Introduction: Optimizing naloxone dosing in the context of increasing fentanyl and ultra-potent opioid (UPO) prevalence is an important consideration for emergency health care providers. The goal of this systematic review was to evaluate the association between initial and cumulative naloxone doses on effective reversal and adverse events in undifferentiated and fentanyl/UPO overdoses. Methods: We searched Embase, MEDLINE, Cochrane Central Register of Controlled Trials, DARE, CINAHL, Science Citation Index, reference lists, toxicology websites, and conference proceedings from July to October 2018 and back to 1972. Our search included pertinent indexing terms for UPOs. We included interventional and observational studies reporting on naloxone administration for opioid toxicity reversal in people ≥12 years old. Additionally, we accessed non-traditional evidence sources (case reports and series) given this rapidly changing field. We conducted inclusion screens, data extraction and quality assessments in duplicate. We summarized study characteristics and where reported, analyzed number of patients with clinical response. Response was defined as not receiving further naloxone doses and remaining alive. Results: We included 174 studies (108 case reports and series, 55 observational, 9 interventional) with 26,660 subjects (median age 35.1; 74.2% male). We observed lower response among patients exposed to fentanyl/UPO versus heroin for initial naloxone doses ≤0.4mg (56.8% versus 80.2%) and > 0.4mg (27.0% versus 82.1%). Mean cumulative doses were higher for fentanyl/UPO (2.10 mg, SD 1.80 mg) versus heroin (1.48 mg, SD 1.68 mg) overdoses. In North American studies the median cumulative dose used was higher for fentanyl/UPO versus heroin overdoses. A dose-response curve for fentanyl/UPO studies showed marked variability in doses among responders, indicating heterogeneity. Adverse events reporting was inconsistent; 10% of subjects experienced withdrawal based on studies in which they were reported. Conclusion: This is the first systematic review to summarize proportion of patients with clinical response by naloxone dose provided. While variable reporting, study quality, heterogeneity, and our outcome definitions limit the conclusions we can draw, it appears that higher initial doses and in some cases, higher cumulative naloxone doses were used and may be necessary to reverse toxicity due to fentanyl/UPO compared to other opioids. High-quality prospective studies assessing effectiveness and safety are needed.