Prescribed Dose of Opioids and Overdose: A Systematic Review and Meta-Analysis of Unintentional Prescription Opioid Overdose.

BACKGROUND: The rate of an unintentional drug overdose involving prescription opioids continues to rise. An understanding of the threshold dose and dose(s) associated with unintentional prescription opioid overdose will help to mitigate this epidemic.
OBJECTIVE: The objective of this systematic review is to systematically synthesise and meta-analyse studies on doses of prescription opioids and ascertain the doses of opioids that are associated with increased risk of severe opioid poisoning or mortality. DATA SOURCES: A search of PubMed, EMBASE, CINAHL and Web of Science from inception to 16 January 2017 was conducted using search strategies and the MeSH (Medical Subject Headings) terms for studies of adult patients using prescription opioids who experienced an accidental overdose. STUDY SELECTION: Of the 1332 studies identified, 117 were selected for full article review. Ten met the inclusion criteria for qualitative analysis, but only seven studies were meta-analysed. The included studies were in English, and participants met predetermined International Classification of Diseases (ICD) codes. Studies were excluded if they included only paediatric participants or the participants met the ICD code for intentional self-harm. DATA EXTRACTION AND SYNTHESIS: Two researchers elaborated and validated a data extraction form. Data were then independently extracted by both reviewers as per this form. We assessed study quality using the Newcastle-Ottawa Scale (NOS) for non-randomised studies in meta-analyses. We performed a meta-regression using a random-effect model and summarised the results using relative risk (RR) and 95% confidence intervals (CIs). The threshold dose for an unintentional overdose is 20 morphine milligram equivalents (MME)/day. There were higher risks with larger doses: (1) ≤ 20 versus ≥ 21 MME/day: RR 2.81, 95% CI 1.09-7.22, p < 0.001; (2) ≤ 50 versus > 50 MME/day: RR 3.87, 95% CI 2.36-6.33, p < 0.001; (3) ≤ 100 versus > 100 MME/day: RR 4.28, 95% CI 2.61-7.1, p < 0.001; and (4) ≤ 50 versus > 50-100 MME/day: RR 3.09, 95% CI 1.84-5.18, p < 0.001). Heterogeneity was explained by the type of overdose event, inpatient or outpatient status, and length of observation. Type of pain (cancer or non-cancer pain) had no impact on heterogeneity. LIMITATIONS: The definition of exposure in studies included in the meta-analysis was heterogeneous. Some studies defined exposure as the filling of a prescription while others defined exposure as the prescription of an opioid to the patient, and all studies assumed that patients took the prescribed opioid. Medications that may contribute to overdose, such as benzodiazepines and other drugs, were not considered.
CONCLUSIONS: A significantly increased risk of inadvertent prescription opioid overdose was found with 20-50 MME/day, with fatality more likely with opioid doses above 50 MME/day, although extensive heterogeneity was found with the dose comparisons. Clinicians should inform patients of this risk and monitor them closely. PROTOCOL REGISTRATION: This protocol was registered with PROSPERO 2017: CRD42017058426.