The purpose of this study is to evaluate whether body composition is associated with a higher risk of mortality in subjects ≥ 65 years.

Study centre: University Hospitals of Geneva Study type: retrospective study Study duration (anticipated): 1 year for collection of data(1.10.2011-1.12.2012)and 3 months for statistical analysis and writing, leading to the completion of the study by march 2013 Study hypothesis: A low fat-free mass (FFM) and a high fat mass (FM) are associated with a higher risk of mortality in subjects ≥ 65 years.

Aims: The primary objective is to evaluate the relationship between body composition and mortality in men and women ≥ 65 years, while taking into account co-morbidities. The secondary objectives are to compare this relationship a) with the relationship of body mass index (BMI) and mortality, b) with the relationship of body composition changes and mortality, c) between healthy subjects and subgroups of patients with chronic diseases Methods: Included are all subjects who had a measurement of body composition by bioelectrical impedance analyses, performed in the HUG from 1990 until december 2011. Excluded are subjects living abroad. Body composition measurements will be retrieved from DPI and 4D which preceded the use of DPI in nutrition and is presently only used for research. Diseases and specific laboratory values at time of body composition measurement will be retrieved simultaneously. Mortality will be retrieved from the population registers for death.

Statistics: The relationship between body composition and mortality, in the total population and subgroups of patients with chronic diseases, will be evaluated by Cox regression, adjusted for age, gender and co-morbidities. The variance of mortality (adjusted R2) related to FFM will be compared to that related to BMI and FFM loss.

Inclusion Criteria:

• All subjects ≥ 65 years included in previous studies on body composition by bioelectrical impedance analysis performed in the Nutrition Unit since 1990 until December 2011.
• All patients ≥ 65 years who had a measurement of body composition by bioelectrical impedance analysis performed in clinical routine by the Nutrition Unit and the dietitians of the University Hospital of Geneva since 1990 until December 2011, while hospitalized or not.

Exclusion Criteria:

• Subjects living in another country than Switzerland at time of body composition measurement

This is a historical cohort study of prospectively collected measurements of body composition performed by bioelectrical impedance analysis, between 1990 and 2011.