An analysis of body composition and its measurement in a sample of Irish adults aged 18-81 years.

Obesity is a global epidemic and is defined as excess fat accumulation to the extent that health may be impaired (WHO 2000). However, body mass index (BMI), the metric used to quantify obesity, does not adequately represent fat tissue mass (FTM). A metric that quantifies obesity according to FTM is required. Methods that accurately measure total and segmental FTM and are suitable for use in large scale studies are necessary. Using dual energy x-ray absorptiometry (DXA) as the reference method of measurement, this thesis documents the body composition, specifically FTM and its distribution, of 1136 Irish adults aged 18-81 years. Fat tissue mass index (FTMI, kg/m2) was chosen as the most appropriate metric to define ‘fat obesity’ and to compare the difference in total body fat according to age and sex. Reference FTMI values were derived from young adult (YA) z-scores. FTMI was found to increase with age in men and women, as did the proportion of FTM deposited abdominally. 40% of men and 45% of women defined as ‘fat obese’ by FTMI were not ‘obese’ by BMI. Compared to DXA, the accuracy of the prediction methods i.e. bioelectrical impedance analysis (BIA), anthropometry and ultrasonography, for the measurement of total and/or segmental body fat has been established. Prediction equations derived from site-specific skinfold and girth measures accurately estimated % body fat in men (r=0.91, standard error of the estimate (SEE) =2.5%) and women (r=0.92, SEE=3.0%). Prediction equations derived from site-specific subcutaneous adipose tissue thickness using ultrasonography accurately estimated % body fat in YA men (r=0.95, SEE=1.9%) and women (r=0.91, SEE=3.0%). BIA was found to underestimate % body fat and did not provide an accurate measure of body composition in YA women or YA men with >25% body fat (p<0.001). This thesis finds fat tissue mass, and its distribution, to be age and sex specific. BMI was found to be an inappropriate measure of ‘fat obesity’ and not indicative of the difference in total and segmental FTM between the sexes or across the lifespan. The thesis concludes that body composition can be accurately predicted from anthropometric and ultrasonography measures in healthy adults and these methods are suitable for use in large scale studies.