The effects of inflammatory bowel disease (IBD) on growth and development in childhood can be severe and profound. In as many as one third of all cases during childhood, growth retardation is seen before the manifestation of gastrointestinal symptoms (1). The causes of sometimes dramatic growth retardation include anorexia, active inflammation, and increased intestinal nutrient losses. Together, these complications of IBD could disturb energy balance, leading to potential weight loss if energy requirements are not met. A key component of energy expenditure, and hence energy requirements, is resting energy expenditure (REE).
Nevertheless, the effect of IBD on REE in childhood is poorly understood. Many children with IBD are anorexic, and studies have shown that REE is reduced in anorexia (2). However, some studies investigating REE in IBD have suggested that energy expenditure is raised (3,4), whereas others have suggested that REE is unchanged (1,5,6). Each scenario can be explained physiologically. It is well known that in anorexia, REE tends to decrease as a sparing mechanism to reduce the physiological impact of the decreased energy intake. However, the inflammatory process associated with the active disease is more than capable of increasing REE above that which would be expected. Indeed, it is possible that both processes may be occurring in any one individual at any one time.
Inherent in any analysis of variance in REE in IBD is the need to fully take into account differences in body size and/or body composition. This need is especially important during childhood, when both factors change significantly during growth and puberty.
Whereas the need to adjust for body size and/or body composition is intuitive, the correct way to achieve this is less obvious, and inappropriate adjustment may be a reason for discrepant findings in assessments of REE in IBD. The relationship between REE and body weight is usually strong; however, in most cases the correlation between an estimate of fat-free mass or body cell mass (BCM) and REE is stronger. This is easily explained by the fact that removing the largely metabolically inert adipose tissue from the analysis will lead to a strengthened relationship. This article considers how to appropriately adjust measurements of REE for differences in body composition between individuals with IBD.
PATIENTS AND METHODS
In our laboratory, as part of ongoing research into the effect of IBD on growth, development, and nutritional status, measurements of REE and body composition are routinely carried out on patients with newly diagnosed and existing Crohn disease and ulcerative colitis. The study was approved by the Human Research Ethics Committee of the University of Queensland and the Royal Children's Hospital, Brisbane. Informed written consent was obtained from the parents or guardian of the children involved, who, in turn, gave assent to be involved in the studies. Measurements of both REE and body composition were available in 41 children and adolescents ages 7.99 to 17.07 years. Of those patients, 30 had a diagnosis of Crohn disease, 9 had a diagnosis of ulcerative colitis, and 2 had a diagnosis of indeterminate colitis.
Resting Energy Expenditure
Resting energy expenditure was measured via the Deltatrac II metabolic cart (Datex Engstrom, Finland). The Deltatrac II metabolic cart was calibrated by gas calibration, pressure calibration, a respiratory quotient calibration, and a flow calibration. A pump drew ambient air through a canopy at a constant rate, and the gases were shunted through a mixing chamber. Resting energy expenditure and the respiratory quotient were derived from the concentration differences between inspired oxygen and expired carbon dioxide.
After resting for a minimum of 30 min, patients were required to lie in a supine position for 30 min with a canopy placed over their heads during the measurement. The majority of them had been fasting for a minimum of 3 h; however, 4 patients were receiving necessary continuous nasogastric feeding or hyperalimentation, which precluded a fasting period. All of these patients had respiratory quotients that were within the normal physiological range. The first 10 min of the measurements were used to ensure that the patient was settled and that the air inside the canopy had equilibrated, and the next 20 min were used to calculate REE.
Body Composition: BCM
Many methods have been used to derive an estimate of body composition in children with IBD, including bioelectrical impedance, total body water, total body potassium (TBK), and skinfold thickness (1,7). In this study we report TBK measurements to derive an estimate of BCM. This method was chosen above other available methods that effectively derive a value for fat-free mass, inasmuch as fat-free mass contains extracellular water, which is known to be expanded in volume in cases of anorexia and Crohn disease (2,6). Using BCM therefore negates the effect of an expanded extracellular water space on estimates of body composition.
The measurement of BCM was calculated via the assessment of TBK using a shadow shield whole-body counter (Accuscan, Canberra Industries, MA). This equipment uses 3 sodium iodide crystal scintillation detectors placed above a movable bed. TBK was measured twice over an 18-minute scan, during which all of the individual's body passed under the detectors. In our laboratory TBK is measured in grams; however, this can be easily converted to millimoles. BCM was then calculated by use of the equation described by Wang et al (8), which has recently been validated for use in children (9):
Equation (Uncited)Image Tools
where 0.00918 is a constant that represents the relationship between the TBK in millimoles found in 1 kg of BCM.
Expressing REE as kilocalories per kilogram of BCM is the equivalent to the expression kcal/kg BCM1. The power function (p) here is 1, but it is certainly not necessarily the most appropriate. The most appropriate value for p can be easily calculated. The simplest approach is to consider the correlation between the natural logarithms of the index REE/BCM and BCM. This index can be rearranged as follows:
Equation (Uncited)Image Tools
The use of logarithms allows the index to be expressed as a linear function of log (REE) and log (BCM), which is then suitable for analysis using linear regression. In this case the value p is the slope of the regression line relating log (REE) to log (BCM) and is the best power function to which BCM should be raised to adjust for differences in BCM between individuals. Hence, raw REE values should be expressed as kcal/kg BCMp. Whereas the data reported in this manuscript were log-transformed to solve for p, all of the REE and BCM data reported in the results section are not log-transformed.
The basic physical characteristics of the patients studied are shown in Table 1. Height, weight, and body mass index z scores were calculated relative to US Centers for Disease Control and Prevention 2000 growth data (10). Figure 1 shows the relationship between REE and BCM before any adjustments to REE were made. As expected, a significant positive correlation was found (r = 0.63; P < 0.001). When REE is expressed per kilogram of BCM (Fig. 2) and plotted against BCM, an inverse relationship emerges. The significant negative correlation (r = −0.62; P < 0.001) found here indicates that simply dividing REE by BCM (kg) does not remove the influence of differences in body composition in these patients.
The results of the log-log regression are shown in Table 2; p has a value of 0.49 with a standard error of 0.10. Thus, a numerically convenient and yet statistically valid power to which to raise BCM to adjust for differences in BCM between individuals is 0.5, equivalent to √BCM. Figure 3 illustrates that when REE is expressed in kcal/kg BCM0.5, the effects of differences in body composition between individuals is removed. Interestingly, when this approach is applied to the relationship between REE and body weight—that is, REE is divided by body weight0.5 (kg)—0.5 remains a statistically valid power to adjust for weight differences between individuals (r = −0.1; P = ns).
This study clearly shows that to adjust for differences in body composition between children with IBD, BCM can be raised to the power of 0.5. It should be noted that the power of 1.0, which is used when using the expression REE/kg BCM, is not within 2 standard errors of the calculated P value of 0.49 and therefore is not statistically valid. A power of 0.5 has been shown to best adjust for body composition differences and also body weight differences in several other animal and human studies (11–18) and these studies have been discussed in more detail elsewhere (19).
We are not the first to illustrate the fact that dividing A by B does not necessarily adjust A for B. The fallacy of this notion goes back >50 years (20). Specifically, in relation to this issue in metabolic studies in IBD, the problem was discussed some 15 years ago by Kushner and Schoeller (21), who elegantly showed the negative relationship between REE expressed relative to fat-free mass and fat-free mass when they studied a cohort of young adults with IBD and a control group. Their study correctly concluded that the increased REE previously reported in patients with IBD is at least partially due to the expected elevation of REE on a per-kilogram body weight or fat-free mass basis because of this mathematical bias.
The relationship between REE/BCM and BCM shown in Figure 2 can potentially explain some of the previous findings in the literature. For example, Azcue et al (6) concluded that even though their cohort of children with Crohn disease was malnourished, no adaptation of REE occurred. Specifically, mean REE/kg lean body mass (LBM) was 40.3 kcal/kg LBM and 39.1 kcal/kg LBM in a group of 24 children with Crohn disease and 22 healthy control children. However, the mean LBM was significantly different between the 2 groups (eg, 30.8 kg vs 44.3 kg, respectively, with LBM being calculated by measurement of total body water). Thus, it would be expected that the children with Crohn disease should have shown a higher REE expressed as REE/kg LBM. The fact that this difference did not occur suggests strongly that the cohort with Crohn disease had, in fact, reduced their REE in the face of anorexia.
Another good example relates to the work of Barot et al (3). In their study of young adults, 3 groups were identified. The first group consisted of healthy control individuals, the second group had IBD with a body weight of >90% of ideal body weight, and finally a group with IBD had a body weight that was <90% of ideal body weight. The mean body weights of the 3 groups were 66.4, 65.0, and 48.5 kg, respectively. Assuming our model shown in Figure 2, we may expect, therefore, the REE/kg to be similar in the control group and Crohn cohort >90% ideal body weight and for both of these groups to have a mean REE/kg lower than that found in the <90% ideal body weight group. Indeed, this was the case, with REE/kg being 21.2, 21.2, and 26.4 kcal/kg, respectively. The difference between the <90% ideal body weight group and the other 2 groups was statistically significant. Body composition was not assessed.
The use of 0.5 as an appropriate power to adjust measures of body composition or body weight is much less than that proposed across species. In a seminal publication in 1947, Kleiber (22) showed that metabolic rate (REE) was best expressed relative to body weight raised to the power of 0.75 in a large group of mammals ranging in weight from 25 g to 600 kg. It should, however, be remembered that across species the correlation between metabolic rate and body size, whether measured by body weight or by body composition, is high and close to 1. In humans that correlation is lower, usually around 0.6 to 0.7, as was found in the current dataset. In our model the slope of a regression line is equal to the ratio of the standard deviation of the X and Y variables, in this case BCM and REE, both logged, multiplied by the correlation between them. Thus, if we accept that the ratio of the standard deviations is 0.75, as suggested by Kleiber (22), and the correlation is, for example, 0.65, the best slope for prediction purposes will be close to 0.5 (0.75 × 0.65).
Clearly, there is a need to adjust for differences in body composition, or at the very least body weight, in any metabolic research. Of course, this translates to studies in children with IBD. It is clear that simply dividing X by Y to adjust for differences in Y is not correct. We suggest that raising BCM, in this case Y, to the power of 0.5 is both a numerically convenient and statistically valid way of achieving this aim, and this approach has been shown to be appropriate in other metabolic studies of health and disease. Under circumstances in which the measurement of BCM is not available, raising body weight to the power of 0.5 remains appropriate. The important issue of whether REE is changed in cases of IBD can then be appropriately addressed.
1. Cormier K, Mager D, Bannister L, et al
. Resting energy expenditure in the parenterally fed pediatric population with Crohn's disease. JPEN J Parenter Enteral Nutr 2005; 29:102–107.
2. Vaisman N, Rossi MF, Goldberg E, et al
. Energy expenditure and body composition in patients with anorexia nervosa. J Pediatr 1988; 113:919–924.
3. Barot LR, Rombeau JL, Feurer ID, et al
. Calorie requirements in patients with inflammatory bowel disease. Ann Surg 1982; 195:214–218.
4. Zoli G, Katelaris PH, Garrow J, et al
. Increased energy expenditure in growing adolescents with Crohn's disease. Dig Dis Sci 1996; 41:1754–1759.
5. Chan AT, Fleming CR, O'Fallon WM, et al
. Estimated versus measured basal energy requirements in patients with Crohn's disease. Gastroenterology 1986; 91:75–78.
6. Azcue M, Rashid M, Griffiths A, et al
. Energy expenditure and body composition in children with Crohn's disease: effect of enteral nutrition and treatment with prednisolone. Gut 1997; 41:203–208.
7. Varille V, Cézard JP, de Lagausie P, et al
. Resting energy expenditure before and after surgical resection of gut lesions in pediatric Crohn's disease. J Pediatr Gastroenterol Nutr 1996; 23:13–19.
8. Wang Z, St-Onge MP, Lecumberri B, et al
. Body cell mass: model development and validation at the cellular level of body composition. Am J Physiol 2004; 286:123–128.
9. Wang Z, Heshka S, Wang J, et al
. Body cell mass: validation of total body potassium prediction model in children and adolescents. Int J Body Compos Res 2005; 3:153–158.
10. Ogden CL, Kuczmarski RJ, Flegal KM, et al
. Centers for Disease Control and Prevention 2000 growth charts for the United States: improvements to the 1977 National Center for Health Statistics version. Pediatrics 2002; 109:45–60.
11. Davies PS, Cole TJ, Lucas A. Adjusting energy expenditure for body weight in early infancy. Eur J Clin Nutr 1989; 43:641–645.
12. Lawrence M. Predicting energy requirements: is energy expenditure proportional to the BMR or to body weight? An analysis of data collected in rural Gambian women. Eur J Clin Nutr 1988; 42:919–927.
13. Millward DJ, Garlick PJ. The energy cost of growth. Proc Nutr Soc 1976; 35:339–349.
14. Kielanowski J. Variation in heat production in growing pigs: some observations on the relationship between feed intake and heat production in pigs fed barley and skim milk. Publ Eur Assoc Anim Prod 1969; 12:289–297.
15. Thorbeck G. Studies on the energy metabolism of growing pigs. Publ Eur Assoc Anim Prod 1969; 12:281–289.
16. Brierem K. Der Energieumsatz bei den Schweinen. Tierernahrang 1939; 11:487–528.
17. Davies PSW, Wells JC, Lucas A. Adjusting milk intake for body size in early infancy. Early Hum Dev 1994; 36:61–67.
18. Zakeri I, Puyau MR, Adolph AL, et al
. Normalization of energy expenditure data for differences in body mass or composition in children and adolescents. J Nutr 2006; 136:1371–1376.
19. Davies PSW, Cole TJ. The adjustment of measures of energy expenditure for body weight and body composition. Int J Body Compos Res 2003; 1:45–50.
20. Tanner JM. Fallacy of per weight and per surface area standards and their relation to spurious correlation. J Appl Physiol 1949; 2:1–15.
21. Kushner RF, Schoeller DA. Resting and total energy expenditure in patients with inflammatory bowel disease. Am J Clin Nutr 1991; 53:161–165.
22. Kleiber M. Body size and metabolic rate. Physiol Rev 1947; 27:511–541.
© 2007 Lippincott Williams & Wilkins, Inc.