20556 Background: The concept of anorexia/cachexia is evolving as its mechanisms are discovered, but the current - simple - diagnostic criteria (weight loss [WL], anorexia) nowadays guide poorly clinical care decisions or clinical trial design. We systematically reviewed factors and systems used to classify cancer cachexia. Methods: Three combined search strings (cachexia, cancer, classification) were applied in MEDLINE, Cochrane, EMBASE, PsycINFO, and CinAhl, from 1976–2007. Three investigators (interrater-reliability before consensus 93%) screened citations, abstracts and papers for inclusion criteria: original work, english/german, and a) systems (≥2 factors) or b) a single factor to classify cancer patients with WL or c) to predict response to anticachexia treatment. Results: Of 9817 citations, 1275 abstracts and 585 papers, 126 papers (7 systems, 112 factors, 7 subgroup analysis [84 anticachexia trials]) were included; representing 7913 patients (50% gastrointestinal, 20% lung cancer). Four systems combined anthropometrics (AM), nutritional intake (NI), symptoms (SY) and inflammation (INF), two systems AM/NI and either SY or INF, one added REE and hormones. All 6 systems challenge the old definition of a sole anorexia/cachexia syndrome. Only 1 study reports a multivariate analysis: 3 factor model (WL, NI, INF). One system classifies patients by dietary patterns. Of the single factors, AM and INF, also muscle strenght and ghrelin classify patients with cachexia in most studies, whereas NI, symptoms, metabolic alterations, REE, and leptin, are variably reported. Few studies describe hormons (testosterone, thyroxin), gastrointestinal dysmotility, and proteolysis-inducing factor. Tumor biology and anti-cancer treatment was not explored. Attempts to define subgroups of responders in anticachexia trials by single factors failed. Conclusions: This novel systematic review supports a multiple factor staging of cancer patients with involuntary weight loss. Future research for validation is required. Main factors used to classify cancer patients with involuntary weight loss Factor Patients (#) Studies (#), (long./c- sect) Classifcation of patients with weight loss by factor Muscle mass/fat mass 1070 9 (2/7) 6 pos, 1 neg, 2 unclear Inflammation (CRP, other) 2316 28 (2/26) 21 pos, 3 neg, 4 unclear Nutritional intake 2429 14 (1/13) 5 pos, 3 neg, 6 unclear Anorexia, early satiety, abdominal symptoms, depression; PG- SGA, FAACT, QOL. 2195 16 (1/15) variable* Muscle strenght 238 2 (0/2) 2 pos Metabolism (glucose/insulin resistance, fat, amino acids, albumine) 533 23 (1/22) variable* Resting Energy Expenditure 1289 9 (2/7) 3 pos, 5 neg, 1 unclear Leptin 642 11 (2/9) 5 pos, 6 neg Ghrelin 497 5 (4/1) 4 pos, 1 neg Pos: positive, neg: negative. Long: longitudinal, c-sect: cross-section. * : see text in results No significant financial relationships to disclose.