Abstract
Aim: To generate reference values for bioelectrical impedance analysis (BIA) in a German collective of healthy pregnant women. Materials and Methods: A total of 90 women with a singleton gestation from 23+0 to 40+6 weeks of pregnancy were enrolled from April 2012 until May 2013. Each week of gestation was represented by the same number of women (n=5). Mean BIA indices were generated for all patients and according to gestational age so that three collectives <30, 30 to 35 and >35 weeks of gestation were formed. Multiple regression analysis was implemented using maternal height, weight, body-mass index, gestational age, hematocrit and abdominal circumference to generate formulae to calculate individualized reference values for resistance, reactance, phase angle, height2/resistance, and total body water adjusted for gestation. Results: Individualised reference values using the new formulae led to results which are more accurate than general mean values or mean values corresponding to the gestational age only. Conclusion: As far as we are aware of the present study is the first to report reference values for BIA indices in a German collective of healthy pregnant women. The new formulae generated by multiple regression analysis enabled for assessment of individualized reference values for BIA indices.
Maternal weight gain in pregnancy is widely considered necessary for an adequate fetal development (1). The increase of maternal weight depends on changes of various components including anatomical feto-maternal structures (fetus, uterus, amniotic fluid, amniotic membranes, placenta), maternal body fat and breast tissue as well as total body water (TBW). Alterations of TBW can be explained by the retention of water in breasts, lower pelvis and blood plasma which ensure for proper development of labour and puerperium (8, 20). Earlier studies have shown that TBW in pregnancy is strictly related to plasma volume (8). Furthermore, various authors describe a relationship between TBW and pregnancy outcome (4, 6, 13, 20). This implicates the clinical relevance of bioelectrical impedance analysis (BIA) in pregnancy, which by measuring TBW provides for example information about the adoption of the maternal organism to gestation. BIA is a standardized technique which is fast, non-invasive and therefore well-tolerated. Its instruments are affordable devices, especially the single-frequency instruments used in this study (12). These have been proven to be eminently suitable for non-laboratory settings (12). The physical properties of BIA, its measurement variables [resistance (R), reactance (Xc), phase angle (PA)], and their clinical significance have been described in many investigations (1, 10, 12). Classic whole-body single-frequency BIA applies a 50 kHz alternating current that penetrates cell membranes and leads to the measurement not only of extracellular but also of intracellular water, hence TBW (18). In various studies, pathological changes of maternal TBW by BIA have been related to gestational maladaptation, such as (gestational) hypertension, edema and pre-eclampsia (3, 20, 21). However, individual results must be interpreted against the background of adequate reference values for the population of interest, as bioelectrical properties and their relationship to body composition are affected by height, weight, hydration status and stage of life (17). Therefore, the aim of this study was to generate reference values for BIA indices of a healthy German collective.
Materials and Methods
A total of 90 pregnant women from 23+0 to 40+6 weeks of pregnancy were enrolled from April 2012 until May 2013. Each week of gestation was represented by the same number of women (n=5). This was an important criterion for proper statistical analysis. This prospective study was approved by the Ethics Committee II of the Heidelberg University, Heidelberg, Germany (2012-401M-MA). Only healthy women were enrolled after written informed consent. Pre-existing maternal diseases possibly affecting BIA, such as diabetes, hypertensive and renal disorders, were exclusion criteria. A standardized questionnaire was used for patient characteristics taking gravidity, parity, age, height, current weight, body mass index (BMI), abdominal circumference (AC), hematocrit and gestational age (GA) into account. A single-frequency BIA device (Biacorpus RX 4000; Medical HealthCare GmbH, Karlsruhe, Germany) was used in this study. This instrument is a fully-digital, phase-sensitive, 4 channel impedance measuring device. Each channel applies 50 kHz alternating current to measure R, Xc and PA. Eight electrodes were attached to the person's hands and feet. The person was then placed supine, limbs slightly abducted and palms pronated flat on the investigator's cot which was covered with a non-conducting surface. After cleaning the areas of the skin where the electrodes were about to be attached with alcohol swabs, the measurement electrodes were placed on the dorsal surface of the wrist and ankle at the level of the process of the radial and ulnar, or fibular and tibial bones. Signal electrodes were attached to the dorsal surface of the third metacarpal bone of hands and feet, so that at least a 5 cm distance was kept between signal and measurement electrodes (2). The resulting measurements are automatically transferred to a computer, where they are duly interpreted by the software. The manufacturer's software (BodyComp V 8.3; Medical HealthCare GmbH, Karlsruhe, Germany) was used. All data were recorded in an Excel datasheet and transferred into the SAS® environment (Statistical Analysis System, Release 9.3; Cary, NC, USA) for statistical analysis. The bioimpedance index (IR) and TBW adjusted for gestation (TBWcor) were calculated using the following formulae according to Lukaski et al. (11):
Quantitative data are presented as the mean and standard deviation, together with the range; for discrete variables, the median and range are given. Three statistical models were used in order to generate the most accurate reference values for each of the parameters R, Xc, PA, IR and TBWcor. In the first model the mean value of each parameter was calculated for all women investigated. Furthermore, the corresponding coefficients of variation (CV) were computed. In the second statistical model, mean values were generated according to GA, forming three collectives: <30, 30 to 35 and >35 weeks of gestation. As a measure of goodness-of-fit, the coefficient of correlation r2 was calculated. In the third model, multiple regression analysis was implemented using maternal height, weight, BMI, GA, hematocrit and AC to generate formulae to assess individualized reference values for R, Xc, PA, IR and TBWcor. Using the forward selection method in the SAS procedure PROC REG, each variable was selected for entry into the model up to a significance level of 0.10.
Results
A total of 90 healthy pregnant women were included in this investigation. Demographic characteristics are shown in Table I. Table II shows the results of all three statistic models and the five parameters implemented. Concerning the first analysis, PA gave fairly good results, with a CV of 12.6%, indicating that these data have a rather low variability. Concerning the second statistical model, results based on IR (r2=20.07) and TBWcor (r2=19.01) were much better than those for R (r2=14.72%) and Xc (r2=14.20 %). PA (r2=0.66%) clearly does not depend on the GA. Table III shows all formulae derived by stepwise regression analysis to predict R, IR, PA, TBWcor and Xc. Comparing the r2 values of the statistical models for each parameter, it is evident that individualized calculation of reference values using the formulae generated by regression analysis (third model) led to better results than using the gestational section only (second model). Regression analysis for IR as a dependent variable based on weight, height and GA (r2=64.40%) seems to be the best option for individual BIA estimates. TBWcor depends on the same variables (r2=60.37%).
Demographic characteristics of all 90 patients.
Discussion
This is the first study, as far as we know, concerning BIA reference values of a healthy German collective in pregnancy. We investigated 23+0 to 41+0 weeks of pregnancy as this is the gestational interval of primary obstetrical interest. Only few earlier studies addressed this topic. Mean values of R and Xc among rural Bangladeshi women in the third trimester were 646 Ω and 64 Ω (17). These values are higher than those reported from two smaller studies in the United States and a larger one in Italy using similar BIA devices with 50 kHz alternating current (4, 9, 11). In those investigations, the mean R and Xc values were lower, at 506-521 Ω and about 60 Ω in the third trimester, respectively. Mean values for R in our collective were 567.8 Ω and 516.5 Ω for patients with 30-35 and >35 weeks of gestation, respectively. Hence third trimester reference values of our collective are similar to those of other investigations in Western populations and are lower than those of the Asian study. An explanation for this inter-population difference might be body volume, which is markedly smaller in Asian than in Western women (17). A decline of R and Xc in late pregnancy was reported in earlier investigations (4, 9, 11, 17). In this context Lukaski et al. reported that in American women, a decline in pregnancy-associated R and Xc predicted gains in TBW, which represents about 50% of total pregnancy-associated weight gain (11). The PA in our investigation remained unchanged throughout the observation period, with a mean value of about 6.2 degrees. Lukaski et al. reported a mean value of 6.6 degrees which also remained unchanged (11). Shaikh et al., on the other hand, observed a decline in PA from 6.1 to 5.7 degrees from early to late pregnancy, possibly indicating a compromised body-cell mass profile in the cause of pregnancy (17). The PA in this context is clinically-interesting as it reflects different electrical properties of tissues that are affected in various ways by hydration status and cellular membrane integrity, without algorithm-inherent errors, or requiring assumptions such as constant tissue hydration (7). The PA is an indicator of the distribution of body water between intra- and extracellular spaces (5). A high PA corresponds to a low ratio of extra- to intracellular water (1). A low PA suggests cell death or decreased cell integrity, whereas higher PAs suggest large proportions of intact cell membranes (16). The PA has been used to predict clinical outcome in various ailments (14, 15, 19). PA, as well as R and Xc, might also be useful as prognostic markers in pregnancy-associated diseases, such as hypertensive disorders. There is an ongoing prospective investigation at our Department analyzing BIA in the assessment of pre-eclampsia, which might reveal new diagnostic options. Concerning the present investigation, the third statistical model, generating individualized reference values by stepwise regression analysis, might be useful for further investigations. To our knowledge, regression analysis generating formulae in order to individually-generate reference values has not been accomplished before. Our results show that this approach leads to a more accurate assessment of BIA indices in pregnant women compared to common stratification by GT.
Results of the three statistical models.
The regression equations to estimate resistance (R), bioimpedance index (IR), phase angle (PA), total body water adjusted for gestation (TBWcor) and reactance (Xc).
Conclusion
To our knowledge the present study is the first to report reference values for a healthy German collective. Reference values for R, Xc, PA, IR and TBW adjusted for gestation can be calculated individually using these new formulae.
Footnotes
-
↵* These Authors contributed equally to this work.
- Received August 20, 2013.
- Revision received October 21, 2013.
- Accepted October 23, 2013.
- Copyright © 2013 The Author(s). Published by the International Institute of Anticancer Research.
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.