SEARCH

SEARCH BY CITATION

Keywords:

  • bone density;
  • elderly;
  • osteoporosis;
  • longitudinal study;
  • risk factors

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Few studies have evaluated risk factors for bone loss in elderly women and men. Thus, we examined risk factors for 4-year longitudinal change in bone mineral density (BMD) at the hip, radius, and spine in elders. Eight hundred elderly women and men from the population-based Framingham Osteoporosis Study had BMD assessed in 1988-1989 and again in 1992-1993. BMD was measured at femoral neck, trochanter, Ward's area, radial shaft, ultradistal radius, and lumbar spine using Lunar densitometers. We examined the relation of the following factors at baseline to percent BMD loss: age, weight, change in weight, height, smoking, caffeine, alcohol use, physical activity, serum 25-OH vitamin D, calcium intake, and current estrogen replacement in women. Multivariate regression analyses were conducted with simultaneous adjustment for all variables. Mean age at baseline was 74 years ± 4.5 years (range, 67-90 years). Average 4-year BMD loss for women (range, 3.4-4.8%) was greater than the loss for men (range, 0.2-3.6%) at all sites; however, BMD fell with age in both elderly women and elderly men. For women, lower baseline weight, weight loss in interim, and greater alcohol use were associated with BMD loss. Women who gained weight during the interim gained BMD or had little change in BMD. For women, current estrogen users had less bone loss than nonusers; at the femoral neck, nonusers lost up to 2.7% more BMD. For men, lower baseline weight and weight loss also were associated with BMD loss. Men who smoked cigarettes at baseline lost more BMD at the trochanter site. Surprisingly, bone loss was not affected by caffeine, physical activity, serum 25-OH vitamin D, or calcium intake. Risk factors consistently associated with bone loss in elders include female sex, thinness, and weight loss, while weight gain appears to protect against bone loss for both men and women. This population-based study suggests that current estrogen use may help to maintain bone in women, whereas current smoking was associated with bone loss in men. Even in the elderly years, potentially modifiable risk factors, such as weight, estrogen use, and cigarette smoking are important components of bone health.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Little longitudinal data exist on bone mineral density (BMD) changes in the elderly. Bone density studies have focused primarily on perimenopausal and early post-menopausal women; however, bone density in elderly persons is highly relevant to the risk of osteoporotic fracture.(1–7) About 26 million white women in the United States have low bone mass, and their lifetime risk of osteoporosis-related fractures exceeds 40%.(8,9) The highest rates of osteoporosis-related fractures occur in elderly women, although 13% of men also will experience such fractures. Low bone density in the femur is a strong predictor of the increased risk for hip fracture. In one of the largest cohort studies of osteoporosis, the Study of Osteoporotic Fractures, women in the lowest quartile of BMD had an 8-fold increased risk of hip fracture compared with the women in the highest BMD quartile.(4) Slowing age-related bone loss may lead to the prevention of a considerable number of fractures. Despite the importance of bone density in the elderly, very little is known about change in bone density of elders and the relative importance of traditional osteoporosis risk factors.

Bone loss, at least at the femur and radius, continues in the elderly years, even past the age of 85 years.(10–18) Our earlier cross-sectional study of femoral and radius BMD in the Framingham cohort found that age was inversely related to BMD in both men and women.(10) Both cross-sectional and longitudinal studies also have reported age-related bone loss.(11–15,17,18) Nevertheless, few studies have examined risk factors for bone loss in elders. It is unknown whether factors affecting bone loss in postmenopausal women are similar for elderly men and women. The population at greatest risk for fracture is the elderly, and thus the greatest public health impact would stem from preventing bone loss in this group.

To evaluate how bone density changes in older persons over time, we examined the longitudinal change in BMD at several anatomic sites for the elderly men and women of the Framingham study, a population-based cohort. Our study further queried whether lifestyle factors and anthropometric variables might be important risk factors for subsequent bone loss. The purpose of our study was to evaluate the relation between baseline risk factors and 4-year BMD loss at several skeletal sites for the now elderly Framingham cohort men and women.

METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Study subjects

The population-based Framingham was established in 1948 with the primary aim of examining risk factors over time for heart disease in 5209 men and women 28–62 years old.(19–21) Subjects, nearly all white, are seen biennially for a physical examination and a battery of questionnaires and tests. Since the establishment of the cohort 50 years ago, nearly two-thirds of the 5209 cohort members have died. The surviving, now elderly, cohort subjects follow the same age- and sex-specific population proportions found in the general population of Framingham, MA.(10) At biennial examination 20 (1988–1989), 1164 cohort members participated in the Framingham Osteoporosis Study, including nursing home residents who were ambulatory. In the osteoporosis study group, 1142 members obtained valid proximal radial shaft scans and 1102 members obtained valid femoral BMD measurements. Because of length of the routine biennial Framingham study clinic examination, subjects were asked to return to a callback examination to obtain BMDs at additional skeletal sites. Lumbar spine and ultradistal radius scans were obtained at examination 20 as a callback to the regular clinic examination with 842 subjects having a lumbar spine scan and 882 subjects having the ultradistal radius scan. Details from the baseline Framingham osteoporosis study at biennial examination 20 have been reported.(10)

As part of their regular clinic visit at biennial examination 22 (1992–1993), 800 subjects (69%) who had baseline BMD assessed at examination 20 had follow-up radial shaft and/or femoral BMD measures approximately 4 years later. All longitudinal scan pairs were evaluated for consistency of anatomic site and quality of analysis by the original technician, and those scans showing inconsistencies were reanalyzed by two experienced investigators (M.T.H. and D.P.K.). Of these 800 subjects with longitudinal scans, 780 cohort members had valid radial shaft BMD measures and 758 subjects had valid repeat hip BMD scans. As in the baseline osteoporosis study, we conducted a callback component to evaluate bone density in the ultradistal radius and spine at biennial examination 22. At this callback examination, 567 members had repeat spine scans and 557 subjects had repeat ultradistal radius scans. Seventeen femur scans and 16 arm scans were performed on the contralateral side at follow-up and thus were not considered valid scans for the longitudinal analysis. Our study was approved by the appropriate institutional review board, and written informed consent was obtained for all study subjects at both examinations.

BMD

Bone density was measured at the 33% radial shaft site and at the ultradistal radius site in grams per centimeter squared using a Lunar SP2 single-photon absorptiometer (Lunar Radiation Corporation, Madison, WI, U.S.A.) at both examinations. The ultradistal radius site includes both radius and ulna bones at the radial/ulna interface per manufacturer. Quality control scans indicated no shift in BMD because of equipment or change in radioactive source over the 4 years of follow-up. BMDs of the proximal right femur (femoral neck, greater trochanter & Ward's area), as well as the lumbar spine, were measured in grams per centimeter squared using a Lunar dual-photon absorptiometer (DP3) (Lunar Radiation Corporation) at biennial exam 20 and a dual X-ray absorptiometry (DPX-L) densitometer at exam 22. The right side was scanned at each exam unless there was a history of previous fracture or hip joint replacement, in which case, the left side was scanned. The lumbar spine BMD represents the average BMD of L2–L4. We used standard positioning as recommended by the manufacturer, including medial rotation of the femur to ensure a clear scan of the femoral neck region. Monthly measurements of a bone phantom over the follow-up period showed no machine drift across time. The CV in normals over the 2 years of the baseline examination for the DP3 was 2.6 (femoral neck), 2.8 (trochanter), and 4.1 (Wards area). The proximal radial shaft CV was 2% using young normal controls; the ultradistal radius CV was 5.7%; and the CV of the lumbar spine was 2.2%. Details of the measurements taken in 1988–1989 at examination 20 were published previously.(10) We used 21 cohort members measured twice with repositioning to evaluate the CV for the DPX-L measurements at examination 22, finding the CV for the femoral neck of 1.7%, for the trochanter 2.5%, for the Ward's area 4.1%, and for the lumbar spine 0.9%. We previously showed high correlations between dual-photon and dual X-ray absorptiometry.(22) However, because of a small but consistent shift in BMD values between the two technologies, femoral BMDs were adjusted for the change in equipment from DP3 to DPX-L technology, using published corrections, based on cross calibrations of the two instruments using our Framingham study subjects.(22) All scans were examined for correct analysis and placement of the region of interest (ROI) analysis box. The specific quality control protocols used to ensure comparability between longitudinal scans included re-examination of all outlier BMD values and re-examination of all longitudinal femur scan pairs with a 5% or greater difference in the area included in the femoral neck ROI box, to ensure that the analyzed region was similarly placed for each pair of longitudinal scans. When necessary, this ROI box was repositioned to more closely approximate the baseline femoral neck ROI area.(22) Any scans with metal or other attenuating material in the region of interest as well as any scans of poor quality were deleted. The quality control protocol resulted in several scans of poor quality being dropped from the analysis, as well as the deletion of nine femoral trochanteric BMD values in which the bone edges for at least one of the paired trochanter scans were cut off during the scan data collection, resulting in an incorrect trochanter BMD value.

Risk factors

We examined the relation of the following factors at baseline (examination 20) to BMD loss at each skeletal site over the 4-year follow-up. The effect of age at baseline was examined in 5-year age groups as well as in continuous units. Weight was measured at examinations 20 and 22 using a standardized balance beam scale. In addition to analyzing weight on a continuous scale, we divided weight at baseline into sex-specific quartiles for analysis. Percent change in weight over the study interval was defined as the difference between exam 20 weights and exam 22 weights, divided by exam 20 weight and multiplied by 100. Percent weight change over the follow-up was categorized into three groups: weight loss greater than 5%, weight loss or gain of no more than 5% (referent group), and weight gain greater than 5%. Height (without shoes) was measured to the nearest one-fourth inch using a stadiometer and analyzed both as continuous and by sex-specific quartiles.

A number of factors were assessed via questionnaire at baseline and preceding examinations. Smoking status was assessed at baseline as current cigarette smoker (smoked regularly in the past year), former smoker, or never smoked. Caffeine use, incorporating coffee and tea intake, was defined as the sum of daily coffee intake (1 cup equals 1 caffeine unit) and daily tea intake (1 cup equals 0.5 caffeine units). Caffeine units were grouped into 0–2 caffeine units consumed per day or more than 2 caffeine units per day based on previous work.(23–26) Current weekly intake of beer, wine, or hard liquor was grouped into current user or nonuser of alcohol, as well as grams of alcohol consumed per week.(27) We also evaluated alcohol in categories based on a previous cross-sectional study by our group that showed an association with BMD: <1.0 oz/week, 1 to <3.0 oz, 3.0 to <7.0 oz, and 7.0 oz/week or more.(28) For women, current use of oral conjugated estrogen, patch, or cream at baseline was examined for its possible effect on bone loss over the follow-up, based on our prior work.(29) Physical activity at baseline was examined using two different aspects. First, we evaluated sex-specific quartiles of the Framingham physical activity index, a weighted 24-h score of typical daily activity, based on hours spent doing heavy, moderate, light, or sedentary activity as well as sleeping.(30,31) The Framingham physical activity score has been used in cardiovascular research to predict heart disease outcome.(32,33) Second, we defined two baseline inactivity variables: first, those subjects who responded affirmatively to “spending most of the day in bed or a chair,” and second, whether the participant reported spending most of the day indoors. We also inquired about general health status using the question “In general, how is your health now?” and categorized respondents into two groups of excellent or good versus fair or poor.

Serum 25-OH vitamin D concentration was determined by a competitive protein-binding assay.(34) Inter- and intra-assay CVs for the 25-OH D assay were 10% and 7%, respectively.(35) Baseline serum levels of (25-OH) vitamin D were available on 715 of the longitudinal subjects and evaluated by categories of low (4 to <20 ng/ml), medium (20-30 ng/ml), and high (>30 ng/ml) based on prior work.(36) Baseline dietary calcium intake, including calcium supplements, was collected from 671 of the subjects based on the Willett 126-item food frequency questionnaire and evaluated by categories of low (145-400 mg/day), moderate (>400-800 mg/day), and high (>800 mg/day) levels of calcium.(37,38)

Data analysis

We examined percent change from baseline BMD to the 4-year follow-up BMD as well as absolute change in bone density. Analyses were conducted separately for men and women. Percent change in BMD was calculated as the difference between baseline and follow-up BMD, divided by baseline BMD, and multiplied by 100. We assessed these components in 5-year age groups by sex. Baseline characteristics were compared using Student's t-tests or χ2-tests as appropriate. We evaluated each BMD site separately, using multiple linear regression to examine the relation of BMD percent change with risk factors. Multivariate regression analyses were conducted with simultaneous adjustment for all variables, except calcium intake and serum vitamin D because they were assessed on only a subset of subjects. Models were adjusted further for calcium intake and serum vitamin D for the 671 subjects with these data, controlling for all other risk factors as well. To examine the possibility of nonlinear associations, we analyzed quartiles of certain risk factors (weight, height, physical activity, calcium, and vitamin D). Where possible, adjusted mean BMDs (least squared means ± SE) are presented for the categories of risk factors. Least squared means are adjusted for the risk factors, that is, they are the means that would be expected for each category if subjects had the same mean values on confounders. No adjustments were made to P-values for multiple comparisons. All analyses were conducted using the SAS statistical analysis package (SAS Institute Inc., Cary, NC, U.S.A.; version 6.12). Similar results were found whether the outcome was the absolute change in BMD or the percent change. We present percent change analyses because they are somewhat easier to interpret and perhaps more intuitive.

Table Table 1.. Comparison of Framingham Cohort Members Attending Both Baseline and Follow-Up Examinations to Those Members Only Attending Baseline Examination
 Attended both exams (n = 800)Attended only baseline exam (n = 341)P-value
  1. Number for attendees is based on BMD for either femur or radius scan.

Mean baseline age (years ± SD)74.5 ± 4.577.8 ± 5.80.0001
(range)(67–90)(67–95) 
Mean baseline weight (lbs ± SD)155.7 ± 32.2151.7 ± 29.30.0846
(range)(87–327)(89–270) 
Percent female63.2%57.4%0.0640
Mean baseline BMD (g/cm2 ± SD):
Radial shaft0.592 ± 0.1350.580 ± 0.1420.1796
Femoral neck0.788 ± 0.1440.757 ± 0.1500.0015
Trochanter0.707 ± 0.1600.691 ± 0.1760.1830
Ward's area0.603 ± 0.1460.584 ± 0.1630.0714
Lumbar spine1.168 ± 0.2341.182 ± 0.2560.4390
Ultradistal radius0.288 ± 0.0870.290 ± 0.0860.8080
Percent current smoker9.4%12.5%0.2620
Former smoker45.7%45.5% 
Never smoker44.9%42.0% 
Mean alcohol at baseline (oz/week ± SD)2.3 ± 3.72.3 ± 4.00.7345
Mean physical activity score (±SD)33.7 ± 5.632.0 ± 5.20.0001
(range)(24.0–63.2)(24.0–51.7) 
Calcium intake (mg/day ± SD)810.53 ± 437.34783.59 ± 415.960.4080
Serum 25-OH vitamin D (ng/ml ± SD)30.3 ± 12.328.8 ± 12.80.0682
Women only
 Percent current estrogen use6.7%2.1%0.0210
 Ever used estrogen38.8%30.8%0.0470

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Of the 1132 subjects who had baseline radius BMD assessed, 764 cohort members (486 women and 278 men) or 67% had valid, repeat BMD measures as part of their regular clinic visit in 1992-1993. From the 1102 baseline femur scans, 741 subjects had valid, longitudinal scans (67%). From the longitudinal subset with callback BMDs available, 567 cohort members had longitudinal spine scans and 557 members had longitudinal wrist scans. The mean age at baseline (examination 20) for those subjects with longitudinal data was 74 years ± 4.5 years, with an age range of 67-90 years. Table 1 compares the 800 cohort members with valid longitudinal femur and/or radial shaft data to those 341 members who only attended the baseline examination and did not obtain follow-up. Nearly half of the nonparticipants died during the follow-up period (43%), 89 (25%) were examined at home or in a nursing home where we were unable to perform BMD scans because of lack of machine mobility, and 110 cohort members failed to attend the follow-up examination. The cohort members without longitudinal data were more likely to be older and male, and their gender-specific mean baseline BMDs were lower than those members who attended follow-up. Nonparticipants were not as likely as participants to have reported good health, had lower physical activity scores, and were more likely to report inactivity at baseline examination (subject spent most of the day in bed or in a chair or spent most of day indoors).

The mean baseline femoral neck BMD for women was 0.732 g/cm2 with a 4-year loss of 0.026 g/cm2 (Table 2). Average percent BMD loss across the 4 years of follow-up for women ranged from 4.8% loss at the radial shaft (1.2% loss per year) to 3.4% loss at the trochanter site (0.86% loss per year). For men, the mean baseline femoral neck BMD was higher than that of women, 0.885 g/cm2 with a 4-year loss of 0.014 g/cm2. Average percent loss during the follow-up for men ranged from 3.6% at the radial shaft (0.9% loss per year) to 0.17% at the trochanter site (0.04% loss per year). As seen in Table 2, the mean 4-year BMD loss was much greater for women than the loss for men at the femur sites but similar at the forearm and spine BMD sites. Annualized rates of BMD loss are seen also in Table 2.

Table Table 2.. Mean Change in BMD (±SE) Over 4-Year Follow-Up for Women and Men
 Baseline BMD (g/cm2)Absolute change (g)Mean percent change (±SE)Mean percent loss/year
  1. Sample size for analysis may vary by site.

Women (n = 486)
 Femoral neck0.732 ± 0.0002−0.026−3.48 ± 0.010.87
 Trochanter0.632 ± 0.0003−0.023−3.42 ± 0.020.86
 Ward's area0.561 ± 0.0003−0.026−4.22 ± 0.021.06
 Radial shaft0.515 ± 0.0002−0.026−4.84 ± 0.021.21
 Lumbar spine1.072 ± 0.0005−0.046−4.48 ± 0.021.12
 Ultradistal radius0.240 ± 0.0002−0.012−4.24 ± 0.041.06
Men (n = 278)
 Femoral neck0.885 ± 0.0005−0.014−1.51 ± 0.030.38
 Trochanter0.844 ± 0.0005−0.002−0.17 ± 0.030.04
 Ward's area0.674 ± 0.0006−0.005−0.63 ± 0.040.16
 Radial shaft0.723 ± 0.0003−0.027−3.59 ± 0.020.90
 Lumbar spine1.328 ± 0.0062−0.006−0.37 ± 0.040.09
 Ultradistal radius0.370 ± 0.0003−0.013−3.09 ± 0.050.77
thumbnail image

Figure FIG. 1.. Distribution of percent of change in BMD for femoral neck and radius.

Download figure to PowerPoint

Figure 1 presents percent change in radial shaft and femoral neck BMD for women and men over the 4 years of follow-up. At the femoral neck site, 41% of women lost between 5% and 33% of their baseline BMD while 23% lost between 1% and 4%, 10% had less than 1% change in bone density, and 16% showed 1–4% gain in BMD from baseline with a further 10% showing a 5–20% gain in femoral neck BMD. At 4-year follow-up 64% of women lost between 1% and 24% of their baseline femoral neck BMD. Men had a similar pattern of percent change in femoral neck BMD although slightly more men had less than 1% change in BMD (14%) and 27% had BMD loss between 5% and 31%. In both sexes there was a trend for bone loss, although some subjects gained BMD over the follow-up. The distributions of percent change at the other BMD sites were similar to the femoral neck site for both men and women.

Fig. 2 presents the mean percent change in femoral neck BMD for women and men across the five-year age groups. Both men and women in all age groups lost BMD on average over the four years of follow-up. Percent loss in femoral neck BMD ranged from 2.2 to 8.1 percent in women and from 1.1 percent to 6.2 percent in men across the age groups. Percent loss at the radial shaft BMD across the 5 year age groups ranged from 3.9 percent to 5.5 percent in women and from 1.0 percent to 4.9 percent in men. Typically, the loss in BMD was similar across the baseline age groups at each BMD site. Indeed, age considered as either a continuous variable or in age groups was not associated with bone loss at any skeletal site (all P-values >0.75) for women or men.

thumbnail image

Figure FIG. 2.. Percent of change in femoral neck BMD for women and men by age group.

Download figure to PowerPoint

Table 3 shows the multivariate adjusted mean percent BMD change for women during the 4 years of follow-up for those risk factors that were associated with bone loss at any site in either gender at P < 0.10. After controlling for the possible effects of other covariates, women in the lower weight quartiles and those women losing 5% or more of their baseline weight had significantly more bone loss (Fig. 3A). As seen in Fig. 3A, women who gained 5% or more of their baseline weight lost less bone or had slight gains in BMD compared with women with less than a 5% change from their baseline weight. Weight considered as a continuous variable was associated with bone loss at all skeletal sites (e.g., for the femoral neck, the β-coefficient is 0.043 with a P value of 0.0009). Women who had baseline alcohol intakes of over 3 oz and over 7 oz had greater bone loss at the trochanter site than those women who had minimal alcohol intake (0 to <1 oz) (Table 3). Similar results were found only at the trochanter site when alcohol was considered as a continuous variable. Current estrogen replacement use appeared to be associated with less BMD loss in women, especially at the femoral neck site where women not using estrogen replacement lost nearly 3% more BMD over follow-up (Fig. 4) than women using estrogen. Surprisingly, caffeine use, physical activity, serum 25-OH vitamin D levels, or calcium intake did not affect bone loss. There were no differences in bone loss between women who were inactive, compared with “active” women using either of the surrogate inactivity measures of spending most of the day in bed or in a chair or most of the day indoors or between those women reporting good health compared with those women reporting fair or poor health. Results for Ward's area BMD and ultradistal radius BMD (not shown) were similar to BMDs displayed in tables.

Table Table 3.. Least Squares Mean Adjusted Percent BMD Change (±SE of Least Squared Mean) at Femoral Neck, Trochanter, Radial Shaft, and Lumbar Spine for Selected Risk Factors for Women
Risk factornFemoral neckTrochanterRadial shaftL2–L4 spine
  1. Models adjusted for age, weight, weight change, height, alcohol intake, current estrogen use, and smoking status; sample size for analysis may vary by site.

  2. * 0.05 < P < 0.10 or borderline.

  3. P < 0.05.

  4. P < 0.01.

  5. §P < 0.001. P < 0.0001.

Weight quartile (lbs)
 Q1 (91–124)117−4.21 ± 1.0−5.78 ± 1.2−5.35 ± 1.1−6.68 ± 1.4§
 Q2 (125–141)117−3.23 ± 0.9§−4.92 ± 1.2−5.58 ± 1.1*−4.51 ± 1.4
 Q3 (142–158)117−1.81 ± 1.0−2.68 ± 1.2−5.36 ± 1.1−1.34 ± 1.5
 Q4 (159–327) referent117−0.04 ± 1.1−0.13 ± 1.3−3.93 ± 1.2−1.73 ± 1.6
Weight change
 5% Loss114−4.25 ± 1.0‡−8.20 ± 1.2−5.85 ± 1.1−5.52 ± 1.4
 No change (referent)286−2.36 ± 0.8−3.10 ± 1.0−5.62 ± 0.9−3.01 ± 1.2
 5% Gain68−0.74 ± 1.1*+1.01 ± 1.4§−3.71 ± 1.2*−2.20 ±1.7
Alcohol intake
 0 to <1 oz (referent)232−2.39 ± 0.8−0.92 ± 1.0−4.21 ± 0.9−3.74 ± 1.2
 1–3 oz163−2.05 ± 0.9−1.54 ± 1.1−5.53 ± 1.0*−2.05 ± 1.3*
 >3-7 oz37−2.28 ±1.4−4.65 ± 1.7−6.18 ± 1.5−3.24 ± 1.9
 >7 oz36−3.09 ± 1.4−6.61 ± 1.7§−4.31 ± 1.5−5.28 ± 2.2
Cigarettes
 Never (referent)243−2.12 ± 0.9−4.39 ± 1.1−4.80 ± 0.9−3.33 ± 1.3
 Former176−2.68 ± 0.9−3.39 ± 1.1−3.53 ± 0.9*−4.71 ± 1.2
 Current49−2.56 ± 1.2−2.51 ± 1.5−6.85 ± 1.3−2.69 ± 1.9
Current estrogen
 No (referent)432−3.74 ± 0.6−4.52 ± 0.7−5.38 ± 0.6−4.86 ± 0.8
 Yes32−1.16 ± 1.4−2.33 ± 1.7−4.77 ± 1.5−2.30 ± 2.0

For men the multivariate adjusted mean percent BMD change across the 4 years of follow-up is shown in Table 4. Men also showed more bone loss in the lower weight quartiles compared with the highest weight quartile and in those men losing 5% or more of their baseline weight compared with men whose weight remained stable across follow-up, although these findings were only statistically significant at the trochanter site (Fig. 3B). Men who were current smokers lost more BMD at the trochanter site than men who never smoked (–4% change compared +0.7% change in BMD; P = 0.02). At the ultradistal radius site, men who reported poor health had greater bone loss than men reporting good health (P = 0.01), although no differences were seen at the other skeletal sites between men who reported good health compared with those men reporting poor health. Men who reported spending most of the day in bed or in a chair had greater bone loss compared with active men at the femoral neck (–5.79% compared with −2.23%, P = 0.0064) and at the trochanter (–3.77% compared with + 1.33%, P = 0.0028). However, those men reporting that they spent most of the day indoors had less bone loss compared with active men, but this was true only at the femoral neck site (–2.68% compared with −5.34% BMD change; P = 0.0123) and not statistically different at the other BMD sites for men. In men, bone loss also was not affected by caffeine use, physical activity, serum 25-OH vitamin D levels, or calcium intake.

thumbnail image

Figure FIG. 3.. (A) Percent of adjusted mean BMD change by weight change group for women. (B) Percent of adjusted mean BMD change by weight change group for men.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

This study presents longitudinal changes in BMDs at the femur, radial shaft, lumbar spine, and ultradistal radius in a population-based study of elderly men and women. Annualized mean bone loss percentages for women ranged from 0.86% to 1.12%, while for men they ranged from 0.04% to 0.90%. During the 4-year follow-up, age-specific mean percent loss in BMD among the skeletal sites ranged from 3% to 5% in women and was somewhat less in men (0.1–4%) across the age groups. Mean percent loss in BMD for women was much greater than the loss for men at all sites. Even given the higher baseline BMD in men, the absolute decline in BMD was greater for women than men. Elderly men continue to lose BMD at all ages but their BMDs remain higher than women's BMDs and their rates of loss at most sites were lower. Thus, our longitudinal results support continued BMD loss through the elderly years in both men and women.

To our knowledge, this study is the first to report variability in the distributions of bone loss (Fig. 1) in addition to mean bone loss. The substantial proportion of elderly women and men experiencing bone loss of 5–33% may offer future opportunity for prevention, while the smaller group with 5–20% gain in bone across the 4-year follow-up may suggest another area for future attention. We attempted to minimize technical errors in that all scans were reviewed and log book entries on patient positioning problems also were reviewed, such that no technical errors could be found with the study scans included in our analyses. These percent differences in bone represent values beyond the CV, indicating losses beyond possible technical measurement error. Nevertheless, because only two measures were used to create the percent change variable, there is a higher variance than if more than two measures were used and a greater likelihood of measurement error. It is possible that the extremes of BMD change would be minimized by having more measures on a subject (e.g., three measures across time or average of two measures at each exam); however, additional measurements are not available for these subjects. Future studies with larger numbers of subjects who lose or gain large amounts of BMD may offer opportunities to understand the factors that explain the extremes of bone changes and may elucidate the underlying pathophysiology and permit targeting of interventions for important subsets of individuals.

thumbnail image

Figure FIG. 4.. Percent of adjusted mean BMD change for women currently using estrogen-replacement therapy and not using estrogen-replacement therapy.

Download figure to PowerPoint

The bone loss rates at the femoral neck in our study are similar to those reported by Jones et al. in the Dubbo study and by Ensrud et al. for women at the femoral neck.(11,12) The Rotterdam study with 2-year femoral neck bone loss for elderly subjects also found a higher bone loss in women compared with bone loss in men.(18) However, their annualized femoral neck bone loss for women was slightly less than our study (–0.6% per year compared with −0.9% per year) and for men it was similar to our findings of −0.4% per year.

The age-specific baseline BMD levels reported in this paper for men and women are approximately 20% higher that those reported by Looker et. al using the National Health and Nutrition Examination Survey III data, although their age groupings were 60-69 years, 70-79 years, and 80+ years and included only the femur site.(39) Additionally, the NHANES data were collected using a Hologic scanner (Hologic, Inc., Waltham, MA, U.S.A.), which may explain some of the apparent discrepancy between our study and Looker's study.(40) Age-specific bone loss rates for women are similar to those reported for these older age groups at the radius by Ensrud et al. and for both women and men at the femoral neck as reported by Jones et al.(11,12) Burger et al. saw bone loss in the Rotterdam study up to age 80 years but no increased rate after age 80 years.(18) We continued to see bone loss in subjects 81-85 years old and 86-90 years old in our study, as did Ensrud et al. in the Study of Osteoporotic Fractures cohort ofwomen.(11)

Table Table 4.. Least Squared Mean Adjusted Percent BMD Change (±SE of Least Squared Mean) at Femoral Neck, Trochanter, Radial Shaft, and Lumbar Spine for Selected Risk Factors, for Men
Risk factornFemoral neckTrochanterRadial shaftL2–L4 spine
  1. Models adjusted for age, weight, weight change, height, alcohol intake, and smoking status; sample size for analysis may vary by site.

  2. * 0.05 < P < 0.10 or borderline.

  3. P 0.05.

  4. P < 0.01.

  5. §P < 0.001.

  6. P < 0.0001.

Weight quartile (lbs)
 Q1 (118–155)68−4.38 ± 1.0*−4.99 ± 1.3−3.13 ± 0.9−7.67 ± 1.1§
 Q2 (156–172)68−2.87 ± 1.1−1.94 ± 1.4‡−3.96 ± 0.9−5.60 ± 1.1
 Q3 (173–193)68−1.74 ± 1.1+0.91 ± 1.4−3.55 ± 1.0−2.23 ±1.2
 Q4 (194–294) referent69−1.93 ± 1.1+2.31 ± 1.5−2.37 ± 1.0−2.79 ±1.2
Weight change
 5% Loss59−4.34 ± 1.0−4.43 ± 1.3−3.43 ± 0.9−6.73 ± 1.1
 No change (referent)185−2.19 ± 0.7−1.10 ± 0.9−3.68 ± 0.7−4.17 ± 0.9
 5% Gain29−1.85 ± 1.4+2.62 ± 1.8−2.30 ± 1.2−3.43 ± 1.4
Alcohol intake
 0 to <1 oz (referent)94−2.68 ± 0.9−2.19 ± 1.2−1.92 ± 0.8−4.87 ± 0.8
 1–3 oz83−2.66 ± 1.0−0.13 ± 1.2−2.12 ± 0.9−3.28 ± 0.9
 >3-7 oz43−2.57 ± 1.3−0.68 ± 1.6−3.70 ± 1.1−4.46 ±1.7
 >7 oz53−3.27 ± 1.1−0.87 ± 1.4−4.79 ± 1.0−6.61 ± 2.0
Cigarettes
 Never (referent)91−1.77 ± 0.9+0.56 ± 1.1−3.91 ± 0.8−4.48 ± 1.0
 Former161−1.19 ± 0.7+0.91 ± 0.9−3.55 ± 0.6−5.87 ± 0.9
 Current21−5.42 ± 1.6−4.37 ± 2.1−1.94 ± 1.5−3.99 ± 1.6
General health
 Good (referent)237−3.26 ± 0.8−2.28 ± 1.0−2.65 ± 0.7−5.98 ± 1.0
 Poor32−4.76 ± 1.5−0.16 ± 2.0−4.66 ± 1.3−4.89 ± 1.7
Most of day in bed/chair
 No (referent)209−2.23 ± 0.9+1.33 ± 1.2−3.86 ± 0.9−4.76 ±1.1
 Yes45−5.79 ± 1.4‡−3.77 ± 1.8‡−3.44 ± 1.2−6.10 ± 1.5

This study reports bone loss at the femur, radial shaft, lumbar spine, and ultradistal radius sites, whereas the other longitudinal studies have focused solely on the femoral neck or radial shaft site. Indeed, few studies have attempted to measure BMD at multiple sites in their subjects. Previous work from the 1980s suggests that among elderly persons, BMD of the ultradistal radius site may remain relatively stable with age whereas bone densities of other sites, especially the calcaneous and femur, may show substantial losses in elderly persons with age.(41–43) We saw bone loss at all sites for both sexes with age. Nevertheless, our results also show a lack of concordance between skeletal sites and the risk factors we studied, possibly because of increased variability at some sites, notably the spine and ultradistal radius sites.

The lumbar spine is well known for presenting difficulties in measurement of elderly persons due to aortic calcifications, osteophytes, and other degenerative changes; yet we included our results for this site.(44–47) Despite the probable contributions to spinal BMD by these artifacts in elderly persons, we found bone loss in lumbar spine BMD for both men and women. The ultradistal radius site also presents difficulty in evaluating longitudinal change. This site often is difficult to measure precisely because of differences in positioning and errors with bone edge detection caused by low bone mass. Similar to our results, other studies have reported little age-related ultradistal radius bone loss in the elderly, and it is likely that there is minimal measurable bone loss in the ultradistal site, especially given the imprecision of this site.(48,49)

Body mass index, physical activity, alcohol, and calcium intake have been shown to affect BMD level.(27,28,50–56) The strongest relation in our study was the link between higher weight and less bone loss as well as the ability to maintain bone in those subjects with a 5% or greater gain in weight over the follow-up period. These results confirm findings from our cross-sectional study that reported that high body weight appeared to protect against low BMD.(50) Further, although in women weight loss had similar effects on bone loss at weight-bearing sites (femur) and nonweight-bearing sites (radius), weight gain appeared to have stronger associations with slowing bone loss at weight-bearing sites. Thus, a loading effect may be implicated as a possible mechanism to maintain bone in the elderly but it is not overwhelmingly clear. Nevertheless, these findings for weight and weight loss highlight the importance of weight on bone health.

We attempted to evaluate surrogate measures of inactivity and found no effect of these factors on change in BMD. Our measures of physical inactivity did not indicate a strong relation with 4-year bone loss; however, these surrogate measures may not sufficiently capture elderly activities and may have insufficient variation in activity across the study groups. Better measures of physical activity that are appropriate to an elderly age group may produce different results.

We found that women whose current alcohol consumption was 7 oz (207 ml) a week or more had greater bone loss at the trochanter than women in the lightest category of intake (less than 1 oz per week). This association was not seen at other BMD sites in women. In men, alcohol consumption at this level was not associated with bone loss compared with the lowest intake. Although other studies have suggested that moderate alcohol intake in women may result in higher BMDs, we did not find this in our study examining current alcohol intake.(55,57,58) In our previous cross-sectional study of Framingham participants we found that women with alcohol intake of 7 oz or more had higher BMDs; however alcohol intake was averaged for typical use over a 20-year period.(28) Other cross-sectional epidemiological studies have found little effect of alcohol intake on bone.(52,59) A longitudinal study by Hansen et al. in post-menopausal women comparing women who drank at least once a week with nondrinkers reported that the drinkers had lower rates of bone loss than nondrinkers.(57) Similarly, Burger et al. report a lower bone loss rate with increasing alcohol intake in men.(18) Our 4-year longitudinal results for alcohol may imply that short-term bone loss may occur in women drinking 7 oz or more of alcohol per week; however, our previous cross-sectional findings imply a positive cumulative effect of alcohol on bone for women. We were unable to confirm a protective effect of alcohol on bone loss in elderly women or men.

Our study found that men who were current smokers lost more BMD at the trochanter site than men who never smoked, but we did not observe any differences between female smokers and nonsmokers. Our previous cross-sectional study in the Framingham population also found cigarette smoking to be a strong risk factor for low BMD in men but not in women.(25) Burger showed higher bone loss in elderly males as well as in females who smoked compared with nonsmokers in the Rotterdam study.(18) The effect of smoking on bone loss in the men of our study, along with similar findings from the Rotterdam study and other cross-sectional studies, provides another rationale for smoking cessation, even among elderly persons.(18,25,60,61)

We found that women who currently used estrogen replacement tended to have less bone loss. A protective effect of estrogen replacement therapy on bone loss was not observed at all BMD sites; however, the lack of consistency may be a result of the low numbers of women (n = 32) who were current estrogen users in our longitudinal study. A protective effect of estrogen and serum estrogen levels on BMD have been reported in cross-sectional and longitudinal studies.(62,63) Prior studies of the Framingham cohort described lower BMD in women not using estrogen replacement therapy, compared with users.(29,64) The Study of Osteoporotic Fractures reported that current estrogen users had 33% lower rates of bone loss than nonusers, similar to the effects in our study, although the only statistically significant difference in our study was at the trochanter site.(11)

Surprisingly, 4-year bone loss in our study was not affected by caffeine, physical activity, serum 25-OH vitamin D, or calcium intake. Many epidemiological studies of calcium intake and BMD in elders do not show a large, if any, impact on bone health, implying that other risk factors may be of greater importance in this age group.(65,66) Lack of association may be a result of influences of other competing risk factors with larger effects on bone than these factors in our study or may relate to a relatively low frequency of exposure decreasing the statistical power to detect an association. Intermittent rather than continuous exposure also is possible for these factors. Further, these factors may only have effects over the long term rather than the 4-year bone loss evaluated in our study.

This study had several limitations. First, we had only two points in time to examine the longitudinal aspect of BMD loss in the cohort. Ideally, several points of BMD evaluation over time would have been preferred to stabilize the pattern of change in BMD. Second, cohort members who were unable to come to the clinic building did not participate in the BMD portion of the Framingham examinations because the Lunar densitometer was not mobile. Nonambulatory institutionalized elderly, often a frail group, were not included in our study, and they may have lower BMDs. Third, different technology assessed femoral and spine BMDs at baseline and at follow-up examinations, although femoral site data were “standardized” using published correction factors. Finally, there are very few nonwhite subjects in the Framingham sample, and our results are not generalizable to this group.

In conclusion, BMDs at the femur, radius, lumbar spine, and ultradistal radius continue to fall with age in elderly women and men in this population-based cohort. Elderly men continue to lose BMD at all ages but their BMDs remain higher than women's BMDs and their rates of loss are lower. Data were presented for each sex by 5-year age groups, thus providing information on men as well as the very elderly. Risk factors consistently associated with bone loss in elders include female sex, thinness, and weight loss of 5% or more, whereas gaining weight (5% or more) appears to protect against bone loss in both men and women. This population-based study suggests that current estrogen use may help to maintain bone in elderly women, while current smoking adversely affects BMD in men. To have an impact on bone loss and elderly fracture rates, risk factors of interest need to be potentially modifiable and of sufficient prevalence in the population, such that instituting a change may prevent bone loss at the population level. Even in the elderly years, potentially modifiable risk factors, such as weight, estrogen use, and cigarette smoking, exist that are important components of bone health.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

We are grateful to the Framingham cohort participants and staff and also thank the densitometer technicians Mimi Brodsky, Mary Hogan, and Cherlyn Mercier. Components of this work were presented in part as concurrent sessions at the 16th and 17th Annual Scientific Meetings of the ASBMR in Kansas City, MO, U.S.A. and in Seattle, WA, U.S.A. as well as the Association of Rheumatology Health Professionals 32nd Annual Scientific meeting in Washington, D.C., U.S.A. This work was supported in part by the National Institutes of Health (NIH) grant RO1-AR/AG 41398, NIH grant RO1-AR20613, and by NIH/NHLBI contract N01–38038.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  • 1
    Cummings SR, Kelsey JL, Nevitt M, O'Dowd KJ 1985 Epidemiology of osteoporosis and osteoporotic fractures. Epidemiol Rev 7: 178208.
  • 2
    Ross PD, Wasnich RD, Davis JW 1990 Fracture prediction models for osteoporosis prevention. Bone 11: 327331.
  • 3
    Black DM, Cummings SR, Genant HK, Nevitt MC, Palermo L, Browner WS 1992 Axial bone density predicts fractures in older women. J Bone Miner Res 7: 633638.
  • 4
    Cummings SR, Black DM, Nevitt MC, Browner W, Cauley J, Ensrud K, Genant HK, Palermo L, Scott J, Vogt TM 1993 Bone density of various sites for prediction of hip fractures. Lancet 341: 7275.
  • 5
    Hui SL, Slemenda CW, Johnston CC Jr 1988 Age and bone mass as predictors of fracture in a prospective study. J Clin Invest 81: 18041809.
  • 6
    Ross PD, Davis JW, Vogel JM, Wasnich RD 1990 A critical review of bone mass and the risk of fractures in osteoporosis. Calcif Tissue Int 46: 14961.
  • 7
    Melton LJ III, Atkinson EJ, O'Fallon WM, Wahner HW, Riggs BL 1993 Long-term fracture prediction by bone mineral assessed at different skeletal sites. J Bone Miner Res 8: 12271233.
  • 8
    Kanis JA and the WHO Study Group 1994 Assessment of fracture risk and its application to screening for postmenopausal osteoporosis: Synopsis of a WHO report. Osteoporos Int 4: 368381.
  • 9
    Looker AC, Johnston CC Jr, Wahner HW, Dunn WL, Calvo MS, Harris TB 1995 Prevalence of low femoral bone density in older U.S. women from NHANES III. J Bone Miner Res 10: 17.
  • 10
    Hannan MT, Felson DT, Anderson JJ 1992 Bone mineral density in elderly men and women: Results from the Framingham osteoporosis study. J Bone Miner Res 7: 547553.
  • 11
    Ensrud KE, Palermo L, Black DM, Cauley J, Jergas M, Orwoll ES, Nevitt MC, Fox KM, Cummings SR 1995 Hip and calcaneal bone loss increases with advancing age. Longitudinal results from the study of osteoporotic fractures. J Bone Miner Res 10: 17781787.
  • 12
    Jones G, Nguyen T, Sambrook P, Kelly PJ, Eisman JA 1994 Progressive loss of bone in the femoral neck in elderly people: Longitudinal findings from the Dubbo osteoporosis epidemiology study. BMJ 309: 691695.
  • 13
    Cali CM, Kiel DP 1995 Age-related bone loss in the “old” old: A longitudinal study of institutionalized elderly. (Abstract) J Bone Miner Res 10(Suppl 1): S467.
  • 14
    Greenspan SL, Maitland LA, Myers ER, Krasnow MB, Kido TH 1994 Femoral bone loss progresses with age: A longitudinal study in women over age 65. J Bone Miner Res 9: 19591965.
  • 15
    Burger H. Van Daele PL, Algra D, van den Ouweland FA, Grobbee DE, Hofman A, van Kuijk C, Schutte HE, Birkenhager JC, Pols HA 1994 The association between age and bone mineral density in men and women aged 55 years and over: The Rotterdam study. Bone Miner 25: 113.
  • 16
    Sowers MF, Clark K, Wallace R, Jannausch M, Lemke J 1991 Prospective study of radial bone mineral density in a geographically defined population of postmenopausal Caucasian women. Calcif Tissue Int 48: 232239.
  • 17
    Steiger P, Cummings SR, Black DM, et al 1992 Age-related decrements in bone mineral density in women over 65. J Bone Miner Res 7: 625632.
  • 18
    Burger H, de Laet C, van Daele P, Weel A, Witteman J, Hofman A, Pols H 1998 Risk factors of increased bone loss in an elderly population: The Rotterdam study. Am J Epidemiol 147: 871879.
  • 19
    Dawber TR, Meadors GF, Moore FE Jr 1951 Epidemiological approaches to heart disease: The Framingham study. Am J Public Health 41: 279286.
  • 20
    Dawber TR 1980 The Framingham Study. Harvard University Press, Cambridge, MA, U.S.A.
  • 21
    Shurtleff D 1970 The Framingham Study: An Epidemiological Investigation of Cardiovascular Disease. KannelWB, GordonT (eds.) U.S. Government Printing Office, Washington D.C., U.S.A.
  • 22
    Kiel DP, Mercier CA, Dawson-Hughes B, Cali C, Hannan MT, Anderson JJ 1995 Effects of scan analysis technique and analytic software on the comparison of dual-x-ray absorptiometry with dual photon absorptiometry of the hip in the elderly. J Bone Miner Res 10: 11301136.
  • 23
    Kiel DP, Felson DT, Hannan MT, Anderson JJ, Wilson PWF 1990 Caffeine and the risk of hip fracture: The Framingham study. Am J Epidemiol 132: 675684.
  • 24
    Cummings SR, Nevitt MC, Browner WS, Stone K, Fox KM, Ensrud KE, Cauley J, Black D, Vogt TM 1995 Risk factors for hip fracture in white women. New Engl J Med 332: 767773.
  • 25
    Kiel DP, Zhang Y, Hannan MT, Anderson JJ, Felson DT 1996 The effect of smoking at different life stages on bone mineral density in elderly men and women. Osteoporos Int 6: 240248.
  • 26
    Harris SS, Dawson-Hughes B 1994 Caffeine and bone loss in healthy postmenopausal women. Am J Clin Nutr 60: 573578.
  • 27
    Felson DT, Kiel DP, Anderson JJ, Kannel WB 1988 Alcohol consumption and hip fracture: The Framingham study. Am J Epidemiol 128: 11021110.
  • 28
    Felson DT, Zhang Y, Hannan MT, Kannel WM, Kiel DP 1995 Alcohol intake and bone mineral density in elderly men and women: The Framingham study. Am J Epidemiol 142: 485492.
  • 29
    Felson DT, Zhang Y, Hannan MT, Kiel DP, Wilson PWF, Anderson JJ 1993 The effect of postmenopausal estrogen therapy on bone density in elderly women. N Engl J Med 329: 11411146.
  • 30
    Hannan MT, Felson DT, Anderson JJ, Naimark A 1993 Habitual physical activity and knee osteoarthritis in the Framingham study. J Rheumatol 20: 704709.
  • 31
    Kannel WB, Sorlie PD 1979 Some health benefits of physical activity: The Framingham study. Arch Intern Med 139: 857861.
  • 32
    Kannel WB, Wilson P, Blair SN 1985 Epidemiological assessment of the role of physical activity and fitness in development of cardiovascular disease. Am Heart J 109: 876885.
  • 33
    Kannel WB 1967 Habitual level of physical activity and risk of coronary heart disease: The Framingham study. Can Med Assoc J 96: 811812.
  • 34
    Preece MA, O'Riordan JLH, Lawson DEM, Kodicek E 1974 A Competitive protein-binding assay for 25-hydroxycholecalciferol and 25-hydroxyergocalciferol in serum. Clin Chim Acta 54: 235242.
  • 35
    Kiel DP, Myers RH, Kong XF, Zhu XH, Ordovas J, Felson DT, Rush D, Wilson PWF, Eisman JA, Holick MF 1997 The BsmI vitamin D receptor restriction fragment length polymorphism (bb) is associated with higher bone density in younger, but not older persons, and influences the effect of calcium intake on bone mineral density. J Bone Miner Res 12: 10491057.
  • 36
    Dawson-Hughes B, Dallal GE, Krall EA, Harris S, Sokoll LJ, Falconer G 1991 Effect of vitamin D supplementation on wintertime and overall bone loss in healthy postmenopausal women. Ann Intern Med 115: 505512.
  • 37
    Willett WC, Sampson L, Stampfer MJ, Rosner B, Bain C, Witschi J, Hennekens CH, Speizer FE 1985 Reproducibility and validity of a semi-quantitative food frequency questionnaire. Am J Epidemiol 122: 5165.
  • 38
    Jacques PF, Sulsky SI, Sadowski JA, Phillips JC, Rush D, Willett WC 1993 Comparison of micronutrient intake measured by a dietary questionnaire and biochemical indicators of micronutrient status. Am J Clin Nutr 57: 182189.
  • 39
    Looker AC, Wahner HW, Dunn WL, Calvo MS, Harris TB, Heyes SP, Johnston CC Jr, Lindsey RL 1995 Proximal femur bone mineral levels of U.S. adults. Osteoporos Int 5: 389409.
  • 40
    Genant HK, Grampp S, Gluer CC, Faulkner KG, Jergas M, Engelke K, Hagiwara S, Van Kuijk C 1994 Universal standardization for dual x-ray absorptiometry: Patient and phantom cross-calibration results. J Bone Miner Res 9: 15031514.
  • 41
    Davis JW, Ross PD, Wasnich RD, Maclean CJ, Vogel JM 1989 Comparison of cross-sectional and longitudinal measurements of age-related changes in bone mineral content. J Bone Miner Res 4: 351357.
  • 42
    Mazess RB, Barden HS, Ettinger M, Johnston C, Dawson-Hughes B, Baran D, Powell M, Notelovitz M 1987 Spine and femur density using dual-photon absorptiometry in U.S. white women. Bone Miner 2: 211219.
  • 43
    Ribot C, Tremolliers F, Pouilles JM, Louvet JP, Guiraud R 1988 Influence of the menopause and aging on spinal density in French women. Bone Miner 5: 8997.
  • 44
    Riggs BL, Wahner HW, Seeman E, Offord KP, Dunn WL, Mazess RB, Johnson KA, Melton LJ III 1982 Changes in bone mineral density of the proximal femur and spine with aging: Differences between the postmenopausal and senile osteoporosis syndromes. J Clin Invest 70: 716723.
  • 45
    Hopkins A, Zylstra S, Hreshchyshyn M 1988 Normal and abnormal features of the lumbar spine observed in dual-photon absorptiometry scans. Clin Nucl Med 14: 410414.
  • 46
    Stutzman ME, Yester MV, Duborsky EV 1987 Technical aspects of dual photon absorptiometry of the spine. J Nucl Med Tech 15: 177181.
  • 47
    Dawson-Hughes B, Dallal GE 1990 Effect of radiographic abnormalities on rate of bone loss from the spine. Calcif Tissue Int 46: 280281.
  • 48
    Hui SL, Wiske PS, Norton JA, Johnston Jr CC 1982 A prospective study of change in bone mass with age in postmenopausal women. J Chron Dis 35: 715725.
  • 49
    Eastell R, Riggs BL, Wahner HW, O'Fallon WM, Amadio PC, Melton LJ III 1989 Colles' fracture and bone density of the ultradistal radius. J Bone Miner Res 4: 607613.
  • 50
    Felson DT, Zhang Y, Hannan MT, Anderson JJ 1993 Effects of weight and body mass index on bone mineral density in men and women: The Framingham study. J Bone Miner Res 8: 567573.
  • 51
    Edelstein Sl, Barrett-Connor E 1993 Relation between body size and bone mineral density in elderly men and women. Am J Epidemiol 138: 160169.
  • 52
    Bauer DC, Browner WS, Cauley JA, Orwoll ES, Scott JC, Black DM, Tao JL, Cummings SR 1993 Factors associated with appendicular bone mass in older women. Ann Intern Med 118: 657665.
  • 53
    Krall EA, Dawson-Hughes B 1994 Walking is related to bone density and rates of bone loss. Am J Med 96: 2026.
  • 54
    Greendale GA, Barrett-Connor E, Edelstein SL, Ingles S, Haile R 1995 Lifetime leisure exercise and osteoporosis. Am J Epidemiol 141: 951959.
  • 55
    Holbrook T, Barrett-Conner E 1993 A prospective study of alcohol consumption and bone mineral density. BMJ 306: 150169.
  • 56
    Dawson-Hughes B, Dallal GE, Krall EA, Sadowski L, Sahyoun N, Tannenbaum S 1990 A controlled trial of the effect of calcium supplementation on bone density in postmenopausal women. N Engl J Med 323: 878883.
  • 57
    Hansen MA, Overgaard K, Riis BJ, Christiansen C 1991 Potential risk factors for development of postmenopausal osteoporosis—examined over a 12-year period. Osteoporos Int 1: 95102.
  • 58
    Laitinen K, Valimaki M 1991 Alcohol and bone. Calcif Tissue Int 49(Suppl): S70S73.
  • 59
    Elders PJM, Netelenbos JC, Lips P, Khoe E, van Ginkel FC, Hulshof KF, van der Stelt PF 1989 Perimenopausal bone mass and risk factors. Bone Miner 7: 289299.
  • 60
    Hollenbach KA, Barrett-Connor E, Edelstein SL, Holbrook T 1993 Cigarette smoking and bone mineral density in older men and women. Am J Public Health 83: 12651270.
  • 61
    Hopper JL, Seeman E 1994 The bone density of female twins discordant for tobacco use. N Engl J Med 330: 387392.
  • 62
    Cauley JA, Gutai JP, Sandler RB, LaPorte RE, Kuller LH, Sashin D 1986 The relationship of endogenous estrogen to bone density and bone area in normal postmenopausal women. Am J Epidemiol 124: 752761.
  • 63
    Cauley JA, Gutai JP, Kuller LH, Scott J, Nevitt MC 1994 Black-White differences in serum sex hormones and bone mineral density. Am J Epidemiol 139: 10351046.
  • 64
    Krall EA, Dawson-Hughes B, Hannan MT, Kiel DP 1997 Postmenopausal estrogen replacement therapy protects against tooth loss. Am J Med 102: 536542.
  • 65
    Kiel DP 1995 Vitamin D, calciumbone: Descriptive epidemiology. In: RosenbergRH (ed.) Nutritional Assessment of Elderly Populations, vol 13. Raven Press, New York, NY, U.S.A., pp. 277290.
  • 66
    Paganini-Hill A, Chao A, Ross RK, Henderson BE 1991 Exercise and other factors in the prevention of hip fracture: The leisure world study. Epidemiology 2: 1625.