Introduction: Patients with suspected opioid overdose frequently require naloxone treatment. Despite recommendations to observe such patients for 4 to 24 hours after naloxone, earlier discharge is becoming more common. This prospective, observational study of patients with presumed opioid overdose examines the safety of early disposition decisions and the accuracy of outcome prediction by physicians 1 hour after the administration of naloxone.
Methods: The study was carried out at St. Paul's Hospital, an inner city teaching centre that cares for most of the injection drug users in Vancouver, BC. Patients were formally assessed 1 hour after receiving naloxone for presumed opioid overdose. Demographics, medical history and physical examination were documented on specific data forms, and physicians recorded their comfort with early discharge. Patients were followed up, and those who required a critical intervention or suffered a pre-defined adverse event (AE) within 24 hours of their 1-hour assessment were identified.
Results: Of 573 patients, 48% were discharged in less than 2 hours, 23% in 2-4 hours and 29% in >4 hours. 94 patients who were held in the emergency department (ED) or admitted required a critical intervention, including supplemental oxygen for hypoxia (74), repeat naloxone (52), antibiotics administered intravenously (IV) (14), assisted ventilations (13), fluid bolus for hypotension (12), charcoal for associated life-threatening overdose (6), IV inotropic agents (2), antiarrhythmics for sustained tachycardia >130 beats/min (1), and administration of bicarbonate for arterial [HCO3] <5 or venous CO2 <5 (1). Physicians predicted adverse events with 94% sensitivity and 59% specificity. No discharged patients suffered a serious AE within 24 hours of ED discharge.
Conclusions: Emergency physicians can clinically identify patients at risk of deterioration after naloxone reversal of suspected opioid overdose. Prolonged observation or hospital admission is not usually required. Selective early discharge of patients with presumed opioid overdose is feasible and appears safe. A clinical prediction rule may be useful in identifying patients eligible for early discharge.