to age, weight, or height. Correlations between spine BMD and age, weight, and muscle strength are presented in Table 2. By stepwise multiple-regression analysis, including weight, back strength, biceps strength, quadriceps strength, and hip adductor strength, only back strength proved to be an independent predictor of spine mineral density (R’ = 0.27; BMD = 0.003 x back strength + 0.53). Regional Hip Measurements: Correlations between muscle strength parameters and BMD of the femoral neck, trochanter, and Ward’s triangle are presented in Table 2. Of the regional measurements, only Ward’s triangle correlated significantly with age, and no significant relationships between regional hip BMD measures and body weight were observed. In stepwise multiple-regression analysis, back strength was the only independent predictor of femoral neck BMD (Fig. l), whereas both back strength (0 weight = 0.45) and biceps strength (p weight = 0.28) predicted trochanter BMD. Independent predictors of Ward’s triangle BMD included back strength (0 weight = 0.45) and age (0 weight = -0.38). The multiple-regression equations were as follows: Femoral neck BMD = 0.002back strength + 0.46 R’ = 0.27 Trochanter BMD = 0.001back strength + 0.002biceps R’ = 0.38 strength + 0.34 Ward’s triangle BMD = 0.002back strength - O.008age + 0.62 R’ = 0.35

Whole-Body Measurements: Whole-body BMD was significantly related to body weight, but not height or age.

TABLE2. CORRELATIONS (R) BETWEEN BONEMINERAL DENSITY AND AGE, WEIGHT, AND MUSCLE STRENGTH (n = 50)

Age Weight Muscle strength Back Biceps Quadriceps Hip adductor Hip abductor Hip flexor Grip ap d 0.01. bp 5 0.001. cp = o.oO01.

Lumbar spine

Femoral neck

Trochanter region

Ward!. triangle

Whole body

0.02 0.38a

0.08

0.08

0.35

-0.37a 0.17

0.002 0.51b

0.08

0.3 1

0.W 0.41a 0.47b 0.39a 0.27 0.02 0.37a

0.56c 0.39a 0.44a 0.34 0.21 0.05 0.38a

0.62c 0.52h 0.W 0.47h 0.31 0.11 0.45a

0.48b 0.38a 0.39a 0.34 0.24 0.12 0.34a

0.62c 0.46b 0.48b 0.44b 0.26 0.13 0.36a

0.53b 0.38a 0.40a 0.35 0.32 0.10 0.39a

Tibia 0.46b

1294

SNOW-HARTER ET AL.

100

120

1.10

160

180

1-00

220

240

260

Back strength (Ibs)

5 100

I20

I30

I60

I80

200

1-20

240

260

Back strength (Ibs)

FIG. 1. Correlation between regional femoral neck density and back strength in 50 healthy men.

FIG. 2. Correlation between whole-body mineral density and back strength in 50 healthy men.

Correlation coefficients are presented in Table 2. In stepwise multiple-regression analysis, independent predictors of whole-body BMD included back strength, leg strength, and body weight (R2= 0.52; BMD = 0.001back strength + 0.001quadriceps strength + 0.001weight + 0.71; p weights = 0.41, 0.26, and 0.26, respectively; Fig. 2).

sity at all skeletal sites measured. Our results further suggest that regular recreational exercise makes a more important contribution to BMD than habitual daily walking. We recognize that this relationship could be independent of muscle strength but suggest that the effect of exercise on bone mineral density may be mediated through mechanical stress conferred on the skeleton by muscular action. This speculation is based on our observation that muscle strength was significantly higher in those men who exercised on a regular basis. Based upon our previous results in young women(6)and other reports in men and ~ o m e n , ( ~ - ' O we) hypothesized that the strength of the biceps and back extensors would independently predict BMD. Our results support this hypothesis. The measurement of back strength isolates primarily the deep back muscles (erector spinae group) of the lumbar area. During the back extension, movement occurs at the lumbar vertebrae. We consider it likely that the correlation between back strength and lumbar BMD comes about by direct attachment of the erector spinae group to the vertebrae. Other groups have found back strength to correlate with spine BMD. Bevier et al.(7)reported a significant correlation between isometric back strength and spine BMD in elderly men. A study by Sinaki et aI.('O) found a similar relationship in postmenopausal women. The positive relationship between back strength and mineral density of the hip, tibia, and whole body, although not site specific in the sense of direct attachment, is not surprising. Given the central role of the back extensors as trunk stabilizers in all vigorous exercise activities, it makes sense that back strength would be related to BMD at peripheral skeletal sites. These results are consistent with our previous study in women in which we observed back strength to independently predict femoral neck mineral density. Biceps strength proved to be an independent predictor of BMD at the femoral trochanter. These data support our earlier work(6)and that of Pocock et al.(9)It appears that, once again, the relationship between muscle strength and BMD is more complex than a simple site-specific model. From a biomechanical standpoint, the relationship be-

Tibia Measurements: Mineral density of the right tibia was significantly related to body weight and strength of the back, quadriceps, biceps, and hip adductors (Table 2). In stepwise multiple-regression analysis, back strength and body weight best predicted mineral density of the tibia (R' = 0.38; BMD = 0.001back strength + O.OO5weight; 0 weights = 0.43 and 0.29, respectively).

Muscle Strength and Bone Mineral Density in Exercisers Versus Nonexercisers The exercisers and nonexercisers were similar for age (38.1 f 7.1 versus 38.9 f 7.5 years), height (178.9 f 6.2 versus 176.7 f 6.2 cm), weight (80 f 7.2 versus 78 9.2 kg), BMI (22.3 f 2.4 versus 22.5 + 2.3), and miles walked per day (3.9 f 1.3 versus 3.8 + l.l), respectively. A comparison of bone mineral density and strength values demonstrated significant differences between these groups. BMD at all sites was significantly greater in the exercisers versus the nonexercisers (Fig. 3). Additionally, the strength of most major muscle groups tested was significantly higher in the exercisers compared to the nonexericsers (Fig. 4).

*

DISCUSSION We conducted this investigation to evaluate the relationship of muscle strength and bone mineral density in young to middle-aged men and to determine the contribution of exercise patterns to skeletal status. The results demonstrate that muscle strength is an independent predictor of bone mineral density. In particular, the strength of the back extensors was the most robust predictor of bone mineral den-

1295

BONE DENSITY AND EXERCISE IN MEN

fore, we consider it very unlikely that the results we observed were due to differences in precision. We observed that men engaged in regular exercise had higher BMD and muscle strength that nonexercising men despite that both groups reported similar daily walking habits. It is probable that the activities in which the exercisers participated conferred a higher magnitude of mechanical loading to the skeleton than is associated with walking activities. We suggest that a large proportion of mechanical loading occurs via muscle pull and that the skeleton adapts to the increased magnitude of loading by depositing mineral. This notion is supported by the results of Colletti and colleagues,(") who reported higher spine and hip density in men performing weight training for 1 year compared with nonexercising men. We emphasize that a cross-sectional analysis does not permit cause-and-effect statements to be made. However, I Slurk I Vttk Tnthrnici Uml. U fhdb RTiliin our findings demonstrate strong correlations between musFIG. 3. Bone mineral density differences in exercising (n cle strength and mineral density in men. We note that the = 27) versus nonexercising (n = 18) men. L Spine, lumbar variance in bone mineral of the spine and hip explained by spine; F Neck, femoral neck; W Body, whole body; R muscle strength was greater in the men studied here Tibia, right tibia. (27-38%) than in the women (15-20%) in our previous study.(6)This difference is attributed to the fact that our group of men had a wider range of both muscle strength and age. We acknowledge that other factors not measured here, that is, genetic endowment, reproductive hormone status, and nutrient intake, may also make significant contributions to bone mineral density.

ACKNOWLEDGMENTS

Biceps

Hip Add

Back I

XI

Leg Lxt

I> Grip

The authors thank Theresa Hutchinson and Tammy Tamanaha for the collection and analysis of the activity level histories and D.B. O'Hara, R.N., manager of the Human Research Facility at NASA Ames Research Center, for monitoring and coordinating the 9 week study. We also thank Dr. Mark Graeber of the Veterans Administration Medical Center, Menlo Park Division, for his continued support. This research was funded, in part, by a grant from NASA Ames Research Center, project number 19926-12-02.

FIG. 4. Muscle strength differences in exercising (n = 27) versus nonexercising (n = 18) men. Add, adductor; Ext, extensor; D, dominant.

tween biceps strength and BMD is supported by the fact that the loads on axial bone by arm activity are high.('*) In our group of men, arm activity may have resulted in increasing loads to the skeleton and simultaneous improvement in back strength. It is possible that differences in precision of our various strength tests may have influenced the multiple-regression results. In repeatability tests, back strength exhibited the lowest measurement error at 1%; errors for hip, quadriceps, and biceps strength were 2, 3, and 6%, respectively. Similar measurement errors were reported in our study of women, however, in which biceps strength proved to be the most robust predictor of spine and hip density. There-

REFERENCES 1. Aloia JF, Cohn SH, Babu T, Abesamis C, Kalici N, Ellis K 1978 Skeletal mass and body composition in marathon runners. Metabolism 27:1793-1796. 2. Dalen N, Olsson KE 1974 Bone mineral content and physical activity. Acta Orthop Scand 45170-174. 3. Jones HH, Priest JD, Hayes WC, Tichenos CC, Nagel DA 1977 Humeral hypertrophy in response to exercise. J Bone

Joint Surg [Am] 59(2):204-208. 4. Huddleston AL, Rockwell D, Kulund DN, Harrison B 1980 Bone mass in lifetime tennis athletes. JAMA 244:1107-1109. 5. Doyle F, Brown J, LaChance C 1979 Relation between bone mass and muscle weight. Lancet 1391-393.

6. Snow-Harter C, Bouxsein M, Lewis B, Charette S, Weinstein P, Marcus R 1990 Muscle strength as a predictor of bone mineral density in young women. J Bone Miner Res 5589595.

12% 7. Bevier WC, Wiswell RA, Pyka G , Kozak KC, Newhall, KM, Marcus R 1989 Relationship of body composition, muscle strength and aerobic capacity to bone mineral density in older men and women. J Bone Miner Res 4:421-432. 8. Block JE, Friedlander AL, Brooks GA, Steiger P, Stibbs HA, Genant HK 1989 Determinants of bone density among athletes engaged in weight-bearing and non-weight-bearing activity. J Appl Physiol 67:1100-1105. 9. Pocock N, Eisman J, Gwinn T, Sambrook P , Kelly P , Freund J Yeates M 1989 Muscle strength, physical fitness and weight but not age predict femoral neck bone mass. J Bone Miner Res 4:441-448. 10. Sinaki M, McPhee MC, Hodgson SF 1986 Relationship between bone mineral density of the spine and strength of back extensors. Mayo Clin Proc 61:116-122. 11. Colletti LA, Edwards J , Gordon L, Shary J, Bell N 1989 The effects of muscle-building exercise on bone mineral density of the radius, spine, and hip in young men. Calcif Tissue Int 45:12-14. 12. Chamay A, Tschantz P 1972 Mechanical influences on bone remodelling. J Biomech 9173-180. 13. Lanyon LE 1984 Functional strain as a determinant for bone remodelling. Calcif Tissue Int (Supp1)36:556-561. 14. Rubin CT, Lanyon LE 1985 Regulation of bone mass by me-

SNOW-HARTER ET AL. chanical strain magnitude. Calcif Tissue Int 37:56-61. 15. Granhad H, Johnson R, Hansson T 1987 The loads on the lumbar spine during extreme weight lifting. Spine 12:146149. 16. U.S. Department of Health and Human Services 1988 The Surgeon General’s report on nutrition and health [DHHS (PHS) Publication No. 88-502101. Washington, DC: U.S. Government Printing Office. 17. Aniansson A, Zitterberg C, Hedberg M 1984 Impaired muscle function with aging. Clin Orthop 191:193-201. 18. Lindh M 1980 Biomechanics of the lumbar spine. In: Frankel VH, Nordin M (eds.) Basic Biomechanics of the Skeletal System. Lea & Febiger, Philadelphia, pp. 255-290.

Address reprint requests to: C. Snow-Harter Department of Exercise and Sport Science Oregon State University Langton Hall 214 Corvallis, OR 97331-3303 Received for publication November 1, 1991; in revised form March 23, 1992; accepted May 21, 1992.