2026 Eastern States Conference

Background/Objective: Midazolam and haloperidol are commonly used for pre-MRI
sedation despite being on the Beers list. The objective of this study is to compare
the efficacy and safety of these medications for pre-MRI sedation in older adults.
Methods: This was a retrospective cohort study from 09/2024 to 01/2026. Patients
were included if they were 65 years or older and ordered pre-MRI sedation through
an order set. Patients were excluded if they had an allergy or intolerance to
benzodiazepines or haloperidol, if they used midazolam or haloperidol 2-24 hours
prior to MRI, if there was deviation from the order set, if they had an anesthesia
consult, or if they were admitted to an intensive care unit (ICU) at the time of MRI. An
MRI was successful if it was completed on the first attempt without use of an
alternative sedation agent. The primary safety outcome was a composite of any of
the following within 24 hours after the pre-MRI sedation agent: fall, intubation or
escalating supplemental oxygen requirement, rapid response team called, use offlumazenil, ICU admission, or death. Chi-square or Fisher’s exact test was used to
compare incidence of MRI success and safety events between midazolam and
haloperidol groups.
Results: There were 39 patients in the haloperidol arm and 81 patients in the
midazolam arm. There was no statistical difference in MRI success between
haloperidol (69.2%) and midazolam (72.9%) [p=0.681]. The primary safety outcome
occurred more frequently in the midazolam arm (30.7% vs. 2.6% (p=0.0002)). New or
escalating use of supplemental oxygen occurred in 21 midazolam patients (25.9%).
Conclusion: No difference in efficacy was observed between haloperidol and
midazolam for pre-MRI sedation in older adults, however, the sample size did not
meet pre-specified power. Midazolam was associated with a higher incidence of
new or escalating oxygen requirement. The use of supplemental oxygen should be
further investigated to determine the clinical significance of this finding.