| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH COMMUNICATION |
1 From the Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA (ABN, JSL, and BHG); the Institute for Research in Extramural Medicine, VU University Medical Center, and the Department of Nutrition and Health, Faculty of Earth and Life Sciences, Vrije Universiteit, Amsterdam, Netherlands (MV); the Division of Gerontology and Geriatric Medicine, Department of Internal Medicine, Wake Forest University School of Medicine, Winston Salem, NC (SBK); the Department of Preventative Medicine, University of Tennessee Health Science Center, Memphis, TN (FAT); the Prevention Sciences Group, University of California, San Francisco, CA (MN); and the Laboratory for Epidemiology, Demography and Biometry, National Institute on Aging, Bethesda, MD (TBH)
2 Supported by National Institute on Aging contracts N01-AG-6-2101, N01-AG-6-2103, and N01-AG-6-2106.
3 Reprints not available. Address correspondence to AB Newman, 130 North Bellefield Street, Room 532, University of Pittsburgh, Pittsburgh, PA 15213. E-mail: newmana{at}edc.pitt.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: The objective was to evaluate the relation between weight loss or weight gain and changes in lean mass and fat mass in older adults.
Design: We observed changes in weight and body composition during a 4-y period in 2163 men (47%) and women (53%) aged 70–79 y in the Health, Aging and Body Composition Study cohort. Whole-body and appendicular bone-free lean mass and fat mass were measured by using dual-energy X-ray absorptiometry.
Results: Weight loss and weight gain were common. In both weight losers and weight gainers, changes in lean mass as a percentage of initial lean mass were substantially smaller than changes in fat mass as a percentage of initial fat mass. However, the difference between the change in lean mass and that in fat mass was more pronounced with weight gain than with weight loss, especially in men. Small amounts of lean loss and fat gain were noted with weight stability. In multivariate models, weight loss was strongly associated with lean mass loss in both men and women, especially in men whose weight loss was concurrent with a hospitalization.
Conclusions: With weight change, a greater proportion of lean mass than of fat mass was conserved, but, especially in older men, significantly more lean mass was lost with weight loss than was gained with weight gain. These findings suggest that weight loss, even with regain, could accelerate sarcopenia in older adults.
Key Words: Body composition • sarcopenia • weight loss • older adults
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Older adults have been shown to be less effective in regulating weight than are younger adults (9). It may be that older adults are less able to conserve lean mass with weight loss than are younger adults. This loss of lean mass in old age (sarcopenia) may be worsened by weight loss if the loss of lean mass is excessive with respect to the available lean mass (10).
In a large, ongoing cohort study of older, community-dwelling men and women, we assessed changes in total, lean, and fat masses over time. Separately in subjects losing weight and in those gaining weight over 4 y, we sought to ascertain whether changes in lean mass and FM were related to the baseline body composition, ie, that before the weight loss or gain. We also examined whether the composition of the weight change differed between those who lost weight and those who gained weight. We hypothesized that older adults might show an excessive loss of lean mass with weight loss.
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The main study cohort consisted of 3075 men and women (black: 41.7%; white: 51.5%). Of these, 263 died and 79 could not be contacted, withdrew from the study, or moved out of the area. A total of 912 participants did not complete the full exam at follow-up. Thus, the analytic sample included 2163 participants (52.5% women, 38.4% black) for whom there were complete data on changes in body composition throughout the 4-y period (70.3% of original cohort). Excluded participants were more likely than were included participants to be black (49.5% and 38.4%, respectively; P < 0.0001) and current smokers (13.9% and 8.9%, respectively; P < 0.0001); they were also less likely than were included participants to be high school graduates (30.7% and 23.0%, respectively; P < 0.0001).
All study participants gave written informed consent. The protocol was approved by the institutional review boards of the participating institutions.
Weight-change category and body-composition changes
Net body weight change through the 4-y period was calculated, and subjects were grouped into 3 weight-change categories: loss (
3% of loss), stability (within 3% weight loss or gain), and gain (3% gain). This 3% cutoff was selected (8) to exceed the CV for dual-energy X-ray absorptiometry (DXA) soft tissue mass (11–13). Total body mass (total weight) and body composition at baseline and 4-y follow-up were measured by using fan-beam DXA (QDR4500A) with DXA software (version 8.21) (both: Hologic, Bedford, MA). The validity and reproducibility of the body-composition data in the Health ABC Study were reported previously (14–17). Quality-assurance measures included the use of a body-composition phantom for calibration and annual assessment for potential site differences or drift over time. Total weight, total-body bone-free lean mass (BFLM), total body FM, and total body fat percentage and lean percentage measured at baseline and 4-y follow-up were used in this analysis. Changes in each body-composition compartment (fat and lean) between baseline and 4-y follow-up were calculated as absolute change (4-y follow-up value minus baseline value) and percentage change [(4-y follow-up value – baseline value) x (100/baseline value)]. The lean composition of the baseline weight and of the weight change itself was also calculated.
Other variables
The Health ABC Study collected extensive information on baseline sociodemographic, lifestyle, and health characteristics of the participants. The sociodemographic factors included were those that are known to influence weight and weight change—ie, age, race, and sex. Race (black, white, or other) was self-designated by the participant. The cohort was stratified by the occurrence of no hospitalizations or 1 hospitalizations in the 4-y period between the first weight and the follow-up weight (interim hospitalization). This measure was used as an indirect indication that the observed weight change may have been unintentional. A hospitalization was defined as an overnight stay (ie, >24 h) in an acute-care hospital, and that occurrence was ascertained at 6-mo intervals by interview at the annual clinic visit or interim phone call and confirmed by review of medical records.
Statistical analysis
Differences in means and proportions of baseline characteristics and body composition by sex were tested by using Student's t tests and chi-square tests. Differences in mean changes in body composition by weight-change category and by sex within each weight-change category over the 4-y period were tested by using analysis of variance (ANOVA) and t tests. When overall differences were significant by ANOVA, post hoc comparisons were performed with Bonferroni adjustment. Analysis of covariance (ANCOVA) was used to ascertain the main effects of sex, race, and weight-change category and the potential interactions between them on changes in total-body BFLM over 4 y after control for potential confounders (including age, race, interim hospitalization, and baseline weight and body composition). There was a significant interaction between the direction of the weight change and sex, but not between the direction of the weight change and race or age, for changes in total-body BFLM (P < 0.0001), which persisted in the fully adjusted models; thus, we present ANCOVA models predicting total-body BFLM loss stratified by sex and by weight-change category. There were no significant interactions between net weight change and baseline weight. There were, however, significant interactions between interim hospitalization and net weight change over the 4-y period in men with weight loss and those with weight stability, and these significant interactions were included in the final models. Other interactions between age, sex, and race and baseline body composition were not significant in sex- and weight-change-category–specific models. All analyses were performed by using SAS software (version 8.0; SAS Institute, Cary, NC) (18).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Of 2163 participants, half were men and 38.4% were black. Baseline body composition was described in terms of absolute lean mass and FM, as well as percentage FM and percentage lean mass. As expected, baseline body composition differed by sex: there was little overlap between men and women, and women had a substantially higher percentage body fat. Hospitalization was common in this age group and was significantly more common in men than in women. Across the follow-up period, changes in weight were common; weight loss of 3% occurred in 31.2% of men and 32.8% of women, and weight gain of 3% occurred in 20.6% of men and 24.1% of women (P < 0.05). The reminder of the subjects maintained their weight within 3% of baseline.
|
). With weight loss, men lost mostly lean mass, but women lost more FM than lean mass. In both men and women, significantly less lean mass was gained in the weight-gain group than was lost in the weight-loss group. However, regardless of the direction of weight change, the changes in the fat compartment were proportionally greater than those in the lean compartment. This proportionally greater percentage change in FM than in lean mass with both weight loss and weight gain is shown in Figure 1. As shown, men losing weight lost 5.8% of their initial lean mass but 10.6% of their initial FM, whereas men gaining weight gained only 2.0% of their initial lean mass but gained 17.9% of their initial FM. Women losing weight lost 5.02% of their initial lean mass and 12.7% of their initial FM, but women gaining weight gained 2.96% of their initial lean mass and 13.7% of their initial FM. All sex differences within weight-change categories were significant (P < 0.05 for all; Student's t tests). Figure 1 also illustrates the mean tendency for both men and women to have lost proportionally more lean mass with weight loss than they have gained lean mass with weight gain.
|
|
), and then we examined the changes by using linear regression models after adjustment for baseline body composition (Table 3). Figure 2 shows that the weight that was lost by men had a composition of 60.2% lean, whereas their baseline weight composition was 68.4% lean (r = 0.23, P < 0.0001); in women, the weight that was lost had a composition of 39.5% lean, whereas their baseline weight composition was 57.3% lean (r = 0.18, P = 0.0005). In weight gainers, a smaller proportion of the weight that was gained was lean (21.4% in men and 24.3% in women). In men, the proportion of lean in the weight gained did not correlate with the baseline body composition (r = –0.05, P = 0.4813), but, in women, the correlation was stronger (r = –0.25, P < 0.0001) and was similar to that seen in women with weight loss.
|
|
). In sex-specific models predicting the loss of lean mass in the weight-loss group, the ß coefficient (±SE) for total weight was 0.325 ± 0.042 in men and 0.257 ± 0.017 in women. In other words, for each kilogram of weight lost, an average of 0.325 and 0.257 kg lean mass was lost in men and women, respectively. In these adjusted models, the magnitude of the loss of lean mass with a given weight loss remained greater than the magnitude of the gain in lean mass with a similar weight gain in men, but not in women (weight-change category x sex interaction, P < 0.001). There was no interaction between baseline weight and weight change in men or women. In other words, the changes in lean mass and total body weight were proportional across the range of baseline weights. These models also show that baseline body composition was related to the composition of the lean mass that was lost in both men and women. Both men and women with greater lean mass or lesser FM lost more lean mass. On a per-kilogram basis, the effect of baseline body composition was much smaller than the effect of weight change on the change in lean mass. In addition, the inclusion of the baseline body composition in the models only slightly attenuated the coefficients for weight loss. With weight gain, the baseline body composition was related to gain in lean mass in the women but not in the men. Black race and older age were not generally related to either the loss or gain in lean mass. Interim hospitalization was also a predictor of loss of lean mass in men, but this was not significant in the women after adjustment for the other factors. In men, there was a significant interaction between weight loss and hospitalization for the loss of lean mass, which suggests that weight loss with interim hospitalization accelerated the loss of lean mass beyond that occurring with weight loss in those without hospitalization (ß ± SE kg lean loss with weight loss: 0.438 ± 0.029 in hospitalized men and 0.325 ± 0.042 in nonhospitalized subjects). When the full model was reexamined with the 2 sexes combined, the interaction term for the direction of change and sex on the loss of lean mass remained significant, which illustrates that the baseline body composition and the total amount of weight loss did not explain the tendency for the men to lose more lean mass with weight loss than they gained with weight gain.
The relations between weight change and lean mass change did not differ significantly in the weight-stable and weight-loss groups. In the weight-stable men and women, there was a slight tendency to lose some weight, and this was the major predictor of the change in lean mass. As was shown with respect to lean mass, the changes in FM were largely related to the net weight change and the baseline body composition in models predicting fat mass change (data not shown). Consistent with the descriptive absolute values for change, the model estimated a larger magnitude of change in FM with a given weight change than was observed in the models predicting change in lean mass with weight change; this indicates that fat is more responsive to weight change than is lean. There was also a significant (P = 0.0009) weight change x sex interaction for changes in FM, so that the gain of FM with weight gain was somewhat greater than the loss of FM with weight loss.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Studies of the changes in body composition in connection with weight cycling in middle-aged dieters show a strong tendency to return to baseline body composition (19, 20). In the rat model, weight cycling resulted in a return to baseline body composition but perhaps also in resistance to weight loss in obese rats (21, 22). If the patterns observed in the Health ABC cohort were to occur with a cycle of loss and subsequent regain in the same older person, the resulting body composition might be expected to become more sarcopenic than if weight had been maintained, at least initially. Future follow-up of this cohort may identify persons with this pattern of change or may ascertain whether the initial changes are maintained over time.
This study provides validation of previous observations in small, select groups of older adults with respect to the importance of weight loss in the loss of lean mass with aging (6, 10). Available longitudinal studies of body composition in older adults show a tendency toward a loss of lean and a gain of fat over time in healthy, weight-stable persons (7, 10). Our data show a similar pattern in the weight-stable subgroup, but they also point out that the magnitude of these changes is fairly small compared with the large loss of lean mass that accompanies weight loss. In cohorts whose weight loss has not been accounted for, the changes in lean mass in older persons may have been overstated.
Baseline body composition was associated with the amount of lean mass lost with weight loss. In both men and women, greater initial lean mass and lesser initial FM predicted a greater loss of lean with weight loss; in men, however, only a small proportion of the greater amount of lean mass lost was explained by their greater initial lean mass or lesser body fat. Conversely, with weight gain, initial body lean and fat mass were related to lean mass gain in the women, but not in the men. These findings support the 2-compartment model of "companion" changes in body composition: ie, lean and fat changes are always in the same direction as weight change, but there are some variations, as reviewed by Forbes (5). To our knowledge, no other study has documented these physiologic relations in large cohort of community-dwelling older adults.
Although baseline body composition as well as weight change explained some of the sex difference in the magnitude of the changes in lean mass and FM, it remained more likely that men, especially those with an interim hospitalization, rather than women would lose lean mass. Men are known to experience age-related weight loss earlier than are women (23). Potentially, there may be other aspects of poorer health in men that we did not account for in these analyses, or perhaps declining concentrations of androgens play a role. There was some suggestion of a slightly greater change in lean in blacks, but the associations between weight loss and loss of lean mass did not differ significantly between blacks and whites.
Strengths of this study include the large sample of community-dwelling men and women (both whites and blacks), the state-of-the-art quantification of lean mass and FM by using DXA with careful calibration across time, and careful follow-up with respect to hospitalization. Several important aspects of the analysis should also be kept in mind in interpreting these data. The weight gainers and weight losers in our cohort were different groups. Our data and those of others (1, 24–26) document that older adults who lose weight generally are sicker than their counterparts who gain weight, although both weight losers and gainers have more disability than do weight-stable older adults. The causes and circumstances of weight change could further influence the composition of change. In addition, the weight-stable group is heterogeneous and may include people with 1 cycles of loss and regain. We are more closely following a subset of subjects with a cycle of weight loss and regain to identify the natural history and body-composition changes that are associated with weight cycling.
It is important to point out that the members of this cohort were functioning well at baseline, and few were at extremes of weight. We did not find significant differences in the association of net weight change with lean change across the spectrum of initial weights in this cohort. Most previous studies focused on obesity, whereas the participants in the current study were mostly of moderate weight. Patterns may differ in persons who are extremely thin or obese or in those who are disabled. In addition, those who dropped out before the follow-up measure may have been the most likely to lose weight and lean mass, so that these findings are probably conservative. Finally, it is important to keep in mind that DXA underestimates lean mass in the setting of fluid overload. Exclusion of the cases with prevalent or incident CHF (<5% of the total) did not change the associations in this cohort.
In addition to weight change associated with acute hospitalization, factors that may influence the loss of lean mass should be examined. These may include sex hormone steroids (27), insulin-like growth factor I (28), protein intake (29), and physical activity (30). Because lean mass is associated with muscle strength (31, 32), bone density (33–35), and physical functioning (36, 37), it appears that avoidance of weight loss would be of substantial value in preventing the loss of lean mass and related disability in old age.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. Mozaffarian, A. Kamineni, R. J. Prineas, and D. S. Siscovick Metabolic Syndrome and Mortality in Older Adults: The Cardiovascular Health Study Arch Intern Med, May 12, 2008; 168(9): 969 - 978. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Fantin, V. D. Francesco, G. Fontana, A. Zivelonghi, L. Bissoli, E. Zoico, A. Rossi, R. Micciolo, O. Bosello, and M. Zamboni Longitudinal Body Composition Changes in Old Men and Women: Interrelationships With Worsening Disability J. Gerontol. A Biol. Sci. Med. Sci., December 1, 2007; 62(12): 1375 - 1381. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. D Hansen, D. A Williamson, T. P Finnegan, B. D Lloyd, J. N Grady, T. H Diamond, E. U. Smith, T. M Stavrinos, M. W Thompson, T. H Gwinn, et al. Estimation of thigh muscle cross-sectional area by dual-energy X-ray absorptiometry in frail elderly patients Am. J. Clinical Nutrition, October 1, 2007; 86(4): 952 - 958. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Thompson Martin, J. Kayser-Jones, N. Stotts, C. Porter, and E. S. Froelicher Nutritional risk and low weight in community-living older adults: a review of the literature (1995-2005). J. Gerontol. A Biol. Sci. Med. Sci., September 1, 2006; 61(9): 927 - 934. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 From the Library Br. J. Ophthalmol., March 1, 2006; 90(3): 394 - 394. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and fat (
) compartments with weight loss and weight gain over a 4-y period in the Health, Aging and Body Composition Study cohort (n = 2163). There were significant (P < 0.05) differences in lean and fat mass changes over the 4 y by weight-change category (one-way ANOVA), by sex (Student's t test), and by sex within each weight-change category (Student's t test). All pairwise comparisons between the 3 weight-change categories were significantly different, P < 0.0167 (t tests with Bonferroni adjustment).