Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Vol. 161, No. 9 Printed in U.S.A. DOI: 10.1093/aje/kwi105 American Journal of Epidemiology Copyright ª 2005 by the Johns Hopkins Bloomberg School of Public Health All rights reserved Determining the Prevalence of Osteoporosis and Related Factors using Quantitative Ultrasound in Vietnamese Adult Women Vu Thi Thu Hien1,2, Nguyen Cong Khan2, Nguyen Thi Lam2, Le Bach Mai2, DucSon NguyenTrung Le1,3, Bui Thi Nhung1,2, Masayo Nakamori1, Daisuke Kunii1, Tohru Sakai1, and Shigeru Yamamoto1 1 Division of International Public Health Nutrition, Institute of Health Biosciences, University of Tokushima Graduate School, Tokushima, Japan. 2 Vietnam National Institute of Nutrition, Hanoi, Vietnam. 3 Nutrition Center of Ho Chi Minh City, Ho Chi Minh City, Vietnam. Received for publication June 30, 2004; accepted for publication December 6, 2004. In 2003, the authors conducted a population-based, cross-sectional survey to determine the prevalence of osteoporosis and related factors in Vietnamese adult women by using quantitative ultrasound at the heel bone (calcaneus). A total of 2,232 adult women aged 20 years, living in Hanoi City, and free of illnesses affecting bone metabolism were randomly selected to participate in the study. Subjects’ bone mass was assessed by speed of sound at the calcaneus, referred to as quantitative ultrasound measurement. The T-score threshold, defined as 1.8, was used to identify subjects with osteoporosis. The crude prevalence of osteoporosis in Hanoi City was 15.4%; after adjustment for age, it was 9.0%. Among premenopausal women, the crude prevalence of osteoporosis was higher in the urban areas compared with the rural areas. By contrast, postmenopausal women in the rural areas had a higher prevalence of osteoporosis. Multiple logistic regression analysis revealed that factors associated with low speed of sound were age, menopause, educational level, lifelong occupation, recreational weight-bearing exercise, number of births, and height. Results suggest that osteoporosis is a noteworthy problem in Vietnam, and intervention strategies should be considered to control it, especially in high-risk populations. cross-sectional studies; osteoporosis; risk factors; ultrasonics; Vietnam Abbreviation: QUS, quantitative ultrasound. Osteoporosis is a growing health problem recognized in both developed and developing countries (1). Osteoporotic fractures, especially hip fractures, are related to considerable mortality and increasingly higher costs of health care. The increasing prevalence of osteoporosis related to more peripheral and vertebral fractures will lead to increased socioeconomic burdens because of the high cost of treatment (2, 3). Thus, for the purposes of prevention and control, there is great interest in conducting epidemiologic surveys of the prevalence of osteoporosis and related risk factors in communities (2–4). The optimal method for diagnosing osteoporosis is to measure bone mineral density by dual-energy x-ray absorp- tiometry at the hip and lumbar spine (5). However, it is very difficult to apply this procedure in community-based studies because of its lack of portability and its cost. Furthermore, the procedure exposes subjects to low, but significant doses of ionizing radiation. Quantitative ultrasound (QUS) measurement, a technique for measuring the peripheral skeleton, has been proposed because it can be performed quickly, is relatively inexpensive, is portable, and involves less radiation. Thus, QUS could be an ideal tool to screen for osteoporosis at the community level (6–12). Although Asians are thought to have lower bone mass than Caucasians because of their smaller body size (13), few population-based studies have been conducted in Asian Reprint requests to Dr. Shigeru Yamamoto, Division of International Public Health Nutrition, Institute of Health Biosciences, University of Tokushima Graduate School, 3 Kuramoto, Tokushima, 770-8503 Japan (e-mail: [email protected]). 824 Am J Epidemiol 2005;161:824–830 Osteoporosis in Vietnamese Women countries. In Vietnam, life expectancy is increasing as the economy improves (14), and, with a longer lifespan, there is concern about an increased prevalence of osteoporosis. However, the prevalence of osteoporosis and its related factors remains unclear. A pilot study in 2000 measuring the bone mass of Vietnamese adult men and women living in Hanoi City estimated that bone density is low in this population (15). However, data were insufficient to characterize the prevalence. Thus, further studies on the epidemiologic distribution of osteoporosis and associated factors in Vietnam are required. Although osteoporosis can affect both men and women of any age, women are four times more likely than men to develop osteoporosis (16). Thus, the aim of the present study was to determine the prevalence of osteoporosis and related factors in Vietnamese adult women. MATERIALS AND METHODS Setting and study subjects This study was approved by the Research and Ethical Committee of the Vietnamese National Institute of Nutrition. It was conducted in Hanoi City from April to May 2003. Hanoi City is the capital of Vietnam, located in the northern area of the country, and has a population of approximately 3 million inhabitants (17). There are 228 wards and communes divided into 102 urban wards and 126 rural communes. A multistage sampling method was used to select subjects. First, 30 of the total 228 wards and communes were chosen by using a probability proportional-to-size sampling method (18). From each of the selected wards and communes, 80 women aged 20 years or older were chosen randomly and were stratified by 5-year age groups (20–24, 25–29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75–79, 80–84, and 85 years). To recruit subjects, a list of all families that included women aged 20 years or older was constructed, and family codes were created. From this list, the first family was selected by randomly choosing a family code. In each family selected, all women aged 20 years or older and free of bone deformities and acute illnesses were invited to participate. After selecting the first family, we used the ‘‘random walking’’ method to approach other families, adding subjects to obtain 80 women stratified by 5-year age groups (18). From 30 selected wards and communes, 2,400 women were selected to participate. This survey was carried out at the local health centers of the selected sites. To maximize participation, the local health staff visited each household to explain the study, obtain written consent from each subject, and remind participants of the time and date of the survey. At the end of the survey, a total of 2,368 participants had been included. Interview All participants completed a structured questionnaire. Included was an assessment of subjects’ medical history, Am J Epidemiol 2005;161:824–830 825 particularly focusing on hyperparathyroidism, gastrectomy, diseases of the kidney, diabetes mellitus, rheumatoid arthritis, chronic liver disease, chronic malabsorptive syndromes, cancer, and jejunoileal bypass, as well as current or past treatment with glucocorticoids and/or thyroid hormone. Subjects were excluded from analysis if they had at least one of the above conditions. Use of oral contraceptive pills was also recorded. Information on fracture history, family history of fracture and osteoporosis, socioeconomic status, lifelong occupation, educational level, and recreational weight-bearing exercise was collected from each subject. The interview also asked about current age and age at menarche and menopause, as well as number of children. Lifelong occupation was defined as the occupation that the subject engaged in most frequently in her life. It was classified as heavy work (farmers, manual workers), office work (office clerks and other sedentary jobs), or domestic work (housewife). Educational level was categorized in three groups, by number of years of schooling: low level (5 years), medium level (6–8 years), and high level (9 years). Recreational weight-bearing exercise was assessed by inquiring about regular weight-bearing exercise during at least the past 12 months. The subjects reported the number of sessions of weight-bearing exercise of at least 30 minutes per week. Active behavior was defined as engaging in more than two sessions per week. Anthropometry measurement Height and weight were measured while subjects were standing, wearing light clothing and no shoes. Body mass index was calculated as the ratio of weight (in kilograms) to height (in meters) squared. Waist circumference was measured at the minimum circumference between the umbilicus and iliac crest, and hip circumference was measured at the widest circumference around the buttocks. Bone mass assessment Bone mass was assessed by speed of sound (m/second) using a QUS device (CM-100; ELK Corporation, Tokyo, Japan). This device is small and portable, with a gel-coupled (dry) system that can measure speed of sound at the calcaneus. Coefficients of variation for the device were measured short term in vivo and in vitro. Precision error (percent coefficient of variation) using the phantom technique was 0.15 percent and, in vivo, was 0.27 percent (19). For all subjects, speed of sound was measured at the right calcaneus. The measurement was taken in a temperaturecontrolled environment and was performed by a trained medical technician only. Standardization and calibration with standards were performed before the first measurement of each survey day. The T-score for each subject was calculated by using the peak speed-of-sound value for a defined population of young adults, and its standard deviation, with the following equation (20): T-score ¼ speed of soundsubject speed of soundpeak value for young adults/ standard deviationpeak value for young adults. 826 Hien et al. We calculated the peak speed-of-sound value for young adults (speed of soundpeak value for young adults) by estimating peak bone mass, which was itself defined as the average maximum bone mass achieved by young, healthy, sex- and race-matched adults (21). For our own population, we defined the young adult population by identifying the age range at which speed of sound reached peak value. Speed of soundpeak value for young adults and standard deviationpeak value for young adults were determined as the mean speed-of-sound value for the young adult population and its standard deviation, respectively. A person was classified as having osteoporosis if her T-score was 1.8 and as normal if the score was >1.8. Statistical analysis In this paper, data are presented as percentage and as mean (standard deviation). Student’s t test (two sided) was applied to examine differences in age, number of children, anthropometric indicators, and speed-of-sound values between women in rural and urban areas. Bonferroni’s t test was used to identify significant differences in speed-ofsound values for 5-year age groups to define the age range for young adults and their peak speed-of-sound value. Chisquare testing was used to examine differences in the prevalence of osteoporosis, active weight-bearing exercise, lowest educational level, and heavy lifelong occupation between women in the rural and urban areas. Multiple logistic regression analysis was used to test several models for the associations between osteoporosis assessed by speed of sound and other variables. Here, data are presented as odds ratios and 95 percent confidence intervals. Associations were considered statistically significant at the p < 0.05 level. All statistical procedures were performed by using SPSS software for Windows, version 10.0 (SPSS, Inc., Chicago, Illinois). RESULTS TABLE 1. Characteristicsy of Vietnamese adult women living in urban and rural areas of Hanoi City in 2003 who participated in a survey of the prevalence of osteoporosis and related factors Urban (n ¼ 1,237) Variable Rural (n ¼ 995) Age (years) 48.8 (16.3) 48.0 (16.6) Weight (kg) 50.7 (7.7) 47.6 (7.1)** Body mass index (kg/m2) 22.0 (3.1) 20.9 (2.7)** Waist/hip ratio 0.83 (0.06) 0.81 (0.06)* Speed of sound at the calcaneus (m/second) 1,516 (33) 1,518 (35) Premenopausal women 153.3 (5.0) 153.1 (4.9)** Postmenopausal women 150.0 (5.5) 148.5 (5.2) Premenopausal women 9.3 11.3 Postmenopausal women 34.8 67.5** Premenopausal women 18.8 67.5** Postmenopausal women 29.1 51.4** Premenopausal women 1.5 (0.8) 1.6 (0.9) Postmenopausal women 3.6 (1.9) 4.5 (2.1)** Premenopausal women 31.8 38.1* Postmenopausal women 60.8 57.7 Height (cm) Lowest educational level (%) Heavy-work lifelong occupation (%) No. of children Active weight-bearing exercise (%) * p < 0.05; ** p < 0.001 (two sided). y Unless otherwise noted, values are expressed as mean (standard deviation). Rural values were compared with urban values by using Student’s t test or the chi-square test. children and a lower educational level than those in the urban areas. Characteristics of the subjects After we reviewed the medical histories, data on 136 subjects with illnesses deemed to affect bone metabolism were excluded from analysis. Thus, data on a total of 2,232 women (1,237 in urban and 995 in rural areas) were available for analysis. Premenopausal women accounted for 46.5 percent of the population (n ¼ 1,038). Table 1 shows the characteristics of subjects; their mean age was 48.5 (standard deviation, 16.5) years. In the rural areas, anthropometric indicators such as weight, body mass index, and waist/hip ratio were significantly lower in both premenopausal and postmenopausal women, while height was significantly lower only in postmenopausal women compared with those in the urban areas. Compared with urban women, most of the women in the rural areas had a lifelong occupation involving heavy work and engaged in more weight-bearing exercise in their leisure time. Postmenopausal women in the rural areas had more Prevalence of osteoporosis Speed-of-sound value reached its peak at age 30–34 years, started to decline at age 35–39 years, and declined significantly from age 45 to 49 years. Because mean speedof-sound values at ages 35–39 years and at ages 40–44 years were similar or even higher than for the age group 20–24 years, we defined women aged 20–44 years as young adults to calculate peak speed-of-sound value (speed of soundpeak value for young adults) and its standard deviation (standard deviationpeak value for young adults) for our own population. The peak speed-of-sound value for the young adults was determined to be 1,536 (standard deviation, 30) m/second. On the basis of this value, the crude prevalence of osteoporosis was estimated at 15.4 percent in the study population. Table 2 shows that the prevalence of osteoporosis in the age groups increased with age (v2 ¼ 669.7; p < 0.001, two-sided). After we adjusted for age (using 1999 Am J Epidemiol 2005;161:824–830 Osteoporosis in Vietnamese Women TABLE 2. Crude prevalence of osteoporosis, by age group, among Vietnamese adult women living in Hanoi City in 2003 who participated in a survey of osteoporosis and related factors Normal (%) Osteoporosis (%) Age group (years) No. 20–29 369 99.2 0.8 30–39 383 97.9 2.1 40–49 407 96.3 3.7 50–59 405 91.6 8.4 60–69 403 69.5 30.5 70–79 217 43.8 56.2 48 20.8 79.2 2,232 84.6 15.4 80 Total Hanoi census data), the estimated prevalence of osteoporosis was 9.0 percent. The prevalence of osteoporosis among premenopausal women was higher in the urban areas than in the rural areas (figure 1). By contrast, among postmenopausal women, it was higher in the rural areas than in the urban areas. Risk factors associated with low speed-of-sound value Multiple logistic regression analysis showed that age, menopause, number of children, height, educational level, lifelong occupation, and recreational weight-bearing exercise acted as significant predictors for osteoporosis (table 3). These associations were adjusted for age at menarche, weight, body mass index, waist/hip ratio, family history of osteoporosis, previous fracture, and oral contraceptive use. Age and number of children were positively related to the risk of osteoporosis, while height was negatively related to 827 TABLE 3. Significant predictors of osteoporosis,* assessed by speed of sound at the calcaneus, for Vietnamese adult women living in Hanoi City in 2003 who participated in a survey of the prevalence of osteoporosis and related factors Age (1-year increment) Odds ratio 95% CIy p value 1.15 1.12, 1.17 <0.001 2.07, 5.52 <0.001 Menopause status Premenopausal 1.00 Postmenopausal 3.38 Educational level Lowest 1.00 Medium 0.54 0.29, 0.98 0.04 Highest 0.45 0.25, 0.79 <0.01 Lifelong occupation Heavy work 1.00 Office work 2.00 1.32, 3.02 0.01 Housewife 1.95 1.28, 2.95 <0.01 Recreational weight-bearing exercise No 1.00 Yes 0.30 0.21, 0.44 <0.001 No. of children (1-quartile increment) 1.75 1.11, 2.76 <0.05 Height (1-quartile decrement) 1.30 1.07, 1.57 <0.01 * Information was obtained by multiple logistic regression analysis. The model was adjusted for age at menarche, weight, body mass index, waist/hip ratio, family history of osteoporosis, previous fracture, and oral contraceptive use. y CI, confidence interval. the risk of osteoporosis. Increased educational level was associated with a significantly reduced risk of osteoporosis. For subjects who engaged in weight-bearing exercise at least three times a week, the prevalence of osteoporosis was three times lower than for those who did not. Regarding menopause status, postmenopausal women had a threetimes higher risk of osteoporosis than premenopausal women. On the basis of lifelong occupation, office workers and housewives had double the risk of osteoporosis compared with women engaged in heavy work. DISCUSSION FIGURE 1. Comparison of the prevalence of osteoporosis among premenopausal and postmenopausal women living in urban and rural areas of Hanoi City, Vietnam, in 2003. Am J Epidemiol 2005;161:824–830 Recently, use of QUS at peripheral sites has been proposed for measuring bone density in large populations (3, 22). In the current study, we attempted to identify the prevalence of osteoporosis and related factors in Vietnamese adult women by using QUS at the calcaneus with our dataderived T-score and by defining osteoporosis as a T-score of 1.8. By using this method, we estimated the prevalence of osteoporosis in Vietnamese adult women in rural and urban communities of Hanoi City. The age-adjusted prevalence of osteoporosis for our participating Hanoi subjects 828 Hien et al. was 9.0 percent. In addition, we found contrasting distributions of osteoporosis between urban and rural areas. Speed-of-sound measurements at the calcaneus can identify persons at risk of osteoporotic fracture as reliably as bone mineral density measurements (23–26) and could be an ideal tool to screen for osteoporosis at the community level (6–12). However, it also has been demonstrated that the T-score threshold of 2.5 may lead to underestimation of the prevalence of osteoporosis when QUS is measured at the heel (21). Recent studies on the Hologic Sahara and the Hologic UBA575þ (Hologic, Bedford, Massachusetts) and on the Osteometer DTUone (Osteometer MediTech, Inc., Hawthorne, California) QUS devices indicate that a T-score threshold of 1.8 identifies the same percentage of persons with osteoporosis as the World Health Organization threshold for bone mineral density measurements (27). The CM100 device used in our study is very similar to the Sahara device and is also used to measure QUS parameters at the calcaneus. The device comes with a manufacturer’s recommended cutoff value for diagnosing osteoporosis, validated by dual-energy x-ray absorptiometry (28). In addition, we found excellent agreement between the manufacturer’s cutoff and our data-derived T-score in classifying osteoporosis (kappa ¼ 0.967, p < 0.001). Therefore, we felt it was reasonable to apply a T-score threshold of 1.8 to define osteoporosis in our population. The age at peak bone mass depends on skeletal sites or technologies used (20). Evidence from previous studies indicates that QUS parameters start to decline from the age of 40–45 years and fall steadily thereafter (29, 30). Our study showed similar results, with the peak value of speed of sound occurring from 20 to 44 years of age, followed by a significant decline thereafter. For these reasons, we concluded that defining peak speed-of-sound value as the mean for women aged 20–44 years and defining osteoporosis by using a T-score threshold of 1.8 is reasonable in screening for osteoporosis in our population when the CM100 device is used. In this study, the prevalence of osteoporosis in Vietnamese women was found to be lower than that in Japanese women in the age group 50–79 years (29.5 percent vs. 51.2 percent) (31) but was similar or higher in comparison to nearby countries such as China (3.7 percent vs. 3.7 percent in the age group 40–49 years, 8.4 percent vs. 3.9 percent in the age group 50–59 years) (32) and Korea (12.8 percent vs. 11.8 percent in women aged 50 years) (33) when the same age range was compared or when the same method was used to classify osteoporosis. When we put the information into context, Japanese women aged 50–79 years were infants and young children during World War II. Their nutrition was poor, and calcium intake in 1946 was only 253 mg per day (34). Therefore, the prevalence of osteoporosis in that age group is very high. In Vietnam, women who are aged 50–79 years were also exposed to war and had poor nutrition. Thus, it is reasonable to compare the prevalence of osteoporosis between Vietnamese and Japanese women aged 50–79 years. The Vietnamese women aged 40–49 years were born when the war had just ended. The situation may be the same with Chinese women in the same age range. This factor might explain why the prevalence of osteoporosis is comparable in Vietnamese and Chinese women aged 40– 49 years. When classified on the basis of the T-score threshold of the World Health Organization, the prevalence of low bone mass among Vietnamese women aged 20 years or older in 2003 increased compared with that in 2000 (6.4 percent vs. 5.7 percent) (15). These comparisons may be slightly weakened by the difference in the measurement, calibration, and standardization methods used in these studies, but our findings indicate that osteoporosis is becoming a noteworthy problem in Vietnam. Factors affecting bone mineral density have been studied extensively (35–38). Recently, the impact of risk factors on QUS parameters has also been given more attention. Our data show that increasing age and postmenopausal status are predictors of low speed of sound. Increased age was associated with a significantly increased risk of osteoporosis, especially when women become postmenopausal. This finding is consistent with those from other studies (10, 15, 39). Our study also agreed with previous research reporting that low educational level was related to increased risk of osteoporosis (40). Low stature and a large number of children are predictors of low speed of sound. These associations were demonstrated in previous studies (41– 43) and were also confirmed in our data. In addition, our survey indicates that engaging in recreational, active weightbearing exercise and having a lifelong occupation involving heavy work can help protect subjects from osteoporosis. Our study also revealed contrasting distributions of osteoporosis between urban and rural areas of the country. To our knowledge, this finding has not been observed in other countries and is likely due to socioeconomic characteristics of Vietnamese. In rural areas, most of the people are farmers and take part in more active weight-bearing exercise, which might explain why the prevalence of osteoporosis among rural premenopausal women was lower compared with that among urban women. However, because rural postmenopausal women have more children, a lower stature, and a lower educational level, the prevalence of osteoporosis among postmenopausal women in the rural areas was higher. In addition, rural women consume fewer dairy products, and eggs/milk consumption in the rural areas was lower than in the urban areas (5.4 g per capita/day vs. 25.5 g per capita/day). Moreover, the postmenopausal women experienced greater problems with access to enough food, including calcium-rich foods, when they were teens and young adults because of war and the early years thereafter (44). Exposed to a long period of inadequate calcium intake, rural postmenopausal women are more likely to suffer from osteoporosis. This finding may partially explain why the prevalence of osteoporosis among postmenopausal women was higher in the rural areas, and it should be confirmed by further studies in order to design an appropriate intervention strategy. The present study has several limitations. First, the T-score threshold for diagnosing osteoporosis by QUS or use of the CM-100 device has not been established. However, in the context of epidemiologic studies at the community level in developing countries, we could only apply QUS and use a T-score threshold of 1.8 to identify subjects with osteoporosis. Many studies support the view of Am J Epidemiol 2005;161:824–830 Osteoporosis in Vietnamese Women using QUS to screen for osteoporosis at the community level (6–12). A previous study also applied a T-score threshold of 1.8 to define osteoporosis using QUS (45). For these reasons, we believe that, with a large sample size and random sampling methods, our data may truly represent the prevalence of osteoporosis in Hanoi City, or they may even underestimate its prevalence. Second, this study was a crosssectional survey, and we could not measure all factors affecting the risk of osteoporosis. Moreover, using a prevalence measure to assess risk factors also has limitations because the factors associated with osteoporosis may not reflect the real cause-effect relation. Risk factors assessed by using prevalence measures can only suggest the hypothesis for a possible cause-effect relation. Thus, a prospective study is needed to confirm any association between low speed of sound and risk factors and to explain the differing patterns of osteoporosis distribution between urban and rural areas. In conclusion, this study shows that the prevalence of osteoporosis in Vietnamese adult women in Hanoi City, determined by QUS, is relatively high compared with that in nearby countries. In addition, our data indicate differing distributions of osteoporosis between rural and urban areas of Hanoi City. These findings suggest that osteoporosis is a problem in Vietnam. Furthermore, using multiple logistic regression, we assessed associations between low speed of sound and risk factors, as well as protective factors. Because osteoporosis is related to considerable mortality and increasingly higher costs of health care, screening for osteoporosis, particularly in high-risk populations, and setting up a national program to prevent and control osteoporosis in Vietnam are urgently needed. REFERENCES 1. Delmas PD, Fraser M. Strong bones in later life: luxury or necessity? Bull World Health Organ 1999;77:416–22. 2. Barrett-Connor E. The economic and human costs of osteoporotic fracture. Am J Med 1995;98(suppl 2A):S3–8. 3. Boyanov M, Popivanov P. Prevalence of low forearm bone density in a Bulgarian female referral population. Osteoporos Int 2002;13:288–95. 4. Liao ER, Wu XP, Deng XG, et al. Age-related bone mineral density, accumulated bone loss rate and prevalence of osteoporosis at multiple skeletal sites in Chinese women. Osteoporos Int 2002;13:669–76. 5. World Health Organization. Assessment of osteoporotic fracture risk and its application to screening for postmenopausal osteoporosis. Geneva, Switzerland: World Health Organization, 1994:29–30. (WHO technical report series no. 843). 6. Hans D, Dargent-Molina P, Schott AM, et al. Ultrasonographic heel measurements to predict hip fracture in elderly women. Lancet 1996;348:511–14. 7. Pluijm SM, Graafmans WC, Bouter LM, et al. Ultrasound measurements for the prediction of osteoporotic fractures in elderly people. Osteoporos Int 1999;9:550–6. 8. Wuster C, Hadji P, Blaul G, et al. Quantitative bone ultrasound (QUS) of the heel bone for diagnosis of osteoporosis in the general community. Zentralbl Gynakol 1999;121:137–42. Am J Epidemiol 2005;161:824–830 829 9. Benitez CL, Schneider DL, Barrett-Connor E, et al. Hand ultrasound for osteoporosis screening in postmenopausal women. Osteoporos Int 2000;11:203–10. 10. Yamaguchi J, Truman G, Cameron ID. Lifestyle factors affecting bone ultrasonometry of the calcaneus in Japanese women. Calcif Tissue Int 2000;66:43–6. 11. Nairus J, Ahmadi S, Baker S, et al. Quantitative ultrasound: an indicator of osteoporosis in perimenopausal women. J Clin Densitom 2000;3:141–7. 12. Kim KI, Han IK, Kim H, et al. How reliable is the ultrasound densitometer for community screening to diagnose osteoporosis in spine, femur, and forearm? J Clin Densitom 2001; 4:159–65. 13. Bhudhikanok GS, Wang MC, Eckert K, et al. Differences in bone mineral in young Asian and Caucasian Americans may reflect differences in bone size. J Bone Miner Res 1996; 11:1545–56. 14. United Nations Development Programme. Basic facts about Viet Nam. UNDP Viet Nam Country Office, Hanoi, Viet Nam, March 2005. (http://www.undp.org.vn/undp/fact/base.htm). 15. Thuy VTT, Chau TT, Cong ND, et al. Assessment of low bone mass in Vietnamese: comparison of QUS calcaneal ultrasonometer and data-derived T-scores. J Bone Miner Metab 2003;21:114–19. 16. Miller GD, Jarvis JK, McBean LD, et al. Handbook of dairy foods and nutrition. 2nd ed. Washington, DC: National Dairy Council, 2000. 17. Hanoi Statistical Office. The population and housing census of Hanoi City 1999. (In Vietnamese). Hanoi, Vietnam: Statistical Publishing House, Hanoi, 2001. 18. The United Nation Children’s Fund. Choosing the sampling. In: Practical handbook for multiple-indicator surveys. New York, NY: UNICEF, 1995;4:1–28. 19. Kishimoto H. Examination of the clinical usefulness of new ultrasonic bone density measurement equipment. (In Japanese). Osteoporos Japan 1997;5:815–17. 20. Cheng S, Fan B, Wang L, et al. Factors affecting broadband ultrasound attenuation results of the calcaneus using a gelcoupled quantitative ultrasound scanning system. Osteoporos Int 1999;10:495–504. 21. Frost ML, Blake GM, Fogelman I. Contact quantitative ultrasound: an evaluation of precision, fracture discrimination, age-related bone loss and applicability of the WHO criteria. Osteoporos Int 1999;10:441–9. 22. Greenspan SL, Cheng S, Miller PD, et al. Clinical performance of a highly portable, scanning calcaneal ultrasonometer. Osteoporos Int 2001;12:391–8. 23. Yamazaki K, Kushida K, Ohmura A, et al. Ultrasound bone densitometry of the os calcis in Japanese women. Osteoporosis Int 1994;4:220–5. 24. Hans D, Dargent-Molina P, Schott AM, et al. Ultrasonographic heel measurements to predict hip fracture in elderly women. Lancet 1996;348:511–14. 25. Bauer DC, Gluer CC, Caley JA, et al. Broadband QUS attenuation predicts fractures strongly and independently of densitometry in older women. A prospective study. Study of Osteoporotic Fractures Research group. Arch Intern Med 1997;157:629–34. 26. Gregg EW, Kriska AM, Salomone LM, et al. The epidemiology of quantitative ultrasound: a review of the relationships with bone mass, osteoporosis and fracture risk. Osteoporos Int 1997;7:789–99. 27. Frost ML, Blake GM, Fogelman I. Can the WHO criteria for diagnosing osteoporosis be applied to calcaneal quantitative ultrasound? Osteoporos Int 2000;11:321–30. 830 Hien et al. 28. Yoh K. Correlation between cut off value determined by using quantitative ultrasound and threshold of fracture in Japanese women. J Bone Miner Res 2002;17(suppl):M106. 29. Langton CM, Langton DK. Male and female normative data for ultrasound measurement of the calcaneus within the UK adult population. Br J Radiol 1997;70:580–5. 30. Stetten EV, Wilson KE, Ouellet H, et al. Caucasian female reference ranges for the Sahara clinical bone sonometer. In: Ring EFJ, Elvins DM, Bhalla AK, eds. Current research in osteoporosis and bone mineral measurement. Vol. V. London, United Kingdom: The British Institute of Radiology, 1998:67–76. 31. Iki M, Kagamimori S, Kagawa Y, et al. Bone mineral density of the spine, hip and distal forearm in representative samples of the Japanese female population: Japanese Population-Based Osteoporosis (JPOS) Study. Osteoporos Int 2001;12:529–37. 32. Liao EY, Wu XP, Deng XG, et al. Age-related bone mineral density, accumulated bone loss rate and prevalence of osteoporosis at multiple skeletal sites in Chinese women. Osteoporos Int 2002;13:669–76. 33. Kim CH, Kim YI, Choi CS, et al. Prevalence and risk factors of low quantitative ultrasound values of calcaneus in Korean elderly women. Ultrasound Med Biol 2000;26:35–40. 34. Ministry of Health and Welfare. Results of the National Nutrition Survey in Japan, 1998. (In Japanese). Tokyo, Japan: Daiichi Shuppan Press, 2000. 35. Cheng S, Suominen H, Heikkinen E. Bone mineral density in relation to anthropometric properties, physical activity and smoking in 75-year-old men and women. Aging Clin Exp Res 1993;5:55–62. 36. Takada H, Washino K, Iwata H. Risk factors for low bone mineral density among females: the effect of lean body mass. Prev Med 1997;26:633–41. 37. Grainge MJ, Coupland CA, Cliffe SJ, et al. Cigarette smoking, alcohol and caffeine consumption, and bone mineral density in postmenopausal women. The Nottingham EPIC Study Group. Osteoporos Int 1998;8:355–63. 38. Uusi-Rasi K, Sievanen H, Vuori I, et al. Associations of physical activity and calcium intake with bone mass and size in healthy women at different ages. J Bone Miner Res 1998; 13:133–42. 39. Gregg EW, Kriska AM, Salamone LM, et al. Correlates of quantitative ultrasound in the Women’s Healthy Lifestyle Project. Osteoporos Int 1999;10:416–24. 40. Varenna M, Binelli L, Zucchi F, et al. Prevalence of osteoporosis by educational level in a cohort of postmenopausal women. Osteoporos Int 1999;9:236–41. 41. Donaldson MMK, McGrother CW, Clayton DG, et al. Calcaneal ultrasound attenuation in an elderly population: measurement position and relationships with body size and past fractures. Osteoporos Int 1999;10:316–24. 42. Inanici-Ersoz F, Gokce-Kutsal Y, Oncel S, et al. A multicenter, case control study of risk factors for low tibial speed of sound among residents of urban areas in Turkey. Rheumatol Int 2002;22:20–6. 43. Saadi HF, Reed RL, Carter AO, et al. Quantitative ultrasound of the calcaneus in Arabian women: relation to anthropometric and lifestyle factors. Maturitas 2003;44:215–23. 44. Vietnamese Ministry of Health–National Institute of Nutrition. General Nutrition Survey 2000. Hanoi, Vietnam: Medical Publishing House, Hanoi, 2003. 45. Frost ML, Blake GM, Fogelman I. Quantitative ultrasound and bone mineral density are equally strongly associated with risk factors for osteoporosis. J Bone Miner Res 2001;16: 406–16. Am J Epidemiol 2005;161:824–830