RESEARCH REVIEW: The Super Power of X-ray vision?

What you can’t measure, you can’t manage, the saying goes – so measuring body fat can be a useful tool for PTs managing clients’ fat loss. But how accurate are the available devices?
Review by Dr Mike Climstein PhD & Joe Walsh

Title: Comparison of multi-frequency bio-electrical impedance and DXA on body composition
Authors: Dr’s Wang and colleagues. (School of Public Health, Bejing, China)
Source: Biomedical and Environmental Sciences (2018) 31(1): 72-75. Click HERE to read.

EARN YOUR FREE CEC or CPD for reading this and two other Research Reviews!

Click HERE to login to your Member Portal, then click the ‘My CEC Courses’ tab and select ‘Research Reviews 2018’

Introduction: As an Accredited Exercise Physiologist (AEP), the clear majority of my patients see me for the purpose of undertaking rehabilitative exercise. Unfortunately, there is still not a big interest in preventative exercise prescriptions from AEPs. That being said, as I am located in a medical centre I receive the majority of referrals from general practitioners and specialists, as the patients typically have chronic diseases and conditions. On the bright side, the apparently healthy individuals are being cared for by personal trainers and group fitness instructors.

We recently had an interesting case which we will present as a case study here. Alan (not his real name) is a middle-aged male with a family history that includes type 2 diabetes mellitus (T2dm, mother, father and both brothers), coronary heart disease (father and mother and paternal and maternal grandparents), hypertension (mother, father and brothers) and dyslipidemia (mother, father and brothers). Alan’s medical history includes obesity (BMI 34.1kg/m2), hypertension (resting blood pressure was 146/88mmHg when we tested), dyslipidemia (total cholesterol 6.1mmol, high density lipoprotein 0.9mmol, low density lipoprotein 3.7mmol) and HbA1c (a test for type 2 diabetes mellitus) reading of 6.3%, which classifies as prediabetes. His prescribed medications include Betaloc (anti-hypertensive) and Lipitor (for cholesterol) to which he is compliant.

Alan’s general practitioner referred him to our clinic, citing the reasons for doing so as being to ‘reduce his likelihood of developing T2dm, weight management, blood pressure control and improved lipids’. These referrals are usually very short and to the point. Alan is employed as an accountant (full time), presented with no musculoskeletal injuries (or history of musculoskeletal injuries) and his current physical activity consists only of incidental walking, specifically to and from his car and moving around his house. We devised an individualised exercise prescription for Alan and he went off to his local gym to embark upon it. Upon follow-up two weeks later, Alan presented his workouts, which he was tolerating well, along with a body fat test that had been conducted using a bio-electrical impedance analyser (BIA). In brief, the results showed his body fat to be an athletic 18%! No way in h*ll – so we referred Alan for a DXA scan to see if that would tell a different story… DXA stands for dual-energy X-ray absorptiometry, a process of conducting a full body scan for segmental body composition. This scan found Alan’s body fat to be much higher, at 37.1%. This leads us into this Research Review, in which we look at Dr Wang and his colleagues’ comparison of the accuracy of a bio-electrical impedance analyser (BIA) to that of a DXA scan.

BIA’s have been around since the mid 1980’s, and admittedly there is a wide range of units available designed for everything from home use to application in the clinical/medical setting. One would therefore expect differing degrees of accuracy between these units. BIA machines have lower accuracy compared to units using four (one per hand and foot) or more electrodes, and poorly insulated wires have been shown to be sensitive to room temperature, thereby affecting the results.

DXA, on the other hand, is considered to be the gold standard (i.e. the diagnostic test that is considered to be the most accurate) for determining body composition and is especially beneficial as it provides segmental results for lean mass (muscle) and fat mass (adipose). In DXA scans, two different energy levels of X-ray pass through the body and measure fat, muscle and bone levels. It also provides specific details about where the fat and muscle are situated on the body.

The authors compared a BIA to DXA in determining body composition in a large cohort (749, males and females) of obese adults. All participants were required to have a body mass index (BMI) > 28kg/m2 and be aged 25-55 years of age. Participants had both BIA and DXA measurements taken following an 8-hour (or more) fast and according to manufacturers’ requirements.

Results: The per cent body fat estimated by the BIA was significantly lower (-4.33%) as compared to the DXA scanner in the males and significantly higher (+0.5%) in the females, which was a real problem. Given these findings, the researchers then created correction equations to improve the accuracy of the BIA device.

The authors concluded that body composition is widely used in clinics, weight loss programs and other health related fields to assess the risk factor of obesity. DXA, as the gold standard, is generally more expensive than BIA assessment and typically not available in most health and fitness facilities.

Pros: This is a good practical study, between BIA and DXA: what was problematic in the findings was that the differences seen were gender specific. Congratulations to the authors for then developing correction equations to improve the accuracy of the BIA equations. For example, the correction equations for per cent body fat (specific to the BIA device used in this study only) for males was 14.098 + 0.694 x % fat (BIA) and females was -3.263 + 1.019 x % fat (BIA).

There is no doubt the technology of the BIAs continues to improve and it is useful to help individuals who are attempting to monitor their adiposity or lean mass (muscle) levels.

It should be noted that when using DXA the subject is irradiated. In the case of those suffering from serious health concerns due to excess weight, suspected poor bone mineral density or other conditions (such as use of medications that might impact bone mineral density), then the use of DXA measurements is supported. This is similar to getting an X-ray to check for a broken bone, the potential benefits gained from correct assessment can greatly outweigh any negatives from a small dose of radiation. While the dose of radiation from DXA is about the same as the level you’d be exposed to by taking a short flight, it still exists. With respect to radiation, there is no threshold dose: any radiation can have some negative effect. It therefore should be cautioned against unnecessary or repeated DXA use when there is not an apparent health benefit. This benefit can arguably be as simple as gaining extra motivation to exercise, however outside of clinical conditions measurement via BIA may be sufficient. As subjects using BIAs are not subjected to the same radiation level as that involved in DXA, improving the accuracy of BIA (as per the authors’ work) is to be commended.

Cons: Unfortunately, BIAs are based on predictive modelling using sample population averages and are sensitive to the different individual characteristics of people. Factors as diverse as hydration levels (body fat overestimated due to dehydration, by as much as 5kg in some research), skin roughness, and whether the subject has recently eaten or exercised can dramatically affect the results. Although the participants in this study were asked to fast for a minimum of 8 hours, there was no mention of avoiding strenuous exercise prior to the testing to ensure they had a normal state of body water content (‘euhydration’ for those of you looking to expand your medical vocabulary!). In fact, due to reduced electrical resistance in the body after exercise, other studies (Khaled, 1988) have shown fat mass as being underestimated by up to 12kg when BIAs have been used after exercise. This finding is particularly noteworthy for those PTs using BIA in the gym, after a client’s training session. Additionally, as females were included as participants there was no mention of whether they were tested during their menses, which could also dramatically affect the accuracy of the results.

If ongoing DXA scans are not a financially viable option for clients and you want to choose the ‘next best thing’ in terms of a BIA, the best thing that you can do is to get yourself a DXA scan, and to then experiment with a number of BIA units to see which one’s results come closest those you got with the DXA. Then, one hour later (having consumed nothing, expelled nothing and not exercised) reassess and see which unit gives the (near) exact same reading as the first time on the BIA monitor.

In fact, I was once asked by a fitness facility I worked with to advise them on their potential purchase of a BIA unit. I recommended that they ask the sales rep to loan the unit to them for 4 hours so that they could carry out the exact same process outlined in the paragraph above with a number of club members of varying body types, from lean to overweight. The rep became defensive at this request and questioned the motives behind it, before refusing to loan the device. This, I assured the facility management, was not the behaviour of someone that was confident in the reliability of his wares. They took my advice on board and did not purchase that particular unit.


Dr Mike Climstein, PhD FASMF FACSM FAAESS AEP is one of Australia’s leading Accredited Exercise Physiologists. He is a faculty member inClinical Exercise Physiology, Sport & Exercise Science at Southern Cross University (Gold Coast). michael.climstein@scu.edu.au

Joe Walsh, MSc is a sport and exercise scientist. As well as working for Charles Darwin and Bond Universities, he is a director of Fitness Clinic in Five Dock, Sydney. fitnessclinic.com.au