Background Medication reconciliation prevents medication errors at care transition points. This process improves communication with general practitioners regarding the reasons for therapeutic changes, allowing those changes to be maintained after hospital discharge. Objective To investigate the impact of medication reconciliation in geriatrics on the sustainability of therapeutic optimization after hospital discharge. Setting This study was conducted in a geriatric unit in a University Hospital Centre in France. Method This was a retrospective study. For 6 months, all patients over 65 years who underwent the process of medication reconciliation performed by a clinical hospital pharmacist and a physician at admission and discharge, were included. A comparison between drug prescriptions at hospital discharge and the first prescription made outside the hospital was made to identify any differences. Main outcome measure The main outcome measures were the provision of the results of the medication reconciliation performed in the hospital to the relevant general practitioner, the subsequent acceptance of that information, the type of medication discrepancies one month after discharge and the therapeutic classes affected by the modifications. Results Among the 112 patients, medication reconciliation allowed us to identify and correct 87 unintentional discrepancies at admission (88% corrected) and 54 at discharge (92% corrected). Patients were discharged to homes or nursing homes (61%), geriatric rehabilitation units (38%) or psychiatric clinics (1%). A general practitioner wrote the first prescription renewal a mean of 36 ± 23 days after discharge, having been made aware of the medication reconciliation in only 24% of the cases (received and taken into account). The impact was a decrease in the number of patients with at least one discrepancy. Twenty-five percent of general practitioners who were aware about the medication reconciliation process accepted all therapeutic changes, while only 7% of those who were not informed did so (p = 0.02). The number of medication discrepancies observed was correlated with the number of medications for which prescriptions were renewed (p < 0.01). Conclusion Medication reconciliation involving therapeutic optimization and the justification of changes is essential to ensure the safety of the prescriptions written for patients. However, its impact after discharge is hampered by the fact that the results are often not received or taken into account by general practitioners. Taking medication reconciliation into account was associated with a significant increase in prescriptions that maintained therapeutic changes made in the hospital, confirming the positive impact of communication between care providers on therapeutic optimization.