The relationship of anthropometry and bio-electrical impedance to dual-energy X-ray absorptiometry in elderly men and women

Age and Ageing,  May, 1998  by Nicola C. chapman,  Elaine Bannerman,  Stephen Cowen,  William J. MacLennan

• E-mail
• Print
• Link

Introduction

One of the important aspects of health in elderly individuals is their nutritional status [1]. One very effective way of defining nutritional status is to measure the fat-free mass (FFM) and fat tissue mass of individuals [2]. In hospital wards and in the community, the simplest way of estimating these is to use equations which relate anthropometric measurements to age and sex [3]. A more complex, but equally practical method is to use a portable bio-electrical impedance analyser to measure bio-electrical impedance in subjects [4]. In elderly individuals there is a marked change in the distribution of body fat and changes in skin elasticity which may compromise the measurement of skinfold thickness [5]. In the case of bio-electrical impedance, readings may be seriously compromised in patients with fluid or electrolyte imbalance [6]. Problems such as these have resulted in concerns over the validity of anthropometry and bio-electrical impedance in elderly patients.

A more accurate but less clinically practical method for estimating body composition, against which anthropometry and bio-electrical impedance have commonly been compared, is dual-energy X-ray absorptiometry (DXA) [7]. Several of these comparisons have been in patients over the age of 60 years [6, 8]. Relatively few of the individuals, however, have been over the age of 75 years, an age beyond which there often is acceleration of many of the physiological changes associated with ageing [9]. It also forms a group in which multiple pathology, disability and poor nutrition are particularly common [10]. There would be particular value, therefore, in defining the accuracy of anthropometric and bio-electrical methods for defining nutritional status in this age group. In addition, anthropometric equations in old people may be compromised by the effects of disability and osteoporosis on height, so that it would be useful to assess the validity of other measures such as knee height or demi-span as an estimate of body stature in such a situation [11, 12].

Materials and methods

Subjects

Seventeen men and 17 women over the age of 75 years of age (mean 82 years; range 76-95 years) were recruited from the age-sex register of a group practice in Edinburgh or from local lunch clubs or day centres. All were living in their own homes and all were able to walk without help. None had dementia or had lower limb oedema. The number of subjects was dictated by the largest number of subjects who could be recruited during the study period and the machine time available to us.

Anthropometry

Subjects wearing a hospital gown and no shoes were weighed to the nearest 0.1 kg using standard electronic scales (Sohnle Digitals, CMS Weighing Equipment). Height was measured using a wall-mounted fixed stadiometer. Demi-span was assessed on the left upper limb to the nearest cm using the technique described by Bassey to measure the distance from the sternal notch to the web between the third and fourth fingers of an outstretched arm [13]. Harpenden callipers were used to measure triceps skinfold thickness on the left upper limb to the nearest 0.2mm. Inter- and intra-observer comparisons were made for these measurements to maximize their reliability [14]. With subjects sitting with their knees and ankles at right ankles, left knee height was measured to the nearest 0.1 cm using a Ross knee-height calliper [15, 11].

Reference population

In terms of age and nutritional status assessed by body mass index neither male and female subjects were significantly different from non-institutionalized individuals aged over 75 years recruited to a much larger study of the nutritional status of old people in Edinburgh [16].

Bio-electrical impedance

Resistance and reactance were measured using a four-terminal impedance analyser (RJL Systems, Detroit, MI, USA) with the subject lying supine and electrodes applied to the left upper and lower limbs [17].

DXA

FFM and total body bone mineral were measured using a Hologic QDR 1000W whole body scanner [18].

Statistical analysis

FFM was calculated from anthropometric measurements using the following age- and sex-specific equations derived from the literature:

FFM = (0.671 x [H.sup.2]/Rx [10.sup.4]) + 3.9 + 3.1S,

where H is height, R is resistance and S is sex (female = 0; male = 1) [6] and

Body fat = 1.4BMI + 0.48T - 25.81,

where BMI is body mass index and T is triceps skinfold thickness [19]. Thereafter, Pearson's coefficient of correlation was used to relate FFM calculated from anthropometric and impedance measurements to FFM assessed by DXA.

Multiple stepwise regression analysis was then performed to generate age- and sex-specific equations using FFM as measured by DXA. Factors offered as possible predictors of FFM in an initial equation were age, height, weight, triceps skinfold thickness, resistance and reactance. The regression analysis was then repeated using only anthropometric measurements. From these, a final equation was derived using age, sex, height, weight and bio-electrical impedance. FFM calculated using our own equations was then related to DXA FFM. After this, knee height and then demi-span were substituted for the equations, which were then again related to DXA FFM and further equations generated.