OBJECTIVETo evaluate the association between the intraoperative administration of midazolam and the incidence of postoperative delirium in patients undergoing cardiac surgery.DESIGNRetrospective observational cohort study.SETTINGThe Japanese Diagnosis Procedure Combination database.PARTICIPANTSPatients aged 65 years and older who underwent cardiovascular surgery (excluding transcatheter surgeries, multiple surgeries per admission, and preoperative delirium) between April 1, 2015, and October 31, 2019.MEASUREMENTS AND MAIN RESULTSPatients who received midazolam (midazolam group) were compared with those who did not receive midazolam (no midazolam group). The primary outcome was the incidence of postoperative delirium. The secondary outcomes were the incidence of postoperative nausea and vomiting, mortality, and duration of intensive care unit stay and hospitalization. Propensity scores were estimated using logistic regression based on the covariates. The outcomes were compared using stabilized inverse probability of treatment-weighting analyses. Among the 16,185 patients analyzed, 10,633 (65.7%) received midazolam. No significant differences were observed in the incidences of postoperative delirium (odds ratio [OR] 0.95; 95% CI 0.87-1.03; p = 0.21) and hospital mortality (OR 0.92; 95% CI 0.76-1.11; p = 0.39) between the groups; however, the midazolam group had slightly longer durations of intensive care unit stay (3.5 [3.5-3.6] v 3.3 [3.3-3.4] days, p < 0.001) and hospitalization (31.5 [31.1-31.9] v 29.4 [28.8-29.9] days, p < 0.001), and slightly lower incidences of postoperative nausea and vomiting (OR 0.92; 95% CI 0.85-0.99; p = 0.03). The sensitivity analyses supported these results.CONCLUSIONSIntraoperative administration of midazolam may not induce postoperative delirium in patients undergoing cardiac surgery.