Rare Obesity Cohorts With Food Behavioral Disorders : Better Diagnosis for Better Treatment

Hypothalamic obesity (HO) is defined as obesity secondary to functional or anatomical alterations of the hypothalamus, the central organ of energy homeostasis. The causes of HO are related either to hypothalamic lesions (eg craniopharyngioma) either to genetic diseases (syndromic obesity such as Prader-Willi syndrome or monogenic non syndromic obesity such as variants on leptin/melanocortin pathway). HO, which accounts for about 5 to 10% of obesity, groups complex disorders characterized by severe obesity associated with eating disorders, cognitive and behavioral disorders, endocrine and metabolic alterations and sometimes a visual deficit, with a major impact on quality of life, morbidity and mortality. There is currently no specific treatment of HO. Clinical management is essentially behavioral, based on daily support of eating behavior and physical activities. HO is characterized by an intense and almost permanent hunger; a satiety disorder and an obsessive interest in food. The education regarding food intake behavior of the caregivers and relateds is critical with advices concerning the control of the access to food and the setting up of a precise food frame on the quantities, with low energetic density, and schedules. HO are complex medical situations, often refractory to current lifestyle therapies. However innovative therapies with molecules targeting the hypothalamus are emerging. The investigator's main hypothesis is that HO have alterations in eating behavior that can be improved by innovative treatments such as, for example, molecule targeting the melanocortin pathway. The response to therapy could depend on hypothalamic origin and especially on the genotype. ObeRar cohort aims to i) improve early diagnosis of HO and ii) characterize the natural history of obesity and eating disorders, the associated phenotypes and "lifestyle" profiles (physical activity, sleep, nutrition) and cardio-metabolic and neuropsychological parameters. Defining profiles will h