Every Two Years, the Fake Xanax Gets a New Name. The Real Trend Is the Naming Pattern Itself.

In March, the DEA put a chemical called bromazolam into Schedule I under emergency powers, the same shelf as heroin, calling it an imminent hazard to public safety. If you didn’t follow that story, here’s the part that matters more than the scheduling itself: bromazolam is not a new drug, and it is not even the interesting part of the trend. It’s the fourth name.

The DEA keeps winning the fight against the newest fake Xanax. It keeps losing the fight against fake Xanax as a category — because scheduling a molecule doesn’t touch the demand for a brand.

The brand survives every compound it’s built from

Start with what “Xanax,” sold on the street or through an unverified online contact, actually is in 2026: usually not alprazolam. A study in the journal Addiction examined 623 police seizures of pills sold under the Xanax name in Victoria, Australia, and found genuine alprazolam in only 30.3% of them. The rest — 60.2% — contained a different, usually unscheduled, designer benzodiazepine sold under the same trusted name and often pressed into tablets carrying real Xanax markings like “B 707” or “GG 249.” Among people who showed up in emergency departments after using something sold to them as Xanax, alprazolam was confirmed in just 19 of 125 cases; the other 78.4% tested positive for a novel psychoactive substance instead.

The specific compound behind the brand has changed on a predictable clock: etizolam displaced older designer benzos around 2020, clonazolam followed, then clobromazolam, and now bromazolam — linked to 201 overdose deaths in DEA’s own count between April 2021 and this February, the large majority alongside fentanyl rather than alone. DEA’s Cheri Oz, announcing the emergency order, framed it as staying ahead of the threat: We will not wait for more lives to be put at risk. Filter reported a more uncomfortable detail buried in DEA’s own numbers: only 4 of those 201 deaths involved bromazolam by itself. The scheduling targets the molecule. The molecule is, by DEA’s own data, rarely the thing that kills someone on its own — fentanyl usually is.

This is the whack-a-mole pattern in its clearest form: etizolam got scheduled, clonazolam filled the gap, clobromazolam followed that, and bromazolam is simply this cycle’s current occupant of a slot that has never gone empty since 2020. Nothing about emergency-scheduling bromazolam changes the economics that keep a new designer benzo arriving roughly every eighteen to twenty-four months to fill the space the last one left. If you take one thing away from a scheduling headline like this, take that: the DEA is correctly identifying and removing each threat, and a new one keeps showing up wearing the old one’s name, because scheduling regulates molecules and the market sells a brand.

Congress has watched this exact race before, and tried to end it with a shortcut forty years ago. Clandestine chemists in the 1980s were producing “designer drugs” — slight molecular tweaks to banned compounds that were, technically, not yet illegal — faster than the DEA could schedule them one at a time. In 1986, Congress passed the Federal Analogue Act, which tried to solve the whole category at once: anything “substantially similar” to a Schedule I or II drug, intended for human use, would be treated as though it were already scheduled. It didn’t end the race. It changed the terrain — pushing the fight into court, where “substantially similar” turned out to be exactly as vague and exploitable as it sounds, and clandestine chemistry kept finding the next unlisted variant. The designer-benzodiazepine rotation now cycling through etizolam, clonazolam, clobromazolam, and bromazolam is the same forty-year-old dynamic, just running on a faster clock and a smaller molecule.

The trend nobody’s tracking is happening in legal prescriptions

While designer benzos rotate through Schedule I, a different and almost entirely separate trend is unfolding inside licensed prescribing — and it’s moving in two directions at once depending on who’s getting the prescription. Overall U.S. benzodiazepine use fell from 4.7% of adults in 2018 to 3.4% in 2022, according to a national study using Medical Expenditure Panel Survey data, driven almost entirely by adults over 56, whose use dropped from 7.2% to 4.7% over the same years. Lead author Mark Olfson’s team found something that should worry prescribers more than the topline decline reassures them: 41.6% of everyone still taking a benzodiazepine was also prescribed a separate central-nervous-system depressant in the same year — rising to 62.9% among people with serious psychological distress and 72.0% among people in fair or poor health. Although benzodiazepine use is declining in U.S. outpatient care, the high rate of co-prescribing with other CNS depressants underscores the need for careful monitoring, Olfson said.

outpatient care, the high rate of co-prescribing with other CNS depressants underscores the need for careful monitoring](https://www.publichealth.columbia.edu/news/benzodiazepine-use-declines-across-u-s-led-reductions-older-adults), Olfson said.

Meanwhile, in Medicare specifically, the line is going the other way. A Frontiers in Medicine analysis of Part D claims found benzodiazepine prescriptions to Medicare beneficiaries rose 80% between 2017 and 2023 — from 1.7 million to 3.1 million — even as national benzo prescribing fell roughly 25% over the same stretch. Average therapy length in that Medicare population ran 108 days per beneficiary per year, well past the 30-day maximum most clinical guidelines recommend. Psychiatry prescribing to this population rose 250%; family practice and internal medicine roughly doubled. Geography matters here too: prescribing in the southeastern U.S. ran far above the northwest, and beneficiaries in New Jersey had two and a half times the odds of an extended-duration prescription compared to those in California.

Put the two studies together and a genuinely underreported story appears: America’s benzodiazepine problem isn’t one trend, it’s two moving in opposite directions, split roughly along a line of age and insurance status, and the “declining use” headline is only true for one of them.

GHB is a smaller, faster-moving line on the same chart

A shorter-duration depressant is following its own accelerating curve. In Australia, GHB-involved deaths rose roughly tenfold between 2012–13 and 2021–22, from fewer than six a year to 52, according to research from UNSW’s National Drug and Alcohol Research Centre; hospitalizations more than tripled over the 2013–2023 decade. Deputy Director Amy Peacock named the specific pharmacology that makes GHB unusually dangerous compared to other depressants: Repeated dosing can quickly lead to overdose because GHB builds up in the body faster than people realise — a narrow window between an effective dose and a fatal one, and a habit of redosing that turns a survivable mistake into an unsurvivable one within the same evening.

The clinical debate underneath the trend lines

None of this trend data is settling the argument happening inside clinical medicine right now about what benzodiazepines even are. Psychiatric Times captured the split this week: some clinicians describe an “iatrogenic epidemic” and call benzos “the new opioids”; others push back that they remain, in the right hands and the right dose, genuinely useful medicines being demonized because extreme framing draws more attention than careful prescribing does. Both things can be true in different rooms. A drug that’s dangerous when it’s counterfeit, dangerous when it’s over-prescribed to someone already on a second depressant, and genuinely helpful when it’s dosed correctly and monitored doesn’t fit cleanly on either side of that fight — which is exactly why the trend data matters more than the slogans. The counterfeit-Xanax pipeline, the Medicare over-prescribing pattern, and appropriate clinical use of the same drug class are three different problems that get flattened into one scary word. Only one of them gets fixed by scheduling a molecule.