Cover Image
close this bookEnergy and Protein Requirements, Proceedings of an IDECG workshop, November 1994, London, UK, Supplement of the European Journal of Clinical Nutrition (International Dietary Energy Consultative Group - IDECG, 1994, 198 pages)
close this folderEnergy requirements of older individuals
View the document(introductory text...)
View the documentIntroduction
View the documentMethodology for determination of total energy expenditure
View the documentTotal energy expenditure of older individuals
View the documentRelationship between total energy expenditure and physical activity
View the documentEffects of energy balance on nitrogen balance
View the documentSummary of energy requirements in older individuals and recommendations for future research
View the documentReferences
View the documentDiscussion

(introductory text...)

SB Roberts

The Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University, 711 Washington Street, Boston, MA 02111, USA

Descriptors: energy metabolism, body composition, isotopes, metabolizable energy

Introduction

Current recommendations on dietary energy intake (FAO/WHO/UNU, 1985; National Research Council, 1989) define expected average amounts of metabolizable energy required for sustaining normal metabolic processes, together with desirable or expected levels of physical activity in healthy individuals. In weight-stable adults, energy requirements are thus equal to total energy expenditure (TEE). In re-evaluating the energy requirements of older individuals, several key issues need to be addressed:

1. Traditionally, TEE has been estimated by a factorial approach involving summation of all the expected components of energy expenditure, taking into account the energy costs of different activities and their durations. Does this approach remain the most suitable method, or alternatively should requirements now be based on more direct measurements of TEE determined by the doubly labeled water technique? An additional methodological issue is whether the current procedure of expressing requirements as multiples of predicted resting energy expenditure (pREE) should be recommended for continuing use, i.e. the PAL (physical activity level) system.
2. Do current recommendations on the energy requirements of older individuals accurately predict usual energy needs for persons in each decade over 50 years of age?
3. What is the relationship between physical activity and energy requirements in older individuals?

Methodology for determination of total energy expenditure

In the past, the lack of an accurate method for the direct determination of TEE has necessitated estimation of TEE either from measured energy intake in healthy populations, as used in the early editions of recommended dietary allowances, or by summing estimates of the principal components of energy expenditure (Roberts, 1995). In this second, 'factorial' approach (which has been in use since the early 1950s; see Roberts, 1995), information on the energy cost of major activities as measured by direct or indirect calorimetry is combined with expected durations of activities to provide an indirect assessment of the TEE over 24 h The primary limitation of the factorial approach to determining TEE is the well-recognized problem that it is hard to categorize and time the many varied activities performed during the course of a typical 24-h period. Thus, estimates of TEE are subject to substantial error and in addition may tend to underestimate TEE because of the tendency to ignore activity not associated with specific tasks.

Correspondence to: SB Roberts.

The doubly labeled water method, which was refined for use in humans during the 1980s (Schoeller, 1988; Roberts, 1989), now provides a more direct assessment of the TEE of free living individuals. Calorimetric validation studies in adults and infants with a wide range of body composition have demonstrated that the method can be accurate as well as reasonably precise. However, there remain several unresolved issues that need to be addressed in relation to using the method for reevaluating energy requirements.

The first concern is that there is currently substantial between-laboratory variability in the analytical capabilities in the increasing number of institutions that now have the capacity to make doubly labeled water measurements of TEE. A recent investigation of this variability in the technique in 19 laboratories from around the world has indicated that the analytical component of the method has the potential to introduce substantial error, resulting in estimates for TEE differing from calorimetrically-measured values by as much as 600% in some laboratories (Roberts et al, 1995). This finding implies that it may be misleading to compare values for TEE between subjects measured in different laboratories unless each of the laboratories has conducted its own validation of the method Either indirect calorimetry or the intake-balance method are a suitable reference method for determining the overall accuracy of the doubly labeled water technique (the key issue for studies of energy requirements in population groups) although indirect calorimetry has the added advantage that the precision of the method can also be assessed.

The second issue concerning the use of the doubly labeled water technique in studies of energy requirements is that it does not measure TEE directly and there is significant potential for influencing the predicted value for TEE through the choice of calculation procedures and correction factors. Of relevance to this issue is the fact that there remains no general agreement on a single calculation method for data processing and several different approaches are used (Prentice, 1990). Although it has been presumed that currently used calculation options should differ only slightly in their effect on predicted TEE, this does not apply to all data sets that have been published. In particular, the question of whether the measured ratio of the dilution spaces (2H2O: H2 18O) or a fixed ratio of 1.03 (2H2O: H2 18O) should be used remains unanswered. In a recent study by Goran & Poehlman (1992), values for TEE calculated using the mean of measured dilution spaces differed by 10.5% from values calculated using a fixed ratio of 1.()3. In addition, it has been suggested on the basis of preliminary data (Roberts et al, 1992) that the correction for isotopic fractionation should be lower in older individuals than in young adults, a factor estimated to influence calculated values for TEE by approximately 2%.

Table 1 Pearson correlation coefficients (R) for the relationship between total energy expenditure and selected variables in 35 young and older mena


Pearson correlation coefficients (R)


Young

Elderly

Young + Elderly

Variable

n = 17

n= 18

n= 35

Measured REE

0.5462*

0.7537***

0.7512***

Predicted REE

0.5089*

0.6638**

0.7071***

Fat free mass

0.6468**

0.7000**

0.6781***

Body fat (% weight)

- 0.3651

0.1737

- 0.4494**

Weight

0.5089*

0.6638**

0.3087

Height

0.3249

0.3535

0.2823

Body mass index

0.3780

0.5661*

0.1768

a SB Roberts, P Fuss & VR Young, includes unpublished data.
* P<0.05; ** P<0.01; *** P<0.001.

Taking into account these considerations on the analytical and calculation aspects of the doubly labeled water method, it appears clear that the technique provides an important new approach to determining energy requirements, but that the following recommendations should be taken into account when individual data sets are examined:

1. It is highly desirable that laboratories producing doubly labeled water data for the purpose of reevaluating requirements should have validated their technique against a reference method.
2. It is also desirable, when studying the effects of aging on energy requirements, that groups of individuals of different ages be included within the same study. In this way, differences due to age should be accurate, even if absolute mean values are subject to error.

Table 2 Number of male (M) and female (F) subjects in doubly labeled water studies including older individuals


Young

Elderly

Comments

Prentice et al (1989)

-

14F

UK; hospitalized mental patients

Goran & Poehlman (1992)

-

7M, 6F

USA; free living

Roberts et al (1992)

14M

15M

USA; free living 'metabolic balance'

Reilly et al. (1993)

-

10F

UK free living

Sawaya et al (1995)

10F

10F

USA; free living

Pannemans & Westerterp (1995)

19M, 10F

16M, 10F

Holland; free living 'metabolic balance'

Another issue related to the methodology for determining energy requirements is whether values for TEE should continue to be expressed as multiples of the pREE (i.e. the PAL index). To evaluate this issue, we examined Pearson correlation coefficients for several potential predictors of TEE in a population of 35 young and older men (Table 1). It can be seen that measured resting energy expenditure (mREE) was the best predictor of TEE in young and older men, and was also a significant predictor in young and older men when they were analyzed as separate groups. The pREE determined from body weight (FAO/WHO/UNU, 1985) was the second best predictor in the two age groups combined as well as a significant predictor in each age group separately. Body weight was an equally good predictor of TEE in young and older men separately (because the predicted REE was determined from body weight) but was not a significant predictor in young and older men combined because of the change in the relationship between REE and body weight that occurs with aging. These observations justify the continuing use of pREE as a means of normalizing TEE data.

Total energy expenditure of older individuals

To our knowledge, there are six studies of TEE in older individuals that have been published (Prentice et al, 1988; Goran & Poehlman, 1992; Roberts et al, 1992; Reilly et al, 1993; Sawaya et al, 1994; Pannemans & Westerterp, 1995). In these studies, a total of 38 free living men, 36 free-living women and 14 hospitalized male and female mental patients have been investigated (Table 2), and these within the younger segment of older individuals. In only three of the six studies (31 men and 20 women) were there measurements in young as well as in older subjects.

Table 3 gives a summary of relevant details of the subjects and data from the studies in older men (mean ages 64-74 years). The values for TEE are those reported by the authors, as well as 'adjusted' mean values (aTEE) which represent our effort to standardize calculation procedures between groups. As described in the legend to the table, the adjusted values all include a 2% increase in TEE to account for the expected lower rate of fractionated water loss in older individuals (Roberts et al, 1992). In addition, data of Goran & Poehlman (1992) were adjusted to take into account the fact that they used a mean measured ratio of the dilution spaces (2H2O: H2 18O) which was higher than the values used in the other studies. Although the necessity of using group-specific dilution space ratios has been considered, there is currently no available evidence to support the validity of this approach and, as noted by Goran & Poehlman (1992), it results in calculated values for TEE 10.5% lower than the use of a standard mean dilution space ratio of 1.03 (Schoeller et al, 1986).

Table 3 Published values for total energy expenditure in free-living older men measured using the doubly labeled water methoda

Authors

Goran & Poehlman (1992)d

Roberts et al (1992)

Pannemans & Westerterp (1995)c

Mean PAL

n

7

15

16


Age (y)

68 ± 6

69 ± 7

71 ± 5


Height (cm)

175 ± 9

177 ± 8

172 ± 9


Weight (kg)

77.1 ± 7.4

75.4 ± 9.7

74.0 ± 12.6


REE (kcal/d)





measured

1715 ± 183

1429 ± 157

1599


predictedb

1526 ± 100

1506 ± 130

1486


TEE (kcal/d)





measured

2675 ± 394

2495 ± 352

2412


adjustedc

3014 ± 444

2495 ± 352

2460


Mean TEE (kcal/kg)





measured

34.7

33.5

32.6


adjustedc

39.1

33.5

33.2


PAL values:





Mean TEE/mREE

1.58

1.75

1.51

1.61

Mean TEE/pREE

1.77

1.66

1.62

1.68

Mean aTEE/mREE

1.78

1.75

1.54

1.69

Mean aTEE/pREE

1.99

1.66

1.66

1.77

a Values are group means with ± s.d.. where data are available.
b Predicted values for REE determined from body weight using FAO/WHO/UNU (1985) equations.
c Adjusted values:
d for Goran & Poehlman (1992) are increased by 12.71% (10.5% for possible underestimation of TEE associated with using high ratio for the
2H: 18O dilution space, and 2% for lower rate of insensible water loss in elderly subjects as predicted by Roberts et al (1992));
e for Pannemans & Westerterp (1995) increased by 2% for lower rate of insensible water loss in elderly subjects as predicted by Roberts et al (1992). Note that the observed negative energy balance in this population is not accounted for in the adjusted TEE.

The reported and adjusted values for TEE in older men are given in Table 3, expressed as absolute values as well as ratios of mREE and pREE (FAO/WHO/ UNU, 1985). It can be seen that both the adjustment to TEE, as well as whether mREE or pREE is the chosen denominator, exert a significant influence on the PAL index in some individual studies. The ratio aTEE/pREE may represent the best estimate of energy requirements in this group. It should be noted that the aTEE/pREE ratio for the study of Goran & Poehlman (1992) differs significantly from data of the other two studies. Excluding the data of Goran & Poehlman (1992), a mean aTEE/pREE of 1.66 is obtained. This value is very slightly higher than the current recommended value of 1.51 for older individuals.

Data on the TEE of older women are given in Table 4. As with the studies in men, there is a relatively narrow range for mean age in these studies, ranging from 64 to 74 years. The reported and adjusted values for TEE are given in the table, expressed as absolute values as well as ratios of the mREE and pREE (FAO/ WHO/UNU, 1985). As seen in the data from men, both the adjustment to TEE and the use of mREE compared to pREE, exert a significant influence on the TEE/PEE ratio in some individual studies (Table 4). The mean aTEE/pREE ratio for all studies combined is 1.71, and in this case data from the study of Goran & Poehlman (1992) do not appear to lie outside the range of the other studies. The mean unadjusted TEE/pREE (1.64) is somewhat lower than the aTEE/pREE, but as with the data from men this value is higher than the current recommended value of 1.51 (FAO/WHO/UNU, 1985).

A combined summary of data from men and women is given in Table 5. Taken together, the studies on men and women provide no evidence of any influence of gender on TEE/pREE in older individuals, although it should be noted that the mean TEE/PEE ratio is indeed lower in women than in men in our own studies, which to our knowledge have involved the largest number of older subjects studied by any one group (Roberts et al, 1992; Sawaya et al, 1995). After excluding data that are influenced by the use of different doubly labeled water calculation procedures and the use of measured vs predicted values for REE, the weighted mean aTEE/pREE ratio for men and women combined is 1.66 and the mean TEE/pREE ratio is 1.64. Both of these values are higher than the current suggested value (1.51 for persons >50 years). It might be hypothesized that these values could be high because of a positive adaptation in energy expenditure to overeating during the TEE measurements. However, our recent overfeeding studies in young and older men have indicated that there is only a small increase in TEE with short term overfeeding (averaging 165 kcal/d) and this is due primarily to an increase in REE associated with increased fat and fat-free mass (Roberts et al, 1990). Thus, the ratio of TEE/PEE does not increase with overeating, and it would appear that adaptive variations in TEE due to overeating should not be of significant concern when using measurements of TEE to examine energy requirements.

Another important question concerning the energy requirements of older individuals is how TEE/PEE changes with age. Although there is little information on this issue, current recommended dietary allowances assume a modest decrease in TEE/pREE up to the age of 50 years (FAO/WHO/UNU, 1985). Our own studies in both young and older men and women (Roberts et al, 1992; Sawaya et al, 1995) support the suggestion of a modest decrease in TEE/pREE between the ages of 25 and 70 years. To quantify this change further, we performed a multiple regression analysis on our combined data set of 55 young and older men and women. The most significant predictor of TEE/pREE was percent body fat (since fat free mass is already accounted for through its close association with REE) and neither gender nor age was significant when included in models with percent fat (data not shown). Thus, the PAL index can be expected to fall with age in association with the usual increase in the proportion of body fat content.

Table 4 Published values for total energy expenditure in free-living older women measured using the doubly labeled water methoda

Authors

Goran & Poehlman (1992)d

Reilly et al (1993)e

Sawaya et al (1993)

Pannemans & Westerterp (1995)f

Mean PAL

n

6

10

10

10


Age (y)

64 ± 5

73 ± 3

74 ± 2

67.6 ± 4.1


Height (cm)

165 ± 3

-

155 ± 5

160 ± 8


Weight (kg)

65.2 ± 7.8

60.0 ± 7.2

58.5 ± 9.9

65.4 ± 5.9


REE (kcal/d)






measured

1472 ± 129

1221 ± 91

1144 ± 103

1219


predictedb

1281 ± 82

1226 ± 75

1210

1283


TEE (kcal/d)






measured

2092 ± 231

2201 ± 354

1852 ± 214

2065


adjustedc

2358 ± 261

2245 ± 354

1852 ± 214

2165


Mean TEE (kcal/kg)






measured

32.1

37.0

31.7

30.9


adjustedc

36.2

37.7

31.7

31.6


PAL values:






Mean TEE/mREE

1.43

1.80

1.62

1.66

1.63

Mean TEE/pREE

1.64

1.80

1.53

1.58

1.64

Mean aTEE/mREE

1.62

1.84

1.62

1.66

1.69

Mean aTEE/pREE

1.85

1.84

1.53

1.61

1.71

a Values are group means with ± s.d. where data are available.
b Predicted values for REE determined from body weight using FAO/WHO/UNU (1985) equations.
c Adjusted values:
d for Goran &: Poehlman (1992) are increased by 12.71% (10.5% for possible underestimation of TEE associated with using high ratio for the 2H: 18O dilution space, and 2% for lower rate of insensible water loss in elderly subjects as predicted by Roberts et al (1992));
e for Reilly et al 1993) increased by 2% for lower rate of insensible water loss in elderly subjects as predicted by Roberts et al (1992);
f for Pannemans & Westerterp (1995) increased by 2% for lower rate of insensible water loss in elderly subjects as predicted by Roberts et al (1992). Note that the observed negative energy balance in this population is not accounted for in the adjusted TEE.

Table 5 Summary of weighted means for PAL indicators in older men and women



Men & women


Men

Women

All

Selected studiesa

TEE/PEE

1.61

1.63

1.62

1.67

TEE/pREE

1.68

1.64

1.66

1.64

aTEE/REE

1.69

1.69

1.69

1.68

aTEE/pREE

1.77

1.71

1.75

1.66

a Excluding data of Goran & Poehlman (1992), which are strongly influenced by the use of data for adjusted TEE and pREE.


Figure 1
Relationship between the ratio of total to resting energy expenditure and reported, self-defined strenuous physical activity in healthy elderly men.

Relationship between total energy expenditure and physical activity

The measurements of TEE described above indicate that the measured TEE/pREE of the groups of subjects studied (mean ages 64-74 years) was higher than the current recommended value for the age group > 50 years. This may indicate that current recommendations for energy (FAO/WHO/UNU, 1985) underestimate the usual energy need of older adults. However, it is alternatively possible that the physical activity of the subjects in those studies was unusually high. The lack of standardization in reporting of physical activity hampers an assessment of this issue. Concerning the studies conducted in our own laboratory, the mean values for self-reported durations of strenuous activity were 29 and 48 min/day for men and women, respectively (Roberts et al, 1992; Sawaya et al, 1995). These values included means of 4 and 7 min/day, respectively, for activities with predicted energy expenditure of >5 × REE. Thus, these groups of subjects did not appear to be unusually active, indicating that the 1985 recommendations on energy requirements may indeed be lower than usual energy needs. However, further research is needed to confirm this suggestion. In the study of Reilly et al (1993), a mean duration of leisure time activity of 40 min/day was reported together with 87 min/day of walking. These mean activity data appear relatively high, in keeping with the high mean PAL index in this group (aTEE/pREE = 1.84).

Another factor to consider in the determination of recommended energy requirements for older individuals is the issue of normal vs desirable levels of physical activity. Physical activity is the major determinant of variability in TEE/PEE between individuals and can potentially have a major effect on energy requirements. Figure 1 illustrates the relationship between strenuous physical activity (defined as the sub-set of self-reported strenuous activity with an expected mean energy cost of >5 × REE) and TEE/pREE (equal to aTEE/pREE) in our population of older men. It can be seen that, as expected, high levels of reported strenuous activity were associated with increased TEE/pREE. Individuals who reported no strenuous physical activity had a mean value for TEE/pREE of 1.70 while those who reported 30 min/day of strenuous physical activity had a mean TEE/PEE ratio of 1.90. Although it is widely accepted that a sedentary existence is associated with increased morbidity and mortality (Paffenbarger et al, 1986; Blair et al, 1989; Helmrich et al, 1991; Paffenbarger et al, 1993), there is no general consensus over what levels and types of physical activity are optimal for long-term health (Paffenbarger et al, 1986; Blair et al, 1989; Helmrich et al, 1991; Paffenbarger et al, 1993). For this reason, it is appropriate to continue to make recommendations for different levels of physical activity, and also to recommend a minimum activity level that makes allowance for some strenuous physical activity for cardiovascular maintenance.

Effects of energy balance on nitrogen balance

Reliable measurement and knowledge of energy requirements of the elderly is essential to the accurate determination of protein requirements. Negative energy balance results in negative N balance (because lean tissue is mobilized along with fat and glycogen), and positive energy balance promotes positive N balance. A summary of data from several N balance studies by Pellett & Young (1992) indicated that even very small discrepancies in energy balance can have a major impact on the outcome of N balance: an energy imbalance of only 1 kcal/kg changes N balance by 1 mg/kg. It is not possible to measure the energy balance of individual subjects to 65 kcal/d in short-term studies ( < 30 days), which means that data on N balance from individual subjects should not be used to judge the adequacy of protein intake. Such data can only be used as the average of groups in which small variations in energy intake above and below actual energy expenditure can be expected to cancel out. Because energy requirements may have been underestimated in many previous studies, there is a potential bias towards overestimation of protein requirements in N balance experiments in older subjects.

Summary of energy requirements in older individuals and recommendations for future research

The successful validation of the doubly labeled water technique (Schoeller, 1988; Roberts, 1989) has provided an improved means for determining TEE and hence energy requirements in different population groups. Currently available data from 74 subjects (mean ages 64-74 years) studied by four research groups indicate

that the energy requirements of older individuals may be higher than indicated in the current recommended energy requirements (FAO/WHO/UNU, 1985), with mean PAL values for both men and women being 1.641.68. This observation should not be taken as a recommendation for increasing energy intake, but rather as an acceptance of the fact that both energy intake and energy expenditure may be higher than anticipated previously in this age group. Data from our own laboratory also indicate that aTEE/pREE ratios for older adults are lower than those for young men and women by approximately 0.2 PAL on average. Based on reported information on physical activity, it appears likely that these high values for TEE/REE reflect a previous underestimation of energy requirements by the factorial approach rather than unusually high levels of physical activity in the subjects studied. It is important to note that these observations are based on studies in a very limited number of older subjects and many more doubly labeled water measurements are required to provide the information needed for a detailed assessment of changes in energy requirements with age. Future studies are needed to address the following issues:

1. Concerning the methodology of determining energy requirements, more detailed calculations of changes in REE with age after 50 years are required in order to normalize measurements of TEE, considering indices of height as well as weight. It is important to use predicted rather than :measured values for REE when determining TEE/REE ratios because otherwise the TEE/PEE ratio can be substantially influenced by between-laboratory differences in measurement of REE as indicated in Tables 3 and 4. In addition, further standardization of doubly labeled water calculation procedures and more detailed between-laboratory standardization of the doubly labeled water method is needed.
2. Individuals over the age of 50 years are not a homogeneous group with regard to energy requirements. Currently available data are restricted to a relatively narrow age range within the younger section of the older population (mean ages 64-74 years), and much more information is required on changes of TEE with age in individuals in each age decade within the range 50-90 years. Both men and women should be investigated in these studies, as well as young control subjects, so that the range of changes in energy requirements with age can be assessed accurately.
3. There is currently no general consensus over the desirable range of physical activity, and hence energy requirements, for long-term health in older individuals. However, it is clear that a sedentary lifestyle is associated with numerous adverse health consequences and should not be recommended. Although further information on desirable levels of physical activity will necessarily come from investigations with much larger numbers of subjects than are feasible for doubly labeled water measurements, additional studies are needed to define the relationship between TEE and simple measures of physical activity that can be used in larger population studies to predict the energy requirements of representative populations.

References

Blair SN, Kohl HW, Paffenbarger RS, Clark DO, Cooper KH & Gibbons LW (1989): Physical fitness and all-cause mortality: a prospective study of healthy men and women. JAMA 262, 2395/2401.

FAO/WHO/UNU (1985): Energy and protein requirements. Report of a joint FAO/WHO/UNU expert consultation. Technical Report Series 724. Geneva: World Health Organization.

Goran Ml & Poehlman ET (1992): Total energy expenditure and energy requirements in healthy elderly persons. Metabolism 41, 744-753.

Helmrich SP, Ragland DR, Leung RW & Paffenbarger RS (1991): Physical activity and reduced occurrence of non-insulin dependent diabetes mellitus. New Eng. J. Med. 325, 147-152.

National Research Council (1989): Recommended Dietary Allowances, 10th edn. Washington, DC: National Academy Press.

Paffenbarger RS, Hyde RT, Wing AL, Lee I-M, Jung DL & Kampert JB (1993): The association of changes in physical-activity level and other lifestyle characteristics with mortality among men. New Eng. J. Med. 328, 538-545.

Paffenbarger RSJ, Hyde PHR, Wing A & Hsieh CC (1986): Physical activity, all-cause mortality, and longevity of college alumni. New Eng. J. Med. 314, 605-613.

Pannemans DLE & Westerterp KR (1995): Energy expenditure, physical activity and basal metabolic rate of eldery subjects. Brit. J. Nutr. 73, 571-581.

Pellett PL & Young VR (1992): The effects of different levels of energy intake on protein metabolism and of different levels of protein intake on energy metabolism: a statistical evaluation from the published literature. In Protein-energy interactions, eds NS Scrimshaw & B Schürch Lausanne, Switzerland: International Dietary Energy Consultative Group.

Prentice AM, Leavesley K, Murgatroyd PR, Schorah CJ, Coward WA & Blaydon P (1988): Is severe wasting in elderly mental patients caused by an excessive energy requirement? Age Ageing 18, 158-167.

Prentice AM (ed.) (1990): The doubly-labelled water method for measuring energy expenditure. Technical recommendations for use in humans. Viena: International Atomic Energy Agency.

Reilly JJ, Lord A, Bunker VW, Prentice AM, Coward WA, Thomas AJ & Briggs RS (1993): Energy balance in healthy elderly women. Br. J. Nutr. 69, 21-27.

Roberts SB (1989): Use of the doubly labeled water method for measurement of energy expenditure, total body water, water intake and metabolizable energy intake in humans and small animals. Can. J. Physiol. Pharm. 67, 1190-1198.

Roberts SB (1995): Influence of age on energy requirements Am. J. Clin. Nutr. 62, (Suppl.), 1053S-1058S.

Roberts SB, Dietz W. Sharp T. Dallal GE, Hill JO (1995): Multiple laboratory comparison of the doubly labeled water method. Obesity Research 3 (Suppl.), 3-14.

Roberts SB, Young VR, Fuss P. Heyman MB, Fiatarone M, Dallal GE, Cortiella J & Evans WJ (1990): Energy expenditure and subsequent nutrient intakes in overfed young men. Am. J. Physiol. 259, R461-R469.

Roberts SB, Young VR, Fuss P. Fiatarone MA, Richard B. Rasmussen H. Wagner D, Joseph L, Holehouse E & Evans WJ (1992): What are the dietary energy needs of elderly adults? Int. J. Obesity 16, 969-974.

Sawaya AL, Saltzman E, Fuss P. Young VR & Roberts SB (1995): Dietary energy requirements of young and older women determined by using the doubly labeled water method. Am. J. Clin. Nutr. 62, 338-344.

Schoeller DA (1988): Measurement of energy expenditure in free living humans by using doubly labeled water. J. Nutr. 118, 12781289.

Schoeller DA, Ravussin E, Schutz Y. Acheson KJ, Baertschi P, Jéquier E (1986): Energy expenditure by doubly labeled water validation in humans and proposed calculation. Am. J. Physiol 250, R823-R830.

Discussion

Definition and classification of 'older' individuals

During the discussion, several attempts were made to define what is meant by 'older' individuals or the 'elderly'. It was recognized that the use of descriptive terms like these can vary enormously from country to country and in different contexts, and that it is therefore preferable to specify the age range when describing such a population. It was further agreed that one could not make general recommendations for adults ranging in age from 60 or 65 to over 100 years, and that this whole range needed to be subdivided into narrower age bands.

The discussion on age ranges and age limits was, however, eclipsed by the realization that, in a discussion of energy and protein requirements, health and life style were of greater importance than age, and that, from that point of view, the elderly were an extremely heterogeneous group: on the one hand they are freed from the constraints imposed by a professional activity and can choose life styles implying widely varying physical activity levels; on the other hand a considerable proportion of them are likely to suffer from various disabilities that restrict their activity or even make them bedridden. This implies that recommendations have to cover a wide range, and that mean values may not offer valid information. Stratifying the elderly by age subgroups and activity level may be the most useful approach.

Formulating energy requirements for the elderly is particularly important since a relatively large proportion of this age group is institutionalized and fed meals planned by dietitians relying on such recommendations. A recent survey in the UK revealed that over 50% of institutionalized elderly subjects ate all the food provided to them and still lost LBM, a clear indication that their dietary requirements were not met (Prentice).

Total energy expenditure (TEE)

Roberts draws her conclusions from 74 measurements of TEE made by DLW in healthy and relatively active and lean individuals from industrialized countries, in an age range from 64 to 74 years; this limits the external validity (generalizability) of her conclusions considerably.

Most discussants find that DLW provides an interesting new approach, but they emphasize that this should not be taken to mean that older data used to infer energy requirements of the elderly are invalid and ought no longer be considered.

Currently available TEE data, obtained by the DLW method and compiled by Prentice, suggest that, throughout the whole life span, energy intakes should be slightly higher than those recommended in the 1985 report. Roberts attributes this to the fact that the 1985 recommendations relied heavily on the factorial method which, according to her, leaves certain activities unaccounted for and therefore leads to an underestimation of TEE. Other participants (e.g. Shetty and Ferro-Luzzi) believe that the differences are more likely to be attributable to a bias in the DLW data base.

Ferro-Luzzi has BMR and energy expenditure data (obtained by observation of activities and multiplication by their energy cost) of some 600 free-living people in Italy. She and others found that mean values of TEE decline with age, due to a decline of both BMR and physical activity.

Basal metabolic rate (BMR)

Throughout her paper, Roberts used the term resting energy expenditure (REE), because she was not sure if the metabolic rate measurements in the studies from which she obtained her data were all carried out under the strict conditions required of BMR measurements. Such precaution may be appropriate in this specific instance, but several discussants re-emphasized that REE is not a well defined concept, and that measuring and using BMR is preferable.

Schofield et al's data base contained only very few BMR values of individuals over 60, and their predictive equations therefore rest on an insecure basis. Henry and Durnin's ongoing re-analysis will include more data points, but their number remains inadequate; in particular, there are only very few data from this age group in less developed countries.

Ferro-Luzzi compared measured with predicted BMR values and found that, in the 60 to 75-year age group, predicted values exceed measured ones by about 10%.

A biologically interesting question is, whether changes in body mass and body composition in the elderly fully explain changes in BMR, as they do in younger adults, or if the aging process involves other metabolic changes that have to be taken into consideration. It is difficult to answer this question, because of the scarcity of body composition data from elderly people. Fat is more internal than subcutaneous, so that skinfolds become poor indicators of fat mass, and for practical reasons densitometric measurements have not been frequently made in this age group. Furthermore, aging results in changes within the fat-free mass that can have energy-metabolic effects, but on which sufficient information is not yet available. BMR per kg LBM seems to decrease slowly with age. This appears somewhat surprising. Since the elderly lose more muscle (with a relatively low metabolic rate at rest) than organ tissue (with a relatively high metabolic rate), one could expect BMR to increase in relation to body weight or LBM, but this phenomenon has actually only been observed in very undernourished, terminal patients.

Practically, the issue of body composition changes may be of lesser importance, since only few investigators will be able to make body composition measurements, and body composition changes that are typical of the elderly will be reflected in the predictive equations for BMR of this age group.

Inter-individual variation in BMR in the elderly is similar to that in older adults: a little over 10% (Ferro-Luzzi).

Physical activity level (PAL)

Roberts took adjusted values of total energy expenditure (TEE) (obtained by DLH method) and divided them by predicted BMR to obtain PAL values, which were higher than the average PAL value used to derive energy recommendations in the 1985 report. Several discussants argued that it would have been better to divide TEE by measured rather than predicted BMR.

Even though the PAL of certain people increases after retirement (Prentice), on average it decreases with age. Roberts reported a drop from 1.85 in 20-year-olds to 1.65 in 65-year-olds; Prentice estimated the mean drop of PAL in the elderly at about 0.1. In certain situations, e.g. in institutionalized people or elderly in poorer countries, energy deficiency should always be taken into consideration as a possible cause of low PALs.

Does the energy cost of activities change with increasing age? According to Durnin and Hautvast increases in energy cost have been observed in activities involving the whole body, like walking on a treadmill, where elderly people show an increasing number of extraneous movements.