Intramuscular Midazolam, Olanzapine, Ziprasidone, or Haloperidol for Treating Acute Agitation in the Emergency Department

Abstract

Agitation in the emergency department (ED) can pose a threat to patient and provider safety; therefore, treatment is indicated. The purpose of this study is to compare haloperidol, olanzapine, midazolam, and ziprasidone to treat agitation. This was a prospective observational study of consecutive patients receiving intramuscular medication to treat agitation in the ED. Medications were administered according to an a priori protocol in which the initial medication given was predetermined in the following 3-week blocks: haloperidol 5 mg, ziprasidone 20 mg, olanzapine 10 mg, midazolam 5 mg, and haloperidol 10 mg. The primary outcome was the proportion of patients adequately sedated at 15 minutes, assessed with the Altered Mental Status Scale. Seven hundred thirty-seven patients were enrolled (median age 40 years; 72% men). At 15 minutes, midazolam resulted in a greater proportion of patients adequately sedated (Altered Mental Status Scale <1) compared with ziprasidone (difference 18%; 95% confidence interval [CI] 6% to 29%), haloperidol 5 mg (difference 30%; 95% CI 19% to 41%), haloperidol 10 mg (difference 28%; 95% CI 17% to 39%), and olanzapine (difference 9%; 95% CI -1% to 20%). Olanzapine resulted in a greater proportion of patients adequately sedated at 15 minutes compared with haloperidol 5 mg (difference 20%; 95% CI 10% to 31%), haloperidol 10 mg (difference 18%; 95% CI 7% to 29%), and ziprasidone (difference 8%; 95% CI -3% to 19%). Adverse events were uncommon: cardiac arrest (0), extrapyramidal adverse effects (2; 0.3%), hypotension (5; 0.5%), hypoxemia (10; 1%), and intubation (4; 0.5%), and occurred at similar rates in each group. Intramuscular midazolam achieved more effective sedation in agitated ED patients at 15 minutes than haloperidol, ziprasidone, and perhaps olanzapine. Olanzapine provided more effective sedation than haloperidol. No differences in adverse events were identified.