Estimates of basal metabolic rate to calculate total energy expenditure

To calculate the energy equivalent of a PAL value, it is multiplied by the BMR. The 1985 FAO/WHO/UNU Expert Consultation endorsed the use of the mathematical equations derived by Schofield, which take into account sex, age and body weight, to estimate a population's mean BMR. Although Schofield revised and modified his equations (Schofield, 1985), those initially published in the FAO/WHO/UNU report on Energy and Protein Requirements are used more often. The two sets of equations give similar values (within ± 1-2%), except for girls 3-10 years old, where the FAO/WHO/ UNU equations give BMR's 6-7% higher than the revised equations. Thus, the PAL of those girls is lower when calculated with the FAO/WHO/UNU equations. In this review we have used the revised equations (Schofield, 1985).

The PAL approach was recommended by the FAO/ WHO/UNU Experts to calculate TEE of adult populations with occupations and lifestyles that involved different PALs. It was used to estimate TEE of children and adolescents 10-18 years old with a pattern of activities that reflected the lifestyle of children in developed countries who spend several hours at school every day (FAO/WHO/UNU, 1985). No calculations were made for those with more energy-demanding lifestyles. This can be corrected, but doubts still remain about the accuracy of the Schofield-FAO/WHO/UNU equations to predict BMR in all races. This has been addressed by authors such as Henry & Rees (1988) and Elia (1992). Table 8 illustrates some of their conclusions about the possibility of over- or underestimating BMR in adults with Schofield's equations.

Accuracy of mathematical estimations of BMR

We explored the accuracy of the Schofield equations to estimate BMR of children and adolescents from various published and unpublished reports. Some studies measured BMR and others measured resting metabolic rate (RMR). The conditions for the latter varied from quasi basal conditions (supine position, 10-12 h fasting, transported by vehicle to the laboratory, resting 30-60 min prior to the measurement) to measurements done in supine, sitting and standing positions, 2-4 h after a light meal and resting for 15-45 min before the test.

The results for measured BMR are shown in Tables 9 and 10. Those results, however, must be interpreted with some caution. For example, Bandini et al (1990b) applied Weir's equation (1948) to correct for the difference in the volumes of inspired and expired air, whereas some of the others apparently did not. When only expired volume is measured and Weir's correction is not applied, BMR is underestimated by about 5%. Some systems that use a ventilated hood and compare the concentration of inhaled and exhaled O₂ and CO₂, such as the diaferometer used by Torun and Viteri (1981a), and the system used by Livingstone et al (1992a), compensate for the difference between inspired and expired air. Saris et al (1989) used a whole body indirect calorimeter that could also have compensated for that difference.

Table 8 Percentage by which Schofield equations overestimate ( + ) or underestimate ( - ) basal metabolic rate in different ethnic groups (18-60 years old)^a

^a Source: Henry & Rees 1988).

With that methodological caveat in mind, Table 9 shows the BMR of boys and girls of different age groups measured in various countries, and compares them with the BMR calculated with Schofield's equations (1985). There seems to be a difference between developed and developing countries, an age-related trend in the data from the latter, and no major effects related to stunting or mild undernutrition. This can be seen more clearly in Table 10 Except for the Colombian underweight preschool aged boys, the difference or coincidence between measured and calculated BMR was similar for boys and girls of the same age groups, either with adequate weight and height, moderately stunted or mildly underweight.

In terms of age and sex, Schofield's equations overestimated the BMR of well-nourished, stunted or underweight Guatemalan, Colombian and Chinese preschoolers by about 10-12% in boys, and by 6-9% in girls. They coincided with measured BMR in boys and girls 7-16 years old in Holland, the UK and the USA, but overestimated the BMR of Colombian boys of that age by about 5%. That overestimation was not observed in their female counterparts, nor in Chinese girls 12-15 years old. By contrast W Wong (personal communication to B Torun) found that Schofield's equations overestimated by about 6% the BMR of 9-12 year-old hispanic and oriental girls living in Houston, Texas. The equations also overestimated by 9% the BMR of Chinese girls 15-18 years old in Guangzhou, China (Table 10).

In addition to those geographic and/or ethnic differences, Henry indicated that BMR in Beninese and Indonesian children is 8-10% lower than in the U.S. and Europe (personal communication).

More evidence about the tendency of current mathematical equations to overestimate BMR of many children and adolescents is derived from measurements of resting metabolic rates that should have been between about 15 and 20% higher than BMR, considering the conditions under which RMR is measured. For example, unpublished studies by Torun and coworkers in 68 Guatemalan 10-12 year-old boys of two economic income groups and repeated measurements in 24 stunted but well nourished girls of that same age, showed that in both sexes the non-fasting mean RMR measured after 15 min in supine, sitting and standing positions was only 7% greater than their BMR calculated with Schofield's equations. This was about 10% less than expected under the prevailing RMR conditions.

Firouzbakhsh et al (1993) reported similar results in 92 boys and 107 girls, 5-16 years old, living in or near Los Angeles, California. RMR measured 2-3 h post-prandial and after resting for 15-30 min. coincided with the calculated BMR within ± 8% in all age groups and either sex.

Table 9 Comparison of measured BMR with BMR calculated from Schofield's equations (1985)

^a + indicates that Schofield's formulas give higher values, and - indicates lower values.
^b Adequate weight but previously malnourished. Height-for-age > 1.5 s.d. below the NCHS median. ^c Weight-for-age and weight-for-height < 95% of Colombian standards (Rueda-Williamson et al, 1969)

Conclusions

Even though there may be some methodological doubts about their interpretation, the preceding observations and the data shown in Tables 9 and 10 indicate that the mathematical equations endorsed in 1985 by FAO/ WHO/UNU to calculate BMR, tend to overestimate the results and, consequently, the TEE of many children and adolescents calculated from estimates of the population's PAL.

It is necessary to decide whether a single set of predictive equations for BMR should be used universally for all children and adolescents, acknowledging an error of certain magnitude in some cases, or whether specific equations must be derived and applied to certain races or to children who live in some parts of the world.

The extensive review of BMR data presently being done by CJK Henry under the auspices of IDECG and with funding from the Nestle Foundation should help to clarify this issue.