State Xylazine Scheduling and Changes in Xylazine and Medetomidine Reports in the U.S. Illicit Drug Supply: A Quasi-Experimental Study
Zhu, D. T.; Oh, S.
Show abstract
Background: Xylazine and medetomidine are veterinary sedatives increasingly detected as adulterants in the U.S. illicit drug supply. In response, several states have scheduled xylazine. Whether these policies are associated with subsequent changes in xylazine and medetomidine detections remains unclear. Methods: We conducted a state-level, semiannual, serial cross-sectional study using National Forensic Laboratory Information System (NFLIS) data from 1999 to 2025. The primary outcomes were xylazine and medetomidine reports per 100,000 NFLIS drug reports. We used staggered difference-in-differences event-study models to estimate changes in report rates after xylazine scheduling. Sensitivity analyses excluded Florida and expanded the treatment definition to include states that criminalized xylazine without formal scheduling. Falsification analyses examined steroid and antidepressant reports as negative-control outcomes. Results: NFLIS recorded 101,987 xylazine reports and 12,085 medetomidine reports. Xylazine scheduling was not associated with a significant change in xylazine report rates (ATT, 2,872.29 per 100,000; 95% CI, -2,024.63 to 7,769.21; p=.250). In contrast, xylazine scheduling was associated with a significant increase in medetomidine report rates (ATT, 1,536.51 per 100,000; 95% CI, 211.14 to 2,861.88; p=.023). Sensitivity analyses produced similar findings. Negative-control outcomes showed no significant changes. Conclusions: State xylazine scheduling was associated with increases in medetomidine reports but no significant change in xylazine reports. These findings suggest that scheduling may be followed by changes in adulterant composition rather than reductions in overall 2-adrenergic agonist involvement. Our study highlights the importance of monitoring the unintended effects of xylazine scheduling and supporting continued investment in public health surveillance, drug checking, and harm reduction services.
Matching journals
The top 6 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- A Mixed-Methods Comparison of Gender Differences in Alcohol Consumption and Drinking Characteristics among Patients in Moshi, Tanzania 88%
- Alcohol Use among Emergency Medicine Department Patients in Tanzania: A Comparative Analysis of Injury Versus Non-Injury Patients 88%
- Adverse Drug Reactions in Tuberculosis Treatment: Incidence, Reporting and Outcomes; Insights from a Mixed-Methods study across eight Indian cities 88%
Similar papers in this journal
- Estimating The Uncertain Effect of the COVID Pandemic on Drug Overdoses 94%
- Commercial Cannabis Product Testing: Fidelity to Labels and Regulations 92%
- Opioid Overdose and Naloxone Administration Knowledge and Perceived Competency in A Probability Sample of Indiana Urban Communities with Large Black Populations 91%
Similar papers in this journal
- Characterizing Declines in US Overdose Deaths Compared to Exponential Predictions 94%
- Using synthetic controls to estimate the population-level effects of Ontario's recently implemented overdose prevention sites and consumption and treatment services 94%
- Changes in Characteristics of Opioid Overdose Death Trends during the COVID-19 Pandemic 93%
Similar papers in this journal
- Changes in injecting versus smoking heroin, fentanyl, and methamphetamine among people who inject drugs in San Diego, California, 2020 to 2023 94%
- Fentanyl, Heroin, and Methamphetamine-Based Counterfeit Pills Sold at Tourist-Oriented Pharmacies in Mexico: An Ethnographic and Drug Checking Study 93%
- Drug Overdose Mortality Rates by Educational Attainment and Sex for Adults Aged 25 to 64 in the United States Before and During the COVID-19 Pandemic, 2015 to 2021 93%
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.