OBJECTIVES: To examine the relationship between key patient variables and variation in naloxone dose (from the standard dose of 1.6 mg IMI) administered by ambulance paramedics in the prehospital management of heroin overdose. METHODS: A retrospective analysis of 7985 ambulance patient care records of non-fatal heroin overdose cases collected in greater metropolitan Melbourne. The main outcome measure was the dose of intramuscular naloxone required to increase the level of consciousness and the respiratory rate in patients presenting with suspected heroin overdose. Key patient variables influencing the dose that were recorded included: age, sex, initial patient presentation and reported concurrent alcohol use. RESULTS: Multinomial logistic regression revealed that patients with higher levels of consciousness and respiratory rates on arrival of the paramedic crew were more likely to receive a less than standard dose of naloxone. Conversely, patients with lower levels of consciousness and low respiratory rates received greater than standard doses of naloxone for resuscitation. Patients who received greater than the standard dose of naloxone were 2.25 (95% CI, 1.83-2.77) times more likely to have been under the influence of alcohol when consuming the heroin that resulted in overdose. CONCLUSIONS: The concurrent use of alcohol with heroin resulted in the use of greater than standard doses of naloxone by paramedics in resuscitating overdose patients. It is possible that the higher dose of naloxone is required to reverse the combined effects of alcohol and heroin. There was also a link between initial patient presentation and the dose of naloxone required for resuscitation. In light of these findings, it would appear that initial patient presentation and evidence of alcohol use might be useful guides as to providing the most effective dose of naloxone in the prehospital setting.