The prognostic significance of platelet count (PC) changes during hospitalization for community-acquired pneumonia (CAP) has not been investigated. For 976 adults, clinical data during hospitalization for CAP and all-cause mortality following discharge were compared according to ΔPC (PC on discharge minus PC on admission): groups A (declining PC, ΔPC < −50 × 109/l), B (stable PC, ΔPC ± 50 × 109/l), and C (rising PC, ΔPC >50 × 109/l), and according to the presence of thrombocytopenia, normal PC, and thrombocytosis on admission/discharge. Groups A, B, and C comprised 7.9%, 46.5%, and 45.6% of patients, respectively. On hospital admission/discharge, thrombocytopenia, normal PC, and thrombocytosis were observed in 12.8%/6.4%, 84.1%/84.4%, and 3.1%/9.2% of patients, respectively. The respective 90-day, 3-year, and total (median follow-up of 54 months) mortality rates were significantly higher: in group A (40.3%, 63.6%, and 72.7%), compared to groups B (12.3%, 31.5%, and 39.0%) and C (4.9%, 17.3%, and 25.4%), p < 0.001; and in patients with thrombocytopenia at discharge (27.4%, 48.4%, and 51.6%), compared to those with normal PC (10.2%, 26.9%, and 35.4%) and thrombocytosis (8.9%, 17.8%, and 24.4%) at discharge (p < 0.001). Mortality rates were comparable among groups with thrombocytopenia, normal PC, and thrombocytosis at admission (p = 0.6). In the entire sample, each 100 × 109/l increment of ΔPC strongly predicted lower mortality (p < 0.001, relative risk 0.73, 95% confidence interval 0.64−0.83). In conclusion, PC changes are common among CAP inpatients. Rising PC throughout hospitalization is a powerful predictor of better survival, while declining PC predicts poor outcome. Evaluation of PC changes during hospitalization for CAP may provide useful prognostic information.