|Energy 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)|
|Energy requirements: general principles|
Since energy requirements are now preferentially derived from data on energy expenditure and expressed as multiples of BMR, most of the discussion dealt with issues related to BMR, the energy cost of physical activity, physical activity levels (PALs) and the factorial method.
Under the auspices of IDECG and with funding from the Nestle Foundation, all currently available BMR data meeting a set of stringent criteria are being reanalyzed under the supervision of Henry and Durnin. The conditions under which BMRs have to be measured were reemphasized. In some earlier studies these were as strictly observed as in more recent ones; publication date is therefore not necessarily a good screening criterion. One of the questions raised was how important it is that subjects sleep at the laboratory where their BMR is measured in the morning. Durnin cited several papers (Bullough & Melby, 1993; Soares et al, 1989; Turley et al, 1993) showing that this did not make any difference. In a group of elderly subjects, Berke et al (1992) even found a significantly higher metabolic rate when they slept in a metabolic ward, than when they slept at home before coming to the laboratory, probably because at the ward they had to sleep in an unfamiliar bed and in a foreign environment. When subjects are brought to the laboratory in the morning, the question arises as to how much resting time is required, before BMR is measured. According to Ferro-Luzzi, experiments have shown that 30 min are enough in persons who have not previously been engaged in heavy physical activity. The current analysis should provide answers to several issues discussed at the meeting and considered in need of further analysis and clarification. In particular it should enable to decide whether predictive equations can remain general or should be population-specific (taking into account ethnicity, geographic regions/climate or body composition), linear or non-linear (depending on which results in the smaller residual values), etc. The importance of taking height into consideration in predictive equations was briefly discussed. Apparently Schofield tested whether including height would make a difference in the prediction of group BMRs and found that it did not.
Even though large numbers of BMR data can be found in the literature, there are age and population groups for which an adequate data base does not yet seem to exist. Underrepresented groups are infants, children, the elderly and almost any age group from developing countries. For the establishment of predictive equations the overrepresentation of individuals with specific characteristics also needs to be avoided.
In his paper, Durnin cited several papers leading to the conclusion that the intra-individual variability of BMR in adults was in the order of 3%. Henry agreed with this figure for men, but argued that in women the menstrual cycle can increase this variability to 8-10%.
In his paper, Durnin expressed the view that, on a population basis and up to a moderate level of fatness (BMI < 30), there was nothing to be gained by expressing BMR per kg lean body mass rather than body weight. In the discussion, reference was made to Garby's work suggesting that most of the inter individual variability in BMR is probably attributable to differences in body composition and that the relative size of organs, which have a high metabolic rate can make a difference, notably in chronically undernourished individuals.
The question was raised, if there are ethnic or geographic differences in BMR. Butte et al studied BMRs of white and black adolescent girls in Houston. In absolute terms, black girls had lower BMRs, but the difference disappeared when the data were corrected for body composition. Sexual maturation is likely to affect body mass, body composition and BMR and may result in differences between adolescent girls going through menarche at different ages.
Shetty et al (1986) found that the Schofield equations overestimated BMRs in Indian subjects. Henry's equations made better predictions of BMRs of populations in tropical regions. In two recent papers (Soares et al, 1993; Piers & Shetty, 1993), Shetty's group documented that BMRs of well-nourished Indian subjects, normalized for body weight, do not differ from BMRs of American subjects, but they seem to differ from Schofield equations and BMRs of certain European populations. He imputed differences, reported in the literature, to differences in ambient temperature which may not have been taken sufficiently into account in earlier studies. Torun suggested that differences in body composition could also provide a partial explanation. On average, BMRs measured in tropical regions are 4-5% lower than BMRs measured in temperate zones.
Differences have also been observed in the same subjects, measured in temperate and tropical regions. Henry suggested that such differences could be due to weight loss or disease. Shetty referred to a study by Mason (1944) who found changes in BMR, even when the latter was normalized for body weight.
A research assistant, under the supervision of Shetty and Durnin, is re-analyzing the information that is currently available on the energy cost of various activities.
The issue as to whether and how much of an energy allowance ought to be made for discretionary activities was discussed at some length. To maintain physical fitness and promote cardiovascular health, the 1985 FAO/WHO/UNU report recommended 20 min of vigorous exercise per day to adults with a sedentary life style and included in its recommendations the energy required for this exercise. Most of the discussants agreed with Durnin that there was only very little scientific evidence relating different levels of physical activity to health, but, for various reasons, the majority felt that these earlier recommendations should nevertheless be maintained.
Prentice collected and analyzed nearly 1000 data on total energy expenditure obtained with the doubly labeled water (DLW) method. More than half of these studies included also BMR measurements, so that PAL factors could be calculated and compared with PAL values obtained with other methods. Only few, mostly institutionalized individuals have PALs below 1.4, but, on average most PAL values obtained with DLW appear somewhat higher than expected and than contained in previous reports. UK dietary reference values, for instance, had assumed that a PAL of 1.4 was representative of light, 1.6 of moderate and 1.7 of heavy occupational activity in an otherwise non-active, non-occupational life style. Average PALs of sedentary people obtained by the DLW method correspond to values that were considered representative of a moderately active life style. Some of the highest values (2.42.6) look suspect because, according to PALs obtained from calorimetry, they presuppose that people spend a considerable amount of time at 70-80% of VO2 max which happens only rarely in real life. It has to be acknowledged (and perhaps provides a partial explanation) that the representativeness and generalizability of the DLW data obtained till now are limited.
The group recognized that injuries and various diseases can affect energy requirements, but was not prepared to deal with such special situations in detail. It recommended a compilation and analysis of information that is currently available in this area, but expressed a preference for presenting such information separately.