Reporting injection of both heroin and methamphetamine during a year does not mean that the drugs were always mixed in one injection. Al-Tayyib and colleagues’ 2017 paper distinguished several patterns among Denver respondents. Preserving those categories changes how its overdose association should be understood.
Who contributed to the analysis?
The study analyzed 592 respondents with complete local questionnaire data from Denver’s 2015 National HIV Behavioral Surveillance cycle. Recruitment used respondent-driven sampling: initial participants referred others from their networks. Participants were adults who reported injection during the preceding twelve months.
The published analysis did not use respondent-driven-sampling weights. Its percentages describe the analyzed sample and should not be presented as representative prevalence estimates for all Denver residents, all U.S. adults or every person who injects drugs.
What did the broad categories mean?
Of the 592 respondents, 173 reported heroin injection only, 123 methamphetamine injection only and 296 injection of both during the previous year. The “both” category therefore concerned an annual history, not chemical testing of one mixture or proof of simultaneous exposure.
Among those reporting both, 130 described usually injecting the drugs separately at different times, 83 described mixing them together and 71 described both approaches. These three reported counts sum to 284, rather than all 296 in the broader group. The remaining twelve should not be assigned to a particular method without supporting information.
Why does a usual pattern not reconstruct an overdose?
A question about usual injection behavior does not identify the exact substances, doses or timing of a particular episode. Someone who usually separates the drugs may also have a different exposure on another occasion.
The study asked respondents about overdose during the preceding twelve months, without explicitly defining overdose for them. These were reported experiences, not independently adjudicated toxicology-confirmed events. No retained specimen established which drug combination was present in each reported episode.
What did the adjusted association establish?
The modeled association between annual use of both drugs and reported overdose, compared with heroin-only injection, was an adjusted prevalence ratio of 2.8 with a 95% confidence interval of 1.7–4.5. Adjustment included demographic and behavioral variables such as injection frequency and homelessness.
A prevalence ratio from this cross-sectional survey is not a measured causal increase in risk from mixing one specified dose. The study could not establish temporal order for every relevant factor or eliminate unmeasured differences between groups.
The authors also reported different overdose frequencies across usual injection patterns. Those observational subgroup comparisons do not establish that one pattern was safe or pharmacologically protective. They may reflect different exposure intensity, histories or other circumstances.
What else limits interpretation?
Interviewer-administered self-report can introduce recall and disclosure error. Recruitment through social networks affects who is reached, and the evidence came from one city in 2015. The study did not characterize present-day drug composition or establish fentanyl exposure in this historical sample.
Its strongest contribution is showing that a broad co-use category contains multiple behaviors. A precise account distinguishes the annual history, usual method and reported outcome instead of turning all three into proof of simultaneous injection at an individual event.
Source
Al-Tayyib A, Koester S, Langegger S, Raville L. Heroin and Methamphetamine Injection: An Emerging Drug Use Pattern. Substance Use & Misuse. 2017;52(8):1051–1058. doi:10.1080/10826084.2016.1271432.
Leave a Reply