Calcified Tissue International

, Volume 91, Issue 2, pp 131–138

Muscle Strength and Body Composition Are Clinical Indicators of Osteoporosis

Authors

    • Bone and Cartilage Research Unit, Mediteknia BuildingUniversity of Eastern Finland
  • Joonas Sirola
    • Bone and Cartilage Research Unit, Mediteknia BuildingUniversity of Eastern Finland
    • Department of Orthopedics, Traumatology and Hand SurgeryKuopio University Hospital
  • Kari Salovaara
    • Bone and Cartilage Research Unit, Mediteknia BuildingUniversity of Eastern Finland
  • Marjo Tuppurainen
    • Bone and Cartilage Research Unit, Mediteknia BuildingUniversity of Eastern Finland
    • Department of Obstetrics and GynecologyKuopio University Hospital
  • Jukka S. Jurvelin
    • Department of PhysicsUniversity of Eastern Finland
  • Risto Honkanen
    • Bone and Cartilage Research Unit, Mediteknia BuildingUniversity of Eastern Finland
  • Heikki Kröger
    • Bone and Cartilage Research Unit, Mediteknia BuildingUniversity of Eastern Finland
    • Department of Orthopedics, Traumatology and Hand SurgeryKuopio University Hospital
Original Paper

DOI: 10.1007/s00223-012-9618-1

Cite this article as:
Rikkonen, T., Sirola, J., Salovaara, K. et al. Calcif Tissue Int (2012) 91: 131. doi:10.1007/s00223-012-9618-1

Abstract

We examined the role of muscle strength, lean tissue distribution, and overall body composition as indicators of osteoporosis (OP) in a pooled sample of 979 Finnish postmenopausal women (mean age 68.1 years) from the Kuopio Osteoporosis Risk Factor and Prevention study. Bone mineral density (BMD) at the femoral neck (FN) and total body composition were assessed by dual-energy X-ray absorptiometry scans. The women (n = 979) were divided into three groups according to WHO criteria, based on FN BMD T score: normal (n = 474), osteopenia (n = 468), and OP (n = 37). Soft tissue proportions, fat mass index (FMI, fat/height²), lean mass index (LMI, lean/height²), and appendicular skeletal muscle mass (ASM, (arms + legs)/height²) were calculated. Handgrip and knee extension strength measurements were made. OP subjects had significantly smaller LMI (p = 0.001), ASM (p = 0.001), grip strength (p < 0.0001), and knee extension strength (p < 0.05) but not FMI (p > 0.05) compared to other subjects. Grip and knee extension strength were 19 and 16 % weaker in OP women compared to others, respectively. The area under the receiver operating characteristic curve was 69 % for grip and 71 % for knee extension strength. In tissue proportions only LMI showed predictive power (63 %, p = 0.016). An overall linear association of LMI (R2 = 0.007, p = 0.01) and FMI (R2 = 0.028, p < 0.001) with FN BMD remained significant. In the multivariate model, after adjusting for age, grip strength, leg extension strength, FMI, LMI, number of medications, alcohol consumption, current smoking, dietary calcium intake, and hormone therapy, grip strength (adjusted OR = 0.899, 95 % CI 0.84–0.97, p < 0.01), leg extension strength (OR = 0.998, 95 % CI 0.99–1, p < 0.05), and years of hormone therapy (OR = 0.905, 95 % CI 0.82–1, p < 0.05) remained as significant determinants of OP. Muscle strength tests, especially grip strength, serve as an independent and useful tool for postmenopausal OP risk assessment. In addition, lean mass contributes to OP in this age group. Muscle strength and lean mass should be considered separately since both are independently associated with postmenopausal BMD.

Keywords

Body compositionBone mineral densityMuscle strengthPostmenopausal osteoporosis

Osteoporosis (OP) and frailty are major health problems in the developed countries. Low amount of lean mass (i.e., sarcopenia), poor physical performance, and low bone mineral density (BMD) are major risk factors for fractures [13]. Dual-energy X-ray absorptiometry (DXA) has been established as the standard test for its predictive value in the assessment of fracture risk, even though some methodological problems remain [46]. Additional parameters measured with DXA, such as body composition, have been suggested as risk factors for OP. However, the associations between frailty, sarcopenia, and OP are not established and need further evaluation. Simple muscle strength measurements have already been shown to indicate frailty and increased fracture risk [7]. This is of special interest in primary health-care centers with limited access to DXA. While they have been shown to be risk factors for OP and fractures, more data comparing strength, body composition, and BMD with standard thresholds needs to be evaluated.

Associations of weight, age, and gender with BMD are well known. Body weight is apparently mostly associated with postmenopausal BMD through weight bearing [8]. However, the constituents of body weight, lean and fat mass, also have their own metabolic and biomechanical associations with BMD and menopause [9, 10]. While contributing to body weight, adipose tissue also elevates circulating estrogen levels during menopause and beyond since hormone concentrations remain high longer in obese postmenopausal women [11]. On the other hand, lean mass is associated with BMD via biomechanical forces through muscle insertions since a larger cross-sectional area in muscle generally produces more powerful contractions. Together they contribute to body mass index (BMI) and can be used to evaluate general weight-bearing load for any individual. However, this approach does not represent proportional composition characteristics accurately at different ages since body composition varies with age, race, and gender [1214].

The aim of this study was to investigate how body composition and muscle strength are associated with postmenopausal OP in Finnish postmenopausal women. We determined the predictive power of simple strength measurements in this cohort and compared their usefulness with relative body composition functions as a tool for screening OP.

Methods

Design and Population

Subjects of this study present a pooled cohort of total-body DXA scans from the Kuopio Osteoporosis Risk Factor and Prevention Study (OSTPRE) densitometry sample and from the OSTPRE-Fracture Prevention Study (OSTPRE-FPS) study baseline cohort. OSTPRE is an ongoing prospective cohort study with 20 years of follow-up, while OSTPRE-FPS is a 3-year randomized subset of OSTPRE, aiming to evaluate the effect of calcium + vitamin D in the prevention of fractures and falls in elderly women. Protocols of both studies have been described previously [15, 16].

The initial target population of OSTPRE included all of the women born in 1932–1941 (n = 14,220) living in Kuopio Province, Finland, in 1989. A random stratified subsample (n = 3,222) of the study population underwent DXA densitometry at baseline and has been followed at 5-year intervals. In addition to the protocol, 506 women underwent a total-body scan from the sample that was invited to the 15-year follow-up densitometry between October 2004 and October 2007. Out of these, valid DXA data and functional measurements were obtained from 391 (77 %) women.

Between August 1 and November 31, 2002, a postal enquiry was sent to a random sample of 5,407 OSTPRE women (>65 years), to enquire about their willingness to participate in a 3-year fracture prevention study. In all, 4,189 women responded and 3,432 were willing to take part in the study. These women were randomized into two groups of equal size (n = 1,718 and 1,714). The supplementation group (n = 1,718) received a vitamin D + calcium regimen (cholecalciferol 800 IU + Ca 1,000 mg/day), while the control group (n = 1,714) did not receive any supplements. The trial was conducted between February 2003 and June 2007. At baseline a subsample of 750 women, 375 from both study groups, was randomly selected and invited to the Bone and Cartilage Research Unit, University of Eastern Finland, for clinical testing. Out of these, valid muscle strength, total-body, and femoral neck (FN) DXA measurements were obtained from 588 (78 %) women.

The pooled OSTPRE and OSTPRE-FPS sample included a total of 979 (319 + 588) women with complete data. Functional tests and questionnaires included leg extension strength, grip strength, quality of life, and food and physical activity diaries, which were acquired between 2003 and 2007. All measurements were performed with the same instruments and protocols throughout both studies. Study population characteristics and DXA and muscle strength measurements were carried out by trained personnel. Basic characteristics known to have an effect on BMD and strength, such as age, a number of chronic health disorders (0–9), use of hormone therapy (HT), number of prescribed medications, dietary calcium intake, physical exercise (sweating, shortness of breath), alcohol consumption, and smoking, were recorded with self-report questionnaires. Dietary calcium intake estimation was based on the question “How many deciliters of dairy products on average (milk, sour milk, yoghurt, whole milk, etc.) you consume daily?” Weight was measured with a calibrated digital scale (Philips, Eindhoven, the Netherlands; HF 351/00) and height with a calibrated, wall-mounted stadiometer. Trials were conducted in accordance with the Declaration of Helsinki, and the study plans of OSTPRE and OSTPRE-FPS have been accepted by the Ethics Committee of the University Hospital of Kuopio. All subjects gave written informed consent.

DXA Measurements

FN and total-body BMD measurements were performed by DXA using the GE Lunar Prodigy (General Electric Medical Systems, Milwaukee, WI). Phantom calibration was performed daily. Scans were done with the proposed default settings for different subjects. In the femur area, a scan was systematically taken from the left side unless an artifact (e.g., prosthesis) forced us to use the right side (n = 4). Appendicular lean mass (arms and legs) was calculated from the total-body DXA imaging software according to the standard anatomic landmarks and regions of interest provided by the manufacturer. According to a previous study, the coefficients of variation (CV) for the GE Lunar Prodigy at the FN and proximal femur are 1.27 and 0.94 %, respectively [17]. Women were classified as OP (T score ≤ –2.5 SD), osteopenic (OPN, T score >2.5 to <1 SD), or normal (T score ≥–1 SD) according to WHO criteria [7]. FN BMD cutoff values were <0.68 g/cm2 for OP and 0.68–0.86 g/cm2 for OPN. Normative values were obtained according to the manufacturer’s reference population of Finnish women at the time (version 110). Lean and fat tissue composition variables were calculated from DXA total-body scans as follows: BMI (kg/height²), lean mass index (LMI, lean mass [kg]/height²), fat mass index (FMI, fat mass [kg]/height²), and appendicular muscle mass (ASM, [arms + legs]/height²).

Muscle Strength Measurements

Strength measurements included grip and isometric knee extension strength measurements. Subjects were given verbal encouragement to maximize their effort in both tests. Grip strength was measured from the nondominant hand while sitting on a bench, with the forearm flexed from the elbow at a 90° angle, near the torso. A total of three attempts were recorded, with approximately 30 s of resting time between the tests. Close attention was paid to make all three attempts in a similar, fixed posture (JAMAR™ handgrip dynamometer; Sammons Preston, Bolingbrook, IL). The best attempt out of the three was recorded as the maximal result. The intraclass correlation coefficient for grip strength measurements was 0.93.

Isometric knee extension strength was measured three times from both legs, with a knee flexion of 65° (dynamometer chair; Metitur Oy, Jyväskylä, Finland). Participants extended the leg against the ankle strap with maximal effort, and peak force was recorded. Between each maximal attempt there was approximately 30 s of rest. The sitting posture was fixed straight with an adjustable backrest and tightened hip belt. The ankle strap was individually adjusted to meet the distal end of the lateral malleolus in the performing leg to minimize anthropometric bias. The average of the highest two out of the three attempts per leg was recorded as the maximum score. The results from both legs were then summed up and divided by 2, forming the overall quadriceps strength score used in the analysis. The intraclass correlation coefficients for the right and left knee extension strength measurements were 0.88 and 0.82, respectively.

Statistical Analyses

Data were analyzed with SPSS, version 19.0 (SPSS, Inc., Chicago, IL). Reported lifestyle risk factors with body soft tissue proportions and muscle strength were estimated for their ability to distinguish OP. Cross-tabulation with the Chi-squared test (Pearson’s χ2) was used to investigate differences between study groups and muscle strength tertiles. Analysis of variance (ANOVA) and Tukey’s honestly significant difference (HSD) test were used to compare characteristics between the groups. Multiple logistic regression was used to investigate the OP group’s characteristics in relation to the OPN group and normal counterparts. The Hosmer–Lemeshow test was used to test goodness of fit in the logistic regression model. The area under the receiver operating characteristic curve (AUROC) and the corresponding confidence intervals (CIs) were calculated with a regression model to estimate body composition and muscle strength test predictive values. The multivariate AUROC model estimated a variable discrimination power and ability to predict the outcome of interest, OP. For further assessment, determinants of OP were investigated in a multivariate regression model using body composition, muscle strength, age, and lifestyle variables (number of medications, alcohol consumption, smoking, dietary calcium, and HT) as covariates. Linear regression was used as a reference between body composition parameters and BMD, while a LOESS model (Epanechikov kernel) set a smoothing curve through a set of data points to investigate possible deviation for cutoff thresholds. p < 0.05 was considered statistically significant in all tests.

Results

Study participants were categorized into three groups according to WHO BMD criteria for FN OP: 3.8 % (n = 37) in the OP group, 47.8 % (n = 468) in the OPN group, and 48.4 % (n = 474) in the normal control group. The mean age (SD) of the study population was 68.1 (±2.4) years (range 62–75). BMI was 28.8 (±4.8, range 18–47). All women were postmenopausal. Years of HT use, age, height, weight, BMI, proportional lean mass, and muscle strength means significantly differed between groups (ANOVA). The study population characteristics are reported in Table 1.
Table 1

Characteristics of the study population (n = 979) by WHO groups

Characteristics

Mean (SD)

Osteoporosis

Osteopenia

Normal

n

979

37

468

474

Age (years)

68.1 (2.4)

69.1

68.4

67.8**

Height (cm)

158.7 (5.4)

155.8

158.3*

159.5***

Weight (kg)

72.6 (12.4)

65.8

70.9*

74.8***

Femoral neck BMD (g/cm²)

0.869 (0.12)

0.641

0.783***

0.971***

Total-body lean mass (kg)

40.2 (4.4)

36.8

40.0***

40.9***

Total-body fat mass (kg)

29.2 (8.9)

26.2

27.9

30.7**

Appendicular lean mass (kg)

16.6 (2.2)

15.0

16.4***

16.8***

Body mass index (kg/m²)

28.9 (4.8)

27.1

28.3

29.4*

Lean mass index (lean/m²)

16.0 (1.6)

15.1

15.9**

16.0**

Fat mass index (fat/m²)

11.6 (3.6)

10.8

11.1

12.1

Appendicular mass (lean/m²)

6.7 (0.7)

6.3

6.7**

6.8**

Max. knee extension (Nm)

648.1 (300.1)

545.2

623.0

680.0*

Max. grip strength (kg)

25.6 (5.4)

21.5

25.9***

26.7***

Number of medications

2.7 (2.1)

3.4

2.7

2.6

Hormone therapy (years)

4.0 (6.9)

1.0

2.7

5.5***

Dairy Ca intake (mg/day)

784 (320)

726

783

790

Currently smoking (%)

4.4

10.8

3.6

4.6

Alcohol intake (g/month)

89 (375)

22

102

81

Physical exercise (h/week)

0.8 (0.9)

0.53

0.76

0.88

Difference between groups (ANOVA (post hoc) vs. osteoporotic): * p < 0.05

** p < 0.01

*** p < 0.001

Maximum handgrip strength and leg extension strength were significantly weaker in the OP group compared with the other groups: grip strength was 17 % lower than in the OPN group and 19 % lower than in the normal group (p = 0.001) (Fig. 1). An identical trend was seen in leg extension strength, with a 12 % lower result in OP vs. OPN and 18 % lower versus normal (p < 0.001). After adjusting for relative tissue variables (LMI, FMI), muscle strength (grip, leg extension), HT, age, number of medications, smoking, alcohol use, and vigorous physical activity, only the effects of HT (p = 0.043), grip strength (p = 0.003), and leg extension (p = 0.028) persisted (Table 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00223-012-9618-1/MediaObjects/223_2012_9618_Fig1_HTML.gif
Fig. 1

Distribution of WHO groups in relation to grip strength tertiles

Table 2

Relationship between femoral neck osteoporosis, muscle strength, soft tissue composition, and lifestyle factors according to logistic regression

FN OP

Regression coefficient

OR

95 % CI

p

Lean index (lean/height²)

0.843

0.63–1.13

0.256

Fat index (fat/height²)

0.921

0.81–1.05

0.202

Grip strength (kg)

0.899

0.84–0.97

0.003

Leg extension strength (Nm)

0.998

0.99–1.00

0.028

Age (years)

1.112

0.96–1.29

0.163

No. of medications

1.096

0.92–1.30

0.299

Alcohol consumption (g/week)

0.995

0.99–1.00

0.197

Smoking (yes/no)

3.157

0.90–11.04

0.072

HT use (years)

0.905

0.82–1.00

0.043

Physical exercise (hr/week)

0.692

0.39–1.22

0.204

R2 = 19.9 %. Hosmer and Lemeshow test p = 0.951

Bold values indicate statistically significant (p < 0.05)

Pearson correlation coefficients for FN BMD with body composition (BMI r = 0.17, p < 0.001; lean mass r = 0.16, p < 0.001; fat mass r = 0.20, p < 0.001) and muscle strength (grip r = 0.14, p < 0.001; leg extension r = 0.08, p = 0.02) were significant. Fat mass alone explained 89 % and lean mass alone 53 % of body weight variation. Both variables together explained 99 %. Differences in LMI and FMI were found between the three groups. The deviation (SD) within FMI of 11.6 (±3.6) was significantly greater than that within LMI, 16.0 (±1.6). Relative lean mass was significantly lower in the OP group than in other groups. The OP group had lower LMI (p < 0.05) but not lower FMI (p = nonsignificant) compared to normal subjects. A difference was also found in the OPN group, with lower FMI (p > 0.05) but not LMI (p = nonsignificant) compared with the normal group. In cross-sectional analysis, the distribution of both LMI and FMI percentiles showed significant differences between the groups (p < 0.05 and p = 0.001, respectively). Distributions among the groups for LMI and FMI tertiles are shown in Figs. 2, 3.
https://static-content.springer.com/image/art%3A10.1007%2Fs00223-012-9618-1/MediaObjects/223_2012_9618_Fig2_HTML.gif
Fig. 2

Distribution of WHO groups in relation to lean mass tertiles

https://static-content.springer.com/image/art%3A10.1007%2Fs00223-012-9618-1/MediaObjects/223_2012_9618_Fig3_HTML.gif
Fig. 3

Distribution of WHO groups in relation to fat mass tertiles

With scatterplot smoothing (LOESS) curves LMI showed a steeper descending trend with BMD below −1 SD (0.744 g/cm2), while FMI showed a drop after OPN BMD values (0.860 g/cm2) and below. An overall linear association of LMI (R2 = 0.007, p = 0.01) and FMI (R2 = 0.028, p < 0.001) with FN BMD was significant, with modest correlation between the observed and predicted values (Fig. 4).
https://static-content.springer.com/image/art%3A10.1007%2Fs00223-012-9618-1/MediaObjects/223_2012_9618_Fig4_HTML.gif
Fig. 4

Change in lean mass (upper, green) and fat mass (lower, blue) according to femoral neck BMD. Vertical cutoff lines (T score), linear reference curve, and LOESS smoothing through the data (Color figure online)

The AUROC was used to evaluate the goodness of muscle strength and body composition in the detection of OP (Table 3). In a five-variable multivariate model, leg extension strength and grip strength were the strongest indicators of OP, with AUROC being 71 and 69 %, respectively. Appendicular (64 %) and total lean mass (63 %) proportions showed lower but statistically significant power, while the relationship with fat mass did not discriminate OP.
Table 3

Area-under-the-receiver-operating-characteristic curve test result variables for osteoporosis

 

Area (SE)

p

Appendicular lean index (lean/height²)

0.638 (0.53)

0.008*

Lean index (lean/height²)

0.625 (0.51)

0.016*

Fat index (fat/height²)

0.564 (0.55)

0.220

Grip strength (kg)

0.693 (0.50)

<0.001*

Leg extension strength (Nm)

0.709 (0.45)

<0.0001*

p < 0.05

Discussion

We performed muscle strength measurements as well as total-body and FN DXA on a total of 979 postmenopausal women, aged 63–75 years. The study analyzed the associations of lean mass, fat mass, and strength measurements with FN BMD. While associations between BMI and BMD have already been established, our further aim was to investigate accurate body composition and muscle strength interactions based on WHO criteria, with a clinical focus on females with OP. The study showed that women with OP had a lower BMI, with significantly lower lean, but not fat, mass proportions than others. However, after adjusting for several confounders, only muscle strength and HT were associated with OP. This suggests that the amount of muscle mass and the ability to produce sufficiently powerful muscle contractions may be stronger determinants of BMD than sheer body weight alone at this age. The finding also supports a plausible relationship between OP, sarcopenia, and frailty. However, even though an association between BMD and amount of lean mass seems evident, the criterion for sarcopenia is currently heterogeneous. Therefore, inclusion of a sarcopenia classification would not bring additional value to the analysis.

The results show a more dominant role of maximal muscle strength than relative body composition alone as an indicator for OP among postmenopausal women. It is noticeable that grip strength, without any biomechanical interaction with FN, gives a comparable result with leg extension strength. The results also suggest that the relationships of fat and lean mass to BMD are different and should be interpreted with caution. Muscle strength is highly related with age, gender, and mortality among different study populations [18, 19]. Also, proportional soft tissue composition accounting for BMI depends largely on the population measured, and associations differ between study groups. In postmenopausal populations, BMI becomes more dependent on the amount of body fat mass but can be altered to some degree by physical activity and proper nutrition [20, 21].

BMI after menopause is more strongly associated with fat mass than with lean mass, whereas in young adults body weight is more strongly associated with lean mass [22]. In addition to diet, these changes during perimenopause can be modified to some degree by HT and exercise [23, 24]. However, it is still unclear to what extent body composition changes can be explained by female hormonal deficiency and by decreases in physical activity with age. The absence of HT use and low muscle strength seem to be independent determinants of OP and contribute to fracture risk [25]. Most probably the changes depend on both [26, 27] and vary according to the origin of the study populations [2830]. Consequently, body composition parameters and muscle strength should be interpreted separately.

Lifestyle, including physical activity, affects strength and body composition even after menopause [31]. Together they are significantly associated with BMD and functional capacity, putting individuals with a sedentary lifestyle at greater risk of OP and fractures [32]. Both grip and leg extension strength measurements were significantly stronger predictors of BMD than body composition. This applies to the OP and OPN groups, with an even steeper negative trend in strength and lean mass compared to the normal category. This independent association of muscle strength and BMD suggests the usefulness of a simple strength measurement in risk assessments. Further, a simple grip strength test alone indicates the overall condition of muscle, assuming that the amount of vigorous physical activity and lean mass reduce with aging to some degree. While lean mass variation was fairly low in the cohort, most of the variation in BMI was explained by fat tissue. The amount of layering fat does not seem to change mean DXA BMD of the FN in humans significantly, but results may become less precise [33]. Our body composition results indicate the classic “frail and skinny” woman with OP. However, relative fat mass did not discriminate the diagnosis. While measuring axial BMD, the assessment of fat and lean mass would not add any clinical value to the decision for medical intervention regarding OP. However, the result highlights the fact that in addition to bone density, the WHO classification of OP differentiates muscle strength and body composition.

The association between muscle strength and BMD is in agreement with our previous findings [34, 35]. However, to our knowledge no study has compared the associations of strength and DXA body composition with OP using the WHO criteria on large samples of postmenopausal women, while controlling for multiple lifestyle and other confounding factors known to affect BMD. This is a population-based cohort with a good response rate. To minimize technical and psychological bias, we maintained the same level of motivation and performance control in strength measurements. The study personnel were trained not to produce any psychological bias by excessive verbal encouragement. Technical and postural variations were minimized as carefully as possible in DXA analysis, and quality assurance was assessed according to the manufacturer. One limitation of the DXA equipment, however, is that it cannot differentiate intramuscular fat, which might affect body composition results in this aging population. A second limitation of the study is the lack of follow-up data after determination of body composition, which might have revealed actual differences in the fracture incidence between groups. Also, muscle strength and fracture risk can be associated with vitamin D status, which was not available in this study and is regarded as a limitation. However, an association between fracture risk and low BMD is well established, and it is reasonable to assume that the relationship exists in this cohort as well to some degree.

The number of women with OP was relatively low. Although direct comparison of BMD (and thus prevalence of OP) is problematic between different populations or equipment due to the fact that different reference ranges give different T score values for the same BMD. However, we can make assumptions regarding our study by looking into OP prevalence with different thresholds. In terms of absolute bone density values, FN BMD within this study cohort was higher (0.869 g/cm2, SD 0.125) compared to previous data measured with Lunar equipment (0.817 g/cm2, SD 0.093) in a cross-sectional study of Finnish women (65–70 years) [36]. By WHO definition (T score), the prevalence of FN OP in Caucasian women of similar age has been reported to be between 8 % [37] and 18 % [38]. In this respect our cohort seems to be healthier, with a lower number (3.8 %) of OP patients. We used Lunar’s Finnish reference population for OP (≤0.68 g/cm2) [36]. Using the NHANES III threshold for OP (≤0.692 g/cm2), the prevalence would have increased from 3.8 % (n = 37) to 5.3 % (n = 52). Despite the reference used, the prevalence of OP in the study cohort remains low. We believe that our study sample represents the postmenopausal female population of eastern Finland. The whole OSTPRE study population is based on birth-year range (1932–1941, n = 14,222) and geographic location of Kuopio Province, including a stratified random sample of DXA follow-up (n = 3,222) for 20 years. The OP prevalence values of the whole OSTPRE study at the 15-year time point (n = 2,457), using Lunar national and NHANES III reference ranges, are 4.3 and 5.7 %, respectively. This is comparable to the prevalence of the subcohort used in this study. The original reference population for Lunar equipment was measured in Finland’s four cities (Helsinki, Kuopio, Oulu, Heinola), while OSTPRE represents a wide area with a relatively rural economy. Other known factors are relatively high BMI [39, 40], frequent use of HRT (45 %), and dietary calcium (784 mg/day), although diary calcium had no association with OP in this model and was excluded from the final analysis. A small number of cases with OP can limit the statistical power and generalization of the data. While the clinical approach is not as powerful as linear regression, it makes fewer assumptions about the normal distribution of the variables and remains focused on diagnosis.

We found that muscle strength tests reveal postmenopausal muscular capacity better than the amount of relative lean mass and that muscle strength was more strongly related to OP than any body composition parameter. This might be due to the fact that a large proportion of motor units are affected by aging prior to the loss of lean mass. Also, reduction in a number of motor units occurs together with the change toward slower type I muscle fibers during aging [37]. Overall, leaner but fit subjects can produce more strenuous muscle contractions and biomechanical forces, maintaining their BMD and thus partly compensating for their lower BMI with muscle strength. In contrast, low muscle power combined with high BMI is typically found in sarcopenic but obese individuals. The European Consensus for Sarcopenia has taken this into account by including both muscle strength and lean mass in the criteria [38]. This nonlinear relationship between strength and body composition is also seen in women with low BMD.

In conclusion, both body composition and muscle strength substantially contribute to BMD in postmenopausal women. While body composition alone does not reveal frailty, both should also be considered independently. Accordingly, muscular strength is a strong predictor of postmenopausal OP. Low muscle strength can be considered an independent risk factor and a useful tool for OP risk assessment. Thus, frailty seems to share a common factor in both sarcopenia and OP.

Acknowledgments

This study was supported by a grant from the Finnish Cultural Foundation, the Eemil Aaltonen Foundation, the Juho Vainio Foundation, the University of Eastern Finland, the Academy of Finland, and the Ministry of Education and Culture as well as an EVO grant from Kuopio University Hospital.

Copyright information

© Springer Science+Business Media, LLC 2012