Multicomponent models in body composition research: opportunities and pitfalls.

This paper presents theoretical, methodological, and statistical considerations in the evaluation of multicomponent body composition models. The opportunities for multicomponent approaches to validate new body composition methods, to provide definitive references for FFB composition values of different populations, and to evaluate changes in various components of the body with age and interventions are well-recognized. The pitfalls of the multicomponent approach, however, limit the success of past research efforts. If future investigations evaluate in the same study design the theoretical, methodological and statistical approaches presented, improved multicomponent approaches will be developed, and present controversies surrounding the variability in composition of FFB as well as the validity of new methodologies such as DXA will be resolved.