ABSTRACT
BACKGROUND: The "obesity paradox" is poorly understood in vulnerable older hospitalized populations. OBJECTIVES: To prospectively analyze the impact of body mass index (BMI) and comorbidities on early (6-week), one- and two-year mortality. DESIGN: Prospective multicenter study with a two-year follow-up of old patients participating in the SAFES cohort study. SETTINGS: Nine university hospitals in France. PARTICIPANTS: Patients aged 75 or older hospitalized in medical divisions through the emergency department. MEASUREMENT: Inpatients' characteristics were obtained through a comprehensive geriatric assessment of inpatients, conducted in the first week of hospitalization. All-cause mortalities at 6-week, one- and two-year were determined using bivariable and multivariable Cox proportional hazard model. RESULTS: The SAFES cohort included 1,306 patients, aged 85±6 years, with a majority of women (65%). One- and two-year mortality were inversely associated with BMI ≥30 kg/m2 while early mortality was not, and positively associated with age, burden of comorbidities, walking disorders, level of dependency and presence of a dementia syndrome. Survival rates between patients in low (< 18.0 kg/m2) and intermediate (18-24.9 and 25-29.9 kg/m2) BMI categories were not significant. CONCLUSION: While our findings seem to confirm the reality of the "obesity paradox" in vulnerable older hospitalized population, the exact understanding of underlying mechanisms and even the truthfulness of this paradoxical relationship are still fraught with considerable methodological, epidemiological and metabolic challenges.