Background: The clinical characteristics and prognostic significance of changes in platelet count (PC) during hospitalization in internal medicine wards have not been well investigated. Methods: Demographic, clinical and laboratory data were collected from 345 patients admitted to an internal medicine ward. Following discharge, all-cause mortality was recorded. These data were compared, according to deltaPC (PC on discharge minus PC on admission): group 1 (drop in PC, deltaPC - 50 x 10(9)/l), group 2 (no significant PC changes, deltaPC up to 50 x 10(9)/l) and group 3 (rise in PC, deltaPC + 50 x 10(9)/l). Results: Groups 1, 2 and 3 comprised 64 (18.5%), 200 (58%) and 81 (23.5%) patients, respectively. Patients from group 3 were younger, more likely admitted for infection and less likely for cardiovascular disorder, and less often presenting with coronary artery disease, complex nursing care and thrombocytosis on admission or thrombocytopenia on discharge than patients from groups 1 and 2. Mean platelet volume was higher in group 2 on admission and lower in group 3 on discharge. During a median follow-up of 25 months, 146 (42.3%) of 345 patients died. The survival rate was higher for group 3 (65.4%) than for groups 1 (45.3%) and 2 (58.5%), p = 0.003. In the entire cohort, each 100 x 10(9)/l increment of deltaPC was a powerful predictor of lower mortality (p = 0.03, relative risk = 0.83, 95% confidence interval = 0.71-0.98). Conclusions: Increased PC throughout hospitalization was associated with better prognosis than a drop or blunted rise in PC. The assessment of PC changes in an internal medicine ward may provide useful prognostic information. (C) 2014 European Federation of Internal Medicine. Published by Elsevier B. V. All rights reserved.