BACKGROUND & AIMS:In hospitalized patients, lack of appetite, i.e., disease-associated anorexia, is the main factor determining insufficient food intake and weight loss, which in turn increase morbidity and mortality. Controversies exist on which tool should be preferred when diagnosing anorexia. Aim of the study was to evaluate in hospitalized medical patients, the performance of 4 different tools [i.e., self-assessment of appetite, FAACT-ESPEN score, visual analog scale (VAS), and the Anorexia Questionnaire (AQ)] in assessing disease-associated anorexia and predicting nutritional and clinical variables. MATERIALS AND METHODS:Hospitalized patients consecutively admitted to the Internal Medicine ward at our institution were considered. After informed consent was obtained, patients were asked to self-assess their appetite vs the previous month. The VAS, the FAACT-ESPEN score and the Anorexia Questionnaire were also submitted. Food intake immediately following the interview was recorded. Nutritional (i.e., body weight, height), functional (i.e., handgrip strength) and clinical variables (i.e., length of stay) were registered upon admission and before discharge. RESULTS:We studied 105 patients (74M:31F; 66.2 ± 16.3 yrs). The prevalence of anorexia as assessed by patients' self assessment, FAACT-ESPEN score, and the Anorexia Questionnaire was 23%, 10% and 48%, respectively. VAS did not show any correlation with food intake. Anorexic patients as identified by the self assessment of appetite showed reduced food intake and weaker handgrip strength than non-anorexic. The FAACT-ESPEN score correlated with body weight, food intake and handgrip strength, but was not related with length of stay. Anorexic patients as identified by the Anorexia Questionnaire showed reduced food intake, lower body weight, weaker handgrip strength and longer hospital stay than non-anorexic patients. DISCUSSION:The prevalence of anorexia significantly varies according to the diagnostic tool used. Except for VAS, all the tested tools identify patients with impaired nutritional and functional variables. However, only the Anorexia Questionnaire identifies patients with longer hospital stay. Our results suggest that in clinical practice, modification of appetite reflects different underlying mechanisms whose impacts on clinical outcome measures may differ. Therefore, an ideal anorexia assessment tool does not appear to exist, but it should be chosen according to the outcome measures to be assessed (i.e., Anorexia Questionnaire to predict length of stay).