Abstract
PURPOSE:The diagnostic value of selective anorexia is debated. Some authors have suggested an association between meat aversion and cancer, but most do not use it as a diagnostic tool. We aimed to characterize anorexia of different diseases to search for an association between selective aversions and diagnostic groups.METHODS:All the patients admitted to three departments of a teaching hospital were included consecutively for 22months if they had more than 10 % weight loss in less than one year. Patients were excluded if history taking was not reliable, or if they suffered from anorexia nervosa. We compiled diagnoses at discharge and validated them six months later. We used logistic regression to identify independent factors associated with selective anorexia.RESULTS:Inclusion criteria were met in 106patients (female 44 %, median age 65years). Most frequent diagnoses were: cancer (36 %), infection (35 %), digestive diseases (19 %), non organic diseases (21 %). Recent selective anorexia was found in 46 % of the cases. It was significantly associated with female gender (P=0.002), marginally with young age (P=0.069) and long duration of weight loss (P=0.079). Opioid use at admission was negatively associated with selective anorexia (P=0.001). No specific diagnostic category was found to be associated.CONCLUSION:Selective anorexia does not appear to be a useful symptom to investigate pathological weight loss. It behaves more like a non-specific reactivation by current disease of earlier latent personal food aversions.