Mechanisms of Changes in Basal Metabolism During Ageing

Mechanisms of changes in basal metabolism during ageing

CJK Henry1*

1Department of Nutrition and Food Science, School of Biological and Molecular Sciences, Oxford Brookes University, Oxford, UK

A considerable number of physiological functions are known to show a gradual decline with increasing age. However, the effects of ageing differ widely between organ systems. It is believed that basal metabolic rate (BMR) falls dramatically with age. These observations, largely based on cross-sectional surveys, are discussed in light of our present understanding of the biology of ageing. This paper reviews both the longitudinal and cross- sectional studies of BMR and presents evidence that the fall in BMR with ageing may be less dramatic than previously perceived. Indeed, some subjects may show an increase in BMR with ageing. The mechanism of changes in BMR during ageing will be discussed. Organ weight changes appear to have a profound impact on BMR. The use of BMR to predict total energy expenditure in the `old elderly' ( > 75 y) is unlikely to be of any practical use due to wide intra- and inter-individual variation in BMR. This wide intra- and inter-individual variation in BMR is due to illness, disease and other metabolic disorders seen in the elderly. Finally, the importance of measuring BMR in elderly populations for its use in clinical medicine will be discussed. Descriptors: aging; basal metabolic rate; body composition European Journal of Clinical Nutrition (2000) 54, Suppl 3, S77±S91

Historical overview The principle embodied in the surface law was that basal metabolism is a simple function of surface area. Hence, if It is no exaggeration to claim that the science of nutrition surface area was computed, basal metabolism could be was founded on the study of energy metabolism. Tata calculated. The surface law was to have a powerful and (1964) de®ned basal metabolism rate (BMR) as the sum dominant hold on the estimation of BMR for over half total of the minimal activity of all tissue cells of the body a century. While the surface law remained in¯uential under steady state conditions. Similarly, Mitchell (1964) amongst many physiologists during the early part of this commented that `the basal metabolism of an animal is the century, it also attracted considerable opposition, notably minimal rate of energy expenditure compatible with life.' from Benedict and his co-workers. Benedict's objections to Magnus-Levy coined the term Grundumsatz or `basal the surface law were both biological and practical. In 1915, metabolism' in 1899. This term was of great value to the Benedict had introduced the concept of `active protoplas- early investigators, as it emphasised the need to conduct mic mass', and approached the problem of metabolism, not experiments under strictly standardized conditions. Krogh from the perspective of heat loss, but heat production. (1916) proposed the term `standard metabolism' to circum- Thus, Benedict proposed that BMR was proportional to vent the erroneous impression that basal metabolism is the protoplasmic mass. However, on a practical level, it soon lowest resting metabolism, ie the lowest that could be became clear that the measurement and estimation of sur- obtained by an individual. Although the term `standard face area was far from easy and satisfactory. By 1916, metabolism' has not gained universal usage, the meaning of DuBois and DuBois were to admit that `the validity of the basal metabolism has been widely accepted. surface law rests on the accuracy of determining surface area in humans'. Despite this cautionary note, they then went on to develop the ®rst reference standards for BMR Review of early work on BMR standards and predictive based on surface area! equations The `basal metabolism', namely the total of heat production DuBois height ± weight formula chart under standardised conditions, is closely related to oxygen consumption. The view that heat is primarily produced in While surface area may be calculated using various anthro- an animal to keep warm was ®rst proposed by Bergman in pometric parameters, DuBois and DuBois (1916) produced 1843 (Lusk, 1928). Forty years later, Rubner presented an equation relating weight and height to surface area as `supportive' evidence by demonstrating that the heat pro- follows: duction in a living organism is directly proportional to the A W 0:425 Â H 0:715 Â 71:84 surface area. This was the origin of the famous `surface rule=law' of Rubner, which he explained in terms of where A surface area (cm2), W weight (kg) and homeothermy. H height (cm). This was derived from surface area measurements in *Department of Nutrition and Food Science, School of Biological and nine subjects ranging in body weight from 6 to 93 kg. Later, Molecular Sciences, Oxford Brookes University, Gipsy Lane, Oxford OX3 0BP, UK. Aub and DuBois (1917), applying the surface law, of heat loss published a table of BMR=m2=h from 14 to 80 y (Table E-mail: [email protected] 1). This table is still widely used despite being based on a Mechanisms of changes in BMR CJK Henry S78 group of only nine subjects and one cadaver. The values socio-economic status and levels of physical activity. consisted of a tabulation for males and females between Longitudinal studies (Rinder et al, 1975) have shown that 14 ± 80 y by 2 y periods from 14 ± 20 y, and by decades many physiological parameters remain relatively constant from 20 to 80 y. All values were expressed as Cal=m2=h. right through middle life, and show a decline at around The major criticism of the Aub and DuBois standards is the 70 ± 75 y. Without considering the additional issue of fact that the values only begin with subjects older than 14 y, sexual dimorphism in the rate and timing of the age-related and are too `general' for ages beyond 20 y. For example, changes in physiology and body composition (Chumlea & the BMR of a 20 and 39 y old man are the same, con®rming Baumgartner, 1989), as a ®rst approximation we may that the standards do not take into account the likely age- classify the elderly as: related changes in BMR. `young elderly' 65 ± 75 y `old elderly' > 75 y BMR and ageing This classi®cation may be a practical consideration Among the earliest to investigate energy exchange in when we attempt to measure and use BMR values to humans at various ages were Magnus-Levy and Falk, estimate total energy expenditure in these subjects. who in 1899, studied seven men. These results were subsequently used by Aub and DuBois (1917) to draw a stylized curve of the average level of metabolism at various Speci®c problems in the assessment of BMR in the ages (Figure 1). elderly At the time of this analysis, these authors commented on Benedict (1938) outlined a comprehensive list of conditions the paucity of data on subjects after the age of 43. In an to be imposed when BMR was to be measured. attempt to rectify this situation, Aub and DuBois (1917) These included the following: studied six men aged between 77 and 83 y. The subjects for the study came from a New York City home for the aged. 1. Absence of gross muscular activity. The authors commented that the subjects were `fairly well 2. Post-absorptive state. nourished, though on a plain and scanty diet'. This raises 3. Zone of thermoneutrality. the question of sample choice and selection of the elderly 4. Minimal emotional disturbance. for studies on BMR ± a point we shall return to later on. 5. Wakefulness. 6. Normal nutritional status. 7. Absence of disease or infection. Problems in the de®nition and classi®cation of the elderly Whilst these conditions may be relatively hard to comply with even in the young and middle aged, it is The term elderly, currently de®ned as 65 y or older, even harder to satisfy all these conditions in the elderly. encompasses a heterogeneous group with varying health, More importantly, conditions (1), (6), and (7) may be nearly impossible to achieve in all elderly subjects. For example, minor muscular movement appears to increase Table 1 DuBois normal standards for BMR (Cal=m2=h) BMR dramatically as shown in Table 2. Age (y) Males Females By de®nition, BMR is measured in the thermoneutral zone where there is no cold or heat stress to in¯uence 14 ± 15 46.0 43.0 metabolism. Little is known about the `thermoneutral zone' 16 ± 17 43.0 40.0 18 ± 19 41.0 38.0 in the elderly. For example Horvath et al (1955) found that, 20 ± 29 39.5 37.0 when young men were exposed to 10 C, this produced 30 ± 39 39.5 36.5 shivering and an increased metabolic rate but the elderly 40 ± 49 38.5 36.0 showed no response at all. Numerous studies have sug- 50 ± 59 37.5 35.0 gested an association between levels of physical activity 60 ± 69 36.5 34.0 70 ± 79 35.5 33.0 and BMR. Indeed, Poehlman and colleagues (Poehlman & Danforth, 1991; Poehlman, 1996; Poehlman, 1998) have

Figure 1 Changes in BMR with age expressed as Cal=m2=h.

European Journal of Clinical Nutrition Mechanisms of changes in BMR CJK Henry S79 Table 2 In¯uence of small muscular movements on oxygen Table 3 Measurements of basal metabolism in subjects over 50 y reported consumption at rest between 1899 ± 1936

O2 consumption Percentage Number of subjects by age group Activity (cm3=min) increase Date Authors Gender 50 ± 69 70 ± 89 > 90 Total Basal 188 ± - Arm movement 218 15.9 1899 Magnus-Levy & Falk M 2 4 0 6 Basal 189 ± - 1916 DuBois M 1 0 0 1 Arm movement 224 18.5 1917 Aub & DuBois M 0 6 0 6 Leg movement 209 10.6 1919 Harris & Benedict M 5 0 0 5 F142016 Adapted from Carpenter (1933). 1922 Boothby & Sandiford M 7 0 0 7 F110011 1928 Benedict M 3 1 0 4 shown that BMR in the active elderly was 6% higher than F5005 the inactive when normalized for FFM. The impact of 1932 Benedict & Meyer F 2 21 0 23 1934 Kise & Ochi M 28 15 1 44 levels of physical activity on BMR is an additional factor F2426050 to be considered especially in the elderly. 1934 Benedict & Root M 0 0 1 1 Freedom from illness is even more dif®cult to achieve. 1934 Matson & Hitchcock M 0 13 1 14 In England and Wales 8% of men and 4% of woman aged F0628 67 ± 74 consult their general practitioners for ischaemic 1935 Benedict M 0 5 0 5 F19010 heart disease in a year (Of®ce of Population Census and 1936 Boothby, Berkson & Dunn M 80 0 0 80 Surveys, 1994). Moreover, in a Swedish study (Svanborg, F 138 0 0 138 1988) 60 ± 70% of elderly subjects had some form of ill health. With increasing prevalence of illness and disease affecting the elderly, how valid is the measurement of BMR in these subjects? It is now well recognized that where h kcal=day; W weight in kg; S stature in cm; illness and disease increase basal metabolic rate dramati- A age in years. cally. Harris and Benedict's analysis marked a signi®cant departure from previous work. First, it introduced for the Early BMR studies in the elderly ®rst time biometric principles in its analysis. Secondly, they used subjects that were maintained under strict experimen- Table 3 presents a chronological sequence of BMR studies tal conditions prior to the measurements. Using partial conducted in the elderly between 1899 and 1935. During correlation coef®cients, they also showed that both stature this period, the only paper from a non-European group and weight have an independent effect on BMR, a point came from Kise and Ochi (1934), who studied 94 Japanese that has since been largely ignored. While these equations males and females aged between 50 and 93 y, and data of were useful and valuable aids to predicting BMR, they subjects over the age of 70 were scarce. were not above criticism. For example, the constant in the Just before World War II, Lewis (1934) presented the equation showed a 10-fold difference between males and most comprehensive study of BMR in the elderly. He females (66 vs 665). Benedict himself later recognized, and studied 100 men, 20 in each decade, between 40 and expressed concern, that the equations overestimated BMR, 89 y. In addition, he also reported the BMR in 7 men `particularly in those young women'. Daly et al (l985) aged 91 ± 101 y. He was the ®rst to systematically compare con®rmed that the Harris and Benedict (1919) equations the regression analyses observed for BMR (Cal=m2=h) vs overestimated BMR by about 10 ± 15%. Despite the wide age with Harris ± Benedict predictive equations (1919), and differences in the constants seen in Harris ± Benedict equa- Boothby and Sandiford (1929) which were derived from a tions, if we extrapolate these regression lines, men and much younger populations. The equation presented by women reach almost the same level of BMR (1200 ± Lewis was: 1280 kcal=day) at the age of 82. BMR C39:138 0:0678A The data presented by Harris and Benedict (1919) came from a series of four studies conducted between 1914 and where C Calories per m2=h; A age in years. 1918. These comprised of 136 men and 103 women aged between 16 and 63 y in men, 15 and 74 y in women. Figure 2 shows the scatter diagrams redrawn from the publication Equations to predict BMR in the elderly A Biometric Study of Basal Metabolism in Man (1919). Mention has already been made of the predictive equation Despite the limited number of subjects older than 50 y, the reported by Lewis (1934). The ®rst researchers to include simplicity of the Harris ± Benedict equation made it a age as an independent variable in computing BMR were popular equation widely used even today by many clin- Harris and Benedict (1919). Their detailed biometric ana- icians in North America (Daly et al, 1985; Roza & Shizgal, lysis of BMR culminated in the publication of their monu- 1984; Mif¯in et al, 1990). Harris and Benedict (1919) went mental work entitled `A Biometric Study of Basal on to say Metabolism in Man'. BMR measurements were made on `The results are as good as or better than those obtain- 136 males and 103 females at the Carnegie Nutrition able from the constant of basal metabolism per square Laboratory. Using rigorous statistical concepts, they devel- metre body surface area. Values obtained by applying oped the following equations to predict BMR: the biometric formulae involve no assumptions concern- for males, h 66.4730 13.7516W 5.0033S 7 6.7750 A ing the derivation of surface area, but are based on for female, h 665.0955 9.5634W 1.8496S 7 4.6756A direct physical measurements'

European Journal of Clinical Nutrition Mechanisms of changes in BMR CJK Henry S80 a signi®cant issue that we shall discuss further. Fleisch These became known as the Mayo foundation standards (1951) graphically presented the result of 24 data sets and are illustrated in Figure 4. A widely held view that (representing several thousand observations) of BMR emerged from these ®gures was that BMR progressively made on subjects at various ages (Figure 3). decreases with age. This view was largely based on an Figure 3 shows that the BMR in early childhood is interpretation of the `BMR curves' drawn over a wide age approximately 53 Cal=m2=h with a sharp decline until the range. For example, the smoothed `Mayo clinic standard age of 20. This is followed by a gradual fall in BMR after curves' (Figure 4) show a rapid fall in BMR between the stage of maturity. A gender difference in BMR appears infancy and childhood, followed by a linear decline from early in life and is retained into senescence. In 1936, 24 to 64 y. The changes in BMR with age have been Boothby, Berkson, and Dunn (Dubois, 1936) made a care- repeatedly presented in the literature without any substan- ful study of the heat production of 639 males and 829 tive biological or physiological support to parallel these females. The children were from a school in Rochester and changes. For example, the changes in BMR with age do not the adults were personnel working at the Mayo Clinic. coincide with any well-known physiological landmark,

Figure 2 Changes in BMR with age expressed as kcal=day. (a) Males, and (b) females.

European Journal of Clinical Nutrition Mechanisms of changes in BMR CJK Henry S81

Figure 3 Changes in BMR with age expressed as Cal=m2=h.

Figure 4 Mayo standards for changes in BMR with age expressed as Cal=m2=h. including the pubertal spurt in height and weight. It appears during ageing. Indeed Keys et al (1973) went on to add `If counter-intuitive to observe a fall in BMR (at puberty) at a the major interest is to evaluate the intensity of metabolism time of rapid growth in lean body mass and adipose tissue. in the individual, some reference unit external to his own For over half a century, body surface area dominated the metabolism is essential'. The estimation of surface area has form of expressing BMR in humans. As early as 1927, an unknown error associated with it. For example, Mitchell DuBois commented that it was necessary to reconsider `the et al (1971) found that the DuBois formulae consistently now outmoded concept that heat loss determines heat underestimated surface area in adult males compared to production'. It is evident that the uncanny agreement true values measured using photometric methods. Sec- observed in applying the surface law to express BMR in ondly, little is known about the nature and extent of error humans (see Figures 3 and 4) appears to have served to associated with using a single equation to predict surface inhibit, rather than stimulate, an interest in de®ning the area in the very young (children and infants) and the very possible differences in metabolic rate between individuals old. Body surface area is usually calculated using the

European Journal of Clinical Nutrition Mechanisms of changes in BMR CJK Henry S82

Figure 5 Changes in BMR with age from the Oxford database expressed as kcal=kg=day (males).

DuBois equation and is rarely measured. One speculation is Basal metabolic rate and body weight that these dramatic age-related changes in BMR (presented It is obvious that body size is a dominant factor in in Figures 3 and 4) are an artefact that arises from expres- determining BMR. Table 4 presents the BMR in humans sing BMR per unit surface area (Dubois, 1927). at various body weights and compares them with animals of The view that heat loss (hence surface area) (DuBois, comparable body weights. The overall intensity of meta- 1919) was the determinant of BMR has now given rise bolism expressed per day or per unit body weight appears to the opinion that it is heat production, and notably the to be similar in man and other homeotherms. This enables organs responsible for heat production, that are the basis for us to make some broad generalizations regarding the effects BMR (Elia, 1992). In view of these criticisms, it is appro- of age on BMR in homeotherms. priate to examine body weight as the reference unit and examine the changes in BMR with age when values are expressed as BMR=kg=day. Figure 5 shows such an analysis. Figure 5 was generated using the Oxford BMR database, which is composed of 8646 data points for males and Is the change in BMR with age unique to man? 5302 for females (from birth to 103 y). The values were collated from the literature, personal communication, and Brody and Proctor, in their publication `Growth and Devel- from our own laboratory measurements. opment' (1932), compared the age changes in basal meta- The most dramatic differences between Figures 3 and 5, bolic rate in humans, pigs, sheep, cattle, rats and chicken. is the sharp fall in BMR (when expressed as BMR=kg=day) They commented that the values for humans, rat and between the ages of 1 and 10 and a `stable' BMR after the chicken represented true senescence, and we use domestic age of 25. It is clear from these ®gures that the fall in BMR fowl as an example. A comparison of Figures 6 and 7 with ageing is less dramatic than has been previously illustrates the point that the fall in BMR with age follows a perceived. At this stage it is important to recognise that regular pattern in homeotherms. A similar analysis by these values were obtained from a cross-sectional analysis. Brody (1945) in humans showed a sharp decline in BMR The use of cross-sectional data to analyse changes in BMR, between 1 and 25 y with little or no change between the speci®cally in the elderly may cloud and hide some years of 40 ± 80. Taken together these results illustrate a important metabolic differences during ageing. `stable' BMR in homeotherms after they reach maturity.

Table 4 Basal metabolic rate and body size in man and other homeotherms

Species Weight (kg) BMR (kcal=day) BMR (kcal=kg=day) Humans Weight (kg) BMR (kcal=day) BMR (kcal=kg=day)

Rabbit 3.5 160 46 Birth 3.5 161 46 Goose 5.0 276 55 Child 3 months 5.5 300 54 Dog 14.9 542 36 Child 4 y 15.0 700 46 Goat 36.1 897 25 Child 10 y 30.0 1070 36 Sheep 45.0 1254 28 Child 12 y 45.0 1173 26 Man 65.0 1667 25 Man 65.0 1667 25

Source: Adapted from Benedict (1938), Holliday et al (1967) and Henry et al (1999).

European Journal of Clinical Nutrition Mechanisms of changes in BMR CJK Henry S83

Figure 6 Changes in BMR with age expressed as Cal=kg=day.

Cross-sectional studies Several cross sectional studies have reported a lower BMR in the elderly (Calloway & Zanni, 1980; Tzankoff & Norris, 1977) compared to the young. An interesting paper by Morgan and York (1983) reported a similar value for fat-free mass (FFM) in the young and old despite Figure 7 Changes in basal metabolism with age of domestic fowl. a lower BMR in the elderly. This raises the possibility that the lower BMR in the elderly may have been due to have shown a much smaller fall in fat free mass with age differences in their organ weight or changes in thyroid (Chumlea et al, 1998) as shown in Table 5. This modest hormone concentration (as these authors speculated). The decline in FFM coincides with the modest fall in BMR seen persistent focus on the changes in FFM with ageing and its during ageing. Recently, Murray et al (1996) reported a impact on BMR appears to have ignored other physiologi- modest fall in BMR and lean body mass over a 65 y period cal factors that may in¯uence BMR. in a group of 22 active men aged 54 ± 72 y. The authors Shock and Yiengst (1955) reviewed the literature on the concluded that `the physically active elderly in good health cross-sectional studies of BMR and ageing in men between in this age range show a very small age related decline in 30 and 89 y and presented an equation to predict BMR in BMR and fat free mass'. Keys et al (1973) also reported the form: that cross-sectional data greatly overestimated the true age affect on BMR. They concluded that the reduction in BMR BMR kcal=m2=h40:22 0:11X where X age. Table 5 Mean values for body composition variables in healthy elderly New Mexico Ageing Process Study participants (Chumlea et al, 1998) Longitudinal studies Age- group 60 ± 70 y 70 ± 80 y 80 y One of the largest longitudinal studies (18 y) on the loss of Men fat-free mass with ageing comes from Flynn et al (1989). n 17 78 37 FFM (kg) 58.2a 55.4b 53.4c They studied loss of fat-free mass and a fall in BMR, TBF (kg) 23.4 20.0 21.8 believed to be characteristic of the ageing process. The Women putative reduction in BMR may be due to (1) loss of fat free n 50 80 51 mass; and (2) an age-related change in cellular metabolic FFM (kg) 40.0a 37.7b 37.3c activity. Cross sectional studies have repeatedly shown a TBF (kg) 26.5a 24.5b 21.4c fall in fat free mass of 3 ± 4 kg=decade in middle age FFM, fat-free mass; TBF, total body fat. (Forbes, 1987; Tzankoff & Norris, 1977). However, long- a,b,cMean values in rows with unlike superscript letters were signi®cantly itudinal studies of changes in BMR and body composition different (P < 0.05).

European Journal of Clinical Nutrition Mechanisms of changes in BMR CJK Henry S84

Figure 8 Changes in BMR with age of EF DuBois over a 22 y period.

attributable to age was no more than 1 ± 2% per decade and steady decline in BMR. Most of the dif®culties related over the age range from 20 to 75 y. to describing the effects of ageing on BMR are due to: Whilst these studies have shed valuable light on the physiological change during ageing, many of them were 1. Problems related to selection bias and cohort effect. conducted in elderly people chosen for convenience, and 2. Dif®culties associated with the measurement of body therefore not necessarily representative of the general composition in older populations. population (Fuller et al, 1996). Most of our current under- 3. Uncertainty surrounding the accuracy and precision of standing of the changes in body composition with age were body composition assessment techniques developed in obtained from studies conducted 30 ± 40 y ago. These younger subjects to predict values in older subjects. studies, along with some anecdotal impressions, provided 4. No validation at present of the best method or `gold us with the view that a dramatic change in body composi- standard' to be used in the elderly to estimate body tion and BMR occurred consistently with ageing. Recent composition. longitudinal studies appear to present a much less dramatic 5. No agreement on the most appropriate unit to express picture. More importantly, we know little about whether BMR during ageing, ie BMR=kg body weight, or 2=h. these changes, reported in predominantly Caucasian sub- BMR=kg fat-free mass, or BMR=m 6. Small sample size. jects, are applicable to other ethnic groups, notably the Afro-Caribbean and Asian communities. Longitudinal studies are expensive, but provide a truer picture of the age changes in metabolism and avoid the Longitudinal changes in BMR in humans problems inherent in cross sectional studies. It has been conventionally assumed that the major components of Keys et al (1973) were the ®rst to express concern at the oxygen consumption at rest are the skeletal muscle mass use of cross-sectional data to infer a dramatic decline in and visceral organs (brain, heart, liver and kidney) (Milber BMR with ageing. With this in mind, these authors pre- & Blyth, 1953). Whilst muscle mass has been reported to sented the results of two studies. First, a study (1947) on a decline, visceral organs are believed not to alter much with group of young students aged (18 ± 26 y) from Minnesota age (Tzankoff & Norris, 1977), a view not entirely sup- who were re-measured after 19 y (Table 6). A second study ported by recent observations. The Baltimore longitudinal was conducted on professionals (aged 44 ± 56 y) in the study provided valuable information on the physiology of metropolitan area of St Paul who had their BMR measure- ageing (Tzankoff & Norris, 1977). Although a decline in fat ments repeated in years 2, 4, 7, 13 and 19 of the study free mass has been documented from cross-sectional and (Table 7). In 1936, DuBois published the ®rst longitudinal longitudinal studies, it is now evident that the magnitude of study of his own BMR spanning 22 y (Figure 8). change in fat free mass is much smaller than previously Finally the authors also presented data from a study perceived. Whilst creatinine excretion measures predomi- conducted 22 y apart in the same group of 87 men, as nantly skeletal muscle, it does not measure visceral organs, shown in Table 8. which are metabolically extremely active. The likely Collectively, all these studies show a very modest fall in impact of changes in organ size during ageing has hitherto BMR with ageing. Indeed, mean values reported in these not been reported. Fukagawa et al (1990) ended their paper longitudinal studies may hide the fact that whilst some by commenting `differences in fat free mass cannot fully subjects may have shown a decrease in BMR, others may account for the lower BMR in the old, suggesting that show an increase in BMR with ageing. Evidence in support ageing per se is associated with an alteration in tissue of this notion comes from the elegant Baltimore long- energy metabolism.' itudinal study of 959 subjects reported by Tzankoff and These new ideas about the process of ageing and its Norris (1978). The increased prevalence of cardiovascular effects on BMR are in marked contrast to the conventional diseases and cancer associated with middle age is known to and simplistic view that the ageing process shows a gradual increase energy expenditure (Obisegan et al, 1997). In the

European Journal of Clinical Nutrition Mechanisms of changes in BMR CJK Henry S85 Table 6 Comparison of BMR in 63 subjects measured 19 y apart based samples and therefore may suffer from sample Mean Mean Mean BMR selection bias and cohort effect. Period age (y) weight (kg) height (cm) (kcal=day)a Several problems are posed in the study of metabolic changes in the elderly. Firstly the methodology chosen to Initial 21.9 73.84 177.6 1707 19 y later 41.3 84.47 177.4 1603 assess body composition may in¯uence the apparent rate of loss of fat-free mass. Secondly, the assumptions made to a Recalculated from O2 consumption. indirectly estimate body compartments may alter with ageing. For example, some of the most commonly used methods to estimate fat-free mass, eg creatinine excretion Table 7 BMR measurements made on 115 subjects ®ve times over a and total body potassium (TBK), are the least precise 17 y period methods for estimating body composition (Jebb et al, Mean Mean Mean BMR 1993). Secondly, the proportion of fat-free mass as water Period age (y) weight (kg) height (cm) (kcal=day)a is assumed to be 0.732, and this constant may vary with age (Lukaski, 1987). For example the estimation of FFM from Year 2 49.8 76.29 176.1 1576 Year 4 51.8 76.67 175.7 1513 total body water assumes a hydration coef®cient of 0.73. Year 7 54.8 77.47 175.7 1499 Baumgartener et al (1991) have shown in the elderly this Year 13 60.8 78.17 175.6 1520 may vary from 0.69 to 0.80. Whilst BMR correlates highly Year 19 66.8 78.43 175.6 1506 with FFM it should be recognized that FFM is composed aRecalculated from O consumption. of metabolically `fast lean tissue' and `slow lean tissue' 2 (Payne & Dugdale, 1977; Garby et al, 1988). The metabolically fast lean tissue includes the brain and visceral organs case of cancer this increase in energy expenditure occurred (liver, heart, kidneys, gut) and the slow lean tissue is made even before any tumour was palpable (Pratt, 1958). The 48 up of the muscle mass. The resting oxygen consumption of subjects who died were classi®ed into the following cate- these organs differs greatly. Some of the energy require- gories: cancer alone, CVD alone, and CVD cancer. Some ments of the liver are used to synthesise glucose and of the patients who died from cancer showed an increase in ketones for the central nervous system. Hence a proportion BMR years before their diagnosis. of the BMR can be directly or indirectly ascribed to the Previous `longitudinal' studies (Keys et al, 1973; Shock demands for speci®c fuels by the brain. Any change in the & Yiengst, 1955) where two point measurements were size of the brain during ageing could have a signi®cant made (without the bene®t of time series measurements) effect on BMR. suggested a gradual decline or stabilisation of BMR. The Disproportionate rates of energy expenditure per unit study reported by Tzankoff and Norris (1978) suggests a weight of the various organs may, in part, explain the marked distinction between two types of metabolic evolu- general observation that body weight correlates well with tion in the elderly. An increase in BMR with ageing is not BMR (Brozek & Grande, 1955). only a real observation, but an ominous one. This observa- Payne and Dugdale (1977) achieved remarkable con- tion is reminiscent of Kleiber's (1975) (1932) pre-mortal cordance with observed BMR when a four compartmental rise in RQ seen in fasting animals. The conclusion that may model was developed to simulate BMR. They achieved this be reasonably reached is that whilst BMR may decline by `grouping' the various components of the body into slowly on a population basis there may be some popula- different tissue metabolic types (Figure 9). The `fast lean tions and individuals that show negligible changes in BMR tissue' consisted of the metabolically active organs such as whilst others show an increase. liver, heart, brain, and gut, and the `slow lean tissue' was the sum of all the rest of the fat-free mass, with muscle being the largest component. BMR and body composition The single largest component (in adults) is the muscle mass, which may account for approximately 40% of body In recent years, methods to measure various components of weight, but only 20% of BMR. The observation that weight the body have been developed (Baumgartner et al, 1991). Body size and composition are in a state of dynamic equilibrium throughout the stages of life. In many ways the changes observed in senescence are analogous to those that occur during early infant development, except that they are in the opposite direction, namely, a decrease in height, fat-free mass and weight of visceral organs. Although these changes may be universal, their magnitude and timing show wide variation. Thus, it is necessary to distinguish between the normal process of ageing, from the process of ageing accompanied by disease. The concept of lean body mass (LBM) was ®rst introduced by Behnke (1953) and subsequently rede®ned as fat- free mass (FFM) by Grande and Keys (1980). The advan- tages of using FFM over body weight to express BMR is out-weighed by methodological errors and approximations. Elia (1992) who reviewed this topic, emphasized that none of the methods are free of assumptions and estimates. In addition, most studies on the elderly are not population- Figure 9 Payne and Dugdale model for energy regulation.

European Journal of Clinical Nutrition Mechanisms of changes in BMR CJK Henry S86 Table 8 Comparison of BMR in men 22 y apart BMR and organ weights Mean Mean Mean BMR a Several attempts have been made to partition the compo- Period age (y) weight (kg) height (cm) (kcal=day) nent organs and tissues contributing to BMR. The partition- Initial 49.4 74.8 176. 0 1540 ing of various organs according to their contribution to total 22 y later 71.4 76.0 175.5 1519 oxygen consumption was ®rst attempted by Grande (1950) aRecalculated from O consumption. and Holliday (Holiday et al, 1967; Holliday, 1971), and 2 recently reviewed by Elia (1992). The physiological partitioning of basal oxygen consumption of the human body points to some major limita- tions in our conventional view of the relationship between Table 9 Percentage contribution of various organs to total BMR (adult body composition and BMR. The physiologist Richet values) erroneously suggested that 90% of the total oxygen consumption was due to muscle control. The misconception Grande Brozek Wade Holliday Organ (1950) (1955) (1962) (1967) seems to still persist in the minds of some investigators who associate musculature with BMR. Brozek and Grande Liver 27 26.4 25 26.1 (1955) commented `The active tissue mass cannot be Brain 19 18.3 20 23.3 identi®ed with the muscle mass'. Naturally the `active Heart 11 9.2 11 10.2 Kidney 4 7.2 7 7.1 metabolic muscle tissue' cannot be associated with skeletal Subtotal 1 61 61.1 63 66.7 muscle mass alone. Muscle 20 25.6 30 28.1 The association between BMR and body composition is Subtotal 2 81 86.7 93 94.8 largely made by measuring oxygen consumption at rest and relating it to active muscle mass measured using creatinine excretion, TBW or TBK. It is important to point out that the measurement of `active muscle mass' does not include the major organs, ie brain, liver, heart, kidney, and gut in its computation, despite the fact that these organs represent alone is a reasonable predictor of BMR is surprising when approximately 80% of BMR (Table 9). Table 10 shows the one recognises that the various organs of the body have brain weight of males at selected ages. Such changes in vastly different rates of energy expenditure per unit weight. brain weight that accompany ageing may have an impact For example, the human brain and liver, which together on BMR. make up about 4% of body weight, contribute 40 ± 45% of BMR. Interestingly, many of the techniques and methods used to estimate body composition focus largely on the assessment of muscle mass (creatinine excretion, total body water, total body potassium) Ð an organ that makes a Contribution of changes in brain size to BMR in ageing modest contribution to overall BMR. This is due to the Post-mortem and in vivo imaging studies have consistently relative simplicity of measuring muscle mass rather than shown a decrease in brain tissue size with ageing (Coffey visceral organ weights. et al, 1998). Moreover, there appears to be a gender

Figure 10 Comparison of BMR calculated using organ weight with observed values.

European Journal of Clinical Nutrition Mechanisms of changes in BMR CJK Henry S87 Table 10 Data on brain weights and age in subjects from birth to age 86 Table 11 Basal oxygen consumption in the normal adult man (body weight 70 kg) (Source: Grande, 1989) Age (y) Brain weight (g) Weight Oxygen 0 380 (percentage consumption 1 970 Organ Weight (g) total body) (cm3=100 g=min) 5 1300 10 1440 Liver 1500 2.1 4.4 30 1440 Brain 1400 2.0 3.3 40 1440 Heart 300 0.43 9.4 50 1430 Kidneys 300 0.43 6.1 55 1410 Skeletal muscle 27,800 39.7 0.23 60 1370 75 1350 80 1330 85 1310 86 1290 calculated BMR may be compared with the values for BMR with age using the Oxford database (Figure 10). These graphs show remarkable similarities providing further support to the hypothesis that organ sizes play a signi®cant role in changes in BMR with age. Strehler difference in the effects of age on brain structure with (1976) noted that, with regard to ageing of the brain, males showing greater age changes than females. `deterioration in function will occur in any semiclosed Organ sizes (except for brain) at different ages were system ... [and is] simply one effect of the general law computed from the values reported by Holliday et al of entropy increase'. (1967), Coppoletta & Wolbach (1933), and Elia (l992). For brain size, values were obtained by Dekaban and Sadowsky (1978) (Table 10) from over 2773 males and 1963 female autopsies. The authors concluded that there Use of BMR to predict energy requirements was a progressive fall in brain weight after 45 ± 50 y. The While BMR measurements were used in clinical diagnosis mean loss in brain weight after 86 y may be up to 150 g during the ®rst part of this century, the ®rst comprehensive (from the value seen at 40 y). Using the organ metabolic study to use BMR as the basis to estimate human energy rates presented by Grande (1989) as shown in Table 11 and requirements, and hence food requirements, was described organ sizes at various ages, the predicted BMR is shown in by Bedale (1923). She studied a group of 45 boys and 55 Figure 10. Notice that the gradual decline in BMR has been girls aged 7 ± 18 y. This was a signi®cant departure from generated by introducing the changes in brain size (Figure previous work, as the early work on BMR was primarily 11) and other organ sizes with age. With the recognition intended to serve as metabolic reference values in clinical that visceral organs contribute a major part of total BMR it nutrition, notably in the diagnosis of the hypo- and is time we also recognised the signi®cant role the changes hyperthyroidism. In 1957, the FAO nutritional studies in brain size plays with the age-related changes in BMR. It publication no. 15 `Calorie Requirements' proposed the is important to emphasize that the changes in muscle mass use of resting energy expenditure to estimate total energy alone cannot account for the changes in BMR with age. The requirements. They presented two formulae:

Figure 11 Comparison of body and brain weight with age.

European Journal of Clinical Nutrition Mechanisms of changes in BMR CJK Henry S88 Table 14 Source of elderly (60 y) from the Oxford database used for Table 12 Number of subjects in each age group used in Schol®eld's BMR analysis analysis Ethnicity Males (n) Females (n) Total (n) Age group (y) Males (n) Females (n)

0 ± 3 162 137 North American and European 492 309 801 Chinese Japanese and Malaysian 81 80 161 3 ± 10 338 413 Aboriginal 5 0 5 10 ± 18 734 575 18 ± 30 2879 829 Indian 1 1 2 30 ± 60 646 372 African American 19 27 46 60 50 38 Total 598 417 1015 Total 4809 2364

the published literature, personal communication, and Formula 1 values obtained from our own laboratories. The number of subjects in each age group in this analysis is listed in E 92:0W 0:73 0:1E 10:88W 236 Table 13. for a 25 y old man at standard level For our present discussion the values of interest are in the 60 age group. of activity A conspicuous feature of Table 14 is the very small number of elderly subjects studied from developing coun- E 82:5W 0:73 0:1E 7:25W 133 tries. Table 15 shows the equations for estimating BMR from weight in males and females using the Oxford for a 25 y old woman at standard level database and compares them with Scho®eld's equations. of activity

Formula II How useful are BMR measurements to estimate total E 152W 0:73 for men E 123:4W 0:73 for women energy expenditure in the elderly? One practical use of BMR is in the estimation of energy where W body weight in kg; E energy in kcal=day. requirements for population groups and subsequently their It is therefore of interest to note that the FAO= food needs. The FAO=WHO=UNU report Energy and WHO=UNU (1985) approach to estimating energy require- protein requirements (1985) makes clear for the ®rst time ments is a re®nement of the method described earlier by the two main purposes of determining energy requirements. Bedale and the FAO publication of 1957. The ®rst is for prescriptive purposes, ie for making recom- The Scho®eld et al (1985) analysis and the FAO= mendations about the level of consumption that ought to be WHO=UNU report (1985) have transformed BMR mea- maintained in a population; the second, for diagnostic surements from being a clinical curiosity into becoming the purposes, ie the assessment of the adequacy or inadequacy basis for estimating energy requirements in humans. BMR of the food situation of a population. In the factorial now plays a central and critical role in estimating human estimation of total energy expenditure (FAO=WHO=UNU, energy metabolism. The Scho®eld database (which formed 1985), a major component was the estimation of BMR. For the basis of the FAO=WHO=UNU (1985) report on energy example James et al (1989) described the total energy and protein requirements) comprised of 114 published expenditure in the elderly as 1.51 times BMR. studies of BMR totalling 7173 data points. Scho®eld's In the factorial estimation of total energy expenditure it analysis has served a signi®cant role in re-establishing the is assumed that the intra-individual variation in BMR is importance of using BMR to predict human energy require- small. Indeed several reports in the literature (see Table 16) ments. con®rm the very small coef®cient of variation in males over It is more than a decade since Scho®eld collated data for a period of weeks, months, and years. Remarkably, we have the analysis. Since then, several laboratories have produced no information on the intra-individual variation in BMR in BMR data for different ages and ethnic groups. These the `old elderly' ( > 75 y). Given the increasing prevalence clearly need to be incorporated into any further analysis. of disease and ill-health in this age group, it is likely that Table 12 shows the breakdown of subjects under each age the intra-individual variation in BMR will be much greater group in Scho®eld's analysis. The number of subjects making BMR measurements less useful in the computation surveyed over 60 y and under 10 y was relatively small. of total energy expenditure. Goran & Poehlman (1992) With a view to expanding and updating the Scho®eld reported a wide inter-individual variation in total energy database, the Oxford BMR database was compiled from expenditure (1856 ± 3200 kcal=day) in the elderly. This may be in part due to the wide inter-individual variation in BMR in addition to the variations in the levels of Table 13 Number of subjects in each age group in the Oxford database physical activity. Age group (y) Males (n) Females (n)

0 ± 3 277 215 3 ± 10 441 515 Conclusion 10 ± 18 1340 1211 Biological ageing is characterized by a reduction in phy- 18 ± 30 4571 1806 30 ± 60 1419 1138 siological activities and a reduced capacity to respond to 60 597 417 environmental challenges. Whilst there is a general decline Total 8645 5302 in physiological function with ageing, the rate of decline differs between subjects. These individual differences

European Journal of Clinical Nutrition Mechanisms of changes in BMR CJK Henry S89 Table 15 BMR equations generated from the Oxford database compared to Scho®eld

Age range (y) BMR (kcal=24 h) BMR (MJ=24 h) n r s.e.

Males Oxford database 30 ± 60 12.7W 734 0.053W 3.07 1419 0.64 187.819 60 12.2W 597 0.051W 2.49 597 0.70 185.150 Scho®eld 30 ± 60 11.472W 873.1 0.048W 3.653 646 0.60 167.2 60 11.711W 587.7 0.049W 2.459 50 0.71 164.1 Females Oxford database 30 ± 60 9.2W 734 0.039 W 3.07 1138 0.64 149.717 60 10.1W 593 0.042 W 2.48 417 0.70 148.060 Scho®eld 30 ± 60 8.126W 845.6 0.034W 3.538 372 0.68 111.2 60 9.082W 658.5 0.038W 2.755 381 0.68 107.8

W: weight (kg)

Table 16 Intra-individual variations in BMR with time

Coef®cient of variation (%)

Authors Sex Age n Days Weeks Months Years

Gessler (1925) M 34 1 3.3 Grif®th (1929) M 34 1 3.67 Benedict (1935) M 62 1 3.2 Garby & Lammert (1981) M 19 ± 50 22 2.4 M 19 ± 50 23 2.2 Lammert et al (1987) M 21 ± 28 7 3.5 4.3 M 21 ± 28 7 4.8 Soares & Shetty (1986) M 18 ± 21 5 3.1 Soares & Shetty (1987) M 19 ± 23 5 2.9 M 19 ± 23 10 2.5 Henry et al (1989) M 19 ± 31 9 4.0

Adapted from Shetty et al (1966); Henry et al (1989).

appear to be much more pronounced in the case of BMR for surement of BMR in the elderly may become a useful the elderly. The elderly are a more heterogeneous group diagnostic procedure to predict morbidity and mortality. than the young in many ways. The search for a universal This would be a ®tting tribute to the pioneers of nutrition, reference unit to compare the changes in BMR with ageing Benedict, DuBois and Lusk. has proved both dif®cult and elusive. The previously reported dramatic declines in BMR with ageing were in References part due to the method of expression (eg Cal=m2=day) and the use of cross-sectional data. The only de®nitive way to Aub JC & DuBois EF (1917): The basal metabolism of old men. Arch. Int. Med. 19, 823 ± 831. determine unequivocally the subtle changes occurring Baumgarter RN, Heyms®eld SB, Lichtman S, Wang J & Pierson RN during ageing is by undertaking longitudinal studies. The (1991): Body composition in elderly people: effect of criterion esti- gradual decline in BMR with advancing age may be mates on predictive equations. Am. J. Clin. Nutr. 53, 1345 ± 1353. attributed to alteration in muscle mass and organ weights. Bedale M (1923): Energy expenditure and food requirements of children at school. Proc. Soc. B 94, 368 ± 402. Notable amongst the organ weights, is the reduction of Behnke AR (1953): The relation of lean body weight to metabolism and brain size with ageing. This may in¯uence the observed some consequent systemisations. Ann. NY Acad. Sci. 1097 ± 1142. decline in BMR. This observation is in keeping with the Benedict FG (1915): Factors affecting basal metabolism. J. Biol. Chem. report by Fukagawa et al (1990) that the loss of fat-free 20, 263 ± 299. Benedict FG (1938): Vital Energetics. A Study on Comparative Basal mass cannot fully explain the lower BMR in older subjects. Metabolism publication no 593. Washington, DC: Carnegie Institute. Whilst the measurement of BMR to predict total energy Benedict, FG & Smith HM (1915): The metabolism of male Chinese in expenditure in `younger' age groups may be appropriate, Manchuria. Chin. J. Physiol. 10, 141 ± 148. the use of BMR in the `old elderly' ( > 75 y) to estimate Boothby WM & Sandiford I (1922): Summary of the basal metabolism total energy expenditure may be less appropriate. This is data on 8614 subjects with special reference to the normal standards for the estimation of the basal metabolic rate. J Biol. Chem. 54, 783 ± 303. largely due to the increased incidence of disease and ill- Boothby WM & Sandiford I (1929): Normal values of basal or standard health in this age group that can elevate BMR. The metabolism. A modi®cation of the DuBois standards. Am. J. Physiol. proposition that the measurement of BMR may have out- 90, 290 ± 291. lived its clinical usefulness can now be challenged. The Boothby WM, Berkson J & Dunn HL (1938): Studies on the energy metabolism of normal individuals: a standard for basal metabolism. Am. estimation of BMR in the elderly appears to be one of the J. Physiol. 116, 468 ± 484. Brody S (1945): Bioenergetics of Growth. most important clinical measurements that can predict New York: Reinhold. impending disease and mortality. In the future, the mea- Brody S (1945): Bioenergetics of Growth. New York: Reinhold.

European Journal of Clinical Nutrition Mechanisms of changes in BMR CJK Henry S90 Brody S & Procter RC (1932): Relation between basal metabolism and Grande F & Keys A (1980): Body weight, body composition, and calorie mature body weight in different species of mammals and birds. Univ. status. In Modem Nutrition in Health and Disease, ed. RS Goodhart & Missouri Agric. Exp. Sta. Res. Bull. no. 166. ME Shils, pp 3 ± 34. Philadelphia, PA: Lea and Febiger. Brozek J (1963): Body composition Ð Part II. Ann. NY Acad. Sci. 110, Harris JA & Benedict FG (1919): A Biometric Study of Basal Metabolism 425 ± 1018. in Man, publication no. 279. Washington, DC: Carnegie Institute. Brozek J & Grande F (1955): Body composition and basal metabolism Henry CJK, Hayter J & Rees DG (1989): The consistency of basal in man: correlation and analysis versus approach. Hum. Biol 27, metabolic rate in free-living subjects. Eur. J. Clin. Nutr. 43, 727 ± 731. 22 ± 31. Henry CJK, Dyer S & Ghusain-Choueiri A (1999): New equations to Calloway DH & Zanni E (1980): Energy requirements and energy estimate basal metabolic rate in children aged 10 ± 15 years. Eur. J. expenditure of elderly men. Am. J. Clin. Nutr. 33, 2088 ± 2092. Clin. Nutr. 53, 134 ± 142. Carpenter TM (1933): Problems in the determination of the Holliday MA (1971) Metabolic rate and organ size during growth from basal metabolism of man and factors affecting it. Ohio J. Sci. 33, birth to maturity and during late gestation and early infancy. Paediatrics 315 ± 434. 47, 169 ± 179 Chumlea WC & Baumgartner RN (1989): Status of anthropometry Holliday MA, Potter D, Jarrah A & Bearg S (1967): The relation of and body composition data in elderly subjects. Am. J. Clin. Nutr. 50, metabolic rate to body weight and organ size. Pediatr. Res. 1, 185 ± 195. 1158 ± 1166. Horvath SM, Radcliffe CE, Huff BK Spurr GB (1955): Metabolic Chumlea WC, Vellas B & Guo SS (1998): Malnutrition or healthy responses of old people to a cold environment. J. Appl. Physiol. 8, senescence. Proc. Nutr. Soc. 57, 593 ± 598. 145 ± 148. Coffey CE, Lucke JF, Saxton JA, Ratcliff G, Unitas LJ, Beilig B & Bryan James WPT, Ferroluzzi A, Waterlow JC (1989): De®nition of chronic N (1998): Sex differences in brain aging. Arch. Neurol. 55, 169 ± l79. energy de®ciency in adults ± report of a working party of the Coppoletta JM & Wolbach SB (1933): Body length and organ weight of International Dieters Energy Consultant Group. Eur. J. Clin. Nutr. 42, infants and children. Am. J. Pathol. 9, 55 ± 69. 969 ± 981. Daly J, Heyms®eld SB & Head CA (1985): Human energy requirements: Jebb SA, Murgatroyd PA & Coward A (1993): In vivo measurement of overestimation by widely used prediction equations. Am. J. Chin. Nutr. changes in body composition. Am. J. Clin. Nutr. 58, 455 ± 462. 42, 1170 ± 1174. Keys A, Taylor HL & Grande F (1973): Basal metabolism and age of adult Dekaban AS & Sadowsky D (1978): Changes in brain weights during the man. Metabolism 22, 579 ± 587. span of human life: Relation of brain weights to body heights and body Kise Y & Ochi T (1934): Basal metabolism of old people. J. Lab. Clin. weights. Ann. Neurol 4, 345 ± 356. Med. 19, 1073 ± 1079. DuBois (1927): Basal Metabolism in Health and Disease, 2nd edn. Kleiber M (1932): Body size and metabolism. Hilgardia 6, 315 ± 353. London: Bailliere, Tindall & Cox. Kleiber M (1975): The Fire of Life. Huntington, NY: Krieger. DuBois D & DuBois EF (1916): Clinical calorimetry. Tenth paper. A Korenchevsky V (1961): Physiological and Pathological Ageing,pp formula to estimate the approximate surface area if height and weight 38 ± 47. New York: Karger. be known. Arch. Intern. Med. 17, 863 ± 672. Krogh A (1916): The Respiratory Exchange of Animals and Man. New DuBois EF (1919): The basal metabolism as a guide in the diagnosis and York: Karger. treatment of thyroid diseases. Med. Clin. North Am p 201. Lewis WH (1934): Changes with age in the basal metabolic rate in adult DuBois EF (1930): Recent advances in the study of basal metabolism Ð men. J. Physiol. 12, 502 ± 517. Part 1. J. Nutr. 3, 217 ± 228. Lukaski HC (1987): Methods for the assessment of human body composi- DuBois EF (1936): Basal Metabolism in Health and Disease, 3rd edn. tion: traditional and new. Am. J. Clin. Nutr. 46, 537 ± 556. London: Bailliere, Tindall & Cox. Lusk G (1928): The Elements of the Science of Nutrition, 4th edn. Elia M (1992a): Energy expenditure in the whole body. In: Energy Philadelphia, Pa: Saunders. Metabolism: Tissue Determinants and Cellular Corollaries, ed. JM Mif¯in MD, St Jeor ST, Hill LA, Scott BJ, Daugherty SA, Koh YO (1990): Kinney & HN Tucker, pp 19 ± 60. Raven Press, New York. A new predictive equation for resting energy expenditure in healthy Elia M (1992b): Organ and tissue contribution to metabolic rate. In: individuals. Am. J. Clin. Nutr. 51, 241 ± 247. Energy Metabolism: Tissue Determinants and Cellular Corollaries, Miller AT & Blyth CB (1953): Lean body mass as a metabolic reference ed. JM Kinney & HN Tucker, pp 61 ± 80. Raven Press, New York. standard. Am. J. Physiol. 5, 311 ± 316. FAO (1957): Calorie Requirements. FAO Nutritional Studies no. 15. Mitchell D, Strydom NB, van Graan CH & van der Walt WH (1971): Rome: FAO. Human surface area: comparison of the Dubois formula with direct FAO=WHO=UNU report (1985): Energy and Protein Requirements. photometric measurement. P¯uÈgers Arch. 325, 188 ± 190. Mitchell HH Report of a joint FAO=WHO=UNL expert consultation. Technical (1964): Comparative Nutrition of Man and Domestic Animals, Vol I. Report Series no. 724. World Health Organization, Geneva. New York: Academic Press. Fleisch A (1951): Le meÂtabolisme basal standard et sa deÂtermination au Morgan JB & York DA (1983): Thermic effect of feeding in relation to moyen du `MeÂtabo calculator'. Helv. Med. Acta 18, 23 ± 44. energy balance in elderly men. Ann. Nutr. Metab. 27, 71 ± 77. Flynn MA, Nolph GB, Sherwood-Baker A, Martin WM & Krause G Murray LA, Reilly JJ, Choudhry M & Durnin JVGA (1996): A long- (1989): Total body potassium in longitudinal study. Am. J. Clin. Nutr. itudinal study of changes in body composition and basal metabolism in 50, 713 ± 770. physically active elderly men. Eur. J. Appl. Physiol. 72, 215 ± 218. Forbes GB (1987): Human Body Composition. New York: Springer. Obisesan TO, Toth MJ & Poehlman ET (1997): Prediction of resting energy Fukagawa NK, Bandini LG & Young JB (1990): Effect of age on needs in old men with heart failure. Eur. J. Clin. Nutr. 51, 678 ± 681. body composition and resting metabolic rate. Am. J. Physiol. 259, Of®ce of Population Censuses and Surveys (1994): Morbidity Statistics E233 ± E238. from General Practice 1991 ± 1992. London: HMSO (OPCS Monitor Fuller NJ, Jebb SA, Laskey MA, Coward WA & Elia M (1992): Four- MB5). component model for the assessment of body composition in humans: Payne PR & Dugdale AE (1977): A model for the prediction of energy comparison with alternative methods, and evaluation of the density and balance and body weight. Ann. Hum. Biol. 4, 525 ± 535. hydration of fat-free mass. Clin. Sci. 82, 687 ± 693. Poehlman ET (1996): Energy intake and energy expenditure in the elderly. Fuller NJ, Sawyer MB, Laskey MA, Paxton P & Elia M (1996): Prediction Am. J. Hum. Biol. 8, 199 ± 206. of body composition in elderly men over 75 years of age. Ann. Hum. Poehlman ET (1998): Effect of exercise on daily energy needs in older Biol. 23, 127 ± 147. individuals. Am. J. Clin. Nutr. 68, 997 ± 998. Garby L, Garrow JS, Jorgensen B, Lammert O, Madsen K, Sorensen P Poehlman ET & Danforth E (1991): Endurance training increases meta- & Webster J (1988): Relation between energy expenditure and body bolic rate and norepinephrine appearance rate in older individuals. Am. composition in man: Speci®c energy expenditure in vivo of fat and fat- J. Physiol. 261, E233 ± E239. free tissue. Eur. J. Clin. Nutr. 42, 301 ± 305. Ravussin E & Bogardus C (1989): Relationship of genetics, age and Goran MI & Poehlman ET (1992): Total energy expenditure and physical ®tness to daily energy expenditure and fuel utilisation. Am. J. energy requirements in healthy elderly persons. Metabolism 41, 744 ± Clin. Nutr. 49, 468 ± 475. 753. Rinder L, Roupe S, Steen B & Svanborg A (1975): 70 year old people in Grande F (1950): El consumo de oxigeno del cerebro. Rev. Clin. Esp. 31, Gothenburg. A population study in an industrialised Swedish city. 1. 1 ± 12. General design of the study. Acta Med. Scand. 37, 198. Grande F (1989): Energy expenditure of organs and tissues. In: Assessment Roza AM & Shizgal HM (1984): The Harris ± Benedict equation re- of Energy Metabolism in Health and Disease, ed. JM Kinney, pp 88 ± evaluate Ð resting energy requirements and the body cell mass. Am. 92. Columbus: Ross Laboratories. J. Clin. Nutr. 40, 168 ± 182.

European Journal of Clinical Nutrition Mechanisms of changes in BMR CJK Henry S91 Scho®eld WN (1985): Predicting basal metabolic rate, new standards Svanborg A (1988): The health of the elderly population: results from and review of previous work. Hum. Nutr. Clin. Nutr. 39C(Suppl 1), 5 ± 41. ongitudinal studies with age-cohort comparisons. In: Research and the Scho®eld WN, Scho®eld C & James WPT (1985): Basal metabolic rate Ð Ageing Population, ed. D Evered & J Whelan. Chichester: Ciba ± John review and prediction, together with an annotated bibliography of Wiley. source maturial. Hum. Nutr. Clin. Nutr. 39C(Suppl 1). Tata JR (1964) Basal metabolic rate and thyroid hormone. ln: Advances in Shetty PS, Henry CJK, Black AE & Prentice AM (1996): Energy Metabolic Disorders, ed. R Levine & R Luft, pp 153 ± 189. New York: requirements of adults: an update of basal metabolic rates (BMRs) Academic Press. and physical activity levels (PALs). Eur. J. Clin. Nutr. 50, S11 ± S23. Tzankoff SP & Norris AK (1977): Effect of muscle mass decrease on age- Shock NW & Yiengst MJ (1955): Age changes in basal respiratory related BMR changes. J. Appl. Physiol. 43, 1001 ± 1006. measurements and metabolism in males. J. Gerontol. 10, 31-40. Tzankoff SP & Norris AH (1978): Longitudinal changes in basal meta- Strehler B (1976): Introduction: aging and the human brain. In: Neuro- bolism in man. J. Appl Physiol. 45, 536 ± 539. biology of Aging, ed R Terry & S Gershon pp 1 ± 22. New York: Raven Wade OL & Bishop JM (1962): Cardiac Output and Regional Blood Flow, Press. p 93. Oxford: Blackwell.

European Journal of Clinical Nutrition