Longitudinal measurements of total body water and body composition in healthy volunteers by online breath deuterium measurement and other near-subject methods

Abstract

Rapid quantification of breath deuterium abundance by flowing afterglow mass spectrometry (FA-MS) enables accurate measurement of total body water (TBW), which combined with other techniques such as bioelectrical impedance analysis (BIA) and anthropometrics enables near-subject assessment of body composition. This study assessed the comparative reproducibility and inter-relationship of these methods in healthy subjects over 12 months. Detailed bedside composition was performed in 22 subjects, (10 male) aged 28-79 with body mass index (BMI) ranging from 21-38 at baseline and again at one year. Techniques included FA-MS deuterium dilution, BIA, skin-fold thickness (SFT) and soft tissue ultrasound measurement of fat and muscle depth. Short-term reproducibility for each method was established. Within and between technique comparisons of measurement were made from Pearson’s linear regression, coefficient of variation (CV) and Bland-Altman analysis. Weight and TBW estimated by FA-MS, BIA and SFT at baseline and one year later were highly correlated (R² = 0.96-0.98), slope 1.02-1.03, CV = 4.5-11.6%. Systematic errors between the different methods in determining TBW were effectively identical at baseline and after one year. There was a tendency for subjects to gain weight during the study period, due to an increase, predominantly in younger women, of body water (FA-MS and SFT) and loss of upper body fat (ultrasound). BIA was relatively insensitive to these changes. It is concluded that over a 12-month period, TBW determined by FA-MS deuterium breath analysis has reproducibility similar to conventional weighing. The stability of between method errors would suggest that these techniques might be used in conjunction with each other in the longitudinal determination of body composition and so detect relatively subtle changes. The value of including an absolute determinant of TBW by FA-MS that is independent of the need to employ population derived equations, appears to be of value in the near-subject determination of body composition as required in clinical practice.

Introduction

Disease and lifestyle changes involving diet, activity level and stress, can cause significant changes in an individual’s muscle, fat, bone and both the quantity and distribution of body water [1]. The ability to measure changes in these components of body composition longitudinally, both accurately and with ease in the clinical situation is, therefore, highly desirable. Bio-electrical impedance analysis (BIA) is increasingly employed to this end as it is quick, non-invasive and highly reproducible [2]. However it predicts rather than measures body water, employing population derived equations to make estimates of body compartments, eg fat free mass, rather than making absolute, independent measurements [3]. The errors in this approach may be higher particularly if more than one component of body composition is changing concomitantly.

Until now, it has been difficult to obtain rapid, precise measurements of total body water (TBW), such as are required in clinical decision-making. We have recently described the development and application of a deuterium breath test that following deuterium ingestion and equilibration enables immediate and accurate determination of TBW from single exhalations [4, 5]. In principle, this method can be combined with other near-patient methods, such as BIA, in association with additional measures such as body weight and anthropometrics to monitor longitudinal body composition. For example, a patient with renal failure who is dialysis dependent may have an emergency admission with breathlessness suggesting fluid weight gain and yet be below their target weight. As BIA often uses also weight to predict body water it may underestimate absolute fluid gain but give the clinician valuable information as to its relative intra verses extra-cellular distribution. Knowledge of both the absolute fluid status combined with relative fluid distribution might be expected to improve patient care provided the information is readily accessible. The purpose of this study was to establish the reproducibility, stability and accuracy of these near-patient methods and their value in combination, in the longitudinal assessment of body composition in a group of healthy subjects in whom it would be expected to be stable.

Design and Methods

Detailed near-patient determination of body composition, including anthropometrics (weight, height, circumferences, skinfold thickness, SFT, ultrasound of fat and muscle), multi-frequency bio-electrical impedance analysis (BIA), and measurement of TBW using deuterium dilution was performed at baseline and 12 months later (‘1 year’) in 24 healthy volunteers. All subjects described themselves as fit and well. At baseline, three subjects were taking antihypertensive medication atenolol, propranolol, verapamil and amiloride although one was taking this for ‘anxiety’ rather than blood pressure, and one woman took hormone replacement. At the second analysis two more subjects had started taking anti-hypertensive medication, (amlodipine and bendrofluazide). Volunteers were selected to represent wide age and weight ranges and equal gender split so as to represent a useful comparator group for future studies of patients with chronic disease. Repeat BIA measurements were also taken 10 minutes after the first measurement and both anthropometrics and BIA were repeated one month after the measurements taken in year 2, in order to establish the observer error and repeatability of these techniques. The purpose of repeating the anthropometrics again after a month was to remove bias due to memory that might occur if being performed after a shorter time interval. All subjects gave their informed consent and the local ethics committee approved the study.

Determination of body water by on-line breath analysis

Body water was determined from the change in breath deuterium abundance, measured from single exhalations using flowing afterglow mass spectrometry (FA-MS), that occurred following administration of a known amount of deuterium oxide employing the dilution principle, as recently described [4-7]. Briefly, base-line levels of deuterium abundance in body water were obtained from subjects and a local control, and found to be ∼156 ppm, the same as local tap water. Each subject then drank a dose of 99.9 % pure D₂O (0.3 g kg^-1 body weight) mixed in 100 ml of tap water. This dose, which is higher than that used in conventional analytical techniques (20 verses 4 g), resulting in a typical incremental rise in breath D of 350 ppm ±3-5, is used to ensure that the overall error of the method is kept below 1%. The actual individual doses of D₂O were accurately weighed. The mixture was washed down with an additional 100 ml of tap water. Post dose breath samples were then taken approximately 3 hours after the D₂O ingestion, after which time a near steady state in the breath deuterium is reached, as our previous studies [4] have shown, implying even dispersal throughout the body water. The TBW was calculated from the increase of the breath deuterium abundance determined from FA-MS analysis of breath (with correction for isotope fractionation in the gas phase) compared to baseline, as ratio to the volume of D₂O ingested, and corrected for isotopic exchange with non-aqueous hydrogen (0.96) [8]. Throughout all of the experiments, an additional, single control subject who did not take D₂O was used to verify that cross-contamination and base-line drift in Deuterium analysis did not occur. Using FA-MS, repeated determinations of body water in peritoneal dialysis patients on the same day, following a known change in body water of 2 kg dialysate fluid, demonstrated an error of <300 ml, or <0.5%, in the estimate of TBW [9].

Bio-electrical impedance analysis (BIA)

BIA was employed to predict both total body water and intra versus extra-cellular water, from measurements of resistance at zero (Re) and infinite (Ri) frequencies as determined from a Cole-Cole plot, using a Xitron Hydra (Model 4200) multi-frequency analyser [10, 11]. The Xitron calculates TBW from the formula, V_TBW = K_Bρ_F.L2/R, where K_B is a factor relating the relative proportions of the leg, arm, torso and height, ρ_F is the resistivity of the fluid, L the body height and R the resistance. Subjects were asked to lie supine on a couch for 10 minutes. After cleaning the skin to ensure good contact, electrodes were attached to the right wrist and ankle in a standardised fashion, with the voltage detector electrodes at the line of the joint and the current-injection electrodes over the metacarpal-phalangeal and metatarsal-phalangeal joints respectively. These were then attached to the Xitron Hydra, which was operated from a laptop computer to initiate and store the readings. The measurement itself took approximately 15 seconds. The measurements of both Re and Ri were closely reproduced 10 minutes later, the standard deviation of the differences being <1%. Repeat measurements after one month gave CV values of 4.9% and 3.8% for Re and Ri respectively. The intra-class correlation coefficients for estimates of TBW, extra-cellular fluid (ECF) and intracellular fluid (ICF) after one month were >0.99.

Anthropometrics

Height was measured using a wall-mounted stadiometer to the nearest 0.1 cm and weight was measured on a Seca scale to the nearest 100g. Both measurements were taken twice and the mean calculated. Subjects wore a hospital gown for all measurements. SFT measurements were taken at five sites (triceps, biceps, subscapular, supra iliac and thigh) on the right side of the body with a Holtain Skinfold caliper according to the methods described by Harrison et al. [12] by a single trained observer (BE). Circumferences of waist, hip, mid arm and mid thigh were taken at each site in rotation and the mean used for subsequent analysis. Body density was calculated from the formula of Durnin and Wormesley and the percentage of total body fat (%TBF) derived from the Siri Equation [13, 14]. TBW was derived from % body fat by calculating 0.73* (weight − TBF) [15].

A portable Aloka 500 Ultrasound (US) machine was used to measure subcutaneous fat and muscle depth of the biceps, triceps and thigh. The patient was seated with their arms relaxed by their sides and knee bent at 90° so that the upper leg was parallel with the floor. Water-soluble gel was applied to the probe to aid transmission of the ultrasonic waves. The centre of the ultrasound probe was placed gently on the limb so as to minimize compression, the image produced was frozen and measured by electronic calipers installed in the machines software. Three measurements were taken for fat and muscle depth and the mean calculated. The maximum range of the Aloka is 8cm.

The short-term reproducibility of anthropometric measurements was assessed from intra-class correlation coefficients (>0.9 for all measurements), and calculation of the technical error of measurement (TEM), and thus the 95% CI for a true change in measurement [16]. This ranged from 0.6%-1.4% for simple measures such as waist, hip and thigh circumferences, 3.6%-6.5% for skin folds, 2.5%-6.8% for ultrasound measurements (with the exception of triceps muscle, 12 %, considered unacceptable by the target set by Gore [17] and 5.1% for %TBF.

Statistics

As indicated above, the short-term reproducibility of the methods was determined from the inter class correlation coefficients and calculation of the TEM. Changes in body composition taking place over the year were measured using paired t-tests, coefficient of variation (C.V.) and Pearson’s correlation coefficient. Comparisons of the different methods of measuring body water were made using with paired t-tests, Pearson’s correlation coefficient and Bland-Altman analysis [18].

Results

Subjects

The initial study involved 12 men and 12 women. However at the follow up study (Year 1) one subject had had a 3-month period of illness associated with weight loss and another subject did not follow the protocol correctly on the day of the deuterium measurements. These two male subjects were excluded from the analysis. The remaining subjects included 12 females and 10 males with an age range of 28-79 years and a BMI range of 21.3 to 38.1 kg/m².

Comparison of Body Composition at baseline and Year 1

Table 1 shows the actual and derived measurements of weight, TBW (FA-MS deuterium dilution, BIA and SFT), ECF, ICF, Re, Ri and %TBF for all subjects and according to gender. Table 2 shows the mean differences over 1 year for these measures, their correlation, coefficient of variation and 95% CI of any change. As would be anticipated, and as depicted in Figures 1a-d, there is a high degree of correlation between all measures of body composition between baseline and 1 year. The reproducibility and stability of the principle measures of weight and body water, in particular FA-MS and BIA, are similar (CV ∼4.5-5%) and without systematic error as indicated by the gradient of the Pearson correlation (see Figs 1a-c), which approached unity in each case. Reproducibility of TBW estimated by SFT is less good (Fig 1d), and it can be seen that the reproducibility of BIA is in part due to the inclusion of weight in the formula, as the estimates of Re and Ri have a slightly higher coefficient of variation. It can also be seen that the mean absolute changes in weight and TBW determined from FA-MS or SFT are very similar, even by gender, whereas the estimates from BIA are somewhat at variance to this. There are clearly important differences between the genders in this study. In general, the mean differences between baseline and 1 year for all measurements were greater for women, and for some measures this was statistically significant. Women tended to put on more weight and had significantly higher body water at 1 year (as determined from FA-MS and SFT). Of interest, however the BIA estimate of TBW was not different in either gender whereas the derived ECF was significantly greater in the women at one year. Taken together, this increased variability of body composition in women seems to be due to a reduction in body fat and an increase in both ICF and ECF (possibly predominant), as opposed to an increase in fat weight in these subjects. This is supported by a significant correlation between the increase in weight and increase in TBW (FA-MS), r=0.45, P=0.035, despite these being entirely independent measurements in methodological terms. These observations were also supported by the ultrasound measures of biceps, triceps and anterior thigh muscle and fat depth (see Table 3). Women lost upper body fat (eg mean biceps fat reduced from 14.5 to 13.1 mm, P=0.001), with a concomitant increase in muscle (mean biceps increased from 24.3 to 27.0 mm, P=0.001), whereas changes were much less marked in men.

Table 1.

Measurements (SD) of weight, TBW and BIA resistance data at baseline and 1 year follow up

	Males		Females		All subjects
M	Baseline	1 year	Baseline	1 year	Baseline	1 year
Weight kg	84.3	85.0	67.9	69.6	75.4	76.6
	(16.1)	(18.2)	(11.4)	(13.7)	(15.8)	(17.4)
TBW	46.0	46.7	32.6	34.2†	38.7	39.9*
FA-MS (kg)	(6.7)	(6.9)	(3.2)	(3.9)	(8.4)	(8.3)
TBW	43.9	44.9	29.4	29.9	36.0	36.7
BIA (kg)	(9.4)	(9.4)	(3.7)	(4.7)	(10.0)	(10.3)
ECF	19.6	19.9	13.4	14.0‡	16.2	16.7¶
BIA(kg)	(3.5)	(3.1)	(1.5)	(1.9)	(4.0)	(3.8)
ICF	24.2	25.0	16.0	15.9	19.7	20.0
BIA (kg)	(6.30	(6.5)	(2.3)	(2.9)	(6.1)	(6.6)
TBW	43.7	44.6	29.6	31.0*	36.0	37.1*
SFT (kg)	(6.6)	(7.2)	(3.6)	(4.6)	(8.8)	(9.0)
% body fat	28.5	27.2	39.7	38.4‡	34.6	33.3‡
(SFT)	(3.7)	(6.0)	(3.7)	(4.1)	(6.7)	(7.5)
Re (ohm)	559.6	544.8	683.1	653.0‡	627.0	604.0¶
	(60.4)	(33.8)	(68.1)	(63.6)	(89.2)	(75.2)
Ti (ohm)	1251.2	1207.9	1574.3	1594.9	1427.4	1419.0
	(242.0)	(248.5)	(231.9)	(262.9)	(283.5)	(318.7)

TBW = Total Body Water, FA-MS = Flowing after glow mass spectrometry, BIA = Bioelectrical impedance analysis, ECF = extracellular fluid, ICF = intracellular fluid, SFT = skin fold thickness, Re and Ri = resistance at zero and infinite frequencies

^*P<0.005

^¶P<0.02

^†P<0.01

^‡P<0.05

comparing 1 year with baseline.

Table 2.

Differences in measurements of weight, TBW and BIA resistance data between baseline and 1 year follow up

	Mean (SD) change over 1 year
	Men	Women	All subjects	95% CI annual change	CV over 1 year	R
Weight (kg)	0.7 (3.5)	1.61 (3.7)	1.21 (3.8)	−0.42 to 2.84	4.56%	0.98
TBW FA-MS (kg)	0.65 (1.6)	1.64 (1.8)	1.19 (1.7)	0.39 to 1.98	5.0%	0.98
TBW BIA (kg)	0.97 (1.9)	0.47 (1.6)	0.70 (1.7)	−0.09 to 1.48	4.75%	0.99
ECF (kg)	0.24 (0.7)	0.59 (0.7)	0.43 (0.7)	0.08 to 0.78	4.78%	0.93
ICF (kg)	0.72 (1.5)	−0.12 (0.9)	0.26 (1.3)	−0.32 to 0.85	6.63%	0.95
Re (Ohm)	−14.7 (35.2)	−29.9 (40.6)	−23.0 (38.2)	−6.0 to −39.9	6.25%	0.81
Ri (Ohm)	−43.3 (81.0)	20.67 (100.5)	−8.4 (95.7)	−50.8 to 34.0	6.73%	0.92
TBW-SFT	0.85 (1.4)	1.32 (1.3)	1.10 (1.3)	0.49 to 1.72	11.6%	0.98

CI = Confidence interval; CV = Coefficient of variation; R = Pearson regression coefficient, all subjects comparing time points; TBW = Total Body Water; FA-MS = Flowing after glow mass spectrometry; BIA = Bioelectrical impedance analysis; ECF = extra-cellular fluid; ICF = intracellular fluid; SFT = skin fold thickness; Re and Ri = resistance at zero and infinite frequencies.

Figure 1.

Correlation between measures of body composition at baseline and one year: (a) weight, (b) TBW from deuterium dilution employing rapid breath test (FA-MS), (c) TBW derived from multi-frequency BIA and (d) TBW derived from anthropometry (SFT).

Table 3.

Measurements (SD) of weight, TBW and BIA resistance data at baseline and 1 year follow up

	Males		Females		All subjects
M	Baseline	1 year	Baseline	1 year	Baseline	1 year
Biceps muscle	31.7	34.8†	24.3	27.0*	27.7	30.6*
(mm)	(6.2)	(5.9)	(4.4)	(5.4)	(6.41)	(1.3)
Triceps muscle	39.0	38.9	27.4	30.1	32.6	34.1
(mm)	(11.4)	(10.2)	(6.6)	(5.2)	(10.6)	(8.9)
Thigh muscle	32.7	33.9	28.3	29.3	30.3	31.4
(mm)	(6.8)	(7.0)	(5.1)	(4.1)	(6.2)	(6.0)
Biceps fat	8.4	7.5	14.4	13.0*	11.7	10.5*¶
(mm)	(2.6)	(2.1)	(4.12)	(4.2)	(4.6)	(4.4)
Triceps fat	9.5	8.8‡	17.6	16.8	13.9	13.2
(mm)	(2.2	(2.1)	(4.1)	(4.9)	(5.2)	(5.6)
Thigh fat	10.1	8.5*	18.7	18.5	14.8	14.0
(mm)	(3.7)	(2.9)	(5.1)	(6.7)	(6.2)	(7.3)

^*P<0.005

^†P<0.01

^‡P<0.05

comparing 1 year with baseline.

Relationships between different methods of measuring body water

Bland-Altman analysis was performed to establish whether the relationship between different estimates of TBW, and in particular the nature of any systematic error, was the same at baseline and year 1 (see summary Table 4). It can be seen that any differences observed between the three methods were very similar at baseline and after 1 year. Both BIA and SFT tend to give, on average, estimates of TBW that are 2 kg less than those obtained using FA-MS. When expressed as the proportion of body weight, these methods give rather low estimates (BIA: men, baseline 52% (±0.04) and at one year 53% (±0.05), women 44% (±0.04) and at one year 43% (±0.05); SFT: men, baseline 52% (±0.03) and at one year 53% (±0.04), women 44% (±0.03) and at one year 45% (±0.03)) compared to estimates by FA-MS: men, baseline 55% (±0.05) and at one year 56% (±0.05), women 48% (±0.04) and at one year 50% (±0.03). In addition to these systematic differences, the methods varied between each other as a function of subject size (see figures 2a-c). This was most marked when one of the methods was BIA that tended, when compared to the other methods, to underestimate TBW in smaller and over-estimate in larger subjects, so that the apparent small difference between BIA and SFT (Table 3) was in fact an average of over and under estimates (Figure 2c). The regression lines in Figures 2a-c for baseline and year 1 comparisons almost superimpose each other in each case and are not significantly different, indicating stability in these systematic and non-systematic differences over the course of the study period.

Table 4.

Summary of the Bland-Altman analysis of different methods estimating TBW (FA-MS, BIA and SFT) at baseline and year 1

TBW methods of measurement		Baseline	Year 1	P value*
FA-MS - BIA	Mean difference Correlation	2.3 (2.76) R=0.58, P=0.005	3.1 (2.73) R=0.76, P=0.001	0.103
FA-MS - SFT	Mean difference Correlation	2.6 (1.74) R=0.18, P=0.401	2.7 (2.16) R=0.34, P=0.12	0.860
BIA - SFT	Mean difference Correlation	−0.01 (3.1) R=0.39, P=0.07	−0.41 (3.1) R=0.43, P=0.046	0.297

^*Paired t-test, comparing 1 year with baseline.

TBW = Total body water; FA-MS = Flowing after glow mass spectrometry; BIA = Bioelectrical impedance analysis.

Figure 2.

Bland and Altman plots for differences between methods in estimating TBW using deuterium dilution (FA-MS), BIA and SFT according to average TBW using both methods (a) FA-MS v. BIA, (b) FA-MS v. SFT and (c) BIA v. SFT. In each case the data at baseline (■) and one year (□) is superimposed along with the linear regression lines (baseline continuous, year one interrupted). It can be seen that the systematic errors between the methods are stable over the study period.

Discussion

The primary objective of this study was to evaluate the stability and reproducibility of a number of near-patient methods of body composition measurement, in particular of our novel rapid FA-MS breath test in the estimation of body water over a 12-month period. Healthy subjects were selected with the assumption that their body composition would remain stable. The high degree of correlation, closeness of the slope of the regression line to unity and the low coefficient of variation suggest that the estimation of TBW by FA-MS is of a similar reproducibility to that of conventional weighing over the same time period. In performing the observations a year apart we have also confirmed the lack of any base-line drift in the technique which can occur in some methodologies. One of the main advantages of FA-MS is the lack of requirement of calibration that is time consuming and undesirable in the clinical environment. This is due to the use of an isotope ratio, rather than absolute measures, in the determination of D abundance. The reproducibility of BIA was similar, although the derived estimate of TBW uses the current body weight in its calculation and is thus heavily influenced by the reproducibility of the weight, whereas the electrical resistance measurements were a little more variable. Most variable of all the measures was the estimate of the TBW from skinfolds, although again the slope of the regression line was still close to unity. This is perhaps not surprising as the estimate of TBW from skin-folds involves five anthropometric measures, with the possibility of multiplying error. These comparisons endorse TBW estimation using our breath test for deuterium abundance as a complimentary near-subject method of body composition analysis, with at least equivalent accuracy and precision when compared to established methods.

Although the body composition of our healthy subjects was very similar over the period of one year, small but statistically significant differences did occur. There was a tendency to increase weight, more marked in women, which both FA-MS and SFT attributed to increased body water rather than fat. BIA was relatively insensitive to this change, although there was, again in women, an increase in the extra-cellular water. These changes were predominant in the upper part of the body and were associated with a reduction of arm fat and an increase in arm muscle by ultrasound, although it should be emphasised that these changes were close in magnitude to the discriminatory power of the techniques employed. Whilst small and not anticipated, there is no reason however to doubt that these differences are real. That BIA and SFT gave different estimates of changing body composition may reflect that fact that these two methods differ in their estimate of percentage body fat (and thus TBW) according to fat distribution [19]. These between-method differences will also vary according to the formulae used to calculate absolute values, an inherent weakness of both these techniques [11]. This reinforces the value of combining more that one method at a time in the assessment of body composition, one of which should be an absolute measurement rather than a value calculated, at least in part from population derived equations. Our goal in developing the FA-MS deuterium breath test as a near-subject and readily available absolute measurement of body water was driven by this problem.

Whilst many cross-sectional studies of body composition are published, those with longitudinal data are found less frequently in the literature. Most of these address the issue of changing body composition in the elderly [20-23] or at times of life change such as the menarche or menopause in women, [24, 25] although there is one large study (Fels Study) of white men and women between 18 and 64 years [26]. Studies in the elderly show a gradual tendency for loss in lean body mass, more marked in men than women and blacks than whites. Rates of loss vary, but are the order of 1% per annum. Estimates of changes in body fat differ between studies from zero over five years using BIA [23] to ∼1% in women using dual energy X-ray absorptiometry [20]. In middle age, body water is relatively well preserved [26]. Our subjects represented a wide age range (50 years), deliberately selected to be broadly representative of the age distribution affected by chronic disease. In fact, the subjects in whom significant changes in body composition occurred were confined to the younger volunteers, ie less than the median age of 68 years (data not shown separately). The observation that the women in our study tended to lose fat mass runs counter to larger population studies. It must be remembered however that the volunteers for our study (they were recruited by advertisement and had expenses reimbursed but where not paid) represent a small and selected group of individuals who were relatively health conscious. The women in particular may have taken steps to avoid fat weight gain by increasing their exercise over the study period.

The differences in estimates for TBW by the three methods were consistent over the course of the study period. Whilst BIA and SFT at first site seem to agree more closely with each other than FA-MS, it is clear from the Bland-Altman analysis that this is only the case for average values, with greater variance at the extremes of TBW measurement. The absolute values for TBW obtained using deuterium dilution FA-MS were much more in keeping with typical values using dilution techniques in the literature and as discussed previously [27, 28]. The finding that these systematic errors between the methods were stable over the course of the study, apart from supporting reproducibility, suggests that these tools might be used together in a different way. Rather than using population derived equations to determine the absolute values for TBW from BIA or SFT it may be more appropriate to develop subject specific equations relating impedance or skin-fold measurements to an absolute measure of TBW, enabling clinicians or investigators to track changes in body composition longitudinally [29].

In summary, we have established that on-line breath analysis of deuterium abundance can be used in conjunction with other near-subject techniques with similar reproducibility to monitor body composition. Whilst in these healthy volunteers body composition was stable over 12 months, the methods combined were able to detect relatively subtle changes, especially in younger women. This is a promising approach for the management of individuals with chronic disease, eg renal failure, in whom maintaining normal body composition is an important indicator of therapeutic success.