Comparison of the FAT-free mass obtained by impedancemetry in patients with amyotrophic lateral sclerosis: direct measurement with device formula versus reference formula.

Abstract : Rationale: Amyotrophic lateral sclerosis (ALS) is a rare disease with a poor prognosis. Body composition is a prognostic factor for survival. During ALS, body composition can be assessed by total body biolelectric impedancemetry analysis (BIA) using a validated reference formula (Desport et al.). This formula is more complex to use than the direct measurement obtained by a BIA device. The aim of the study was to compare the agreement for the Fat-Free Mass (FFM) obtained by BIA with reference formula or with the formula of the device. Methods: The data from January 2015 to January 2019, were extracted from our database and from the BIA devise Bodystat Quadscan 4000® used in consultation. Each ALS patient had a measurement of FFM (FFM) by direct measurement of the device (FFMdev) and by the reference formula (FFMref) also requiring the measurement of the tricipital skin fold. A simple regression test was applied between the two values of FFM and the agreement was determined using the method of Bland and Altman. The repeatability coefficient was studied for the impedance at 50 kHz, from 6 consecutive measurements in 15 patients. Results: The study population included 343 ALS patients. The mean age was of 66.5 years ± 11.0 years, the male / female sex ratio was of 1. The FFMref was of 43.9 ± 9.8 kg and the FFMdev was of 41.7 ± 11.0 kg. The repeatability coefficient of the impedance at 50 kHz was of 0.53 ± 1.28%. The correlation between the two types of measures was very good (R at 0.90, p <0.0001). On the other hand, the limits of agreement on the Bland and Altman graph were + 7.6 and -11.9 kg, with an average risk of error of 12.3% which is not acceptable in clinical practice. FFMdev underestimated the FFM on average with 2.2 ± 5.0 kg (5.0%) compared to the FFMref. Conclusion: Measuring the FFM directly on the BIA device is associated with a too important risk of error to be able to be used in clinical practice, despite simple use and good repeatability of the measurements. For the measurement of FFM duri