Moderate midazolam sedation is often used in gastrointestinal endoscopy to induce stress-free conscious sedation. Conversely, flumazenil can reverse midazolam-induced sedation and cause temporary awakening and resedation. However, the effects of flumazenil disuse on the incidence of inpatient falls are unknown. In this study, we performed a retrospective cohort analysis of the incidence of falls in inpatients who underwent gastrointestinal endoscopy under midazolam-induced sedation with or without flumazenil. This study included 1,424 procedures, of which 559 involved flumazenil use. The frequency of inpatient falls did not significantly differ between the flumazenil and nonflumazenil use groups (2/559 episodes [0.36%] vs. 2/865 episodes [0.23%], P = 0.648). The inverse probability of treatment weighting analysis could not determine the association of flumazenil disuse with the incidence of inpatient falls (odds ratio, 0.57; 95% confidence interval, 0.08-4.14; P = 0.58). Our results indicate that the association between flumazenil disuse and the incidence of inpatient falls remain unclear.