Buprenorphine or its metabolite appeared in 55 of 2,369 selected unintentional opioid-death records in a Marion County study. That 2.3% figure described detection among people who had died. It was not the probability that a person prescribed buprenorphine would die, nor proof that buprenorphine caused each of the 55 deaths.
How the study population was assembled
Del Pozo and colleagues retrospectively analyzed coroner, death-certificate and toxicology records from Marion County, Indiana, covering January 2015 through December 2021. Testing for buprenorphine and its metabolite began in the contracted laboratories in 2015, so earlier records were excluded.
Among 2,951 unintentional overdose-death records, 2,475 involved detection of an opioid. The authors excluded 101 records when testing was not comprehensive or the testing laboratory was unknown, leaving uncertainty about buprenorphine coverage. Five children younger than 14 were also excluded. The final analytical sample was 2,369 deaths.
These exclusions were designed to make testing coverage more dependable. They also define the population to which the percentage applies. It cannot be relabeled as a percentage of all deaths, all overdoses, all opioid users or all buprenorphine-treated patients.
What counted as a buprenorphine finding?
The 55 cases involved detection of buprenorphine or norbuprenorphine. Describing every case as a measured parent-drug concentration would erase that distinction. The study used blood and urine findings and laboratory reporting thresholds; it did not present a uniform specimen collection site, death-to-collection interval, storage history or analyte-specific validation dataset.
No common quantitative concentration distribution was provided for these 55 cases. The analysis therefore cannot supply a buprenorphine lethal cutoff, compare paired blood sites, measure redistribution, or reconstruct the drug concentration before death. Laboratory detection and the study’s use of the word “involvement” should remain separate from causal attribution.
Most detections occurred with other substances
Of the 55 cases, 51—92.7%—included other substances the authors considered potent and commonly associated with overdose. Fentanyl and related findings were present in 28 cases (50.9%), benzodiazepines in 24 (43.6%), amphetamine in 23 (41.8%) and methamphetamine in 21 (38.2%). These groups overlapped; adding their percentages would not count distinct people.
Co-detection complicates attributing a mixed-drug death to one compound. It does not automatically establish that every detected drug contributed, that one drug was harmless, or that relative pharmacological potency alone identifies the decisive mechanism in an individual case. The authors themselves state that their study could show substances present after death, rather than establish the causes of the overdoses.
Four cases lacked the other potent substances emphasized by the investigators. One involved a reported encephalopathy attribution, another had diabetic ketoacidosis as a contributing cause, one included hydroxyzine and one antidepressants. These were not four experimentally demonstrated deaths caused solely by buprenorphine.
Why a low detection proportion is not a treated-person risk
The study did not count prescriptions, treated patients, dose, adherence or time receiving treatment. Without those exposure denominators, it cannot calculate the risk of death per treated person or compare that risk directly with another treatment. Low prescribing volume could partly explain a low proportion of detections, a possibility the authors acknowledged.
Nor can the data determine whether a particular detected exposure came from prescribed or non-prescribed medication, why it was taken, or whether treatment was protective for that individual. The paper’s broader treatment and policy discussion draws on additional literature; this record analysis alone does not answer every such question.
Scene descriptions were contextual evidence
Case narratives became available from 2016 onward. Mentions of needles, paraphernalia or suspected heroin provided context, but did not chemically confirm the substance used during the fatal incident. The paper’s narrative counts and table entries are not fully consistent, so they should not be promoted into a precise independently reconciled exposure total.
The study also classified some morphine-and-codeine findings without 6-monoacetylmorphine as heroin under its stated rule. That was an operational classification for analysis, not proof that every such result uniquely established heroin as its source.
What the findings support in death review
Buprenorphine-related findings were uncommon in this selected seven-year regional series and usually occurred with other drugs. The useful forensic lesson is to distinguish what was tested, what was detected, which cases entered the denominator and what the complete investigation attributed as contributory.
Detection alone establishes neither impairment nor an isolated cause of death. A related distinction arises when a study applies a “main intoxicant” classification rule, or when death certificates and integrated forensic review produce different poisoning counts.
Primary source
Del Pozo B, Atkins D, Andraka-Christou B, Wightman R, Clark MH, Huynh P, Ray B. Buprenorphine involvement in opioid overdose deaths: A retrospective analysis of postmortem toxicology in Marion County, Indiana, 2015–2021. Drug and Alcohol Dependence Reports. 2023;6:100131. doi:10.1016/j.dadr.2023.100131.
Leave a Reply