OBJECTIVE We performed this study in adult patients with chronic primary immune thrombocytopenia to explore the long-term efficacy and safety of splenectomy. METHODS Data of 174 patients who underwent splenectomy in our hospital from 1994 to 2014 were analyzed. RESULTS After splenectomy, 126 (72.4%) patients achieved a complete response (CR) and 28 (16.1%) achieved a response (R). Thirty-two (20.8%) responders relapsed with a median time of 24 months. Compared with non-responders and recurrent patients, the stable responders were younger and had higher preoperation and postoperation peak platelet count, later peak platelet count emergence time, and more megakaryocytes. Corticosteroid-dependent patients were more likely to response to splenectomy than those refractory to corticosteroid. We performed a relapse-free survival analysis among the 154 responders. In univariate analyses, corticosteroid dependent and time from diagnosis to splenectomy ≤24 months showed predictive value to persistent response. But only corticosteroid dependent was a significant predictor in multivariate analysis. The 30-d complication rate after the surgery was 25.9%. There were five (2.9%) patients experienced thrombosis and three (1.7%) refractory patients died during follow-up. CONCLUSIONS Splenectomy was a safe treatment with a cure rate of 58.0%. Corticosteroid dependent showed predictive value to persistent response.