Journal of Gerontology: MEDICAL SCIENCES 2006, Vol. 61A, No. 1, 92–96
Copyright 2006 by The Gerontological Society of America
Endogenous Hormones, Muscle Strength, and Risk of Fall-Related Fractures in Older Women Sarianna Sipila¨,1,2 Eino Heikkinen,1,2 Sulin Cheng,1 Harri Suominen,1 Pa¨ivi Saari,1 Vuokko Kovanen,1 Markku Ale´n,1 and Taina Rantanen1,2 1
Department of Health Sciences, and 2The Finnish Centre for Interdisciplinary Gerontology, University of Jyva¨skyla¨, Finland.
Background. Among older people, fracture-causing fall often leads to health deterioration. The role of endogenous hormone status and muscle strength on fall-related fracture risk is unclear. This study investigates if, after adjustment for bone density, endogenous hormones and muscle strength would predict fall-related limb fracture incidence in older community-dwelling women followed-up over 10 years. Methods. As a part of a prospective population-based study, 187 75-year-old women were investigated. Serum estradiol, testosterone, sex hormone binding globulin, and dehydroepiandrosterone sulfate concentrations were analyzed, and isometric muscle strength and bone mineral density were assessed. Fall-related limb fractures were gathered from patient records. Results. Serum estradiol concentration was a significant predictor of fall-related limb fractures. Women with serum estradiol concentrations less than 0.022 nmol/L had a 3-fold risk (relative risk 3.05; 95% confidence interval, 1.26–7.36), and women with estradiol concentrations between 0.022 and 0.066 nmol/L doubled the risk (relative risk 2.24; 95% confidence interval, 0.97–5.19) of fall-related limb fracture compared to the women with estradiol concentrations above 0.066 nmol/L. Adjustment for muscle strength and bone mineral density did not materially change the risk estimates. High muscle strength was associated with a low incidence of fall-related limb fractures. Conclusions. This study showed that in 75-year-old women higher serum estradiol concentration and greater muscle strength were independently associated with a low incidence of fall-related limb fractures even after adjustment for bone density. Our results suggest that hormonal status and muscle strength have their own separate mechanisms protecting from fall-related fractures. This finding is of importance in developing preventive strategies, but calls for further study.
A
MONG older people, falls resulting in an injury often lead to health deterioration. An earlier study showed an almost 20-fold risk of institutionalization among older people who had sustained an injurious fall compared to the no-falls group (1). In particular, among older people with low bone density, a typical consequence of an injurious fall is a peripheral bone fracture (2,3). Another major contributory factor to fracture risk is muscle strength, which has been found to independently predict fall incidence (4–6). Evidence has been found to show that muscle and bone loss share a common etiology. Changes in the endogenous hormone milieu lead to muscle and bone tissue atrophy (7). Menopause-related coincident decreases in bone density, muscle mass, and muscle strength may, at least to some extent, be reversed through hormone replacement therapy (HRT) (8–10). The most prominent hormonal change following menopause is rapid decline in ovarian estradiol (E2) production (11). Postmenopausal women also have slightly lower serum testosterone concentration compared to their premenopausal level (11). In older women, the circulating E2 and testosterone are mainly produced by peripheral conversion of adrenal steroids such as dehydroepiandrosterone sulfate (DHEA-S; 12). After age 65, DHEA-S concentration shows an average annual decline of 4% (13). Whereas serum E2, T, and DHEA-S concentrations decline with age, the concentration of sex hormone binding globulin (SHBG) increases, resulting in lower levels of free sex hormone concentrations (14).
92
The role of endogenous hormone status on risk of fractures remains unclear. A case–control study by Cummings and colleagues (15) showed that postmenopausal women with undetectable serum E2 (,0.018 nmol/L) and a high SHBG concentration had an increased risk of osteoporotic fracture, whereas two other studies (16,17) failed to show a significant independent association between sex hormones and fracture risk. Endogenous hormonal status may have an effect on the incidence of fall-related fractures through its effect on muscle strength and bone density. However, no earlier research has prospectively studied the association between hormonal milieu, muscle strength, and fall-related fracture incidence in a representative sample of older women. This study was undertaken to investigate if, after adjustment for bone density, endogenous hormones and muscle strength would predict fall-related limb fracture incidence in older community-dwelling women followed-up over 10 years. METHODS
Participants Data for this report were collected as part of the Evergreen Project, a prospective population-based study on the health and functional capacity of older people. A detailed description of the baseline cohort has been published earlier (18). Briefly, all 75-year-old community-dwelling women in the city of Jyva¨skyla¨, Finland in autumn 1989 composed the
HORMONES AND RISK OF FRACTURE IN WOMEN
Table 1. Body Composition in 75-Year-Old Women According to Hormonal Status Body Mass, kg
LBM, kg
Fat, %
BMD, gcm3
SHBG, nmol/L Low: ,42.6 Middle High: .91.5
75 6 13 67 6 10 61 6 10*
47 6 4 45 6 4 43 6 4*
36 6 7 33 6 7 30 6 8y
0.150 6 0.036 0.125 6 0.035 0.118 6 0.033*
E2, nmol/L Low: ,0.022 Middle High: .0.066
66 6 10 68 6 12 68 6 14
44 6 4 45 6 4 45 6 5
32 6 7 34 6 7 33 6 7
0.127 6 0.030 0.128 6 0.040 0.132 6 0.036
E2/SHBG Low: ,0.0003 Middle High: .0.0011
66 6 10 66 6 11 72 6 14z
44 6 4 44 6 4 46 6 4z
32 6 7 32 6 7 35 6 7
0.127 6 0.032 0.124 6 0.037 0.141 6 0.038
T, nmol/L Low: ,0.867 Middle High: .2.045
65 6 9 67 6 13 72 6 11z
44 6 4 45 6 5 46 6 4
32 6 6 33 6 8 36 6 6z
0.125 6 0.032 0.124 6 0.037 0.143 6 0.035z
T/SHBG Low: ,0.0110 Middle High: .0.0398
62 6 10 68 6 13 74 6 10*
43 6 4 45 6 4 46 6 4y
30 6 8 33 6 7 37 6 6y
0.115 6 0.033 0.128 6 0.035 0.146 6 0.036y
DHEA-S, lmol/L Low: ,0.915 Middle High: .1.875
68 6 10 68 6 13 65 6 10
45 6 4 45 6 5 44 6 4
34 6 7 33 6 7 32 6 6
0.121 6 0.032 0.131 6 0.035 0.122 6 0.030
Variable, Cut Point
Notes: Low n ¼ 38–44, Middle n ¼ 77–90, High n ¼ 38–44. p for trend: *p , .001, yp .01, zp .05. LBM ¼ lean body mass; BMD ¼ bone mineral density; SHBG ¼ sex hormone binding globulin; E2 ¼ estradiol; T ¼ testosterone; DHEA-S ¼ dehydroepiandrosterone sulfate.
target population (n ¼ 261). Four of the women died before the baseline measurements, and 66 women refused or were unable to participate in the study. Due to an aneurysm, one woman was unable to participate in the strength tests. Hormones were not analyzed from three of the participants due to the small blood sample volume. Consequently, data from 187 women were used for this analysis. The study was approved by the Ethics Committee of the University of Jyva¨skyla¨, and all participants gave their written informed consent prior to entering the study.
Health Assessment A multiphase health assessment comprised a face-to-face interview carried out in the participants’ homes, a health questionnaire, and physician’s examination. During the examination, the presence of chronic conditions and use of medication were established according to the questionnaire, current prescriptions, and medical records. Two of the participants were taking postmenopausal HRT and had serum E2 concentrations of 0.064 nmol/L and 0.031 nmol/L. These values did not differ from those obtained from the others in the study. Consequently, the data of these individuals were included in the analysis. Physical Activity According to the scale by Grimby and colleagues (19), the participants were classified as sedentary, moderately active, or active. Those reporting no other activity but light walking were rated as sedentary. Those reporting moderate physical activities at least three times a week, but no exercise more
93
intensive than that, were rated as moderately active. If a participant reported moderate or vigorous exercise at least three times per week she was rated as active.
Hormone Analysis From all participants, venous blood samples were drawn between 8:15 and 11 AM. Serum was separated for the E2, T, and SHBG analyses, while EDTA plasma was taken for the DHEA-S analyses. Both serum and plasma were stored at 808C until analysis. E2, T, and SHBG were analyzed by time-resolve fluoroimmunoassays (DELFIA; Wallac Oy, Turku, Finland). DHEA-S concentrations were measured from plasma samples by enzyme-linked immunosorbent assay (ELISA) kit (IBL [Immuno-Biological Laboratories], Hamburg, Germany). Serum E2, T, SHBG, and DHEA-S concentrations were available from 156, 153, 157, and 178 women, respectively. Fifteen of the women had serum E2 concentration below detection level (,0.001 nmol/L). Because SHBG determines the proportion of the biologically active sex steroids in circulation, E2/SHBG and T/SHBG ratios were calculated. Interassay coefficients of variation (CVs) ranged between 8.0% and 17.5% for E2, 3.4% and 7.9% for T, 1.4% and 2.9% for SHBG, and 5.7% and 10.7% for DHEA-S. Table 1 describes the study population according to their endogenous hormone concentration levels. Muscle Strength Maximal isometric muscle strength for elbow flexion and knee extension was measured on the side of the dominant hand of the participant in a sitting position using an adjustable dynamometer chair (20). Elbow flexion strength was measured at an elbow angle of 908, and knee extension strength at a knee angle of 608 from full extension. The participants were allowed two to three practice trials followed by three maximal efforts separated by a 1-minute rest. The highest value was taken for the analysis. The CV between two consecutive measurements is 4%–6% (21). Bone Mineral Density Bone mineral density (BMD) measurement was performed from the right calcaneus bone by a single-energy photon absorption method (SPA) as described earlier (22). Our previous study showed that calcaneus BMD predicts fracture risk in older populations. The CV between repeated measures is less than 2%. Anthropometry Body mass and body height were measured using standard procedures. Lean body mass and total body fat percentage were assessed using bioelectrical impedance (Spectrum II; RJL Systems, Detroit; MI). The CV between two consecutive measurements was less than 2% for lean body mass and less than 3% for body fat percentage (21). Surveillance of Fall-Related Fractures Injurious accidents were followed-up for 10 years. Selfreports concerning accidents were collected by interviews and questionnaires in 1991, 1993, and 1995. Patient records held by the area health care centers and the central hospital were reviewed in 1995 and 2000 to validate self-reports and to collect new data. All fractures were diagnosed by using radiographs. Injuries were classified according to the
SIPILA¨ ET AL.
94
Table 2. Physical Activity Level, Smoking History, and Chronic Conditions in the 75-Year-Old Women With and Without Fall-Related Limb Fracture Over 10-Year Follow-Up No Fractures
Fracture
Relative Risk (95% CI) Variable
Variable
N
%
N
%
Physical activity level Sedentary Moderate Active
31 59 34
25 48 27
11 27 22
18 45 37
Current smoking Daily Occasionally No
4 2 117
3 2 95
3 0 57
5 0 95
Smoking, y Never smoked 1–10 11–30 31–60
109 4 3 5
90 3 3 4
51 1 2 2
90 2 4 4
58 38 9 14 4 44 1 8 10 2
46 30 7 11 3 35 1 6 8 2
26 16 7 7 3 20 1 6 2 0
43 26 11 11 5 33 2 10 3 0
Chronic conditions Atherosclerotic diseases Hypertension COPD Diabetes Hypertyreosis Osteoarthritis Osteoporosis Cancer Depression Parkinsonism
Table 3. Predictors of Fall-Related Limb Fractures in 75-Year-Old Women Over 10-Year Follow-Up
Unadjusted
Adjusted*
E2 referent, .0.066 nmol/L ,0.022 nmol/L 0.022–0.066 nmol/L
3.05 (1.26–7.36) 2.24 (0.97–5.19)
2.53 (1.03–6.19) 2.00 (0.86–4.63)
SHBG, referent .91.5 nmol/L ,42.6 nmol/L 42.6–91.5 nmol/L
0.79 (0.35–1.81) 1.10 (0.57–2.13)
1.65 (0.67–4.05) 1.34 (0.66–2.69)
E2/SHBG, referent .0.0011 ,0.0003 0.0003–0.0011
2.08 (0.88–4.92) 1.60 (0.73–3.54)
1.46 (0.61–3.52) 1.06 (0.47–2.40)
T, referent .2.045 nmol/L ,0.867 nmol/L 0.867–2.045 nmol/L
1.22 (0.55–2.73) 1.04 (0.51–2.12)
1.17 (0.50–2.76) 0.91 (0.41–2.00)
T/SHBG, referent .0.0398 ,0.0110 0.0110–0.0398
1.36 (0.62–3.00) 1.16 (0.57–2.37)
1.01 (0.42–2.40) 1.02 (0.48–2.19)
DHEAS, referent .1.875 nmol/L ,0.915 nmol/L 0.86 (0.43–1.75) 0.915–1.875 nmol/L 0.68 (0.37–1.25)
0.93 (0.46–1.91) 0.82 (0.44–1.52)
Muscle strength, PC score
0.71 (0.52–0.97)y
0.71 (0.56–0.90)
Notes: CI ¼ confidence interval; SHBG ¼ sex hormone binding globulin; E2 ¼ estradiol; T ¼ testosterone; DHEA-S ¼ dehydroepiandrosterone sulfate; PC ¼ principal component. *Adjusted for muscle strength, body height, and bone mineral density. y Adjusted for E2, body height, and bone mineral density.
Note: COPD ¼ chronic obstructive pulmonary disease.
International Classification of Diseases, 10th Revision (ICD-10). These data were used to identify those participants who had suffered a fall-related limb fracture (ICD-10 categories S42, S52, S62, S72, S82, and S92)
Data Analysis Pearson correlation coefficients were calculated to investigate associations between serum hormone concentrations. Knee extension and elbow flexion results were condensed into one score using principal component analysis. On the basis of the distribution of each hormone measure, the participants were divided into four equal groups. Two midquartiles were then combined into one middle group for more detailed examination in the extreme ends of distributions. The mean differences in body composition between the low, middle, and high serum hormone concentration groups were analyzed by one-way analysis of variance. The mean difference in muscle strength principal component between the high, middle, and low hormone group was analyzed using univariate analysis of variance with body height as a covariate. Relative risk of fall-related limb fracture according to baseline muscle strength and serum hormone concentrations was computed using Cox regression analysis using the number of days from the baseline to the first limb fracture-sustaining fall as an outcome variable. Data on the participants without fracture-sustaining fall were censored at the time of death or at the end of the follow-up. RESULTS Of the 187 women, 65% suffered at least one injurious fall during the 10-year follow-up. Of these injuries, 55% were
fractures of which 91% were in the upper or lower limbs. The fracture sites were femur (n ¼ 22), lower leg or foot (n ¼ 11), humerus (n ¼ 5), forearm (n ¼ 16), and wrist or hand (n ¼ 7). The average time to event was 1767 (standard deviation 1105) days from baseline. The baseline characteristic of women who suffered a fall-related limb fracture versus those women who did not suffer a fracture are presented in Table 2. Serum E2 concentration was a significant predictor of fallrelated limb fractures (Table 3). Women with E2 serum concentrations below 0.022 nmol/L (low group) at the age of 75 had a 3-fold risk of fall-related limb fracture compared to the women with E2 concentrations above 0.066 nmol/L (referent). The risk of fall-related limb fracture in women with E2 concentrations between 0.022 and 0.066 nmol/L was double the risk of the high concentration group. After adjustment for muscle strength, body height, and BMD, women with E2 serum concentrations below 0.022 nmol/L still had a nearly 3-fold risk of fall-related limb fractures compared to the high E2 concentration group. The other serum hormone concentrations analyzed in this study did not predict fracture incidence. However, low SHBG and high DHEA-S concentrations were associated with high serum E2 concentration (r ¼0.161, p ¼ .046 and r ¼ 0.189, p ¼ .022, respectively). High muscle strength was associated with low incidence of fall-related limb fractures (Table 3). Adjustment for serum E2 concentration, body height, and BMD did not materially change risk estimates. As shown in Figure 1, muscle strength was associated with endogenous hormone concentrations. Women in the lowest SHBG and highest E2 and E2/SHBG ratio groups were significantly stronger than those in the highest SHBG and lowest E2 and E2/SHBG ratio groups.
HORMONES AND RISK OF FRACTURE IN WOMEN
Figure 1. Isometric muscle strength (principal component including knee extension and elbow flexion strength) adjusted for body height in 75-year-old women in different endogenous hormone concentration groups. White bars indicate low, light gray middle, and dark gray high values.
DISCUSSION This study showed that in 75-year-old women, higher endogenous serum E2 concentration and greater muscle strength were independently associated with a low incidence of fallrelated limb fractures even after adjustment for BMD. Our results suggest that hormonal status and muscle strength have their own separate mechanisms affording protection from a fall-related limb fracture. To the best of our knowledge, no previous studies have reported on the association between endogenous hormone concentration and fall-related fracture incidence in a population-based sample of older women. The independent association between E2 concentration and fall-related fracture risk may be partly explained by the widespread and diverse effects of estrogens. In addition to bone and muscle tissue, estrogens target the central nervous system and affect performance in spatial tasks (23) and motor behavior such as coordination and reaction time (23,24). Therefore, it is possible that the mechanism behind the association between E2 and fall-related fracture risk lies in the central nervous system through which women with higher E2 concentrations are better protected from fracture-sustaining falls. The observed association between endogenous serum E2 concentration and fall-related fracture risk may be modified by HRT. The case–control study by Michae¨lsson and colleagues (25) showed that HRT in women older than 50 years protects from hip fracture and that every year of therapy decreases the overall risk by 6%. The debate on the consequences of HRT on women’s health, however, continues, and both favorable and unfavorable long-term effects have been suggested. This study also showed that women with greater muscle strength were less likely to sustain a fall-related limb fracture compared to the women with poor muscle strength. This result is in accordance with the study by Nguyen and colleagues (26) which showed that muscle strength is an independent predictor of fracture incidence. Similarly, the case–control study by Luukinen and colleagues (5) suggested that home-dwelling
95
older adults with poor knee extension strength have increased risk of fracture-sustaining fall. One potential explanation is that women with higher muscle strength have a greater amount of soft tissue (27), which absorbs energy during a fall. Larger muscle mass, in older people, is also beneficial for several health-related conditions due to its protein-rich and metabolically active nature. Moreover, older people with higher muscle strength have a lower risk of mobility limitation (28) thus decreasing the probability of a fall. The significance of the association observed between muscle strength and fall-related fracture risk is of great importance for designing preventive strategies. Although strength declines with increasing age (27), progressive resistance training may delay some of the consequences of strength impairment. A number of earlier studies have shown that 2–4 months of progressive resistance training improved mobility (29,30) and balance (30,31), especially in frail older people, potentially also decreasing the risk of injurious falls. Experimental trials are, however, needed to ascertain the effects of resistance training on fall-related fracture risk. This study showed that high E2 and low SHBG concentrations are associated with muscle strength, suggesting that the endogenous hormonal milieu is a significant factor in muscle performance in older women. This result is supported by earlier studies, although conflicting results have also been reported. Until the year 2003, the pathway for E2 action on skeletal muscle tissue in humans was unclear. Quite recently, Lemoine and colleagues (32) detected the messenger RNA transcript of the nuclear estrogen receptor alpha in female skeletal muscle. This finding suggests that skeletal muscle cells are targets for estrogens and that the effects of the hormone are at least partly mediated through direct nuclear hormone-receptor interaction. In this study, low serum SHBG concentration was associated with high muscle strength but not with fallrelated fracture risk. Blood DHEA-S concentration was associated with neither muscle strength nor fall-related fracture risk. It is, however, possible that the interrelationship between SHBG, DHEA-S, and E2 contributes to fracture risk. Low SHBG concentration is beneficial for E2 action, because SHBG binds sex steroids and makes them biologically inactive. DHEA-S, in contrast, is converted into more active sex hormones in peripheral tissues; therefore, circulating E2 concentration is affected by DHEA-S. The major strength of this study is a true prospective design including over 60% of all 75-year-old women living in the same urban location. All the participants measured at baseline were included in the data analysis. Fall-related fracture surveillance was carried out over 10 years, and data were collected from the patient records. The association between muscle strength, E2 concentration, and fall-related fracture risk was observed in older community-dwelling women who were able to participate in the laboratory measurements. The association would most likely have been stronger had more disabled and weaker women been included in the analysis. A potential limitation of the study is the use of SPA for BMD measurements of the calcaneus bone. SPA was an appropriate technique in the 1980s but is currently considered outdated for bone assessments. However, calcaneus
96
SIPILA¨ ET AL.
BMD is known to predict fracture risk in older populations (21), and it correlates with dual-energy x-ray absorptiometry (DXA) (33). Nevertheless, some residual confounding of BMD may have remained in the models. Fall-related fractures are a major public health burden that significantly contributes to the increasing number of older people with disability. This prospective study shows that, in older community-dwelling women, higher muscle strength and serum E2 concentrations greater than 0.066 nmol/L were associated with low incidence of fall-related limb fractures independent of BMD. This finding is of great importance in developing sustainable preventive and therapeutic strategies, but calls for further studies. ACKNOWLEDGMENTS The Evergreen Project was supported by the Academy of Finland, Social Insurance Institution, Ministry of Education, Ministry of Social Affairs and Health, City of Jyva¨skyla¨, and The Association of Finnish Lions Clubs. E. Heikkinen is the project leader responsible for the design and implementation of the study, organization of funding, and finalization of the manuscript. He initiated this study, while S. Sipila¨ formulated the study question, analyzed the data, and wrote the manuscript. S. Cheng was involved in the hormone and bone assessments. H. Suominen was involved in the bone, muscle strength, and anthropometry measurements and contributed to the finalization of the manuscript. P. Saari contributed to the fracture surveillance. V. Kovanen contributed to hormone analysis, and M. Ale´n to health and clinical assessments, hormone analysis, and finalization of the manuscript. T. Rantanen was involved in the muscle strength measurements, and contributed to the data analysis and finalization of the manuscript. Address correspondence to Sarianna Sipila¨, PhD, Department of Health Sciences, University of Jyva¨skyla¨, P.O. Box 35 (Viv), FIN-40014 Jyva¨skyla¨, Finland. E-mail:
[email protected] REFERENCES 1. Tinetti ME, Williams CS. Falls, injuries due to falls, and the risk of admission to a nursing home. N Engl J Med. 1997;337:1279–1284. 2. Cauley JA, Robbins J, Chen Z, et al. Effects of estrogen plus progestin on risk of fracture and bone mineral density: the Women’s Health Initiative randomized trial. JAMA. 2003;290:1729–1738. 3. Cummings SR, Nevitt MC, Browner WS, et al. Risk factors for hip fracture in white women. Study of Osteoporotic Fractures Research Group. N Engl J Med. 1995;332:767–773. 4. Campbell AJ, Robertson MC, Gardner MM, Norton RN, Tilyard MW, Buchner DM. Randomised controlled trial of a general practice programme of home based exercise to prevent falls in elderly women. BMJ. 1997;315:1065–1069. 5. Luukinen H, Koski K, Laippala P, Kivela SL. Factors predicting fractures during falling impacts among home-dwelling older adults. J Am Geriatr Soc. 1997;45:1302–1309. 6. Stel VS, Smit JH, Pluijm SM, Lips P. Balance and mobility performance as treatable risk factors for recurrent falling in older persons. J Clin Epidemiol. 2003;56:659–668. 7. Proctor DN, Balagopal P, Nair KS. Age-related sarcopenia in humans is associated with reduced synthetic rates of specific muscle proteins. J Nutr. 1998;128:351S–355S. 8. Cauley JA, Petrini AM, LaPorte RE, et al. The decline of grip strength in the menopause: relationship to physical activity, estrogen use and anthropometric factors. J Chronic Dis. 1987;40:115–120. 9. Cheng S, Sipila S, Taaffe DR, Puolakka J, Suominen H. Change in bone mass distribution induced by hormone replacement therapy and high-impact physical exercise in post-menopausal women. Bone. 2002;31:126–135. 10. Sipila¨ S, Taaffe D, Cheng S, Puolakka J, Toivanen J, Suominen H. Effects of hormone replacement therapy and high-impact physical exercise on skeletal muscle in post-menopausal women: a randomized placebo-controlled study. Clin Sci. 2001;101:147–157. 11. Rannevik G, Carlstro¨m K, Jeppsson S, Bjerre B, Svanberg L. A prospective long-term study in women from pre-menopause to postmenopause: changing profiles of gonadotrophins, oestrogens and androgens. Maturitas. 1986;8:197–307.
12. Couzinet B, Meduri G, Lecce M, et al. The postmenopausal ovary is not a major androgen-producing gland. J Clin Endocrinol Metab. 2001;86:5060–5066. 13. Mazat L, Lafont S, Berr C, et al. Prospective measurements of dehydroepiandrosterone sulfate in a cohort of elderly subjects: relationship to gender, subjective health, smoking habits, and 10-year mortality. Proc Natl Acad Sci U S A. 2001;98:8145–8150. 14. Newcomb PA, Klein R, Klein BE, et al. Association of dietary and lifestyle factors with sex hormones in postmenopausal women. Epidemiology. 1995;6:318–321. 15. Cummings SR, Browner WS, Bauer D, et al. Endogenous hormones and the risk of hip and vertebral fractures among older women. N Engl J Med. 1998;339:733–738. 16. Barrett-Connor E, Mueller JE, von Mu¨hlen DG, Laughlin GA, Schneider DL, Sartoris DJ. Low levels of estradiol are associated with vertebral fractures in older men, but not women: Rancho Bernardo Study. J Clin Endocrinol Metab. 2000;85:219–223. 17. Chapurlat RD, Garnero P, Bre´art G, Meunier PJ, Delmas PD. Serum estradiol and sex hormone-binding globulin and risk of hip fracture in elderly women. J Bone Miner Res. 2000;15:1835–1841. 18. Heikkinen E. Functional capacity and health of elderly people–the Evergreen project. Background, design and methods of the project. Scand J Soc Med Suppl. 1997;53:1–18. 19. Grimby G. Physical activity and muscle training in the elderly. Acta Med Scand Suppl. 1986;711:233–237. 20. Viitasalo J, Era P, Leskinen AL, Heikkinen E. Muscular strength profiles and anthropometry in random samples of men aged 31–35, 51–55, and 71–75 years. Ergonomics. 1985;28:1563–1574. 21. Sipila¨ S, Multanen J, Kallinen M, Suominen H. Effects of strength and endurance training on isometric muscle strength and walking speed in elderly women. Acta Physiol Scand. 1996;156:457–464. 22. Cheng S, Suominen H, Sakari-Rantala R, Laukkanen P, Avikainen V, Heikkinen E. Calcaneal bone mineral density predicts fracture occurrence: a five-year follow-up study in elderly people. J Bone Miner Res. 1997;12:1075–1082. 23. McEwen BS. Clinical review 108: The molecular and neuroanatomical basis for estrogen effects in the central nervous system. J Clin Endocrinol Metab. 1999;84:1790–1797. 24. Joseph JA, Roth GS. Hormonal regulation of motor behavior in senescence. J Gerontol. 1993;48: Spec No:51–55. 25. Michae¨lsson K, Baron JA, Farahmand BY, et al. Hormone replacement therapy and risk of hip fracture: population based case-control study. The Swedish Hip Fracture Study Group. BMJ. 1998;316:1858–1863. 26. Nguyen T, Sambrook P, Kelly P, et al. Prediction of osteoporotic fractures by postural instability and bone density. BMJ. 1993;307:1111–1115. 27. Kenny AM, Dawson L, Kleppinger A, Iannuzzi-Sucich M, Judge JO. Prevalence of sarcopenia and predictors of skeletal muscle mass in nonobese women who are long-term users of estrogen-replacement therapy. J Gerontol A Biol Sci Med Sci. 2003;58:M436–M440. 28. Rantanen T, Guralnik JM, Foley D, et al. Midlife hand grip strength as a predictor of old age disability. JAMA. 1999;281:558–560. 29. Fiatarone MA, O’Neill EF, Ryan ND, et al. Exercise training and nutritional supplementation for physical frailty in very elderly people. N Engl J Med. 1994;330:1769–1775. 30. Timonen L, Rantanen T, Ryynanen OP, Taimela S, Timonen TE, Sulkava R. A randomized controlled trial of rehabilitation after hospitalization in frail older women: effects on strength, balance and mobility. Scand J Med Sci Sports. 2002;12:186–192. 31. Messier SP, Royer TD, Craven TE, O’Toole ML, Burns R, Ettinger WH Jr. Long-term exercise and its effect on balance in older, osteoarthritic adults: results from the Fitness, Arthritis, and Seniors Trial (FAST). J Am Geriatr Soc. 2000;48:131–138. 32. Lemoine S, Granier P, Tiffoche C, Rannou-Bekono F, Thieulant ML, Delamarche P. Estrogen receptor alpha mRNA in human skeletal muscles. Med Sci Sports Exerc. 2003;35:439–443. 33. Mole PA, McMurdo ME, Paterson CR. Evaluation of peripheral dual energy X-ray absorptiometry: comparison with single photon absorptiometry of the forearm and dual energy X-ray absorptiometry of the spine or femur. Br J Radiol. 1998;71:427–432. Received March 16, 2005 Accepted June 3, 2005 Decision Editor: Luigi Ferrucci, MD, PhD