Previous - Next
Initially, the discussion dealt with the selection of studies and data that form the basis of the paper and recommendations of Torun et al. For methodological reasons, Ferro-Luzzi questioned the validity of older energy expenditure data based on heart rate measurements. She also argued that the life-styles of children are changing so rapidly, also in developing countries, that time allocation studies made more than 10 years ago may no longer reflect today's situation. Torun explained and justified the selection criteria that were used in greater detail and argued that, in general, there was a reasonably good fit between data obtained by different methods at different points in time. Shetty supported this view and emphasized the relatively good fit even with data that were collected in the 1970s and formed the basis of the 1985 report.
There was general agreement that energy requirements should be based on data on energy expenditures of normal children. Some of the data in the paper of Torun et al were from children from developing countries who were of low height-for-age (stunted), but otherwise 'normal', healthy, and adequately nourished (in terms of BMI and weight-for-height) at the time the measurements were taken. There was some debate as to whether data of such children should be included in the data base or not, particularly since energy expenditures of stunted children, expressed per kg body weight, are higher than those of non-stunted children of the same weight. In the end, the prevailing view was that in making recommendations one could not ignore that large proportions of children in developing countries are stunted, but otherwise healthy and normal, and that it seems therefore justified to include such data (Town, Scrimshaw). Another argument was that recommending the feeding of stunted but otherwise healthy children additional energy to catch up in height (i.e. by determining their dietary energy requirements on the basis of ideal rather than actual weight, as one does with protein) will only tend to make them obese (Scrimshaw, Torun), and as long as dietary recommendations to satisfy energy requirements are made by age groups in absolute terms (i.e. energy units per day) and not by kg body weight or lean body mass, there will be no substantial difference between stunted and non-stunted children.
Several discussants speculated on the reasons why energy expenditures per kg body mass or LBM tend to be higher in stunted children than in children of normal height. Behavioral and life-style differences could be responsible for some of the differences found in DLW studies. This argument is supported by heart-rate and time-motion studies tending to show that stunted children from lower socioeconomic classes are likely to have a different life-style. Using minute-by-minute records and estimating the time spent in various categories of activities. Torun found that poorer children spent less time in sedentary and more time in light activities. Spurr found no differences in activity between well nourished and marginally nourished Colombian children during the school year, but the better nourished children were more active in leisure activities during the summer holidays. Comparing heart rates of children from the UK and developing countries, Prentice found that UK children had higher heart rates.
Torun argued in his paper and in the discussion that, in general, studies show differences in life-style requiring more energy-demanding activities from children of rural populations in developing countries. To Schüch this conclusion appeared to be a reflection of the assumptions made rather than an inference from empirical data. Table 14 shows that it is assumed that children living in rural areas of developing countries spend a greater proportion of their time in domestic chores and production activities requiring greater effort. These assumption; are used to calculate mean daily EE in terms of PAL factors from which it is then concluded that energy expenditures are higher in rural than in urban and industrial environments. This argument seems to a large extent circular.
Body mass and LBM are both very heterogeneous, i.e. composed of different tissues with different energy requirements. If their proportions differ between children of low and normal height this could explain some of the differences observed in calorimetric studies (Young). The results of a few small studies that tried to test this hypothesis remain inconclusive, and more work is needed to clarify this issue. Differences in body proportions have been observed: stunted Peruvian children, for instance, tend to have reduced limb length relative to the size of their trunks (Reeds), and the secular trend in Japanese to become taller reflects primarily an increase in leg length (Butte).
Several participants (e.g. Ferro-Luzzi & Torun) emphasized how difficult it was to measure the energy cost of activities in children and argued for the development of instruments that are better adapted to children.
Much of the discussion dealt with the introduction of PAL indices and their values reflecting the different lifestyles of different groups of children. Torun integrated a section on this issue into the final version of his paper. He also defined cut-off points in terms of PAL values. Dietary intake data, for instance, lying below or above these cut-offs were considered physiologically improbable and excluded from the data base.
Some discussants expressed concern about the current trend towards increasingly sedentary life-styles not only in adults, but also among children. Should proposed PAL values reflect actual or desirable levels of activity? Butte and Durnin expressed doubts that we have enough information to be prescriptive; others feared that by recommending higher PAL values without being able to ensure that children actually do increase their level of physical activity may well lead to recommendations that are inappropriately high.
Hautvast commented on the dietary intakes presented in Figure 10, and showing that energy intakes appear to be higher than energy expenditures in infants and lower than energy expenditures in adolescents. This appears quite plausible considering that dietary intakes during the first years of life are mainly based on reports of caretakers, who are more likely to over-report intakes, whereas adolescents, trying to stay slim, may under-report intakes. Torun also finds this interpretation quite plausible, particularly because adolescent girls reported intakes that were further below predicted values than the intakes reported by adolescent boys. Torun added further that the discrepancy between intakes and expenditures tends to be greater in data from the US (Dietz) than from the UK (Livings/one). Prentice emphasized that only a few of the columns in Figure 10 exceed the range representing the estimated limits of precision, given the inaccuracies of both methods.
Waterlow raised the question of how accurate recommendations needed to be. The answer obviously depends on the use to which recommendations are to be put. To assess the adequacy of food available to countries or populations, FAO and other organizations use food balance sheets and estimates of population structure and food wastage which are notoriously inaccurate; in this context it appears unnecessary to try to formulate recommendations with an accuracy below 5 kcal/kg. Of greater concern are cumulative effects of errors. Reed pointed out that the energy equivalent of an excess of 5 kcal/kg/d for a year in a 14 year old weighing 40 kg amounts to about 8 kg. Clugston mentioned that WHO is often approached by governments and agencies who use these recommendations to calculate energy requirements of populations. Since children are often 40% of the population of poor countries, the accuracy of requirements is quite important for such calculations. Since a large proportion of children in poor countries are undernourished, Clugston encouraged IDECG to examine further the energy (and other dietary) requirements of stunted and wasted children.