Comparison between two methods of hydrostatic weighing without head submersion in morbidly obese females

Comparison of estimation accuracy of body density between different hydrostatics weighing methods without head submersion

This study compared the accuracy of body density (Db) estimation methods using hydrostatic weighing without complete head submersion (HW(withoutHS)) of Donnelly et al. (1988) and Donnelly and Sintek (1984) as referenced to Goldman and Buskirk's approach (1961). Donnelly et al.'s method estimates Db from a regression equation using HW(withoutHS), moreover, Donnelly and Sintek's method estimates it from HW(withoutHS) and head anthropometric variables. Fifteen Japanese males (173.8+/-4.5 cm, 63.6+/-5.4 kg, 21.2+/-2.8 years) and fifteen females (161.4+/-5.4 cm, 53.8+/-4.8 kg, 21.0+/-1.4 years) participated in this study. All the subjects were measured for head length, width and HWs under the two conditions of with and without head submersion. In order to examine the consistency of estimation values of Db, the correlation coefficients between the estimation values and the reference (Goldman and Buskirk, 1961) were calculated. The standard errors of estimation (SEE) were calculated by regression analysis using a reference value as a dependent variable and estimation values as independent variables. In addition, the systematic errors of two estimation methods were investigated by the Bland-Altman technique (Bland and Altman, 1986). In the estimation, Donnelly and Sintek's equation showed a high relationship with the reference (r=0.960, p<0.01), but had more differences from the reference compared with Donnelly et al.'s equation. Further studies are needed to develop new prediction equations for Japanese considering sex and individual differences in head anthropometry.

Bioelectric impedance and hydrostatic weighing with and without head submersion in persons who are morbidly obese

OBJECTIVE

To compare hydrostatic weighing with and without head submersion and bioelectric impedance analysis (BIA) for measurement of body composition of persons who are morbidly obese.

DESIGN

Body composition was determined using 3 methods: hydrostatic weighing with and without head submersion and BIA. Residual volume for the hydrostatic weighing calculation was determined by body plethysmography.

SUBJECTS

Subjects were 16 morbidly obese men (142.5 kg mean body weight) and 30 morbidly obese women (125.9 kg mean body weight) living in the Salt Lake County, Utah, area. Morbid obesity was defined as 40 kg or more over ideal weight.

STATISTICAL ANALYSIS

One-way, repeated-measures analysis of variance was followed by Scheffé post hoc tests; body-fat measurement method served as the repeated variable and percentage of body fat as the dependent variable. Men and women were analyzed separately. In addition, degree of agreement between the 3 methods of determining body composition was determined. A regression equation was used to calculate body density for hydrostatic weighing without head submersion. Two new BIA regression equations were developed from the data of the 16 men and 30 women.

RESULTS

Values for percentage body fat from hydrostatic weighing with and without head submersion (41.8% vs 41.7%, respectively) were the same for men but differed for women (52.2% vs 49.4%, respectively, P < .0001). Values for body fat percentage measured by BIA were significantly lower for men (36.1%) and women (43.1%) (for both, P < .0001) compared with values from hydrostatic weighing methods. BIA underpredicted percentage body fat by a mean of 5.7% in men and 9.1% in women compared with the traditional hydrostatic weighing method.

APPLICATIONS/CONCLUSIONS

BIA tended to underpredict the measurement of percentage body fat in male and female subjects who were morbidly obese. Hydrostatic weighing without head submersion provides an accurate, acceptable, and convenient alternative method for body composition assessment of the morbidly obese population in comparison with the traditional hydrostatic weighing method. In population screening or other settings where underwater weighing is impractical, population-specific BIA regression equations should be used because general BIA equations lead to consistent underprediction of percentage body fat compared with hydrostatic weighing.