Medicine & Science in Sports & Exercise:
SPECIAL COMMUNICATIONS: Methods
Use of bioelectrical impedance analysis to estimate body fluid compartments after acute variations of the body hydration level
KOULMANN, NATHALIE; JIMENEZ, CHANTAL; REGAL, DAMIEN; BOLLIET, PHILIPPE; LAUNAY, JEAN-CLAUDE; SAVOUREY, GUSTAVE; MELIN, BRUNO
Unité de Bioénergétique et Environnement, Unité de Thermophysiologie, Centre de Recherches du Service de Santé des Armées “Emile Pardé”, 38702 La Tronche Cedex, FRANCE
Submitted for publication February 1999.
Accepted for publication June 1999.
Address for correspondence: Nathalie Koulmann, Unité de Bioénergétique et Environnement, Centre de Recherches du Service de Santé des Armées “Emile Pardé,” BP 87-24 avenue des Maquis du Grésivaudan, 38702 La Tronche Cedex, France. E-mail: email@example.com.
KOULMANN, N., C. JIMENEZ, D. REGAL, P. BOLLIET, J.-C. LAUNAY, G. SAVOUREY, and B. MELIN. Use of bioelectrical impedance analysis to estimate body fluid compartments after acute variations of body hydration level. Med. Sci. Sports Exerc., Vol. 32, No. 4, pp. 857–864, 2000. Physiological measurements including body mass, plasma osmolality, natremia, plasma volume measured by Evans Blue dilution, and total body water (TBW) and extracellular water (ECW) volumes estimated by bioelectrical impedance analysis (BIA) were recorded in eight healthy young Caucasian subjects before and after acute variations of their body hydration state on four separate occasions: 1) euhydration or control trial (C); 2) heat-induced dehydration of 2.8% body mass (D); 3) exercise-induced dehydration of 2.8% body mass (E); and 4) glycerol-hyperhydration (H). Heart rate, rectal and mean skin temperatures were also recorded throughout the experiment. The main result of the study is that BIA only half predicted the body water loss after exercise, although conditions were standardized (electrode placement, side of the body, limb position, posture, and ambient temperature). Differences in body temperatures cannot explain such an unexpected result, nor did the study of plasma osmolality and sodium concentration. If BIA appears to adequately predict changes in TBW after heat-induced dehydration and glycerol hyperhydration, further studies including measures of TBW and ECW by dilution tracer methods would be necessary to establish the validity of using the BIA method to measure such changes and to interpret ECW variations.
When exposed to heat stress, whatever it is, metabolic and/or thermal heat, men may lose a considerable amount of body water through sweating, induced by the regulation of body temperature. As soon as 1947, Adolph et al. (1) reported that body fluid compartments contribute in different extent to the water loss. However, only a few data are available concerning the distribution of the water loss among body fluid compartments; the alterations in different compartmental fluid status would depend on both the kind of heat stress applied (12) and the level of hypohydration (25). On the other hand, hyperhydration induced by glycerol ingestion before exposure to dehydrating events appears to be a potential countermeasure to the deleterious effects of hypohydration (15,20). Only one study (23) reported the changes in the interstitial fluid volume due to the increase of total body water. During those studies, body fluid compartments were determined using tracer methods, including radioactive tracers. The dilution techniques were indeed the method used until recently to assess total body water and fluid compartment volumes. However, repeated measurements using the dilution methods were difficult, because they required either a waiting period while the tracers cleared from the body, or the use of higher doses. Moreover, the repeated use of radioactive tracers appeared now to be questionable, especially in young healthy subjects for experimental design.
Recently, there has been considerable interest in bioelectrical impedance as a safe, convenient, and portable method of estimating body composition, associated with high reliability in normal individuals (2) and the ability to perform frequent, rapid, and noninvasive measurements. The aim of the present study was then, using bioelectrical impedance analysis (BIA), to estimate changes in body fluid compartments after different acute changes in the body hydration level—heat-induced dehydration, exercise-induced dehydration, and glycerol hyperhydration. Because a great variety of factors other than body composition, as electrode placement, side of the body, limb position (2), body posture (21,24), ambient and skin temperatures (3,8), plasma osmolality and sodium concentration (8,21) could, at least potentially, alter BIA measurements, we have particularly taken care to standardize our experimental design and to respect the manufacturer’s recommendations.
Eight male Caucasian volunteers gave their informed consent to the experimental conditions after the details of the protocol had been explained to them. This study was approved by an Ethics Committee (Comité Consultatif pour la Protection des Personnes dans la Recherche Biomédicale, Grenoble, France). The selection of the subjects was based on a normal clinical investigation that comprised a detailed medical history, a physical examination, and general blood screening. All the participants were regularly trained for endurance and unaccustomed to heat. Their mean anthropomorphic characteristics were: age 27 ± 1 yr, body mass 73 ± 2 kg, height 1.77 ± 0.01 m. Their physical fitness was estimated from the maximal oxygen uptake (V̇O2max) measured during a progressive treadmill test using a breath-by-breath automated gaz exchange system (MedGraphics CPX/D, Medical Graphics Corporation, St. Paul, MN); the average result was 56.6 ± 1.5 mL·min−1·kg−1. The experiment took place in winter and spring.
Deuterium dilution method for determining total body water (TBW).
TBW was measured once with deuterium dilution at the beginning of the study (in January), using the method described by van Marken Lichtenbelt et al. (16). Subjects received an orally administered dose of 10-mL D2O (99.8% of D, group CEA, CE Saclay, Gif-sur-Yvette, France) diluted with mineral water to 0.075 L for intake. The dose bottle was washed out, and the rinse water (0.075 L) was also ingested by the subject to ensure that all deuterium was consumed. D2O enrichment in the body fluid was measured in urine. The dose was given at night before bedtime (2230 h) immediately after the background sample of urine. After an equilibrium interval of 10 h (16), the final sample of urine was taken from the second voiding next morning at 0830 h (first voiding at 0730 h) to avoid dilution of labeled urine in the bladder. No food or drink was permitted until the final sample.
Isotope abundance in urine was measured using nuclear magnetic resonance spectroscopy in Eurofins Laboratories (Nantes, France). The accuracy of this method was ± 0.2 ppm for a half-hour-analysis, and the repeatibility was 2–3% (9). TBW was calculated as the deuterium dilution space divided by 1.04, correcting for exchange of the deuterium label with nonaqueous H of body solids (22).
Between months of February and May, each subject performed four trials, each separated by at least 15 d. Experiments were conducted in a randomized crossover design. For 3 d before each trial, the subjects were asked to refrain from strenuous exercise and to drink at least 2 L of water per day in order to be normohydrated.
The subjects arrived at the laboratory at 0830 h after a standard breakfast. They emptied their bladder, wore shorts, and were weighed (Sebag-Pesage 286 scale, Sebag-Pesage, Villeurbanne, France; precision ± 20 g). Then, they were asked to lie down and the different probes (catheters, heart rate recorder, thermistances, and electrodes) were attached to them.
For each trial, the protocol included three phases (see Fig. 1). During phase 1, the subjects remained in semirecumbent posture for 90 min in a thermoneutral environment (dry bulb temperature (Tdb) = 25–26°C, relative humidity (RH) = 40–60%); the first 30-min period was necessary to stabilize hemodynamic conditions. A first measure of plasma volume (PV) was made during the next 60-min period, with the method described by Jimenez et al. (11). TBW and extracellular water volume (ECW) were determined by BIA during the last 15 min of phase 1. During that period, subjects were requested to make minimal movements while maintaining a relaxed and comfortable position, with the trunk slightly tilted: “semirecumbent” position means that the angle between the trunk and lower limbs was 155° instead of 180° for a fully recumbent position. The same degree of recumbency with each measurement was achieved because subjects were lying down on a specific bed for medical examination.
After phase 1, a 30-min period was then spared to stabilize hemodynamic conditions in the posture held during phase 2 before the beginning of it. Phase 2 consisted in the variation of the body hydration level and was different according to the four trials. 1) The control experiment (C) was an euhydrated trial during which the subjects remained in sitting posture for 2 h. Then they were weighed and the body mass loss noted was compensated by an equivalent water intake in order to maintain an euhydrated state. 2) In the heat-induced dehydration trial (D), a passive heating session was conducted in a climatic chamber until the subjects had lost 2.8% of their body mass. The method used to dehydrate the subjects was derived from the controlled hyperthermia technique described by Henane and Valatx (10). Briefly, the subjects were asked to lie down on a balance (TESTUT 9009, Béthune, France; sensitivity ± 3 g) to measure the sweat loss. A copper-constantan thermocouple was inserted and insulated in the auditory canal. Climatic parameters were then adjusted (successively Tdb = 45°C, RH = 70% and Tdb = 50°C, RH = 30%) to clamp the auditory canal temperature at 38°C. This passive heating session was immediately stopped when the target loss of body mass was achieved which generally took 2 h. 3) In the exercise-induced dehydration trial (E), the subjects exercised on a treadmill at 60% V̇O2max in controlled thermal conditions (Tdb = 25–26°C, RH = 35–50%, wind speed modulated to maintain core temperature under 39°C). The exercise was briefly stopped every 30 min to verify the subject’s body mass loss, which had to reach 2.8% of their total body mass at the end of the 2 h of that phase. 4) In the hyperhydration experiment (H), the subjects remained in sitting posture for 2 h. Hyperhydration state was induced by ingestion of glycerol and water, using the method described by Lyons et al. (15). At the beginning of phase 2, the subjects had to drink 1.1 g·kg−1 body mass of glycerol which corresponded in mean to 128 ± 4 mL of a pharmaceutic flavoured solution (Glycérotone®, Laboratoires H. Faure, Annonay, France) half-diluted with mineral water. Thereafter, the subjects had to drink water containing 1.2 g·L−1 NaCl in three equal intakes (mean 465 ± 13 mL) every 30 min of phase 2 so that the total intake of water corresponded to 21.4 mL·kg−1 body mass.
Phase 3 was the same as phase 1: the subjects remained in semirecumbent posture for 90 min in thermoneutral environment; a 30-min period was also spared to stabilize hemodynamic conditions and then a second measure of PV was made during the next 60-min period as described by Jimenez et al. (11). TBW and ECW were determined by BIA during the last 15 min of phase 3.
At the end of each phase, the subjects were asked to empty their bladder and were weighed on the same scale (Sebag-Pesage 286). Throughout the experimental session, heart rate (HR) was recorded every minute with a telemetry system (Polar Sport Tester PE 4000®, Polar Electro Oy, Kempele, Finland) and rectal (Tre) and four cutaneous temperatures were recorded every minute with thermistances (Ysi series 400, Yellow Springs, OH). Mean skin temperature (¯Tsk) was calculated using the equation of Ramanathan (19). Results of HR, Tre, and ¯Tsk are presented as means calculated every 10 min.
During C, D, and H, the oxygen consumption (V̇O2) was measured during 10 min at the beginning of phase 2 using the breath-by-breath automated gas exchange system MedGraphics CPX/D. During E, the intensity of the exercise was checked during 10 min every 30 min by the measurement of V̇O2. Total body sweat loss during D and E was calculated from the difference in nude body mass before and after phase 2, adjusted for metabolic and water respiratory mass losses according to Mitchell et al. (17).
Bioelectrical impedance analysis.
Impedance was measured by using the ANALYCOR 3 bioimpedance analyser (Eugédia, Chambly, France) and by applying the right-sided unilateral tetrapolar-electrode method, according to the manufacturer’s instructions. After the skin was shaved and cleaned with alcohol, Ag/AgCl electrodes (Silver Sircuit®, Sentry Medical Products, Irvine, CA) were placed to the right hand and to the ipsilateral foot. Specifically, the two current injector electrodes were placed just below the phalangeal-metacarpal joint of the third finger in the dorsal side of the hand and just below the transverse arch on the superior side of the foot midway between the second and third metatarsal. The two voltage-sensing detector electrodes were placed on the posterior side of the right wrist, at the line bisecting the styloïd processes of the ulna and radius and on the anterior side of the ankle, over the axis of the medial and lateral malleoli. The electrodes were applied at the beginning of each trial and remained until the end of the experiment during C and H. During D and E, the electrodes were removed between both measures; the skin was cleaned with alcohol at the sites of electrodes before they were replaced, and their placement were accurately reproduced by drawing in ink their outline. Impedance values were measured at 5 and 100 kHz in triplicate and were converted to an estimate of TBW and ECW using the algorithm supplied by the manufacturer.
A polyethylene catheter (Angiocath 20GA 2in, Becton Dickinson, Sandy, UT) was inserted in an antecubital vein of each arm at the beginning of phase 1 to allow injection of dye and blood samples. Plasma volume was measured twice, before and after the variation of the body hydration level, using an Evans blue dye method extensively described by Jimenez et al. (11). Briefly, for each measure an exact quantity of tracer was injected via the catheter in the left arm and the serynge was rinsed 3 times with 5-mL isotonic saline to wash out residual dye. Four blood samples (5 mL) were drawn from the catheter of the opposite arm: the first one at time 0 just before the injection for the reference value (blank plasma sample), and three others 30, 45, and 60 min after dye injection. After centrifugation, the plasma was stored at +4°C until Evans blue analysis. Plasma samples were analyzed using a Lambda 2UV/VIS spectrophotometer (Perkin-Elmer, Überlingen, Germany) at four wavelengths against the blank plasma sample according to the method and calculations described by Farjanel et al. (6).
Additional blood (1.5 mL) was sampled at time 90 min (t 90) of phase 1, just before the beginning of phase 2 after the stabilization of hemodynamic conditions (t 120), at t 150, t 180, and t 240 during phase 2, and at t 300 and t 360 during phase 3 (see Fig. 1). Blood samples were transferred into tubes containing lithium-heparin and centrifugated at +4°C. Plasma osmolality was measured by freezing point depression (AUTOCAL 13/13 DR® Osmometer, Roebling Messtechnnik, Berlin, Germany). Plasma sodium concentration was measured by flame photometry (Corning 480 Flame Photometer, Ciba Corning Diagnostics, Cergy Pontoise, France).
Data analysis was performed with the Statistica® package (Statsoft Inc, Tulsa, Oklahoma, USA). Statistical differences were established by using a repeated-measures analysis of variance design; when an overall difference was found, individual stages were compared by using the Tukey post hoc test. Appropriate linear regression analysis was performed to compare TBW measured by deuterium dilution method and TBW estimated by BIA when subjects were normohydrated for one part, and to compare variations in total body water (ΔTBW) and changes in body mass for the other part. The null hypothesis was rejected when P < 0.05.
Heart Rate and Body Temperatures
HR was similar during phase 1 whatever the trial, and throughout the experiment during C and H (Fig. 2). During phase 2, HR was higher during E than during the other trials (P < 0.001) and also higher during D than during C and H (P < 0.001). From t 260 to the end of the experiment, HR was similar during D and E but was still higher (P < 0.001) than during C and H until t 300. From t 300 to the end of phase 3, HR did not differ significantly whatever the trial, and during the last half-hour of the experiment, HR values were similar to those measured during phase 1.
Tre had the same time course during phase 1 whatever the trial (Fig. 3). During phase 2, Tre was slightly lower during H than during C (P < 0.05 at t 240 and t 250). During dehydration sessions, Tre increased sooner during E (P < 0.05 between t 130 and t 140) than during D (P < 0.05 between t 150 and t 160) with higher values (P < 0.001 from t 140 to t 250). Values during E were also higher than during C and H until t 310 (P < 0.001), and then were similar up to the end of exercise. During D, Tre was higher (P < 0.001) than during C and H from t 160 up to the end of the experiment; values recorded during phase 3 were higher than those measured during phase 1, whereas during the other trials Tre was similar during phase 1 and the last hour of phase 3.
Mean skin temperature.
¯Tsk was similar during phase 1 whatever the trial, and throughout the experimental session for C and H (Fig. 4). During D, ¯Tsk markedly increased at the beginning (P < 0.001) and decreased at the end of phase 2 (P < 0.001), with higher values than during the others trials from t 120 to t 260 (P < 0.001). During E, ¯Tsk only slightly increased at the beginning of phase 2 (P < 0.01) with higher values than during C and H only from t 140 to t 160 (P < 0.01). During phase 3, ¯Tsk was the same whatever the trial.
Body Mass Variations
The hyperhydration protocol induced a gain of body mass which half-persisted at the end of phase 3 (Table 1). That body mass gain was different (P < 0.001) to the mean body mass loss observed during C at the end of the experiment.
During both heat-induced and exercise-induced dehydration, the body mass loss was similar and reached 2.8% of body mass at the end of phase 2 (Table 1). However, the distribution of mass losses was different between D and E according to the different level of metabolic rate (Table 2): metabolic and respiratory mass losses were higher while sweat and urinary volumes were lower during E compared with D (P < 0.05).
Body Fluid Compartments
Total body water.
TBW was first estimated by BIA during all trials at the end of phase 1 when subjects were normohydrated (TBW 1;Table 3). The mean coefficient of variation for trial-to-trial intraindividual impedance (Ω) measurements at 100 kHz was 4.0%, which induced a mean reproducibility of TBW estimations of 3.0%. The precision of those estimations was 3.9% compared with TBW previously measured by deuterium dilution. The linear regression (Fig. 5) between both direct and indirect methods of measure of TBW was highly significant (P < 0.01). TBW was estimated once again by BIA at the end of phase 3 (Table 3): TBW 2 decreased during C, D, and E compared with TBW 1 whereas it increased during H (P < 0.05). The variation of total body water (ΔTBW = TBW2 − TBW1) was different according to each trial: the largest loss observed during D was twice as important as the loss observed during E (P < 0.05) which was higher than C changes (P < 0.05). If compared to variations of body mass observed at the end of phase 3 (Δbody mass;Table 1), the linear regression between ΔTBW and Δbody mass was highly significant (Y = 1.10 x, R = 0.92, P < 0.001).
Extracellular water volume.
ECW estimated by BIA was similar whatever the trial at the end of phase 1 (ECW 1;Table 4). The mean reproducibility of those values was 2.8%. After the variation of the body hydration level, ECW (ECW 2) decreased during C, D, and E (P < 0.05), whereas it increased during H (P < 0.05). The loss observed during D was twice as important as that observed during E (P < 0.05); the gain observed during H was different than the slight loss noted during C (P < 0.05).
PV did not change during the control experiment, whereas it only slightly decreased during E without significant differences between PV 1 and PV 2 (Table 5). After heat-induced dehydration, PV markedly decreased (P < 0.05) and plasma volume changes (ΔPV) reached −11.4%. PV increased after hyperhydration (P < 0.05) with higher values than during C (P < 0.05) (Table 5).
Plasma Osmolality and Natremia
Plasma osmolality (Posm).
Posm was stable throughout C (Fig. 6), whereas it markedly increased at the beginning of phase 2 during H (P < 0.001 between t 120 and t 150). Those values were higher than during all the other trials (P < 0.001 compared with C, D, and E) even if Posm progressively decreased after Phase 2 (P < 0.05 between t 180 and t 240 and between t 300 and t 360). The time course of Posm was similar during D and E with higher values than during C from t 180 up to the end of the experiment (P < 0.001) (Fig. 6).
[Na] was stable throughout C (Fig. 7), although it progressively decreased during H with lower values (P < 0.001) than during C from t 240 up to the end of the experiment. The time course of [Na] was similar during both dehydration sessions with higher values than H values at t 180 (P < 0.001) and C and H values from t 240 up to the end of the experiment (P < 0.001) (Fig. 7).
To evaluate the distribution of water changes among body fluid compartments, we used bioelectrical impedance analysis to estimate TBW and ECW before and after the variation of body hydration level of young healthy subjects. To our knowledge, such an estimation had never been studied to differentiate the influence of heat-induced versus exercise-induced dehydration nor has the influence of glycerol-induced hyperhydration. BIA is an indirect method to measure body fluid volumes, and to test its reliability, we have first compared measures made before any variation of the hydration level. The mean reproducibility for trial-to-trial intraindividual impedance measurements at 5 kHz (coefficient of variation: 4.4%) and at 100 kHz (coefficient of variation: 4.0%) was slightly larger than that reported for week-to-week (2.2%) by Kushner and Schoeller (13) even though the slight variation in body weight over both studies was the same (1.6% on an average). However, in the present study, each subject’s trial was separated to each other by at least 2 wk and sometimes 1 month; so the overall experiment had lasted 6 wk as a minimum and 4 months as a maximum. Thus, the most important trial-to-trial variability was observed when the overall experiment had lasted the longest, whereas that variability was minimal for the subject whose experiment had taken place in only 6 wk. We have also compared BIA-predicted total body water with TBW measured by deuterium dilution: the precision (3.9%) was within the range of a prediction error of about 5% reported by Deurenberg and Schouten (4). Those preliminary verifications make us confident that BIA estimates for TBW are accurate, even if they do not establish the validity of using the BIA method to measure changes in hydration status under dehydrated or hyperhydrated conditions. Nevertheless, the main result of the present study is that BIA only half predicted the body water loss after exercise-induced dehydration, whereas it appears to adequately predict changes in TBW after heat-induced dehydration and glycerol-hyperhydration when compared with variations in body mass. Indeed, although the number of subjects might be too small for significant differences to be accurately detected, if such a difference between TBW-estimated changes and variations of body mass after heat-induced dehydration and glycerol hyperhydration does exist, it might be small compared with the large error of prediction observed after exercise-induced dehydration. The assumption that variations in body mass are only due to variations in total body water seems to be reasonable in all experimental conditions of the present study, even during exercise-induced dehydration as the mass of carbon (average 82 g) lost during the 2 h of treadmill exercise may be considered negligible in regard to the total body mass loss (average 1961 g; see Table 1). Thus, there is a great discrepancy between variations in body mass and TBW changes estimated by BIA after exercise-induced dehydration. However, during exercise water is lost but produced as well, the latter mainly as a result of glycogen breakdown and the subsequent release of the water that had been bound to the glycogen matrix. We have calculated that endogenuous water produced during E would represent 760 mL, shared between water produced by metabolic oxidations (170 mL) and bound water released from glycogen breakdown (590 mL). Considering overall glycogen stores, this bound water potentially represents about 2000 mL in young trained subjects (2.7 g water per g glycogen, 25 g glycogen per kg muscle, and 25 kg muscle in whole body for an approximation). If it might be speculated that this bound water would be “indetectable” by BIA, nevertheless, we cannot verify such an assumption. The reason why BIA only half predicted the body water loss during E may be elsewhere as well. Several factors are known to affect impedance measurements (2): electrode placement, side of the body, limb position, posture (21,24), ambiant and skin temperatures (3,8), plasma osmolality, and sodium concentration (8,21). Some of those factors had been previously accurately controlled during our experiment and cannot be involved in such a difference. Indeed, because the side on which impedance is measured must match the side measured during the development of the predictive equation (2), we always measured impedance on the right side, according to the manufacturer’s recommendations. During both dehydration trials, we changed the electrodes between measures made before and after the variation of the body hydration level; however, we had previously accurately marked the sites of electrode placement to reduce inter-observer differences (5). The subject’s limbs were positioned so that they were not in contact with each other, and they were slightly separated from the trunk, with the same position every time. In the same way, because resistance measured with tetrapolar body impedance is unstable during the first 60 min of recumbency (21,24), we took particularly care of the subject’s posture: they were lying in a thermoneutral environment for at least 75 min before each measure.
Nevertheless, some among the above-mentioned factors known to affect impedance measurements depend on the dehydration procedures. If we succeeded to induce by heat exposure and treadmill exercise the same body mass loss within the same duration, we also aimed to perform each BIA measurement at similar values of body temperatures (Tre, ¯Tsk). That was clearly achieved (see Figs. 3 and 4) during control trial, glycerol hyperhydration, and exercise-induced dehydration; so differences in body temperatures cannot explain the large error of prediction of body mass loss after exercise. If after heat-induced dehydration a slightly higher value of Tre than before was observed, it seems to have no effect on BIA prediction. The study of time courses of HR, Tre, and ¯Tsk is of great interest because it emphasizes the fact that our experimental design was perfectly controlled, with similar values during both measures of BIA—before the variation of the body hydration level (phase 1) and during the recovery (phase 3)—although dehydration sessions (phase 2) induced very different physiological responses according to each condition. Indeed, metabolic heat is produced by active muscles as soon as exercise begins, and the blood convection rapidly homogenizes the implied increase in core temperature. Conversely, during the passive heating session, the external thermal load is first applied to the skin, then transfered by conduction from skin area to core and finally distributed by blood convection, which might explain the delay observed before the increase of Tre during D compared with E. During the recovery period after dehydration (phase 3), Tre slowly decreased during D, whereas during E Tre fell at the end of exercise to return to control and preexercise values, probably because plasma volume was rather maintained during E, whereas it was dramatically decreased during D. On the contrary, during both dehydration trials plasma osmolality and sodium concentration had very similar time courses, and those factors can neither be involved in the unexpected result of BIA prediction after exercise. Finally, as suggested by Monnier et al. (18), the influence of physical exercise on measures of BIA might be explained by a nonsteady state of body fluid volumes: their redistribution to active muscles, here the lower limbs, would induce a relative increase of the state of hydration of those segments, which would partly conceal the decrease of fluid volumes occurring mainly in the trunk and the upper limbs. It would be interesting in further studies, to determine whether waiting for a longer time than 2 h as in the present study could allow BIA to accurately predict body water changes after exercise.
With regard to the distribution of the water loss among body fluids compartments, after heat-induced dehydration, BIA would predict a decrease in ECW, which would represent 37% of the TBW decrease; the loss of plasma volume would take the main part of that ECW decrease. Because to our knowledge the study of Kozlowski and Saltin (12) is the only one that describes the distribution of the water loss after a dehydration of 4% body mass induced by sauna, it is particularly interesting to verify that present results of BIA predictions are in accordance with their determinations by dilution tracers. After exercise-induced dehydration, BIA would predict a decrease in ECW of the same relative range than during D (36% of TBW estimated by BIA), in which the decrease in plasma volume would represent half of the decrease in ECW; however, taking into account the fact that BIA half predicted the loss in TBW (see above), it seems to be very difficult to interpret those results. Thus, in order to further investigate the validity of BIA to predict ECW changes in our experimental conditions, it will be necessary to compare, during the same protocol design, changes in ECW predicted from variations in low-frequency BIA with changes in ECW measured by a dilution tracer method.
The ingestion of glycerol and water induced a real state of hyperhydration as attested by the gain of body mass at the end of phase 2. That gain represented 77 ± 6% of the volume of water ingested by the subjects which is consistent with results from Riedesel et al. (20) and Lyons et al. (15). At the end of phase 3, the body mass gain was no more than 37 ± 11% of the ingested water volume, but one subject had lost in that phase an important volume of urine (1025 mL) compared with the others (mean: 359 mL). If we excluded the value of that subject, then the body mass gain at the end of the experiment was still 47 ± 6% of the ingested water volume, which agrees with results reported by Riedesel et al. (20) 3 h after a similar protocol of ingestion of glycerol and water. Such a hyperhydration occurred with an increase in ECW and hypervolemia, which were expected because glycerol is known to be evenly distributed throughout TBW (7,14).
We concluded that after exercise-induced dehydration BIA only half predicted the TBW loss whereas the main factors usually known to affect BIA predictions appear to be not involved in an unexpected result. Further studies, including measures of TBW and ECW by dilution tracer methods are necessary to establish the validity of using the BIA method to measure TBW and ECW under dehydrated or hyperhydrated conditions.
The authors thank Dr P. Monnerot, the S.A des Eaux Minérales d’Evian and its scientific committee for their helpful contribution. The technical assistance of A. Guinet, A. M. Hanniquet, and Y. Besnard is also gratefully acknowledged. The authors extend their thanks to the volunteers whose participation made this study possible.
1. Adolph, E. F., A. H. Brown, D. R. Goddard, et al. Physiology of Man in the Desert. New York: Interscience, 1947.
2. Baumgartner, R. N., W. C. Chumlea, and A. F. Roche. Bioelectric impedance for body composition. Exerc. Sports Sci. Rev. 18:193–224, 1990.
3. Caton, J. R., P. A. Mole, W. C. Adams, and D. S. Heustis. Body composition analysis by bioelectrical impedance: effect of skin temperature. Med. Sci. Sports Exerc. 20:489–491, 1988.
4. Deurenberg, P., and F. J. M. Schouten. Loss of total body water and extracellular water assessed by multifrequency impedance. Eur. J. Clin. Nutr. 46:247–255, 1992.
5. Elsen, R., M-L. Siu, O. Pineda, and N. W. Solomons. Sources of variability in bioelectrical impedance determination in adults. In: In Vivo Body Composition Studies, K. J. Ellis, S. Yasumura, and W. D. Morgan (Eds.). Upton, NY: The Institute of Physical Sciences in Medicine, Brookhaven National Laboratory, 1987, pp. 184–188.
6. Farjanel, J., C. Denis, J-C. Chatard, and A. Geyssant. An accurate method of plasma volume measurement by direct analysis of Evans blue spectra in plasma without dye extraction: origins of albumin-space variations during maximal exercise. Eur. J. Appl. Physiol. 75:75–82, 1997.
7. Freund, B. J., S. J. Montain, A. J. Young, et al. Glycerol hyperhydration: hormonal, renal, and vascular fluid responses. J. Appl. Physiol. 79:2069–2077, 1995.
8. Gudivaka, R., D. Schoeller, and R. F. Kushner. Effect of skin temperature on multifrequency bioelectrical impedance analysis. J. Appl. Physiol. 81:838–845, 1996.
9. Guillou, C., M. Trierweiler, and G. Martin. Repeatability and reproductibility of site-specific isotope ratios in quantitative 2
H NMR. Magn. Res. Chem. 26:491–496, 1988.
10. Henane, R., and J. L. Valatx. Thermoregulatory changes induced during heat acclimatization by controlled hyperthermia in man. J. Physiol. (Lond.) 230:255–271, 1973.
11. Jimenez, C., B. Melin, N. Koulmann, A. M. Allevard, J. C. Launay, and G. Savourey. Plasma volume changes during and after acute variations of body hydration level. Eur. J. Appl. Physiol. 80:1–8, 1999.
12. Kozlowski, S., and B. Saltin. Effect of sweat loss on body fluids. J. Appl. Physiol. 19:1119–1124, 1964.
13. Kushner, R. F., and D. A. Schoeller. Estimation of total body water by bioelectrical impedance analysis. Am. J. Clin. Nutr. 44:417–424, 1986.
14. Lin, E. C. C. Glycerol utilization and its regulation in mammals. Annu. Rev. Biochem. 46:765–795, 1977.
15. Lyons, T. P., M. L. Riedesel, L. E. Meuli, and T. W. Chick. Effects of glycerol-induced hyperhydration prior to exercise in the heat on sweating and core temperature. Med. Sci. Sports Exerc. 22:477–483, 1990.
16. Marken Lichtenbelt, W. D. van, K. R. Westerterp, and L. Wouters. Deuterium dilution as a method for determining total body water: effect of test protocol and sampling time. Br. J. Nutr.
17. Mitchell, J. W., E. R. Nadel, and J. A. J. Stolwijk. Respiratory weight losses during exercise. J. Appl. Physiol. 32:474–476, 1972.
18. Monnier, J. F., E. Raynaud, J. F. Brun, and A. Orsetti. Influence of meal consumption and exercise on the assessment of body composition by bioelectrical impedance. Sci. Sports, 12:256–258, 1997.
19. Ramanathan, N. L. A new weighting system for mean surface temperature of the human body. J. Appl. Physiol. 19:531–533, 1964.
20. Riedesel, M. L., D. Y. Allen, G. T. Peake, and K. Al-qattan. Hyperhydration with glycerol solutions. J. Appl. Physiol. 63:2262–2268, 1987.
21. Roos, A. N., R. G. J. 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, 1992.
22. Schoeller, D. A., E. Van Santen, D. Peterson, W. Dietz, J. Jaspan, and P. Klein. Total body measurement in humans with 18
O and 2
H labeled water. Am. J. Clin. Nutr. 33:2686–2693, 1980.
23. Seifert, J. G., M. J. Luetkemeier, W. Schrage, and M. Coles. The effects of glycerol ingestion on fluid compartment volumes, physiological function, and cycling performance. Med. Sci. Sports Exerc. 27:S19, 1995.
24. 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. 69:461–463, 1994.
25. Singh, M. V., S. B. Rawal, G. Pichan, and A. K. Tyagi. Changes in body fluid compartments during hypohydration and rehydration in heat-acclimated tropical subjects. Aviat. Space Environ. Med. 64:295–299, 1993.
DEHYDRATION; EXERCISE; HEAT EXPOSURE; GLYCEROL-HYPERHYDRATION; DEUTERIUM DILUTION; TOTAL BODY WATER; EXTRACELLULAR WATER; PLASMA VOLUME
©2000The American College of Sports Medicine
What does "Remember me" mean?
By checking this box, you'll stay logged in until you logout. You'll get easier access to your articles, collections,
media, and all your other content, even if you close your browser or shut down your
To protect your most sensitive data and activities (like changing your password),
we'll ask you to re-enter your password when you access these services.
What if I'm on a computer that I share with others?
If you're using a public computer or you share this computer with others, we recommend
that you uncheck the "Remember me" box.
Highlight selected keywords in the article text.
Data is temporarily unavailable. Please try again soon.