Calcif Tissue Int (1997) 60:332–337

© 1997 Springer-Verlag New York Inc.

Correlation of Bone Density to Strength and Physical Activity in Young Men with a Low or Moderate Level of Physical Activity P. Nordstro¨m,1 G. Nordstro¨m,2 R. Lorentzon1 1 2

Sports Medicine Unit, Department of Orthopaedics, Umeå University, Sweden, S-901 85 Umeå, Sweden Departments of Geriatric Medicine and Prostethic Dentistry, Umeå University, Sweden, S-901 85 Umeå, Sweden

Received: 5 February 1996 / Accepted: 24 September 1996

Abstract. The objective of this study was to evaluate the relationship among bone mineral density (BMD), physical activity, muscle strength, and body constitution, in young men with a low or moderate level of physical exercise. Another aim was to investigate whether the head is unaffected by physical activity. The subjects consisted of 33 Caucasian healthy men, mean age 24.8 ± 2.3 years. BMDs of the total body, lumbar spine (L2-L4), femoral neck, Ward’s triangle and trochanter, humerus, and head were measured using dual-energy-X-ray absorptiometry (DXA). Bivariate correlations were measured among the different BMD sites and age, weight, height, body mass index (BMI), fat mass, lean body mass, amount of physical activity (hours/week), hamstrings strength, and quadriceps strength. Significant predictors were found for all BMD sites except the head. Using all these variables, only 6% of the variation in BMD of the head could be explained, whereas 46% (total body), 31% (humerus), 17% (lumbar spine), 38% (femoral neck, Ward’s), and 41% could be explained for the trochanter. Physical activity and muscle strength were found to be independent significant predictors of BMD of the total body and the sites at the proximal femur. These results suggest that at the time of peak bone mass attainment, physical activity is an important predictor of the clinically relevant proximal femur in young men with a low or moderate level of physical activity. Furthermore, since head BMD was not related to the level of physical activity, we suggest that head BMD may be used as an internal standard, to control for selection bias, in studies investigating the effect of physical activity on bone mass.

bone mass in important in determining the future risk for osteoporosis, studies have investigated the predictors of bone mass during adolescence [7, 8] and the potential determinants of peak bone mass [9]. Accordingly, in children, self-reported physical activity has been associated with BMD at several sites [7]. Hendersson et al. [10] demonstrated body weight and muscle strength to be independent significant predictors of bone mass in 18-year-old women. Welten et al. [9] suggested that body weight and regular weight-bearing exercise were of key importance in reaching the highest possible bone mass in the lumbar spine at 27 years of age. Studies have also demonstrated that the effects of weight-bearing exercise is site specific. In a recent study [11], we demonstrated a significantly higher bone mass of the tuberositas tibiae in adolescent boys subjected to a high amount of weight-bearing exercise and with a high quadriceps muscle strength, although there was no significant difference in bone mass of the head. Karlsson et al. [12] found significantly higher BMD in weight lifters of the proximal femur, lumbar spine, and total body, but not of the head. This may indicate that the head is not affected by weightbearing loading activities. In the present study, we used parameters well established to be related to bone mass, to find predictors of both weightbearing and nonweight-bearing BMD sites in young men at the time of peak bone mass attainment. Another purpose was to evaluate whether head BMD is affected by environmental factors such as physical activity, or could be used as an internal standard in studies investigating the effect of weight-bearing exercise on BMD.

Key words: Bone mineral density — Head — Physical activity — Muscle strength — Body constitution.

Material and Methods

Osteoporosis is a disease characterized by a low bone mass with a subsequent increased risk for fragility fractures [1]. The hip fracture is perhaps the most devastating manifestation resulting in an increased mortality and debility in both men and women [1]. The incidence of hip fractures has increased since the 1950s in both women and men, with a faster rate of increase in men [2, 3]. Since there is no effective cure in established osteoporosis there is growing emphasis for osteoporosis prevention. The bone mineral density (BMD) accumulates during childhood and adolescence to peak in the mid-20s [4, 5], or at some sites perhaps before 20 years of age [6]. Since peak

Subjects Thirty-three Caucasian young men, age 24.8 ± 2.3 (mean ± SD) were recruited by advertising. Most of the subjects were students (31/33). In clinical interviews, the subjects reported a physical exercise level of no more than 3 hours per week during the past year. Physical activity was defined as physical exercise combined with sweating or breathlessness. The average amount of physical activity was 1.5 ± 1 (mean ± SD) hours per week (range 0–3). Most of the time was spent playing floorball and volleyball, and doing distance running, cross-country skiing, and weight training. None of the subjects smoked, had any bone or muscle disease, or were using medication known to affect bone metabolism. The subjects gave their written consent to participate in the study which was approved by the Ethical Committee of the Medical Faculty, Umeå University. BMD Measurements

Correspondence to: R. Lorentzon

BMD of the total body, lumbar spine (L2–L4), left femoral neck

P. Nordstro¨m et al.: Correlation of Bone Density to Strength and Physical Activity

(neck), Ward’s triangle, and greater trochanter of the hip (trochanter), was measured using a dual energy X-ray absorptiometer, Lunar DPX-L, software version 1.3, (Lunar Co, Wisconsin, USA). The accuracy and precision of this method has previously been discussed in detail by others [13, 14]. BMD of the head, defined as all skeletal parts of the skull and the first four cervical vertebrae, was derived from the total body scan using the region of interest program. BMD of the left humerus was also derived in this manner. The coefficient of variation (CV) (mean/standard deviation) for a total body scan in our laboratory is 0.7% [15]. To evaluate the region of interest program one person was scanned (total body) 8 times during a short period of time, with repositioning between the scans. To maximize the precision, the centering option was used and scaling was set to about 200. Accordingly, the CV values were estimated to be 2.2% (head) and 2.5% (humerus). Fat mass and lean body mass were also derived from the total body scan, and the CV values were 0.9% (lean body mass) and 2.6% (fat mass) [15]. The CV values for the femur software and spine software were 0.8% (neck), 1.2% (Ward’s), 1.5% (trochanter), and 0.6% (lumbar spine). Isokinetic Muscle Strength Isokinetic muscle strength of the left quadriceps femoris and hamstrings muscles were measured in Newton-meters (Nm) using a Biodex isokinetic dynamometer (Biodex Co, New York, USA). The subject sat at a 120° hip angle with the lever attached just above the ankle. The dynamometer’s axis of rotation was aligned with the knee joint and the angular movement of the knee joint was 90°. Each subject made five maximal consecutive repetitions at 90°/second and 10 at 225°/second. The subjects were allowed to rest for 30 seconds between change of velocities. The highest peak torque for each velocity was used in the correlation analysis. Clinical Measurements Height and weight were measured in stockinged feet and underwear using standardized equipment. BMI was calculated (weight/ height2). Statistical Analysis Bivariate correlations between the different BMD sites and age, quadriceps strength, and hamstrings strength at 90 and 225°/ second; lean body mass, height, weight, BMI, fat mass, and physical activity were calculated using Pearson’s coefficient of correlation. A multivariate analysis was then conducted to estimate the variation explained at each BMD site by these variables and to find the independent predictors of each BMD site. It was hypothesized that a BMD site can be explained from the multiple regression model: BMD = c + a + bc + ms + ph + e where c 4 constant, a 4 age, bc 4 body constitution (i.e., weight, height, BMI, and fat mass), ms 4 muscle strength (i.e., quadriceps strength, hamstrings strength, and lean body mass), ph 4 physical activity, and e 4 error term. The error term consists of measurement errors, genetic effects on BMD, and the environmental factors not investigated in this study (e.g., calcium intake). Using this model, it was assumed that the error term was not correlated with any of the other independent variables. It was also assumed that the relationships between the independent variables and the dependent variable are both linear and additive. Since several of the body constitutional factors and muscle strength measurements were found to be highly intercorrelated (r > 0.8, P < 0.001), a principal components analysis [16] with rotation of the component loadings (Varimax) was then conducted on the explanatory variables. This procedure was conducted to avoid the consequences of multicollinearity, i.e., imprecise regression parameter estimates due to highly correlated explanatory variables.

333

Table 1. Anthropometric data, results of the muscle peak torque, and different BMD measurements

Age, years Height, cm Weight, kg BMI, kg/m2 Fat mass, kg Lean body mass, kg Strength, Nm Quadriceps 90°/second Quadriceps 225°/second Hamstrings 90°/second Hamstrings 225°/second Bone mineral density g/cm2 Total body Head Humerus Femoral neck Ward’s Trochanter Lumbar spine

Mean ± SD

Range

24.8 184 77.4 22.9 13.3 61.4

± ± ± ± ± ±

20.7–29.7 170–198 62.5–98.1 19.9–26.1 5.3–25.0 52.6–76.5

221 149 119 100

± 38 ± 27 ± 18 ± 17

1.23 ± 2.11 ± 1.17 ± 1.14 ± 1.08 ± 1.02 ± 1.26 ±

2.3 6 7.6 1.6 4.3 5.0

0.07 0.13 0.10 0.14 0.16 0.16 0.14

149–329 100–224 89–166 70–160 1.09–1.38 1.76–2.36 0.95–1.37 0.86–1.47 0.76–1.38 0.74–1.38 1.05–1.64

The principal components (PCs) formed from the original variables were used as ordinary explanatory variables in the multiple regression model to explain the variation of each BMD site. Principal components analysis is a statistical technique that linearly transforms an original set of variables into a substantially smaller set of uncorrelated variables. Principal components analysis searches for a few linear combinations of the original variables that capture most of the information (variance) of the original variables. Geometrically, the first PC is the line of closest fit to n observation in the multidimensional variable space. The second PC is the line of closest fit to the residuals from the first PC, and so on. A rotation of the PCs is sometimes done if the unrotated PCs are difficult to interpret. The most often used orthogonal rotation is Varimax [16]. The rotated component loadings are the PC’s correlation with the original variables. To combine a principal components analysis with a multiple regression model is statistically accepted [16]. In this study, each PC was characterized by its contents of rotated component loadings with scores as high and close to 1.0 as possible. The SPSS package (SPSS Inc., Chicago, USA) for Personal Computer was used for the statistical analysis. A P value less than 0.05 was considered significant. Results

The subjects age, physical characteristics, and results of muscle strength and BMD measurements are summarized in Table 1. Using bivariate correlations, the amount of physical activity (hours/week) was found to correlate with BMD of the total body (r 4 0.38, P 4 0.03) and femoral trochanter (r 4 0.38, P 4 0.03). Physical activity also showed a tendency to correlate with BMD of the femoral neck (r 4 0.34, P 4 0.05) and BMD of the Ward’s triangle (r 4 0.33, P 4 0.06) (Table 2). BMD of the head was not significantly correlated to physical activity. Bivariate correlations between BMD of the total body and different regional BMD sites and the explanatory variables age, weight, height, BMI, lean body mass, fat mass, hamstrings strength, and quadriceps strength at 90 and 225°/second were tested (Table 2). No variable was significantly correlated to BMD of the head, and only hamstrings strength at 90°/second was found to be significantly corre-

334

P. Nordstro¨m et al.: Correlation of Bone Density to Strength and Physical Activity

Table 2. Bivariate correlations between each BMD site and muscle strength, body constitution, physical activity (hours/week), and age BMD site Explanatory variables Quadriceps strength 90°/second Quadriceps strength 225°/second Hamstrings strength 90°/second Hamstrings strength 225°/second Lean body mass Weight Height Body mass index Fat mass Physical activity Age

Total body a

0.54 0.52a 0.53a 0.51a 0.52a 0.34 0.05 0.37b −0.06 0.38b 0.10

Head −0.20 −0.23 −0.04 −0.05 0.01 −0.13 −0.16 −0.03 −0.17 0.24 0.02

Humerus b

0.40 0.41b 0.40b 0.33 0.48a 0.24 −0.09 0.39b −0.15 0.27 0.04

Spine 0.30 0.26 0.38b 0.28 0.34 0.15 0.01 0.17 −0.19 0.22 0.05

Neck

Wards a

b

0.45 0.45a 0.48a 0.47a 0.48a 0.30 −0.01 0.33 −0.07 0.34 −0.09

0.40 0.38b 0.50a 0.52a 0.51a 0.27 0.02 0.25 −0.16 0.33 −0.21

Trochanter 0.50a 0.50a 0.52a 0.58a 0.49a 0.33 0.12 0.28 −0.02 0.38b 0.09

a

P value less than 0.01 P value less than 0.05

b

lated to lumbar spine BMD. To find the independent contributors and estimate the explained variation at each BMD site, all explanatory variables were then transformed into four PCs. Altogether, 85.7% of the variation explained by the original explanatory variables were explained by these 4 PCs. The dominant content of the first PC was quadriceps strength, hamstrings strength, lean body mass, and height, and this PC was interpreted as muscle strength. The dominant content of the second PC was weight, BMI, and fat mass, and this PC was interpreted as body constitution. The dominant content of the third and fourth PC was physical activity and age, respectively. The variation explained by each PC and the rotated component loadings for the dominant variables of each PC are shown in Table 3. These four PCs were then used in seven different multiple regression models to explain the variation of each BMD site (Fig. 1). Only 6% of the variation could be explained for head BMD (P > 0.05). The total explained variation (R2 values) of the other BMD sites was estimated to be 46% for total body (P < 0.01), 31% for humerus (P < 0.05), 17% for lumbar spine (P > 0.05), 38% for Ward’s triangle and femoral neck (P < 0.01), and 41% for trochanter (P < 0.01) (Fig. 1). Physical activity was found to be an independent, significant (P < 0.05) predictor of all BMD sites but the head and lumbar spine. Muscle strength, that is the first PC, was an independent predictor of total body BMD and the sites at the proximal femur, but not of the head, lumbar spine, and humerus (Fig. 1). All BMD sites measured but the head were highly significantly correlated with each other (r 4 0.68–0.91, P < 0.001); none of the BMD sites was significantly correlated with head BMD. However, when BMD of the total body and the regional BMD sites were adjusted for the best significant predictor of that site (Table 2), significant correlations (P < 0.05) appeared between head BMD and BMD of total body (Beta 4 0.44) and lumbar spine (Beta 4 0.41). Head BMD also showed a tendency towards a significant correlation with Ward’s (Beta 4 0.38, P 4 0.07) and the femoral neck (Beta 4 0.34, P 4 0.10). Table 4 shows the relationship between the weightbearing BMD sites and physical activity when adjusting for the influence of BMD of the head. Discussion

The aim of the present study was to evaluate possible pre-

Table 3. Results of the principal components analysis

Dominant content of principal component Quadriceps strength 90°/second Quadriceps strength 225°/second Hamstrings strength 90°/s Hamtrings strength 225°/s Lean body mass Height Weight Fat mass Body mass index Physical activity Age

Variance explained by each principal component (%)

Rotated component loadings of dominant variables

50.3

0.86 0.83 0.90 0.93 0.78 0.80 0.75 0.84 0.93 0.88 0.98

16.5 10.6 8.3

The dominant content of each principal component (PC), the variation explained by each of the PCs, and the rotated component loadings are shown. The PCs formed were then used in a multiple regression analysis to explain the variation of each BMD site

dictors of weight-bearing and nonweight-bearing BMD sites in young men with a low or moderate exercise level, and especially to determine whether bone mass of the head is related to the level of physical activity or can be used as an internal standard in studies investigating the effect of physical activity on bone mass. Age, body constitutional parameters, e.g., weight, BMI and fat mass, muscle strength and physical activity are wellknown predictors of bone mass [6, 10, 17–19]. Using all these parameters, only 6% of the variation for BMD of the head could be explained whereas 17% (lumbar spine) to 46% (total body) of the variation were explained for the other BMD sites. Physical activity or muscle strength were independent predictors of the BMD sites of the proximal femur, humerus, and total body, but not of the head. Using bivariate correlations, significant predictors were found for all BMD sites except the head. These results indicate that BMD of the head is not affected by environmental factors such as physical activity and could thus be used as an internal standard in studies investigating the effect of physical activity on bone mass. It should be realized that the varia-

P. Nordstro¨m et al.: Correlation of Bone Density to Strength and Physical Activity

335

Fig. 1. Results of the multiple regression analyses. The four ellipses represents the explanatory principal components (PCs). The dominant content of each PC is written inside the ellipse it refers to. The dependent variables, that is BMDs of the head, total body, humerus, lumbar spine, and the proximal femur sites, are described as boxes. Significant independent relationships are delineated as arrows from the explanatory PCs to the dependent variables. The regression coefficients for the significant correlations and the P values are expressed at each arrow. One multiple regression analysis was made for each dependent variable. R2 values for each multiple regression analysis are indicated at the top of each box of the corresponding dependent variable.

Table 4. The relationship between physical activity and the weight-bearing BMD sites when adjusting for the influence of head BMD Regression coefficients ± SEM Physical activity Total body BMD Lumbar spine BMD Femoral neck BMD Ward’s BMD Trochanter BMD

0.027 ± 0.013 (P 0.022 ± 0.026 (P 0.045 ± 0.026 (P 0.048 ± 0.029 (P 0.060 ± 0.028 (P

4 4 4 4 <

0.05) 0.39) 0.10) 0.11) 0.05)

tion in physical activity level was low to moderately high in the present study (0–3 hours), and it is possible that a greater range in physical activity also would have affected bone mass of the head. Previous studies concerning this matter are not conclusive. In a recent study, Karlsson et al. [12] demonstrated that male weight lifters had significantly higher BMD at all sites investigated (total body, lumbar spine, and proximal femur), except the head. In another study [20] a significantly lower BMD of the head (11%) was found in active female ballet dancers than in agematched controls. The female dancers also had a signifi-

cantly lower BMD of the arms (7%) and BMI (15%). It was suggested that the dancers start with a lower BMD, but because of weight-bearing exercise they increase BMD of the arms and legs to the same level as the controls. We [11] have recently compared adolescent ice hockey players training for about 10 hours/week with adolescent boys training at most 3 hours/week (mean 1.6 ± 1.1) concerning site-specific bone mass changes. The groups were matched according to age, pubertal stage, and weight. Although there was a 10% difference in BMD of the tuberositas tibiae from high physical loading and quadriceps muscle stress, there was no significant difference in head BMD (0%). Experimentally, Rawlinson et al. [21] demonstrated that limb but not calvarian organ cell cultures responded to applied dynamic mechanical strain. It was suggested that this difference in mechanical responsiveness may be due to an inherent property of these apparently similar cells. These results support our suggestion that head BMD is not affected by physical activity and muscle stress. However, the head is probably affected by factors such as body size, age, malnutrition, and hormone deficiency [20] in a similar fashion as other BMD sites. In this context it should be noted that Eisman et al. [22] previously have suggested that environmental factors interact to allow or prevent the genetic factors that influence the bone mass. The use of head BMD as an internal standard is depen-

336

P. Nordstro¨m et al.: Correlation of Bone Density to Strength and Physical Activity

dent upon the assumption that all BMD sites are determined by the same genetic factors. Previously, Pollitzer and Andersson [23] reviewed studies in which genetic contributions to bone mass were examined and found that most studies suggested that the same genetic factors determined both weight-bearing and nonweight-bearing bones. Pocock et al. [24] concluded from their twin study that one gene or a single set of genes determine bone mass at all skeletal sites. Additionally, in a recent study [15], quadriceps strength was found to be an independent predictor of BMD of the total body, head, humerus, spine, femur, and tibia/ fibula in adolescent boys (mean age 15.9 ± 0.3) with a low or moderate exercise level, indicating that all BMD sites are determined by the same genetic factors. In the present study, however, head BMD was not significantly related to any of the BMD sites investigated, whereas these other sites were highly intercorrelated. We hypothesized that this was due to the fact that all BMD sites except the head, in the present study, had been influenced by environmental factors such as physical activity for several more years than in the previous study [15]. This was supported by the fact that an adjustment of each BMD site for the best significant predictor of that site revealed significant correlations between head BMD and BMD of the total body and lumbar spine. The fracture of the proximal femur is probably the most devastating manifestation of osteoporosis. Women are most commonly affected by these fractures, but the incidence of hip fractures seem to increase predominantly in men [2, 3]. Maximizing peak bone mass has been suggested to be an important strategy in preventing osteoporosis later in life [22, 25]. Slemenda et al. [7] demonstrated physical activity to be an important predictor of bone mass in 118 children, aged 5–14 years, and suggested that a moderate increase in physical activity is related to important increments in peak bone mass. These results were adjusted for the fact that older children were more often engaged in different forms of physical exercise, but not for the possibility that generally stronger children may be more prone to participate in physical exercise. To exclude this possibility in the present study, it was assumed that BMD of the head is not affected by physical activity and therefore a result of bone mass heritability and those factors that may promote or prevent expression of the genetic factors that influence the bone mass [22]. Each BMD site was therefore also adjusted for the influence of head BMD before investigating the relationship with physical activity. This procedure changed the significant bivariate relationship between physical activity and BMD of the total body. However, physical activity was found to be an independent predictor of the BMD sites of the proximal femur, humerus, and total body when adjusting for the influence of age, muscle strength, and body constitution. Together these results clearly indicate that physical activity is an important determinant of the proximal femur at the time of peak bone mass attainment [5, 6] in men with a low or moderate level of physical activity. In the present study, we have only evaluated BMD of the whole head. This measure might be biased from loss of teeth, dental fillings, and parodontitis [26]. However, this will be a problem especially in an elderly population. Furthermore, the neck muscles may have some impact on bone mass of the lower part of the skull, although a recent experimental study [21] suggested that mechanical strain does not affect calvarian bone. Previous studies have evaluated both the whole head [11, 12] and the upper half of the head [26]. The upper half of the head is a smaller region containing a smaller amount of tissue measuring points, most probably leading to a greater measurement error.

Cross-sectional and longitudinal studies have demonstrated positive effects of physical activity on bone mass [17, 27]. Because stronger individuals may more often participate in sports activities, there might be a selection bias when investigating the effect of physical activity on bone mass in cross-sectional studies. In conclusion, the present and other studies [15, 23, 24] indicates that the same genetic factors determine bone mass of the whole skeleton. Our study has demonstrated that physical activity level and muscle strength cannot predict bone mass of the head. Considering these facts we suggest that head BMD can be used as an internal standard, to control for selection bias, when investigating the effect of physical activity on bone mass. The present study has also demonstrated that the level of physical activity is an important predictor of the clinically relevant proximal femur at the time of peak bone mass attainment in young men on a low or moderate exercise level.

Acknowledgments. We would like to thank Hans Stenlund, associate professor, Department of Statistics, Umeå University, for statistical advice. This study was supported by Grants (no. 100/94) from the Swedish Research for Sports. References 1. Consensus Development Conference (1993) Diagnosis, prophylaxis and treatment of osteoporosis. Am J Med 94:646– 650 2. Boyce WJ, Vessey MP (1985) Rising incidence of fracture of the proximal femur. Lancet 1:150–151 3. Obrant KJ, Bengner U, Johnell O, Nilsson BE, Sernbo I (1989) Increasing age-adjusted risk of fragility fractures: a sign of increasing osteoporosis in successive generations? Calcif Tissue Int 44:157–167 4. Teegarden D, Proulx WR, Martin BR, Zhao J, McCabe P, Lyle RM, Peacock M, Slemenda C, Johnston CC, Weaver CM (1995) Peak bone mass in young women. J Bone Miner Res 10:711–715 5. Hui SL, Johnston CC, Mazess RB (1985) Bone mass in normal children and young adults. Growth 49:34–43 6. Bonjour J-P, Theintz G, Buchs B, Slosman D, Rizzoli R (1991) Critical years and stages of puberty for spinal and femoral bone mass accumulation during adolescence. J Clin Endocrinol Metab 73:555–563 7. Slemenda CW, Miller JZ, Hui SL, Reister TK, Johnston CC (1991) Role of physical activity in development of skeletal mass in children. J Bone Miner Res 11:1227–1233 8. Ruiz JC, Mandel C, Garabedian M (1995) Influence of spontaneous calcium intake and physical exercise on vertebral and femoral bone mineral density of children and adolescents. J Bone Miner Res 10:675–682 9. Welten DC, Kemper HCG, Post GB, Van Mechelen W, Twisk J, Lips P, Teule GJ (1994) Weight-bearing activity during youth is a more important factor for peak bone mass than calcium intake. J Bone Miner Res 7:1089–1096 10. Hendersson KN, Price RI, Cole JH, Gutteridge DH, Bhagat CI (1995) Bone density in young women is associated with body weight and muscle strength but not dietary intakes. J Bone Miner Res 10:384–392 11. Nordstro¨m P, Nordstro¨m G, Thorsen K, Lorentzon R (1996) Local bone mineral density, muscle strength and exercise in adolescent boys—a comparative study of two groups with different muscle strength and exercise levels. Calcified Tissue Int 58:402–408 12. Karlsson M, Johnell O, Obrant K (1993) Bone mineral density in weight lifters. Calcif Tissue Int 52:212–215 13. Orwoll ES, Oviatt SK, Biddle JA (1993) Precision of dualenergy x-ray absorptiometry: development of quality control

P. Nordstro¨m et al.: Correlation of Bone Density to Strength and Physical Activity

14. 15.

16. 17. 18.

19.

rules and their application in longitudinal studies. J Bone Miner Res 8:693–699 Sieva¨nen H, Oja P, Vouri I (1992) Precision of dual-energy x-ray absorptiometry in determining bone mineral content of various skeletal sites. J Nucl Med 33:1137–1142 Nordstro¨m P, Thorsen K, Nordstro¨m G, Bergstro¨m E, Lorentzon R (1995) Bone mass, muscle strength, and different body constitutional parameters in adolescent boys with a low or moderate exercise level. Bone 17:351–356 Dunteman G (1989) Principal components analysis. Sage University Paper series on Quantitative Applications in the Social Sciences, pp 07–069 Gutin B, Kasper MJ (1992) Can vigorous exercise play a role in osteoporosis prevention? A review. Osteoporosis Int 2:55– 69 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 bone neck mass. J Bone Miner Res 4:441–448 Reid IR, Ames R, Evans MC, Sharpem S, Gamblem G, Francem JT, Lim T, Cundy TF (1992) Determinants of total body and regional bone mineral density in normal postmenopausal women—a key role for fat mass. J Clin Endocrinol Metab 75:45–51

337

20. Karlsson MK, Johnell O, Obrant KJ (1993) Bone mineral density in professional ballet dancers. Bone Miner 21:163– 169 21. Rawlinson SC, Mosley JR, Suswillo RF, Pitsillides AA, Lanyon LE (1995) Calvarial and limb bone cells in organ and monolayer culture do not show the same early responses to dynamic mechanical strain. J Bone Miner Res 10:1225–1232 22. Eisman JA, Kelly PJ, Morrison NA, Pocock NA, Yeoman R, Birmingham J, Sambrook PN (1993) Peak bone mass and osteoporosis prevention. Osteoporosis Int (suppl) 1:S56–60 23. Pollitzer WS, Andersson J (1989) Ethnic and genetic differences in bone mass: a review with a hereditary vs environmental perspective. Am J Clin Nutr 50:1244–1259 24. Pocock NA, Eisman JA, Hopper JL, Yeates MG, Sambrook PN, Eberl S (1987) Genetic determinants of bone mass in adults. A twin study. J Clin Invest 80:706–710 25. Hui SL, Slemenda CW, Johnston CC (1990) The contribution of bone loss to postmenopausal osteoporosis. Osteoporosis Int 1:30–34 26. Karlsson MK, Harrerius R, Nilsson JÅ, Obrant KJ (1995) Bone mass and density of the head. Eur J Exp Musculoskel Res 4:51–55 27. Souminen H (1993) Bone mineral density and long-term exercise. Sports Med 16(5):316–330