The influence of skin thickness on the determination of the percentage of body fat (skinfolds and ultrasound)

[摘要] ENGLISH SUMMARY: There are several ways to assess the body composition of young men in alaboratory setting. However, due to the invasiveness, expense, time,specialised equipment, extensive instrumentation and level of skilled personnelrequired to perform these methods, they are generally not used in a clinicalsetting. New methods are being developed for clinical use, which may offer thepotential for non-invasive, not too costly in terms of time and equipment,reliable and easy body composition estimate.These methods involve the development of formulae which relate skinfold orcircumference measurements, or a combination of both, to indirect estimates ofbody fat (Drinkwater, 1984: 17). Since many of these methods involve thedetermination of the adipose tissue in vivo, most techniques rely on externalbody measurements through the skin.The importance of the influence or contribution of the actual skin thickness onthe determination of the caliper skinfold thickness has rarely been investigated.Ultrasound or sonar scans proved to be a fairly non-invasive and inexpensive,not too time consuming, yet accurate method to determine the skin andsubcutaneous fat layer individually in human beings.Various tests were performed on the subjects, such as hydrostatic weighing,lung functions, bio-electrical impedance analysis and a vast range ofanthropometrical measurements including weight, stature, diameter, girth andskinfold (caliper), skin and fat (sonar) on fourteen body locations wererecorded.The body density values (skinfolds), according to Durnin & Rahaman (1967)and Jackson & Pollock (1978), were individually applied to the formulae ofBrozek (1953) and Siri (1961) to calculate the percentage body fat.Correlations with hydrodensitometry were high, ranging from r=O,81499 (forDurnin & Rahaman and Brozek) to r=O,82338 (Jackson & Pollock and Siri) (seeFigures 10 to 13).The same procedure was followed with the sonar measurements and differentcombinations such as fat, 1x skin + fat and 2x skin + fat, were used.Correlations with hycfrodensitometry were remarkably high, seeing that theseformulae were developed for the use of skinfold calipers, and ranged fromr=O,77198 (Durnin &' Rahaman and Brozek) to r=O,84545 (Jackson & Pollockand Siri) (see Figures 14a to 17c). When applying the sonar (fat) measurements to the formula of Jackson &Pollock, it resulted in a relatively higher correlation (r=0,8455), as whencompared to other combinations of sonar measurements, namely sonar (1 xskin + fat) and sonar (2x skin + fat), which resulted r=0,83697 and r=0,81987respectively. When substituting the supra iliac (anterior) and medial thigh bodylocations with the supra iliac (posterior) and anterior thigh values respectively, agood correlation (r=0,83575) was found.Although all of these variations yielded good results, it can be seen that theformula of Jackson & Pollock was developed for the use of skinfolds(r=0,89224). When substituting the supra iliac (anterior) and anterior thigh bodylocations with the supra iliac (posterior) and medial thigh values respectively, agood correlation (r=0,86924) was found. This indicated that more attentionshould be given to these areas of fat deposits formerly not investigated,especially the medial thigh for females, and the posterior supra iliac for males.Each of the 14 body locations was individually correlated with body density viathree different methods, namely skinfold measurements (Harpenden caliper),fat measurements (sonar) and 2x Skin and fat: (sonar). Skinfolds correlated thehighest with body density, ranging from r=0,827 (abdomen) to r=0,615 (bicep).The second highest correlations were found to be that of fat thickness, varyingfrom r=0,791 (tricep) to r=0,237 (chin). The sonar measurements (2x skin + fat)correlated third highest ranging from r=0,754 (chest) to r=0,239 (chin).Subsequently no ultrasound formulae were available to compare this populationgroup and a new regression equation, using seven sonar fat values, wasdeveloped to indirectly estimate the body density. These findings werecompared to and correlated with hydrodensitometry (r=0,86784) (see Fig, 29).A further regression equation was developed, using the sum of seven skinfoldsas measured by caliper, in order to predict body density. This equation was alsocompared to hydrodensitometry and yielded a correlation of r=0,86936 (Fig 30).Either method of measurement and accompanying formula yielded good resultswhen compared to hydrodensitometry providing that the subject qualifies for the18 to 30 year old endo-mesomorphic category. This study provides the healthprofessional with alternatives where a choice between sonar and skinfold measurements can be made, depending on the preference of the patient andclinician or the time and apparatus available.

[发布机构] Stellenbosch University