Formules de prédiction de l’adiposité chez la femme – contrôle de qualité

Translated title of the contribution: Adipose Tissue prediction equations in women – quality control

S. Provyn, J. Wallace, A. Scafoglieri, B. Sesboue, M. Marfell-Jones, I. Bautmans, J. P. Clarys

    Research output: Contribution to journalArticle


    Percent whole body fat is widely used in Public health sectors and in sports medicine. In addition, skinfolds are the most common laboratory and field anthropometric technique to determine percent body fat resulting into more than 600 prediction equations during more than half a century. The purpose of this study is to investigate the suitability of anthropometric-based equations for estimating percent adipose tissue of 54 female subjects (age: 30.9 ± 8.5 years) with a variety of different lifestyles compared with Dual Energy X-ray Absorptiometry (DXA).

    Whole-body DXA scan and 15 skinfolds, 14 girths, four breadths were measured in a randomized order. Hundred equations were eligible for application to the study sample, from which only 34 provided percent adipose tissue values with good correlation (r ≥ 0.70, p < 0.05) and without significant differences (p > 0.05) compared to DXA. However, Bland and Altman plots show acceptable to very good mean differences with DXA with a range from −1.9 to +1.8 %, and a 95 % limits of agreement from −10.5 to +10.8 %.

    These differences suggest that the majority of formulae are not valid for practical use on age-matched and activity individuals.
    Translated title of the contributionAdipose Tissue prediction equations in women – quality control
    Original languageFrench
    Pages (from-to)291–303
    JournalScience and Sports
    Issue number6
    Publication statusPublished - Dec 2010


    • Quality control
    • DXA
    • Adipose tissue
    • Body fat
    • Prediction equations
    • Anthropometry
    • Women


    Dive into the research topics of 'Adipose Tissue prediction equations in women – quality control'. Together they form a unique fingerprint.

    Cite this