P. Deurenberg

The validity of predicted body fat percentage from body mass index and from impedance in samples of five European populations

Objectives: To test and compare the validity of a body mass index (BMI)-based prediction equation and an impedance-based prediction equation for body fat percentage among various European population groups.Design: Cross-sectional observational study.Settings: The study was performed in five different European centres: Maastricht and Wageningen (The Netherlands), Milan and Rome (Italy) and Tampere (Finland), where body composition studies are routinely performed.Subjects: A total of 234 females and 182 males, aged 18–70 y, BMI 17.0–41.9 kg/m2.Methods: The reference method for body fat percentage (BF%REF) was either dual-energy X-ray absorptiometry (DXA) or densitometry (underwater weighing). Body fat percentage (BF%) was also predicted from BMI, age and sex (BF%BMI) or with a hand-held impedance analyser that uses in addition to arm impedance weight, height, age and sex as predictors (BF%IMP).Results: The overall mean (±s.e.) bias (measured minus predicted) for BF%BMI was 0.2±0.3 (NS) and−0.7±0.3 (NS) in females and males, respectively. The bias of BF%IMP was 0.2±0.2 (NS) and 1.0±0.4 (P<0.01) for females and males, respectively. There were significant differences in biases among the centres. The biases were correlated with level of BF% and with age. After correction for differences in age and BF% between the centres the bias of BF%BMI was not significantly different from zero in each centre and was not different among the centres anymore. The bias of BF%IMP decreased after correction and was significant from zero and significant from the other centres only in males from Tampere. Generally, individual biases can be high, leading to a considerably misclassification of obesity. The individual misclassification was generally higher with the BMI-based prediction.Conclusions: The prediction formulas give generally good estimates of BF% on a group level in the five population samples, except for the males from Tampere. More comparative studies should be conducted to get better insight in the generalisation of prediction methods and formulas. Individual results and classifications have to be interpreted with caution.European Journal of Clinical Nutrition (2001) 55, 973–979

How fat is obese

Abstract.The aim of the study was a comparison between body fat measurements and body mass index. We analyzed the data of 890 subjects, 596 females and 294 males, ranging in age from 18 to 83 years, in body mass index (BMI) from 14 to 54 kg/m2, and in body fat percentage (BF%) from 4% to 57%. A considerable number of subjects, both males and females, could not be classified as obese based on their BMI alone. Such a misclassification is undesirable, especially in general practice, and it calls for diagnostic criteria other than the BMI alone to be used for obesity.

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

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 7983±1007 kJ/24 h (males) and 6127±907 kJ/24 h (females). Measured RMR and predicted RMR values using various equations from the literature were significantly different in males and females, except for the Harris–Benedict equation and the Schofield 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 Schofield 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 Schofield formulas and thus was not further explored.Conclusions: The Harris–Benedict formula and the Schofield 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.European Journal of Clinical Nutrition (2001) 55, 208–214

Determination of intracellular water by multifrequency bioelectrical impedance.

Body composition was measured in 57 healthy males by 40K measurements and by multifrequency bioelectrical impedance. Intracellular resistance (Ricf) was calculated from the impedance values using the Cole-Cole model. From total body potassium, intracellular water (ICW) was calculated. In addition, in 14 subjects, total body water (TBW) and extracellular water (ECW) was measured using deuterium oxide dilution and bromide dilution, respectively. Prediction formulas from the literature from impedance at different frequencies were used to predict TBW and ECW, and ICW was calculated as the difference of predicted TBW and ECW. Predicted and measured values of TBW, ECW and ICW did not always show the same values as measured. Generally the ICW from potassium was well correlated with calculated ICW from impedance values, correlation coefficients varying from 0.68 to 0.79, depending on the used prediction formula. From the impedance index H2/Ricf, body weight and age, ICW from potassium was predicted with stepwise multiple regression. The prediction formula was ICW = 0.37065.H2/Ricf - 0.132.age + 0.105.weight + 12.2. The prediction error was 1.9 kg and the explained variance 0.69. The residuals of this prediction formula were dependent on the level of ICW as measured by potassium. The same phenomenon was observed when ICW was calculated as the difference of predicted TBW and ECW. The results show that multifrequency impedance is able to predict ICW, however, the prediction is influenced by body water distribution.

Multi-frequency bioelectrical impedance: a comparison between the Cole-Cole modelling and Hanai equations with the classical impedance index approach

Total body water and extracellular water were measured by deuterium oxide and bromide dilution respectively in 23 healthy males and 25 healthy females. In addition, total body impedance was measured at 17 frequencies, ranging from 1 kHz to 1350 kHz. Modelling programs were used to extrapolate impedance values to frequency zero (extracellular resistance) and frequency infinity (total body water resistance). Impedance indexes (height2/Zf) were computed at all 17 frequencies. The estimation errors of extracellular resistance and total body water resistance were 1% and 3%, respectively. Impedance and impedance index at low frequency were correlated with extracellular water, independent of the amount of total body water. Total body water showed the greatest correlation with impedance and impedance index at high frequencies. Extrapolated impedance values did not show a higher correlation compared to measured values. Prediction formulas from the literature applied to fixed frequencies showed the best mean and individual predictions for both extracellular water and total body water. It is concluded that, at least in healthy individuals with normal body water distribution, modelling impedance data has no advantage over impedance values measured at fixed frequencies, probably due to estimation errors in the modelled data.

Assessment of changes in extra-cellular water and total body water using multi-frequency bio-electrical impedance.

Multifrequency bioelectrical impedance is a new approach in impedance methodology. Theoretically an alternating current with a low frequency is not able to penetrate the cell membrane because of the high capacitive resistance of the membrane. Consequently the measured impedance at low frequency is a measure of extracellular water (ECW) only. At high frequency the capacitive resistance diminishes. Hence the measured impedance at high frequency will be related to total body water (TBW). Aim of this study was to develop prediction formulas for ECW and TBW at different frequencies and to check the validity of these formulas in the assessment of body water compartments and changes in body water compartments.

Resting metabolic rate in Italians : relation with body composition and anthropometric parameters

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 their 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 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.

Impedance Ratio as a Measure of Water Shifts

Total and segmental body compositions (left arm and left leg) were measured by bioelectrical impedance analysis at 5 and 100 kHz and by dual-energy X-ray absorptiometry (DXA) in 14 healthy young males (body mass index, mean +/- SD, 23.5 +/- 2.7 kg/m2) every 20 min for a period of 100 min. During the measurements the subjects remained in the supine position on the examination table, except for the time between the last two measurements, where they got up to walk around. This study focuses on the impact of orthostatic fluid shifts on impedance ratios in body segments and the total body. After 20 min of lying supine lean tissue (DXA) was slightly but significantly lower for the total body and the left leg, but all other consecutive DXA measurements at different times did not differ. Total body impedance and leg impedance increased during the time the subjects were in the recumbent position. Impedance changes in the leg were more pronounced than in the total body, and at 5 kHz the changes were more pronounced compared to 100 kHz. After the subjects got up the impedance values again decreased. Impedance ratios (Z5/Z100) increased for both the leg and to a lesser extent for the total body during the time lying supine, but again decreased after getting up. The results of this study indicate a fluid shift from the legs to the trunk after lying supine and show that this fluid shift is mainly extracellular water. DXA measurements were not able to detect these changes, probably because the magnitude of these changes is below the detection level for lean tissue with DXA methodology. The observations have important implications in the interpretation of impedance measurements in the clinical situation when measurements are made in patients who have been reclining for some time.

Comparison of the Body Composition of Age-Matched Italian, Ukrainian, and Dutch Children

In the present study, the body composition was measured in 35 Ukrainian children, 14 boys and 21 girls, aged 8-12 years, by dual-energy X-ray absorptiometry (DEXA), anthropometry, and bioelectrical impedance. The results were compared with those obtained from a sample of the same number of Italian and Dutch children of the same sex and range of age. The Ukrainian children were slightly smaller, had a lower body weight, and had a lower body fat percentage, as measured by skinfolds, as compared with the Italian and Dutch children, as well as a lower calculated fat-free mass (FFM) by impedance. Comparison between FFM as determined by different methods in the Ukrainian children showed that all assessed values differed significantly. In particular, all methods overestimated the FFM as compared with DEXA, but the differences between DEXA and the other methods were not dependent on the level of FFM as determined by DEXA. Furthermore, the difference between predicted FFM and FFM values obtained by DEXA was correlated for skinfolds and body mass index (r = 0.77, p < 0.01), but not for body mass index and impedance, or skinfolds and impedance. The mean differences were 1.8 +/- 1.5 for body mass index, 1.0 +/- 1.4 for impedance, and 3.6 +/- 1.6 for skinfolds. Some differences in the body composition have also been found between boys and girls. The boys had higher values of lean tissue (23.3 +/- 3.1 vs. 20.7 +/- 3.3; p < 0.01) and FFM (24.5 +/- 3.3 vs. 21.8 +/- 3.5; p < 0.01) as compared with the girls and slightly higher values of bone mineral content. It should be noted, however, that the differences between different groups and different methods are always small, even when they are statistically significant.