| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
a Jean Mayer U.S. Department of Agriculture Human Nutrition Research Center on Aging, Tufts University, Boston, Massachusetts
Susan B. Roberts, Energy Metabolism Laboratory, Jean Mayer U.S.D.A. Human Nutrition Research Center on Aging at Tufts University, 711 Washington St., Boston, MA 02111 E-mail: SRoberts{at}hnrc.tufts.edu.
Decision Editor: Jay Roberts, PhD
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
AGING is associated with progressive changes in body composition that have important effects on physical function, overall health status, and pharmacokinetic processes. There is typically a doubling in adipose tissue from early adult life (20–30 years) to age 50–60 years that is associated with impaired glucose tolerance, morbidities such as coronary heart disease and stroke, and premature death (1) (2) (3). In contrast, after age 65–70 years, body fat content tends to decrease even in healthy individuals (4) (5), and unexplained weight loss leading to protein-energy malnutrition becomes increasingly common (6) (7). Among institutionalized older persons, as many as 30% to 50% are reported to suffer from protein–energy malnutrition (8) (9). This loss of body weight, fat-free mass, and fat late in life is associated with premature death, micronutrient deficiencies, frailty, increased hospital admission, and increased risk of disability from falls and is also known to delay recovery from injury (10) (11) (12). However, the underlying causes of weight loss in old age (>65 years) remain unclear.
Functional disabilities such as a diminished sense of taste and smell, poor dentition, impaired vision, dementia, psychological problems (e.g., depression, bereavement, and alcoholism), and social changes (e.g., poverty, social isolation, and living alone) are often suggested as promoting inadequate energy intake leading to weight loss (13) (14). In addition, data from Roberts and colleagues (15) suggested a specific impairment in the regulation of food intake. Following a period of enforced overeating or undereating, healthy elderly individuals, in contrast to young subjects, did not exhibit appropriate compensatory responses in energy intake. Thus, body weight remained depressed following undereating, and weight remained elevated over a follow-up period of 7 weeks following overeating. In relation to this observation, it should be noted that overeating and undereating occur routinely during the course of normal life. Thus, abnormalities in the subsequent energy intake response of elderly individuals to both weight gain and weight loss may have an important impact on health.
We therefore conducted a study to further investigate the regulation of food intake in older subjects. Specifically, we tested the hypothesis that there is long-term impairment in body weight regulation following underfeeding. We also examined the question of whether this impairment can in part be traced to a reduced perception of hunger.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
|
|
The following 6 weeks of the study were designated Phase 2, and during this time energy intake was decreased by approximately 1000 kcal/d relative to Phase 1 (see below). Fecal collections were again obtained over 3 days to assess fecal energy losses, and in addition, two 3-day complete urine collections were obtained for measurement of osmolar excretion rate and assessment of dietary compliance (17). Fasting body weight was measured, and body composition was determined by underwater weighing at the end of Phase 2.
During a 6-month follow-up period (designated Phase 3), subjects were free to consume as much or as little food as they wished. Food was provided by the research center during the initial 2-week period, after which subjects prepared their own food at home. They returned to the center at approximately 6-week intervals for measurements of fasting body weight and body composition, but were not given instructions on further weight loss or prevention of weight regain.
Diets
Meals consisted of normal food and beverage items divided between three meals, plus a snack each day. There were three menus, which were provided on a 3-day rotating basis. Coffee and tea were available in fixed daily amounts if they formed part of the subject's normal diet. Sodium intake of provided foods was fixed in Phase 1, and remained constant throughout Phases 1 and 2, and subjects were advised to add no salt to the provided foods. In addition, a standard multivitamin/mineral supplement was given throughout Phase 2 to ensure no deficiency in any essential nutrient when energy intake was decreased. At least one meal per day was consumed in the research center 5 days per week during the first study week and at least 3 days per week subsequently. Other meals were consumed either at the research center or at the subject's residence according to individual preference.
The nutrient content of the diets provided during Phase 1 was designed to mimic a typical American diet, providing 35% of energy from fat, 13% of energy from protein, and 52% from carbohydrates. During Phase 2, protein intake was maintained constant and energy intake was decreased by 1000 kcal/d relative to Phase 1 in all but seven subjects (who had initial energy intakes that were low and for whom a 1000-kcal reduction was considered too severe; in these cases, an 800–900 kcal/d deficit was employed). The energy reduction was achieved by a proportional reduction in carbohydrate and fat. In addition, all subjects were randomly assigned within gender- and age-group categories to receive either a diet containing rolled oats (45 g/1000 kcal) or a diet that had the same percentage of energy from fat, protein, carbohydrate, and the same insoluble fiber per 1000 kcal but with an equivalent amount of wheat products substituted for oats. There was no effect of the oats versus nonoats diet on body composition or change in body weight, as described elsewhere (18).
Dietary Compliance
Subjects were instructed to consume all the food and beverages provided by the research center (except during periods when food returns were permitted) and to eat no other food or energy-containing beverages. To assess dietary compliance during Phase 2 we used the osmolar excretion rate technique (OER)
(17). The theoretical basis of this method is the comparison of measurements of the urinary OER with values for OER predicted from dietary nitrogen, sodium, and potassium intakes. Briefly, the urinary osmol load consists primarily of compounds derived from consumed foods, in particular nitrogen-containing compounds and sodium and potassium salts. Thus, the urinary osmol load can be predicted from dietary protein, sodium, and potassium, taking into account obligatory nonurinary losses of these nutrients and correcting for other nutrients contributing to urine osmolality. A measured osmol load significantly greater than predicted indicates the consumption of illegal foods, whereas a measured osmol load significantly less than predicted suggests either that provided foods are not being consumed or that the 24-hour urine collection is incomplete. The theoretical urinary osmolar load for the Phase 2 diet was calculated as
 |
The measured urinary osmolar excretion rate (mOER) in Phase 2 was calculated as the product of urine volume and urine osmolality (Model 3MO Micro-osmometer, Advanced Instruments Inc, Needham, MA) for each of the six 24-hour periods when urine was collected. These values were then expressed as a percentage of tOER, and values were averaged over 6 days (i.e., the two 3-day collections combined).
Anthropometry and Body Composition
As described elsewhere
(19), body weight and height were measured to ±100 g and 0.1 cm, respectively, and body density was measured by hydrostatic weighing after a 12-hour overnight fast. Body density was calculated with the residual lung volume predicted by Quanjer's equations
(20). Hydrostatic measurements were repeated until at least three were within 1% body fat of each other, and then the average of three tests was used for analysis. Fat mass and fat-free mass were calculated using the equation of Siri
(21).
Energy Excretion
For determining fecal energy excretion, 3-day stool collections were obtained during both Phase 1 and Phase 2. The collections were weighed, homogenized, freeze-dried to constant weight (Virtis Model 6203-3006-OG freeze drier, Virtis Co, Gardiner, NY), ground to a fine powder, and analyzed in duplicate for fecal energy by isoperibol bomb calorimetry (Model 1261, Parr Instrument Co, Moline, IL).
Hunger and Satiety Ratings
Due to a technical error, only a subgroup of subjects (7 men and 12 women) completed a questionnaire at the end of Phases 1 and 2 to quantify changes in hunger, satiety, thirst, and constipation. Subjects were asked to quantify the frequency of hunger, satiety, thirst, and constipation using a 5-point scale (where 1 = never or no change, 2 = rarely or once or twice only, 3 = occasionally or 3–6 times in a phase, 4 = frequently or 2–3 times per week, 5 = very frequently or more than once per day).
Statistical Analyses
Data are expressed as mean ± SEM unless otherwise specified. Data were analyzed using SPSS, version 7.5 (SPSS Inc, Chicago, IL), Systat (version 7.0, SPSS Inc.), and SAS for Windows, version 6.12 (SAS Institute, Inc, Cary, NC). Statistical significance was accepted at the .05 level. Gender was examined as a potential independent variable in initial analysis and was not significant. For this reason, data are not separated by gender in Results.
Within study phases, group means were compared using analysis of variance (ANOVA). We then used two approaches to analyze the effect of the intervention on body weight in the three groups (YNW, YOW, OLD). First, two-way, repeated-measures ANOVAs were performed with body weight as the dependent variable, and time (Phase 1 end, Phase 2 end, Phase 3 week 6, Phase 3 week 12, Phase 3 end) and group as independent variables. The time by group interaction was also examined. When the time by group interaction was significant, repeated measures ANOVAs were subsequently performed separately for each of the three groups. Tukey's post hoc multiple comparison procedure was used to determine which means differed significantly.
The second procedure we performed to examine the effect of the intervention on body weight was to calculate and use body weight change scores as dependent variables in analysis of covariance (ANCOVA). This was done to assess group differences in weight change. Four weight change scores were calculated as follows: Phase 2 end - Phase 1 end (change during Phase 2); Phase 3 week 6 - Phase 2 end (change in first 6 weeks of Phase 3); Phase 3 week 12 - Phase 3 week 6 (change in second 6 weeks of Phase 3); and Phase 3 end - Phase 3 week 12 (change in second half of Phase 3). In four separate ANCOVA procedures (one for each change score), group was the independent variable and initial body weight (body weight at the beginning of each weight change score time interval) was included as a covariate to control for the influence of initial weight on weight change.
We also performed the above ANCOVA procedure on fat mass and fat-free mass changes. However, because these body composition parameters were measured less often than body weight, we were only able to calculate three change scores: Phase 2 end - Phase 1 end (change during Phase 2); Phase 3 end - Phase 2 end (change during Phase 3); and Phase 3 end - Phase 1 end (change during Phases 2 and 3 combined).
Finally, we used ANOVA to assess potential differences among the groups in associations among parameters (e.g., change in body weight with change in fat-free mass). We did this by including an interaction term between group and the other independent variable in the analysis. When there was no difference among the three groups in the association, a pooled within-group correlation coefficient was calculated by using GLM univariate analysis. Change in weight and change in fat-free mass were dependent variables, and group was the independent variable.
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Drop-out rate and compliance (as assessed by the urinary osmolar excretion rate technique) are shown in Table 2 . The three groups generally had good compliance, with relatively few subjects providing urine collections with more than 130% or less than 70% of expected (more than 130% and less than 70% equate approximately to 2 SD from expected values in the original validation study; see reference 17). There was no significant difference among groups in these measures, although fewer older subjects appeared to be noncompliant. Note that data from subjects with urinary osmolar excretions of more than 130% expected, together with subjects who dropped out, were excluded from the other study analyses described below because of expected bias in the results; however, subjects with expected excretions less than 70% were not excluded because of the possibility that low excretion rates were due to incomplete urine collection rather than dietary noncompliance.
|
|
|
|
The relationships between weight change and fat-free mass change in Phases 2 and 3 are shown in Fig. 3 and Fig. 4. In both phases, the relationship between weight change and fat-free mass change was significant (pooled within group r = .41 and .71 for Phases 2 and 3, respectively, p < .05). A single line is shown for each relationship because ANOVA indicated that the interaction between the change in fat-free mass and group was not significant (p = .598 and .622 in Phases 2 and 3, respectively). In other words, the relationship between change in fat-free mass and change in weight did not differ significantly among the three groups for either Phase 2 ( Fig. 3) or Phase 3 ( Fig. 4).
|
|
|
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Previous studies by our group (15) (23) suggested that an alteration in the ability to accurately regulate energy intake and energy balance in older individuals might underlie multiple apparent causes of weight loss. In particular, older men were found to compensate less well for imposed changes in energy balance than younger men. While younger men reacted to an experimental 3-week decrease in energy intake with a subsequent compensatory increase in voluntary energy intake and regain in weight, the older subjects did not exhibit these responses within the 7-week study period and maintained a significantly lower body weight following underfeeding. However, that study did not address the question of whether the elderly subjects might have experienced delayed, rather than absent, energy regulation, thus returning to baseline weight over an extended period of time. Furthermore, that study did not address the question of whether the observed group differences were due to alterations in biological mechanisms of energy regulation or to psychological differences between the young and older subjects.
In the present study, we assessed body weight and composition changes over 6 weeks of underfeeding and a 6-month follow-up period, using both normal weight and overweight young subjects as controls, so that differences between the groups could be examined in relation to both age and body composition. In addition, we documented changes in hunger and satiety during the underfeeding phase of the study. One of the primary findings of this study was that both OLD subjects and YOW subjects lost significantly more weight during underfeeding than YNW subjects. Fat loss during underfeeding was not different among the groups, but this may have been due to the fact that measurements of body fat content are less precise than body weight, because the trends in fat were consistent with the trends in weight. The finding of increased weight loss in response to underfeeding in OLD subjects is consistent with our previous finding that elderly adults are less able to reduce energy expenditure during periods of negative energy balance (24), perhaps because their lower muscle mass makes compensation more difficult, thus increasing the effective negative energy balance for a given deficit in energy intake.
Another finding in this study was that YNW subjects subsequently gained back the weight they lost during underfeeding, whereas YOW and OLD subjects did not. The finding that the older subjects did not regain weight is again consistent with our previous results (15) and is consistent with the suggestion of a loss of the regulation of food intake in old age. However, the fact that OLD and YNW subjects had similar body weight responses to this underfeeding protocol suggests that the increase in body fat mass that occurs with aging may be in part responsible for the loss of control of food intake. It is also possible that different factors were responsible for the same lack of weight regain in OLD and YOW subjects and that further studies in this area are needed. For example, there may have been an age-specific loss of the control of food intake in OLD subjects or different social circumstances that inhibited energy intake and a greater desire to prevent weight regain in YOW subjects.
In relation to this latter suggestion, we also observed that although weight and fat loss were greater in older subjects than YNW subjects, the older subjects experienced significantly less frequent hunger than did both young groups. This finding strongly suggests that an impairment in the biological mechanism of hunger was present in our older subjects and may have contributed to their lack of weight regain following underfeeding. Moreover, this result in OLD subjects was obtained despite the fact that our older subjects were healthy and reported having no problems with their appetite. Anorexia is also common in nursing home populations, and our results are consistent with the suggestion that a lack of hunger precedes weight loss and is a contributing factor (14) (25). Most of the reports of anorexia in older individuals attribute the problem to social factors, functional limitations, and psychological disorders. Depression is suggested to be the most common treatable cause of anorexia in older persons (13), but in fact other reports have noted that depression is associated with weight loss only in older persons and instead causes weight gain before age 50 (26). For this reason, we previously suggested that weight loss in older persons was most likely due to a combination of impaired biological mechanisms of energy regulation combined with adverse circumstances such as poverty, depression, and loss of teeth that provided an impediment to eating in the absence of strong biological signals encouraging food intake. Consistent with this view, Clarkston and colleagues (27) reported that carefully screened older persons tended to be less hungry than younger persons after a standardized overnight fast, and after a standard meal, older persons reported a greater degree of satiation than did younger persons that was correlated with reduced gastric emptying.
Concerning possible underlying causes of a decreased ability to regulate food intake in old age, Morley (13) summarized several potential candidate mechanisms, and research is needed to identify which may be quantitatively important. Nitric oxide was postulated to have significant effects on food intake at central and peripheral sites (28), and a decrease in messenger ribonucleic acid for nitric oxide synthase occurs with aging in older animals (29). Cholecystokinin and insulin have long been suggested to be satiety hormones, and circulating concentrations increase with age (30) (31). Rodent studies have also shown a decrease in the ability of opioids to drive food intake in older animals (32). Neuropeptide Y (NPY) is a potent anorectic agent and NPY concentrations decline with aging (33). Finally, circulating glucose is thought to be one of the signals for hunger in young adults (34), and detection of hypoglycemia is known to be impaired in older individuals (35).
In summary, the results of this study suggest that older individuals did not significantly regain weight lost due to underfeeding within 6 months, and a decreased perception of hunger was identified as a potential causal factor. Thus, both intentional and unintentional weight loss may be easier for older individuals than young adults. This may both facilitate weight management in elderly individuals needing to lose weight and put individuals who do not need to lose weight at increased risk of undesirable unintentional weight loss.
|
Acknowledgments |
|---|
Received August 25, 1999
Accepted May 18, 2000
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. Yukawa, D. S. Weigle, C. D. Davis, B. T. Marck, and T. Wolden-Hanson Peripheral ghrelin treatment stabilizes body weights of senescent male Brown Norway rats at baseline and after surgery Am J Physiol Regulatory Integrative Comp Physiol, May 1, 2008; 294(5): R1453 - R1460. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. P. Hays, G. P. Bathalon, R. Roubenoff, M. A. McCrory, and S. B. Roberts Eating behavior and weight change in healthy postmenopausal women: results of a 4-year longitudinal study. J. Gerontol. A Biol. Sci. Med. Sci., June 1, 2006; 61(6): 608 - 615. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. W Krieger, H. S Sitren, M. J Daniels, and B. Langkamp-Henken Effects of variation in protein and carbohydrate intake on body mass and composition during energy restriction: a meta-regression 1 Am. J. Clinical Nutrition, February 1, 2006; 83(2): 260 - 274. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. S. Lee, S. B Kritchevsky, T. B Harris, F. Tylavsky, S. M Rubin, and A. B Newman Short-term weight changes in community-dwelling older adults: the Health, Aging, and Body Composition Weight Change Substudy Am. J. Clinical Nutrition, September 1, 2005; 82(3): 644 - 650. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Mizuno, I-W. Shu, H. Makimura, and C. Mobbs Obesity Over the Life Course Sci. Aging Knowl. Environ., June 16, 2004; 2004(24): re4 - re4. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. E. Morley Anorexia and Weight Loss in Older Persons J. Gerontol. A Biol. Sci. Med. Sci., February 1, 2003; 58(2): M131 - 137. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
determination of fasting body weight and body composition; and 
fecal and urine collections were obtained over 3 days.
, young overweight (YOW) subjects are represented by 
, and older (OLD) subjects are represented by 
. The symbols indicate significant differences between groups at individual time points: p < .05 (* denotes YNW subjects vs YOW subjects, and 
