Objective Naloxone is an established antidote for the treatment of heroin poisoning; however, dosing regimens vary widely, with a current trend towards small titrated intravenous dosing. This study aims to characterise naloxone use in the treatment of patients presenting with suspected heroin poisoning. Methods This was a retrospective review of poisoned patients presenting to a clinical toxicology unit in Brisbane from January 2015 to December 2017. Patient demographics, clinical effects, naloxone dosing, observation periods and complications were extracted from the patient's medical records. Results There were 117 presentations accounted for by 108 patients. Prehospital naloxone was provided to 57 (49%) patients, 46 of which received a standardised 1.6 mg i.m. dose. The remaining 60 (51%) patients received their first naloxone in hospital, with 58 (97%) receiving this by titrated i.v. doses. A subsequent naloxone infusion was required significantly more often in those treated with i.v. titrated naloxone compared to i.m. dose (27/69 [39%] vs 5/48 [10%], P = 0.0006). The need for parenteral sedation to manage acute behavioural disturbance following naloxone provision was rare (3/117 [3%]). Conclusions In this retrospective observational study, a single large i.m. dose of naloxone reversed the toxicity of suspected heroin overdose in the majority of patients. In addition, patients were less likely to require repeated intermittent doses or naloxone infusion than those treated solely with i.v. naloxone. Further comparison in a prospective study is warranted to validate these observations in confirmed heroin overdose. Requirement for sedation secondary to acute behavioural disturbance was rare regardless of the route.