Type 2 diabetes: bioelectrical impedance analysis does not replace DXA for assessing muscle mass

Cheaper and more accessible, can bioelectrical impedance analysis replace dual-energy X-ray absorptiometry, or DXA, for assessing muscle mass in adults with type 2 diabetes? A meta-analysis of seven studies concludes that the two methods are not interchangeable. Their average results are similar, but their individual estimates diverge, with low to very low certainty of evidence.

Comparison of bioelectrical impedance analysis and DXA for assessing body composition.
Gadgets & TechEditorial illustration, not study data.

Two methods compared in 646 adults

The review brings together seven studies conducted in six countries, involving 646 participants with type 2 diabetes. Each measured muscle mass using bioelectrical impedance analysis, known as BIA, and DXA. The researchers primarily examined agreement between measurements, then their ability to classify a person as having low muscle mass.

This distinction matters: two devices can produce strongly correlated results without providing values close enough to guide the same decision for a given person.

A small average difference masks large discrepancies

For whole-body fat-free mass, assessed in three studies, the mean difference between DXA and BIA was +0.82 kg. Its 95% confidence interval ranged from −0.69 to +2.34 kg: no clear average difference emerged. This measurement concerns fat-free mass, not exclusively muscle.

But the individual limits of agreement ranged from −6.4 to +8.0 kg. In other words, a reassuring average at the group level does not guarantee similar estimates for each participant. Limb lean mass indices also showed substantial variability across studies.

The correlation between estimates reached 0.90. The authors stress that this describes an association between the results, not their interchangeability.

Classification can change depending on the device

In three studies, the proportion of participants classified as having low muscle mass was 3.6 percentage points higher with BIA. However, the confidence interval ranged from −8.6 to +15.7 percentage points: this estimate remains uncertain and does not guarantee that the same people are identified.

The authors consider that standardized BIA can support an initial assessment when DXA is unavailable. However, they recommend confirmation with DXA if the result would change clinical management, particularly near a diagnostic threshold. This conclusion concerns adults with type 2 diabetes, not everyone who uses bioelectrical impedance scales.

Limitations to keep in mind

The certainty of the evidence is low to very low, with substantial heterogeneity in limb lean mass indices. Only the PubMed abstract was consulted, not the full text. Funding and conflicts of interest were not checked.

What we can take away

Similar average results do not make two methods interchangeable. In adults with type 2 diabetes, bioelectrical impedance analysis can contribute to an initial assessment, but it does not replace DXA when a decision depends on the result.

Sources and provenance

Partager cet article

X f in

An error or a source to report? Contact the editorial team.