Background: Undifferentiated severe agitation (SA) in the prehospital environment is a commonly encountered problem that represents a safety issue for both the patient and their caregivers. When rapid sedation is indicated, controversy exists regarding the ideal agent. We hypothesized ketamine (K) 5 mg/kg intramuscular (IM) would be superior to haloperidol (H) 10 mg IM for SA in the prehospital environment, with time to adequate sedation as our primary outcome measure. Methods: This is a prospective open label Waiver of Consent study (45 CFR 46.116) of all patients in our EMS system needing chemical restraint for SA that were subsequently transported to our ED. From October 2014 to February 2015, all patients in our EMS system with SA were treated with H. Our standard treatment for prehospital SA in February of 2015 was subsequently changed to K. All paramedics in our EMS system were trained in the Altered Mental Status Scale, a validated ordinal scale of agitation. Paramedics carried stopwatches and measured time to adequate sedation after injection. Secondary outcomes included additional sedatives given, ethanol concentration, intubation, vomiting, dystonia, akathisia, emergence reaction, laryngospasm, or hypersalivation. Results: 89 subjects have thus far been enrolled; 64 received H, 25 received K. Median age of the H arm was 32 (range 18.69); median age for K was 36 (range 20.55). For gender, 32/63 (51%) were male in the H arm; 15/25 (60%) were male in the K arm. Twelve subjects in the H arm required another medication prehospital for sedation; all were given midazolam 5 mg IM. No subjects in the K arm required additional sedation prehospital. In the H arm 38/64 (59%) achieved adequate sedation prehospital; in the K arm 24/25 (96%) achieved adequate sedation prehospital (p ðu 0.001). Median time to adequate sedation in the H arm was 19.6 min (range 3.8.84); median time to adequate sedation in the K arm was 5.5 min (range 1.6.15) (p ðl 0.0001). Regarding intubation, 2/64 (3%) of subjects in the H arm were intubated versus 12/25 (48%) of subjects in the K arm (p ðl 0.001). Complication rate including vomiting, dystonia, akathisia, emergence reaction, laryngospasm, or hypersalivation were higher in the ketamine group (3/55, 5% in the H arm versus 10/22, 45% in the K arm). Complications in the H arm included only dystonia and vomiting. All other complications were seen in the K arm. Median breath ethanol in the H arm was 0.16 g/dL (n ðu 53, range 0.0.42), in the K arm it was 0.18 g/dL (n ðu 10, range 0.0.34). Conclusion: For severe prehospital agitation, ketamine 5 mg/kg IM is superior to haloperidol 10 mg IM regarding time to adequate sedation. Ketamine is, however, associated with a significantly higher complication rate.