Measured and predicted resting metabolic rate in Italian males and females, aged 18 ± 59 y

Transcription

1 (2001) 55, 208±214 ß 2001 Nature Publishing Group All rights reserved 0954±3007/01 $ Measured and predicted resting metabolic rate in Italian males and females, aged 18 ± 59 y A De Lorenzo 1,2 *, A Tagliabue 3, A Andreoli 1, G Testolin 2, M Comelli 3 and P Deurenberg 1,4 1 Department of Human Physiology and Nutrition, University `Tor Vergata', Rome, Italy; 2 International Center for the Assessment of Body Composition, University of Milan, Milan, Italy; 3 Department of Applied Health Science, University of Pavia, Pavia, Italy; and 4 Department of Human Nutrition and Epidemiology, Wageningen University, Wageningen, The Netherlands, and Nutrition Consultant in Singapore Objectives: To determine the resting metabolic rate in a sample of the Italian population, and to evaluate the validity of predictive equations for resting metabolic rate (RMR) from the literature in normal and obese subjects. Design: Cross-sectional observational study. Settings: Department of Human Physiology and Nutrition, University `Tor Vergata', Rome. Subjects: A total of 320 healthy subjects, 127 males and 193 females, aged 18 ± 59 y. Methods: Weight, height and resting metabolic rate by indirect calorimetry were measured. Resting metabolic rate was also predicted using equations from the literature. Results: Resting metabolic rate (mean s.d.) in normal weight subjects was kj=24 h (males) and kj=24 h (females). Measured RMR and predicted RMR values using various equations from the literature were signi cantly different in males and females, except for the Harris ± Benedict equation and the Scho eld equations. Also, in overweight and obese subjects the prediction error was generally larger compared to normal-weight subjects for all formulas except for the Harris ± Benedict and Scho eld formulas. In overweight and obese males but not in females, RMR was lower than in normal-weight subjects after correcting for weight and age differences. Stepwise multiple regression of resting metabolic rate against weight, height and age in males and females did not reveal a prediction formula with a lower prediction error than the Harris ± Benedict or Scho eld formulas and thus was not further explored. Conclusions: The Harris ± Benedict formula and the Scho eld formula provide a valid estimation of resting metabolic rate at a group level in both normal-weight and overweight Italians. However, the individual error can be so high that for individual use a measured value has to be preferred over an estimated value. Descriptors: resting metabolic rate; prediction equations; validity; normal weight; overweight; obesity; Italians (2001) 55, 208±214 Introduction Resting metabolic rate (RMR) is an important parameter in the assessment of nutritional status in patients and is used for example to calculate the energy needs of a patient who needs parenteral or enteral nutrition (Brandi et al, 1988; MacFie, 1984). In addition information on resting energy expenditure is necessary to calculate energy needs at a population level. For this, the FAO=WHO=UNU (1985) has published prediction formulas for the assessment of resting metabolic rate. *Correspondence: A De Lorenzo, Human Nutrition Unit, University Tor Vergata, Via di Tor Vergata, Rome, Italy. delorenzo@uniroma2.it Guarantor: A De Lorenzo. Contributors: ADL and GT: general project management; AT and AA: daily project management and measurements; MC: data analyses; PD: data analyses and intrepretations. Received 2 May 2000; accepted 6 November 2000 The RMR is de ned as the energy expenditure 10 ± 12 h after a meal, the subject lying supine and completely at physical and mental rest in a thermoneutral environment. It can be measured by direct (heat exchange) or indirect calorimetric (gas exchange) techniques (Garrow & James, 1993), where the latter technique is easier. The accuracy of RMR measurement procedures is fairly good, as the withinsubjects coef cient of variation is about 5% (Weststrate et al, 1989). This is a crucial point, since any bias in RMR assessment would amplify the calculation errors of estimated total energy requirements of individuals. Variability ranges of RMR assessment have to be carefully evaluated if total energy expenditure (TEE) is calculated as a multiple of RMR. RMR is the component of energy expenditure that explains the largest proportion of TEE in individuals, but the contribution of a low RMR to the aetiology of obesity is controversial. Given the high prevalence of obesity in many countries (WHO, 1998) and the strong relationship of obesity with several diseases, information on energy meta-

2 bolism in individuals as well as in population groups may be important to combat obesity (Seidell, 1997). Variations in RMR are known to be related to body weight, fat-free mass (FFM), fat mass (FM), age, sex, ethnicity and environmental factors such as temperature, and these factors must be taken into account (Weststrate et al, 1990). For this reason predictive equations for RMR based on body composition formulas are generally population-speci c. There are quite a number of studies in which prediction formulas for RMR have been validated (Hayter & Henry, 1994; Heshka et al, 1993; Weinsier et al, 1992), but there are no recent studies carried out in Italians, or they are in very speci c population groups (Scal et al, 1993; De Lorenzo et al, 1999,2000) or they are outdated and in young subjects with an active life style (see Hayter & Henry, 1994). The aim of the present study was to measure resting metabolic rate in a relatively large population group and to re-evaluate the validity of prediction equations from the literature for Italians. Subjects and methods Three-hundred and twenty subjects, 127 males (18 ± 59 y) and 193 females (19 ± 59 y), participated in the study. None of the subjects had any disease or was taking any medications known to affect resting metabolic rate. The mean body mass index in different age categories was similar to the 51st percentile of the Italian population in 1994 (ISTAT, 1997). Table 1 gives some descriptive statistics of the subjects. The subjects were invited to the Human Nutrition Unit at University `Tor Vergata' in Rome in the early morning after an overnight fast. They were requested to refrain from any unnecessary physical activity prior to the measurements. The study protocol was approved by the Ethical Committee of the University of `Tor Vergata', and written informed consent was obtained from each subject. Body weight was measured in underwear to the nearest 0.1 kg and body height was measured without shoes to the nearest 0.1 cm. Body mass index (BMI) was calculated as weight=height 2 (kg=m 2 ). Subjects were categorised based on their BMI according to the WHO (1998) into normal weight (BMI < 25 kg=m 2 ), overweight (25 kg=m 2 BMI Table 1 Characteristics of the population Mean s.d. Range Mean s.d. Range Age (y) ± ** ± 59 Height (cm) ± ** ± Weight (kg) ± ** ± BMI (kg=m 2 ) ± * ± 39.4 RMR (kj=24 h) ± ** ± BMI, body mass index (weight=height squared). RMR, resting metabolic rate. *P < 0.05; **P < < 30 kg=m 2 ) and obese (BMI 30 kg=m 2 ). FFM was calculated using the gender-speci c equation of Moore et al (1963). Body surface area was calculated using the formulas of Dubois and Dubois (1916). Prior to the RMR measurements, the subjects lied supine for 25 ± 30 min in a quiet room at an ambient temperature of 22 C. Then oxygen consumption (VO 2 ) and carbon dioxide production (VCO 2 ) were measured for a 30 min period by an open circuit indirect calorimeter using a face mask (Sensormedic 2900, California, USA). Daily calibration of the calorimeter was conducted following the instructions of the manufacturer. For additional quality control two different certi ed oxygen=carbon dioxide gas mixtures (SIAD Ltd Co, Rome, Italy) were used. The accuracy of the gas measurements was within 4.5% of the true value and the reproducibility of the measurements over time was within 3.5% (unpublished results). RMR was calculated from oxygen consumption and carbon dioxide production according to the formula of Weir (1949). For the calculation of RMR, only data of subjects in apparently steady-state conditions (ie VO 2 and VCO 2 did not vary more than 5% from the mean value of the 30 min measurement period) were used. In addition to measured values, RMR was predicted using the equations formulated by Harris and Benedict (1919), Robertson and Reid (1952), Scho eld (1985), Pavlou et al (1986), Owen et al (1986,1987), Mif in et al (1990) and Cunningham (1991). The equations are given in Table 2. Statistical analyses were performed using the SPSS software program (SPSS, 1997). Analysis of (co-)variance was used to test for differences in parameters between males and females and to test differences in RMR between normal-weight and overweight subjects. A paired t-test was used to test differences between measured and predicted values. Correlations are Pearson's product ± moment correlations. Stepwise multiple regression was used to explore the relationship of RMR with weight, height and age within the gender groups. Values are presented as mean and standard deviation (s.d.) unless otherwise stated. A P-value less than 0.05 is considered as signi cant. Results Characteristics of the study group are shown in Table 1 for males and females separately. As expected, males were taller (P < 0.001) and had higher body weights (P < 0.001) compared to females. Also, males had a signi cantly (P < 0.03) lower BMI than females and were signi cantly younger (P < 0.001). After correcting for the age difference between males and females the BMI did no longer signi cantly differ between the sexes. The difference among the sexes in weight and height remained unchanged. In Table 3 measured and predicted RMR (kj=24 h) as well as the difference between measured and predicted RMR (with the 95% con dence interval) is given for males and females separately. The correlations between measured and predicted RMR values (Table 3) were generally high 209

3 210 Table 2 Used prediction formulas for resting metabolic rate (kj=24 h) Author(s) Formula a Population in which formula was developed Cunningham ( FFM)4.18 Meta analyses, normal-weight and obese individuals Harris & Benedict ( W 5H 7 6.8A)4.18 Normal-weight individuals Mif in et al b (9.99W 6.25H A 5)4.18 Normal-weight and obese individuals Pavlou et al c ( H 16.81W 7 8.9A AIBW)4.18 Obese individuals Pavlou et al d ( BMR Harris ± Benedict )4.18 Obese individuals Scho eld e (63W 2896) Normal-weight and obese individuals Scho eld f (48W 3653) Normal-weight and obese individuals Scho eld e (63W H 2953) Normal-weight and obese individuals Scho eld f (48W H 3670) Normal-weight and obese individuals Owen ( W)4.18 Normal-weight and obese individuals Robertson & Reid Calculated from body surface according to Heshka et al (1993) Normal-weight individuals Cunningham ( FFM)4.18 Meta analyses, normal-weight and obese individuals Harris & Benedict ( W 1.9H 7 4.7A)4.18 Normal-weight individuals Mif in et al b (9.99W 6.25H A 7 161)4.18 Normal-weight and obese individuals Scho eld e (62W 2036) Normal-weight and obese individuals Scho eld f (34W 3538) Normal-weight and obese individuals Scho eld e (57W 11.84H 411) Normal-weight and obese individuals Scho eld f (34W 0.06H 3530) Normal-weight and obese individuals Owen ( W)4.18 Normal-weight and obese individuals Robinson & Reid Calculated from body surface according to Heshka et al (1993) Normal-weight individuals a W, weight in kg; H, height in cm; A, age in years, FFM, fat-free mass in kg. b Sex speci c. c AIBW, percentage above ideal weight. d BMR Harris ± Benedict, BMI as predicted by Harris & Benedict. e Based on weight only. f Based on weight and height. Table 3 Differences and correlation coef cients a between measured and predicted resting metabolic rate (kj=24 h) in males and females (mean, s.d., 95% con dence interval) RMR Difference Pearson Mean s.d. correlation* Mean s.d. 95% con dence interval Measured Cunningham * Harris & Benedict NS Mif n et al * Owen * Pavlou et al b * Pavlou et al c * Scho eld d NS Scho eld e NS Robertson & Reid * Measured Cunningham * Harris & Benedict NS Mif in et al * Owen * Scho eld d NS Scho eld e NS Robertson & Reid * NS, not signi cant;.*p < a Correlation between measured an predicted value. b Formula based on percent above ideal weight. c Formula based on Harris ± Benedict. d Based on weight only. e Based on weight and height.

4 and ranged in females from (Scho eld, weight alone, P < 0.001) to (Harris ± Benedict, P < 0.001) and in males from (Owen et al, P < 0.001) to (Pavlou et al, P < 0.001). However, all prediction formulas signi cantly underestimated RMR, except for the Harris ± Benedict and the Scho eld equations in both sexes. The sometimes large difference between measured and predicted value is notable, for example in males and females for the Owen equations. In overweight (BMI 25 kg=m 2 ) and obese subjects (BMI 30 kg=m 2 ) there was generally a larger prediction error with all equations but the estimates from Harris ± Benedict and Scho eld formulas were only borderline signi cant in the obese groups. In obese females both formulas slightly underestimated the RMR (P ranging from 0.03 to 0.07) and in overweight males the formulas had the tendency (P ˆ 0.07) to overestimate the RMR (results not shown). The dependency of predicted values on the BMI could be observed for most prediction formulas in females but not in males (see Table 4), indicating that in females the underestimation of RMR tends to be higher in obese subjects. Table 5 shows the metabolic rate in normal-weight (BMI < 25 kg=m 2 ) and overweight and obese subjects (BMI 25 kg=m 2 ). The uncorrected value of metabolic rate was higher in the obese subjects. However, after for weight and age differences, the difference between normal-weight and overweight subjects disappeared in females, whereas in males the RMR in overweight and obese was lower compared to normalweight males. Figures 1 and 2 show the difference between measured and predicted RMR in normal-weight, overweight and obese subjects in males and females, respectively. 211 Table 4 Correlation of bias a of predicted metabolic rate with BMI Formula Correlation P-value Correlation P-value Cunningham Harris ± Benedict Mif in et al Owen Pavlou et al b Ð Ð Pavlou et al c Ð Ð Scho eld d Scho eld e Robertson & Reid a Bias: predicted minus measured RMR. b Formula based on percent above ideal weight. c Formula based on Harris ± Benedict. d Age speci c formula based on weight only. e Age speci c formula based on weight and height. Table 5 Resting metabolic rate (kj=24 h) in normal weight and overweight males and females, before and after for weight and age differences Before After Before After mean s.e. Mean s.e. Mean s.e. mean s.e. Normal-weight Overweight * *P < 0.01 compared to normal-weight males. Normal-weight, BMI < 25 kg=m 2 ; overweight, BMI 25 kg=m 2. Figure 1 Differences between measured and predicted resting metabolic rate in normal weight, overweight and obese males.

5 212 Figure 2 Differences between measured and predicted resting metabolic rate in normal weight, overweight and obese females. Stepwise multiple regression revealed as prediction equation for RMR in females: RMR ˆ 46:322 weight 15:744 height 16:66 age 944 (r 2 ˆ 0.597, SEE 581 kj=day) and for males: RMR ˆ 53:284 weight 20:957 height 23:859 age 487 (r 2 ˆ 0.597, SEE 650 kj=day). These prediction equations have a comparable error compared to the Harris ± Benedict and the Scho eld equations (see s.d. of difference between measured and predicted values in Table 3). Discussion The aim of the study was to measure RMR in a large number of not specially selected subjects and to verify the validity of existing prediction equations from the literature in healthy normal-weight and obese adults. Although the present study population can not be regarded as representative for the Italian population, weight and height are around the median value of the Italian population (ISTAT, 1997). The RMRs found in this study are higher than values reported in other, comparable populations (Weinsier et al, 1992; Scho eld, 1985; Mif in et al, 1990) and comparable to the relatively high RMR values reported earlier in Italians (Pepe, 1938; Hayter & Henry, 1994). However, mean weight, height and body mass index in the present study population are higher compared to other populations (for example Scho eld, 1985; Mif in et al, 1990), which could easily explain the higher RMR values found. The fact that, for example, the Scho eld (1965) equations provide valid mean estimates of RMR supports this explanation. The results also show that, among the published RMR predicted equations used in this study, most equations grossly underestimate RMR in males and in females, whereas the Harris and Benedict (1919) and the Scho eld (1965) equations result in rather accurate mean predicted values. In overweight and obese subjects the mean underestimation was generally higher in females but slightly lower in males. In both normal-weight and obese subjects the Harris ± Benedict and the Scho eld formula estimated mean RMR well with only slight underestimations in females and slight overestimations in males. Overweight and obese subjects in the present population have a higher absolute RMR compared to normal-weight subjects, but after adjustment (ANCOVA) for body weight and age RMR was not different between normal-weight and overweight females. In overweight males the corrected RMR was slightly lower (P < 0.01) than in normal weight males (Table 5). A lower RMR in obese subjects after for weight can be expected on the bases of their body composition, as obese subjects are likely to have less metabolically active tissue (fat free mass, organ mass) per kg body weight (Weinsier et al, 1992). Ideally RMR data have to be corrected for differences in FFM (Deurenberg, 1994), although it can be argued that such a is only a very crude one. Gallagher et al (1998) recently showed that the FFM is not one entity in that different components of the FFM have different contributions to the RMR. Based on their data it can be argued that obese subjects will have a lower RMR per kg body mass or even per kg fat-free mass, as their organ contribution per kg body weight or per kg FFM will be relatively lower. One of the causes for becoming (or being) obese may be a relatively low metabolic rate, resulting in an easy weight

6 gain despite a modest energy intake compared to others, who have a relatively higher metabolic rate. Astrup et al (1999) showed that lower RMR values are usually found in (normal-weight) post-obese subjects compared to normalweight subjects and concludes that the obese state is at least partly the consequence of this lower metabolic rate. However, in a study of Wyatt et al (1999) no evidence was found that weight loss results in a lower than expected RMR, suggesting that many people will have a normal RMR after weight loss. It is well known that, although many obese individuals are able to lose weight, most cannot maintain the weight loss for longer periods. A major point of controversy is whether the high degree of recidivism after weight loss is due to biological (eg relatively low metabolic rate) or due to behavioural factors. Some investigators conclude that weight regain is inevitable as a result of strong biological pressures to return the subject to an obese body weight. Others believe that the inability to maintain substantial lifestyle changes over time is the main culprit in weight regain. The results of Astrup et al (1999) suggest that a low RMR may contribute to weight regain in some formerly obese subjects. This should be taken into account in the treatment of obesity. The individual differences between measured and predicted values of RMR in this study were in a range normally found also by other authors (Heshka et al, 1993; Taaffe et al, 1995; Ferro-Luzzi et al, 1997). These individual differences are large and may be even too large to make any prediction formula useful for individual use. The development of a prediction equation in this study population did not result in better individual estimates as the SEE of the prediction equation was comparable with the s.d. of the difference between measured and predicted values as obtained from formulas from the literature. As the Harris and Benedict (1919) formula and the Scho eld (1965) formula provide valid estimates of RMR with comparable estimation error (see Table 3), such a prediction equation would also not contribute in any way, and no further in-depth analyses on this were performed. Conclusions This study con rms data from the literature suggesting that Italians have relatively high RMR compared to other populations, but it is likely that this higher RMR is due to differences in body composition. Prediction formulas from the literature generally underestimated RMR, but the Harris ± Benedict and the Scho eld formulas provided a valid mean estimate of RMR in both normal-weight and overweight subjects. Individual differences between measured and predicted values were high, making it necessary that in circumstances where reliable individual values are required, RMR should be measured rather than predicted. References Astrup A, Gotzsche PC, van de Werken K, Ranneries C, Toubro S, Raben A & Buemann B (1999): Meta-analysis of resting metabolic rate in formerly obese subjects. Am. J. Clin. Nutr. 69, 1117 ± 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. Cunningham JJ (1991): Body composition as a determinant of energy expenditure: a synthetic review and a proposed general prediction equation. Am. J. Clin. Nutr. 54, 963 ± 969. De Lorenzo A, Bertini I, Candeloro N, Piccinelli R, Innocente I & Brancati A (1999): A new predictive equation to calculate resting metabolic rate in athletes. J. Sports Med. Phys. Fitness 39, 213 ± 219. De Lorenzo A, Andreoli A, Bertoli S, Oriani G, Testolin G & Deurenberg P (2000): : relation with body composition and anthropometric parameters. Acta Diabetologica 37, 1±7. Deurenberg P (1994): Assessment and classi cation of obesity. In Obesity in Europe 93, ed. H Ditschuneit, FA Gries, H Hauner, V Schusdziarra & JG Wechsler, pp 83 ± 88. London: John Libbey. DuBois D & DuBois EF (1916): A formula to estimate the approximate surface area if height and weight be known. Arch. Intern. Med. 17, 863 ± 868. FAO=WHO=UNU (1985): Expert consultation on energy and protein requirements. Geneva: WHO. 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 ± Gallagher D, Belmonte D, Deurenberg P, Wang Z, Krasnow N, Pi-Sunyer FX & Heyms eld SB (1998): Organ-tissue mass measurement allows modeling of REE and metabolically active tissue mass. Am. J. Physiol. Endocrinol. Metab , E249 ± 258. Garrow JS & James WPT (1993): Human Nutrition and Dietetics. 9th edn. Edinburgh: Churchill Livingstone. Harris JA & Benedict FG (1919): A Biometric Study of Basal Metabolism in Man. (publication no. 279), pp 1 ± 266. Washington, DC: Carnegie Institute of Washington. 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. Heshka S, Feld K, Wang M-U, Allison D & Heyms eld SB (1993): Resting energy expenditure in the obese: a cross-validation and comparison of prediction equations. J. Am. Diet. Assoc. 93, 1031 ± ISTAT (1997): Condizioni di salute e ricorso ai servizi sanitari. Indagine Multiscopo sulle Famiglie. Anno Istituto Nazionale di Statistica (Central Bureau of Statistics) no 54. MacFie J (1984): Active metabolic expenditure of gastro-enterological surgical patients receiving intravenous nutrition. J. Parenter. Enteral. Nutr. 8, 371 ± 376. Mif in MD, St Jeor ST, Hill LA, Scott BJ, Daugherty SA & Koh YO. (1990): A new predictive equation for resting energy expenditure in healthy individuals. Am. J. Clin. Nutr. 51, 241 ± 247. Moore F, Olesen K, McMurray J, Parker V, Ball M & Boyden C (1963): The Body Cell Mass and its Supporting Environment. Philadelphia, PA: WB Saunders. Owen OE, Kavle E, Owen RS, Polansky M, Caprio S, Mozzoli MA, Kendrick ZV, Bushman MC, Boden G (1986): A reappraisal of caloric requirements in healthy women. Am. J. Clin. Nutr. 44, 1 ± 19. Owen OE, Holup JL, D'Alessio DA, Craig ES, Polansky M, Smalley KJ, Kavle EC, Bushman MC, Owen LR, Mozzoli MA, Kendrick ZV & Boden GH (1987). A reappraisal of the caloric requirements of men. Am. J. Clin. Nutr. 46, 875 ± 885. Pavlou KN, Hoefer MA & Blackburn GL (1986): Resting energy expenditure in moderate obesity. Predicting velocity of weight loss. Ann. Surg. 203, 136 ±

7 214 Pepe M (1938): Contributo alla conoscenza del metabolismo di base degli Italiani. Nota VII. II metabolismo di base di soggeti dai 18 ai anni. Quadd. Nutr. 5, 206 ± 214. Robertson JD & Reid DD (1952): Standards for the basal metabolism of normal people in Britain. Lancet i, 940 ± 943. Scal L, Coltorti A & Sapio C (1993): Predicted and measured resting energy expenditure in healthy young women. Clin. Nutr. 12, 1 ± 7. Scho eld WN (1985): Predicting basal metabolic rate, new standard and review of previous work. Hum. Nutr. Clin. Nutr. 39c(Suppl 1), 5 ± 41. Seidell JC (1997): Time trends in obesity: an epidemiological perspective. Horm. Metab. Res. 47, 155 ± 199. SPSS (1997): SPSS=Windows V7.5.2 Manuals. Chicago, IL, SPSS Publishing. Taaffee DR, Thompson J, Butter eld G & Marcus R (1995): Accuracy of equations to predict basal metabolic rate in older women. J. Am. Diet. Assoc. 95, 1387 ± 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. Weir JBD (1949): New methods for calculating metabolic rate with special reference to protein metabolism. J. Physiol. London 109, 1 ±9. Weststrate JA, Weys PJM, Poortvliet EJ, Deurenberg P & Hautvast JGAJ (1989): Diurnal variation in postabsorptive resting metabolic rate and diet-induced thermogenesis. Am. J. Clin. Nutr. 50, 908 ± 914. Weststrate JA, Dekker J, Stoel M, Begheijn L, Deurenberg P & Hautvast JGAJ (1990): Resting energy expenditure in women: impact of obesity and body-fat distribution. Metabolism 39, 11 ± 17. WHO (1998): Obesity Ð preventing and managing the global epidemic. Report of WHO consultation on obesity, Geneva, 3 ± 5 June 1997 (WHO=NUT=NCD=98.1). Geneva: WHO. Wyatt HR, Grunwald GK, Seagle HM, Klem ML, McGuire MT, Wing RR & Hill JO (1999): Resting energy expenditure in reduced-obese subjects in the National Weight Control Registry. Am. J. Clin. Nutr. 69, 1189 ± 1193.