Voluntary Weight Reduction in Older Men Increases Hip Bone Loss: The Osteoporotic Fractures in Men Study

Abstract

To test the hypothesis that weight loss in older men is associated with increased rates of hip bone loss regardless of adiposity and intention to lose weight, we measured body weight, body composition, hip bone mineral density (BMD), and intention to lose weight in a cohort of 1342 older men enrolled in the Osteoporotic Fractures in Men (MrOS) study and followed them prospectively for an average of 1.8 yr for changes in weight and BMD. The adjusted average rate of change in total hip BMD was 0.1%/yr in men with weight gain, −0.3%/yr in men with stable weight, and −1.4%/yr in men with weight loss (test for trend, P < 0.001). Higher rates of hip bone loss were observed in men with weight loss regardless of category of body mass index, body composition, or intention to lose weight. Even among obese (body mass index, ≥30 kg/m²) men trying to lose weight, those with documented voluntary weight reduction experienced an increase in hip bone loss (average rate of change in total hip BMD, 0.5%/yr in those with weight gain, −0.1%/yr in those with stable weight, and −1.7%/yr in those with weight loss; test for trend, P < 0.001). Older men who experience weight loss have increased rates of hip bone loss, even among overweight and obese men undergoing voluntary weight reduction.

BODY WEIGHT IS strongly associated with bone density at multiple skeletal sites in older men and women. Several cross-sectional studies (1–4) have reported that thinner men have lower bone density, especially at weight-bearing sites such as the hip and spine. In addition, prior epidemiological studies (5–7) that measured changes in weight and bone density in older men have reported higher rates of hip bone loss in men losing weight. However, it is uncertain whether older overweight and obese men undergoing voluntary weight reduction lose bone.

To test the hypothesis that weight loss is associated with an increase in the rate of hip bone loss in older men regardless of adiposity and intention to lose weight, we measured body weight, body composition, hip bone density, and intention to lose weight in a cohort of 1342 older men enrolled in the Osteoporotic Fractures in Men (MrOS) study and followed them prospectively for changes in weight and bone density.

Subjects and Methods

Participants

From March 2000 through April 2002, 5995 men at least 65 yr of age were recruited for participation in the baseline examination of the prospective MrOS study. (8) Men were recruited from population-based listings in six areas of the United States: Birmingham, AL; the Monongahela Valley near Pittsburgh, PA; Minneapolis, MN; Palo Alto, CA; San Diego, CA; and Portland, OR. (9) We excluded men with a history of bilateral hip replacement and those who were unable to walk without the assistance of another person.

From September 2002 through May 2003, men enrolled at the Birmingham, AL, center and those enrolled at the Portland, OR, center were invited to participate in a MrOS ancillary study to identify determinants of periodontal disease in older men. Of the 1976 men enrolled at the baseline MrOS examination at these two centers, 1353 (68%) attended the dental (second) examination an average of 1.8 yr later. Of these, 1342 men (99%) had technically adequate measurements of hip BMD at the baseline and second examinations and are the subject of this analysis. The institutional review board at each center approved the study protocol, and written informed consent was obtained from all subjects.

Weight change

Body weight (in indoor clothing with shoes removed) was recorded with a scale (balance beam scale at Birmingham center and digital scale at the Portland center) at both the baseline and second examinations (mean ± sd, 1.8 ± 0.4 yr between examinations). The scale at each clinical center was calibrated every month during both examinations. Weight change was calculated by subtracting weight at the baseline examination from weight at the second examination and was expressed as a percentage of the baseline value.

Measurement of bone mineral density (BMD)

BMD and bone mineral content of the total hip and three subregions was measured at the baseline and second examinations (mean ± sd, 1.8 ± 0.4 yr between examinations) using fan beam dual energy x-ray absorptiometry with Hologic QDR-4500W scanners (Hologic, Inc., Bedford, MA). Standardized procedures for participant positioning and scan analysis were executed for all scans. All measurements were performed on the right hip unless a participant reported a right hip replacement or metal objects in the right leg, in which case the measurement was performed on the left hip. All densitometry operators at the centers were certified on the basis of an evaluation of scanning and analysis techniques. To assure adherence to standardized protocol, densitometry technicians at the coordinating center (University of California, San Francisco) reviewed a random sample of all scans, scans with exceptionally high or low BMD, and problematic scans flagged at the clinic. Cross-calibration studies performed before the baseline MrOS examination found no linear differences across scanners, and the interscanner coefficient of variation was 0.9%. Longitudinal quality control using daily scan data for standardized phantoms indicated no shifts or drifts in scanner calibration. The change in bone density was expressed as the percent change in BMD per year, an annualized percentage of the initial value.

Other measurements

Participants completed a questionnaire and were interviewed at the baseline examination by trained and certified clinical staff, who asked about self-reported health, smoking status, and alcohol use. Physical activity was assessed using the Physical Activity Scale for the Elderly (PASE) (10). A selected medical history was obtained, including a history of physician diagnosis of stroke, diabetes, hyperthyroidism, hypothyroidism, Parkinson’s disease, coronary heart disease, congestive heart failure, chronic obstructive lung disease, and nonskin cancer. Dietary calcium intake from foods and supplements was estimated by using a modified Block semiquantitative food frequency questionnaire developed specifically for MrOS by Block Dietary Systems (Berkeley, CA). (11) Total calcium intake was calculated by summing dietary calcium intake (milligrams per day) and daily dosage of calcium supplements (milligrams per day). Intention to lose weight (whether or not the participant was trying to lose weight during the past 12 months) was assessed by completion of a mailed follow-up questionnaire an average (±sd) of 2.0 ± 0.4 yr after the baseline examination.

Height was measured using a standard held-expiration technique with a wall-mounted Harpenden stadiometer (Holtain, Pyred, UK). Height and weight were used to calculate the standard body mass index, a measure of weight in kilograms divided by the square of height in meters. Isometric quadriceps maximum power was measured using the Nottingham Power Rig (12) and was scored as the maximum in up to nine trials for the strongest leg, in watts.

Statistical analysis

We categorized percent weight change between the baseline and second examinations (average of 1.8 yr between examinations) on the basis of current beliefs about clinically relevant changes in weight in older adults and the availability of sufficient numbers of participants in each category. Weight loss was defined as a decrease of 5% or more since the baseline examination, stable weight was defined as less than a 5% change from baseline weight, and weight gain was defined as an increase of 5% or more since the baseline examination. χ² tests of homogeneity, analyses of variance, and Kruskal-Wallis tests were used to compare characteristics at the baseline examination by category of weight change.

To examine the association between weight change and rate of change in BMD at the total hip, the adjusted mean of the annualized change in total hip BMD and its 95% confidence interval were calculated by category of weight change. A P value for linear trend in annualized percent change in total hip BMD between weight change groups was calculated. Initial analyses were adjusted only for age. Factors related to percent weight change and/or rate of change in total hip BMD at P ≤ 0.20 were included in multivariate models as potential confounders of the association between weight change and rate of change in hip BMD. Because results from multivariate models did not differ from those of age-adjusted models, findings from multivariate analyses are presented.

We stratified participants by characteristics measured at the baseline examination, including body mass index (underweight or normal weight, body mass index <25 kg/m²; overweight, body mass index between 25 and 29.9 kg/m²; obese, body mass index ≥30 kg/m²); lean mass (median, <57.3 vs. ≥57.3 kg/m²), fat mass (median, <20.6 vs. ≥ 20.6 kg/m²), age (<75 vs. ≥ 75 yr), total calcium intake (median, <962 vs. ≥962 mg/d), physical activity level (median PASE score, <143 vs. ≥ 143), total hip BMD (median, <0.950 g/cm²vs. ≥0.950 g/m²), and intention to lose weight (trying to lose weight vs. not trying to lose weight). We tested for interactions between weight change and these variables. To examine whether voluntary weight reduction in obese elderly men was associated with increased rates of hip bone loss, we stratified our analyses by both body mass index and intention to lose weight.

Additional analyses were performed with substitution of absolute weight change (expressed as a continuous variable in kilograms) for percent weight change and expression of percent weight change as a continuous variable. We also repeated our analyses using percent rate of change in bone mineral content as the outcome variable. Finally, we performed analyses using rate of change in bone density at three subregions of the hip (femoral neck, intertrochanteric region, and trochanter) as our outcome measure. Because the findings from these secondary analyses were similar to those of our primary analyses, we only present the primary results.

To determine the distribution of T-scores at the hip at baseline and second examinations within category of baseline body mass index in men with weight loss, we calculated male specific T-scores using reference values for males 20–29 yr of age (13).

Results

Characteristics of the study population

Of the 1976 men who had measurement of hip bone density at the baseline examination, 1342 (68%) were enrolled in the dental cohort and had measurements of hip bone density at both the baseline and second (dental) examinations. On the average, the 1342 men with initial and follow-up hip bone density measurements were slightly younger (72.8 vs. 73.9 yr; P < 0.001), more likely to report good to excellent health (87% vs. 75%; P < 0.001), and had slightly higher average total hip bone density (0.957 vs. 0.945 g/cm²; P = 0.096) compared with the 634 men who completed hip bone density measurement at the baseline examination only, although body mass index did not differ between the two groups (27.4 vs. 27.3 kg/m²; P = 0.758).

The characteristics of the 1342 participants at baseline according to category of percent weight change between the baseline and the second examination (average, 1.8 yr later) are shown in Table 1. Compared with the 1087 men (81%) with stable weight and the 115 men (9%) with weight gain, the 140 men (10%) with weight loss were more likely to report poorer health status and tended to have less leg power. Among the 510 participants who reported trying to lose weight during the past 12 months (on the follow-up questionnaire, an average of 2.0 yr after the baseline examination), the average measured weight change between the baseline and subsequent examination was +0.1 kg, whereas among the 822 participants who reported not trying to lose weight on the questionnaire, the average change in weight was 0.0 kg between examinations. The proportion of men reporting trying to lose weight did not differ by category of measured weight change.

TABLE 1.

Characteristics of 1342 participants by category of percent weight change

Variable	Category of % weight change			P value
	Weight loss (≤−5.0%; n = 140)	Stable weight (−4.9% to 4.9%; n = 1087)	Weight gain (≥5.0%; n = 115)
Mean age ± sd (yr)	73.4 (5.6)	72.8 (5.5)	72.0 (5.1)	0.127
Caucasian race (%)	91	89	89	0.907
Self-reported health status (%)
Excellent or good	77	89	82	<0.001
Fair/Poor or very poor	23	11	18
Mean PASE score ± sd	149 (79)	149 (69)	153 (71)	0.873
Current smoker (%)	6	3	6	0.128
Mean alcoholic drinks/wk	4.0 (8.3)	4.1 (6.6)	3.4 (5.8)	0.118
Mean total calcium intake (mg/d)	1105 (562)	1108 (573)	1162 (614)	0.636
Any fracture since age 50 y (%)	23	25	26	0.825
≥1 selected medical conditions (%)1	58	49	51	0.122
Weight loss intention (%)
Trying to lose weight	40	37	44	0.323
Not trying to lose weight	60	63	56
Mean body mass index ± sd (kg/m2)	27.6 (4.3)	27.4 (3.6)	26.7 (3.9)	0.141
Mean lean mass ± sd (kg)	57.0 (7.2)	57.6 (7.2)	57.3 (7.4)	0.602
Mean fat mass ± sd (kg)	22.3 (7.8)	21.5 (6.9)	21.3 (7.2)	0.449
Mean % weight change ± sd (%)	−7.7 (2.7)	0.3 (2.4)	7.5 (3.0)	<0.001
Mean absolute weight change ± sd (kg)	−6.6 (3.0)	0.2 (2.0)	6.2 (2.5)	<0.001
Mean maximum Nottingham leg power (watts)	205 (57)	219 (60)	209 (59)	0.014
Mean total hip bone density ± sd (g/cm2)	0.93 (0.15)	0.96 (0.14)	0.96 (0.14)	0.126

All characteristics measured at the baseline examination; change in weight calculated between baseline and second exam (average 1.8 yr between examinations). Intention to lose weight assessed by follow-up questionnaire.

¹

History of one or more selected medical conditions, including stroke, diabetes, hyperthyroidism, hypothyroidism, Parkinson’s disease, coronary heart disease, congestive heart failure, chronic obstructive lung disease, and nonskin cancer.

At baseline, only one participant was classified as underweight (body mass index, <18.5 kg/m²), 360 were classified as normal weight (body mass index, 18.5–24.9 kg/m²), 703 were classified as overweight (body mass index, 25–29.9 kg/m²), and 276 were classified as obese (body mass index, ≥30 kg/m²). Among the 53 overweight and obese men who reported trying to lose weight on the questionnaire and who had a documented decrease of 5% or more in body weight between examinations, 26% used dieting alone, 8% used exercise alone, and 60% used a combination of diet and exercise. During an average follow-up of 1.2 yr after completion of the second examination, 22 men died, including three of the 140 men (2.1%) with weight loss between examinations, 17 of the 1087 men (1.6%) with stable weight, and two of the 115 men (1.7%) with weight gain.

Weight change and rate of hip bone loss

Men who lost weight had the greatest average rate of bone loss at the hip, even after adjustment for multiple potential confounders at the baseline examination, including age, health status, physical activity level, smoking status, alcohol use, total calcium intake, medical conditions, body mass index, lean mass, leg power, and total hip BMD (Table 2). The adjusted average rate of change in total hip bone density was 0.1%/yr in men with weight gain, −0.3%/yr in men with stable weight, and −1.4%/yr in men with weight loss (test for trend, P < 0.001).

TABLE 2.

Mean annualized rate of change in hip BMD by category of percent weight change and body mass index (BMI)

Category of weight change	Mean annualized % change in total hip bone mineral density (95% confidence interval)
	Overall cohort1 (n = 1213)	BMI2
		BMI <25.0 kg/m2 (n = 330)	BMI 25.0–29.9 kg/m2 (n = 637)	BMI >30 kg/m2 (n = 246)
Weight loss (n = 124)	−1.4 (−1.6, −1.2)	−1.4 (−1.9, −0.9)	−1.2 (−1.5, −0.9)	−1.7 (−2.2, −1.2)
Stable weight (n = 980)	−0.3 (−0.4, −0.3)	−0.4 (−0.5, −0.2)	−0.4 (−0.5, −0.3)	−0.3 (−0.5, 0.0)
Weight gain (n = 109)	0.1 (−0.1, 0.4)	0.2 (−0.3, 0.7)	0.2 (−0.2, 0.5)	−0.2 (−0.8, 0.5)

¹

Adjusted for age; health status; physical activity level; smoking status; alcohol use; total calcium intake; history of one or more select medical conditions, including stroke, diabetes, hyperthyroidism, hypothyroidism, Parkinson’s disease, coronary heart disease, congestive heart failure, chronic obstructive lung disease, and nonskin cancer; BMI, lean mass; leg power; and total hip bone density (base model). A total of 1213 men are included in the multivariate model for the overall cohort; 129 men are excluded due to missing data on a covariate (108 missing data on leg power).

²

Base model minus BMI.

Higher rates of hip bone loss were observed in men with weight loss regardless of body mass index (Table 2). Among underweight or normal weight (body mass index <25 kg/m²), overweight (body mass index 25–29.9 kg/m²), and obese (body mass index ≥30 kg/m²) men, the tests for trend were significant (P < 0.001 within each subgroup). There was no evidence of an interaction between weight change and body mass index for the prediction of rate of change in total hip bone density (P = 0.952). Results were similar when lean or fat mass was substituted for body mass index in the analyses (results not shown).

In addition, higher rates of hip bone loss were observed in men with documented weight loss regardless of whether they were trying to lose weight (Fig. 1). Among men trying to lose weight (P < 0.001) and those not trying to lose weight (P < 0.001), the tests for trend were significant. There was no evidence of an interaction between weight change and intention to lose weight for the prediction of rate of change in total hip bone density (P = 0.213).

Fig. 1.

Mean annualized rate of change in total hip BMD by category of percent weight change and intention to lose weight.

In analyses stratified by both body mass index and intention to lose weight, documented weight loss was associated with increased rates of hip bone loss among underweight to normal weight, overweight, and obese men regardless of whether the loss was intentional or not (Table 3). In general, regardless of category of body mass index or intention to lose weight, men with weight gain had a slight increase in BMD, those with stable weight had a small decrease in BMD, and those with weight loss had rates of bone loss exceeding 1%/yr. For example, among overweight men trying to lose weight, the average rate of change in total hip bone density was 0.2%/yr in those with weight gain, −0.5%/yr in those with stable weight, and −1.1%/yr in those with weight loss (for trend, P < 0.001). The only exception to this pattern was among obese men not trying to lose weight, in whom rates of hip bone loss were similarly increased in men with weight loss and those with weight gain.

TABLE 3.

Association between weight change and rate of change in hip BMD stratified by body mass index (BMI) and intention to lose weight

	Mean annualized % change in total hip BMD (95% confidence interval)1
	Trying to lose weight	Not trying to lose weight
BMI <25.0 kg/m2
Weight loss	−2.6 (−3.9, −1.3)	−1.2 (−1.7, −0.7)
Stable weight	−0.5 (−0.8, −0.2)	−0.3 (−0.5, −0.1)
Weight gain	1.3 (−0.8, 3.4)	0.1 (−0.4, 0.6)
BMI 25.0–29.9 kg/m2
Weight loss	−1.1 (−1.5, −0.7)	−1.4 (−1.9, −1.0)
Stable weight	−0.2 (−0.3, −0.1)	−0.5 (−0.6, −0.3)
Weight gain	0.2 (−0.2, 0.6)	0.2 (−0.4, 0.7)
BMI ≥30 kg/m2
Weight loss	−1.7 (−2.4, −1.1)	−1.4 (−2.3, −0.5)
Stable weight	−0.1 (−0.4, 0.1)	−0.5 (−0.8, −0.2)
Weight gain	0.5 (−0.3, 1.3)	−1.7 (−2.4, −0.5)

Among the 328 men with BMI <25 kg/m, there were 45 trying to lose weight and 283 not trying to lose weight. Among the 635 men with BMI 25–29.9 kg/m, there were 274 trying to lose weight and 361 not trying to lose weight. Among the 242 men with BMI ≥30 kg/m, there were 151 trying to lose weight and 91 not trying to lose weight.

¹

Adjusted for age; health status; physical activity level; smoking status; alcohol use; total calcium intake; history of one or more select medical conditions, including stroke, diabetes, hyperthyroidism, hypothyroidism, Parkinson’s disease, coronary heart disease, congestive heart failure, chronic obstructive lung disease, and nonskin cancer; BMI, lean mass; leg power; and total hip bone density.

Higher rates of bone loss at the hip were also consistently observed in men with weight loss in additional analyses stratified by age (<75 vs. ≥75 yr), total calcium intake (median, <962 vs. ≥ 962 mg/d), physical activity level (median PASE score, <143 vs. ≥143), and total hip BMD (median, <0.950 g/cm²vs. ≥0.950 g/m²). Within each of these subgroups, the test for trend reached significance (test for trend, P < 0.001).

Among the 140 men with documented weight loss, the distribution of male specific T-scores at the hip at baseline and second examinations within a given category of baseline body mass index is shown in Table 4. Among these men with weight loss, a greater proportion of underweight to normal weight men than of overweight or obese men had T-scores between −1.1 and −2.4 and at or below −2.5 at each examination. Among overweight and obese men who subsequently lost weight between the baseline and second examinations, the proportion with T-scores −1.0 or greater decreased, whereas the proportion with T-scores between −1.1 and −2.4 increased.

TABLE 4.

Distribution of T-scores at the hip at baseline and second examinations within category of body mass index (BMI) at baseline in men with weight loss

Total hip BMD1	BMI at baseline in men with weight loss
	Underweight to normal weight (≤25 kg/m2; n = 40)	Overweight (25.0–29.9 kg/m2; n = 65)	Obese (≥30 kg/m2; n = 35)
Male specific T-score ≥−1.0 (%)
Baseline	23	57	89
Second examination	23	49	83
Male specific T-score −2.4 to −1.1 (%)
Baseline	68	42	11
Second examination	65	49	17
Male specific T-score ≤−2.5 (%)
Baseline	10	2	0
Second examination	13	2	0

¹

T-Score calculated using reference values for males aged 20–29 yr (13).

Discussion

We found that older men with weight loss have higher rates of hip bone loss, even among overweight and obese men trying to lose weight. These results indicate that sustained modest weight loss in older men is a strong clinical risk factor for hip bone loss. In addition, our findings suggest a need for randomized trials in overweight people to evaluate the effectiveness of interventions to prevent bone loss during weight reduction.

To our knowledge, three prior studies (5–7) including older men have measured changes in weight and changes in hip bone density occurring over the same time period and have reported that weight loss is a strong risk factor for hip bone loss. In agreement with these investigations, we found that men with weight loss experienced a rate of hip bone loss 3- to 4-fold higher in magnitude than the overall rate of loss observed among our and other cohorts of community-dwelling older men. Although weight loss in older people may be a marker of other conditions, such as underlying illness that increases bone loss, the majority of older men with weight loss in our study reported good or excellent health and experienced a modest degree of weight loss (average, 8%) over the 2-yr duration of the project. In addition, the effect of weight loss in older men on rates of hip bone loss we observed in our investigation was independent of other reported risk factors for bone loss in elderly men, including advanced age, poor health status, inactivity, smoking, alcohol use, low calcium intake, medical comorbidities, low body weight, decreased lean mass, poor lower extremity strength, and low baseline hip bone density. The association was also consistent across risk subgroups defined by age, body mass index, calcium intake, physical activity level, and bone density. Finally, although only a small number of deaths occurred during the relatively short follow-up period, the proportion of men who died during follow-up did not vary across weight change category.

If the relationship between weight loss and bone loss we noted in our study was explained in large part by occult illness (14), one might expect that the rate of loss would be higher in men with involuntary weight loss compared with that in men with voluntary weight reduction. However, to the contrary, we found that both intentional and unintentional weight losses were similarly associated with increased rates of hip bone loss in our cohort of community-dwelling older men. Even among overweight and obese men in our study, voluntary weight reduction, achieved primarily through a combination of diet and exercise, increased rates of hip bone loss. These findings are consistent with those of a prior large prospective study (15) in older women and smaller randomized trials of 12- to 18-month duration evaluating behavioral interventions aimed at producing modest weight loss in overweight younger men (16) and perimenopausal (17) and postmenopausal women (18). Because the prognostic significance of excess weight in elderly people is controversial (19), our results suggest that randomized trials in overweight and obese elderly subjects evaluating the effectiveness of interventions aimed at weight reduction in reducing disease outcomes should consider including bone density and fracture as end points. In addition, trials should be conducted in overweight subjects undergoing weight reduction to determine whether the addition of exercise regimens aimed at loading bone and/or supplementation with calcium and vitamin D are beneficial in preventing bone loss due to weight loss.

The rate of hip bone loss in older men may increase with weight loss for a number of reasons. Weight reduction may result in a decline in muscle mass and strength that decreases muscle force/strain on basic multicellular units in bone, leading to disuse mode remodeling and bone loss (20). The decline in mechanical load on the weight-bearing skeleton may also result in increases in bone turnover (21, 22), leading to bone loss. In addition, weight loss may be associated with a decrease in nutrient intake, including calcium and protein or may be linked to changes in levels of hormonal factors, such as lower levels of endogenous sex steroids (23, 24) including estradiol and testosterone; increases in serum glucocorticoids (25); or disturbances in biochemical factors modulated by adipose tissue, such as leptin (26) and adiponectin. (27) Any or a combination of these alterations might mediate increases in the rates of bone loss observed with weight reduction. Finally, weight loss may be a marker of other conditions that increase the risk of bone loss. However, the increased rates of hip bone loss in our older men with weight loss were not explained by known risk factors, including advanced age; poor health status and a higher prevalence of medical conditions; lifestyle behaviors, including inactivity, smoking, alcohol use, and low calcium intake; and decreases in body weight, muscle mass, lower extremity strength, and hip bone density.

Due to our limited follow-up period, we are uncertain whether the higher rates of hip bone loss observed in men with weight loss in our cohort result in an increased risk of hip fracture, especially among heavier men, who have greater initial hip bone density. However, after adjustment for multiple factors, including body mass index, a previous study in older men (28) found that men who had lost 10% or more of their body weight between age 50 yr and old age had an increased risk of hip fracture. In addition, a previous large prospective study in older women (15) reported that modest weight loss in later years was an independent risk factor for subsequent hip fracture, even among overweight women trying to lose weight. Of interest, the average rate of change in total hip bone density among the older women with weight loss (−0.9%/yr) in the study (15) was lower than that (−1.4%/yr) observed among older men with weight loss in this cohort.

Our study has a number of limitations. The participants were older, predominantly Caucasian, men, and our results may not apply to other ethnic groups. Small decreases in bone density that accompany weight loss or increases in bone density that accompany weight gain may be related to measurement artifacts associated with changes in the amount of soft tissue surrounding bone between scans. However, other studies in older men (28) and women (15) reported that weight loss is an independent risk factor for subsequent hip fracture, suggesting that accelerated bone loss does occur. A substantial proportion of the men in our study did not enroll in the dental cohort and did not have repeat hip bone density measurements. Because of their older age and poorer health status, the men who were not remeasured most likely had even greater rates of bone loss and weight loss than those who returned. Thus, our estimates of the rates of bone loss attributable to weight loss in older men may be conservative estimates of the true rate of loss.

We conclude that older men with weight loss are at increased risk of hip bone loss, even among overweight and obese men with voluntary weight reduction. Given the robustness of this association, our findings indicate that health care providers should take into account weight loss when evaluating older men for risk factors for osteoporosis and related fractures. Randomized trials in overweight older people undergoing voluntary weight reduction should be conducted to evaluate the effectiveness of interventions to prevent bone loss due to weight loss.

Acknowledgments

Investigators in the Osteoporotic Fractures in Men (MrOS) Research Group: Coordinating Center (University of California-San Francisco and California Pacific Medical Center Research Institute): S. R. Cummings (principal investigator), M. C. Nevitt (coinvestigator), D. C. Bauer (coinvestigator), K. L. Stone (coinvestigator), D. M. Black (coinvestigator), P. M. Cawthon (project director), R. Fullman (research associate), R. Benard, T. Blackwell, J. Diehl, S. Ewing, C. Fox, M. Jaime-Chavez, E. Kwan, S. Litwack, L. Y. Lui, A. Mills, L. Palermo, J. Schneider, R. Scott, D. Tanaka, C. Yeung; Administrative Center (Oregon Health and Sciences University): E. Orwoll (principal investigator), L. Marshall (coinvestigator), J. Babich Blank (project director), L. Lambert, B. Chan, D. Neevel, J Mougey, L. Press; University of Alabama-Birmingham: C. E. Lewis (principal investigator), J. Shikany (coinvestigator), P. Johnson (project director), E. Clavino, C. Oden, N. Webb, K. Hardy, S. Felder, P. Grayson, J. Wilkoff, J. King, T. Johnsey, J. Thompson; University of Minnesota: K. Ensrud (principal investigator), H. Fink (coinvestigator), D. King (program manager), N. Michaels (assistant program manager), N. Nelson (clinic coordinator), C. Bird, D. Blanks, F. Imker-Witte, K. Moen, M. Paudel, M. Slindee; Stanford University: M. Stefanick (principal investigator), A. Hoffman (coinvestigator), E. Moore (project director), K. Kent, B. Malig, S. Wong; University of Pittsburgh: J. Cauley (principal investigator), J. Zmuda (coinvestigator), M. Danielson, L. Harper (project director), L. Buck (clinic coordinator), M. Nasim, D. Cusick, D. Moore, M. Gorecki, D. Lee, N. Watson, C. Bashada, C. Newman, G. Engleka; University of California-San Diego: E. Barrett-Connor (principal investigator), T. Dam (coinvestigator), M. L. Carrion-Petersen (project director), P. Miller, N. Kamantigue, S. Szerdi, G. Reno.

The MrOS Study and its ancillary study, Determinants of Periodontal Disease in Older Men, are supported by National Institutes of Health funding. The following institutes provided support: the National Institute of Arthritis and Musculoskeletal Skin Diseases, the National Institute on Aging, the National Institute of Dental and Craniofacial Research, and the National Cancer Institute under the following grant numbers: UO1-AR-45580, UO1-AR-45614, UO1-AR-45632, UO1-AR-45647, UO1-AR-45654, UO1-AR-45583, UO1-AG-18197, RO1-DE-014386, and M01-RR-000334.

First Published Online January 25, 2005

Abbreviations:

 
• BMD,

  Bone mineral density;

 
• PASE,

  Physical Activity Scale for the Elderly.