Twentyfour hour metabolic rate measurements utilized as a reference to evaluate several prediction equations for calculating energy requirements in healthy infants
 Russell Rising^{1}Email author
DOI: 10.1186/147528911014
© Rising; licensee BioMed Central Ltd. 2011
Received: 27 August 2010
Accepted: 2 February 2011
Published: 2 February 2011
Abstract
Background
To date, only shortduration metabolic rate measurements of less than four hours have been used to evaluate prediction equations for calculating energy requirements in healthy infants. Therefore, the objective of this analysis was to utilize direct 24hour metabolic rate measurements from a prior study to evaluate the accuracy of several currently used prediction equations for calculating energy expenditure (EE) in healthy infants.
Methods
Data from 24hour EE, resting (RMR) and sleeping (SMR) metabolic rates obtained from 10 healthy infants, served as a reference to evaluate 11 lengthweight (LWT) and weight (WT) based prediction equations. Six prediction equations have been previously derived from 50 shortterm EE measurements in the Enhanced Metabolic Testing Activity Chamber (EMTAC) for assessing 24hour EE, (EMTACEELWT and EMTACEEWT), RMR (EMTACRMRLWT and EMTACRMRWT) and SMR (EMTACSMRLWT and EMTACSMRWT). The last five additional prediction equations for calculating RMR consisted of the World Health Organization (WHO), the Schofield (SCHLWT and SCHWT) and the Oxford (OXFORDLWT and OXFORDWT). Paired ttests and the Bland & Altman limit analysis were both applied to evaluate the performance of each equation in comparison to the reference data.
Results
24hour EE, RMR and SMR calculated with the EMTACEEWT, EMTACRMRWT and both the EMTACSMRLWT and EMTACSMRWT prediction equations were similar, p = NS, to that obtained from the reference measurements. However, RMR calculated using the WHO, SCHLWT, SCHWT, OXFORDLWT and OXFORDWT prediction equations were not comparable to the direct 24hour metabolic measurements (p < 0.05) obtained in the 10 reference infants. Moreover, the EMTACEELWT and EMTACRMRLWT were also not similar (p < 0.05) to direct 24hour metabolic measurements.
Conclusions
Weight based prediction equations, derived from shortduration EE measurements in the EMTAC, were accurate for calculating EE, RMR and SMR in healthy infants.
Background
There are many prediction equations currently in use to calculate energy requirements in healthy infants [1–4]. These are popular among many health care practitioners due to their ease of use. For many of these equations only length and weight of the infants need to be measured prior to calculations. Some of the most commonly utilized equations for infants include those from the World Health Organization [1], Schofield [2] and Oxford [3]. However, some of these prediction equations [1, 2] were based on limited data obtained over 80 years ago utilizing nonstandardized techniques [3, 5]. Derivation of all of these prediction equations [1–4] were based on shortterm metabolic measurements. For example, only 3045 minute measurements of resting metabolic rates in individuals were utilized for the derivation of the World Health Organization [1], Schofield [2] and Oxford equations [3]. In an attempt to improve the accuracy of calculating energy requirements for infants, new prediction equations derived and published from our laboratory [4] were based on 50 fourhour morning metabolic measurements in the Enhanced Metabolic Testing Activity Chamber (EMTAC).
The EMTAC was applied in the first ever direct 24hour measurement of energy expenditure, resting and sleeping metabolic rates in both healthy [6] and in those infants recovering from malnutrition [7]. The aim of this analysis was to use data from previously published direct 24hour metabolic measurements in healthy infants [6] as a reference to evaluate 11 prediction equations for calculating 24hour energy expenditure, resting and sleeping metabolic rates [1–4].
Methods
Subjects
Anthropometric, growth performance and metabolic data (Mean ± SD) for the healthy reference infants [6]
Parameter  N = 10 

Males/Females  7/3 
Age (months)  5.0 ± 0.8 
Length (cm)  68.8 ± 2.8 
BMI (kg/m^{2})  15.5 ± 1.5 
Lengthforage percentile  81.5 ± 14.3 
Weightforage percentile  65.0 ± 21.0 
Weightforlength percentile  27.5 ± 23.1 
24h Energy expenditure (kcal/kg/d)  78.7 ± 8.4 
Resting metabolic rate (kcal/kg/d)  66.0 ± 3.5 
Sleeping metabolic rate (kcal/kg/d)  65.0 ± 3.4 
Direct measurement of 24hour metabolic rate
Each of the 10 reference infants spent 24hours in the Enhanced Metabolic Testing Activity Chamber (EMTAC) for measurement of energy expenditure, resting and sleeping metabolic rate as previously published from our laboratory [6]. However, a brief description of the methodology is provided. Prior to each metabolic measurement the EMTAC was calibrated with standard gases with a known concentration of oxygen and carbon dioxide. Furthermore, parents were given instruction on how to interact with their infants and were given time to practice using the hand access ports prior to metabolic testing. Each infant was placed in the EMTAC for 24hours from 9:30 AM till 9:29 AM the following day for continuous measurements of energy expenditure (EE; kcal/min), physical activity (PA; oscillations in weight/min/kg body weight) and the respiratory quotient (RQ:VCO_{2}/VO_{2}). Any supplies such as diapers, formula, baby food or toys were placed in the EMTAC in hanging bags before the start of the test. Parents continued to formula feed their infants at their discretion during metabolic testing.
Energy expenditure (kcal/min) was continuously calculated during metabolic testing according to the method of Jequier [9] and summarized every five minutes as described previously [10].
There were no restrictions in regards to room lighting, feeding or interaction of the infant or with any of the activities of the family during the entire testing procedure. One of the four investigators (RR, MC, DD and SV) acted as observers on rotating eight hour shifts and recorded all infant activities such as infant feedings, periods of observed sleep and amount of parental interaction during the entire 24hour testing period.
At the conclusion of each metabolic test, all metabolic data were corrected for parental interaction, prior to the calculation of resting (RMR; kcal/kg/d) and sleeping metabolic rates (SMR; kcal/kg/d). This involved eliminating any fiveminute EE summary period where parents interacted with their infants. Thereafter, RMR was calculated by regressing EE on PA, multiplying the resulting yintercept by 1440 (minutes in 24hours). Twentyfour hour SMR was calculated by retaining all EE periods between 11:30 PM and 5:30 AM where the index of PA was less than or equal to 1.5 and the infant was observed to be asleep. The mean of these EE periods was multiplied by 1440. This is similar to the methodology used for calculating SMR in adults [11]. All metabolic results were expressed as kcal/kg/d.
Calculations
Current prediction equations for infants 03 years of age used in our analysis
Source  Sex  Prediction equation (kcal/kg/d) 

EMTACEEWT  Males or females  (98.1 × WT )  121.7 
EMTACEELWT  Males or females  (10.7 × L) + (73.3 × WT)  635.1 
EMTACRMRWT  Males or females  (84.5 × WT)  117.3 
EMTACRMRLWT  Males or females  (10.1 × L) + (61.0 × WT)  605.1 
EMTACSMRWT  Males or females  (73.3 × WT)  72.6 
EMTACSMRLWT  Males or females  (7.6 × L) + (55.6 × WT)  440.7 
WHO  Males  (60.9 × WT)  54.0 
Females  (61.0 × WT)  51.0  
SCHWT  Males  (59.5 × WT)  30.3 
Females  (58.3 × WT)  31.1  
SCHLWT  Males  (0.2 × WT) + (1517.4 × (L/100))  617.6 
Females  (16.3 × WT) + (1023.2 × (L/100))  413.5  
OXFORDWT  Males  (61.0 × WT)  33.7 
Females  (58.9 × WT)  23.1  
OXFORDLWT  Males  (28.2 × WT) + (859.0 × (L/100))  371.0 
Females  (30.4 × WT) + (703 × (L/100)) 287.0 
Statistical Analysis
All data were analyzed with SPSS software (v13; Chicago, IL) and expressed as Mean ± Standard Deviation (SD) at the 5% level of probability (p < 0.05). The 24hour data from the reference infants was normally distributed as previously published [6] therefore parametric statistics was utilized for the data analysis. Moreover, the sample size was appropriate for this analysis based on the results from previous short [4, 10, 12] and long term [6, 7] metabolic studies with the EMTAC instrument.
Two different statistical analyses were conducted utilizing the anthropometric and sex data from the 10 healthy reference infants. First, paired ttests were performed to compare the metabolic results calculated using the EMTACEELWT, EMTACEEWT, EMTACRMRLWT, EMTACRMRWT, EMTACSMRLWT, EMTACSMRWT, WHO, SCHLWT, SCHWT, OXFORDLWT and OXFORDWT prediction equations [1–4] to the respective 24hour measured metabolic parameter from the EMTAC [6]. Second, the Bland and Altman limit analysis [13] was also performed to determine agreement between the reference metabolic parameter determined for 24hours in the EMTAC to that calculated with each of the prediction equations being evaluated [13]. This involves taking the average between each reference and respective calculated value ((reference + calculated value/2) for each infant and comparing the mean result across all 10 infants to the mean differences (reference  calculated value) between that particular reference and respective calculated value. Mean differences with a close proximity to zero (no differences between the reference and calculated values) suggests good agreement while mean differences close to two standard deviations from zero (large differences between reference and calculated values) suggest poor agreement [13].
All metabolic data were expressed as kcal/kg/d where necessary by dividing the results from each of the prediction equations by the infant's body weight in kg.
Results
Statistical analysis for prediction equations that are in agreement with direct 24hour metabolic measurements
Prediction  Value  Difference^{1}  P  Agreement^{2} 

equation  (Mean ± SD)  (%)  (Paired Ttest)  (Bland & Altman ± 2SD) 
EMTACEEWT  81.3 ± 1.8  4.5 ± 12.0  0.42  2.6 ± 19.6 
EMTACRMRWT  65.0 ± 3.9  4.0 ± 7.9  0.16  2.4 ± 9.8 
EMTACSMRWT  63.3 ± 1.1  2.2 ± 7.4  0.23  1.7 ± 9.6 
EMTACSMRLWT  66.7 ± 2.4  3.0 ± 8.2  0.29  1.8 ± 10.0 
Statistical analysis for prediction equations that are not in agreement with direct 24hour metabolic measurements
Prediction  Value  Difference^{1}  P  Agreement^{2} 

Equation  (Mean ± SD)  (%)  (Paired Ttest)  (Bland & Altman ± 2SD) 
EMTACEELWT  86.6 ± 3.3  11.2 ± 12.7  <0.01  7.9 ± 19.2 
EMTACRMRLWT  73.3 ± 3.2  11.4 ± 9.3  <0.01  7.3 ± 11.2 
WHO  53.6 ± 0.8  18.6 ± 5.5  <0.01  12.4 ± 8.2 
SCHWT  54.9 ± 0.8  16.6 ± 5.1  <0.01  11.1 ± 7.6 
SCHLWT  57.9 ± 5.6  12.2 ± 9.0  <0.01  8.1 ± 11.8 
OXFORDWT  56.2 ± 0.6  14.7 ± 5.2  <0.01  9.9 ± 7.6 
OXFORDLWT  58.1 ± 3.1  11.9 ± 6.2  <0.01  8.0 ± 8.6 
Discussion
This is the first time were direct 24hour energy expenditure measurements in healthy infants with a standardized methodology [6], was used as a reference to test the accuracy of several previously published prediction equations [1–4] for calculating 24hour energy expenditure, resting and sleeping metabolic rates. In this comparison the weight based prediction equations for calculating 24hour energy expenditure, resting and sleeping metabolic rates, derived from the shortduration metabolic measurements in the EMTAC, agreed with their respective reference values. Moreover, the lengthweight based prediction equation for sleeping metabolic rate, derived from similar metabolic measurements in the EMTAC, also agreed with its respective reference value. However, neither of the lengthweight based prediction equations for calculating resting and sleeping metabolic rate, as derived from shortduration metabolic measurements in the EMTAC, were in agreement with their respective reference values. Finally, the World Health Organization, Schofield or Oxford prediction equations for calculating resting metabolic rate were not in agreement with the respective reference values. Some of the problems encountered in the derivation of these early equations included data obtained from measurements utilizing closed circuit indirect calorimetry [3]. There were many problems associated with the closed circuit technique including the absorption of carbon dioxide not allowing for the calculation of the respiratory quotient [14], hyperventilation due to the subject knowledge of air being recirculated and no direct measurement of oxygen. Furthermore, most of the laboratory technicians did not record whether the subject was post absorptive and/or in a relaxed state prior to resting metabolic rate measurements. Moreover, many of the early measurements of resting metabolic rate were not conducted in a thermoneutral environment where the room temperature was kept between 2227 degrees C [15]. Finally, a lot of the data were obtained in a limited number of ethnic groups. For example, much of the data utilized to derive the Schofield equations included a disproportionately large number Italians who have been found to have a higher resting metabolic rate per kg body weight [16]. As a result, the Schofield equations tended to overestimate resting metabolic rate in many tropical ethnic groups by as much as 25% [5]. The minor ethnic group differences in body composition might also contribute to the World Health Organization [1] and Schofield [2] equations over estimating resting metabolic rate in many ethnic groups today [17].
In a previous study in our laboratory [4] we derived new prediction equations for calculating 24hour energy expenditure, resting and sleeping metabolic rates in healthy infants utilizing the EMTAC instrument. Moreover, all metabolic measurements were conducted under standard conditions [4] at the same time in the morning between 9:00 AM and 1:00 PM. It is possible that variations in energy expenditure over the course of 24hours, as shown by the presence of the metabolic circadian rhythm [6], might contribute to inherent inaccuracies when utilizing the World Health Organization [1], the Schofield [2] and the Oxford [3] prediction equations. Moreover, the metabolic measurement period of less than one hour might have also contributed to the inherent inaccuracies in these equations [1–3]. Despite using only the first four hours of metabolic data, the fact that measurements were conducted at the same time of day and were run at least three additional hours, as compared to the length of measurement when the World Health Organization [1], Schofield [2] and Oxford [3] prediction equations were derived, might have improved the consistency and accuracy of the metabolic data in the derivation of our new weightbased equations [4]. This is most likely due to the inclusion of some periods of increased physical activity and sleep in the infants [4]. This is further substantiated by the fact the infants have a two to four hour sleepwake cycle between birth and six months of age [18].
These preliminary results suggest that the World Health Organization [1], the Schofield [2] and the Oxford [3] prediction equations may not be suitable for calculating caloric requirements in infants. Furthermore, both of the lengthweight equations derived with the EMTAC instrument [4] for calculating energy expenditure and resting metabolic rate, respectively, were also not suitable for use in healthy infants. Moreover, these results suggest that additional 24hour metabolic measurements need to be conducted in a greater number of infants from various ethnic groups. This will allow derivation of new equations that will be accurate for calculating energy requirements in healthy infants, accounting for all the metabolic variations that occur over a 24hour period. Moreover, infants with various clinical disorders also need to be included such as those from our prior study in infants suffering from primary and secondary malnutrition [7].
In general both the lengthweight prediction equations derived with the EMTAC instrument tended to overestimate their respective metabolic parameters. This might be due to the fact that metabolic measures were performed in the morning [10, 12], possibly representing the infants most active part of the day. This is further verified by the direct 24hour metabolic measurements that showed a lower energy expenditure and physical activity during the evening and early morning hours [6]. However, the World Health Organization [1], the Schofield [2] and the Oxford [3] prediction equations greatly underestimated resting metabolic rate. The lack of standardized methods, limited number of subjects less than sixmonths old and some of the data being obtained over 80 years ago probably contributed to errors in their derivation and consistent under estimates in resting metabolic rate when utilized in today's infants.
Conclusions
This is the first time actual 24hour metabolic measurements in the Enhanced Metabolic Testing Activity Chamber (EMTAC) were used as a reference to evaluate several previously published prediction equations. We found those prediction equations by the World Health Organization, Schofield and Oxford, as well as the two lengthweight based prediction equations from the EMTAC instrument, were inaccurate. However, the weight based prediction equations derived from our previous shortterm metabolic measurements of fourhours in the EMTAC were accurate for calculating energy requirements in healthy infants up to six months of age.
Abbreviations
 BMI:

Body Mass Index
 EE:

Energy expenditure
 EMTAC:

Enhanced metabolic testing activity chamber
 F:

Females
 Kg:

Kilograms
 L/100:

Conversion of L from meters to centimeters prior to calcuations
 LWT:

Length and weight
 M:

Males
 OXFORD:

Oxford
 RMR:

Resting metabolic rate
 SCH:

Schofield
 SD:

Standard Deviation
 SMR:

Sleeping metabolic rate
 WHO:

World Health Organization
 WT:

Weight
 24h EE:

Twentyfour hour extrapolated energy expenditure.
Declarations
Acknowledgements
This study was supported in part by a National Institutes of Health grant (#1R43HD/DK3818001A4). I will also like to thank Dr. Luisa Borrell of Lehman College for her invaluable assistance with the statistical analysis and editing of this manuscript.
Authors’ Affiliations
References
 World Health Organization (WHO): Energy and protein requirements, Report of a joint FAO/WHO/UNU Expert consultation. 1985, Geneva: WHO Technical Report Series 724Google Scholar
 Schofield WN: Predicting basal metabolic rate, new standards and review of previous work. Hum Nutr Clin Nutr. 1985, 39C: 5S41S.Google Scholar
 Henry CJ: Basal metabolic rate studies in humans: measurement and development of new equations. Public Health Nutr. 2005, 8: 11331152. 10.1079/PHN2005801.View ArticlePubMedGoogle Scholar
 Duro D, Rising R, Cole C, Cedillo M, Valois S, Lifshitz F: New equations for calculating metabolic rate in infants. J Ped. 2002, 140: 534539. 10.1067/mpd.2002.123308.View ArticleGoogle Scholar
 Henry CJK, Rees DG: New predictive equations for the estimation of basal metabolic rate in tropical peoples. Eur J Clin Nutr. 1991, 45: 177185.PubMedGoogle Scholar
 Rising R, Duro D, Cedillo M, Valois S, Lifshitz F: Daily metabolic rate in healthy infants. J Ped. 2003, 143: 180185. 10.1067/S00223476(03)003627.View ArticleGoogle Scholar
 Rising R, Sonmez GT: Energy Expenditure and physical activity in recovering malnourished infants. J Nutr & Met. 2010, Article ID 171490Google Scholar
 CDC Growth Charts: National Center for Health Statistics. Health, United States, 2000. 2000, Hyattsville, Maryland: Public Health ServiceGoogle Scholar
 Jequier E: Longterm measurement of energy expenditure in man: direct and indirect calorimetry. In Recent advances in obesity research. Edited by: Bjoerntorp P, Cairella M, Howard A. 1981, London: Libbey, 3: 130135.Google Scholar
 Cole C, Rising R, Mehta R, Hakim A, Choudhury S, Sundaresh M, Lifshitz F: Comprehensive assessment of the components of energy expenditure in infants using a new infant respiratory chamber. J Am Coll Nutr. 1999, 18: 233241.View ArticlePubMedGoogle Scholar
 Ravussin E, Lillioja S, Anderson TE, Christin L, Bogardus C: Determinants of 24hour Energy expenditure in man: Methods and results using a respiratory chamber. J Clin Invest. 1986, 78: 15681578. 10.1172/JCI112749.View ArticlePubMedPubMed CentralGoogle Scholar
 Rising R, Lifshitz F: A lower metabolic rate in infants from obese biological mothers. Nut J. 2008, 7: 1510.1186/14752891715.View ArticleGoogle Scholar
 Bland JM, Altman DG: Statistical methods for assessing agreement between two Methods of clinical measurement. Lancet. 1986, 327: 307310. 10.1016/S01406736(86)910081.View ArticleGoogle Scholar
 Durnin JVGA: Basal Metabolic Rate in Man. Working paper submitted to the Joint FAO/WHO/UNU Expert Consultation on Energy and Protein Requirements. 1981, Food and Agriculture Organization: RomeGoogle Scholar
 Wilkerson JE, Raven PB, Horvath SM: Critical temperature of unacclimatized male Caucasians. J Appl Phy. 1972, 33: 451455.Google Scholar
 Hayter JE, Henry CJK: A reexamination of basal metabolic rate predictive equations: the importance of geographic origin of subjects in sample selection. Eur J Clin Nutr. 1994, 48: 702707.PubMedGoogle Scholar
 Norgan NG: Population differences in body composition in relation to the Body Mass Index. Eur J Clin Nutr. 1994, 48: S10S27.PubMedGoogle Scholar
 BalfourLynn IM, Valman HB: Practical Management of the Newborn. 1993, London England: Blackwell Scientific PublicationsGoogle Scholar
Copyright
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.