📄 medicine

Agreement between bioelectrical impedance analysis and dual-energy X-ray absorptiometry in Parkinson’s disease and controls

While bioelectrical impedance analysis (BIA) demonstrates strong correlation and consistency with dual-energy X-ray absorptiometry (DXA) for assessing body composition in Parkinson's disease, its systematic underestimation of lean mass and wide limits of agreement preclude its use as a direct substitute for DXA in individual clinical settings, though it remains valuable for population-level screening.

Original authors: Wêndel de Oliveira, Danielle Lima, Walter Junior, João de Luna, Antônio Negreiros, Fábia Lopes, Antonio Junior, Adriano Lopes, Anna Fernandes, Pedro Neto, Heitor Virgínio, Paulo Gonçalves, Gabriel de

Published 2026-07-07

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Wêndel de Oliveira, Danielle Lima, Walter Junior, João de Luna, Antônio Negreiros, Fábia Lopes, Antonio Junior, Adriano Lopes, Anna Fernandes, Pedro Neto, Heitor Virgínio, Paulo Gonçalves, Gabriel de Britto, Antônio de Oliveira, Catarina D'Alva, Jarbas Filho, Virgínia Fernandes, Renan Junior

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine your body is a complex house. To understand its condition, you need to know how much of it is "furniture" (muscle and lean tissue) and how much is "storage boxes" (fat).

This study is like a head-to-head comparison between two different ways of measuring the contents of that house: a high-tech, expensive scanner (called DXA) and a cheap, easy-to-use scale (called BIA). The researchers wanted to see if the cheap scale could be trusted to give the same answers as the expensive scanner, specifically for people with Parkinson's disease and for healthy people.

Here is the breakdown of their findings in simple terms:

The Two Measuring Tools

The Gold Standard (DXA): Think of this as a super-precise 3D X-ray scanner. It's the "referee" that everyone trusts to get the exact truth about how much muscle and fat you have. However, it's expensive, hard to find, and not something you can keep in your doctor's office for a quick check-up.
The Everyday Tool (BIA): This is the bioelectrical impedance scale. You step on it, and it sends a tiny, harmless electrical signal through your body. Since muscle conducts electricity well and fat doesn't, the machine guesses your body composition based on how hard the electricity had to push through. It's cheap, fast, and available almost everywhere.

The Experiment

The researchers gathered 190 people: 95 with Parkinson's disease and 95 healthy controls. Everyone stepped on the "cheap scale" (BIA) and then got scanned by the "super scanner" (DXA). They compared the results to see if the two tools agreed.

What They Found

1. The "Trend" Matched Perfectly
When they looked at the data, the cheap scale and the super scanner were like two friends walking in step. If the super scanner said a person had a lot of muscle, the cheap scale also said they had a lot of muscle. If the super scanner said someone had high fat, the cheap scale agreed. They were very consistent in ranking people from "lean" to "heavy."

2. The "Exact Numbers" Were Off
Here is the catch: While they agreed on the ranking, they didn't agree on the exact numbers.

Fat: The two tools were almost identical when measuring fat. They gave very similar numbers.
Muscle (Lean Mass): This is where the cheap scale got it wrong. It consistently underestimated how much muscle people had. It told the patients they had about 1 to 1.25 kg (2.5 to 3 lbs) less muscle than the super scanner actually found. It did this for both the Parkinson's group and the healthy group.

3. The "Margin of Error" Was Too Wide
The researchers used a special graph (Bland-Altman) to see how much the numbers could vary. They found that while the average difference was small, the "wiggle room" was huge.

Analogy: Imagine two weather forecasters. One says it will be 70°F, and the other says 72°F. They are close. But if the second forecaster says, "It could be anywhere between 60°F and 85°F," you can't trust their specific prediction for your day, even if they are right on average for the whole city.
Because of this wide "wiggle room," you cannot swap one tool for the other if you need to know the exact muscle mass of a single individual.

Why Did the Cheap Scale Get the Muscle Wrong?

The paper suggests this isn't because Parkinson's disease confuses the machine. The healthy people had the same problem.

The Shape Problem: The cheap scale assumes your body is a perfect cylinder (like a soda can). But as we age, or if we have Parkinson's, our bodies change shape (more fat in the middle, less in the limbs). The electricity gets confused by this shape change.
The "Hidden" Fat: As we age, fat can sneak inside the muscles (like oil soaking into a sponge). The cheap scale thinks this "oily sponge" is less conductive than pure muscle, so it guesses there is less muscle than there actually is.
The Recipe: The machine uses a "recipe" (math formula) built on healthy, young, white people. When it tries to apply that same recipe to older people or people with Parkinson's, the math doesn't quite fit.

The Bottom Line

Can you use the cheap scale? Yes, but with caution. It is great for looking at a whole group of people (like a population study) or for a quick screening to see if someone is generally losing weight.
Can you use it to diagnose a specific person? No. Because the cheap scale consistently underestimates muscle and has a wide margin of error, you cannot use it to make a precise medical diagnosis (like confirming sarcopenia/muscle loss) for an individual patient. You still need the "super scanner" (DXA) for that.
Does Parkinson's break the machine? No. The study found that the cheap scale worked just as "badly" (or well) for people with Parkinson's as it did for healthy people. The disease itself didn't mess up the results; the limitations of the technology did.

In short: The cheap scale is a good "rough draft" of your body's story, but if you need the "final, polished version" with exact numbers, you still need the expensive scanner.

Technical Summary: Agreement between Bioelectrical Impedance Analysis and Dual-Energy X-ray Absorptiometry in Parkinson's Disease and Controls

Problem Statement
Body composition alterations, including reductions in fat and muscle mass, are prevalent in Parkinson's disease (PD) and are linked to adverse clinical outcomes such as physical decline, falls, and mortality. While Dual-Energy X-ray Absorptiometry (DXA) is the reference standard for assessing body composition, its limited availability restricts routine clinical application. Bioelectrical Impedance Analysis (BIA) offers a low-cost, accessible alternative; however, its accuracy and interchangeability with DXA in the specific context of PD remain under-investigated. This study addresses the need to evaluate whether BIA can reliably replace or supplement DXA for body composition assessment in individuals with PD compared to healthy controls.

Methodology
This was an observational, cross-sectional study conducted at the Universidade Federal do Ceará, Brazil, involving 190 participants: 95 individuals with PD (stages 1–3 on the Hoehn and Yahr scale) and 95 age-matched controls.

Assessment Tools: Participants underwent body composition assessment using two methods:
1. DXA: The reference standard (Lunar Prodigy Advance), measuring total fat mass (FM), FM percentage, lean mass, appendicular lean mass (ALM), and appendicular lean mass index (ALMI).
2. BIA: A multifrequency segmental device (InBody 770) measuring the same variables.
Statistical Analysis: The study employed a comprehensive statistical approach to compare the two methods:
- Correlation: Spearman's coefficient ( $\rho$ ).
- Consistency: Intraclass Correlation Coefficient (ICC, two-way fixed-effects, consistency model).
- Agreement: Bland–Altman analysis to determine mean bias and 95% limits of agreement (LoA), alongside tests for proportional bias.
- Group Comparisons: Mann–Whitney U tests and regression models adjusted for age were used to compare groups and assess the impact of PD status on method agreement.

Key Results

Correlation and Consistency: BIA demonstrated strong correlations ( $p < 0.001$ ) and excellent consistency (high ICC values) with DXA for all body composition parameters (FM, FM%, lean mass, ALM, ALMI) in both the PD and control groups.
Systematic Bias:
- Fat Mass: No statistically significant differences were found between BIA and DXA for total FM or FM percentage. However, Bland–Altman analysis revealed small mean biases with wide limits of agreement and significant positive proportional bias (BIA underestimated lower values and overestimated higher values).
- Lean Mass: BIA significantly underestimated lean mass, ALM, and ALMI compared to DXA in both groups.
  - PD Group: Underestimation of 1.09 kg (lean mass), 1.18 kg (ALM), and 0.48 kg/m² (ALMI).
  - Control Group: Underestimation of 1.25 kg (lean mass), 1.29 kg (ALM), and 0.56 kg/m² (ALMI).
Disease Impact: The pattern of agreement and bias was similar between the PD and control groups, indicating that the presence of Parkinson's disease did not substantially alter the performance or agreement of the BIA device relative to DXA.

Significance and Claims
The paper concludes that while BIA shows strong correlation and excellent consistency with DXA, the two methods are not interchangeable at the individual level due to systematic underestimation of lean mass-related parameters and wide limits of agreement.

Clinical Utility: The authors assert that BIA remains a valuable tool for clinical screening and population-level assessment when DXA is unavailable, particularly because the disease state (PD) does not appear to confound the measurement agreement.
Diagnostic Caution: The study emphasizes caution when using BIA for precise individual diagnoses, such as classifying sarcopenia, where absolute values are critical. The systematic bias suggests that BIA-derived values should not be directly equated to DXA values without adjustment.
Methodological Insight: The findings suggest that the observed discrepancies are likely due to inherent methodological differences (e.g., BIA's reliance on electrical impedance and predictive equations derived from healthy populations) rather than specific pathophysiological characteristics of PD. The study highlights that while BIA tracks changes parallel to DXA, it does not provide equivalent absolute values.

In summary, the paper positions BIA as a practical, accessible surrogate for DXA in broad assessments but explicitly limits its use for individual diagnostic precision without acknowledging the systematic underestimation of lean mass.