| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research Communications |
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: The purpose of this study was to compare bioelectrical impedance with metabolic activity in healthy and critically injured patients. If bioelectrical impedance accurately determines body composition during critical illness, the slope between body-composition variables and oxygen consumption would be the same in critically injured and healthy subjects.
Design: There is a strong linear relation between body composition and metabolic activity. In the present study, body composition (fat-free mass and body cell mass) was determined by using bioelectrical impedance and resting metabolic activity (metabolic rate and oxygen consumption) by using gas exchange analysis in a group of healthy and critically injured subjects. The relation between these variables was compared by using linear regression to a similar relation established by hydrostatic weighing in a large historical control group.
Results: The slope of the line relating fat-free mass to resting metabolic rate was the same in the healthy and critically ill groups (P = 0.62) and each was similar to the slope of the line for the control group. However, in 37% of the critically injured group, overhydration contributed to an increase in fat-free mass, disturbing the relation with resting metabolic rate. The slope of the line relating body cell mass to oxygen consumption in our healthy and critically ill groups was almost identical.
Conclusion: These results support the use of bioelectrical impedance to determine body cell mass in healthy and critically ill subjects.
Key Words: Bioelectrical impedance • body composition • critical illness • fat-free mass • body cell mass • resting metabolic rate • trauma • humans
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Bioelectrical impedance is a noninvasive, inexpensive, portable method of body-composition analysis, which is appealing for both research and clinical practice (1). Current techniques of bioelectrical impedance measure electrical resistance and reactance across one or more signal frequencies. Resistance is proportional to the fluid and electrolyte content of the body, whereas reactance is thought to be a measure of the capacitance of cell membranes (for review, see reference 1). Mathematic models constructed from comparisons of resistance measurements with other methods of body-composition analysis have been used to develop equations to compute total body water, body fat, and fat-free mass in healthy subjects. If reactance is measured, tissue masses can be further compartmentalized into body cell mass, intracellular water, and extracellular water, again, through the application of mathematic models constructed from comparison of resistance and reactance measurements with other body-composition methods in healthy persons.
Although bioelectrical impedance has been validated in healthy populations and certain disease states (2–6), it has not been tested extensively in critically ill patients (7, 8). This is important because previous studies have shown that bioelectrical impedance measurements should be verified in the population of interest (9). Furthermore, a recent National Institutes of Health consensus statement cited critical illness as an area requiring further research into the application of the bioelectrical impedance technique (10).
Previous studies have shown a linear relation between body composition and energy metabolism (energy expenditure and oxygen consumption) (11). In the current study, the slope of the line relating fat-free mass (measured by bioelectrical impedance) to resting metabolic rate in a group of healthy volunteers was compared with the same relation in a control group whose fat-free mass was determined by hydrostatic weighing. We then compared the same relation between a group of critically injured subjects (fat-free mass measured by bioelectrical impedance) and 2 healthy groups (11). Because of the potential for alterations in extracellular water (and therefore fat-free mass) during critical illness, a final analysis was performed comparing the relation of body cell mass with oxygen consumption in the critically injured and healthy groups.
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Subjects
Critically injured group
Critically injured patients were recruited from the surgical intensive care unit of our institution, which is a level I trauma center. Subjects were eligible for the study if they had a blunt or penetrating injury (exceptions: spinal cord injury, burn injury, and isolated brain injury), were expected to require mechanical ventilation for 
5 d, were expected to require nutritional support, were not receiving steroids, and had no history of neuromuscular paralysis or impairment, diabetes mellitus, renal disease, or hepatic disease.
Enteral or parenteral nutrition was initiated within 48–72 h of injury. Feeding continued for 4 d, at which time whole-body electrical resistance and reactance were measured with bioelectrical impedance and gas exchange was measured with indirect calorimetry. Entrance criteria and the timing of studies were planned to minimize variability in resting metabolic rate and oxygen consumption from factors other than body composition (eg, feeding state, sepsis, and paralysis).
Healthy group
The healthy group consisted of persons who responded to posted advertisements for the study. Subjects were selected for the study if they had no current infection, were not receiving steroids, were ambulatory, and had no history of neuromuscular paralysis or impairment, diabetes mellitus, renal disease, or hepatic disease. Subjects arrived at the test site in the morning after having fasted the night before. Weight and height were measured and then. as subjects rested in a supine position, gas exchange and bioelectrical impedance analyses were performed.
Methods
Gas exchange was measured with an open-circuit indirect calorimeter (Deltatrac MB101 Metabolic Monitor; SensorMedics, Anaheim, CA). All critically ill subjects were mechanically ventilated and gas collection was achieved through the ventilator. For healthy subjects, gas collection was accomplished by placing a rigid, clear plastic canopy over the subject's head while the subject breathed normally. A test period consisted of 30 consecutive minutes during which the subject was observed to be at rest and for which the CVs for oxygen consumption, carbon dioxide production, and minute ventilation (in the mechanically ventilated subjects) was 
10%. The data collected in the first 5 min were automatically discarded. If resting conditions could not be maintained for 30 min, a 5-min test period with CVs 5% was accepted (12). Resting metabolic rate was calculated from oxygen consumption and carbon dioxide production by using the modified de Weir equation (13) and then converted to kJ/min.
Bioelectrical impedance measurements were performed with the BIA-101 Quantum Plethysmographer (RJL Industries, Detroit), a single-frequency device (50 kHz). Nutrition and hydration were administered continuously in the critically injured group. In contrast, the healthy group was fed intermittently and fasted for 12 h before the measurements were made (to maintain steady state conditions in both groups). The first step in the bioelectrical impedance technique was to measure the subjects' height and weight. For the bedridden, critically injured subjects, height was measured with a tape measure while the subject was stretched out in a supine position in bed. Weight was measured with a sling-type bed scale. Heights and weights of the healthy subjects were measured while subjects were in a standing position with a stadiometer and balance scale, respectively. All subjects were placed in the supine position for 10 min before bioelectrical impedance measurements were made. Arms and legs were abducted 
30° from the body. A tetrapolar electrode arrangement (dorsal aspects of the hands and feet) was used (2, 4, 5). Impedance measurements (resistance and reactance) were made in triplicate and averaged. Resistance, reactance, height, weight, and sex were used to compute fat-free mass, body cell mass, total body water, extracellular water, and intracellular water with proprietary software (RJL Industries). A parallel model is used by this manufacturer for calculating body-composition variables and the equations were developed from data obtained in healthy adults.
As an alternative marker of body composition, 24-h urinary creatinine excretion was obtained in all critically injured subjects and 16 healthy subjects. Meat-free diets were consumed for 4 d before the urine collection period. Urine was collected by bladder catheterization in the critically injured group.
Historical data
Poelhman and Toth (11) published results of a large study (n = 719) of subjects who had fat-free mass determined by hydrostatic weighing and resting metabolic rate by open-circuit indirect calorimetry. Subjects were measured in a fasted state after an overnight stay at a research center. Regression equations were published for these data. These data were chosen for comparison with our data because of the large sample size and because the authors' statistical analysis of the relation between fat-free mass and resting metabolic rate allowed direct comparison with our results by using bioelectrical impedance.
Statistics
Correlation and linear regression analyses were performed by using MINITAB release 9.2 (Minitab Corp, State College, PA). Regression lines (slopes and intercepts) were compared with a t test. Student's t test was used to compare group means. Data were reported as means ± SEMs. Statistical significance was defined as P < 0.05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, demographic characteristics of the critically ill, healthy, and control groups were not significantly different. Although the critically injured women were somewhat older and heavier than the healthy and control women, only body weight was significantly higher. Both resting metabolic rate and fat-free mass were significantly greater in the critically injured group (men and women). Body mass index (in kg/m2) did not exceed 40 in any critically injured (26 ± 1) or healthy (25 ± 1) subject.
|
Resistance, reactance, body water distribution, and body cell mass measured by using bioelectrical impedance in the critically injured and healthy groups are shown in Table 2. Resistance and reactance values were significantly lower in the critically injured group. The critically injured group had significantly more extracellular water than the healthy group but there were no significant differences in intracellular water or body cell mass between the 2 groups.
|
In the healthy group, correlations with resting metabolic rate and oxygen consumption were similar for body cell mass (r2 = 0.84, P < 0.0001) and fat-free mass (r2 = 0.82, P < 0.0001). Significant correlations were also noted between oxygen consumption and body weight (r2 = 0.72, P < 0.0001) and urinary creatinine (r2 = 0.72, P < 0.0001).
Regression analyses
In the first regression analysis (Figure 1), the linear relation (slope) between fat-free mass and resting metabolic rate (11) was compared among the healthy, critically injured, and control groups. The slopes of the regression lines for these 3 groups were not significantly different. The critically injured group had a significantly higher resting metabolic rate than the healthy group (P = 0.02) (Table 1
). Consequently, the y intercept for the critically injured group was significantly higher than that of the healthy (P = 0.014) and control groups. Complete regression equations are as follows:
|
| (1) |
|
| (2) |
|
| (3) |
|
|
| (4) |
|
| (5) |
The intercepts of these regressions were not significantly different (P = 0.155), but the slopes were significantly different (P = 0.015).
To circumvent the alterations in body composition caused by the expansion of extracellular water in a portion of the critically injured subjects, we examined the linear relation between body cell mass and oxygen consumption (Figure 2). The slopes of the lines relating body cell mass to oxygen consumption for the healthy and critically injured groups were not significantly different (P = 0.24) and there was no stratification by intracellular water (as was found for fat-free mass and extracellular water). Again, as a result of the hypermetabolic response to injury in the critically injured group, y intercepts were significantly different (P = 0.005). Body cell mass and oxygen consumption were not given in the study from which we derived the control data (11), nor could the values be computed from the data; therefore, data for the control group could not be used in our analysis.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In the current study, the established slope of the linear relation between fat-free mass (by hydrostatic weighing) and resting metabolic rate (11) was preserved in healthy and critically injured subjects whose body composition was measured with bioelectrical impedance (Eqs 1–3). However, a more detailed analysis showed that expanded extracellular water led to expanded fat-free mass in 37% of the critically injured subjects, which significantly altered the relation between fat-free mass and resting metabolic rate (Eqs 4 and 5). Consequently, the use of fat-free mass was problematic in this population of critically injured patients. In contrast, body cell mass is independent of extracellular water and thus does not have the same limitations as fat-free mass. The slope of the linear relation between body cell mass and oxygen consumption established in our healthy subjects was maintained in all of the critically injured subjects, without need for stratification by hydration status. The correlation between body cell mass and oxygen consumption was not explained by hidden cross-correlation between body weight and oxygen consumption. We therefore concluded that bioelectrical impedance accurately estimates body cell mass in critically injured patients. The significant correlation between urinary creatinine and oxygen consumption (r2 = 0.61) and the correlation between urinary creatinine and body cell mass (r2 = 0.69) further supports our conclusion.
Body impedance is more directly a measure of body water, and possibly capacitance (17), than a measure of fat-free mass and body cell mass. The mathematic impedance models for computation of fat-free mass and body cell mass were developed in healthy persons. Our results suggest that these mathematical models estimate body cell mass in critically ill patients as well. However, we cannot completely exclude the possibility that bioelectrical impedance is actually estimating some other tissue mass that is highly correlated with body cell mass. This type of uncertainty is common to many body-composition methods because of the indirect nature of the techniques.
The validity of bioelectrical impedance as a measure of body composition in healthy persons has been well established in studies using isotope dilution and hydrostatic weighing (2–5). Studies in malnourished and obese subjects have also shown the bioelectrical impedance method (especially body cell mass) to be a valid measure of body composition (6). Body composition in patients with AIDS and chronic renal failure has been measured successfully with bioelectrical impedance (18–20). However, only 2 studies have examined bioelectrical impedance as a method of measuring fat-free mass and body cell mass in critically ill patients (7, 8). Schroeder et al (7) compared bioelectrical impedance with in vivo neutron activation in surgical intensive care patients. A strong correlation was found between the methods for total body water (r2 = 0.94) and fat-free mass (r2 = 0.86), but 95% prediction limits for fat-free mass were wide (the authors felt that assumptions of constant tissue density of fat-free mass may not be valid in critical illness). Note that these investigators did not measure reactance and thus did not compute or test body cell mass, which may be more germane to critical illness. In a second validation study, Fearon et al (8) compared bioelectrical impedance analysis with total body water by isotope dilution and body cell mass by total body potassium scanning in intensive care general surgery patients. There was a strong correlation between the methods for total body water (r2 = 0.86) and body cell mass (r2 = 0.96). Systematic error for the bioelectrical impedance measurement of body cell mass was 6%. The error by the reference method for body cell mass (total body potassium scanning) was 2–4%; therefore, the actual error in the bioelectrical impedance method of body cell mass determination was estimated at 3%. The authors concluded that bioelectrical impedance analysis of body cell mass was reasonably accurate in critically ill surgical patients.
A single-frequency bioelectrical impedance device was used in the current study. Multiple-frequency plethysmographs are also available and may more accurately measure extracellular water and intracellular water compartments. This is because at lower frequencies, the impedance signal travels predominantly through the extracellular space, whereas high-frequency signals travel through extracellular and intracellular spaces. However, in human trials, single-frequency bioelectrical impedance performed similarly to multiple-frequency bioelectrical impedance (9, 21). Additionally, our measures of water distribution in critically ill patients with single-frequency bioelectrical impedance are consistent with results from others who used muscle biopsy specimens (22). The total water content of biopsy samples from critically ill patients was significantly greater than that from healthy subjects. The increase in total water content was due to significant expansion of extracellular water; there was no significant change in intracellular water (22). This pattern of water distribution is the same as that found in the present study, which used whole-body bioelectrical impedance rather than biopsy material.
The critically injured patients qualifying for the current study met specific entrance criteria, which were designed to achieve a critically ill but relatively homogeneous study group. The goal was to measure a group of patients whose variation in oxygen consumption was due mostly to body composition. The regression equations described in this paper were not intended to, and should not be used to, predict the metabolic rate in other patient populations.
The results of the current study indicate that body-composition analysis by bioelectrical impedance is valid in both critically ill patients and in healthy persons, which agree with the results of other studies that used different methods of validation. In critically ill populations, body cell mass is more appropriate than fat-free mass as an index of the metabolically active tissue mass, primarily because fat-free mass includes extracellular water, which often fluctuates in critically ill people, whereas body cell mass includes only intracellular water.
|
FOOTNOTES |
|---|
2 Supported by Clinical Research Council grant no. 93-16 and Dean's Feasibility Grant, Pennsylvania State University.
3 Address reprint requests to DC Frankenfield, Department of Clinical Nutrition (H124), Milton S Hershey Medical Center, PO Box 850, Hershey, PA 17033.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and healthy (
) subjects in whom fat-free mass was determined with bioelectrical impedance and in male (
) and female (
) control subjects in whom fat-free mass was determined by hydrostatic weighing (with data from reference 11). Coefficients of determination (r2) were 0.47 for the critically injured subjects, 0.84 for the healthy subjects, and 0.48 for the control subjects combined.
, n = 46) and healthy (