| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Nutritional Sciences, University of Wisconsin, Madison, Wisconsin 53706
Submitted 26 July 2002 ; accepted in final form 14 September 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
magnetic resonance imaging; simulated microgravity; caloric restriction; body composition
As space missions lengthen from weeks to years, the need for in-flight methods to monitor the muscle mass of astronauts becomes critical to health maintenance. In-flight measurements of muscle mass loss have been limited to anthropometric methods such as leg circumferences (28, 30). However, these methods have a relatively poor sensitivity and can be affected by the fluid shifts that occur with adaptation to a microgravity environment (7). More sophisticated measurement techniques (e.g., MRI, computed tomography scanning) are not possible in-flight because of the limitations on mass, size, and electricity posed by spaceflight.
One method that would meet the practical requirements for spaceflight as well as potentially have the sensitivity to detect muscle mass loss is bioelectrical impedance analysis (BIA). BIA employs a single-frequency (50 kHz), alternating-current signal that is applied across the body at very low (safe) currents that are undetectable. The ionic media in fat-free mass conduct the current while the body fat's insulating properties provide resistance to current flow. The length2/resistance of the body is highly correlated to total body water and fat-free mass (16, 20, 26).
Several studies have investigated the potential of singlefrequency (50 kHz) BIA in measuring arm and/or leg muscle volume (5, 24, 25) or the volume of discrete segments of these limbs (9, 10, 19, 21, 22). Most studies have attempted to correlate the impedance index (segment length2/resistance) with muscle volume predicted by dual-energy X-ray absorptiometry or MRI. Although BIA's accuracy (mean differences of 5–11%) has been fairly good, the precision has been relatively poor (9, 10) except in a case where the BIA was systematically (57 ± 1%) lower (19).
Although these are promising results, it is likely that BIA would fail to predict muscle volume well under conditions of spaceflight. Because the primary current path at 50 kHz is extracellular (12), any disruption in the ratio of extracellular water (ECW) to intracellular water (ICW) would increase the error of body composition prediction (4, 12). This clearly poses a problem for analysis during spaceflight in which ECW is redistributed from the periphery to the trunk and substantial extracellular losses occur (6, 23).
A method that does have the ability to distinguish intracellular and extracellular fluid spaces is bioelectrical impedance spectroscopy (BIS). BIS measures the resistance and reactance at various frequencies, typically ranging from 5 kHz to 1,000 kHz. When the data are modeled in a Cole-Cole plot (resistance vs. reactance over the frequency spectrum), the resistance of the intracellular fluid (Ri) and extracellular fluid (Re) can be determined (4) and ECW and ICW can be differentiated (1, 31).
In addition, BIS can be used to predict the ECW, ICW, and total water volumes of a limb segment if it roughly approximates a cylinder. The equation V = 
L2/R, where is the specific resistivity of the fluid medium (
·cm), L is the length of the segment (cm), and R is the resistance (), can be used to predict the total limb segment fluid volume (V) in liters. By using emprically derived resistivities (8), the ECW and ICW, and total water content of a segment can be calculated.
The purpose of this study was to validate BIS with ColeCole modeling for the estimation of appendicular muscle volume and muscle volume changes during 72 h of simulated microgravity and caloric restriction. We expected that the 6° head-down tilt position would cause headward fluid shifts, which would produce losses in muscle volume from the calf, and smaller losses from the arm. To further simulate early phases of spaceflight, subjects were placed on a short-term, energy- and protein-restricted diet. We expected that this dietary restriction would cause additional depletion of muscle volume in both regions. We hypothesized that the total segmental fluid volume by BIS would be highly correlated to the segmental muscle volume by MRI and that BIS would track the changes in MRI muscle volume with accuracy and precision.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Protocol. Subjects were admitted to the University of Wisconsin General Clinical Research Center on the evening of day 1 and remained on the unit until all procedures were completed on the evening of day 4. On the evening of day 1, subjects received BIS and MRI scans of their upper arm and lower leg. Later that evening, subjects began 
72 h of bed rest in the 6° head-down tilt position to simulate the fluid shifts that occur in microgravity. Starting on the morning of day 2 and progressing to the end of the study on day 4, subjects were placed on a low-protein, hypocaloric research diet designed to produce negative protein and energy balance. The diet was a low-salt, clear liquid-type diet containing 800 kcal, 4 g protein, 5 g fat, and 185 g carbohydrate per day. Subjects were required to consume at least 30 ml water·kg-1·day-1 and all research diet foods. On the evening of day 4, subjects received the posttreatment BIS and MRI scans of their upper arm and lower leg. When all procedures were completed, subjects were brought to an upright position and given food and fluids. Subjects were discharged when they demonstrated orthostatic tolerance.
Nitrogen balance. Nitrogen balance during the 72-h span between BIS and MRI scans was estimated from the difference of dietary protein intake and urinary nitrogen losses. Because few subjects had bowel movements during this period, no corrections for fecal losses were made. Dietary protein intake for the defined hypocaloric diet was calculated by using the Diet Planner program (University of California Clinical Research Center, San Francisco, CA). During hospitalization, all urine was stored under refrigeration in plastic collection jugs. Hydrochloric acid was added to the jugs to ensure retention of ammonia. Within 24 h of discharge, all samples were pooled and an aliquot was taken for analysis. Urinary nitrogen was measured in quadruplicate by using an automated nitrogen analyzer (Antek, Houston, TX).
Definition of arm and leg segments. A 13.3-cm section of the upper right arm and a 16.1-cm section of the right leg were modeled by using both MRI and BIS. For the arm segment, the elbow was bent at a 90° angle, and a straight edge was applied to the underside of the forearm. A ruler edge was applied to the backside of the upper arm so that it bisected the straight edge at a 90° angle. A line was drawn 10 cm up from the tip of the ventral side of the elbow. The arm was straightened and then abducted so that it was parallel to the floor, and the mark was extended to circumscribe the arm in a circle that was perpendicular to the floor. Another circumscribing line was drawn 13.3 cm proximal to the first line. For the leg segment, a mark was drawn 2 cm distal to the tibial tuberosity. While subjects were standing, the mark was extended to circumscribe the leg in a circle that was parallel to the floor. Another circumscribing line was drawn 16.1 cm distal to the first line.
MRI procedure. Gel beads (containing oil) were applied to the dorsal side of the leg and the ventral side of the upper arm at the sites of the circumscribing lines to allow for identification of defined limb segments in the scans. By using a wrap coil for the arm and an extremity coil for the leg, oblique axial scans of the arm and leg were obtained by a 1.5-T GE MRI system with LX platform software (GE Medical Systems, Milwaukee, WI). Longitudinal relaxation timeweighted, spin-echo, consecutive multislice scans with a slice thickness of 7 mm, a time to echo of 25 ms, and a time to repeat of 600 ms were used to differentiate between muscle, fat, bone, and air. By using the gel beads as a guide, 23 leg slices (7 mm thick) and 19 arm slices (7 mm thick) were selected for analysis. The slices were processed with the custom-designed three-dimensional modeling package of an Advantage Windows workstation (software version 2.0.20). In brief, thresholds were manually set to eliminate air, bone, and subcutaneous fat from the scans. The remaining tissue, called "muscle," included muscle tissue, intramuscular fat, intramuscular vasculature, and intramuscular connective tissue. Then, the software integrated the consecutive slices of muscle to produce a muscle volume estimate for the entire set of slices. The muscle volume modeling procedure was conducted in a blind fashion by the same investigator on three separate occasions. For outlying values, the procedure was repeated a fourth time. The coefficient of variation (CV) for repeated (triplicate) analyses of the same scan by one investigator in this study was 8.1% for the calf and 5.6% for the arm. The average volume of triplicate measures was used as the criterion measure of muscle volume for further statistical analysis.
BIS procedure. The areas of the arm and leg containing the circumscribing lines were cleaned with isopropyl alcohol. Then 0.5-cm-wide gel electrode tape with lead-attachment "tags" (Electro-Diagnostic Instruments, Valencia, CA) was placed directly on the skin and wrapped around the arm and leg directly on the circumscribing lines for voltage detection. Standard current injection electrodes patches (7.6 x 1.9 cm) with self-adhesive conducting gel (Xitron Technologies, San Diego, CA) were placed on the skin 8 cm away from the electrode tape. BIS scans of the arm and leg segments were completed in triplicate with a Xitron Hydra 4200 BIS instrument (Xitron Technologies, San Diego, CA) by the same investigator. In this study, the CV for triplicate measures was 1.8% for Re and 4.0% for Ri in the arm and 1.7% for Re and 2.4% for Ri in the calf. The laboratory's day-to-day variability (CV) in Re and Ri values of muscle segments without bed rest was 4.0% for Re and 6.9% for Ri in the arm and was 6.9% for Re and 6.0% for Ri in the calf, in a convenience sample (n = 5 men). The average of Re and Ri was used for muscle volume calculations. The equation ECW volume = pL2/Re, where p is the specific resistivity of the ECW (·cm), L is the length of the segment (cm), and Re is the ECW resistance (), was used to predict the ECW volume (liters) in the segment. The empirically determined whole body Re values are 40.5 and 39 ohm·cm for men and women, respectively (8). The equation ICW volume = pL2/Ri, where p is the specific resistivity of the ICW (ohm·cm), L is the length of the segment (cm), and Ri is the ICW resistance (), was used to predict the ICW volume (in liters) in the segment. The empirically determined whole body Ri values are 265 and 274 ·cm for men and women, respectively (8).
Statistics. Descriptive data are presented as means ± SD except where noted. Statistical differences between pre- and posttreatment and between methods were detected by using paired t-tests. Simple linear regression was used to test the relationship between MRI muscle volume and total segmental water volume, as well as the change in muscle volume and the change in water volume. Analysis of the error between methods included the bivariate correlation (r) and standard error of the estimate (SEE).
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
|
The measured muscle volume and predicted water volumes of the calf and arm segments at both time points are given in Table 2. Strong correlations were observed between MRI muscle volume and BIS-predicted total water volume of the calf (r = 0.93) and arm (r = 0.84) segments at baseline (Fig. 1). For the arm, the relationship was as follows: segmental skeletal muscle (ml) = 1.23 (BIS water volume) - 14.6; r2 = 0.93, SEE = 61 ml, and P < 0.001. For the calf, the relationship was as follows: segmental skeletal muscle (ml) = 1.26 (BIS water volume) - 68.3; r2 = 0.87, SEE = 89 ml, and P < 0.001. The slopes (P = 0.88) and intercepts (P = 0.36) were not statistically different between the calf and arm data sets. The combined regression equation was as follows: segmental skeletal muscle (ml) = 1.19 (BIS water volume) - 6.7; r2 = 0.97, SEE = 76 ml, and P < 0.001. High correlations between MRI muscle volume and BIS-predicted total water volume of the calf (r = 0.84) and arm (r = 0.96) segments were also seen on day 4. A Bland-Altman plot comparing BIS and MRI methods is presented in Fig. 2. For the combined arm and calf data set, the regression equation was as follows: difference MRI - BIS (ml) = -0.20 (average volume) + 26; r2 = 0.43, SEE = 69 ml, and P < 0.001.
|
|
|
The change in muscle volume, ECW, ICW, and total water of the segments produced by the treatment is shown in Fig. 3. According to MRI measurements, the calf segment volume decreased by 49 ± 68 ml, which was an 6% decrease (P = 0.02). BIS-predicted changes in the calf included a net loss of 38 ± 61 ml in total water (4% decrease; P = 0.02), which was made up of a loss of 52 ± 19 ml loss in ECW (28% decrease; P < 0.001) and nonsignificant gain of 13 ± 61 ml in ICW (1% increase). MRI-measured arm segment volume decreased by 18 ± 23 ml, which was an 4% decrease (P = 0.02). BIS-predicted arm changes included small, nonsignificant decreases in total water content (loss of 11 ± 24 ml, 2% decrease), ECW (loss of 1 ± 7 ml, 1% decrease), and ICW (loss of 10 ± 23 ml, 3% decrease). Mean changes in MRI muscle volume and BIS total water were not significantly different for the calf or arm (P value for calf = 0.75 and for arm = 0.61). The relationship between MRI- and BIS-predicted volume changes for the arm and calf are shown in Fig. 4. The correlation between methods was fair for the calf (r = 0.70) and poorer for the arm (r = 0.22). The relative prediction error for change was 6% for both the calf and arm.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Assumptions of the BIA model, including constancy in cross-sectional area, cylindrical shape, and muscle pennation, appear to make little difference in the strength of this relationship (21). Our finding that the slopes of the regression lines for MRI muscle volume vs. BIS total water volume were strikingly similar for both the calf and arm (see Fig. 1) supports this conclusion. This is notable given that most subjects had conical-shaped upper arm segments and cylindrical-shaped calf segments and that the cross-sectional area was more variable for the arm segment than the leg segment. We cannot be sure, however, that this would be the case if patch electrodes were used in place of the circumferential electrodes used in the present study, particularly in light of the difference in resistivities of muscles when current flow is not parallel to the muscle fibers. In addition, we have used whole body specific resistivities, which may or may not apply to the specific limbs segments measured in this study. The error between techniques may, in part, be due to the difference between whole-body and limb specific resistivities as well as interindividual differences in limb specific resistivities.
The bed rest and underfeeding protocol produced an average of 2.1-kg loss over the 72 h. If this weight loss consisted entirely of water, it would correspond to 5.5% of total body water, which is substantially more than short-term bed rest protocols providing adequate nutrition (11). In addition, the nitrogen balance data suggest that subjects were substantially underfed. During the 72 h of treatment, mean nitrogen losses were 37 ± 8 mg·kg-1·day-1 (22 ± 8 g), which corresponds to a loss of 135 ± 49 g of protein and 540 g water during the interval (13).
The MRI measured loss of muscle volume of the arm was 17 ml, or 4.3% the segment volume (P < 0.05). This volume loss is consistent with data from three male astronauts during the Skylab 4 mission (29) who had very minimal arm volume changes (<200 ml for the entire left arm) during the mission. The calf volume losses (6%; P < 0.05) are within the range described in MRI and anthropometric studies of astronauts after short spaceflight missions (2, 14, 18). However, it is unclear from these studies how much of the volume decrease is due to fluid shifts vs. true muscle atrophy. It has been hypothesized that the loss in leg muscle volume in the first 3 days of microgravity are due to fluid shifts (e.g., ECW) and that the later losses are due to atrophy (ICW) (7). However, these assumptions cannot be tested by MRI, which works on the tissue-system level of organization (32). In this model, the body is compartmentalized into skeletal muscle, adipose tissue, bone, blood, and residual components. This tissue-system model can measure a loss in muscle volume, but it cannot discern whether that loss is due to fluid shifts or a loss of body cell mass.
In contrast, BIS works on the cellular level of organization, in which the body is divided into extracellular fluid, extracellular solids, and body cell mass (which contains ICW and associated solids) (32). Thus BIS is able to differentiate losses in ECW and ICW. This approach could theoretically be used to determine whether muscle volume losses are due to the headward shifts of ECW or actual body cell mass loss, which is indicative of malnutrition and muscle atrophy (15). BIS could serve as an adjunct technique for qualifying muscle changes in flight when combined with more standard ground-based techniques such as MRI.
We expected that MRI muscle volume and BIS total water volume of the segments would be not only highly correlated, but also fairly similar. Given that the water volume of the segment is found in the fat-free mass compartment and that skeletal muscle makes up most of the fat-free mass of the limb segment, we expected the MRI muscle volume and BIS total water volume to be correlated. In addition, most of the ECW and ICW in the limb segment are likely to be found in tissue that the MRI instrument identified as skeletal muscle. This MRI-identified muscle volume included the limb segment's skeletal muscle, intramuscular connective tissue, intramuscular vasculature, and intramuscular fat depots. However, we assumed that there would be two limitations with this approach. First, the tissues excluded from the muscle volume (skin, bone, and subcutaneous fat) contain a significant volume of ECW and ICW. The water contribution of these tissues would cause BIS-predicted total segmental water volume to overestimate MRI muscle volume. Second, skeletal muscle is only 75% water, with the remainder being proteins, salts, pigment, and metabolic substrates. Thus the segmental total water volume could potentially underestimate MRI muscle volume. In our data, the BIS total water volume exceeded muscle volume by an average of 19% for the arm and leg. In addition, the Bland-Altman analysis (Fig. 2) revealed significant bias between the techniques such that as segmental volume increased the discrepancy between techniques increased. BIS water volume overestimates MRI muscle volume as the total segmental muscle and water volume increases. This may relate to inconstancy in the ratio of muscle structure (e.g., protein and inclusive metabolites) to ECW, ICW, and total segmental water.
We also hypothesized that the BIS-predicted water volume changes would be similar to MRI-measured muscle volume changes. Indeed, the mean change in total water and muscle volumes was remarkably similar (Fig. 3). However, the BISpredicted changes in segmental water did show some ability to discern between volume changes due to fluid shifting and volume changes due to atrophy. The 72 h of caloric restriction and bed rest was designed to produce losses in the ICW of the arm and both the ECW and ICW of the calf. We expected that the calf would lose ECW in response to the fluid shifts that occur with the 6° head-down tilt position and that it would lose body cell mass (measured by ICW changes) in response to underfeeding. We also expected that the arm would lose ICW in response to the dietary treatment but that it would not experience the ECW losses seen in the lower limb.
Indeed, the BIS-predicted ECW changes in the calf and arm are consistent with expectations for the head-down tilt treatment. The ECW content of the calf was decreased by 28%, although there was essentially no change in ECW in the arm (Fig. 3). In contrast, it appears as though the dietary treatment did not cause a sufficient change in segmental ICW levels to be detected by the BIS instrument. The change in ICW levels of both limb segments did not reach statistical significance. This was despite a significant decrease in whole body protein stores, as evidenced by the 72-h urinary nitrogen losses (Table 1).
One explanation for the inability to produce a substantial loss in calf and arm segmental ICW may be that the majority of ICW losses resulting from the short-term dietary treatment were confined to the trunk region. Regional (elbow to trunk, trunk, and knee to trunk) BIS scans suggest that the majority (55%) of body water lost was intracellular and that 65% of the ICW losses were localized to the trunk region (3). The change in arm region total water was estimated to be just 0.6%. Thus the combination of underfeeding and bed rest produced significant changes in trunk ICW, but the treatment was not of sufficient duration or intensity to produce arm and leg ICW losses within the limits of detection by the BIS instrument.
As can be seen in Fig. 4, the relationship between change in muscle volume and change in total segmental water was reasonably good for the calf (r = 0.70) but was poorer for the arm (r = 0.22). One reason the correlations were fairly low and the total error of prediction was high relates to the precision of measurement for both BIS and MRI. The standard deviations for change in muscle volume and change in total water volume are nearly identical. Thus we can conclude that both the MRI and BIS methods contributed to the variability in measuring change and that the noise in the data was due to both methods. In addition, the arm had such a small decrease in volume (11-ml average by MRI) that the range was too small to adequately assess the relationship between individual change in muscle volume and change in water volume given the limits of precision for both techniques in our hands.
In summary, this study examined the potential of multifrequency BIS with Cole-Cole modeling in evaluating skeletal muscle volume and muscle volume changes of calf and upper arm segments with MRI as a reference. BIS-predicted total segmental water volume was highly correlated to MRI-measured segmental muscle volume in both the calf and arm. In addition, the BIS method detected the ECW shifts in the calf that were a result of the head-down tilt treatment. Significant individual variability in both BIS and MRI measurements combined with limited volume changes made it difficult to assess the individual accuracy and precision of BIS-predicted changes in volume. Additional BIS validation research should include longer bed rest protocols that produce more significant limb muscle change.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
This work was supported by National Aeronautics and Space Administration Grant NAG9-1039, National Institutes of Health General Clinical Research Centers Program of the National Center for Research Resources Grant M01 RR-03186, and a training grant from the National Institutes of Health.
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  N. I. Tanaka, M. Miyatani, Y. Masuo, T. Fukunaga, and H. Kanehisa Applicability of a segmental bioelectrical impedance analysis for predicting the whole body skeletal muscle volume J Appl Physiol, November 1, 2007; 103(5): 1688 - 1695. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Stahn, E. Terblanche, and G. Strobel Modeling upper and lower limb muscle volume by bioelectrical impedance analysis J Appl Physiol, October 1, 2007; 103(4): 1428 - 1435. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Ishiguro, H. Kanehisa, M. Miyatani, Y. Masuo, and T. Fukunaga Applicability of segmental bioelectrical impedance analysis for predicting trunk skeletal muscle volume J Appl Physiol, February 1, 2006; 100(2): 572 - 578. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
