When athletes, coaches, performance staff, and doctors have their first experience measuring body composition with dual X-ray absorptiometry (DXA), they are commonly surprised about the different numbers they receive compared to other body composition methods. Within this blog we will outline:
- Why the values for DXA are higher than other body composition methods
- How these differences can be explained so everyone is on the same page
Why are DXA values higher than other body composition methods?
Body composition is describing the type and distribution of body mass for each person. Depending on the method used to measure body composition these masses are a combination of fat mass, fat-free mass, bone mass, lean skeletal mass, and water. To measure body composition, a device must divide the body into two or more of these components. The simplest model is what is typically called a two-component model of body composition. In this model, the body is divided into fat mass and fat-free mass components. This means that bone, muscle, organs, and water are put together to form the fat-free mass component and fat is left over as the other component.
Two-component methods make assumptions about bone mass. We know that bone mass varies greatly between individuals and can change through our lifespan. When we are young the density of bone is lower than the density of our bone as young adults. Bone density also starts to decline when we get to middle age. The two-component model also assumes that everyone has the same bone density, which we now know is not correct. Bone density not only varies by age (Jones et al., 1994; Ensrud et al., 1995), but also by different racial groups (Cauley et al., 2005; George et al., 2003). All devices that use the two-component model (i.e., skinfolds, air displacement, bioelectrical impedance, underwater weighing, etc.) assume bone density does not change and therefore have this error built into their determination of body composition. Additionally, the fat mass and fat-free mass are not directly measured by two-component methods, but are calculated from equations or algorithms. The accuracy of the measurement of body composition is dependant on the population in which these equations and algorithms were developed. Differences from the population being tested and the population that the equations or algorithms are based upon may lead to additional errors in the accuracy of the estimation of body composition.
A more complex model of body composition is a three-component model. This model places an individual’s mass into three components: fat, lean skeletal and bone masses. DXA directly measures bone density and allows you to separate body composition into three components. This significantly reduces the margin of error because DXA doesn’t calculate body fat percentage from body density equations or require extra calculations that use non-specific equations like two-component methods. More importantly, DXA determines lean skeletal, fat and bone masses in various regions of the body versus assuming body composition is the same throughout the entire body as is done in the two-component method. Therefore, by determining regional body compositions as well as determining lean skeletal, bone and fat masses rather than making assumptions about bone and muscle, the percent fat is a little higher than two-component methods of determining body composition.
Accuracy and precision of body composition methods
Finally, a couple of other factors to consider regarding the different methods of determining body composition are the accuracy and precision (reliability) of the measurement. The image below demonstrates the differences between accuracy and precision. Generally speaking, DXA is more accurate and precise compared to two-component methods. In particular, when dealing with athletes with non-traditional body types (i.e., football, basketball, etc), the accuracy and precision of two-component methods tends to get worse, whereas DXA remains similar (Raymond et al. 2018).
Accuracy can be determined by examining the standard error of the measurement. For body fat the standard error of the measurement for DXA is 2.5-3.5% while the other methods utilizing a two-component model (i.e., skinfolds, underwater weighing, air displacement or bioelectrical impedance) are between 3.5-6% (Gatterer et al., 2017). We have observed in our studies (Raymond et al. 2018), that the differences can be upwards of 10-20 pounds in some cases. Thus, it is significant that DXA has greater accuracy than two-component methods of measuring body composition.
Precision is determined by the coefficient of variation between measurements. A measurement is more precise if the limits on the coefficient of variation are low and narrow (see figure above). Precision, both when tracking an individual and comparing multiple individuals, is critical if you plan to compare athletes. For example, you could have two athletes that are 20% body fat on DXA, but using a two-component method one is 12% and the other one is 17%. The decisions made with these athletes from a training or nutrition perspective may differ based on the results of the two-component method, but the more accurate and precise DXA methodology indicates the fat mass is actually the same.
How do I communicate these differences to athletes, coaches and other performance staff that are used to seeing lower numbers?
When we started working with athletes we knew this would be one of the biggest challenges. The issues include a negative connotation with percent body fat and explaining to coaches and athletes what these numbers mean and why they are different. If individuals are used to seeing values from skinfolds, or another two-component method, there can be a little bit of “sticker shock” when discussing the body composition results from a DXA. One analogy used to explain this is the differences between hand-timed forties to electronic-timed forties. When teams first made this switch, their athletes didn’t suddenly get slower (~.2 higher electronic times), just as when body composition is measured by DXA athletes don’t suddenly gain more fat. It’s likely your fastest athletes are still the fastest, and it's also likely athletes with the lowest percent body fat, still have the lowest percent body fat. The difference is in the method, and the values are all relative. The goal should be to have accurate and reliable information, which can then be used to both track an individual athlete over time as well as compare athletes to one another.
Finally, when reviewing the results, it is good to use ranges rather than averages. These ranges should be based on a set percentage for that position. For example, 50% of all our defensive backs have a percent body fat between 12-15%. Individual variation may affect the limits of each player’s body type and 15% (height 74 inches) for one athlete may be comparable to 12% for another athlete (height 68 inches). Understanding the variability within position rather than focusing on positional “norms” provides an individual focus on body composition that stresses an optimal body type for that athlete and not a norm that may be unachievable.
When measuring body composition be sure to understand exactly how the numbers are generated. Accuracy and precision are critical pieces to understanding the differences between methods. More importantly, individual variation can affect the accuracy and precision of most methods, particularly, two component methods based on assumptions and reference populations. When communicating the results of DXA, use positional ranges in order to demonstrate that the values are in-line (or in some cases not) with previous measurements done with other methods.
Cauley JA, Lui L-Y, Ensrud KE, Zmuda JM, Stone KL, Hochberg MC, et al. Bone mineral density and the risk of incident non spinal fractures in black and white women. JAMA. 293:2102–2108, 2005.
Ensrud KE, Palermo L, Black DM, et al. Hip and calcaneal bone loss increase with advancing age: longitudinal results from the Study of Osteoporotic Fractures. J Bone Miner Res 10:1778–1787, 1995.
Gatterer H, Schenk K, Burtscher M. Assessment of human body composition: methods and limitations. In Lukaski HC (ed.), Body Composition: Health and Performance in Exercise and Sport. Boca Raton, FL: CRC Press, pp. 13-26, 2017.
George A, Tracy JK, Meyer WA, Flores RH, Wilson PD, Hochberg MC. Racial differences in bone mineral density in older men. J Bone Miner Res. 18:2238–2244, 2003.
Jones G, Nguyen T, Sambrook P, Kelley PJ, Eisman JA. Progressive loss of bone in the femoral neck in elderly people: longitudinal findings from the Dubbo Osteoporosis Epidemiology Study. Br Med J. 309:691–695, 1994.
Raymond, CJ. Dengel, DR, Bosch, TA. Total And Segmental Body Composition Examination In Collegiate Football Players Using Multifrequency BIA And DXA. JSCR. 32(3):772-782, 2018.
About the Author: Tyler Bosch, PhD is a Research Scientist in the College of Education and Human Development at the University of Minnesota, and is a co-founder of Dexalytics.
About the Author: Donald Dengel, Ph.D., is a Professor in the School of Kinesiology at the University of Minnesota and is a co-founder of Dexalytics. He serves as the Director of the Laboratory of Integrative Human Physiology, which provides clinical vascular, metabolic, exercise and body composition testing for researchers across the University of Minnesota.