Accurate assessment of energy expenditure (EE) is important for the study of energy balance and metabolic disorders. Combined heart rate (HR) and acceleration (ACC) sensing may increase precision of physical activity EE (PAEE) which is the most variable component of total EE (TEE).To evaluate estimates of EE using ACC and HR data with or without individual calibration against doubly-labelled water (DLW) estimates of EE.23 women and 23 men (22-55 yrs, 48-104 kg, 8-46%body fat) underwent 45-min resting EE (REE) measurement and completed a 20-min treadmill test, an 8-min step test, and a 3-min walk test for individual calibration. ACC and HR were monitored and TEE measured over 14 days using DLW. Diet-induced thermogenesis (DIT) was calculated from food-frequency questionnaire. PAEE (TEE ÷ REE ÷ DIT) and TEE were compared to estimates from ACC and HR using bias, root mean square error (RMSE), and correlation statistics.Mean(SD) measured PAEE and TEE were 66(25) kJ·day(-1)·kg(-1), and 12(2.6) MJ·day(-1), respectively. Estimated PAEE from ACC was 54(15) kJ·day(-1)·kg(-1) (p<0.001), with RMSE 24 kJ·day(-1)·kg(-1) and correlation r = 0.52. PAEE estimated from HR and ACC+HR with treadmill calibration were 67(42) and 69(25) kJ·day(-1)·kg(-1) (bias non-significant), with RMSE 34 and 20 kJ·day(-1)·kg(-1) and correlations r = 0.58 and r = 0.67, respectively. Similar results were obtained with step-calibrated and walk-calibrated models, whereas non-calibrated models were less precise (RMSE: 37 and 24 kJ·day(-1)·kg(-1), r = 0.40 and r = 0.55). TEE models also had high validity, with biases <5%, and correlations r = 0.71 (ACC), r = 0.66-0.76 (HR), and r = 0.76-0.83 (ACC+HR).Both accelerometry and heart rate may be used to estimate EE in adult European men and women, with improved precision if combined and if heart rate is individually calibrated.

This study assessed soldier's physical demands and energy balance during the Section Commanders' Battles Course (SCBC). Forty male soldiers were monitored during the 8-week tactics phase of the SCBC. Energy expenditure was measured using the doubly labeled water method. Cardiovascular strain (heart rate) and physical activity (using triaxial accelerometer) were also monitored. Average sized portions of meals were weighed, with all recipes and meals entered into a dietary analysis program to calculate the calorie content. Energy expenditure averaged 19.6 ± 1.8 MJ · d(-1) in weeks 2 to 3 and 21.3 ± 2.0 MJ · d(-1) in weeks 6 to 7. Soldiers lost 5.1 ± 2.6 kg body mass and body fat percent decreased from 23 ± 4% to 19 ± 5%. This average weight loss equates to an estimated energy deficit of 2.69 MJ · d(-1). The Army provided an estimated 14.0 ± 2.2 MJ · d(-1) in weeks 2 to 3 and 15.7 ± 2.2 MJ · d(-1) in weeks 6 to 7. Although this provision adheres to the minimum requirement of 13.8 MJ · d(-1) set by Army regulations, soldiers were in a theoretical 5.6 MJ · d(-1) energy deficit. The physical demands of SCBC were high, and soldiers were in energy deficit resulting in loss in body mass; primarily attributed to a loss in fat mass.

OBJECTIVE We sought to assess the association between maternal gestational physical activity and insulin action and body composition in early infancy. RESEARCH DESIGN AND METHODS At 28–32 weeks' gestation, pregnant women participating in an observational study in Sweden underwent assessments of height, weight, and body composition, an oral glucose tolerance test, and 10 days of objective physical activity assessment. Thirty mothers and infants returned at 11–19 weeks postpartum. Infants underwent assessments of weight, length, and body composition. RESULTS Early insulin response was correlated with total physical activity (r = −0.47; P = 0.007). Early insulin response (r = −0.36; P = 0.045) and total physical activity (r = 0.52; P = 0.037) were also correlated with infant fat-free mass. No maternal variable was significantly correlated with infant adiposity. CONCLUSIONS The relationships between maternal physical activity, insulin response, and infant fat-free mass suggest that physical activity during pregnancy may affect metabolic outcomes in the mother and her offspring.

Background: Meal-Q and its shorter version, MiniMeal-Q, are 2 new Web-based food frequency questionnaires. Their meal-based and interactive format was designed to promote ease of use and to minimize answering time, desirable improvements in large epidemiological studies. Objective: We evaluated the validity of energy and macronutrient intake assessed with Meal-Q and MiniMeal-Q as well as the reproducibility of Meal-Q. Methods: Healthy volunteers aged 20-63 years recruited from Stockholm County filled out the 174-item Meal-Q. The questionnaire was compared to 7-day weighed food records (WFR; n=163), for energy and macronutrient intake, and to doubly labeled water (DLW; n=39), for total energy expenditure. In addition, the 126-item MiniMeal-Q was evaluated in a simulated validation using truncated Meal-Q data. We also assessed the answering time and ease of use of both questionnaires. Results: Bland-Altman plots showed a varying bias within the intake range for all validity comparisons. Cross-classification of quartiles placed 70%-86% in the same/adjacent quartile with WFR and 77% with DLW. Deattenuated and energy-adjusted Pearson correlation coefficients with the WFR ranged from r=0.33-0.74 for macronutrients and was r=0.18 for energy. Correlations with DLW were r=0.42 for Meal-Q and r=0.38 for MiniMeal-Q. Intraclass correlations for Meal-Q ranged from r=0.57-0.90. Median answering time was 17 minutes for Meal-Q and 7 minutes for MiniMeal-Q, and participants rated both questionnaires as easy to use. Conclusions: Meal-Q and MiniMeal-Q are easy to use and have short answering times. The ranking agreement is good for most of the nutrients for both questionnaires and Meal-Q shows fair reproducibility.

BackgroundIncreased use of the Internet provides new opportunities for collecting data in large studies. The aim of our new Web-based questionnaire, Active-Q, is to assess total physical activity and inactivity in adults. Active-Q assesses habitual activity during the past year via questions in four different domains: (1) daily occupation, (2) transportation to and from daily occupation, (3) leisure time activities, and (4) sporting activities. ObjectiveThe objective of our study is to validate Active-Q’s energy expenditure estimates using the doubly labeled water (DLW) method, and to assess the reproducibility of Active-Q by comparing the results of the questionnaire completed by the same group on two occasions. MethodsThe validity and reproducibility of Active-Q were assessed in a group of 37 individuals, aged 20 to 65 years. Active-Q was distributed via email to the participants. The total energy expenditure of the participants was assessed using DLW for 11 consecutive days. ResultsThe median time to complete Active-Q was 6.1 minutes. The majority of participants (27/37, 73%) reported that the questionnaire was “easy” or “very easy” to answer. On average, Active-Q overestimated the total daily energy expenditure by 440 kJ compared with the DLW. The Spearman correlation between the two methods was r = 0.52 (P < .001). The intraclass correlation coefficient for total energy expenditure between the results of Active-Q completed on two occasions was 0.83 (95% CI 0.73-0.93). ConclusionsActive-Q is a valid and reproducible method of assessing total energy expenditure. It is also a user-friendly method and suitable for Web-based data collection in large epidemiological studies.

This study compared the physical demands and progression of basic training for male and female British Army recruits in single-sex platoons. Thirty male and 30 female recruits were monitored for energy expenditure (EE) (doubly labeled water), physical activity (3-dimensional accelerometry) and cardiovascular strain (percent heart rate reserve) during 6 weeks over the 14-week course. First time pass rate was similar for male (60%) and female (57%) recruits. Average daily percent heart rate reserve (female 31 +/- 4%; male 32 +/- 5%), physical activity levels (female 2.2 +/- 0.2; male 2.3 +/- 0.2) and percentage improvements in 2.4-km run time (female 10 +/- 4%; male 10 +/- 5%) were similar for both sexes (p > 0.05), although male recruits had 12% higher physical activity counts (p < 0.01). Although the absolute physical demands of basic training were greater for male recruits, the relative cardiovascular strain experienced was similar between sexes.

Background Overweight and obesity during pregnancy raise the risk of gestational diabetes and birth complications. Lifestyle factors like physical activity may decrease these risks through beneficial effects on glucose homeostasis. Here we examined physical activity patterns and their relationships with measures of glucose homeostasis in late pregnancy compared to non-pregnant women. Methods Normal weight and overweight women without diabetes (N = 108; aged 25-35 years) were studied; 35 were pregnant (in gestational weeks 28-32) and 73 were non-pregnant. Insulin sensitivity and β-cell response were estimated from an oral glucose tolerance test. Physical activity was measured during 10-days of free-living using a combined heart rate sensor and accelerometer. Total (TEE), resting (REE), and physical activity (PAEE) energy expenditure were measured using doubly-labeled water and expired gas indirect calorimetry. Results Total activity was associated with reduced first-phase insulin response in both pregnant (Regression r 2 = 0.11; Spearman r = -0.47; p = 0.007) and non-pregnant women (Regression r 2 = 0.11 Spearman; r = -0.36; p = 0.002). Relative to non-pregnant women, pregnant women were estimated to have secreted 67% more insulin and had 10% lower fasting glucose than non-pregnant women. Pregnant women spent 13% more time sedentary, 71% less time in moderate-to-vigorous intensity activity, had 44% lower objectively measured total activity, and 12% lower PAEE than non-pregnant women. Correlations did not differ significantly for any comparison between physical activity subcomponents and measures of insulin sensitivity or secretion. Conclusions Our findings suggest that physical activity conveys similar benefits on glucose homeostasis in pregnant and non-pregnant women, despite differences in subcomponents of physical activity.

BACKGROUND:Few studies have compared the validity of objective measures of physical activity energy expenditure (PAEE) in pregnant and non-pregnant women. PAEE is commonly estimated with accelerometers attached to the hip or waist, but little is known about the validity and participant acceptability of wrist attachment. The objectives of the current study were to assess the validity of a simple summary measure derived from a wrist-worn accelerometer (GENEA, Unilever Discover, UK) to estimate PAEE in pregnant and non-pregnant women, and to evaluate participant acceptability. METHODS:Non-pregnant (N = 73) and pregnant (N = 35) Swedish women (aged 20-35 yrs) wore the accelerometer on their wrist for 10 days during which total energy expenditure (TEE) was assessed using doubly-labelled water. PAEE was calculated as 0.9×TEE-REE. British participants (N = 99; aged 22-65 yrs) wore accelerometers on their non-dominant wrist and hip for seven days and were asked to score the acceptability of monitor placement (scored 1 [least] through 10 [most] acceptable). RESULTS:There was no significant correlation between body weight and PAEE. In non-pregnant women, acceleration explained 24% of the variation in PAEE, which decreased to 19% in leave-one-out cross-validation. In pregnant women, acceleration explained 11% of the variation in PAEE, which was not significant in leave-one-out cross-validation. Median (IQR) acceptability of wrist and hip placement was 9(8-10) and 9(7-10), respectively; there was a within-individual difference of 0.47 (p<.001). CONCLUSIONS:A simple summary measure derived from a wrist-worn tri-axial accelerometer adds significantly to the prediction of energy expenditure in non-pregnant women and is scored acceptable by participants.

OBJECTIVE To quantify the health, fitness, and physiological responses to military training of Officer Cadets from a Gulf Cooperation Council country. METHODS One hundred and nineteen Officer Cadets volunteered; body composition, core body temperature, aerobic fitness, hydration status (urine osmolality), cardiovascular strain, physical activity (3-dimensional accelerometry), and energy expenditure (doubly labelled water) were measured over 5-days of Basic Training (BT), Army Training (AT), Navy Training (NT), and Air Force Training (AFT). RESULTS There were no differences between courses for body mass index (mean all courses: 24.1 +/- 4.1 kg x m2) or peak core body temperature (mean all courses: 38.1 +/- 0.4 degrees C) (p > 0.05). AT body fat (19.8 +/- 3.6%) and BT VO2 max (36.8 +/- 11.6 mL x kg(-1) x min(-1)) were lower than the other courses (BT, 26.1 +/- 8.1; NT, 26.0 +/- 6.0; AFT, 24.7 +/- 6.1%) and (AT, 44.8 +/- 9.6; NT, 45.0 +/- 7.5; AFT, 44.6 +/- 5.2 mL x kg(-1) x min(-1)), respectively (p < 0.05). NT urine osmolality (979 +/- 90 mOsmol x kg(-1)) was similar to BT (946 +/- 181 mOsmol x kg(-1) p > 0.05) but lower in AT (868 +/- 144 mOsmol x kg(-1), p < 0.05) and AFT (883 +/- 121 mOsmol x kg(-1), p < 0.05). Cardiovascular strain during NT (22 +/- 5% HRR) was lower than other courses (range, 25 +/- 4-29 +/- 3% Heart Rate Reserve) (p < 0.05). Physical activity level during AFT (1.70 +/- 0.18 AU) was lower than other courses (range, 1.86 +/- 0.21-1.92 +/- 0.18 AU) (p > 0.05). CONCLUSION Positive developments were apparent from BT leading into other courses. Potential exists to increase physical training volume on all courses, which may improve participants' aerobic fitness, body composition, and health.

BACKGROUND The increasing burden of non-communicable diseases in sub-Saharan Africa (SSA) warrants rigorous studies of contributing lifestyle factors. Combined heart rate (HR) and movement monitoring make it possible to objectively measure physical activity in free-living individuals. We examined the validity of a combined HR and motion sensor in estimating physical activity energy expenditure (PAEE) in free-living adults in rural and urban Cameroon compared with doubly-labelled water (DLW) as criterion. METHODS PAEE was measured in 33 free-living rural and urban dwellers by DLW over 7 consecutive days. Simultaneously, the combined sensor recorded HR and uni-axial acceleration. Individual HR vs PAEE calibration was done by a step test. Branched equation modelling was used to estimate PAEE from HR and acceleration. Validity and accuracy of prediction were expressed as mean bias and root mean square error (RMSE). Agreement was analysed using Bland and Altman limits of agreement (LOA). RESULTS There was no significant mean bias between PAEE estimated from the combined sensor or measured by DLW [mean bias (standard error): -5.4 (5.1) kJ/kg/day; P = 0.3; RMSE = 29.3 kJ/kg/day]. The bias doubled for group compared with individual calibration of HR [-9.1 (5.0) kJ/kg/day, P = 0.08]. PAEE prediction was more accurate in urban compared with rural volunteers. The 95% LOAs between predicted and measured PAEE were ∼50-60 kJ/kg/day above or below perfect agreement. CONCLUSIONS Combined HR and movement sensing is a valid method for estimating free-living PAEE on group level in adults in SSA.

The Section Commanders' Battles Course (SCBC) is a 15 week infantry skills course for junior non-commissioned officers, renowned for its high physical demands. PURPOSE: To determine if the energy requirements of the SCBC are commensurate with the energy provision from Army catering. METHODS: Forty male soldiers (age 25 ± 3 yr; stature 176 ±7 cm; body mass 81 ± 10 kg) were monitored for five weeks (weeks 2-3 and 6-8) during the final eight weeks on the tactics phase of SCBC. The doubly labelled water (DLW) method was used to measure the daily energy expenditure (EE) of 30 soldiers, averaged over two separate 10 day periods. Total body water calculations were used to determine body composition at the start and end of the eight weeks. Sample meals were weighed in the cook-house and in the field (container meals). Recipes and meals were entered into a dietary analysis programme (Microdiet, Downlee Systems Ltd, UK) to calculate the energy and macro-nutrient content. RESULTS: Energy expenditure averaged 4693 ± 424 kcal.day-1 in weeks 2-3 and 5094 ± 471 kcal.day-1 in weeks 6-7. Soldiers lost 5.1 ± 2.6 kg body mass and body fat reduced from 23 ± 4 % to 19 ± 5 % over the eight weeks (n=40). Body composition estimated for the 30 soldiers in the DLW sample showed that the majority of weight loss was due to loss of fat mass (4.1 kg) rather than fat free mass (0.8 kg) (n=30). These body composition changes equate to an estimated energy deficit of 644 kcal.day-1. The Army provided soldiers with an estimated 3350 ± 529 kcal.day-1 and 3746 ± 517 kcal.day-1 in weeks 2-3 and 6-7, respectively. Although this provision adheres to the minimum requirement of 3300 kcal.day-1 set by Army regulations, the provision places soldiers in a theoretical energy deficit of 1340 kcal.day-1. Actual energy deficit was lower than this presumably because the soldiers supplemented their Army provisions from other sources. Macro-nutrient content for carbohydrate, protein and fat averaged 50 %, 15 % and 35 % for cookhouse meals, and 52 %, 15 % and 38 % for container meals. CONCLUSION: The EE during SCBC was high and in excess of energy provision. Macro-nutrient content of the meals was not optimal; carbohydrate and protein content should be increased and fat content decreased in meals provided by the Army. This work was conducted by Optimal Performance Limited and funded by the UK Ministry of Defence.

Physical activity (PA) has rarely been quantified in adolescent populations undergoing economic transition; therefore relationships with disease still remain uncertain. This study assessed whether absolute PA energy expenditure (PAEE), PAEE/kg, and PAEE/kg FFM could be accurately estimated using accelerometry and a questionnaire in Indian adolescents and how these values compared to those of other populations. PAEE was assessed using doubly labeled water (DLW) in 30 adolescents from Chennai, India, over seven consecutive days, simultaneous with the measurement of PA using accelerometry and a previous‐week recall questionnaire. Accelerometry counts (regression analysis) and questionnaire data were used to estimate PAEE; estimates were cross‐validated using the Bland‐Altman method. Accelerometry data and DLW‐derived PAEE were visually compared to values from four North American and European populations. For boys, 49% of the variance in DLW‐derived PAEE was explained with an equation including accelerometry counts and fat‐free mass (FFM). Questionnaire‐derived estimates did not contribute to the explained variance in DLW derived PAEE. The group‐level PA of these Indian adolescents was successfully assessed using accelerometry, but not questionnaire. DLW‐derived PAEE/kg FFM (mean (s.d.): 53.0 (27.5) kJ/kg FFM /day) was lower in this group than other adolescent populations in Europe and similar to those in North America. Additionally, four boys and none of the girls accumulated ≥60 min/day of accelerometry‐derived moderate intensity activity, indicating low levels of PAEE and PA in these adolescents. Further research is necessary to investigate the association between PA and health outcomes in Indian adolescents.

Lack of physical activity may be an important etiological factor in the current epidemiological transition characterized by increasing prevalence of obesity and chronic diseases in sub-Sahara Africa. However, there is a dearth of data on objectively measured physical activity energy expenditure (PAEE) in this region. We sought to develop regression equations using body composition and accelerometer counts to predict PAEE. We conducted a cross-sectional study of 33 adult volunteers from an urban (n = 16) and a rural (n = 17) residential site in Cameroon. Energy expenditure was measured by doubly labeled water (DLW) over a period of seven consecutive days. Simultaneously, a hip-mounted Actigraph accelerometer recorded body movement. PAEE prediction equations were derived using accelerometer counts, age, sex, and body composition variables, and cross-validated by the jack-knife method. The Bland and Altman limits of agreement (LOAs) approach was used to assess agreement. Our results show that PAEE (kJ/kg/day) was significantly and positively correlated with activity counts from the accelerometer (r = 0.37, P = 0.03). The derived equations explained 14-40% of the variance in PAEE. Age, sex, and accelerometer counts together explained 34% of the variance in PAEE, with accelerometer counts alone explaining 14%. The LOAs between DLW and the derived equations were wide, with predicted PAEE being up to 60 kJ/ kg/day below or above the measured value. In summary, the derived equations performed better than existing published equations in predicting PAEE from accelerometer counts in this population. Accelerometry could be used to predict PAEE in this population and, therefore, has important applications for monitoring population levels of total physical activity patterns.

BACKGROUND:It is unclear whether it is possible to accurately estimate physical activity energy expenditure (PAEE) by self-report in youth. OBJECTIVE:We assessed the validity and reliability of 4 self-reports to assess PAEE and time spent at moderate and vigorous intensity physical activity (MVPA) over the previous week in British young people between 4 and 17 y of age. DESIGN:PAEE and MVPA were derived from the Children's Physical Activity Questionnaire, Youth Physical Activity Questionnaire, and Swedish Adolescent Physical Activity Questionnaire; a lifestyle score indicative of habitual activity was derived from the Child Heart and Health Study in England Questionnaire. These data were compared with criterion methods, PAEE, and MVPA derived from simultaneous measurements by doubly labeled water and accelerometry in 3 age groups: 4-5 y (n = 27), 12-13 y (n = 25), and 16-17 y (n = 24). Validity was assessed by using Spearman correlations and the Bland-Altman method, and reliability was assessed by using intraclass correlation coefficients. RESULTS:The strength of association between questionnaire and criterion methods varied (r = 0.09 to r = 0.46). Some questionnaires were able to accurately assess group-level PAEE and MVPA for some age groups, but the error was large for individual-level estimates throughout. Reliability of the Youth Physical Activity Questionnaire and Child Heart and Health Study in England Questionnaire was good (intraclass correlation coefficient: 0.64-0.92). CONCLUSIONS:Absolute PAEE and MVPA estimated from these self-reports were not valid on an individual level in young people, although some questionnaires appeared to rank individuals accurately. Age (the outcome of interest) and whether individual or group-level estimates are necessary will influence the best choice of self-report method when assessing physical activity in youth.

By the repeated use of the doubly labeled water method (DLW), this study aimed to investigate (1) the extent of changes in energy expenditure and physical activity level (PAL) in response to increased agricultural work demands, and (2) whether the seasonal work demands induce the changes in the fairly equitable division of work and similarity of energy needs between men and women observed in our previous study (Phase 1 study; Kashiwazaki et al., 1995 : Am J Clin Nutr 62: 901–910). In a rural small agropastoral community of the Bolivian Andes, we made the follow‐up study (Phase 2, 14 adults; a time of high agricultural activity) of the Phase 1 study (12 adults; a time of low agricultural activity). In the Phase 2 study, both men and women showed very high PAL (mean±SD), but there was no significant difference by sex (men; 2.18 ± 0.23 (age; 64 ± 11 years, n = 7), women; 2.26 ± 0.25 (63 ± 10 years, n = 7)). The increase of PAL by 11% ( P = 0.023) in the Phase 2 was equally occurred in both men and women. The factorial approach underestimated PAL significantly by ≈15% ( P < 0.05). High PAL throughout the year ranging on average 2.0 and 2.2 was attributable to everyday tasks for subsistence and domestic works undertaking over 9–11 h (men spent 2.7 h on agricultural work and 4.7 h on animal herding, whereas women spent 7.3 h almost exclusively on animal herding). The seasonal increase in PAL was statistically significant, but it was smaller than those anticipated from published reports. A flexible division of labor played an important role in the equitable energetic increase in both men and women. Am. J. Hum. Biol., 2009. © 2009 Wiley‐Liss, Inc.

Objective: To evaluate a novel quantitative magnetic resonance (QMR) methodology (EchoMRI-AH, Echo Medical Systems) for measurement of whole-body fat and lean mass in humans.Methods and Procedures: We have studied (i) the in vitro accuracy and precision by measuring 18 kg Canola oil with and without 9kg water (ii) the accuracy and precision of measures of simulated fat mass changes in human subjects (n = 10) and (iii) QMR fat and lean mass measurements compared to those obtained using the established 4-compartment (4-C) model method (n = 30).Results: (i) QMR represented 18kg of oil at 40 degrees C as 17.1 kg fat and 1 kg lean while at 30 degrees C 15.8kg fat and 4.7kg lean were reported. The s.d. of repeated estimates was 0.13 kg for fat and 0.23 kg for lean mass. Adding 9 kg of water reduced the fat estimates, increased misrepresentation of fat as lean, and degraded the precision. (ii) the simulated change in the fat mass of human volunteers was accurately represented, independently of added water. (iii) compared to the 4-C model, QMR underestimated fat and over-estimated lean mass. The extent of difference increased with body mass. The s.d. of repeated measurements increased with adiposity, from 0.25kg (fat) and 0.51 kg (lean) with BMI <25kg/m(2) to 0.43 kg and 0.81 kg respectively with BMI >30 kg/m(2).Discussion: EchoMRI-AH prototype showed shortcomings in absolute accuracy and specificity of fat mass measures, but detected. simulated body composition change accurately and with precision roughly three times better than current best measures. This methodology should reduce the study duration and cohort number needed to evaluate anti-obesity interventions.

Measurements of body composition are crucial in identifying HIV-infected patients at risk of malnutrition. No information is available on the validity of indirect body composition methods in African HIV-infected outpatients. Our first aim was to test the validity of fifteen published equations, developed in whites, African-Americans and/or Africans who were or not HIV-infected, for predicting total body water (TBW) from bioelectrical impedance analysis (BIA) in HIV-infected patients. The second aim was to develop specific predictive equations. Thirty-four HIV-infected patients without antiretroviral treatment and oedema at the beginning of the study (age 39 (sd 7) years, BMI 18·7 (sd 3·7) kg/m2, TBW 30·4 (sd 7·2) kg) were measured at inclusion then 3 and 6 months later. In the resulting eighty-eight measurements, we compared TBW values predicted from BIA to those measured by 2H dilution. Range of bias values was 0·1–4·3 kg, and errors showed acceptable values (2·2–3·4 kg) for fourteen equations and a high value (10·4) for one equation. Two equations developed in non-HIV-infected subjects showed non-significant bias and could be used in African HIV-infected patients. In the other cases, poor agreement indicated a lack of validity. Specific equations developed from our sample showed a higher precision of TBW prediction when using resistance at 1000 kHz (1·7 kg) than at 50 kHz (2·3 kg), this latter precision being similar to that of the valid published equations (2·3 and 2·8 kg). The valid published or developed predictive equations should be cross-validated in large independent samples of African HIV-infected patients.