A declining percentage can coexist with a growing death toll. In a historical Los Angeles County study, cardiovascular involvement accounted for a smaller share of methamphetamine-related deaths as opioid co-involvement increased. Reading that pattern as evidence that methamphetamine had become less harmful to the heart would go beyond what the records establish.
Which deaths entered the study?
Shover and colleagues analyzed 6,125 medical-examiner cases from January 1, 2012, through June 30, 2021. Methamphetamine had to be listed in a cause-of-death, contributing-factor or injury-description field. A positive toxicology finding alone, without a listing in those fields, did not qualify a case for inclusion.
This was a retrospective observational analysis of recorded death determinations. It was not an experiment exposing people to drugs, and it was not a study of every person with detectable methamphetamine. The case definition selected deaths in which investigators had already attributed some relevance to methamphetamine.
How did the total and the composition change?
The annual number increased from 238 in 2012 to 1,385 in 2020, the last complete calendar year available. The 2021 data contained 671 cases through June 30. That half-year count should not be labeled a full-year total or mechanically compared with a complete year as though the observation periods were equal.
| Recorded feature | 2012: 238 deaths | January–June 2021: 671 deaths |
|---|---|---|
| Cardiovascular involvement | 113, approximately 47% | 172, approximately 26% |
| Opioid involvement | 37, approximately 16% | 363, approximately 54% |
| Fentanyl involvement | 0 | 311, approximately 46% |
These categories were not mutually exclusive. A death could involve cardiovascular disease, methamphetamine and an opioid. Adding their percentages would therefore misrepresent the classification. The full study period included 2,331 deaths with cardiovascular involvement, or approximately 38% of the 6,125 cases.
Why does the denominator change the interpretation?
The cardiovascular percentage used methamphetamine-related deaths in that period as its denominator. It did not use all people taking methamphetamine, the number of exposure episodes, or person-years of follow-up. If deaths involving another mechanism expand rapidly, the cardiovascular fraction can fall without demonstrating a lower cardiovascular risk among exposed people.
The count comparison itself also resists a reassuring interpretation: 172 cardiovascular-involved deaths appeared in the partial final year, versus 113 in the complete first year. This observation is not an annualized rate calculation. It simply illustrates why a smaller fraction cannot be substituted for a smaller number or a lower exposure-specific risk.
Does opioid co-involvement prove that methamphetamine was contaminated?
No. The records showed an increasing contribution of opioids, particularly fentanyl, to these deaths. They did not establish whether each person knowingly used separate drugs, knowingly used a combined product, or unknowingly encountered an opioid. The authors explicitly acknowledged that co-occurrence could not distinguish intentional from unintentional co-use.
A death record is not a chemical analysis of the product consumed, a reconstruction of ingestion order, or a direct measurement of intent. The population pattern can motivate further investigation of changing supply and use practices. It cannot determine an individual product’s composition or a decedent’s knowledge.
Could examination practices affect the organ-system trend?
Yes. The authors noted that full autopsies had generally been performed in suspected overdoses before 2020, with rare exceptions. During the increased pandemic caseload, detailed postmortem examinations may have been less frequent in some cases with other evidence pointing to a drug-related death.
Conditions requiring internal examination could consequently have been under-recorded. That is a measurement limitation, not proof that the entire cardiovascular trend was an artifact. It nevertheless prevents treating changes in the recorded organ-system distribution as a direct readout of changing biological toxicity.
What other uncertainty remains?
The classification depended on death-certificate descriptions and the medical examiner’s determinations. Manual coding had high agreement in a sampled double-coding exercise, but agreement between coders does not independently validate every underlying cause-of-death attribution. Recorded housing status was also partly subjective, and its misclassification could operate in either direction.
The study described one jurisdiction and a defined historical period. Its proportions should not be presented as current national figures. It also did not estimate survival probabilities or individual drug-concentration thresholds.
What is the useful conclusion for a case review?
Request the actual case definition, calendar window and denominator behind a reported trend. Separate methamphetamine detection from attributed involvement, distinguish overlapping causes, and account for changes in examination practices. This study establishes a major shift in the composition of recorded Los Angeles deaths. It does not establish a protective change in cardiovascular effects or explain the mechanism and intent in every individual death.
Scientific source
Shover CL, Friedman JR, Romero R, et al. Longitudinal changes in co-involved drugs, comorbidities, and demographics of methamphetamine-related deaths in Los Angeles County. Journal of Substance Use and Addiction Treatment. 2023;151:209101. doi:10.1016/j.josat.2023.209101. Historical observational medical-examiner record study.
Leave a Reply