Body composition assessment of undernourished older subjects by dual-energy x-ray absorptiometry and bioelectric impedance analysis

Abstract

Background

The prevention and treatment of diseases related to changes in body composition require accurate methods for the measurement of body composition. However, few studies have dealt specifically with the assessment of body composition of undernourished older subjects by different methodologies.

Objectives

To assess the body composition of undernourished older subjects by two different methods, dual energy x-ray absorptiometry (DXA) and bioelectric impedance (BIA), and to compare results with those of an eutrophic group.

Design

The study model was cross-sectional; the study was performed at the University Hospital of the School of Medicine of Ribeirão Preto, University of São Paulo, Brazil.

Participants

Forty-one male volunteers aged 62 to 91 years. The groups were selected on the basis of anamnesis, physical examination and nutritional assessment according to the Mini Nutritional Assessment (MNA) score. Body composition was assessed by DXA and BIA.

Results

Body weight, arm and calf circumference, body mass index (BMI), fat free mass (FFM) and fat mass (FM) were significantly lower in the undernourished group as compared to the eutrophic group. There were no significant differences between FFM and FM mean values determined by DXA and BIA in both groups, but the agreement between methods in the undernourished group was less strong.

Conclusion

Our results suggest caution when BIA is to be applied in studies including undernourished older subjects. This study does not support BIA as an accurate method for the individual assessment of body composition.

Introduction

The understanding of age-related changes in body composition and their effects on health is extremely important for health care and nutritional support for the older population (1). Increasing age causes a number of physiological changes that associated with a sedentary life style, inadequate food intake and poor health, may lead to an increased risk of malnutrition (2).

Despite the growing understanding about the pathophysiology of undernutrition in the aged, its prevalence remains high. Prevalence rates vary substantially in different studies, but figures between 40–50% have been reported depending on the tools apllied for the evaluation of nutritional status and population studied. The prevalence is especially high among those that have been hospitalized, live alone or in resting and nursing homes (3, 4).

Although undernutrition is seldom regarded as a primary medical concern, it is strongly associated with a number of poor medical outcomes. Unintentional weight loss and undernutrition are associated with increased mortality and morbidity, increased risk of infections and pressure ulcers development (5), increased risk of falls, decreased ability to perform self-care activities, low mobility and dependence (6).

Different methods have been applied for the evaluation of body composition but not all of them can be applied to older people (7). The most frequently used methods for the assessment of body composition establish a quantitative relation between fat mass (FM) and fat-free mass (FFM) according to a two-compartment model. Among them, Bioelectric Impedance Analysis (BIA) is extensively used in clinical practice (8) because of its reproducibility and clinical applicability. BIA is a portable noninvasive, easy to use and relatively inexpensive method that can be applied in a rapid and precise manner (9). Specific BIA equations have been developed for older persons (10).

Dual-energy X-ray absorptiometry (DXA) is a safe and non-invasive method that allows the evaluation of whole body composition and also the composition of separate body segments (limbs and trunk), with high precision and accuracy (11). For this reason, DXA has been applied as a reference method for the assessment of body composition in various studies. This method, however, is not portable and not available in many services, due to the high cost of equipment, operation and maintenance.

Although BIA has been recognized as a reliable method for the assessment of body composition in young populations, there are concerns about its accuracy in older people, due to changes in hydration and fat distribution (12). Although some studies comparing these methods in older persons are available, studies on diseased and undernourished older persons are scarce, further studies being clearly needed (13).

The objective of this study was to assess the body composition of eutrophic and undernourished older subjects by BIA and compare the results with hose obtained by DXA, employed here as a reference method. We intended to verify if BIA remains precise in undernourished older people, a considerable proportion of the older population in daily clinical practice.

Methods

Studied Population

Patients of the Geriatric Medicine and Clinical Nutrition outpatient clinics of the University Hospital of the School of Medicine of Ribeirão Preto, University of São Paulo were invited to participate. The study was approved by the local Human Research Ethics Committee and all subjects gave written informed consent prior to participating.

Each volunteer was submitted to detailed clinical and nutritional assessment by a physician and a nutritionist to identify acute and chronic diseases and to diagnose nutritional status. Exclusion criteria included severe frailty, obesity, alcoholism, kidney, liver and gastrointestional diseases, diabetes, malignant disease, advanced dementia, edema, ascites, presence of heart pacemaker, orthopedic prostheses and osteosynthesis.

Experimental Design and Methods

The study model was cross-sectional. A convenience sample of male volunteers was established, so that two groups, one with 20 eutrophic volunteers and one with 21 undernourished volunteers were constituted.

The groups were selected according to clinical anamnesis, physical examination and Mini Nutritional Assessment (MNA) score (14).

The MNA is a simple and user-friendly 18 items tool, widely employed for the identification of undernutrition and risk of undernutrition that can be applied to both independent and dependent older subjects. This instrument is composed by four sections: anthropometric assessment (weight, height, mid-upper arm circumference, calf circumference and a question about weight loss); general assessment (lifestyle, use of medication, presence of pressure ulcers, presence of psychological stress or neuropsychological problems and mobility); dietary assessment (number of meals, food and fluid intake and autonomy for feeding) and subjective assessment (self-perception of health and nutrition status). The maximum score is 30.

Subjects with a MNA score of 23.5 points or over and no clinical signs of undernutrition (dry flaky skin, sparse, easily pluckable, dyspigmented hair; muscle wasting and peripheral oedema) were classified as eutrophic and subjects with a MNA score of less than 17 points and clinical signs of malnutrition were considered to be undernourished. Subjects with an MNA score between 17 and 23.5 (classified as “at risk of malnutrition” by the MNA) and clinical signs of undernutrition were also included in the undernourished group. Subjects with MNA score in this range but no signs of undernutrition were excluded from the study.

Anthropometric measurements

Body mass index (BMI) was calculated as weight (kg) divided by height (m2). Body weight was measured by an electronic platform scale (Filizola ID 500, São Paulo, Brazil) to the nearest 0.1 kg, with the subject lightly dressed. Height was measured to the nearest centimeter by a stadiometer at head level with the subject standing barefoot, with feet together.

The mid-upper arm circumference and calf circumference (used in the calculation of the MNA score) were measured to the nearest 1 mm with a non-elastic flexible plastic tape. Mid-upper arm circumference was measured in the middle point of the right arm (mean distance between the tip of the acromion process and the tip of the olecranon process) extended alongside the body. Calf circumference was measured with the patient in the supine position. The left knee was raised until a right angle between the thigh and calf was reached. The greatest circumference of the calf was measured without compression of the subcutaneous tissue (15).

Body Composition

Whole body composition was evaluated by DXA using a total body analysis model (Hologic, QDR 4500 W, Waltham, MA, USA). Fat Free Mass (FFM) was defined as the sum of bone mineral content (BMC, g) and lean body mass (LBM, g). All metal items were removed from the subjects to ensure the accuracy of the measurement. Scans were performed by a trained technician with the subjects in the supine position. The coefficient of variation of DXA was 0.49% for body weight, 1.48% for BMC, 1.99% for FFM, and 4.71% for FM in the eutrophic group and 0.34% for body weight, 1.44% for BMC, 1.79% for FFM and 7.98% for FM in the undernourished group.

Body composition was also evaluated by BIA, performed by a single investigator. A single-frequency 50 kHz bioelectrical impedance analyzer (BIA 101 Q-RJL Systems, Detroit, MI, USA) was employed. Proximal electrodes were placed on the ipsilateral wrist and ankle lines on the right half of the body and the distal electrode placed on the metacarpal and carpal line. BIA measurements were obtained after an 8-hour overnight fast. Subjects were instructed to avoid any exercise in the day before and to empty their bladder before the test. All volunteers were asked to lie in the supine position for 10 minutes before the evaluation in a temperature-controlled room at 24 to 26 °C. During data acquisition, limbs were maintained in abduction to avoid interferences to the instrument (16). The equation of Roubenoff et al. (17) was used for the estimation of FFM.

Statistical Analysis

Data are reported as mean and standard deviation. The STATA Statistic Software (Release 10.0. Special Edition, Stata Corporation, College Station, TX, USA) was employed for statistical analysis. The unpaired Student's t-test was employed for the comparison of anthropometric and body composition parameters between the eutrophic and undernourished groups. The St. Laurent coefficient and Bland-Altamn plots were employed to verify the agreement between BIA and DXA. For statistical purposes, DXA was adopted as the gold standard method (18, 19). The level of significance was set at 5% (20).

Results

There were no significant differences between groups in age and height (Table 1). Body weight, arm and calf circumferences, MNA and BMI scores were lower in the undernourished group. FFM and FM measured by DXA were higher in the eutrophic group as compared with the undernourished group (p < 0.005).

Table 1.

General characteristics of the eutrophic and undernourished older subjects

	Eutrophic (N=20)		Undernourished (N=21)
	Range	Mean (SD)	Range	Mean (SD)
Age (years)	62 – 81	73 (5.46)	66 – 91	76.47 (7.04)
Height (cm)	156 – 175	165.15 (5.75)	158 – 17	165.03 (4.00)
Weight (kg)	54 – 84	69.66 (8.40)	42 – 62	50.91 (6.16)*
Arm circumference (cm)	26 – 34	29.47 (2.61)	20.5 – 29	23.67 (2.39)*
Calf circumference (cm)	33 – 38	35.42 (1.70)	27 – 35	30.54 (2.34)*
MNA (range 0-30)	24 – 30	26.77 (1.69)	11 – 23.5	17.97 (3.55)*
BMI (kg.m-2)	22 – 30	25.68 (2.23)	16 – 21	18.66 (1.85)*
FFM DXA (kg)	39 – 58	50.78 (5.37)	38 – 49	41.83 (4.68)*
FM DXA (kg)	9 – 28	18.44 (4.63)	5 – 14	8.44 (2.50)*

N: number of volunteers; MNA: Mini Nutritional Assessment; BMI: Body Mass Index; FFM: fat free mass measured by DXA; FM: fat mass measured by DXA; *: p<0.005.

DXA was highly accurate for the determination of body weight, as compared with the values obtained by the scale (r = 0.99, p < 0.0001; data not shown). Although there were no significant differences in mean FFM and FM as measured by DXA and BIA (p > 0.05, Table 2), BIA tended to underestimate FFM in eutrophic volunteers and to overestimate FFM in undernourished ones (Table 2).

Table 2.

Body composition of the eutrophic and undernourished volunteers, as measured by BIA and DXA

	Eutrophic		Undernourished
	FFM	FM	FFM	FM
BIA	50.31 (3.82)	19.35 (5.74)	43.42 (4.22)	7.49 (3.58)
DXA	50.78 (5.37)	18.44 (4.63)	41.83 (4.68)	8.44 (2.50)
ABIA-DXA	-0.47	0.91	1.59	-0.95
St L. Coefficient	0.85	0.75	0.79	0.54
95% CI	(0.75 - 0.90)	(0.58 - 0.88)	(0.64 - 0.88)	(0.32 - 0.73)

FFM: fat free mass mass (kg); FM: fat mass (kg); ΔBIA-DXA (kg): BIA minus DXA; St L. Coefficient: St Laurent Coefficient; 95% CI: St. Laurent Coefficient 95% confidence interval.

In the eutrophic group, there was a strong positive agreement between the two methods in FFM (r = 0.85) and FM (r = 0.75). In the undernourished group, the correlation remained significant, although less strong (FFM, r = 0.79 and FM, r = 0.54). The St. Laurent coefficient of agreement was also lower for both FFM and FM in the undernourished group, as compared to the eutrophic group.

Figure 1 shows the Bland-Altman plots comparing the two methods. Although there were no significant differences in mean FFM and FM between the two groups, individual variation was high, with a tendency of BIA to underestimate FM in the lower range and overestimate in the higher range in undernourished volunteers.

Figure 1.

Bland–Altman plots comparing FFM by BIA and DXA of the eutrophic (A); FFM by BIA and DXA of the undernourished (B); FM by BIA and DXA of the eutrophic (C) and FM by BIA and DXA of the undernourished (D) older subjects. Dotted lines show the mean difference and limits of agreement (mean difference ± 2SD) between the two methods; the plain line represents equality between the methods

Discussion

This study showed that the measurement of FFM and FM of older persons by BIA may be affected by nutritional status, so that in undernourished subjects the correlation of BIA measurements with those obtained by DXA (used as gold-standard in this study) is less strong than in eutrophic subjects. Moreover, there is a tendency of BIA to overestimate FFM and underestimate FM in undernourished subjects.

BIA is based on the different behavior of biological tissues in response to an applied electrical current (21). There is controversy about the applicability of BIA in old subjects, because aging is associated with changes in height, weight and fat distribution (22). Undernourishment in old people could add further variability by determining a different bioelectrical model, as weight loss and muscle atrophy lead to changes in peripheric bioelectrical resistance (23).

Despite the aspects discussed above, there was good agreement between the two methods in both groups studied, especially for fat-free mass. These results are in accordance with those obtained by Lupoli et al. (22) when they assessed the body composition of underweight older subjects using the same methods as those employed in the present study. Those authors also found a strong correlation between FFM assessed by DXA and BIA. Verifying the reliability of eight BIA equations, they observed that the Roubenoff equation (used in the present sudy) estimated FFM with high accuracy when compared to DXA.

The group of eutrophic older subjects also showed agreement between FFM and FM estimated by BIA and DXA, confirming the results reported by Bolanowski and Nilsson (23) who observed a positive association of FFM and FM assessed by DXA and BIA in volunteers aged 16 to 78 years, and those reported by Sun et al. (24) for 591 healthy individuals aged 19 to 60 years.

In conclusion, this study showed that although mean values for FM and FFM as measured by BIA and DXA did not differ significantly in undernourished older people, the correlation between the two methods in this group was less strong than in the eutrophic one, suggesting caution when BIA is to be applied in studies including undernourished older subjects. Variability was high between individuals, so our results do not support BIA as a reliable method for the individual assessment of body composition.

Financial Disclosure: K.H.C. Vilaça: During the research, K. Vilaça was supported by a scholarship from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior -CAPES. The institution did not interfere with the research design, development or results. F.J.A. Paula, E. Ferriolli, N.K.C. Lima, J.S. Marchini and J.C. Moriguti: no support or other form of conflicts of interest.

Author Contributions: K.H.C. Vilaça: research design, data collection, data analysis, article writing. F.J.A. Paula: data collection, data analysis, article revision. E. Ferriolli: data analysis, article writing and revision. N.K.C. Lima: data analysis, article revision. J.S. Marchini: data analysis, article revision. J.C. Moriguti: research design, data collection, data analysis, article writing and revision.