A new method for assessing fatness from an anthropometric study on 8799 British adults

• Main Author: McKay, Frances Carol
• Published: University of Glasgow 1983
• Subjects: 612; QP Physiology : RA Public aspects of medicine
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.236560

The aim of this study was to develop a method for measuring an individual's fat content, which was both simple and inexpensive and could therefore be used by relatively inexperienced researchers in large scale field studies. At present the most popular field methods for assessing 'overweight' are weight for height tables based on Insurance Company data, and weightheight indices. The methods chapter points out the major limitations of these methods and describes how they cannot differentiate between weight due to bone, muscle, water or fat. Another popular field method is to measure skinfolds at a few predefined si tes and convert these to a fat content using regression equations. Al though this method allows 'fatness' as opposed to 'overweight' to be assessed in the individual, it has the disadvantage that the observer requires some training, which is not· always feasible, and carefully calibrated skinfold calipers are essential. It is for these reasons that a new field method, requiring minimal training and equipment was sought. This study was carried out on a group of 6,495 males and 2,304 females aged 16-64y, selected, as described in Chapter 2, from both the British Armed Forces and the civilian population. The measurements taken from each individual were height, weight, 4 circumferences, 4 boney diameters and 4 skinfolds. Using the equations of Durnin and Womersley (1974) and Siri (1956) the skinfolds were converted into a value for percent body fat, and fat free mass (FFM) was calculated by subtracting fat mass from body weight. The height and weight results were compared with the results of the Office of Population Censuses and Surveys (OPCS), 1981, UK survey. Since the OPCS survey was believed to be representative of the UK population, the comparison allowed an assessment of possible sampling errors. Variations in' anthropometric results related to geographical origins and social class (SC) were also examined, within Chapter 3, together with age related changes. Within the Forces, civilians and OPCS samples respectively, mean height had values of 175.9cm, 175.6cm and 173.8cm. Within the female samples, these 3 values were 163.6cm, 162.4cm and 160.7cm. The differences between the 3 populations were due mainly to the facts that the Forces selection procedure includes a minimum height cutpoint for many occupations and that the civilian selection was not very random. When predicting percent fat or FFM however, these differences appeared to be relatively unimportant. Although height appeared to vary little with age, it did vary in relation to geographical region. In general, -the northern regions had slightly smaller means for most of the anthropometric measurements, when compared to the southern regions. In addition, there was a slight tendency fof height to decrease with se. Mean weight increased with age from 65.5kg in the Forces male 16y olds to 80.0kg in the 5O-56y olds. The Forces and civilian females kept their .weight around 61 and 57kg respectively, between 17 and 29y, after which it rose steadily. Most of these weight increases were due to increases in fat content, since between the 16-17y and over 50y olds, mean percent fat rosefrom 13.4% to 27.2% and from 28% to 35.7% in the Forces males and females respectively. FFM also varied slightly with age, especially in the male sample. In the male Forces it averaged 56.5kg, 61.8kg and 59.6kg in the 16y, 25-29y and 50-56y olds. The initial rise was mainly reflecting growth in the younger subjects. The subsequent changes are discussed in detail in Sections 3.2.10 and 3.2.11. When matched for height and age the Forces males had FFM values on average 2.5kg larger than the civilians and this reflected a larger mean 'build'. This had to be taken into account in order to produce prediction equations applicable to both populations. There was little difference in fat content between the 2 groups. The Forces females were of a similar 'build' to the civilians, but on average 1-2% of body weight fatter. regression equations. This made no difference to the Section 3.4. describes the calculation of regression equations which predicted fat content and FFM. Although initially both FFM and percent fat were used as dependent variables, the prediction of FFM was the more accurate and therefore it was used in preferance. The males were ini tally divided into height, weight then age groups but since the regressions predicting FFM in age groups were the most accurate, age was chosen as the final grouping variable in both sexes. The number of age groups depended on the similarities between different ages, and was calculated using a F-test. Using the BMDP package of computer programmes, the variables height, weight, calf circumference and ulnar diameter were chosen from those measured as the 'best' to predict FFM in the male sample. In the females, the 'best' variables were height, weight and upperarm circumference. The regression equations are in Tables 90 and 91. The final 7 male and 2 female age related regression equations were ini tially calculated from the Forces data, and cross validated on the ci vUian sample. The range of standard errors of the estimates (SEE) in· both samples was 1.54-2.39kg in the males and 1.44-1.80kg in the females. Approximately 95% of the prediction errors would lie within! 2xSEE• Overall, FFM and hence percent fat could be predicted with greater accuracy using these regression equations than using weight-height indices or tables. The method is also simple enough to be used by untrained observers, in field studies.