Technical Issues
Long-Term Precision of Dual-Energy X-ray Absorptiometry Body Composition Measurements and Association With Their Covariates

https://doi.org/10.1016/j.jocd.2013.09.010Get rights and content

Abstract

Few studies have described the long-term repeatability of dual-energy X-ray absorptiometry scans. Even fewer studies have been performed with enough participants to identify possible precision covariates such as sex, age, and body mass index (BMI). Our objective was to investigate the long-term repeatability of both total and subregional body composition measurements and their associations with covariates in a large sample. Two valid whole-body dual-energy X-ray absorptiometry scans were available for 609 participants in the National Health and Nutrition Examination Survey 2000–2002. Participants with scan-quality issues were excluded. Participants varied in race and ethnicity, sex, age (mean 38.8 ± 17.5; range 16–69 yr), and BMI (mean, 26.9 ± 5.2; range 14.1–43.5 kg/m2). The length of time between scans ranged from 3 to 51 days (mean, 18.7 ± 8.4). Precision error estimates for total body measures (bone mineral density, bone mineral content, lean mass, total mass, fat mass, and percent body fat) were calculated as root mean square percent coefficients of variation and standard deviations. The average root mean square percent coefficients of variation and root mean square standard deviations of the precision error for total body variables were 1.12 and 0.01 g/cm2 for bone mineral density, 1.14 and 27.3 g for bone mineral content, 1.97 and 505 g for fat mass, 1.46 and 760 g for lean mass, 1.10 and 858 g for total mass, and 1.80 and 0.59 for percent body fat. In general, only fat and lean masses were impacted by participant and scan qualities (obesity category, sex, the magnitude of the body composition variables, and time between scans). We conclude that long-term precision error values are impacted by BMI, and sex. Our long-term precision error estimates may be more suitable than short-term precision for calculating least significant change and monitoring time intervals.

Introduction

Total body and subregional bone mineral density (BMD) and bone mineral content (BMC), as well as soft-tissue measurements such as fat mass and lean mass, can be measured with dual-energy X-ray absorptiometry (DXA) (1). DXA is a safe, fast, and efficient method for bone and body composition assessment and thus its use has become increasingly popular in both clinical and research settings. DXA is most commonly used for measuring BMD, and most of the literature on the characteristics of DXA systems is related to bone measures. The precision of BMD and BMC measures is affected by body size (2), bone density, and by system make and model 3, 4. However, much less is known about the accuracy and precision of total body soft-tissue measures.

To determine whether a change between 2 measurements is statistically significant, the precision error of the measurements must be known. The International Society for Clinical Densitometry states in its position paper on precision (5) that precision analysis is performed by calculating the root mean square standard deviation (RMS-SD) as an absolute measure in the units of the measure, and the root mean square percent coefficients of variation (RMS-%CV) as an expression of precision error in percent relative to the mean value of the population. The manufacturer's precision error values should not be used because of differences among study populations, technologist skill, and confidence intervals used.

Although there are many benefits to in the use of DXA to assess bone health and body composition, maintaining precision between repeated scans is a challenge. Repeatability is important in maintaining the consistency of DXA measurements in longitudinal studies. It is recommended that consecutive examinations be taken on the same system or a system from the same manufacturer to allow results to be comparable over time (6). Previously, Lohman et al (7) reported excellent precision for total body and subregional measurements. However, this was a short-term study in which repeat scans were acquired on the same day on an all-male population. Long-term precision errors for bone variables measured in spine and hip scans 8, 9, based on scans that have been repeated after several days have passed, are larger than short-term estimates, which is due to differences in scan acquisition caused by patient repositioning problems and scanner calibration stability. Little is known about the relationship of sex, body mass index (BMI), age, and measurement magnitude to long-term body composition precision error values. The purpose of this study was to use repeated scan data collected in the National Health and Nutrition Examination Survey (NHANES) to estimate total body and subregional body composition precision measures and to determine their associations with potential covariates.

Section snippets

Participants

NHANES is a continuous program designed to assess the health and nutrition status of a representative sample of noninstitutionalized adults and children in the United States (10). Interviews as well as physical examinations, including DXA examinations, are used to obtain demographic, socioeconomic, dietary, and health-related information for different population groups. During the survey years in which replicate DXA scans were collected (2000–2002), African Americans, Mexican Americans,

Results

Of the 609 participants who were eligible for the present study, 8 participants were excluded because of files from either the primary or secondary visit that had been corrupted by the time the study was conducted. Selected characteristics of the study population are shown in Table 1. The population included both sexes and was diverse in race and Hispanic origin; age ranged from 16 to 69 yr. Mean length of time between scans was approximately 19 d, but it varied between 3 and 51 d (Table 1).

Discussion

To our knowledge, this study based on 3 yr of NHANES data is one of the largest long-term precision studies for whole-body DXA available to date. The long-term precision error values observed in this study are somewhat worse than previous reports of short-term precision error (discussed further below) but were generally stable over the time difference studied. In fully adjusted models, several factors were significantly associated with the precision error of total-body and subregional DXA

Acknowledgments

The whole-body scans for the study were provided through contract no. 200-2005-11219 with the Centers for Disease Control and Prevention. The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.

References (21)

There are more references available in the full text version of this article.

Cited by (0)

View full text