Acta Diabetol (2000) 37:77-81 Springer-Verlag 2000 ORIGINAL A. De Lorenzo A. Andreoli S. Bertoli G. Testolin G. Oriani P. Deurenberg Resting metabolic rate in Italians: relation with body composition and anthropometric parameters Received: 19 January 2000 / Accepted in revised form: 3 July 2000 Abstract The objectives of this study were to obtain values for resting metabolic rate in Italians in relation to parameters of body composition, and to compare them to predicted values using the FAO/WHO/UNU equation. We performed a cross-sectional observational study of 131 healthy subjects (46 males and 85 females) at the Human Nutrition Unit, University Tor Vergata, Rome. Body composition was determined by dual energy X-ray absorptiometry (DXA) and resting metabolic rate was calculated using the Weir formula. Resting metabolic rate was 1865 ± 234 kcal/day in males and 1354 ± 154 kcal/day in females. These values decreased slightly with age. The relationships with weight and age were stronger than that with lean mass from DXA as independent variables in multiple regression analysis. Mean resting metabolic rates predicted with FAO/WHO/UNU and Harris-Benedict formula were not A. De Lorenzo ( ) A. Andreoli Human Nutrition Unit University Tor Vergata Via di Tor Vergata 135, I-00173 Rome, Italy and Scientific Institute S. Lucia, Rome A. De Lorenzo S. Bertoli G. Testolin International Centre for the Assessment of Body Composition Department of Food and Microbiological Sciences and Technologies University of Milan, Milan, Italy G. Oriani Department of Animal and Plant Sciences University of Molise, Campobasso, Italy P. Deurenberg Division of Human Nutrition and Epidemiology Wageningen Agricultural University Wageningen, The Netherlands significantly different from measured values except for the Harris-Benedict value for females (p < 0.01). Individual differences between measured and predicted values were notably high. The measured values were higher than those reported in the literature. The prediction of resting metabolic rate is more accurate with simple anthropometric parameters than with fat-free mass obtained by DXA. The individual error in the predicted values can be so high that for individual use a measured value is preferred over an estimated value. Key words Resting metabolic rate Body composition Italians Humans Introduction Resting metabolic rate (RMR) is an important parameter in the assessment of nutritional status in a patient: it is used, for example, to calculate the energy requirements of a patient who needs parenteral or enteral nutrition [1, 2]. In addition, information on resting energy expenditure is necessary to calculate energy needs at a population level. For this, the Food and Agriculture Organization, World Health Organization and United Nations-UNICEF (FAO/WHO/ UNU) have established formulas for predicting the resting metabolic rate [3]. Several studies have validated these prediction formulas for RMR [4, 5], but until now no such study has been performed in Italy. Generally, there is a lack of data on metabolic rate in the Italian population. There are some rather old, extensive publications from the 1930s [6-12], in which RMR was studied in relatively young population groups. In addition, the study subjects typically had an active life style (e.g. soldiers, nurses). This could explain the relatively high metabolic rate reported for Italians, as mentioned by Hayter and Henry [4].
78 A. De Lorenzo et al.: Resting metabolic rate in Italians The aim of the present paper is to provide data on RMR in Italian subjects and to relate the measured RMR with estimated values using prediction equations from the literature, especially those of FAO/WHO/UNU [3] and Harris and Benedict [13]. In addition, the relation of RMR with body composition and anthropometric parameters was studied. differences in parameters between males and females and between different age groups. Paired t test was used to test differences between measured and predicted values. Stepwise multiple regression was used to explore the relationship of RMR with other variables. The Bland and Altman [17] technique was used to study the difference between measured and predicted values. Correlations were determined as Pearson s product-moment correlations. Values are presented as mean and standard deviation (SD). Subjects and methods We studied 131 healthy subjects, including 46 males and 85 females. The patients were disease-free and were not taking any medications known to affect body composition or resting metabolic rate. The subjects came to the Human Nutrition Unit in the early morning after an overnight fast and were requested to refrain from any unnecessary physical activities. The study protocol, approved by the Ethical Committee of the University Tor Vergata, was performed in accordance with the standards laid down in the 1964 Declaration of Helsinki. Written informed consent was obtained from each subject. The subjects were measured in the fasting state in the morning, after voiding. For each subject, anthropometric and body composition measurements were taken, and resting metabolic rate was determined on the same morning. Body weight was measured to the nearest 0.1 kg in all subjects clothed in only underwear, and height was measured without shoes to the nearest 0.1 cm. Body mass index (BMI) was calculated as weight/height 2 (kg/m 2 ). At 8.30 a.m. the subjects rested supine for 25-30 min in a quiet room with an ambient temperature of 22 C. Afterwards, oxygen consumption (VO 2) and carbon dioxide production (VCO 2) were assessed by an open circuit indirect calorimeter (Sensormedics 2900, California, USA) for 45 min. The values of accuracy for the measurement of VO 2 and CO 2 were respectively 10.6 ml/min (4.3%) and 7.1 ml/min (4.7%) while those of precision were 7.6 ml/min (3.1%) and 6.0 ml/min (3.9%) respectively. Calibration of the calorimeter was conducted following the manufacturer s instructions. Particularly, two different cylinders containing certified mixtures (SIAD, Rome) were employed. One mixture contained 26.00% CO 2 and 74.00% N 2; the other was composed of 16.00% O 2, 4.00%, CO 2 and 80.00% N 2. The resting metabolic rate (RMR) was calculated from the oxygen consumption and carbon dioxide production according to the formula of Weir [15]. In the calculation of RMR, only VO 2 and VCO 2 values recorded when subjects were in steady-state (SS) conditions were used (SS = when values of VO 2 and VCO 2 were not varying 5%). In addition to measured values, RMR was predicted using the equations formulated by Harris and Benedict [13], and the FAO/WHO/UNU [3]. Dual energy X-ray absorptiometry (DXA) measurements were performed with a total body scanner (Model DPX, Lunar, Madison, WI, software version 3.6) that uses a constant potential X-ray source at 12.5 fj and a K-edge filter to achieve a congruent beam of stable dual-energy content (40 and 70 kev). Fat-free mass (FFM) was calculated as the sum of lean mass and bone mineral content. The variability of the instrument was reported in our previously published paper [15]. Statistical analyses were performed using the SPSS software program (SPSS, Chicago). Analysis of variance was used to test for Results We studied 131 subjects, including 46 men of mean age 30.2 years (SD = 13.1 years) and 85 women of mean age 45.3 years (SD = 7.4 years). The male and female populations differed significantly (p < 0.01) in anthropomorphic and bodycomposition measures, with the exception of BMI (Table 1). Values of RMR were determined separately for males and females (Table 2). Among males, there was no significant difference between clinically determined and predicted RMR. In contrast, the RMR predicted for females by the Harris-Benedict formula was significantly (p < 0.01) higher than that predicted by the FAO/WHO/UNU equation as well as that determined clinically. RMR was lower in the higher age groups (data not Table 1 Anthropomorphic values and body composition of the subjects. Values are means (SD) Male Female p value (n = 46) (n = 85) Age (years) 30.2 (13.1) 45.3 (13.7) < 0.01 Weight (kg) 80.1 (10.8) 63.8 (7.4) < 0.01 Height (cm) 177.4 (6.8) 160.4 (5.8) < 0.01 BMI (kg/m 2 ) 25.4 (2.7) 24.8 (2.6) ns Body fat (%) a 12.2 (8.3) 23.2 (6.2) < 0.01 Lean mass (kg) a 61.8 (8.2) 37.5 (4.1) < 0.01 BMI, body mass index; ns, not significant a Determined using dual energy X-ray absorptiometry (DXA) Table 2 Clinically determined and predicted values of resting metabolic rate (RMR), by gender. Values are means (SD) RMR (kcal/day) Male Female Measured 1865.5 (234.9) 1354.1 (154.9) Predicted Harris-Benedict formula 1851.3 (200.4) 1440.6 (148.2)* FAO/WHO/UNU formula 1862.1 (178.2) 1374.4 (91.6) *p < 0.01 compared to the measured value for women
A. De Lorenzo et al.: Resting metabolic rate in Italians 79 shown); these differences were not significant in females (p = 0.08) but they were in males (p < 0.001). The differences remained after correction for differences in weight and height between the age groups. After corrections for fat-free mass, the difference was only apparent in females. The difference between measured and predicted RMR was analyzed using the Bland and Altman technique (Table 3) [16]. Table 3 Differences between measured and predicted RMR in males and females. Value are 95% confidence intervals for the higher and lower limit of agreement, as determined by the by Bland and Altman technique RMR (kcal/day) Males Females a Mean (SD) Mean (SD) Measured vs. 14.0 (135) -86.0 (124)* predicted (HB) Measured vs. predicted 3.5 (156) -18.0 (133) (FAO/WHO/UNU) RMR, resting metabolic rate; HB, Harris-Benedict formula *p < 0.01 Only the difference between measured RMR and that predicted using the Harris-Benedict equation was significant (p < 0.001) in females. Pearson s correlation coefficients between measured and predicted RMR values were high: 0.75 and 0.82 in males and 0.52 and 0.66 in females for FAO/WHO/UNU and HB, respectively. Resting metabolic rate expressed per kilogram FFM was 34.2 (SD = 3.2) and 28.6 (SD = 1.9) kcal/kg day in females and males, respectively (p < 0.001). Figure 1 shows the individual differences between measured and predicted values of RMR plotted against the measured values. It is obvious that individual differences can be high. In Table 4, the coefficients of the stepwise multiple regression of RMR with weight and age are given for males and females separately. Height did not reach the level of significance (p < 0.05) to be entered in the regression model, in males (p = 0.4) nor in females (p = 0.3). The prediction of RMR using lean mass from DXA as independent variable did not result in better prediction equations, neither in males (R 0.75; SEE, 118 kcal) nor in females (R 0.38; SEE, 122 kcal). Fig. 1a,b Individual differences between measured and predicted values of RMR plotted against the measured values. a For RMR values predicted using the Harris- Benedict (HB) formula. b For RMR values predicted according to FAO/WHO/UNU. WEIR, RMR values measured according to Weir [15] Females Males b Females Males
80 A. De Lorenzo et al.: Resting metabolic rate in Italians Table 4 Stepwise multiple regression of RMR as dependent variable and weight and age as independent variables fl Constant R 2 SEE (kcal) (kcal) Weight (kg) Age (years) Coeff. (SE) Coeff. (SE) Coeff. (SE) Females 9.74 (2.02) - - 732 (129) 0.21 137 10.90 (1.76) - 5.11 (0.95) 890 (115) 0.42 118 Males 12.8 (2.6) - - 841 (214) 0.34 192 12.5 (1.8) - 10.6 (1.5) 1186 (154) 0.69 133 fl, regression coefficient for the mentioned variable and the standard error; R 2, explained variance; SEE, standard error of estimate of the prediction formula Discussion The subjects measured in this study were not specially selected. Most of them came to the Human Nutrition Unit for various studies on body composition and energy expenditure. The mean values for RMR in males and females are higher compared with values from other studies [5, 17, 18], both as absolute values as well as expressed per kilogram fat-free mass. Relatively high values of RMR have been reported for Italians. Hayter and Henry [4] reported, in a comparison of RMR data of Italian, North Europeans, Americans, Indians and Chinese, that the Italian regression equations over the entire range of body weights were higher compared to other populations groups. We applied the Italian prediction equations of Hayter and Henry [4] to this population and found that the predicted values for males were 1979 ± 185 kcal/day and for females 1565 ± 146 kcal/day. These values were significantly higher compared to the measured values in this study of 1865 ± 235 kcal/day and 1354 ± 155 kcal/day in males and females, respectively. The high values of the Italian subjects as reported by Hayter and Henry [4] can be explained by the fact that the data they used were collected mainly in young people with relatively high activity levels. Predicted values using an equation specific for obese subjects [19] resulted in highly underestimated values of 1179 ± 115 kcal/day and 1532 ± 203 kcal/day in females and males, respectively. The values predicted using the Harris-Benedict equation [13] or the FAO/WHO/UNU [3] equation did, however, result in fairly good mean estimates of RMR. The individual differences between measured and predicted values were in a range normally found also by other authors [20, 21]. These individual differences may be too large to make any prediction formula useful for individual use. In circumstances where individual values are required, the measurement instead of the prediction of RMR is highly recommended. The observed lower values of RMR at higher ages, as apparent from the age effect in the regression equation, is a normal observation. Apart from a lower fat-free mass in older subjects, mainly due to a lower muscle mass, organ mass is also lower at older ages [22, 23]. When RMR was regressed against fat-free mass as measured by DXA, the explained variance in RMR by lean mass and age was not better than the explained variance when weight and age were used (Table 4). A prediction equation based on weight and age has a wider applicability compared to a prediction equation which requires the use of expensive instruments such as DXA. Only a better explained variance and a much better estimate would justify the use of such a prediction equation. The relatively low explained variance in RMR from fat-free mass and age may be due to the fact that RMR is mainly determined by the organ masses of heart, liver and brain, and only a relatively small contribution is from muscle [5, 24, 25]. Muscle, however, is the largest part of the fat-free mass. The fact that RMR expressed per kilogram FFM in males is lower compared to that in females may have its reason in a lower organ fraction in the FFM in males. Resting metabolic rate in Italian subjects is found to be relatively high in comparison with international data in comparable age groups, although the values are slightly lower than those reported earlier. The reason for this remains unclear. The prediction of RMR from fat-free mass is not better compared to that from weight and age. The individual error of the prediction is too high to be of practical use in individuals. References 1. Brandi LS, Oleggini M, Lachi S, Frediani M, Bevilacqua S, Mosca F, Ferrannini E (1988) Energy metabolism of surgical patients in the early postoperative period: a reappraisal. Crit Care Med 16:18-22
A. De Lorenzo et al.: Resting metabolic rate in Italians 81 2. MacFie J (1984) Active metabolic expenditure of gastroenterological surgical patients receiving intravenous nutrition. J Parenter Enteral Nutr 8:371-376 3. FAO/WHO/UNU (1985) Expert consultation energy and protein requirement N 724. WHO, Geneva 4. Hayter JE, Henry CJK (1994) A re-examination of basal metabolic rate predictive equations: the importance of geographic origin of subjects in sample selection. Eur J Clin Nutr 48:702-707 5. Weinsier RL, Schutz Y, Bracco D (1992) Re-examination of the relationship of resting metabolic rate to fat-free mass and to the metabolically active components of fat-free mass in human. Am J Clin Nutr 55:790-794 6. Felloni G (1936) Contributo alla conoscenza del metabolismo di base degli Italiani. Nota I. Il metabolismo di base di soggetti dai 19 ai 25 anni. Quad Nutr 3:419-421 7. Granati A, Busca L (1941) Il metabolismo di base di un gruppo di lavoratori. Quad Nutr 8:411-421 8. Lafratta E (1937) Contributo alla conoscenza del metabolismo di base degli Italiani. Nota VI. Il metabolismo di base di soggetti (maschi) dai 14 ai 20 anni. Quad Nutr 4:410-416 9. Lenti C (1937) Contributo alla conoscenza del metabolismo di base degli Italiani. Nota IV. Il metabolismo di base di soggetti dai 20 ai 25 anni. Quad Nutr 4:192-225 10. Occhiuto A, Pepe M (1939) Contributo alla conoscenza del metabolismo di base degli Italiani. Nota VIII. Il metabolismo di base di soggetti (femmine) di oltre 20 anni. Quad Nutr 6:189-198 11. Occhiuto A, Pepe M (1940) Contributo alla conoscenza del metabolismo di base degli Italiani. Nota X. Il metabolismo di base di soggetti (maschi) di oltre 20 anni. Quad Nutr 7:142-154 12. Pepe M (1938) Contributo alla conoscenza del metabolismo di base degli Italiani. Nota VII. Il metabolismo di base di soggetti dai 18 ai 24 anni. Quad Nutr 5:206-214 13. Harris JA, Benedict FG (1919) A biometric study of basal metabolism in man. Carnegie Institute of Washington, Washington DC, pp 1-266 (Publ. no. 279) 14. Weir JBD (1949) New methods for calculating metabolic rate with special reference to protein metabolism. J Physiol (London) 109:1-9 15. De Lorenzo A, Andreoli N, Candeloro (1997) Within-subject variability in body composition using dual energy X-ray absorptiometry. Clin Physiol 17(2):383-388 16. Bland JM, Altman DG (1986) Statistical methods for assessing agreement between two methods of clinical measurement. Lancet i:307-310 17. Schofield WN (1985) Predicting basal metabolic rate, new standard and review of previous work. Hum Nutr Clin Nutr 39[Suppl. 1]:5-41 18. Weststrate JA (1993) Resting metabolic rate and diet-induced thermogenesis: a methodological reappraisal. Am J Clin Nutr 58:592-601 19. Bernstein RS, Thornton J, Yang MU, Wang J, Redmond A, Pierson RN, Pi-Sunyer FX, Van Itallie TB (1983) Prediction of the resting metabolic rate in obese patients. Am J Clin Nutr 37:595-602 20. Taaffee DR, Thompson J, Butterfield G, Marcus R (1995) Accuracy of equations to predict basal metabolic rate in older women. J Am Diet Assoc 95:1387-1302 21. Ferro-Luzzi A, Petracchi C, Kuriyan R, Kurpad AV (1997) Basal metabolism of weight-stable chronically undernourished men and women: lack of metabolic adaptation and ethnic differences. Am J Clin Nutr 66:1086-1093 22. Tzankoff SP, Norris AH (1977) Effect of muscle mass decrease on age-related BMR changes. J Appl Physiol 43:1001-1006 23. Vaughan L, Zurlo F, Ravussin E (1991) Aging and energy expenditure. Am J Clin Nutr 53:821-825 24. Deurenberg P (1994) Assessment and classification of obesity. In: Ditschuneit H, Gries FA, Hauner H, Schusdziarra V, Wechsler JG (eds) Obesity in Europe 93. John Libbey, London Paris Rome, pp 83-88 25. Gallagher D, Belmonte D, Deurenberg P, Wang Z-M, Krasnow N, Pi-Sunyer FX, Heymsfield SB (1998) Organ tissue mass measurement by MRI allows accurate in vivo modeling of REE and metabolic active tissue mass. Am J Physiol 275(2 Pt 1):E249-E258