SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

Objective

To evaluate the precision of the iDXA for total body composition and fat distribution measurements in severely obese patients.

Design and Methods

Sixty-five severely obese participants with a mean age of 46 ± 11years, BMI of 49 ± 6 kg/m2, and a mean body mass of 137.3 ± 20.9 kg took part in this investigation. Two consecutive iDXA scans with repositioning of the total body were conducted for each participant. The coefficient of variation (CV), the root-mean-square (RMS) averages of standard deviations of repeated measurements, the corresponding 95% least significant change, and Intraclass Correlations (ICC) were calculated.

Results

Precision expressed as % CV, for total body bone mineral content, fat free mass, total body fat, total body lean, and % total body fat were 1.08%, 0.94%, 0.90%, 1.00%, 0.79%, respectively. Precision was 1.44% for gynoid fat distribution and 1.64% for android fat (AF) distribution. The ICCs in all DXA measurements were 0.99 with % AF having the lowest at 0.96.

Conclusions

The GE Lunar iDXA™ demonstrated excellent precision for total body composition assessments and is the first study to assess reproducibility in severely obese individuals.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

Dual energy X-ray absorptiometry (DXA) is accepted as a convenient and useful diagnostic tool that provides precise, low-dose radiation assessment of body composition [1]. The assessment of body composition has become increasingly important for research and health professionals in order to quantify tissue changes in response to nutrition, exercise, surgical, and pharmaceutical interventions. There are several methods for assessing body composition including: bioelectrical impedance, skinfold calipers, air displacement plethysmography, and under water weighing, which all have significant limitations when assessing severely obese populations [2].

With advances in DXA technology, it has become an established technique for the measurement of body composition in research and clinical practice. A recent model, the Lunar iDXA™ by GE (GE Healthcare, USA), has been redesigned and can accommodate larger patients who weigh up to 204 kg. In comparison to previous models, the scanning field is wider and the scanner arm is higher with a width of 52 inches and a height of 49 inches. In addition, the iDXA with the accompanying enCORE™ software allows investigators to measure android (AF) and gynoid fat (GF) distribution. The accumulation of fat in the abdominal area and its links to cardiovascular risk and metabolic health highlight the need for precise measurement of this area [3].

With the rise of obesity rates and associated co-morbidities in most developed countries [4], there is an increased focus on obesity research including interventions to help obese individuals manage their weight. First generation DXA scanners had weight limitations that would not allow severely obese patients to be assessed. In addition, when obese individuals who did not fit the scanning area were assessed, there was questionable accuracy of body composition because either tissue was compressed together with straps and/or only half body scans were performed and then combined together [5]. Given the importance of assessing body composition of severely obese patients, the iDXA enables researchers and clinicians to measure individuals who exceed the width parameters of the scanning area. The iDXA with the accompanying enCORE™ software enables technicians to place participants on the scanning bed in such a way that allows for the maximal amount of real body tissue in each scan to be analyzed. In this way, the entire right side and the most of the left side possible will be measured depending on the individual's morphology. The new enCORE software then estimates the missing left side based on the results of the right side.

The short-term precision of the iDXA was recently found to be excellent in a population with BMIs ranging between 16.7 and 42.7 kg/m2 whereby only 1 participant had a BMI over 40 kg/m2 [6]. In order to assess the effectiveness of weight management interventions, there is a clear need to determine the precision of the iDXA in severely obese populations (Class III; BMI > 40 kg/m2). Therefore, the purpose of this investigation was to determine the precision of the GE Lunar iDXA™ for total body composition, including the three compartments of total body bone mineral content (TBBMC), total body fat mass (TBF), and total body lean mass (TBL) in a severely obese population. Furthermore, AF and GF distribution measurements were also assessed for precision error.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

Patients consisted of 65 severely obese (Class III; BMI > 40 kg/m2) adults with a mean age of 46 ± 11 years, BMI of 49.0 ± 6 kg/m2, and average body mass of 137.3 ± 20.9 kg. There were 24 males and 41 females and ethnic group breakdown was: 86% white, 6% Hispanic, and 8% black. Pregnant or breastfeeding women were excluded.

The GE procedure manual has clear guidelines on the placement of individuals within normal weight ranges. In this study, however, participants were severely obese and therefore followed a different procedure than normal weight individuals. Participants were instructed to sit on the side of the scanning bed facing the technician and were asked to turn and place their legs onto the table while remaining in a seated position. The DXA technician then assessed how far to the right side of the scan table the participant should be placed to ensure that the right arm is at the limit on the right side; that there is enough space between the right arm and trunk to ensure that analysis of different regions of interest in the body can be performed with the software. With severely obese individuals, the left arm is placed outside of the scanning area and, when possible, the left side of the trunk is kept within this area. The portion of the participant's body outside of the scan window is estimated from the right side using the mirror imaging software. During the scan, the supine participant's ankles were placed together using the Lunar ankle supports. Two consecutive total body scans with complete repositioning between scans were conducted for each participant. With this population, a total body scan can be uncomfortable, as they are not allowed to move during the entire scan. It was found that a 5 minute rest between scans in the seated position on a chair alleviated the back pain and allowed a second scan to be performed with minimal discomfort. The scans were all performed in Thick Mode™, determined by GE's Lunar software enCORE™, which operates at a slower speed of 80mm/sec compared to 153 mm/sec in the Standard Mode. For this reason, the total body scan takes 13 minutes compared to the Standard 7 minutes which allows for the photon energy to penetrate the larger body size. One technician (TC) performed all scans for the study and ensured that calibration of the DXA was checked and passed using the GE Lunar calibration phantom on a daily basis before each scanning session.

The coefficients of variations (%CV) and square root of the mean of the sum of squares of differences between the two measurements (RMS) were calculated. The least significant change (LSC) at 95% confidence interval and Intraclass Correlations (ICC) to determine the amount of technique variance were calculated using SPSS Ver 20 [7].

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

Table 1 displays the results presented as means, ranges, %CV, and ICCs for each measurement. Precision error was less than 1.08% for all total body measures. Errors were slightly higher for the %AF (1.64%) and %GF (1.44%) regions. The ICCs in all DXA measurements were 0.99 with %AF having the lowest at 0.96. In addition, with this severely obese population and the extra 6 cm of width for the Lunar iDXA scanning area (compared to previous models) it is normal to expect that a large percentage of the participants will exceed the width scanning area. In our study, 51% of the cohort had more than just the left arm outside of the scan area and, therefore, arm, trunk, and legs were estimated from the right side using the software's mirror imaging. When comparing the precision, using an ANOVA, between having just the left arm versus more of the body outside the scan area, no significant difference was found for TBBMC (0.82% vs. 1.13%), TBF (0.72% vs. 0.90%), and TBL (all P's > 0.05). Although, for TBL (0.78% vs. 1.10%), differences approached significance (P = 0.06). These trends were similar when data were adjusted for age and gender.

Table 1. In vivo short-term precision of total body composition measurements and fat distribution in severely obese (BMI > 40 kg/m2) using GE Lunar iDXA (n = 65)
 Measurement 1 Mean (SD)RangeMeasurement 2 Mean (SD)RangeRMS (SD)CV (%)LSC (95% CI)ICC
RMS (SD)CV (%)
  1. SD, standard deviation; TBBMC, total body bone mineral content; FFM, fat-free mass; TBF, total body fat; TBL, total body lean; AF, android fat; GF, gynoid fat; RMS, root mean square; CV, coefficients of variation; CI, confidence interval; ICC, intra-class correlations; LSC, least significant change.

TBBMC (kg)2.992 (0.53)1.60–4.472.986 (0.52)1.60–4.440.0331.080.0912.980.99
FFM (kg)67.02 (12.76)45.01–96.0266.92 (12.69)44.79–94.860.6630.941.8372.620.99
TBF (kg)68.56 (12.97)47.67–106.1968.28 (13.03)47.34–106.470.6650.901.8422.480.99
TBL (kg)64.03 (12.37)42.92–93.1563.93 (12.31)42.67–90.950.6701.001.8572.780.99
%TBF51.73 (5.56)38.8–62.251.64 (5.64)38.5–61.60.4040.791.1182.180.99
%AF59.41 (4.75)46.8–71.159.35 (4.85)46.5–70.10.9691.642.6844.540.96
%GF51.65 (7.59)30.3–63.151.39 (7.75)29.5–63.10.7101.441.9673.980.99

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

An important finding in this study was that the level of precision of the iDXA in a severely obese population was excellent for total body composition measurements, with most %CVs <1%, and all ICCs close to one (0.96-0.99). In a similar study assessing short-term in vivo precision of the iDXA™ in a normal weight population cohort (n = 52) with a mean BMI of 25.8kg/m2 (16.7-42.7 kg/m2) [6], precision for TBF was 0.82% compared to our result of 0.90% in severely obese; and %TBF precision was 0.86% compared to our value of 0.79%. In TBL, precision in normal weight adults was 0.51% [6] compared to our result of 1.00%. The precision error for TBBMC was 1.08% in severely obese participants compared to 0.6% previously reported in normal weights [6]. The previous Prodigy DXA model was determined to have a precision error, with normal weight adults, of 1.3% for TBBMC, 2.7% for %TBF, 2.5% for TBF, and 0.8% for TBL [8]. It has been observed that precision error for TBL was lower compared to TBF [6, 8], however, our study resulted in similar values of precision for both lean and fat tissue with %CVs close to 1. In fact, in a recent review, in-vivo precision values for total body composition measurements in normal weight populations were similar to our results in a severely obese population [1].

Little has been done in assessing the reproducibility of fat distribution measurement by DXA, mainly because in past models the analysis was performed manually. With the iDXA, and accompanying software, the AF and GF distributions are automatically computed. The AF patterning is of clinical interest due to its association with visceral fat and the metabolic syndrome [3]. The iDXA now assesses fat distribution of severely obese individuals and can track changes in visceral adiposity during the course of treatment. We found a good level of precision for %AF of 1.65% which compared favorably to the precision in normal weights of 2.32% [6]. For %GF, precision was found to be 1.44% compared to 0.96% in normal weights [6].

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

This is the first study to determine the precision of new DXA technology in severely obese populations (BMI > 40kg/m2). The GE Lunar iDXA™ provided excellent precision for total body measurements of body composition and fat distribution in severely obese individuals (BMI: 40.2-65 kg/m2). The results of this study suggest that investigators may now precisely assess total body and regional body composition in severely obese individuals using the iDXA scanner.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References