The effect of a meal on measures of impedance and percent body fat estimated using contact-electrode bioelectrical impedance technology

European Journal of Clinical Nutrition (2013) 67, 950 955 & 2013 Macmillan Publishers Limited All rights reserved 0954-3007/13 www.nature.com/ejcn ORIGINAL ARTICLE The effect of a meal on measures of impedance and percent body fat estimated using contact-electrode bioelectrical impedance technology CB Dixon 1, B Masteller 1 and JL Andreacci 2 BACKGROUND/OBJECTIVES: To determine the effect of a meal on impedance and percent body fat (%BF) determined using contact-electrode bioelectrical impedance analysis (BIA) technology. SUBJECTS/METHODS: Forty-three adults (23 women and 20 men) volunteered to participate in this study (age ¼ 20.5±1.1 years; body mass index ¼ 24.1±3.8 kg/m 2 ). Body composition was assessed using three BIA analyzers: leg-to-leg (LBIA), segmental (SBIA) and multi-frequency (MFBIA), on two separate occasions. After a baseline measurement, subjects consumed a meal or received nothing, which served as the control (CON). Subjects were reassessed 20, 40 and 60 min following (POST) the baseline measure in each condition. RESULTS: Twenty minutes after eating (3847±900 kj), body mass (LBIA ¼ 0.8 kg, SBIA ¼ 0.8 kg, MFBIA ¼ 0.7 kg, Po0.05), impedance (LBIA ¼ 6.0 O, SBIA ¼ 17.9 O, MFBIA ¼ 27.1 O, Po0.05) and %BF (LBIA ¼ 0.9%, SBIA ¼ 1.7%, MFBIA ¼ 0.8%, Po0.05) increased significantly and remained elevated at 60 min POST. During the CON trial, a consistent body mass reduction (60 80 g) and impedance increase (4 9 O) was observed over time resulting in a small increase in %BF (0.3 0.7%) 60 min POST (Po0.05). CONCLUSIONS: Twenty minutes after eating, %BF increased due to elevations in impedance and body mass. As such, when precision is critical, we recommend adhering to the pretest fasting guidelines to avoid meal-induced alterations in %BF estimates. In addition, use of a consistent testing schedule may minimize normal %BF variation over time. European Journal of Clinical Nutrition (2013) 67, 950 955; doi:10.1038/ejcn.2013.118; published online 3 July 2013 Keywords: bioelectrical impedance analysis; BIA; body composition; eating INTRODUCTION The utilization of bioelectrical impedance analysis (BIA) technology to assess body composition has increased in recent years. Partially responsible for the popularity may be the development of the easy to use contact-electrode BIA analyzers. The contactelectrode analyzers, which measure impedance as an individual stands on a scale-like platform, differ significantly from the traditional BIA method that requires the accurate placement of gel electrodes at specific anatomical locations. Currently, several different types of BIA analyzers are available including leg-to-leg (LBIA), segmental (SBIA) and multi-frequency (MFBIA), each named after the electrical current and/or electrical pathway used to measure the impedance of body tissues. Collectively, the BIA analyzers are portable, fast and a relatively inexpensive method of evaluating body composition, thereby increasing their appeal in clinical and health/fitness settings. 1 3 During the assessment, the BIA analyzers introduce an electrical current into the body and measure the impedance or resistance to current flow. Fat-free mass, due to its high water and electrolyte content, is highly conductive, whereas adipose tissue contains little water and is therefore a poor conductor (that is, higher impedance). Body composition measurements, such as percent body fat (%BF) and fat-free mass, are automatically calculated by the analyzer using preprogrammed prediction equations that combine impedance and body mass measurements with height, gender and age information. 4 The accuracy of the body composition measurements calculated by the analyzer is dependent upon the accuracy and precision of the impedance measurement. 5 Previous research has demonstrated that impedance is affected by factors that produce shifts in body fluids or electrolytes. 6 11 Therefore, controlling pretest behaviors that may alter hydration state is recommended when using BIA technology. 4,5 For instance, not eating or drinking 4 h before testing is a traditional BIA pretest guideline; 4 however, few studies have examined the impact that consuming a meal has on contactelectrode BIA body composition measures. If necessary, this restriction significantly reduces the practicality of utilizing BIA analyzers in settings where controlling pre-assessment behavior may be difficult or unrealistic. Recently, we examined the effect of acute fluid consumption on LBIA and SBIA-determined impedance, and %BF measures in adults 9,10 and children. 11 Collectively, fluid-induced elevations in body mass and impedance resulted in small, but statistically significant, increases in %BF (mean change ¼ 0.5 1.0%) in those investigations. 9 11 However, the previous studies examined only 1 Department of Health Science, 165 Willis Health Professions Center, Lock Haven University, Lock Haven, PA, USA and 2 Department of Exercise Science, Bloomsburg University, Bloomsburg, PA, USA. Correspondence: Dr CB Dixon, Department of Health Science, 165 Willis Health Professions Center, Lock Haven University, Lock Haven, PA 17745, USA. E-mail: cdixon@lhup.edu Contributors: CBD and BM were responsible for study design, data collection and analysis, and writing of the manuscript. JLA was responsible for statistical analysis and assisted in writing of the manuscript. Received 6 February 2013; revised 23 May 2013; accepted 24 May 2013; published online 3 July 2013

the impact that drinking had on contact-electrode BIA measurements post consumption. Previously, the influence of a breakfast meal on BIA-determined body composition measurements was examined in young adults 12 and elderly persons. 13 Gallagher et al. 12 reported a significant decrease in body impedance 2 h after the consumption of a 2300 kj breakfast meal; body composition measurements were not examined in this study. Conversely, Vilaca et al. 13 found no difference in fat-free mass or fat mass between fasting and 1 h after consuming a 1250 kj breakfast meal in 41 elderly men. Of importance, the aforementioned studies of Gallagher et al. 12 and Vilaca et al. 13 utilized the traditional BIA method rather than the increasing popular contact-electrode BIA technology. Clearly, several operational differences exist between the traditional approach and many of the contact-electrode devices that are manufactured today. Most notably, differences are related to the skin-electrode interface (gel vs plate), electrical frequency used, anatomical assessment region (for example, lower body, upper body and whole body), body position (supine vs upright) and prediction equation selection. 4 To our knowledge, the effect of food ingestion on body composition measurements determined by contact-electrode BIA technology has yet to be examined. As such, the purpose of this investigation was to determine the effect of a meal on impedance and %BF determined using contactelectrode BIA technology (LBIA, SBIA and MFBIA). It is anticipated that our findings may further clarify whether the pretesting fasting restriction is necessary when using the contact-electrode BIA technology to assess %BF. MATERIALS AND METHODS Subjects Forty-three adults (23 women and 20 men) between 19 and 23 years of age volunteered to participate in this study. The Lock Haven University Institutional Review Board approved the study protocol and methods. All subjects signed an informed consent form before participation. Study procedures Each subject reported to the body composition laboratory for testing on two separate days: an experimental (EXP) and control (CON) trial. Subjects were instructed to adhere to the following traditional BIA and manufacturer-recommended pretest guidelines: 4 (a) no food or drink within 4 h of the test, (b) no exercise within 12 h of test, (c) no alcohol consumption within 48 h of the test, (d) empty bladder within 30 min of the test and (e) no diuretic medications within 7 days of the test. Subject compliance to these guidelines was confirmed before each trial. After an initial LBIA, SBIA and MFBIA body composition measurement, subjects consumed a meal (EXP trial) or received nothing, which served as the CON trial. The meal consisting of a beverage, a sandwich, salad or wrap and a bag of chips or pretzels was self-selected from a provided menu. Drinks were served cold (3 5 1C) to increase the gastric emptying rate from the stomach. 14 Subjects were permitted 15 min to consume the meal. The treatment order for each subject (EXP vs CON) was determined using a counterbalanced assignment. The BIA measurements were reassessed 20, 40 and 60 min following (POST) the baseline measure in each condition. During the 60-min assessment duration, subjects sat quietly. Laboratory temperature was maintained at a constant 23 1C for all assessments. Height was determined using a Seca 240 wall-mounted mechanical measuring rod (Seca Corp., Hanover, MD, USA). Leg-to-leg BIA LBIA measurements were determined using a Tanita body fat analyzer; model TBF-300 A (Tanita Corporation of America, Inc., Arlington Heights, IL, USA). Each subject, wearing only a t-shirt and shorts, stood erect with bare feet placed properly on the contact electrodes of the LBIA instrument. As previously described, 2 the LBIA system consists of four contact electrodes (two anterior and two posterior) that are mounted on the surface of a platform scale. During the measurement, an electrical signal at a frequency of 50 khz is passed through the anterior electrode on the scale platform and the voltage drop is measured on the posterior electrode. Lower-body impedance and body mass were measured simultaneously while the subject stood on the LBIA platform. The LBIA analyzer, using preprogrammed proprietary equations developed by the manufacturer, automatically calculated %BF. Segmental BIA The Tanita BC-418 8 contact electrode analyzer (Tanita Corporation of America, Inc.) was used to collect SBIA measurements. Each subject stood erect holding the hand electrodes with bare feet placed properly on the contact electrodes of the SBIA instrument. Arms were held in the straightdown position without touching their sides. As previously described, 15 the SBIA system consists of four contact electrodes (two anterior and two posterior) that are mounted on the surface of a platform scale and each extremity handgrip has an anterior and posterior electrode. All measurements are carried out using a constant single-frequency current (50 khz). Whole-body impedance was measured using an ipsilateral foot hand electrical pathway. The SBIA analyzer automatically calculated %BF using preprogrammed proprietary equations developed by the manufacturer. Multi-frequency BIA The MFBIA measurements were measured using the InBody 520 (Biospace Co., Beverly Hills, CA, USA). The InBody 520 measures the direct segmental impedance across both legs, arms and the trunk at multiple frequencies (5, 50 and 500 khz). Like SBIA, the system s 8-point tactile electrodes contact the body at two points in each hand and foot. Body mass and five segmental impedance measurements (right arm, left arm, trunk, right leg and left leg) are automatically measured while the subject stands erect holding the hand electrodes with bare feet placed properly on the contact electrodes of the MFBIA scale-like platform. As recommended by the manufacturer, arms were held in the straight-down position without touching the sides of the trunk. At each of the three electrical frequencies, the five segmental impedance measurements were summed to create a total segmental impedance value to be used for statistical comparison. The MFBIA analyzer automatically calculated %BF using preprogrammed proprietary equations developed by the manufacturer. Statistical methods Data were analyzed using SPSS 19.0 for Windows (SPSS, Inc., Chicago, IL, USA). All values are expressed as mean±s.d. unless otherwise noted. Dependent variables were analyzed using a two-factor repeated measures analysis of variance. The factors were meal condition with two levels (CON or EXP) and time with four levels (baseline, 20, 40 and 60 min). Pairwise comparisons were conducted using the Bonferroni approach to discriminate between means when analysis of variance yielded significant results. Bland Altman plots were constructed and correlation coefficients were computed to determine whether body mass affected the magnitude of change POST. 16 Statistical significance was established a priori at Po0.05 for all analyses. RESULTS Characteristics of the 43 adults who participated in this investigation are presented in Table 1. The average duration for completing both trials was 6.0±1.5 days. The body mass (46.6 116.1 kg) and the body mass index (18.4 35.2 kg/m 2 ) range for the sample were fairly large. The average energy content of the meal consumed during the EXP trial was 3847±900 kj. LBIA analyzer data between treatment conditions and over time are presented in Table 2. LBIA-determined %BF, body mass and impedance increased 20 min after eating, and remained elevated above baseline at 40 and 60 min POST (Po0.05; Table 2). During the CON trial, %BF and impedance increased (60 min), and body mass decreased (40 and 60 min) over time (Po0.05; Table 2). However, the EXP trial %BF change at 60 min was significantly greater (Po0.05) than that observed during the CON trial (Table 2). SBIA analyzer data between treatment conditions and over time are presented in Table 3. SBIA-determined %BF, impedance and body mass increased 20 min after eating, and remained elevated above baseline at 40 and 60 min POST (Po0.05; Table 3). During 951 & 2013 Macmillan Publishers Limited European Journal of Clinical Nutrition (2013) 950 955

952 the CON trial, %BF (60 min) and impedance (40 and 60 min) increased, whereas a slight reduction in body mass was observed (40 and 60 min) over time (Po0.05; Table 3). Once again, the EXP trial changes were significantly greater (Po0.05) than those observed during the CON trial (Table 3). MFBIA analyzer data between treatment conditions and over time are presented in Table 4. MFBIA-determined %BF, impedance (5, 50 and 500 khz) and body mass increased 20 min after eating and remained elevated above baseline at 40 and 60 min POST (Po0.05; Table 4). During the CON trial, %BF (40 and 60 min) and impedance (20, 40 and 60 min) also increased over time (Po0.05; Table 4). The EXP trial changes were significantly greater (Po0.05) than CON variability at 20 min (%BF and impedance) and 40 min (impedance), respectively. However, at 60 min, the %BF and impedance changes from baseline observed during the EXP trial were not different than the CON trial (P40.05, Table 4). Table 1. Subject characteristics Women (n ¼ 23) Men (n ¼ 20) Mean±s.d. Range Mean±s.d. Range Age (years) 20.2±1.0 19.0 23.0 20.9±1.0 19.0 23.0 Height (cm) 164.9±4.0 154.9 172.7 177.8±6.5 168.9 190.5 Body mass (kg) 63.5±9.4 46.6 85.0 80.1±11.8 52.4 116.1 BMI (kg/m 2 ) 23.0±3.0 19.0 30.7 25.4±4.2 18.4 35.2 Body fat (%) 24.9±6.7 11.7 40.1 14.0±8.5 5.5 39.7 Abbreviation: BMI, body mass index. Individual differences in body mass influenced the SBIA determined %BF and the SBIA and MFBIA-determined impedance change POST. A consistent pattern of change was observed over time and across conditions, therefore, we have selected to present the 20 min data. According to the Bland Altman analysis (Figure 1), the change in %BF at 20 min POST was influenced by body mass for SBIA (r ¼ 0.39, P ¼ 0.010), but not for LBIA (r ¼ 0.25, P ¼ 0.100) and MFBIA (r ¼ 0.11, P ¼ 0.481). As shown in Figure 1, %BF increased in the majority of the subjects (LBIA ¼ 88%, SBIA ¼ 95% and MFBIA ¼ 79%); for most subjects, the magnitude of change was o2.0 %BF (LBIA ¼ 100%, SBIA ¼ 66% and MFBIA ¼ 91%). The total sample increases in %BF from baseline to 20 min POST (mean±s.d.) were 0.9±0.6, 1.7±1.1 and 0.8±0.8%BF for LBIA, SBIA and MFBIA, respectively. As shown in Figure 2, the change in impedance at 20 min POST was influenced by body mass for SBIA (r ¼ 0.35, P ¼ 0.022) and MFBIA (r ¼ 0.33, P ¼ 0.028), but not LBIA (r ¼ 0.24, P ¼ 0.119). In addition, 20 min after eating, impedance increased in the majority of the subjects (LBIA ¼ 72%, SBIA ¼ 88% and MFBIA ¼ 93%) regardless of the analyzer used for the assessment. Furthermore, the total sample increases in impedance from baseline to 20 min POST (mean±s.d.) were 6.0±9.5, 17.9±17.3 and 27.1±19.2 O for LBIA, SBIA and MFBIA, respectively. DISCUSSION Eating a meal before using contact-electrode BIA technology to assess body composition resulted in an artificial increase in the %BF estimate. The %BF value, as calculated by the BIA analyzers, is derived from proprietary equations combining body mass and Table 2. Leg-to-leg bioelectrical impedance analysis measurements during the two treatment conditions Baseline 20 min D 40 min D 60 min D Body fat (%) CON 20.3±9.3 20.3±9.2 0.03±0.5 20.4±9.4 0.16±0.6 20.6±9.3 a,b 0.3±0.7 EXP 19.8±9.5 20.7±9.3 a 0.9±0.6 c 20.7±9.4 a 0.9±0.7 c 20.8±9.5 a 0.9±0.7 c Body mass (kg) CON 71.2±16.0 71.1±16.0 0.03±0.05 71.1±16.0 a 0.06±0.05 71.1±16.0 a 0.06±0.06 EXP 70.9±15.9 71.8±16.0 a 0.8±0.2 c 71.7±16.0 a 0.7±0.2 c 71.7±16.0 a 0.7±0.2 c Impedance (ohms) CON 523.0±58.2 523.7±58.9 0.7±7.4 526.5±58.8 3.5±10.1 529.0±61.0 a,b 6.0±11.7 EXP 518.0±60.0 524.0±60.4 a 6.0±9.5 c 524.6±61.3 a 6.6±10.2 524.3±60.8 a 6.3±11.2 Abbreviations: D, mean change from baseline; CON, control trial, EXP, experimental trial. Po0.05. All data presented as mean±s.d. a Significantly different from baseline. b Significantly different from 20 min. c Significantly different from CON at same time point. Table 3. Segmental bioelectrical impedance analysis measurements during the two treatment conditions Baseline 20 min D 40 min D 60 min D Body fat (%) CON 20.7±8.4 20.6±8.3 0.04±1.0 20.9±8.5 0.3±0.6 21.1±8.3 ab 0.5±0.6 EXP 19.7±8.7 21.3±8.5 a 1.7±1.1 c 21.4±8.6 a 1.7±1.1 c 21.4±8.5 a 1.7±1.0 c Body mass (kg) CON 71.6±16.1 71.6±16.1 0.05±0.10 71.5±16.0 a 0.08±0.1 71.5±16.0 a 0.08±0.1 EXP 71.4±15.9 72.2±16.0 a 0.8±0.2 c 72.2±16.0 a 0.8±0.2 c 72.1±16.0 a 0.7±0.2 c Impedance (ohms) CON 615.5±92.4 619.1±95.4 3.6±7.9 622.0±95.8 a 6.4±8.6 624.3±94.3 a 8.8±10.4 EXP 602.8±90.6 620.8±96.2 a 17.9±17.3 c 621.8±97.3 a 18.9±15.0 c 621.3±96.6 a 18.5±16.0 c Abbreviations: D, mean change from baseline; CON, control trial; EXP, experimental trial. Po0.05. All data presented as mean±s.d. a Significantly different from baseline. b Significantly different from 20 min c Significantly different from CON at same time point. European Journal of Clinical Nutrition (2013) 950 955 & 2013 Macmillan Publishers Limited

Table 4. Multi-frequency bioelectrical impedance analysis measurements during the two treatment conditions Baseline 20 min D 40 min D 60 min D 953 Body fat (%) CON 21.6±8.7 22.0±8.8 0.3±0.7 22.3±9.1 a 0.7±1.6 22.3±8.9 a 0.7±0.9 EXP 21.6±8.7 22.4±8.7 a 0.8±0.8 b 22.3±8.8 a 0.7±0.9 22.3±8.8 a 0.6±0.9 Body mass (kg) CON 72.2±16.0 72.1±16.0 0.07±0.2 72.1±16.0 0.06±0.3 72.1±16.0 0.07±0.3 EXP 72.0±15.9 72.7±16.0 a,b 0.7±0.3 b 72.7±16.0 a,b 0.7±0.3 b 72.7±16.0 a,b 0.7±0.3 b Impedance (5 khz ohms) CON 1348.1±176.7 1360.1±183.2 a 12.0±17.0 1364.7±188.6 a 16.6±37.6 1377.2±186.8 a,c 29.2±25.8 EXP 1333.8±183.4 1369.7±191.3 a 36.0±24.1 b 1369.7±190.0 a 35.9±24.7 b 1368.4±190.2 a 34.6±26.4 Impedance (50 khz ohms) CON 1179.2±175.6 1188.4±181.2 a 9.2±13.0 1191.5±184.5 a 12.3±26.6 1201.3±184.2 a,c 22.4±20.1 EXP 1169.0±182.0 1196.2±188.1 a 27.1±19.2 b 1194.7±187.3 a 25.5±20.0 b 1194.0±186.5 a 24.8±21.5 Impedance (500 khz ohms) CON 1008.7±158.9 1017.6±164.3 a 8.9±11.4 1019.9±167.1 a 11.2±22.7 1029.3±167.4 a,c 20.6±17.3 EXP 1000.4±165.4 1024.7±170.3 a 24.2±16.3 b 1023.8±169.5 a 23.3±17.8 b 1022.8±168.3 a 22.3±18.7 Abbreviations: D, mean change from baseline; CON, control trial; EXP, experimental trial. Po0.05. All data presented as mean±s.d. a Significantly different from baseline. b Significantly different from CON at same time point. c Significantly different from 20 and 40 min. impedance measurements with height, gender and age information. As such, any %BF alterations are a resultant of changes in the body mass and/or impedance measurement. Overall, the consumption of the meal increased both body mass and impedance in this study, thus resulting in a slight %BF overestimation. The contribution that each variable had on the overall magnitude of %BF change cannot be specifically determined without access to the manufacturer s proprietary equations. Most importantly, meal-induced elevations in body mass and impedance are interpreted by contact-electrode BIA analyzer s preprogrammed equations as higher %BF readings. To our knowledge, previous research examining the effect of eating a meal on contact-electrode BIA technology is non-existent; however, data does exist using the traditional gel-electrode BIA procedure. 12,13 Previously, the influence of eating a breakfast meal on BIA-determined body composition measurements was examined in young adults 12 and elderly persons. 13 Gallagher et al. 12 reported a significant decrease in body impedance 2 h after eating, while, Vilaca et al. 13 found no difference in fat-free mass or fat mass 1 h after the ingestion of 500 ml of orange juice and a roll with butter. In addition to the operational differences between gel-electrode and contact-electrode BIA technology, a possible explanation for the different results between the aforementioned and present study may include differences in the meal consumed. More specifically, the volume and energy content (3847±900 kj) of the present meal far exceed that of the breakfast meals consumed previously, 2300 12 and 1250 kj, 13 respectively. As expected, the consumption of a larger meal resulted in a comparatively larger increase in body mass and impedance, two variables that are directly utilized to calculate the %BF estimate. Previously, we examined the effect of acute fluid consumption on contact-electrode BIA measures using a LBIA and SBIA analyzer in adults. 9,10 In two separate studies, we found significant increases in %BF estimates (LBIA ¼ 0.5% and SBIA ¼ 1.0%, respectively) after the consumption of a 591 ml beverage. 9,10 The present meal-induced %BF changes (LBIA ¼ 0.9%, SBIA ¼ 1.7% and MFBIA ¼ 0.8%) found after eating compare favorably with the aforementioned studies. However, as demonstrated by these data, the addition of food resulted in a comparatively larger increase in %BF than previously observed after fluid ingestion. Most likely, this finding was a resultant of the higher body mass gain (800 vs 500 g) and larger impedance increases (LBIA 6 O vs no change; SBIA 18 vs 14 O) observed presently in comparison with the previous beverage-only studies. 9,10 Data from the CON trial were examined to explore for the normal hourly variability in body composition measures. When using the LBIA, SBIA and MFBIA analyzers for the assessment, a consistent body mass reduction (mean range ¼ 60 80 g) and impedance increase (mean range ¼ 4 9 O) were observed over time, thus resulting in a small, but statistically significant, increase in %BF (mean range ¼ 0.3 0.7%) at 60 min. For LBIA and SBIA, the increase in %BF observed during the CON trial at the 60 min time point was significantly less than the meal-induced changes observed during the EXP trial at the same time point. However, the MFBIA analyzer produced similar %BF increases during the CON and EXP trials at both the 40 min (0.7 and 0.7%) and 60 min (0.7 and 0.6%) assessment points. As such, when using MFBIA to assess body composition after eating, the %BF increases observed over time appear to be no different than those that would have occurred had the subject been assessed in a fasting state. As shown in Tables 2 4, impedance and %BF increased significantly above baseline in both the EXP and CON trials; however, the pattern of change differed upon whether food was consumed or not. In agreement with our previous work on fluid consumption, 9,10 the impedance and the %BF changes after eating had two distinct phases: (1) there was a rapid initial increase in both variables at 20 min after eating followed by (2) a leveling off, which lasted for the 60-min testing period. As suggested by Gomez et al., 17 a redistribution of blood from the periphery to the core in response to a volume of food and fluid entering the stomach and gastrointestinal tract may have contributed to the initial impedance elevation. Thereafter, the absorption of the beverages and subsequent dilution of body fluid electrolytes may have caused impedance to remain significantly elevated above baseline for up to 60 min post consumption. 17 During the CON trial, impedance and %BF also increased over time; however, the increase was gradual as opposed to the rapid meal-induced elevation observed 20 min after eating. Whole-body impedance has been previously reported to increase if the body water content in the upper and lower limbs decreases due to a shift in body water to the trunk. 18,19 Our subjects walked to the laboratory for testing and then sat quietly for 60 min, excluding the brief time (o1 min) required to stand on the analyzer for the assessment. A redistribution of body water from the active skeletal & 2013 Macmillan Publishers Limited European Journal of Clinical Nutrition (2013) 950 955

954 Figure 1. Scatter plots exploring individual differences in %BF by BIA analyzer (a ¼ LBIA, b ¼ SBIA and c ¼ MFBIA). The differences between baseline and 20 min %BF is plotted against body mass for the women (K) and men (J). Negative values indicate an increase from baseline. The mean difference is represented by the solid line and the dashed lines represent±2 s.d. from the mean. Figure 2. Scatter plots exploring individual differences in impedance (IMP) by BIA analyzers (a ¼ LBIA, b ¼ SBIA and c ¼ MFBIA). The differences between baseline and 20 min IMP is plotted against body mass for the women (K) and men (J). Negative values indicate an increase from baseline. The mean difference is represented by the solid line and the dashed lines represent±2 s.d. from the mean. muscle in the extremities to the torso may have resulted in the gradual impedance increase over time during the CON trial. Collectively, it is apparent that eating caused impedance and %BF to increase abruptly at 20 min, a response that was not observed in the CON trial. The variables then remained elevated at 40 and 60 min; however, in some cases, the increase was not significantly different than the normal hourly variability observed during CON trial. Despite this similarity between trials at these specific assessment points, it is likely that different mechanisms were responsible for the observed alterations. Although group mean comparisons are often reported and discussed in the literature, these data do not necessarily give accurate estimates for all individuals within the group. As such, the Bland Altman method was used to explore individual variability within the sample. Overall, eating caused an increase in the LBIA, SBIA and MFBIA-determined impedance, and %BF for most subjects. In some instances, the greatest %BF (SBIA) and impedance (SBIA and MFBIA) increases were observed in the lightest subjects, whereas the heavier subjects demonstrated less of a change. All subjects, regardless of body size, consumed similar volumes of food and beverage during the meal. Clearly, the volume of food/beverage had a greater impact on the SBIAdetermined %BF estimates in individuals that had a lower-body mass. When using this technology to assess body composition of populations that vary in body mass, technicians should recognize the larger effect that eating may have on lighter individuals. In summary, the ingestion of a meal before LBIA, SBIA and MFBIA body composition assessment significantly increased %BF estimates (LBIA ¼ 0.9%, SBIA ¼ 1.7%, MFBIA ¼ 0.8%). When one considers the observed meal-induced %BF increases with the European Journal of Clinical Nutrition (2013) 950 955 & 2013 Macmillan Publishers Limited

inherent prediction error of BIA, it is apparent that eating may further reduce the accuracy of this technology. When using BIA analyzers in research or clinical trials, this degree of error may be considered unacceptable. In those instances, we recommend adhering to the fasting guideline in order to avoid a meal-induced elevation in the %BF value. In addition, a relatively small, but statistically significant, %BF increase (0.3 0.7%) was also observed during the CON trial over time. To control for normal hourly variability, we recommend using a consistent testing schedule, preferably conducting the body composition assessment soon after the subject arrives at the testing location to minimize variation over time. Obviously, stringent pretest guidelines do limit the practicality of utilizing these analyzers in clinical and health/fitness settings. For the majority of the subjects, the %BF increase after eating was o2.0 %BF. In certain instances, such as a physical examination where the assessment of body fatness is used to determine health risk, variance of this magnitude may have little practical significance, and therefore controlling food consumption before the assessment may not be essential. CONFLICT OF INTEREST The authors declare no conflict of interest. ACKNOWLEDGEMENTS We gratefully acknowledge all subjects for their participation in this investigation. This investigation was supported by a Small Campus Research Grant from Lock Haven University. REFERENCES 1 National Institutes of Health Technology Assessment Conference (NIHTAC) (1996). Bioelectrial impedance analysis in body composition measurement. Am J Clin Nutr 1996; 64: 524 532. 2 Nunez C, Gallagher D, Visser M, Pi-Sunyer FX, Wang Z, Heymsfield SB. Bioimpedance analysis: evaluation of leg-to-leg system based on pressure contact foot-pad electrodes. Med Sci Sports Exerc 1997; 29: 524 531. 3 Lukaski HC. Assessing regional muscle mass with segmental measurements of bioelectrical impedance in obese women during weight loss. Ann NY Acad Sci 2000; 904: 154 158. 4 Heyward VH, Wagner DR. Applied Body Composition Assessment, 2nd edn. Human Kinetics: Champaign, IL, USA, 2004, pp 87 98. 5 Roche AF, Heymsfield SB, Lohman TG. Human body composition. Human Kinetics: Champaign, IL, USA, 1996, pp 79 108. 6 Deurenberg P, Weststrate JA, Paymans I, van der Kooy K. Factors affecting bioelectrical impedance measurement in humans. Eur J Clin Nutr 1988; 42: 1017 1022. 7 Gomez T, Mole PA, Collins A. Dilution of body fluid electrolytes affects bioelectrical impedance measurements. Sports Med Training Rehab 1993; 4: 291 298. 8 Kushner RF, Gudivaka R, Schoeller DA. Clinical characteristics influencing bioelectrical impedance analysis measurements. Am J Clin Nutr 1996; 64: 423 427. 9 Dixon CB, Lovallo SJ, Andreacci JL, Goss FL. The effect of acute fluid consumption on measures of impedance and percent body fat using leg-to-leg bioelectrical impedance analysis. Eur J Clin Nutr 2006; 60: 142 146. 10 Dixon CB, Ramos L, Fitzgerald E, Reppert D, Andreacci JL. The effect of acute fluid consumption on measures of impedance and percent body fat estimated using segmental bioelectrical impedance analysis. Eur J Clin Nutr 2009; 63: 1115 1122. 11 Dixon CB, Fitzgerald E, Makuta L, Andreacci JL. The effect of acute fluid consumption on measures of impedance and percent body fat estimated using leg-to-leg and segmental bioelectrical impedance analysis in children. Int J Body Comp Res 2012; 10: 21 28. 12 Gallagher M, Walker KZ, O Dea K. The influence of a breakfast meal on the assessment of body composition using bioelectrical impedance. Eur J Clin Nutr 1998; 52: 94 97. 13 Vilac a KH, Ferriolli E, Lima NK, Paula FJ, Moriquti JC. Effect of fluid and food intake on the body composition evaluation of elderly persons. J Nutr Health Aging 2009; 13: 183 186. 14 Costill DL, Saltin B. Factors limiting gastric emptying during rest and exercise. J Appl Physiol 1974; 37: 679 683. 15 Pietrobelli A, Rubiano F, St-Onge M-P, Heymsfield SB. New bioimpedance analysis system: improved phenotyping with whole-body analysis. Eur J Clin Nutr 2004; 58: 1479 1484. 16 Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1: 307 310. 17 Gomez T, Mole PA, Collins A. Dilution of body fluid electrolytes affects bioelectrical impedance measurements. Sports Med Train Rehabil 1993; 4: 291 298. 18 Slinde F, Bark A, Jansson J, Rossander-Hulthen L. Bioelectrical impedance variation in healthy subjects during 12 h in the supine position. Clin Nutr 2003; 22: 153 157. 19 Oshima Y, Shiga T. Within-day variability of whole-body and segmental bioelectrical impedance in a standing position. Eur J Clin Nutr 2006; 60: 938 941. 955 & 2013 Macmillan Publishers Limited European Journal of Clinical Nutrition (2013) 950 955