Validity and reproducibility of resting metabolic rate measurements in ...

European Journal of Clinical Nutrition (2005) 59, 651–657

& 2005 Nature Publishing Group All rights reserved 0954-3007/05 $30.00 www.nature.com/ejcn

ORIGINAL COMMUNICATION Validity and reproducibility of resting metabolic rate measurements in rural Bangladeshi women: comparison of measurements obtained by Medgemt and by Deltatract device DS Alam1, PJM Hulshof2*, D Roordink2, M Meltzer2, M Yunus1, MA Salam1 and JMA van Raaij2 1 Centre for Health and Population Research, Dhaka, Bangladesh; and 2Division of Human Nutrition, Wageningen University, Wageningen, The Netherlands

Objective: To assess reproducibility and validity of resting metabolic rate (RMR) of Bangladeshi women as measured with the MedGem device and using the Deltatrac metabolic monitor as a reference; and (2) to evaluate the FAO/WHO/UNU basal metabolic rate (BMR)-prediction equations. Design: In each of two sessions, resting oxygen consumption was measured in triplicate by MedGem and in triplicate by Deltatrac device. Setting: Matlab area, the rural field research area of the Centre for Health and Population Research, Bangladesh (ICDDR,B). Subjects: A total of 37 nonpregnant, nonlactating women, aged 27.674.5 y, BMI 20.873.1 kg/m2 participated. Results: The difference in oxygen consumption by MedGem and Deltatrac device was significantly level dependent. Withinsubject within-session variations (expressed as CV) were 9.0 and 3.0% (Po0.01) and within-subject between-session variations were 8.2 and 4.5% (Po0.01) for MedGem and Deltatrac, respectively. Mean RMR measured by Deltatrac (5.1770.51 MJ/day) was not significantly different from the BMR predicted by the FAO/WHO/UNU equations (5.1670.42 MJ/day) in the second session and only 0.19 MJ/day higher than predicted in the first session (Po0.05). Conclusion: Reproducibility and validity of the MedGem device was poor compared to the Deltatrac reference method. The FAO/WHO/UNU BMR-prediction equations give a good estimation of the BMR of rural, nonpregnant, nonlactating Bangladeshi women of 18–35 y. Sponsorship: Wageningen University (The Netherlands) and ICDDR,B (Bangladesh).

European Journal of Clinical Nutrition (2005) 59, 651–657. doi:10.1038/sj.ejcn.1602122 Published online 30 March 2005 Keywords: indirect calorimetry; validity; reproducibility; FAO/WHO/UNU BMR-prediction equations; MedGem; Deltatrac; rural Bangladeshi women

Introduction *Correspondence: PJM Hulshof, Division of Human Nutrition, Wageningen University, PO Box 8129, 6700 EV Wageningen, The Netherlands. E-mail: [email protected] Guarantor: PJM Hulshof. Contributors: DSA, JvR and PH conceptualized and supervised the study and developed the analysis plan. MM, DR, MY and MAS assisted with the study design. MM and DR led the data collection and analysis and provided support for the manuscript preparation. All authors critically revised the first draft for content and contributed to the final draft. No authors had any financial or personal interest in the research conducted. Received 22 June 2004; revised 1 December 2004; accepted 20 December 2004; published online 30 March 2005

In Bangladesh, a large proportion of the population, especially children and women, suffer from micronutrient malnutrition and/or chronic energy deficiency (FAO, 1999). Bangladeshi women, who enter pregnancy with an insufficient nutritional status, have a greater risk of inadequate pregnancy performances and enter lactation with inadequate body stores (Osendarp et al, 2001; Alam et al, 2003). Poor maternal nutritional status and low weight gains during pregnancy are both indicators for low birth weight (ACC/ SCN, 1997). In Bangladesh, the rate of low birth weight is nearly 50% and is one of the major causes of neonatal deaths (Fauveau, 1994). Therefore, Bangladesh has a variety of

RMR measurements in Bangladeshi women DS Alam et al

652 nutritional intervention programmes aiming to improve health and nutritional status of mothers and infants (ICDDR,B, 2002). The FAO/WHO/UNU Expert Committee on Energy and Protein Requirements adopted the approach to base energy requirements on energy expenditure as a multiple of basal metabolic rate (BMR), rather than on energy intake (FAO/WHO/UNU, 1985). The widely used FAO/WHO/UNU BMR-prediction equations to estimate energy requirements are based on the work of Schofield et al (1985), who, however, have been criticized for their overestimation of BMR in a number of non-Western populations (Henry & Rees, 1991; Shetty et al, 1996). Indirect calorimetry, either by Douglas bag or by ventilated hood technique, is the most commonly used approach for measuring BMR. However, these techniques are often not very convenient for large-scale use in rural settings. Recently, the MedGem, a portable and easy to use handheld device intended for measuring resting metabolic rate (RMR), was introduced into market (Healthetech, 2001a). The MedGem is an indirect calorimetric metabolic monitor able to monitor airflow and oxygen concentration in the inspired and expired air and hence allows estimation of energy expenditure when a fixed respiratory quotient is assumed. Recently Nieman et al (2003) described the validity of the BodyGem, a device that is identical to the MedGem, except that the former device only displays RMR and not oxygen consumption. According to these researchers, the BodyGem is an accurate and reliable device for measuring oxygen consumption and RMR. The current paper describes the reproducibility and validity of RMR in 37 Bangladeshi women from a rural setting as measured with the MedGem device using the Deltatrac metabolic monitor as a reference. In order to judge the basics of the energy requirement assessment for women from rural settings, the FAO/WHO/UNU BMR-prediction equations based on body weight (Schofield et al, 1985) were evaluated by comparing measured RMRs with predicted BMRs.

Subjects and methods Subjects Subjects were recruited from five villages (Charmukundi, Dhakirgaon, Nabakalash, Dighaldi and Kaladi) in the Matlab area, the rural field research area of ICDDR,B: Centre for Health and Population Research, Bangladesh. Matlab has a total population of approximately 220 000 people and is located 55 km southeast of Dhaka, Bangladesh. For reasons of feasibility, the selected villages were within 3 km from the ICDDR,B clinic in Matlab, the location where the measurements took place. A total of 37 nonpregnant nonlactating women aged 18–35 y were randomly selected from the Health and Demographic Surveillance System (HDSS) database of ICDDR,B. Before enrollment, an informed written consent was obtained from each woman. The characteristics of the study subjects are shown in Table 1. European Journal of Clinical Nutrition

Table 1

Characteristics of the 37 women in the study

Age (y)a Body weight (kg)a Height (cm)a BMI (kg/m2)a Body fat (%)a Midupperarm circumference (cm)a Waist-to-hip ratioa

27.674.5 49.577.9 154.575.7 20.873.1 2875.1 27.373.0 0.8070.04

Religionb Muslim Hindu

32 (86%) 5 (14%)

Education (y)b 0 1–5 6–10 11–16 Unknown

7 12 12 5 1

Marital statusb Unmarried Married

3 (8%) 34 (92%)

Woman’s occupationb Housewife Other

30 (81%) 7 (19%)

Man’s occupationb Unskilled Skilled Businessman Farmer Services Unknown

2 2 17 4 9 3

(19%) (32%) (32%) (14%) (3%)

(5%) (5%) (46%) (11%) (24%) (8%)

a

Mean7s.d. Frequency (%).

b

Study design Measurements were performed during two sessions, with a time interval of two and a half weeks between the sessions. The measurements started at 7 or 8 O’clock in the morning. Oxygen consumption was measured in supine position with alternating use of the MedGem and Deltatrac device according to the time schedule shown in Figure 1. For each subject, the same time protocol was followed in both measurement sessions. The women were instructed to abstain from foods and drinks during 10–12 h preceding the measurements in the morning. To prevent them from heavy exercise prior to the measurements, they were picked up from their homes by rickshaw and transported to the clinic. On arrival in the clinic, the subjects were instructed to lay down for at least 25 min before the start of the measurements. Next, subjects were equilibrated for 10 min under the canopy of the Deltatrac device. This time period was sufficient to reach steady-state levels in the subjects. Deltatrac readings from this time period were not included in the data analysis. After the oxygen consumption measurements by Deltatrac and Medgem, anthropometric measurements (body weight, height, body circumferences and


653 Subject 1

Resting Equilibrating

Resting

Deltatrac1

MedGem1

Deltatrac2

MedGem2

Deltatrac3

MedGem3

Equilibrating

MedGem1

Deltatrac1

MedGem2

Deltatrac2

MedGem3

Deltatrac3

under the hood Subject 2

Resting Resting

under the hood Time (min)

5

15

25

35 37

47 49

59 61

71 73

83 85

95

Figure 1 Time schedule of each measurement session; two subjects were simultaneous measured; after 25 min the RMR measurements started; RMR was measured 6 times: 3 times by MedGem and 3 times by Deltatrac device; subsequently anthropometric measurements were taken.

skinfolds) were made. At the end of the first session, socioeconomic and demographic data were collected by a structured questionnaire. Three local Bengali women with 12 grade or higher education were trained to assist in this study. They were responsible for inviting the subjects to participate, to assist the investigators (MM, DR) in explaining the measurement procedures and in filling in the questionnaire. The study was conducted in March and April 2003, the spring/summer season in Bangladesh. Room temperature at the clinic in this period was between 24 and 301C.

Measurement of oxygen consumption Oxygen consumption was measured using a Deltatrac II metabolic monitor (Datex Ohmeda, Helsinki, Finland) and by a handheld MedGem device (HealtheTech Inc., Golden, CO, USA). The Deltatrac metabolic monitor is an open system indirect calorimeter (ventilated hood) for measurement of both oxygen consumption and carbon dioxide production. It consists of a flow meter, a paramagnetic oxygen sensor and an infrared sensor for carbon dioxide measurement. The Deltatrac ventilated hood system was used in canopy mode with a flow rate of 40 l/min. Each morning, the system was calibrated using a mixture of oxygen/carbon dioxide (QuickCalt, Datex-Ohmeda, Helsinki, Finland). The device was set to record oxygen consumption (ml/min) and carbon dioxide production (ml/min) every 1 min. To check the validity of the device, before and after the study period 5 ml 96% ethanol (v/v) was burned from which oxygen consumption and carbon dioxide production was measured. The MedGem is a handheld open circuit indirect calorimeter consisting of an oxygen sensor and a flow sensor to measure oxygen consumption from which energy expenditure is estimated. The MedGem can be used in combination with a facemask or mouthpiece. It is a self-calibrating device. The oxygen consumption is measured untill it reaches a stable flow, this may take 5–10 min. Oxygen consumption in this study was measured using a mouthpiece and a nose clip to prevent breathing through the nose. The MedGem was auto-calibrated prior to each measurement. Within each session, a single disposable mouthpiece was used for each woman. RMR was calculated for both the Deltatrac data and the MedGem data by using the following equation: EE (kJ/ min) ¼ 16.3 VO2 (ml/min) þ 4.6 VCO2 (ml/min) (Weir, 1949).

For the MedGem data, a constant respiratory quotient of 0.85 was used to estimate carbon dioxide production.

Anthropometry Height of the women was measured to the nearest 0.1 cm using a stadiometer. Weight was measured to the nearest 0.1 kg using a digital weighing scale (SECA, Hamburg, Germany). Both height and weight were measured in duplicate without shoes. For body weight a correction of 0.5 kg was made for clothes. Biceps, triceps, subscapular and suprailiac skinfolds were measured in triplicate to the nearest 0.2 mm using Holtain calipers (Holtain Ltd., Crymych, UK). Body fat percentages were calculated from these measurements using the Durnin– Womersley tables (Durnin & Womersley, 1974). Waist, hip and upper arm circumference were measured to the nearest 0.1 cm using a flexible nonstretch tape. Hip circumference was taken over the petticoats of the women.

Prediction of BMR BMR was predicted from the FAO/WHO/UNU equations based on body weight: BMR ¼ 0.062 body weight (kg) þ 2.036 for women of 18–29 y, or BMR ¼ 0.034 body weight (kg) þ 3.538 for women of 30–59 y (FAO/WHO/UNU, 1985).

Data screening Extreme ranges in the oxygen measurements (range of three measurements) were evaluated by testing homogeneity of variation (Wernimont, 1985). Outlying mean oxygen values (mean of three measurements) were evaluated against the distribution of all mean values. Oxygen values more than 3s.d. away from the grand mean were considered as outliers. Normality of the distribution of the mean oxygen consumption of the three measurements was determined using the Kolmogorov–Smirnov test.

Statistical analysis Analysis of variance for repeated measurements was used to determine whether a time effect within or between sessions was present and to assess reproducibility. Within- and between-session reproducibility was determined by comparing European Journal of Clinical Nutrition


654 the within-subject variances of the oxygen consumption measured by MedGem and Deltatrac. Systematic differences in oxygen consumption measured by the two devices were determined using the statistical approach outlined by Bland and Altman (1986). The same procedure was used to compare the measured RMR with the predicted BMR. All statistical analyses were performed using SPSS, version 11.0 for Windows (SPSS Inc., Chicago, IL, USA).

increases for higher values of oxygen consumption (Figures 2 and 3). Within-subject variation in oxygen consumption is shown in Table 3. Within-subject within-session reproducibilities presented as coefficient of variation of the MedGem and Deltatrac device were around 9 and 3%, respectively

Results

Table 2 Resting oxygen consumption (ml/min) measured with MedGemt and Deltatract device MedGemt (n ¼ 36)a

Measurement 1 Measurement 2 Measurement 3 Mean

Deltatract (n ¼ 37)

Session 1

Session 2

Session 1

Session 2

b

200734 199728 201727 200726

183720 184720 185720 184720

177718 178720 181718* 178718**

200732 198736 202732 200730

a

One missing value. Mean7s.d. *Significantly different from the first measurement in the same session (P ¼ 0.013). **Significantly different from Deltatrac value in session 1 (P ¼ 0.007).

b

European Journal of Clinical Nutrition

MedGem minus Deltatrac VO2 (mL/min)

50.00

25.00

0.00

-25.00 175.00

200.00

225.00

250.00

Mean MedGem and Deltatrac VO2 (mL/min)

Figure 2 Bland and Altman plot of oxygen consumption measured with the MedGemt and Deltatract for session 1. VO2MedGem minus Deltatrac ¼ 79.32 þ 0.51 mean VO2MedGem and Deltatrac.

0.00

25.00

50.00 MedGem minus Deltatrac VO2 (mL/min)

Data were collected from 37 women using the Deltatrac device and the MedGem device. Screening of the data revealed no outlying values, so all data were included into the statistical analysis. From one woman no MedGem data were collected. A small but statistically significant within-session time effect was observed for the Deltatrac measurement in session 2 (Table 2). Because of the small size of the change in oxygen consumption (4 ml/min, 2%), we still decided to use the mean of the three measurements of session 2 for further analysis. Mean oxygen consumptions measured with the Medgem device were 200730 and 200726 ml/min for session 1 and 2, respectively (Table 2). Mean oxygen consumptions measured with the Deltatrac device for these sessions were 184720 and 178718 ml/min, respectively. As a result of the small but statistically significant difference in the Deltatrac measurements of session 1 and 2, it was decided to analyse the data for both sessions separately. Pearson’s correlation coefficients (r) of oxygen consumption measured by Deltatrac and MedGem were 0.80 (Po0.01) for session 1 and 0.75 (Po0.01) for session 2. Figures 2 and 3 show Bland and Altman plots for measurement of oxygen consumption by the two devices. These plots reveal a leveldependent bias for both session 1 (Po0.01) and session 2 (Po0.01), indicating that the difference in oxygen consumption measured with the two devices is dependent on the level of oxygen consumption: the MedGem measures higher values than the Deltatrac (Table 2) and the difference

-25.00 175.00

200.00

225.00

250.00

Mean MedGem and Deltatrac VO2(mL/min) Figure 3 Bland and Altman plot of oxygen consumption measured with the MedGemt and Deltatract for session 2. VO2MedGem minus Deltatrac ¼ 61.89 þ 0.45 mean VO2MedGem and Deltatrac.


655 (Po0.01). Within-subject between-session reproducibilities were 8.2 and 4.5% for the MedGem and Deltatrac device, respectively (Po0.01). RMRs calculated from Weir’s formula and predicted from Schofield equations are shown in Table 4. In agreement with the measured differences in oxygen consumption, RMR measured by MedGem is higher (5.82 MJ/day) compared to the Deltatrac measurements (5.35 and 5.17 MJ/day, respectively). For the MedGem, the difference between measured and predicted RMR was dependent on the RMR level as shown by the regression coefficients obtained by Bland and Altman analysis: b1 ¼ 0.97 (P ¼ 0.000) for session 1, and b1 ¼ 0.68 (P ¼ 0.000) for session 2. For the Deltatrac, no leveldependent difference between measured RMR and predicted RMR was observed. The mean RMR measured by Deltatrac was not different from the BMR estimated by the FAO/WHO/ UNU prediction equations for session 2. For session 1, the RMR measured by the Deltatrac were slightly higher than the predicted BMR; however, this was only 3.7% (0.19 MJ/day; P ¼ 0.031).

Discussion The main aims of the study were firstly to assess reproducibility and relative validity of oxygen consumption using the MedGem metabolic monitor in a rural setting and secondly to evaluate the FAO/WHO/UNU equations in adequately predicting BMR in rural Bangladeshi women.

Table 3 Within-subject variation in resting oxygen consumption, measured within and between sessions MedGem (n ¼ 36)

Deltatrac (n ¼ 37)

s.d. (ml/min)/(CV%)

s.d. (ml/min)/(CV%)

Within session 1 Within session 2 Between sessions

a

19.2 (9.6) 16.6 (8.3) 16.3 (8.2)a

P-value

4.5 (2.5) 6.3 (3.6) 8.1 (4.5)b

0.000 0.000 0.000

a

n ¼ 35. Taking into account the small but statistically significant difference in oxygen consumption between measurement 1 and 3 (see Table 2).

b

The women participating in this study had a mean body mass index (BMI) of 20.8 kg/m2 and a mean height of 1.54 m. In all, 11 women (30%) had a BMI o18.5 kg/m2, while two of them had a BMI o17 kg/m2, indicating occurrence of mild malnourishment. These women came from villages within 3 km of the ICDDR,B hospital. The socioeconomic and healthcare situation in that area is probably better developed than in the rest of the Matlab area. It cannot be excluded that the nutritional status of the women in our study is better than the actual nutritional status in the whole Matlab area. However, degree of leanness has no influence on the comparison between the two devices made in this study, since the subjects were subdued to the same standardized protocol.

Relative validity of the MedGem Oxygen consumption measured in the present study by MedGem was on average 10% higher compared to the Deltatrac measurements (200 ml/min vs 184 and 178 ml/ min, Table 2). This difference was dependent on the level of oxygen consumption and the difference increased at higher levels of oxygen consumption (Figures 2 and 3). In previous studies, the Deltatrac metabolic monitor was found to be a valid instrument to assess oxygen consumption (Takala et al, 1989; Weissman et al, 1990; Tissot et al, 1995). During the measurement sessions, the Deltatrac device was calibrated each day using a well-defined mixture of oxygen/carbondioxide. In addition, before and after the study period an alcohol calibration was performed, showing that performance of the Deltatrac measurement was stable during the study period and that the release of carbon dioxide and consumption of oxygen were within the expected limits. Nevertheless, we found a slight but significant increase of 2% in oxygen consumption (4 ml/min) for the Deltatrac measurements within session 2. No clear explanation was found for this time effect. We consider the relevance of this difference to be limited; therefore, the mean of the three Deltatrac measurements in session 2 was used in statistical analyses. In session 2, the oxygen consumption measured by Deltatrac was lower than in session 1. This is probably due

Table 4 RMR measured with the Deltatrac and MedGem device and BMR predicted from Schofield equation MedGem (n ¼ 36) RMRa (MJ/day) Session 1 Session 2

5.8270.85 5.8270.76

Deltatrac (n ¼ 37) RMRa (MJ/day) 5.3570.54 5.1770.51

Prediction equation

RQ 0.85 0.83

BMRa (MJ/day) 5.1670.42

b

MedGem minus prediction equation (n ¼ 36) Differencea (MJ/day) 0.6870.82 0.6870.58

P* c

NA NAc

Deltatrac minus predicion equation (n ¼ 37) Differencea (MJ/day)

P*

0.1970.52 0.0170.45

0.031 0.869

*P-value of paired t-test. a Mean7s.d. b BMR (18–29 y) ¼ 0.062 body weight þ 2.036; BMR (30–59 y) ¼ 0.034 body weight þ 3.538, (FAO/WHO/UNU, 1985). c Not applicable because the difference is dependent on the level of oxygen consumption (as indicated by Bland and Altman plot; not presented).



656 to familiarization with the study procedure. It is conceivable that the women were more relaxed and less anxious during the second session, which could have caused the energy expenditure to decrease. A reason why such a familiarization was not observed for measurements with the Medgem device could be due to the lack of sensitivity of the Medgem device, expressed as a relative high coefficient of variation. Nieman et al (2003) evaluated the BodyGem metabolic monitor against the Douglas bag technique in 43 women and 20 men with a body-mass of 26.576.6 kg/m2 (BMI range: 19.1–56.2 kg/m2). The BodyGem is a portable indirect calorimeter identical to the Medgem but it shows only RMR on the display. Correlation coefficients for oxygen consumption between BodyGem and Douglas bag technique ranged from 0.81 to 0.87 for single tests during the two sessions. In our study, correlation coefficients between MedGem and Deltatrac were 0.80 and 0.75 for session 1 and 2, respectively. In the study of Nieman and co-workers, no significant differences in oxygen consumption were reported between BodyGem (mean: 241 ml/min) and Douglas bag technique (mean: 240 ml/min) over the whole range of oxygen consumption and across the three BMI categories (o25, 25–30, Z30 kg/m2). The difference in the absolute oxygen consumption measured by the MedGem and Deltatrac in our study is not acceptable in applications where accurate estimates of RMR are necessary such as in estimations of energy requirements of individuals and populations.

Reproducibility of the MedGem Within-subject variation (within and between sessions) was significantly lower for the Deltatrac measurements compared to the Medgem measurements (Table 3). The observed variation in measuring RMR using the Deltatrac (range: 2.5– 4.5%) is comparable with other studies using the Deltatrac (Ventham & Reilly, 1999) and in agreement with the generally observed measurement variability in RMR (Shetty et al, 1996). The observed within-subject coefficient of variation (range: 8.2–9.6%) of the Medgem device was higher compared to the value reported by the manufacturer of the BodyGem device: in a mechanical ventilation study a within-day coefficient of variation of 1.45% was reported (Healthetech, 2001b). All women who participated in our study reported that the measurements made with the Deltatrac were more comfortable than the measurements made with the MedGem. The discomfort most women experienced using the Medgem were pain of the nose because of the noseclip, a dry throat, difficulties with breathing and inconvenience holding the MedGem device while lying in a supine position. This could partly explain the higher variation and increased oxygen demand during measurements with the MedGem device. However, this may not explain the finding that the oxygen consumption measured with the MedGem was dependent on the level of oxygen consumption. The reproducibility of oxygen measurements and thus RMR has implications for the number of European Journal of Clinical Nutrition

RMR measurements required to get a precise estimate and for the ability to detect changes in RMR (Figueroa-Colon et al, 1996). In studies using a MedGem, multiple assessments of RMR would be necessary or more subjects need to be included to detect physiologically relevant changes in RMR, compared to measurements with the Deltatrac metabolic monitor. In addition, the discomfort reported by the women in this study population may limit application of the MedGem device in critically ill subjects.

Prediction of RMR In the present study, RMR was measured after an overnight fast, with the women in supine position after 25 min of rest early in the morning, so RMR was measured close to BMR condition. Comparison of RMR measured by Deltatrac with the FAO/WHO/UNU prediction equations (FAO/WHO/UNU, 1985) showed that the measured BMR was accurately predicted by the FAO/WHO/UNU equations in session 2 and slightly underestimated by 0.19 MJ/day in session 1. Because of a level-dependent bias of the RMR measurements using the MedGem device, only the RMR measurements by Deltatrac device were compared with the FAO/WHO/UNU equations. The FAO/WHO/UNU equations are based on an extensive literature review by Schofield et al (1985) involving more than 100 published studies of BMR obtained from mainly European and North American subjects. It was adopted as a new approach to base energy requirements on energy expenditure and to express these requirements in terms of multiples of BMR. The use of FAO/WHO/UNU equations to predict BMR of tropical populations has been challenged by Henry and Rees (1991) and Hayter and Henry (1994) among others. Henry and Rees (1991) reviewed BMR studies conducted in tropical countries and concluded that the Schofield equations overestimate BMR in adults by 3.8% (women, 18–30 y) to 11.2% (men, 30–60 y). It has been suggested that metabolic adaptation of subjects may occur due to chronic marginal or severe energy deprivation leading to a lower BMR and hence overestimation by prediction equations (Shetty, 1984). However, several studies performed during the last decade support the view that BMRs of people in the tropics are not different from those in temperate regions (Shetty et al, 1996). Ferro-Luzzi and co-workers studied the effect of chronic energy undernutrition on energy turnover in 88 men and 90 women from Bangalore (India) and found no evidence for metabolic adaptation in subjects with chronic energy deficiency. The measured BMR was accurately predicted by the FAO/WHO/UNU equations, regardless of nutritional status and so this study did not confirm a discrepancy between measured and predicted BMR in subjects from tropical countries (Ferro-Luzzi et al, 1997). It can be concluded that the observed reproducibility of the Medgem applied in a rural setting is poor and the validity of the instrument for measuring the RMR of Bangladeshi women is questioned. Although the MedGem is an affordable handheld device, the current reliability impedes its use


657 as a first choice instrument for measuring RMR. Improvement of the technology deserves further attention. We also conclude that the FAO/WHO/UNU prediction equations give a good estimation of the BMR of rural, nonpregnant, nonlactating Bangladeshi women of 18–35 y old, although it should be realized that the women in this study may not belong to the most disadvantaged groups in the Matlab area. Acknowledgements This research was supported by the Poverty and Health grant of the Department for International Development (DFID), UK (Grant Number 00230) to ICDDR,B: Centre for Health and Population Research. ICDDR,B acknowledges with gratitude the commitment of DFID to the Centre’s research effort. We are grateful to the ICDDR, B staff that gave full support to this study, and to Shamsun Naher, Rokeya Akter, Rashida Akter and Ramjan Ali for their valuable contribution to data collection.

References ACC/SCN (1997): Nutrition and Poverty Nutrition policy paper no. 16. Kathmandu: ACC/SCN. Alam DS, van Raaij JMA, Hautvast JGAJ, Yunus M & Fuchs GJ (2003): Energy stress during pregnancy and lactation: consequences for maternal nutrition in rural Bangladesh. Eur. J. Clin. Nutr. 57, 151–156. Bland JM & Altman DG (1986): Statistical methods for assessing agreement between two methods of clinical measurement. Lancet i, 307–310. Durnin JVGA & Womersley J (1974): Body fat assessed from total body density and its estimation from skinfold thickness: measurements on 481 men and women aged from 16–72 years. Br. J. Nutr. 32, 77–97. FAO (1999): Bangladesh Nutrition country profiles. Rome: FAO. FAO/WHO/UNU (1985): Energy and Protein Requirements. Technical Report Series no. 724. Geneva: WHO. Fauveau V (1994): Neonatal and perinatal health in Matlab. In Matlab: Women, Children and Health, ed. V Fauveau, pp 139–160. Dhaka: ICDDR,B. 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. Figueroa-Colon R, Franklin FA, Goran MI, Lee JY & Weinsier RL (1996): Reproducibility of measurement of REE in pre-pubertal girls. Am. J. Clin. Nutr. 64, 533–536.

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. Healthetech (2001a): MedGem Indirect Calorimeter. Operator’s manual. Golden: Healthetech. Healthetech (2001b): Comparison of the BodyGem to a Mechanical Simulation Device. Healthetech technical report 002.http:// www.healthetech.com/corp/downloads/technical002.pdf(accessed 15 November 2002). Henry CJK & Rees DG (1991): New predictive equations for the estimation of basal metabolic rate in tropical peoples. Eur. J. Clin. Nutr. 45, 177–185. ICDDR,B (2002): Nutrition. In ICDDR,B: Centre for Health and Population Research. Annual Report 2001 pp 36–43. Dhaka: ICDDR,B. Nieman DC, Trone GA & Austin MD (2003): A new handheld device for measuring resting metabolic rate and oxygen consumption. J. Am. Diet. Assoc. 103, 588–593. Osendarp SJM, van Raaij JMA, Darmstadt GL, Baqui AH, Hautvast JGAJ & Fuchs GJ (2001): Zinc supplementation during pregnancy and effects on growth and morbidity in low birthweight infants: a randomised placebo controlled trial. Lancet 357, 1080–1085. Schofield WN, Schofield C & James WPT (1985): Basal metabolic rate—review and prediction, together with an annotated bibliography of source material. Hum. Nutr. Clin. Nutr. 39C (Suppl 1), 5–41. Shetty PS (1984): Adaptive changes in basal metabolic rate and lean body mass in chronic under nutrition. Hum. Nutr. Clin. Nutr. 38C, 443–451. Shetty PS, Henry CJK, Black AE & Prentice AM (1996): Energy requirements of adults: an update on basal metabolic rates (BMRs) and physical activity levels (PALs). Eur. J. Clin. Nutr. 50 (Suppl 1), S11–S23. Takala J, Keina¨nen O, Va¨isa¨nen P & Kari A (1989): Measurement of gas exchange in intensive care; laboratory and clinical validation of a new device. Crit. Care Med. 17, 1041–1047. Tissot S, Delafosse B & bertrand O (1995): Clinical validation of the Deltatrac monitoring system in mechanically ventilated patients. Intern. Care Med. 21, 149–153. Ventham JC & Reilly JJ (1999): Reproducibility of resting metabolic rate measurement in children. Br. J. Nutr. 81, 435–437. Weir JB de V (1949): New methods for calculating metabolic rate with special reference to protein metabolism. J. Physiol. 109, 1–9. Weissman C, Sardan A & Kemper M (1990): In vitro evaluation of a compact metabolic measurement instrument. J. Parenter. Enteral. Nutr. 14, 216–221. Wernimont GT (1985): Use of Statistics to Develop and Evaluate Analytical Methods: Interlaboratory Evaluation of an Analytical Process. Arlington: AOAC International.
