See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/7490153

Basal metabolic rate studies in humans: Measurement and development of

new equations

Article  in  Public Health Nutrition · November 2005

DOI: 10.1079/PHN2005801 · Source: PubMed

CITATIONS READS

473 7,465

1 author:

Jeya Henry
Singapore Institute for Clinical Sciences
196 PUBLICATIONS   5,960 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Article Substituting brown rice for white rice on diabetes risk factors in India: A randomized controlled trial Article Effect of a Novel High Fiber Rice Diet on 24-Hour
Glycemic Responses in Asian Indians Using Continuous Glucose Monitoring: A Randomized Clinical Trial View project

Article Association of whole grains, dairy and dietary fibre with neonatal outcomes in women with gestational diabetes mellitus: The WINGS project (WINGS – 12) Article
Finger millet (Eleusine coracana L.) and white rice diets elicit similar glycaemic response in Asian Indians: Evidence from a randomised clinical trial using continuous
glucose monitoring View project

All content following this page was uploaded by Jeya Henry on 23 June 2015.

The user has requested enhancement of the downloaded file.

 Public Health Nutrition: 8(7A), 1133–1152 DOI: 10.1079/PHN2005801

Basal metabolic rate studies in humans: measurement

and development of new equations
CJK Henry*
School of Biological and Molecular Sciences, Oxford Brookes University, Gipsy Lane Campus, Oxford OX3 0BP, UK

Abstract
Objective: To facilitate the Food and Agriculture Organization/World Health
Organization/United Nations University Joint (FAO/WHO/UNU) Expert Consultation
on Energy and Protein Requirements which met in Rome in 1981, Schofield et al.
reviewed the literature and produced predictive equations for both sexes for the
following ages: 0–3, 3–10, 10–18, 18–30, 30–60 and . 60 years. These formed the
basis for the equations used in 1985 FAO/WHO/UNU document, Energy and Protein
Requirements.
While Schofield’s analysis has served a significant role in re-establishing the
importance of using basal metabolic rate (BMR) to predict human energy
requirements, recent workers have subsequently queried the universal validity and
application of these equations. A survey of the most recent studies (1980 –2000) in
BMR suggests that in most cases the current FAO/WHO/UNU predictive equations
overestimate BMR in many communities. The FAO/WHO/UNU equations to predict
BMR were developed using a database that contained a disproportionate number –
3388 out of 7173 (47%) – of Italian subjects. The Schofield database contained
relatively few subjects from the tropical region.
The objective here is to review the historical development in the measurement and
application of BMR and to critically review the Schofield et al. BMR database
presenting a series of new equations to predict BMR.
Design: This division, while arbitrary, will enable readers who wish to omit the
historical review of BMR to concentrate on the evolution of the new BMR equations.
Setting: BMR data collected from published and measured values.
Subjects: A series of new equations (Oxford equations) have been developed using a
data set of 10 552 BMR values that (1) excluded all the Italian subjects and (2) included
a much larger number (4018) of people from the tropics.
Results: In general, the Oxford equations tend to produce lower BMR values than the
current FAO/WHO/UNU equations in 18–30 and 30 –60 year old males and in all
Keywords
females over 18 years of age. Universal validity
Conclusions: This is an opportune moment to re-examine the role and place of BMR Basal metabolic rate
measurements in estimating total energy requirements today. The Oxford equations’ Energy metabolism
future use and application will surely depend on their ability to predict more Energy requirements
accurately the BMR in contemporary populations. Body mass index

Introduction This division, while arbitrary, will enable readers who

wish to omit the historical review of BMR to concentrate on
Since the last Food and Agriculture Organization/World the evolution of the new BMR equations.
Health Organization/United Nations University Work concerning energy metabolism may be traced
(FAO/WHO/UNU) Expert Committee on Energy and back to 1783 and the classical experiments of Lavoisier and
Protein Requirements met in 1981, a considerable amount Laplace. The principles of calorimetry laid down by these
of work has been reported on the use and validity of the founding fathers over 200 years ago are still valid today.
FAO/WHO/UNU1 equations to predict basal metabolic rate The development and subsequent apparatus used to
(BMR). This paper is divided into two parts – one will measure respiratory exchange were based on the
review the historical development in the measurement and principles of calorimetry. The term ‘basal’ was used to
application of BMR; and the second will critically review the distinguish between the energy expended while perform-
Schofield BMR database and then present a series of ing physical activity and being at rest. BMR represents the
new equations (Oxford equations) to predict BMR. integration of minimal activity of all the tissues in the body

*Corresponding author: Email jhenry@brookes.ac.uk q The Author 2005

1134 CJK Henry
under steady state conditions. It is usually expressed as Conditions to be met while measuring BMR
heat production or oxygen consumption per unit body
size. A more succinct definition of BMR was presented by The concept of basal metabolism arose from the need to
Mitchell2 who said, ‘Basal metabolism of an animal is the standardise measurements so that accurate comparisons
minimal rate of energy expenditure compatible with life’. could be made between individuals. This is achieved by
In order to begin our analysis, it is appropriate to briefly measuring a minimum rate of heat production free of the
review the historical developments in the study of BMR. effects of any consumption of food and ‘extreme’ physical
This approach will enable us to appreciate the primary environments6.
purpose of the early measures of BMR and how its All BMR measurements must therefore meet the
application has evolved with time. following conditions:
BMR is the daily rate of energy metabolism an individual
1. The subject should be completely rested, both before
needs to sustain in order to preserve the integrity of vital
and during the measurements. They should be lying
functions. It must be measured under conditions, which,
down and fully awake.
as far as possible, avoid the influence of the external
2. The subjects should be fasted for at least 10 –12 hours
environment, for example heat, or cold, physical move-
before the measurements are taken.
ment and the effects of food or drugs. If these conditions
3. The environment in which the measurements are taken
are observed, the result of the measurement is considered
should be thermo-neutral (22–268C) so that there is no
to represent the physiological and biochemical integrity of
thermoregulatory effect on heat production.
the individual concerned. In normal free-living persons,
4. The subject should be free from emotional stress and
the amount of energy provided by their diet must cover the
familiar with the apparatus used.
demands of basal metabolism plus additional amounts
needed for the physical activity associated with essential
bodily needs, and also those imposed by the physical
work involved in occupation, in engaging in social Ambient temperature during BMR measurements
exchanges and in leisure activities.
The ambient temperature at which energy expenditure is at a
minimum was termed the ‘critical temperature’ by Rubner at
Historical background: origins of the term ‘basal the turn of the 20th century. The more commonly used term
metabolism’ was ‘zone of thermal neutrality’. This was defined as the
ambient temperature above or below which resting
Magnus-Levy coined the term Grundumsatz or ‘basal metabolism of subjects begins to rise. The lowest ambient
metabolism’ in 1899. This term was of great value to the temperature at which an organism can maintain ‘resting’ or
early investigators, as it emphasised the need to conduct basal metabolic rate (without an increase in energy
the experiments under strictly standardised conditions. expenditure) is called the lower critical temperature. Work
These included the following: (1) absence of gross carried out on humans suggests the lower critical
muscular activity; (2) post-absorptive state; (3) minimal temperature to be between 22 and 278C7,8. Numerous
emotional disturbance; (4) wakefulness; (5) normal published works on BMR were conducted at temperatures as
nutritive condition; (6) absence of disease or infection; low as 9–158C9. Indeed many of the early studies paid little
and (7) thermo-neutral environment. In practice, however, attention to maintaining the subjects at thermoneutrality.
it was impossible to impose all of the above conditions.
For example, many of the early studies in humans Clinical and physiological standards
reported by DuBois, Lusk and Rubner during the years
1900 –1920 did not strictly meet the requirement of a During the early studies on BMR, there were two schools
thermoneutral environment, leading to a slightly elevated of thought on how BMR values should be represented.
BMR. Moreover, many of the values reported by Aub and One group, called the ‘clinical standard’, assembled data
DuBois3,4 were obtained in anxious, untrained subjects. on first tests on supposedly ‘normal’ subjects. As is now
For this reason, the Aub –DuBois standards tended to be well known, first tests are usually higher in untrained
higher than other BMR standards. subjects. Therefore, these values and standards led to
The term ‘basal metabolism’ is often misunderstood to values that were usually higher than those of the
imply the lowest level of energy expenditure, which it ‘physiological standards’. On the basis of extensive review
clearly is not. During sleep and in conditions of under- and observations of BMR values at that time, Roth and
nutrition, metabolism may be lower than that observed Buckingham10 made the point succinctly as follows, ‘more
under basal conditions. To avoid this confusion, Krogh5 than one authority has stated on the basis of extensive
coined the term ‘standard metabolism’. In order to secure travel and observation, that as many as 70% of the basal
comparable results, the imposition of strict conditions for metabolism reports made today by the average operator
the measurement of BMR is essential. may not be worth the paper on which they are written’.
Basal metabolic rate studies in humans 1135
Early use of BMR measurements equations used to predict BMR today. It is instructive to
record that the concept of using energy expenditure to
BMR measurement in humans attracted considerable estimate energy requirements historically originated from
interest during the early part of the 20th century. They the 1957 report of the FAO13.
were primarily used for the diagnosis of hypo- and
hyperthyroidism. BMR tests marked a new era in clinical Current views on the use and application of BMR
medicine. It is interesting to note that until the mid-1950s
the only reliable measure to diagnose thyroid dysfunction The resurgence of interest in BMR can be directly ascribed to
was to determine a subject’s BMR. The study of basal a confluence of factors, notably two events. Firstly, a desire
metabolism represents an example of the early collabor- to understand the biology and aetiology of obesity and,
ation between physiologists and clinicians. BMR measure- secondly, the publication of the FAO/WHO/UNU document
ments were routinely made during clinical examinations Energy and Protein Requirements in 1985 that proposed, for
and were believed to be instrumental in the diagnosis of the first time, the use of energy expenditure (hence BMR)
thyroid disease, diabetes and leukaemia11. rather than food intake to calculate energy requirements.
This new approach to estimate energy requirements
Use of BMR to predict energy requirements emphasised the need to estimate accurately BMR in
populations living under various climatic and environmental
While BMR measurements were used in clinical diagnosis conditions. Under- or overestimation of BMR would severely
during the early part of the 20th century, the first affect the overall estimation of energy requirements.
comprehensive study to use BMR as the basis to estimate If BMR measurements are to be used in estimating
human energy requirements and hence food requirements energy requirements, it is important to have some details
was described by Bedale12. She studied a group of 45 boys on the apparatus used and the techniques adopted by
and 55 girls aged 7–18 years. This was a significant various investigators during the past 80 –90 years. The
departure from previous work as the early work on BMR following section will, therefore, review the methods and
was primarily intended to serve as metabolic reference apparatus used to measure BMR between 1900 and 2000.
values in clinical nutrition, notably in the diagnosis of
hypo- and hyperthyroidism. It is, therefore, of interest to Description of methodology: development of
note that the 1985 FAO/WHO/UNU1 approach to the apparatus to measure BMR
estimation of energy requirements is a refinement of the
method first described by Bedale12. With the growing importance of determining BMR in the
diagnosis and treatment of endocrine disorders (notably
thyroid disorders), the demand and use of calorimetry
FAO studies on calorie requirements
rapidly expanded between 1910 and 1950. It was
customary to use indirect calorimetry to measure BMR.
The FAO nutrition studies No. 1513 published in 1957,
The methods available to measure BMR may be divided
entitled Calorie Requirements, represented a landmark
into two types: closed and open circuit methods. In the
both in approach and analysis. Many of the previous
closed circuit methods, the CO2 produced is absorbed
reports on energy requirements proposed by Voit in 1890,
within the system. Oxygen is added to maintain the
Atwater in 1895, Lusk in 1918 and the NRC (USA)
volume of the gas constant. Benedict in 1918 initially
published in 1943 and revised in 195313, were based on
devised a method where the amount of CO2 absorbed by
food intake. The FAO13 publication proposed for the first
soda-lime was carefully replaced by O2 which could be
time the use of energy expenditure to calculate energy
measured. Later, Krogh5 and Roth14 developed an
requirements.
instrument that measured O2 consumption from the
Two simplified empirical equations to predict energy
reduction in the volume of the gas by using a spirometer.
requirements were presented:
An interesting feature related to the pioneering studies
Males: E ¼ 152W0.73 on BMR was that, until 1919, calorimetry was confined to
Females: E ¼ 123.47W0.75 experimental laboratories under the control of highly
trained scientists and technicians. With the advent of the
This was further simplified to: Benedict-Roth spirometer (in the 1920s), which was a
simple, portable calorimeter, the use and application
Males: E ¼ 815 þ 36.6W
spread widely. While this portable calorimeter was
Females: E ¼ 580 þ 31.1W
encouraged by some15, others were more critical. Roth
where, E represents total energy requirements and Buckingham10 commented ‘many of these technicians
(kcal day21) and W represents weight. educated overnight merely to man the machine, lacked
These simple linear equations to predict total energy training and experience necessary to face the multiplicity
requirements bear close resemblance to the linear of problems to be otherwise encountered. . .Their work
1136 CJK Henry
and reports were generally unquestioned because there period. If not, the changes in spirometer volume will
were few capable of checking and passing judgement’. represent not only the changes in O2 consumption but also
While it is impossible to gauge the impact such practices changes in lung volume.
had on BMR measurements at that time, it nevertheless In a series of studies comparing the closed to the open
reinforces the need to examine more carefully and circuit method to estimate BMR, Lewis et al.19 – 21 reported
critically the methods used to estimate BMR during the that the closed circuit overestimated BMR in 12 out of 25
early days of BMR collection. It is important to recognise adults and showed no difference in BMR measurements in
that closed circuit was the more widely used method to the other subjects. Willard and Wolf22 reported another
measure BMR in the early days of its study. source of error with closed circuit calorimetry. This
The most commonly used closed circuit apparatus to involved changes in chest volume during respiration.
measure BMR during the period 1910 –1950 were the Changes in the expiratory position of the chest occurring
following: during the experimentation had a marked effect on the
slope of the tracing. Thus, a change in chest volume by its
1. Krogh spirometer impact on the slope of the spirogram may lead to a falsely
2. Benedict spirometer (universal apparatus) high apparent metabolic rate. A further reason for the
3. Benedict-Roth spirometer elevated BMR values reported in the early literature was
4. Knipping apparatus due to the usual practice at that time to record values from
5. Fleish metabometer-metabograph the first test. An additional source of error relates to the
sample collection of O2 for a very short period of time
The major drawback of the closed circuit method was that (5 –10 minutes). If hyperventilation had occurred (which is
the absorption of CO2 did not allow the value of RQ to be quite common in untrained subjects) an overestimation of
calculated. As a rule, a value of 0.82 –0.85 was assumed, energy expenditure would occur16. Moreover, these early
leading to an error of up to ^ 6% since the food studies did not maintain standard BMR conditions prior to
undergoing oxidation had not been defined16. testing. It must be remembered that the early BMR studies
were used for clinical diagnostic purposes and not for any
Problems related to the use of closed circuit other metabolic usage.
apparatus BMR measurements using open circuit calorimetry show
little difference due to type of equipment used23 – 26 using
The accurate determination of BMR requires firstly, that the a range of methods, notably Douglas Bag, Oxylog, HB
subject be in the basal resting state (for either open or metabolator, ventilated hood, canopy and whole body
closed circuit calorimetry) and secondly for the rate of O2 calorimeter, showed marginal difference in BMR between
consumption to be measured accurately. In closed circuit methods used (see Table 1).
tests, the subject rebreathes from a spirometer that contains In contrast to the numerous studies comparing different
a CO2 absorber. It also contains O2 at a partial pressure methods of measuring BMR using open circuit calorimetry,
much greater than the atmosphere. The period of few studies have compared closed circuit with open circuit
rebreathing is usually 5–10 minutes as subjects become calorimetry. Such comparisons as there are suggest that
uneasy if prolonged. The rate of O2 consumption in the closed circuit calorimetry usually overestimates BMR (see
subject is calculated from the average rate of decrease in Table 2).
volume from the spirometer. Numerous difficulties in More recently, Clark and Hoffer27 who measured
obtaining accurate values using the spirometer have been BMR in 18–30 year old men either using a Deltatrac
reported. These include (1) completeness of CO2 absorbed, (open circuit) or a 9 litre respirometer (similar to a
(2) volumetric calibration, and (3) Kymographic accuracy. Benedict-Roth closed circuit apparatus), found that the
Both Krogh5,17 and Benedict18 commented on the ventilated hood measurements produced a BMR of
importance of the lung volume remaining effectively 6.87 ^ 0.619 MJ/24 h (1643 ^ 148 kcal/24 h) compared
a constant at the beginning and end of the experimental to 7.19 ^ 0.606 MJ/24 h (1721 ^ 145 kcal/24 h) in the

Table 1 Measurement reliability and reproducibility using indirect calorimetry

Whole-body
Oxylog HB metabolator Ventilated hood Ventilated tent calorimeter Power
(kcal day21) (kcal day21) (kcal day21) (kcal day21) (kcal day21) Difference (%) of t test

Protocol 1 (n 6) 1386.23 ^ 83.65 – – 1367.11 ^ 81.26 – þ 1.6 ^ 2.5 0.07

Protocol 2 (n 6) – 1460.33 ^ 64.53 1515.30 ^ 93.21 – – 23.1 ^ 2.5 0.15
Protocol 3 (n 6) 1364.72 ^ 59.75 1379.10 ^ 62.14 – – – 20.9 ^ 1.9 0.05
Protocol 4 (n 6) – 1412.52 ^ 59.75 1410.13 ^ 74.09 – – þ 0.4 ^ 2.5 0.03
Protocol 5 (n 10) – – 1321.70 ^ 45.41 – 1367.11 ^ 57.36 23.1 ^ 2.8 0.15
Basal metabolic rate studies in humans 1137
Table 2 Comparison between closed and open circuit BMR from 14 to 80 years of age (Table 3). These formulae are
still widely used despite being based on a group with only
Authors No. of subjects Difference
nine subjects and one cadaver!
Krogh and 5 of 19 6– 11% of open circuit
Rasmussen (1922) values
Hunt (1926) 20 of 25 12% of open circuit Harris–Benedict standards
values
Lewis et al. (1943) 13 of 25* 5% of open circuit
values While the ‘surface law’ remained a dominant concept since
Willard and Wolf (1951) 8 of 18* 10% of open circuit its introduction early in the 20th century, it was
values nevertheless strongly challenged by Harris and Benedict11
Fowler et al. (1957) 52* SD 7% of open circuit
values who embarked on a detailed biometric analysis of BMR
Harmin (1953) – SD 7% of open circuit which culminated in the publication of their monumental
values work entitled A Biometric Study of Basal Metabolism in
Abbreviation: BMR – basal metabolic rate. Man. BMR measurements were made on 136 males and
Source: Consolazio et al.26 103 females at the Carnegie Nutrition Laboratory in
*Adults.
Boston. Using rigorous statistical concepts, they devel-
oped the following equations to predict BMR:
closed circuit respirometer. These values indicate that
the closed circuit apparatus produced a BMR 5.6% males h ¼ 66:4730 þ 13:7516W þ 5:0033S 2 6:7750A
higher than the open circuit.
Summary females h ¼ 665:0955 þ 9:5634W þ 1:8496S 2 4:6756A
1. Closed circuit calorimetry was widely used in the
measurement of BMR during the first half of the 20th where, h ¼ kcal day21; W ¼ weight in kilograms; S ¼
century. stature in centimeters; A ¼ age in years.
2. Closed circuit calorimetry tended to produce higher Harris and Benedict’s analysis marked a significant
BMR values. departure from previous work. Firstly, it introduced for the
3. The elevated values were ascribed to: first time biometric principles in its analysis. Secondly,
(a) Small leaks causing a larger error in the closed they used subjects that were maintained under strict
circuit calorimetry28. experimental conditions prior to the measurements. Using
(b) The respiration of pure oxygen tended to elevate partial correlation coefficients, they also showed that both
BMR16. stature and weight have an independent effect on BMR.
(c) Changes in chest volume during respiration While these equations were useful and valuable aids to
tended to alter the slope of the spirogram reading, predicting BMR, they were not above criticism. For
leading to an apparently higher BMR value22. example, the constant in the equation showed a ten-fold
difference between males and females (66 versus 665).
Benedict himself later recognised and expressed concern
Brief review of BMR standards and predictive that the equations overestimated BMR, ‘particularly in
equations those young women’. Daly et al.31 confirmed that the
Harris –Benedict equations overestimated BMR by about
This section briefly reviews the predictive equations for 10 –15%. Despite this, the simplicity of the Harris –
BMR in man beginning with the ‘surface area law’ at the Benedict equation made it a popular equation in wide
turn of the 20th century to the more recent analysis by use. Even today, many clinicians in North America use it
Schofield et al.29 routinely32.

DuBois height –weight formula chart

Table 3 DuBois normal standards for BMR (Cal m22 per hour)
While surface area may be calculated using various
Age (y) Males Females
anthropometric parameters, DuBois and DuBois30 pro-
duced an equation relating weight and height to surface 14– 15 46.0 43.0
16– 17 43.0 40.0
area as follows: 18– 19 41.0 38.0
20– 29 39.5 37.0
A ¼ W0:425 £ H0:715 £ 71:84 30– 39 39.5 36.5
40– 49 38.5 36.0
where, A ¼ surface area in cm2; W ¼ weight in kilograms 50– 59 37.5 35.0
and H ¼ height in centimeters. 60– 69 36.5 34.0
70– 79 35.5 33.0
Later, Aub and DuBois3,4, applying the surface law
principle to man, published a table of BMR m22 per hour Abbreviation: BMR – basal metabolic rate.
1138 CJK Henry
Boothby and Sandiford or ‘Mayo standards’ Table 4 The percentages by which the FAO/WHO/UNU
equations overestimate (þ) or underestimate (2) the actual BMR
in different ethnic groups
Scientists at the Mayo clinic commenced collecting BMR
data systematically in a variety of subjects from 1917. The Age group (y) Mean % No. of subjects
investigators used a combination of normal, free-living Males (all ethnicities)
subjects and ‘hospital normal’ subjects. While the subjects 3– 10 þ 1.9 196
admitted to hospital were not seriously ill, they never- 10– 18 þ 7.1 409
18– 30 þ10.3 1174
theless highlight the point that the subjects were not all 30– 60 þ11.2 274
‘normal’ free-living subjects33,34. Boothby et al.15 made a 3– 60 þ 9.0 2053
careful study of BMR in 639 males and 828 females. Females (all ethnicities)
3– 10 þ 1.5 88
10– 18 þ 7.6 233
Quenouille standards 18– 30 þ 3.8 350
30– 60 þ 9.7 98
3– 60 þ 5.4 769
Quenouille et al.’s35 analysis in 1951 was the first All ethnicities, all ages, both sexes þ 8.0 2822
comprehensive survey of all the available BMR studies
Abbreviations: FAO/WHO/UNU – Food and Agriculture Organization/World
conducted and represented over 8600 subjects (4300 aged Health Organization/United Nations University; BMR – basal metabolic
between 17 and 39 years, 800 over 40 years and 3520 less rate.
Source: Henry and Rees37.
than 1 year of age). We need to pay tribute to these
investigators who were ‘pioneers’ in the systematic
collection of BMR and they statistically analysed the data comprised 114 published studies of BMR, totalling 7173
prior to the advent of computers. Quenouille et al.’s35 data points. Although their database comprised almost
extensive review of the early literature on BMR has also 11 000 BMR values (including group mean values), most of
been a major source of valuable information for both the the results were obtained from European and North
Schofield and Oxford databases. For the first time, American subjects. An interesting feature that emerged
Quenouille et al.35 also included BMR measurements from their analysis was that the BMR of Asiatic Indians was
from people living in the tropics. Their analysis attempted overestimated by 10–11% by their equations. This issue
to examine the role of ethnicity and climate on BMR. This was further highlighted by the FAO/WHO/UNU report. At
made it the first large-scale study of the world literature on the time of their analysis, there was insufficient data to
BMR. While they considered temperature and humidity as ascertain whether the effect noted in Indians was unique or
important factors in predicting BMR, sadly their equations whether it reflected a general pattern of metabolism in
were not used extensively. Given below is an example of tropical peoples. Indeed, the observation that BMR may be
their equation for men in Northern Europe. different in peoples living in the tropics was first reported
by de Almeida36. He showed that BMR in Brazilians was
M ¼ 2:975H þ 8:90W þ 11:7S þ 3:0h 2 4:0t þ 293:8 approximately 24% lower than the Aub–DuBois standards.
where, M ¼ kcal day21; W ¼ weight in kilograms; H ¼ Subsequently, Henry and Rees37 showed that the FAO/
height in centimeters; S ¼ surface area from DuBois;
t ¼ temperature and h ¼ humidity.
Table 5 The percentage by which the FAO/WHO/UNU equations
Schofield equations (FAO/WHO/UNU equations): overestimate (þ ) or underestimate (2) BMR in different ethnic
groups by sex, all ages 3 –60 years
issues and analysis
Male Female
Note here that the term FAO/WHO/UNU equations and Ethnicity Mean % Sample size Mean % Sample size
Schofield equations will be used interchangeably. To
Philippino þ9.5 172 þ 1.1 31
facilitate the 1981 FAO/WHO/UNU expert consultation on Indian þ 12.8 50 þ12.9 7
Energy and Protein Requirements, Durnin16 surveyed the Japanese þ5.8 202 þ 4.6 152
literature on BMR and assembled BMR values and South American þ9.4 941 þ 4.8 227
Chinese þ7.6 274 þ 3.8 190
anthropometric data on 2238 subjects. Durnin16 presented Malayan þ9.3 62 No data
tables to predict BMR based on body weight, age and Javanese þ5.0 86 No data
gender. Subsequently, the FAO/WHO/UNU requested Mayan þ1.5 76 No data
Ceylonese þ 22.4 125 þ12.5 100
Schofield et al.29 to extend this analysis and produce a African þ6.5 20 No data
series of predictive equations. Schofield et al.29 reviewed Hawaiian þ7.2 19 þ 4.5 62
the literature and produced predictive equations for both Samoan þ3.3 21 No data
All þ9.0 2053 þ 5.4 769
sexes for the following ages: 0–3, 3–10, 10 –18, 18 –30,
30 –60 and .60 years. These formed the basis for the Abbreviations: FAO/WHO/UNU – Food and Agriculture Organization/World
Health Organization/United Nations University; BMR – basal metabolic
equations used in the FAO/WHO/UNU document Energy rate.
and Protein Requirements1. The Schofield database Source: Henry and Rees37.
Basal metabolic rate studies in humans 1139
WHO/UNU equations overestimated BMR in a range of over- and underestimation in BMR using the FAO/WHO/
tropical populations (see Tables 4 and 5). UNU equations observed in children aged 2.5–18 years was
While Schofield’s analysis has served a significant role in reported by several investigators50 – 53. These children were
re-establishing the importance of using BMR to predict studied in a range of countries, including Columbia,
human energy requirements, recent workers have sub- Guatemala and China. While all of the above studies
sequently queried the universal validity and application of reported an overestimation of BMR in these children, the
these equations38 – 43. Other authors44,45 have questioned studies by Livingstone et al.54 reported an underestimation
the equations’ continued use in present day populations and Bandini et al.55 no significant difference when the FAO/
with their secular changes in body weight and body WHO/UNU equations were used. BMR studies conducted
composition46 – 48. In contrast to the previous observation more recently have also shown that the predicted values
that the FAO/WHO/UNU equations overestimated BMR in using the FAO/WHO/UNU equations overestimate BMR in
tropical people49, further analysis shows that the FAO/ Asian and Chinese subjects56,57. For example, in the study
WHO/UNU equations tend to overestimate BMR in most by Leung et al.57 the FAO/WHO/UNU equation over-
populations and these overestimations are not small or estimated BMR from the measured values by up to
insignificant (see Tables 6 and 7) . A detailed analysis of the 456 ^ 67 kJ (measured BMR 5.025 ^ 0.967 MJ; predicted

Table 6 Comparison of BMR values from the literature: observed values vs. predicted values using the Schofield equations

Investigator Year Subjects Age range (y) Male Female Total % difference

Owen et al. 1986 Mixed Race 18– 65 – 34 34 þ11

Americans
Owen et al. 1987 Mixed Race 18– 82 48 – 48 þ5
Caucasians
African-American
Oriental
Soares and Shetty 1988 Indian 18– 30 123 123 þ 9.3
Urban upper socio-econ 47 þ 6.6
Urban lower socio-econ 36 þ12.9
Rural 40 Range þ 5.5–12.6
Mifflin et al. 1990 Mixed Race 19– 78 251 247 498 þ 6.2 (males)
Americans þ 2.2 (females)
Spurr et al. 1992 Colombian 2– 16 153 93 246 Range
Males 3– 10 þ 2.7–9.4
10– 18 þ 2.5
Females 3– 10 þ 9.4
10– 18 þ 1.4
Maffeis et al. 1993 Italian 6– 10 62 68 130 No
Males
Non-obese 48 – Difference
Obese 14 – þ15.3
Females
Non-obese – 49 þ 8.0
Obese – 19 þ20.0
Rieper et al. 1993 German 14– 15 – 11 11 No difference
Valencia et al. 1993 Mexican 18– 40 32 – 32 þ 8.2
Fontville and Ravussin 1993 Mixed Race 7– 12
Caucasian 21 21 42 No difference
Pima Indian 22 21 43 þ 6.7
Piers and Shetty 1993 Indian 18– 30 – 60 60 þ 9.2
Past Indian Study – 52 þ10.5
European/American – 52 þ 4.1
Molnar et al. 1995 Hungarian 10– 16 193 178 371
Males
Non-obese 116 – þ 7.8
Obese 77 – þ13.4
Females
Non-obese 119 – þ 8.0
Obese 59 – þ 5.8
Liu et al. 1995 Chinese 20– 78 102 121 223 þ15.1 (males)
þ17.9 (females)
Wong et al. 1996 Mixed Race 8– 17 – 118 118
Caucasian – 76 No difference
African-American – 42 þ8
Piers et al. 1997 Australian 18– 30 39 89 128 þ 5.3 (males)
þ 2.2 (females)

Abbreviation: BMR – basal metabolic rate.

1140 CJK Henry
Table 7 Contribution of Italian subjects to Schofield database similarity in body composition exists between the surveyed
database and the test population applied to. It is increas-
Percentage
Gender Age (y) Italians (n) Schofield (n) of total ingly clear that the subjects from whom the BMR database
(originally from data gathered over 80–90 years ago) was
Males 0 –3 0 162 0.0
3 –10 158 338 46.7 assembled had a different body composition to that seen
10–18 472 734 64.3 today (Norgan)61. The discussion and debate surrounding
18–30 1740 2879 60.4 the use and application of the BMR equations were
30–60 392 646 60.7
60þ 0 50 0.0 summarised by Soares et al.44 as follows:
Total 2762 4809 57.4 ‘Whatever the reason it is becoming increasingly evident
Females 0 –3 0 137 0.0 that the equations of Schofield derived from measure-
3 –10 220 413 53.3
10–18 167 575 29.0 ments made over 60 years ago are not at present valid for
18–30 135 829 16.3 the precise prediction of BMR of population groups
30–60 106 372 28.5 worldwide’.
60þ 6 38 15.8
Total 634 2364 26.8
Contribution of Italian subjects to the Schofield
database
5.481 ^ 0.845 MJ). Recently, Piers et al.58 reported that the
FAO/WHO/UNU equations overestimated BMR in male A closer examination of the Italian data points in the
and female Australians. Schofield database reveals some issues of major concern.
A survey of the most recent studies (1980–2000) in BMR The Italian group represented 3388 BMR data points from
suggests that in most cases the current FAO/WHO/UNU a total of 7173 values in the Schofield database. These 3388
predictive equations overestimate BMR in many commu- BMR values for Italians came from just nine papers that
nities. The few exceptions to this general trend are the were published between 1936 and 1942 and contributed
reports by Bandini et al.,55 Livingstone et al.,54 Ferro-Luzzi 60–64% of the Schofield database, depending on the age
et al.59 and Yamauchi et al.60. These authors either showed group under consideration (see Table 7).
an underestimation or good agreement with the FAO/ A significant feature of the Schofield database was that it
WHO/UNU equations. Contrary to the popular view that shared a large proportion of the original database initially
only people in the tropics may have lower BMR, the recent identified by Quenouille et al.35 Table 8 illustrates this
study by Wong et al.43 showed that BMR in African- point.
American children aged 8–17 years was 7% lower than The Schofield database contained the same nine Italian
that predicted by the FAO/WHO/UNU equations. What- papers first identified by Quenouille et al.35 Quenouille
ever the reasons for these varied observations, it is clear analysis35 delineated four populations: North European
that the present FAO/WHO/UNU equation tends to and Americans; Italians; Asians; and a ‘residual mixed
overpredict BMR in many communities. group’. Within these groups, Italians had the highest BMR
Another significant feature of the Schofield database per kilogram. Schofield et al.29 also noted a similar higher
was that for males aged between 10 and 60 years, over BMR per kg in the Italians. Since the 18 –30 year old male
3000 (50%) data points come from Italian subjects. The group in the Schofield database contained the largest
Italian group appear to have a higher BMR per kilogram number of Italian subjects (1740 out of a total of 2879), this
than any other Caucasian group49,41. More importantly, group may be further analysed when matched for body
the inclusion of this disproportionately large Italian group size and body mass index (BMI).
with a higher BMR per kilogram may have artificially Table 9 shows weight, height, BMR and BMR per
elevated the Schofield predictive equations. Indeed this kilogram in BMI-matched Italian and North European and
view was first expressed by Schofield29. He wrote: American subjects in the Schofield database. What is clear
‘The equation for adult males and females (18 –30) were is that the Italian subjects show an elevated BMR
recalculated excluding Italian subjects. The new equations (MJ day21 or kJ kg21 per day) when also matched for
were: BMI62.
Males: BMR(MJ/24 h) ¼ 0.0582W þ 3.2399
SEE ¼ 0.6148
Table 8 Papers and data points shared by Schofield29 and
Females: BMR(MJ/24 h) ¼ 0.0545W þ 2.5135 Quenouille et al.35
SEE ¼ 0.4813
Number Number of
When these equations were used to predict BMR for Investigator of papers data points
Italian subjects there was a highly significant lack of fit for Quenouille 89 7434
both males and females’. Schofield 114 7173
Papers common to Quenouille 50 6124
The applicability and use of body weight to predict BMR
and Schofield
in various populations depends on the assumption that a
Basal metabolic rate studies in humans 1141
21
Table 9 Wilcoxon U-tests of weight, height, BMR and BMR kg Summary
in BMI matched Italian and North European and American
(NE&A) subjects in the Schofield database
A need to re-assess the FAO/WHO/UNU equations to
BMR predict BMR is prompted by the following:
Weight Height BMR (kJ kg21
21
BMI range Group n (kg) (m) (MJ day ) per day) 1. The current FAO/WHO/UNU equations appear to
18.0–18.9 NE&A 36 54.8 1.72 6279 115 overestimate BMR in many populations (both in
Italian 29 54.0 1.70 6476 120 tropical and temperate regions). Some studies,
19.0–19.9 NE&A 63 58.4 1.73 6467 111 however, show good agreement with the FAO/
Italian 78 56.5* 1.70* 6890*** 122***
20.0–20.9 NE&A 78 61.8 1.74 6635 107 WHO/UNU equations.
Italian 218 59.6*** 1.70*** 6966** 117*** 2. If BMR values are to be used to estimate total energy
21.0–21.9 NE&A 85 66.2 1.75 6932 105 requirements for prescriptive and diagnostic pur-
Italian 313 62.1*** 1.70*** 7007 113***
22.0–22.9 NE&A 59 68.1 1.74 6932 102 poses (in both individuals and populations), any
Italian 314 65.3*** 1.71*** 7204* 110*** overestimation in BMR is likely to produce mislead-
23.0–23.9 NE&A 38 71.8 1.75 7259 101 ing estimates of energy requirements.
Italian 298 67.9*** 1.70*** 7259 107
24.0–24.9 NE&A 18 74.7 1.75 7217 97 3. The FAO/WHO/UNU equations to predict BMR were
Italian 155 70.6* 1.70 7480 106** developed using a database that contained a
25.0– 25.9 NE&A 8 73.9 1.70 7208 98 disproportionate number – 3388 out of 7173 (47%)
Italian 53 72.3 1.69 7468 103
– of Italian subjects.
Abbreviations: BMR – basal metabolic rate; BMI – body mass index. 4. These Italian subjects had a higher BMR per kilogram
Significant difference: *P , 0.05; **P , 0.01; ***P , 0.001.
Source: Hayter and Henry63. than any other group in the Schofield database.
5. The numerical dominance of Italians in the Schofield
database and their apparent elevated BMR recorded
The results of this analysis suggest that previous studies may largely explain why the Schofield equations
appear to have either overlooked or ignored such large overestimate BMR in most populations today.
population differences in BMR per kilogram observed in 6. It is recommended that the Italian database be
the Italians. The nine papers from the Italian investigators removed from any future analysis as (1) they have
used in the Schofield analysis are shown in Table 10. unusually high BMR values (even when normalised
All of the Italian studies used the Benedict – Roth for body weight and BMI); (2) they are over-
spirometer, an indirect closed circuit method. Our earlier represented in the Schofield database; and (3) their
review has suggested that closed circuit methods are more over-representation has meant by definition an
likely to lead to higher BMR values compared to open under-representation of other world populations.
circuit. The subjects used in the Italian studies were 7. The Schofield database contained relatively few
primarily young, the males tending to lead physically active subjects from the tropical region (322 Indian and 615
lives. This is especially so in the case of labourers and tropical residents) making it a poor representation of
miners; these occupations are known to have very high the global population.
rates of daily energy expenditure. The work of Poehlman 8. The primary purpose of collecting BMR measures
et al.64 indicates that physically active individuals have during the early part of the 20th century was to
higher BMRs compared to sedentary lifestyles. Whatever diagnose hypo- and hyperthyroidism.
the reasons for the elevated BMR noted in the Italians, their 9. The primary purpose of collecting BMR measures in
numerical dominance in the Schofield database may largely recent years has been to estimate total energy
explain why the Schofield equations (FAO/WHO/UNU requirements or to better understand energy
equations) overestimate BMR in present day populations. regulation.

Table 10 Italian data used in the Schofield database29

Study n Gender Age (y) Subject details

Felloni (1936) 532 Male 19– 25 Students of the Royal Fascist Academy
Granti and Busca (1941–1942) 186 Male 16– 55 Labourers and miners on shift work
Lafratta (1937) 213 Male 14– 20 Students of Naples Royal Military College
Lenti (1937) 525 Male 20– 25 Military servicemen
Occhiuto and Pepe (1939) 247 Female 20– 67 Different social groups
Occhiuto and Pepe (1940) 571 Male 22– 54 Police officers
Pepe (1938) 252 Male 18– 24 Students of Royal Naval Academy
Pepe and Perrelli (1937) 267 Male 5– 16 No details
235 Female 5– 12 No details
Pepe and Rinaldi (1936) 217 Male 6– 16 No details
143 Female 5– 12 No details
Total 3388
1142 CJK Henry
10. This change of emphasis and role of BMR has placed MEDLINE. As papers began to accumulate, it rapidly
it within a different nutritional paradigm. became clear that the quantity of data gathered was
11. If BMR equations are to be used and applied uneven. The research methods presented the description
worldwide, the database must contain a more of subjects and conditions varied from complete details to
representative sample of the world population. no information. To preserve uniformity and to meet the
criteria defining basal metabolism, it was decided to
The objective of all consultations is to scientifically progress include data for further analysis only if the following
and identify fresh ideas and issues. Many of the information was provided in the papers:
technological advances that have emerged during the past
1. Age, weight, gender of subjects
few years have enabled the measurement of BMR to be
2. Description of experimental conditions and equipment
conducted with ease and reproducibility. This is an
used to measure BMR.
opportune moment to re-examine the role and place of
3. Post-absorptive, rested subjects.
BMR measurements in estimating total energy requirements
4. Subjects that were described as ‘healthy’ (i.e. not
today, using a more representative world population base.
suffering from any illness).
Development of the Oxford database 5. Location/ethnicity of subjects.

Initial selection criteria for BMR data in the Oxford Reasons for rejecting data for further analysis included the
Database: preliminary screening following:
There has been considerable disagreement in the literature 1. BMR presented only in terms of surface area (no height
as to the ‘best’ way in which BMR data should be selected or weight provided), therefore BMR/24 hours could
and collected. not be calculated.
The four methods used to ‘accept’ values for BMR 2. BMR presented as a percentage deviation of other
reported in the literature include: standards (usually Harris – Benedict and DuBois).
1. Mean of all determinations (the BMR of the subject was 3. BMR measured on malnourished or sick subjects.
taken as the mean of all determinations taken on the 4. BMR measured below 188C.
subject). 5. BMR measured at high altitudes.
2. First determination (only the first observation was 6. BMR measured in subjects who had eaten or drunk
considered). coffee.
3. Lowest of all determinations (only the lowest value
was chosen). It was not possible to obtain information on the ambient
4. Mean of the lower of three duplicates (the lower values temperature at the time of BMR measurements in all papers.
in each of the 3 days were averaged and taken as the However, in papers where temperature was described, if
BMR of the subject). the ambient temperature was below 188C, such papers
were rejected. In most papers, BMR was expressed as
While an element of training has been considered by kcal day21, kJ day21, kcal kg21 per day or kcal m22. When
several investigators as an important factor in BMR BMR was expressed as O2 consumption, with no RQ values
determinations, the BMR report by Robertson and Reid65 reported, an energy equivalence of 4.9 was used. By the
from a large study in Britain has been excluded from wide time the compilation was complete, data were available for
use as the investigation used the lowest values recorded 10 552 subjects (5794 males and 4702 females). The data
after several bouts of collection. Durnin16 suggested that came from 166 separate investigations. In this analysis, only
no significant effect was produced by the method of individual data points were used. Several studies that
selection of BMR data with the exception of including presented data as group means were excluded.
lowest BMR results. Apart from excluding BMR values that In the cases of researchers such as Benedict18,66 – 75,
were reported as lowest values, all other BMR values were Lewis20,76,77, Mason78 – 81, Nakagawa82 – 85, Wardlaw86 – 88
initially considered for inclusion in the Oxford survey, and Wang90 – 93, even if descriptive details were not
prior to further screening. provided in all their papers, they were included for further
analysis as their protocol was detailed (and acceptable) in
Computation of new BMR equations (Oxford the first of their papers.
equations) The Oxford database also excluded all the Italian
subjects due to their unusually high BMR values. To ensure
Oxford data compendium quality data for the equations to estimate BMR, further
Once such papers were identified (as outlined above), the screening took place. All individual data was screened to
literature search produced numerous papers on BMR. identify errors of data input and transcription. Screening
Much of the identification of the early studies on BMR was also allowed outlying or extreme cases to be identified and
done by hand search. This was supplemented by using removed, if appropriate, from the database. As well as
Basal metabolic rate studies in humans 1143

Fig. 1 Basal metabolic rate (BMR) vs. body weight – males Fig. 3 Basal metabolic rate (BMR) vs. body weight – females
18– 30 years 18–30 years

screening data on an individual basis, screening also took subjects for illustrative purposes. It is evident that the
place at study level. Italians once again show considerable difference with the
The value of a large database that draws on information Oxford trendline (Italian trendline – top line on figures;
collected by a wide range of investigators rests on the Oxford regression – top left-hand corner on figures).
assumption that all investigators adopted a ‘standard’ To further substantiate why the Italian subjects have
practice to measure BMR – which clearly they did not. been excluded from the Oxford database, Table 11 shows
Although strict inclusion criteria had been used to develop descriptive statistics between the Italian subjects and the
the Oxford database, very similar to those adopted by rest of the Oxford database. The age bands 10– 18 years,
Schofield et al.29, the fact that such criteria must, of 18–30 years and 30 –60 years only were chosen for
necessity, rely on published reports of methods and analysis as they contain the largest number of Italian
protocol needs to be recognised. subjects. It is evident that the Italian subjects show
significant differences in BMR, even when expressed as
BMR day21 or BMR/kg/body weight.
Computation of equations from Oxford database Table 12 contains the equations for predicting BMR from
weight alone and descriptive statistics for the Oxford
A series of plots of BMR against body weight were equations.
performed at six different age groups (0– 3 years, 3– 10 Equations to predict BMR from weight for six separate
years, 10– 18 years, 18– 30 years, 30–60 years and . 60 age groups and gender are presented in Table 13, along
years) for males and females. Representative examples are with the FAO/WHO/UNU equations for comparison.
shown in Figs 1–4 . These represent BMR vs. body weight Given that a reasonably large number of BMR values
in the Oxford database and compares them with the Italian from elderly subjects were available, it was decided to

Fig. 2 Basal metabolic rate (BMR) vs. body weight – males Fig. 4 Basal metabolic rate (BMR) vs. body weight – females
30– 60 years 30–60 years
1144 CJK Henry
Table 11 Comparison of descriptive statistics (mean ^ SD ) for Italian data vs. all data in Oxford database

Gender Age (y) Source Weight (kg) Height (m) BMI BMR (kJ kg21 per day) BMR (MJ day21)

Males 10– 18 All1 40.0 ^ 12.5 1.49 ^ 0.146** 17.7 ^ 2.65*** 143 ^ 22.6*** 5.51 ^ 1.11***
Italian 41.5 ^ 15.4 1.47 ^ 0.171 18.4 ^ 3.10 151 ^ 26.8 5.89 ^ 1.27
18– 30 All1 61.0 ^ 11.4*** 1.70 ^ 0.0872 20.9 ^ 2.84*** 106 ^ 12.8*** 6.36 ^ 1.00***
Italian 64.8 ^ 6.83 1.70 ^ 0.0658 22.4 ^ 1.79 111 ^ 11.9 7.15 ^ 0.767
30– 60 All1 65.3 ^ 13.0 1.69 ^ 0.0942* 22.8 ^ 3.2* 98.7 ^ 13.6*** 6.35 ^ 1.03***
Italian 65.6 ^ 10.3 1.68 ^ 0.0538 23.3 ^ 3.35 109 ^ 14.5 7.04 ^ 0.772
Females 10– 18 All1 43.4 ^ 12.9*** 1.50 ^ 0.113*** 18.8 ^ 3.64*** 126 ^ 24.1*** 5.20 ^ 0.797***
Italian2 28.4 ^ 4.42 1.33 ^ 0.0690 16.0 ^ 1.55 155 ^ 17.5 4.36 ^ 0.475
18– 30 All1 53.2 ^ 10.0*** 1.60 ^ 0.0755*** 20.7 ^ 3.18*** 99.8 ^ 12.7*** 5.24 ^ 0.786***
Italian 58.0 ^ 10.3 1.57 ^ 0.0612 23.4 ^ 3.97 106 ^ 14.3 6.04 ^ 0.688
30– 60 All1 59.1 ^ 13.7 1.59 ^ 0.0792*** 23.3 ^ 4.48*** 92.0 ^ 14.3*** 5.30 ^ 0.804***
Italian 60.7 ^ 13.0 1.56 ^ 0.605 25.0 ^ 5.00 98.4 ^ 17.8 5.72 ^ 0.605

Significant difference: *P , 0.05; **P , 0.01; ***P , 0.001.

1
Excludes Italians
2
Italian girls all just over 10 years.

break down the elderly group into ‘young elderly’ and adults the greatest differences in BMR were found in
‘older elderly’ (see Table 14). 18–30 and 30 –60 year age groups within the lower
Figures 5–8 illustrate the relationship between body body weight ranges (, 60 kg) in males. There was a
weight and BMR in the elderly groups. steadily increasing difference in BMR with decreasing
While Table 13 contains equations for predicting body weight in these two age groups (18 –30 and
BMR from weight alone, Table 15 contains equations 30–60 years). For females, differences in BMR were
using height and weight. To enable the comparison of the seen in most of the body weight range notably in the
effect of including height as a second variable for males age groups 18– 30, 30–60 and .60 years. In the lower
and females, the equations were re-calculated using age groups (0 –3 years), the FAO/WHO/UNU equations
weight/height as independent variables across the entire appear to underestimate BMR, both in males and
age range. Table 16 gives mean values and standard females. At 3– 10 years of age, the FAO/WHO/UNU
deviations for the various age groups in the Oxford equations underestimate BMR in males and over-
analysis. estimate BMR in females. The differences in BMR
Any improvement in using height and weight to predict assume significance when one considers that the lower
BMR was tested. Table 17 shows that no significant body weights (50 –60 kg) are those that are commonly
advantage was afforded in predicting BMR with the observed in many developing countries. Moreover, the
inclusion of height. ages (18 –60 years) that show the greatest differences in
BMR are ages at which most populations are in a
productive stage of occupational activity. Over- and
Applications of new BMR equations under-prediction of BMR in these groups may have
significant influences on estimating their energy
Using the Oxford equations, BMR was calculated for a requirements and hence food needs. Table 20
range of body weights and ages (0 –3, 3–10, 10 –18, summarises these differences in BMR at various ages.
18 –30, 30–60, . 60 years) for both males and females It is significant to note that the ages at which differences
(Tables 18 and 19). It is interesting to note that in in BMR were recorded between the new Oxford equations

Table 12 Descriptive equations and statistics (mean ^ SD ) of Oxford predictive equations for BMR

Gender Age (y) MJ day21 kcal day21 SE n r

Males 0– 3 0.255W 2 0.141 61.0W 2 33.7 0.255 277 0.954

3– 10 0.0937W þ 2.15 23.3W þ 514 0.328 289 0.827
10– 18 0.0769W þ 2.43 18.4W þ 581 0.566 863 0.861
18– 30 0.0669W þ 2.28 16.0W þ 545 0.652 2821 0.760
30– 60 0.0592W þ 2.48 14.2W þ 593 0.693 1010 0.742
60 þ 0.0563W þ 2.15 13.5W þ 514 0.685 534 0.776
Females 0– 3 0.246W 2 0.0965 58.9W 2 23.1 0.242 215 0.960
3– 10 0.0842W þ 2.12 20.1W þ 507 0.360 403 0.820
10– 18 0.0465W þ 3.18 11.1W þ 761 0.525 1063 0.752
18– 30 0.0546W þ 2.33 13.1W þ 558 0.564 1664 0.700
30– 60 0.0407W þ 2.90 9.74W þ 694 0.581 1023 0.690
60 þ 0.0424W þ 2.38 10.1W þ 569 0.485 334 0.786

Abbreviation: BMR – basal metabolic rate.

Basal metabolic rate studies in humans 1145
Table 13 New Oxford equations with FAO/WHO/UNU equations
for comparison

BMR Oxford
Gender Age (y) (MJ day21) BMR FAO (MJ day21)

Males 0 –3 0.255W 2 0.141 0.255W 2 0.226

3 –10 0.0937W þ 2.15 0.0949W þ 2.07
10–18 0.0769W þ 2.43 0.0732W þ 2.72
18–30 0.0669W þ 2.28 0.0640W þ 2.84
30–60 0.0592W þ 2.48 0.0485W þ 3.67
60þ 0.0563W þ 2.15 0.0565W þ 2.04
Females 0 –3 0.246W 2 0.0965 0.255W 2 0.214
3 –10 0.0842W þ 2.12 0.0941W þ 2.09
10–18 0.0465W þ 3.18 0.0510W þ 3.12
18–30 0.0546W þ 2.33 0.0615W þ 2.08
30–60 0.0407W þ 2.90 0.0364W þ 3.47
60þ 0.0424W þ 2.38 0.0439W þ 2.49 Fig. 7 Basal metabolic rate (BMR) vs. body weight – females
60–70 years
Abbreviations: FAO/WHO/UNU – Food and Agriculture Organization/World
Health Organization/United Nations University; BMR – basal metabolic
rate.

Table 14 Descriptive equations and statistics (mean ^ SD) of

Oxford predictive equations for BMR in the elderly

Gender Age (y) MJ day21 kcal day21 SE n r

Males 60– 70 0.0543W þ 2.37 13.0W þ 567 0.697 270 0.766

70 þ 0.0573W þ 2.01 13.7W þ 481 0.667 264 0.779
Females 60– 70 0.0429W þ 2.39 10.2W þ 572 0.476 185 0.798
70 þ 0.0417W þ 2.41 10.0W þ 577 0.518 155 0.746

Abbreviation: BMR – basal metabolic rate.

Fig. 8 Basal metabolic rate (BMR) vs. body weight – females

70þ years

and FAO/WHO/UNU equations coincide with the ages at

which a disproportionate number of Italian subjects were
included in the Schofield database.
Practical examples of how the use of Oxford BMR
equations influence energy requirements

On the basis of the new BMR equations, it is now possible

to predict energy requirements in subjects performing
Fig. 5 Basal metabolic rate (BMR) vs. body weight – males different tasks.
60– 70 years
Tables 21 and 22 predict total energy requirements at
light, moderate and heavy physical activity levels in
18–30 year old males and females. It is apparent, when
applying the Oxford equation for BMR, that a reduction
in total energy requirements ranges from 396 kJ (95 kcal)
to 841 kJ (201 kcal) per day for males, and 202 kJ
(48 kcal) to 863 kJ (206 kcal) per day for females. As
illustrative examples, Tables 23–28 show the effect an
alteration in BMR will have on total energy requirements
and cereal needs in various occupations. In the case of a
subsistence farmer, moderately active, weighing 58 kg,
the reduction in energy needs per day turns out to be
676 kJ (162 kcal) and a reduction in cereal requirement
of 41 g day21 or 15 kg per year. In the case of a 35-year
Fig. 6 Basal metabolic rate (BMR) vs. body weight – males 70þ old male, engaged in heavy physical activity, weighing
years 65 kg, the reduction in energy needs become 1064 kJ
1146 CJK Henry
Table 15 Descriptive equations and statistics (mean ^ SD ) of Oxford prediction equations for BMR using height
and weight

Gender Age (y) MJ day21 kcal day21 SE n r

Males 0 –3 0.118W þ 3.59H 2 1.55 28.2W þ 859H 2 371 0.246 246 0.959
3 –10 0.0632W þ 1.31H þ 1.28 15.1W þ 74.2H þ 306 0.322 289 0.835
10–18 0.0651W þ 1.11H þ 1.25 15.6W þ 266H þ 299 0.562 863 0.864
18–30 0.0600W þ 1.31H þ 0.473 14.4W þ 313H þ 113 0.648 2816 0.764
30–60 0.0476W þ 2.26H 2 0.574 11.4W þ 541H 2 137 0.678 1006 0.756
60 þ 0.0478W þ 2.26H 2 1.07 11.4W þ 541H 2 256 0.668 533 0.789
Females 0 –3 0.127W þ 2.94H 2 1.20 30.4W þ 703H 2 287 0.232 201 0.964
3 –10 0.0666W þ 0.878H þ 1.46 15.9W þ 210H þ 349 0.357 403 0.825
10–18 0.0393W þ 1.04H þ 1.93 9.40W þ 249H þ 462 0.521 1063 0.758
18–30 0.0433W þ 2.57H 2 1.18 10.4W þ 615H 2 282 0.542 1655 0.724
30–60 0.0342W þ 2.10H 2 0.0486 8.18W þ 502H 2 11.6 0.564 1023 0.713
60 þ 0.0356W þ 1.76H þ 0.0448 8.52W þ 421H þ 10.7 0.472 324 0.805

Abbreviation: BMR – basal metabolic rate.

Table 16 Descriptive statistics (mean ^ SD ) of data in Oxford database

Gender Age (y) Height (m) Weight (kg) BMI BMR (MJ day21)

Males
0 –3 0.4 ^ 0.62 0.65 ^ 0.13 6.3 ^ 3.20 15.1 ^ 2.01 1.47 ^ 0.86
3 –10 6.6 ^ 2.04 1.17 ^ 0.13 21.4 ^ 5.14 15.5 ^ 1.26 4.17 ^ 0.58
10–18 12.7 ^ 2.07 1.49 ^ 0.15 40.0 ^ 12.48 17.7 ^ 2.65 5.51 ^ 1.11
18–30 22.7 ^ 2.87 1.70 ^ 0.09 61.0 ^ 11.40 20.9 ^ 2.84 6.36 ^ 1.00
30–60 40.8 ^ 8.72 1.69 ^ 0.09 65.3 ^ 12.98 22.8 ^ 3.24 6.35 ^ 1.03
60 þ 70.9 ^ 7.60 1.70 ^ 0.09 71.3 ^ 14.94 24.6 ^ 4.13 6.17 ^ 1.09
Females
0 –3 0.5 ^ 0.71 0.65 ^ 0.14 6.7 ^ 3.40 15.0 ^ 2.31 1.54 ^ 0.87
3 –10 7.1 ^ 1.77 1.22 ^ 0.14 23.6 ^ 6.14 15.7 ^ 1.59 4.10 ^ 0.63
10–18 13.0 ^ 2.35 1.50 ^ 0.11 43.4 ^ 12.91 18.8 ^ 3.64 5.20 ^ 0.80
18–30 22.4 ^ 3.01 1.60 ^ 0.08 53.2 ^ 10.04 20.7 ^ 3.18 5.24 ^ 0.79
30–60 41.6 ^ 8.18 1.59 ^ 0.08 59.1 ^ 13.65 23.3 ^ 4.48 5.31 ^ 0.80
60 þ 69.8 ^ 6.88 1.56 ^ 0.09 60.0 ^ 14.52 24.3 ^ 4.78 4.93 ^ 0.78

Abbreviations: BMI – body mass index; BMR – basal metabolic rate.

(254 kcal) per day with a reduction in cereal needs of 18–30 and 30 –60 year old males and in all females over 18
approximately 65 g. As a final comparison, the energy years of age. In the examples cited above (where the
requirements of a rural woman in a developing country, newly calculated BMR was used to estimated energy
weighing 50 kg are presented. requirements), the Oxford BMR equation produced a
In summary, the Oxford equations produced lower significant reduction in total energy and cereal require-
BMR values than the FAO/WHO/UNU equations in the ments per day. Possible explanations for these differences
in BMR when applying the Oxford equations may be
Table 17 Prediction of BMR from weight and height or weight because the Oxford database (1) did not include any of the
alone (mean ^ SD ) elevated BMR values of the Italian subjects and (2)
BMR (MJ day21)
included a much larger number of people from the
tropical region (see Table 29).
%
Gender Age (y) Weight alone Weight þ height difference P

Males 0 –3 1.474 ^ 0.86 1.564 ^ 0.82 26.12 0.169 Discussion and areas for future research
3 –10 4.168 ^ 0.58 4.168 ^ 0.49 20.02 0.465
10–18 5.506 ^ 1.11 5.505 ^ 0.96 þ 0.01 0.767 One practical use of BMR is in the estimation of energy
18–30 6.364 ^ 1.00 6.366 ^ 0.77 20.02 0.364
30–60 6.347 ^ 1.03 6.349 ^ 0.78 20.01 0.808 requirements for population groups and subsequently
60þ 6.173 ^ 1.09 6.178 ^ 0.86 20.08 0.615 their food needs. The FAO/WHO/UNU report Energy
Females 0 –3 1.544 ^ 0.87 1.598 ^ 0.84 23.56 0.245 and Protein Requirements1 made clear for the first time
3 –10 4.100 ^ 0.63 4.096 ^ 0.52 0.12 0.617
10–18 5.202 ^ 0.80 5.199 ^ 0.60 0.05 0.995 two main purposes of determining energy requirements.
18–30 5.239 ^ 0.79 5.232 ^ 0.57 þ 0.14 0.800 The first was for prescriptive purposes, i.e. for making
30–60 5.306 ^ 0.80 5.307 ^ 0.57 20.02 0.658 recommendations about the level of consumption that
60þ 4.931 ^ 0.78 4.934 ^ 0.64 20.05 0.640
ought to be maintained in a population; the second, for
Abbreviation: BMR – basal metabolic rate. diagnostic purposes, i.e. the assessment of the
Basal metabolic rate studies in humans 1147
Table 18 Comparison of Oxford and FAO/WHO/UNU BMR Table 19 Comparison of Oxford and FAO/WHO/UNU BMR
equations in males at various ages (MJ day21) equations in females at various ages (MJ day21)

Difference Difference from

Age Weight (kg) BMR (Ox) BMR (FAO) from FAO (%)a Age Weight (kg) BMR (Ox) BMR (FAO) FAO (%)a

0 –3 3 0.624 0.5390 215.77 0– 3 3 0.6415 0.5510 216.42

5 1.134 1.0490 28.10 5 1.1335 1.0610 26.83
10 2.409 2.3240 23.66 10 2.3635 2.3360 21.18
12 2.919 2.8340 23.00 12 2.8555 2.8460 20.33
15 3.684 3.5990 22.36 15 3.5935 3.6110 þ 0.48
18 4.449 4.3640 21.95 18 4.3315 4.3760 þ 1.02
20 4.959 4.8740 21.74 20 4.8235 4.8860 þ 1.28
3 –10 12 3.2744 3.2088 22.04 3– 10 12 3.1304 3.2192 þ 2.76
15 3.5555 3.4935 21.77 15 3.383 3.5015 þ 3.38
18 3.8366 3.7782 21.55 18 3.6356 3.7838 þ 3.92
22 4.2114 4.1578 21.29 22 3.9724 4.1602 þ 4.51
25 4.4925 4.4425 21.13 25 4.225 4.4425 þ 4.90
30 4.961 4.9170 20.89 30 4.646 4.9130 þ 5.43
35 5.4295 5.3915 20.70 35 5.067 5.3835 þ 5.88
45 6.3665 6.3405 20.41 45 5.909 6.3245 þ 6.57
50 6.835 6.8150 20.29 50 6.33 6.7950 þ 6.84
10– 18 25 4.3525 4.5500 þ 4.34 10–18 25 4.3425 4.3950 þ 1.19
30 4.737 4.9160 þ 3.64 30 4.575 4.6500 þ 1.61
35 5.1215 5.2820 þ 3.04 35 4.8075 4.9050 þ 1.99
45 5.8905 6.0140 þ 2.05 45 5.2725 5.4150 þ 2.63
55 6.6595 6.7460 þ 1.28 55 5.7375 5.9250 þ 3.16
65 7.4285 7.4780 þ 0.66 65 6.2025 6.4350 þ 3.61
75 8.1975 8.2100 þ 0.15 75 6.6675 6.9450 þ 4.00
95 9.7355 9.6740 20.64 95 7.5975 7.9650 þ 4.61
105 10.5045 10.4060 20.95 105 8.0625 8.4750 þ 4.87
18– 30 50 5.6250 6.0400 þ 6.87 18–30 50 5.060 5.1550 þ 1.84
55 5.9595 6.3600 þ 6.30 55 5.333 5.4625 þ 2.37
60 6.294 6.6800 þ 5.78 60 5.606 5.7700 þ 2.84
65 6.6285 7.0000 þ 5.31 65 5.879 6.0775 þ 3.27
75 7.2975 7.6400 þ 4.48 75 6.425 6.6925 þ 4.00
85 7.9665 8.2800 þ 3.79 85 6.971 7.3075 þ 4.60
95 8.6355 8.9200 þ 3.19 95 7.517 7.9225 þ 5.12
105 9.3045 9.5600 þ 2.67 105 8.063 8.5375 þ 5.56
30– 60 50 5.440 6.0950 þ 10.75 30–60 50 4.935 5.2900 þ 6.71
55 5.736 6.3375 þ 9.49 55 5.1385 5.4720 þ 6.09
60 6.032 6.5800 þ 8.33 60 5.342 5.6540 þ 5.52
65 6.328 6.8225 þ 7.25 65 5.5455 5.8360 þ 4.98
75 6.92 7.3075 þ 5.30 75 5.9525 6.2000 þ 3.99
85 7.512 7.7925 þ 3.60 85 6.3595 6.5640 þ 3.12
95 8.104 8.2775 þ 2.10 95 6.7665 6.9280 þ 2.33
105 8.696 8.7625 þ 0.76 105 7.1735 7.2920 þ 1.63
60 þ 55 5.2465 5.1475 21.92 60 þ 55 4.712 4.9045 þ 3.92
60 5.528 5.4300 21.80 60 4.924 5.1240 þ 3.90
65 5.8095 5.7125 21.70 65 5.136 5.3435 þ 3.88
75 6.3725 6.2775 21.51 75 5.56 5.7825 þ 3.85
85 6.9355 6.8425 21.36 85 5.984 6.2215 þ 3.82
95 7.4985 7.4075 21.23 95 6.408 6.6605 þ 3.79
100 7.78 7.6900 21.17 100 6.62 6.8800 þ 3.78

Abbreviations: FAO/WHO/UNU – Food and Agriculture Organization/World Abbreviations: FAO/WHO/UNU – Food and Agriculture Organization/World
Health Organization/United Nations University; BMR – basal metabolic Health Organization/United Nations University; BMR – basal metabolic
rate. rate.
a a
þ indicates that FAOs formulae give higher values, and 2 indicates þ indicates that FAOs formulae give higher values, and 2 indicates
lower values. lower values.

adequacy or otherwise of the food needs in a The Schofield database comprised 114 published
population. In the factorial estimation of total energy studies of BMR, totalling 7173 data points. These formed
expenditure (FAO/WHO/UNU)1, a major feature and the basis for the equations used in the FAO/WHO/UNU
component was the estimation of BMR. The measure- document Energy and Protein Requirements1. While
ment and prediction of BMR thus took on a greater Schofield’s analysis has served a significant role in
significance. It is important to recognise that the re-establishing the importance of using BMR to predict
primary purpose of the early measures of BMR human energy requirements, recent workers have
(1900 –1950) was to diagnose hypo- or hyperthyroidism, subsequently queried the universal validity and appli-
not to estimate energy requirements. cation of these equations. A survey of the most recent
1148 CJK Henry
Table 20 Summary of difference between the Oxford and FAO equations

Sex and age band Oxford equation Schofield (1985) equation

(male/female, years) (MJ day21) (MJ day21) Difference

Males, 0– 3 0.225W 2 0.141 0.255W 2 0.226 Large difference at low body weights
Females, 0–3 0.246W 2 0.0965 0.255W 2 0.214 Large difference at low body weights
Males, 3– 10 0.0937W þ 2.15 0.0949W þ 2.07 No significant difference
Females, 3–10 0.0842W þ 2.12 0.0941W þ 2.09 Difference at high body weights
Males, 10–18 0.0769W þ 2.43 0.0732W þ 2.72 Difference at low body weights
Females, 10– 18 0.0465W þ 3.18 0.0510W þ 3.12 Difference at high body weights
Males, 18–30 0.0669W þ 2.28 0.0640W þ 284 Large difference at body weight , 75 kg
Females, 18– 30 0.0546W þ 2.33 0.0615W þ 2.08 Large difference at body weight , 75 kg
Males, 30–60 0.0592W þ 2.48 0.0485W þ 3.67 Very large difference at body weight , 75 kg
Females, 30– 60 0.0407W þ 2.90 0.0364W þ 3.47 Large difference at body weight , 65 kg
Males, 60 þ 0.0563W þ 2.15 0.0565W þ 2.04 No significant difference
Females, 60 þ 0.0424W þ 2.38 0.0439W þ 2.49 Difference across body weights

Abbreviation: FAO – Food and Agriculture Organization.

studies (1980–2000) in BMR suggests that in most cases larger number (4018) of people from the tropics. In general,
the current FAO/WHO/UNU predictive equations over- the Oxford equations tend to produce lower BMR values
estimate BMR in many communities. than the current FAO/WHO/UNU equations in 18– 30 and
It is concluded that the over-representation of BMR 30–60 year old males and in all females over 18 years of age
values obtained from Italian subjects – 3388 out of 7173 – (see pages 33–42 for detailed discussion). The objective of
(who had a higher BMR kg21) in the Schofield1 database all consultations is to scientifically progress and identify
may have resulted in the FAO/WHO/UNU predictive fresh ideas and issues. This is an opportune moment to re-
equations to overestimate BMR in contemporary popu- examine the role and place of BMR measurements in
lations. estimating total energy requirements today. The Oxford
A series of new equations (Oxford equations) have been equations use and future application will surely depend on
developed using a data set of 10 552 BMR values that (1) their ability to predict more accurately the BMR in
excluded all the Italian subjects and (2) included a much contemporary populations.

Table 21 Differences in total EE for light, moderate and Table 22 Differences in total EE for light, moderate and
high activity levels in males (18–30 years) using Oxford and high activity levels in females (18– 30 years) using Oxford and
FAO/WHO/UNU equations (MJ day21) FAO/WHO/UNU equations (MJ day21)

BMR BMR EE EE Difference BMR BMR EE EE Difference

Weight (Ox) (FAO) (Ox) (FAO) from FAO (kJ) Weight (Ox) (FAO) (Ox) (FAO) from FAO (kJ)

Light Activity (1.55) Light activity (1.56)

55 5.9595 6.3600 9.237225 9.858 2620.77 55 5.3330 5.4625 8.32 8.5215 2202.02
60 6.2940 6.6800 9.7557 10.354 2598.30 60 5.6060 5.7700 8.75 9.0012 2255.84
65 6.6285 7.0000 10.27418 10.85 2575.82 65 5.8790 6.0775 9.17 9.4809 2309.66
75 7.2975 7.6400 11.31113 11.842 2530.88 75 6.4250 6.6925 10.02 10.4403 2417.3
85 7.9665 8.2800 12.34808 12.834 2485.93 85 6.9710 7.3075 10.87 11.3997 2524.94
95 8.6355 8.9200 13.38503 13.826 2440.98 95 7.5170 7.9225 11.73 12.3591 2632.58
105 9.3045 9.5600 14.42198 14.818 2396.02 105 8.0630 8.5375 12.58 13.3185 2740.22
Moderate Activity (1.76) Moderate activity (1.64)
55 5.9595 6.3600 10.48872 11.1936 2704.88 55 5.3330 5.4625 8.75 8.9585 2212.38
60 6.2940 6.6800 11.07744 11.7568 2679.36 60 5.6060 5.7700 9.19 9.4628 2268.96
65 6.6285 7.0000 11.66616 12.32 2653.84 65 5.8790 6.0775 9.64 9.9671 2325.54
75 7.2975 7.6400 12.8436 13.4464 2602.80 75 6.4250 6.6925 10.54 10.9757 2438.7
85 7.9665 8.2800 14.02104 14.5728 2551.76 85 6.9710 7.3075 11.43 11.9843 2551.86
95 8.6355 8.9200 15.19848 15.6992 2500.72 95 7.5170 7.9225 12.33 12.9929 2665.02
105 9.3045 9.5600 16.37592 16.8256 2449.68 105 8.0630 8.5375 13.22 14.0015 2778.18
High Activity (2.10) High activity (1.82)
55 5.9595 6.3600 12.51495 13.356 2841.05 55 5.3330 5.4625 9.71 9.94175 2235.69
60 6.2940 6.6800 13.2174 14.028 2810.60 60 5.6060 5.7700 10.20 10.5014 2298.48
65 6.6285 7.0000 13.91985 14.7 2780.15 65 5.8790 6.0775 10.70 11.06105 2361.27
75 7.2975 7.6400 15.32475 16.044 2719.25 75 6.4250 6.6925 11.69 12.18035 2486.85
85 7.9665 8.2800 16.72965 17.388 2658.35 85 6.9710 7.3075 12.69 13.29965 2612.43
95 8.6355 8.9200 18.13455 18.732 2597.45 95 7.5170 7.9225 13.68 14.41895 2738.01
105 9.3045 9.5600 19.53945 20.076 2536.55 105 8.0630 8.5375 14.67 15.53825 2863.59

Abbreviations: EE – energy expenditure; FAO/WHO/UNU – Food and Abbreviations: EE – energy expenditure; FAO/WHO/UNU – Food and
Agriculture Organization/World Health Organization/United Nations Univer- Agriculture Organization/World Health Organization/United Nations Univer-
sity; BMR – basal metabolic rate. sity; BMR – basal metabolic rate.
Basal metabolic rate studies in humans 1149
Table 23 Energy requirement of a subsistence farmer (moderate activity work) using FAO/WHO/UNU
equations (age: 25 years, weight: 58 kg, height: 1.61 m, BMI: 22.4)

Hours kcalth kJ

In bed at 1.0 £ BMR 8 520 2170

Occupational activities at 2.7 £ BMR 7 1230 5150
Discretionary activities:
– Socially desirable and household tasks at 3.0 £ BMR 2 390 1630
– Cardiovascular and muscular maintenance –not needed if moderately active –
For residual time, energy needs at 1.4 £ BMR 7 640 2680
Total ¼ 1.78 £ BMR 2780 11630

Abbreviations: FAO/WHO/UNU – Food and Agriculture Organization/World Health Organization/United Nations Univer-
sity; BMI – body mass index; BMR – basal metabolic rate.
Source: FAO/WHO/UNU1.
Estimated BMR: 65 kcalth (273 kJ)/h.

Table 24 Energy requirement of a subsistence farmer (moderate activity work) using Oxford equations
(age: 25 years, weight: 58 kg, height: 1.61 m, BMI: 22.4)

Hours kcalth kJ

In bed at 1.0 £ BMR 8 491 2052

Occupational activities at 2.7 £ BMR 7 1160 4849
Discretionary activities:
– Socially desirable and household tasks at 3.0 £ BMR 2 368 1539
– Cardiovascular and muscular maintenance– not needed if moderately active –
For residual time, energy needs at 1.4 £ BMR 7 601 2514
Total ¼ 1.78 £ BMR 2621 10 954

Abbreviations: BMI – body mass index; BMR – basal metabolic rate.

Reduction in energy requirements per day ¼ 676 kJ (162 kcal).
Reduction in cereal requirements per day ¼ 41 g (assuming energy value of 16.3 kJ g21 – raw rice (McCance and Wid-
dowson94))
Reduction in cereal requirements per year ¼ 15.0 kg.
Estimated BMR: 61.5 kcalth (257 kJ) per hour.

Areas for further research

Table 25 Energy requirement for a male engaged in heavy work
using the Oxford equations (age: 35 years, weight: 65 kg, height:
1.72 m, BMI: 22)

Hours kcalth kJ

In bed at 1.0 £ BMR 8 545 2280 1. It is recommended that a more detailed analysis of BMR
Occupational activities at 3.8 £ BMR 8 2070 8660 in children aged between 10 and 18 years and from
Discretionary activities at 3.0 £ BMR 1 205 860 different communities be undertaken. A break down of
For residual time, maintenance energy needs 7 670 2800
at 1.4 £ BMR the age band into more physiologically acceptable
Total ¼ 2.14 £ BMR 3490 14 580 ranges, e.g. 10 – 12, 12 – 15 and 15 – 18 years, is
Abbreviations: BMI – body mass index; BMR – basal metabolic rate.
recommended.
Source: FAO/WHO/UNU1. 2. There is an urgent need to develop age and gender
Estimated BMI: 68 kcalth (284 kJ) per hour.
specific BMR equations taking into account the
stages in pubertal development (Tanner rating).
Table 26 Energy requirement for a male engaged in heavy work Because of the rapid changes during puberty
using the Oxford equations (age 35 years, weight 65 kg, height (changes in body composition, hormone levels,
1.72 m, BMI 22) growth), even small age differences may cause large
Hours kcalth kJ changes in metabolic rate.
3. There is a glaring absence of BMR data from mainland
In bed at 1.0 £ BMR 8 505 2112
Occupational activities at 3.8 £ BMR 8 1920 8025 China and Africa. It is recommended that BMR values
Discretionary activities at 3.0 £ BMR 1 189 792 are collected from China and other developing
For residual time, maintenance energy needs 7 619 2587 countries, especially from young children and the
at 1.4 £ BMR
Total ¼ 2.14 £ BMR 3233 13 516 elderly.
4. While BMR data collection in the elderly living in
Abbreviations: BMI – body mass index; BMR – basal metabolic rate.
Reduction in energy requirements per day ¼ 1064 kJ (254 kcal).
developing countries should be encouraged, the
Reduction in cereal requirements per day ¼ 65 g (assuming energy value present age band for the elderly should be further
of 16.3 kJ g21 – raw rice (McCance and Widdowson94).
Reduction in cereal requirements per year ¼ 23.7 kg.
refined to the following groups: 60–75, 76 –85 and
Estimated BMR: 63 kcalth (264 kJ) per hour. .85 years.
1150 CJK Henry
Table 27 Energy requirement of a rural woman in a developing Helen Lightowler are also gratefully acknowledged.
country using FAO/WHO/UNU equations (age: 35 years, weight: Special thanks to Drs. Butte, Chen, Goldberg, Jones,
50 kg, height: 1.6 m, BMI: 19.5)
Muhilal, Poehlman, Pullicino, Reilly, Soares, Shetty,
Hours kcalth KJ Westerterp and Wong for sending their raw data on BMR
In bed at 1.0 £ BMR 8 425 1780 so readily.
Occupational activities:
– Housework, preparing food, etc, 3 430 1800
at 2.7 £ BMR References
– Working in fields, at 2.8 £ BMR 4 595 2490
Discretionary activities at 2.5 £ BMR 2 265 1110 1 FAO/WHO/UNU. Energy and Protein Requirements. Report
For residual time, energy needs 7 520 2180 of a Joint FAO/WHO/UNU Expert Consultation. Technical
at 1.4 £ BMR Report Series No. 724. Geneva: World Health Organization,
Total ¼ 1.76 £ BMR 2235 9360 1985.
2 Mitchell HH. Comparative Nutrition of Man and Domestic
Abbreviations: FAO/WHO/UNU – Food and Agriculture Organization/World
Health Organization/United Nations University; BMI – body mass index; Animals. Vol. I. New York: Academic Press, 1964.
BMR – basal metabolic rate. 3 Aub JC, DuBois EF. Clinical calorimetry: the basal
Source: FAO/WHO/UNU1. metabolism of dwarfs and legless men with observations
Estimated BMR: 53 kcalth (220 kJ) per hour. on the specific dynamic action of protein. Archives of
Internal Medicine 1917; 19: 840–64.
4 Aub JC, DuBois EF. The basal metabolism of old men.
Table 28 Energy requirement of a rural woman in a developing Archives of Internal Medicine 1917; 19: 823– 31.
country using Oxford equations (age: 35 years, weight: 50 kg, 5 Krogh A. Determination of standard (basal) metabolism of
height: 1.6 m, BMI: 19.5) patients by recording apparatus. Boston Medical Surgery
Hours kcalth KJ Journal 1923; 189: 313–7.
6 Blaxter K. Energy Metabolism in Animals and Man.
In bed at 1.0 £ BMR 8 392 1640 Cambridge: Cambridge University Press, 1989.
Occupational activities: 7 DuBois EF, Ebaugh FG, Hardy JD. Basal heat production and
– Housework, preparing food, etc, 3 397 1660 elimination of thirteen normal women at temperatures from
at 2.7 £ BMR 22 –358C. Journal of Nutrition 1952; 48: 257– 93.
– Working in fields, at 2.8 £ BMR 4 549 2296 8 Wilkerson JE, Raven PB, Horvath SM. Critical temperature of
Discretionary activities at 2.5 £ BMR 2 245 1025 unacclimatized male Caucasians. Journal of Applied Physi-
For residual time, energy needs 7 480 2009 ology 1972; 33: 451– 5.
at 1.4 £ BMR 9 Pi-Suner J. Studies in racial metabolism: basal metabolism of
the Araucanian Mapuches. American Journal of Physiology
Total ¼ 1.76 £ BMR 2063 8630 1933; 105: 383–8.
Abbreviations: BMI – body mass index; BMR – basal metabolic rate.
10 Roth P, Buckingham PE. The criteria of a dependable basal
Reduction in energy requirements per day ¼ 730 kJ (175 kcal). metabolism report. American Journal of Clinical Pathology
Reduction in cereal requirements per day ¼ 45 g (assuming energy value 1939; 9: 79– 92.
of 16.3 kJ g21 – raw rice (McCance and Widdowson94) 11 Harris JA, Benedict FG. A Biometric Study of Basal
Reduction in cereal requirements per year ¼ 16.4 kg Metabolism in Man. Publication No. 279. Washington:
Estimated BMR: 49 kcalth (206 kJ) per hour.
Carnegie Institute of Washington, 1919.
12 Bedale EM. Energy expenditure and food requirements of
children at school. Proceedings of the Royal Society of
London 1923; 94: 368– 404.
Table 29 Comparison between Oxford database and Schofield 13 FAO. Calorie Requirements. Report of the Second Commit-
database tee on Calorie Requirements. Nutritional Studies No. 15.
Rome: Food and Agriculture Organization, 1957.
Number Number of
14 Roth PR. Portable respiration apparatus for clinical use.
of papers data points %
Boston Medical Surgery Journal 1922; 186: 457.
Oxford database 166 10 552 15 Boothby WM, Berkson J, Dunn HL. Studies of the energy
Common to Schofield and 77 4039 metabolism of normal individuals: a standard for basal
Oxford database metabolism, with a nomogram for clinical application.
New in Oxford database 89 6513 American Journal of Physiology 1936; 116: 468– 84.
Tropical subjects in 937 16 Durnin JVGA. Basal Metabolic Rate in Man. Working paper
Schofield database submitted to the Joint FAO/WHO/UNU Expert Consultation
Percentage of tropical subjects 13 on Energy and Protein Requirements. Rome: Food and
in Schofield database Agriculture Organization, 1981
Tropical subjects in Oxford 4018 17 Krogh A. Respiratory Exchange of Animals and Man.
database New York: Longmans, 1916.
Percentage of tropical subjects 38 18 Benedict FG. Portable respiration apparatus for clinical use.
in Oxford database
Boston Medical Surgery Journal 1918; 178: 667.
19 Lewis RC, Iliff A, Duval AM. Further consideration of the
effect of altitude on basal metabolism. The Journal of
Acknowledgements Nutrition 1943; 26: 175– 85.
20 Lewis RC, Iliff A, Duval AM. The comparative accuracy of
the closed circuit bedside method and the open circuit
The author wishes to thank IDECG Switzerland, Nestle
chamber procedure for the determination of basal metab-
Foundation and FAO Rome, for supporting this analysis. olism. Journal of Laboratory Clinical Medicine 1943; 28:
The efforts of David Martineau, Darren Massey and 1238 – 45.
Basal metabolic rate studies in humans 1151
21 Lewis RC, Duval AM, Iliff A. Standards for the basal 40 Piers LS, Shetty PS. Basal metabolic rates of Indian women.
metabolism of children from 2 to 15 years of age, inclusive. European Journal of Clinical Nutrition 1993; 47: 586–91.
Journal of Pediatrics 1943; 23: 1– 18. 41 Hayter JE, Henry CJK. A re-examination of basal metabolic
22 Willard HN, Wolf GA Jr. A source of error in the rate predictive equations: the importance of geographic
determination of basal metabolism by the closed-circuit origin of subjects in sample selection. European Journal of
technique. Annual International Medicine 1951; 34: Clinical Nutrition 1994; 48: 702– 7.
148– 62. 42 Valencia ME, Moya SY, McNeill G, Haggarty P. Basal
23 Segal KR. Comparison of indirect calorimetric measurement metabolic rate and body fatness of adult men in northern
of resting energy expenditure with a ventilated hood, face Mexico. European Journal Clinical Nutrition 1994; 48:
mask, and mouthpiece. American Journal of Clinical 205–11.
Nutrition 1987; 45: 1420 –3. 43 Wong WW, Butte NF, Hergenroeder AC, Hill RB, Stuff JE,
24 Shetty PS, Soares MJ, Sheela ML. Basal Metabolic Rates of Smith EO. Are basal metabolic rate prediction equations
South Indian Males. Bangalore: Food and Agriculture appropriate for female children and adolescents? Journal of
Organization, 1986. Applied Physiology 1996; 81: 2407 – 14.
25 Soares MJ, Sheela ML, Kurpad AV, Kulkarni RN, Shetty PS. 44 Soares MJ, Francis DG, Shetty PS. Predictive equations for
The influence of different methods on basal metabolic rate basal metabolic rates of Indian males. European Journal of
measurements in human subjects. American Journal of Clinical Nutrition 1993; 47: 389– 94.
Clinical Nutrition 1989; 50: 731 –6. 45 Cruz CM, daSilva AF, dos Anjos LA. A taxa metabica basal e
26 Consolazio CF, Johnson RE, Pecora LJ. Physiological superestimade pelas equa co es preditivas em universitaras
Measurements of Metabolic Function in Man. New York: do Rio de Janeiro, Brasil. Archivio Latino americano
McGraw-Hill, 1963. Nutricion 1999; 49: 232–7.
27 Clark HD, Hoffer LF. Reappraisal of the resting metabolic 46 Dietz WH, Bandini LG, Schoeller DA. Estimates of metabolic
rate of normal young men. American Journal of Clinical rate in obese and non-obese adolescents. Journal of
Nutrition 1991; 53: 21– 6. Pediatrics 1991; 118: 140–9.
28 Fowler WS, Blackburn CM, Helmholz HF Jr. Determination 47 Cunningham JJ. A reanalysis of the factors influencing basal
of basal rate of oxygen consumption by open and closed- metabolic rate in normal adults. American Journal of
circuit methods. Journal of Clinical Endocrinological Clinical Nutrition 1980; 33: 2372 –3374.
Metabolism 1957; 17: 786 – 96. 48 Cunningham JJ. Body composition as a determinant of
29 Schofield WN, Schofield C, James WPT. Basal metabolic rate energy expenditure: a synthetic review and a proposed
– review and prediction, together with an annotated general prediction equation. American Journal of Clinical
bibliography of source material. Human Nutrition Clinical Nutrition 1991; 54: 963– 9.
Nutrition 1985; 39C: 5 –96. 49 Hayter J, Henry CJK. Basal metabolic rate in human subjects
30 DuBois D, DuBois EF. The measurements of the surface area migrating between tropical and temperate regions – a
of man. Archives of Internal Medicine 1915; 15: 868– 75. longitudinal study and review of previous work. European
31 Daly JM, Heymsfield SB, Head CA, Harvey LP, Nixon DW, Journal of Clinical Nutrition 1993; 47: 724– 34.
Katzeff H, Grossman GD. Human energy requirements: over 50 Ho Z-C, Zi HM, Bo L, Ping H. Energy expenditure of pre
estimation by widely used prediction equation. The school children in a subtropical area. World Reviews in
American Journal of Clinical Nutrition 1985; 42: 1170 – 4. Nutrition Dietetics 1988; 57: 75– 94.
32 Alpers PH, Stevenson WF, Bier DM. Manual of Nutritional 51 Min Q, Ho Z-C. The basal metabolism rate of adolescent girls
Therapeutics. Philadelphia: Lippincott Williams and Wilkins, in the sub-tropical areas of China. Acta Nutrimenta Sinica
1995. 1991; 13: 252– 8.
33 Boothby WM, Sandiford I. A comparison of the DuBois and 52 Spurr GB, Reina JC, Hoffmann RG. Basal metabolic rate of
the Harris and Benedict normal standards for the estimation Colombian children 2 – 16 years of age: ethnicity and
of basal metabolic rate. Journal of Biological Chemistry nutritional status. American Journal of Clinical Nutrition
1922; 54: 767– 81. 1992; 56: 623– 9.
34 Boothby WM, Sandiford I. Summary of the basal metabolic 53 Torun B, Davies PSW, Livingstone MBE, Paolisso M, Sackett
data on 8614 subjects with special reference to the normal R, Spurr GB. Energy requirements and dietary energy
standards for the estimation of the basal metabolic rate. recommendations for children and adolescents 1 to 18 years
Journal of Biological Chemistry 1922; 54: 783 – 803. old. European Journal of Clinical Nutrition 1996; 50(Suppl.
35 Quenouille MH, Boyne AW, Fisher WB, Leitch I. Statistical 1): S37 – S81.
Studies of Recorded Energy Expenditure of Man. Basal 54 Livingstone MBE, Coward WA, Prentice AM, Davies PSW,
Metabolism Related to Sex, Stature, Age, Climate, and Race. Strain JJ, McKenna PG, Mahoney CA, White JA, Stewart CM,
Commonwealth Bureau of Animal Nutrition. Technical Kerr MJJ. Daily energy expenditure in free living children:
Communication No. 17. Aberdeen: Commonwealth Agri- comparison of heart-rate monitoring with the doubly
cultural Bureau, 1951. labelled water (2H18 2 O) method. American Journal of
36 de Almeida AO. L’émission de chaleur. Le métabolisme basal Clinical Nutrition 1992; 56: 343– 52.
et le métabolisme mimimum de l‘homme noir tropical. 55 Bandini LG, Schoeller DA, Dietz WH Jr. Energy expenditure
Journal de Physiologie et Pathologie Géneral 1921; 18: in obese and non-obese adolescents. Pediatric Research
958– 64. 1990; 27: 198– 203.
37 Henry CJK, Rees DG. New predictive equations for the 56 Case KO, Brahler CJ, Heiss C. Resting energy expenditures in
estimation of basal metabolic rate in tropical peoples. Asian women measured by indirect calorimetry are lower
European Journal of Clinical Nutrition 1991; 45: 177– 85. than expenditures calculated from prediction equations.
38 de Boer JO, van Es AJH, Voorrips LE, Blockstra F, Vogt JE. Journal of American Dietetic Association 1997; 97: 1288 – 92.
Energy metabolism and requirements in different ethnic 57 Leung R, Woo J, Chan D, Tang N. Validation of prediction
groups. European Journal Clinical Nutrition 1988; 42: equations for basal metabolic rate in Chinese subjects.
983– 97. European Journal of Clinical Nutrition 2000; 54: 551–4.
39 Arciero PJ, Goran MI, Gardner AM, Ades PA, Tyzbir RS, 58 Piers LS, Diffey B, Soares MJ, Frandsen SL, McCormack LM,
Poehlman ET. A practical equation to predict resting Lutschini MJ, O’Dea K. The validity of predicting the basal
metabolic rate in older females. Journal of American metabolic rate of young Australian men and women.
Geriatric Society 1993; 41: 389 – 95. European Journal of Clinical Nutrition 1997; 51: 333–7.
1152 CJK Henry
59 Ferro-Luzzi A, Petracchi C, Kuriyan R, Kurpad AV. Basal 77 Lewis WH. Changes with age in the basal metabolic rate in
metabolism of weight-stable chronically undernourished adult men. American Journal of Physiology 1934; 12:
men and women: lack of metabolic adaptation and ethnic 502– 17.
differences. American Journal of Clinical Nutrition 1997; 78 Mason ED, Benedict FG. The basal metabolism of South
66: 1086 –93. Indian Women. Independent Journal of Medical Research
60 Yamauchi T, Ohtsuka R. Basal metabolic rate and energy 1931; 19: 75– 98.
costs at rest and during exercise in rural- and urban-dwelling 79 Mason ED, Benedict FG. The effect of sleep on human basal
Papua New Guinea Highlanders. European Journal of metabolism, with particular reference to South Indian
Clinical Nutrition 2000; 54: 494 – 9. women. American Journal of Physiology 1934; 108: 377–83.
61 Norgan NG. Population differences in body composition in 80 Mason ED, Jacob M. Variations in basal metabolic rate:
relation to the Body Mass Index. European Journal of responses to changes between tropical and temperate
Clinical Nutrition 1994; 48(Suppl. 3): S10– S27. climates. Human Biology 1972; 44: 141– 72.
62 Shetty PS, Henry CJK, Black AE, Prentice AM. Energy 81 Mason ED. The basal metabolism of European women in
requirements of adults: an update on basal metabolic rates South India and the effect of change of climate on European
(BMRs) and physical activity levels (PALs). European Journal and South Indian Women. Journal of Nutrition 1934; 8:
of Clinical Nutrition 1996; 50(Suppl. 1): S11 – S23. 695– 713.
63 Hayter J, Henry CJK. Basal metabolic rate in human subjects 82 Nakagawa I. Growth and basal metabolism. Basal metab-
migrating between tropical and temperate regions – a olism of preschool children. American Journal of Diseases in
longitudinal study and review of previous work. European Childhood 1934; 47: 963– 9.
Journal of Clinical Nutrition 1993; 47: 724 – 34. 83 Nakagawa I. Growth and basal metabolism. Basal metab-
64 Poehlman ET, Melby CL, Badylak SF. Resting metabolic rate olism of elementary school children. American Journal of
and post-prandial thermogenesis in highly trained and Diseases in Childhood 1934; 48: 34– 8.
untrained males. American Journal of Clinical Nutrition 84 Nakagawa I. Growth and basal metabolism. Changes in the
1988; 47: 793– 8. basal metabolism of children during a year. American
65 Robertson JD, Reid DD. Standards for the basal metabolism Journal of Diseases in Childhood 1935; 49: 1232– 9.
of normal people in Britain. The Lancet 1952; 1: 940– 9. 85 Nakagawa I. Growth and basal metabolism. Basal metab-
olism of high school children. American Journal of Diseases
66 Benedict FG, Garven HSD. The basal metabolism of male
in Childhood 1937; 51: 985– 9.
Chinese in Manchuria. Chinese Journal of Physiology 1936;
86 Wardlaw HSH, Davies HW, Joseph MR. Energy metabolism
10: 141–8.
and insensible perspiration of Australian Aborigines.
67 Benedict FG, Meyer MH. The basal metabolism of American-
Australian Journal of Experimental Biology 1934; 12: 63 –73.
born Chinese girls. Chinese Journal of Physiology 1933; 7:
87 Wardlaw HSH, Horsley CH. The basal metabolism of some
45– 60. Australian Aborigines. Australian Journal of Experimental
68 Benedict FG, Miles WR, Roth P, Monmouth Smith H. Human Biology 1928; 5: 263– 72.
Vitality and Efficiency Under Prolonged Restricted Diet. 88 Wardlaw HSH, Lawrence WJ. Further observations on the
Washington: Carnegie Institution of Washington, 1919, 280. basal metabolism of Australian Aborigines. Australian
69 Benedict FG, Talbot FB. The Gaseous Metabolism of Infants. Journal of Experimental Biology 1932; 10: 157–65.
Washington: Carnegie Institution of Washington, 1914, 201. 89 Wang CC, Hawks JE. Basal metabolism of twenty-one
70 Benedict FG. Age and basal metabolism of adults. Journal of Chinese children reared or born and reared in the United
Biological Chemistry 1928; 80: 650 – 64. States. American Journal of Diseases in Childhood 1932; 44:
71 Benedict FG. Basal metabolism data on normal men and 69 –80.
women (series II) with some considerations on the use of 90 Wang CC, Kaucher M, Wing M. Metabolism of adolescent
prediction standards. Journal of Biological Chemistry 1928; girls. Basal metabolism and energy exchange. American
8: 607–20. Journal of Diseases in Childhood 1936; 51: 801–15.
72 Benedict FG. Factors affecting basal metabolism. Journal of 91 Wang CC, Kern R, Frank M, Hays BB. Metabolism of
Biological Chemistry 1915; 20: 263 – 99. undernourished children. Basal metabolism. American
73 Benedict FG. Metabolism and Growth from Birth to Puberty. Journal of Diseases in Childhood 1926; 32: 350–9.
Washington: Carnegie Institution of Washington, 1921, 302. 92 Wang CC. Basal metabolism and preformed total creatinine
74 Benedict FG. Old age and basal metabolism. New England in urine of seventy children. American Journal of Diseases in
Journal of Medicine 1935; 212: 1111 –23. Childhood 1939; 57: 838– 50.
75 Benedict FG. The basal metabolism of girls 12 to 17 years of 93 Wang CC. Basal metabolism of American-born Chinese girls
age. Proceedings of the National Academy of Science 1921; and of American girls of the same age. American Journal of
VII: 10– 13. Diseases in Childhood 1934; 48: 1041.
76 Lewis RC, Kinsman GM, Iliff A. The basal metabolism of 94 Holland B, Welch AA, Unwin ID, Buss DH, Paul AA,
normal boys and girls from two to twelve years old, Southgate DAT. McCance and Widdowson’s the Composition
inclusive. American Journal of Diseases in Childhood 1937; of Foods, 5th edn. Cambridge: Royal Society of Chemistry,
53: 348–428. 1991.

View publication stats