Background: Many activities contribute to reduce drug-related problems. Among them, the medication reconciliation (MR) is used to compare the best possible medication history (BPMH) and the current admission medication order (AMO) to identify and solve unintended medication discrepancies (UMD). This study aims to assess the impact of the implementation of admission MR by clinical pharmacists on UMD. Method: This prospective study was carried out in two units of general medicine and infectious and tropical diseases in a 1844-bed French hospital. A retroactive MR performed in an observational period was compared to a proactive MR realized in an interventional period. We used a logistic regression to identify risk factors of UMD. Results: During both periods, 394 patients were enrolled and 2,725 medications were analyzed in the BPMH. Proactive MR reduced the percentage of patients with at least one UMD compared with retroactive process (respectively 2.1% vs. 45.8%, p<0.001). Patients with at least one UMD during both periods were older compared to patients without UMD (79 vs. 72, p<0.005) and had more medications at admission (7 vs. 6, p<0.0001). UMD occur 38 times more often when there is no clinical pharmacist intervention. Among the 226 UMD detected in both periods, 42% would have required monitoring or intervention to preclude harm, and 10% had potential harm to the patient and 2% were life threatening. Conclusion: Proactive MR performed by clinical pharmacists is an acute process of detection and correction of UMD, but it requires a lot of human resources.