Dynamics of segmental extracellular volumes during changes in body position by bioimpedance analysis

Transcription

1 Dynamics of segmental extracellular volumes during changes in body position by bioimpedance analysis FANSAN ZHU, DANIEL SCHNEDITZ, 1 ERJUN WANG, AND NATHAN W. LEVIN 1 1 Renal Research Institute and Division of Nephrology and Hypertension, Department of Medicine, Beth Israel Medical Center, New York, New York 1018 Zhu, Fansan, Daniel Schneditz, Erjun Wang, and Nathan W. Levin. Dynamics of segmental extracellular volumes during changes in body position by bioimpedance analysis. J. Appl. Physiol. 85(): , Extracellular volume (ECV) of arms, trunk, and legs determined from segmental bioimpedance data in 11 healthy men ( yr) obtained at the end of a 30-min equilibration phase in the supine body position was compared with ECV determined from whole body measurements (ECV WB ). ECV was calculated from extracellular resistance (R ECV ) identified from the bioimpedance spectrum for a range of 10 frequencies. Whole body R ECV ( ) was equal to the sum of R ECV in the arms, trunk, and legs ( , , and , respectively). The sum of equilibrated ECV in arms ( liters), trunk ( liters), and legs ( liters) was smaller than ECV WB ( liters). In six subjects who changed from a standing to a supine body position, ECV decreased in arms (.59.51%, P NS) and legs ( %, P 0.05) but increased in the trunk ( 4. 3.%, P 0.05). ECV WB also decreased ( %, P 0.05). However, the sum of segmental extracellular volumes remained unchanged ( %, P NS). The sum of segmental ECVs is not sensitive to changes in body position, which otherwise interferes with the estimation of ECV in bioimpedance analysis when ECV WB is used. segmental extracellular resistance; multifrequency bioimpedance analysis; automatic digital switch; fluid shifts; whole body impedance ONE APPROACH TO DETERMINE body composition and body hydration in humans is based on the measurement of electrical characteristics of biological tissue (1, 14, 3, 35, 37). However, recent studies have shown that changes in body position lead to changes in whole body bioimpedance (8, 31, 33) and to apparent changes in extracellular volume (ECV) of up to 3.8 liters (3). This is a significant error for many clinical applications, such as the restoration of fluid balance during treatments with the artificial kidney, where the measurement of changes in ECV is of value in the estimation of adequate body hydration (1,, 5 7, 11, 15 18, 0,, 7, 30, 34, 36). It has been suggested that the apparent changes in ECV calculated from whole body bioimpedance measurements were related to erroneous model assumptions (31), which are outlined as follows: It was assumed that the geometry of the body could be approximated by a single equivalent cylinder and that the ECV was a homogeneous compartment. Usually, measurements were taken from one body segment with wellcharacterized geometry, such as the lower leg or the upper arm (segmental bioimpedance measurement), or from the whole body, for which electrodes were placed on the wrist and the ankle (so-called whole body bioimpedance measurement). With the assumption of homogeneous distribution of ECV, data obtained in one segment or in the whole body were then extrapolated for total ECV computation. Because measures in the limb segments do not adequately reflect total ECV with changes in body position, paradoxical results, such as an apparent increase in ECV with orthostasis and a decrease with supine body position, have been obtained (31). Although the distribution between intra- and extracellular fluid volumes is known to be unaffected by changes in body position (5), there are considerable changes in the regional distribution of ECV (4, 9), so that the ECV is heterogeneous with respect to local tissue resistance. A stable measurement of ECV is the prerequisite for bioimpedance monitoring in applications with variations in regional fluid distribution because of changes in body position. Local changes in ECV require considerable time and a constant body position to equilibrate between different parts of the body. A technique that can account for changes in regional fluid distribution would be of great advantage in clinical applications, such as hemodialysis, where considerable amounts of fluid are removed from the body by ultrafiltration. It was the aim of this study to observe differences in ECV estimation between whole body and segmental analysis after changes in body position. It was hypothesized that estimation of ECV via segmental analysis would be less sensitive to changes in body position, because it is not based on the assumption that the body can be treated as a uniform, equivalent cylinder with a homogeneous distribution of ECV throughout the body. Glossary R ECV R SSR R WBR ECV ECV WB ECV SSV ECV arm,(trunk,leg) Whole body bioimpedance Segmental bioimpedance Modeled resistance of the extracellular compartment Sum of segmental extracellular resistance Whole body extracellular resistance Extracellular volume of the whole body Extracellular volume estimated from whole body (wrist to ankle) extracellular resistance ECV estimated from sum of segmental ECV ECV in arm, trunk, or leg determined from segmental extracellular resistance Bioimpedance measured between the ipsilateral wrist and ankle Bioimpedance measured in one segment of the body such as arm, trunk, or leg /98 $5.00 Copyright 1998 the American Physiological Society 497

2 498 SEGMENTAL BIOIMPEDANCE AND CHANGES IN BODY POSITION MATERIALS AND METHODS Theory. The electrical resistance (R) of a body segment is given by R L A (1) where L is the length and A is the cross-sectional area of the cylinder and is the specific resistivity of the tissue. At low frequencies, the resistivity is equal to the resistivity of extracellular volume ( ECV ), which is assumed to be 47 cm (10). Expansion of the right side of Eq. 1 yields an expression that relates R to volume (V) R L V () However, A must be uniform over the distance L. This condition is violated for whole body bioimpedance measurements when electrodes that inject the current (I) and electrodes that sense (S) the voltage are mounted on the wrist and on the ankle. To obtain a uniform A, the parts of the body measured by whole body bioimpedance can be divided into three segments: the arm, the trunk, and the leg. Usually, spot electrodes are used to inject the current from the body surface. Therefore, the distribution of current in the segment depends on the ratio of L to A. This ratio must be large to obtain a homogeneous distribution of the current. The ratio is large in the arm and in the leg, but not in the trunk. Estimation of trunk volume by use of Eq. leads to much smaller values than expected. However, the nonuniform distribution can be corrected using a weighting factor (k). In a distributed model, ECV is given by ECV ECV SSV ECV head,neck,hands,feet (3) where ECV SSV is the sum of all segmental ECVs ECV SSV (ECV arm ECV leg ) ECV trunk (4) L arm ECV arm ECV R arm (5) L leg ECV leg ECV (6) R leg L trunk ECV trunk k 1 ECV R trunk (7) and where ECV head,neck,hands,feet refers to the ECV in the head, neck, hands, and feet. This volume is not measured by bioimpedance analysis but can be obtained from anthropometric relations, where the head accounts for 5.6% and the hands and feet for 1.8% of the body weight (9). In this study, k 4 was assumed for the calculation of ECV trunk from symmetrical considerations. This is similar to using a higher specific resistivity of the trunk (4). Measurements. Bioimpedance was measured for a logarithmic spectrum of 10 frequencies ranging from 5 to 500 khz (model 4000B analyzer, Xitron Technologies, San Diego, CA) with use of disposable electrodes ( cm ) supplied with the instrument. Two injecting electrodes for applying the alternating current were placed on the dorsal surfaces of the hand (I1) and the ankle (I) on the same side of the body. Sensing electrodes were placed on the wrist (S1), the shoulder (acromion, S), the upper anterior iliac spine (S3), and the ankle (malleolus, S4) (Fig. 1). Fig. 1. Measurement system and arrangement of electrodes. DS, digital switch; I1 and I, injecting electrodes; S1 S4, sensing electrodes; 4000B, bioimpedance measuring system; PC, personal computer. To make use of current technology that derives the bioimpedance signal from only two sensing electrodes, electrodes S1 S4 were connected to a digital switch that was developed for the purpose of segmental measurements (39). The switch automatically controlled and transferred the signal measured between electrodes to the bioimpedance-measuring device (model 4000B, Xitron). A personal computer was used for data acquisition and data storage. The different positions of the switch permitted the measurement of body segmental bioimpedance between S1 and S ( arm ), S and S3 ( trunk ), and S3 and S4 ( leg ) and whole body bioimpedance between S1 and S4 ( whole body ) (Fig. 1). After data acquisition had been completed in one segment, the connection of the sensing electrodes was automatically switched to the next position. Thus sequential measurements of three body segments (arm, trunk, leg) and of the whole body were obtained. The duration for a sweep of frequencies for the measurement of bioimpedance in one segment was 15 s. The duration for a complete cycle of measurements was 1 min. Data analysis. Acquisition, storage, and analysis of data were performed with the software supplied with the bioimpedance analyzer. Resistance of the extracellular compartment (R ECV ) for each segment and for the whole body was identified by fitting the spectrum of bioimpedance data to the Cole-Cole model by use of the software supplied with the bioimpedance device (8, 10). The correlation coefficient of the data fit was better than 0.99 for all measurements. Noise in the resistance data related to physiological processes such as breathing and movements and other noise related to the interface between the electrode and the skin were reduced by a digital low-pass filter. ECV in the different segments and ECV SSV were calculated from Eqs Extracellular volume (ECV WB ) was also calculated from whole body extracellular resistance (R WBR ) measured between the wrist and the ankle with use of relations that have

3 SEGMENTAL BIOIMPEDANCE AND CHANGES IN BODY POSITION 499 been derived elsewhere (10) ECV WB k ECV 1 H ⅔ W R WBR (8) where R WBR is whole body extracellular resistance, W is body weight (in kg), H is body height (in cm), and k ECV is a function of ECV (in cm), body density (D, in kg/m 3 ), and K B, which relates H to limb and trunk size k ECV 1 1,000 1 K B ECV D ⅓ (9) Subjects and procedure. Eleven male subjects (31.6 7yr, kg, cm) who had given informed consent were studied. The study protocol was approved by the Committee on Scientific Activities of Beth Israel Medical Center. Body weights were measured to the nearest 0.1 kg with an electronic scale. Body heights, limb lengths, and limb circumferences were measured by flexible tape to the nearest 0.1 cm. All limb measurements were taken on the right side of the body with the subject standing erect and the arms hanging freely. The mean of duplicate circumference measurements was used. The room temperature was 1 C. Bioimpedance was measured in the 11 subjects during an equilibration phase of 30 min, while they rested in a relaxed supine body position with arms and legs slightly abducted and with forearms pronated. In 6 of 11 subjects, measurements were also obtained during a 30-min standing phase, which preceded the subsequent supine phase. Each phase lasted at least 30 min. Arms were kept in a dependent position during the standing phase. Statistical analysis. Values are means SD. Paired t-tests were used for comparison of extracellular resistance and ECV during the different phases. P 0.05 was assumed to reject the null hypothesis. Equilibrated extracellular resistance data were obtained by fitting serial data measured during the duration of the equilibration phase to exponential functions and extrapolating the value of the dependent variable to the end of the equilibration phase. The precision of the resistance measurements and of the volume calculations was obtained from the SD and from the coefficient of variation (CV) of fitted and raw data. RESULTS Physical and electrical characteristics of subjects. Physical characteristics of the 11 subjects are summarized in Table 1. A summary of extracellular resistance for the different segments and for the whole body obtained from the bioimpedance spectrum is given in Table. For Table, extracellular resistance was obtained at the end of each observation phase with subjects standing or lying down. In all subjects the resistance of the arm ( ) was larger than the resistance of the leg ( ). Among all segments, the trunk had the lowest resistance ( ). The CV of serial data in different segments was comparable in the arm, leg, and whole body and higher in the trunk (Table 3). An important theoretical identity was validated: the sum of resistance measured in the arm (45.7%), trunk (9.4%), and leg segment (44.9%) was equal to 99.9% of the whole body resistance measured between the wrist and the ankle. This strong correlation may be attributed to the measuring technique in which leads are switched automatically and Table 1. Physical characteristics of the subjects Mean SD Range Age, yr Height, cm Body wt, kg Length, cm Arm Trunk Leg Circumference, cm Arm Trunk Leg Cross-sectional area, cm Arm Trunk ,003. Leg n 11. where the four different segments are measured within a short period of time. ECV estimation. ECV of different segments (Eqs. 5 7), ECV SSV (Eq. 4), and ECV WB (Eq. 8) are given in Table. Data were obtained after 30 min in the supine body position. The precision in the measurement of ECV was highest for the sum of segmental volumes, followed by a lower and comparable precision for arm, leg, and whole body volumes and a reduced precision in the estimation of trunk volumes (CV %; Table 3). ECV SSV was 3 1% smaller than ECV WB. This corresponds to a mean difference of 6.7 liters. Part of this difference is due to the exclusion of head, neck, hands, and feet from bioimpedance measurements. However, whole body bioimpedance measurements make use of anthropometric relationships to calibrate ECV WB to ECV, as determined by gold standard techniques such as indicator dilution (Eq. 8). Figure shows the relationships between ECV SSV determined from the sum of segmental volumes, ECV WB determined from whole body bioimpedance analysis, and body weight. An excellent linear relation and a high correlation coefficient were obtained for both relationships. Changes in body position. The change from a standing to a supine body position led to a marked increase in R WBR and in sum of segmental resistance (R SSR ; Fig. 3A). Linear regression showed that R WBR and R SSR were identical throughout the observation phase (Fig. 3B). Table. Extracellular resistance and ECV from segmental and whole body bioimpedance Resistance, ECV, liters Mean SD %Whole body Mean SD %Whole body Arm Trunk Leg Sum of segments Whole body Data were obtained after 30 min of supine body position; n 11. ECV, extracellular fluid volume.

4 500 SEGMENTAL BIOIMPEDANCE AND CHANGES IN BODY POSITION Table 3. Precision of segmental and whole body extracellular resistance and ECV measurement However, when analyzed for different segments, leg and arm resistance increased whereas trunk resistance decreased during the supine phase (Fig. 3C). A representative time course of the changes in ECV WB and ECV SSV is shown in Fig. 4. Whereas ECV SSV remained virtually constant throughout the entire observation phase, ECV WB increased during the standing phase and sharply decreased when the subject changed from a standing to a supine body position. In this study the amplitude of the ECV change was 1.5 liters, and no steady state was reached within 30 min of the supine observation phase. The mean change of ECV in all subjects was ECV SSV liter, with ECV trunk liter, ECV arm liter, and ECV leg liter (Table 4). Analysis of the time course in the different segments is shown in Fig. 5. Relative changes in the three segments were calculated as (ECV t /ECV 0 1) 100%, where indexes 0 and t refer to initial and subsequent measurements, respectively. The relative change was especially large in the leg. In both phases the trunk change was opposite to the change in the leg. A summary of changes in ECV with changes in body position is given in Table 5. DISCUSSION Resistance ECV SD CV, % SD CV, % Arm Trunk Leg Sum of segments Whole body Values are means SD; n 11. CV, coefficient of variation. In this study, ECV WB was compared with ECV SSV. Bioimpedance was continuously measured over a 1-h Fig.. Extracellular fluid volume (ECV) and body weight. ECV was calculated from sum of segmental volumes (solid line, ECV SSV, Eq. 3) and from whole body bioimpedance analysis (dashed line, ECV WB, Eq. 8). Linear regressions between ECV and body weight (W) are as follows: ECV SSV 0.0W 0.68 (r 0.86) and ECV WB 0.18W 7.65 (r 0.79). period, during which subjects changed from a standing to a supine body position. The change in body position caused significant changes in whole body and segmental bioimpedance. ECV WB significantly changed by as much as 1.5 liters in a single subject ( liters, P 0.05), whereas ECV SSV remained unchanged ( liter, P NS) during the change in body position. Even though ECV must be assumed to remain unchanged during a change in body position (5), ECV WB is apparently very sensitive to changes in body position: ECV WB increases with orthostasis and decreases with supine body position (31). These changes can be explained by changes in regional fluid distribution in different parts of the body, which lead to changes in whole body bioimpedance and, hence, to spurious estimations of ECV. During orthostasis, extracellular fluid is shifted from the upper to the lower parts of the body (3, 4). First, there is a rapid translocation of blood volume to the venous system of the lower body. Second, the increase in intravascular pressure in the lower extremities causes a change in the Starling equilibrium. The pressure gradient for fluid filtration increases because of the rise in capillary pressure, and fluid is shifted from the intravascular to the extravascular space in the lower body. Thus extracellular fluid increases in the leg but decreases in the trunk. Extracellular resistance. In this study the extremities accounted for 90.6% of whole body resistance, and 44.9% of this value was due to the high resistance of the legs (Table ). These ratios were almost identical to data reported for a group of 51 normal and overweight women (4) and to results from a detailed study of segmental bioimpedance in 00 healthy, white adult subjects (6). The contribution of the trunk to whole body extracellular resistance was only 9.4%. The change from a standing to a supine body position led to an increase of both whole body and sum of segmental extracellular resistance (Fig. 3, A and B). R ECV increased in the arm and in the leg but decreased in the trunk (Fig. 3C). The change in whole body resistance can be clarified in a thought experiment with use of two cylindrical segments (indexes a and b) where volume is shifted from segment a to segment b (see APPENDIX). Assume that the volume change in segment a ( V a )is V a V a,t V a,0 (10) where indexes 0 and t refer to conditions before and after the fluid shift, respectively. The sum of volumes in both segments remains constant. If L remains constant in both segments, then 1 1 R S R S,t R S,0 V a 1 A b,t A b,0 A a,t A a,0 (11) where R S is the sum of resistance in both segments. If A is larger in segment a (A a ) than in segment b (A b ) before and after the fluid shift, the expression in parentheses in Eq. 11 is positive. Thus the change in

5 SEGMENTAL BIOIMPEDANCE AND CHANGES IN BODY POSITION 501 Fig. 3. Resistance and changes in body position. Continuous recording of segmental and whole body resistance is shown for subject 7. At 30 min, subject changed from a standing to a supine body position. A: sum of segmental (R SSR ) and whole body (R WBR ) resistance of extracellular compartment (R ECV ). B: linear relationship between R WBR and R SSR : R SSR 1.07R WBR 3 (r 0.97). C: segmental R ECV in arm, leg, and trunk. resistance ( R S ) is negative for a volume shift from segment a to segment b; otherwise it is positive (see APPENDIX). Extracellular volume. In contrast to the small contribution of the trunk to whole body R ECV, 48.% of the extracellular fluid computed from Eq. 8 was located in the trunk (Table ). In other words, 48.% of the extracellular fluid located in the trunk only contributed to 9.4% of whole body resistance. When liter of fluid was shifted from the leg to the trunk when body position was changed from standing to supine, trunk resistance changed by. 1.5, but leg resistance changed by and arm resistance changed by (Table 4). As a consequence, whole body resistance decreased and ECV WB computed from Eq. 8 increased during orthostasis (Fig. 4). Because ECV change in segments contributes to whole body resistance with different weights, segments must be measured separately and segmental volumes must be calculated for each segment with use of Eqs ECV SSV is the sum of segmental ECVs (Eq. 4). During changes in body position, changes in ECV SSV are much smaller than changes in ECV WB. This is important from a practical and a theoretical point of view. When ECV is calculated from the sum of segmental volumes, neither rigorous control of body position nor an equilibration phase may be required for the assessment of fluid status. In applications such as hemodialysis, during which patients change from a standing to a sitting and often to a head-down tilt position, regional fluid shifts are Table 4. Extracellular resistance and ECV with changes in body position Resistance, ECV, liters Standing Supine Standing Supine Fig. 4. ECV and changes in body position. Registration of changes in ECV SSV (see Eq. 3) and ECV WB (see Eq. 8) is shown for subject 7.At30 min, subject changed from a standing to a supine body position. Arm Trunk * * Leg * * Sum * Whole body * * Values are means SD; n 6. *P Sum of segments.

6 50 SEGMENTAL BIOIMPEDANCE AND CHANGES IN BODY POSITION Fig. 5. Relative changes of segmental ECV (ECV rel ) and whole body ECV with changes in body position. Experiment is the same as that shown in Fig. 4 but was separated for arm, trunk, and leg segments. likely to affect the ECV as measured by conventional whole body technique. Such a change in body position will lead to considerable redistribution of blood and ECV between peripheral and central body compartments. If measured by whole body bioimpedance technique, this redistribution will lead to an apparent change in ECV. Peripheral sequestration of blood and extracellular fluid is likely to contribute to an apparent increase in ECV WB during hypotension (38). On the other hand, sequestration of ECV in the trunk leads to an underestimation of ECV WB by whole body impedance technique (7). Others have used one segment, such as the lower leg (19, 1) or the forearm (13), to measure bioimpedance and to relate changes in segmental measurements to changes in body hydration during hemodialysis and ultrafiltration. It follows from the data presented in this study that measurements in one segment will be subject to the same inaccuracies as measurements for which the whole body technique is used (Fig. 5). Measurements in the leg can be assumed to be especially susceptible to changes in body position. Differences of and liters were found between ECV WB and ECV SSV during the standing and supine phases, respectively (Table 4). The reasons for the difference in estimated volumes may be as follows: 1) ECV WB has been found to overestimate ECV measured by indicator-dilution technique by.7.0 Table 5. Changes in extracellular resistance and ECV with changes in body position x Resistance, x/x standing 100 x ECV, liters x/x standing 100 Arm Trunk Leg Sum* Whole body Values are means SD; n 6. x (x supine x standing ), absolute change; x/x standing, relative change. *Sum of segments. liters (10). ) Neither method includes head, neck, hands, and feet by direct measurement. However, the contribution of these segments is included in the ECV WB value with anthropometric relations built into the derived results. The contribution of these segments to the ECV is 1 liter for a body weight of 70 kg. 3) ECV WB may not have reached a steady state within 30 min in the supine body position (31). If the ECV WB is corrected for these effects, the difference between ECV SSV and ECV determined by indicator-dilution technique reduces to less than liters. Because the sum of segmental resistance was identical to whole body resistance, the difference between ECV WB and ECV SSV must have been related to differences in the calculation of ECV from R ECV in segmental (Eqs. 5 7) and whole body calculations (Eq. 9). The segmental technique also needs to be calibrated, which requires measurements of ECV in different body segments and introduction of correction factors in Eqs Unfortunately, standard approaches such as the inulin-distribution technique cannot be used for the measurement of segmental and regional extracellular volumes. The measurement of regional volumes with radiolabeled tracers and body scanning techniques surpasses the scope of this contribution but remains to be considered should segmental measurements gain wider interest. The advantage of the new technique is an improved spatial resolution of ECV, i.e., separation of ECV in the limbs from ECV in the trunk, and the fact that ECV SSV is relatively independent of changes in body position. However, the new approach requires two more electrodes, an automatic switch (39), and more anthropometric measures such as segmental lengths and circumferences. The additional sensing electrodes on the trunk (S and S3 in Fig. 1) could be replaced by electrodes on the contralateral wrist and ankle (6). Our results are obtained with the assumption that the resistivity of the extracellular fluid is constant (47 cm) and that the trunk can be approximated by four normal cylinders with uniform current distribution. ECVs are then calculated from simple equations (Eqs. 4 7). In conclusion and in contrast to measurements using whole body bioimpedance technique, ECV SSV is almost independent of changes in body position. This is a considerable improvement in the application of bioimpedance technique. It also is a prerequisite for clinical applications such as continuous fluid monitoring in hemodialysis patients, in whom orthostatic fluid shifts have obscured changes in regional fluid distribution. Such changes in regional fluid distribution may be caused by cardiovascular compensation and by osmotic transcellular gradients. The analysis of these processes should be possible in future studies with the new segmental technique. APPENDIX Assume two cylindrical segments a and b with crosssectional areas A a and A b and with constant lengths L a and L b. Also assume that A a A b before and after the volume shift and that L a L b. The resistance in each segment before the

7 SEGMENTAL BIOIMPEDANCE AND CHANGES IN BODY POSITION 503 shift (index 0) is and R a,0 L a A a,0 (A1) R b,0 L b (A) A b,0 In a serial arrangement of segments, the resistance of both segments (R S ) is given by the sum of the individual resistances. Therefore R S,0 L a L b A a,0 A b,0 before the shift (index 0), and R S,t L a L b A a,t A b,t after the shift (index t). A change in volume in segment a ( V a ) is defined as V a V a,t V a,0 The volume changes in segments a and b are given by V a V b 0 The change in resistance ( R S ) after the fluid shift is 1 1 R S R S,t R S,0 V a 1 A b,t A b,0 A a,t A a,0 Because of the initial assumption (A3) (A4) (A5) (A6) (A7) (A8) A b,t A b,0 A a,t A a,0 Therefore, R S depends on the direction of the fluid shift. If V a is negative (from a to b), comparable to a fluid shift in an upright body position, R S is also negative. If V a is positive (from b to a), comparable to a shift from the leg to the trunk in the supine body position, R S is also positive (Fig. 3A). Because R S changes with fluid shifts between segments, an apparent change in volume is computed if segments are lumped into a single compartment. The authors thank Raphael Recanati for support of the Renal Research Fellowship for the Division of Nephrology and Hypertension of Beth Israel Medical Center and Xitron Technologies (San Diego, CA) for supplying the bioimpedance analyzer. Address for reprint requests: D. Schneditz, Renal Research Institute, Yorkville Dialysis, rd Ave., New York, NY Received 8 September 1997; accepted in final form 7 March REFERENCES 1. Abrahamsen, B., T. B. Hansen, I. M. Hogsberg, F. B. Pedersen, and H. Becknielsen. Impact of hemodialysis on dual X-ray absorptiometry, bioelectrical impedance measurements, and anthropometry. Am. J. Clin. Nutr. 63: 80 86, Biasioli, S., L. Petrosino, L. Cavallini, A. Cesaro, S. Fazion, A. Zambello, and A. Mazzali. Cardiovascular stability during the haemodialysis session: relationship between modelling and impedance parameters. Nephrol. Dial. Transplant. 5: , Bjerkhoel, P., and J. Lundvall. Rapid and large plasma volume decrease upon short-term quiet standing. Acta Physiol. Scand. 150: , Bracco, D., D. Thiebaud, R. L. Chiolero, M. Landry, P. Burckhardt, and Y. Schutz. Segmental body composition assessed by bioelectrical impedance analysis and DEXA in humans. J. Appl. Physiol. 81: , Bradbury, M. G., S. W. Smye, and J. T. Brocklebank. Assessment of the sensitivity of bioimpedance to volume changes in body water. Pediatr. Nephrol. 9: , Cha, K., G. M. Chertow, J. Gonzalez, J. M. Lazarus, and D. W. Wilmore. Multifrequency bioelectrical impedance estimates the distribution of body water. J. Appl. Physiol. 79: , Cha, K., A. G. Hill, J. D. Rounds, and D. W. Wilmore. Multifrequency bioelectrical impedance fails to quantify sequestration of abdominal fluid. J. Appl. Physiol. 78: , Cole, K. S., C. L. Li, and A. F. Bak. Electrical analogues for tissues. Exp. Neurol. 4: , Croney, J. Anthropometry for Designers. New York: Van Nostrand Reinhold, 1981, p De Lorenzo, A., A. Andreoli, J. R. Matthie, and P. O. Withers. Predicting body cell mass with bioimpedance by using theoretical methods: a technological review. J. Appl. Physiol. 8: , De Lorenzo, A., P. Deurenberg, A. Andreoli, G. F. Sasso, M. Palestini, and R. Docimo. Multifrequency impedance in the assessment of body water losses during dialysis. Renal Physiol. 17: 36 33, Deurenberg, P., A. Tagliabue, and F. J. Schouten. Multifrequency impedance for the prediction of extracellular water and total body water. Br. J. Nutr. 73: , De Vries, P. M. J. M., J. H. Meijer, K. Vlaanderen, V. Visser, P. L. Oe, A. J. M. Donker, and H. Schneider. Measurement of transcellular fluid shift during haemodialysis.. In vitro and clinical evaluation. Med. Biol. Eng. Comput. 7: , Foster, K. R., and H. C. Lukaski. Whole-body impedance. What does it measure? Am. J. Clin. Nutr. 64: S388 S396, Ho, L. T., R. F. Kushner, D. A. Schoeller, R. Gudivaka, and D. M. Spiegel. Bioimpedance analysis of total body water in hemodialysis patients. Kidney Int. 46: , Jabara, A. E., and R. L. Mehta. Determination of fluid shifts during chronic hemodialysis using bioimpedance spectroscopy and an in-line hematocrit monitor. ASAIO J. 41: M68 M687, Jaffrin, M. Y., M. Maasrani, B. Boudailliez, and A. Le Gourrier. Extracellular and intracellular fluid volume monitoring during dialysis by multifrequency impedancemetry. ASAIO J. 4: M533 M538, Johnson, D. W., B. J. Thomas, S. J. Fleming, J. Westhuyzen, D. Moran, and L. C. Ward. Monitoring of extracellular and total body water during haemodialysis using multifrequency bio-electrical impedance analysis. Kidney Blood Press. Res. 19: 94 99, Kanai, H., M. Haeno, and K. Sakamoto. Electrical measurement of fluid distribution in legs and arms. Med. Prog. Technol. 1: , Katzarski, K. S., B. Charra, G. Laurent, F. Lopot, J. C. Divino-Filho, J. Nisell, and J. Bergström. Multifrequency bioimpedance in assessment of dry weight in haemodialysis. Nephrol. Dial. Transplant. 11: 0 3, Kouw, P. M., C. G. Olthof, P. M. ter Wee, L. P. Oe, A. J. M. Donker, H. Schneider, and P. M. J. M. de Vries. Assessment of post-dialysis dry weight: an application of the conductivity measurement method. Kidney Int. 41: , Kushner, R. F., P. M. J. M. de Vries, and R. Gudivaka. Use of bioelectrical impedance analysis measurements in the clinical management of patients undergoing dialysis. Am. J. Clin. Nutr. 64: 503S 509S, Kushner, R. F., A. Kunigk, M. Alspaugh, P. T. Andronis, C. A. Leitch, and D. A. Schoeller. Validation of bioelectricalimpedance analysis as a measurement of change in body composition in obesity. Am. J. Clin. Nutr. 5: 19 3, Lundvall, J., P. Bjerkhoel, S. Quittenbaum, and P. Lindgren. Rapid plasma volume decline upon quiet standing reflects large

8 504 SEGMENTAL BIOIMPEDANCE AND CHANGES IN BODY POSITION filtration capacity in dependent limbs. Acta Physiol. Scand. 158: , Maw, G. J., I. L. Mackenzie, and N. A. Taylor. Redistribution of body fluids during postural manipulations. Acta Physiol. Scand. 155: , Organ, L. W., G. B. Bradham, D. T. Gore, and S. L. Lozier. Segmental bioelectrical impedance analysis: theory and application of a new technique. J. Appl. Physiol. 77: 98 11, Piccoli, A., B. Rossi, L. Pillon, and G. Bucciante. A new method for monitoring body fluid variation by bioimpedance analysis: the RXc graph. Kidney Int. 46: , Roos, A. N., R. G. Westendorp, M. Frölich, and A. E. Meinders. Tetrapolar body impedance is influenced by body posture and plasma sodium concentration. Eur. J. Clin. Nutr. 46: 53 60, Rowell, L. B. Reflex control during orthostasis. In: Human Cardiovascular Control. New York: Oxford University Press, 1993, p Scanferla, F., S. Landini, A. Fracasso, P. Morachiello, F. Righetto, and G. Bazzato. On-line bioelectric impedance during haemodialysis: monitoring of body fluids and cell membrane status. Nephrol. Dial. Transplant. 5: , Scharfetter, H., M. Monif, Z. Laszlo, T. Lambauer, H. Hutten, and H. Hinghofer-Szalkay. Effect of postural changes on the reliability of volume estimations from bioimpedance spectroscopy data. Kidney Int. 51: , Scharfetter, H., G. H. Wirnsberger, Z. László, H. Hinghofer- Szalkay, H. Hutten, and H. Holzer. Impedance-spectroscopy after postural changes (Abstract). Nephrol. Dial. Transplant. 11: A13, Shirreffs, S. M., and R. J. Maughan. The effect of posture change on blood volume, serum potassium and whole body electrical impedance. Eur. J. Appl. Physiol. 69: , Smye, S. W., H. M. Norwood, T. Buur, M. Bradbury, and J. T. Brocklebank. Comparison of extra-cellular fluid volume measurement in children by 99Tcm-DPTA clearance and multifrequency impedance techniques. Physiol. Meas. 15: 51 60, Sutcliffe, J. F. A review of in vivo experimental methods to determine the composition of the human body. Phys. Med. Biol. 41: , Thompson, C. M., C. H. Kong, C. A. Lewis, P. D. Hill, and F. D. Thompson. Can bio-electrical impedance be used to measure total body water in dialysis patients? Physiol. Meas. 14: , Van Marken Lichtenbelt, W. D., K. R. Westerterp, L. Wouters, and S. C. M. Luijendijk. Validation of bioelectricalimpedance measurements as a method to estimate body-water compartments. Am. J. Clin. Nutr. 60: , Zaluska, W., D. Schneditz, A. M. Kaufman, A. T. Morris, and N. W. Levin. Intercompartmental fluid shifts monitored by multifrequency bio-impedance spectroscopy (Abstract). J. Am. Soc. Nephrol. 7: 144, Zhu, F., D. Schneditz, E. Wang, and N. W. Levin. Continuous measurement of segmental and whole body bio-impedance. Proc. IEEE 19: , 1997.