Predicting basal metabolic rate, new standards and review of previous work

After reviewing the literature on basal metabolism, this paper discusses and reviews recent attempts to predict BMR from age, sex and anthropometric measurements. Criticism is made of the scientific and statistical integrity of a widely used table of standard metabolic rates for weight. The statistical screening of data from the literature of the past 50 years is described and equations computed from these screened data are presented. In these equations, BMR is predicted simply from weight or from weight and height with sex and age taken into account. Information is given on error, and tables estimating error for predictions on new data both for individuals and for means of groups of subjects are included. A table of BMRs for weights from 3 to 84 kg for males and females separately is also included. Cross-validation techniques are used to estimate possible threats to validity from various sources including, for example, different procedures of early workers. It was found that in the data available subjects from developing countries not only were smaller and had lower metabolic rates (as was expected) but also had lower rates per unit body weight than European or North American subjects. It is argued that at an individual level the error of prediction must be high since the global operationalisation of BMR confounds separate effects known to participate in complex relations with sex, age and anthropometric indices. The work reported is aimed at meeting a practical need for equations which are simple to apply. However, it was found that little was gained by the use of more complex equations, although they remain of scientific interest.