Objective: The use of generic drugs is promoted to reduce medical costs and copayments. However, tumor agents are expensive and generic drugs are not widelyused. Thus, it is necessaryto evaluate the safetyof generic drugs in more detail. We compared the incidence of adverse events between the original drug (Gemzar®: GEM) and generic drug (Gemcitabine [Sandoz]: GE-GEM) using propensityscore (PS) matching.Methods: We investigated adverse events in patients who received one course of GEM or GE-GEM. The patient background (age,sex, BSA, cancer type, stage, metastasis, surgical history, and radiotherapy) and administration status (administration route and RDI) were used to calculate the PS.Results: Among all patients (GEM: 51, GE-GEM: 54), a significantlygreater number in the GE-GEM group had cancer metastasis. On comparison of adverse events, there were significantlymore cases of vascular pain (p＜0.05) in the GEM group, and manycases of nausea (p＝0.08) and rash (p＝0.08). Fortypatients in each group were extracted byPS matching. There were no significant differences in the patient background between the groups, and on comparison of adverse events, the two groups did not significantly differ.Conclusion: Our studysuggested that there is no difference in side effects between Gemzar® and gemcitabine [Sandoz]. To compare the incidence of adverse events, it is useful to use PS matching in clinical practice.