Refining cause of death attribution among opioid, opioid‐stimulant and stimulant acute toxicity deaths

Abstract Background and aims Deaths attributed to a combination of opioids and stimulants have risen dramatically in recent years, largely attributed to fentanyl, yet little is understood about which drug class is primarily responsible. Attributing death to acute substance toxicity is complex and lacks clear standards. We aimed to determine whether additional causes of death and other significant conditions among deaths attributed to fentanyl were similar regardless of stimulant involvement, and distinct from deaths involving stimulants without opioids. Design Cross‐sectional analysis using records from the California Electronic Death Registration System. Setting and cases Unintentional acute toxicity deaths involving fentanyl or stimulants (methamphetamine or cocaine) occurring in San Francisco, USA, during 2013–2023. Measurements We compared demographic characteristics and causes of death or other significant conditions (cardiovascular, cerebrovascular, other medical cause, or no other additional cause) among five mutually exclusive groups of deaths: stimulants without opioids (stimulant‐only), fentanyl with stimulants (fentanyl‐stimulant), fentanyl without stimulants (fentanyl‐only), other opioids (e.g., heroin, oxycodone) with stimulants (“other opioid‐stimulant”), and other opioids without stimulants (“other opioids‐only”). We conducted separate unadjusted and adjusted multivariable logistic regression models for each outcome (cardiovascular, cerebrovascular, other medical, or no additional cause). The primary analysis included results for the fentanyl groups. Findings Of 4475 deaths attributed to acute opioid and/or stimulant toxicity, 24% involved stimulants‐only, 45% fentanyl‐stimulants, and 12% fentanyl‐only; the remaining 20% involved other opioids. Stimulant‐only decedents were the oldest (mean age 54 years), followed by fentanyl‐stimulant (47 years) and fentanyl‐only (44 years; p < 0.001 for all). The adjusted odds of having cardiovascular, cerebrovascular, or other medical causes of death (adjusted odds ratios [aORs] from 0.03 to 0.52, with 95% confidence intervals [CIs] from 0.01 to 0.68) were lower and the odds of no additional cause of death (aORs from 2.53 to 3.31, with 95% CIs from 2.00 to 3.40) were higher for both groups of deaths involving fentanyl compared with deaths attributed to stimulants‐only. There were no statistically significant differences in causes of death when comparing fentanyl‐only with fentanyl‐stimulant deaths. Findings were similar for other opioid deaths. Conclusion In San Francisco, USA, causes of death and other significant condition characteristics among deaths attributed to fentanyl appear to be similar regardless of the involvement of stimulants, but are markedly different from deaths involving stimulants without opioids. When reporting on drug‐related mortality and developing interventions, deaths attributed to a combination of fentanyl and stimulants may be appropriately considered in the context of opioid overdose prevention, while deaths attributed to stimulants without opioids may require a response focused on preventing and treating underlying chronic medical conditions.