| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research Communication |
1 From the Center for Nutrition and Department of Family and Community Medicine, Meharry Medical College, Nashville, TN (MSB); the Department of Medicine (MSB and KYC) and General Clinical Research Center (DB), Vanderbilt University Medical Center, Nashville, TN; and St Jude Hospital, Memphis (WCW).
2 Supported in part by NIH grant HBL 003530 (to MSB), General Clinical Research Center grants RR-00095 (to Vanderbilt University) and RR-1179204 (to Meharry Medical College), Clinical Nutrition Research Unit grant DK-26657 (to Vanderbilt University), and a Meharry Center of Excellence grant (to MSB). 3 Address reprint requests to MS Buchowski, Center for Nutrition, Meharry Medical College, 1005 DB Todd Boulevard, Nashville, TN 37208. E-mail: maciej.buchowski{at}mcmail.vanderbilt.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONREFERENCES |
|---|
Objective: Our objective was to develop a clinically useful equation to estimate resting energy expenditure (REE) in adolescents with SCA.
Design: REE and other components of total energy expenditure were measured in adolescents with SCA (n = 37) and in control subjects (n = 23) for 24 h in a whole-room indirect calorimeter. Multiple linear regression analysis was used to describe the relations of REE with independent variables such as sex, weight, height, fat-free mass, fat mass, age, and hemoglobin concentration in adolescents with SCA. The Bland-Altman comparison technique was used to compare values predicted by existing equations with measured REE values.
Results: Mean (±SD) measured REEs were 7746 ± 974 and 6332 ± 869 kJ/d in the male and female subjects with SCA, respectively, and these values were 16% higher than those in the healthy control subjects. Standard equations underestimated REE by 12% (P 
0.001) in adolescents with SCA and overestimated REE by 4% in control subjects (P = 0.01–0.29). Several REE regression equations for SCA were developed. The most parsimonious equation for males was REE (kJ/d) = 5461 + 77.7·weight (kg) - 233.2·hemoglobin (g/dL), or REE (kcal/d) = 1305 + 18.6·weight (kg) - 55.7·hemoglobin (g/dL). For females, the most parsimonious equation was REE (kJ/d) = 4603 + 55.6·weight (kg) - 126.2·hemoglobin (g/dL), or REE (kcal/d) = 1100 + 13.3·weight (kg) - 30.2·hemoglobin (g/dL).
Conclusion: The new equations have better agreement with the basal metabolic needs of adolescents with SCA than do standard predictive equations.
Key Words: Resting energy expenditure • sickle cell anemia • prediction equations • adolescents • African Americans
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONREFERENCES |
|---|
50% of the higher resting metabolic rate in patients with SCA could be accounted for by the enhanced energy cost of total protein turnover (7), including 20% caused by higher bone protein turnover (2). The clinical consequences of this hypermetabolic state are the slower growth rate and delayed sexual development often seen in children and adolescents with SCA (8). Therefore, it is important that individual requirements are determined and that an optimal amount of energy is provided to children and adolescents with SCA.
In 1981, the FAO/WHO/UNU Committee on Protein and Energy Requirements recommended that daily energy requirements be estimated by using measurements of energy expenditure (9). This recommendation substantially underscores the importance of measuring REE, which is defined as the energy required to sustain the body’s functions at rest and which accounts for 65% of total daily energy expenditure (10). REE is calculated from the resting metabolic rate, which equals the basal metabolic rate only if the measurement is performed in the postabsorptive state and under precisely specified conditions, which include the position of the subject during measurement, temperature, humidity, and degree of light and sound in the room at the time of measurement, duration and quality of sleep or rest interval before the measurement period, and duration of the measurement (10). Such rigorous conditions are not required in the measurement of REE, which is adequate in most clinical applications. Moreover, as pointed out by others, accurate measurement of total daily energy requirements requires that REE be recorded over a protracted period (11–13). Nevertheless, measurement of REE requires specialized equipment, and it is not practical to obtain an REE measurement on every patient. Therefore, several equations and modified equations were developed to predict age- and sex-adjusted basal or resting energy expenditure for children and adolescents by using variables such as sex, age, height, weight, and body surface area (14–17). Numerous studies showed that the major factor explaining the variation in metabolic rate in humans is fat-free mass (FFM), a heterogeneous component that can be partitioned into muscle mass and nonmuscle mass (10). It was also shown that FFM is lower in children and adolescents with sickle cell disease than in healthy control subjects (4, 6). Because the predictive equations commonly used in clinical practice were generated by using healthy populations and did not take into consideration the hypermetabolic state observed in SCA and differences in body composition, it is reasonable to assume that these equations would not precisely predict REE in subjects with SCA. Kopp-Hoolihan et al (18) showed that standard equations are unreliable in patients with SCA and proposed that clinically useful formulas for predicting REE in patients with SCA should be developed.
The objective of our study was to develop clinically useful equations for the estimation of REE in adolescents with SCA. The null hypothesis was that in adolescents with SCA, REE calculated by using existing standard equations would not differ from REE measured in a previously validated whole-room indirect calorimeter under strictly controlled conditions. To test this hypothesis, we measured REE and other components of total energy expenditure in 37 adolescents with SCA and in 23 control subjects during a 24-h stay in a room calorimeter and compared the results with basal energy expenditure values calculated from standard equations (9, 19, 20). Furthermore, we assumed that if the null hypothesis were incorrect, our experimental data would be sufficient to generate a new simple equation for predicting REE in adolescents with SCA. Simple potential predictive factors such as weight, FFM, fat mass, height, age, and hemoglobin concentration were chosen for testing in this study.
SUBJECTS AND METHODS
Subjects
A group of 37 African American adolescents (14–18 y of age) with homozygous SCA were identified and screened for participation in the study at the sickle cell clinics of the Comprehensive Sickle Cell Center at Meharry Medical College in Nashville, TN, and of the Pediatric Hematology Center of Memphis. The group included 18 males and 19 females. Each participant’s hemoglobin genotype was determined by using standard electrophoretic methods (21) to confirm the diagnosis of SCA, in which both ß globin genes code for hemoglobin S (22). Additionally, 23 African American adolescents who did not carry the hemoglobin S gene or have any other hemoglobinopathy were matched for sex, age, FFM, and fat mass.
Before participating in the study, subjects provided a medical history and underwent a complete physical examination. The participants were free of any metabolic, skeletal, hepatic, or renal dysfunction as confirmed by blood tests. They were not taking drugs known to affect energy metabolism and were nonsmokers. Female subjects were not pregnant as determined by a serum pregnancy test and were studied between days 3 and 12 after the onset of menses (follicular phase) to eliminate the influence of menstrual function on energy expenditure (23, 24). Subjects with SCA were studied in the steady state; ie, they were not experiencing a sickle cell crisis during the study and had not experienced a painful crisis for 28 d before the study.
Prospective participants received written and verbal information about the nature and purpose of the study, and those who agreed to participate signed an informed consent form. The consent form was approved by the Institutional Review Boards of both Meharry Medical College and Vanderbilt University School of Medicine for procedures to be performed at the Vanderbilt University General Clinical Research Center.
Study protocol
All participants reported to the General Clinical Research Center after a 10-h overnight fast. After initial admission, they were transferred to the whole-room indirect calorimeter, where they were asked to engage in a 24-h protocol involving standard daily activities and a 2-h nonintensive exercise protocol. Oxygen consumption (
O2) and carbon dioxide production (CO2) and the flow rate, temperature (inside and ambient), barometric pressure, and humidity of the air were sampled 60 times/s and integrated at the end of each minute to calculate energy expenditure (25). The participants were free to view television, read, write, walk, and perform personal care activities and to exercise by using the stationary bicycle, step platform, and aerobic tapes as much as they would in their normal daily routine. They received 5 meals but were not allowed to eat or drink after 2100 and were asked to go to bed between 2130 and 2200.
Methods
Anthropometry
Body weight was measured to the nearest 0.05 kg with a digital scale. Height was measured to the nearest 1 cm with a stadiometer. All anthropometric measurements were performed by one investigator (MSB).
Body composition
Fat mass and FFM were determined by hydrodensitometry. The subjects were weighed underwater, and their residual lung volumes were measured by using the nitrogen dilution technique while the subjects were submerged in water to chest level (26). Percentage of body fat was calculated from body density by using Schutte’s equation (27), and fat mass and FFM were calculated from body mass.
Resting energy expenditure
REE (kJ/min) was defined as the average baseline energy expenditure excluding physical activity and sleeping during a 24-h stay in a room calorimeter. Physical activity was detected by a large precision force platform, and energy expenditure above baseline during periods of physical activity was subtracted for REE calculation (28). An automatic algorithm in data processing was used to detect the baseline level of energy expenditure during these nonsleeping, nonexercising, and nonmoving periods. The lowest 60-min period selected by a computer program by sorting was used to obtain the REE. The detailed methodology of the room calorimeter was previously reported (25, 28). Regression models of REE based on the predictors FFM and fat mass were used rather than division of REE by FFM because division by FFM does not account for nonzero intercepts typically observed in these regressions (29).
Prediction equations for basal metabolic rate
The equations formulated by Harris and Benedict (19), FAO/WHO/UNU (9), and Schofield (20) were used to predict the REE of the subjects (Table 1
). The predictive power of existing equations was assessed by comparing measured and predicted REE values.
|
level of 0.005, we calculated that 10 participants in each sex subgroup would provide 90% statistical power to detect differences between the adolescents with SCA and the control adolescents within each sex. We used a conservative number of 
12 adolescents in each SCA sex subgroup because of uncertainty regarding the effect of other variables. The Bland and Altman method (30) was used to assess agreement between the REE values predicted by the equations and the REE values measured by whole-body indirect calorimetry. Differences between predicted and measured REE were plotted against the averages of predicted and measured REE. Regression analysis was used to test the relation between these differences and means. If the slope relating the differences and means was significant, the 95% limits of agreement were estimated as 2 SD of the estimate around the regression line. Differences between subjects in continuous variables were evaluated by using independent-sample, two-tailed t tests. Multiple linear regression analysis was performed on REE by using a backward elimination approach. The covariates that were considered for inclusion in the modeling were SCA status (0 = healthy control subject, 1 = subject with SCA), sex (0 = male, 1 = female), FFM, fat mass, age, hemoglobin concentration, and plausible interactions among these. A multiple linear model (analysis of variance) was used to test for differences between estimated and measured REE values. Plots of REE estimated from the new equations against measured REE were used to assess the general agreement between paired values of estimated and measured REE. The constancy of estimation across the range of predicted REE values was assessed by plotting these values against residuals. A P value < 0.05 was used for inclusion of terms in the regression and to indicate statistical significance. Statistical analyses were performed with SPSS version 10.0 (SPSS Inc, Chicago).
RESULTS
Subject characteristics
Descriptive data for the study participants are shown in Table 2. Body weight, age, FFM, and fat mass were not significantly different between the adolescents with SCA and the control adolescents. Mean (±SD) hemoglobin concentrations were significantly (P < 0.001) lower in the subjects with SCA than in the control subjects (8.9 ± 1.8 compared with 13.3 ± 1.1 g/dL). Plasma concentrations of albumin and thyroid-stimulating hormone and red blood cell concentrations of folate were not significantly different between the subjects with SCA and the control subjects. There were no significant differences between the subjects with SCA and the control subjects in plasma concentrations of testosterone in the males and of estradiol in the females.
|
.
|
. The prediction equations underestimated measured REE in the SCA group and overestimated it in the control group. Detailed comparisons between values predicted by using 3 standard equations and measured values are shown in Figure 2. The mean (±SD) bias in REE in adolescents with SCA was negative for all equations and ranged from -771 ± 479 kJ/d for the FAO/WHO/UNU equation by weight (Figure 2A) to -930 ± 521 kJ/d for the Harris-Benedict equation (Figure 2B). For the FAO/WHO/UNU equation by weight and height and the Schofield equation by weight, the mean biases were -834 ± 467 kJ/d (P = 0.101, r = -0.274) and -724 ± 482 kJ/d (P = 0.627, r = -0.083), respectively. In the control subjects, the mean bias was positive for all equations and ranged from 171 ± 367 kJ/d for the FAO/WHO/UNU equation by weight to 82 ± 360 kJ/d for the Harris-Benedict equation. Ratios of predicted to measured REE expressed in percentages are shown in Table 4. On average, the prediction equations underestimated the measured values for adolescents with SCA by 11.30%, and the differences in REE values were significant (P < 0.001) in both the males and the females for all equations. A comparison between measured REE and REE obtained by using the FAO/WHO/UNU equation by weight is shown in Figure 3 (top). Overestimation of REE in the control subjects ranged from 1.8% in the Harris-Benedict equation to 5.4% in the Schofield equation by weight. Detailed analysis showed that the bias was significant in the female control subjects for the FAO/WHO/UNU equation by weight (P = 0.008; t test), the Schofield equation by weight (P = 0.003; t test), the Schofield equation by weight and height (P = 0.035; t test), and the Harris-Benedict equation by weight (P = 0.036; t test). No significant bias was found in the male control subjects (P > 0.05 for all; t test).
|
|
|
|
. FFM and body weight were the most important alternative variables in developing regression equations for calculation of REE in males (R2 = 0.805 and 0.702, respectively) and females (R2 = 0.757 and 0.825, respectively).
|
0.0001, overall R2 = 0.879) was
 | (1) |
0.0001, overall R2 = 0.760) for males were
 | (2) |
 | (3) |
0.0001, overall R2 = 0.855) were
 | (4) |
 | (5) |
(bottom).
DISCUSSION
In this study, we examined the validity of standard equations used in clinical practice to predict REE in adolescents with SCA and found that these equations underestimated REE by 10–15%. We propose new, practical, simple equations based on 24-h measurements of REE in a whole-room indirect calorimeter.
We chose to evaluate the prediction equations of Harris and Benedict (19) and of FAO/WHO/UNU (9) because they are the most commonly used equations in clinical settings. We also examined the equations of Schofield (20) because they were developed for adolescents of similar age to that of our study population. These equations were examined previously in independent samples (31–35). For example, the Harris-Benedict equations were shown to overestimate the basal metabolic rate in healthy female adolescents by 12% and to underestimate it by 12% in healthy male adolescents (31). Dietz et al (16) reported good agreement between the results predicted by FAO/WHO/UNU equations and the measured REE values of 54 adolescents. Finan et al (32) concluded that the FAO/WHO/UNU equations are reasonably accurate for predicting REE in healthy children. In contrast, these and other prediction equations overestimated REE in 33 Italian children (17). Also, Stallings and Pencharz (31) showed that the FAO/WHO/UNU equation underestimated REE by 9 ± 13% (
± SD) in obese adolescents.
On the basis of the comparisons shown in Tables 3 and 4, it appears that none of the equations are appropriate for estimation of REE in adolescents with SCA. Although some of these equations predicted average REE values in the control subjects similar to the measured values, detailed analysis in adolescents with SCA indicated a significant underestimation of 9–13% with the FAO/WHO/UNU equation and of 11–15% with the Harris-Benedict equation. Previously, we (1, 2, 7) and others (3, 4) showed that REE is 15–20% higher in patients with SCA than in healthy control subjects. In the present study, the difference in REE between the adolescents with SCA and healthy control adolescents (adjusted for sex, FFM, fat mass, and age) averaged 21%. Therefore, it was not surprising that the concurrently used prediction equations underestimated measured REE. Detailed comparisons (Figure 2) indicated that REE in individual adolescents with SCA could be underestimated by as much as 1700 kJ/d with the HarrisBenedict equation and 1410 kJ/d with the FAO/WHO/UNU equation.
Only 10% of the REE values predicted by standard equations were within 2 SD of the measured values. Several factors may account for these predicted REE values in adolescents with SCA. Some equations are not appropriate for African American adolescents (12, 33). More importantly, however, all standard equations were defined for healthy individuals with normal energy metabolism. Metabolic perturbations in SCA include elevated cardiac output necessitated by the low oxygen carrying capacity of the blood (3, 5) and increased protein turnover (1, 3, 7). We estimated that 50% of the increase in resting metabolic rate is due to increased protein synthesis (1) and that <20% of this increase is caused by greater hemoglobin synthesis (7). Hemoglobin synthesis is related to the degree of anemia. Our adolescents with SCA had chronic hemolytic anemia, a typical feature of SCA (36). Although the degree of anemia is extremely variable among patients with SCA, the hemoglobin concentration is relatively constant in any given patient. Therefore, it is reasonable to predict that changes in hemoglobin concentration would cause changes in energy requirements. Indeed, including hemoglobin concentration increased the precision of our predictive equations, and hemoglobin concentrations explained 10% and 12% of the variation in REE in males and females, respectively. This inverse relation between REE and hemoglobin concentration in SCA may be related to increased cardiac output with tachycardia, increased myocardial oxygen demand, high protein turnover, and changes in cell metabolism due to hypoxia (37–40).
In the control subjects, REE was overestimated by 100–400 kJ/d depending on the equation. Several explanations have been offered for the lack of concordance between observed and predicted REE. In a study by Wong et al (15), differences in lean body mass, age, and sexual maturation were factors leading to overestimation of REE in healthy African American children and adolescents. A similar conclusion was reached by Morrison et al (33) in a cross-sectional study of 47 African American and 56 white girls aged 6–16 y. Therefore, we can assume that these factors could have resulted in an overestimation of REE in the control adolescents in the present study.
In the formulation of our new equations, we found that body weight predicted 74% and 80% of the variation in REE in the males and females, respectively. This finding is consistent with those of other studies that reported that weight was the most important variable in predicting REE in girls and boys (11). FFM may be considered a better predictor of REE in adolescents (14, 15, 34), and it explained 83% of the variation in the present study. However, the limited improvement in R2 and potential problems with the accurate measurement of FFM in clinical practice make weight a more appropriate single variable to estimate REE in adolescents with SCA. In some studies (17, 28, 35), including height improved REE estimation in adolescents. In the present study, however, the addition of height to the equations did not contribute significantly to the accuracy and precision of the predicted REE values. Regression equations derived by combining the 2 sexes were less accurate than were single-sex equations. In particular, the REE of females with SCA was overestimated, probably because of the differences between the sexes shown in Table 2 and Figure 1. Consequently, single-sex equations are recommended.
Our data must be interpreted with caution. The predictions of basal energy expenditure in our new equations are dependent on body weight, which in adolescents with SCA may not be an accurate reflection of metabolically active tissue. Part of the variation in REE cannot be explained by body size or body composition even in healthy adolescents (10, 41, 42). The nutritional status of the subjects with SCA at the time of REE measurement, use of medications and other therapies, and hormonal status may also affect REE. Although we attempted to control for most of these conditions, not all clinically relevant variation could be eliminated. Nevertheless, our proposed prediction equations were examined for both accuracy and precision. Accuracy was assessed by the slope and intercept of the plot of measured REE versus predicted REE (Figure 3). The precision of the equations was assessed by using R2 values (Table 5). An alternative approach is to use correction factors for existing standard equations. We calculated that the correction factors for the FAO/WHO/UNU and Harris-Benedict equations are 1.14 and 1.19, respectively, for male adolescents with SCA and 1.11 and 1.13, respectively, for female adolescents with SCA. Both these approaches could lead to a valid estimation of energy needs. Using correction factors, however, does not account for individual variation in REE caused by variation in hemoglobin concentrations.
Finally, our study population was limited to 14–18-y-old adolescents. Although including age in our analyses may have explained some of the association between tested variables and REE, including prepubertal and pubertal adolescents in future studies would allow formulation of more a universal predictive equation.
In conclusion, the REE data collected in the present study were used to produce regression equations for the estimation of REE in 14–18-y-old males and females with SCA. Standard prediction equations underestimate REE by 10–15% in adolescents with SCA. On an individual basis, REE may be underestimated by 800–2000 kJ/d (190–525 kcal/d) depending on the equation. Although accurate, measurement of REE by using indirect calorimetry is cumbersome, expensive, and unavailable in many clinical situations. By contrast, body weight and hemoglobin concentration are 2 easily measurable predictive factors of REE. We recommend that these new simple equations be used in clinical practice for predicting energy requirements in adolescents with SCA.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  G. S. Metsios, A. Stavropoulos-Kalinoglou, V. F. Panoulas, Y. Koutedakis, A. M. Nevill, K. M. J. Douglas, M. Kita, and G. D. Kitas New resting energy expenditure prediction equations for patients with rheumatoid arthritis Rheumatology, April 1, 2008; 47(4): 500 - 506. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
; n = 18; R2 = 0.82) and female (
; n = 19; R2 = 0.77) adolescents with sickle cell anemia (SCA) and in healthy male (
; n = 12; R2 = 0.60) and female (•; n = 11; R2 = 0.41) control adolescents. On average, REE was 1377 kJ/d higher in the adolescents with SCA than in the control adolescents after adjustment for sex and FFM (P < 0.001). On average, REE was 576 kJ/d lower in the females than in the males after adjustment for SCA status and FFM. No interactions involving sex were significant. A positive relation between REE and FFM was observed in both the adolescents with SCA and the healthy control adolescents after adjustment for sex. The overall regression equation (R2 = 0.846) was REE (kJ/d) = 2366 + 82.94 · FFM - 576 · sex + 1377 · SCA. The relation between REE and FFM in the male adolescents with SCA (upper solid line) was significantly greater than that in the male control adolescents (lower solid line) (P < 0.001). The relation between REE and FFM in the female adolescents with SCA (upper dashed line) was significantly greater than that in the female control adolescents (lower dashed line) (P < 0.001).
) and adolescents with sickle cell anemia (SCA) (