| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Division of Nutrition, Department of Biomedicine and Surgery, University of Linkoping, SE-58185 Linkoping, Sweden
Submitted 22 August 2003 ; accepted in final form 14 November 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
extracellular water; gestation; intracellular water; total body water
Measurements of body composition are fundamental in all studies where the nutritional status of population groups is assessed and especially so during pregnancy when the body fat content changes rapidly. Another interesting variable during pregnancy is the amount of extracellular water (ECW) in the body. This is because human pregnancy may be complicated by preeclampsia or eclampsia, a condition with an incompletely understood etiology. Women suffering from this condition show signs of edema and have a reduced plasma volume expansion compared with healthy pregnant women (17). This is important because poor plasma volume expansion has been associated with poor fetal growth (21). Considering these facts, the so-called bioimpedance spectroscopy (BIS) technique is of interest. This technique, which is based on the conductive properties of body tissues, is able to estimate the amounts of ECW and intracellular water (ICW) in the body and hence the total body water (TBW), calculated as the sum of ECW and ICW (7). The latter is important because estimates of body fat are commonly based on assessments of TBW. Several related techniques based on the conductive properties of the body have been used to study TBW and ECW of pregnant women. For example, Valensise et al. (26) reported that the multifrequency bioimpedance technique can be used to monitor variations in body water compartments in normal pregnancy and to detect abnormal changes in body water distribution in women with gestational hypertension. Furthermore, Ghezzi et al. (10) showed that bioelectrical impedance indexes of pregnant women were significant predictors of birth weight. Although relationships between estimates obtained by means of bioimpedance and body water compartments have not been fully established during pregnancy, these results do indicate that techniques based on the conductive properties of tissues have a potential for use in studies of human pregnancy.
The BIS technique is based on the principles that tissues containing water and electrolytes are more conductive than bone and fat and that the volume of a conductive tissue can be deduced from its resistance (2). By measuring resistance at low and high frequencies, BIS provides estimates of ECW as well as of ICW, because, at low frequencies, the flow of the current through the cell is blocked, whereas at higher frequencies it flows through both the extracellular and intracellular space (7). The technique is noninvasive, very easy to use, may be repeated several times in the same subject, involves no health hazards, and, therefore, represents an ideal method for studies of pregnant women. The convenience of the technique is of special interest, because simple and accurate methods for assessing TBW, ECW, and ICW during pregnancy in the clinical situation are presently lacking. However, although several studies have shown that BIS is able to provide accurate results in healthy adults (28, 29), the method has so far been evaluated in only one group of pregnant women. In that study, average values of TBW and ECW obtained by means of BIS in gestational weeks 24–26 and 34–36, as well as 4–6 wk postpartum, were in good agreement with corresponding results obtained by means of reference methods. However, no comparable data for the first one-half of pregnancy have been reported. Furthermore, the potential of BIS to assess ICW during reproduction has not been studied, and no data regarding the ability of BIS to assess changes in ECW and ICW, or in TBW, during pregnancy are available.
In the BIS technique, so-called resistivity coefficients are needed to calculate tissue volumes from resistance, and commonly used coefficients are intended for the average healthy adult. However, resistivity coefficients are influenced by the body composition of the subject under study (6). This point was noted by van Loan et al. (27) when evaluating BIS in pregnant women. Thus these authors calculated resistivity coefficients, specific for their population of women, which they used when evaluating BIS during pregnancy. These population-specific resistivity coefficients made it possible to successfully predict both ECW and TBW before, during, and after pregnancy. The BIS results were apparently calculated by the same model on all occasions. Thus ECW, ICW, and TBW could be calculated in pregnant women by a model developed for nonpregnant subjects, if population-specific resistivity coefficients were used. The study was, however, based on a limited number of women, and the authors emphasized the need for more research in the area (27).
The aims of this study were to evaluate the potential of BIS, by using a model developed for the nonpregnant body; to estimate ECW, ICW, and TBW before conception, during early and late pregnancy, and postpartum, as well as to evaluate its potential to estimate changes in these variables during the reproductive cycle; and to compare results for ECW, ICW, and TBW as obtained by means of BIS, by using resistivity coefficients calculated for our specific population, with the corresponding results obtained when generally applied resistivity coefficients, as provided by the manufacturer of the BIS analyzer, were used.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Protocol. The women were studied before conception, in gestational weeks 14 and 32, as well as 2 wk postpartum. The following procedure was repeated on each occasion. After an overnight fast, the subject arrived at the hospital by car. After drawing a venous blood sample, the subject was given an oral dose of sodium bromide, and her ECW and ICW were measured by using BIS. An oral dose of stable isotopes was then given to the subject. Another blood sample was obtained 4 h after the dose of sodium bromide, making it possible to estimate the subject's ECW by means of bromide dilution (ECWref). During the next 14 days, the subject collected five urine samples to estimate TBW by means of isotope dilution (TBWref).
Isotope dilution. Each subject was given an accurately weighed dose of 2H2O and H2 18O (0.05 and 0.15 g/kg body wt, respectively) after collection of two or three background urine samples during a 2- to 7-day period before dosing. Another five urine samples were collected 1, 4, 8, 11, and 15 days after the day of dosing. The date and time of day when a sample was collected was always noted. The same procedure was repeated at all measurements, but only 2H2O was given at the postpartum measurement. Urine samples were stored in glass vials with internal aluminum-lined screw cap sealing at +4°C until sample collection was completed. They were then stored at -20°C until analyzed. Isotopic enrichments of dose and urine samples were analyzed by an isotopic ratio mass spectrometer fitted with a CO2/H2/H2O equilibrium device (Deltaplus XL, Thermoquest, Bremen, Germany). The procedure described by Thielecke and Noack (25) was followed, except that equilibration time was 180 and 840 min for H2 and CO2, respectively. Isotope dilution spaces (ND and NO) were calculated from zero-time enrichments obtained from the exponential isotope disappearance curves that provided estimates for the rate constants, kD and kO, for 2H and 18O, respectively. The mass spectrometric response was standardized using Vienna standard mean ocean water. Dose and urine samples from the same subject were always analyzed on the same occasion within the same equilibrium device when a linear mass spectrometric response was also confirmed for both isotopes. Analytic precision in the measurement range, for results expressed as mole fraction, was 0.44 ppm for 2H and 0.15 ppm for 18O. ND/NO for our 21 subjects was 1.037 ± 0.007, 1.029 ± 0.008, and 1.027 ± 0.008 before pregnancy and in gestational weeks 14 and 32, respectively. Before pregnancy and in gestational weeks 14 and 32, TBWref was the average of ND/1.04 and NO/1.01, whereas it was ND/1.04 postpartum.
Bromide dilution. Each subject was given an accurately weighed dose of sodium bromide (40 mg bromide/kg body wt). Bromide concentration in dose and serum samples was analyzed by using a spectrophotometric procedure based on the forming of a gold bromide complex (24). Bromide space was calculated as dose of bromide/(the concentration of bromide in serum at 4 h postdose - the concentration of bromide in the baseline serum sample). ECWref was calculated as bromide space x 0.90 x 0.95 x 0.94 (3).
BIS. The measurements were performed with careful attention to the instructions provided by the manufacturer of the equipment used (Hydra ECF/ICF model 4200 with Hydra BIS 4200 Software Utilities, Xitron Technologies, San Diego, CA). Thus, after body weight was recorded, the subject was placed in a supine position with her arms abducted from her body and her legs separated. Two electrodes were placed 5 cm apart on the dorsal surface of the right hand and another two on the right foot. The software estimated the resistance of ECW and of ICW on the basis of measurements of resistance and reactance at 50 frequencies between 5 kHz and 1 MHz with the Cole model (5). The accuracy of the fit to the Cole model was rated as excellent for all measurements by the software. ECW and ICW were calculated, by means of the software, from their resistances using equations based on the Hanai mixture theory (12). For these calculations, general resistivity coefficients, as provided by the manufacturer (39.00 and 264.90 
·cm for ECW and ICW, respectively), as well as resistivity coefficients calculated especially for the women in our study, were used. The latter were calculated by using the software described above, using TBWref, ECWref, and the resistance of ECW and ICW, respectively, as measured by means of the Hydra equipment, for our 21 subjects before conception. These population-specific resistivity coefficients were 39.61 ·cm for ECW and 269.57 ·cm for ICW. TBW was calculated as the sum of ECW and ICW. TBW, ECW, and ICW calculated using the resistivity coefficients provided by Xitron Technologies will be referred to as TBWXT, ECWXT, and ICWXT, respectively, whereas the corresponding figures obtained using the population-specific resistivity coefficients are labeled TBWPS, ECWPS, and ICWPS, respectively.
Body weight and height. The body weight of the women in light underwear was recorded at the different measurements on a high-precision scale (KCC 150, Mettler-Toledo). Their body weights were also recorded in the delivery room before childbirth. Height was measured by a wall stadiometer.
Statistics. Values given are means ± SD. Significant differences between group averages were identified by repeated ANOVA and subsequent post hoc analysis using Tukey's multiple-comparison test (13). Paired t-test (13) was also used. Agreement between results obtained with BIS and the appropriate reference methods was examined according to Bland and Altman (4). Linear regression and correlation analyses were performed as described by Hassard (13). Significance was accepted at the P < 0.05 level. All statistical analyses were carried out by using Statistica software, 6.0 version (StatSoft, Scandinavia, Uppsala, Sweden).
|
RESULTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
|
|
ECW. Table 2 shows ECWXT, ECWPS, and ECWref before, during, and after pregnancy. ECWXT was significantly lower than ECWref in gestational weeks 14 and 32, as well as 2 wk postpartum, whereas ECWPS was significantly lower than ECWref in gestational week 32 and 2 wk postpartum. Average ECWPS was similar to average ECWXT at all four measurements. ECWXT is compared with ECWref, according to Bland and Altman, at the different stages of reproduction in Table 3 and in Fig. 1. The mean differences between ECWXT and ECWref were small before conception and in gestational week 14 (-0.13 and -0.84 kg, respectively). However, in gestational week 32, the mean difference was as high as -3.12 kg for ECWXT vs. ECWref. Moreover, in gestational week 32, ECWXT was lower than ECWref for all subjects, as shown in Fig. 1. In addition, a significant linear relationship between the average of ECWXT and ECWref (x) and ECWXT - ECWref (y) was found (y = -0.3427x + 2.800; r = -0.570; P < 0.05). At the measurement 2 wk postpartum, the mean difference was small, and ECWXT - ECWref was only -0.84 kg (Table 3). As also presented in Table 3, limits of agreement (±2 SD) for ECWXT - ECWref ranged from ±2.08 to ±3.46 kg at the four different measurements. Results obtained for ECWPS (data not shown) were very similar to those described above for ECWXT. Table 2 also shows that average values for ECWref, ECWXT, and ECWPS were all higher in gestational weeks 14 and 32, as well as postpartum, compared with the corresponding values obtained before pregnancy. However, Table 4 shows that increases obtained by means of BIS, i.e., using ECWXT, were lower than the corresponding increases obtained by means of the reference method, i.e., using ECWref, in gestational weeks 14 and 32, as well as 2 wk postpartum, and that these increases in ECWXT and ECWref were significantly different in gestational week 32 and 2 wk postpartum. The corresponding results for ECWPS (data not shown) were very similar to those shown in Table 4 for ECWXT.
|
|
|
ICW. Table 2 shows ICWXT, ICWPS, and ICWref before, during, and after pregnancy. ICWXT and ICWPS were both similar to ICWref at all measurements, and neither ICWXT nor ICWPS was significantly different from ICWref at any of the measurements. When ICWXT was compared with ICWref, according to Bland and Altman, the mean differences (ICWXT - ICWref) were small at all four measurements (-0.19, -0.30, -0.63, and 0.13 kg before pregnancy, in gestational weeks 14 and 32, and 2 wk postpartum, respectively). However, the corresponding limits of agreement (±2 SD) were large: ±4.32, ±4.14, ±6.52, and ±3.76 kg, respectively. No significant relationship between the average of ICWXT and ICWref (x) and ICWXT - ICWref (y) was found at any of the measurements. When substituting ICWPS for ICWXT in these comparisons, very similar results were found (data not shown). With the use of the data presented in Table 2, it can be calculated that, compared with the average prepregnant value, average ICWXT, ICWPS, and ICWref had all changed very little in gestational week 14, but they had increased significantly by 
2 kg in gestational week 32. Postpartum average ICWXT, ICWPS, and ICWref were all 1 kg higher than the corresponding value before conception, and the increases for ICWXT and ICWPS were significant. Table 4 shows that increases vs. the prepregnant value obtained by means of BIS, i.e., using ICWXT, were not significantly different from the corresponding increases obtained by means of reference methods, i.e., using ICWref, either in gestational weeks 14 or 32 or 2 wk postpartum. The corresponding results for ICWPS (data not shown) were very similar to those shown in Table 4 for ICWXT.
TBW. Table 2 shows TBWXT, TBWPS, and TBWref before, during, and after pregnancy. Average TBWPS was similar to average TBWXT at all four measurements. TBWXT was significantly lower than TBWref in gestational week 14, whereas TBWPS and TBWXT were both significantly lower than TBWref in gestational week 32. Table 5 compares TBWXT with TBWref at the different stages of reproduction, according to Bland and Altman. As shown in Table 5, the mean difference between TBWXT and TBWref was small before conception, in gestational week 14, and 2 wk postpartum (ranging from -0.33 to -1.13 kg). However, in gestational week 32, the difference vs. TBWref was as high as -3.75 kg. Limits of agreement (±2 SD) for TBWXT - TBWref ranged from ±4.66 to ±5.88 kg at the four different measurements (Table 5). No significant linear relationships between the average of TBWXT and TBWref (x) and TBWXT - TBWref (y) were found at any of the measurements. Results obtained for TBWPS (data not shown) were very similar to those described above for TBWXT. Table 2 also shows that average values for TBWref, TBWXT, and TBWPS were all higher in gestational weeks 14 and 32, as well as 2 wk postpartum compared with the corresponding values obtained before pregnancy. However, as shown in Table 4, such increases obtained by means of BIS, i.e., using TBWXT, were lower than the corresponding increases obtained by means of the reference method, i.e., using TBWref, in gestational weeks 14 and 32, as well as 2 wk postpartum, and these two ways of measuring increases in TBW produced significantly different results in gestational week 32. The corresponding results for TBWPS (data not shown) were very similar to those shown in Table 4 for TBWXT.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
7 kg of TBW during the first 32 wk of pregnancy, which is somewhat higher than the corresponding estimate made by Hytten (14). Furthermore, our subjects had gained almost 5 kg of ECW in gestational week 32, which is slightly higher than the corresponding estimate made by Hytten, but in agreement with data from the only longitudinal study where ECW has been measured before and during pregnancy in the same women by using bromide dilution (27). In the present study, average ECWXT and ECWPS, as well as average TBWXT and TBWPS, agreed well with corresponding results obtained by means of the reference methods when our subjects were measured before pregnancy. The Bland and Altman comparison showed, however, that the limits of agreement for TBW and ECW (Tables 3 and 5, respectively) were somewhat wider than those reported by van Marken Lichtenbelt et al. (29) in a group of healthy young women. This is probably due to the fact that our subjects were more variable with respect to body weight and body mass index than the subjects studied by van Marken Lichtenbelt et al., since a recent report has indicated that the degree of body fatness affects the accuracy of BIS (6).
The underestimation of TBW and ECW in late pregnancy when BIS was used, as found in the present study, is in contrast to the results of van Loan et al. (27) in American women. It is, therefore, of interest to consider the reference methods used in the two studies. The bromide method as well as the isotope dilution method that were used in both studies represent well-established methods for estimating ECW and TBW (22). However, some minor methodological differences should be pointed out. First, in contrast to van Loan et al. (27), we multiplied bromide space by 0.94 to correct for the water content of serum. However, even without this correction, our BIS measurements underestimated TBW and ECW by 3 kg in week 32 of pregnancy. Second, we used both 2H and 18O to estimate TBWref, whereas van Loan et al. used only 2H. Recalculation of our data using only the 2H results changed TBW by <0.4% and did not affect the conclusions presented in this paper. Third, we calculated TBWref using the back-extrapolation approach, whereas van Loan et al. used both the back-extrapolation and the plateau methods to calculate TBW, although it is not quite clear at which stages of reproduction the respective methods were used. When we recalculated our data using 4-h postdose isotope enrichment results to obtain data comparable to the plateau method, we obtained 2% higher TBW. Use of these higher values for TBW did not affect the results and conclusions of this study in any important way. Thus we are unable to show that methodological differences explain the different results obtained in our study compared with the study by van Loan et al. The populations in the two studies were also different. Thus our subjects varied more with respect to body fat content before pregnancy and also had higher gestational weight gains. However, exclusion of the subjects with the highest body fat content (>39%) or subjects with the lowest fat content (<18%) did not change our results. Furthermore, when the eight subjects with the highest gestational weight gains (>23 kg) were excluded, BIS still underestimated ECW and TBW to the same extent as reported in Tables 2 and 5. Finally, it is conceivable that the difference in results, as observed in the two studies, is related to the resistivity coefficients used. In our study, the population-specific resistivity coefficients were similar to the general resistivity coefficients provided by the manufacturer of the equipment used. Consequently, ECWPS was similar to ECWXT, ICWPS was similar to ICWXT, and hence TBWPS was similar to TBWXT. These findings are in contrast to those of van Loan et al. (27), in which the use of population-specific resistivity coefficients was essential for the good agreement between BIS and the reference methods before and during pregnancy. Unfortunately, van Loan et al. did not present any values either for their population-specific resistivity coefficients or for the resistivity coefficients provided by the manufacturer of the equipment used.
This study provides some interesting aspects concerning the possibility of assessing ICW during the reproductive cycle. Although we found that the limits of agreement were large for the differences between ICW obtained by means of BIS and the reference methods before, during, and after pregnancy, we also found that average estimates of ICW, obtained by means of BIS, were in good agreement with the corresponding reference data. The large limits of agreement are probably related, at least to some extent, to the relatively large random error in the reference estimates when ICW is assessed as the difference between TBWref and ECWref (22). This imprecision in the reference estimates of ICW limits the possibility of assessing the potential of BIS for studying ICW during human reproduction, especially when evaluating its potential for measuring ICW in individual women. However, our data certainly support the conclusion that BIS, based on a model developed for nonpregnant subjects, may be useful for estimating average ICW, as well as changes in ICW, in groups of women during reproduction. In this context, it is relevant to note that Earthman et al. (8) provided evidence showing that BIS could accurately assess changes in body cell mass in patients with human immunodeficiency virus infection. Increases in ICW during pregnancy are likely to include changes in the maternal body, for example, increases in mammary and uterine tissues, as well as growth of the fetus and placenta, whereas increases due to maternal blood volume expansion and accumulation of amniotic fluid, for example, would not be included in such estimates. Obtaining valid estimates of ICW may, therefore, represent a new possibility to explore the physiology of human pregnancy in relation to health and disease.
Our results clearly show that BIS, as applied in the present study, is unable to produce valid estimates of ECW in women in gestational week 32. It is, therefore, relevant to consider the model used to calculate body water compartments in this study. This model was developed for the nonpregnant body, which differs in a number of respects from the pregnant body. For example, the model (30) includes a factor allowing for the proportions of the body that may obviously be altered during pregnancy. It is also relevant to note that, as shown in a study on women in gestational weeks 18 and 36 (9), electrolyte concentrations in ECW and ICW change during pregnancy. Nevertheless, in the present study, fairly accurate estimates of ECW and ICW in gestational week 14 and ICW in gestational week 32 were obtained by using the model developed for nonpregnant subjects. This seems to indicate that modifying this model in a way that makes it theoretically valid for pregnant women may not be particularly complicated. In this context, it should also be emphasized that our results showing that the applied model (30) can accurately assess ECW and ICW in gestational week 14 as well as ICW in gestational week 32 are based on empirical evidence only. Until a new model is available that takes the physiological and anatomic situation of pregnant women properly into account, estimates of ECW and ICW during gestation obtained by means of BIS should be interpreted with caution.
It is also important to point out that the model we used to assess body water compartments by means of BIS is based on wrist-ankle measurements of resistance and assumes that the body consists of five cylinders, the arms, the legs, and the trunk, connected in a series, in which conductors with the smallest cross-sectional area (i.e., arms and legs) will determine most of the resistance, whereas the part with the largest cross-sectional area (i.e., the trunk) will contribute less (2). This model may not be the most appropriate during pregnancy when a substantial amount of the water retained is located in the trunk. This suggestion is supported by reports that have pointed out limitations of a technique based on principles similar to those of BIS, i.e., multiple bioimpedance analysis, in which the wrist-ankle approach has been used to estimate ECW in patients with ascites (16) and in patients receiving peritoneal dialysis (31). Moreover, Zhu et al. were able to measure changes in ECW in patients during peritoneal dialysis by combining whole body measurements with segmental measurements of the trunk (31). Whether this approach is able to improve the accuracy of BIS in late pregnancy could be the topic for future studies.
In conclusion, when based on a model developed for nonpregnant subjects, the BIS technique produced fairly accurate but imprecise estimates of ECW, ICW, and TBW before pregnancy, in early pregnancy, and postpartum. In late pregnancy, BIS underestimated ECW and TBW while ICW remained accurate. Furthermore, BIS underestimated pregnancy-related increases in ECW and TBW but was able to estimate changes in ICW accurately, which may represent a new possibility for studying the physiology of human pregnancy. All of these findings were the same whether general resistivity coefficients, as provided by the manufacturer of the BIS analyzer, or population-specific resistivity coefficients were used. However, general application of BIS for assessing body water compartments during gestation requires the development of a new model that takes into account the physiological and anatomic changes occurring during pregnancy.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
GRANTS
The study was supported by the Swedish Research Council (project no. 12172), the Swedish Nutrition Foundation, the County Council of Ostergotland, the Health Research Council in the Southeast of Sweden, and Knut and Alice Wallenberg's Foundation.
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. Earthman, D. Traughber, J. Dobratz, and W. Howell Bioimpedance Spectroscopy for Clinical Assessment of Fluid Distribution and Body Cell Mass Nutr Clin Pract, August 1, 2007; 22(4): 389 - 405. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
