Article Versions
Export Article
Cite this article
  • Normal Style
  • MLA Style
  • APA Style
  • Chicago Style
Original Article
Open Access Peer-reviewed

How Well Are We Predicting the Resting Energy Expenditure in Underweight to Obese Brazilian Adults?

Natália Tomborelli Bellafronte , Daiane Leite da Roza, Paula Garcia Chiarello
Journal of Food and Nutrition Research. 2019, 7(1), 19-32. DOI: 10.12691/jfnr-7-1-4
Received November 14, 2018; Revised December 24, 2018; Accepted January 16, 2019

Abstract

Resting energy expenditure (REE) measurement is costly and rarely feasibly. Alternatively, is easily predicted by predictive equations but accuracy varies across ethnicity, body size and body composition. The purpose was to evaluate the validity of REE predictive equations between adult from Brazil and across body mass index (BMI). We also developed a new equation. It was a cross-sectional observational study in which predicted REE was tested for agreement with indirect calorimetry. Brazilian men and women (n=367) age from 20 to 40 years were analyzed from October 2016 to October 2017. Participants were from underweight to obese (BMI from 17 to 40kg/m2). REE was measured (mREE) with indirect calorimetry and predicted (pREE) by REE predictive equations. Each equation was compared for accuracy (pREE= ±10%mREE). Bias between mREE and pREE with 95% confidence intervals, Bland-Altman and scatter plots graphs were also analyzed. The root mean squared error measured the performance of the equations. Stepwise multiple-regression analyses were applied for developing a new equation. As a result, the new equation (r2= 0.46; SEE= 226.92kcal/day) presented a good performance among normal, overweight and obese subjects (51, 51 and 45% of accuracy, respectively). The best performance of REE equations evaluated were from Anjos (39% of accuracy) among underweight subjects; Anjos and Sabounchi among normal weight (38% of accuracy for both) and overweight (53 and 48% of accuracy, respectively) individuals; Anjos and Owens (48 and 44% of accuracy, respectively) among obese subjects. No equation was able to accurately predict REE in 60% of the time and all tended to overestimate. In conclusion, Anjos, Sabounchi and Owens equations showed the best performances. Therefore, all predictive equation presented large individual errors. The new equation had a good performance from normal to obese individuals. However, indirect calorimetry is still necessary for the most accurate information about energy needs.

1. Introduction

As recently reported 1, 2, the proportion of adults with excess weight has been increasing in high to low-income countries, and ranges from 40 to 50% worldwide 1, 2. In Brazil, a fast increase in adult obesity has been occurring: the latest national survey on nutritional status 3 showed that half of the adults are overweight and 15% are obese. On the other hand, undernutrition still exists in the world, mainly in poverty 4 and in low-income countries such as Brazil 3. Excess weight and obesity are considered one of the main causes of mortality and metabolic diseases in the modern and affluent world 5 and malnutrition still has a great influence in morbimortality 4. Therefore, weight management has become a major global health challenge.

Young adults are a large segment of the population in need of weight management. As long-term small inadequacies in energy balance negatively interfere on weight management 6, 7, once under and overestimation of energy needs could predispose to the development of malnutrition and obesity-comorbidities, the assessment of energy needs is the first step of a personalized dietary plan. The FAO Reports 8, 9 established that the energy requirement should be based on total daily energy expenditure, of which 60 to 80% is represented by resting energy expenditure (REE) 10. Thus, the energy content of a diet is usually based on the REE.

REE can be measured by indirect calorimetry 7, but the methodology is costly and rarely feasibly in clinical practice. As an alternative, REE is usually and more easily calculated by predictive equations. On the other hand, its accurate determination is a challenge since it relies in the accuracy and bias of the chosen equation and varies across individuals and population groups. In addition, choosing the most adequate equation among the existing ones is another great challenge. The FAO Reports 8, 9 provide a set of formulas to estimate REE based on easily assessed parameters (sex, age, body weight, and height) to guide clinicians. Thus, the validation of REE prediction equations is a way to marginally provide utility and information about calculation of energy needs in various populations.

As ethnicity 11, 12 and body composition 12, 13 play an important role in the variation of REE, they are a central issue in the performance of REE predictive equations. As recently observed by an expert panel from the American Dietetic Association 6, clinicians should be cautious when applying predictive equations to groups of subjects who are underrepresented in development and validation studies, as the application to participants that fall outside of the original sample might not be appropriate resulting in inappropriate REE estimation.

Taking this into account, the accuracy and reliability of REE prediction equations in different populations remain inconsistent and need further validation: the evaluation of usually applied equations has been limited since the majority of the studies were done with Europeans and North Americans, are restricted across body size and body composition, and scarce in young healthy population. As a gap of investigation, studies with Latin American populations are rare.

Therefore, the purpose of the present study was to examine the accuracy and bias of common REE predictive equations when applied in Brazilian adults from underweight to obese to assess the equations ‘utility in a population with different body sizes. In addition, three newer equations were evaluated: one with a specific ethnicity fit derivate from a metanalysis 14, another recently developed from a Mexican data base 15, and one developed from Brazilian population 16. We also developed and validated a new equation.

With this study we pretend to help clinicians in choosing the best REE prediction equations for application in the target population, also in the planning of public health policies and dietary interventions, helping in weigh management and prevention of malnutrition and excess weight.

2. Materials and Methods

This cross-sectional observational study was conducted at the University of São Paulo at Ribeirão Preto, with a group of 367 non-smoker Brazilians (hospital or university staff, under or graduate students), 193 females and 174 males, aged between 20 and 40 years, being 3% of Asian and 9% of African descent. Subjects ranging from underweight to obese without acute or chronic diseases (exception: obesity) were recruited specifically for this study and exclusion criteria were pregnancy, lactation, use of drugs that influence energy metabolism and body composition, menopause, presence of amputation and metal objects in the body as protheses or pacemakers, being on a diet for weight loss, unintentional or intentional weight loss greater than 10% in the last 6 months, and others that could interfere in bioelectrical impedance analysis and indirect calorimetry results.

The study was conducted in accordance to the Declaration of Helsinki and received the approval of the Local Ethical Committee (number: 1076550). All participants provided written informed consent prior to the initiation of the study procedures. Convenience sampling was applied to contact and screen potential candidates: all subjects interested in participating in the research during the study period (from October 2016 to October 2017) and that met the eligibility criteria were evaluated.

This study followed the STROBE guidelines.

The same professional performed all measurements consecutively in a temperature-controlled (23-27°C) room, early in the morning after a fasting period of 12 h with abstention from beverage or foods containing alcohol or caffeine and rigorous physical activity for 36 h prior to evaluation 7, 17. All participants completed health-related interview to ensure adherence to study protocol.

Anthropometric data. Body weight and height were measured in duplicate in light clothing without shoes using a platform beam scale with a built-in stadiometer (FilizolaEletrônica ID 1500, Filizola Corporation) and the mean of the values was used. Nutritional status was defined according to BMI: underweight (BMI<18.5 kg/m2), normal weight (18.5<BMI<25 kg/m2), overweight (25<BMI<30 kg/m2), and obese (BMI≥30 kg/m2).

Body Composition data. Body composition was evaluated by tetra-polar wrist-to-ankle multifrequency bioelectrical impedance spectroscopy (BCM, Fresenius Medical Care) in accordance to published protocol 17: measurements were carried out on the right side of the body after the volunteer being in supine position for 20 min and data about fat free mass (FFM) and fat mass (FM) were collected.

REE data. REE was measured according to a recommended protocol 7 in a thermoneutral (20-25°C), humidity-controlled (45-55%), and quiet environment and in a supine position using a validated indirect calorimetry system 18 (VO2000, Aerograf). Before starting the test, the device was heated for 30 min and the barometric and gasometric calibration were performed, both automatically and manually using Aerograf software, according to manufacturer's instructions. Subjects laid down on the bed for a 25-min adaptation period, proved to allow accurate REE measures 7. Afterwards, oxygen consumption and carbon dioxide production were measured for 30 min; the first 5 min were eliminated and REE was calculated from minute-to-minute values averaged over the remaining time of steady state 7 using the Weir equation 19 and reported in kcal/day. Data were excluded from analysis if the respiratory quotient was outside the physiological range (0.70-1.00) and if steady state was less than 15 min. Participants were monitored for activity, hyperventilation, or napping. Fifty participants (25 women) repeated the REE measures on different days with an interval of 3 to 5 days. Mean REE values between days were not significantly different (mean difference = 7 kcal±15; p=0.43), the two measures were highly correlated (r=0.94), and the mean intra-individual coefficient of variation of measured REE (mREE) was 4.7%.

Evaluation of REE predictive equations. The performance of predictive equations commonly applied in clinical practice 8, 9, 13, 20, 21, 22, 23, 24, 25, 26, 27, 28 and three new equations 14, 15, 16 were analyzed (see Table 1, supplementary material). We applied all equations according to the sex, age, weight, height and BMI status. Performance of predicted REE (pREE) and bias between pREE and mREE at group and individual levels were evaluated for the total sample and for each BMI subgroup. The CI95% for mean bias was evaluated. The root mean square error (RMSE) was used to measure how well the equations fitted the mREE at group level. A predictive equation was considered accurate at the individual level if pREE and mREE were within ±10%. Under (pREE<90% mREE) and over predictions (pREE>110% mREE) were also calculated. Bland-Altman plots were used in the total sample to assess agreement at the individual level between pREE and mREE. Correlation between bias of pREE and mREE with BMI were analyzed in total sample by scatter plots.

Development and evaluation of a new REE equation. The sample was randomly split into design (n=184) and validation (n=183) groups and compared with Student's t-test. For the generation of the new equation, correlation analysis (Pearson or Spearman correlation) and scatter plots were used to select the covariates (sex, men = 1 and women = 0; age (years); weight (kg); height (cm); FFM (kg and %); and FM, (kg and %)) that would be entered into the multiple linear regression model to observe the linearity between continuous independent variables and the dependent variable (mREE). The correlation coefficient and the independent association of each covariate were analyzed, demonstrating the collinearity between them. The interactions between the variables were verified; no significant interaction was found. In order to identify the variables, either alone or in combination, with the highest correlations and predictors of mREE, a stepwise multiple regression analysis (α=0.15) was used to predict REE (kcal/day) in the sample. The adjusted coefficient of determination (r2), the slope (α), and the intercepts (β) were analyzed, and the fit of the model was performed by selecting the best model. For the final fit of the prediction model, the normality of the residuals was tested by observing the plot between the residuals and the predicted values, presenting constant variance and zero mean (application of normal Q-Q plot and Shapiro-Wilk normality test, with p>0.05). The statistical analyses described in the previous paragraph were applied in the evaluation of the new equation.

Statistical analysis. Continuous variables are presented as mean ± SD and minimum and maximum values, and categorical variables as frequencies and percentages. Q-Q plots were applied to analyze whether data were normally distributed. For variables that were not normally distributed, appropriate logarithmic transformations were performed. The Mann-Whitney test was used for differences between males and females, and ANOVA for comparing data between BMI subgroups. Pearson correlation was used to assess the association between pREE and mREE. Statistical significance was considered when p<0.05. Data analysis was performed using MINITAB 2013.

3. Results

Descriptive data are shown in Table 2. Age was similar between sexes. Men had a higher body weight, height, FFM, and FFMI, but FM and FMI were higher in women. Both were classified on average as overweight. REE was greater in men than women, but when corrected for body weight, no difference was observed. Between BMI subgroups, women and men were equally distributed; underweight was an exception due to a greater participation of women. Weight, FM, FMI, and REE were significantly different and increased with increasing BMI classification. However, REE per body weight decreased with increasing BMI classification.

Design and validation groups were similar (see Table 3, supplementary material). The new equation was named Bellafronte-Chiarello: REE (kcal/day) = 546.3 + 158.0(sex) + 8.8(weight) (Model R2 = 0.46; SEE = 226.92), weight in kg and sex, men = 1 and women = 0. The equation was derived from simple parameters providing an easily available equation for clinical settings. This equation was unbiased at group level (see Table 4). At individual level, a good performance was found for normal weight, overweight, and obese; underweight subjects presented a low accuracy and a high underestimation rate.

The statistics of the REE predictive equations evaluated are shown in Table 5 (total sample) and Table 6, Table 7, Table 8, Table 9 (BMI subgroups). Bland-Altman plots for bias between pREE and mREE in the total sample are presented in Figure 1 (supplementary material): large limits of agreement towards overestimation were observed with greater bias in extremes of REE. Scatter plots between pREE and mREE bias with BMI for the total sample are presented in Figure 2 (supplementary material): a positive correlation was observed between them. Only for Owens, Livingston-Kohlstadt#2, Muller#2, Anjos, and NEQ equations no correlation was seen (p>0.05).

  • Figure 1. Bland-Altman plots of the bias between REE determined by indirect calorimetry and by sixteen prediction equations against the mean of the two methods. Analysis performed with the total sample. Bias (kcal/d) for each REE prediction equation was determined by subtracting mREE from pREE. The solid black line represents the mean measurement difference and the dashed black lines correspond to the 95% limits of agreement (±2 SD). The gray solid line is the regression line for the measurement bias plotted against the mean of the two methods for female and the gray dashed line for male subjects. Separate plots are presented for each equation. The name of each equation ploted is in the graphic and is referee to as the same named presented in Table 1. For NEQ equation, underweight subjects were not evaluated as the equations developed are only for normal weight to obese subjects. For Bellafronte-Chiarello equation, the equation was evaluated using subjects form validation group (see Table 3)
  • Figure 2. Scatter plots of the bias between REE determined by indirect calorimetry and by sixteen prediction equations against the mean of body mass index. Analysis performed with the total sample. Bias (kcal/d) for each REE prediction equation was determined by subtracting mREE from pREE. The gray solid line is the regression line for the measurement bias plotted against the body mass index for female and the gray dashed line for male subjects. Pearson correlation test was applied for measurement the correlation between bias and body mass index and its value is presented in the graphic. Separate plots are presented for each equation. The name of each equation ploted is in the graphic and is referee to as the same named presented in Table 1. For NEQ equation, underweight subjects were not evaluated as the equations developed are only for normal weight to obese subjects. For Bellafronte-Chiarello equation, the equation was evaluated using subjects form validation group (see Table 3)

In the total sample (Table 5), all REE predictive equations were biased: the lowest RMSE and the highest accuracy rates were from Anjos, Sabounchi, and Owens equations. For the underweight sample (Table 6), Muller#1 was the only equation unbiased at the group level: the lowest RMSE was from Anjos, Muller#1, and Sabounchi, and the highest accuracy rates from Anjos, Henry-Rees, and Owens. In the normal weight sample (Table 7), all REE predictive equations were biased: the lowest RMSE and the highest accuracy rates were from Anjos, Sabounchi, and Owens. In the overweight sample (Table 8), all REE predictive equations were biased, with the lowest RMSE and highest accuracy rates from Anjos and Sabounchi. In the obese sample (Table 9), Sabounchi was the only equation unbiased at the group level: the lowest RMSE were from Sabounchi and Anjos, and the highest accuracy rates from Anjos and Owens equations.

In almost all REE predictive equations evaluated, a decrease in RMSE was observed with an increase in BMI classification, except in the obese group. The exceptions were FAO/WHO/ONU#1, Ireton-Jones, and Muller#1 and #2: the RMSE did not present a pattern of behavior.

Schofield and FAO/WHO/ONU#1 and #2 equations maintained their accuracy rates without major changes regardless of increase in BMI classification and with the highest accuracy rate in the underweight subgroup. Harris-Benedict equation had a similar pattern, but the underweight subgroup had the lowest accuracy rate. For Ireton-Jones, the accuracy rate was always less than 10%. For Sabounchi and Anjos, the accuracy rate increased with increasing BMI classification and decreased in the obese group. For the other equations, there was a tendency to a higher accuracy rate as the classification of BMI increased.

Concerning underprediction rate, in general, the equations maintained the underestimation rate as BMI classification increased. For Owens, an increase in the underprediction rate was observed as BMI classification increased. For Muller#1 and #2, a considerable underprediction rate was observed in the underweight group, especially for Muller#1. For Sabounchi, the obese group had a high underprediction rate.

Almost all equations had a tendency to decrease the overestimation rate as BMI classification increased. Muller#1 and #2 began to decrease the overprediction rate from normal weight individuals. The Ireton-Jones and Harris-Benedict equations maintained the overestimation rate regardless of BMI classification. FAO/WHO/ONU#1 and #2 and Schofield increased the overprediction rate as BMI classification increased.

4. Discussion

There has been limited research concerning the accuracy of predictive models in estimating REE in young healthy adults, across BMI spectrum and in a population from Latin America. The results of this study showed large errors with commonly applied equations, leading to severely inaccurate recommendations.

In general, Sabounchi, Anjos, and Owens predictive equations presented the best performances. Unfortunately, none of the equations evaluated were able to accurately predict REE in more than 60% of the time, in the total sample or BMI subgroups. In addition, all equations tended to overestimate REE; underestimation usually occurred in a small sample (except for the underweight group). For almost all equations, bias increased with increasing BMI and Bland-Altman plots presented large limits of agreement with fixed and systematic errors: the equations tended to overestimate mREE values at lower means and to underestimate values at higher means. Thus, weight management still benefits from indirect calorimetry assessment.

The equations of Muller 13 were generated from a German database. The BMI-weight-specific Muller#1 equation for underweight was developed with older (32.0±12.2 years) and more underweight (BMI: 16.3±1.6kg/m²) subjects than our sample. Probably, for this reason it presented low accuracy rate and large underprediction in underweight subgroup being not recommended.

The equation of Anjos 16, developed from 529 Brazilian adults (20 to 80 years old), was the one with the best performances. The original sample used for Anjos equation, despite being older (female: mean age of 45 years; male: mean age of 43 years), had similar BMI (25 kg/m2 for both sexes) and REE (female: 1137kcal; male: 1395kcal), justifying the high accuracy rates and low bias. In its validation study 29, the accuracy rates were 77.5 and 59.3% for women and men, respectively, much greater than our results. As the validation sample was more similar to the original sample than our data (34 and 32 years old; BMI: 23.8 and 24.6 kg/m2; REE: 1155 and 1500 kcal, for women and men, respectively), the improved accuracy is explained.

REE was estimated using 1 of the 20 Sabounchi meta-regression equations 14, a weight and height-based equation applicable to white men and women aged from 18 to 57 years and with the highest r2 (0.804) of this group, thus better fitting our study subjects as they were adults and the majority was white. The ethnicity fit promoted by the meta-regression 14 improved the performance of pREE as this equation presented the best accuracy rates for normal and overweight subjects. A study developed with older adults (38.2±14.9 years old) of both sexes from USA 30 showed an accuracy rate of about 70% from this equation. As the studies used in the meta-regression are mostly derived from USA and Europe population, improved performance is explained.

The NEQ equation 15 was developed from a Mexican sample. Both equations had low accuracy rates and high bias toward overestimation, probably because they were derived from an older, more obese and with higher REE subjects than our sample.

Owens equation 22, 23 has a good performance, presenting the best accuracy rates in overweigh and obese subjects. Recently, a study 15 evaluated the performance of several REE predictive equations in a Mexican population and found an accuracy rate around 43 and 39% in overweight and obese individuals, respectively, for Owens equation, values close to our results. In turn, an Italian study 31 showed higher accuracy rates of 72, 66, and 55% in normal, overweight, and obese subjects, respectively. However, they were older (48 years), more obese (BMI: 34.3 kg/m2; 67% of obese participants), and with a higher REE (1609 kcal), characteristics that were more similar to the original Owens sample.

Henry-Rees equation 26 was developed using the FAO/WHO/ONU equations database with the exclusion of Italian soldiers and the addition of subjects from tropical regions, with the intention of having a more representative ethnic diversity. Despite this, it still had bias towards overestimation with fixed and systematic errors.

The data from the Livingston-Kohlstadt equation 27 derived from the original Harris-Benedict 20 and Owens data 22, 23 with 367 individuals added to the sample. These authors utilized an allometric model based on the relationship between REE and body size, supposing it would confer a better ability to predict REE in obese individuals. In our study, Livingston-Kohlstadt equation was better than Harris-Benedict but not than Owens equation in overweight and obese subjects. Two studies 28, 31 observed that accuracy rate of Livingston-Kohlstadt equations decreased as BMI increased, and fall dramatically 28 within those with BMI>50 kg/m2, showing that they are not appropriate for the morbidly obese.

The FAO/WHO/ONU #1 and #2 equations 8, 9 had the worst performance, with bias toward overestimation in overweight and obese Mexicans 15. In a study with Brazilian obese woman 32, FAO/WHO/ONU and Harris-Benedict had accuracy rates of 35 and 40%, higher than our results, but, as emphasized by the authors, Bland-Altman plots showed poor agreement, limiting their use in clinical practice for weight management.

Evidence analysis by the American Dietetic Association 6 has lent support to Mifflin-St Jeor equation 25 as the most appropriate for predicting REE in non-obese and obese subjects. A study with Brazilian subjects 29 showed a better performance for Mifflin-St Jeor equation between overweight to obese compared to leaner subjects, in agreement with our results once this equation increased accuracy rates as BMI classification increased. But, although the 2005 Evidence Working Group 6 recommended the use of Mifflin-St. Jeor equation for predicting REE, the results of our study and others 29, 32 suggest that it is inappropriate for use in young Brazilian adults.

The Schofield equation 21, recommended for international use by the World Health Organization, resulted in overestimation of REE and low accuracy rates, as seen by others 16, 29, 32. Therefore, although recommended by a respected international association, Schofield equation was found inappropriate for adults in Brazil.

The Frankenfield equations 28 tended to have lower RMSE with increasing BMI classification and its highest accuracy rate was observed among obese subjects. However, the high overprediction rates compromises the use of these equations in young adults, as observed by others 33.

The Harris-Benedict equation 20 is still widely used in clinical settings. However, its predictive capacity has been shown to be low with a tendency for overestimation, as seen by Anjos and colleagues 16, with accuracy rates around 20% when applied in Brazilian adults. The tendency for overprediction of REE was also observed by other studies 33. Based on our findings, we do not recommend the use of Harris-Benedict equation for predicting REE among Brazilian adults.

The Ireton-Jones equation 24 presented the worst accuracy rates and performance, probably because it was developed from subjects before bariatric surgery with a much higher obesity degree than our participants.

Besides the use of REE predictive equations in their original form, the substitution of the current body weight for an adjusted one is a common practice for predicting REE even with worse accuracy rates 28 as a result. Clinicians also commonly apply a fixed REE of 25 or 30 kcal/kg/day or 2000 kcal for females and 2500 kcal for males: as can be seen in Table 2, regardless of sex and BMI classification, the mean mREE was always below 1600 kcal and the highest REE in women and in men was less than 2000 and 2500 kcal, respectively; in addition, mean mREE/weight was less than 20 kcal/kg/day. Therefore, the use of fixed values is not recommended.

The developed new predictive equation showed low bias and good performance from normal to obese subjects. Because the large limits of agreement seen by Bland-Altman plots, the equation needs to be extensively tested before a recommendation could be made.

This study has several strengths and limitations. As strengths, the REE was measured by indirect calorimetry, a reference method, with a validated system 18; its reproducibility, evaluated in a subsample, was proved to be high. All measurements were done by the same trained and experienced dietitian at the same time, in standardized conditions for all groups and the adherence to the protocol was verified prior to measurements. The inclusion of subjects with BMI ranging from undernourished to obese allowed a more comprehensive analysis of REE predictive equations performance. Thus, the analyses performed can be generalized to individuals with diverse body composition.

Once not only one REE equation can be used for prediction of energy needs of subjects from different populations, it is necessary to evaluated the validation of REE prediction equations in different races, mainly in the marginally represented, as an example of Latin American populations. This is the first study to evaluate the predictive power of numerous REE equations, including three newer equations 14, 15, 16, in a large sample of Brazilians across BMI spectrum. Furthermore, the application of statistical tests detailing the accuracy at group and individual levels allowed a detailed analysis of pREE and the limitations of each predictive equation.

As limitations, the applicability of the results is restricted to adult individuals and cannot be extrapolated to children or elderly. Also, as Brazil is a multiethnic country, the low participation of African descendants in our sample is a negative point.

5. Conclusions

In conclusion, this study showed a wide variation in the accuracy of predictive equations when applied in healthy Brazilian adults. The new equation presented a good performance from normal weight to obese subjects but needs to be extensively tested before a recommendation could be made. The predictive equations with the best performance were Anjos among underweight, Anjos and Sabounchi among normal and overweight, and Anjos and Owens among obese. However, the poor accuracy rates highlight a caution application: until a more accurate formula is developed, the equations cited above may be recommended for estimating energy needs in Brazilian health adults respecting BMI subgroups as a starting point in nutritional therapy and the energy content may need to be later adjusted based on clinical signs and judgment. Indirect calorimetry is still necessary when the most accurate information about energy needs is needed.

Taking all these findings into account, the development of low cost and portable indirect calorimetry equipment, as well as the extensive validation of these new devices in comparation with a reference method, are a broad field of research, once REE prediction equations still lack desired accuracy.

Acknowledgements

This work was supported by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) under grant finance code 001. CAPES had no role in research project and article development. We sincerely appreciate all participants for their kind cooperation in our study. We also wish to acknowledge all of the investigators in our study and the hospital staff members.

Statement of Competing Interests

The authors have no competing interests.

List of Abbreviations

BMI: body mass index; FFM: fat free mass; FFMI: fat free mass index; FM: fat mass; FMI: fat mass index; mREE: measured REE by indirect calorimetry; pREE: predicted REE by prediction equations; REE: resting energy expenditure; RMSE: root mean square error; SEE: standard error of the estimate.

Financial Support

This work was supported by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) under grant finance code 001. CAPES had no role in research project and article development.

Transparency Declaration

The lead author affirms that this manuscript is an honest, accurate and transparent account of the study. No important aspects of the study were omitted and any discrepancies from the study as planned was explained. The reporting of this work is compliant with STROBE guideline.

Acknowledgements

We sincerely appreciate all participants for their kind cooperation in our study. We also wish to acknowledge all of the investigators in our study and the hospital staff members.

References

[1]  (NCD-RisC), NRFC, “Trends in adult body-mass index in 200 countries from 1975 to 2014: a pooled analysis of 1698 population-based measurement studies with 19.2 million participants”, Lancet, 387 (10026). 1377-1396. April 2016.
In article      View Article
 
[2]  Ng, M., Fleming, T., Robinson, M., Thomson, B., Graetz, N., Margono, C., et al, “Global, regional, and national prevalence of overweight and obesity in children and adults during 1980-2013: a systematic analysis for the Global Burden of Disease Study 2013”, Lancet, 384 (9945). 766-781. Aug 2014.
In article      View Article
 
[3]  Instituto Brasileiro de Geografia e Estatística, Pesquisa Nacional de saúde 2013: Percepção do estado de saúde, estilo de vida e doenças crônicas. Brasil, Grandes Regiões e Unidades da Federação, IBGE, 2013. http://www.ibge.gov.br/home/estatistica/populacao/pns/2013/. Published 2014. Accessed 05 April, 2018.
In article      
 
[4]  Uauy, R., Garmendia, M. L., Corvalán, C, “Addressing the double burden of malnutrition with a common agenda”, Nestle Nutr Inst Workshop Ser, 78. 39-52. Jan 2014.
In article      View Article  PubMed
 
[5]  Kopelman, P.G., “Obesity as a medical problem”, Nature, 404 (6778). 635-43. April 2000.
In article      View Article  PubMed
 
[6]  Frankenfield, D., Roth-Yousey, L., Compher, C., “Comparison of predictive equations for resting metabolic rate in healthy non-obese and obese adults: a systematic review”, J Am Diet Assoc, 105 (5). 775-789. May 2005.
In article      View Article  PubMed
 
[7]  Fullmer, S., Benson-Davies, S., Earthman, C.P., Frankenfield, D.C., Gradwell, E., Lee, P.S., et al, “Evidence Analysis Library Review of Best Practices for Performing Indirect Calorimetry in Healthy and Non–Critically Ill Individuals”, J Acad Nutr Diet, 115 (9). 1417-1446. September 2015.
In article      View Article  PubMed
 
[8]  “Energy and protein requirements. Report of a Joint FAO/WHO/UNU Expert Consultation”, in World Health Organ Tech Rep Ser. 724:1-206. 1985.
In article      PubMed
 
[9]  “Human energy requirements. Report of a Joint FAO/WHO/UNU Expert Consultation”, in World Health Organ Tech Rep Ser. 1. Roma: FAO, 2004.
In article      
 
[10]  Psota, T., Chen, K.Y., “Measuring energy expenditure in clinical populations: rewards and challenges”, Eur J Clin Nutr, 67 (5). 436-442. May 2013.
In article      View Article  PubMed  PubMed
 
[11]  Adzika Nsatimba, P.A., Pathak, K., Soares, M.J., “Ethnic differences in resting metabolic rate, respiratory quotient and body temperature: a comparison of Africans and European Australians”, Eur J Nutr, 55 (5). 1831-1838. August 2016.
In article      View Article  PubMed
 
[12]  Hasson, R.E., Howe, C.A., Jones, B.L., Freedson, P.S., “Accuracy of four resting metabolic rate prediction formulas: effects of sex, body mass index, age, and race/ethnicity”, J Sci Med Sport, 14 (4). 344-351. July 2011.
In article      View Article  PubMed
 
[13]  Müller, M.J., Bosy-Westphal, A., Klaus, S., Lührmann, P.M., Neuhäuser-Berthold, M., et al., “World Health Organization equations have shortcomings for predicting resting energy expenditure in persons from a modern, affluent population: generation of a new reference standard from a retrospective analysis of a German database of resting energy expenditure”, Am J Clin Nutr, 80 (5). 1379-1390. November 2004.
In article      View Article  PubMed
 
[14]  Sabounchi, N.S., Rahmandad, H., Ammerman, A, “Best-fitting prediction equations for basal metabolic rate: Informing obesity interventions in diverse populations”, Int J Obes, 37 (10). 1364-1370. October 2013.
In article      View Article  PubMed  PubMed
 
[15]  Orozco-Ruiz, X., Pichardo-Ontiveros, E., Tovar, A.R., Torres, N., Medina-Vera, I., Prinelli, F., et al., “Development and validation of new predictive equation for resting energy expenditure in adults with overweight and obesity”, Clin Nutr, S0261-5614 (17)31395-X. November 2017.
In article      
 
[16]  Anjos, L.A., Wahrlich, V., Vasconcellos, M.T., “BMR in a Brazilian adult probability sample: the Nutrition, Physical Activity and Health Survey”, Public Health Nutr, 17 (4). 853-860. April 2014.
In article      View Article  PubMed
 
[17]  Kyle, U.G., Bosaeus, I., De Lorenzo, A.D., Deurenberg, P., Elia, M., Manuel Gómez, J., et al., “ESPEN. Bioelectrical impedance analysis – parte I: utilización in clinical practice”, Clin Nutr, 23 (6). 1430-1453. December 2004.
In article      View Article  PubMed
 
[18]  Wahrlich, V., Anjos, L.A., Going, S.B., Lohman, T.G., “Validation of the VO2000 calorimeter for measuring resting metabolic rate”, Clin Nutr, 25 (4). 687-692. August 2006.
In article      View Article  PubMed
 
[19]  Weir, J.D.V., “New methods for calculating metabolic rate with special reference to protein metabolism”, J Physiol, 109 (1-2). 1-9. August 1949.
In article      View Article  PubMed  PubMed
 
[20]  Harris, J.A., Benedict, F.G., “A biometric study of human basal metabolism. Proceedings of the National Academy of Sciences of the United States of America”, Proc Natl Acad Sci USA, 4 (12). 370-373. December 1918.
In article      View Article  PubMed
 
[21]  Schofield, W.N., “Predicting basal metabolic rate, new standards and review of previous work. Human Nutrition”, Clin Nutr, 39 (1). 5-41. 1985.
In article      
 
[22]  Owen, O.E., Kavle, E., Owen, R.S., Polansky, M., Caprio, S., Mozzoli, M.A., et al., “A reappraisal of caloric requirements in healthy women”, Am J Clin Nutr, 44 (1). 1-19. July 1986.
In article      View Article  PubMed
 
[23]  Owen, O.E., Holup, J.L., D’Alessio, D.A., Craig, E.S., Polansky, M., Smalley, K.J., et al., “A reappraisal of the caloric requirements of men”, Am J Clin Nutr, 46 (6). 875-885. December 1987.
In article      View Article  PubMed
 
[24]  Ireton-Jones, C.S., “Evaluation of energy expenditures in obese patients”, Nutr Clin Pract, 4 (4). 127-129. August 1989.
In article      View Article  PubMed
 
[25]  Mifflin, M.D., St Jeor, S.T., Hill, L.A., Scott, B.J., Daugherty, S.A., Koh, Y.O., “A new predictive equation for resting energy expenditure in healthy individuals”, Am J Clin Nutr, 51 (2). 241-247. February 1980.
In article      View Article  PubMed
 
[26]  Henry, C.J., Rees, D.G., “New predictive equations for the estimation of basal metabolic rate in tropical peoples”, Eur J Nutr, 45 (4). 1777-1785. April 1991.
In article      
 
[27]  Livingston, E.H., Kohlstadt, I., “Simplified resting metabolic rate-predicting formulas for normal-sized and obese individuals”, Obes Res, 13 (7). 1255-1262. July 2005.
In article      View Article  PubMed
 
[28]  Frankenfield, D., “Bias and accuracy of resting metabolic rate in non-obese and obese adults”, Clin Nutr, 32 (6). 976-982. December 2013.
In article      View Article  PubMed
 
[29]  Wahrlich, V., Teixeira, T.M., Anjos, L.A.D., “Validity of a population-specific BMR predictive equation for adults from an urban tropical setting”, Clin Nutr, 37 (1). 208-213. February 2018.
In article      View Article  PubMed
 
[30]  Flack, K.D., Siders, W.A., Johnson, L., Roemmich, J.N., “Cross-validation of resting metabolic rate prediction equations”, J Acad Nutr Diet, 116 (9). 1413-1422. September 2016..
In article      View Article  PubMed
 
[31]  Bedogni, G., Bertoli, S., Leone, A., De Amicis, R., Lucchetti, E., Agosti, F., et al., “External validation of equations to estimate resting energy expenditure in 14952 adults with overweight and obesity and 1948 adults with normal weight from Italy”, Clin Nutr, S0261-5614 (17). 31409-7. November 2017.
In article      
 
[32]  Poli, V.F., Sanches, R.B., Moraes, A.D., Fidalgo, J.P., Nascimento, M.A., Andrade-Silva, S.G., et al., “Resting energy expenditure in obese women: comparison between measured and estimated values”, Br J Nutr, 116 (7). 1306-1313. October 2016.
In article      View Article  PubMed
 
[33]  Willis, E.A., Herrmann, S.D., Ptomey, L.T., Honas, J.J., Bessmer, C.T., Donnelly, J.E., et al., “Predicting energy expenditure in young adults”, Obes Res Clin Pract, 10 (3). 304-314. May 2016.
In article      View Article  PubMed  PubMed
 

Published with license by Science and Education Publishing, Copyright © 2019 Natália Tomborelli Bellafronte, Daiane Leite da Roza and Paula Garcia Chiarello

Creative CommonsThis work is licensed under a Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Cite this article:

Normal Style
Natália Tomborelli Bellafronte, Daiane Leite da Roza, Paula Garcia Chiarello. How Well Are We Predicting the Resting Energy Expenditure in Underweight to Obese Brazilian Adults?. Journal of Food and Nutrition Research. Vol. 7, No. 1, 2019, pp 19-32. http://pubs.sciepub.com/jfnr/7/1/4
MLA Style
Bellafronte, Natália Tomborelli, Daiane Leite da Roza, and Paula Garcia Chiarello. "How Well Are We Predicting the Resting Energy Expenditure in Underweight to Obese Brazilian Adults?." Journal of Food and Nutrition Research 7.1 (2019): 19-32.
APA Style
Bellafronte, N. T. , Roza, D. L. D. , & Chiarello, P. G. (2019). How Well Are We Predicting the Resting Energy Expenditure in Underweight to Obese Brazilian Adults?. Journal of Food and Nutrition Research, 7(1), 19-32.
Chicago Style
Bellafronte, Natália Tomborelli, Daiane Leite da Roza, and Paula Garcia Chiarello. "How Well Are We Predicting the Resting Energy Expenditure in Underweight to Obese Brazilian Adults?." Journal of Food and Nutrition Research 7, no. 1 (2019): 19-32.
Share
  • Figure 1. Bland-Altman plots of the bias between REE determined by indirect calorimetry and by sixteen prediction equations against the mean of the two methods. Analysis performed with the total sample. Bias (kcal/d) for each REE prediction equation was determined by subtracting mREE from pREE. The solid black line represents the mean measurement difference and the dashed black lines correspond to the 95% limits of agreement (±2 SD). The gray solid line is the regression line for the measurement bias plotted against the mean of the two methods for female and the gray dashed line for male subjects. Separate plots are presented for each equation. The name of each equation ploted is in the graphic and is referee to as the same named presented in Table 1. For NEQ equation, underweight subjects were not evaluated as the equations developed are only for normal weight to obese subjects. For Bellafronte-Chiarello equation, the equation was evaluated using subjects form validation group (see Table 3)
  • Figure 2. Scatter plots of the bias between REE determined by indirect calorimetry and by sixteen prediction equations against the mean of body mass index. Analysis performed with the total sample. Bias (kcal/d) for each REE prediction equation was determined by subtracting mREE from pREE. The gray solid line is the regression line for the measurement bias plotted against the body mass index for female and the gray dashed line for male subjects. Pearson correlation test was applied for measurement the correlation between bias and body mass index and its value is presented in the graphic. Separate plots are presented for each equation. The name of each equation ploted is in the graphic and is referee to as the same named presented in Table 1. For NEQ equation, underweight subjects were not evaluated as the equations developed are only for normal weight to obese subjects. For Bellafronte-Chiarello equation, the equation was evaluated using subjects form validation group (see Table 3)
[1]  (NCD-RisC), NRFC, “Trends in adult body-mass index in 200 countries from 1975 to 2014: a pooled analysis of 1698 population-based measurement studies with 19.2 million participants”, Lancet, 387 (10026). 1377-1396. April 2016.
In article      View Article
 
[2]  Ng, M., Fleming, T., Robinson, M., Thomson, B., Graetz, N., Margono, C., et al, “Global, regional, and national prevalence of overweight and obesity in children and adults during 1980-2013: a systematic analysis for the Global Burden of Disease Study 2013”, Lancet, 384 (9945). 766-781. Aug 2014.
In article      View Article
 
[3]  Instituto Brasileiro de Geografia e Estatística, Pesquisa Nacional de saúde 2013: Percepção do estado de saúde, estilo de vida e doenças crônicas. Brasil, Grandes Regiões e Unidades da Federação, IBGE, 2013. http://www.ibge.gov.br/home/estatistica/populacao/pns/2013/. Published 2014. Accessed 05 April, 2018.
In article      
 
[4]  Uauy, R., Garmendia, M. L., Corvalán, C, “Addressing the double burden of malnutrition with a common agenda”, Nestle Nutr Inst Workshop Ser, 78. 39-52. Jan 2014.
In article      View Article  PubMed
 
[5]  Kopelman, P.G., “Obesity as a medical problem”, Nature, 404 (6778). 635-43. April 2000.
In article      View Article  PubMed
 
[6]  Frankenfield, D., Roth-Yousey, L., Compher, C., “Comparison of predictive equations for resting metabolic rate in healthy non-obese and obese adults: a systematic review”, J Am Diet Assoc, 105 (5). 775-789. May 2005.
In article      View Article  PubMed
 
[7]  Fullmer, S., Benson-Davies, S., Earthman, C.P., Frankenfield, D.C., Gradwell, E., Lee, P.S., et al, “Evidence Analysis Library Review of Best Practices for Performing Indirect Calorimetry in Healthy and Non–Critically Ill Individuals”, J Acad Nutr Diet, 115 (9). 1417-1446. September 2015.
In article      View Article  PubMed
 
[8]  “Energy and protein requirements. Report of a Joint FAO/WHO/UNU Expert Consultation”, in World Health Organ Tech Rep Ser. 724:1-206. 1985.
In article      PubMed
 
[9]  “Human energy requirements. Report of a Joint FAO/WHO/UNU Expert Consultation”, in World Health Organ Tech Rep Ser. 1. Roma: FAO, 2004.
In article      
 
[10]  Psota, T., Chen, K.Y., “Measuring energy expenditure in clinical populations: rewards and challenges”, Eur J Clin Nutr, 67 (5). 436-442. May 2013.
In article      View Article  PubMed  PubMed
 
[11]  Adzika Nsatimba, P.A., Pathak, K., Soares, M.J., “Ethnic differences in resting metabolic rate, respiratory quotient and body temperature: a comparison of Africans and European Australians”, Eur J Nutr, 55 (5). 1831-1838. August 2016.
In article      View Article  PubMed
 
[12]  Hasson, R.E., Howe, C.A., Jones, B.L., Freedson, P.S., “Accuracy of four resting metabolic rate prediction formulas: effects of sex, body mass index, age, and race/ethnicity”, J Sci Med Sport, 14 (4). 344-351. July 2011.
In article      View Article  PubMed
 
[13]  Müller, M.J., Bosy-Westphal, A., Klaus, S., Lührmann, P.M., Neuhäuser-Berthold, M., et al., “World Health Organization equations have shortcomings for predicting resting energy expenditure in persons from a modern, affluent population: generation of a new reference standard from a retrospective analysis of a German database of resting energy expenditure”, Am J Clin Nutr, 80 (5). 1379-1390. November 2004.
In article      View Article  PubMed
 
[14]  Sabounchi, N.S., Rahmandad, H., Ammerman, A, “Best-fitting prediction equations for basal metabolic rate: Informing obesity interventions in diverse populations”, Int J Obes, 37 (10). 1364-1370. October 2013.
In article      View Article  PubMed  PubMed
 
[15]  Orozco-Ruiz, X., Pichardo-Ontiveros, E., Tovar, A.R., Torres, N., Medina-Vera, I., Prinelli, F., et al., “Development and validation of new predictive equation for resting energy expenditure in adults with overweight and obesity”, Clin Nutr, S0261-5614 (17)31395-X. November 2017.
In article      
 
[16]  Anjos, L.A., Wahrlich, V., Vasconcellos, M.T., “BMR in a Brazilian adult probability sample: the Nutrition, Physical Activity and Health Survey”, Public Health Nutr, 17 (4). 853-860. April 2014.
In article      View Article  PubMed
 
[17]  Kyle, U.G., Bosaeus, I., De Lorenzo, A.D., Deurenberg, P., Elia, M., Manuel Gómez, J., et al., “ESPEN. Bioelectrical impedance analysis – parte I: utilización in clinical practice”, Clin Nutr, 23 (6). 1430-1453. December 2004.
In article      View Article  PubMed
 
[18]  Wahrlich, V., Anjos, L.A., Going, S.B., Lohman, T.G., “Validation of the VO2000 calorimeter for measuring resting metabolic rate”, Clin Nutr, 25 (4). 687-692. August 2006.
In article      View Article  PubMed
 
[19]  Weir, J.D.V., “New methods for calculating metabolic rate with special reference to protein metabolism”, J Physiol, 109 (1-2). 1-9. August 1949.
In article      View Article  PubMed  PubMed
 
[20]  Harris, J.A., Benedict, F.G., “A biometric study of human basal metabolism. Proceedings of the National Academy of Sciences of the United States of America”, Proc Natl Acad Sci USA, 4 (12). 370-373. December 1918.
In article      View Article  PubMed
 
[21]  Schofield, W.N., “Predicting basal metabolic rate, new standards and review of previous work. Human Nutrition”, Clin Nutr, 39 (1). 5-41. 1985.
In article      
 
[22]  Owen, O.E., Kavle, E., Owen, R.S., Polansky, M., Caprio, S., Mozzoli, M.A., et al., “A reappraisal of caloric requirements in healthy women”, Am J Clin Nutr, 44 (1). 1-19. July 1986.
In article      View Article  PubMed
 
[23]  Owen, O.E., Holup, J.L., D’Alessio, D.A., Craig, E.S., Polansky, M., Smalley, K.J., et al., “A reappraisal of the caloric requirements of men”, Am J Clin Nutr, 46 (6). 875-885. December 1987.
In article      View Article  PubMed
 
[24]  Ireton-Jones, C.S., “Evaluation of energy expenditures in obese patients”, Nutr Clin Pract, 4 (4). 127-129. August 1989.
In article      View Article  PubMed
 
[25]  Mifflin, M.D., St Jeor, S.T., Hill, L.A., Scott, B.J., Daugherty, S.A., Koh, Y.O., “A new predictive equation for resting energy expenditure in healthy individuals”, Am J Clin Nutr, 51 (2). 241-247. February 1980.
In article      View Article  PubMed
 
[26]  Henry, C.J., Rees, D.G., “New predictive equations for the estimation of basal metabolic rate in tropical peoples”, Eur J Nutr, 45 (4). 1777-1785. April 1991.
In article      
 
[27]  Livingston, E.H., Kohlstadt, I., “Simplified resting metabolic rate-predicting formulas for normal-sized and obese individuals”, Obes Res, 13 (7). 1255-1262. July 2005.
In article      View Article  PubMed
 
[28]  Frankenfield, D., “Bias and accuracy of resting metabolic rate in non-obese and obese adults”, Clin Nutr, 32 (6). 976-982. December 2013.
In article      View Article  PubMed
 
[29]  Wahrlich, V., Teixeira, T.M., Anjos, L.A.D., “Validity of a population-specific BMR predictive equation for adults from an urban tropical setting”, Clin Nutr, 37 (1). 208-213. February 2018.
In article      View Article  PubMed
 
[30]  Flack, K.D., Siders, W.A., Johnson, L., Roemmich, J.N., “Cross-validation of resting metabolic rate prediction equations”, J Acad Nutr Diet, 116 (9). 1413-1422. September 2016..
In article      View Article  PubMed
 
[31]  Bedogni, G., Bertoli, S., Leone, A., De Amicis, R., Lucchetti, E., Agosti, F., et al., “External validation of equations to estimate resting energy expenditure in 14952 adults with overweight and obesity and 1948 adults with normal weight from Italy”, Clin Nutr, S0261-5614 (17). 31409-7. November 2017.
In article      
 
[32]  Poli, V.F., Sanches, R.B., Moraes, A.D., Fidalgo, J.P., Nascimento, M.A., Andrade-Silva, S.G., et al., “Resting energy expenditure in obese women: comparison between measured and estimated values”, Br J Nutr, 116 (7). 1306-1313. October 2016.
In article      View Article  PubMed
 
[33]  Willis, E.A., Herrmann, S.D., Ptomey, L.T., Honas, J.J., Bessmer, C.T., Donnelly, J.E., et al., “Predicting energy expenditure in young adults”, Obes Res Clin Pract, 10 (3). 304-314. May 2016.
In article      View Article  PubMed  PubMed