Xylazine Went National While the Dose Went Down. Don’t Call That Good News.

You give someone naloxone. Their breathing comes back. They still don’t wake up.

If you’ve been in a room where that happened, you may have spent the next ten minutes convinced you did it wrong — wrong dose, too late, bad batch. You didn’t do it wrong. Naloxone reverses opioids. Xylazine, a veterinary sedative that has been riding along in the illicit fentanyl supply since roughly 2019, is not an opioid, and naloxone does not touch its sedation. For years that scene belonged to Philadelphia, then to the Northeast. A study published July 9 in JAMA Network Open says it now belongs to everyone.

Xylazine has finished going national, and the falling dose is not a reprieve — it’s the supply making room for something worse.

What they actually looked at

Olivia Sugarman and colleagues at Johns Hopkins pulled 42,307 urine specimens that tested positive for fentanyl, collected between March 2023 and September 2025 through Millennium Health, a national clinical lab. Then they asked two questions most surveillance conflates: how often xylazine showed up, and how much was there.

Xylazine turned up in 16.91% of fentanyl-positive samples overall — roughly one in six. The geography is the story. In the East, positivity climbed from 11.11% to 19.63% over the study window. In the West, it went from 0.0% to 13.46% — from a substance that was effectively absent to one present in more than one in eight fentanyl-positive tests, in under three years. The West still had 79% lower odds of xylazine detection than the East after adjustment (adjusted odds ratio 0.21) — “adjusted” here just means they accounted for age, sex, testing setting and other differences between the samples so the regional gap isn’t an artifact of who got tested. But a 79% gap that used to be a 100% gap is a gap closing fast.

Meanwhile the concentration fell everywhere. Normalized to creatinine — a standard way of correcting for how dilute someone’s urine is, so a hydrated person doesn’t look like a low-dose person — Eastern samples dropped from 61.77 to 24.00 ng/mg. Western samples fell from 28.62 to 20.64 ng/mg. Peak concentrations were back in early 2023, and it has been a downhill slope since.

Two other findings deserve a line. Samples also positive for heroin had 2.2 times the odds of containing xylazine, which fits the read that xylazine tracks the older injection-drug supply. And each tenfold increase in fentanyl concentration came with 55% higher odds of xylazine — the two are being cut together, not independently.

The limits, said plainly

These are people who were already in care. Nearly 89% of the xylazine-positive specimens came from substance use treatment settings — a picture of people connected to a clinic, not of people using in an encampment or a car. Concentration in urine also depends heavily on when someone last used before the cup, so a falling average could partly reflect testing timing rather than a changing supply. And the authors are explicit that they never tested for medetomidine, the sedative displacing xylazine in some cities. This is a big, national, carefully built dataset that cannot tell you what is in the drugs of the people furthest from care — exactly the population that needs the answer most.

Why a smaller dose isn’t a smaller problem

Here is where the “concentrations are down” headline falls apart.

First, xylazine does not clear the way animal data predicted. A June 2026 paper in Drug and Alcohol Dependence drew serial blood samples from 13 people who came to two emergency departments after non-fatal overdose and calculated a xylazine half-life of 345 ± 145 minutes — close to six hours, against the 22-to-50-minute range reported in sheep and horses. Small sample, wide variance, and it is the best human pharmacokinetic data anyone has. If you are the person on that stretcher, a lower dose of a drug that lingers for the better part of a day is not the same thing as a lower dose of something that washes out.

Small sample, wide variance, and it is the best human pharmacokinetic data anyone has.

Second, exposure is mostly not a choice. CDC’s MMWR analysis of nearly 44,000 adults entering treatment assessment found 12.7% of those whose primary substance was fentanyl or heroin reported ever using xylazine — and the authors note their instrument never asked whether people used it on purpose or simply got it. Most people don’t want it. They get it anyway.

Third, and most important: the invited commentary running alongside the study, by Janet Childerhose, names the replacement problem directly. Medetomidine is now showing up in the illicit supply with and without xylazine. And medetomidine is not a milder cousin. A Philadelphia clinical series published in 2025 detected it in 58 of 58 tested samples and described a withdrawal syndrome with median blood pressures of 190/110, two-thirds of patients requiring ICU admission, and a third developing encephalopathy or seizures — often not responding to the standard oral medications clinicians reach for first.

What to do with this

If you use, assume xylazine is possible wherever you are now, not just on the East Coast. Test strips are worth carrying. Still give naloxone — the opioid part is what stops breathing, and that is the part that kills. Just expect that someone may stay heavily sedated afterward and needs to be positioned on their side and watched, not left alone because “the Narcan worked.” Check skin regularly, and get wounds looked at early rather than when they’re deep.

If you’re a clinician, Childerhose’s recommendation is blunt and correct: test fentanyl-positive specimens for xylazine whether or not the patient reports it. People often genuinely don’t know.

The thing worth sitting with is what this study proves about method. Urine testing caught xylazine crossing the country while it was happening — years faster than death certificates would have. We built an entire national surveillance apparatus around counting bodies. This is a demonstration that we don’t have to wait that long, and it arrives at the exact moment the supply is handing us the next adulterant to find.