Improving risk factor identification for opioid overdose deaths in Tennessee

Abstract

ObjectiveTo examine specific drugs present based on postmortem toxicology for prescription opioid, heroin, and fentanyl overdoses classified based on ICD-10 coding. To compare drugs identified from postmortem toxicology with those listed on the death certificate for opioid overdoses.IntroductionUsing death certificates alone to identify contributing substances in drug overdose deaths may result in misclassification and underestimation of the burden of illicit and prescription opioids and other drugs in drug-related deaths. To enable timely and targeted prevention in Tennessee (TN), the identification and monitoring of new drugs and trends in use should utilize toxicology and medicolegal death investigation data directly, as recommended by others 1-3. These data can inform mortality outcome definitions for improved surveillance and risk factor identification 4-7. To our knowledge, this is the first analysis to use statewide linked toxicology and death certificate data in TN.MethodsWe identified 615 opioid involved overdose deaths in TN of unintentional (underlying ICD-10 codes: X40-X44) or undetermined (underlying ICD-10 codes: Y10-Y14) intent during June 1st to December 31st 2017. Utilizing the Interim Medical Examiner Database (I-MED), we identified postmortem toxicology reports for 454 cases, which were from one of three national laboratories used by a state Regional Forensic Center. Toxicology data were abstracted and independently verified by two co-authors and linked to the TN death statistical file that included cause of death information (literal text and ICD-10 codes) and demographics. The analysis focuses on cases with an available toxicology report.ResultsWe identified 171 prescription opioid overdoses, 221 fentanyl overdoses, and 113 heroin overdoses. Table 1 displays postmortem toxicology profiles for major drugs/classes. For prescription opioid deaths (excluding fentanyl and heroin), positive toxicology results for prescription opioids were as follows: methadone (11%), buprenorphine (14%), hydrocodone (14%), oxycodone (36%) and oxymorphone (also a metabolite, 47%). Benzodiazepines were present in close to 58% of prescription opioid overdoses; stimulants (cocaine, amphetamines, methamphetamines) in about 25%. For fentanyl and heroin deaths, prescription opioids were detected in about 26% and 34%, respectively; stimulants in about 57.9% and 52.2%, respectively, and benzodiazepines 36-37%. Fentanyl was present on toxicology in about half of heroin overdoses, and 6–monoacetylmorphine in 72.6%.ConclusionsUsing medical examiners’ data, including toxicology data, improves estimation of contributing drugs involved in opioid deaths. This analysis provides jurisdiction-specific data on drugs that can help with monitoring trends and informs risk factor identification. Future work includes adding information on prescribed opioid and benzodiazepines using TN’s Prescription Drug Monitoring Database and evaluating demographic variation in contributing drugs between toxicology and DC data to identify susceptible populations.