Background: Although intravenous midazolam and flunitrazepam are frequently administered for primary insomnia in Japan, there is no empirical study on their efficacy and safety.
Design and subjects: To compare the efficacy, safety, and cost-effectiveness of midazolam and flunitrazepam, a multicenter retrospective audit study was performed on 104 and 59 patients receiving midazolam and flunitrazepam, respectively, from 18 certified palliative care units.
Results: Median administration periods were 6 days and 9 days for midazolam and flunitrazepam, respectively. The median initial and maximum doses were 10 mg per night and 18 mg per night for midazolam, and 2 mg per night and 2 mg per night for flunitrazepam, respectively. There were no significant differences in the efficacy (91% in the midazolam group versus 81% in the flunitrazepam group, p = 0.084), hangover effect (34% versus 19%, p = 0.094), delirium at night (12% versus 10%, p = 1.0) and delirium next morning (11% versus 15%, p = 0.33), treatment withdrawal (4.8% versus 1.7%, p = 0.41), and treatment-related death (0% versus 0%, p = 1.0). Flunitrazepam caused respiratory depression defined as physician or nurses records such as apnea, respiratory arrest, decreased respiratory rate, and respiratory depression significantly more frequently than midazolam (17% versus 3.8%, p = 0.0073). The maximum dose was more highly correlated with the administration period in the midazolam group than in the flunitrazepam group (ρ = 0.52, versus ρ = 0.39), and, for patients treated for 14 days or longer, the daily escalation dose ratio required for maintaining adequate sleep was significantly higher in the midazolam group than in the flunitrazepam group (11% versus 2.6%, p = 0.015). The costs of the initial and maximum administration were significantly higher in the midazolam group than in the flunitrazepam group (p < 0.001).
Conclusion: Intravenous midazolam and flunitrazepam appeared to be almost equal about efficacy and safety for primary insomnia, but flunitrazepam is less expensive and shows lower risk of tolerance development. A future prospective comparison study is necessary.
